All Experimenters Meeting Fermilab, February 21, 2011 COLLIMATION STUDIES WITH HOLLOW ELECTRON BEAMS G. Stancari This talk is about low-energy electron beams and how they interact with protons and antiprotons in the Tevatron. This project is leading to a new kind of collimator for high-intensity beams in storage rings and colliders. We are dealing with beams with hollow current density profiles. In a hollow electron beam collimator, electrons surround the circulating beam. Their electric charge kicks halo particles transversely. If their distribution is axially symmetric, the core of the circulating beam does not experience any electric field. For typical parameters, the kick experienced by 980-GeV protons is of the order of 0.2 microradians. The electron beam is generated with a pulsed electron gun and transported with strong axial magnetic fields, in an arrangement similar to electron cooling or to the existing Tevatron electron lenses. What is unique about this system? In a conventional two-stage collimation scheme, a primary collimator (target) imparts a random transverse kick due to multiple scattering. The affected particles have increasing oscillation amplitudes and a large fraction of them is caught by the secondary collimator (absorber). These systems offer robust shielding of sensitive components. They are also very efficient in reducing beam losses at the experiments. The classic two-stage system does have limitiations. In high-power accelerators, no material can be placed too close to the beam. The minimum distance is limited by instantaneous loss rates, radiation damage, and by the electromagnetic impedance of the device. Another problem is beam jitter. The orbit of the circulating beam oscillates due ground motion and other vibrations. Even with active orbit stabilization, the beam centroid can oscillate by tens of microns. This translates into periodic bursts of losses at aperture restrictions. With the hollow electron beam collimator we are trying to address these limitations. We are studying whether this concept is viable as a complement to conventional systems. A magnetically confined electron beam is stiff. And it can be placed very close to, and even overlap with, the circulating beam. The transverse kicks are small, so that it acts more like a "soft collimator" or a "diffusion enhancer", rather than a hard aperture limitation. After some preliminary modeling and simulations, we decided to test this concept experimentally. A 15-mm-diameter hollow electron gun was designed and built in 2009. It is a tungsten dispenser cathode with a 9-mm-diameter hole. The gun was tested and characterized in the electron lens test stand, located in the lower linac gallery. The maximum yield of this gun is 1 A at 5 kV. We installed the gun in one of the Tevatron electron lenses in August 2010. There are 2 electron lenses in the Tevatron, TEL1 and TEL2. The electron beams are pulsed and can be timed with any bunch or group of bunches. The maximum confining field is 6 T. TEL1 is used during normal operations for cleaning the abort gap. TEL2 is a backup for TEL1 and it was used for studies. Experiments began last October. We have gained a lot of experience with this system. We measured its effects under various experimental conditions: beam current, alignment, pulsing pattern, collimator configurations. I will focus here on some examples of the electron beam acting on antiproton bunches. The first question we addressed is the following: can these small, 0.2-urad transverse kicks really remove particles from the beam? In the experiment on Slide 8, the electron lens was turned on the second antiproton bunch train about 1 hour after the beginning of a regular collider store. The size of the hole was 4.5 sigma and 5 sigma, respectively. The plot on Slide 8 shows the intensity of each bunch train as a function of time. The black trace is the TEL2 voltage. One can already see that train number 2 is being scraped. To isolate the effect of the hollow beam, one can look at the ratio between the intensity of the affected train and the other two trains, shown on Slide 9. One clearly sees the smooth scraping effect: 2.5%/h with the 4.5 sigma hole, and 0.32 %/h with a larger hole. The other important question is whether there are any adverse effects on the core of the circulating beam. Does the hollow beam introduce noise and cause emittance growth? Are particles really removed from the halo? This is a concern because the overlap region is not a perfect hollow cylinder. We can approach the problem from three points of view. First, by looking at the evolution of the emittances. In the plot on Slide 11, the evolution of emittances for the affected bunch train are shown. The hollow beam does not produce additional emittance growth. If there is emittance growth, it is much smaller than the one driven by the usual factors: beam-beam, intrabeam scattering, beam-gas scattering, and rf noise. Secondly, we can compare beam scraping with the corresponding decrease in luminosity. Luminosity is proportional to the product of antiproton and proton populations, and inversely proportional to the squared rms of the overlap region. If antiprotons are removed and the other factors are unchanged, luminosity should decrease by the same relative amount. If emittance growth or proton loss are caused by the hollow beam, luminosity should decrease even more. A smaller relative change, or no change, in luminosity is a clear indication of halo scraping. In the plot on Slide 12, one can see how the luminosity for the affected bunch changed with time relative to the other bunch trains. During the first scrape with small hole, intensity was reduced by 1.4%, whereas luminosity only decreased by 0.57%. Even clearer was the second experiment with 5 sigma hole: a 0.39% reduction in intensity was accompanied by no detectable change in luminosity (less than 0.05%). From this and other experiments, we conclude that it is possible to remove particles from the halo without affecting the core. The third approach is to try to measure the halo population directly. This can be done by scanning a collimator in small steps and observing the corresponding beam loss. The time evolution of losses can also be used to estimate the diffusion rate as a function of amplitude. On Slide 13, one can see the preliminary results of a collimator scan. The hollow beam was acting on the second bunch train with a 3.5 sigma hole. About 1% of the beam was scraped by the hollow-beam. The vertical antiproton collimator at F49 was moved downward in 50-micron steps. All other collimators were retracted. The plot on the left on Slide 13 shows how much beam was lost at each step for each of the 3 bunch trains. One can see that there is a region in which the population of the affected train is about 40% lower than the other trains. As expected, populations tend to be equal towards the beam axis and far away from it. The plot on the right on Slide 13 is the ratio between the fractional beam loss of the affected train and the average of the other two, to appreciate the same effect in more detail. For comparison, the ratio of train 1 with train 3 is also plotted. The ratio of populations for these two control trains appears to be constant. We plan to continue the experimental study in the next few months, if possible. We want to compare diffusion measurements with and without electron lens; study the capture efficiency as a function of hole size in a collider store; measure the effect on scraped-down protons in a special proton-only store. In parallel, we are designing a larger cathode for use with protons. A guest scientist joined us this week to work on modeling and simulations.