Electrochemical and Solid-State Letters, 7 共3兲 C31-C34 共2004兲

C31

0013-4651/2004/7共3兲/C31/4/$7.00 © The Electrochemical Society, Inc.

Electrochemical Impedance Spectroscopy Investigation of Bath Aging in Damascene Process Chemistries C. Gabrielli,a,*,z P. Mocoteguy,a H. Perrot,a,* A. Zdunek,b,* P. Bouard,c and M. Haddixd,* a UPR15 du CNRS, Physique des Liquides et Electrochimie, Universite´ Pierre et Marie Curie, 75252 Paris Cedex 05, France b Air Liquide, Chicago Research Center, Countryside, Illinois 60525, USA c Centre de Recherche Claude Delorme, 78354 Jouy en Josas, France d Dallas Research Laboratory, Dallas, Texas 75243, USA

Electrochemical impedance measurements were performed on copper bath solutions during copper deposition and subsequent aging of the bath. The results show a correlation between the low-frequency capacitive loop of the electrochemical impedance and copper deposit roughness. The deposit became smoother as the bath aged and reached an optimal value, after which the deposit became rougher. The optimal surface smoothness correlated with a minimum in the low-frequency capacitive loop diameter of the impedance. The results suggest that electrochemical impedance measurements may be used to monitor changes in the copper bath that affect copper deposit quality and superfilling capability. © 2004 The Electrochemical Society. 关DOI: 10.1149/1.1643793兴 All rights reserved. Manuscript submitted July 8, 2003; revised manuscript received August 6, 2003. Available electronically January 22, 2004.

Copper bath chemistry for electroplating copper interconnects on integrated circuits requires special additives to achieve ‘‘bottom-up’’ or ‘‘superfilling’’ of the trenches and vias. The ability to perform such superconformal deposition depends strongly on the organic additives added to the copper sulfate/sulfuric acid electroplating bath. Monitoring and controlling the concentration of these additives as well as their by-products is crucial to guaranteeing the properties of the copper interconnect deposit.1 Electrolyte formulations for superconformal deposition have been proposed in the literature,2-6 including several commercial electroplating baths for which the composition is proprietary and not known. All bath electrolytes, either commercial or in the public domain, contain copper sulfate and sulfuric acid with the addition of chloride ions. In addition, an inhibitor, often poly共ethylene glycol兲 共PEG兲,7-9 and an accelerator, generally a thiol,2,10 are also added to the bath. The relationships between deposition kinetics, deposit morphology, and the concentration and influence of the additives, however, are not fully understood. Recent studies on superconformal deposition have focused on correlations between simple current-potential curves and effective filling of the patterned substrate.2,5,6,10-12 Only a few studies have dealt with the interactions between the additives, the deposition process, the additive consumption, or bath degradation.13-15 Electrochemical impedance spectroscopy 共EIS兲9,16-20 and quartz crystal microbalance 共QCM兲8 are promising experimental tools which may help improve the understanding of the deposition process and, more particularly, provide information on the specific additive action during superconformal deposition. In this article, the influence of solution aging on the copper deposit morphology and the impedance of the copper-deposited electrode was studied for an industrial copper bath containing three additives. Furthermore, the feasibility of a new type of sensor to follow bath quality and copper deposit morphology for the damascene process by impedance measurements was also explored. Experimental Copper bath aging and impedance measurements were carried out from a commercial electrolyte solution from Enthone OMI, Inc. The bath electrolyte consisted of copper sulfate, sulfuric acid, and three additives added to the bath in two separate groups: the first one containing a proprietary suppressor and accelerator 共named additive B here兲 and the second containing a proprietary leveler 共named additive C兲. Preliminary impedance measurements were also per-

* Electrochemical Society Active Member. z

E-mail: [email protected]

formed on fresh industrial baths of varying additive concentration before any deposition was conducted. Figure 1 illustrates the bath aging and impedance measurement setup. The bath solution was aged by depositing copper at a constant current density of 15.88 mA/cm2 using two, 68 cm2 copper plates connected to a stabilized power supply. Aging time is defined as the electric charge consumed per solution volume, or ampere-hour per liter 共Ah/L兲. Five aging experiments were carried out at progressively increasing aging times, up to 12 Ah/L for the last experiment. Additive depletion compensation was not performed during bath aging. The electrochemical impedance was measured after specific amounts of electric charge were passed through the bath aging circuit. The impedance was measured by a circuit independent of the bath aging system and consisted of a galvanostat 共Sotelem PZstat兲 and a frequency-response analyzer 共Solartron 1250兲 connected to a three-electrode cell. The counter electrode was a copper anode containing phosphorus, the working electrode a 0.2 cm2 copper rotating disk electrode 共RDE兲, and the reference electrode a sulfate-saturated electrode. The measurements were carried out at a constant deposition current density of 25 mA/cm2. The frequency-response analyzer imposed a current of 0.8 mA magnitude with a frequency range of 0.001 to 100,000 Hertz. The RDE rotation speed was maintained at 2000 rpm during the experiments. Copper deposit morphology as a function of bath aging time was characterized using scanning electron microscopy 共SEM兲 for the five aging experiments. After performing an impedance measurement at a given aging time, the copper rotating disk electrode was taken out of the solution, rinsed with deionized water and analyzed by SEM for surface roughness or other features. A new copper RDE was used for each aging experiment.

Results Preliminary impedance measurements are shown in Fig. 2 for fresh industrial solutions with different concentrations of the additives. Figure 2a shows the impedance diagram for a solution with nominal additive concentrations necessary to obtain superconformal deposition. Figure 2b and c illustrated the impedance measured when the concentration of the component B, and component C are halved, respectively. The impedance scans in the three figure parts generally show three capacitive loops and one inductive loop in the lowest frequency range. The high frequency capacitive loop, labeled 共a兲 in Fig. 2a, is related to the charge-transfer resistance of copper deposition in parallel with the double-layer capacity. The second capacitive loop 共labeled 共b兲兲 is related to the chloride component of the deposition solution. The two, low frequency loops, one capaci-

C32

Electrochemical and Solid-State Letters, 7 共3兲 C31-C34 共2004兲

Figure 1. Experimental setup for bath solution aging and impedance measurements.

tive 共c兲 and one inductive 共d兲, are related to the organic additives.20 These preliminary measurements demonstrate that the impedance is sensitive to the additive concentration. Figure 3 shows the impedance and SEM micrograph results from the aging experiments using a copper bath of nominal concentration and after deposition with various quantities of electrical charge. The SEM image in Fig. 3-1 indicates that the surface of the deposit is somewhat rough after deposition from a fresh, non-aged bath solution. As aging of the bath progresses, the deposit becomes smoother and reaches an optimal level, as seen in Fig. 3-2 and 3-3. After further aging of the bath, the deposit again becomes rougher 共Fig. 3-4 to 3-5兲. Practical experience in copper interconnect deposition has shown that when a new bath solution is used, an induction time is necessary before obtaining a deposit with good superfilling properties.21 However, after some time, these properties disappear because of consumption or degradation of the additives. Observations of the impedance diagrams during aging of the solution in the observed time scale indicate that the two high frequency capacitive loops do not change with time. However, the sizes of the two low frequency loops do change during aging. To characterize this change, the diameter of the low-frequency capacitive loop is taken as a representative parameter to follow solution aging. The diameter is defined as D 3L ⫽ Max Re共 Z 兲 ⫺ Re共 Z 兲 1Hz

关1兴

where Max Re(Z) is the maximum value of the real part of the impedance from the diagram, and Re(Z)1Hz is the value of the real part of the impedance at 1 Hz. Figure 4 shows a graph of D 3L vs. aging time. The numbers 1-5 on the graph correspond to D 3L values obtained from the impedance diagrams numbered in Fig. 3. Comparison between Fig. 3 and 4 indicates a good correlation between the change in D 3L values and copper deposit morphology. A smoother deposit was obtained when D 3L was at a minimum value of about 0.4-0.6 ⍀ cm2 in Fig. 4. At higher values of D 3L , the copper deposit was much rougher in appearance. The results suggest that electrochemical impedance measurements can be used to monitor when copper bath degradation begins to affect copper deposit quality, such as the deposit roughness. Furthermore, surface roughness may be able to indicate an increase or decrease in the superfilling capability of a bath solution due to similarities between leveling theory and superconformal deposition mechanisms. It is known in the literature that surface roughness occurs in metal electrodeposition when additives are not present.22,23 Surface roughness results when more metal is deposited on surface peaks rather than in surface recesses. In terms of current distribution, the recess current, i R , is less than the peak current, i P , due to diffusion and ohmic limitations. Thus, following the change in surface roughness may indicate whether bath aging 共additive consumption or degradation兲 has occurred and is affecting the superfilling mechanism.

Figure 2. Impedance diagrams for fresh industrial solutions without charging, 共a兲 nominal concentrations of component B and C, 共b兲 nominal concentration of C with B diluted to 1:2, 共c兲 nominal concentration of B with C diluted by 1:2.

Furthermore, since the electrochemical impedance measurements correlate with surface roughness, impedance may be used to characterize copper deposit morphology and aging of the copper bath.

Conclusion The influence of copper bath aging on copper deposit morphology and electrochemical impedance of the copper-deposited electrode was studied and indicates that electrochemical impedance measurements can be used to monitor when changes in the copper bath begin to affect copper deposit roughness. The diameter of the

Electrochemical and Solid-State Letters, 7 共3兲 C31-C34 共2004兲

C33

Figure 3. SEM micrographs and impedance diagrams of copper deposition in an industrial solution at different aging times.

third capacitive loop of the impedance measurement appears to be an attractive candidate for characterizing the copper deposit morphology and aging of the copper deposition bath. This parameter could be the basis of a sensor for monitoring the solution aging due,

in particular, to additive consumption or degradation. Research is in progress to further understand the kinetics of copper deposition in the presence of additives and to explain changes due to solution aging.

C34

Electrochemical and Solid-State Letters, 7 共3兲 C31-C34 共2004兲 References

Figure 4. The diameter of the third capacitive loop, D 3L , vs. aging time for a bath aging experiment using an industrial bath solution. The numbers on the graph are related to the SEM micrographs and impedance diagrams in Fig. 3.

Acknowledgment The authors acknowledge Altis Semiconductor, Corbeil-Essones, France, for providing copper bath samples for the experiments. CNRS assisted in meeting the publication costs of this article.

1. Z.-W. Sun and G. Dixit, Solid State Technol., 44, 97 共2001兲. 2. T. P. Moffat, J. E. Bonevich, W. H. Huber, A. Stanishevsky, D. R. Kelly, G. R. Stafford, and D. Josell, J. Electrochem. Soc., 147, 4524 共2000兲. 3. J. J. Kelly and A. C. West, Electrochem. Solid-State Lett., 2, 561 共1999兲. 4. J. J. Kelly, C. Tian, and A. C. West, J. Electrochem. Soc., 146, 2540 共1999兲. 5. S. Y. Chiu, J. M. Shieh, S. C. Chang, K. C. Lin, B. T. Dai, C. F. Chen, and M. S. Feng, J. Vac. Sci. Technol. B, 18, 2835 共2000兲. 6. P. Taephaisitphongse, Y. Cao, and A. C. West, J. Electrochem. Soc., 148, C492 共2001兲. 7. D. Stoychev and C. Tsvetanov, J. Appl. Electrochem., 26, 741 共1996兲. 8. J. J. Kelly and A. C. West, J. Electrochem. Soc., 145, 3472 共1998兲. 9. J. J. Kelly and A. C. West, J. Electrochem. Soc., 145, 3477 共1998兲. 10. J. Reid, Jpn. J. Appl. Phys., Part 1, 40, 2650 共2001兲. 11. T. P. Moffat, D. Wheeler, W. H. Huber, and D. Jossell, Electrochem. Solid-State Lett., 4, C26 共2001兲. 12. K. R. Hebert, J. Electrochem. Soc., 148, C726 共2001兲. 13. E. G. Eddings and T. A. Ring, PC Fab, 35 共Dec 1990兲. 14. L. T. Koh, G. Z. You, S. Y. Lim, C. Y. Li, and P. D. Foo, Microelectron. J., 32, 973 共2001兲. 15. T. F. Moffat, B. Baker, D. Wheeler, and D. Josell, Electrochem. Solid-State Lett., 6, C59 共2003兲. 16. R. Wiart, Electrochim. Acta, 35, 1587 共1990兲. 17. J. J. Kelly, K. M. Anisur Rahman, C. J. Durning, and A. C. West, J. Electrochem. Soc., 145, 492 共1998兲. 18. S. Krzewska, Electrochim. Acta, 42, 3531 共1997兲. 19. S. Goldbach, W. Messing, T. Daenen, and F. Lapicque, Electrochim. Acta, 44, 323 共1998兲. 20. C. Gabrielli, J. Kittel, P. Mocoteguy, H. Perrot, A. Zdunek, P. Bouard, M. Haddix, L. Doyen, and M. C. Clech, in Actes du 15eme Forum Impe´dances Electrochimiques, p. 159 共2002兲. 21. M. C. Clech, Internal communication, Altis Semiconductor, Corbeil-Essones, France, Dec 2002. 22. M. Paunovic and M. Schlesinger, Fundamentals of Electrochemical Deposition, The Electrochemical Society Monograph Series, p. 180, John Wiley & Sons, New York 共1999兲. 23. J. Dini, Electrodeposition, The Material Science of Coatings and Substrates, p. 203, William Andrews Publ., Norwich, NY 共1993兲.