Determination of Additives and Byproducts in an Acid Copper Plating Bath by Liquid Chromatography

Significance of topic

Copper electroplating baths play a critical role in semiconductor manufacturing by enabling high-quality copper deposition on wafers. Proprietary additives control deposit characteristics, while byproducts formed during plating can degrade performance. Accurate quantification of individual additives and byproducts is essential for process monitoring, bath maintenance, and yield optimization.

Objectives and overview of study

This work presents a liquid chromatography method using a polymeric reversed-phase column and absorbance detection to separate, identify, and quantify key additives and their byproducts in acid copper plating baths. Two gradient programs, steep and shallow, were developed to resolve Accelerator, Leveler, and electrolysis byproducts.

Methodology and instrumentation

• System: Dionex DX-600 LC with GP50 gradient pump, AD25 absorbance detector or PDA-100 diode array detector, and AS50 PEEK autosampler.
• Fluid path: PEEK tubing and sample loops (10 and 500 μL) to resist corrosion from sulfuric acid and copper sulfate matrix.
• Columns: IonPac NS1 analytical (4×250 mm, 10 μm) with NG1 guard (4×35 mm, 10 μm), capable of pH 0–14 operation.
• Eluents: 100 mM and 150 mM sulfuric acid, acetonitrile gradients (5 to 90 or 0.25 to 9 %) in sulfuric acid.
• Gradient methods: Steep gradient (5 to 90 % acetonitrile over 20 min, 1 mL/min) and shallow gradient (0.25 to 9 % over 18 min, 2 mL/min).

Main results and discussion

• Steep gradient resolved Accelerator peaks at 5.8 and 6.5 min and Leveler at 17.3 min, despite a large matrix peak from 2 to 5 min. Detection at 190 nm provided clear chromatograms.
• Shallow gradient improved separation of two Accelerator components from the copper matrix; detection at 246 nm gave optimal signal to noise.
• Calibration for Accelerator (1.67 to 11.1 mL/L) and Leveler (0.23 to 1.52 mL/L) showed r²>0.99 and precision RSD<10 % over multiple days.
• Aged bath samples injected at 10 μL revealed two byproduct peaks at 3.4 and 6.2 min not present in fresh bath. Spiking with synthesized standards confirmed their identity.

Benefits and practical applications

• Enables individual quantitation of additives and byproducts, unlike cyclic voltammetric stripping which measures only combined electrochemical activity.
• Inert PEEK fluid path prevents corrosion issues with acidic, copper-rich samples common in standard HPLC systems.
• High sensitivity and reproducibility support routine quality control, bath recharge scheduling, and additive optimization.

Future trends and potential applications

• Integration of diode array or mass spectrometric detection for wider analyte coverage and structural characterization of unknown byproducts.
• Automation and inline sampling for real-time bath monitoring in manufacturing environments.
• Development of miniaturized or high throughput LC platforms tailored to plating bath analysis.

Conclusion

A robust LC method using IonPac NS1/NG1 and PEEK-based fluidics has been demonstrated for selective separation and quantitation of copper bath additives and byproducts. The approach offers high sensitivity, reproducibility, and resistance to corrosion, making it suitable for routine monitoring of acid copper plating baths in semiconductor production.

References

1. Hurtubise R Too E Cheng CC Future Fab Intl 1998 243–245
2. Lin XW Pramanik D Solid State Technol 1998 41 (10) 63–79
3. Plieth W Electrochim Acta 1992 37 (12) 2115–2121
4. Kelly JJ Tian C West AC J Electrochem Soc 1999 146 2540–2545
5. Tench D Ogden C J Electrochem Soc 1978 125 194
6. Ogden C Tench D Plat Surf Fin 1978 66 30
7. Haak R Ogden C Tench D Plat Surf Fin 1982 69 62
8. Freitag WO Ogden C Tench D White J Plat Surf Fin 1983 70 55
9. Reid JD Plat Surf Fin 1988 75 108–112
10. Heberling S Campbell D Carson S PC Fab 1989 12 (8) 72
11. Taylor T Ritzdorf T Lindberg F Carpenter B Solid State Technol 1989 41 (11) 47–57
12. Weiss JJ J Chromatogr 1986 353 303–307