Transmission efficiency measurement at the FNAL 4-rod RFQ
J.-P. Carneiro, F.G. Garcia, J.-F Ostiguy, A. Saini, R. Zwaska, B. Mustapha, P. Ostroumov
TTransmission Efficiency Measurement at the FNAL 4-rod RFQ ∗ J.-P. Carneiro † , F. G. Garcia, J.-F. Ostiguy, A. Saini, R. Zwaska, Fermilab, Batavia, IL 60510B. Mustapha, P. Ostroumov, ANL, Argonne, IL 60439 Abstract
This paper presents measurements of the beam transmis-sion performed on the 4-rod RFQ currently under operationat Fermilab. The beam current has been measured at theRFQ exit as a function of the magnetic field strength in thetwo LEBT solenoids. This measurement is compared withscans performed on the FermiGrid with the beam dynamicscode
TRACK . A particular attention is given to the impact,on the RFQ beam transmission, of the space-charge neu-tralization in the LEBT.
INTRODUCTION
A new injector has been in operation since 2012 on theFNAL 400 MeV Linac as a part of the Proton Improve-ment Plan whose primary goal is to increase the protonflux in the booster to ultimately 2.25 × protons perhour. This new injector, composed of an ion source, a LowEnergy Beam Transport line (LEBT), a 4-rod Radio Fre-quency Quadrupole (RFQ) and a Medium Energy BeamTransport (MEBT) is presented in details in Ref.[1]. Thebeam transmission in the injector, from the ion source tothe MEBT exit, has routinely been measured since the startof its operation ranging from 40% to 50%. This measuredtransmission is significantly lower than the expected one,which according to computer simulations should be closeto 100%. After a brief description of the injector, this pa-per presents a measurement of a beam transmission at theMEBT exit as a function of the LEBT solenoid fields. Thismeasurement is compared to numerical simulations fromthe code TRACK [2]. The simulations reveal that the spacecharge neutralization pattern, which is unlikely to be homo-geneous along the LEBT, plays a crucial role in the injectortransmission.
THE FNAL LINAC INJECTOR
A layout of the injector is depicted in Fig. 1. A mag-netron source produces 35 keV H − bunches of typically100 µ s long, at a repetition rate of 15 Hz with an averagecurrent ranging from 50 to 70 mA. The LEBT comprisestwo solenoids that match the H − beam produced by thesource into the RFQ entrance. The beam is further acceler-ated to keV by a 4-rod RFQ operating at 201.25 MHz.At the exit of the RFQ, the MEBT matches the beam intothe first Drift-Tube Linac Tank (DTL ∗ Work supported by Fermilab Research Alliance, LLC under ContractNo. DE-AC02-07CH11359 with the United States Department of Energy. † [email protected] Figure 1: Layout of the FNAL Linac Injector.RFQ to chop the first 20 µ s of the pulse. A current moni-tor which is used in the measurement described in the nextsession is located 8.25 cm from the downstream face of thelast MEBT quadrupole, i.e at the DTL TRANSMISSION MEASUREMENT
An experiment has been performed on the linac injectorwhich consisted in varying, in a systematic way, the currentin the two LEBT solenoids and measuring the beam inten-sity on the current monitor located at the MEBT exit. Thismeasurement is represented in Fig. 2.Figure 2: Measurement of the beam intensity at the currentmonitor located at the exit of the MEBT as a function of theLEBT solenoid currents. The cross represents the actualpoint of operation of the LEBT and the ellipse mimics thearea of favorable transmission.For this measurement, the LEBT solenoids have beenscanned from 300 A to 500 A with a step of 5 A, all other a r X i v : . [ phy s i c s . acc - ph ] N ov arameters in the injector being kept at their optimal val-ues. During the measurement, the vacuum on the LEBTwas measured to be 4 × − Torr using a cold cathodegauge located in the middle of the LEBT. The maximumbeam intensity measured at the MEBT exit and reported inFig. 2 is 27.3 mA for a measured current at the ion sourceof 47.5 mA, which represents a transmission of 57.5%.
BEAM DYNAMICS SIMULATION
An attempt to numerically reproduce, with the beam dy-namics code
TRACK , the experimental scan presented inFig. 2 has been undertaken at Fermilab. The code
TRACK ,developed at ANL, has been selected for this work becauseof its ability to use an external 3D field map to simulatean RFQ. The EM fields were extracted from a MicroWaveStudio model of the 4-rod RFQ built by S. Kurennoy atLANL. The advantage of using 3D fields to simulate theRFQ lies on the fact that, as discussed in [3], the 4-rod RFQpresents a field asymmetry due to the stems which can leadto emittance increase and beam losses along the RFQ. Ascript launches 10201
TRACK runs on the FermiGrid, repre-senting a scan of the two LEBT solenoids from 0 to 500 A,with 5 A steps. Another script analyses for each run thepredicted beam transmission at the position of the currentmonitor in the MEBT exit. A 4D Waterbag has been usedas input distribution with 5 × macro-particles, 3D fieldsmaps have been generated for the LEBT solenoids and theMEBT buncher has been modeled using a simulated ax-ial electric field. The radius of aperture has been kept at5 cm all along the LEBT but in the last 8 cm where it wasreduced to 2.2 cm to take into account the Einzel lens aper-ture. In the TRACK model, a 1 cm aperture hole at the RFQentrance was assumed. The vane aperture has been imple-mented in the code along the RFQ and in the MEBT theradius of aperture has been kept at 2 cm.
LEBT neutralization
An important parameter to take into account in the sim-ulations is the space charge neutralization factor in theLEBT. In fact, upon exciting the ion source, the H − beaminteracts with the H molecules present in the residual gascreating H +2 ions and electrons. The H +2 ions are thentrapped in the beam potential well and counteract the beamspace charge field while the electrons are ejected to thebeam pipe wall. In TRACK the neutralization factor is mod-eled by a simple reduction of the beam intensity. We con-sidered in our simulation three scenarios: a uniform spacecharge neutralization factor in the LEBT, a region at the en-trance of the RFQ which is un-neutralized and an hypothet-ical space charge neutralization pattern along the LEBT.
Uniform LEBT neutralization factor
We performed with
TRACK a scan of the LEBT solenoidcurrents for a uniform neutralization pattern along theLEBT (from the simulation starting point to the RFQ en-trance hole) ranging from 0% (full space charge) to 100%(no space charge), with steps of 10% in the neutralization (a)(b)
Figure 3: Simulated beam intensity at the MEBT exit as afunction of the LEBT solenoid currents for a uniform neu-tralization factor along the LEBT of (a) 100% and (b) 60%.From
TRACK . The ellipse mimics the measured area of fa-vorable transmission reported in Fig. 2.factor. Figure 3(a) and 3(b) show respectively the cases100% and 60%, where the simulated beam current is re-ported at the location of the current monitor at the MEBTexit. In these figures is also reported the ellipse whichdepicts the region of favorable transmission measured inFig. 2. Clearly, the agreement between the simulated andmeasured area of favorable transmission is poor for thesetwo cases as for all others not reported in this document.Our conclusion from these studies is that a model based onuniform neutralization is too simplistic. The neutralizationis, in fact, unlikely to be uniform.Figure 4 reports for each neutralization factor scanabove-mentioned the maximum transmission predicted by
TRACK . We believe that, as the neutralization factor de-creases, the beam develops along the LEBT non linearspace charge effects which inhibit a proper matching intothe RFQ resulting in beam losses. For a beam fully neu-tralized,
TRACK predicts a beam transmission of 98% whichdrops to 60% (close to the measured value) for a beam neu-tralized at 60%.igure 4: Maximum simulated transmission at the MEBTexit as a function of a uniform neutralization factor alongthe LEBT. From
TRACK . Un-neutralized region at the LEBT end
We performed another set of LEBT solenoid scans with
TRACK taking a 100% fully neutralized beam all along theLEBT but in the last few centimeters at the LEBT end.Fig. 5 shows that few centimeters of un-neutralized regionat the LEBT end has a significant impact in the beam trans-mission. We could speculate that the level of neutralizationis inversely correlated with the beam size and that at thisregion the neutralization could be less effective. For in-stance, if we consider the LEBT 100% neutralized but thelast 8 cm, the beam transmission at the MEBT exit drops to63%.Figure 5: Maximum simulated transmission at the MEBTexit as a function of an un-neutralized region at the LEBTend, the remaining part neutralized at 100%. From
TRACK . Non-uniform LEBT neutralization factor
Figure 6(a) shows a scan performed with
TRACK takinga non-uniform neutralization factor along the LEBT, as de-picted in Fig. 6(b). We considered a neutralization pro-file with gradual linear increase and decrease respectivelyat the upstream and downstream extremities of the LEBT. The last 6 cm of the LEBT were considered un-neutralized.With this pattern, the simulations presented in Fig. 6(a)show better agreement with the measurement presented inFig. 2, particularly concerning the size and location of theareas of favorable beam transmission. Yet, the predictedtransmission from
TRACK is 20% higher than the measuredone. (a)(b)
Figure 6: (a) Simulated beam intensity at the MEBT exitas a function of the LEBT solenoid currents for (b) a non-uniform neutralization along the LEBT. From
TRACK . Theellipse mimics the measured area of favorable transmissionreported in Fig. 2.
CONCLUSION
The simulations results indicate that it is likely that theneutralization profile in the LEBT can account for the mea-sured transmission. More work will be needed in order tounderstand and better characterize the neutralization pat-tern.
ACKNOWLEDGEMENT
The author would like to thank C.Y Tan for sharing in-formation about the injector, B. Pellico and V. Shiltsev forcontinuing support.
REFERENCES [1] C.Y. Tan, “FNAL Beams-doc-3646-v16”, 18 Feb. 2011.[2] V.N. Aseev et al.et al.