ECI Technology | We Keep Your Chemistry Right

Leading Provider of Chemical Management Systems for Semiconductor, PV, and PCB Industries

Articles

In situ monitoring of components to control dilute wet chemistries

Materials loss per cleaning cycle must be limited in advanced semiconductor process flows. Using dilute chemistries is necessary to achieve this goal. Fab engineers need to be able to analyze the concentrations of each component in processing solutions in real time to ensure that the chemistries are always in a narrowly defined process window. This article describes how multivariate methods can be applied to analyze chemistries of concentrations ranging from conventional to ultra-dilute levels.

Material loss during cleaning is particularly problematic for advanced technology generations where there are very shallow junctions at the drain and source areas, and any excess removal of silicon can be detrimental to the resulting transistors. The International Technology Roadmap for Semiconductors (ITRS) therefore limits the materials loss of each cleaning cycle at 0.5Å for both oxide and silicon for the 65nm technology node, and 0.3Å for the 45nm technology node [1].

To achieve minimal materials loss, leading R&D groups have resorted to using dilute chemistries for cleaning processes. However, if the solution becomes overly dilute, then cleaning becomes ineffective. If the solution is too concentrated, then materials loss will exceed that permitted by the process specification. Therefore, concentrations of all components in a cleaning solution must be precisely controlled at all times. Effective in situ methods are required to monitor the concentrations of cleaning solutions in order to make necessary timely adjustments by spiking or diluting key components.

Currently adapted monitoring methods such as resistivity/conductivity techniques, though widely used for single component solutions (HF, for example), cannot differentiate contributions from different components in multicomponent solutions, such as SC1, SC2, DSP, etc. Titration and other laboratory techniques, though highly accurate and fully capable of differentiating co-existing components in the same solution, are relatively slow and not ideal for real-time control of solution concentrations. Conventional optical methods, typically measuring absorbance at discrete wavelengths corresponding to the absorption peaks of each component, worked well for legacy concentration ranges such as 1:1:5 or 1:1:20 SC1. But as the solutions become more dilute, the absorption signals are very weak even at the peaks, and the corresponding signal-to-noise ratio becomes very small. In some cases, absorption peaks of key components overlap each other. Such conventional optical methods lack the sensitivity mandatory for analyzing dilute chemistries. An improved method is necessary to analyze highly dilute solutions, such as 1:1:100 SC1 and 1:200 HF for advanced technology nodes.

To reliably analyze ultra-dilute multicomponent wet processing solutions in real time, the method must be noninvasive to both chemicals and wafers, meeting the automatic operation requirements for wafer fabs. The corresponding analytical results can then be fed back to the wet processing equipment through closed-loop control, so that the concentrations are maintained within process windows.

Broadband monitoring

A broadband absorption monitoring hardware system can analyze optical absorption signals of semiconductor wet processing solutions in the near-infrared (NIR, 700-2500nm) wavelength range. However, unlike simpler optical systems that measure at only a very limited number of wavelengths corresponding to absorption peaks, this hardware measures the absorption spectra of the chemical solutions across a wide range of wavelengths at very small intervals. The corresponding spectrum, which includes several hundreds of data points, is then analyzed by a proprietary software routine using a multivariate method [2] to resolve the limited number of unknowns.

In SC1 for example, the only unknowns are the concentrations of NH4OH and H2O2 balanced by DI water. The abundance of data points ensures high accuracy of the analysis; the redundancy of information compensates for noises associated with signal collection. Figure 1 shows examples of as-collected spectra by the hardware before being fed into the microprocessor for further processing.

The solutions were buffered HF, and each widely separated and differentiable spectrum corresponds to a distinctly different NH4F concentration. The two wavelength bands show examples of the discrete ranges typically used by conventional optical methods.

In many real fab cases, particularly for dilute solutions, minute changes in concentrations cannot be easily differentiated from raw data, rendering the conventional optical methods ineffective. In such cases, further processing of the spectrum by a multivariate method is mandatory to achieve the desired accuracy to determine the component concentrations. A schematic of such a setup is shown in Fig. 2a.

Light emitted from a high-stability broad-wavelength-range light source is diverted by a multiplexer into multiple channels through optical fibers to individual measuring modules. The setup is capable of measuring eight or more samples/points of different compositions.

Within each measuring module, piping (in a re-circulation loop or at point-of-use) of the wet processing equipment is encased by a flow-through cell (Fig. 2b).

Figure 02b. The flow-through cell hardware used to take the measurements

Incoming light is partially absorbed by the solution inside the piping before being collected by the optical fiber at the other side of the tubing. Light signals from multiple locations are then detected in sequence by a diode array detector before being forwarded to the computer for signal processing. The flow-through cell blocks stray light from the cleanroom and holds optical fibers onto precision locations and orientations, with respect to piping, to minimize measurement noises.

The flow-through cell is an integral part of the fluid path and does not alter the diameter from upstream or downstream piping. This eliminates concerns about optical distortion due to compression of fluid or release of dissolved air. No optics are in physical contact with the solutions in the piping and the reliability of the setup is expected to be high. Each analysis takes 1.5 sec, including software processing and collecting multiple full spectra at 0.4 millisec/spectrum.

Experimental results

One significant improvement of the current work over the authors’ earlier results [3, 4] is temperature correction. Small temperature variations at measurement points are unavoidable because of temperature changes in the solution and in the ambient, among other factors. Optical absorption signals are known to be extremely sensitive to minute temperature changes. For example, measurements of a dilute HF solution prepared at 620ppm varied perceptibly due to slight changes in temperature despite real-time correction. To resolve this problem, new spectral temperature correction software was developed that fundamentally minimizes measurement variability. The new spectral approach ensures that the measurement is insensitive to temperature variations over several degrees Celsius.

Multivariate measurements of NH4OH in ultra-dilute 1:1:500 SC1 solutions correlated better than 0.995 with the known concentrations of as-prepared solutions. Statistics of measuring H2O2 in SC1 showed similar reliability.

Figure 3 shows concentration changes of the two key components of an originally 1:1:50 SC1 mixture over a 16-hr period in a recirculation loop without active concentration control.

Figure 03. SC1 concentration changes over a 16-hr period in a recirculation loop without spiking/dilution, showing the ability to clearly resolve the components

NH4OH decreases much more slowly than H2O2, thus the latter should be replenished more frequently into the tank. Such different trends cannot be resolved by resistivity/conductivity analytical techniques since contributions from ammonium hydroxide and hydrogen peroxide cannot be differentiated.

The multivariate technique measured several different concentrations of dilute HF between 1000 and 3000ppm (corresponding to approximately 1:570 and 1:190 dilutions, respectively) over an extended period of time. The temperature of the measuring environment was not actively controlled, varying between 20° and 25°C. Results show that all datasets were within 2.5% of the known as-prepared concentrations, and the repeatability was better than 0.7% in all cases. In general, when measuring HF of a 50,000ppm concentration, a standard deviation of 200ppm (0.4% σ) is achievable.

We monitored the concentration of a dilute HF solution as it changed from 1500ppm to 8350ppm by spiking and then back to 1500ppm by dilution. Because of the short 1.5 sec analytical time of each measurement, several data points during mixing were captured in the graph (Fig. 4).

The effectiveness of the current method to measure solutions in transition can help fabs precisely control tank or point-of-use concentrations associated with process recipe changes.

Other cleaning solutions, including BEOL organic solvents, aqueous solutions, HF/H2SO4 blends, as well as many proprietary mixes were measured using the same setup, and similar accuracies as the previously mentioned examples were achieved.

Conclusion

A proven multivariate method measures ultra-dilute single- and multi-component wet processing solutions in real time. Simpler optical methods to measure concentrations often have problems associated with unavoidable small temperature variations. The multivariate method and system can be used to help control the concentrations of critical processing solutions in fabs, vital for necessary wafer cleaning without excessive material removal.

Authored by

Chenting Lin, Peter Bratin, Guang Liang, Michael Schneider, Eugene Shalyt
ECI Technology, Totowa, NJ

Reference

1. “Surface Preparation Technology Requirements: Near-term Years,” Table 68a, Front End Process, ITRS, 2005 Edition.
2. I.T. Jolliffe, Principal Components Analysis, 2nd ed., Springer, 2002.
3. E. Shalyt, et al., “Real-time Monitoring of Dilute Multicomponent Wet Processing Chemistries,” pp. 163-172, 25th SPWCC Proceedings, 2005.
4. Y. Shekel, et al., “Real-time Chemical Monitoring by NIR Spectroscopy,” pp. 245-250, Proceedings of the 208th ECS Meeting, Los Angles, CA, 2005.

Control of Electroless Nickel Baths

This paper reports the authors’ work on developing methods to analyze in-situ key parameters, including pH, nickel concentration and reducing agent concentration of electroless nickel baths. The purpose is to enable production line automated control of the deposition process. The analytical techniques developed within the scope of this work are discussed and their inclusion into an automated chemical monitoring system described. Thousands of data points have been collected to evaluate the system’ s performance. The corresponding results are presented in the context.

Introduction

As PCB manufactures comply with the requirements of lead-free regulations, alternative finishes such as ENIG (Electroless Nickel/Immersion Gold), immersion tin, and immersion silver have become widely adapted. Among those, ENIG provides a highly solderable surface that does not tarnish nor discolor – ensuring a relatively long storage time as compared to other alternative finishes. In addition, ENIG is known as an effective barrier preventing copper diffusion and maintaining solderability of the PCB pads.

One possible drawback of the ENIG finish is the probable nickel corrosion during immersion gold deposition – a defect commonly referred to as “black nickel” or “black pad” [1]. Black pads typically cause solderability failure and therefore need be avoided. The structure and the phosphorous content of the nickel layer are among the key factors in determining the subsequent formation of black pads. Those factors are, in turn, related to the composition and pH of the bath during electroless nickel deposition. Control of the electroless nickel bath is therefore key to defect-free ENIG processes. Bath properties change as solution ages, through consumption of components as well as creation of byproducts. To be able to control the bath, one must know the properties of the bath, and then make adjustment accordingly. Being able to analyze bath properties is thus the first step towards effective control of baths.

Unlike electroplating in which an external circuit provides the electrons to reduce metal ions into metal deposits, an electroless process must use reducing agents to provide the electrons. The most commonly used reducing agent in an electroless nickel bath is sodium hypophosphite. This paper reports the authors’ approaches to analyze divalent nickel (Ni2+) and sodium hypophosphite concentrations as well as pH in the bath.

Analytical Methods and Results

Analyses of pH and nickel concentration were conducted by a Quali-Stream inline chemical monitoring system (ECI Technology, Inc.), Figure 1.

Figure 01. Quali-Stream inline bath analyzer used in this work for controlling electroless nickel baths.

The system samples and analyzes solutions alternately from two production tanks based on pre-set schedules, and the solutions are automatically returned to the original production tanks after analysis. Solution inlets and outlets for multiple tanks are located on the left side panel of the system.

Analysis of nickel concentration is by spectroscopic method, based on Beer’s Law. Incoming light is partially absorbed by the solution under analysis. The higher the nickel concentration, the stronger the absorbance is, resulting in a weaker outgoing optical signal in the corresponding wavelength ranges. The outgoing light is collected by fiber optics and brought to an internal high-performance detector for analysis to ensure sensitivity and accuracy. A calibration curve is built by measuring the absorbance of solutions of known different nickel concentrations (carefully prepared ahead of time), Figure 2.

Figure 02. Calibration curve of nickel concentration

Absorbance of tank solution is measured and the corresponding nickel content is determined by mapping the absorbance to the calibration curve (an automatic process performed by software, while eliminating contribution from other species). Figure 3 shows the performance test results of the system measuring divalent nickel concentration.

Figure 03. Performance check of nickel concentration analysis

More than 4,000 data points were collected over a 3-day period, with the spectrometer automatically calibrating itself periodically. As can be seen from the figure, while analyzing the same standard solution of 6.0 g/l nickel concentration, the analytical method achieved very high accuracy – with the highest reading during the test period being of 6.094 g/l and the lowest 5.922 g/l. Statistical analysis of this data set further affirmed the high reliability of the method, with relative standard deviation at 0.86%.

Measurement of pH was conducted by a pH meter that has been built into the Quali- Stream analyzer. Figure 4 shows the long-term performance test of the system on measuring pH.

Figure 04. Performance check of pH analysis

More than 4,000 data points were taken at the same time as the aforementioned nickel concentration test was performed. The pH output reading had been maintained in a very narrow range throughout the 3-day period, with max at 4.727 pH unit, and min at 4.740 pH unit. Statistical analysis of this data set also showed a small relative standard deviation of only 0.06%. The accuracy of the system’s pH measurement was further affirmed by conducting an additional set of test. In this 2nd set of performance test, pH of one buffer solution was measured at several different temperatures. It’s been well documented that pH readings, even for the same solution, changes with the solution’s temperature. The pH vs. temperature results of this work, presented in figure 5 (blue data set), matched very closely with published data (pink data set), affirming the performance of the system.

Figure 05. Results of measuring pH buffer at multiple temperatures

Analysis of reducing agents was conducted by CVS (Cyclic Voltammetric Stripping) technique, the most widely adapted method to determine organic components concentrations in a copper electroplating bath [2]. The system used in this work to analyze sodium hypophosphite concentration was a Qualilab QL-5 plating bath analyzer (ECI Technology, Inc.). CVS technique applies a cyclic potential onto a platinum working-electrode that is immersed in the working solution (containing copper ions as well as precisely diluted bath sample from the process tank under analysis). The cyclic potential swings between pre-determined positive and negative limits. Copper is deposited onto the working electrode during the negative potential portion of the cycle and then completely stripped away during the positive potential portion of the cycle. The concentrations of additives in the solution affect the rate of copper plating onto the working electrode. When measuring reducing agents, the authors found that the deposition rate of copper in the working solution (note that Cu is the working metal used in the CVS analysis, though the reducing agent concentration in electroless nickel solution is being analyzed) increases monotonically with the increase of reducing agent concentrations.

Figure 6 illustrates the effect of hypophosphite concentration on voltammogram (I-V diagram monitoring the progress of CVS).

Figure 06. Effect of hypophosphite concentration (#1 < #2 < #3 < #4) on Voltammogram

Four carefully as-prepared test solutions of different hypophosphite concentrations gave distinct I-V curves during voltage scan. The enclosed areas under the curves, referred to as ‘peak area’ or Ar, correspond to the integration of current against the applied voltage and are therefore proportional to how fast plating/stripping occurred. A calibration curve, figure 7, plotting peak area vs. hypophosphite concentration can thus be built to compare with the peak area of an unknown solution and accordingly determine the hypophosphite concentration of the unknown solution.

Figure 07. Calibration curve of CVS peak area vs. hypophosphite concentration in the solution

Long-term statistics showed that using CVS to measure hypophosphite concentration could achieve better than 4% relative accuracy and 3.5% repeatability.

Summary and Conclusion

Methods for analyzing pH and nickel concentration in electroless nickel baths have been developed. Engineering efforts based on instrumentation know-how’s have integrated the developed methods into one automated system, enabling PCB production environments to analyze tank solutions in real time. The corresponding long-term results demonstrated both high accuracy and repeatability of the measurements. Closed-loop dosing based on the analytical results can ensure the stability of the electroless nickel bath and give production engineers full control of their parts quality. Additionally, reducing agent in the electroless nickel solution can be measured by CVS technique using a separate lab analyzer, although in this case sampling from the tank needs be performed manually.

The authors have established similar analytical approaches to analyze palladium activation solution, electroless copper solution and electroless cobalt solution, achieving comparable accuracies. Discussions of some of those topics have been published elsewhere [3].

Authored by:

Eugene Shalyt, Semyon Aleynik, Michael Pavlov, Peter Bratin, Chenting Lin
ECI Technology, Totowa, New Jersey, USA

Reference

1. George Milad and Jim Martin, “Electroless Nickel/Immersion Gold, Solderability and Solder Joint Reliability as Functions of Process Control,” CircuiTree, October 2000.
2. D. Tench and C. Ogden, “A New Voltammetric Stripping Method Applied to the Determination of the Brightener Concentration in Copper Pyrophosphate Plating Baths,” J. Electrochem. Soc. n. 125, p. 194 (1978).
3. P. Bratin, et. Al., “Development of Chemical Metrology for Electroless Deposition Baths,” ISTC Proceedings, March 2006.

Detection of Accelerator Breakdown Products in Copper Plating Baths

The mercaptopropylsulfonic acid (MPS) breakdown product of the bis(sodiumsulfopropyl)disulfide (SPS) additive used in acid copper plating baths accelerates copper electrodeposition and can be detected by cyclic voltammetric stripping (CVS) analysis. In the presence of oxygen, MPS decomposes rapidly in acid copper sulfate baths so that the CVS stripping peak area (Ar) decreases on successive cycles. The slope of a plot of Ar vs. log of the CVS cycle number (or time) provides a measure of the initial MPS concentration.

INTRODUCTION & BACKGROUND

Acid copper sulfate baths are employed in the “Damascene” process (1) to electrodeposit Cu within fine trenches and vias in dielectric material on semiconductor chips. Two organic additives are required to provide bottom-up filling of the Damascene features. The “suppressor” additive, which is typically high-molecular-weight polyalkene glycol (e.g., PEG), adsorbs strongly on the Cu cathode surface, in the presence of chloride ion, to form a film that sharply increases the overpotential for Cu deposition. The “anti-suppressor” or “accelerator” additive counters the suppressive effect of the suppressor to provide the accelerated deposition within trenches and vias needed for bottom up filling.

Close organic additive control needed for Damascene plating is provided by Cyclic Voltammetric Stripping (CVS) analysis, which involves alternate plating and stripping of Cu at a Pt rotating disk electrode. The additives are detected from the effect that they exert on the electrodeposition rate measured via the Cu stripping peak area (Ar). The accelerator concentration is typically determined by the linear approximation technique (LAT) or modified linear approximation technique (MLAT) described by Bratin (2). During Damascene Cu plating, however, additive species break down into breakdown products that may interfere with the electrodeposition process. These breakdown products need to be controlled to ensure that high quality Damascene deposits are obtained. A method for detecting suppressor breakdown products was described in our earlier publications (3,4).

This paper describes a CVS method for detecting breakdown products of accelerator additives that are widely used for Damascene copper plating. Results are presented for the 3-mercaptopropylsulfonic acid (MPS) species, which is a breakdown product of the bis(sodiumsulfopropyl)disulfide (SPS) additive (5).

EXPERIMENTAL DETAILS

CVS measurements were made using a Qualilab QL-10® plating bath analyzer (ECI Technology, Inc.) with a polyethylene beaker cell containing 50 mL of solution (open to the atmosphere). For some experiments to verify that oxygen plays a role in MSA decomposition, the cell was partially sealed and deaerated via nitrogen bubbling (stopped during the CVS measurements). The supporting electrolyte contained 40 g/L Cu (added as CuSO4 . 5 H2O), 10 g/L H2SO4, 50 ppm chloride ion, and 2.0 g/L of 5000 molecular weight (MW) polyethylene glycol (Aldrich). The SPS and MPS materials were purchased from Raschig Chemical (Germany). The working electrode was a Pt rotating disk (4 mm diameter, 2500 rpm). Unless otherwise noted, the potential was scanned at 100 mV/s between a positive limit of +1.575 V and a negative limit of either -0.225 and -0.325 V vs. SSCE-M (standard silver-silver chloride electrode modified by replacing the solution with a saturated AgCl solution also containing 0.1 M KCl and 10 volume% sulfuric acid). The counter electrode was usually a stainless steel rod (6 mm diameter). During CVS measurements, the solution temperature was controlled at 24°C within +0.5°C. Specimens of MPS and SPS were injected into the cell at the positive limit in the CVS cycle. The effects of the commercial Viaform™ (Enthone Inc.) and Ultrafill™ (Shipley, Inc.) additives (at normal concentrations) were also investigated.

RESULTS & DISCUSSION

MPS Analysis Method

Figure 1 shows that Ar measured on the first CVS cycle after addition of MPS to the acid copper electrolyte varies monotonically with the MPS concentration. However, a simple Ar measurement cannot be used for MPS analysis since organic additives and other species present in plating baths also affect Ar values.

Fig. 1 Plot of Ar for first CVS cycle as a function of initial MPS concentration in acid copper electrolyte (-0.225 V limit).
Fig. 2 Plots of Ar as a function of CVS cycle number for acid copper supporting electrolyte containing SPS or various concentrations of MPS (–0.225 V limit).

Figure 2 shows that Ar remains constant for acid copper baths containing only SPS, but decreases monotonically with potential cycling in the presence of the MPS breakdown product. Both compounds tend to accelerate the copper deposition rate but the accelerating effect of MPS is stronger and much more time-dependent. After addition of the Viaform™ and Ultrafill™ accelerator additives at the normal concentrations, the Ar values were also constant (not shown) but were smaller (1.5 and 1.4, respectively) than the value of 2.2 observed for SPS at 1.0 ppm concentration. For the potential scan rate and limits used, a CVS cycle corresponded to 38 seconds and copper deposition occurred over a time frame of about 6 seconds. Since the MPS and SPS specimens were injected at the positive potential limit, the copper deposition rate measurement for cycle number 1 began after about 16 seconds and ended at about 22 seconds (onset for copper deposition is about 0.0 V vs SSCE-M). It is clear from these data that MPS decomposes rapidly when its concentration is high, and much more slowly as its concentration decreases.

Figures 3 and 4 illustrate the effects of delaying CVS cycling (after addition of the MPS sample) and interrupting CVS cycling.

Fig. 3 Effects of delays and interruptions in CVS cycling on plots of Ar as a function of CVS cycle number for acid copper supporting electrolyte containing 1.0 ppm MPS (–0.325 V limit).
Fig. 4 Plots of Ar vs. CVS cycle number for which a 3-minute delay was taken into account by shifting the first data point to the 5th cycle (conditions same as Fig. 3).

Note that a relatively negative potential scan limit (-0.325 V) was used to enhance measurement sensitivity. It is evident that Ar continues to decrease unabated even when no voltage is applied to the working and counter electrodes, indicating that MPS decomposes chemically in the bath. When the beginning of the curve for a 3-minute delay was shifted to the 5th CVS cycle (corresponding to 3.2 minutes), the 3-minute delay curve practically coincided with the curve for which cycling had been interrupted for 2 cycles (Fig. 4). Potential cycling appears to actually slow MPS decomposition, possibly because of SPS formation.

Figure 5 illustrates that an exposed copper counter electrode tends to increase the rate of MPS decomposition compared to a stainless steel electrode (or a copper electrode partitioned from the electrolyte via a double-junction glass frit). This effect is relatively small and may result from adsorption of MPS on the relatively large copper counter electrode.

Fig. 5 Effects of counter electrode on plots of Ar as a function of CVS cycle number for acid copper supporting electrolyte containing 1.0 ppm MPS (–0.325 V limit).

Figure 6 shows that the decrease in Ar after MPS additions is exponential since a plot of Ar vs. Log (CVS cycle number) is linear.

Fig. 6 Plot of Ar vs. Log (CVS cycle number) as a function of initial MPS concentration for acid copper supporting electrolyte containing 1.0 ppm MPS (–0.225 V limit).

The theoretical basis for this empirical relationship would be difficult to ascertain from the present data since the measured copper deposition rate is a composite for a range of potentials, and both electrochemical and chemical processes may be involved in the decomposition process. Nonetheless, as shown in Fig. 7, the slope of such plots provides a measure of the initial MPS concentration.

Fig. 8 Effect of SPS concentration on Ar vs. CVS cycle number curve for acid copper supporting electrolyte with 1.0 ppm MPS added.
Fig. 9 Effect of SPS concentration on Ar vs. Log (CVS cycle number) curve for acid copper supporting electrolyte with 1.0 ppm MPS added.

Figure 10 illustrates the effect of negative potential scan limit on plots of Ar vs. Log (CVS cycle number). Linear plots are obtained in all cases, although the slopes vary. Figure 11 shows the dependence of the slope on the negative potential limit. Obviously, the negative potential limit must be held constant for the MPS analysis.

Fig. 10 Plot of Ar vs. Log (CVS cycle number) as a function of negative potential limit for acid copper supporting electrolyte containing 1.0 ppm MPS.
Fig. 11 Plot of the slopes from Fig. 10 as a function of CVS negative potential limit.

Previous work by Healy et al. (6) has shown that SPS and MPS undergo complicated chemical and electrochemical reactions in acid copper plating baths. The SPS species, also known as 4,5-dithiaoctane-1,8-disulphonic acid, is slowly oxidized chemically in the bath but only in the presence of copper metal, although the oxidation rate is accelerated in the presence of oxygen. A complex involving Cu(I) and MPS, e.g., Cu(I)SCH2CH2CH2SO3H, apparently plays a key role as an intermediate in electrochemical reduction of SPS and oxidation of both SPS and MPS in acid copper baths. Nonetheless, Healy et al. conclude that oxidation of MPS via this intermediate does not lead to regeneration of the SPS species. However, Moffat et al. (5) provide convincing evidence that decomposition of MPS eventually results in formation of SPS in acid copper baths.

Figure 12 shows that the CVS stripping peak area (Ar) for a 1.0 ppm MPS acid copper electrolyte decays (after about two days) to a constant value corresponding to that for a 1.0 ppm SPS electrolyte.

Fig. 12. Effect of aging on the Ar values for a 1.0 ppm MPS acid copper electrolyte compared to the constant Ar value for a 1.0 ppm SPS electrolyte.

Since the concentration of the two species based on weight was the same (1.0 ppm), the molar concentration of the MPS electrolyte was initially double that of the SPS electrolyte. The equivalent Ar values observed for the aged MPS electrolyte and the fresh SPS electrolyte indicate that the MPS dimerized to SPS, resulting in the same molar concentration of SPS in both solutions. Thus, our results support the conclusion of Moffat et al. (5) that MPS dimerizes to form SPS in acid copper baths. Our results also indicate that this process is reversible (under some conditions) since the initial Ar value for the MPS solution aged for one day was somewhat greater than the final Ar value from the previous day (fresh MPS solution).

Oxygen also plays a role in the decomposition of MPS in acid copper baths since partial deaeration of the solution was found to significantly reduce the decomposition rate (slow the decrease in Ar value with time). On the other hand, stirring of the solution had little effect, indicating that mass transport is not an important factor. Likewise, removal of the Pt rotating disk electrode from the solution had no effect. Future studies will determine the effects of acidity and copper ion concentration. The goal of this work is to provide metrology that helps tool manufacturers, chemical suppliers, and users to better control acid copper plating processes.

CONCLUSIONS

The mercaptopropylsulfonic acid (MPS) breakdown product of the bis(sodiumsulfopropyl)disulfide (SPS) additive used in acid copper plating baths can be detected by cyclic voltammetric stripping (CVS) analysis. Decomposition of MPS in acid copper baths apparently involves dimerization to SPS, which is accelerated in the presence of oxygen.

Authored By

M. Pavlov, E. Shalyt, P. Bratin and D. M. Tench
ECI Technology, Inc.
60 Gordon Drive, Totowa, NJ 07512

REFERENCES

1. P. C. Andricacos, Electrochem. Soc. Interface, p. 32, Spring 1999.
2. P. Bratin, Proc. AES Analytical Methods Symp., Chicago, IL (March 1985)
3. P. Bratin, G. Chalyt, A. Kogan, M. Pavlov, J. Perpich and M. Tench, “Detection of Suppressor Breakdown Contaminants in Copper Plating Baths”, 203rd ECS Meeting, Paris, France (Apr. 27 – May 2, 2003)
4. M. Pavlov, E. Shalyt and P. Bratin, Solid State Tech. 46(3), 57 (2003)
5. T. P. Moffat, B. Baker, D. Wheeler and D. Josell, Electrochem. Solid State Lett. 6(4), C59 (2003)
6. J. P. Healy, D. Pletcher and M. Goodenough, J. Electroanal. Chem. 338, 167 (1992)

Control of Tin/Lead Solutions for Electrodeposition of Bumps

Ingredients of commercial SolderOn SC and SolderOn BP chemistry as well as samples of aged solutions were provided by Shipley Company. Reagents used were from Aldrich.

Abstract

Typical bumping process includes formation of bumps through UBM copper electrodeposition, followed by deposition of tin/lead coating. Quality control of the tin or tin/lead electroplating solutions is critical to meet demands on the properties of the plated deposit, cost, and environmental issues. Even though lead is being phased-out as enemy of environment and many replacements are being tested, it is still widely used. As all electrochemical processes, tin/lead plating is a dynamic system; unless precisely controlled, concentration of consumable components and breakdown products soon gets outside of acceptable range resulting in manufacturing problems and eventually rejects.

Copper Electrodeposition process for bumping is similar and somewhat less challenging than Damascene copper electrodeposition process used for the interconnects. The control of the Damascene copper process has been a focus of research by all major semiconductor companies in recent years. We have previously demonstrated an on-line controller for complete analysis of copper electroplating solutions1,2. This paper will focus on automated on-line determination of tin/lead coatings, by reviewing analysis of up to 6 components in a commercial tin/lead electrodeposition bath. While the benchtop analysis of tin/lead solutions has been utilized for a number of years, many obstacles exist when transitioning such procedures into automated on-line system due to steps such as gravimetric or extraction procedures.

While determination of metals and acid is reasonably common, measurement of proprietary organic additives remains a challenging task. Organic additives are the key ingredients in the plating solution that influence the properties and quality of the deposits. Cyclic Voltammetric Stripping (CVS) method is an established analytical technique that has long been demonstrated to be applicable to analysis of tin and tin/lead additives in various plating solutions (fluoroborate, sulfate, MSA, and PSA). Detailed description was published in an earlier paper3. This paper will describe an on-line analysis of all components generally used in a tin/lead bumping process. The total automated analysis takes about 40-60 minutes with accuracy better than 10% and reproducibility better than 5%.

Experimental

Experiments were performed using Qualilab QL-10® CVS/Titration benchtop analyzer from ECI technology, Qualiline QLC-7500® on-line analyzer from ECI technology, S2000 Spectrometer from Ocean Optics. Titration was performed using pH and ORP combination electrodes from Thermo Orion. Analysis of three inorganic components (Methanesulfonic acid – H-MSA, Tin (II), and Lead) and three organic components (Primary additive, Secondary additive, and anti-oxidant) was tested and validated using both, benchtop and on-line, analyzers.

A. Analysis of Inorganic components

1. Methanesulfonic Acid (H-MSA)

As for many acids, an acid-base titration of H-MSA with NaOH is a recommended procedure. Manual analysis typically employs a color indicator; use of pH electrode to follow the titration allows for easier automation of this analysis, as well as improved accuracy and reproducibility.

Regardless whether color indicator or pH electrode is employed, titration with NaOH measures “total” acid, which is a sum of tin and H-MSA concentrations, because tin hydroxide formation is at pH values too close to the acid-base titration end-point, making it difficult to separate these two species. To measure “Free” H-MSA, one recommended procedure3 is to add masking agent, such as magnesium EDTA, which ties tin from forming hydroxide or shifts hydroxide formation away from the acid-base endpoint. This allows measurement of free H-MSA. Figure 1 shows titration curve for the free acid.

Figure 1. Determination of Free Acid

2. Determination of Sn (II)

There are two approaches that we have tested for analysis of tin (II) in the tin/lead plating solutions: (a) iodometric titration in acidic solution using starch (manual) or ORP electrode (on-line) as end-point indicators and (b) acid-base titration with NaOH with and without Mg-EDTA masking agent and subsequent extraction of tin from the two experiments.

The acid-base titration with NaOH requires two experiments; first titration of the solution under test with NaOH, where endpoint is proportional to the sum of concentrations of tin (II) and free acid (“total acid value”), and second experiment where Mg-EDTA is added prior to titration with NaOH, resulting end-point volume being proportional to concentration of free acid only. The difference of the two end volumes then yields the concentration of tin (II). This approach yields results with accuracy of 10% and reproducibility of 5%. This method, however, looses accuracy and precision with “high lead/ low tin” formulations, when tin concentration falls below 10 g/l. At these low tin levels the difference between total and free acid is small, and falls within the experimental error for determination of end-point.

The industry recommended determination of Sn (II) is based on its oxidation by iodine in acidic media with starch as color indicator for detection of equivalence point. This method is employed with most benchtop units. Our on-line approach utilizes ORP electrode to automate the equivalence point detection without need for any color indicator. Figure 2 shows curves obtained with this method.

Figure 2. Determination of Tin

Experimental data indicates that “iodometric” titration yields better accuracy and reproducibility than set of acid-base titrations, while the latter has an advantage of simpler analysis, less electrodes, less reagents, improved reliability, and thus lower cost and cost of ownership.

3. Determination of Pb

Two approaches are described in literature:

  • Determination by AAS
  • Oxidation of Sn(II) followed by titration with EDTA

Both procedures contain multiple steps, which are very difficult, if not impossible, to automate.

The technique implemented on the on-line unit is a second step in the analysis of total acid, which is the approach used for analysis of Copper in the Damascene systems. After titration of the solution under test with NaOH to determine total acid (acid + tin), excess of EDTA is added. EDTA reacts with lead releasing additional acid proportional to amount of lead in tested solution, which is then further titrated by NaOH using pH electrode as an indicator. The second equivalence point is then proportional to concentration of lead. Resulting curve is shown in Figure 3.

Figure 3. Determination of Total Acid and Lead

For higher tin concentrations, complete set of 4 analyses (free acid, total acid, tin, and lead) can be determined using set of two titrations with but single electrode (pH) and 3 reagents: Magnesium EDTA, Sodium EDTA, and NaOH. Needless to say, pH electrode is the most robust ion selective electrode, thus yielding highly reliable methods.

B. Analysis of Organic components (additives)

1. Determination of Primary Additive

Primary additive is the key component of the Tin/Lead electrodeposition solution, responsible for grain size and other properties of deposit. It is Primary suppressor and is typically monitored by CVS (Cyclic Voltammetric Stripping) technique. In this technique, activity of the primary additive is measured from its effect on the plating rate of the metal, in this case tin or tin/lead. CVS method employed for the Primary additive is Dilution Titration, where the Supporting electrolyte is “titrated” with solution under test, causing suppression (reduction of plating rate). Titration curve of unknown solution is then compared to that of calibration standard. Details of this method as well as its application to the analysis of the Primary Additive in Tin/Lead systems have been previously described3,5. Figure 4 shows curves obtained for 100mL/L of the Primary Additive with different levels of the Secondary Additive component.

Figure 4. Determination of Primary

There is a wide variety of commercially available Sn/Pb formulations. All additives tested to date display different degrees of the same suppression behavior, and can therefore be monitored using the same technique. The very same technique is used for analysis of suppressors in the copper electroplating processes.

2. Determination of Secondary Additive

Secondary Additive component provides “Secondary” suppression. The suppression effect of this component is much weaker than that of Primary additive. At standard conditions, effect of Secondary component can be seen, but is not quantitatively analyzable due to variation of other components in the solution, even if they are within process spec.

Secondary additive contains dye and most chemical suppliers recommend analysis using a photometric method with extraction, in order to avoid interferences from other ingredients. Extraction technique can be tricky, and is hardly viable for automation.

We have equipped the analyzer with spectrophotometric capability and optimized experimental conditions in order to provide on-line photometric analysis for the Secondary additive component without need for extraction. Figure 5 shows spectral curves obtained for low, target, and high levels of Secondary additive, as well as effect (or better non-effect) of other solution components. Calibration curve is shown in Figure 6.

Figure 5. Spectrometric determination of Secondary
Figure 6. Determination of Secondary

Results with standard and aged solutions indicate that on-line technique without extraction gives results that are consistently 5-10% higher than benchtop extraction technique.

3. Determination of “Active” Secondary Additive

During the electrodeposition additive undergoes multiple transformations, forming a variety of breakdown products. The photometric methods measure analytical concentration of a compound or a group of compounds that absorb in the same frequency region, thus they might or might not include effect of the breakdown products. Since Secondary additive affects plating through its suppression of the electroplating process rather than color, analysis of secondary by photometry might not be the most appropriate analytical method. CVS technique, on the other hand, is an electroanalytical procedure, which monitors concentration of “active” Secondary additive, or its activity.

As mentioned above, the suppression effect of Secondary Additive is quite weak at standard parameters. By optimization of parameters and use of Pulse (CPVS) technique, sensitivity of analysis can be significantly improved. Figure 7 shows response curve for the Secondary Additive utilizing the optimized conditions.

Figure 7. Determination of Active Secondary

Results obtained for various chemistries indicate that in aged solutions, the activity of the Secondary Additive as determined by CVS is about 20% lower than total Secondary Additive concentration as obtained with photometric methods, with activity decreasing with increasing bath age.

During these study, we also compared results of Photometric and Electrochemical analysis obtained from standard solutions, which did not go through electro-plating process. The results obtained for Secondary additive were the same, which confirms the ability of electrochemical (CVS) technique to monitor its “active” concentration. This approach, when two independent techniques are used for analysis of the plating bath, allows to characterize the conditions of plating solutions and its transformations during electroplating process.

4. Determination of Anti-oxidant

Recommended analysis of Antioxidant is performed by photometric analysis after extraction and precipitation. Needless to say, such procedures, similar to the previously mentioned procedure for the Secondary Additive, are very difficult, if not impossible, to automate. We have optimized the analytical methods to allow use of spectrophotometric measurement without need for extraction and precipitation steps, thus allowing us to automate the analysis. Figure 8 shows spectral responses for low, target, and high concentrations of anti-oxidant and effect of other solution components. Figure 9 shows linearity of the response.

Figure 8. Determination of Antioxidant
Figure 9. Linearity of Antioxidant Response

C. Validation of Results

All analytical procedures described above were validated by:

  • repeated analysis of standard solutions with various ratios of components to establish parameters, reproducibility and accuracy
  • repeated analysis of aged sample to establish reproducibility
  • recovery of known dilution of aged sample to establish accuracy and reproducibility
  • recovery of known spike with component of interest into aged sample to establish that we are measuring component of interest, as well as accuracy
  • effect of spike with other components in the matrix into aged sample to evaluate interference from these components
  • correlation of results to recommended benchtop methods

The results obtained for standard solutions and recovery of spike and dilution were within 10% of expected values, as were correlations with recommended benchtop methods.

D. On-Line Analysis

All of the above mentioned analyses were tested individually in the on-line analyzer. Analyzer employs 2 cells; one cell is used for inorganic titrations and second cell for analysis of organic additives by CVS and UV-VIS Photometry. Standard qualification included 10 data points for each of 3 compositions: Target, Low end of process window, and High end of process window. Accuracy in all experiments was within 10% and relative standard deviation within 5%. Cycle time for complete analysis is 40 min (without CVS measurement for activity of Secondary Additive) or 60 min with activity of Secondary Additive. Picture of the online analyzer is shown in Figure 10.

Figure 10. On-line Analyzer for Sn/Pb Electrodeposition Process

Authored By

P. Bratin, E. Shalyt, M. Pavlov, J. Berkmans
ECI Technology, 60 Gordon Drive, Totowa, NJ 07512

References

1. P. Bratin, E. Shalyt, M. Pavlov. “Automated On-line Control of Plating Bath Additives Increases Wafer Yield”, Semiconductor Fabtech –14th Edition, pp. 205-207
2. M. Pavlov, E. Shalyt, P. Bratin. “A New Generation of CVS Monitors Cu Damascene Plating Bath”, Solid State Technology, March 2003, vol. 46, N 3, pp. 57-60
3. P. Bratin, E. Shalyt, M. Pavlov, and R.Gluzman “Control of Tin and Tin/Lead Electroplating Solutions” Proc. IPC’96 , pp S16 1-14
4. Atotech Technical Spec Data Sheets for Sulfotech-M process.
5. P. Bratin “New developments in use of CVS for analysis of plating solutions”, Proc. AES Analytical Methods Symposium, March 1985

A New Generation of CVS Monitors Cu Damascene Plating Baths

Overview

With damascene processing gaining acceptance in high-volume IC manufacturing, reliable analysis of copper plating baths is increasingly important. Cyclic voltammetric stripping has shown that it is an accurate and precise method for the control of the organic additives. The latest generation adds reliability, simplicity, speed, and acceptable cost of ownership.

In copper electro-deposition processing, it is critical to achieve uniformity of the copper film across the wafer surface, the desired mechanical film properties, and super-filling characteristics in trenches and vias. These important properties are typically achieved using multicomponent plating solutions that include organic and inorganic components. There is a need for accurate, sensitive, easy-to-use, and cost-effective on-line monitoring metrology.

table 01. comparisons of previous and new generations of CVS analysis

Copper electroplating baths are normally formulated using highly stable electrolytes containing copper sulfate and sulfuric acid. Copper concentration in these electrolytes is 14 to 60 g/1 and sulfuric acid 1 to 240 g/1. Other components added into the bath are present in relatively small amounts. These components are organic additives and chloride ions. The organic additives, depending on the concentration and chemical composition, affect the properties of electro-deposited copper, including uniformity, hardness, ductility, and tensile strength. The control of the additives is known to be critical in maintaining the desired properties of the deposits.

Organic additives added to copper electroplating baths fall into three major categories:

  • Accelerators are compounds that contain pendant sulfur atoms that locally accelerate deposition where they are adsorbed.
  • Suppressors are polymers, such as polyethylene glycols (PEGs), that have the ability to form a current-suppressing film on the entire wafer surface, especially in the presence of chloride ions. (Chloride enhances the suppression.)
  • Levelers are secondary suppressors and work only on protruding surfaces where mass transfer mechanisms are most efficient.

Organic components are consumed during an electroplating process at various rates, thus requiring individual control. Accelerators, for example, are consumed faster than the other organic species. There are other interactions between the organic additives and inorganic compounds during the electroplating process that cause decomposition and modification of initial organic compounds. These by-products can affect the deposition process, sometimes significantly more than the initial organic compounds. This means that monitoring the pure analytical concentrations of the original organic additives is insufficient to control copper electroplating bath performance. The presence of multiple components of the additives and breakdown products make it necessary to monitor total effective concentration (“activity”) of individual bath components, which affect suppression, acceleration, and leveling properties. At the same time, the accumulation of breakdown products might increase to a level where they can seriously affect the plating process. Monitoring of bath aging is therefore another important aspect in the analysis and control of copper electroplating solutions.

The CVS method

Among the wide range of analytical approaches that have been used to monitor the concentrations of organic additives in electroplating solutions, only cyclic voltammetric stripping (CVS) has demonstrated the ability to reliably monitor the activity of a wide range of components in plating baths. In this method, the classic three-electrode cell is used, where the main indicator is a platinum (Pt) rotating-disc electrode [1]. CVS involves cycling the potential of the rotating-disk electrode so that metal is alternately plated and stripped at its surface. Organic additives are detected by CVS from the effect that they exert on the electro-deposition rate measured via the metal stripping peak area.

Figure 1. Performance comparisons of previous and new generations of CVS analysis.
Figure 2. Endpoint CVS dilution titration volume for -0.225V and -0.400V limits as a function of the weight fraction of 900 MW PEG (in mixtures with 10,000 MW PEG).

This approach has become the most popular method for controlling damascene copper deposition processes. The initial method was to adapt bench-top analyzers that have been used for more than 20 years by PC-board manufacturers. IC manufacturers soon discovered that the far more critical requirements of their process often created a need for non-stop on-line measurement of the plating solution.

On-line plating bath analysis

As electroplated copper deposition has become an integral part of the latest semiconductor manufacturing process, CVS chemical monitoring systems have established themselves as well, forming a closed-loop system with the plating tools.

The first generation of CVS-based analyzers was developed in 1997 and controlled only two components of organic additives. These systems were used in several research locations and produced valuable results for understanding and improving the plating process.

One of the lessons from these early monitoring systems was that both organic and inorganic components of plating solutions needed to be monitored, and organic additives need to be significantly modified from their original formulations. More organic components were introduced into the bath and concentrations of inorganic components were modified.

A second-generation analyzer — capable of controlling six components of a plating solution — was introduced in 1999 and became a popular system for analysis of copper electroplating solutions for high-volume manufacturing by virtually every major semiconductor facility.

Today, we have a new, third-generation CVS-based monitoring system that has been designed to handle the challenges of 300mm wafer processing. This system was developed to answer the challenge, based on experiences collected from more than 100 analyzers running worldwide at various semiconductor facilities. The major differences of this analyzer from the previous design include its dual-cell design and significant advance in its software, giving improvements in accuracy, precision, reliability, cost of ownership, and time of analysis.

Today’s CVS

With the dual-cell configuration, organic additives are analyzed in one cell and inorganic analysis is performed in the second cell. This significantly enhances analytical capabilities. The biggest benefit of dual-cell design is the reduction in analysis time from 100 min using a previous single cell configuration to <40 min/cycle for all six components in the dual-cell configuration. Faster analysis is achieved by simultaneous analyses and by use of new algorithms designed to reduce total time of analysis, chemical consumption, and waste.

The overall benefits of the shorter time to first analysis yields a decreased time delay in bringing a plating tool up after prolonged downtime, a decrease in cross-contamination between analyses, allowing for reduction in rinsing and cell preparation, and significant reduction in conversion from one chemistry to another, including method development if required.

Figure 3. CVS endpoint volume for standard and aged production solutions using two parameter sets (a and b).

Improved accuracy and precision is another important aspect of dual-cell design (see table). Single-cell analyzers have successfully passed all specifications set by users and tool manufacturers, but these specifications are continuously changing due to demands on plating processes (i.e., the need to fill finer features on larger wafer areas). Today, these demands are being achieved through decreased chemical interference and thus decreased effect on electrodes during an analysis cycle.

Since organic additives and inorganic components are analyzed simultaneously and in separate cells, pre-conditioning and cleaning of various electrodes before analysis is significantly reduced. This is very beneficial for stability of analytical results during a long process run, as well as for the reliability of the analyzer. Better stability of electrodes results in less frequent preventive maintenance procedures and thus lower cost of ownership.

Even with the increased performance of this generation of CVS, the hardware is more straightforward, with improved solution delivery and electronics. Improvements include such critical parts as syringe modules, valves, electrochemical circuitry, and electrodes. Operation of the system is also easier via new user-friendly software. Changes include new analytical procedures for accelerator and suppressor, new rinsing and conditioning procedures, new drivers, and new analysis parameters. A new combined procedure for suppressor and accelerator completely eliminates the need to replace solutions, as well as to rinse and condition electrodes between analyses, producing a drastic reduction in consumption of reagents and waste generation, as well as analysis time and maintenance frequency (Fig. 1).

Suppressor breakdown products

To detect suppressor breakdown products with this new generation of CVS, we used CVS dilution titration analyses of a copper-plating bath using two parameter sets. The volume needed to reach endpoint using a first parameter set with a -0.4V limit yields a direct measure of the active suppressor component concentration, since the suppressor breakdown contaminants are not effective at suppressing the copper deposition rate with these parameters. The volume needed to reach endpoint using a second parameter set with a -0.225V limit yields a measure of the concentration of the active suppressor plus the suppressor breakdown products. The ratio of the two endpoint volumes provides a relative measure of the suppressor degradation.

To confirm the presence and behavior of the breakdown products, we performed CVS dilution titrations with -0.400V and -0.225V limits for solutions containing 1.0 g/L of PEG polymer comprising mixtures of the 900 and 10,000 molecular weight (MW) species. Figure 2 shows plots of the CVS endpoint volumes for the two potential limits as a function of the weight fraction of the 900 MW species. For the -0.400V limit, the required endpoint volume increases (activity decreases) as the fraction of the 900 MW species increases, since the concentration of the 10,000 MW species is decreasing and the 900 MW species is relatively inactive as a suppressor for this potential limit. For the -0.225V limit, the endpoint volume exhibits a slight decrease as the weight fraction of the 900 MW species increases, since both PEG species are effective suppressors at this potential limit and the overall PEG molar concentration is increasing.

In addition, we performed CVS dilution titrations for solutions containing fresh standard and aged production solution, using both parameter sets (Fig. 3). Results show that the response with parameter set 1 was the same for both solutions, while with parameter set 2, the fresh solution showed higher activity (less volume needed to suppress VMS).

We also looked at how decreasing molecular weight of PEGs changes their CVS response and therefore their suppressing activity (Fig. 4). As a bath ages, more PEGs are cleaved into shorter (lower MW) fractions, thus affecting their performance.

Figure 4. CVS response of decreasing molecular weight of PEGs.

Recently, possible mechanisms for functioning and breakdown of accelerators have been presented at the Electrochemical Society meetings. We plan to publish our work describing analysis of breakdown products of accelerators and levelers in the near future.

Authored by

Michael Pavlov, Eugene Shalyt, Peter Bratin
ECI Technology, Totowa, New Jersey

Reference

1. P. Bratin, et al., Semicondutor Fabtech, 14th Edition, Summer 2001.

 

Michael Pavlov received his MS in electrochemistry from the Moscow Institute of Steel and Alloys. He is product manager at ECI Technology, 60 Gordon Drive, Totowa, NJ 07512; ph 973/890-1114, fax 973/890-1118, mpavlov@ecitechnology.com.

Eugene Shalyt received his PhD in electrochemistry from Mendeleev University of Chemical Technology in Moscow. He is R&D manager at ECI Technology.

Peter Bratin received his PhD in electroanalytical chemistry from CUNY in New York. He is VP at ECI Technology.

Control of Damascene Copper Processes By Cyclic Voltammetric Stripping

Use of electroplated copper for onchip metallization in ultra-largescale integrated circuits (ULSI) devices is gaining momentum because of the low cost and high throughput of the process. Electroplated lines and trenches with submicron dimensions, however, are strongly affected by changes in the composition of the plating solution, thereby creating a high demand for control techniques. The most dynamic ingredients of electroplating solutions are organic additives. Even a small imbalance between components of the additive system can cause various defects in the filling of the trenches and properties of electroplated copper. On-line monitoring and control of these additive components is therefore desirable. Cyclic voltammetric stripping (CVS) analysis has long been used for just such a purpose in the manufacturing of printed circuits.

Numerous chemical and electrochemical processes (i.e. thermal decomposition, cathodic reduction, occlusion in the deposit, anodic oxidation, drag-out) contribute to additive consumption, as does timedependent chemical decomposition in the bulk solution. Additive dosing is often done based on time or amp hours. Some companies have developed proprietary algorithms that predict additive consumption. However, even the best prediction algorithm requires an analytical tool to fine-tune the algorithm’s coefficients, and it cannot predict changes due to equipment failure and human errors. There is a variety of analytical equipment for monitoring the analytical concentration of plating additives. However, only CVS can monitor activity of individual components in a whole range of proprietary organic additives. Over the years, CVS has become a standard for monitoring these additives in copper plating solutions in the printed circuit industry, and is currently being tested by the majority of companies involved with Damascene processes.

With the CVS technique, the potential of the inert electrode is cycled at a constant rate in the electroplating bath, so that a small amount of Copper is alternately deposited on the electrode surface and stripped off by anodic dissolution. The measured area under the stripping peak is proportional to the plating rate, which strongly depends on active concentration of plating additives and their components (Fig. 1).

A modification of the CVS technique, cyclic step voltammetric stripping (CSVS) employs a series of potential steps, instead of the linear sweep, to measure the effect of additives. CVS and CSVS scans, combined with proprietary analytical techniques, allow independent determination of up to three components (brightener, carrier and leveler) of the additives, even in cases of premixed additive system (Figs. 2-3).

Additional applications of the CVS technique include monitoring of plating bath ageing. While most PWB manufacturers use bench-top analyzers, the requirements of the semiconductor industry are much more demanding. The small volume of plating solution creates very dynamic changes in chemical component activity that, when combined with a narrow operating window, generates demand for continuous on-line, closed-loop analytical systems. Other requirements adding to the value of on-line units are: (1) high frequency of analysis; (2) sample size limitation; (3) clean room requirements; and (4) automation concerns.

Demand Has Caused Changes

The demands of the semiconductor industry have led to the recent redesign of the proprietary on-line acid copper bath controller* to meet the particular requirements of the semiconductor industry (Fig. 4).

The unit, supported by all major additive suppliers, samples up to four plating tanks and standard solutions. It incorporates analysis for brightener, leveler, carrier and contamination level, and it replenishes the plating solutions as needed. The accuracy and reproducibility of on-line analysis are significantly better than lab instruments because the on-line unit includes temperature control, automatic calibration, reproducible conditioning and it eliminates human variations. Table 1 summarizes results of online analysis of several Damascene plating solutions at high, middle and low levels of additive specifications.

Results indicate very good accuracy and reproducibility of analysis for all components. Figure 5 illustrates the history of one plating solution followed for 160 measurements.

Before measurements #79 and #130, the customer performed carbon treatment procedures. In both cases, treatment effectively removed brightener, but left intact most parts of carrier the component.

Early Warning

Under ideal conditions, Damascene plating is performed at steady-state conditions, and readings in production solution are almost as stable as in standard solution. However, an online analyzer provides an early indication of process malfunctioning– e.g., dosing pump failure (Fig. 6).

To test response of CVS to changes of additive levels, a series of additive spikes to a production solution was performed and measured. Figure 7 shows the results obtained for these spikes.

Because of the extremely high cost of wafers, a significant amount of work is being performed to correlate CVS results to the performance of the plating baths with different levels of additives and inorganic components. Although organic additives present the main challenge in controlling plating solution, one needs to monitor inorganic components as well. Therefore, a new generation of CVS units is capable of monitoring both organic and inorganic components (copper, acid and chloride). All . components are analyzed by potentiOmetric titration using proprietary procedures. Table 2 summarizes the results of analysis for standard solution.

At the preparation time of this paper, the analyzer was tested at 10 facilities for an overall period of seven years, including four facilities performing Damascene plating for a 12-month period. Testing included several 5,000-wafer marathons. Use of the analyzer allowed manufactures to keep the process in a very tight window, resulting in a high yield of wafers. Results of the analyzer are in excellent correlation with results of bench-top analyzers. Results of longterm testing of the analyzer at customer facilities are now available.

Authored by

Peter Bratin, Eugene Shalyt & Michael Pavlov

Terms and Conditions - Privacy Policy