Chromatography of bis‐(3‐sulfopropyl) disulfide and its breakdown products by HPLC coupled with electrochemical detection

The most widely used accelerators for microvia filling by copper electroplating are 3-mercapto-1-propanesulfonate (MPS) and bis(3-sulfopropyl) disulfide (SPS). In this study, a new accelerator, 3,3-thiobis(1-propanesulfonate) (TBPS), which has neither a thiol group (e.g., MPS) nor a disulfide group (e.g., SPS) but rather contains a thioether group and two sulfonic acid groups, was investigated and formulated to perform microvia filling by copper electroplating. The accelerating effect of TBPS on copper electrodeposition and its chemical interaction with chloride ions and H 2 SO 4 were characterized by galvanostatic measurements. Electrochemical analyzes showed that the acceleration of TBPS on copper electrodeposition strongly depended on the chloride ion and H 2 SO 4 concentrations. An optimal copper plating solution composed of CuSO 4 , polyethylene glycol (PEG), TBPS, chloride ions, and H 2 SO 4 was developed and used to perform bottom-up copper filling of microvias. Although TBPS contains a thioether group instead of a thiol group, a trace amount of TBPS can strongly accelerate copper electrodeposition in the presence of proper chloride ion and H 2 SO 4 concentrations.

mentioning

confidence: 99%

Use of 3,3-Thiobis(1-propanesulfonate) to Accelerate Microvia Filling by Copper Electroplating

Chan

Chiu

Dow

et al. 2013

“…low ppb levels). 36 A few efforts have been described in the literature for analysis of SPS and its byproducts by HPLC with various detection schemes, such as UV spectroscopy, 13,37 mass spectrometry, 17 and electrochemical detection. 38 However, we are not aware of studies that show an adequate separation of SPS by-products, unambiguously assign chromatographic peaks, and have detection sensitivity in the nano-molar range.…”

Section: Spsmentioning

confidence: 99%

“…Experimental HPLC-ED.-All chromatographic analyses were performed on an in-house developed system at Atotech, Inc. (Berlin, Germany). A more detailed description of the methodologies is found in work by Volov et al 36 The set-up consisted of a K-120 gradient pump (Knauer, Inc.) connected to a PEEK pulsation damper (Metrohm, Inc.), which, over a T-piece, led to a Triathlon Spark auto-sampler (Metrohm, Inc.). At the T-piece, an eluent stream split in two streams: one half was sent to the auto-sampler and the other half was passed through an equivalent back pressure column, to maintain the electrochemical cell under flow when the matrix from an analytical column was directed to waste.…”

Section: Spsmentioning

confidence: 99%

“…The retention times have been established based on peak matching relative to standards and were found to be as follows: $11 min for mono-oxide-SPS, $16.5 min for MPS, $19 min di-oxide-SPS, and $64 min for SPS. Further, details on the method development can be found in work by Volov et al 36 Electrochemical.-The electrochemical cell consisted of a platinum rotating disk electrode (diameter 3 mm), Ag|AgCl reference electrode (BASi Re-5), and a platinum wire as a counter electrode. Prior to every experiment, the platinum RDE was pretreated in 0.5 M H 2 SO 4 solution by cycling the electrode potential repeatedly from À0.173 to 1.2 V at 1 V/s until a characteristic cyclic voltammogram was observed.…”

Section: Spsmentioning

confidence: 99%

See 1 more Smart Citation

Interaction between SPS and MPS in the Presence of Ferrous and Ferric Ions

Volov

West

2011

Self Cite

The ferric/ferrous redox couple interacts with bis-(3-sulfopropyl) disulfide (SPS), leading to increased amounts of 3-mercaptopropyl sulfonate (MPS) in solution. SPS is an essential organic additive in the electrochemical deposition of copper from acidic cupric sulfate electrolytes onto printed circuit boards and integrated circuits. Electrochemical studies show that the accelerating action of SPS is dependent on the ferric/ferrous concentration ratio, with weakening dependence at increasing cupric sulfate concentration. HPLC and electrochemical results suggest that an increased concentration of MPS in the presence of the redox couple is the primary cause of the increased acceleration. Equilibrium estimates of MPS generation by the introduced redox chemistry partially explain these observations; the standard reduction potential of SPS to MPS reduction is estimated to be in a range between 0.3 and 0.4 V vs. SHE. However, the amount of acceleration goes through a maximum at an intermediate ferric-ferrous ratio, possibly because SPS may be converted to MPS via both an oxidative and a reductive pathway.

“…[11][12][13] Chloride in these electrolytes is known to be one of the critical constituents enabling defect-free filling of surface features. [14][15][16][17][18][19][20] The main challenge for depositing silver from copper-plating electrolytes, which contain about 50 ppm of chloride, is the low solubility of silver in the presence of chloride ions (the solubility product of AgCl in water at 25…”

mentioning

confidence: 99%

Pulse-Plating of Copper-Silver Alloys for Interconnect Applications

Volov

Swanson

O'Brien

et al. 2012

Self Cite

The electrodeposition of Cu-Ag alloys was studied as a possible application for interconnect technology, where Cu-Ag alloys may be less susceptible to electromigration than Cu alone. The presence of chloride in state-of-the-art copper plating electrolytes limited the solubility of Ag. However, pulse-plating approach enabled a wide range of Cu-Ag alloy compositions at substantial chloride concentration levels. The deposition of Ag was driven by the displacement reactions between the metallic copper and ionic silver during the off-time. Measured alloy compositions were consistent with theoretical estimates at various electrolyte concentrations, electrode rotation speeds, pulse frequencies and duty cycles. However, organic additives decreased incorporation of Ag into the alloy. It was also discovered that CuSO 4 · 5H 2 O from a number of major chemical suppliers contained Ag as an impurity. The roughness of the films was significant when produced by pulsed plating, but was shown to be substantially reduced in the presence of a leveling agent. Additionally, the concentration of chloride in the electrolyte was shown to significantly affect surface quality of the deposited Cu-Ag thin films. With the continuing miniaturization of microelectronics, electromigration effects in copper interconnect systems are becoming a major factor in determining device lifetime and reliability.1-3 Accordingly, there is a need for interconnect materials with improved electromigration resistance, while maintaining adequate electrical resistivity. It has been shown that co-deposits of Cu with small amount of other metals (such as Ag, Sn, Co, and Mg) can potentially mitigate both electro-and stress-migration.4-8 However, the resistivity of the interconnect increases with the addition of foreign metals. The International Technology Roadmap of Semiconductors (ITRS) has set the resistivity criterion at ρ < 2.2 μ · cm, 9 only slightly above the bulk resistivity of pure copper at 1.68 μ · cm. Alloying copper with silver (ρ Ag = 1.59 μ · cm) was shown to increase the resistivity of electrochemically deposited Cu-Ag films the least when compared to other copper alloys.10 For Ag content between 0.17 and 3.2 wt% the resistivity ranges 1.8 to 3.1 μ · cm.10 For this reason Cu-Ag alloys, especially at the lower Ag weight percentages, are potential candidates for the fabrication of interconnects in microelectronic devices.Acidic copper-sulfate electrolytes containing chloride have been successfully applied for many years to the electrochemical fabrication of copper interconnects. [11][12][13] Chloride in these electrolytes is known to be one of the critical constituents enabling defect-free filling of surface features.14-20 The main challenge for depositing silver from copper-plating electrolytes, which contain about 50 ppm of chloride, is the low solubility of silver in the presence of chloride ions (the solubility product of AgCl in water at 25• C is 1.8×10 We demonstrate that the application of a pulsating current instead of a direct current permits ...