PEPPo Based Polarized Positron Beam Formation for JLEIC

PEPPo Based Polarized Positron Beam Formation for JLEIC Jiquan Guo 2017 International Workshop on Physics with Positrons Jefferson Lab, Newport News, VA, USA September 12 -15, 2017 1

Outline Introduction – Overview of JLEIC – Basics of JLEIC electron ring injection – Challenges for PEPPo based JLEIC positron beam formation with an electron accumulator RF harmonic kicker for accumulator ring extraction Summary 2

Introduction: JLEIC overview Electron complex – 3 -10(12) Ge. V electron collider storage ring – CEBAF as full energy injector Ion complex 3 -12 Ge. V – Ion source – SRF linac (285 Me. V/u for protons) – Booster (8 Ge. V for protons) – Figure 8 Ion collider ring Optimum detector location for minimizing background 8 -100(400) Ge. V 8 Ge. V SRF Linac & 2015 ar. Xiv: 1504. 07961 May 17 update: https: //eic. jlab. org/wiki/index. php/Main_Page 3

JLEIC Electron Collider Ring Circumference 2255. 4 -2256. 1 m (variate to sync with ions of wide velocity range) Reuse SLAC’s PEP-II NCRF cavities in the initial phase, frequency tuned to 476. 3 MHz (1497× 7/22), Nh=3584 (28*128), and will upgrade to 952. 6 MHz SCRF cavities later (Nh=7168). – Bunch reprate needs to have flexible choices of 119 MHz, 238 MHz, 476 MHz, and maybe 952 MHz after the upgrade, depending on luminosity optimization results at different energy Two electron bunch trains with opposite polarization in the ring; each section ~1047 m with two gaps of ~80 m (128 RF buckets of 476. 3 MHz) each. – Gaps matching ion ring due to ion beam formation process; also required for injection/abort kicker rise time Beam lifetime a few hours; polarization lifetime 0. 3 hr (10 Ge. V) to 66 hrs (3 Ge. V), continuous topoff required at high energy JLEIC e-ring Beam Current Cavity Number 35 Limited by cavity impedance 10 MW SR 3, 5 limit 3 Cavity Number 30 25 2, 5 2 Beam current using PEP-II cavity 20 15 1, 5 Beam current w/o impedance limit 10 1 Down polarized bunch train 0, 5 5 0 0 3 5 7 Energy (Ge. V) 952. 6 MHz SCRF cavity under design 476. 3 MHz cavity (NCRF PEP-II) Beam Current (A) 3 A admin limit 40 Up polarized bunch train gaps 9 Bunch train pattern in JLEIC electron ring 4

JLEIC e-ring Injection with CEBAF Need multiple injection cycles to accumulate higher charge per bunch in the ring. During each injection cycle, only one bunch train of ~3. 6μs (1/2 ring, one polarization) from CEBAF can be injected. The injected bunch will be in 68. 05 MHz (or 34/17 MHz) buckets, which is a common factor of CEBAF’s 1497 MHz (1/22) and the ring’s bunch reprate 476. 3 MHz (1/7). The 7 groups of buckets in the e-ring can be filled in different cycles. During each cycle, kickers will be on for ~3. 6μs to minimize the distance (in x direction as shown) between the freshly injected bunches and the circulating bunches in the electron storage ring, and then wait for about 2× transverse synchrotron radiation damping time ( E-4, 6 -400 ms for JLEIC) for the two bunches to fully merge. However, during that waiting time, we can also inject the other half ring one time. If the damping time is shorter than the kicker recovery time (assumed 20 ms), the waiting time will be the kicker recovery time. Duty factor of CEBAF linac for this operation mode is really low, 10 -5 -10 -3 e-ring P+ current Injected beam, first turn Injected beam after several turns circulating beam, kickers on circulating beam, kickers off e-ring P- current kicker amp (arb) septum -4000 Example of x injection: Kickers on for few μs every 2 x damping time or 2 x kicker recovery time(a few ms) to form an orbit bump 1000 time (ns) 6000 JLEIC e-ring kicker pulse: 134 ns rise/fall time, 3628 ns flat top, 2 -50 Hz 5

Projected JLEIC electron injection time 50 45 40 35 30 25 20 15 10 5 0 projected injection time Beam current in e-ring (no impedance limit) 3 5 7 Energy (Ge. V) 9 11 5 4, 5 4 3, 5 3 2, 5 2 1, 5 1 0, 5 0 JLEIC e-ring beam current (A) Estimated injection time (minutes) JLEIC Electron Injection Time The total injection time for JLEIC electron is Iinj is the pulsed beam current from CEBAF τwait is either 1× SR damping time or kicker recovery time The maximum ~4μs pulsed beam current CEBAF can supply is mainly limited by the voltage drooping in the cavities when the cavity stored energy is taken away by the strong beam loading. – – Without compensating the beam loading with pulsed klystron power, CEBAF’s maximum ~4μs pulsed beam current is ~1 m. A; By pulsing the klystron power, we can increase CEBAF’s maximum ~4μs pulsed beam current to ~1. 8 m. A, resulting an injection time of 22 min for the worst case. 6

JLEIC e-ring Top-off for Polarization Retention JLEIC e+/e- polarization has a limited lifetime proportional to the SR damping time at low energy, and drops even faster beyond 8 Ge. V due to Sokolov–Ternov effect. Continuous top-off injection (and removing the same amount of recirculating particles) to keep the polarization Equilibrium polarization is determined by To retain ≥ 80% of the injected polarization, the initial injection time needs to be less than 1/4 of depolarization time 7

Challenges for PEPPo Based JLEIC Positron Injection See L. Cardman’s talk Polarized Electrons for Polarized Positrons (PEPPo) Concept S 1 Pe. E. G. Bessonov, A. A. Mikhailichenko, EPAC (1996) A. P. Potylitsin, NIM A 398 (1997) 395 e- T 1 PEPPo Experiment D (PEPPo Collaboration) D. Abbott et al. , Phys. Rev. Lett. 116 (2016) 214801 S E. A. Kuraev, Y. M. Bystritskiy, M. Shatnev, E. Tomasi-Gustafsson, PRC 81 (2010) 055208 D The low e+/e- yield (10 -5 -10 -3) of PEPPo, in addition to the low duty factor injection beam (10 -5 -10 -3 again) required by conventional e+/e- storage rings, will pose a major challenge to the JLEIC positron bunch formation 2 PT Pe+ e+T Calorimeter 2 J. Dumas, Doctorate Thesis (2011) Yield of electron-positron conversion at 10 -100 Me. V Assuming Δp/p<10%, 75% electron polarization transfer – Even with the lowered luminosity goal, the required electron beam current and charge per bunch is beyond the state-of-the-art of the polarized sources – The required 17 -68 MHz beam will increase the required electron charge per bunch further. A natural idea to address the challenge is to compress a longer bunch train from the gun into a shorter one, with an accumulator ring 8

Positron Production Scheme Polarized Electron 50 Me. V Injector Accumulator Ring (35. 6 m) 500 -turn phase-space painting Bunch Management Positron Conversion/Collection Efficiency ~ 10 -5 - 10 -3 60% e- polarization transfer to CEBAF Harmonic kicker Extraction 50 Me. V polarized e 2 n. C @ 17 MHz, ~3. 5μs bunch train @2 -50 Hz 50 Me. V polarized e 4 p. C @ 748. 5 MHz 2 n. C bunches ~60μs bunch train @2 -50 Hz @ 748. 5 MHz e- bunch train, 44500 bunches, 60μs Iave up to 3 m. A 4 p. C/bunch ………. . 17 MHz e+ bunch train, 60 bunches, 3. 5 us Iave up to 6 n. A 1 -2 p. C …… 20 -85 ms, up to 50 Hz 9 ~25 Me. V Polarized e+ 1 p. C @ 17 MHz 1050 m bunch train @2 -50 Hz

Positron Production Scheme Polarized Electron 50 Me. V Injector Accumulator Ring (35. 6 m) 500 -turn phase-space painting Bunch Management Positron Conversion/Collection Efficiency ~ 10 -5 - 10 -3 60% e- polarization transfer to CEBAF Harmonic kicker Extraction 50 Me. V polarized e 4 p. C @ 748. 5 MHz 2 n. C bunches ~60μs bunch train @2 -50 Hz @ 748. 5 MHz JLEIC CEBAF ~25 Me. V Polarized e+ 1 p. C @ 17 MHz 1050 m bunch train @2 -50 Hz 50 Me. V polarized e 2 n. C @ 17 MHz, ~3. 5μs bunch train @2 -50 Hz Pol. e- source Accumulator ring e- at converter Pol. e+ source Comments 10 Me. V test bed e+/e- =3 e-5 0. 2 p. C @ 748. 5 MHz 0. 15 m. A DF = 0. 3% 0. 1 n. C @ 748. 5 MHz 75 m. A DF = 0. 3% 0. 1 n. C @ 17 MHz 1. 7 m. A DF = 0. 018% 3 f. C @ 17 MHz 0. 051 μA DF = 0. 018% • 50 Me. V production e+/e- =5 e-4 4 p. C @ 748. 5 MHz 3 m. A DF = 0. 3% 2 n. C @ 748. 5 MHz 1. 5 A DF = 0. 3% 2 n. C @ 17 MHz 34 m. A DF = 0. 018% 1 p. C @ 17 MHz 17 A DF = 0. 018 A% • • • Iave~3 n. A @ 50 Hz Modest neutron radiation 150 k. W peak beam power in the e- injector linac 100 Me. V production e+/e- =1 e-3 4 p. C @ 748. 5 MHz 3 m. A DF = 0. 3% 2 n. C @ 748. 5 MHz 1. 5 A DF = 0. 3% 2 n. C @ 17 MHz 34 m. A DF = 0. 018% 2 p. C @ 17 MHz 34 A DF = 0. 018% • • • Iave~6 n. A @ 50 Hz Strong neutron radiation 300 k. W peak beam power in the e- linac, high extraction kicker voltage 4 -40 p. C @ 250 MHz 0. 1 – 1 m. A (cw) Not necessary 4– 40 p. C @ 250 MHz 0. 1 – 1 m. A (cw) 0. 4 -4 f. C @ 250 MHz 10 n. A – 100 n. A (cw) • Remove the accumulator ring • Iave~0. 009 n. A @ 50 Hz bunch train reprate No neutron radiation Positron yield roughly scaled from Dumas’ thesis, assuming Δp=1 Me. V and 60% e- polarization transfer (resulting ~43% polarization multiplied with 90% polarization in e - source and 80% polarization retention in JLEIC e-ring) 10

Accumulator Ring: Phase Painting Injection Concept: an orbit bump created near a septum and then slowly reduced as beam being injected (phase-space painting) x > Septum thickness + bunch width Accumulator ring x Injected beam A number of painting schemes have been developed Process can also be simultaneously occurring in vertical and longitudinal dimensions CERN’s LEIR has a design for 75 -turn injection of Pb 54+, We plan to push this number to ~500 in the accumulator using low electron emittance Main injection system components – Magnetic or electrostatic septum – Four bumper magnets with ~60 s fall time for 50 Me. V, reasonably fast rise time and ~10 mrad maximum deflection 11

Accumulator ring: Harmonic RF Kicker Extraction Example: A ring using a harmonic kicker to extract a bunch train 5× longer from a 11 bucket ring 9 11 2 7 n Empty gap buckets to accommodate kicker burst rise time n Empty ring buckets already extracted to the transfer line n Ring buckets still occupied by bunches 5 4 Empty transfer line buckets 3 6 8 1 3 n Transfer line buckets occupied by extracted bunches 2 1 10 Harmonic kicker that kicks every 5 th bunch Ideal kicker pulse JLEIC injection requires 17 MHz (or 34/68 MHz) beam in bunch trains with ~1050 m in length (60 bunches for 17 MHz). An accumulator ring of 1050 m circumference at ~100 Me. V is not practical. However, we can build a ring containing 60 bunches with 748. 5 MHz spacing, so the bunch train length in the ring can be reduced to 24 m. We can use a kicker to extract the beam at 17 MHz (1/44 of 748. 5 MHz), kicking every 44 th bunch out. To leave N continuous gaps in the ring to accommodate the kicker burst rise time, the ring’s harmonic number needs to be picked at 44 N+1. In our case, the ring harmonic number will be 89. The ring circumference will be 35. 6 m, with two gaps of ~6 m (20 ns) each, and two 12 m bunch trains. Harmonic stripline RF kickers will be the right technology choice to provide <1 ns pulse rise time, <1 ns pulse width and ~20 ns burst rise time, with ~60 pulses in one burst. 12

JLEIC Positron Injection Performance Estimated JLEIC e-ring positron beam current (A) 1 0, 8 Base line (50 Me. V e-with JLEIC damping wiggler) 0, 6 50 Me. V e- w/o JLEIC damping wiggler 0, 4 100 Me. V e- w/o damping wiggler 0, 2 100 Me. V e-with JLEIC damping wiggler 0 3 4 5 6 7 8 9 10 JLEIC positron energy (Ge. V) The beam current is limited by the long damping time at low energy end; at high energy, the limit is the kicker wait time longer than damping time, as well as a faster drop in polarization lifetime. – JLEIC positron current peaked at 8 Ge. V, as coincidentally both the 1/4 depolarization time and the SR damping time reach the limit. With damping wigglers in the JLEIC e-ring, low energy performance can be improved. The chart above assumes that we can shorten the damping time for <5 Ge. V to the 5 Ge. V level (85 ms). Higher e- conversion energy helps to increase e+ yield, as well as to suppress instabilities (such as space charge tune shift) in the accumulator ring, but also increases the peak power required in the e - linac and the extraction kicker. 13

RF Harmonic Kickers RF harmonic kickers have been studies intensively for JLEIC’s circulating cooler ring (CCR) – Harmonic combination is the most efficient way to produce repetitive short pulses Two technical options have been compared. – QWR Cavity RF harmonic kickers are more efficient, and became the baseline for the CCR. However, it can’t provide the 20 ns burst rise time for the accumulator ring with the high-Q (high efficiency) design. – Stripline kickers can have fast burst rise time, and the power requirement for extracting a 50 Me. V beam seems tolerable, especially when operated in low duty factor bursts. Different mode combination schemes (kicks every 10 th bucket) Y. Huang PRAB 19, 084201 Mode amplitude for different schemes (for kicking 748. 5 MHz beam at 17 MHz) 2, 5 Mode RF voltage (Arb) 5 -harmonics, copper prototype QWR kicker cavity, Yulu Huang, IMP/JLab Ph. D Thesis, 2016 PRAB 19, 084201 / PRAB 19, 122001 0 -gradient Least Mode Equal Amp 2 1, 5 1 0, 5 0 0 500 Millions 1000 All 3 schemes gives 0 amplitude at the center of the bunches not to be kicked 14

RF Harmonic Stripline Kicker Port 4, to matched load Port 2 Port 3, to matched load 2 -D E-field simulated with superfish E=2 V/g/d Port 1 3 -D CST model of the PEP-II feedback kicker Z 0, 50 -100Ω Based on the PEP-II feedback kicker design – – Travelling wave TEM mode RF kicker Geometric factor g=0. 84, kicker length L=0. 63 m, d=70 mm +V -V d The accumulator ring can use a similar design with further optimization – – Need to adjust the length of the kicker to optimize the efficiency (or transverse impedance) for desired combination of modes Other optimizations such as HOM suppression can be in synergy with the JLEIC e-ring feedback kicker development. 15 J. Guo, COOL’ 15, TUPF 10

Preliminary Accumulator Ring Stripline Kicker Optimization For stripline kickers, the “Least mode scheme” is the most efficient, as the modes are in lower frequency and the kicker can be designed longer. However, the “ 0 -gradient scheme” is only ~15% worse in efficiency, and gives much better waveform quality For a PEP-II stripline kicker shortened to L=0. 184 m, the effective shunt impedance with 043 rd harmonics of 17 MHz in the 0 -gradient scheme is optimized to 163 kΩ. – To provide 50 k. V kick (1 mrad at 50 Me. V), 15 k. W peak power (not include DC) is needed. 1 0, 8 Mode voltage Rt 0, 6 0, 4 0, 2 0 0 500 Millions 10 9 8 7 6 5 4 3 2 1 0 1000 Synthesized waveform (Arb) Mode RF voltage (Arb) 1, 2 Thousands Mode voltage and transverse shunt impedance for the L=0. 184 m PEP-II style stripline kicker (0 -gradient scheme) Waveform synthesized with the 0 -9 th harmonics V=0, d. V/dt=0 at the centers of bunch 1 -9 (0 -gradient) 60 50 40 30 Synthesized waveform Desired waveform 20 Bunches not to be kicked 10 Bunches to be kicked 0 -10 16 0 2 4 Z 0 (λ) 6 8 10

Summary We preliminarily developed a positron bunch formation scheme for JLEIC, based on PEPPo. The key element of the scheme is a 500 turn e- accumulator ring, including its extraction kicker. – With a small (35. 6 m) ring and an RF harmonic stripline kicker that converts the 748. 5 MHz beam into a 17 MHz beam, we can increase the extracted pulsed e- current by a factor of 11, and the extracted bunch train length matches the JLEIC e-ring bunch train length. As a result, we can form a (40 -45%) polarized positron beam current of ~0. 2 A in the JLEIC e-ring, making it possible to achieve ~1033 luminosity The work presented was mainly completed with a small study group, including Joe Grames, Fanglei Lin, Vasiliy Morozov, and Yuhong Zhang. We also thank all the JLEIC accelerator R&D colleagues for their support. 17