A primary electron beam facility at CERN -- eSPS Conceptual design report
M. Aicheler, T. Akesson, F. Antoniou, A. Arnalich, P. A. Arrutia Sota, P. Bettencourt Moniz Cabral, D. Bozzini, M. Brugger, O. Brunner, P. N. Burrows, R. Calaga, M. J. Capstick, R. Corsini, S. Doebert, L. A. Dougherty, Y. Dutheil, L. A. Dyks, O. Etisken, L. Evans, A. Farricker, R. Fernandez Ortega, M. A. Fraser, J. Gall, S. J. Gessner, B. Goddard, J-L. Grenard, A. Grudiev, E. Gschwendtner, J. Gulley, L. Jensen, R. Jones, M. Lamont, A. Latina, T. Lefevre, R. Lopes, H. Mainaud Durand, S. Marsh, G. Mcmonagle, E. Montesinos, R. Morton, P. Muggli, A. Navascues Cornago, M. Nonis, J. A. Osborne, Y. Papaphilippou, A. M. Rossi, C. Rossi, I. Ruehl, S. Schadegg, E. Shaposhnikova, D. Schulte, S. Stapnes, M. Widorski, O. E. Williams, W. Wuensch
118 September 2020
Conceptual Design Report - eSPS
M. Aicheler f , a , T. Akesson b , ∗ , F. Antoniou a , A. Arnalich a , P. A. Arrutia Sota a , c ,P. Bettencourt Moniz Cabral a , D. Bozzini a , M. Brugger a , O. Brunner a , P. N. Burrows d ,R. Calaga a , M. J. Capstick a , R. Corsini a , S. Doebert a , L. A. Dougherty a , Y. Dutheil a ,L. A. Dyks d , a , O. Etisken a , e , L. Evans a , A. Farricker a , R. Fernandez Ortega a , M. A. Fraser a ,J. Gall a , S. J. Gessner a , B. Goddard a , J-L. Grenard a , A. Grudiev a , E. Gschwendtner a ,J. Gulley a , L. Jensen a , R. Jones a , M. Lamont a , A. Latina a , T. Lefevre a , R. Lopes a ,H. Mainaud Durand a , S. Marsh a , G. Mcmonagle a , E. Montesinos a , R. Morton a , P. Muggli a ,A. Navascues Cornago a , M. Nonis a , J. A. Osborne a , Y. Papaphilippou a , A. M. Rossi a ,C. Rossi a , I. Ruehl a , S. Schadegg a , E. Shaposhnikova a , D. Schulte a , S. Stapnes a , ∗ ,M. Widorski a , O. E. Williams a , W. Wuensch a a CERN, Switzerland, b Lund University, Sweden, c John Adams Institute, Royal Holloway, University ofLondon, UK, d John Adams Institute, University of Oxford, UK, e Ankara University, Turkey, f University of Helsinki, Finland
Abstract
The design of a primary electron beam facility at CERN is described. It re-enables theSPS as an electron accelerator, and leverages the development invested in CLIC technologyfor its injector and as accelerator R&D infrastructure. The facility would be relevant forseveral of the key priorities in the 2020 update of the European Strategy for Particle Phys-ics, such as an electron-positron Higgs factory, accelerator R&D, dark sector physics, andneutrino physics. In addition, it could serve experiments in nuclear physics. The electronbeam delivered by this facility would provide access to light dark matter production sig-nificantly beyond the targets predicted by a thermal dark matter origin, and for natures ofdark matter particles that are not accessible by direct detection experiments. It would alsoenable electro-nuclear measurements crucial for precise modelling the energy dependenceof neutrino-nucleus interactions, which is needed to precisely measure neutrino oscillationsas a function of energy. The implementation of the facility is the natural next step in thedevelopment of X-band high-gradient acceleration technology, a key technology for com-pact and cost-effective electron/positron linacs. It would also become the only facility withmulti-GeV drive bunches and truly independent electron witness bunches for plasma wake-field acceleration. A second phase capable to deliver positron witness bunches would makeit a complete facility for plasma wakefield collider studies. The facility would be used forthe development and studies of a large number of components and phenomena for a futureelectron-positron Higgs and electroweak factory as the first stage of a next circular colliderat CERN, and its cavities in the SPS would be the same type as foreseen for such a futurecollider. The operation of the SPS with electrons would train a new generation of CERNstaff on circular electron accelerators. The facility could be made operational in about fiveyears, could start already in LS3, and would operate in parallel and without interference withRun 4 at the LHC. c (cid:13) ∗ [email protected], [email protected] a r X i v : . [ phy s i c s . acc - ph ] S e p ontents Contents ontents ontents Introduction
Figure 1.1: Schematic of the primary electron beam facility.Recent interest in light dark matter searches using GeV electrons has stimulated a new study of howsuch a beam could be provided at CERN [1]. The basic requirement is a very long spill of low intensityelectrons to a missing-energy/missing-momentum experiment [2, 3] and as described in Sec. 2.1.The present proposal as illustrated in Fig. 1.1 is to use the Super Proton Synchrotron (SPS) simultan-eously as an accelerator and as a very long pulse stretcher to provide such a beam. The SPS has, in thepast, accelerated electrons and positrons from 3.5 GeV to 22 GeV when it was used as the injector tothe Large Electron-Positron collider (LEP) [4], although most of the equipment required has now beendismantled. The electron injector would be replaced by a 3.5 GeV compact high-gradient linac basedon Compact LInear Collider (CLIC) [5] technology injecting pulses of 200 ns duration into the SPS,filling the ring at 100 Hz on a 700 msec duration plateau. The beam would then be accelerated to 16 GeV,using a 800 MHz SC RF system, similar to what is needed for FCC-ee. SPS-LSS6 is the preferred loc-ation for the RF system in order to exploit the existing infrastructure from the crab cavity installation.The electrons would then be extracted at 16 GeV using slow resonant extraction. The extracted beam istransported along an existing beamline to an experimental area on the Meyrin site.Use of the bypass for the SPS RF system requires a 10 min changeover period. If a more rapid change-over from protons to electrons is interesting or necessary, for example during a SPS supercycle, anextended pulsed bypass beamline can be considered in this area.The spacing of the SPS RF buckets is 1.25 ns given its 800 MHz SC RF system. Therefore, thefilling of the SPS has to be a multiple of this frequency. This multiple is determined by the matchingof the S-band injector linac to the SPS RF frequency, and is thus limited to a multiple of 5 ns betweenbunches. The default number of extracted electrons per filled bucket in the SPS is 1–10, depending onthe experimental capabilities. Around 2/3 of the ring is filled, and the filling and acceleration time isnegligible compared to the extraction time. Since the beam can be distributed over a relatively large areaup to 30 × and the SPS RF bucket is significantly wider than the timing resolution of the detectorthere is scope of being able to deal with several extracted electrons per filled bucket. The experiment canbe provided with 10 electrons in around 10 seconds of beamtime assuming an average of 6 electrons5 Introduction per delivered bunch spaced with 5 ns.Changes to the S-band linac would allow to fill all 1.25 ns spaced filled buckets in the SPS giving afactor 4 higher rate. The RF system and the beamline for delivery to the experiment is compatible witha beam up to 18 GeV.A fast extraction is also possible, when the whole beam is extracted from the machine in one revolution(23 µ s) to feed a possible beam dump type experiment. This could be repeated every 2 seconds if suchan operation was to be given priority.During the eSPS operation the beam from the linac is used for less than 5% of the time for injectioninto the SPS and is available for other uses the rest of the time. Two experimental areas will be availablefor a broad range of accelerator R&D using the injector and/or the full linac beam. The possibilitiesrange from studies very relevant for future Higgs-factories (linear or circular) to providing a uniqueplasma acceleration R&D facility. All these possibilities are described in Sec. 6. The full range of beamsavailable to experiments, from injector to extracted beam from the SPS, are summarised in Tab. 1.1.Most of the hardware, apart from the 3.5 GeV linac and the 800 MHz SC RF cryomodule in the SPS,already exist. The main components, shown in schematically in Fig. 1.1, are introduced below. A 3.5 GeV Compact Linac would be built using the technology developed for CLIC at 12 GHz, butwith klystrons as power source instead of the two-beam acceleration method proposed for CLIC. Anideal location would be in Transfer Tunnel 4 and 5 (TT4 and TT5) at the entrance to the West Hall(building 180), and connected to the SPS by the TT60 tunnel complex. A 0.2 GeV S-band photo-injectorwould provide a 200 ns pulse with a very flexible bunch structure inside the pulse. Both of these elementsare detailed in Sec. 3.
Transfer and Injection.
The beam from the linac is transferred via TT61, previously used to transportprotons to the west area, with the electrons being injected in the opposite direction to the protons. A new3.5 GeV kicker with a flat top of 200 ns and a 100 Hz repetition frequency would be installed at the SPSinjection point in Long Straight Section 6 (LSS6). The ring would be filled in about 700 milliseconds.Sec. 4.1 presents in detail the scheme.
Acceleration in the SPS.
A new vacuum sector in the SPS-LSS6 region was created to test superconduct-ing crab cavities with proton beams for HL-LHC. This sector comprises two Y-shaped vacuum chambersarticulated by mechanical bellows, the circulating proton beam line and the beam bypass consisting ofthe RF module. The mechanical bypass is equipped with a movable table to move the cryomodule inand out of the circulating beam. A dedicated RF system, cryogenic system and general infrastructurewere put in place on the surface and in the tunnel, and then successfully operated with beam during2018. The mechanically movable bypass allows for an installation of a RF module operating at 800 MHzfor a dedicated mode of operation with electrons. This configuration alleviates the strong constraints ofimpedance requirements for the high intensity proton beams and other modes of SPS operation. Thechangeover time to use the stage is around 10 minutes. Recent studies in the framework of the FCCstudy have led to design of several 800 MHz cavities ranging from single-cell for high beam currents tofive-cell structure for the high energy. Two five-cell cavities housed in a cryomodule will be suitable foracceleration of the electrons from 3.5 GeV to 18 GeV.
Resonant Extraction of the Beam from the SPS at 16 GeV would be done by exciting the third integerresonance using existing sextupoles (see Sec. 4.3.4). Quantum excitation due to synchrotron radiationwould provide a powerful method of pushing particles into resonances with no dynamic variation of thetune being required as is the case for the proton beam. The intensity and duration of the spill can becontrolled down to very low currents by adjusting the distance in tune of the beam from the resonance.A new extraction channel using similar hardware to proton extraction must be installed in LSS1. Fast6
Introduction extraction through the same channel could be performed using existing SPS kickers if needed for a beamdump experiment. Preliminary simulation results supporting this scheme are shown in Sec. 4.3.
Beam Transport to the Detector.
The beam would be extracted into TT10, the beamline which is alsoused to inject protons into the SPS. The TT10/TT2 switchyard magnets would not be powered duringelectron extraction so that the beam can be directed to a new experimental area. Quadrupoles wouldbe used to blow up the beam to match the detector requirements. This is presented in Sec. 4.3.4. Theonly civil construction required for the whole project would be a new short connection tunnel and anexperimental hall housing one or two experiments, located just outside TT2 and shown in Sec. 4.2.Table 1.1: Summary of the parameters of available experimental beams after each accelerating section,the S-band linac, the X-band linac and the SPS. Note that single quantities listed here corres-pond to the upper limit and can be reduced for operation.
Parameter Accelerating SectionS-band linac (Sec. 3.2)
X-band linac (Sec. 3.3)
SPS (Sec. 4.2)Energy [GeV] 0.05–0.25 3.5 3.5–18Electrons per bunch 10 – 10 – 10 Bunch length [ns] 10 − – 4 × − − – 2 . × − × Cycle length [s] 0.02 0.02 0.02The CERN Council adopted an update of the European Strategy for Particle Physics [6] in June 2020.The conceptional design report presented here, describes an infrastructure relevant for several of its keypriorities, like an electron-positron Higgs factory, accelerator R&D, dark sector physics, and neutrinophysics. • The implementation of eSPS would make excellent use of the investment made in the CLIC pro-gramme and is the natural next step in the development of X-band high-gradient accelerationtechnology , a key technology for compact and cost-effective electron/positron linacs. • The multi-GeV electron beam from the linac would drive wakefields in the non-linear regimeand would, with an independent electron witness bunch, demonstrate the applicability of plasmawakefields for high-gradient acceleration . The facility would be unique in its ability to studycollider related challenges, as the only facility with multi-GeV drive bunch and truly independentelectron witness bunch. Addition of a positron witness bunch would make it a complete facilityfor collider studies. • The 800 MHz super-conducting cavities for the eSPS would be the same type as foreseen for afuture electron-positron Higgs and electroweak factory as the first stage of a next circular colliderat CERN. The eSPS would be used for the development of and studies of a large number ofcomponents and phenomena for this circular collider. The operation of SPS with electrons wouldtrain a new generation of CERN staff on circular electron accelerators. • The electron beam would open a dark sector physics programme and in particular provide sensit-ivity to light dark matter production significantly beyond the targets predicted by a thermal darkmatter origin, and for the nature of dark matter particles that are not accessible by direct detectionexperiments. 7
Introduction • The future neutrino physics need to precisely measure neutrino oscillation probabilities as a func-tion of energy. This critically relies on the ability to model neutrino-nucleus interactions, andthis in turn requires input data on electro-nuclear reactions; the beam from this facility would beexcellent for such measurements.eSPS could be made operational in about five years, and serve the programmes above. It could startalready in LS3 and would operate in parallel and without interference with Run 4 at the LHC.8
Physics Potential and Requirements on the Electron Beam
The following section is a summary of the physics case presented in the Expression of Interest [3] sub-mitted to the SPSC in October 2018.
Only 15% of the observed matter is made of particles described by the Standard Model [7] (SM). Theevidence for dark matter (DM) does not give much direct guidance on the masses of its constituents,which could be anywhere from a tiny fraction of an eV up to many solar masses. More constraintscan be obtained by focusing on likely scenarios for how the primordial DM was created. The moststraightforward and simplest scenario is the thermal origin, in which DM arose as a thermal relic fromthe hot early Universe. This scenario only requires small non-gravitational interactions between darkand Standard Model matter, and is viable over the MeV to TeV mass range. The mass region ∼ MeV to ∼ GeV is largely unexplored. The fact that most stable forms of ordinary matter are found in this range,argues in favour of exploring this mass range.A thermal origin for DM requires an interaction between DM and familiar matter, and if there is aninteraction of light dark matter (LDM) with ordinary matter, (Fig. 2.1(a)) then there necessarily is aproduction mechanism in accelerator-based experiments (Fig. 2.1(b)), both in the minimal framework ofa four-particle contact interaction and in realistic ultraviolet completions of this scenario by the additionof a new force carrier. The most sensitive way to search for this reaction is to use an electron beamto produce DM in fixed-target collisions, making use of missing energy and/or momentum to identifythis process [8, 9]. Dedicated searches for these production reactions thereby provide sensitivity to DMcouplings to the Standard Model. e + e − χ ¯ χ ( a ) e − e − χ ¯ χZ ( b ) Figure 2.1:
Left:
Diagrammatic representation of a contact interaction between DM and electrons.
Right:
Production of DM in electron-nucleus fixed target interactions, guaranteed by theexistence of this contact interaction. As discussed below, models that resolve this contactinteraction through the introduction of a new mediator at experimentally accessible energyscales can also be detected through related production modes.The strength of this interaction determines when the DM froze out of equilibrium, therefore, the resid-ual DM abundance. This production mechanism, together with the observed DM density thus motivatesa precise interaction strength for any given DM mass.Relative to other experimental techniques used to search for DM, such as direct detection scatteringin underground experiments or indirect detection searches with satellite experiments, fixed-target ac-celerator experiments are the only technique available that can probe the DM interaction at momentum9
Physics Potential and Requirements on the Electron Beam scales comparable to those governing freeze-out in the early Universe. This technique is, therefore, nothindered by some of the common challenges faced by, for example, direct detection experiments, wheremass threshold or velocity suppression can severely inhibit signal rates in the non-relativistic limit as isshown in the left panel of Fig. 2.2. This makes fixed-target probes of DM both complementary to otherterrestrial techniques, and especially robust in exploring thermal freeze-out. Models of thermal DM inthe MeV – GeV mass range require that the interaction governing freeze-out have a cutoff scale belowthe weak scale. This is, essentially, a simple generalisation of the Lee-Weinberg bound [10, 11], withtwo important consequences: • Light Forces:
There must be new force carriers at the GeV-scale or below to mediate an efficientannihilation rate for thermal freeze-out. • Neutrality:
Both the DM and the mediator must be singlets under the full SM gauge group;otherwise they would have been produced and detected at LEP or at hadron colliders [12].These properties single out the hidden sector scenario highlighted in [13, 14], which is the focus ofconsiderable experimental activity.For the remainder of this chapter, we will use one of the simplest and most representative hidden sectormodels in the literature – a DM particle charged under a U ( ) gauge field (i.e. “dark QED”). We definethe LDM particle to be χ , the U ( ) gauge boson A (cid:48) (popularly called a “dark photon” mediator), and ε as the kinetic mixing parameter.This framework permits two qualitatively distinct annihilation scenarios in the early universe, depend-ing on the A (cid:48) and χ masses. • Direct Annihilation:
A mediator with m A (cid:48) > m χ generates the effective contact interaction fornon-relativistic DM particles. In the resolved theory, annihilation proceeds via χ χ → A (cid:48)∗ → f f toSM fermions f through a virtual mediator. This scenario is quite predictive, because the SM- A (cid:48) coupling ε must be large enough, and the A (cid:48) mass small enough, in order to achieve the thermalrelic cross-section. Depending on the mass of the mediator, on-shell mediator production withdecay to DM or production of DM through an off-shell mediator may be the dominant signal ina missing momentum experiment. In each case, the observed DM abundance implies a minimumDM production rate at accelerators. Constraints on this scenario can be extracted from CMB data,but are only relevant for some combinations of DM and mediator spin and couplings. This casewill be the focus of the remaining discussion. • Secluded Annihilation:
For m A (cid:48) < m χ , a new annihilation channel becomes kinematically al-lowed, and generically dominates. In this case, DM annihilates predominantly into A (cid:48) pairs [15].This annihilation rate is independent of the SM- A (cid:48) coupling ε . The simplest version of this scenariois robustly constrained by CMB data [16].Since the Feynman diagram that governs direct annihilation can be rotated to yield a scattering processof SM particles, the direct detection cross section is uniquely predicted by the annihilation rate in theearly universe for each choice of DM mass. Thus, direct annihilation models define thermal targets inthe σ e vs. m χ plane. The left panel of Fig. 2.2 shows how the non-relativistic direct detection crosssections can be loop or velocity suppressed in many models, and, therefore, these thermal targets varyby dozens of orders of magnitude in some cases. However, these vast differences in the direct detectionplane mask the underlying similarity of these models in relativistic contexts where both the scatteringand annihilation cross sections differ only by order-one amounts. To study all direct annihilation modelson an equal footing, we follow conventions in the literature (see [13]), and introduce the dimensionlessinteraction strength y as σ v ( χ χ → A (cid:48)∗ → f f ) ∝ ε α D m χ m A (cid:48) = ym χ , y ≡ ε α D (cid:18) m χ m A (cid:48) (cid:19) . (1)10 Physics Potential and Requirements on the Electron Beam A s y mm e t r i c F er m i o n E l a s t i c S c a l a r I n e l a s t i c S c a l a r (cid:72) s m a ll s p li tt i n g (cid:76) M a j o r a n a F er m i o n P s e ud o (cid:45) D i r a c F er m i o n (cid:72) s m a ll s p li tt i n g (cid:76) (cid:45) (cid:45) (cid:45) (cid:45) (cid:45) (cid:45) (cid:45) (cid:45) (cid:45) (cid:45) (cid:45) (cid:45) (cid:45) m DM Σ e Thermal and Asymmetric Targets for DM (cid:45) e Scattering
Figure 2.2:
Left:
Thermal targets for DM plotted in terms of the electron-recoil direct detection crosssection σ e vs. mass m DM . The appropriate thermal freeze-out curve for each scenario dif-fers by many orders of magnitude in the σ e plane due to velocity suppression factors, loop-level factors, or spin suppression, any of which are significant for non-relativistic scattering. Right:
By contrast, the dimensionless couplings (captured by y ) motivated by thermal freeze-out do not differ by more than a couple of orders of magnitude from one another, as shownin the y vs. m χ plane. Probing couplings at this magnitude is readily achievable using ac-celerator techniques, which involve DM production and/or detection, as well as mediatorproduction, all in a relativistic setting. Both plots above are taken from [14] and also showa target for asymmetric fermion DM, a commonly discussed variation on the thermal-originframework.This is a convenient variable for quantifying sensitivity because for each choice of m χ there is aunique value of y compatible with thermal freeze-out independently of the individual values of α D , ε and m χ / m A (cid:48) . The right panel of Fig. 2.2 shows the thermal targets in the y vs. m χ plane. A measured (orupper limit on the) production cross section ( σ ) for the process shown in Fig. 2.1(b), is translated to y as, y ∝ (cid:18) m χ m A (cid:48) (cid:19) α D m χ σ (2)From (2) it is seen that for a fixed values of α D and of the m A (cid:48) to m χ ratio, a measured (or limit on) σ would translate to a parabola in the y versus m χ plane in the right panel of Fig. 2.2. As also can be seenfrom (2), small m χ to m A (cid:48) ratios, and small α D values, would result in stronger experimental reaches. Weconservatively chose m A (cid:48) = × m χ , and α D = .
5, when estimating the experimental reach described inthis Conceptional Design Report, and in the Expression of Interest (EoI) [3] that was submitted to theSPSC in 2018.More discussions on reach for such parameter settings, including off-shell production of A (cid:48) and whenapproaching the resonance region of m A (cid:48) ≈ × m χ , are in [17].The thermal targets for various direct annihilation models shown on the right panel of Fig. 2.2 in the y vs m χ plane, are the same models as shown on the left panel, but the parameterisation in y and m χ re-11 Physics Potential and Requirements on the Electron Beam veals the underlying similarity of these targets and their relative proximity to existing accelerator bounds(shaded regions). Reaching experimental sensitivity to these benchmarks for masses between MeV andGeV would provide nearly decisive coverage of this class of models.Reaching the sensitivity to find events as shown in Fig. 2.1(b), or to establish the absence of such pro-duction, sets two contradictory requirements on the beam:1. A large number of electrons on target (EOT) since the interaction strength is weak. More quantit-atively, to reach the thermal targets for all possible natures of DM particles not excluded by CMBdata [16], would require from ∼ to ∼ EOT for Scalar and Pseudo-Dirac DM particlesrespectively [2].2. A low current to be able to measure individual electrons entering the target, and to match themwith potential signal electrons leaving the target. In addition, one has to be able to ensure theabsence of any bremsstrahlung photons, but without rejecting signal events due to such photonsproduced by unrelated beam electrons.
Only a primary electron beam can deliver ∼ to ∼ EOT with low current and high dutycycle at 5–20 GeV energy.
The higher end of this energy scale gives advantages on both signal production and background re-jection (where the latter in particular requires discovering photo-nuclear events from bremsstrahlungphotons). On the one hand, the signal cross sections increase with energy, improving the sensitivity, par-ticularly in the high mass region (several hundred MeV) as is shown in Fig. 2.3. On the other hand, therates of certain backgrounds decrease with higher energy, e.g. that of the exclusive 2-body photo-nuclearreactions scales as E − γ , and the products from these reactions carry more energy and are hence more vis-ible in a detector. Similarly, in-flight decays within a detector, e.g. of charged kaons from photo-nuclearreactions, have a lower rate and more detectable products at higher energy. (cid:72) MeV (cid:76) Y i e l d E nhan c e m en t (cid:72) vs . G e V on W t a r ge t (cid:76) Figure 2.3: Both beam energy and target material affect the dark photon production cross-section at thehigher end of the mass range of interest. This figure illustrates how increasing the beamenergy to 8 or 16 GeV, and/or switching from a Tungsten to an Aluminium target (fixed at0.1 X ), impacts the signal production cross-section for different dark photon masses. Weassume the kinematic selection E recoil < . E beam . The plot is taken from [3].As mentioned above, an Expression of Interest [3] was submitted to the CERN SPSC in 2018. Itdescribes the potential of a missing-momentum experiment like the Light Dark Matter eXperiment12 Physics Potential and Requirements on the Electron Beam P s e u d o (cid:45) D i r a c F e r m i o n M a j o r a n a F e r m i o n E l a s ti c & I n e l a s ti c S c a l a r P s e u d o (cid:45) D i r a c F e r m i o n M a j o r a n a F e r m i o n E l a s ti c & I n e l a s ti c S c a l a r LDMX: 4 10 EOT (cid:72) (cid:76) on 10 (cid:37) W (cid:72) Μ e (cid:61) (cid:76) Extended: 1.6 10 EOT (cid:72) (cid:76) on 40 (cid:37) Al (cid:72) Μ e (cid:61) (cid:76) Future: 1.6 10 EOT (cid:72)
16 GeV (cid:76) on 40 (cid:37) Al Α D (cid:61) A' (cid:61) Χ (cid:45) (cid:45) (cid:45) (cid:45) (cid:45) (cid:45) (cid:45) (cid:45) (cid:45) m Χ (cid:64) MeV (cid:68) y (cid:61) Ε Α D (cid:72) m Χ (cid:144) m A ' (cid:76) Extended LDMX Sensitivity
Figure 2.4: The experimental reach compared to the thermal targets for various direct annihilation mod-els shown on the right panel of Fig. 2.2. The blue line is the sensitivity from the refer-ence study discussed in [2], that conservatively assumes 0.5 background events for 4 × EOT at 4 GeV. A scaling estimate of the sensitivity of the configuration for the mass range150 ≤ M χ <
300 is illustrated by the solid red line. The dashed red line represents a similarestimate of the projected reach for µ e ∼
12 and roughly 3 years of running. For the latter twoexamples we have again assumed low background, consistent with reductions in yields of po-tential background sources, and better rejection, while increasing the effective luminositiesto 1 . × and 1 . × EOT, respectively. The plot is taken from [3].(LDMX), using the primary electron beam delivered by the accelerator complex described in this Con-ceptual Design Report. The red dashed line in Fig. 2.4 shows the expected performance compared withthe thermal targets. The potential to cover a large fraction of sub-GeV mass range for all natures of DMparticles, is clear.As was discussed above, each coloured line in Fig. 2.4, corresponds to one value of a measured (orthe limit on) σ . But, more information can be extracted if signal events are found. As demonstratedin [18], the background to the signal, can be fully rejected without making use of the p T (or deflectionangle) of the signal-electrons. The electron p T (or deflection angle) distribution can, therefore, be usedas an independent signal hypothesis test, and as a DM mass estimator. Figure 2.5 shows how much thesedistributions change for different DM masses.Although the emphasis in this chapter has been on various models of DM with direct annihilationthrough a dark photon mediator A (cid:48) , the missing momentum technique can probe multiple other mediatorscenarios with equally powerful sensitivity to the corresponding theoretical targets. For example, bothdark and visible matter could be directly charged under a new U ( ) group which gauges an anomaly-free combination of SM quantum numbers (e.g. baryon minus lepton number). Such new forces canalso mediate DM direct annihilation to SM particles with thermal targets analogous to those presentedin Fig. 2.2. Some of these are discussed in [3]. Furthermore, missing momentum techniques can alsoprobe strongly interacting dark sectors [19], millicharged particles, minimal dark photons, minimal U ( ) gauge bosons, axion like particles, and light new leptophilic scalars [17].13 Physics Potential and Requirements on the Electron Beam
10 MeV100 MeV200 MeV500 MeV 1000 MeV 1500 MeVInclusiveSingle e - Background ⟵ - - - [ deg. ] E v e n t F r ac ti on / e g . Electron Angular Distributions, 50 MeV < E e <
20 MeV300 MeV 2000 MeV - - - [ deg. ] E v e n t F r ac ti on / e g . Angular Distributions, E beam =
16 GeV, 50 MeV < E recoil < Figure 2.5: Electron deflection angle for DM pair radiation process, at various dark matter masses forelectron beam energy of 4 GeV (left) and 16 GeV (right). The distributions are for electronsrequiring 50
MeV < E < . E beam . In both panels, the numbers next to each curve indicate A (cid:48) mass.The plots are taken from [3]. The missing momentum search is sensitive to a range of other new-physics scenarios, potentially unre-lated to dark matter.
BaBarNA64 ( g - ) e ( g - ) μ + σ LDMX Extended LDMX -
10 10 - - - - - - - - - - m A ' [ MeV ] ϵ Invisibly Decaying Dark Photon ( Laboratory Bounds Only ) - - - - - - - - - m Z ¢ @ GeV D g B L Minimal B - L n exp.beam dumps BaBar L D M X E x t e nd e d L D M X Figure 2.6: Sensitivity to invisibly decaying dark photons (left) and B-L gauge bosons (right). The blueand the red lines correspond to the full blue and red lines in Fig. 2.4. The plots are takenfrom [3].Figure 2.6 illustrates the sensitivity to invisible dark photons and to minimal B-L Z (cid:48) gauge bosons, viatheir invisible decays to neutrino final states. Figure 2.7 illustrates the sensitivity of a missing momentumsearch to production of millicharged particles. Millicharge production occurs through off-shell photonexchange, and particles with sufficiently small millicharge Q χ / e have no additional interactions in thedetector.A primary electron beam facility described in this CDR, opens more possibilities than the missingmomentum searches described in Sec. 2.1 and above in this section. Section 5.5, therefore, outlines ascenario with two beamlines into the experimental hall, with space for two experiments.The majority of the electrons remain in the SPS after the extraction of the long low current electron14 Physics Potential and Requirements on the Electron Beam - - - - - - m c @ MeV D Q c ê e Millicharged Fermion E D G E S SN1987ASLAC MilliQ L D M X E x t e nd e d L D M X n e x p . colliders Figure 2.7: Sensitivity to millicharge fermion particles with charge Q χ / e vs mass. Production occursthrough an off-shell photon. Grey regions are existing constraints. The green shaded regionrepresents parameter space where a millicharged dark matter subcomponent can accommod-ate the 21 cm absorption anomaly reported by the EDGES collaboration [20–23]. The blueand the red lines correspond to the full blue and red lines in Fig. 2.4. The plot is takenfrom [3].spill. These 1 − × electrons could be dumped in the other beamline in a 23 µ s spill. This couldallow the accumulation of more than 10 electrons in a year. More than that in fact, since if priority wasgiven to such an operation, the cycle fill-accelerate-dump could be repeated every two seconds.As can be seen in [18], photo-nuclear and electro-nuclear reactions are major background sources fora dark matter missing momentum experiment. However, to measure such reactions is also important tounderstand neutrino-nuclear response. The future neutrino long baseline experiments need to preciselymeasure neutrino oscillation probabilities as a function of energy. This critically relies on the ability tomodel neutrino-nucleus interactions, and this in turn requires input data on electro-nuclear reactions; thebeam from this facility would be excellent for for this purpose. Such measurements could maybe bedone by a missing-momentum experiment, however, given the importance to understand such reactionsfor neutrino physics, there may be a case for a dedicated experiment. The physics of this is describedin [3].There is a broad usage of electron beams for the study of hadrons and their underlying structures, likethe momentum and spin distributions of sea quarks and gluons in the nucleons, the study of excitationspectra of nucleons and hyperons, and the prospect to produce mesons with exotic composition and/orexotic quantum numbers. The facility presented in this CDR would extend the energy range, but couldnot reach the beam intensity currently available, at Jefferson laboratory in the USA where the requestsfor beam go beyond what is available. 15 Linac
The electron linac produces the electron beam and accelerates it to an energy of to 3.5 GeV, the energyrequired for injection into the SPS. The linac consists of two parts; the injector, that produces the electronbunches with the required time structure, emittance and charge and brings them at an energy of about200 MeV, and the high-gradient X-band linac that further accelerates them to 3.5 GeV. Each beam pulseconsists of 40 bunches, separated by 5 ns. Such bunch spacing corresponds to every 4 th RF bucket in the800 MHz SPS RF system and the number of bunches in turn depends by the optimised pulse length inthe linac. The pulse structure and the 100 Hz linac repetition rate allow to fill the SPS ring with 3000bunches, the maximum given its diameter, within 1 second.After the injector, a low-energy beam line branches out from the main beam line, bending the beam by180 ◦ to an experimental area. This area can be used to perform independent experiments at a maximumbeam energy of 250 MeV, similar to what is presently ongoing in the CLEAR user facility [24]. At theend of the linac, another independent beam line can be used for experiments requiring a higher beamenergy. The beam dynamics in the eSPS linac are dominated by the impact of wakefields, due to the small irisaperture of the X-band accelerating structures. The linac optics are based on a FODO lattice that usesthe two quadrupoles of each pair of consecutive RF modules as focusing and defocusing magnets. Theaverage β function in the linac is less than 5 metres, and the phase advance per cell is 90 degrees. Thissetup is chosen to apply strong focusing in the transverse planes and to increase beam stability. At thestart of the linac the beam energy is about 200 MeV; the linac provides acceleration to the final energyof 3.5 GeV. The beamline consists of 24 RF units and has a length of approximately 68 m, with a fillingfactor of 86%.Start-to-end simulations of the linac have been performed in order to assess the beam quality perform-ance in the presence of various imperfections. The simulations were carried out using the code PLACET,a particle tracking code developed at CERN for linear collider studies, that can simulate the transport ofparticle beams through a linear accelerator under the effects of various imperfections [25]. PLACET en-ables the simulations of beam correction techniques, such as Beam-Based Alignment (BBA), to evaluatethe beam quality after correction, and implements multi-bunch effects. The wakefields generated in the X-band structures induce both emittance growth in the transverse planeand an increase in the correlated energy spread of the bunch. The effect on the energy spread can bepartly compensated by running off crest in the phase of the accelerating structures with the optimal RFphase depending on both the charge and the bunch length. The requirement on the energy spread of 0.1%at the end of the linac, for injection in the SPS, limits the maximum bunch length allowed for a givenspecific charge. Figure 3.1 shows the minimum energy spread achievable at the end of the linac as afunction of the RMS bunch length, for different bunch charges. A bunch with charge 50 pC and length150 µ m reaches its minimum energy spread when the RF structures operate nearly on crest. A bunchcharge of 1000 pC can reach at best 0.2% energy spread at a bunch length of approximately 200 µ m foran RF phase of -7 degrees.Transverse misalignment of the elements also induce emittance growth, through two mechanisms:spurious dispersion due to off-axis quadrupoles, and transverse wakefield kicks due to the beam travellingoff-axis through the accelerating structures. BNS damping can be considered to stabilise the beam,but it wasn’t deemed necessary given the relatively large emittance requirements and the overall goodrobustness of the linac to such effects. 16 Linac
Figure 3.1: Minimal energy spread achievable at the linac end as a function of the RMS bunch length,for different bunch charges. The black line indicates the required value of 0.1% for injectionin the SPS.
Considerable advancement in the understanding and control of the effects of long-range wakefields onthe beam stability has been made throughout the last few decades, for example within the context ofthe CLIC Study, supported by experimental verification at SLAC/FACET [26]. The tools developed forthe CLIC Study have been used to evaluate the maximum bunch charge that can be transported throughthe 3.5 GeV electron linac. The natural suppression of the long range wakefields, due to the taperingof the iris aperture, also guarantees beam stability in multi-bunch operation for all 40 bunches in thenominal 200 ns-long train. A bunch charge of up to 300 pC can be transported with negligible emittancegrowth through the linac without high-order mode damping. Semi-analytical estimations show that theaverage amplification factor of incoming beam offsets as a function both of the bunch charge and of thestructure’s Q factor, in case of a coherent offset of all bunches in the train: even with no damping (thatis, small Q ), a charge of 300 pC is transported through the linac with a small amplification factor of lessthan 2, as shown in Fig. 3.2. Four operational modes have been studied: • nominal: for a missing momentum experiment, trains of 40 bunches of 50 pC, spaced by 5 ns areaccelerated; • high-charge: trains of 40 bunches with charge up to 300 pC and bunch length about 200 µ m areaccelerated; • very-high-charge: single bunches with 1 nC charge and length about 750 µ m are accelerated; • plasma R&D: for plasma acceleration R&D, single bunches of 1.7 nC and length less than 250 µ mare accelerated. 17 Linac
Figure 3.2: Multi-bunch operation: average beam offset amplification due to long-range wakefields as afunction of bunch charge and the structure’s Q factor, due to a coherent incoming offset of thebunch train. Bunch charges below 300 pC show a factor 2 amplification even without HOMdamping.In the latter case the goal is to achieve the required high electron density at the plasma cell. Longitud-inal bunch length compression would need to be implemented though a conventional 4-bend magneticchicane. The initial and final bunch parameters, of all cases, are listed in Tab. 3.1. Other operationalscenarios could be considered, like filling all buckets of 800 MHz RF system in the SPS: 160 bunches,50 pC (or 50/4), ≈ Parameter Nominal High-charge Very-high-charge Plasma R&D
Bunch charge [pC] 50 300 1000 1700Bunch length [ µ m] 150 200 750 250Bunches per train [ µ m] <100Initial energy [MeV] 200Initial relative energy spread [%] 1Final energy [MeV] 3500Final relative energy spread [%] < .
1% 0 .
2% 3% 0.5%
In order for the eSPS linac to operate in each of the four operational modes mentioned in Sec. 3.1.3, theinjector must be flexible enough to provide beams of several different types, the parameters of whichare summarised in Table 3.1. The beam dynamics in the injector are dominated by space-charge effectswhen the beam is at a low energy in the gun and bunching cavity. The beam parameters can be adjustedby altering the pulse length and spot size of the laser, the electric field strength and phase of the RFgun and bunching cavity and, the magnetic field strength of the gun and buncher solenoids. The current18
Linac instillation of the CLEAR injector uses a laser with a fixed pulse length of 4.7 ps, and fixed electric fieldstrengths of 80 MV/m and 18 MV/m in the gun and bunching cavity respectively.The beam in the injector was simulated using the code ASTRA, to calculate the effect of space-chargeforces [27]. Any wakefield effects in the S-band linac were considered negligible so were not includedin simulation. Any misalignments and losses were also omitted. To verify the validity of the ASTRAmodel, the simulated beam was compared to experimental measurements of the beam in the CLEARuser facility. Different bunch charges of up to 2000 pC with laser spot sizes of an RMS radius of up to1.2 mm were investigated.When accelerated on the maximum energy phase in the bunching cavity the bunch length stays aroundthe 4.7 ps ( ∼ µ m) of the laser pulse. To achieve shorter bunches the phase of the buncher is reduced,with maximum compression achieved around zero crossing, -90 ◦ below crest. The simulated bunchlength at different phases of the buncher gave reasonable agreement with experiment up to a charge of500 pC on CLEAR, shown in Fig. 3.3. The minimum bunch length achievable depends on the bunch − − − − −
20 0 20Buncher Phase [degrees]12345 R M S B un c h L e n g t h [ p s ] Q = 0.45nCQ = 0.5nCQ = 0.55nCExperimental Data
Figure 3.3: A comparison of simulated and experimentally measured bunch lengths at different phasesof the bunching cavity on CLEAR for a laser spot size of 0.6 mm. The phase, 0 ◦ , is definedas the phase of maximum energy gain.charge, due to the increased space-charge forces in higher charge bunches. To reduce the effect of thespace charge forces, thus to maximise compression, the size of the laser spot can be increased. Theminimum lengths achievable in simulation for different laser spot sizes and bunch charge are shown inFig. 3.4. In these simulations it is assumed that the RF gun is operating at the peak energy phase. Byde-phasing the gun by up to ∼ ◦ , shorter bunch lengths can be achieved for bunches of charge less than ∼
500 pC. For the nominal 50 pC case, a bunch length of less than 0.1 ps (30 µ m) can be achieved in bothsimulation and experiment. Without any gun de-phasing the bunch lengths required for the high-chargeand very-high-charge modes can be satisfied. The bunch length for the Plasma R&D beam may be ableto be achieved with velocity bunching alone, but the tolerance will be very fine. Increased compressionwith a magnetic chicane may be desirable for this operational mode.When beams are accelerated with the crest energy phase in the buncher the energy spread of eachbunch simulated is less than the required 1%. When compressing, the energy spread grows. Whenoperating at maximum compression phase in the buncher, the energy spread of the beam after the buncheris less than 4%, for each bunch up to 2000 pC. After the two following accelerating structures, the energyspread is reduced to less than 1% for each bunch charge. It must be noted that for compression phases inthe buncher between the peak energy and maximum compression phases the energy spread is larger than1%, with a maximum of ∼
4% at the end of the injector.The emittance tolerance for the eSPS linac in each of the operational modes is quite large at 100 µ m.19 Linac . . . . . . . . . σ laser [mm]0 . . . . . . . . M i n i m u m R M S B un c h L e n g t h [ p s ] q = 0.01 nCq = 0.1 nCq = 0.5 nCq = 1.0 nCq = 1.5 nCq = 2.0 nC Figure 3.4: Simulated bunch lengths in CLEAR, at maximum compression for different laser spot sizesand different charges.When accelerating on crest in the bunching cavity, the emittance for bunches up to a charge of 2000 pCremains below 20 µ m. When operating at a compression phase in the buncher, the emittance growsrelative to the peak energy phase by around a factor ∼
3. This growth can be suppressed by optimisingthe strength of the buncher solenoid. After an optimisation, the emittance growth is reduced to a factor ∼ µ m for a 2000 pC beam produced with a laser spot of 1.2 mm.There are likely to be several small upgrades made to the CLEAR injector for use as the eSPS linacinjector. It is proposed to use two independent klystrons to power the gun and bunching cavity, instead ofthe current setup of one. Therefore, the electric fields in each would be able to be adjusted independently.A higher field in the gun would reduce the emittance growth and bunch lengthening due to space-chargeeffects. The gun would also be more able to operate at a lower phase for high charge bunches allowingcompression in the gun. The added klystron would also allow the optimisation of the buncher field tomaximise compression. The use of a new laser, with a shorter laser pulse, would also enable the creationof shorter electron bunches. The emittance of the beam could also be significantly reduced by optimisingthe drift length between the gun and the buncher. In the presence of imperfections beam-based alignment techniques must be applied in order to preservebeam quality. The alignment procedure follows the experience made in the context of linear colliderstudies, and relative experimental tests, and consist of three steps to be applied in cascade:1. Orbit correction, where dipole correctors are used to centre the beam through the BPMs;2. Dispersion-Free Steering (DFS), where the dipole correctors are used to remove residual dispersionintroduced by misaligned quadrupoles;3. Wakefield-Free Steering (WFS), where the dipole correctors are used to remove wakefield kicksintroduced by misaligned accelerating structures.In order to perform trajectory correction, it has been assumed that all quadrupoles are equipped with abeam position monitor, and corrector coils to deflect the beam transversely in both horizontal and verticaldirection. Table 3.2 lists the RMS imperfections that have been considered. All simulation results shownare the average of 100 randomly misaligned machines, corrected with beam-based alignment. Figure 3.520
Linac shows the relative emittance growth in presence of static imperfections for the four operational scenariosdescribed. Table 3.2: List of imperfections.
Imperfection RMS Value
Quadrupole offset [ µ m] 100Quadrupole pitch [ µ rad] 100Structure offset [ µ m] 100Structure pitch [ µ rad] 100BPM offset [ µ m] 100BPM resolution [ µ m] 10Given the relatively large initial emittance, the impact of errors is small in both the nominal andhigh-charge cases; in the case of very-high-charge or plasma R&D case the application of beam-basedalignment techniques is necessary to reduce the emittance growth. The CERN Linear Electron Accelerator for Research (CLEAR) injector operating at CERN can be usedbasically unchanged to produce the electron beam with the specifications required for missing momentumand beam dump experiments as well as for many other potential applications [24].
The present installation is about 25 m long and its schematic layout is shown in Fig. 3.6. The electronbunches are generated on a Cs Te photo-cathode by a pulsed UV laser. This allows the generation ofa beam with arbitrary time structure with the minimum bunch spacing of 0.33 ns given by the 3 GHzRF frequency of the gun (at present the spacing is limited to multiples of 0.66 ns by the laser system,a modification will be needed to be able to get multiples of 0.33 ns). The gun is followed by threeLEP Injector Linac (LIL) 4.5 m-long accelerating structures which are used for beam bunching andacceleration. The first structure can be used to vary the bunch length from 0.3 to 1.2 mm r.m.s. by meansof velocity bunching. The gun, buncher and first accelerating structure are inside solenoid magnetswhich provide tuneable focusing and space charge compensation. A matching section with three tuneablequadrupoles and a spectrometer line complete the injector. The range of beam parameters which can beobtained at the end of the CLEAR injector are summarised in Table 3.3.Table 3.3: Beam parameters at the end of the CLEAR injector.
Parameter Value range Value for eSPS
Energy [MeV] 50 to 250 200Bunch charge [nC] 0.001 to 1.5 0.05Norm. emittance [ µ m] ∼ ∼
20 for 0.4 nC/bunchBunch length rms [mm] 0.3 to 1.2 0.8Energy spread rms [%] below 0.2 0.1Number of bunches 1 to 200 40Micro-bunch spacing [ns] multiple of 0.33 521
Linac
Q = 50 pC (nominal)
Q = 300 pC (high-charge)
Q = 1000 pC (very-high-charge)
Q = 1700 pC (plasma R&D)
Figure 3.5: Relative emittance growth along the linac in presence of static imperfections, before and afterbeam-based alignment. In the nominal and high-charge cases, the emittance growth remainsbelow 10%. In the case of very-high-charge and plasma-R&D, WFS is needed to reduce theemittance growth to below 20% and 40%, respectively.Figure 3.6: Layout of the CLEAR injector. The electron beam travels from right to left [24].22
Linac
The present installation is located in B2010, with some technical equipment being hosted in adjacentbuilding. All the equipment will be relocated in TT5 and in B183. A few upgrades and modificationsof the CLEAR linac are also planned, the main being the implementation of a new Modulator/Klystronstation, in order to provide more operational flexibility. At present, the RF gun and the first accelerationstructure, used as a buncher, are fed by a single klystron (see Fig. 3.6). In the new configuration, theywill be independently fed by two klystrons, giving the possibility of completely independent adjustmentof the RF amplitude and phase, the use of compression only in the buncher, and slightly increasingthe maximum obtainable beam energy. The present laser system should also be modified in order toprovide pulses with a spacing of 0.33 ns instead of the present value of 0.66 ns. A likely better alternativeis to substitute it with a new system, simplified and adapted to the eSPS requirements. Other minorupgrades and rearrangements will likely concern beam diagnostics and ancillary equipment. Some ofthese upgrades are under study for the next few years of CLEAR operation, therefore, the details willdepend on their advancement at the time of installation in TT5.In the new implementation layout (see Fig. 3.7) a 10 m free space has been reserved after the CLEARinjector and before the linac, to host dedicated beam diagnostics and possibly a three bends chicane,which will add flexibility for bunch compression especially at high bunch charge. In this space, a beamFigure 3.7: TT5 Layout.line will branch off the main one and give the possibility to send the beam back, through a 180 degreesachromatic arc cell, towards the low energy experimental beamline, hosted in an independently shieldedzone within TT5.
Unlike the injector where S-band (3 GHz) RF structures are used at a relatively low accelerating gradientof about 15 MV/m in order to obtain the desired beam parameters, high gradient X-band (12 GHz) RFaccelerating structures are used in the linac in order to make it compact and accelerate the beam from0.2 GeV up to 3.5 GeV within 70 m. A high gradient is required due to limited space in the TT4/TT5 areaat CERN. The X-band high-gradient RF technology has been developed in the framework of the CLICstudy and is now being widely adopted. The high-gradient X-band RF systems have been extensivelyoperated at CERN Xbox1, 2 and 3, as well as in several linac based light sources: SwissFEL, FERMI,and LCLS, where it is used for beam manipulation and beam diagnostic purposes.23
Linac
For the klystron-based option of the first stage of CLIC at 380 GeV [28], an average loaded accelera-tion gradient of 75 MV/m (95 MV/m unloaded) has been chosen as a compromise between making themain linac as short as possible and reducing the required peak power and associated number of klys-trons. About one klystron per metre is necessary to feed the klystron-based CLIC main linac. This veryambitious specification requires development of a special compact modulator unit accommodating twohigh-efficiency klystrons. Going to slightly lower gradients for the eSPS reduces the peak power permetre and simplifies the integration in the available space in TT4/TT5 area. This makes it possible to usecommercially available klystron/modulator units used, for example, in XBOX2 facility at CERN. Theschematic layout of the RF unit is presented in Fig. 3.8. The total length of one RF unit is 2650 mmFigure 3.8: Schematic layout of one X-band RF unit. Dimensions are given in mm.including 2400 mm for four accelerating structures and 250 mm for magnetic quadrupole, BPM and in-terconnections. Sec. 3.4 provides more details on the integration of different components into an eSPSmodule and how the R&D done for the klystron-based option of CLIC has been used. In particular, forthe RF waveguide (WG) network connecting X-band klystron with four accelerating structures compon-ents designed for CLIC has been used. Based on this, total power loss from the klystron to the input ofthe accelerating structures has been estimated to add up to about 13%. Table 3.4 summarises losses ineach component and also indicates whether single or double height is used.Table 3.4: Power loss in the X-band RF WG network.
Component Double height [Y/N] Power loss [%]
Long straight with bend Y 4.5CCC Y 2.3BOC Y 2.23dB-Hybrid Y 0.3Directional coupler Y 0.3Straight with 2 bends Y 2.23dB-Hybrid N 0.5Straight with bend N 0.5Total 12.8The RF WG network includes an RF pulse compression system designed for klystron-based CLICthat includes SLED-type pulse compressors based on Barrel Open Cavity (BOC) with Correction Cavity24
Linac
Chain (CCC). Power gain versus compression ratio curve calculated for CLIC and a pulse shape with aflattop are shown in Fig. 3.9.Figure 3.9: Power gain versus compression ratio for pulse compression system based on BOC + CCC isshown.For eSPS it is envisaged to use commercially available CPI VKX-8311A tubes. One tube can generate50 MW, 1500 ns RF pulses at 100 Hz repetition rate. Based on this klystron pulse length a compressionratio of five has been identified as good compromise between achievable power gain (3.2) and com-pressed pulse length (300 ns) available for filling acceleration structure (AS) (100 ns) and acceleratingthe train of bunches (200 ns).The total flange to flange length of one accelerating structure is 600 mm. Using the CLIC designfor structure interconnection, an active length of 575 mm per structure can be used for the RF design.With these two constraints on the filling time and the active length. The accelerating structure has beendesigned as a quasi constant gradient structure with linear tapering of the iris radius. In Fig. 3.10, thedistribution of surface field quantities, accelerating gradient, and power along the structure for a loadedgradient of 60 MV/m averaged over the structure active length are shown (solid lines) for the nominaltrain of 40 bunches of 50 pC every 5 ns for a missing momentum experiment. The dashed lines show theunloaded case, which is very close to the loaded one since the peak beam current in the train is very low(10 mA) and the beam loading effect is very weak and hardly visible on the plot.The parameters of the X-band linac for an energy gain of 3.3 GeV are summarised in Table 3.5 for twocases: the nominal case with 24 RF units in operation and the sub-nominal case where one RF unit isdefective and the corresponding energy gain is compensated by increasing the gradient in the remaining23 RF units.In both cases there is a large enough operational margin in the required klystron peak power foran overall energy gain of 3.3 GeV since it is significantly lower than the klystron maximum specifiedpower of 50 MW. For operation in single bunch mode the compressed pulse is shorter (approx. 100 ns)and higher compression factors can be used. However, since the RF pulse compression system andaccelerating structure parameters (summarized in Table 3.6) are optimised for multi-bunch operation andcannot be adjusted for the single bunch mode the improvement in the effective power gain is limited andestimated to go up to 3.5 only. This results in a rather small increase in the energy gain in the X-bandlinac for the same klystron input power. Keeping the same operational margin for nominal parameters inTable 3.5, the beam energy goes up from 3.3 GeV to 3.45 GeV only.25
Linac
Figure 3.10: Distribution of accelerating gradient (red), surface electric field (green), S c (magenta),power (black) and pulse surface temperature rise (blue) along the accelerating structure isshown for 60 MV/m average loaded accelerating gradient (solid lines). The unloaded caseis also shown in dashed lines.Table 3.5: Parameters of the X-band linac for energy gain of 3.3 GeV. Parameter Nominal Fault
Klystron pulse length [ns] 1500Power loss in WG [%] 13Compression factor 5Power gain 3.2Number of RF units 24 23Energy gain per RF unit [MeV] 137.5 143.5Average acc. gradient [MV/m] 59.8 62.4Required peak power per klystron [MW] 43.5 47.4The structure is designed without any higher order mode suppression features, except the detuningcaused by the iris tapering. The frequency of the first dipolar Higher Order Mode (HOM) changes from16 GHz in the first cell to 17.2 GHz in the last cell. The difference in frequency of 1.2 GHz results ina suppression of the wakefield amplitude by an order of magnitude within 1 ns, resulting in a very lowvalue of the wakefields at the position of the second bunch. Since the structure has 69 cells, wakefieldre-coherence takes place and the amplitude increases every 60 ns. Nevertheless, ohmic losses in copper26
Linac
Table 3.6: Parameters of the X-band accelerating structure.
Parameter First cell Last cell
Aperture iris radius [mm] 3.7 2.7Iris thickness [mm] 1.35 1.35Q-factor 7090 7020Group velocity [% of c ] 3.6 1.34 R (cid:48) / Q [k Ω /m] 14.3 17.5Number of cells 69Active length [mm] 575Input power for <60MV/m> [MW] 30.5Rise time (1/bandwidth) [ns] 12Filling time [ns] 88walls help to reduce the wakefield amplitude on this longer time scale. The amplitude of the transverselong range wakefield at the positions of all 40 bunches is presented in Fig. 3.11 demonstrating goodwakefield suppression.Figure 3.11: Magnitude of the transverse wakefields in AS is shown for the first dipole mode at theposition of the bunches with 5 ns spacing (circles) and 1.25 ns (dots). Dashed line is theenvelop of Q=5000 mode for comparison. As described in Sec. 3.3, four 0.6 m long accelerating structures are assembled in one eSPS Module(see Fig. 3.12). Each of the structures is aligned by adjustable supports that are envisaged to provide analignment precision of the accelerating structures within 10 µ m rms with respect to the beam axis. The27 Linac
Figure 3.12: Side view of the eSPS Module, with longitudinal dimensions.Module is also supporting a quadrupole magnet and its associated Beam Position Monitor (BPM). In thiscase it has been chosen to adopt an adjustable support for the BPMs based on a platform developed by theSurvey Group that can be equipped with motors, thus allowing realignment of the quadrupole during theaccelerator operation or using it to introduce beam steering. The possibility to adopt integrated steeringdipoles is also considered in order to avoid the complication introduced by the motorised displacementof the quadruple magnet. The accelerating structures and BPMs are sitting on a steel girder, which issupported and aligned by means of jacks that have already been used for Linac4 installation.A total of 24 Modules is required to achieve the final electron energy of 3.5 GeV. Due to the limitedavailable space in the installation area an as compact as possible layout has been adopted (see Fig. 3.13).Each Module is fed by a single 12 GHz klystron delivering a 43.5 MW, 1.5 µ s RF pulse to a pulse com-pression system. A compressed RF pulse 300 ns long is delivered to each accelerating structure with30.5 MW peak power. The power delivery system adopts waveguide components that have been de-veloped for the CLIC accelerator and have been built and tested at the Xbox test stations. We estimatethat 13 % of the power is lost in the power distribution system. The average power dissipation is 0.9 kWfor each TW accelerating structure and 0.6 kW in the output load, if the 100 Hz operation is assumed.The accelerating structure temperature stabilisation is achieved by demineralised water circulation atthe reference temperature of 28 ◦ C. The water consumption is estimated in 6 litres/min per acceleratingstructure.
The source and injector will be reusing the existing CLEAR facility along with its current beam instru-mentation. The latter will benefit from the most recent development in beam instrumentation performedon CLEAR for low beam charge using inductive BPMs [29]. The beam position monitors along the X-band linac will be based on RF cavities also developed for the CLIC main linac [30]. There will be onemonitor per quadrupole, i.e. 24 BPMs in total for the full linac. The beam energy, emittance and bunchlength will be measured at the entrance and the exit of the linac to monitor the evolution of beam prop-erties before and after acceleration. Beam emittance and energy measurements will be based on Optical28
Linac
Figure 3.13: Two eSPS Modules in the accelerator tunnel.Transition Radiation (OTR) beam-imaging systems (BTV). Emittance monitoring is performed at theentrance and the exit of the linac using optical systems capable of providing resolution better than 10 µ m[31]. The beam energy is measured in spectrometer lines equipped with optimised screen materials andshapes developed for the CLIC test facility 3 (CTF3) [32, 33]. Non-invasive bunch length measurementswill be performed non-invasively using coherent radiation [34] or electro-optical techniques [35]. There is an enormous interest in studying positron production and acceleration for future lepton colliderprojects. Worldwide only a few positron sources are still operated and usually part of facilities withlittle time for R&D. Efficient positron production typically requires a primary electron beam of a fewGeV which is then converted into a positron beam using a dense target. The availability of a 3.5 GeVelectron beam, therefore, presents an excellent opportunity to study positron production and possibly re-acceleration. A first step could be to implement a converter target area into the facility which allows thestudy of target technologies and their limitations, a very critical area in the design of a reliable positronsource. The positron yield which can be achieved is a figure of merit for these sources. A second criticaldevice is the so called adiabatic matching device, a strong pulsed magnet which focuses and collects thelow energy positrons for re-acceleration. A promising place for such a positron source R&D area couldbe the first alcove within the TT61 transfer tunnel. The primary beam could be separated with a doglegfrom the main SPS injection line allowing the installation of a positron target, a capturing device andsome diagnostics including a beam dump. The location in the underground tunnel would not requireexcessive shielding to be able to perform such experiments. Figure 3.14 shows a schematic of the CLICpositron source consisting of a hybrid target and an adiabatic matching device.The CLIC positron source[28, 36] represents a typical example of such a source and is similar to theone proposed for FCC-ee. In order to provide high quality, high energy positrons for experiments, thepositrons have to be captured and re-accelerated to a few 100 MeV and cooled in a damping ring. Afterthat they can be transported back to the beginning of the linac and finally accelerated to 3.5 GeV. Theyield of such a system would be of the order of one, which would result in positron bunches relevant forlepton collider studies. These beams could be used for plasma based acceleration studies with positrons,29
Linac
Figure 3.14: Schematic layout of a positron production target consisting of a tungsten crystal and anamorphous tungsten target followed by a adiabatic matching device.which is still an unsolved problem for future linear colliders based on plasma acceleration, see Sec. 6.6.Another possibility would be to inject the positrons into the SPS for further acceleration and for dampingring studies or Muon collider studies based on the LEMMA scheme. It is, however, not straightfor-ward to upgrade the proposed eSPS linac facility to a high energy positron source without additionalcivil engineering. Nevertheless such an integration could be feasible as shown by the proposed FACETII facility which includes a positron production target, damping ring and re-acceleration all within themoderately sized SLAC main tunnel[37]. To summarise, the implementation of an area for positron pro-duction studies seems straightforward and would require only minor modifications of the beam transferline down to the SPS in TT61 while a full energy positron beam available for experiments represents alarger integration challenge. More details about the potential research program of such an upgrade of thefacility can be found in Sec. 6.6.
The possibilities and the potential for plasma-based acceleration Research and Development taking ad-vantage of the 3.5 GeV electron linac are described in Sec. 6.4. For these kind of experiments the highenergy 3.5 GeV linac with its excellent beam quality is used to accelerate the drive bunch. The witnessbunch has to be truly independent to be able to probe the wakefields at variable distances without com-promising the beam quality of the witness bunch. This can only be achieved with a second independentinjector. Therefore, to fully exploit the potential of such experiments, a second injector could be addedin an upgrade of the facility. An energy between 100 and 200 MeV would be sufficient for this purpose.Further requirements as explained in Sec. 6.4 are a very short bunch length of the order of 140 fs (42 µ m), a moderate charge of 100 pC and a small emittance to be able to match the witness beam to theplasma wakefields. Such an injector has been studied in detail for the AWAKE project[38] and could befitted into the facility at the beginning of the TT61 tunnel. The injector consists of an S-band RF-gun fol-lowed by X-band structures for velocity bunching and acceleration. The beam produced by the injectorcould be connected with a achromatic dogleg onto the main beam axis at the end of the linac. After thedogleg an approximately 2 m long experimental area would be needed to accommodate a plasma cell andcorresponding diagnostics. Finally a spectrometer capable of beam energies of up to 10 GeV is neededat the end of the beam line. A schematic drawing of such an implementation can be seen in Fig. 3.15.Space for two more modulators units can be found extending slightly in the klystron gallery.The injector itself could use a slightly modified X-band acceleration module as described for the mainlinac. This would provide up to 160 MeV of acceleration when powered by a single X-band klystron.According to the AWAKE study the witness beam injector could produce the required beam parameters,see Table 3.7.Beside the space for the experimental area at the end of the X-band linac, no hardware modificationsare necessary within the main linac and injector. However, the drive beam requires a high charge of1.7 nC which has to be accelerated and finally compressed to a bunch length of 0.8 ps (240 µ m). Beamdynamics studies have shown that the acceleration of such a bunch is feasible with the linac describedin Sec. 3.1. Such a facility would represent unique opportunities to study beam driven plasma wakefield30 Linac
Figure 3.15: Schematic layout for a second injector dedicated for plasma wakefield acceleration research.The injector could be installed at the beginning of TT61 and connected via a dogleg to anexperimental area at the end of the linac.Table 3.7: Tentative Beam parameters for a witness beam injector.
Parameter Value
Energy [MeV] 160Bunch charge [pC] 100Norm. emittance [ µ m] 1Bunch length rms [fs] 140Energy spread rms [%] 0.1acceleration (PWFA). For the time being, there are no facilities in the world providing truly independentdrive and witness bunches, which is essential to study emittance preservation, beam loading and energyspread compensation in detail and with high precision for collider applications.31 SPS and Transfer Lines
From 1989, and for more than a decade, the SPS was used as injector for the LEP. It accelerated electronsand positrons coming from the PS from 3 . Electrons from the linac in Transfer Tunnel 4 (TT4) will be transported through TT61 to the injectionsystem in SPS Long Straight Section 6 (LSS6). Figure 4.1 presents a conceptual design of these transferline optics. Beam size is represented using a 4 σ envelope with the beam parameters mentioned inTable 3.3 and includes maximum trajectory offsets of 2 mm.Figure 4.1: Beam envelopes and optical functions from the linac to the SPS. The linac matching section is composed of 6 independently powered quadrupoles. Those quadrupolesof type QTN allow matching of the beam envelope coming from the linac to larger sizes that are moreconvenient for the long and mostly straight transport in the TT61 tunnel. Bending of the trajectory inboth horizontal and vertical planes is needed to transfer the beam from the linac to the axis of the TT61.A total of 4 water-cooled dipoles of type BH2 type2 will be used in the linac matching section.A long FODO straight section uses QTR type quadrupoles in a regular arrangement of 2 familiespowered in series. Those low-field quadrupoles provide sufficient focusing strength, large aperture andpresent the advantage of being air-cooled. The use of air-cooled magnets in TT61 for around 400 mminimises the costs by forgoing the need to supply cooling water.From the end of the FODO section the beam joins the TI2 and TT61 lines currently used for transferof the proton beam from the SPS to the LHC. A further 3 dipole magnets are used to align the trajectory32
SPS and Transfer Lines of the electron beam with the TI2 line. The existing magnets and powering scheme in TI2 are used toprovide the optics and beam sizes shown in Fig. 4.1. However, this study does not consider power supplystability or remnant fields at the very low currents the magnets of the TI2 and TT61 lines will need tobe operated. Table 4.1 lists the existing magnets that are considered to be re-used in the design of theelectron transport line, only needing at most a refurbishment before installation.Table 4.1: List of existing magnets used in the linac to SPS beam transport line design.
Name Type Quantity Max. current [A] Max. integrated strength
QTN Water-cooled quadrupole 6 150 2 . . . Matching of the beam to the SPS lattice is imperfect despite the many independent quadrupoles availablein TI2/TT61. In particular, a mismatch of the horizontal dispersion causes a large horizontal beam sizein the SPS ring. A more careful design of the transfer line optics and FODO section may cure thismismatch but is not critical for our application. As long as the beam fits within the acceptance of theSPS, the synchrotron radiation damping will ensure that beam characteristics quickly converge towardstheir equilibrium values.The injected bunch structure will depend on the linac injector and the SPS RF system frequency. Weconsider here that the linac will produce 200 ns trains of bunches. A fast bunch-to-bucket fast injectionscheme, very similar to the one in used during LEP era [4], will synchronise the injection of the buncheswith the SPS buckets. The number of bunches per train will depend on the frequency of the SPS RFsystem and the experimental requirements. It could be as little as 40 bunches for a 5 ns spacing betweenbunches or as much as 160 bunches to fill the SPS with an 800 MHz RF system.Figure 4.2 shows the injection trajectory through the side channel of the quadrupole QDA61910 fol-lowed by bending by the MSE septa and deflection onto the SPS orbit by a new fast kicker. As duringFigure 4.2: Horizontal electron injection trajectory in the SPS LSS6 region.LEP operation in LSS1, the MSE will be powered at low field using a new low current power supply andthe MST will be powered-off but demagnetised using another power supply.Short trains of 200 ns coming from the linac require a kicker system with fast rise or fall time, to allowmaximum filling of the ring. A transmission line of magnets composed of two 50 cm tanks could achieve33
SPS and Transfer Lines a rise time of 100 ns and the desired flat-top of 200 ns to provide the required kick of 500 µrad. The useof solid state switches will allow a 100 Hz repetition rate. The SPS will be filled by up to 70 trains withinless than a second. The new kicker system has to be designed and built but will largely make use ofexisting technologies and CERN competencies.
Fourteen new inductive beam position monitors [29] are required to measure the beam trajectory throughthe transfer line, with two new Optical Transition Radiation Beam TeleVision (OTR-BTV) systems [31]foreseen to measure the average beam size at the beginning and the end of the transfer line.In addition, two new OTR-BTV systems [31] will also be required to measure the average beam sizeand position upstream and downstream of the new injection kickers to be installed in the SPS LSS6section.
After injection into the SPS, the electron bunches should be accelerated from 3 . th harmonic of the injectorfrequency. Higher frequencies, up to 800 MHz, open the possibility for increased number of bunches(x4) and give reduced beam loading for high current beams. Frequencies of 200, 400 and 800 MHz arealready in use at the CERN accelerator complex with significant operational experience of both normaland superconducting RF systems. Although different options at these frequencies were considered, onlythe use of the normal conducting 200 MHz cavities from the LEP era and a superconducting option at800 MHz are discussed below. The LEP-era 200 MHz cavity design and the RF system are described in [39–41]. Although 13 of LEP-era 200 MHz RF cavities were stored in good conditions and are still available for re-installation into theSPS ring, only a few of the RF auxiliaries are still available. Most of the HOM dampers, fundamentalpower couplers and tuners have to be rebuilt. In addition, a new high power RF system including 60 kWpower amplifiers for each cavity has to be built and installed in the SPS together with the new LLRFsystem. In general, the 200 MHz RF system, accelerating leptons in the SPS used as a LEP injector, wasvery robust. It worked for 12 years and had a very little impact on the down time, also during the protonoperation for fixed target experiments.All RF systems for lepton acceleration were removed from the SPS, along with many other items,when the LEP was decommissioned. This allowed to reduce the impedance of the SPS ring which wasnecessary for the production of high intensity (nominal) proton beams for the LHC. To double the bunchintensity for HL-LHC, a broad impedance reduction campaign is underway in the framework of the LIUproject [42]. Re-installation of 12 accelerating cavities for electron acceleration in the SPS ring will re-quire a detailed study of the impact of their impedance on beam stability for proton operation includingboth their low-frequency inductive part ImZ/n and narrow-band HOMs. The so-called low-frequencyImZ/n, will be increased by 0 . Ω , comparable with overall SPS impedance after LIU upgrades. Thispreliminary estimate was done using the values published in [43]. A 200 MHz normal conducting RFsystem designed for beam capture in the LHC [44] could also be a potential alternative, but the compat-ibility with HL-LHC beams in the SPS have to be assessed, similar to the LEP-era cavities.34 SPS and Transfer Lines
In the framework of the HL-LHC project, a new vacuum sector in the SPS LSS6 region was createdto test superconducting crab cavities with proton beams [45]. This sector comprises two Y-shaped va-cuum chambers articulated by mechanical bellows, the circulating proton beam line and the beam bypassconsisting of an RF module. The mechanical bypass is equipped with a movable table to move the cryo-module in and out of the circulating beam path. A dedicated RF system, cryogenic system and generalinfrastructure were put in place on the surface (BA6) and in the tunnel (see Fig. 4.3), and then success-fully operated with beam during 2018. The mechanically movable bypass allows the installation of anFigure 4.3: Schematic view of the tunnel and surface installation of the cryogenic, mechanical and RFinfrastructure in the SPS [45].RF module operating at 800 MHz for a dedicated mode of operation with electrons. This configura-tion alleviates the strong constraints of impedance requirements for the high intensity proton beams andother modes of SPS operation. The remotely operated bypass with 2 K helium, allows a rapid exchangebetween the dedicated mode and regular operation in about 10 minutes, already demonstrated in 2018.An alternative method could foresee a fixed dog-leg at the same location using the SPS magnets (or ad-ditional elements) to allow the electrons to circulate through the bypass with the RF cryomodule whilethe protons circulate through the empty vacuum chamber. Feasibility of such a configuration requires adetailed study on magnetic requirements, integration and aspects related to synchrotron radiation in thedog-leg region.Recent studies in the framework of the FCC study have led to design of the 800 MHz cavities rangingfrom single-cell for high beam currents to five-cell structures for the high energy and lower beam cur-rents [46, 47]. A schematic of two five-cell cavities housed in a cryomodule and suitable for accelerationof the electrons from 3 . / m per cavity to keep theFigure 4.4: Two-cavity concept for a cryomodule using five-cell 800 MHz cavities for the bypass option.RF power to moderate level and be compatible with the existing crab cavity RF chain. A five-cell cavityin bulk niobium was built in collaboration with Jefferson lab and obtained an accelerating gradient of30 MV / m [48].The length of such a cryomodule is approximately the same as the crab cavity module presently in-stalled. The two-cavity option is also compatible with the rest of the infrastructure needs. Due to the35 SPS and Transfer Lines
Table 4.2: Relevant parameters of the RF systems in the SPS for electrons and, where applicable, forprotons in parentheses.
Unit LSS6 bypass Inline Inline
Cavity type SC NCFrequency [MHz] 801.58 200.39Number of cells 5 1 1Number of cavities 2 4 12Voltage/cavity [MV] 5.0 2.5 ( < Ω ] 196 45 ≈ ≤ ≤
20 ( ≈ th harmonic (800 MHz) RF system. The two IOTs were modified toprovide power at 400 MHz for the crab cavity tests and can be readily transformed back to their originalconfiguration at 800 MHz. First estimates of the RF power requirements are listed in Table 4.2, but exactpower requirements and compatibility with the existing IOTs will need further studies. Any additionalpower will have to be compensated with an extra IOT per cavity. The power is transferred to the cavitiesin LSS6 zone via RF coaxial lines of approximately 150 m in length. Taking into account the RF losses inthe coaxial lines, it is estimated that only 40 kW-CW might be available at the cavity input. Two speciallydesigned high-power V-shaped RF lines with rotating connections provide the required movement tofollow the table movement of 510 mm. The cavity control (aka LLRF) is located in BA6 next to the IOTswith its own Faraday cage, electromagnetically shielding it from neighbouring equipment.A new cryogenic system, including a mobile refrigerator, provides the required cooling capacity in theSPS.The warm compressor unit, located at the surface (see Fig. 4.5), is connected to the cold-box under-ground by two warm lines for high and low pressure gas. The cryogenics system is capable of providinga total cooling capacity of 3 . / s at 30 mbar (saturation at 2 K). This translates to approximately a totalheat load (static and dynamic) of 48 W at 2 K or equivalent to 2 g / s including a safety factor of 1 . SPS and Transfer Lines and very strong damping of HOMs using single-cell cavities. A schematic of four single-cell cavitiescomprising a cryomodule to provide the 10 MV voltage is shown in Fig. 4.6. Table 4.2 includes someFigure 4.6: Four-cavity concept for a cryomodule using one-cell 800 MHz cavities for an in-line option.relevant RF parameters for both five-cell and single-cell cavities. The present infrastructure in BA6 hasto be increased by a factor of two or more including the change to higher power RF amplifiers and RFwaveguides to be able to operate in this configuration. The long-delay loop for the RF control will bevery challenging to maintain strong feedback. The cavities will have to be operated at fixed frequencydue to the slow tuning capabilities of SC cavities. The compatibility of a such a system with the SPShigh intensity proton operations requires a detailed study for the multitude of the SPS beams.
In the inline option, the eSPS will operate in parallel with the proton operation, supplying beam to boththe fixed target experiments and to the HL-LHC. In order to provide the increased intensity requiredby the HL-LHC (compared to the nominal LHC) the LIU project has been executed [42]. One majorcomponent of this was an impedance reduction campaign in the SPS encompassing improved HOMdamping in the 200 MHz RF cavities, vacuum flange shielding and optimization of the electrostatic septa.The 200 MHz RF cavities themselves were shortened and rearranged to cope, together with upgradedLLRF system, with high beam loading of HL-LHC beam. In addition, equipment exposed to the LIUbeams must be able to withstand the high intensities if parasitic energy loss is caused by the device.This type of loss often manifests itself in the form of reduced peak accelerating voltage or damage to theequipment itself.In addition to the RF system, equipment required for injection, extraction and specific beam instru-mentation must also be installed. For injection there will be two additional kickers and for extraction oneelectrostatic and one magnetic septa. Experience during LHC runs I and II shows that exposing thesetypes of equipment to the proton beam can result in excessive heating [49] and possible sparking [50]. Inorder to mitigate these effects the broadband impedance of any kicker magnet installed must be reducedeither through minimizing the use of ferrites or shielding with serigraphy as has been done successfullyfor the SPS extraction kickers. For the septa the low frequency resonant impedance must be minimizedto reduce the magnitude of beam induced fields that can cause sparking. An implementation using thedetailed knowledge gained from the ZS replacement will achieve this [50].Aperture changes which are introduced will require tapered transitions where steps in aperture cannotbe avoided. In keeping with the other tapers in the SPS, a maximum angle of 15 degrees is enforced.Further, the installation of kickers, septa and RF cavities may require additional sectorisation in the SPSbeam line. Additional sector valves should match the neighbouring apertures as well as possible andhave RF shields installed [51].
The electron beam is injected at 3 . . SPS and Transfer Lines long damping time which corresponds to 9 s at 3 . (a) (b) Figure 4.7: Transverse damping times (a) and energy loss per turn (b) versus electron energy in the SPS.At this energy, particular attention has to be given for the first turn beam threading and establishing theclosed orbit. In fact, the SPS dipoles have to operate at a very low current of around 40 A, as comparedto their flat top current which reaches 6 kA at 450 GeV. During the LEP era, special control loops had tobe included in the SPS power converters in order to guarantee the required current stability [52, 53]The SPS is a 6-fold symmetric ring, based on a FODO lattice. It was traditionally tuned to phaseadvances of π /
2, which provide an integer tune of 26 (Q26-optics) and a total arc phase advance multipleto 2 π for dispersion suppression in the straight sections. This was the working point used for injectingleptons in the SPS for LEP. The quadrupoles at the injection energy of 3 . ◦ ( i.e. π / π . This is not strictly necessary for the present application but can open the way to lowemittance ring R&D [54].Figure 4.8: Equilibrium emittance versus energy for the nominal SPS working point (red) and for lowemittance optics (blue).In particular at injection but also through the cycle (although less critical due to the increase of ra-38 SPS and Transfer Lines diation damping), it is necessary to mitigate collective instabilities. The classical "head-tail" instabilitycan be suppressed by imposing positive chromaticity, primarily using the SPS sextupoles (two families),throughout the cycle. For establishing a good chromaticity control, effects such as remnant fields in themain dipoles and sextupoles and eddy currents have to be taken into account. Regarding the TransverseMode Coupling Instability (TMCI), and based on a broad-band resonator model for the impedance andtypical SPS parameters, the threshold at injection is found at 2 × electrons / bunch, considering abunch length of 1 mm and RF voltage of 1 MV. This provides a comfortable margin with respect tothe injected bunch population. Some emittance blow-up may be observed due to intrabeam scattering atinjection, but this will heavily depend on the injected beam parameters, whose brightness can be tunedto be relatively low, in particular by injecting large transverse and longitudinal emittances. Ion instabilit-ies can be mitigated using trains with gaps and standard multi-bunch transverse feedback systems, as inmodern electron storage rings.The acceleration cycle of 0 . / sachieved for the lepton cycles in the SPS. At 16 GeV, the dynamics are dominated by synchrotron radi-ation damping, with transverse damping times of around 94 ms and energy loss per turn of 7 . / m for 10 electrons in the ring. Collective Effects.
Collective effects can limit the ultimate performance of any accelerator. In this re-spect, an analytical estimation of intensity thresholds and impedance budgets have been performed for theeSPS, studying the impact of: space charge (SC), longitudinal microwave instability (LMI), transversemode coupling instability (TMCI), ion effects, intra-beam scattering (IBS) and coherent synchrotron ra-diation (CSR) [55–68]. Two optics options (Q26 and Q40) for the SPS are investigated and the resultsare summarised in Table 4.3.Based on the analytical estimations, no major limitations are expected from SC, LMI, CSR, TMCIand IBS. On the other hand, the only critical point would be the control of the fast ion instability (FII)with rise times of 15 t rev and 6.7 t rev for the aforementioned options. These rise times can be mitigatedwith a feedback system, provided that the SPS achieves average vacuum pressures of 10 − mbarr. Theinstability rise times can be also enhanced by large clearing gaps in the bunch train structure. The SPS orbit and trajectory system will be upgraded during LS2 with newelectronics read-out over optical fibres [69]. No changes are foreseen for the pickups, consisting of204 single plane electrostatic shoebox monitors and 12 dual plane electromagnetic strip-line monitors.The strip-line monitors are directional and hence additional cabling and electronics will be necessary toobserve the counter-rotating electron bunches (with respect to proton bunches) using these 12 monitors.The nominal linac bunch intensity of 3 × electrons is at the lower operational limit of the electronics,but the orbit system should still function for a single bunch train containing the nominal 40 bunchesspaced by 5 ns. A lower number of bunches or increased bunch separation may prove problematic fororbit measurement. The 100 µ m turn-by-turn position resolution requested should be achievable whenthe SPS is filled with the nominal 3080 bunches but may be significantly worse when operating with alower number of bunches or with reduced bunch intensity. Intensity Measurement.
The SPS Direct Current Current-Transformer (DCCT) has a resolution of ∼ µ A which limits the detectable intensity change to 7 × electrons. The nominal requested resol-ution of 0.1% should, therefore, be achievable for total beam intensities above 7 × electrons. Thewall current transformer [70], used to compute bunch-by-bunch intensity in the SPS, is adapted to meas-ure LHC type beams separated by 25 ns. Its bandwidth is, therefore, intentionally limited to ∼
100 MHz.39
SPS and Transfer Lines
Table 4.3: Collective effects estimations for the eSPS.
Parameters and thresholds Q40 Q26
Phase advance (h/v) 135 ◦ / ◦ ◦ / ◦ Eq. hor. emittance @ injection. [nm.rad] 1.62 3.56Eq. hor. emittance @ extraction [nm.rad] 34.7 74Eq. Bunch length [mm] 10 14Injected hor. emittance [nm.rad] 0.43 e − per bunch 10 Bending radius [m] 70.1Average chamber radius [m] 0.04Longitudinal impedance [ Ω ] Ω /m] 9.77SC tune shift @ injection [10 − ] 4 4SC tune shift @ equilibrium [10 − ] 8 3SC tune shift @ extraction [10 − ] 1 . . [ Ω ] [ Ω ] [ Ω ]
682 1294TMCI impedance threshold @ injection [M Ω /m] 5060 6900TMCI impedance threshold @ equilibrium [M Ω /m] 506 690TMCI impedance threshold @ extraction [M Ω /m] 1121 1612Tune shift due to ions [10 − ] 4 4FII rise time [ t rev ] × charges, implying that averaging overmany turns would be required to obtain any meaningful data from this system. It should, nevertheless, beable to provide a qualitative intensity profile along the injected trains that could help with commissioningand optimisation. If this is considered not sufficient, a new monitor with an adequate electronic read-outsystem can be envisaged to provide better sensitivity. Profile Measurement.
A new technique will be required to measure the beam profile of electrons inthe SPS, as the existing instrumentation is not designed to work with such small beam sizes ( < µ min the horizontal plane and < µ m in the vertical plane). Equipping the existing wire-scanners withadditional detectors (to deal with the change of direction compared to the proton beams) may allowcomparative, qualitative measurements to be made in the horizontal plane, but would certainly not allowany measurement in the vertical plane. The existing beam gas ionisation profile monitors also do notpresently offer the required resolution to measure such small beam sizes. The synchrotron radiationemitted by the highly relativistic electron beam provides a way to measure such small beam sizes, withseveral techniques developed over the last 20 years for such purposes using x-ray optic imaging systems[71] or interferometric techniques in the visible range [72]. Recent studies using randomly distributedinterferometric targets indicate that such techniques could also be considered for use in the eSPS [73, 74].At least one such system should therefore be envisaged in the SPS ring if the electron beam transverseprofile is to be monitored. If fitted with a streak camera, such a system would also be able to measurethe longitudinal profile of the beam. The existing SPS synchrotron monitor in LSS5 is designed for a40 SPS and Transfer Lines clockwise rotating beam and a new light extraction system in a suitable location, i.e. with minimumdispersion in both planes, would need to be designed, installed and equipped in order to use synchrotronlight monitors for electrons.
Following acceleration in the SPS, the 16 GeV electron beam extraction and transfer to the experimentalarea are discussed in this section. The concept also proposes a novel slow extraction scheme to reach therequired experimental beam parameters.
An internal dump of the electron beam needs to be considered, both for machine development and formachine protection purposes. For instance, a beam instability may cause losses that necessitate to quicklydispose of the circulating beam.The existing SPS internal dump (the so called TIDVG) makes use of fast kicker magnets followed byabsorption blocks to quickly channel and dilute the circulating beam into an internal absorber, below thecirculating beam path. The rise time of the system is in the order of 1 µs, faster than the revolution periodof around 23 µs. Therefore, the absorbing material has to be downstream of the fast kicker magnets,which is no longer the case with electrons circulating counter-clockwise in the SPS [75].An internal target limiting the SPS horizontal aperture in a dispersive region can make use of thenatural energy loss of the electrons and serve as dump system. The existing TIDP collimator is alreadymeant to intercept off momentum trajectories. The 4 m long device certainly has the stopping power toeasily absorb 16 GeV electrons. The average maximum electron beam power on the dump with 2 s cycleperiod and up to 10 electrons in the ring will be around 1 . . − at injection and 0 . × − at 16 GeV. We consider thatthe beam will be completely intercepted by the collimator when drifting horizontally inwards by 1 cm.Without RF acceleration the beam will drift at this location by 1 cm within around 4000 turns or 90 msat 3 . . SPS and Transfer Lines
Resonant slow extraction allows the delivery of beam over long periods of time and using low fluence,which is particularly close to the experimental requirement discussed in Sec. 2.1. Control of the extrac-tion rate and the method used to bring particles into resonance are critical. Experimental requirementsand specificities of the electron beam at 16 GeV suggest two promising methods to drive the extractionprocess. • Using RF noise and chromatic extraction. Band-limited noise forces particles to exit the bucketon one side and start drifting in energy along a longitudinal separatrix. The advantage is that thecontrol of RF noise power controls the diffusion rate of particles out of the bucket. This can providea very low extraction rate and may reduce the effects of machine ripples on the spill structure [76]; • Using quantum excitation and amplitude extraction. Zero amplitude transverse tune is set at a dis-tance δ from the resonant tune. Amplitude detuning makes all particles beyond a certain amplituderesonant. Quantum excitation diffuses particles from the core to the resonant and unstable region.The extraction rate can be controlled by changing the distance to the resonant tune δ .Regardless of the method used for slow extraction, the feasibility of reaching the very low rates re-quired for the detector has already been demonstrated in other facilities [77].Of the two methods presented above, the latter was investigated in more detail. At 16 GeV the electronbeam dynamics in the SPS are dominated by synchrotron radiation effects with characteristic dampingtimes below 100 ms in all planes. This may prove problematic for conventional slow extraction methodsas particles can spontaneously jump on and off resonance. We take advantage of the quantum excitationto propose an extraction method that will use the existing SPS lattice while providing the capability foran arbitrarily low spill rate.Figure 4.9(a) shows the horizontal phase space during the extraction process simulated using theMADX thin tracking with synchrotron radiation. The core of the beam is stable while large amplitude (a) (b) Figure 4.9: Extraction process in the horizontal phase space at the electric septum (a). Particle trajectoriesin the transverse horizontal space in the extraction region with apertures and the injected14 GeV / c proton beam envelope in grey (b).particles are trapped on the third order resonance, as the triangle shape of the beam shows. A new electricseptum placed at a specific distance from the beam core channels particles reaching an excursion of more42 SPS and Transfer Lines than 45 mm towards the extraction channel. This simulation allowed a first estimation of the parametersof the extracted beam shown in red on Fig. 4.9(a) and in particular a geometrical emittance of 10 − m rad. Extraction Rate Control.
The main experimental program considered here requires a very low buthighly constant flux of particles, making the control of the extraction rate critical. The scheme presentedabove addresses the control of the spill rate but machine and in particular power supplies stability iscritical to ensure constant spill rate.Regulation of the SPS dipole magnets at low field relied on a switch of the gain settings of the AnalogControl board during the LEP era. This system has since been removed but the current regulation is nowimplemented in a virtual FGC within the SPS Mugef Control system. The Current Loop is now fullydigital hence the resolution problem does no longer exist and stability at low field should not be an issue.The quadrupole converter hardware was modified during LS2 in order to improve performances for theproton mode. The stability of these converters at very low currents is no longer obvious, hence requestsfor machine developments time in the SPS at low energy shall be booked in 2021.In view of the LEP era, all other SPS auxiliary power converters were modified to allow operation atlow energy. However, during the SPS consolidation campaign at the end of the 1990’s, these converterswere modified or replaced: operation of the SPS at low energy is not possible anymore today. Consider-ing the ageing state of these converters and the ongoing consolidation program that will last until LS4,the requirements for the operation at low energy should considered within the framework of this programand a budget should be granted accordingly.
Extraction Trajectories.
Slowly extracted particle trajectories were investigated in LSS1 of the SPStowards the Meyrin site. LSS1 is used for injection of the proton beam coming from the Proton Synchro-tron (PS). The scheme shown in Fig. 4.9(b) requires two new elements. • An electric septum of length 2 . / m to provide a deflection of the extractedparticles of around 780 µrad; • A thin magnetic septum with a blade of 10 mm and a field of 150 mT will provide an additionaldeviation of 4 mrad, to reach the existing MSI septa and enter the TT10 line.Further synergies with possible upgrades of the proton injection scheme are also considered.
A fast extraction scheme was not studied in detail here. However, the slow extraction orbit bump canbe combined with a non-local fast kicker to achieve complete extraction of the beam within a singleturn. A preliminary study indicates that non-local extraction using an MKE kicker in LSS4 would becapable of quickly extracting the beam in the TT10 channel with a Q53, or double tune, SPS optics.Similar schemes have already been studied for high energy protons [78]. A fast extraction scheme forelectrons will be feasible, should not present any specific technical challenge and will come at little tono additional costs.
The TT10 line is used to transport protons from the PS to the SPS. The transport in the opposite directionof the extracted electron beam comes naturally. The polarity of magnetic elements will remain the sameand the required strengths will be similar to proton operation, as the extracted beam rigidity is in thesame range as the injected beam coming from the PS.Of particular interest is the powering scheme and optics of this line. The first part of the line is poweredby only two power supplies in a regular FODO scheme. The FODO lattice was designed to match the43
SPS and Transfer Lines injected proton beam into the SPS lattice. In the case of slow extraction the beam parameters are verydifferent from the ring ones. Therefore, the extracted beam is not matched to the FODO lattice of theTT10 line. However, transporting the beam in this mismatched transfer line is still possible due mainlyto its small emittance.Figure 4.10 shows the evolution of the 4 σ beam envelope with the beam parameters estimated earlier,a momentum spread of 10 − and a maximum trajectory offset of 2 mm. It is clear from the evolution ofFigure 4.10: Beam sizes and optical functions from the SPS to the experimental target.the beta functions in TT10 that the beam is not correctly matched. However, due to the particularly lowgeometrical emittance of the extracted beam, the beam size along the line remains reasonable and wellwithin the 108 mm aperture of the quadrupoles.Starting s =
650 m the powering scheme is composed of two independent FODO cells followed by5 independent quadrupoles. In the conceptual design presented here, this section is used to match theincoming beam to the new line. The dispersion created by both horizontal and vertical bends in TT10 isalso maintained below around 10 m.Figure 4.12(b) shows the switchyard region at the end of the TT10 line. This region is already filledby magnetic elements due to the split of the TT2 line coming from the CPS into the TT10 and n_ToFlines. The design proposed here aims at maintaining the proton trajectories while allowing the electronbeam to branch off the existing line towards a new line. Many factors where included in the design andintegration of all the elements and ancillary systems was decisive (see Sec. 5.4.2).Currently two MAL-type dipoles bend the proton beam towards the TT10 line. In Fig. 4.11 thosedipoles are replaced by three HB2 C-type dipoles. The deviation of the proton beam coming from thePS is around 50 mrad for the first two dipoles and around 17 mrad for the last one. This peculiar choiceof strengths preserves the proton trajectories outside of these three dipoles, in TT2 and TT10. Thequadrupole immediately before the wall on the TT10 side and its downstream beam monitor will have tobe shifted downstream by 30 cm (see Sec. 5.4.2), with very small effect on the optics. Two new powersupplies will have to be wired to those three dipoles. This new layout for the transport of the proton beamwill be able to transport protons up to 26 GeV but further studies on the HB2 magnets at higher fieldscould allow operation at higher rigidity.Electrons coming from TT10 will experience a deviation of 50 mrad in the first BH2 dipole. The beam44
SPS and Transfer Lines
Figure 4.11: Layout of the junction region between the TT2, TT10 and n_ToF lines with the proposedelectron trajectory in cyan.will travel through the opening of the following two dipoles before crossing the n_ToF line. The beampipe and vacuum chamber in this region will become particularly complex, but the careful integrationstudy ensured that this concept is within the capabilities of CERN experts. This solution is only oneamong several explored. A complete bypass of this area with a new beamline above those magnets couldbe studied in case a technical design of the concept presented here identifies show stoppers. Howeverthe concept discussed here was developed in close coordination with 3D integration and other groups toensure its feasibility (see Sec. 5.4.2 and 5.4.4).The new beam line starts with a single MCW C-type dipole as close to the region of crossing withthe n_ToF line as possible. The following drift space up to the wall allows for passage of personneland equipment, if the beam pipe is removed. There follows a penetration through the existing tunnelwall towards the new tunnel. The concept presented here for the new tunnel was directed by severalconstraints: • The two experimental areas cannot overlap with existing structures; • The two experimental areas should be capable of receiving beam successively; • The beamline should make use of existing and available magnets due to the high cost of newmagnets manufacturing; • The design should be capable of transporting 16 GeV electrons with possible upgrade to at least18 GeV; • The beam delivered to the primary experimental area should be manipulated to reach large trans-verse sizes of up to a few tens of centimetres.The bending section of the new line is built around a simple FODO arrangement, shown in Fig. 4.12(a).It is designed around two MCW dipoles of 3 . ◦ and alignthe beam trajectory with the experimental area. The beam is directed towards the secondary experimental45 SPS and Transfer Lines (a) (b)
Figure 4.12: Schematic view of the base cell used in the new line with important dimensions in mm orrad (a) and layout of the new line in the junction region (b), with electron trajectory in cyan.area by switching the powering of the last two bending cells to the secondary line, around 5 m besidesthe primary one. This is possible due to the opening on the C-type dipoles used for the bending cells.The very large total angle of the beam trajectory from the SPS extraction is a concern since the highenergy electrons may radiate a non- negligible fraction of their energy. From the extraction of the SPSto the end of the TT10 line the cumulative angle is quite small and the average relative energy lost byelectrons at 16 GeV is in the order of 10 − . The large angle of 51 ◦ between the TT10 line and theexperimental area (see Fig. 4.12(b)) causes a relative average loss of energy of 1 . × − for electronsof at 16 GeV. The change of the average beam energy along the new line can be compensated by splittingthe powering of the cells in different circuits. A technical design will also need to account for the increasein momentum spread associated with stochastic energy loss.The beam delivery system to a missing momentum experiment must be capable of provide a particu-larly large beam size. A 12 m straight section after the last dipole is dedicated to the manipulation of thebeam size. The design studied here is very simple and limited to the use of quadrupoles that are avail-able and stored at CERN. The synoptic of this Final Defocus scheme is shown in Fig. 4.10. The beamsizes achieved here range from 10 × . to 30 × . . Figure 4.10 presents the optics leading to abeam size on the experimental target of 20 × . . This concept demonstrates an approximate rangeof beam sizes that could be provided. However, a more detailed design can be established together withprecise experimental requirements with an experiment design optimised for large beam sizes.The design makes use of existing and available magnets to minimise the cost. Table 4.4 lists thosemagnets with some of their most important characteristics.46 SPS and Transfer Lines
Table 4.4: List of existing magnets used in the junction area and the new line.
Name Type Quantity Max. current [A] Max. integrated strength
QTN Water-cooled quadrupole 9 150 2 .
052 TQTS Water-cooled quadrupole 2 416 37 TQ200 Water-cooled quadrupole 1 750 22 .
05 TMCW Water-cooled dipole 15 1000 4 .
63 T mMCB Water-cooled dipole 3 880 4 .
44 T mIn this section we discussed how to produce and deliver a low intensity beam to a missing momentumexperiment. The intensity considered will be around 10 to 10 electrons on target per second, consid-erably below the maximum number of electrons in the SPS ring, in the order of 10 . A dump-typeexperiment making use of that beam could be housed at the end of the secondary new line and beside themissing momentum experiment, as shown in Fig. 4.12(b). The instrumentation for the SPS extraction system and TT10 transfer line to the experimental area willneed to be able to deal with two modes of operation. The first mode is based on a fast extraction forsetting-up the line, where the full beam intensity is extracted in one SPS turn. The second one is a slowextraction performed while running the experiment, where only a fraction of electron beam circulatingin the SPS is extracted over 10 seconds giving, on average, a few electrons per 5 ns period.For the fast extraction, the 10 existing OTR-BTV systems can be used to monitor the beam, providedsome modifications are made to their light extraction systems. The 10 existing strip-line beam positionmonitors in TT10 would also be capable of giving the trajectory of such beams, but would require ad-ditional cabling and electronic systems to be installed due to the directionality of these monitors. Inaddition, some 6 additional BPMs will need to be installed in the new transfer line that will send thebeam to the experiment hall.The case of the slow extraction is much more challenging for beam instrumentation as it would requireextremely high sensitivity to be able to detect the very few particles sent to the experiment. A new typeof monitor will need to be developed to cope with the extremely low beam fluence. One candidate tomeasure the position of the extracted beam is to detect the Cherenkov light emitted by the electronsas they propagate close to the surface of a long dielectric [79]. This method, recently pioneered in theframework of the CLIC study on the Cornell Electron Synchrotron Ring (CESR) [80], would neverthelessrequire an exhaustive R&D phase to assess the sensitivity limit of such monitors. One candidate tomeasure transverse beam profile is an ultra-high vacuum (UHV) version of the scintillating fibre monitordeveloped for the CERN Neutrino Platform, capable of detecting particle rates as low as 100 Hz [81].Significant development work will nevertheless need to be carried out to adapt the existing monitor andmake it UHV compatible, probably requiring replacement of the scintillating fibres by Cherenkov fibres.If successful, one can assume that 16 Cherenkov beam position monitors and 10 optical fibre profilemonitors would then be built to monitor the extracted low charge beam while running the experiment.47
Infrastructure and Civil Engineering
Civil engineering (CE) and infrastructure generally represents a significant proportion of the total budgetfor projects at CERN. On that basis, a significant amount of attention has been given to the CE andother infrastructure required to implement eSPS. This chapter will discuss the infrastructure and civilengineering considerations for the project, the optioneering process undertaken and the key factors inchoosing the selected design options. Figure 5.1 below shows the main elements of the scheme. Thescope of the study in this area includes the infrastructure and civil engineering required for the following: • Beam injection in TT5,TT4 (shown in red); • Transfer of beam via TT61, TCC6 and TT60 into the SPS (shown in green); • Extraction of the beam via TT2, TTL2 and a new extraction tunnel (shown in dark yellow); • Construction of a new experimental hall (shown in light yellow).The section breaks the study down into general considerations (applicable to some or all parts of thescheme) and the specific considerations divided into the areas noted above.Figure 5.1: Overview of main elements of project.The scope of this chapter also specifically excludes any changes required as part of the followingwhich will be covered elsewhere in the CDR: • Acceleration of the beam in the SPS (shown in blue); • Extraction of the beam in TT10 (shown in purple).48
Infrastructure and Civil Engineering
Figure 5.2: Location of the planned facilities at the CERN Meyrin site showing Franco-Swiss border.
Location.
The eSPS scheme would span the CERN site, housed within a mixture of new and existinginfrastructure as depicted in Fig. 5.1. The areas within the CE scope of this document are located entirelywithin CERN land in France as shown in Fig. 5.2.
Geology.
The location is well suited from a CE point of view since the ground conditions are relat-ively stable and well understood with detailed geological records in and around the area, reducing thegeotechnical risk. New construction would generally be founded within the Molasse rock, which is veryclose to the surface in the relevant areas. The Molasse is composed of an alternating sequence of marlsand sandstones (and formations of intermediate compositions) and is generally considered good rock fortunnelling. This rock is overlain by Quaternary glacial moraines related to the Wurmien and Rissienglaciations and in many places on the Meyrin site by made ground.
Civil Engineering Layout.
The layout required for each part of the project and the CE works neededto achieve it are detailed in the appropriate sections of this chapter. The main elements of CE could besummarised as follows: • Refurbishment, minor structural modifications and provision of shielding block arrangement inB183, TT5 and TT4; • Monitoring and possible maintenance of TT61 to allow re-installation of a transfer beam line; • Construction of a new 55 m long, 3.57 m wide (varying to 9.09 m) extraction tunnel;49
Infrastructure and Civil Engineering • Construction of a 20 m wide by 40 m long experimental hall and detector pit as well as associatedinfrastructure.
Integration studies have been performed to evaluate the feasibility of siting the eSPS facility withinCERN’s Meyrin site. The infrastructure requirements were defined and integrated within the civil en-gineering layout. The overall layout was optimised in terms of radiation protection, general safety,accessibility, and practicality. Figure 5.1 shows the layout of the overall facility.The integration of the facility is divided into the following four areas: • Beam injection: Building 183, Transfer tunnels TT5 and TT4: this houses the CLEAR injector,the linac beam line and services; • SPS transfer: Transfer tunnel TT61 and junction cavern TCC6: this houses the SPS transfer beamline; • SPS extraction: Transfer tunnels TT2, TTL2 and the new extraction tunnel: this houses the SPSextraction beam line; • Experimental area: Surface and underground experimental halls: to house and install the experi-ment and to provide the services required for the operation of the detector.The following sections describe in detail the integration studies performed for each of the four areaslisted above. Table 5.1 contains the SmarTeam numbers of the integration models (ENOVIA SmarTeamis a product data management tool that enables organisations to manage and collaborate on componentinformation). Table 5.1: SmarTeam references of the integration 3D models.
Description SmarTeam number
Injection ST1153166SPS transfer ST1283509SPS extraction ST1222240Experimental area ST1222240
The CERN electrical network is composed of a transmission and a distribution level. The transmissionlevel transmits the power from the existing source of the European Grid to the different CERN locationsincluding the Meyrin campus and the SPS complex where the eSPS is planned to be constructed. Thetransmission network operates at high voltage levels of 400 kV, 66 kV and 18 kV. The distribution leveldistributes the power from the transmission level to the end users at medium and low voltage levelscomprised of between 400 V and 18 kV. A schematic view of the CERN transmission and distributionnetwork is shown in Fig. 5.3. The yellow ellipse in Fig. 5.3 shows the approximate geographical locationof the eSPS according to the existing CERN electrical infrastructure.50
Infrastructure and Civil Engineering
Figure 5.3: Schematic view of the CERN transmission and distribution network. Approximate locationof eSPS according to the existing CERN electrical infrastructure is shown in yellow.The concept for the design of the eSPS electrical network is driven by four factors:1. The estimated electrical power requirements;2. The status existing electrical network infrastructure;3. The location and type of equipment to be supplied;4. The electrical load class and load types. eSPS power requirements.
Table 5.2 summarises the electrical power loads for each of the threeeSPS electrical areas. Power loads include values received as of April 2020 by concerned stakeholders.51
Infrastructure and Civil Engineering
In the case of unknown values such as for the new experimental area, an estimate has been derived fromequivalent and comparable CERN accelerators and utilities infrastructures. A total electrical power needof 4800 kW is considered at conceptual level. The exact values of the electrical power requirements willhave to be confirmed during the technical design phase.Table 5.2: eSPS power requirements shown by electrical area.
Electrical area Nominal power (kW)
Linac 3600SPS including injection and extraction 400Experimental area 800
Electrical network infrastructure status and requirements.
From an electrical supply point ofview the eSPS accelerator is in areas where several electrical substations are available, operational, andequipped to cover the additional eSPS demand in terms of power and energy consumption. Three elec-trical zones, which correspond to the accelerator description in this chapter are defined for the eSPS. Thethree zones can be considered independent from an electrical point of view and are : The linac (B183,TT4 and TT5), the SPS ring including the injection and extraction of the SPS (TT61, TT10, TT2 and thenew extraction tunnel) and the new experimental area.Figure 5.4: U0-A01 area hosting UPS infrastructure for the n_TOF experiment.
Linac -
The linac is located nearby Building 112 where a primary electrical substation, hereaftercalled ME59, is located. Considering the expected power requirements given in Table 5.2, the powerrequired for the linac can be supplied form the substation ME59. The existing electrical infrastructureof the buildings where the linac will be installed dates from the seventies. A full refurbishment of thegeneral services is necessary to replace obsolescent electrical equipment, comply with applicable safetyrules and engineering standards. In TT4 the construction of the shielding wall and the installation of the52
Infrastructure and Civil Engineering beam dump will require the displacement of the UPS electrical installation in the area which supplies,among other users, the n_TOF experiment. The concerned area U0-A01A is shown in Fig. 5.4. A newsuitable location will have to be identified.
The SPS ring including the injection and extraction -
The electrical infrastructure of the SPScomplex has recently been renovated and sufficient power is available for the needs of the eSPS includingthe needs for the new extraction tunnel. The general services network of the existing SPS tunnels arein a good state. Changes or modifications might be necessary according to eSPS requirements. Theadditional loads for the injection and extraction of the SPS will be supplied from existing substationslocated in points SPS-6 and SPS-7. The additional loads for the new extraction tunnel can be suppliedfrom either the SPS infrastructure or from the new experimental area.
The new experimental area -
The new experimental area will be constructed on the Meyrin campusas shown in Fig. 5.2. The new experimental area will be supplied from the electrical substation ME59which is also the source for the linac zone.
Electrical load class and load types
The electrical supply concept for eSPS is based on the require-ment to keep essential parts of the accelerator infrastructure operational if the normal power source fails.Emphasis is put on loads related to personnel and machine safety during degraded situations. The variousload classes and types can be characterised as shown in Table 5.3. The main ranking parameters are theacceptable duration of the power interruption and whether the load is part of a personnel or acceleratorsafety system. Table 5.3: Load classes and main characteristics.
Load class Load type (non-exhaustive list) Maximum duration ofpower unavailabilityMachine
Power converters, cooling and ventilationmotors, radio frequency, klystrons Until return of main supply
General Services
Lighting, pumps, vacuum, wall plugs Until return of main or sec-ondary supply
Secured Personnel safety:
Lighting, pumps, wallplugs, lifts 10 – 30 seconds
Uninterruptible Personnel safety: evacuation and anti-paniclighting, fire-fighting system, oxygen defi-ciency, evacuation Interruptions not allowed,continuous service mandat-ory
Machine safety: sensitive processing andmonitoring, beam loss, beam monitoring,machine protectionMachine loads do not have a second source of supply, therefore, in case of an upstream electricalpower cut the equipment will be cut. The general service loads typically accept power cuts betweenseveral minutes and several hours, sufficiently long to commute to the second source or to wait untilthe main source is restored. Both the machine and general services loads do not include personnel ormachine safety equipment or systems. Secured loads include personnel and machine safety equipmentor systems that can sustain short power cuts up to a duration of 30 seconds. In a degraded situation orupstream electrical power cut a level backup is provided by the diesel power station, which typicallystarts up within 10 seconds. Uninterruptible loads include personnel and machine safety equipment53
Infrastructure and Civil Engineering or systems that require continuous and stable power supply. Wherever necessary, the uninterruptiblenetwork is created locally by installing an uninterruptible power supply (UPS) powered from the securednetwork. Figure 5.5: eSPS simplified electrical supply schematic of the distribution network.
Distribution network proposal.
Considering the information gathered at the time of writing on theeSPS layout and electrical requirements a simplified conceptual electrical supply scheme of the distri-bution network is proposed and shown in Fig. 5.5 from [82]. The scheme describes how the four typesof load classes are made available at the level of medium voltage and low voltage to the three eSPSelectrical areas. For simplicity of the scheme, the 400 kV and 66 kV transmission network are not shownbut are available in the drawing GENEM0033 [82].The part of the diagram shown in black correspondsto the existing infrastructure while the part in blue represents the new infrastructure to be installed. Theproposed scheme allows margin for optimisation as soon as the technical details become available duringthe technical design phase.
The Cooling and Ventilation section is composed of three parts: the first presents the piped utilities, thesecond the air-related systems and the third addresses safety and environmental protection aspects.The design herewith presented is based on the user requirements available in February 2020. A safetymargin of approximately 20% was applied to the heat loads, given the early stage of the project. Ad-ditionally, the infrastructure was designed to be flexible and allow for future upgrades and extensions.Energy efficiency is considered a priority and different strategies are adopted to enhance it.
Piped Utilities.
The piped utilities are mainly dedicated to the cooling of the accelerator equipmentand the related infrastructure such as power converters, electronic racks, cables, etc., as well as to theventilation and air-conditioning of underground and surface premises. In addition, specific systems areforeseen to cover other general needs such as fire extinguishing plant, drainage and sumps (for bothsurface and below-ground areas), drinking water and compressed air for ancillary equipment (vacuumpumps, valves, dampers, etc.). The main circuit typologies are: • Demineralised and industrial water: for the cooling of accelerator equipment (excluding: magnets,power converters, etc.) and infrastructure if needed;54
Infrastructure and Civil Engineering • Chilled water: at present, used exclusively in ventilation and air-conditioning plant; • Firefighting systems; • Wastewater: water from underground and surface premises that is to be rejected; • Drinking water: for sanitary purposes and make up of industrial water circuits; • Compressed air: mainly for ancillary equipment such as pneumatic valves or vacuum pumps.Two different typologies of water cooling circuits are defined according to their equipment and work-ing temperatures: demineralised water circuits and chilled water ones. The first are cooled by primarycircuits (industrial water) and the latter by chillers, that are able to produce water at lower temperatures.Figure 5.6 illustrates a simplified layout for the two types of circuits, although they can assume differentconfigurations, depending on the requirements.The primary water circuits, cooled by open wet cooling towers, are configured in closed loops tominimise water consumption. Drinking water make-up is used to compensate for evaporation, leaksand blowdown. A continuous water treatment against Legionella, scaling and proliferation of algae isforeseen.Generally, demineralised water circuits are used to refrigerate equipment and have a maximum con-ductivity of 0.5 µ S/cm. A set of demineralisation cartridges is foreseen for the cooling circuits, in orderto locally control the conductivity. The temperature of the water leaving the cooling station is set to 27 ◦ C( ± . ◦ C).Chilled water is used to condition air in air-handling units. Where dehumidification is required, theleaving water temperature is approximately 6 ◦ C, otherwise it is usually raised to 14 ◦ C.Figure 5.6: General diagram for cooling circuits.Most electromechanical cooling components, such as cooling towers, pumps and chillers are foreseenwith N+1 redundancy. However, no redundancy is foreseen for electrical cubicles, control cubicles andheat exchangers. Under present conditions, there is no need to provide a secure power supply for themajority of the cooling plant. In case of power failure, all accelerator related equipment would stop andwould not require any cooling.The drinking water network is used for sanitary purposes and as make-up water for the cooling towers.The existing drinking water network at CERN is extended and upgraded to satisfy the requirements inthe new premises. 55
Infrastructure and Civil Engineering
The existing firefighting water is upgraded to cover areas that require such systems. For the under-ground infrastructure, such as TT4 and TT5, rigid pipelines are installed. They are kept dry to avoidcorrosion and water stagnation. In case of fire, the fire brigade opens manual valves at the surface tosupply water to the concerned area.The reject water is split in two distinct networks - one for clear water and one for sewage; thesenetworks are connected to the corresponding existing systems in each area. For safety reasons, each sumpis equipped with sensors providing alarms for high and too high level. Thorough control mechanismsare implemented before releasing reject water to existing natural water lines. Relevant parameters suchas pH and temperature are monitored.The user requirements for compressed air are not defined at this stage of the project. However, existingnetworks are present in CERN premises. If their capacity is not sufficient, or if they cannot be used (forany other reason), new production stations will be installed. An N+1 redundancy is foreseen for thecompressors.
Heating, Ventilation and Air-Conditioning.
The heating, ventilation and air conditioning (HVAC)plants are designed to provide fresh filtered air to people and to purge (where required) contaminants andpollutants that might be harmful, allowing for safe access. For certain facilities, the exhaust air is filteredbefore being rejected to the atmosphere. Moreover, HVAC plants provide the desired indoor temperatureand/or humidity in underground and surface premises. Finally, these plants provide smoke extraction inplaces where it is required, such as for TT4 and TT5.
Safety and Environmental Protection.
The safety of people and environmental protection are prior-ities in the early design herewith described. Collaboration with CERN’s Occupational Health & Safetyand Environmental Protection (HSE) Unit is established to ensure the satisfaction of the regulations andapplication of best practice in these domains.Two safety aspects were considered in the context of the cooling and ventilation infrastructure - fireprotection and radiation protection. For the first, firefighting water networks are put in place, as de-scribed in the Piped Utilities section. Additionally, hot smoke extraction systems are foreseen for certainselected areas, as mentioned in the Heating, Ventilation and Air-Conditioning section. Other safety re-lated mechanisms like pressurisation of safe zones, particularly in the new detector region, are likelyto be required. However, this will be defined in the future, during a more advanced stage of the study.For radiation protection, pressure cascades are evaluated, as well as requirements concerning differentoperational modes, such as purge and access.Environmental protection is addressed from two different perspectives - mitigation of impacts relatedto waste fluids and energy/water efficiency of cooling and ventilation plant. Strategies are put in place todesign systems that treat and prepare fluids that are to be rejected. For instance, filters are foreseen at theair outlets of certain premises and water treatment is put in place for the cooling towers blowdown.Water conservation is considered central in the design of the cooling and ventilation infrastructure.For instance, a new system to treat and reuse cooling tower waste water is currently being developed atCERN and is foreseen to be implemented. This allows for large savings in drinking water used as make-up. Energy efficient measures are taken to minimise electricity consumption. For instance, air is recycledas much as possible, avoiding the energy costs associated with treating fresh air. For TT4 and TT5, thisrequires long ducts; however, the additional fan power is largely offset by the savings in ventilation loads.A detailed study to optimise energy efficiency is required for future more detailed stages of the presentstudy. 56
Infrastructure and Civil Engineering
To mitigate risks associated with ionising radiation, CERN radiation protection rules and proceduresmust be applied [83]. Risks resulting from ionising radiation must be analysed to develop mitigationapproaches. Design constraints ensure that the doses received by personnel working next to the beaminstallations as well as the public will remain below regulatory limits under all operational conditions. Areliable radiation monitoring system coupled with an effective early warning and emergency stop systemsare important parts of the radiation protection infrastructure.Radiation protection is concerned with two aspects: protection of personnel operating and maintainingthe installations and the potential radiological environmental impact. The facilities were optimised basedon general radiation protection guidelines and specific studies on prompt and residual dose rates and airactivation. To assess the radiation protection aspects, extensive simulations were performed with theFLUKA Monte Carlo radiation transport code [84–86].The radiological hazards mainly arise from the accelerated electron beam with energies of up to16 GeV and a beam power of up to 1 .
28 kW and the prompt stray radiation generated by beam losses.This hazard represents the major challenge for the shielding design given the construction constraintsin the existing infrastructure. The basic parameters used for the radiation protection study are listedin Table 5.4. They provide sufficient margin compared to the projected range of parameters as shownin Table 3.1, except for the high charge mode, for which the repetition rate would need to be loweredaccordingly.By the nature of electron beams, which have a considerably lower activation potential compared toproton beams of similar power, the activation of materials, liquids and air represent a lesser concern.Radiation from induced activity will be limited by actual admissible beam losses driven by the shieldingin place. The radiological environmental impact of the eSPS installations is negligible given the proposeddesign and beam parameters.The operation of klystrons to produce high-power RF and generating parasitic X-rays is a furthersource of radiation. Standard prescriptive methods to mitigate these risks are well established and allowan efficient control.A more detailed radiation protection study was conducted for the linac part to demonstrate feasibilityregarding the envisaged beam parameters and operational and shielding constraints. The injection, ac-celeration and extraction from the SPS have no significant radiological impact. The design of the newexperimental hall and cavern have been covered in less detail but no limiting constraints have been iden-tified for their foreseen implementation. More details on the radiological study presented here can befound in [87].All generally accessible areas inside buildings which are dedicated to the exclusive operation of thetherein installed accelerators shall not be classified higher than Supervised Radiation Area during nom-inal beam operation. The areas outside, which are generally classified as non-designated public domainareas shall be protected such to comply with the applicable limits. The design target values were chosena factor 3 lower than the actual limits to include a margin for uncertainties in the models, calculationsand imperfections in the construction of the infrastructure. The corresponding dose limits and designvalues are shown in Table 5.5. 57
Infrastructure and Civil Engineering
Table 5.4: Operational scenarios and beam parameters used for the radiation protection study. Thesescenarios basically cover the intended Linac beam modes as shown in Table 3.1.
S-Band Linac X-Band Linacmulti bunch single bunch multi bunch single bunch
Max. energy [GeV] 0 .
250 0 .
250 3 .
55 3 . − / s] 1 . × . × . × . × Avg. beam current [nA] 3000 500 200 200Avg. beam power [W] 750 125 710 730Repetition rate [Hz] 10 100 100 100Pulse charge [nC] 300 5 2 2Pulse charge [e − ] 1 . × . × . × . × Bunch charge [nC] 2 5 0 .
05 2Bunch charge [e − ] 1 . × . × . × . × Bunches per pulse 150 1 40 1dE per module [MeV] 137 . . . . .
65 2 . SPS Missing Momentum Experiment
Max. energy [GeV] 16 16Avg. beam intensity [e − / s] 5 . × . × Avg. beam power [W] 1280 8 . Area classification Permanent stay Low occupancy Optimisation Design targetthreshold
Non-designated 0 . − . −
100 µSv a − .
05 µSv h − Supervised 3 µSv h −
15 µSv h − - 1 µSv h − Simple 10 µSv h −
50 µSv h − - 3 µSv h − Limited Stay - < − - -High Radiation - <
100 mSv h − - -Prohibited - >
100 mSv h − - -Reference group 0 . −
10 µSv a − < . − Many standard industrial hazards will need to be considered for working in theeSPS, such as noise, lighting, air quality, and working in confined spaces, which can be satisfied byfollowing existing CERN safety practices and the Host State’s regulations for workplaces.The ambient temperature in the tunnels will be of particular interest for access and work; following aguidance document provided by the French INRS [89] a limit of 28 ◦ C is recommended for manual workwithin the tunnels (across the full range of relative humidity values) during extended access periods. At58
Infrastructure and Civil Engineering temperatures above this where manual work is required, care must be taken to monitor the condition ofworkers, ensuring reduced working time and increased rest and hydration in proportion to any increasedheat induced stress.
Fire Safety.
All buildings, experimental facilities, equipment and experiments installed at CERN shallcomply with CERN Safety Code E and other fire safety related instructions and notes listed at [90]. Inview of the special nature of the use of certain areas, in particular underground, and the associated firerisks, CERN’s HSE unit is to be considered the authority for approving and stipulating special provisions.As the project moves to the technical design report stage, finalising layouts and interconnecting vent-ilation systems, detailed fire risk assessments will have to be made for the eSPS complex. At this stage,general fire safety considerations have been outlined, and implemented in the civil engineering, integra-tion and ventilation designs, based upon the latest fire safety strategies employed at CERN. As such, thecomplex has been considered to apply an extension of the fire safety concept being implemented in theSPS during CERN’s LS2 period and beyond.The most efficient protection strategy is one that uses complementary “safety barriers”, with a bottom-up structure, to limit fires at the earliest stages with the lowest consequences, thus considerably limitingthe probability and impact of the largest events.In order to ensure that large adverse events are possible only in very unlikely cases of failure of manybarriers, measures at every possible level of functional design need to be implemented: • In the conception of every piece of equipment (e.g. materials used in electrical components, circuitbreakers, etc.); • In the grouping of equipment in racks or boxes (e.g. generous cooling of racks, use of fire-retardantcables, and fire detection with power cut-off within each rack, etc.); • In the creation and organisation of internal rooms (e.g. fire detection, power cut-off and fire sup-pression inside a room with equipment); • In the definition of fire compartments; • In the definition of firefighting measures.The key fire safety strategy concepts are set out below.
Access and Egress - • Ensure that the evacuation distance and path width is within the acceptable range for undergroundfacilities; • Ensure the evacuation routes are practicable; • Ensure two ways in and out for the fire service intervention.
Reasonable Fuel Load and Fire Ignition Sources - • Reduce additional fuel load, beyond that of the oil in the klystrons, especially due to storage ofmaterial or equipment; • Minimise the number of possible fire ignition sources, avoiding workshops within the same firecompartment. 59
Infrastructure and Civil Engineering
Compartmentalisation -
Compartmentalisation impedes the propagation of fire and potentiallyactivated smoke through a facility, allowing occupants to escape to a comparatively safe area much morequickly than otherwise, as well as facilitating the effective fighting of the fire, and evacuation of victimsby the CERN Fire and Rescue Service. The following requirements have been set: • All ventilation doors must be fire doors EI90; • Isolate communicating galleries with fire doors EI90; • Isolate neighbouring surface facilities with fire doors EI120; • Normally opened fire doors to be equipped with remote action release mechanism, monitoringposition and self-action thermal fuse.
Fire Detection, Fire Alarm, Safety Action Integration -
An early fire detection system, integratedinto the safety action system is a crucial component of the SPS fire safety WP2 technical solution,which shall also be implemented in the eSPS. Early detection is such that it allows evacuation (lastoccupant out) before untenable conditions are reached. The system must be capable of transmitting analarm, triggered upon fire detection, action on evacuation push buttons, CERN Fire Brigade action out ofCERN FB SCR/CCC or BIW (Beam Imminent Warning) situations. Evacuation push buttons shall coverall premises. The fire detection and evacuation push buttons must also be integrated with safety actionssuch as compartmentalisation, ventilation stop and other machine functions according to a predefinedfire protection logic. The existing fire safety system shall be dismantled. The key features of the systemare: • Fire detection and fire alarms throughout facility: smoke detection by air sampling, and manualcall points; • Standard ventilation interlocked with fire detection: command & control of fire dampers, fire doorsand ventilation stops; • Voice alarm system : broadcast of audible signals with loudspeakers (evacuation signal and beamimminent Warning signal); • Triggering of an evacuation alarm in the fire compartment of origin and the adjacent fire compart-ments; • Triggering of a Level 3 alarm in the CSAM system, which alerts the CERN Fire Brigade controlroom, and results in crews being dispatched immediately.The eSPS fire safety system will be connected to the SPS fire safety system by including the controlcabinets in the optical fibre loops of the SPS system. Space allocation shall be made in an accessiblearea (not subject to radiation) in the TT5 Hall, for the evacuation and fire detection control panels.The eSPS system shall have the same functionalities as the SPS system: manual command of safetyfunctions, broadcast of instructions with safety microphones for the CERN Fire and Rescue Service use,and microphones for SPS OP use.
Smoke Extraction -
Smoke extraction facilitates both the safe evacuation of occupants, and theeffective intervention of the CERN Fire and Rescue Service to locate victims and prevent the furtherspread of a fire. A buildup of smoke can also result in lasting damage to the sensitive and valuableequipment present, an effect that can be limited through early extraction. This will be addressed in detailfor each area of the eSPS complex. 60
Infrastructure and Civil Engineering
Fire Suppression -
The CERN Fire Brigade need adequate means of fighting a fire on arrival. Thiswill be addressed in detail for each area of the eSPS complex.
Electrical Safety.
The electrical hazards present for this project are considered standard for such aninstallation. There will be a number of high voltage systems, including the klystrons, modulators andmagnets. The supply infrastructure represents additional hazards frequently found at CERN, such asuninterruptible power supplies, transformers and power converters; these shall be mitigated throughsound design practice and execution. The CERN Electrical Safety Rules, alongside NF C 18-510, shallbe followed throughout the design process; where exceptions are required, this shall be subject to anappropriate level of risk assessment to evaluate the residual risk, and determine the mitigation strategiesrequired. NF C 18-510 compliant covers, interlocks preventing access to high voltage equipment, andrestriction of access to the electrical rooms to those with the appropriate level of CERN electrical trainingand certification shall be used to protect personnel from any electrical hazards present.For all electromagnets, appropriate grounding measures shall be implemented for the magnet yokes,and all live parts protected to a minimum of IP2X for low voltage and IP3X for high voltage circuitsor locked-out for any intervention in their vicinity. Interventions may only be carried out by person-nel with the necessary training and certification, after following the work organisation procedures andauthorisation of the facility coordinator (VICs, IMPACT etc.).
Electromagnetic Safety.
Radiofrequency (RF) components purchased from industry are required tobe CE marked and to comply with the EU emission norms for industrial environments. RF equipmentbuilt or installed at CERN shall be “leak-tight”, and tested against EU industrial norms for RF emissionsin situ. In case of a RF-related accident (for example the breaking of a waveguide), the mismatch inreflected power shall be detected, and trigger a cut in the electrical supply.The magnetic fields from the proposed electromagnets in the current eSPS design represent a hazardsimilar to that found in many of the facilities at CERN, and shall be handled with standard mitigationstrategies. The facility shall follow the Directive 2013/35/EU on the occupational exposure of workers,alongside CERN Safety Instruction IS 36 and its Amendment. Any activity inside the static magneticfield shall be subject to risk assessment and ALARA. Personnel shall be informed about the hazardsand appropriately trained. Areas with magnetic flux densities exceeding 0.5 mT shall be delimited (usepacemaker warning signs), while areas with magnetic flux densities exceeding 10 mT shall be renderedinaccessible to the public.
Mechanical Safety.
Currently the TT4 and TT5 galleries are equipped with electrical overhead trav-elling (EOT) cranes, installed in the 1970s; these will be replaced with modern EOT cranes. The newexperimental building will be equipped with an EOT crane and a lift for equipment and personnel. Spe-cial lifting beams will also be required for the installation of the modulators and accelerators components.As the project moves into the Technical Design Report stage, the mechanical Safety aspects shall bereviewed in greater detail.
Civil Engineering Safety.
The installation of the eSPS facility will involve civil works, including tun-nel modifications, new concrete structures, and a new building. At CERN, all infrastructure shall bedesigned in accordance with the applicable Eurocodes to withstand the expected loads during construc-tion and operation, but shall also consider accidental actions, such as seismic activity, fire, release ofcryogens (if any – not currently foreseen) and the effect of radiation on the concrete matrix and othertunnel construction fabric.
Fire Resistance -
New structures and infrastructure shall be designed and executed to guaranteea mechanical resistance for 120 minutes of exposure to the design fire. Eventual passive protection61
Infrastructure and Civil Engineering systems, e.g. intumescent paintings and plasters, will be foreseen only for those elements that are unableto respect such a requirement. The structural assessment will need to be carried out in accordance withEN 1991-1-2, EN 1992-1-2 and EN 1993-1-2.
False Floors -
A number of considerations must be taken into account when deciding the type offloor to install and the space required to meet the above criteria. Each phase of the floor’s life must alsobe designed for and must allow for safe installation, utilisation and finally decommissioning of the floor.Key design criteria include: • Minimise the area of floor that is freely removable (this reduces the risks of damage and deterior-ation of the floor from multiple interventions); • Where possible provide a full height access to the cable trays and pipework; • Identify cable pulling routes for the current layout and where possible future modifications and tryto fix smaller access points for pulling the cables; • Identify points of access and egress from the floor for maintenance purposes that have the minimumimpact on the normal walkways through the building; • Detail the pulling points and access points such that there is a rigid barrier integrated into theaccess trap (access via a trap door which in the open position is supported by barriers that protectthe opening); • Provide a suitable ladder or steps to access under the floor; • Consult all groups who may need to pull cables to ensure the layout chosen is suitable and ifnecessary pre-equip the cable trays with pulleys and pulling wires; • Consult with transport to identify the method for transport of foreseen equipment into the buildingand verify the floor loadings; • Develop a plan of the building showing all access and egress points for daily use, access under thefloor for works, fire escape routes and transport routes. This can be used to quickly validate theloadings and mark the floor areas with the permissible loads; • Once the basic layout of the floor and structures is made, the removable parts can be either designedin-house or bought in as a system. In either case it is highly recommended to have a system thatfixes the position of the tiles with a solid frame to prevent creep of the tiles when repeatedly liftedand replaced. It is also worth considering a numbering system or pattern to help ensure tiles arereplaced with the correct orientation and position.
Environmental Safety.
Many efforts have been made in the conception of the eSPS design to econom-ise on the use of energy and water, particularly in the design of the cooling and ventilation system forTT4 and TT5. With the refurbishment of this linac area, along with the construction of the new exper-imental area, waste, and particularly the excavation of soil are foreseen to be significant considerationsin the delivery of this project. As the design of the experimental area is still at a preliminary stage, thegeneral requirements for environmental protection have been set out here, with a full study to follow asthe project moves into detailed design. The new and refurbished facilities for this project are locatedentirely within French territory, and French regulations are, therefore, to be applied.62
Infrastructure and Civil Engineering
Air -
Atmospheric emissions shall be limited at the source and shall comply with the relevant tech-nical provisions of the following regulations: • Arrêté du 02 février 1998 relatif aux prélèvements et à la consommation d’eau ainsi qu’aux émis-sions de toute nature des installations classées pour la protection de l’environnement soumises àautorisation, Articles 26, 27, 28, 29, 30.
The design of exhaust air discharge points shall comply with the requirements of the section 5.1.3of the CERN Safety Guideline C-1-0-3 - Practical guide for users of local exhaust ventilation (LEV)systems.
Water -
The project leader shall ensure the rational use of water. The discharge of effluent waterinto the CERN clean and sewage water networks shall comply with the relevant technical provisionscontained in the following regulations: • Loi no 2006-1772 du 30 décembre 2006 sur l’eau et les milieux aquatiques; • Arrêté du 02 février 1998 relatif aux prélèvements et à la consommation d’eau ainsi qu’aux émis-sions de toute nature des installations classées pour la protection de l’environnement soumises àautorisation.
The direct or indirect introduction of potentially polluting substances into water, including their infilt-ration into ground is prohibited. Applicable emission limit values for effluent water discharged in theHost States’ territory are defined in the following regulations: • Arrêté du 02 février 1998 relatif aux prélèvements et à la consommation d’eau ainsi qu’aux émis-sions de toute nature des installations classées pour la protection de l’environnement soumises àautorisation Art. 31 and Art. 32.
Retention measures for fire extinguishing water are required for any CERN project in which largequantities of hazardous, or potentially polluting substances are used or stored. As the project moves tothe technical design report stage, through discussions with the environmental protection specialists fromHSE, it will be determined whether the following guidance document shall be applied (in accordancewith the French
Code de l’Environnement ): • Energy -
The use of energy shall be done as efficiently as possible. For the entire facility, adequatemeasures shall be taken to comply with the relevant technical provisions contained in the followingregulations: • Loi no 2010-788 du 12 juillet 2010 portant engagement national pour l’environnement (GrenelleII).
In addition, construction of new buildings sited in France shall comply with the relevant technicalprovisions relating to thermal efficiency contained in the following regulation: • Décret no 2012-1530 du 28 décembre 2012 relatif aux caractéristiques thermiques et à la per-formance énergétique des constructions de bâtiments; Infrastructure and Civil Engineering • Arrêté du 26 octobre 2010 relatif aux caractéristiques thermiques et aux exigences de perform-ance énergétique des bâtiments nouveaux et des parties nouvelles de bâtiments and the FrenchRéglementation Thermique 2012 (RT 2012); • NF EN 15232 Performance énergétique des bâtiments - Impact de l’automatisation, de la régula-tion et de la gestion technique.
Soil -
The natural physical and chemical properties of the soil must be preserved. All the relevanttechnical provisions related to the usage and/or storage of hazardous substances to the environment shallbe fulfilled to avoid any chemical damage to the soil. Furthermore, the excavated material shall behandled adequately and prevent further site contamination. All excavated material must be disposed ofappropriately in accordance with the associated waste regulations.
Waste -
The selection of construction materials, design and fabrication methods shall be such thatthe generation of waste is both minimised and limited at the source. Waste shall be handled from itscollection to its recovery or disposal according to: • Code de l’environnement, Livre V: Titre IV -Déchets; • LOI no 2009-967 du 3 août 2009 de programmation relative à la mise en œuvre du Grenelle del’environnement (1), Art. 46.
The traceability of the waste shall be guaranteed at any time.
The eSPS access process will be based on building access controlSystem and a personnel protection system (PPS) with dedicated access points that will regulate thepersonnel access to the “controlled” Machine areas. The access point will be made of one personnelaccess booth and an optional material booth, and will regulate with the highest safety level the accessaccording to the operating modes managed remotely by the CERN control centre (CCC). It will performseveral checks: such as user’s identification, biometric authentication, access and training validity, properwork authorisation (IMPACT) as well as the use of Operational dosimeters (DMC).The PPS controls a number of independent beam zones divided into access sectors equipped with vari-ous safety elements such as sector doors, end-of-zone doors, mobile shielding, crinoline, patrol boxes,etc. The access sectors are particularly important to organise the patrol of the machine and to minimisethe radiation exposure by means of radiation veto handled locally or remotely by the RP responsibleperson. Several access modes are foreseen: ‘General’, ‘Restricted’ with a ‘Safety token’ and ‘EquipmentTest’, automatically managed by the PPS or controlled by human operator, locally or remotely from theCCC. The eSPS being a facility installed both in the PS and the SPS complex, it would probably beoperated by the SPS operators from the CCC for the most part.In the machine, the safety is ensured by access or beam “Important Element for Safety (EIS)”. Eachbeam zone has its own independent access conditions. The absence of beam in each independent beamzone is guaranteed by at least two beam safety elements, with at least one passive element (e.g. a movablestopper) and one active element (e.g. magnet power converter interlock). These safety measures areactivated and interlocked by the PPS: the access status can make a zone unsafe for operation with beam,or forbid the access to the Machine if the status of a safety element is unsafe. The proposed zoning hasbeen designed in accordance with the RP classification.64
Infrastructure and Civil Engineering
The following beam areas have been considered: • Electron linac; • CLEARER experimental area; • TT5 Hall; • TT4 experimental area and the upper part of the TT61; • TT2-TT10; • eSPS experimental area. The following systems would be implemented for survey and alignment: •
24 modules to be aligned along a straight line of 70 m; • On each module, 4 accelerating structures (length ∼ • The overall tolerance for alignment of each component axis (RF electromechanical of AS andmagnetic axis of quadrupole): would be ± σ ), for roll, pitch andyaw with respect to a straight line.Schedule to be based on preparation during a run with installation during long shutdown. Due to uncer-tainties, a schedule is provided later in the document which is from any given start point and not rootedin the CERN LHC and injector timetable. Survey and Alignment Strategy Proposed.Introduction -
The tolerances of alignment requested for the eSPS are tighter than for current ac-celerators at CERN but larger than in lepton colliders (CLIC or FCC) under study. As a matter of fact,in the LHC, each fiducial, on top of any cryostat has to be aligned within ± σ ),while for CLIC, each reference axis of component has to be included within a cylinder with a radius of 14 µ m over a sliding length of 200 m. The survey and alignment strategy proposed for eSPS will integratespecific steps to cover these tighter tolerances.The configuration of the tunnel in which eSPS components will be installed is very particular: • Narrow and straight: 2.10 m in width (including the components to align) and 2.5 m in height, oncethe shielding wall and roof are in place; • The main components in the “tunnel” will be linked to modulators and klystrons located on theother side of the shielding wall by RF waveguides; • The shielding roof will be put in place once all the components are installed. Components will bepre-aligned before shielding walls are installed and their alignment will be controlled again oncethe shielding has been added; • TT4 and TT5 halls should be a very stable in terms of ground movement due to their age, however,this is to be confirmed by regular measurements before the installation of components.65
Infrastructure and Civil Engineering
Steps of Survey and Alignment Strategy Proposed -
First, a geodetic network will be installedin all the transfer tunnels required for eSPS and the position of points w.r.t. to the existing undergroundnetwork and components (located in TCC6) re-determined. The new points installed in TT4 and TT5,will be located preferably on the ground floor (to limit the impact of the shielding roof), close to theexisting wall, in the “transport” area. Additional targets will be installed on the permanent walls.The fiducialisation and initial alignment of components on the same module will be a very importantstep in the overall alignment. If such an “assembly” alignment on each module is performed within ± µ m (1 σ ), this will allow use of standard means of alignment later in the tunnel for each module.Each accelerating structure will have to be fiducialised independently, e.g. its mechanical axis will bedetermined w.r.t. external targets, at the metrology lab or using laser tracker. In both cases, this meanswe will align them w.r.t. to their mechanical axis, which does not correspond to their RF axis. For a moreaccurate determination, it is recommended to use a stretched wire setup (see [91]): the stretched wire willbe put at the mean RF axis and then the position of the wire will be determined w.r.t. external fiducials.The same for the determination of the quadrupole magnetic axis (see [92]). Once all the components arefiducialised, they will be pre-aligned on a common support, using universal adjustment platforms. Themean axis of components will be determined in the support coordinate system, materialised by fiducialsor sensors interfaces, depending on the solution chosen for their smoothing. Some components comingfrom the CLEAR injector will need to be re-fiducialised as well, by a combination of laser tracker andRomer arm measurements.Once all the MAD-X files and all element drawings are available, the marking can take place in thetunnels, including the beam line, and the projection of the jacks’ position on the ground floor.Once the modules and the components are ready, they can be transferred and pre-aligned w.r.t. thegeodetic network using laser tracker measurements, within ± σ ). Different options are possibleconcerning the smoothing (final relative alignment) of the modules within ± σ ): • Option 1:
Use of standard means of measurements to position them in vertical and in radialposition (offsets measurements w.r.t. a stretched wire for radial alignment, levelling measurementsfor vertical alignment), combined with inclinometer measurements. Such a smoothing operationwill have to be carried out at least twice: before the installation of the shielding and after theinstallation of the shielding; • Option 2:
Use of temporary WPS sensors, measuring w.r.t. a permanent wire stretched betweentwo metrological platforms (located at each extremity) with an absolute position determined in thegeodetic network of the tunnels. The two metrological platforms will be equipped with permanentWPS sensors and HLS sensors, while the modules will be equipped with permanent sensors inter-faces, determined in the coordinate system of the module. The configuration of temporary sensorsper module would consist of two wires, four WPS sensors, providing a redundant and controlleddetermination of the position of each module. Each module would be aligned one after the otherw.r.t. the extremity reference platforms, using four sensors plugged temporary. Such a solutionprovides the possibility to upgrade later on each module with permanent sensors if needed; • Option 3:
Use of WPS sensors installed permanently, with the same options concerning theirconfiguration as before.In all cases, the standard fiducials (or the sensors interfaces) shall be located above the adjustmentmeans to avoid the level arms effect during the adjustment process and to both facilitate and quicken theprocess of alignment. Adjustment solutions shall be studied and installed below each module as well,with the possibility of motorisation (if needed). 66
Infrastructure and Civil Engineering
The injection infrastructure is to be housed in the existing B183 (B183), Transfer Tunnel5 (TT5) and Transfer Tunnel 4 (TT4). The area which would be used to house eSPS is currently usedas low-level radioactive material storage. The existing stored materials would need to be relocated inadvance of re-purposing for eSPS. This is not covered as part of this feasibility study although the subjectis discussed further in Sec. 5.2.8. Figure 5.7: B183 floor plan.
Existing Structures.B183 -
B183 is an existing structure built in 1971 as an experimental hall forming part of the ’Westarea’. The building is constructed in a mixture of reinforced concrete and steel frame construction witha truss roof structure. The building is predominately used as part of the magnet assembly building butthe area (designated R-002) abutting TT5 is a continuation of the low-level radioactive material storagearea as shown on Fig. 5.7. This area is surrounded by and delineated by shielding blocks and requires nocivil engineering works to be used for the scheme.
TT5 -
TT5 is a buried reinforced concrete structure built in 1971 to house three transfer beamlinesdelivering beam from the SPS to the West area. The main part of the structure comprises deep stripfoundations of variable depth sat on the Molasse rockhead with a 620 mm deep reinforced concrete slababove a further 600-800 mm thick layer of improved structural fill spanning between founds. The slab ismade up of a 300 mm lean concrete base below a 300 mm reinforced concrete slab finished with a layerof PVC waterproofing and a 2 mm layer of screed. The walls are 7.32 m tall, 800 mm thick reinforcedconcrete columns with a 1.8 m wide, 600 mm deep footing supported on the strip foundations. Thereinforced concrete roof slab is cast integrally with the columns in nine bays and varies in depth from1.2 m at the outside of columns to 1.7 m at the centre. The typical section is shown in Fig. 5.9.The building in plan measures 47.53 m long by 16 m wide. There are a number of channels which runthe full width of the hall with metal plate covers. Channels are 400 mm wide and at irregular spacingvarying between 2.98 m to 7.9 m but typically at around 4.5 m spacing. The channels have drains at themid-point of the hall with a carrier drain below running along the centre line of the building.As-built drawings for the structure exist and are quite detailed, including reinforcement and detailingdrawings. Details of the existing structure are shown in Figs. 5.9, 5.10 and 5.11.67
Infrastructure and Civil Engineering
Figure 5.8: View within existing TT5 tunnel looking east towards TT4.Figure 5.9: Existing TT5 typical section.TT5 also has several technical galleries beneath it to carry services and provide access for maintenancewhich are shown in the tunnel long-section Fig. 5.11 . Additional tunnels provide level access to the mainbuilding from the surrounding hard-standing. 68
Infrastructure and Civil Engineering
Figure 5.10: Existing TT5 plan view showing channels and drains in floor slab.Figure 5.11: Existing TT5 long-section.
TT4 -
TT4 is another buried reinforced concrete structure built in 1971 as a continuation of thestructures housing transfer beamlines bringing beam from the SPS to the West area. TT4 abuts theeastern end of TT5 as shown in Fig. 5.13, connecting it with TT61 and TT3 to the east. TT4 comprisesvariable depth lean mass concrete strip foundations which extend from the Molasse rockhead to the baseof the box section footings as shown in Fig. 5.12. TT4 has a concrete slab 500 mm deep with a PVCwasterproofing layer and 2mm screed on top. Figure 5.12 also shows the significant depth of earth abovethe structure as well as the retaining wall running alongside TT4 retaining the higher ground above
RouteNord , a general purpose road.The reinforced concrete tunnel structure above spans between 1 m deep, 1.8 m wide footings. Thewalls are 600 mm wide and 6m tall while the integral roof slab is 800 mm thick with a span of 8 m overthe majority of the structure. The layout in plan and in terms of floor levels is more complex at itsjunction with TT61 and TT3/TT2A. TT4 has channels in every way similar to those found in TT5, againat regular spacings. Some channels have been filled in with concrete.Drawings for TT4 are less detailed, potentially making future works to modify the structure morechallenging but still achievable.As part of the study, 3D scans of TT5 and TT4 were carried out by the SMB-SE-DOP section toallow an accurate model to be produced as shown in Fig. 5.14. This has been used as the basis of anew integration layout. In general the scans showed good agreement with the as-built drawings, givinga fair degree of confidence in the drawings’ quality, although there were some areas which requiredadjustments.All of the structures were found to be in generally good condition with no major issues apparent from69
Infrastructure and Civil Engineering a civil engineering perspective.Figure 5.12: Existing TT4 typical cross section.Figure 5.13: Existing TT4 long-section showing junction with TT5.Figure 5.14: Interim scan output during processing.70
Infrastructure and Civil Engineering
Civil Engineering Enabling Works.
Limited civil engineering works are required in order to facilitatethe implementation of the eSPS scheme. The integration study carried out focused on fitting the requiredinjection and associated infrastructure into the existing space within B183, TT5 and TT4 and throughcareful planning and optimisation between all disciplines, this has been achieved. Some limited civilengineering enabling works are still required and these are summarised here.
B183 -
In B183 there are minimal enabling works required to implement the scheme. The mainelements are as follows: • A dividing wall between B183 and TT5 will be needed to provide compartmentalisation for firesafety and to provide some additional shielding to equipment located in B183. The requirementsfor the wall are set by applicable fire safety standards and shielding requirements detailed in otherchapters. The wall must be 2 m thick, 4.5 m high constructed in normal density concrete; • Although the form of construction is not yet finalised, rooms will need to be constructed to accom-modate ancillary infrastructure as shown in Fig. 5.15: – ×
10 m Control room; – × – × Infrastructure and Civil Engineering
TT5 -
In TT5, the layout required is shown in Fig. 5.16. The main civil engineering contribution isthe required shielding walls and roof slabs over the injector and linac. The shielding has been defined bythe RP study which has identified the thickness of shielding needed along with the grade: • Normal density concrete 2.4 g/cm ; • High density concrete 3.9 g/cm ; • Iron 7.9 g/cm . Figure 5.16: TT5 layout.The concrete shielding walls will be formed from a number of temporary blocks to allow greater flex-ibility for any future changes in layout and greater speed of construction. The blocks will be cast offsiteand transported into position at a suitable time during refurbishment. The arrangement of temporaryshielding blocks will be designed to allow blocks to be standardised as far as possible. Blocks will,however, need to interlock and ensure there is no shine path for radiation to pass meaning there will beslightly reduced opportunity for modularisation.The concrete mix for normal density concrete is flexible and needs only to support itself and any blocksabove. The concrete mix for high density concrete has been selected based on a density high enough tobe suitable for radiation protection criteria while being practical for standard concrete suppliers withoutbeing cost prohibitive. The concrete mix is based on using high density magnetite, an ore of iron, asaggregate in the concrete:Formula per cubic meter given as an indication for a density of 3.9 g/cm , • MD8S 0/6 mm: 1180 kg; • MD20S 0/16 mm: 2360 kg; • CEM II type cement: 310 kg; • Water (including moisture from aggregates): 160/170 l; • High water reducing super-plasticiser: 1.5 to 2% of the weight of the cement.In addition to the shielding walls, the higher levels of radiation from the injector and linac necessitatea shielding roof. This is to be removable to allow access to equipment by bridge cranes. The roofwill be formed of a number of separate slabs. Each slab unit must be lifted and transported via the72
Infrastructure and Civil Engineering crane so the length of has been selected to ensure they remain within the crane’s safe lifting capacity.For radiation protection, slabs must be 800 mm thick and have been sized (in terms of width) to spanbetween the central shielding wall and a support beam to be fixed to the wall of TT5 (and also TT4). Anoutline design of the support beam has been carried out by the EN-AC-INT section using conservativeassumptions which confirmed this would be feasible. The initial design was based on a new steel beam200mm wide such as HEB 200 with HEB 140 supports every metre attached to the existing walls withheavy duty anchorages such as Hilti HSL-3-G.An initial assessment of floor loading for the shielding block arrangement has been carried out by theSMB-SE-DOP section which found the existing slab and foundations will be sufficient.In several places in TT5, cores or breakouts in the existing concrete will be required to allow CV ductsto enter and leave the structure. Initial study shows there should be no issue with feasibility, however,further study will be needed to design and detail these penetrations. Experience has shown reinforcementdrawings cannot be absolutely relied upon so ground penetrating radar (GPR) scans will be needed aswell as assessment and design to ensure there is no detrimental reduction on the structural capacity ofwalls, roof slabs or floor slabs. To ensure the water tightness of the structure is not impacted, a suitablewaterproofing joint detail will be used with hydrophillic sealant to prevent water ingress around servicepenetrations.
TT4 -
In TT4, the layout required is shown in Fig. 5.17. The enabling works will practically be thesame as those required in TT5 with the addition of some minor works to modify access arrangementsto n_TOF at the junction with TT3. Equipment transport and access will not be affected, other than ashielding wall will need to be moved to allow access. Options for this include a movable wall on railsor a wall composed of blocks which can be moved and replaced when infrequent access is required. Forpersonnel access which is far more frequent, a set of steel stairs or access ladder can be provided beyondthe shielding wall which will allow access via TT3 and the ISR.Significant enabling is also required for CV services. Above TT4 and TT5, several ducts and a servicebuilding must be sited as shown in Fig. 5.20. Although this layout shows the ducts and building sitedimmediately above the building, it is likely all ducts and building will be sited at or just below existingground level. Some deep trench-box excavations will be needed to enable duct connections to be formed.Lagged and trace heated duct-work can be installed at ground level. As there is a considerable earthmound above the tunnel here, this will be more cost effective than removing all earthworks and will, inaddition, leave the considerable earth shielding in place for the future in case it is needed for radiationprotection.The overall load of the building and any applied loads when considered at the level of the tunnel shouldnot be problematic, although a more detailed study will be required to confirm this at the next stage ofdevelopment. Figure 5.17: Layout of TT4.73
Infrastructure and Civil Engineering
B183 is an old experimental hall connected to transfer tunnel TT5. The control room, meetingroom and laser room for the eSPS facility are located at the western end of B183 next to transfer tunnelTT5, in which this area is currently used for storing shielding blocks. A 2 m thick shielding wall willseparate B183 from TT5. The size of the control room and meeting room is like that of CERN’s existingCTF3 control and meeting room in B2008. The laser room is a controlled room with an area like like thatof CERN’s existing laser room in B2013. There is an entrance and a laser transport pipe with a diameterof 50 mm goes through the shielding wall to the CLEAR RF gun in TT5 approx. 5 m away. As the totaldistance from the laser output to the cathode needs to be matched to the focal length of the lenses inuse, a revised layout will be studied at the technical design stage of the project. The vehicle access intoTT5 is via B183 in which vehicles enter through the existing lorry access door and into TT5 through thefire/ventilation badge-controlled access door as shown in Fig. 5.15.
Transfer Tunnel TT5.
Transfer tunnel TT5 is an underground tunnel connected to TT4. Historically,three beamlines coming from TT4 were focused on primary targets located in TT5 after which secondarybeams were sent to experiments in B180. The tunnel is currently used to store low-level radioactivemagnets and shielding blocks.TT5 will house the CLEAR injector, modulators, the experimental beamline and the first 10 m of thelinac as well as some racks for the different infrastructure as shown in Fig. 5.16.The CLEAR injector is an existing injector at CERN housed in the B2010. The start of the injectorwill be located 2 m from the shielding wall separating B183 from TT5. This allows personnel to easilyaccess the injector and it is also positioned away from an existing technical gallery access shaft of whichwill be permanently closed. For radiation protection the injector is surrounded by concrete shieldingwith an clear width of 2.8 m to allow for the installation and maintenance of the injector. With respect tothe beam direction starting on the left-hand side the breakdown is as follows: •
200 mm allowance for safety lighting and services; •
900 mm allowance for personnel access; •
700 mm allowance for the width of the injector; • ◦ into an experimental area. The layoutof which was specified by the equipment owners. Personnel access to the injector and experimental areais via the access chicanes as shown in Fig. 5.16. The width of the chicanes is 1.4 m to ensure that firefighters can access the area with a stretcher and have sufficient room to turn.The control racks for the access control, fire detection, beam instrumentation, vacuum and the powerconverters for the facility were specified by the equipment owners. The racks are standard SPS 45Uracks used at CERN. The required number of racks for each sub-system is listed in Table 5.6. Transfer Tunnel TT4.
Transfer tunnel TT4 is an underground tunnel connected to transfer tunnel TT61.Historically, the SPS beamline transferred via TT61 was split into three proton beamlines in TT4 andcontinued into TT5. The tunnel is currently used to store low-level radioactive magnets and shieldingblocks.TT4 will house the linac beamline and modulators as well as an experimental beam dump area atthe downstream end of TT4. The linac is approx. 65 m in length of which is composed of 24 moduleseach 2.65 m long with a gap of 200 mm every 4 th module for the installation of the sector valve, vacuum74 Infrastructure and Civil Engineering
Table 5.6: Required number of racks for each sub-systems.
Equipment Number of racks
Access control 3Fire detection 3Experimental beamline vacuum control 3Experimental beamline beam instrumentation 2Experimental beamline power converters 7Experimental beamline spare 3Linac vacuum control 6Linac beam instrumentation 6Linac power converters 5Linac spare 3Transfer line beam instrumentation and vacuum control 1Transfer line power converters 3Transfer line spare 1Total 46pumping port and for the connection of the primary pumping. The beamline splits at the downstream endof TT4 with the SPS transfer beamline bending into TT61 towards the SPS and the beam dump beamlinecontinuing straight as shown in Fig. 5.17.Personnel access to the linac is via the fire/ventilation door at the downstream end or via the CLEARinjector chicane entrance in TT5. Access to the modulators at the downstream end is via a concrete chi-cane and access gate or via the access gate in TT5. Because of the restricted geometry of transfer tunnelTT4, a detailed breakdown of the cross-section was undertaken to ensure that the linac and modulatorscould fit within the existing tunnel whilst adhering to the radiation protection and general safety re-quirements specified by CERN experts. The dimensions required for equipment and personnel/transportaccess, as shown in Fig. 5.18 are: •
200 mm allowance for safety lighting and services; •
905 mm allowance for personnel access; •
645 mm width for the linac; •
550 mm allowance for personnel access for maintenance of the linac and cooling equipment; •
800 mm width of shielding wall; •
600 mm allowance for personnel access; • • •
200 mm allowance for safety lighting and services.As the modulators are quite large, the access around the modulators is restricted because the spacebetween them is controlled by the spacing of the linac beamline as detailed above. The longitudinaldistance between each unit is 900 mm, apart from every 4 th unit in which this distance is increased to1100 mm. There is a movable rack between every two units as shown in Fig. 5.19. Due to the restricted75 Infrastructure and Civil Engineering
Figure 5.18: Cross-section of TT4.cross-sectional geometry, the access on the side close to the shielding wall is only 600 mm. There is afalse floor surrounding the modulators to allow the cabling and piping to be distributed underneath themodulators allowing unobstructed access all around the modulators.
Cooling and Ventilation Layout.
A CV study was undertaken for the facility and the infrastructurerequirements integrated into the existing layout. A new CV building will be located on top of TT4 ideallylocated such that the length of cooling piping and ventilation ducts was optimised throughout the facility.The CV layout of the facility is shown in Figs. 5.20, 5.21 and 5.22. For further details on the CV studyundertaken see Sec. 5.2.3. 76
Infrastructure and Civil Engineering
Figure 5.19: Integration of linac and modulators in TT4Figure 5.20: External cooling and ventilation layout for TT4 and TT5.
Radiation Protection Layout.
Radiation protection was one of the key aspects in the overall layoutof the facility. Detailed radiation protection studies have been undertaken to determine the material andthickness of the shielding walls and roofs in the facility (see Sec 5.2.4). Different densities of shieldingmaterial have been utilised to fit within the existing geometry of facility as shown in Figs. 5.23, 5.24and 5.25. 77
Infrastructure and Civil Engineering
Figure 5.21: Internal cooling and ventilation layout for TT5.Figure 5.22: Internal cooling and ventilation layout for TT4.
Access Control Layout.
The access control systems have been defined in accordance with the accessrequirements of the different equipment owners. The access control layout for transfer tunnels TT4 andTT5 are shown in Fig. 5.26 and Fig. 5.27. Access rules to the different areas was studied in detail, inparticular to the n_ToF target [93].
The cooling and ventilation conceptual design for TT4 and TT5 is herewith presented. It is based onheat loads provided by machine experts in February 2020 and is strongly dependent on these numbers. A78
Infrastructure and Civil Engineering
Figure 5.23: Shielding wall layout in TT5.Figure 5.24: Shielding roof layout in TT5.Figure 5.25: Shielding wall layout in TT4.79
Infrastructure and Civil Engineering
Figure 5.26: Access control layout in TT4.Figure 5.27: Access control layout in TT5.80
Infrastructure and Civil Engineering safety factor of 1.2 is applied to the heat loads to account for uncertainties and future facility expansions.The premises are divided in three regions according to operational requirements. The first encom-passes the accelerator and injector, the second includes the klystrons and the low energy experimentalarea and the third comprises the area TT4/TT61. These regions, or compartments, are presented inFig. 5.28, along with the heat load sources.Figure 5.28: Top view of the studied compartments and heat load sources.
Piped Utilities.
Demineralised water circuits are foreseen to cool equipment installed in TT4, TT5and TT4/TT61, as portrayed in Figs. 5.29 and 5.30. The list of corresponding heat loads is presented inTable 5.7. Local fine temperature stabilisation is provided for the pulse compressors (in the acceleratorcompartment), using electrical resistances, placed on the shielding wall.Existing cooling towers (B280) are used in the primary circuit to cool the demineralised water loop.Pipes extend from the cooling towers to the CV building, sited on top of TT4, as shown in Fig. 5.31.Two cooling stations are foreseen to be installed there - one for the linac compartment (lines 1 and 2in Table 5.7) and one for the klystrons compartment and TT4/TT61 region (lines 3 to 5 in Table 5.7).The pumps in cooling stations have an N+1 redundancy, contrary to heat exchangers, for which noredundancy is considered. Details concerning the primary and demineralised water circuits can be foundin Table 5.8 and in the simplified diagram presented in Fig.5.32.Table 5.7: Water heat loads.
Component Heat Load (kW) Safety Margin Final Load (kW)
Accelerator 178.4 1.2 214.1Injector 42.1 1.2 50.5Klystrons 615.5 1.2 737.5Injector’s klystrons 102.4 1.2 123.0Low energy exp. area 10.0 1.2 12.0High energy exp. area 10.0 1.2 12.0Magnets in TT4/TT61 9.5 1.2 11.4Table 5.8: Primary and demineralised water circuits.
Circuit Heat (kW) Design flow (m /h) Piping (mm) dP (bar) Pumps (kW e ) Primary 1160.5 131 DN200 - -Accelerator Comp. 264.6 30 DN100 5 2 x 6Klystrons Comp. 895.9 101 DN150 5 2 x 1881
Infrastructure and Civil Engineering
Figure 5.29: Top view of demineralised water circuits.Figure 5.30: Integration of the demineralised water pipes in the accelerator and klystron compartments.Figure 5.31: Chillers and CV building positioning.Two air cooled chillers (one is redundant) with a maximum refrigeration capacity of 150 kW (each) areforeseen to generate chilled water for air handling units (AHU). The two chillers are installed outdoors,close to the CV building, where the AHUs are placed, as displayed in Fig. 5.31. In the future, duringa more detailed stage of the study, one should investigate whether water cooled chillers, that tend to bemore efficient, are to be used instead. The details concerning the chilled water plant and circuits can be82
Infrastructure and Civil Engineering
Figure 5.32: Simplified diagram of the primary and demineralised water circuits.found in Table 5.9 and Table 5.10, as well as in the diagram presented in Fig. 5.33. The chilled waterpumps have an N+1 redundancy, to ensure proper operation in case of technical failure of a pump.It is often energy efficient to vary the chilled water temperature as a function of the loads. In par-ticular, when they are reduced, the water temperature is increased, resulting in lower chiller electricityconsumption. This strategy has a negative impact on pump consumption (in variable flow systems), asthe flow rate is higher. However, for short distribution systems, as in this case, the gains greatly offsetthe losses. A detailed study on this topic would be of interest for a future technical design report.Table 5.9: Chilled water plant.
Equipment Refrigeration Capacity (kW) Design water flow (m /h) Piping (mm) Chiller 1 150 19 DN80Chiller 2 150 19 DN80Table 5.10: Chilled water circuits.
Circuit Heat (kW) Design flow (m /h) Piping (mm) dP (bar) Pumps (kW e ) Main branch 150 19 DN80 1 2 x 650AHU L1 55 8 DN50 - -AHU L2 55 8 DN50 - -AHU K1 75 11 DN65 - -AHU K2 75 11 DN65 - -83
Infrastructure and Civil Engineering
Figure 5.33: Simplified diagram of the chilled water plant.A dry line for firefighting is put in place in the klystron fire compartment, where large fire heat loadsare present. A DN100 pipe is foreseen to transport 60 m /h (enough to supply two firefighting teams)of industrial water in the compartment. It has an outlet for connection every 20 m. In case of fire,valves can be opened at the surface in order to provide the desired flow. Figure 5.34 shows the two firecompartments and the DN100 pipe.Figure 5.34: Fire compartments and firefighting water pipe. Heating Ventilation and Air-Conditioning.
The HVAC plant is designed to guarantee the requiredindoor conditions for the most demanding scenario. Three operational modes are foreseen - run mode,access mode and smoke extraction. In run mode, the linac is running and people are not allowed inthe accelerator compartment. On the other hand, they are allowed in the klystron compartment. Accessmode is activated when the machine is switched off and people have to work within the accelerator84
Infrastructure and Civil Engineering compartment. Smoke extraction is engaged by the fire brigade in case of need. The system is designedto handle both cold and hot smoke.A push-and-pull ventilation system is adopted for the three operational modes. During run and accessmodes, air is supplied at 20 ◦ C and 22 ◦ C to the klystron and accelerator compartments respectively. Themaximum supply dew point temperature is 11 ◦ C for both regions and modes so that condensation isavoided on surfaces with temperatures lower than 12 ◦ C. Air can be partially recycled in each compart-ment, as ducts are installed to drive the extracted flow back to the AHUs, sited in the CV building, asillustrated in Fig. 5.31. This allows for great energy savings, as treating fresh air is minimised. Dur-ing smoke extraction, fresh air is supplied to the fire compartments, displayed in Fig. 5.34 without anytreatment.The heat loads for run mode are presented in Table 5.11, and correspond to values provided by machineexperts for the desired ambient temperature. In access mode, equipment is switched off and the loads arebased on the maximum foreseen occupancy - 10 people in the accelerator compartment and 20 people inthe klystron region undertaking hard physical work. The respective loads are presented in Table 5.12.Table 5.11: Air heat loads during run mode.
Component Sensible (kW) Margin Final Sens. (kW) Latent (kW)
Accelerator 17.3 1.2 21.0 0.0Injector 4.1 1.2 5.0 0.0Klystrons 44.9 1.2 54.0 0.0Injector’s klystrons 7.5 1.2 9.0 0.0Low energy exp. area 3.0 1.2 4.0 0.0High energy exp. area 3.0 1.2 4.0 0.0Magnets in TT4/TT61 0.5 1.2 0.6 0.0Table 5.12: Air heat loads during access mode.
Compartment Sensible (kW) Latent (kW)
Accelerator Comp. 1.5 2.7Klystron Comp. 3.1 5.4The ventilation layout during run mode is shown in Fig. 5.35. During this mode, air is mostly recir-culated to avoid the introduction of fresh air heat loads. The flow rates are 25000 m /h and 35000 m /hin the accelerator and klystron compartments respectively. They are selected according to the desiredinlet-outlet temperature difference, whilst providing reasonable longitudinal velocities. The inlet-outlettemperature difference in the accelerator compartment is smaller than the standard values. This is toensure a relatively small temperature gradient in the area, which is relevant for the beam alignment sys-tems. On the other hand, a higher difference can be taken for the klystron region, where there are noconstraints on this parameter. The low energy experimental area is ventilated (200 m /h, correspondingto 1 air change per hour) with air transferred from the klystron compartment to the accelerator one (froman area with lower to higher radiation). This implies a minimum fresh air flow of 200 m /h in the klystrongallery during run mode, which is desired, given that people can access this area during run mode (non-etheless, more people are expected to work in the space during access mode). The loads in the low andhigh energy experimental areas are treated by local direct expansion units. Moreover, two systems areadded to avoid stagnant air or recirculating flow patterns in the injector’s klystron area and TT4/TT61.The flow rates are defined to ensure at least three air changes per hour in these regions.85 Infrastructure and Civil Engineering
Figure 5.35: Ventilation layout during run mode.The ventilation layout during access mode is presented in Fig. 5.36. The design flow rate is 12000 m /hfor both the accelerator and klystron regions. However, they are calculated based on different paramet-ers. For the accelerator compartment, a minimum longitudinal velocity of 0.7 m/s is taken, whilst for thelatter, two air changes per hour are ensured. Most of the air is recirculated; however, minimum flowsof 400 m /h and 750 m /h of fresh air are guaranteed in the accelerator and klystron compartments re-spectively, to provide proper air quality for the maximum number of people foreseen in these regions.The systems aiming at enhancing air mixing in the injector’s klystron and TT4/TT61 regions run duringaccess mode, as portrayed in Fig. 5.36.Figure 5.36: Ventilation layout during access mode.During smoke extraction, a fully fresh air ventilation is adopted. The layout is illustrated in Fig. 5.37,where the selected flow rates are shown: 30000 m /h for the accelerator region and 60000 m /h for theklystron compartment. The flow rates were defined by analogy, using French regulations for buildings.The system is able to extract both cold and hot smoke. Hence, the extraction ducts and the extractionunits (EXUs) are fire rated. The EXUs are not part of the integration model displayed in Fig. 5.31, butthey are foreseen to be installed close to the extraction points, to avoid the installation of long fire ratedducts. According to radiation protection studies, the smoke can be freely released to the atmospherewithout any sort of treatment or filtration.The HVAC infrastructure, presented in Table 5.13, is designed to withstand the most demanding oper-ational conditions. Moreover, most of the equipment has an N+1 redundancy to ensure normal operationin case of failure. The air handling units dedicated to the accelerator compartment, AHU L1 and L2 (oneis redundant), have to provide a maximum flow of 30000 m /h during smoke extraction mode. Simil-arly, the maximum flow rate for the AHUs treating the klystrons compartment, AHU K1 and K2 (one86 Infrastructure and Civil Engineering
Figure 5.37: Ventilation layout during smoke extraction.is redundant), is 60000 m /h. The smoke extraction units match the flows of the supply ones: EXU L1and EXU L2 (one is redundant) are able to extract 30000 m /h whilst EXU K1 and EXU K2 (one isredundant) extract 60000 m /h. The maximum cooling and heating capacity of AHU L1 and L2 are isrequired for access mode, due to the dehumidification and reheating needed to treat 400 m /h of freshair. For AHU K1 and K2, the maximum heating is also defined during access mode, particularly for thereheating stage after dehumidification of 750 m /h of fresh air. However, the maximum cooling happensduring run mode, to deal with the technical sensible heat loads, as well as to condition 200 m /h of freshair.At the moment, the heating is foreseen to be provided by electrical resistances. However, the energyreleased by the chiller’s condenser might well be recovered for reheating after dehumidification. Thiswould greatly reduce the energy consumption. A possible alternative is the use of solar panels to heatthe water for reheating air. A detailed study on these and other energy efficiency aspects is needed at thenext stage of design. Table 5.13: HVAC Infrastructure. Equipment Cooling (kW) @ flow (m /h) Heating (kW) @ flow (m /h) Max flow (m /h) AHU L1 55 @ 12000 50 @ 12000 30000AHU L2 55 @ 12000 50 @ 12000 30000AHU K1 75 @ 35000 40 @ 12000 60000AHU K2 75 @ 35000 40 @ 12000 60000EXU L1 - - 30000EXU L2 - - 30000EXU K1 - - 60000EXU K2 - - 60000
The linac, including the two CLEARER experimentalareas, will be installed in the existing infrastructure of TT5 and TT4. The linac will operate with varyingbeam power depending on the beam destination. Areas in TT4 and TT5 will be separated into differentaccess zones to ensure personnel safety during the operation of the accelerator and experimental areas.Figure 5.51 gives an overview on the zoning and sectorisation concept.Access to classified radiation areas must be controlled through an access control system. Figure 5.38summarises the access status to all areas depending to the operational conditions. The table includes aswell the radiological area classification. All generally accessible areas will be classified as Supervised87
Infrastructure and Civil Engineering
Figure 5.38:
Access conditions to different areas, depending on the operational scenarios. Radiation area classi-fication: ND: Non-designated area, ZS: Supervised area, ZCS: Simple controlled area, ZSL: Limitedstay area, ZHR: High radiation area.
Radiation Area. An exception is the sector of the klystron gallery next to the X-Band linac (Sector AK).This part will be accessible during RF conditioning in the X-Band linac and will be classified as a SimpleControlled Radiation Area. During X-Band linac beam operation the area will be closed and access isprohibited. The high energy experimental area and the upper part of TT61 will remain closed duringX-Band RF conditioning, injector operation for plasma studies (see Sec. 3.7) and X-Band linac beamoperation. Access to the low energy experimental area will be possible independently of the operationalmodes of the linac.The area in TT4 adjacent to TT3 and the n_TOF Experimental Area 1 in TT2 is classified as SupervisedRadiation Area and will remain accessible during all operation modes. This will allow personnel accessto n_TOF independent of eSPS operation. Material access through the access gallery 852 and TT4will only possible without X-Band linac beam operation, injector operation for plasma studies or RFconditioning.The exact location of the separation door between the SPS and the X-Band linac in TT61 has not yetbeen defined. It will be placed such as to allow access to each part of the TT61 tunnel while the adjacentzone is in operation. During X-Band linac operation without ejection to the SPS, the lower part of TT61shall remain accessible, depending only on the operation mode of the SPS. Access to the upper part ofTT61 shall be possible while the SPS is in operation with ejection to HiRadMat or the LHC.
Prompt Radiation.
The shielding design in the linac part is challenging due to the increasing beamenergy, high beam intensities, different beam destinations, operational conditions and space constraints.To maximise the available space a high density iron-ore based concrete (magnetite type with density of3 . / cm ) is used as shielding material in most of the locations. This type of concrete is used throughoutfor the lateral and top shielding on both, the S-Band linac including its experimental area and the X-Bandlinac. A general wall and roof thickness of 1 m is used on the S-Band linac and 0 . Infrastructure and Civil Engineering
The proposed layout limits access to areas in forward direction of the beam. One exception is the S-Band linac beam, which is bent by 180 ◦ to deliver beam to the low energy experimental area. Shieldingwalls which could be hit directly by beam losses with angles close to 0 ◦ are made of 1 . • Distributed beam loss along the S-Band linac over a distance of 15 m between 5 MeV and 250 MeVand 2% of the nominal beam intensity; • A point-like beam loss in the return chicane towards the low energy experimental area, pointingperpendicular to the linac beam direction at 250 MeV and 1% of the nominal beam intensity; • Distributed beam loss along the the low energy experimental area beamline over a distance of ∼
10 m between 250 MeV and 3 .
65 GeV and 2% of the nominal beam intensity; • Distributed beam loss along the X-Band linac over a distance of ∼
70 m between 250 MeV and3 .
65 GeV and 2% of the nominal beam intensity; • Point-like beam losses in the vacuum pipe after the X-Band linac at 3 .
65 GeV and each with 1%of the nominal beam intensity; • Nominal beam operation on both beam dumps in the low energy experimental area at 250 MeVand on the linac main dump at 3 .
65 GeV.Currently not included was a potential future upgrade for plasma related studies as described inSec. 3.7. This additional injector may have an impact on the location of the inter-machine door inTT61 and the shielding wall towards the n_TOF area.Figure 5.39 shows the combined loss scenarios at nominal parameters with maximum admissible beamloss intensity. For such limiting operation scenarios at maximum nominal intensities to all destinations(except the injection towards the SPS), dose rate levels in accessible areas in TT5 and TT4 (n_TOF side)remain within the defined design limits.
RF Conditioning X-Band Linac.
High-energy X-ray radiation and neutrons may be generated fromthe dark current which is produced, captured and accelerated in the RF structures during conditioning.During beam operation the klystron gallery next to the X-Band linac will remain inaccessible, whileaccess is needed during RF conditioning periods. Dose rates must hence comply with the defined areaclassification. The dark current source term, both in terms of absolute current and energy distribution,has a large uncertainty. An approximate source term input has been derived from measurements andsimulations at the XBOX facility during conditioning for CLIC structures [94]. It is assumed that thedark current can be accelerated over one module consisting of four RF structures to limit the current andthe maximum energy gain. A dark current of 76 nA per module and a continuous X-ray spectrum up to140 MeV was considered in this study. The dark current must be effectively stopped in between modules,e.g. through closed vacuum valves or dephasing of the RF between modules.RF wave guides will connect from the klystrons to the RF structures through ducts roughly every 2 m.The wave guide cross section is relatively small, so the effective duct diameter will be much smaller thanillustrated in the integration drawings. Shielding inserts around the wave guides are required to reducethe radiation streaming through the ducts. 89
Infrastructure and Civil Engineering
Figure 5.39: Dose rate levels for different loss scenarios. From top to bottom: S-Band linac opera-tion, low energy experimental area operation, X-Band linac operation to the high energyexperimental area, X-Band linac RF conditioning, simultaneous beam operation to bothexperimental areas with maximum nominal beam losses.90
Infrastructure and Civil Engineering
The high energy experimental area in TT4/TT61 must remain inaccessible because to the radiationpropagating from the RF structures down the linac tunnel. However, with a modified access scheme atthe end of the linac, e.g. a passage towards TT61 and an additional shielding wall, the area could bemade accessible also during RF condition of the X-Band linac.Figure 5.40 shows the expected dose rate in the klystron gallery during RF conditioning in the X-Bandlinac with the assumed parameters.Figure 5.40: Effective dose equivalent rate from RF conditioning, generating a dark current of 76 nA anda continuous spectrum up to 140 MeV in each of the 24 modules.
Activation.
Induced activity and the resulting remnant radiation levels are usually less critical in elec-tron accelerators compared to proton accelerators. The admissible beam losses in the linac are very muchconstrained by the available shielding against stray radiation. The permanent beam losses hence remainrelatively low. The radiation levels after 180 days of operation with maximum admissible nominal beamlosses along the linac remain in the order of 10 µSv h − after 1 day of cool-down, Fig. 5.41.The exposure of workers to activated air intervening directly in the accelerator after beam stop hasbeen evaluated as well and is insignificant. Reference inhalation doses for a one hour stay after beamstop are negligible well below 1 µSv. Environmental Impact.
Accelerator installations may generate a direct radiological impact on the en-vironment from exposure to stray radiation during operation and releases through air and water. Theexposure path through water has not yet been studied due to lack of detailed implementation of accel-erator components. It is however considered as negligible regarding the activation potential of the linacin its planned configuration. The stray radiation impact is negligible considering the existing shield-ing of TT5 and TT4. The main environmental impact from the linac would be induced by releases ofradioactivity in the air.The production rates for radioisotopes in the accelerator air volume were calculated for a nominaloperation scenario. The ventilation scheme is considered to operate mainly in a recycling mode with apermanent air extraction flow rate of 200 m / h during beam operation and 400 m / h during access mode.The production rate, released activity, dose conversion coefficients and effective dose to a referencegroup for each relevant radioisotope are detailed in [87]. The annual committed effective dose wouldarise to about 13 nSv. In case of a non-recycling mode, the effective dose would increase by about a factor10. These dose levels are acceptable from an environmental impact perspective and are insignificant intheir contribution to the overall impact by the installations on the CERN Meyrin Site. Radiation Monitoring System.
A monitoring system measuring the ambient equivalent dose rates inaccessible areas is needed to protect against high radiation levels in case of unrestricted beam losses in91
Infrastructure and Civil Engineering
Figure 5.41: Effective dose equivalent rate from induced activity after 1 day of cool-down following 180days of beam operation at a 2% distributed beam loss rate at nominal intensity in the S-Band(1 . × e − / s) and X-Band (1 . × e − / s) linac.the various parts of the accelerator. A rather dense network of detectors is required to cover the differentloss scenarios and operation modes of the accelerator.The shielding was designed to limit doses from full beam losses, but dose rates may exceed admissiblelimits by up to a factor 100. The mixed field monitors and X-ray monitors need to be interlocked withthe accelerator. A number of detectors is required to monitor the radiation levels from induced activityalong the accelerator.Figure 5.42 illustrates the proposed positions of different monitor types for X-ray, gamma and neutronradiation in TT5 and TT4. Further studies.
In a more detailed technical design of the eSPS project, further optimisation of theshielding design can be achieved. As the design of the accelerator is more refined and more is knownabout beam loss terms, the shielding design can be adapted, both in terms of volume and shieldingmaterials. • The magnet design in the return chicane to deliver beam to the low energy experimental area mayhave a positive impact to optimise the lateral shielding; • As the beam dump designs will be developed, dump shielding and integration will potentiallyhave an impact on the surrounding shielding wall. The current design is considered as sufficientlyconservative such that the overall integration layout will not be jeopardised; • A shielding separation to be introduced between the linac and the TT4/TT61 area will allow access92
Infrastructure and Civil Engineering
Figure 5.42: Proposed locations of radiation monitors: Mixed field monitors (red), induced activity mon-itors (orange), X-ray monitors (green).to the latter area in case of RF conditioning operation. The feasibility in terms of safety, ventilationand operation modes has to be studied; • More detailed knowledge about beamline equipment between the X-Band linac and the beam dumpwill allow to tailor the shielding wall towards n_TOF. More detailed technical solutions have to bestudied and implemented to allow the passage of cable trays and ventilation through this shieldingwall; • The location of the separation door between the SPS and the linac area in TT61 has to be defined.This must be based on more detailed studies on the radiation scattering down TT61 and upstreamfrom possible stray radiation from protons extracted from the SPS towards the HiRadMat installa-tion and TI2;
Currently the TT4 tunnel is equipped with an electrical overhead trav-elling (EOT) crane CRPR-00173, shown in Fig. 5.43, that was installed in 1975 with a safe working load(SWL) of 15 t. Figure 5.43: Current EOT crane installed in TT4.This EOT crane needs replacing with a new EOT crane with a reduced SWL of 10 t but an increasedcoverage zone and lifting height in order to install the eSPS machine components in the TT4 tunnel. The93
Infrastructure and Civil Engineering
Table 5.14: New TT4 EOT crane characteristics.
SWL Span Lifting height Power
10 t 7500 mm 4500 mm 15 kWFigure 5.44: A 3D view of the new EOT crane for TT4.94
Infrastructure and Civil Engineering design of the new EOT crane is shown in Fig. 5.44 and its characteristics are given in Table 5.14.
TT5 Handling Infrastructure.
Currently the TT5 tunnel is equipped with an EOT crane CRPR-00144that was installed in 1971 with a SWL of 20 t, as shown in Fig. 5.45. This obsolete EOT crane needs tobe replaced by a new EOT crane with the characteristics shown in the Table 5.15.Figure 5.45: Current EOT crane in TT5.Table 5.15: Current TT5 EOT crane characteristics.
Lifting capacity Span Lifting height Power
20 t 15300 mm 5000 mm 35 kWThe overall dimensions of the new EOT crane will remain roughly the same as the existing one. Thetwo cranes in TT4 and TT5 overlap in order to transfer equipment’s at the junction of the two tunnels, asshown in red in Fig. 5.46.
TT4 Installation.
The delivery of eSPS machine components for installation in TT4 takes place byaccessing via the TTA4 access gallery shown in Fig. 5.47. The new EOT crane will pick up the eSPSmachine components at the junction of the TTA4 access gallery and the TT4 tunnel and then will liftthem to their final installation position.The imposed strategy for the installation of the modulators is to install them starting from the junctionbetween TT4 and TT5, and working towards TT61. This is due to the limited headroom for the passageof modulators on top of each other, which is not possible. It is not foreseen to change the completemodulator unit for maintenance or repair purposes. The SWL and the extended coverage zone of the newEOT crane will allow installing all other eSPS machine components in TT4. Special lifting beams will berequired to optimise the height clearance during lifting operations for the installation of the modulatorsand accelerator components as shown in Fig. 5.48.
TT5 Installation.
The access to TT5 via B183 is limited due to the massive concrete beam that reducesthe clearance to 4.5 m, Fig. 5.49. There is no direct crane access to TT5 via B183. All the eSPS machinecomponents needs to be transferred on a trailer via B183 to TT5. The SWL and coverage zone of the newEOT crane in TT5 will then allow installation of all the eSPS machine components. As in TT4, speciallifting beams will also be required in TT5 to optimise the height clearance during lifting operations forthe installation of the modulators and accelerators components.95
Infrastructure and Civil Engineering
Figure 5.46: Crane overlap region between the EOT cranes in TT4 and TT5.Figure 5.47: Access gallery TTA4 entering TT4.The current access routes, via B183 and via the TT4A access gallery, Fig. 5.50, are suitable to deliverequipment into both TT4 and TT5.
The following requirements were outlined and implemented (as set out in Secs. 5.1.6 and5.2.3) for the TT4 & TT5 sectors:
Access and Egress -
A push-and-pull ventilation system is adopted the klystron and acceleratorcompartments. There is fire fighter access from two tunnels, upstream of the fire, into the fresh air. Dueto the restricted cross-sectional geometry, the access on the side of the modulators close to the shielding96
Infrastructure and Civil Engineering
Figure 5.48: Example of cavity handling in TT4.Figure 5.49: Height restriction to enter TT5 via B183.97
Infrastructure and Civil Engineering
Figure 5.50: Access points in TT4 and TT5 used to deliver equipment.98
Infrastructure and Civil Engineering wall is only 600 mm. This is deemed acceptable as any casualties can be extracted into the transportaccess way for evacuation. For the TT4 and TT5 areas, it is additionally important that fire equipmentcan be freely moved through the "chicanes" without any problem, especially for a rescue operation. Thishas been discussed with CERN’s Fire and Rescue Service.
Compartmentalisation -
Two fire compartments are foreseen, one fully containing the linac, andone fully containing the klystrons (as shown in Fig. 5.34). As shown in Fig. 5.51, smoke-proof sectordoors (rated to EI90 or greater) will be required at the following points (normally open for the air flow): • The entrance to the linac compartment; • The entrance to the Klystron compartment.A 2 m thick dividing wall between B183 and TT5 will be constructed to provide compartmentalisationfor fire safety and as additional radiation shielding to equipment located in B183. The entrance door toTT5 via B183 shall be rated to EI120. Foam shall be used to prevent smoke leakage from TT5 betweenthe concrete blocks. This will need to be replaced when the blocks are removed.Figure 5.51: Safety and access elements of the linac, adjacent areas and CLEARER experiment.
Fixed Suppression Means -
A fixed fire suppression system is not necessary for life safety; theproject has opted not to install such a system for the protection of the equipment within the fire compart-ment through discussion of the costs and benefits. The cost estimation made for this has, therefore, beende-scoped from the CDR cost estimation.A dry riser/standpipe is foreseen for the full lengths of the klystron gallery, with outlets every 20 m,as shown in Fig. 5.34. This is the area where the fuel load is the most significant. The flow rate forfire attack shall be 60 m per hour. There shall be an outlet very close to the entry point of the linac forimmediate use by the CERN Fire and Rescue Service. In case of fire, valves can be opened at the surfacein order to provide the desired flow. The full requirements for this equipment are summarised in [95]. Smoke Extraction -
Hot smoke extraction is foreseen, allowing the CERN Fire and Rescue Serviceto employ this tool in line with their strategy to tackle a fire. During smoke extraction, fresh air issupplied to the fire compartments, displayed in Fig. 5.37 without any treatment. Radiation Protectionstudies show that this smoke can be released to the atmosphere safely without control or filtration.99
Infrastructure and Civil Engineering
Electrical Safety.
The HV oil tank will be in place to insulate the klystrons and modulators, with pulsequality monitoring to identify any breakdown in the oil. NF C 18-510 compliant covers, and restrictionof access to those with the appropriate electrical training and certification shall also be in place.At the conceptual stage of design, it is known that the existing electrical infrastructure in TT4 and TT5is aged, and will require rejuvenation work to bring it into an acceptable level of conformity suitable forthe project.
Laser Safety.
B183 will house a pulsed UV laser, used to generate the required electron bunches. Thiswill be located in a controlled access laser room, and the safety system will therefore need to meet theappropriate IEC required control measures: • IEC 60825-1: Safety of laser products - Part 1: Equipment classification and requirements; • IEC/TR 60825-14: Safety of laser products – Part 14: A user’s guide.
Noise Safety.
Noise generated at CERN shall comply with the safety requirements provided in thefollowing Safety rules: • GSI-SH-4: Protection of workers against noise; • SG-SH-4-0-1: Noise at the workplace.Emissions of environmental noise related to neighbourhoods at CERN, on Swiss or French territory,shall comply with the requirements provided in the following safety rule: • Noise footprint reduction policy and implementation strategy [90].As the cooling and ventilation systems will use a pre-existing cooling tower, the increase in noise levelsat the CERN site as a result of the additional cooling load is expected to be very low. An additional chillerwill be installed outside for cooling the air for the beam injection, which based on the current cooling andventilation specification is expected to be approximately 150 kW, with only one chiller in operation atany time. An initial simulation of the noise impact predicts a total increase in noise at the CERN borderonto the Route de Meyrin of 1 dBA, which is expected to be acceptable; this will need to be confirmedas the project moves into detailed design.
Chemical Safety.
The most significant hazard identified in the chemical domain is the large quantityof oil in the HV tanks for the klystron and modulator assemblies. There will be 800-1000 litres of oil permodulator assembly. This is currently expected to be a highly refined mineral oil, capable of withstandingthe high voltages. The Safety Data Sheet (SDS) must be obtained from the supplier, a copy of whichmust be available for consultation by users. The oil must be registered in the CERN Chemical database,CERES, with the quantity, location and any hazards recorded. A copy of the up-to-date SDS must beuploaded into the database and a Chemical risk assessment performed using CERES, if appropriate. Themitigation strategies will consist of a removable retention basin for each HV tank, with sufficient capacityfor the entire quantity of oil stored in the tank. Coupled with this, an oil level detector shall be used toindicate any drop in level. As a further measure, consideration shall be given to the potential leak pathshould oil be spilled outside of the retention basins. Appropriate procedures should be put in place toprevent spillage during filling and draining of the tanks.
Environmental Safety.
Infrastructure and Civil Engineering
Greenhouse Gases -
Accelerating Klystron systems attached to waveguides can contain the fluor-inated greenhouse gas SF6. The replacement with SF6 by dry air, vacuum or any other alternative shall beconsidered during the detailed design phase. If not possible, the installation of leak control systems and arecovery system for any maintenance operation shall be foreseen and personnel involved in the activitiesshall be trained and certified according to the regulation in force at the time of the construction/operationof the project.
According to the RP assessment, the radiation levels of the electron beam in thelinac tunnel and in its adjacent technical galleries, as well as the CLEARER experimental area, forbidthe presence of people during beam operation. Therefore, new safety chains dedicated to the linac andto CLEARER will be integrated and managed by the SPS Personnel Protection System currently underrenovation.The access to the ‘electron linac’ zone will be done through one Personnel Access Device. Duringbeam operation, the personnel access will be forbidden in the two TT4 technical underground galleriesand the Klystron area, which will be closed by new end-of-zone doors. As there is no access point toaccess to these sectors, a patrol procedure should be initiated before passing the zone into beam mode ifone of the end-of-zone door delimiting the sector has been opened.To ensure the protection of the users accessing the zone from radiation hazard, it is considered to acton the following Important Safety Elements (EIS-b): • Two Beam Stoppers able to stop the electrons coming from the electron gun; • The S-band and X-band RF preventing beam acceleration.In case of intrusion during operation with beam or if the safe position of a safety element is lost duringaccess, a safety interlock will act on the electron gun. As it is foreseen to use lasers inside the linactunnel, the laser lines should be equipped with shutters, whose positions should be monitored by thePPS, thus protecting the personnel from the lasers hazards.If needed, additional local safety functions could be considered such as an ‘RF Test’ mode or a ‘LaserAlignment’ mode.
CLEARER Experimental Area.
The CLEARER experimental area will be accessible through an ac-cess point made of one personnel access device. Aside the access point an "End-of-Zone" door will beinstalled to allow the exit of people in case of emergency.To protect people accessing the zone from radiation hazard, one bending magnet and two beam stop-pers will be used as EIS-b.In case of intrusion during beam operation or if the safe position of a safety element is lost duringaccess, the safety chain will act on the upstream zone: the electron linac.
TT5 Hall.
In order to control the access to the TT5 hall and to the buffer zone, the three doors inthe following figure will be equipped with an electrical lock controlled by a dosimeter badge reader, asshown in Fig. 5.52.
The installation of the linac in TT5 and TT4 has an impact on the current access path to the n_TOFExperimental Area 1 (EAR1) situated at the extension of TT4 and upper part of TT2. Requiring theintegration of the linac dump in TT4, space for accelerator R&D experiments and allowing sufficient101
Infrastructure and Civil Engineering
Figure 5.52: Access controlled doors to the TT5 Hall.clearance for material access towards TT61, no solution was found for a local shielding which wouldhave been compatible with an independent access to EAR1 at the same time.A new potential access path to EAR1 could be provided via ISR T7 and TT3. This access requiresadditional fencing in the ISR, modifications of the access control system, securing the path with barriersand marking, and a new staircase towards EAR1. The access path would be only suitable for personneland light equipment. Heavy material access to EAR1, which is required only a few times per year, couldstill be made through the access gallery 852 and TT4. For that purpose a part of the shielding wall inTT4 must be removable to allow material passage towards EAR1. Typically, such interventions wouldbe scheduled during extended accelerator shutdown periods.The area in B183 adjacent to TT5 is currently used for the storage of radioactive shielding blocks.To free this area for the required installation of the eSPS control and laser room, an alternative indoorstorage location has to be found. It is assumed an area of 1500 m is needed to accommodate displacedmaterial from TT4,TT5 and B183. This is a condition of the project and if such space is not available inthe flex-storage building or similar, then an additional cost will be incurred to re-provide this space. TT61 is a horseshoe shaped tunnel with horizontal invert over the majority of itslength, measuring 4.5 m wide (4 m wide at floor slab level) and 3 m tall at the crown. The tunnel wasconstructed predominately by mining into the Molasse rock. The tunnel’s primary lining consists ofsteel centring at regular spacings with sprayed concrete infill and a secondary lining of cast in situ mass102
Infrastructure and Civil Engineering concrete. A PVC waterproofing layer separates the primary and secondary lining. The tunnel invert isformed with a PVC waterproofing layer below a lean-mix concrete base 100 mm thick, above that thereis a 220 mm thick cast in situ slab with a 50 mm screed forming the invert. A 300 mm diameter drainruns along the length of the tunnel, located centrally below the invert construction. The typical crosssection in the Molasse is shown in Fig. 5.53 with an additional stone and corrugated steel preliminaryliner where non-cohesive ground was found.Figure 5.53: TT61 typical cross section.Constructed in 1975, TT61 was originally part of TT60 until works to enable the LHC upgrade re-quired construction of junction cavern TCC6 to link with the new TI2 injection tunnel, splitting TT60into two parts (now named TT60 and TT61). TT61 is now 488 m long connecting TT4 and TCC6. Thetunnel also has several alcoves dating from the original construction which housed beamline equipment.These may be put back into service to house off-axis experiments but this is to be confirmed. The tunnelhouses some monitoring equipment for the HiRadMat experiment which is located in the adjacent TNtunnel. The monitoring equipment and shielding takes up a substantial portion of the tunnel cross-section.The need to relocate this is discussed later in the chapter.The tunnel was found to be in very poor condition at the latest inspection with significant cracksrunning along the centre of the invert up to 30 mm wide and numerous drainage issues. The invert isalso suffering from ground heave with some areas showing uplift of up to 100 mm. This is likely dueto the failure of tunnel drainage and the fact the tunnel descends through the moraines into the Molasse,forming a water pathway into the normally dry Molasse. The Molasse is known to swell on contact withwater. This, in combination with the flat invert which provides little or no resistance to ground heave,has left the tunnel in quite poor condition as shown in Figs. 5.54 and 5.55.103
Infrastructure and Civil Engineering
Figure 5.54: TT61 typical cracking in invert.Figure 5.55: TT61 area of significant floor heave next to alcove.
Enabling Works.
In order to enable the eSPS project to go ahead, no works would be needed to changethe geometry of TT61 which is suitable for replacement of infrastructure to re-implement a beam transferline.Capital maintenance of the existing tunnel would still be important. Existing drainage systems should,as a minimum, be cleaned, surveyed with cameras to check their condition and repaired as required.Further work should also be carried out to investigate the causes of cracking and floor heave in thetunnel as well as a campaign of monitoring to determine to what extent the condition is deteriorating. Ifthe rate of uplift is continuing this could pose problems for the overall stability of TT61 as well as forbeam alignment during use as a transfer line to the SPS.104
Infrastructure and Civil Engineering
At this stage, maintenance of TT61 has not been studied in detail. An assumption has been madeat this stage that superficial repairs to the invert and minor drainage repairs will be required. This iseffectively a provisional sum and must be re-valuated in full once monitoring and investigation workshave been carried out.
Transfer tunnel TT61 is an underground tunnel that was used to extract the SPS beamline to B180. It isconnected to the TT4 transfer tunnel at one end and the TCC6 junction cavern at the other end.
Connection with Transfer Tunnel TT4.
The eSPS beamline is bent from Transfer Tunnel TT4 intoTransfer Tunnel TT61 by horizontal and vertical BH2 dipole magnets as shown in Fig. 5.56.Figure 5.56: eSPS beamline transfer to the SPS from transfer tunnel TT4.The beamline goes through the wall dividing TT4 and TT61 via an existing hole. This hole will haveto be enlarged as the height of the eSPS beamline is different than that of the hole.
Transfer Tunnel TT61.
Currently in TT61 there is some existing beamline equipment remaining fromwhen the SPS beamline was extracted to B180. TT61 also houses some equipment for the HiRadMatfacility of which includes removable shielding blocks at the TCC6 end of the tunnel. The tunnel is usedby the transport group to access the LHC Injection Tunnel TI2.TT61 will house the SPS transfer beamline from TT4 to TT61. The beamline is approximately 450 mlong and consists of a few BH2 vertical dipoles and quadrupoles. As the tunnel’s cross-section remainsuniform, a model of the typical cross-section was developed as shown in Fig. 5.57.105
Infrastructure and Civil Engineering
The cross-section includes: • Allowance for safety lighting and services; • Transport volume show in green as specified by the transport group; • Width of the dipole; • Allowance for personnel access for maintenance of the cable trays; • Width of the cable trays. Figure 5.57: Cross-section of TT61.The positions of the dipoles and quadrupoles along the length of the tunnel were not studied at thisstage if the project and does not pose any specific issues regarding the overall integration.
Junction Cavern TCC6.
TCC6 is a junction cavern housing the TI2 LHC transfer beamline and theTT66 HiRadMat beamline.TCC6 will house the eSPS SPS transfer beamline, the concept looks at merging this electron line withthe TI2 LHC transfer in TCC6. Due to the low energy of the eSPS electron beam, a dipole is placed at theentrance of TCC6 and the beamline bent towards the TI2 beamline with a removable vacuum chamberallowing the transport to TI2 from TCC6 with minimal obstruction a shown in Fig. 5.58 and Fig. 5.59.Within this section of eSPS beamline in TT66, quadrupoles and beam diagnostic equipment will beplaced, its exact location is to be determined at the detailed design stage. Detailed integration studies ofthis area have not been undertaken as there has not been any integration studies of this area for severalyears. New scans of the area should be undertaken in the detailed stage of the project.106
Infrastructure and Civil Engineering
Figure 5.58: Plan view of the integration layout of TCC6.Figure 5.59: Integration layout of TCC6.
The existing cooling and ventilation infrastructure for TT61 is not modified because there are no heatdissipating equipment added to this tunnel. Some magnets are installed in TT4/TT61, as displayed in Fig.5.60; however, they dissipate only 0.5 kW of heat to air. Hence, the existing flow that ventilates TT61 issufficient to deal with this load. These magnets are also water cooled, dissipating around 9.5 kW to thedemineralised water infrastructure dedicated to the klystron compartment, as displayed in Fig. 5.29.Figure 5.60: Magnets in TT4/TT61.107
Infrastructure and Civil Engineering
The interface between the linac and the SPS is situated in TT61. Access until this inter-machine doorshall be possible from both sides while beam operation is ongoing on the respective other side. The exactlocation of the inter-machine separation door is depending on the potential beam loss scenarios in TCC6by the LHC or HiRadMat beams and the radiation coming from electron beam losses in the Linac inTT4.During electron beam injection from TT61 via TCC6 and TT60 into the SPS at TS65, the adjacentareas to TCC6, namely the access to the HiRadMat facility via BA7 through TJ7, TA7 and TNC must beprohibited. The same applies to the upper part of TI2 and the lower part of TT70. Radiation levels fromthe high energy electron beam in these areas are prohibitive for the presence of persons.A more detailed radiological impact study of beam loss scenarios impacting the areas in vicinity ofthe beam transfer tunnel from TT4 to the SPS is required during the technical design phase.The radiological impact of the high energy electron beam in the SPS has not been addressed in detailfor this report. It is expected that the impact in terms of activation is negligible compared to the ongoingSPS proton operation with much more intense and radiologically relevant beams.In case a dedicated internal electron beam dump needs to be implemented in the SPS, an appropriateradiological risk assessment shall be included in the technical design study. Similarly, all modificationworks in the SPS will have a radiological impact due to residual radiation levels and must be properlyplanned and optimised.
To access TT61, equipment needs to be delivered through the access gallery TTA4 (804-U0-201, 804-U0-202). The transport volume in TT4 has been studied to allow the passage of the largest convoy forthe installation of the magnets in TT61 and TCC6. This path, shown in Fig. 5.61, should remain free asit is the only path for the exchange of magnets for the transfer lines from SPS to TI2 (LHC injection)and from SPS to HiRadMat. For the installation of magnets in TT61 (heaviest magnet weight 1800 kg) aforklift equipped with a lifting jib with a capacity of 2 t is required. The transport volume design takes inconsideration the largest required transport and handling equipment (special handling machine Dumontand tow tractors) for the maintenance of all transfer lines located in TCC6 (HiRadMat and TI2).Figure 5.61: 3d view of the path to enter TT61 with handling equipment (green volume)108
Infrastructure and Civil Engineering
In the SPS accelerator, the following new components needs to be installed: • two kickers in the SPS LSS6; • one electrostatic septum and one thin magnetic septum in the SPS LSS1.Those components will be installed with an existing custom-built forklift side loader (called PRATT)using the same procedure as for similar equipment installed in the SPS. TT61, as a pre-existing tunnel without an active accelerator or experiment, has not yethad a new fire safety study carried out on it; a full fire safety study will, however, need to be to bemade as the project moves into detailed design. The new installation in TT61 will also interface withageing facilities: TCC6, T60 and BA7 and the associated infrastructure. The performance based firesafety design approach required for these areas shall have coherence with the SPS Fire Safety Study. Itis not yet clear whether the implementation of the fire safety requirements for TT61 would fall underthe scope of the eSPS project, or whether it would form part of a separate CERN consolidation project.The cost estimation made for this work (including the dismantling of the old fire safety installation) has,therefore, been de-scoped from the CDR cost estimation, but a study will need to be done, and an updatedcost estimation made, at the detailed design stage.
Superconducting Cavities.
The proposed installation of the superconducting cavities will use the pre-existing SPS crab cavities test stand in LSS6. This test stand was subject to checks by CERN’s HSE Unitat its commissioning, and is, therefore, deemed to be compliant without any significant modifications.This shall be reviewed as the project moves into detailed design.
Figure 5.62: Safety and access elements of the TT4 and TT61.According to the RP assessment, radiation levels from the high-energy electron beam in the TT4experimental area and the upper part of the TT61 tunnel forbid the presence of personnel during beamoperation. Thus, a new safety chain dedicated to the TT4 and the TT61 will be integrated and managedby the SPS access safety system.To control the access to the TT4 and TT61 tunnels, the current TT61 access point, composed of a PADand a MAD, will be moved at the entry of the gallery 852. A new “sector” door will be installed betweenthe TT4 and TT61 tunnels. Moreover, three new “inter-zone” doors will be added to delimit these sectorsas shown in Fig. 5.62. 109
Infrastructure and Civil Engineering
To protect people accessing the zone from radiation hazard, one bending magnet and two beam stop-pers will be used as EIS-beam as shown in Fig. 5.63. The magnetic element being not fail-safe, it wouldbe beneficial to review the EIS of this beamline in a future stage of the project.Figure 5.63: Safety elements of the TT4 and TT61.Moreover, in case of intrusion during beam operation or if the safe position of a safety element is lostduring access, the safety chain will act on the upstream zone: the electron linac.The zoning of the current TT61 has to be adapted due to the removal of TT61 access point. Theexact location of the “inter-zone” door inside TT61 will be defined according to RP calculations anddepending on the potential beam loss scenarios in TCC6 from the LHC or HiRadMat beams and theradiation coming from electron beam losses in the TT4 linac.In order to protect people against radiation hazards coming from the injection of the electron beamfrom the eSPS injection into the SPS ring, new EIS-beam must added into the ‘West Extraction’ and the‘SPS Ring’ interlock chains. To do that, it is foreseen to use one bending magnet and two beam stoppers,located just before the TT61 tunnel as shown in Fig. 5.64.Figure 5.64: Safety elements to avoid high energy electron beam injection into the SPS.110
Infrastructure and Civil Engineering
A new extraction tunnel will be required to implement eSPS, located on the CERN Meyrin site betweenthe AD building (B193) and the magnet assembly facility in B181 as shown in Fig. 5.65.Figure 5.65: Birdseye visualisation of the proposed facilities in red showing surface and buried struc-tures.
Existing Site and Infrastructure.
The site is currently characterised by a large earth mound around7 m above surrounding road levels with a covering of grass, shrubbery and trees. The proposed site isclose to B181 and B193 (as well as its extensions and service Buildings: B93, B8854, B393) along withTunnels TT2 and TT7. The site intersects the existing TTL2 tunnel, also known as ’The Ear’ due to itsdistinctive shape as shown in Fig. 5.66.The civil engineering works interface with the existing TT2 and TTL2 tunnels. To enable the conceptdesign, a recent 3D scan of the area was used to update the models of the existing infrastructure to ensureit provides suitably accurate base for the design.The existing TT2 and TTL2 tunnels are both in good condition with no major defects. TT2 wasconstructed in 1969 as a transfer tunnel between the PS and the West area and has remained in use eversince. TTL2 was subsequently constructed in 1980 to allow the beam to be turned around in a loop andhas been recently refurbished.The planned works are also very close to TT7 which formerly housed a beam dump with steel shieldingimmediately beyond the tunnel end as shown in Fig. 5.67. This has also been modelled to ensure it canbe taken into account in the planning of works.
Civil Engineering Layout.
This section covers the civil engineering required to enable the extractionfrom the SPS to a new experimental area.The new tunnel, approximately 55 m long will form a link to deliver beam between the existing TT2tunnel and the new detector underground hall as shown in Fig. 5.66.A beamline core will be drilled through the wall of the existing TT2 tunnel, allowing the beam to passwithin a vacuum tube. 111
Infrastructure and Civil Engineering
Figure 5.66: 3D visualisation of the underground infrastructure showing the new infrastructure in red.Figure 5.67: Plan view of TT7 and steel shielding block arrangement.Connections will be formed (as shown in Fig. 5.68) with the existing TTL2 tunnel.The tunnel has been designed at a constant 3.57 m wide internal cross section over the first 36 mfollowed by a tapered section of 19 m length widened to 9.09 m at the connection with the building toallow beam delivery to two separate experiments as shown in Fig. 5.69. The tunnel is effectively flat,with no slope over the full length to accommodate beamline requirements.One section will consist of a horseshoe cross-section while the ’cut and cover’ section will be rectan-gular, both 3.75 m wide an the minimum as shown in Fig. 5.71.The horseshoe shaped mined tunnel section will be constructed using a sprayed concrete primarylining around the arch incorporating rock bolts followed by a cast in situ mass concrete secondary lining.In this case, the curved invert slab is 600 mm thick at the minimum and 950 mm at the maximum caston a 100 mm blinding of lean-mix concrete. Drains would be included at the base of the arch with freedraining material surrounding a perforated drainage pipe located on a concrete plinth.The cut and cover rectangular tunnel section will be reinforced cast in situ concrete comprising aninvert slab 300 mm thick, walls 400 mm thick and a roof slab varying from 300 mm at the edge to 350 mmdeep at the centre. The floor slab is haunch thickened to 450 mm deep directly below walls and toaccommodate shear forces efficiently, the joints between walls and roof slab have additional 250 by250 mm haunches. 112
Infrastructure and Civil Engineering
Figure 5.68: 3D visualisation cut at tunnel invert level showing new extraction tunnel layout in relationto existing infrastructure.Figure 5.69: Plan view showing geometry of proposed CE layout and existing infrastructure.Along the full length of the tunnel, a tunnel waterproofing system will be provided to prevent wateringress. 113
Infrastructure and Civil Engineering
Figure 5.70: Section 7-7 showing extraction tunnel abutting TT2, junction with TTL2 and flat profilethrough to its connection with the detector underground hall.A drain will run along the centre of the tunnel, with a fall of at least 1% to ensure water flows withoutcreating blockages due to sediment build-up. The fall will be artificially created in the drain alone, withthe tunnel invert remaining horizontal as shown in Fig. 5.70. The drain will connect into the existingtunnel drain in TTL2. Therefore, the section of extraction tunnel between TT2 and TTL2 will drain tothe south while the remainder of the extraction tunnel will drain to the north.Figure 5.71: Typical Cross-sections of the tunnel showing the the cut and cover section (left) and themined section (right).The tunnel width will allow sufficient space to accommodate the machine, services and suitable spacereservations for safe transport and maintenance.
Civil Engineering Design Considerations.
Tunnel cross-sections have been based on recent exper-iences and designs used for the High-luminosity LHC upgrade which is located relatively nearby at theLHC Point 1. The cross section design will need to be reviewed following detailed ground investigationand adjusted in the case ground conditions vary significantly from expectation.The new extraction tunnel will be constructed to abut TT2 without forming a junction between tunnels.The exact structural form here will need to be determined following assessment of the structural capacityof TT2. The most likely options are either a free standing tunnel with significant reinforcement in thetunnel lining to avoid imposing any additional loads on TT2; or some form of heavy duty structuralconnection between the crown of the new tunnel and TT2 to allow some transfer of loads, supporting thenew tunnel.Connections between TTL2 and the new extraction tunnel are necessary for several reasons : • To facilitate connections with electrical services; • To allow ventilation of the tunnels and avoid any ’dead leg’ area without sufficient air movement; • For transportation of some large magnets via TTL2 transport shaft, the operation of the ear needsto be preserved; 114
Infrastructure and Civil Engineering • Although not a requirement, in line with best practice, this also allows an alternative means ofemergency access and egress via the TTL2 transport shaft.The tunnel will be constructed entirely within the Molasse. To enable the CE tunnelling works, sig-nificant earthworks must first be carried out to clear the site down to a construction platform level. Thishas been set at a suitable level to allow the finished works to tie into the surrounding road levels to enableaccess to the site.The section excavated as part of the cut and cover tunnel will be back-filled with excavated materialto reduce the amount of spoil which must be taken off site as shown in Fig. 5.73. During excavation inthe Molasse, both for the tunnel and experimental hall, rock bolts and shotcrete sprayed concrete will beused to provide temporary excavation face support to maximise the angle of excavation and minimise thevolume of material. An indicative detail is shown in Fig. 5.72.Figure 5.72: Indicative Molasse face during excavation.Figure 5.73: Section 4-4 through cut and cover tunnel showing likely excavation profile and backfill.115
Infrastructure and Civil Engineering
Transfer tunnel TT10 is an 800 m long tunnel that houses the TT10 beamlinethat runs from transfer tunnel TT2 to the SPS. The eSPS beamline is extracted from the SPS via TT10.There are no integration changes in the tunnel for the eSPS project.
Transfer Tunnel TT2.
Transfer tunnel TT2 is a 400 m long tunnel that houses the FT16, FTA, TT10,FTN and FT12 beamlines. The eSPS beamline crosses through the tunnel from TT10 into the newextraction tunnel at an angle of approximately 20 ◦ as shown in Fig. 5.74.Figure 5.74: Integration layout of TT2.In TT2, approximately 10.5 m of the TT10 beamline will be modified of which includes removingsome beamline equipment and replacing it with three MCB dipoles. An additional MCW dipole willbe installed in the TT2 transport zone for which the distance between the wall and the magnet is 2.2 m,allowing enough space for transport vehicles to pass by. The removable vacuum chamber will passfrom the MCW dipole into the extraction tunnel via a hole in the TT2 wall. The vacuum system willbe modified such that a new four-way separation chamber will be installed as well as additional sectorvalves.The MCB and MCW dipoles will be installed via the TT2 access shaft. This shaft is currently usedto install the existing MCB magnets. As the MCW dipoles are longer, a transport integration study wasundertaken to ensure that the magnets could be installed (see Fig. 5.76). Extraction Tunnel.
The extraction tunnel (shown in Fig. 5.75(a)) is a new tunnel that houses the eSPSbeamline and its corresponding services. It allows the beamline to deviate away from the existing beam-lines in TT2 to the experimental hall. The beamline services include cooling and ventilation, cable trays,and lighting, among other things.During technical stops and long shutdowns of the beamline, personnel and transport vehicles mayaccess the extraction tunnel for maintenance purposes, using the transport access way located on theright-hand side of the tunnel (with respect to the beam direction). Personnel and transport access into theextraction tunnel is via the experimental hall connected to the downstream of the extraction tunnel.The tunnel has a length of approximately 60 m and is shaped with an internal height of about 3 mand a width of about 4 m. The width and height of the tunnel are based on the dimensions required for116
Infrastructure and Civil Engineering (a) (b)
Figure 5.75: 3D layout (a) and cross-section drawing (b) of the extraction tunnel.equipment and personnel/transport access, as shown in Fig. 5.75(b): • • •
500 mm allowance for personnel access for maintenance; •
300 mm allowance for cable trays; •
50 mm allowance for cabling and the cable tray support structure.
Analogue to the injection of protons from TT2 via TT10 into the SPS, the extraction of electrons in theopposite direction requires the transfer tunnels to be in a safe state. Extraction must be allowed only ifthe receiving side, the experimental cavern, is closed and ready to receive beam. The safety elementsallowing access to the experimental cavern must be located sufficiently far away in TT2 to limit anaccidental exposure of persons in the experimental cavern in case of a faulty ejection from the SPS.The failure modes of extracting the electron beam accidentally via TT2 towards the PS must be studiedand, if required, mitigation measures must be taken to prevent such an event. The location of the inter-machine door in TT10 must be confirmed to be compatible with electrons circulating and dumped in theSPS, as access is possible from the TT2 side, while the SPS is operating with electrons.The impact on the n_TOF target area (TT2A) in case of exceptional beam losses in the upper part ofTT10 or in TT2 must be studied. However, considering the configuration, beam directions and installedshielding, it is not expected that the n_TOF target area will be impacted by the electron beam extractionthrough TT10. 117
Infrastructure and Civil Engineering
At the junction region between TT10 and TT2 two magnets need to be removedand four new magnets installed. All of these magnets enter the tunnel with the help of a mobile craneoperating via the shaft located next to B193, a simulation of this procedure is shown in Fig. 5.76. Theinstallation of the magnets on the beamline will be done with existing motorised bogies (Tortue).Figure 5.76: 3D simulation of one of the new magnets being lowered in the shaft.
Transfer Line TT2 Experimental Area.
The magnets installed in this new tunnel will enter via theexperimental hall by means of the existing EOT crane with 25 t SWL of the building. The installation ofthe magnets on the beamline will either be done using existing forklifts with lifting jibs or with existingmotorised bogies (Kouba).
To allow the extraction of electron beams from SPS ring via TT10 when the TT2 chain is ready forbeam, one bending magnet and two beam stoppers must be included as safety elements (EIS-b) on theTT2 safety chain, as shown in Fig. 5.77.In addition, the new transfer tunnel to the eSPS experimental area must be included in the sector S2 ofthe TT2 personnel protection system. An "inter-zone" door will be installed between this transfer tunneland the eSPS experimental area to close this sector, as shown in Fig. 5.78.118
Infrastructure and Civil Engineering
Figure 5.77: Safety elements to protect TT2 from electron beam extraction via TT10.Figure 5.78: Sector S2 with transfer tunnel to the eSPS exp. area included.119
Infrastructure and Civil Engineering
The CE study has covered the creation of the space required for a detector underground hall, surfacebuilding and associated drainage, access and parking which are detailed here.Following on from the optioneering exercise carried out at the feasibility stage of development, thechosen option has been further developed to create a full concept design.The study is based on that the experimental area should host two experiments. While one of them,the missing momentum experiment, is approximately known [2] with the physics potential described inSec. 2, not much information is available at this stage about the design of a second possible experiment.To model the most demanding scenario, this is assumed to be a beam dump experiment briefly discussedin [3] Sec. III.B.Figure 5.79: View of the proposed site looking north towards the existing earth mound.
Existing Site and Infrastructure.
The existing site is effectively the same as the extraction tunnel. Thesite is again close to the existing buildings described in Sec. 5.4.1. When discussing the experimentalhall specifically, the site is dominated by the large mound shown in Fig. 5.79 which is believed to be spoilfrom the construction of B181 and the ISR. The composition of the mound is not currently known andwill need to be confirmed. This mound is also known to be home to a number of orchid species whichare protected.
Civil Engineering Layout.
The proposed layout consists of a surface hall, underground detector halland associated drainage, parking and access.All infrastructure has been designed and sized based on discussions with eSPS and the missing mo-mentum experiment (LDMX) project leaders. A full 3D-model has been created in conjunction withCERN’s integration team. Space reservations have been estimated and agreed between parties wheredetails were not known. The layout of the building has been developed to accommodate sufficient spacefor installation, assembly, operation and maintenance of the two experiments and incoming beam lineequipment.The underground detector hall dimensions are 11.5 m wide by 14.5 m along the beam axis (internally),12.74 m tall at the maximum pit depth but 11.47 m to the extraction beam tunnel invert level with thearrangement illustrated in Fig. 5.81. Adjacent to the hall is a shaft housing the stairs and lift to accessthe underground area. The shaft measures 4.6 m by 4.95 m (internally) and extends from the upper120
Infrastructure and Civil Engineering
Figure 5.80: Plan layout of the experimental building at surface hall level (left) and detector undergroundhall level (right).Figure 5.81: Sections through Experimental surface building, detector pit and tunnel along beam axis(left) and transverse to it (right).underground detector hall floor level to the surface hall as shown in Fig. 5.80. The functional layout ofthe surface hall is discussed in detail in the integration section so is not covered here.The underground detector hall is formed from 1 m thick reinforced concrete walls with a 1 m floor slabcast onto a concrete blinding. A full width opening is provided to accommodate access to the extraction121
Infrastructure and Civil Engineering tunnel. The upper portion of the floor slab is at the same level as the extraction tunnel invert while thereis a lowered section which is a further 1.27 m deep to allow for the depth of the detectors, associatedinfrastructure and support structures.Figure 5.82: 3D cut-away view of stair and lift shaft.The lift shaft provisionally comprises 0.4 m thick reinforced concrete walls with an internal dividingwall further separating the lift shaft as shown in Fig. 5.82. Standard dimensions are used for stair treads,risers and landings, conforming to safety requirements. The lift shaft is designed based on geometry toaccommodate lifts used elsewhere around the CERN site for consistency.The building above ground will be constructed as a steel portal frame structure with exterior claddingfor insulation and weather-tightness. The Fig. 5.83 shows views of the building with and without claddingto reveal the preliminary steel frame design.Figure 5.83: 3D renderings of the experimental area surface building showing isometric view with clad-ding and access doors (left) and cut away view with underlying steelwork structure (right).The equipment shaft used for transport access to the underground area is covered with 1 m thickpre-stressed, pre-cast reinforced concrete beams, which span the opening to provide radiation shieldingduring beam operation. Beams are supported on a recessed plinth 0.5 m wide on either side of theopening. Beams therefore sit at floor slab level when in position. When removed for access the buildinglayout allows sufficient space to stack the beams to the side of the shaft opening. A removable socketed122
Infrastructure and Civil Engineering
Figure 5.84: 3D section view through experimental surface hall, detector pit and tunnel.1.1 m tall guardrail will be installed around the opening for safety while the shaft is open. The beamsare handled by the 25 t crane within the building which is also used for transport of heavy equipmentwithin the surface hall and underground areas. Fig. 5.84 shows the arrangement of shaft, beams andcrane support structure.Figure 5.85: Section 2-2 through building showing parking, access and delivery area sloping down toexisting road level. Dashed outline indicates the existing mound to be removed.The northern most portion of the hall floor will have an additional 400 mm high false floor to allowfor distribution of services. A technical gallery will be constructed just below slab level to carry servicesfrom the electrical and CV equipment on the false floor to the underground area via the stair and lift shaft.The technical gallery will be recessed in the floor slab: 1 m deep, 2 m wide with removable concrete slabs,400 mm thick as covers during normal operation.An additional technical gallery to bring services to the hall may be required but at this stage, nothinghas been confirmed. An allowance is made for costing purposes but this is a provisional sum and willneed to be re-costed once confirmed.The existing earth mound shown in Fig. 5.79 will need to be removed to enable the construction of123
Infrastructure and Civil Engineering the building and to allow the finished site levels to tie in with surrounding infrastructure. It is assumedthe levels of existing surrounding roads cannot be adjusted since they serve existing buildings and sothese have effectively set the floor slab level. Outside the building, a parking and access area has beenallowed for, although the space has not been further defined at this stage (see Fig. 5.85). The amountof earth to be removed has been minimised based on leaving slopes of 1 in 3 (as shown in Fig. 5.70)for the remaining mound. This could be further optimised following ground investigation by increasingthe slope to the characteristic friction angle of the material when confirmed, or by introducing retainingstructures to increase slope angles, for example soil nailed slopes or mechanically stabilised earth walls.As part of the construction, a significant quantity of earthworks will be required in order to reduce theexisting ground levels. A permanent storage site will be required for the spoil. Ideally this should bewithin CERN land and as close as possible to the existing site although a site has not been chosen at thisstage. The composition of the earth mound will need be confirmed as soon as possible and any additionalcost implications and constraints identified.
Construction Methods.
An indicative construction sequence covering the extraction tunnel and ex-perimental area has been envisaged as part of the cost estimation and in order to consider the timescalesnecessary to deliver the CE works. Initially, earthworks would need to be carried out with site facilitieshoused on a closed area of carriageway since there is no available accessible land in the vicinity. Prior toearthworks, and at a suitable time of year, the orchids would need to be relocated to an alternative site.Once earthworks were complete, the site compound, offices etc could be housed on the site itself. Asthe Molasse rock is so close to the surface, it will be possible to use standard techniques such as rockbolting, shotcreting and traditional CE methods to provide temporary face support during excavationin the Molasse. Once a space has been excavated and support put in place to work safely within, thereinforced concrete basement and shaft would be cast in situ using proprietary formwork systems or evenslip-formed. Construction of the building thereafter would be straightforward using industry standardmethods to erect and clad a steel portal frame building.The widened section of tunnel will need to be constructed in reinforced concrete cast in situ via opencut while a rock breaker would be used to advance the final 36 m of extraction tunnel using the NewAustrian Tunnelling Method. The open cut section would be used as the construction access for tunnel-ling due to schedule constraints requiring fit-out of the experimental hall to begin as soon as possibleto reduce the overall duration. That requirement would mean the open cut section of tunnel would beconstructed last then back-filled on completion.A core would be drilled using a wall mounted coring rig either from TT2 or the works side of the wall.In either case, a "sas" or airlock would be erected in TT2 to contain any dust produced during coring.Tunnelling works close to existing infrastructure will have to be undertaken outside of beam operationperiods both to avoid radiation close to existing tunnels carrying beam and to prevent vibration causingissues to the machine. In addition where works are required within existing tunnels, equipment will needto be removed to allow space for works and prevent contamination of equipment with dust. For thisreason, the schedule will be planned around a long shutdown to give sufficient time to carry out: • Coring into TT2; • Construction abutting TT2; • Demolition and construction of joints with TTL2; • Work within close proximity of the AD building.The site is very constrained in terms of space as can be seen from as shown in Fig. 5.86, so furtherstudy will be required to ensure the impact on surrounding buildings, accesses and existing operations is124
Infrastructure and Civil Engineering
Figure 5.86: Overview showing siting of Experimental surface building with existing adjacent infrastruc-ture and contour plan of topography after proposed works.minimised. The required off-sett to allow works during normal beam operations will need to be definedthrough study.Radiation issues will also need to be carefully considered as part of the construction of the basementstructure due to the minimal distance to TT7 and its shielding blocks. Further testing of soil activationand to confirm the shielding block arrangement will be needed in advance of works. The followinghierarchy of considerations will be applied for works planning of this area specifically:1. Avoid disturbing shielding blocks if possible;2. Review possibility of casting concrete against shielding if required, giving consideration to futuremaintenance and demolition;3. Remove if unavoidable, ideally reusing shielding in an area which will become activated in anycase.All temporary facilities needed for the civil engineering work have been included in the cost estimate,but any temporary areas/buildings needed for machine or detector assembly/installation were not. It isassumed no existing roads will require diversion.
The eSPS experimental area is composed of a surface hall located next to B393 and the undergroundexperimental hall located below the surface hall next to transfer tunnel TT7 as shown in Fig. 5.83.The purpose of these structures includes the following: • Install and house the missing momentum and beam dump experiments; • Houses the control room, clean room and work area for the experiments; • Service storage and distribution for the experiments and the extraction tunnel;125
Infrastructure and Civil Engineering • Beamline equipment access to the extraction tunnel; • Personnel access to the experiments and extraction tunnel.The equipment includes magnets, beam instrumentation, vacuum equipment, and general electricaland safety equipment. The services include electrical services, cooling, and ventilation.The surface building has a finished floor level of 444.5 m above sea level at ground level. The under-ground hall is 11.5 m below the surface building and is the same level as the extraction tunnel. Thereis second floor level 1.2 m below this to allow the large detectors of the missing momentum and beamdump experiments to be centred on the beam line. The surface hall’s east wall is approx. 10 m fromB393 and experimental hall’s east wall is approx. 1.2 m from transfer tunnel TT7’s steel shielding wall.The surface hall houses the power supply, control racks, and the cooling and ventilation equipmentfor the extraction tunnel and the experimental hall as well as the detector control room, clean room (forwork on silicon detector) and work area as shown in Fig. 5.87. Within the building is the personnelaccess shaft and the equipment access opening. The experimental hall houses the missing momentumand beam dump experiments. The proposed integration layout is shown in Fig. 5.88.Figure 5.87: Integration layout of the surface building.The cable trays, cooling piping, and ventilation ducts are distributed to the experimental hall andextraction tunnel via the personnel shaft, from which services pass underneath the false floor in thesurface building and into the technical gallery that runs into the personnel shaft. The technical gallery is2.0 m wide and 0.8 m deep with a removable ceiling to allow personnel to access it from the floor levelof the surface hall.The access control for the experimental hall and the extraction tunnel is housed inside the surface halland is composed of a material access door (MAD), a personnel access door (PAD), and an emergencydoor. Through the access control there is a personnel lift and stairs, which bring personnel and lightequipment down to the experimental hall. A buffer zone for storing material brought up from the tunneland experimental hall during maintenance work is located next to the access control. Located inside thesurface hall is a large equipment access opening that allows equipment to be lowered directly onto theextraction tunnel floor and the experimental hall floor.During the equipment installation, transport vehicles, including a 40 t semi-trailer, a 19 t truck, anda forklift, will move the equipment into the access building, from which a 25 t overhead crane will126
Infrastructure and Civil Engineering
Figure 5.88: Integration layout of the experimental hall.lift and move equipment throughout the building as well as lower the equipment into the equipmentaccess opening. A handling vehicle inside the extraction tunnel will collect the equipment for positioningalong the beam line whilst the detector equipment will be manoeuvred into its final position using theoverhead crane. Inside the building, there is a personnel platform at the height of the crane rail to allowpersonnel to access the crane. To access the platform, there is a ladder attached to the side wall. Theequipment access opening will be filled with 20 concrete shielding beams during beam operation becauseof RP considerations, and surrounding the equipment shaft is a guard rail for personnel safety. Duringinstallation and de-installation of equipment in the tunnel, the shielding beams will be removed from theshaft and stored at the back of the building.The personnel shaft is located inside the surface hall next to the access control. The shaft houses apersonnel lift and a metallic personnel staircase, and the services (ventilation, cooling, and cable trays).The shaft brings personnel from the FFL of the surface hall to the FFL of the experimental hall. Theservices are brought into the personnel shaft from the technical gallery in the surface hall to a hole in thepersonnel shaft that allows the services to be distributed in the experimental hall.The surface hall is a single-storey steel portal frame industrial building with a width of 19 m betweenthe inner edges of the columns. The building has a length of 40 m between the inner edges of the gablecolumns with a longitudinal centre-to-centre column spacing of approximately 5 m. The ceiling-to-floorheight is approximately 8.8 m, and there is a 0.4 m deep false floor on the north side of the building witha technical gallery running from the false floor to the personnel shaft.The width of the structure (19 m) is based on a personnel accessway (1.0 m), upstand (0.2 m), beamsupport (0.5 m), the equipment access opening (11.5 m), beam support (0.5 m), upstand (0.2 m), the liftshaft (2.65 m) and the stairs shaft (2.45 m). The length of the structure (40 m) is based on the clear-ance of the overhead crane (3.365 m), the shielding blocks (3.285 m), a personnel accessway (1.0 m),upstand (0.2 m), the equipment access opening (14.5 m), upstand (0.2 m), an assembly area (3.75 m), a127
Infrastructure and Civil Engineering semi-trailer (2.5 m), a personnel accessway (1.0 m), length of power converter (4.8 m), minimum powerconverter edge distance (1.4 m), control room (4 m).The equipment access opening is 14.5 m long and 11.5 m wide, for transporting heavy equipment intoand out of the experimental hall and extraction tunnel. The 14.5 m is composed of the detector area is10 m long and the extraction tunnel area is 4.5 m long (the largest magnets are approximately 3.5 m longallowing for 0.5 m clearance either side). The 11.5 m is composed of the two detector widths (6 m eachwith a 0.5 m overlap).The height of the structure (8.765 m) is based on the height of the lorry access door (4.7 m), theclearance between the top of the door and the underside of the crane (0.5 m), the height of the crane(2.765 m), the clearance between the top of the crane and the underside of the services (0.5 m), and anallowance for lighting and services (0.3 m). To access the overhead crane for maintenance, there is apersonnel platform, for which the crane is offset from the wall/column by 0.8 m. On the other side, thecrane is offset by 0.35 m to provide the minimum crane clearance. On the roof, there is a 1.1 m tallparapet wall for personnel access safety.The clean room, work area and control rooms sizes were specified by the users of the facility.
The cooling and ventilation conceptual design for the experimental area of eSPS is presented below. Ithas been designed in accordance with the heat loads estimated by each user group. It should be statedthat since the design is conceptual and estimates have been made conservatively, no safety factor wasapplied to the heat loads. This will be discussed at a later stage when more detailed studies can be made.The scope of the cooling and ventilation needs of eSPS has excluded the SPS tunnel, TT10 and TT20 inline with the ICE scope. It is assumed that present cooling and ventilation systems in these areas whichare sufficient for proton operation in SPS, can provide for the reduced needs of eSPS.The experimental area premises are divided in three zones: • Experimental hall, i.e. surface building; • Detector hall, i.e. underground area; • New tunnel, linking the detector hall to TT2 and crossing TTL2
Piped utilities.
The cooling system for the detector is composed of a primary circuit connected to theAD cooling towers and two secondary circuits: • One with demineralised water for the power converters and beam lines, • A second circuit with chilled water for detector cooling and air cooling. • A third system composed by glycol water is foreseen for the cooling of part of the detector.Table 5.16 summarises the heat loads for the different equipment. The cooling circuits are, therefore,sized with the operational parameters indicated in Table 5.17.128
Infrastructure and Civil Engineering
Table 5.16: Water heat loads for cooling circuits.
Demin. water Chilled water Glycol water
Heat loads [kW] Experimental hall 150 - -Detector hall - 7 5.1New tunnel 730 - -TT10, TT20 200 - -Air conditioning - 60 -
Total heat load 1080 67 5.1
Table 5.17: Working parameters for cooling circuits in experimental area.
Cooling capacity Flow rate Temperature Inlet pressure[kW] [m /h] [ ◦ C] [bar]
Primary circuit 1200 130 24 - 32 5Demin. water 1100 120 25 - 33 12Chilled water 80 11 12 - 18 10Glycol water 5.5 1 0 - 5 5The primary circuit shall be connected to the cooling towers of the AD experimental area, close to thefuture building. This connection is not studied yet and will be defined at a later stage.The chilled water shall be produced by a dedicated chiller to be located outside the new building. Inthis case it is foreseen to have a redundant chiller to ensure continuity of operation for the cooling ofequipment. This could be omitted at a later stage of project development if not considered essential inorder to minimise costs. A second possibility would be to foresee a backup supply from the future chilledwater station that will be installed during Long Shutdown 4 (currently planned for 2025-2026) for theventilation needs in the AD complex.All sections of the cooling plant would be installed in the surface building, close to control room 1 andthe power converter area. An area measuring 13 m in length and 6 m in width is needed. A false floorbelow the station approximately 1 m deep is needed to allow the routing of pipes.Dry riser pipes made from stainless steel and ND100 for fire extinguishing purposes, will be installedin the underground areas, i.e. the detector hall and the tunnel up to TT2. Initially it is estimated that 20connections will be made available to allow fire brigade interventions along the length of the pipeline.They will be connected to the Meyrin site hydrant distribution network. At the next stage of development,when the routing and length of the pipe will be defined, the available pressure will need to be checkedand, if needed, an overpressure pumping system can be included. At the present level of the study sucha system is not foreseen. Similarly, the installation of hydrants outside the building and fire hoses insidewill be discussed at a future stage of the study.Compressed air will be provided by the Meyrin site distribution network. The precise connection pointand the distribution network in the experimental area will be defined in the future.
Heating, ventilation and air conditioning.
Ventilation in the experimental area is provided by severaldifferent systems in order to cover the functionalities required in the different areas for operation duringrunning and maintenance periods. The different ventilation areas are: • Ventilation of the experimental hall; • Ventilation of the detector hall; • Ventilation of control room no. 1; 129
Infrastructure and Civil Engineering • Ventilation of control room no. 2; • Ventilation of clean room; • Pressurisation of the lift cage; • Ventilation of the new tunnel and TTL2; • Smoke extraction from the experimental hall; • Gas extraction from the detector hall.For the time being, it is not foreseen to allow for different flow rates for run and access modes since therequested values during run do not warrant it. In order to save energy, air re-circulation is foreseen and amaximum of 20% fresh air can be provided in each installation with the exception of the pressurisationof the lift cage which will be fully supplied with fresh air.All other considerations made for the ventilation systems of the injector and transfer areas are validfor the experimental area, in particular the redundancy level.The ventilation of the experimental hall is performed via an air handling unit (AHU) located inside thebuilding, with supply ducts and diffusers along the perimeter at ground floor level and an extraction linein the upper part of the building to recycle the air. Smoke extraction from the building will be allowedvia 4 smoke extractors located on the ceiling of the building.The control rooms and the clean rooms in the surface building will have a dedicated air conditioningsystem each, with fresh air and exhaust air ducts connected to the exterior of the building.The lift cage and the stairs will have an overpressure of at least 40 Pa above the underground detectorhall with the pit closed. It is essential that an effective air-tight seal is created with the shielding blocksclosing the pit. For this, an AHU will be located on the roof of the lift cage and will distribute fresh air tothe top and the bottom of the lift cage thus ensuring an overpressure corresponding to the door locations.The ventilation of the detector hall is effected using: • AHUs located close to those dedicated to the ventilation of the building. • A supply duct running through the pit in the underground area and supplying air via diffusors atfloor level • An extraction duct in the upper part of the underground hall will allow the recycling of the air viadedicated fans.The extraction duct and the fans can also be used for smoke extraction in case of fire with the pitclosed by shielding beams. If a fire occurs with the pit open, smoke extraction will be via the smokeextractors in the building. Although the type of gas used has not been finalised, the detector will requirethe presence of some gas (Helium or carbon dioxide) and, therefore, a gas extraction system is foreseenfrom the detector hall venting to the roof of the building. The layout of this system is not defined andwill be studied when more detailed information will be available.Finally, ventilation of the new tunnel and TTL2 will be implemented using AHUs located in thesurface building, with a supply duct linking AHUs, via the detector hall, to the far end of TTL2 and thenew extraction tunnel. In order to reach one of the ends of TTL2, a hole will be drilled to allow the ductto be installed through the rock or tunnel lining between one end of TTL2 and the end of the new tunnel.Air will be blown at this end and at the beginning of the new tunnel, close to the detector hall. Then, airis recuperated via a duct extracting air at the level of the crossing between the new tunnel and TTL2 andcirculated back to ventilation units on the surface.Table 5.18 shows the heat loads to be removed by the ventilation systems. The operational parametersof the ventilation systems are in Table 5.19. 130
Infrastructure and Civil Engineering
Table 5.18: Technical loads on ventilation for the experimental area
Area [m ] Volume [m ] Technical load Experimental hall 808 7430 60 kWUnderground Detector hall 176 1922 25 kWControl room 1 32 96 10 PCs, 10 screens, 10 peopleControl room 2 32 96 10 PCs, 10 screens, 10 peopleClean room 15 45 5 kWLift / stair cage 23 251 noneNew tunnel 316 778 37 kWSmoke extraction 808 - 3 m /s per 100 m Gas extraction - - not availableTable 5.19: Working parameters for ventilation systems in the experimental area
Technical load Cooling Heating Flow rate Number AHUs[kW] [kW] [kW] [m /h] Experimental hall 60 20 47 30000 2Detector hall 25 10 23 15000 2Control room 1 4.5 2 5 3000 1Control room 2 4.5 2 5 3000 1Clean room 5.6 1 2 10000 1Lift / stair cage 0 10 23 3000 2New tunnel 37 15 34 22000 2Smoke extraction 0 0 0 100000 4Gas extraction n.a. n.a. n.a. 5000 2
Access to the experimental cavern, where the detectors for the dark matter andthe beam dump experiment are installed, will be prohibited during beam operation. The experimentalhall on top of the cavern is classified as non-designated area, with exception of the fenced area directlyon top of the shaft.Access to the experimental cavern itself shall be possible while proton beam operation is continu-ing in the TT2 transfer tunnel for beam delivery to n_TOF and the SPS. The zoning, local shieldingand distances are defined such to assure that this access is possible without impacting the proton beamoperation.
Prompt Radiation.
While the design of the missing momentum experiment is approximately known,and is discussed in [2], not much information is available at this stage about the design of a second pos-sible experiment. To model the most demanding scenario, this is assumed to be a beam dump experimentabsorbing all electrons that eSPS can deliver. In order to plan with a minimum shielding on top of thecavern, a large amount of shielding is considered around the beam dump experiment target. This shield-ing is required as well to reduce the radiation background in the missing momentum experiment. Note,however, that a dumped beam into a second experiment, will not arrive at the same time as an electronbeam is delivered into the missing momentum experiment. A large part of the prompt radiation from theinteracting electrons is absorbed in the detectors of the experiment themselves and the bulk shielding.The protection of the experimental hall is achieved by a shielding on top of the experimental cavern madeof 1 m thick removable concrete slabs. The shielding design must be confirmed once more details are131
Infrastructure and Civil Engineering known about the design of this second experiment.Figures 5.89 and 5.91 show the expected dose rates inside the cavern and on top of the cavern at theground level of the experimental hall. The dose rates in the accessible areas in the hall remain below thedesign target of 0 .
15 µSv h − for a non-designated area.Figure 5.89: Effective dose rate inside the experimental cavern from beam on the missing momentumexperiment (3 . × e − / s) and the beam dump experiment (5 × e − / s). (a) (b) Figure 5.90: Si 1 MeV neutron equivalent fluence inside the experimental cavern from (a) beam onthe missing momentum experiment (3 . × e − / s) and (b) the beam dump experiment(5 × e − / s).Three underground infrastructures are located in the vicinity to the experimental cavern. TT7 is locatedlaterally and TT1/TT6 in forward direction of the beam, as shown in Fig. 5.92. Fig. 5.93 show verticalcuts aligned with the beam line directed on the experiments and illustrate the distances and levels withrespect to TT1/TT6, TT7 and TT2.The unused tunnel TT7 is located at only about 1 m lateral distance from the cavern, but after thecavern itself. The massive steel shielding at the end of TT7 protects against radiation inside TT7 suchthat it is of no concern during beam operation (see as well Fig. 5.89).132 Infrastructure and Civil Engineering
Figure 5.91: Effective dose rate in the experimental hall (on top of the cavern) from beam on the missingmomentum experiment (3 . × e − / s)and the beam dump experiment (5 × e − / s).Figure 5.92: Location of the experimental hall (red) on the Meyrin Site with respect to other undergroundinfrastructure.TT1 and the junction with TT6 are located in forward direction of the beams in the experimentalcaverns and at the same level. The distance is about 38 m. Muons produced by the high-energy electronbeam will penetrate and potentially reach the TT1/TT6 tunnel, which is visible in Fig. 5.89. The radiationtransport calculations show that the radiation levels drop at about 30 m distance from the end of theexperimental cavern (Fig. 5.94). Access to TT1/TT6 during beam operation should hence be safe. Impact from Proton Operation in TT2.
The experimental cavern shall remain accessible during pro-ton operation in TT2. The dose rate contribution inside the cavern from proton beam operation shall atleast remain to Supervised Radiation Area design target values. This is to avoid unjustified exposure of133
Infrastructure and Civil Engineering
Figure 5.93: Location of the TT1, TT6, TT7 and TT2 tunnels with respect to the experimental cavern.Vertical cuts parallel to the beamline directed on the experiment, through the experimentalcavern (bottom) and with some offset towards TT7 (top).Figure 5.94: Dose rates in beam forward direction behind the missing momentum experiment and beamdump experiment for nominal beam intensity of 5 × e − / s, mainly coming from penet-rating muons. 134 Infrastructure and Civil Engineering persons inside the experimental cavern whose work is not linked to the proton operation of the acceleratorcomplex.Figure 5.95 shows the effective dose rate impact on the experimental cavern from proton opera-tion in TT2. Two scenarios were considered: Permanent beam operation with 4 . × protons / s(=10 W) point-like losses in TT2 and a full beam loss localised at the most penalising location with2 × protons, which corresponds to two high-intensity PS cycles destined to the SPS for fixed targetoperation.In both scenarios the dose rates and doses are compliant with the design target inside the experimentalcavern leading to less than 1 µSv h − and 1 µSv per 2 × protons lost respectively. From actual exper-ience, transmission in TT2 is much cleaner, hence the dose rates should be well within these values. Aradiation monitor will be required at the zone delimitation to detect potential proton beam losses whileaccess is possible to the experimental cavern. 135 Infrastructure and Civil Engineering
Figure 5.95: Effective dose rates from 14 GeV proton operation in TT2 with 2 ×
10 W localised beamloss points, corresponding to 4 . × protons / s. Residual Radiation.
The beam intensity on the missing momentum experiment represents only a frac-tion of the total beam power sent to the experimental area. The majority of electrons and the resulting136
Infrastructure and Civil Engineering secondary radiation from the interaction with the target will be absorbed in the massive shielding aroundthe beam dump experiment.The residual dose rates in the experimental cavern from beam operation on the missing momentumexperiment and the beam dump experiment remain hence at a very low level, well below 1 µSv h − . Theexpected dose rates after 180 d of operation and 1 d cool-down are shown in [87]. Environmental Impact.
As noted in the previous paragraph on the low level of activation of the exper-imental cavern, the activation of air, the principal path of production of radioactivity which could impactthe environment, will remain low. Compared to the activation in the linac in TT5/TT4, the levels will beconsiderably lower in the sub-nanosievert per year range for the committed dose to the reference group.
Radioactive Waste.
Under nominal operation conditions the activation levels in the experimental cav-ern will be low as shown above. A more detailed study of the activation of the experiments can beconducted once the detailed design and material compositions are known. A full life-time analysis of theexperiments is required to estimate volume, mass and production rate of radioactive waste. Such detailsare expected to be known and addressed during the technical design phase of the project.
Radiation Monitoring System.
A small number of detectors will be required to monitor the surround-ings of the experimental area. One mixed-field detector will be required at the interface between theexperimental cavern and the TT2/TT2L tunnel connection. This detector will be conditioned by the ac-cess mode to the experimental cavern. Another mixed-field monitor will be required in the experimentalhall on top of the shaft to detect potential beam losses potentially occurring just before the experimenttargets.
The new building will be equipped with an EOT crane, shown in Fig. 5.96, with the characteristicsdescribed in the Table 5.20. The EOT crane will allow the construction, the installation of the detectorsin the pit and the lowering of all magnets for the new beam line.Table 5.20: Experimental hall EOT crane characteristics.
SWL Span Lifting height Power
25 t 1500 mm 19000 mm 50 kWIn addition, the building will be equipped with a lift with the characteristics described in Table 5.21below being capable to lower the size of a Euro-palett including the pallet truck.Table 5.21: Experimental hall lift characteristics.
Capacity Cabin size Door width
25 t 1500 mm 19000 mm137
Infrastructure and Civil Engineering
Figure 5.96: 3D view of the new EOT crane inside the experimental hall. • The lift and stairs shall be protected against fire, not connected to the general electrical circuit (i.e.,can be used at any time); • A safe area, with an over-pressure relative to the surroundings shall be available at the base ofthe lift and stairs. The size of this area shall be commensurate with the number of occupants,in addition to the time taken for evacuation, and shall be determined as the project moves to thetechnical design report stage.
Compartmentalisation -
A fire compartment must be established between the new transfer tunnelto the eSPS and the eSPS experimental area. This will be subject to further analysis as the project movesinto detailed design;
Fire Suppression Means -
A dry riser is foreseen for the experimental area. The full requirementsfor this equipment are summarised in [95].
Smoke Extraction -
Hot smoke extraction is foreseen, allowing the CERN Fire and Rescue Serviceto employ this tool in line with their strategy to tackle a fire.
Chemical Safety.
No chemical agents or gases are currently foreseen for use within the missing mo-mentum experiment; should any additional chemicals be proposed for use in the facility, the chemicalspecialists within the HSE group must be consulted.Lead is not currently part of the shielding design for the experimental area. However, as the designis still at a preliminary stage, it is important to note that lead can present significant hazards. Care mustbe taken that the necessary procedures are followed for purchasing, shipping, storing and handling of138
Infrastructure and Civil Engineering the blocks to limit the dangers of lead poisoning or exposure to activated materials. In particular, blocksshould arrive at CERN pre-painted or adequately protected by equivalent means, to ensure that risksfrom dust are contained. Should lead be required, it must be registered in the CERN Chemical database,CERES, with the quantity, location and hazards recorded. A copy of the up to date SDS must be uploadedinto the database and a Chemical risk assessment performed using CERES, if required. The followingsafety form provides guidance for the safe handling, storage and use of lead: • Safety Guideline C-0-0-3 – Lead.
Electromagnetic Safety.
The analysing magnet for the missing momentum experiment is intended tobe operated with a 1.5 T central field. The restrictions around this magnet are expected to be enforcedas a natural result of the radiation levels around the magnet when it is in operation. The expected riskis, therefore, only foreseen for commissioning, testing, and maintenance of the magnet, which shall bemitigated by ensuring that these operations are only carried out by appropriately qualified personnel.This shall be reviewed as the project moves into detailed design.
Preservation of the Natural Environment.
The mound in which the experimental area is to be locatedis known to be home to a number of orchid species which are protected under the following regulation;they shall be protected, restored or adequately replaced: • Code de l’environnement, Art. L411-1 . A new safety chain dedicated to the eSPS experimental area will be integrated in the personnel protectionsystem of the PS complex.The access to the eSPS Experimental area will be made through an access point composed of onepersonnel access device and one material access device. Aside the access point an ‘end-of-door’ zoneis installed to allow the exit of people in case of emergency. The protection of the experimental hallis ensured through a specific shielding on top of the experimental cavern. An operational procedureor a specific interlock shall check the presence of this shielding before authorising beam operation. Anew ‘inter-zone’ door will be installed to close the end of the TT2 tunnel before entering the eSPSexperimental area. The access control layout is shown in Fig. 5.97.To protect people accessing the eSPS experimental area from radiation hazard, it is proposed to useone bending magnet and two beam stoppers in the TT2 tunnel as EIS-beam. In case of intrusion duringbeam operation or if the safe position of a safety element is lost during access, the safety chain will alsosecure the upstream zone: the TT2.The entrance to the eSPS experimental hall will be equipped with an electrical lock controlled by abadge reader. 139
Infrastructure and Civil Engineering
Figure 5.97: Safety elements and access equipment of the eSPS Experimental area.140
Accelerator Facility Research and Development
The proposal for the eSPS accelerator described in Sec. 3 and Sec. 4 was initiated as an electron beamfacility optimised for a missing momentum experiment to reach the physics goals described in Sec. 2.However, in addition the 3 . . Studies with Relevance for Future Facilities.
The X-band linac and operation of the SPS with elec-trons open key strategic possibilities for future e + e − facilities at CERN. First of all, the construction andoperation of the linac is in itself a natural next step for the X-band high-gradient technology towards aCompact LInear Collider (CLIC) [5]. The resources needed are similar to those invested annually duringthe last decade in the CLIC study. In addition the FCC-ee [96], LEP3 [97] and LHeC [98] conceptsrely on electron and positron injectors with important challenges that could be addressed with the eSPSfacility. In particular the SPS RF system proposed would serve as a prototype for FCC-ee, see Sec. 4.2.2.Studies of positron production at the end of the X-band linac are an interesting future addition whichwould benefit any future e + e − collider, be it linear or circular, inside or outside CERN. For any futuree + e − facility, positron production is one of the most important parts. New ideas being discussed in thiscontext include an implementation of the LEMMA [99]) (Low EMittance Muon Accelerator) conceptfor a muon collider to be implemented at CERN by re-introducing positrons at 45 GeV in the LEP/LHCtunnel. The muon production scheme could in this case be studied in parallel for example with LEP3operation. While such a phased approach to a muon collider is well beyond the scope of this CDR andalso current planning, positron production and target studies will be both relevant and crucial for thispossible avenue. The relevance of eSPS to the future facilities mentioned is described in Sec. 6.3 andSec. 6.6 , while the additional studies made possible by positron production are described in Sec. 6.6. Plasma Acceleration.
The proposed facility opens the possibility of a significantly broader plasmaacceleration programme at CERN, in line with European priorities for this field, primarily addressingkey challenges as needed for the potential use of the technology in colliders. Such a programme wouldnaturally build on or be an extension of the existing AWAKE collaboration. Use of the linac electronbeam is considered for plasma acceleration as both driver and probe. As mentioned, positron productionstudies are of vital interest for any e + e − machine, as well as new muon collider ideas, but such a positronbeam would also provide unique opportunities for studies of plasma acceleration of positrons. Therelevant possibilities are described in Sec. 6.4 and Sec. 6.6. General Accelerator R&D.
The facility will provide a test bed for accelerator R&D covering a widerange of topics potentially serving an important user community in its own right. Many of the generalaccelerator studies that can be envisaged are natural continuations of the studies currently carried out inthe CLEAR facility at CERN, see Sec. 6.5. Examples of potential R&D studies include, high gradientand plasma lens studies, instrumentation and impedance studies, medical accelerator developments forexample for VHEE [100] irradiation, component irradiation, THz acceleration, and educational activities.141
Accelerator Facility Research and Development
With the eSPS linac parameters, the capabilities are significantly increased with respect to what can bedone today in the CLEAR facility. The possibilities are described in Sec. 6.5.
The SPS Electron Beam.
The use of the SPS electron beam is also possible but will be in competitionwith other users of the SPS. For the SPS beam one can take advantage of small equilibrium emittanceoptics using the existing SPS lattice. Low emittance ring studies and the use of SPS as a damping ringcan be pursued, as have been considered for both linear and circular future e + e − machines at CERN.One could also consider pursuing final focus studies beyond the ATF2 [101] if this becomes a priority.A realistic implementation of the latter is not studied at this point. The possibilities offered by the SPSelectron beam are briefly outlined in Sec. 6.3.Overall, the facility can significantly extend the strategic possibilities for future machines at CERNand also provide a platform to carry out important studies for future facilities outside CERN as part ofa European contribution to such facilities. The possibility of achieving this at a resource level which iscompatible with that already existing, profiting of the investment made in the last decades for CLIC, andelectrons in the SPS during the LEP era, while performing physics studies that are essential to pave theway towards future larger machines, is very attractive and a unique opportunity for the organisation atthis time. The Accelerator Community Involved as Developers or Users.
The scientific and technical com-munity capable of contributing to and interested in the X-band machine development and construction, aswell as the accelerator user community, are both large. The potential links between the eSPS facility andX-band accelerator developments outside CERN are many, in particular the on-going design study foran X-band based FEL, CompactLight [102], and the collaboration with INFN-LNF for building a 1 GeVX-band linac [103] as part of the LNF EuPRAXIA [104] efforts. These ongoing developments alreadyprovide a network of 25 collaboration partners, many of which are developers and users of X-band tech-nology in their local facilities. One can also expect many additional CLIC collaboration partners toactively direct their collaboration efforts towards studies and technology developments directly applic-able to the eSPS linac. As shown above, it is not only the CLIC collaboration partners that will engagein this facility, AWAKE [105] collaboration partners and CLEAR users would be ready to pursue thefacility build-up and its scientific programme in the area of accelerator R&D. Both these communitiesare large, in particular the very large novel accelerator technology community consider test-facilities atCERN crucial for making these technologies applicable for high energy colliders. The relevance of thefacility for FCC-ee RF and more general circular e + e − accelerator studies, as well as possible studies fornovel muon colliders, bring in additional groups of potential users. The project implementation plan for CLIC foresees an initial 5 year preparation phase prior to a potentialconstruction start. The preparation phase will focus on further technical and industrial developments aswell as production and preparation of key components, system verifications - not necessarily at CERN,strategic developments focused on risk, cost and power reduction, as well as developments towardsthe Technical Proposal for the detector. The governance structure and the international collaborationagreements for the construction and operation will be set up during this time. Site authorisations will alsobe established during this period and site/civil engineering and infrastructure preparation will becomeincreasingly detailed. All of these considerations will be folded into the final design and parameters forthe first CLIC stage. 142
Accelerator Facility Research and Development
The technical developments needed for the preparation phase of CLIC have a large overlap with whatis needed for the X-band linac of the eSPS facility described in Sec. 3 to Sec. 5. In Table 6.1 the keyCLIC accelerator programme for the next phase are considered in view of the technical overlap with theconstruction of an 3.5 GeV X-band linac for eSPS.Table 6.1: Main CLIC related activities and their relation to the 3.5 GeV linac for eSPS.
Details Purpose eSPS Equivalent CommentMain linac modules
Build ten prototypemodules in qualifiedindustries, two beamand klystron versions Final technical design,qualify industrypartners, verifyperformance 12 X-band klystronmodules Covered by eSPS butadaptations to twobeam modules need tobe considered
Accelerating structures
Around 50 structuresincl. for modulesabove Industrialisation,manufacturing andcost optimisation Same number needed Programmesoverlapping
Operating X-band test-stands, high efficiency RF
X-band test-stands atCERN andcollaboratinginstitutes, costoptimised X-band RF X-band componenttest, validation andoptimisation, costreduction andindustrially availableRF units Similar test capacityneeded for eSPS, 24X-band RF unitsneeded for eSPS Programmesoverlapping
Technical components
Magnets,instrumentation,alignment, stability,vacuum Luminosityperformance, costs andpower,industrialisation These components arealso needed for eSPS eSPS specificationsless stringent, howeversignificant advantageto implement insmaller completesystem
Design & Parameters
Beam dynamicsstudies, parameteroptimisation, costs,power Luminosityperformance, risk,costs and powerreduction Needed for eSPS linac Specific studies forCLIC needed but goodreality checkAdditional studies are needed for CLIC, the most prominent is the drivebeam front end optimisationand system tests to around 20 MeV. The purpose is a careful verification of the most critical parts ofdrivebeam concept and to develop the industrial capabilities for L-band RF systems as needed for theCLIC drivebeam. Another area where the CLIC studies do not overlap is in the area site and civilengineering studies. These activities are, however, not expected to be resource consuming until closerto the beginning of construction when increased effort will be needed. Furthermore, system tests in lowemittance rings, FELs, etc. should be pursued and are relevant for eSPS and CLIC. These tests are notresource intensive and are usually carried out in collaboration with outside institutes.The programme outlined for the preparation phase of CLIC above, has, for the most resource demand-ing parts, a large overlap with constructing a 3.5 GeV linac for eSPS. Hence, from the CLIC developmentpoint of view, constructing a 3.5 GeV X-band linac for the eSPS facility as a next stage is a very attract-ive possibility. It enables important physics and accelerator studies of vital interest for CERN, and with143
Accelerator Facility Research and Development limited additional CLIC specific studies, also moves the CLIC accelerator preparation efficiently and ina timely manner towards a more technically and industrially mature state.In addition, collaboration with external groups, for example INFN for the SPARC 1 GeV X-band linac,a potential energy upgrade of the Clara facility at Daresbury and future CompactLight FEL implementa-tions, provide a network of important external collaborators that need a linac module and components ona similar timescale as that of the eSPS. More generally, the CLIC collaboration partners would feel thatthese concrete construction projects, as well as R&D opportunities offered by the linac, would providethem with completely new possibilities for contributions and participation in the coming phase(s).
The ILC in Japan is being considered for implementation in the coming decade [106]. The ILC is a250 GeV e + e − linear collider based on super-conducting RF technology. In the scenario of the ILCbeing constructed in the near future it is expected that Europe will make a contribution to the project.Possible contributions from CERN in relation to the eSPS include a CERN platform for injector, positronproduction, stability, alignment and high efficiency RF studies. Longer term there will be a push toincrease the ILC energy and luminosities, and CERN along with its collaborators would be in a goodposition to provide testing grounds for such developments using the eSPS installations. In addition, thedamping ring studies mentioned above, possibly extracting a low emittance beam for final focus studies,are relevant in this scenario. It it scientifically very attractive to combine an in-kind activity at CERNtowards the development of the ILC with a strong R&D effort for higher energies and higher luminositiesfor the ILC, as well as any future linear collider. Circular Machine studies at CERN such as FCC-ee [96] and LEP3 [97] would rely on cost-effectiveelectron and positron production and injection. FCC-ee studies are currently an important R&D activity,providing an initial option for a future ring based accelerator facility at CERN. The RF system proposedfor eSPS is a prototype of an FCC-ee RF system as discussed in 4.2.2. Design, construction and operationof such a system can provide important experience in all the aspects mentioned above. Beyond thesetwo straightforward considerations, more elaborate possibilities exist. The possible use of the SPS as adamping ring, building on the LEP experience and expanding on the studies made in the context of CLICand modern low emittance rings, for these machines have been described in [54]. Studies consideredin this context are tests of super-conducting wigglers, kickers, vacuum/coating, instrumentation, beamprofile monitors (synchrotron light), halo monitoring, BPMs, bunch-by-bunch and turn-by-turn feedback(LARP), RF systems (800 MHz cavities in particular), beam dynamics issues related to optics, non-lineardynamics, intra-beam scattering, instabilities, e − cloud (for e + ) and ions (for e − ). Electron beam driven plasma-based accelerators, known as plasma wakefield accelerators (PWFA) (seeFig. 6.1), have shown accelerating gradients in excess of 50 GeV/m sustained over a metre-long plasmaleading to a 42 GeV energy gain [107] and the acceleration of a witness electron bunch with narrow en-ergy spread (few %-level) and good energy transfer efficiency ( ∼ • Preservation of the incoming emittance of the accelerated bunch; • Acceleration with a narrow final relative energy spread (%-level) using beam loading. This alsocontributes to the preservation of the beam emittance;144
Accelerator Facility Research and Development • Matching of the beam to the plasma strong focusing, including plasma density ramps at the en-trance and exit of the plasma; • Possibility of shaping the drive bunch to reach a transformer ratio larger than two, the currentmaximum reachable, with a symmetric current profile bunch; • Operating with a plasma source with characteristics suitable to preserve emittance and energyspread (density uniformity). © P. Muggli
P. Muggli, PBC 07/10/2018 € E WB = m e c ω pe e = m e ce n e e ε m e ∝ n e € k pe σ z = n e . k pe σ z € k pe σ r ≤ σ r € n b n e = Ν € n b = π ( )
3/ 2 N σ z σ r cm -‐3 =241µm =171µm =1.1x10 + + + + + + + + + + + + + + + + + + + + + + + + + + + + + ----- ----------------------- ----------- -------- --------------------------- ------- --- -- - ------- - -- ------ - -- ------ - -- - - - --- --- -- - - - - - -------- -- --- + + + + + + + + + + ++ + + + + + + + + + + + + + ++ + + + + + + + + + + + + + ++ + + + + + + + + + + + + + + - - - -- - - - - e - Driver + + + + + + + + + + + + + + ++ + + + + + + + + + + + + + + e - Witness + + + + + + + + + + +++ + e - bunch € σ rW = σ rD € σ zW < σ zD € N W ~ N D / 3 € ε W ~ ε D € W W << W D Need simulaHons Include a W-‐bunch in the structure:
Figure 6.1: Schematic of the PWFA: the drive electron bunch expels the plasma electrons that are attrac-ted back towards the axis by the positive charge of the ion column and overshoot (blue arrowsin the bunch frame of reference). The plasma electron density perturbation sustains the lon-gitudinal wakefields that are decelerating inside the drive bunch (green arrow in the secondbubble) and accelerating within the witness bunch. It also sustains transverse focusing fieldsinside the bubble (red arrows).The use of the eSPS facility for PWFA studies could directly address some of the challenges relatedto the first electron acceleration PWFA stage of a possible linear collider. They are related to the accel-eration and the quality of the electron witness bunch. Though the energy of the drive bunch (3.5 GeV) islower than that envisaged for such a collider stage ( ∼
25 GeV), the challenges are not (or weakly) energydependent.An experimental facility with a multi-GeV electron beam capable of driving wakefields in the non-linear regime and with an independent electron witness bunch is necessary to demonstrate the applicab-ility of the concept to HEP. Challenges related to the quality of acceleration of a positron witness bunchwill be addressed if/when positron bunches become available.Such a facility for PWFA experiments would be unique in the world to study collider related chal-lenges. It may be the only facility in the world with a multi-GeV drive bunch and truly independent elec-tron witness bunch, absolutely necessary features to tackle these challenges. The addition of a positronwitness bunch would make the eSPS the facility for PWFA collider studies.
Considering the potential eSPS linac drive bunch energy of 3.5 GeV, the beam and plasma parametersnecessary to conduct a PWFA experiment in the non-linear regime can be determined from the following"recipe" for the plasma and drive bunch parameters. Reaching the blow-out regime is necessary to be ableto preserve emittance and narrow relative energy spread over long plasma lengths and in the presence oflarge energy gains.One can target full energy depletion over a metre scale plasma, i.e., an accelerating field of ∼ E WB = m e c ω pe / e ∼ = (cid:113) n e [ cm − ] [V/m] . (3)145 Accelerator Facility Research and Development
Here ω pe = (cid:16) n e e ε m e (cid:17) / is the plasma electron angular frequency in a plasma of electron density n e . Thisdetermines n e . To effectively drive wakefields, the rms length of the drive bunch σ z must be of the orderof the cold plasma skin depth c / ω pe , which is expressed as: σ z / (cid:0) c / ω pe (cid:1) ∼ = √
2. The transverse drivebunch rms size σ r must also be smaller than c / ω pe to avoid filamentation of the bunch: σ r / (cid:0) c / ω pe (cid:1) <
1. In order to reach E WB and the nonlinear regime of the PWFA, the bunch density n b = ( π ) / N σ z σ r must exceed n e : N > ( π ) / σ z σ r n e . This "recipe" yields the parameters listed in Table 6.2. TheTable 6.2: Parameters obtained from PWFA "recipe" for the plasma and drive bunch. Parameter Symbol Value
Bunch electrons energy W WB e × cm − Plasma length L p σ z µ mBunch rms radius σ r µ mBunch population N > × Bunch charge Q > Because the plasma density envisaged here (n e =10 cm − ) is similar to that used in the AWAKE ex-periment, witness bunch parameters can be similar. Therefore, the design of the linac that produces thisbunch can be similar to that for AWAKE. The witness bunch can be of lower energy than the drive bunch.An estimate for its minimum energy is obtained by requiring that with its incoming energy it does notde-phase with respect to the wakefields by more than a quarter plasma wavelength (i.e., it remains in theaccelerating and focusing phase of the wakefields) over the plasma length. With a 1 m-long plasma withdensity 10 cm − (plasma wavelength of 530 µ m), the incoming witness bunch relativistic factor mustexceed 31. This corresponds to an energy of 16 MeV.Its normalised emittance ε N may be small enough to match the witness bunch to the ion column fo-cusing force. In this case, the bunch and plasma parameters must satisfy: σ m = (cid:16) π r e ε N γ n e (cid:17) / , σ m thematched transverse size. Here r e is the classical electron radius. Figure 6.2 displays the beam matched ra-dius as a function of the beam relativistic factor γ for two plasma densities. It shows that for a low energywitness bunch ( γ ∼ = ∼ =
250 and 400 µ m. Upon(adiabatic) acceleration, the transverse size reduces to less than 200 µ m at the final energy ( γ ∼ = Accelerator Facility Research and Development MatchedBeamSize n e =10 (cm -3 )n e =10 (cm -3 ) σ r m ( µ m ) γ ε N =10mm-‐mrad Figure 6.2: Transverse bunch size σ m matched to the pure plasma ion column as a function of the beamrelativistic factor γ for two plasma densities n e =10 and 10 cm − and a normalised emit-tance ε N =10 mm.mrad.and alignment between the drive and the witness bunch must be possible. Plasma and wakefield parameters are similar to those used in the AWAKE experiments. Witness bunchparameters will, therefore, be similar. AWAKE studies showed that a witness bunch length of the orderof ∼
5% of the wakefield’s period is suitable, 42 µ m (or 140 fs) for parameters of Table 6.2. For matchingof the bunch to transverse wakefields, its beta function must be of the order of 5 mm. This then leads toa transverse beam size at the waist, located at the plasma entrance, in the order of a few to a few tensof microns. Producing such small and short bunches carrying the hundreds of picocoulombs necessaryto load wakefields and limit the final relative energy spread requires relativistic electrons. Experienceshows particle energy must be in the region of 150 MeV. The injector could thus be similar to that whichwill used for AWAKE, CLEAR and EuPRAXIA@SPARC_LAB (see Sect. 3.7). It consists of an S-band gun, followed by an X-band structure for velocity bunching and one or two X-band acceleratingstructures. With two accelerating structures the energy can reach 160 MeV. The witness beam producedin this accelerator needs to be merged with the 3.5 GeV drive beam. An achromatic and synchronousdog-leg was developed for AWAKE. A similar design could be adopted to bring the bunch from the lowenergy witness bunch linac to the plasma entrance. Integration of the witness injector is described inSec. 3.7 and its integration in Fig. 3.15. Typical parameters are given in Table 3.7. At this point, numerical simulations must be used to determine the parameters more precisely. Thisincludes drive and witness bunch parameters. In particular, the drive bunch needs to be shorter to ac-commodate the witness bunch in the accelerating cavity and the witness bunch parameters need to beoptimised to reach optimum beam loading.Figure 6.3 shows snapshots of the drive bunch (-0.1 ≤ ξ ≤ Accelerator Facility Research and Development into the plasma, the electrons at the back of the drive bunch have lost almost all their energy (see bottom-right panel) and have de-phased towards the back of the bubble ( ξ ∼ = -0.40 mm, top-right panel). This isevidence of energy depletion of the drive bunch. Dephased e -‐ e -‐ gaining energy e -‐ re-‐lose energy z=0.05m z=0.62m z=0.05m z=0.62m p z ( G e V / c ) r ( mm ) ξ (mm) ξ (mm) Figure 6.3: Top panels: snapshots of the drive bunch (-0.1 ≤ ξ ≤ e /n e ) at two locations along the plasma (z=5 and62 cm). Bottom panels: corresponding snapshots of the drive bunch electrons longitud-inal momentum. The simulation parameters used were: n e =5.6 × cm − , σ z =100 µ m, σ r =70 µ m, N=4.3 × (n b /n e =1), thus E WB =7.24 GeV/m.The detailed parameters of the drive and witness bunch will be determined from similar simulationswith bunch parameters consistent with the capabilities of the injector(s), as described for example inSec. 3. We note here that the PWFA requires short bunches, which seem to be less sensitive than longerbunches for example to misalignment (see Sec.6.6). The drive electron bunch parameters can also be ad-justed to drive wakefields in other regimes (e.g.quasi-linear, hollow channel) for example for accelerationof a positron bunch on the wake driven by the electron bunch (see Sec. 6.6). The plasma parameters of Table 6.2 call for either a laser-ionised alkali metal vapour source [109], orfor a discharge source. Each source has its pros and cons. Both require isolating vacuum windows ordifferential pumping. The vapour source is attractive because it is a passive and very reproducible systemthat can be finely tuned, but it requires an ionising laser system (excimer for lithium or Ti:sapphire forrubidium). The discharge source operates with a noble gas (argon), which simplifies operation, but itrequires a (rather simple) discharge system.
Reaching drive bunch parameters suitable for the PWFA experiments requires further compression ofthe linac bunch by a chicane compressor. The main components of the experiment itself are a finalfocus system, the plasma source (metre-long), a magnetic imaging spectrometer capable of measuringup to ∼ Accelerator Facility Research and Development
A PWFA experimental facility driven by the 3.5 GeV electron bunch would play an essential role to-wards the development of a beam-driven, plasma-based, more compact and affordable linear collider, byfocusing on collider-specific issues. Such a facility would most likely be the only one dedicated to thisimportant topic. The availability of a positron witness bunch (see Sec. 6.6) would make it a true andcomplete plasma-based collider research facility. This facility would also be complementary to the highenergy CLEARER facility. Possible extensions could include multiple two or more drive bunches andplasma sections to address the staging challenge.
CLEAR [110, 111] is a user facility at CERN, running in parallel with the main CERN acceleratorcomplex, with the primary goal of enhancing and complementing the existing accelerator R&D andtesting capabilities at CERN. A workshop on the conversion of the probe beam line of the former CLICTest Facility (CTF3) into a new test-bed was held in October 2016 with participation from 80 peoplecovering a broad science community. The scientific and strategic goals set out were the following: • Providing a test facility at CERN with high availability, easy access and high quality bunchedelectron beams; • Performing R&D on accelerator components, including beam-based impedance measurements,innovative beam instrumentation prototyping and high gradient RF technology advancement withrealistic beam tests; • Providing a radiation facility with high-energy electrons, e.g. for testing electronic components incollaboration with ESA or for medical purposes, possibly also for particle physics detectors; • Performing R&D on novel accelerating techniques - electron driven plasma acceleration and THzacceleration. In particular developing technology and solutions needed for future particle physicsapplications, e.g. beam emittance preservation for reaching high luminosities; • Maintaining CERN and European expertise for electron linacs linked to future collider studies (e.g.CLIC and ILC, but also AWAKE), and providing a focus for strengthening collaboration in thisarea; • Using CLEAR as a training infrastructure for the next generation of accelerator scientists andengineers.The CLEAR facility was approved in December 2016 with the first beam being set up in August 2017.After only a few weeks of commissioning, stable and reliable electron beams with energies between 60and 220 MeV in single or multi bunch configuration at 1.5 GHz could be provided to users.
Irradiation Studies.
Irradiation tests are being performed mainly in a dedicated spectrometer beamline (Very energetic Electron facility for Space Planetary Exploration missions in harsh Radiative envir-onments - VESPER). The initial aim of VESPER was to characterise electronic components for operationin a Jovian environment - as foreseen in the JUpiter Icy Moon Explorer mission (JUICE) of the EuropeanSpace Agency (ESA), in which trapped electrons of energies up to several hundred MeV are present withvery large fluxes. Measurements in VESPER showed the first experimental evidence of electron-inducedsingle event upsets (SEU) on electronic components in this kind of environment, thus justifying further149
Accelerator Facility Research and Development studies. A dependency of SEU cross-section with energy and no dependency on radiation flux was ob-served, excluding prompt dose effects [112]. Since then, a wider range of devices have been tested,showing a strong dependency on the device process technology.Interest in further measurements, both at low and high beam intensities and over a wide energy range isquite strong in the space community - a contact with NASA has been established recently and the possib-ility of a dedicated measurement is being explored. In parallel, the local CERN group studying radiationdamage of electronics components in an accelerator environment is interested in the continued use of thefacility. A first request for irradiation studies on electronics components for detector applications wasalso received from Uppsala University, and the first measurements have been completed. Other internalCERN groups have investigated the effect of radiation on components for various detector materials andcomponents.
Medical Accelerator Studies.
The scope of VESPER has been further extended to medical applic-ations. The recent advances in compact high-gradient accelerator technology, largely prompted by theCLIC study, renewed the interest in using very-high energy electrons (VHEE) in the 50 - 250 MeV en-ergy range for radiotherapy of deep-seated tumours. In order to understand the dosimetry of such beamsand assess their viability for treatment, a group from the University of Manchester carried out energydeposition studies in VESPER using a set of EBT3 Gafchromic films submerged in water. The measureddose deposition agreed with simulations, and the longitudinal dose profiles with and without inserts ofdifferent materials showed that the electron beam is relatively unaffected by both high-density and low-density media, thus indicating the potential of VHEE to be a reliable mode of radiotherapy for treatingtumours also in highly inhomogeneous and mobile regions such as lungs [100]. Other groups, from theUK National Physical Laboratory and from Strathclyde University, also showed interest in performingVHEE studies in CLEAR and completed preliminary tests there. Further studies on the dose distributionof a converging beam as opposed to a parallel wide beam are in preparation.
Novel Accelerator Research.
CLEAR’s first exploration of new accelerator technologies has been onactive plasma lenses, promising as strong focusing devices in novel accelerators, due to their compactsize. Transverse field uniformity and beam excitation of plasma wakefields appear to be the more signi-ficant potential limitations. A collaboration lead by the University of Oslo and including CERN, DESYand the University of Oxford was set up to develop a novel low-cost, scalable plasma lens and assess suchlimitations. The setup, described in [113], consists of a 1 mm diameter, 15 mm long sapphire capillary.The capillary can be filled with He, Ar or other gases at a controllable pressure. The gas is ionised by a500 A peak current discharge with a duration of up to a few hundred ns, provided by a 20 kV spark-gapcompact Marx bank generator. The longitudinal discharge current creates the transverse focusing forcein both transverse planes. The experimental set-up was installed in CLEAR in September 2017 and aclear focusing effect was rapidly observed. Extensive measurements, including transverse position scansof a pencil beam, revealed gradients as high as 350 T/m and beyond, compatible with use in a stagedplasma accelerator. More studies measured the uniformity of the field and beam emittance preservationemploying different gas species and gave clear indications on the origin of transverse non-linear effectsand emittance growth, and on conditions for linear and non-linear regimes linked to emittance growth[114]. Evidence of non-linear self-focusing at relatively high bunch charge (about 50 pC/bunch) wasalso observed on the beam after the plasma-forming discharge. This opens up another branch of studieson passive plasma lenses that will be further developed.
Novel Production of Terahertz Radiation.
Another novel technology being explored at CLEAR isthe production of terahertz radiation. Apart from accelerator applications, this technology has a strongimpact in many other areas of research, spanning the quantum control of materials, plasmonics, andtunable optical devices based on Dirac-electron systems to technological applications such as medical150
Accelerator Facility Research and Development imaging and security. The aim at CLEAR is to characterise a linac-based THz source, exploiting relativ-istic electron bunches that emit coherent radiation in the THz domain. For such a source, sub-picosecondelectron bunches are needed. This triggered a study and optimisation of the CLEAR injector in collab-oration with LAL-Orsay, thanks to which sub-ps bunches down to 100 fs rms have been demonstratedin the machine, paving the way to the THz radiation generation. Current studies, carried out in collab-oration with Rome University La Sapienza, CEA/CESTA and Royal Holloway University of London,are focused on the production of (sub-)THz radiation by Coherent Transition Radiation (CTR) in thinmetal foils, by Coherent Cherenkov Radiation (CCR) mechanism in targets of different materials andshapes, and by Coherent Smith-Purcell Radiation (CSPR) in open periodic structures. A peak powerof about 0.1 MW at 0.3 THz has been measured so far, in agreement with theoretical expectations, andtest set-ups to reach the 1-10 MW level are in preparation. Moreover the coherent radiation has beenused for longitudinal bunch diagnostics, providing reliable bunch length values consistent with otherdiagnostics. Transverse shaping of the radiation source has been also demonstrated via control of thesize and divergence of the electron bunch at the source plane. The problem of electromagnetic shadow-ing has been investigated in view of experiments of (sub-)THz radiation production/utilisation involvingmultiple sources and/or compact setups.
General Accelerator R&D.
Ongoing CLEAR activities allow the continuation of R&D and beam testverifications of CLIC critical technologies, for example by measuring the resolution of CLIC cavity BPMprototypes and by checking the behaviour of the wakefield monitors installed in the CLIC acceleratingstructures.An intense R&D activity on beam instrumentation is also ongoing in the CLEAR facility, led bythe BI group at CERN in conjunction with many external collaborators. For example, in the last fewyears, several studies were performed on high-precision BPM systems based on RF cavities [30] or onstudying the limitations in Optical Transition Radiation (OTR) beam imaging systems due to shadowingeffects [115] that occur when using two consecutive OTR screens. CLEAR has also been used to validatebeam instruments prior to their installation on other machines, like the testing of AWAKE BPMs and theAWAKE spectrometer screens.
An obvious synergy exists between the CLEAR facility and the electron beam facility proposed in thisdocument. As already mentioned, the CLEAR injector, or some of its components, could be transportedto and used as an injector for the 3.5 GeV linac. It has also been noted that the linac itself would havea strong impact on further developments of the X-band high-gradient technology and help to build-upan industrial basis for its extensive use in future projects, including CLIC and the ILC. The linac wouldalso be a test bed for other technical components, for example diagnostics, or operational tools and pro-cedures, developed for CLIC, and thus constitute an ideal extension of part of the CLEAR experimentalprogramme.Other experimental activities ongoing in CLEAR could also be made part of the programme of the3.5 GeV linac, by sharing the beam time available when not used for SPS injection. Most of these activ-ities would profit hugely from the enhanced beam characteristics provided by such facility with respect tothe present CLEAR installation. In particular, a 3.5 GeV and better quality beam will extend the scope ofbeam diagnostics R&D dramatically (e.g., for diffraction radiation studies for beam profile monitors andhigh-resolution BPMs). Having access to electron beam energies as high as 3 GeV at the end of the linacand 16 GeV in the SPS and its extraction line will open new opportunities in testing beam instrumenta-tion methods for ultra-relativistic beams. For example, synchrotron radiation (SR) in the ring will allowdevelopment of further innovative techniques using randomly distributed interferometric targets in orderto improve the performance of SR based monitors. At the end of the linac or in the extraction line, R&D151
Accelerator Facility Research and Development on non-invasive beam diagnostics could easily be pursued, for example instruments based on the detec-tion of Cherenkov diffraction radiation (ChDR), an innovative scheme that was recently demonstrated.ChDR refers to the emission of Cherenkov radiation by charged particles travelling parallel to the surfaceof a dielectric material. This combines the already well-known advantages of Cherenkov radiation with anon-invasive technique, and would find applications as beam imaging systems or high directivity BPMsfor a wide range of accelerators, i.e. from high energy lepton or hadron colliders to light sources (3rdgeneration synchrotron light sources or free electron laser facilities). This would also be a very attractivesolution for novel accelerating technologies, like plasma or dielectric wakefield accelerators, into whichstandard beam instruments cannot be easily integrated.The same is true for the studies on novel accelerator technologies. THz radiation sources at muchhigher power than in CLEAR would be possible, with direct applications to the acceleration and manip-ulation of charged particle beams. Higher beam energy will make easily accessible extended interactionregions and staging for both THz and plasma-based acceleration schemes. Potentially important issuesto be explored in order to make even more attractive novel accelerations studies are the single bunch cap-ability of the linac, positron production and acceleration, and the use of two sources with independenttiming for drive/probe type experiments.
Future lepton colliders such as CLIC, ILC, FCC-ee, or a Muon collider based on high-energy positrons,rely on challenging positron source designs which are still at a conceptual level. The CLIC positronsource design can be seen as a reference for electron-positron colliders [28, 36], it was adopted for FCC-ee and serves as an alternative for ILC. A Muon collider based on the LEMMA scheme is of coursemuch more demanding with respect to positron production but the fundamental challenges are similar.Positron source R&D is urgently needed in the areas of target construction,cooling and survival, magneticfield concentrators, and photon production enhancement through channelling radiation. Such challengescould be addressed with a simple positron production area as described in Sec. 3.6.
The availability of intense positron beams with an energy of a few GeV would be vital to a large numberof projects and experiments. For example, future lepton colliders will require both intense electronand positron beams, PWFA based colliders will need to overcome substantial challenges that exist inpreserving the quality of the accelerated positron witness bunch and the interaction of positron beamswith crystals to generate photon beams and their performance against electron beams should be studied.Yet, during the last decades, almost no work in these areas has been done due to a lack of availablepositron beams, since most of the limited number of positron beams currently generated around the worldare serving high-energy physics experiments and are rarely available, if at all, for other experiments oraccelerator R&D. By providing electron and positron beams, the eSPS facility would therefore enableabsolutely unique and crucial research in at least three areas: • positron production R&D for future lepton colliders; • plasma wakefield acceleration; • photon production in crystals using positrons.The proposed 3.5 GeV electron linac could be used to produce positrons using a conventional fixedtungsten-target or a CLIC-like hybrid-target scheme followed by a capture linac. The CLIC-like hybrid-target is a highly innovative scheme, as the primary electron beam impinges on a single crystal target152 Accelerator Facility Research and Development which is used as a radiator of intense gamma-rays, followed by an amorphous converter target placeddownstream of the crystal. Such a scheme, baseline for the CLIC positron source as well as FCC-ee,promises high-intensity positron bunches. Other challenges related to positron production and capturewould require experimental validation, e.g. the adiabatic damping device and the capture linacs, whichultimately define the initial positron beam parameters.In order to provide high-quality positron bunches with small transverse emittances, the low energypositron beam should be transported to a damping ring to reduce the transverse emittance and to makethem suitable for re-acceleration. The design of such a positron damping ring could be based on theFACET II facility design [37]. Possible parameters of the positron production and beam are collectedin Table 6.3. These parameters have been checked with beam dynamics simulations and are within thecapabilities of the linac described in Sec. 3.Summarising, to operate for production of high-quality high-current positron beams, the facility wouldneed: a positron production target, a capture linac, a positron return line, and a small damping ring. Sucha scheme is thought to be technically feasible but would represent serious integration challenges orrequire some civil engineering to provide the appropriate space for the different components.Table 6.3: Possible parameters for positron production.
Parameter Symbol Value
Electron drive bunchEnergy W σ z µ mPositron bunchEnergy W > σ z µ mCapture energy W c
335 MeVFinal emittance ε <
20 mm mrad
The LEMMA [99] muon collider concept has many attractive features. One possible implementationroute of such a facility at CERN is a phased approach in three stages: • Phase 1: eSPS tests of positron production and targets, and injector studies for LEMMA; • Phase 2: LEP3 where 45 GeV positrons would be available in the LEP/LHC tunnel providing atest ground for muon production with positrons; • Phase 3: A final booster and storage ring for muons which could be the SPS or a larger tunneladapted to the physics requirements pertaining.Alternatives to phase 3 would be the higher energy stages of CLIC or using novel accelerator schemes toreach higher energies in e + e − collisions. The attractiveness of this is that at all three stages high-priorityphysics studies can be performed, and decisions about the future phases will be based on physics resultsand accelerator studies in the previous stages. The initial stage physics at the eSPS is the main subject ofthis section, while the necessity of exploring the Standard Model in detail in e + e − collisions is widelyrecognised. More information about the studies for LEMMA that could be made at the eSPS and aboutthe scenario above can be found in [116]. 153 Accelerator Facility Research and Development
LEMMA is a novel scheme to produce the muons for a muon collider. It uses a positron beam that issent through a target to produce muon pairs. The produced beams have much smaller emittances thanmuon beams produced via pion decay. This could avoid the complexity of cooling the muon beamsand could lead to high luminosities with small muon beam currents. Such a technology could have thepotential to reach very high lepton collision energies.One of the challenges that this approach has to face is the production of very high average positroncurrents, well in excess of what is needed for linear colliders. Positron source R&D is thus instrumentalfor this approach.A key issue of the LEMMA scheme is the stress in the muon production target due to the impingingpositron beam. Different technologies could be considered to overcome this. They range from conven-tional targets from robust materials to crystals or liquid targets. Experimental studies of these targets areof great importance in order to establish whether they are practical for a muon collider.Another challenge that has to be faced is the control of the positron and muon beam emittance. Thebeams pass through the muon production target repeatedly, each time increasing emittance by multiplescattering. Experimental studies of the scattering of positrons (and also electrons) will improve thereliability of the predictions of the emittances of the produced muon beams. It will also allow the testingof specific target shapes and materials that promise improved performances.
As introduced in Sec. 6.4, PWFA is under consideration as a possible high-gradient accelerator techno-logy for an Advanced Linear Collider [117, 118]. The electron beam-driven experiments described inSec. 6.4 are needed to show that the quality of the electron witness bunch can be preserved throughoutthe acceleration process, thus demonstrating a suitable first stage of a plasma-based linear collider (oradvanced linear collider, ALIC).However, a linear collider requires both electron and positron beams, and there are substantial chal-lenges in preserving the quality of the accelerated positron witness bunch in a plasma.
Plasma wakefield acceleration is the only charge-asymmetric acceleration mechanism currently underconsideration as a future accelerator technology. The asymmetry arises from the fact that plasma iscomposed of light electrons and heavy ions. The ion-electron mass ratio ranges from 1836 for hydrogen(commonly used in laser-driven plasma acceleration [119]) to 155,800 for rubidium used in the AWAKEexperiment [105]. When an intense drive bunch propagates into neutral plasma, it induces an oscillationin the mobile plasma electrons, while the plasma ions remain stationary on the time scale of an electronoscillation period (typically measured in femtoseconds or picoseconds depending on the plasma density).For electron beam-driven PWFA in the non-linear regime, this results in complete expulsion of the plasmaelectrons from the region surrounding the drive beam (referred to as blow-out). This region, filled witha uniform background of plasma ions, provided a strong focusing force on the transiting electron beam.The blow-out wake has attractive properties for accelerating a trailing bunch of electrons, namely thelarge gradients and strong focusing fields, which when combined allow for substantial energy gain inthe plasma over distances much larger than the beam vacuum beta-function. Finally, since the ionsare uniformly distributed in the blow-out region, it is possible to exactly calculate the parameters ofa matched witness beam, as shown in Sec. 6.4 and in principle to also maintain the emittance of theaccelerated electron bunch.Such a a pure ion column situation, ideal for an electron bunch, does not exist for a positron bunch.Despite this complication, significant experimental progress has been made over the years that demon-strates the possibility of accelerating positron beams in plasma [120–122], while also characterising theeffects of nonlinear focusing and transverse wakefields on the quality of the accelerated beam [123,154
Accelerator Facility Research and Development
Innovative ideas to use crystals as undulators to produce gamma-rays would need positron beams tovalidate the concept. Positron beams with energy of a few GeV offer unique possibilities in investigatingvarious effects related to interactions between high-energy particle beams and oriented crystals. Besideschannelling radiation and coherent bremsstrahlung, radiation in crystalline undulators reaches its optimalcharacteristics in this energy range [126]. It is expected that a sub-millimetre crystal undulator will allowone to generate a few MeV photon energy radiation with few % spectrum width and the efficiency of onephoton per thousand positrons. Planar channelling and coherent bremsstrahlung in crystals shows a highdegree of linear polarisation, which can be exploited for application related to nuclear physics [127, 128].In case of axial orientation, the emitted radiation becomes harder (10 MeV and more), easily tunable, withnarrow spectrum and can be made circularly polarised [129]. All these innovative sources of radiationcould open new possibilities for studies of both nuclei excitation and disintegration competing withthe large infrastructures, in particular with inverse compton sources, such as ELI-NP, in both size andspectrum width and hardness. Availability of a linac providing a low emittance beam with energy of afew GeV combined with properly designed crystals looks a very promising source of intense high-energyradiation.Furthermore, it could be interesting to exploit bent crystals for GeV positron beam manipulation,for beam extraction, focusing etc. Indeed, in this energy range, studies with bent crystals at INFN-Ferrara have already demonstrated their capability to steer electron beams. In the case of electrons,themaximum deflection efficiency obtained up to now is above 35% at 0.855 GeV [130]. One may expecta much higher deflection efficiency in case of positrons that are less affected by de-channelling thanelectrons. Indeed, being negatively charged, channelled electrons repeatedly oscillate across the nucleiof the crystal, leading to an increase of particle de-channelling compared to positive charges.155
Implementation
The work for this CDR has been done during 2019 and the first months of 2020. Compared to theExpression of Interest [3] in 2018, major work has been done in three main areas, infrastructure andcivil engineering for the linac, final beamline and experimental hall, as well as a revised SPS RF systemsolution. Other changes, to the linac RF layout and SPS specific parts, are also important but the mainsolutions were already presented in the Expression of Interest. The CDR is now a complete descriptionof an eSPS facility and the physics and accelerator R&D potential of it.
A first schedule for the project has also been established and is shown in Fig. 7.1. The schedule driversare: the project approval time; linac construction and commissioning; SPS SRF system design andproduction; construction of the experimental hall/area; and connection to it from the extraction tunnel.The schedule is technically based, but not linked to a specific starting date. This makes the proposedschedule flexible even though some aspects of the work, such as the new transfer tunnel to the experi-mental area, will need to be organised around shutdown periods. Most of the other aspects of the project,in particular in the SPS ring, can be done during short or extended end of year shutdowns.Figure 7.1: Possible eSPS implementation schedule.This program assumes that important investments for the project implementation can be made fromyear 1 with significant deliveries and payments from year 2.
The eSPS facility, i.e. the linac, transfer lines, SPS adaptations and experimental area, has been pre-liminarily costed. These cost estimates were produced by CERN’s groups based on their expertise andexperience from previous projects. The methodology, the level of detail and the resulting uncertaintyon the estimated cost is not completely homogeneous across all aspects included. In particular the riskson the cost of the elements discussed Sec. 7.4 are not included. However, it has been the goal of thisexercise to achieve an uncertainty in the order of ±
30 % on the costs presented here.The top level cost summary is shown in Fig. 7.2. The facility cost is dominated by the linac and source.Ancillary systems account for around a fifth of the cost of the project, with the majority accounting forheating, cooling and ventilation costs. The next largest item of this summary includes the beam transport156
Implementation
Item cost[MCHF]
Source and linac 49.8SPS transfer,acceleration and extraction 23.4Civil engineering 14.0Ancillary systems 23.8
Sum
The linac costs are based on extensive prototype experience within the CLIC project and additionalindustrial quotes for the main components of the volume needed for eSPS, for example klystrons, modu-lators and accelerator structures. With further prototyping of high efficiency klystrons the linac costs canbe reduced by around 10%. If a new source optimised for 800 MHz is constructed the linac costs willincrease by around 15%.
Item cost[MCHF]
RF cavities 11.52Klystrons 11.42Modulators 10.32Waveguides 6.53Vacuum 3.12Controls 2.47Magnet 1.66Instrumentation 1.41Source 1.00Miscellaneous 0.35
Sum
The cost estimates for the electron beam transport and acceleration from 3 . Implementation the 800 MHz superconducting cavities as well as all the supporting systems such as cryostats or poweramplifiers.The power supplies category includes an estimate of the magnet power converters. The powering ofthe new line to transport the beam to the experimental area from the TT10 line is the main contributordue to the large power required. Next are the power converter costs for the SPS injection and extractionsystems, in particular the solid-state generators of the new fast injection kickers. Powering of the linacto SPS transport line is relatively economical due to the low energy and required magnet currents of theline magnets.
Item cost[MCHF]
RF Cavities 7.75Power supplies 3.37Magnets 3.16Instrumentation 2.63Cabling 1.97Vacuum 1.83Miscellaneous 1.50Controls 1.10
Sum . The cost estimate has been based on the layout presented in this CDR and includes costs for detaileddesign, construction and construction management, excluding personnel costs for CERN resources. Theestimate does not include development costs, permits, materials or personnel costs in advance of detaileddesign and construction.A detailed bottom up cost estimate has been carried out for civil engineering work required to im-plement eSPS. The cost estimation process has involved building up costs from labour, equipment andmaterials for each item. The cost for each item is multiplied by the quantity of that item to obtain a totalcost. The total cost has then been compared with other similar projects carried out by CERN in recentyears to ensure the costs are adjusted as accurately as possible to current market rates. The costs are allstated in terms of Q2 2020 prices.In order to produce a cost estimate at this stage of the project, a certain number of assumptions had tobe made: 158
Implementation
Item cost[MCHF]
Injection 4.76Extraction 2.74Experimental Area 6.53
Sum • Ground conditions have been based on 33% poor rock and 67% good rock; • Costs have been based on spoil disposal on CERN land within 20 km of the site with no tippingand disposal costs. If this were to change, the cost increase could be significant; • The proposed drainage can be connected into existing tunnel drainage without significant capacityenhancement or repairs; • Maintenance of TT61 has not been studied in detail. An assumption has been made at this stage thatsuperficial invert and drainage repairs will be required. This must be re-valuated once monitoringand investigation works have been carried out; • The cost of bringing electrical or other service supplies to the experimental hall will be evaluatedfrom scratch with the estimate replacing the provisional sum included to date; • The existing earth mound has been assumed to be composed of unactivated, uncontaminated ’ac-ceptable’ material; • A space requirement of 1500 m is needed to accommodate displaced material from TT4,TT5 andB183. This is a condition of the project and if such space is not available in the flex-storagebuilding or similar, then an additional re-provision cost will be incurred; • Costs of moving or disposal of the material stored in TT4,TT5 and B183 is assumed to be coveredby the owners of the material; • No contingency, risk allowance or optimism bias have been included; • Prices are stated correct for Q2 2020 based on the stated schedule, if there are any delays, thencost inflation must be applied.The accuracy of the estimate is Class 4 - study or feasibility, which could be 15-30% lower or 20-50%higher (in line with [131] as has been used for previous CERN projects). Given the level of uncertaintyand project development, the error bars for this estimate have been set as -20% to +30%). A full list ofassumptions for the costing is noted in the detailed cost estimate.Additional ’below the line’ costs for civil engineering for the maintenance of TT61 in order to put thetunnel back into service are estimated as shown in Fig. 7.5 but excluded from the civil engineering andoverall project cost estimates. 159
Implementation
Cost estimates are required for each discipline included within the infrastructure and civil engineeringparts of the study namely: • Integration; • Heating cooling and ventilation (HVAC); • Electrical; • Radiation protection; • Transport and Handling; • Safety Engineering; • Access Control; • Survey and alignment;Civil Engineering costs are separated out as they represent a large proportion of costs. For each of theother areas, subject experts have carried out individual estimates.Electrical cost estimates have been necessarily derived from comparable projects and considering theconcept design at this stage of the project.It should be noted the following items do not form part of the electrical infrastructure estimate andshall be charged to equipment owners budget estimates: • DC cables for the (inter)connection of power converters and magnets; • Control cables and optical fibres for accelerators components and experimental areas; • Control cables and optical fibres for the access system and radiation monitoring;
Item cost[MCHF]
Cooling and ventilation 13.18Electrical 3.79Access control 3.26Transport 2.02Safety 0.73Radiation protection 0.49Survey 0.36
Sum
Implementation
Table 7.1: Table of estimated non-project ancillary systems costs.
Item cost[MCHF]
TT61, TCC6 and T60 safety systems 0.82BA7 and related infrastructure safety systems 0.59
Sum
The low energy electron beam from the linac or the injector can be used independently from the rest of theaccelerator complex. Any experimental program carried in the linac area or in the TT61 tunnel will runseparately from the other CERN accelerators. Integration and radiation protection studies demonstratedin particular the possibility of concurrent operation of the linac and access to the n_ToF target (seeSec. 5.2.2 and 5.2.4).For higher energy operation the beam will be accelerated in the SPS and the operation constraintsdepend on the RF system selected (see Table 4.2). The superconducting cavity bypass system proposedin this study, imposes a switch time between nominal proton operation and electron operation of at least10 min (see Sec. 4.2.2). The delivery of the high energy electron beam to the experimental area will needto be done alternately with SPS proton operation but could realistically proceed behind LHC physicsstores. The proton physics program making use of the PS or PSB beams will not be impacted by the highenergy electron beam production.
SPS Electron Beam Handling.
Injection of the electron beam at 3 . . Civil Engineering Work Required at the Next Stage of Project Development.
For the project toprogress, additional studies will be required as follows: • Monitoring and investigation of TT61; • Ground penetrating radar (GPR) scans in B183, TT5, TT4 and TTL2 where cores and break-throughs are needed; 161
Implementation • Ground investigation around the new experimental area, new extraction tunnel and at the locationof the CV service building above TT4. • Study to minimise the impact on adjacent beam operations and experiments during constructionwork. • A GPR survey to confirm the location of services around the new experimental area; • A survey and study of existing drainage systems; • An RP study of the area immediately around TT7 and the associated shielding blocks;162
Conclusion
The CERN Council adopted in late spring 2020, an update of the European Strategy for Particle Phys-ics [6]. Some of its priorities are: • Accelerator R&D including plasma wakefield acceleration and other high-gradient acceleratingstructures, bright muon beams, energy recovery linacs ; • An electron-positron Higgs factory is the highest-priority next collider ; • The option of a circular electron-positron Higgs and electroweak factory is possibly a first stageof CERN’s next circular collider, i.e. an accelerator requiring CERN increasing its in-house com-petences on circular electron accelerators; • The quest for dark matter including dark sector candidates, and that such experiments should besupported in laboratories in Europe. • Europe should support future neutrino experiments in Japan and the United States.This conceptional design report describes an infrastructure relevant for all these priorities. The imple-mentation would make excellent use of the investment made in the CLIC programme and is the naturalnext step in the development of X-band high-gradient acceleration technology. The multi-GeV electronbeam from the linac would drive wakefields in the non-linear regime and would with an independentelectron witness bunch demonstrate the applicability of plasma wakefields for high-gradient acceler-ation . The facility would become unique in the world to study collider related challenges, as the onlyfacility with multi-GeV drive bunch and truly independent electron witness bunch. Addition of a positronwitness bunch would make it a complete facility for collider studies. The 800 MHz super-conductingcavities for the eSPS would be the same type as foreseen for a future electron-positron Higgs and elec-troweak factory as the first stage of a next circular collider at CERN. eSPS could be used for the develop-ment and studies of a large number of components and phenomena. The operation of SPS with electronswould train a new generation of CERN staff on circular electron accelerators.The electron beam delivered by this facility would open a dark sector research programme and inparticular provide sensitivity to light dark matter production significantly beyond the targets predictedby a thermal dark matter origin, and for nature of dark matter particles that are not accessible by directdetection experiments. The future long baseline neutrino physics studies need to precisely measureneutrino oscillation probabilities as a function of energy. This critically relies on the ability to modelneutrino-nucleus interactions, and this in turn requires input data on electro-nuclear reactions; the beamfrom this facility would be excellent for such measurements. In addition, it could serve experiments in nuclear physics . The eSPS could be made operational in about five years, and serve a programme asoutlined above. This could start already in LS3 and would operate in parallel and without interferencewith Run 4 at the LHC. 163 eferences
References [1] T. Åkesson et al.,
A primary electron beam facility at CERN (2019),
URL : http://arxiv.org/abs/1905.07657 .[2] T. Akesson et al., Light Dark Matter eXperiment (LDMX) (2018),
URL : https://arxiv.org/abs/1808.05219 .[3] T. Akesson et al., Dark Sector Physics with a Primary Electron Beam Facility at CERN ,tech. rep., CERN-SPSC-2018-023. SPSC-EOI-018, 2018,
URL : http://cds.cern.ch/record/2640784 .[4] LEP design report , CERN, Geneva, 1983,
URL : http://cds.cern.ch/record/98881 .[5] P. N. Burrows et al., eds., Updated baseline for a staged Compact Linear Collider , 2016,
DOI : .[6] The CERN Council, , tech. rep.,2020, URL : https://indico.cern.ch/event/924500/contributions/3884837/subcontributions/308163/attachments/2060582/3456333/CERN-ESU-013.pdf .[7] P.A.R. Ade et al., Planck 2015 results. XIII. Cosmological parameters ,Astron. Astrophys. (2016) A13,
DOI : .[8] E. Izaguirre et al., Testing GeV-Scale Dark Matter with Fixed-Target Missing Momentum Experiments ,Phys. Rev.
D91 (2015) 94026,
DOI : .[9] E. Izaguirre et al., Analyzing the Discovery Potential for Light Dark Matter ,Phys. Rev. Lett. (2015) 251301,
DOI : .[10] B. W. Lee, S. Weinberg, Cosmological Lower Bound on Heavy Neutrino Masses ,Phys. Rev. Lett. (1977) 165, DOI : .[11] C. Boehm, P. Fayet, Scalar dark matter candidates , Nucl. Phys.
B683 (2004) 219,
DOI : .[12] ALEPH Collaboration et al., Precision Electroweak Measurements and Constraints on the Standard Model (2010),
URL : http://arxiv.org/abs/1012.2367 .[13] J. Alexander et al., Dark Sectors 2016 Workshop: Community Report , 2016,
URL : http://inspirehep.net/record/1484628/files/arXiv:1608.08632.pdf .[14] M. Battaglieri et al., US Cosmic Visions: New Ideas in Dark Matter 2017: Community Report (2017),
URL : http://arxiv.org/abs/1707.04591 .[15] M. Pospelov, A. Ritz, M. B. Voloshin, Secluded WIMP Dark Matter ,Phys. Lett.
B662 (2008) 53,
DOI : .[16] M. S. Madhavacheril, N. Sehgal, T. R. Slatyer, Current Dark Matter Annihilation Constraints from CMB and Low-Redshift Data ,Phys. Rev.
D89 (2014) 103508,
DOI : .[17] A. Berlin et al., Dark matter, millicharges, axion and scalar particles, gauge bosons, and othernew physics with LDMX , Physical Review D (2019) 075001, ISSN : 2470-0010,
DOI : , URL : https://link.aps.org/doi/10.1103/PhysRevD.99.075001 .164 eferences [18] T. Åkesson et al., A high efficiency photon veto for the Light Dark Matter eXperiment ,Journal of High Energy Physics (2020),
ISSN : 10298479,
DOI : , URL : http://arxiv.org/abs/1912.05535 .[19] A. Berlin et al., Cosmology and Accelerator Tests of Strongly Interacting Dark Matter ,Phys. Rev.
D97 (2018) 55033,
DOI : .[20] J. D. Bowman et al., An absorption profile centred at 78 megahertz in the sky-averaged spectrum ,Nature (2018) 67,
DOI : .[21] R. Barkana, Possible interaction between baryons and dark-matter particles revealed by the first stars ,Nature (2018) 71,
DOI : .[22] A. Berlin et al., Severely Constraining Dark Matter Interpretations of the 21-cm Anomaly ,Phys. Rev. Lett. (2018) 11102,
DOI : .[23] E. D. Kovetz et al., Tighter limits on dark matter explanations of the anomalous EDGES 21 cm signal ,Physical Review D (2018) 103529, ISSN : 2470-0010,
DOI : , URL : https://link.aps.org/doi/10.1103/PhysRevD.98.103529 .[24] D. Gamba et al., The CLEAR user facility at CERN , Nucl. Instrum. Meth. (2018),
DOI : .[25] A. Latina et al., Evolution of the tracking code placet ,IPAC 2013: Proceedings of the 4th International Particle Accelerator Conference, 2013,p. 1014,
ISBN : 9783954501229,
URL : https://cds.cern.ch/record/1574813 .[26] H. Zha et al., Beam-based measurements of long-range transverse wakefields in the CompactLinear Collider main-linac accelerating structure , Phys. Rev. Accel. Beams (2016), DOI : http://dx.doi.org/10.1103/physrevaccelbeams.19.011001 .[27] K. Floettmann, A Space Charge Tracking Algorithm , tech. rep., Hamburg: Desy, 2000,
URL : .[28] M. Aicheler et al., The Compact Linear Collider (CLIC) - Project Implementation Plan ,tech. rep., 2019,
DOI : .[29] M. Gasior, An Inductive Pick-Up for Beam Position and Current Measurements , Dipac,January 2003, 2003, p. 53,
URL : http://accelconf.web.cern.ch/accelconf/d03/papers/CT01.pdf .[30] F. J. Cullinan et al., Long bunch trains measured using a prototype cavity beam position monitorfor the Compact Linear Collider ,Physical Review Special Topics - Accelerators and Beams (2015) 112802, ISSN : 1098-4402,
DOI : , URL : https://link.aps.org/doi/10.1103/PhysRevSTAB.18.112802 .[31] B. Bolzon et al., Very high resolution optical transition radiation imaging system: Comparisonbetween simulation and experiment ,Physical Review Special Topics - Accelerators and Beams (2015) 082803, ISSN : 10984402,
DOI : .[32] C. Welsch, E. Bravin, T. Lefevre, Optimization of OTR screen surface materials and OTR screen geometry at CTF3 ,Reviews on Advanced Materials Science (2007) 73, URL : .165 eferences [33] M. Olvegârd et al., Beam profile monitoring at the test beam line at the Compact Linear Collider test facility ,Phys. Rev. ST Accel. Beams (2013) 82802, DOI : .[34] K. Lekomtsev et al., Coherent Diffraction Radiation as a Tool for Longitudinal Beam Profile Diagnostics at CTF3.
Proceedings, 25th International Linear Accelerator Conference, LINAC2010: Tsukuba, Japan,September 12-17, 2010, 2011, TUP099,
URL : http://accelconf.web.cern.ch/AccelConf/LINAC2010/papers/tup099.pdf .[35] R. Pan et al., Reprint of: Coulomb field strength measurement by electro-optic spectral decodingsystem at the CALIFES beam line , Nuclear Instruments and Methods in Physics Research,Section A: Accelerators, Spectrometers, Detectors and Associated Equipment (2016) 545,
ISSN : 01689002,
DOI : .[36] M. Aicheler et al., A Multi-TeV Linear Collider Based on CLIC Technology ,CERN Yellow Reports: Monographs (2012),
DOI : .[37] V. Yakimenko et al., FACET-II Accelerator Research with Beams of Extreme Intensities ,Proceedings, 7th International Particle Accelerator Conference (IPAC 2016): Busan, Korea,May 8-13, 2016, 2016, TUOBB02,
DOI : .[38] K. Pepitone et al., The electron accelerators for the AWAKE experiment at CERN—Baseline andFuture Developments , Nuclear Instruments and Methods in Physics Research, Section A:Accelerators, Spectrometers, Detectors and Associated Equipment (2018) 102,
ISSN : 01689002,
DOI : .[39] W. Herdrich, H. P. Kindermann, RF power amplifier for the CERN SPS operating as LEP injector ,IEEE Transactions on Nuclear Science (1985) 2794, ISSN : 15581578,
DOI : , URL : https://cds.cern.ch/record/160202 .[40] P. E. Faugeras et al., The new RF system for lepton acceleration in the CERN SPS ,Proceedings, Particle Accelerator Conference (PAC1987), 1987, p. 1719,
URL : https://cds.cern.ch/record/177291 .[41] P. E. Faugeras et al., Installation and operation of the new RF system for Lepton acceleration in the CERN SPS ,CERN-SPS-87-19-ARF, 1989, p. 106,
DOI : , URL : https://cds.cern.ch/record/177291 .[42] H. Damerau et al., LHC Injectors Upgrade, Technical Design Report , tech. rep., CERN, 2014,
URL : https://cds.cern.ch/record/1976692 .[43] T. P. R. Linnecar, E. N. Shaposhnikova, Resonant impedance in the SPS , SL-Note 96-49 RF,SL-Note-96-49-RF, CERN, Geneva, 1996,
URL : https://cds.cern.ch/record/703728 .[44] D. Boussard et al., Design Considerations for the LHC 200 MHz RF System , tech. rep., 2000,
URL : https://cds.cern.ch/record/424847 .[45] R. Calaga, O. Capatina, G. Vandoni, The SPS Tests of the HL-LHC Crab Cavities (2018) TUPAF057. 4 p,
DOI : , URL : https://cds.cern.ch/record/2649807 .166 eferences [46] R. Calaga, SRF for Future Circular Colliders FCC Study Scope (2015) 1,
DOI : , URL : http://cds.cern.ch/record/2288285 .[47] S. Zadeh, Accelerating cavity and HOM doupler design for the Future Circular Collider ,PhD thesis, 2019.[48] F. Marhauser,
Recent results on a multi-cell 800 MHz bulk Nb cavity , FCC WEEK 2018, 2018,
URL : https://indico.cern.ch/event/656491/contributions/2932251/ .[49] M. J. Barnes et al., Measurement and analysis of SPS kicker magnet heating and outgassingwith different bunch spacing (2009) 4 p,
URL : https://cds.cern.ch/record/1199759 .[50] B. Balhan et al., Improvement of the CERN SPS electrostatic septa ion traps , 2016 27thInternational Symposium on Discharges and Electrical Insulation in Vacuum (ISDEIV), vol. 2,2016, p. 1,
DOI : .[51] T. Kaltenbacher, J. Repond, C. Vollinger, Characterization of resonant impedances of CERN-SPS gate valves ,Journal of Physics: Conference Series, vol. 874, 1, 2017, 012073. 7 p,
DOI : , URL : https://cds.cern.ch/record/2276051 .[52] K. Cornelis, R. Schmidt, Commissioning of the SPS as LEP injector (1988) 4 p,
URL : https://cds.cern.ch/record/188928 .[53] K. Cornelis, R. Schmidt, The performance of the SPS as LEP injector (1990),
URL : https://cds.cern.ch/record/217377 .[54] Y. Papaphilippou, R. Corsini, L. Evans, The SPS as an Ultra-Low Emittance Damping Ring Test Facility for CLIC (2013) 3 p,
URL : https://cds.cern.ch/record/1588122 .[55] A. Chao, Physics of collective beam instabilities in high energy accelerators , 1993,
URL : https://cds.cern.ch/record/246480 .[56] F. Zimmermann et al., Collective effects in the CLIC damping rings ,Proceedings of the IEEE Particle Accelerator Conference, vol. 2005, 2005, p. 1312,
ISBN : 0780388593,
DOI : .[57] M. Ferrario, M. Migliorati, L. Palumbo, Wakefields and Instabilities in Linear Accelerators ,Proceedings of the CAS-CERN Accelerator School: Advanced Accelerator Physics (2016) 357,
DOI : , URL : http://arxiv.org/abs/1601.05216 .[58] G. Rumolo et al., Collective effects in the CLIC damping rings , tech. rep., 2008,
URL : https://cds.cern.ch/record/1122957 .[59] A. Wolski, Beam dynamics: In high energy particle accelerators , Imperial College Press, 2014,p. 1,
ISBN : 9781783262786,
DOI : .[60] E. Koukovini-Platia et al., Impedance effects in the CLIC damping rings , tech. rep., 2011,p. 1111,
URL : https://infoscience.epfl.ch/record/170542 .[61] R. Nagaoka, Ions , CERN Yellow Reports: School Proceedings, vol. 3, November, CERN, 2017,p. 519,
ISBN : 9789290834625,
DOI : , URL : https://doi.org/10.23730/CYRSP-2017-003.519 .167 eferences [62] J. Byrd et al., Ion instability experiments on the ALS , AIP Conference Proceedings, vol. 395,AIP, 1997, p. 175,
DOI : , URL : http://aip.scitation.org/doi/abs/10.1063/1.52931 .[63] M. Ferrario, M. Migliorati, L. Palumbo, Space Charge Effects , ISBN : 978-92-9083-411-3,
DOI : , URL : http://dx.doi.org/10.5170/CERN-2014-009.331 .[64] F. Zimmermann, Estimates of CSR Instability Thresholds for Various Storage Rings , tech. rep.,2010,
URL : https://cds.cern.ch/record/1306778 .[65] K. L. Bane, Y. Cai, G. Stupakov, Threshold studies of the microwave instability in electron storage rings ,Physical Review Special Topics - Accelerators and Beams (2010) 104402, ISSN : 10984402,
DOI : , URL : https://journals.aps.org/prab/abstract/10.1103/PhysRevSTAB.13.104402 .[66] A. W. Chao et al., Handbook of accelerator physics and engineering, second edition ,World Scientific Publishing Co., 2013, p. 1,
ISBN : 9789814415859,
DOI : , URL : .[67] M. Lonza, H. Schmickler, Multi-bunch feedback systems ,CERN Yellow Reports: School Proceedings, vol. 3, November, CERN, 2017, p. 471,
ISBN : 9789290834625,
DOI : .[68] D. Brandt et al., Beam dynamics effects in the CERN SPS used as a Lepton injector ,Publ by IEEE, 1989, p. 1205,
DOI : .[69] M. Wendt et al., Upgrade of the CERN SPS Beam Position Measurement System , Proceedings,9th International Particle Accelerator Conference (IPAC 2018): Vancouver, BC Canada, 2018,WEPAF085,
DOI : .[70] M. Krupa, M. Gasior, The Wall Current Transformer - a New Sensor for PreciseBunch-by-Bunch Intensity Measurements in the LHC , Proceedings, 5th International BeamInstrumentation Conference (IBIC 2016): Barcelona, Spain, September 11-15, 2016, 2017,WEAL02,
DOI : .[71] S. Takano, M. Masaki, H. Ohkuma, X-ray imaging of a small electron beam in a low-emittance synchrotron light source ,Nuclear Instruments and Methods in Physics Research, Section A: Accelerators, Spectrometers,Detectors and Associated Equipment (2006) 357,
ISSN : 01689002,
DOI : .[72] T. Naito, T. Mitsuhashi, Very Small Beam-Size Measurement By a Reflective Synchrotron Radiation Interferometer ,Phys. Rev. ST Accel. Beams (2006) 122802, DOI : .[73] M. Siano et al., Characterizing temporal coherence of visible synchrotron radiation withheterodyne near field speckles , Physical Review Accelerators and Beams (2017), ISSN : 24699888,
DOI : .[74] S. Mazzoni, Progress on transverse beam profile measurement using the Heterodyne Near FieldSpeckles method at ALBA ,Proceedings, 7th International Beam Instrumentation Conference, Shanghai, China, 2018,
DOI : .[75] C. Heßler et al., JACoW : Beam Simulation Studies for the Upgrade of the SPS Beam Dumping System (2018) TUPAF031,
DOI : .168 eferences [76] H. Stockhorst et al., Beam shaping for the ultra slow extraction and stochastic cooling at COSY ,Nonlinear and stochastic beam dynamics in accelerators: A challenge to theoretical andcomputational physics. Proceedings, Workshop, Lueneburg, Germany, September 29-October 3,1997, 1997, p. 356,
URL : https://inspirehep.net/literature/456048 .[77] F. Frommberger et al., Single Electron Extraction at the ELSA Detector Test Beamline ,Proceedings, 7th International Particle Accelerator Conference (IPAC 2016): Busan, Korea,May 8-13, 2016, 2016, THPOW029,
DOI : , URL : http://inspirehep.net/record/1470674/files/thpow029.pdf .[78] F. Velotti et al., Non-local Fast Extraction from the CERN SPS at 100 and 440 GeV ,Proceedings, 4th International Particle Accelerator Conference (IPAC 2013): Shanghai, China,May 12-17, 2013, 2013, MOPFI050,
URL : http://jacow.org/IPAC2013/papers/mopfi050.pdf .[79] T. Lefevre et al., Non-invasive beam diagnostics with Cherenkov diffraction radiation , Proc. 9thInternational Particle Accelerator Conference (IPAC’18), Vancouver, BC, Canada, Apr. 4, 2018,
DOI : , URL : https://cds.cern.ch/record/2649944 .[80] R. Kieffer et al., Direct Observation of Incoherent Cherenkov Diffraction Radiation in the Visible Range ,Phys. Rev. Lett. (2018) 54802,
DOI : http://dx.doi.org/10.1103/PhysRevLett.121.054802 .[81] I. Ruiz, A multipurpose scintillating fibre beam monitor for the measurement of secondarybeams at CERN ,Proceedings, 7th International Beam Instrumentation Conference, Shanghai, China, 2018,
DOI : , URL : https://inspirehep.net/literature/1736073 .[82] EDMS, Drawing GENEM0033 , URL : https://edms5.cern.ch/cedar/tmpfiles/preview16004_em0033.pdf .[83] CERN, Safety Code F – Radiation Protection , 2006,
URL : https://edms.cern.ch/document/335729 .[84] G. Battistoni et al., Overview of the FLUKA code , Annals of Nuclear Energy (2015) 10, ISSN : 03064549,
DOI : , URL : https://linkinghub.elsevier.com/retrieve/pii/S0306454914005878 .[85] T. Böhlen et al., The FLUKA Code: Developments and Challenges for High Energy and Medical Applications ,Nuclear Data Sheets (2014) 211,
ISSN : 00903752,
DOI : , URL : https://linkinghub.elsevier.com/retrieve/pii/S0090375214005018 .[86] V. Vlachoudis, FLAIR: A powerful but user friendly graphical interface for FLUKA ,Proceedings of the International Conference on Mathematics, Computational Methods &Reactor Physics (M&C 2009), American Nuclear Society, Saratoga Springs, New York, 2009,
URL : https://inis.iaea.org/search/search.aspx?orig_q=RN:42064858 .[87] M. Widorski, Radiation protection studies for the eSPS project , tech. rep., CERN, 2020,
URL : https://edms.cern.ch/document/2363125 .[88] D. Forkel-Wirth, T. Otto, Area classification , tech. rep., CERN, 2006,
URL : https://edms.cern.ch/document/810149 .169 eferences [89] Travail à la chaleur. Ce qu’il faut retenir - Risques - INRS , URL : .[90] CERN-HSE, Politique de Réduction de l’Empreinte Sonore et Stratégie d’Implémentation |Document 2102605 (v.2.00) , URL : https://edms.cern.ch/ui/ .[91] N. Galindo Munoz et al., JACoW : Pre-Alignment Techniques Developments and MeasurementResults of the Electromagnetic Center of Warm High-Gradient Accelerating Structures (2017) TUPIK076,
DOI : .[92] D. Caiazza et al., New solution for the high accuracy alignment of accelerator components ,Physical Review Accelerators and Beams (2017) 083501, ISSN : 24699888,
DOI : .[93] L. A. Dougherty, Electrons in the SPS - 5th ICE Working Group: Integration Update , URL : https://indico.cern.ch/event/824438/contributions/3643987/ .[94] D. Banon-Caballero et al., Dark Current Analysis at CERN’s X-band Facility ,Proc. 10th International Particle Accelerator Conference (IPAC’19),Melbourne, Australia,19-24 May 2019, International Particle Accelerator Conference 10,JACoW Publishing, Geneva, Switzerland, 2019, p. 2944,
ISBN : 978-3-95450-208-0,
DOI : doi:10.18429/JACoW-IPAC2019-WEPRB059 , URL : http://jacow.org/ipac2019/papers/weprb059.pdf .[95] Common fire protection requirements on water network and ventilation , 2017,
URL : https://edms.cern.ch/document/1854596/1 .[96] M. Koratzinos, FCC-ee accelerator parameters, performance and limitations ,Nucl. Part. Phys. Proc. (2016) 2326,
DOI : .[97] A. Blondel, F. Zimmermann, A High Luminosity e+e- Collider in the LHC tunnel to study the Higgs Boson (2011),
URL : http://arxiv.org/abs/1112.2518 .[98] F. Zimmermann, O. Bruning, M. Klein, The LHeC as a Higgs Boson Factory , Proceedings, 4thInternational Particle Accelerator Conference (IPAC 2013): Shanghai, China, May 12-17, 2013,2013, MOPWO054,
URL : http://jacow.org/IPAC2013/papers/mopwo054.pdf .[99] M. Boscolo, J. P. Delahaye, M. Palmer, The future prospects of muon colliders and neutrino factories ,Submitted to: Rev. Accel. Sci. Tech. (2018),
DOI : , URL : http://arxiv.org/abs/1808.01858 .[100] A. Lagzda et al., Very-High Energy Electron (VHEE) Studies at CERN’s CLEAR User Facility ,Proceedings, 9th International Particle Accelerator Conference (IPAC 2018): Vancouver, BCCanada, 2018, MOPML023,
DOI : .[101] B. I. Grishanov et al., ATF2 Proposal (2005),
URL : https://cds.cern.ch/record/855957 .[102] A. Latina et al., CompactLight Design Study (2018) 85,
DOI : .170 eferences [103] M. Ferrario et al., EuPRAXIA@SPARC_LAB Design study towards a compact FEL facility at LNF ,Nuclear Instruments and Methods in Physics Research, Section A: Accelerators, Spectrometers,Detectors and Associated Equipment (2018),
ISSN : 01689002,
DOI : .[104] M. Diomede et al., RF Design of the X-band Linac for the EuPRAXIA@SPARC_LAB Project ,Proceedings, 9th International Particle Accelerator Conference (IPAC 2018): Vancouver, BCCanada, 2018, p. 4422,
DOI : .[105] E. Adli et al., Acceleration of electrons in the plasma wakefield of a proton bunch ,Nature (2018) 485,
DOI : , URL : .[106] T. Behnke et al., The International Linear Collider Technical Design Report - Volume 1: Executive Summary (2013),
URL : http://arxiv.org/abs/1306.6327 .[107] I. Blumenfeld et al., Energy doubling of 42 GeV electrons in a metre-scale plasma wakefield accelerator ,Nature (2007) 741,
ISSN : 14764687,
DOI : .[108] M. Litos et al., High-efficiency acceleration of an electron beam in a plasma wakefield accelerator ,Nature (2014) 92,
ISSN : 14764687,
DOI : .[109] P. Muggli et al., Photo-ionized lithium source for plasma accelerator applications ,IEEE Transactions on Plasma Science (1999) 791, ISSN : 0093-3813,
DOI : .[110] R. Corsini et al., First Experiments at the CLEAR User Facility , Proc. 9th International ParticleAccelerator Conference (IPAC’18), Vancouver, BC, Canada, Apr. 4, 2018, p. 4066,
DOI : .[111] K. N. Sjobak et al., Status of the CLEAR electron beam user facility at CERN (2019),
DOI : .[112] M. Tali et al., High-energy Electron Induced SEUs and Jovian Environment Impact ,IEEE Transactions on Nuclear Science (2017) 1, ISSN : 0018-9499,
DOI : , URL : http://ieeexplore.ieee.org/document/7944600/ .[113] C. A. Lindstrøm et al., Overview of the CLEAR plasma lens experiment ,Nuclear Instruments and Methods in Physics Research, Section A: Accelerators, Spectrometers,Detectors and Associated Equipment (2018) 379,
ISSN : 01689002,
DOI : .[114] C. A. Lindstrøm et al., Emittance Preservation in an Aberration-Free Active Plasma Lens ,Physical Review Letters (2018) 194801,
ISSN : 10797114,
DOI : , URL : https://journals.aps.org/prl/abstract/10.1103/PhysRevLett.121.194801 .[115] R. Kieffer et al., Experimental Observation of "shadowing" in Optical Transition Radiation ,Physical Review Letters (2018),
ISSN : 10797114,
DOI : .171 eferences [116] M. Boscolo et al., Low emittance muon accelerator studies with production from positrons on target ,Phys. Rev. Accel. Beams (2018) 61005, DOI : , URL : https://link.aps.org/doi/10.1103/PhysRevAccelBeams.21.061005 .[117] B. Cros, P. Muggli, Towards a Proposal for an Advanced Linear Collider , ICFA, 2017,
ISBN : 9789290834687,
URL : https://inspirehep.net/literature/1643957 .[118] Advanced Accelerator Development Strategy Report: DOE Advanced Accelerator ConceptsResearch Roadmap Workshop , tech. rep., USDOE Office of Science (SC) (United States), 2016,
DOI : , URL : .[119] W. P. Leemans et al., Multi-GeV Electron Beams from Capillary-Discharge-GuidedSubpetawatt Laser Pulses in the Self-Trapping Regime ,Physical Review Letters (2014) 245002,
ISSN : 0031-9007,
DOI : , URL : https://link.aps.org/doi/10.1103/PhysRevLett.113.245002 .[120] B. E. Blue et al., Plasma-Wakefield Acceleration of an Intense Positron Beam ,Phys. Rev. Lett. (2003) 214801, DOI : , URL : https://link.aps.org/doi/10.1103/PhysRevLett.90.214801 .[121] S. Corde et al., Multi-gigaelectronvolt acceleration of positrons in a self-loaded plasma wakefield ,Nature (2015) 442,
ISSN : 14764687,
DOI : , URL : http://dx.doi.org/10.1038/nature14890 .[122] Gessner Spencer et al., Demonstration of a positron beam-driven hollow channel plasma wakefield accelerator ,Nature Communications (2016) 11785, URL : http://dx.doi.org/10.1038/ncomms11785 .[123] P. Muggli et al., Halo Formation and Emittance Growth of Positron Beams in Plasmas ,Phys. Rev. Lett. (2008) 55001,
DOI : , URL : https://link.aps.org/doi/10.1103/PhysRevLett.101.055001 .[124] C. A. Lindstrøm et al., Measurement of Transverse Wakefields Induced by a Misaligned PositronBunch in a Hollow Channel Plasma Accelerator , Physical Review Letters (2018) 124802,
ISSN : 0031-9007,
DOI : , URL : https://link.aps.org/doi/10.1103/PhysRevLett.120.124802 .[125] E. Adli et al., A Beam Driven Plasma-Wakefield Linear Collider: From Higgs Factory to Multi-TeV ,Proceedings of the Snowmass Process CSS2013, SLAC-PUB-15426, 2013,
ISBN : 9783954501328,
DOI : , URL : .[126] V. V. Tikhomirov, A Benchmark Construction of Positron Crystal Undulator ,e-Print arXiv:1502.06588v1 (2015),
URL : http://arxiv.org/abs/1502.06588 .[127] D. Lohmann et al., Linearly polarized photons at MAMI (Mainz) ,Nuclear Instruments and Methods in Physics Research Section A: Accelerators, Spectrometers,Detectors and Associated Equipment (1994) 494,
ISSN : 01689002,
DOI : , URL : https://linkinghub.elsevier.com/retrieve/pii/0168900294902305 .172 eferences [128] C. A. Paterson et al., Photoproduction of Λ and Σ ,Physical Review C (2016) 065201, ISSN : 2469-9985,
DOI : , URL : https://link.aps.org/doi/10.1103/PhysRevC.93.065201 .[129] V. Guidi, A. Mazzolari, V. V. Tikhomirov, How to make coherent bremsstrahlung circularly polarized , Nuova Cimento C. 55, 2011, p. 55,
URL : https://inis.iaea.org/search/searchsinglerecord.aspx?recordsFor=SingleRecord&RN=43053370 .[130] A. I. Sytov et al., Steering of Sub-GeV electrons by ultrashort Si and Ge bent crystals ,The European Physical Journal C (2017) 901, ISSN : 1434-6044,
DOI : , URL : http://link.springer.com/10.1140/epjc/s10052-017-5456-7 .[131] P. Christensen et al., Cost Estimate Classification system-as applied in engineering,procurement, and construction for the process industries , AACE, Inc (2005),
URL :