Electrodeposition of Cu from Acidic Sulfate Solutions in the Presence of Bis-(3-sulfopropyl)-disulfide (SPS) and Chloride Ions B. Bozzini, * ,z L. D’Urzo, V. Romanello, and C. Mele Dipartimento di Ingegneria dell’Innovazione, Università di Lecce, I-73100 Lecce, Italy This paper reports an in situ Raman study of Cu electrodeposition from an acidic sulfate solution in the presence of bis-3- sulfopropyl-disulfide Na salt SPS. In the absence of chloride, in situ surface-enhanced Raman spectra scarcely show few labile features in a narrow range of cathodic potential. When Cl - ions are added to the deposition bath, several features are clearly visible in the spectra, showing that SPS is adsorbed on the copper surface in a wide potential window during Cu electroplating. © 2006 The Electrochemical Society. DOI: 10.1149/1.2172555All rights reserved. Manuscript submitted June 10, 2005; revised manuscript received December 22, 2005. Available electronically February 24, 2006. The widely investigated processes of Cu electrodeposition for ultralarge scale integration ULSIfabrication rely heavily on the use of complex blends of organics. Notwithstanding the consider- able number of papers published on the subject and the relatively widespread use of such bath chemistries in the industry, the role of the single additives and the reasons underlying the well-document synergistic actions developing among them have not yet been un- ravelled. The organic additives in acid Cu plating baths are commonly categorised as: icarriers, iibrighteners or accelerators, and iii levellers. Carriers and levellers are sometimes denominated suppres- sors, implying the fact that they hinder the deposition rate and con- sequently enhance the electrodeposition overvoltage. Usually, i carriers are polyethylene glycols, iibrighteners are molecules with thiol RSH and disulfide RSSR bonds and sulfonic acid groups, and iiilevellers are molecules with amine functionality or aromatic rings. A host of different chemistries has been proposed in the lit- erature, a review of which is beyond the scope of this paper. In a typical bath chemistry, bis-3-sulfopropyl-disulfide Na salt SPS: NAO 3 SCH 2 3 SSCH 2 3 SO 3 Naacts as the accelerator. 1-28 Accelerators are believed to adsorb on the growing Cu surface and to participate in charge transfer; accelerators coad- sorbed with suppressors would thus offer growth sites on the cath- ode surface, otherwise occupied by other additives exhibiting inhib- iting action. 29 Estimates of the total surface coverage with cathode-blocking species have been evaluated on the basis of electrochemical mea- surements in Ref. 9 for baths containing polyethylene glycol PEG, SPS, and Janus Green B JGB; SPS was shown to lower the total cathode coverage. This scavenging action was reported to be positively correlated with the SPS concentration and to be a function of the potential. If SPS is the sole additive, preferential deposition at the peaks of a saw-tooth profile was found, but use of both SPS and JGB brings about higher deposition rates at the valleys. 3 Enhance- ment of the depolarization effect of SPS in recesses has been dem- onstrated in Ref. 18 by systematic use of through-mask cathodes with different aspect ratios. The accelerating effect has been shown to depend on the pres- ence of Cl - and to be positively correlated with SPS concentration. 18 The use of SPS in the electrodeposition of Cu for ULSI applications has been described in many papers. 1-25,29 Two chief mechanisms have been proposed for the action of SPS in acidic Cu 2+ baths. iSPS is believed to undergo cathodic reduc- tion to MPS. MPS is able to reduce Cu 2+ to CuIthiolate complex. 14,15,17,18,20,25,27,28 In this process, MPS has been reported to be able to reduce back to SPS. 28 SPS would thus take part in repeated redox cycles, causing cathodic depolarization, because a chemical path for the reduction of Cu 2+ to Cu + becomes available. SPS has been proved not to be able to reduce Cu 2+ to Cu + in aque- ous solutions. Formation of MPS from SPS has been suggested to go on, in addition to electrochemical reduction, also by chemical reac- tion with metallic Cu, giving rise to Cu oxidation and SPS reduction to MPS. 27,28 . Positive evidence of this fact has been recently pro- vided by electron paramagnetic resonance EPR. 29 iiSPS has been proposed to form a film of a CuIcomplex, possibly CuI-MPS, at sites where intense mass transport to the cathode takes place; 26,27 electrochemical reduction of this film would yield metallic Cu. Mechanisms iand iican be essentially reconciled apart from details on the exact sequence of RSSR cleavage and CuIcoordi- nation, and the adsorption behavior of the CuIcomplexes. No positive evidence seems to be available regarding these processes, to the best of the authors’ knowledge. The empirical formulation of successful additive systems has been accompanied by the development of several theories of their action—essentially based on integral kinetic equations and mass- transport relationships—predicting current density distributions; re- viewing this topic is beyond the scope of the present paper. In par- ticular, as far as the action of the accelerator is concerned, a curvature-enhanced mechanism accounting for higher deposition rates at the bottom of the features 11 and an accelerator-adsorption model able to predict bump formation 30 were proposed. To the best of the authors’ knowledge, the only papers treating the effects of SPS as the sole additive in Cu baths are Ref. 16 and 25. It has been noted that adding SPS or MPS alone to the plating bath without suppressor and/or Cl - causes polarization of Cu elec- trodeposition. In Ref. 25 the synergistic action of SPS and MPS and Cl - on Cu 2+ has been studied by EPR. Apart from previous work in our group, 31-34 SERS experiments on Cu have been reported in the absence 35-37 and presence of or- ganic additives. 26,38-43 In this paper we describe SERS measurements during Cu ECD from a bath containing SPS and Cl - . Experimental The electrodeposition bath composition was CuSO 4 ·5H 2 O 20 mM, H 2 SO 4 0.5 M, SPS Rasching GmbH, Germany6 ppm. We used analytical-grade chemicals and ultrapure water with a re- sistivity of 18.5 Mcm from a Millipore Milli-Q system. The elec- trodeposition bath was not deaerated for two reasons: ito match the typical industrial conditions and iibecause the current density values recorded during Cu electrodeposition are not measurably af- fected by oxygen reduction. In order to minimize organic contami- nation, all the items in contact with the working electrolyte were rinsed with concentrated HNO 3 before and after each experiment and stored in 20 vol % HNO 3 diluted with ultrapure water. SERS measurements were performed with a LabRam confocal Raman system. Excitation at 633 nm was provided by a 12-mW * Electrochemical Society Active Member. z E-mail: benedetto.bozzini@unile.it Journal of The Electrochemical Society, 153 4C254-C257 2006 0013-4651/2006/1534/C254/4/$20.00 © The Electrochemical Society C254