C-AD Accelerator R&D Division

eRHIC R&D Group

Coherent Electron Cooling

CeC Group Head: Igor Pinayev

Coherent Electron Cooling is an exciting concept towards cooling ion beams through amplification of charge distributions in electron beams. This method promises extremely fast cooling virtually independent of energy of the ions. Recently, the performance of ERLs allows the use of the FEL amplification mechanism to realize Coherent Electron Cooling at high ion energies.


eRHIC R&D Group

Coherent Electron Cooling

I. Goal and Scope

The goal of the experiment is to demonstrate longitudinal (energy spread) cooling in CeC mode before expected CD-2 for eRHIC. Present scope of the experiment is to cool longitudinally a single bunch of 40 GeV/u Au ions in RHIC. The CeC experiment expected to be carried out during RHIC Runs 15 and 16 (and possible Run 17), with early commissioning of the accelerator during Run 14.

In addition to use for CeC demonstration experiment, the e-beam will be used for studying novel aspects of beam-beam effects in eRHIC by colliding the electron and hadron beams. These tests would be carried out after completion of CeC experiment.

II. Location

The experiment will be located at IP2 (formerly the Brahms IP) and some equipment for the experiment will be located in a portion of Service Bldg. 1002A and 1002B.


Fig. 1. Approximate location of CeC PoP experiment at IP2

Local control will be conducted, most likely, from a former Brahms Counting house, i.e. a trailer located at IP2. Detailed layout of the equipment allocation is not available at this time.

III. Schematic of the CeC

Fig. 2 is a schematic of a coherent electron cooler comprised of a modulator, a FEL-amplifier, and a kicker. The figure also depicts some aspects of coherent electron cooling. In CeC, the electron and hadron beams co-propagate in vacuum along a straight line in the modulator and the kicker and have the same velocity, v:


Since the electrons are about 2,000 fold lighter than a nucleon their energy is about the same factor less. For cooling Au ions with energy of 40 GeV/u we will use electron with energy of 21.8 MeV. Since the rigidity of the ion beam is in addition amplified by A/Z =197/79 ~ 2.5 fold, the effect of magnetic elements designed to transport electron beam has minuscule effect on the ion beam. In our case that magnetic elements (dipoles, trims and quadruples) affect the trajectory of electron beam 5,000 times stronger than that of the Au ion beam. This feature allows us to use common elements in the IP2 and to optimize lattice of the electron beam transport channel without affecting ion beam in RHIC.

The CeC works as follows: In the modulator, each positively charged hadron (with charge Z and atomic number A) induces a density modulation in electron beam that is amplified in the high-gain FEL; in the kicker, the hadrons interact with the electric field of the electron beam that they have originated, and receive energy kicks toward their central energy. The process reduces the hadrons’ energy spread, i.e., cools the hadron beam.


Fig. 2. Schematic of the CeC – economic option (see [1] for more details)

In practice, the scheme will look like that sketched in Fig.3 using most of the available space between two DX magnets at IP2. A 21.8 MeV linear accelerator will provide an electron beam, which will be merged with yellow Au ion beam circulating in RHIC. Both beams will co-propagate through the 14-meter (45-feet) long CeC system after which the electron beam will be separated from the yellow Au ion and damped.


Fig. 3. Schematic of the CeC experiment at IP2.

Table 1 shows major electron and hadron beam parameters we plan to use for the experiment.

Table 1. Main beam parameters for CeC experiment


Species in RHIC

Au ions, 40 GeV/u

Number of particles in bucket


Electron energy

21.8 MeV

Charge per e-bunch

0.5-5 nC


78.3 kHz

Average e-beam current

0.4 mA

Electron beam power

8.5 kW

Table 2. Electron beam and FEL parameters


RMS Energy Spread

≤ 1×10-3

Normalized Emittance

≤ 5 μm.rad

Peak Current

60-100 A


Wiggler Length

7 m

Wiggler Period

0.04 m

Wiggler Strength, aw


FEL Wavelength

13 μm

Requirements for the quality of the electron beam and some FEL parameters are listed in the Table 2.

IV. Layout and the Lattice

Fig. 4 show a schematic of the right-hand side of the CeC beam-line: the 100-300 picoseconds long, 2 MeV electron beam generated the 112 MHz gun goes through the two 300 kV max 500 MHz bunching cavities, where electron acquire energy chirp. Alternatively the electrons can be generated off-crest to utilize distance between gun and bunching cavities for compression.

Fig. 4. Schematic of the CeC beam-line – not to scale

After bunching to about 10-20 psec in the drift section, electron are accelerated to 21.8 MeV in 5-cell SRF cavity and pass through an achromatic dog-leg to join the yellow Au beam. After passing through the CeC section, electron beam is separated from the ion beam by another dogleg, which directs it into the beam dump.

V. Electron Accelerator

Fig. 5 shows more details of 21.8 MeV electron linear accelerator.


Fig. 5. 3D rendering of the 21.8 MeV accelerator.

Electrons are generated at the CsKSb photo-cathode, which is inserted into the 112 MHz quarter-wave SRF cavity from the back of the cryomodule (see Fig. 6).


Fig. 6. The 112 MHz photo-injector with cathode stalk insertable from the back of the cryomodule. The cathode is illuminated by a green light from a laser (not shown in the figure). The cavity has a quarter-wave structure, with coaxial fundamental power coupler, which also serves as a fine tuner. Here electrons travel from left to right.

The laser will deliver 100 ps to 500 ps long pulses of green light to the surface of the photocathode and generate near-flat top e-beam profile. This beam will be accelerated up to 2.5 MeV (total energy) by 112 MHz SRF field.

The 112 MHz SRF gun will be repackaged by Niowave Co., while the cathode stalk and transport system is under production at Stony Brook University. The fundamental power coupler (FPC) for this cavity should be designed and manufactured.

Transmitter for the gun is ordered from Tomco Technology, and the circulator is ordered from Ferrite Microwave.

The 4°K helium will be provided by a heat exchanger from the CeC PoP cryosystem (see attachment).

Two 500 MHz room temperature cavities (see Fig. 7) will give energy chirp to the electron beam. Each cavity can provide up to 300 kV RF voltage to provide the necessary chirp. The chirp creates the velocity difference, with the electrons at the head of the bunch having lower velocities that that at the tail.

The cavities will be fed by a transmitter (under purchase from Thomson Broadcast Co) via an AFT microwave’s circulator and a power splitter.

Both 500 MHz cavities require room temperature water-cooling.



Fig. 7. One of two 500 MHz room-temperature RF cavities arrived from Daresbury Lab.

Fig. 8. Top: Design of 5-cell 704 MHz BNL3 SRF linac (without the cryostat). Bottom: Cu model of the BNL3 5-cell cavity during the tests.

The main 20 MeV accelerator will be a 5-cell SRF linac, named BNL3, working at 704 MHz. The BNL3 704 MHz SRF linac cryostat is in a preliminary design stage, the niobium 5-cell SRF structure is in the production at AES. Present approach calls for using BNL1 type cryostat and to place the order for the package with outside company.

The 2°K helium will be provided from CeC PoP cryosystem.

All necessary clock and frequencies (78 kHz revolution clock, 112 MHz, 500 MHz, 704 MHz) will be generated locally from the 100 MHz reference frequency brought from IP4.

VI. The Electron Beam Transport

Three focusing solenoids – identical in the design to that used in R&D ERL - will be used to focus low energy electron beam in the beamline between the gun and the 20 MeV linac. The first solenoid will be mounted on the gun. The other two solenoids, placed between bunching cavities and accelerator, will be used as a pair fired in the opposite direction. Three power supplies for these solenoids are identical to that used in R&D ERL.

The rest of the optics provides for matching of the electron beam through the doglegs and into the FEL wiggler. As we mentioned in the introduction, quadrupoles located at the common trajectory of the ion and electron beams would focus electrons while being almost invisible for heavy ions.

The only common elements whose effect on the ion beam requires compensation are dipoles where the electron and ion beams are merged and separated. Two dipoles, identical to the merging and separating once, but with opposite direction of the magnetic field, are located on the ion-beam trajectory to null the effect on the ion beam. All focusing of the ion beam is provided by RHIC super-conducting quadrupoles located out-side of CeC section. The β-functions of the ion beam are identical in x- and y- directions, are symmetric with respect to the wiggler center where β-functions have minimum of β*=5.5.

In contrast with steady behavior of ion beam size, the electron beam has to be matched into β=1m the helical wiggler as well as having a good overlap with ion beam in the kicker and the modulator sections.


Fig. 9. Desirable β-functions for the ion and electron beams in the IP2 for our CeC experiment.

Two dipoles form a dogleg to bring the e-beam to co-propagate with Yellow beam. The second dogleg takes the electron beam into the dump. Two additional dipoles, energized in opposite direction, compared with their neighbors, serve to make the CeC optics transparent of RHIC beams. Dipole magnets should have the same geometry and the same coil as 60-degree chevron magnets for R&D ERL, except that the gap should be increased to accommodate pipe of 2" ID (e.g. a gap ~ 5.6 cm). Dipole should provide 0.2 T magnetic field. All six dipoles will be fed in series by a single power supply. Each dipole will have individual 1% trim-coil. The dipole magnets will need water cooling.



Fig. 10. General view of the quadrupole magnet.

We are using 16 quadrupoles to focus the 20 MeV electron beam. The design of the quadrupoles will be identical to that of R&D ERL, the quadrupole with ID of 2.362" (60 mm). The power supplies for the quadrupoles can be identical to that of that used in R&D ERL.

Quadrupole will provide gradient up to 0.3 kGs/cm, and its magnetic length of about 16 cm. The field quality requirement is that 12-pole integral ratio is below 1.6×10-4 at a radius of 2 cm.

The beam steering of the low energy electron beam will be performed with three dual plane corrector magnets. The maximal deflection angle is ±3 mrad (± 200 G cm). The 20 MeV electron beam will be steered using corrector coils in the quadrupoles.

Preliminary optics functions of the e-beam are shown in figures 11 and 12. Fig. 11 shows beam optic of the beamline from the end of the 5-cell accelerator to the entrance of the wiggler. Three quadrupoles are used to match the e-beam into the first dogleg. Three dogleg quads provide for its achromaticity as well as a proper beam size at the entrance of the modulator section (β~6m). Four quads keep the beam-size of the e-beam through the modulator section and then focus it to required β=1m at the entrance of the helical wiggler.


Fig. 11. 21.8 MeV e-beam optics from the exit of accelerator to the entrance to the wiggler.

After the wiggler, the e-beam matched by four quadrupoles into the kicker section. As seen in Fig. 12, the following dogleg with two quadrupoles serves as a beam spreader by blowing the e-beam size 10-fold. This expanding beam is absorbed in the dump.


Fig. 12. 21.8 MeV e-beam optics from the exit of the wiggler to the beam dump.

VII. Helical Wiggler

BINP (Novosibirsk, Russia) has designed and manufactures a prototype of the helical wiggler. The specifications for the wiggler are:

Wiggler type Helical

Wiggler period 4 cm

Prototype length 0.50 m

Production unit length 2×3.75 m

Type permanent magnets

Gap 3.2 cm

Aw value for 3.2 cm gap ~ 0.52 (peak field 0.14 T)

Phase errors < ±1 degrees

First integral < 25×10-6 T m

Second integral < 7×10-6 T m2

Helical geometry provides equal focusing in both transverse directions and gives matched β=1m. The matched beam was used to simulate the amplification process in the wiggler. The wiggler is designed with adjustable gap and can be reused for full size CeC cooler in RHIC/eRHIC operations.

At the end of the wiggler, we will install as simple electromagnetic compensated 3-pole wiggler to adjust the e-beam path length for about 20 microns. This wiggler should be designed and manufactured. Focusing of this weak device is negligible even for the electron beam.


Fig. 13. Views of helical wiggler prototype.


VIII. The vacuum system

The vacuum system close to SRF cavities requires particulate-free vacuum system. The rest of the transport beam-line should have 10-9 Torr vacuum. The vacuum system of the CeC PoP beam-line should have two automatic valves at the ends of the doglegs to shut them off from RHIC vacuum in the case of unlikely vacuum accident.

The transport lines should have the pipe-size equal to that in the gun-to-linac (i.e. 2" ID). The only exception is a 4" pipe, which will connect the beam dump to the exit dogleg.

The helical wiggler opening has a square aperture with 32 mm opening. The wiggler will be installed with the 45-degree tilt to maximize the vertical aperture available for beam injection in RHIC.


Fig. 14. Profile of the vacuum chamber in the helical wiggler. The vacuum chamber will be pressure-formed and the corners of the vacuum chamber will be rounded with about 4 mm radius.

IX. The Beam Diagnostics

The electron beam diagnostics is very similar to that used for G5 test at R&D ERL. Preliminary list of the diagnostics is shown below:

  1. Two in-flange Bergoz integrating current transformers (ICT) with beam charge monitors
  2. Fluorescent screens (R&D ERL type) will be used for beam profile and position monitoring
  3. Low energy emittance measurements system will utilize pepper-pot set-up
  4. High energy emittance measurement system will utilize quadrupole scan
  5. Beam Position Monitors
  6. RHIC wall current monitor to monitor ion beam profile
  7. Spectrum analyzer to monitor ion beam spectrum evolution
  8. Infrared Radiation Diagnostics for FEL tuning utilizes 13 microns radiation from wiggler and include intensity detector and/or monochromator
  9. Beam loss monitors to observe ion losses in the wiggler

We are discussing a possibility of using a streak camera and OTR radiation.

The ion beam position will be monitored with existing RHIC pick-up electrodes (strip-lines). The electron beam position will be monitored with button type BPMs.

The flag resolution is 50 microns, which is sufficient to measure emittance of the beam. The flag will placed on the extension of the beam-line after the first dipole. In such configuration the dispersion zero and does not affect emittance measurement. The flag in the dogleg will be used for energy spread measurement, and when acceleration off crest for the bunch length measurement. This flag will allow the measurement of sliced emittance. For this purpose the solenoids will be fed with opposite currents keeping the focusing the same but rotating beam in the XY plane.

The coarse synchronization between ion and electron bunches will be performed by observation signals form the RHIC and electron pick-ups on the oscilloscope.

The matching of the beam velocities (and the fine tuning of the synchronization) will be done by observation of the spontaneous radiation from the wiggler. The ions with significantly larger charge will produce larger shot noise coming from the electron beam due to the Debye screening.

The cooling effect will be observed by modification of the longitudinal profile of ion bunch. We expect to see a growth of the short peak with subnanosecond duration on the top of the 5-nsec profile of the ion bunch. The early detection of the cooling (or heating) of the central part can be detected by observing ion-bunch spectrum in the 1-3 GHz range. The ion beam-induced signal should come from the wall current monitor with nominal bandwidth from DC to 6 GHz.

We are considering a possibility of the port for launching of alignment laser beam for IR diagnostics.



[1] V. N. Litvinenko, Ya. S. Derbenev, Coherent Electron Cooling, Physical Review Letters 102, 114801 (2009), http://link.aps.org/abstract/PRL/v102/e114801.

[2] V.N. Litvinenko et al., Proof-of-Principle Experiment for FEL-based Coherent Electron Cooling, Proceedings of 2011 Particle Accelerator Conference, New York, NY, http://accelconf.web.cern.ch/AccelConf/PAC2011/papers/thobn3.pdf

[3] V.N. Litvinenko et al., Coherent Electron Cooling Demonstration Experiment, Proc. of Second International Particle Accelerator Conference, San Sebastian, Spain, http://accelconf.web.cern.ch/AccelConf/IPAC2011/papers/thps009.pdf


For more details see presentations bellow and papers under the link.



Top of Page