* Corresponding author. Tel.: 0041 22 7676387; fax: 0041 22 7678380; e-mail: chiggiato@cern.ch Vacuum 53 (1999) 317—320 Surface cleaning efficiency measurements for UHV applications C. Benvenuti, G. Canil, P. Chiggiato*, P. Collin, R. Cosso, J. Gue´rin, S. Ilie, D. Latorre, K.S. Neil EST division, SM Group, CERN European Laboratory for Particle Physics, CH-1211 Geneva 23, Switzerland Abstract High-energy particle accelerator UHV systems subjected to energetic charged particle or photon bombardment are particularly sensitive to surface contamination. Well-established cleaning procedures employing solvents such as Freon and perchloroethylene to prepare vacuum system components have been used at CERN for many years. The recently adopted legislation which prescribes or strongly limits the use of some chemicals has led CERN to undertake a qualification programme using electron stimulated desorption, Auger spectroscopy and radioactive tracers to compare the cleaning efficiency of a wide range of alternative cleaning agents. It is found that detergents are preferable to solvents for cleaning UHV components. Some solvents have been found to be as good, if not better, than Freon when detergents cannot be used.  1999 Elsevier Science Ltd. All rights reserved. 1. Introduction Beam lifetime, which largely depends on the gas den- sity in the vacuum beam pipe, is one of the most impor- tant parameters for particle accelerators [1]. Under static conditions, the residual partial pressures of individual gas species are determined by the static thermal outgass- ing of the vacuum chamber walls and the available pumping speed. In the presence of circulating beams, dynamic degassing also occurs due to desorption induced by beam via synchrotron radiation and/or ion bombard- ment [1]. The continual trend to reduce the aperture of the vacuum chamber together with an increase of the circula- ting beam current [2] results in higher gas loads with lower pumping conductance. This has motivated several developments in vacuum system design. Over the last few years, the available pumping speed has been increased by adopting new linearly distributed pumping solutions [3—5]. A significant effort has also been made to reduce the vacuum chamber degassing by appropriate surface [6] and thermal [7] treatments. The main step for surface conditioning is chemical cleaning which, for certain applications, can be com- pleted by other more sophisticated treatments such as UV [8], RF plasma [9] or glow discharge [10, 28] cleaning. In 1987, the United Nations Environment Pro- tection (UNEP) office approved the Montreal protocol [11], which called for a freeze on the production and use of chemicals which may deplete the atmospheric ozone layer. Three years later the protocol was significantly strengthened by what is known as the London Amend- ments. One of the main effects of the new regulations has been the ban of two widely used solvents [12], namely Freon (CFC-based) and 1-1-1-Trichlorethane (methyl chloroform). Many cleaning substitutes have since been elaborated and commercialised by major chemical companies for a wide range of applications. Some high-energy physics laboratories have launched characterisation programmes [13—18] to test the efficiency of new products in the specific UHV accelerator environment. The present strat- egy is to use detergents whenever possible, solvents being limited either to components for which final water rins- ing is not recommended, for instance bellows and ceram- ics, or as a pre-cleaner before a final treatment with detergents. This trend has also been adopted in other technological fields where surface cleaning is mandatory [19]. The aim of the present work has been to characterise the cleaning efficiency of new chemicals and compare it with that of the standard CERN cleaning procedure [20] 0042-207X/99/$ — see front matter  1999 Elsevier Science Ltd. All rights reserved. PII: S 0 0 4 2 - 2 0 7 X ( 9 8 ) 0 0 3 6 4 - 9