BEAM INDUCED GAS DESORPTION IN THE CERN INTERSECTING STORAGE RINGS

The maximum beam intensity achieved in the ISR has been limited up to now by the beam induced pressure rise which builds up with the stacked proton current. This pressure increase can be explained by ion induced gas desorption from the vacuum chamber. The pressure P, as a function of the stacked proton current I, can be described in good approximation by P = Pa/p-(kn/S)I], where P, is the pressure without beam, S the pumping speed, n the net gas desorption coefficient in molecules per incident gas ion and k is a constant. This concept shows the existence of a critical current at which the pressure goes to infinity. The desorption coefficient depends on the primary ion energy, the type of ions and above all, on the surface conditions. Surface treatments yielding low and even negative values of n are discussed together with experimental results obtained.


Summary
The maximum beam intensity achieved in the ISR has been limited up to now by the beam induced pressure rise which builds up with the stacked proton current.This pressure increase can be explained by ion induced gas desorption from the vacuum chamber.The pressure P, as a function of the stacked proton current I, can be described in good approximation by P = Pa/p-(kn/S)I], where P, is the pressure without beam, S the pumping speed, n the net gas desorption coefficient in molecules per incident gas ion and k is a constant.This concept shows the existence of a critical current at which the pressure goes to infinity.
The desorption coefficient depends on the primary ion energy, the type of ions and above all, on the surface conditions.Surface treatments yielding low and even negative values of n are discussed together with experimental results obtained.

Introduction
The present limit on beam current in the CERN Intersecting Storage Rings (ISR) has been determined above all by the performance of the vacuum system.The average pressure along the 1 km vacuum pipe which has been in the low lo-lo torr range (and a current improvement program is reducing this even further to about 2 x 10-l' torr), yields low beam decay and blow-up rates up to moderate currents.However, when stacking 'to high currents one or more localised pressure rises may appear which lead to a rapid decay and, if stacking is continued, the beam gets lost due to the various types of beam instabilities or the internal dumping mechanism triggered by the vacuum interlocks.Fig. 1 shows a localised pressure bump and Fig. 2   4 ii).This is attributed to the number of gas ions being buried in the chamber wall exceeding the number of molecules released -in effect an ion pump.
For completeness it may be added that electron desorption and desorption by beam loss have been observed but both effects play only a secondary role.

Theoretical
For   One may therefore expect an approximate overall proportionality between n and 6 for high degrees of contamination with a limiting value of -1 for vanishing 0.
At moderate and high energies (e.g.Ei > 500 eV) the desorption term "1 dominates and n will be roughly proportional to the primary ion energy2.At lower energy, especially for clean chambers, n will be determined by both "I and ~1~.The term "2) depending on various phenomena such as ion adsorption, backscattering and ion burial, may account for the systematically observed rapid rise of n at low energies as seen on Fig. 4  and 5.The potential of the proton beam and hence the ion energy Ei depends on beam density,beam-chamber geometry and further it is roughly proportional to the beam current.
Typically one finds between 100 and 200 volts per ampere for the ISR. with PI the equilibrium pressure before dumping the beam.In practice this method ceases to work for small pressure bumps and at very low pressures due to the excessively long response time of the gauge control units.Therefore a second method is used, based on equation ( 5) for the equilibrium pressure.
Here the knowledge of the effective pumping speed Seff is required which can be obtained either from calculation, or by measurement of the dynamic pressure behaviour when the above limitation does not apply.The more rigorous bake to 300° C for 24 hours reduced n in the worst places by at least a factor of two and beam currents of up to 10 A could be obtained.
Repeated bake-outs seem to give still further improvement but the effect is marginal.
Also there are practical limits to lengthening the time to excess of 24 hours and unfortunately only at some places it is possible to increase the temperature above 300° C. Bake-outs at 340° C prove indeed to be very effective as is shown in Fig. 4, comparing o as a function of ion energy for one 300° C and a subsequent 340' C bake-out.
The improvement in terms of n is about one order of magnitude.
In  A proportion of the cont.aminntionSo far argon at 10 -I? torr has been used as the discharge gas since it does not adsorb chemically on the surface and any physisorhed layer can be removed by a subsequent bake-out.
The ion flux can be made very high -best results are obtained with a few l@I8 ions cmm2.The second important effect of the gas discharge is shown in Fig. 6 Experiments in this direction have so far been unsuccessful because ci the low inn flux whic!l can be obtained (about 101? ion cm-' set-'j. iii n(x,o) cos ivx dx, -1, where n(x,o) is the initial gas distribution.The term containing the beam desorption is F = cX2 -cni.Since a finite equilibrium pressure can 0x1~ cx?st for ne- Fig.3Cross-section of the experimental arrangement to study ion induced gas desorption effects. Fig.4 ion bombardment of tile beam.R.e:ucing t!lc ;.Jrface contamination by ion bcmbardmcnt in a pas disc'l3rgc, Khich !imulatrs and aczelerates the effect of the beam, has been successfully stlidied in the lahorntory4.
Fig. 5The dependence of t'le desorption coefficient n on ion energy Ei, (lj before and (2) after .In argon glow discharge tre.ltment :neasured at 6 and 9,5 A respestivcly.r\nother method <of surface cle.lning h:? iOn hw?:b:lrdment consists in using directly t:lc !~cam produced icns (beam cleaning).Experiments in this direction have so far been unsuccessful because ci the low inn flux whic!l can be obtained (about 101? ion cm-' set-'j.