Control of sonoluminescence signal in deionized water using carbon dioxide S. Kumari a , M. Keswani a , S. Singh b , M. Beck c , E. Liebscher c , P. Deymier a , S. Raghavan a,⇑ a Department of Materials Science and Engineering, The University of Arizona, Tucson, AZ 85721, USA b Department of Chemistry and Biochemistry, The University of Arizona, Tucson, AZ 85721, USA c Product Systems Incorporation, Campbell, CA 95008, USA article info Article history: Received 10 May 2010 Received in revised form 24 August 2010 Accepted 28 October 2010 Available online 5 November 2010 Keywords: Megasonic cleaning Wafer damage Sonoluminescence Carbon dioxide Acoustic cavitation Cavitation threshold abstract Megasonic cleaning is routinely employed in semiconductor industry for cleaning of wafers. However, the method also results in damage to wafer features and such damage has been proposed to arise from tran- sient, imploding cavities formed during megasonic processing. Transient cavitation is associated with the release of light, a phenomenon called sonoluminescence (SL) and the extent of damage has been shown to correlate with the intensity of SL. Control of sonoluminescence may therefore allow control of damage during megasonic processing of wafers. In this study, the ability of carbon dioxide to quench sonolumi- nescence generation in deionized water exposed to megasonic field of varying power density and duty cycle has been systematically investigated. It has been found that CO 2 is not only incapable but also a potent inhibitor of sonoluminescence, providing a potential means for selective alleviation of the violent effects of transient cavitation in process fluids. A novel chemical method has been established for in situ release of CO 2 from NH 4 HCO 3 through a pH induced shift in the carbonic acid equilibria in deionized water. Using this method, a precisely controlled, progressive decrease in SL of air saturated deionized water through addition of NH 4 HCO 3 has been demonstrated. It has been determined that 130 ppm of released CO 2 is sufficient for complete inhibition of sonoluminescence generated in air saturated deion- ized water. Published by Elsevier B.V. 1. Introduction Megasonic cleaning has been traditionally used for cleaning of wafers and photomasks in semiconductor processing [1,2]. Mega- sonic agitation is believed to remove contaminant particles pri- marily though exertion of physical forces generated by acoustic streaming and acoustic cavitation [3–5]. However, cavitation also results in wafer damage [6–9]. With the advent of the sub-45 nm technology nodes, the tolerable size of contaminant particle has decreased while wafer features have become increasingly fragile, placing stringent requirements on megasonic cleaning perfor- mance. It has been suggested that damage to wafer features pri- marily arise from violent implosions of transient cavities while cleaning is enhanced by the streaming forces generated by stable cavitation [5]. Transient cavitation is associated with the release of light, a phenomenon known as sonoluminescence (SL) [10]. Con- ditions that generate high SL also lead to increased pattern damage [11,12]. Therefore, selective control of transient cavitation or the associated phenomenon of sonoluminescence may allow develop- ment of damage-free megasonic cleaning of sub-45 nm structures [7,13]. Sonoluminescence has been a subject of extensive research ever since its discovery in 1930s [14–18]. Though it is generally ac- cepted that light is released when high temperature and pressure conditions are reached during the implosion of transient cavities, a rigorous theoretical understanding of the underlying mecha- nisms in SL generation and the origins of light emissions, remain elusive till date [17,19,20]. Numerous factors such as nature and concentration of dissolved gases [15,21], solutes [10,22–24], pres- sure [24], temperature [25], intensity and frequency of sound waves [16,24,26], second order effects arising from interaction of sound waves with solid interfaces [4,27,28], are known to influ- ence cavitation, making it difficult to control SL generation or to deconvolute the role of individual factors in SL generation. The successful development of a portable and UV light tight cavitation threshold (CT) cell that allows the measurement of SL under precisely controlled conditions, has recently been reported. The CT cell has been used to demonstrate the control of SL gener- ation in air saturated deionized (DI) water through dissolved O 2 scavenging [13]. In the present work, the use of CT cell to system- atically investigate the role of dissolved gases such as air, N 2 ,O 2 , CO 2 and Ar, in the generation of SL in DI water is reported. Consis- tent with previous reports, these gases, together comprising 99.99% (V/V) of air, were all found to generate SL signal with the sole exception of CO 2 . Based on the inability of CO 2 to support 0167-9317/$ - see front matter Published by Elsevier B.V. doi:10.1016/j.mee.2010.10.036 ⇑ Corresponding author. E-mail address: srini@email.arizona.edu (S. Raghavan). Microelectronic Engineering 88 (2011) 3437–3441 Contents lists available at ScienceDirect Microelectronic Engineering journal homepage: www.elsevier.com/locate/mee