Multipactor suppression in dielectric-assist accelerating structures via diamond-like carbon coatings
MMultipactor suppression in dielectric-assist accelerating structures via diamond-likecarbon coatings
Shingo Mori ∗ and Mitsuhiro Yoshida † KEK Accelerator department, Tsukuba, Ibaraki 305-0801, Japan
Daisuke Satoh ‡ Radiation Imaging Measurement Group Research Institute for Measurement and Analytical Instrumentation,National Metrology Institute of Japan, National Institute of Advanced Industrial Science and Technology (AIST) (Dated: September 29, 2020)Multipactor discharges are widely observed in the accelerator field, and their suppression hasbeen studied to improve accelerator performance. A dielectric-assist accelerating (DAA) cavity isa standing-wave accelerating cavity that attains a Q-value of over 100,000 at room temperature byusing the reflection of the dielectric layer; the DAA cavity is expected to be used with a small RFpower source and high-duty operation. Thus far, the maximum accelerating field of DAA cavitieshas been limited to a few MV/m by the multipactor. By applying a diamond-like carbon (DLC)coating to reduce the secondary electron emission coefficient without sacrificing the Q-value of thecavity, we have demonstrated that a 6 [ µ s] RF pulse can be injected into a DAA cavity with a fieldof more than 10 [MV/m] while suppressing the multipactor. Multipactors are a widespread phenomenon in the ac-celerator field and have appeared in the developmentof RF windows, accelerator tubes, and beam pipes. Inthe multipactor phenomenon, the number of electronsincreases exponentially owing to repeated resonant col-lisions of electrons on the surface of a material in analternating electric field; it occurs when the secondaryelectron yield (SEY) of the material is greater than one.There are the several classes of multipactor suppressionmethods that aim to reduce the SEY through applica-tion of a coating of low SEY material such as TiN [1, 2]or amorphous carbon [3–5], or by grooving [6], reducingthe tangential electric field on the surface, or applying asolenoidal magnetic field [7, 8].The RF accelerating cavity is widely used to generatehigh-energy beams in numerous scientific and technolog-ical applications. One of the figures of merit of an RFcavity is the Q-value, which represents the power effi-ciency of the cavity and is determined according to theconductivity of the material and the structure. The su-perconducting cavity has a high Q-value of greater thanorder O (10 ) at low temperature T < . O (10 ) work-ing at room temperature, as determined by the ohmic lossof the metal. Recently, it was proven that a dielectric-assist accelerating (DAA) structure could attain a highQ-value of O (10 ) at room temperature [9, 10] througha reduction in the surface magnetic field using the reflec-tion of a dielectric layer exceeding the limits of the Q-value determined by the ohmic loss. Figure 1 (a) showsa five-cell DAA cavity with a Q ∼ . × , which is oneorder of magnitude larger than the typical Q-value of a ∗ Electronic address: [email protected] † Electronic address: [email protected] ‡ Electronic address: [email protected] normal conducting cavity.The maximum accelerating field of a dielectric-basedaccelerating cavity is limited by multipactors and break-downs. [11] Numerous high-power tests of X-banddielectric-loaded accelerating structures (DLA) havebeen performed, and accelerating fields of up to 20 MV/mhave been achieved by applying an axial magnetic field tosuppress multipactors. [8] The study of photonic bandgapcavities using sapphire rods shows a maximum electricfield of 19.1 MV/m, which is limited by breakdowns intriple junctions. [12] In both examples, short RF pulses of100 [ns] to 500 [ns] in the X-band are fed in. In contrast,in a DAA cavity, relatively long RF pulses of several µs are fed in at a lower frequency, the C-band. In the caseof our DAA cavity, the maximum value of the acceler-ating field was limited to approximately 1 or 2 MV/mwithout the SEY countermeasure. Attempts have beenmade to suppress the multipactor effect in DLA struc-tures by lowering the SEY with TiN coating. However,in the case of DLA with TiN coating, it is not possible tocompletely suppress the multipactor. [13] Recently, it hasbeen shown that a diamond-like carbon (DLC) coatingis relatively more resistant to multipacting than TiN asa coating for the inner surface of a metal standing-waveaccelerating cavity in the X-band. [14].The ideal coating for a DAA cavity will be a materialthat has both very low electrical conductivity and SEY.A coating of a material with high electrical conductivity,such as the TiN coating, reduces the Q-value of the DAAcavity by increasing the ohmic loss on the surface of thedielectric cells. In fact, as shown in Table I, when aTiN coating with a thickness of 10 [nm] was applied toboth sides of the ceramic cells of the five-cell DAA cavityfabricated in our previous work [10], the reduction in theQ-value was approximately 40 [%]. A TiN coating of afew nanometers in thickness will not suffice because it willbe displaced by the electrical discharge occurring duringconditioning. [13] a r X i v : . [ phy s i c s . acc - ph ] S e p Coating f [MHz] β Q w/o coating (set 1) 5708 .
29 1 . .
01 0 .
79 64000w/o coating (set 2) 5717 .
10 0 .
93 113000DLC (set 2) 5717 .
07 1 . . µ m] was applied on both sides of allcells in set 2. The Q-value was measured via the couplerand mode converter shown in Figure 1 (c) from S using theAgilent N5230A network analyzer. DLC has a low SEY and low electrical conductivity,and is composed of sp (graphene-like) and sp (diamond-like) bounds of carbons. The SEY of DLC coating is lessthan 1.5. [15] We also measured the SEY of the DLCcoating on an MgO sample with a thickness of 0 . µ m]using an SEM-based device at KEK [16]. Then, in theincident energy range between 0 . . . .
7, and the SEY was δ < . . µ m] thickness on both sides of the ceramiccells of the five-cell DAA cavity resulted in a Q-valuechange of approximately 3 [%], which is within the errorof low-power measurements.Figure. 1 (e) shows the MgO cell after application ofthe DLC coating. We applied DLC coating processedby the plasma CVD method with a thickness of 0 . µ m],hydrogen content of 5 ∼
50 [%], and sp / (sp + sp ) ratioof 20 ∼
50 [%].In this study, we performed a high-power test of thetwo-cell DAA cavity with low-SEY and low-conductivityDLC coating. The various conditions are described asfollows.To focus on the effect of the surface treatment on theaccelerating gradient, we fabricated a DAA test cavitywith two cells. In order to implement the translationalsymmetric field configuration shown in Figure 1 (b) andreduce the tangential electric field on the surface of theDAA cells, we did not adopt the end cell structure tomitigate the power loss at the copper endplate, as dis-cussed in [9]. The Q-value of our fabricated test cavitywas Q ∼ . × , which is approximately 25 [%] lowerthan the Q-values of the simulations assuming electricalconductivity ρ = 1 . × − [Ω / m] and dielectric losstan δ = 6 × − . This is due to the insufficient contactand surface treatment of the metal cavity and end-plates.As shown in Figure 1 (c), the test cavity is composed of a copper end-plate, copper cylinder, coupler, the modeconverter, and dielectric cell. Figure 1 (d) shows theMgO cell after machining with tolerances from 0 .
01 [mm]to 0 .
07 [mm]. The MgO ceramics had the same charac-teristics as those used in the previous work [10], i.e., arelative permittivity of (cid:15) r = 9 .
64 and a loss tangent oftan δ = 6 × − .Figure 1 (f) is an RF schematic showing the powersource, waveguide, acceleration cavity, and measurementsystem that comprise the C-band test stand. The C-bandRF was produced by a Tektronix TSG 4106A signal gen-erator (SG) and amplified by a solid-state amplifier andthe klystron. The DAA cavity was installed in the vac-uum chamber and connected to the rectangular waveg-uide via the mode converter. We placed a Faraday cupin front of the mode converter to measure the dark cur-rent emitted from the beam hole in the coupler side. Weplaced a pickup antenna composed of semi-rigid coax ca-ble outside the copper endplate, along the beam axis.The stored power P c coupled to the antenna was trans-mitted through the cable, the attenuator (Lucas Wein-schel Model 47), and a Keysight 8472B crystal diode,which generated a rectified voltage to be monitored byan oscilloscope (Iwatsu DS-5514). The measurement us-ing the signal generator indicated that the attenuatordamped the input RF power by 41 [dB].In the following, we describe the in situ measurementsof the Q-value and the transmission coefficient repre-senting the coupling of the antennas. As shown in Fig-ure 1 (f), the 4-port circulator is equipped with two vac-uum windows, and the section between them is at atmo-spheric pressure and can be removed. Thus, we couldperform low-power measurements of the cavity throughthe waveguide from the position of the vacuum window.The transmission coefficient S from the waveguide tothe antenna and coaxial cable (Huber & Suhner Sucoflex104) could be measured directly to obtain information onthe coupling between the antenna and the cavity, includ-ing the transmission loss of the coaxial cable. Neglectingthe transmission loss by the rectangular waveguide, wecan obtain the ratio of the power observed by the an-tenna via the coaxial cable and that stored inside thecavity from the transmission coefficient S . Coating f [MHz] Q L ( × ) β Q ( × ) S [dB]uncoated 5736 .
81 1 . . . . .
72 1 . . . . .
65 0 .
88 2 . . . Table II shows the results of the low-power measure-ment described above. The rows represent the coat-ing conditions, i.e., without coating (uncoated) and withDLC coating with a thickness of 0 . µ m] (DLC1, DLC2).Different pairs of dielectric cells were used in each condi-tion, and the dimensional errors in the cells shifted the FIG. 1: (a) Schematic representation of the original DAA cavity [10]. (b) Cross-section of the electric field of the acceleratingmode calculated using
SuperFish . (c) Development view of the two-cell cavity. (d) MgO cell without coating and (e) withDLC coating with a thickness of 0 . µ m]. (f) RF system of the high-power test from the klystron to the measurement of storedpower through the antenna. resonant frequency and coupling. f is the resonant fre-quency of the cavity. Q L = Q / (1 + β + β a ) is the loadedQ-value, where the coupling between the cavity and theantenna satisfies β a (cid:28) β is the coupling throughthe iris of the coupler.The stored power P c can be reconstructed via two inde-pendent methods using the measurements from the low-power test.In the first method, the transient model is used to com-pute P c ( t ) from the input waveforms P in ( t ), and the mea-sured loaded Q value Q L . We assume that there is noloss due to time-varying multipactors in the cavity andthen Q L is constant. By using the relation of power con-servation dW c /dt = P in − P ref − P Bc , Q = ωW c /P B c , P in = G wg V , P ref = G wg ( V c − V in ) , P Bc = ( G wg /β ) V ,the cavity voltage V c , and waveguide voltage follow T f dV c dt + V c = 2 β β V in , (1)where T f = Q L / ( πf ) and G wg denotes the waveguideadmittance. By substituting the measured time variationof the incident power P in ( t ) into Eq. (1), we obtain thestored power P Bc ( t ). The second method reconstructs the stored power P Ac using the transmission coefficients S , attenuator, andcrystal diode. Using the signal generator, we measuredthe relation between the input power and the output volt-age, f dBm (ln V ), for the crystal diode and RF power me-ters, where V is the voltage measured by the oscilloscope.Then, we can reconstruct the stored power P Ac as P c = 1 [mW] · f dBm (ln V ) / · − S [dB] / · ∆ Att / , (2)where ∆ Att ∼
41 is the power reduction measured by theattenuator.From the simulation using
Superfish , the fraction ofshunt impedance per unit length in the Q-value is ob-tained as
Z/Q = 4 .
25 [kΩ / m], which is independentof the properties of the materials. To obtain the shuntimpedance per unit length Z , we multiplied the mea-sured Q by the fraction. Combining the above quan-tities, we can obtain the peak accelerating voltage as E z, max = (cid:112) Z · P c /L , where we use a cavity length of L = 52 .
485 [mm].The repetition rate of the RF pulses is 50 [Hz]. Duringthe experiment, the RF frequency, magnitude of the in-cident power, and pulse length of the RF input from theklystron to the DAA cavity can be changed at any time.We investigated the upper limit of the stored power by in-creasing the incident power from approximately 100 [W]to 100 [kW]. In this process, the RF frequency and pulselength were repeatedly adjusted to maximize the powerwaveform in the cavity measured by the antenna. In thecase of DLC1 and DLC2, the conditioning time to reachan accelerating field of 10 MV/m was approximately 6 h,which is shorter than the conditioning time of the normal-conducting cavity.
Coating P Ac [kW] P Ac /P Bc P Ac /P Cc E Az [MV / m] N shot ( × )uncoated 0 .
89 0 .
010 0 .
013 1 . . .
92 0 .
74 11 1 . . . . P Cc =max( P in ( t ) − P ref ( t )). E Az is the peak accelerating gradientcalculated by P Ac . DLC1 and DLC2 denote the DLC-coatedcells, with a thickness of 0 . µ m] on both sides. In Figures 2, 3, and 4, we show the waveforms of themeasured incident power (blue), reflected power (red),and the stored power reconstructed by the antenna (yel-low), P Ac , and by the incident power (green), P Bc .As shown in Figures 3 and 4, power equivalent to anaccelerating electric field of over 10 [MV / m] could be in-put when approximately 0 . µ m] of DLC was applied.Table III shows each measured maximum stored powervalue and its conversion to the maximum acceleratingelectric field corresponding to these waveforms. In theDLC1 test, when the accelerating electric field was in-creased further, a large discharge occurred and the pulsesstopped coming in. In the DLC2 test, the pulse lengthwas extended to approximately 7 [ µ s] while maintaining10 [MV / m], resulting in a waveform similar to that ofDLC1.In an uncoated MgO cell, the stored power is saturatedwhen the accelerating electric field is at the maximumelectric field E z, max ∼ / m]. In this case, the satu-ration value of the stored power does not change even if P in increases with the same pulse length T p . When thestored power is saturated, extending the pulse length T p can extend the region where the stored power waveformis flattened to 10 [MV / m]. The reflected waveform is notconsistent with the transient model; above 1 [MV / m],the waveform is maintained at the same depth after a slight absorption.When DLC coating was applied, the stored power in-creased in stages until a breakdown occurred in the cav-ity. Once the breakdown occurred, the stored powerwaveform became unstable and did not improve with fur-ther conditioning. After the test, some traces of electricdischarge remained at the cell contact area and at thecontact area between the cell and the metal end-plates,as shown in Figure. 6. Therefore, we consider that themicrogap discharges at the cell junctions determine themaximum accelerating electric field in the DLC-coatedDAA cavity. In order to obtain a higher accelerating field,countermeasures against breakdowns at the cell junctionsare required. For example, methods of separating the di-electric cells from each other or separating the dielectriccells from the metal end-plates have been proposed in[18, 19].In conclusion, By applying a DLC that lowers the SEYwithout reducing the Q-value, we have demonstratedthat accelerating fields above 10 [MV / m] can be appliedfor pulses as long as 6 [ µ s], thus suppressing the multi-pactor that has limited the accelerating field of the DAAcavity. We have also demonstrated that, owing to thevery high resistance of the DLC coating, even a 0 . µ m]-thick DLC coating on both sides of a dielectric cell doesnot reduce the Q-value O (10 ) of a five-cell DAA cavity.One possible application of the DAA cavity is as a com-pact power source for X-ray and RI production, as DAAcavities have a power efficiency one order of magnitudehigher than that of a normal-conducting cavity. In thefuture, if the anti-breakdown treatment is successful andthe accelerating field can be further extended to tens ofMV/m, a linear accelerator facility such as an FEL maybe able to operate at room temperature without RF pulsecompression by the SLED cavity. Acknowledgements
We would like to acknowledge Hiroyasu Ego forhelpful discussions. This work was supported byJSPS KAKENHI Grant-in-Aid for Scientific Research(A) (No. 16H02134), JSPS KAKENHI Grant Number19K20609, and Mitsubishi Heavy Industries MachinerySystems, Ltd.The data that support the findings of this study areavailable from the corresponding author upon reasonablerequest. [1] A. Nyaiesh, E. Garwin, F. King, and R. Kirby, Journalof Vacuum Science & Technology A: Vacuum, Surfaces,and Films , 2356 (1986).[2] K. Kennedy, B. Harteneck, G. Millos, M. Benapfl,F. King, and R. Kirby, in Proceedings of the 1997 Parti-cle Accelerator Conference (Cat. No. 97CH36167) , Vol. 3 (IEEE, 1997) pp. 3568–3570.[3] C. Y. Vallgren, G. Arduini, J. Bauche, S. Calatroni,P. Chiggiato, K. Cornelis, P. C. Pinto, B. Henrist,E. M´etral, H. Neupert, et al. , Physical Review SpecialTopics-Accelerators and Beams , 071001 (2011).[4] C. Y. Vallgren, G. Arduini, J. Bauche, S. Calatroni, FIG. 2: uncoated cells with multipactor FIG. 3: DLC1 FIG. 4: DLC2FIG. 5: Waveforms of the measured incident power (blue), reflected power (red), and the stored power reconstructed by theantenna (yellow), P Ac , and by the incident power (green), P Bc . For uncoated cells, typical waveforms with multipactors areshown. For DLC1 and DLC2, the waveforms show stable stored power with an accelerating gradient exceeding 10 MV/m.FIG. 6: DLC-coated MgO cell after a breakdown in a high-power test. Black marks were observed at the corners of thecylindrical part of the cell. Black marks were also observed atthe contact point between the dielectric cell and the adjacentdielectric cell and at the contact point between the dielectricand the copper end-plate.P. Chiggiato, K. Cornelis, P. C. Pinto, E. M´etral, G. Ru-molo, E. Shaposhnikova, et al. , in , Vol. 10 (2010).[5] E. Shaposhnikova, G. Rumolo, K. Cornelis, J. Axen-salva, J. Jim´enez, M. Taborelli, P. Chiggiato, S. Cala-troni, G. Arduini, C. Yin Vallgren, et al. , Experimentalstudies of carbon coatings as possible means of suppress-ing beam induced electron multipacting in the CERN SPS ,Tech. Rep. (2009).[6] V. Baglin, I. Collins, and O. Gr¨obner,
Photoelectronyield and photon reflectivity from candidate LHC vac-uum chamber materials with implications to the vacuumchamber design , Tech. Rep. (1998).[7] A. Kulikov, A. Fisher, S. Heifets, J. Seeman, M. Sullivan,U. Wienands, and W. Kozanecki, in
PACS2001. Proceed- ings of the 2001 Particle Accelerator Conference (Cat.No. 01CH37268) , Vol. 3 (IEEE, 2001) pp. 1903–1905.[8] C. Jing, S. Gold, R. Fischer, and W. Gai, Applied PhysicsLetters , 193501 (2016).[9] D. Satoh, M. Yoshida, and N. Hayashizaki, Physical Re-view Accelerators and Beams , 011302 (2016).[10] D. Satoh, M. Yoshida, and N. Hayashizaki, Physical Re-view Accelerators and Beams , 091302 (2017).[11] J. Power, W. Gai, S. Gold, A. Kinkead, R. Konecny,C. Jing, W. Liu, and Z. Yusof, Physical review letters , 164801 (2004).[12] J. Zhang, B. J. Munroe, H. Xu, M. A. Shapiro, and R. J.Temkin, Physical Review Accelerators and Beams ,081304 (2016).[13] C. Jing, W. Gai, J. G. Power, R. Konecny, W. Liu, S. H.Gold, A. K. Kinkead, S. G. Tantawi, V. Dolgashev, andA. Kanareykin, IEEE transactions on plasma science ,1354 (2010).[14] H. Xu, M. Shapiro, and R. J. Temkin, Physical ReviewAccelerators and Beams , 021002 (2019).[15] Y. Zhang, Y. Wang, X. Ge, B. Zhang, W. Wei, S. Wang,B. Zhu, J. Shao, W. Li, and Y. Wang, in Journal ofPhysics: Conference Series , Vol. 1350 (IOP Publishing,2019) p. 012177.[16] Y. Yamamoto, E. Cenni, A. Four, E. Kako, S. Michizono,and Y. Okii, in \ textsuperscript { th } Linear Accelera-tor Conf.(LINAC’18), Beijing, China, 16-21 September2018 (JACOW Publishing, Geneva, Switzerland, 2019)pp. 905–907.[17] B. Meyerson and F. Smith, Journal of Non-CrystallineSolids , 435 (1980).[18] S. Mori, M. Yoshida, and J. D. Satoh, in10th Int. PartileAccelerator Conf.(IPAC’19), Melbourne, Australia, 19-24 May 2019