Web http://www.jcse.org/

Volume 1 Paper 14


Tarnish Process of Silver in 100ppb H2S Containing Environments

H. Kim* and J.H. Payer+
* Technical Research Lab. Pohang Iron & Steel Co., P. O. Box 36, pohang 790-785, Korea
+Dep. of Materials Sci. & Eng., Case Western Reserve Uni., 10900 Euclid Ave. Cleveland, OH 44106

Abstract

The effects of sub-ppm levels of H2S, oxygen and the adsorbed water on the atmospheric corrosion of silver were studied. The concentration of H2S and the relative humidity affecting the tarnish rate were carefully controlled by permeation tube and monitoring the humidity. The quartz crystal microbalance (QCM), an in situ mass measuring technique, was used in order to measure the tarnish rate of silver in various exposure conditions and varying amounts of adsorbed water depending on relative humidity. The tarnish films were analyzed by XPS, the analyzed data were compared to the thermodynamic calculation used to predict the stable phases and the equilibrium potentials such as f Ag+/Ag, f O/O2- and pH. QCM results showed that oxygen was required for the sulfidation of silver exposed to 100ppb H2S environments. This agrees with the predicted results that silver was not oxidized by hydrogen ion but by dissolved oxygen in the condensed water. In addition, the water adsorbed from the highly humidified environments enhanced the tarnish rate. This could be explained by the fact that the adsorbed water provided a medium for the oxygen, which oxidizes silver metal.

Introduction

Silver has been used for electronic connector materials, solder, batteries and decorating materials because of its high electrical and thermal conductivities. Hence, the atmospheric corrosion of silver has been studied with an emphasis on the effect of humidity, the effect of gas and the kinetic mechanism.

Fiaud and Guinement1 conducted indoor laboratory studies on the corrosion of silver in atmospheric environments of controlled relative humidity and pollutant gases, reporting that the corrosion product was silver sulfide and that the kinetic behavior conformed to a parabolic law. This does not agree with D. Simon’s result2 that the kinetic behavior conformed to a linear law. Kuhn and Kelsall3 calculated the stable phase in Ag - H2S - O2 and Ag - H2S - O2 - H2O with their thermodynamic data, indicating that silver sulfate was stable instead of silver sulfide. This difference between Kuhn & Kelsall’s result and Fiaud & Guinement’s experimental result was explained by a large barrier for the formation of sulfate from sulfide.

In addition, the reaction process is controversial. Abbott4 proposed that the sulfidation mechanism of silver exposed to hydrogen sulfide with ambient air was as follows:

According to Graedel5, the sulfidation mechanisms were proposed to depend on the humidity. At low relative humidity, H2S gas reacted with silver metal directly, but at high relative humidity, dissolved species (HS-, S2-) from H2S gas reacted with silver (i.e., Ag + H2S Ag2S + H2(g)). Volpe and Peterson15 exposed silver to dry air containing H2S (100ppb~ 400ppb) in flowing gas and proposed the following reaction:

In addition, the dominant species for ionic conduction across the film is the silver ion since the diffusivity of the silver ions in silver sulfide is higher than that of the sulfide ions.13, 14

However, according to Phipps and Rice6, at low relative humidity (R.H. < 40%, the adlayer thickness < 1 monolayer), it is difficult to hydrate ions, but at high relative humidity (R.H.< 65%, the adlayer thickness < 3 monolayers), the behavior of the adsorbed water layer approaches that of a bulk water layer and electrochemical processes are possible.

In this study, to get a better understanding of the sulfidation process of silver in H2S environments, the effect of relative humidity, oxygen or H2S was studied with a quartz crystal microbalance (QCM). In addition, electrolyte thermodynamics were applied by assuming that the bulk thermodynamical properties are the same as the surface in order to explain the tarnishing process.

Experimental

Silver was deposited on the front surface of a crystal by evaporation at Midwest Research Technologies, Inc. to a thickness of 5000 . Before the deposition, the back surface was masked with crystal bond. Gold, previously deposited on the front side was stripped by aquaregia solution, cleaned with distilled water and acetone alternatively three times and dried with N2 gas. The silver coated crystal was kept in desiccant before and after exposure. Before the exposure, the surface of silver was analyzed with XPS. Adsorbed species, such as adventitious hydrocarbon, (OH)ad and adsorbed oxygen were detected, but the oxide peak was not detected.

The sub-ppm level of concentration was obtained using a permeation tube. In order to get reproducible results, the concentrations of H2S, the relative humidity and the temperature in the chamber were monitored during the exposure. The temperature of the chamber was kept at 25 0.1 C and the flow rate was 1 l /min. Purified air supplied by an air supply, or nitrogen (with 99.99% purity) from nitrogen cylinder was used as a carrier gas. The concentration of H2S was maintained at 5% during the exposure, while the concentrations of undesired gases in the purified air were less than 1ppb for SO2, 0.1ppb for H2S, 1ppb for NO, 1ppb for NO2, 1ppb for O2, 1.8 103ppb for CO2 and 100ppb for CO. The details of the mixed flowing gas system were described in other papers18.

The quartz crystal microbalance, an in situ mass measurement technique with high mass sensitivity (@ 10ng/cm2), was employed to monitor the mass change during exposure to part per billion pollutants and during the transition between wet and dry environments. The block diagram of the circuitry elements and the description of the particular circuitry employed are given elsewhere17. The circuitry designed for the oscillation of a 5 MHz quartz crystal was modified for the oscillation of a 6 MHz quartz crystal. In addition, the newly designed QCM was calibrated by the potentiostatic method with the following experimental conditions7: 0.2M HClO4 + 10-3 M AgNO3 at 0.0V (SHE). The result (12.21ng/Hz-cm2) agreed well with the theoretically calculated value (12.26ng/Hz-cm2). The holder was made of polychlorotrifluoroethylene to minimize the mass change due to the chemical reaction between the holder, silver and gases. Viton O-rings were placed on each side of the crystal to ensure proper sealing and to minimize the residual stress since the frequency can be affected by pressure as well as mass change.

The tarnish film after the exposure was analyzed on a PHI 5400 XPS system (Perkin Elmer Corp.) equipped with a water cooled Mg-anode (Mg-Ka : 1253.6eV and FWHM (full width at half maximum): 0.7eV) at total power dissipation of 400 watts (15kV, 26.7mA). The sample stage was positioned such that the photoelectron emitted at an angle of 45 degrees to the plane of silver surface entered the analyzer through a 1 mm 4 mm aperture. The spectrometer was calibrated with two reference points (Ag-3d5/2: 368.3eV and C-1s: 284.5eV) to minimize the error due to the drift of the spectrometer with assumption that the shift of energy due to charging effects was negligible. In order to analyze the inside of the tarnish film, the differentially pumped argon ion gun (4kV, 2mA ions) of the XPS system was utilized. The calibration of the removal rate was accomplished by sputtering through the thickness of a Ta2O5 standard (1.8nm/min).

The surface morphology after exposure was observed by Field Emission Scanning Electron Microscopy (Model S-4500, Hitachi., Ltd.) with an accelerating voltage of 5kV to provide information on the tarnishing process. The sample stage was positioned such that the incident electron collided with the silver surface at an angle of 40 degrees to the plane of the surface to maximize the spatial resolution. The morphology could be observed with a maximum magnification of 100kX because of utilization of field emission as the electron source.

Experimental Results

Analysis of tarnish film with XPS

Figure 1 shows the XPS spectrum of the tarnish film formed on silver exposed to 100ppb H2S with air containing 75% R.H. for 72 hours. The peaks of sulfur-2p were detected at 161.8 and 161.0eV from the film surface indicating sulfide. No peak indicating sulfite or sulfate was detected before or after sputtering. The Auger parameter (724.8eV) calculated from the silver-3d5/2 peak (B.E. = 368.2eV) and Ag-M4VV (K.E. = 357.7eV) detected correspond to that of silver sulfide. The Auger parameter is shifted to that for silver metal as the intensity of sulfur-2p peak reduced with sputtering. The broad peak of oxygen-1s on the surface disappeared after sputtering. This peak is thought to result from (OH)ad, water and adsorbed oxygen. No oxygen-1s peak indicating the presence of oxide was detected before or after sputtering. Adventitious carbon also appeared on the surface and disappeared after sputtering. The XPS spectra of the tarnish film formed on silver exposed to 100ppb H2S with air containing 15% R.H. for 90 hours were similar to the above except for the sulfur-2p shown in Figure 2. The silver peaks and sulfur peaks identify the tarnish film as silver sulfide and the adsorbed species on the surface were also the same as the above. A small amount of sulfite (or sulfate) was detected on the surface. However, it was not detected after the exposure to 100ppb H2S with air containing 15% R.H. for 40 hours. Hence, the sulfite (or sulfate) may be produced after longer exposures than 40 hours. The XPS spectra of the tarnish film formed on the silver exposed to 100ppb H2S with nitrogen gas containing 75% R.H. for 72 hours were similar to the above except for the silver peak shown in Figure 3. The Auger parameter of silver did not match that of silver or silver sulfide and was positioned in the middle of these Auger parameters. Hence, the mismatch between the Auger parameter obtained and that of silver sulfide is due to the fact that the film is thinner than the escape depth of electrons having the kinetic energy of approximately 360eV ( 2nm)16.

Hence, the tarnish film after exposure to environments containing H2S consisted of only silver sulfide covered by adventitious carbon, adsorbed water and possibly adsorbed oxygen and (OH)ad.

figure1a.gif (10506 bytes)figure1b.gif (12649 bytes)

FIG. 1 XPS spectra from silver exposed 100ppb H2S and 75% R.H. air for 72 hours ( dot line: surface, solid line: after sputtering as deep as 3.8nm): (a) Ag-3d, (b) Ag-M4VV, (c) O-1s, (d) C-1s, (e) S-2p

figure2.gif (4804 bytes)

FIG. 2 Sulfur-2p spectrum from the surface of silver exposed 100ppb H2S and 15% R.H. air for 90 hours

figure3.gif (13687 bytes)

FIG. 3 XPS spectra from silver exposed 100ppb H2S and 75% R.H. N2 for 72 hours (dot line: the surface, solid line: after sputtering as deep as 0.6nm): (a) Ag-3d5/2, (b) Ag-M4VV, (c) S-2p

Measurement of mass gain with QCM

Figure 4 shows the mass gain with time when silver coated crystals were exposed to environments containing hydrogen sulfide. The mass loss occurred after the exposure condition changed from 100ppb H2S with 75% R.H. air (or N2) to 15% R.H. air, indicating water desorption (120ng/cm2) which was 1.7 times greater than that of adsorbed water (70ng/cm2) on as-received silver metal. The reason that water was more adsorbed may be due to the difference of hydrophilicity19 between silver metal surface and the surface covered by H2S which is soluble in water. The mass gain after exposure of 100ppb H2S with 75% R.H. air for 72 hours was two times higher than after 100ppb H2S with 15% R.H. air, indicating that adsorbed water enhanced the reaction rate. In addition, the mass gain in 100ppb H2S with 75% R.H. N2 was negligible, indicating that the presence of oxygen is required for the sulfidation of silver. It agreed with the XPS result that the observed Auger parameter of silver was between the Auger parameter of elemental silver and that of silver sulfide. The curve - (a) and (b) in Fig. 4 were reploted versus time1/2 scale instead of time scale. The mass gain in 100ppb H2S with 75% R.H. air increased linearly until 8 hours after the exposure, but its rate was reduced after this period, conforming to parabolic behavior. However, in 100ppb H2S with 15% R.H. air, the mass gain increased only parabolically without an initial linear region. In 100ppb H2S with air containing 75% R.H., the kinetic behavior is due to the change of the rate determining step from surface reaction or mass transport in the gas phase or the mixed step to mass transport through the tarnish film. The reaction mechanism or transport mechanism through the tarnish film will be discussed in the next section.

figure4.gif (6272 bytes)

FIG. 4 Mass gain vs. time when silver coated crystals were exposed to environments containing 100ppb H2S: (a) 75% R.H. air, (b) 15% R.H. air, (c) 75% R.H. N2

figure5.gif (5476 bytes)

FIG. 5 Mass gain vs. time1/2 when silver coated crystals were exposed to environments containing 100ppb H2S: (a) 75% R.H. air, (b) 15% R.H. air

Observation of morphology with FESEM

The surface morphologies of the tarnish films after exposure to various environments containing 100ppb H2S were the same as the surface morphology of the silver coated crystal at 100k magnification. Hence, the tarnish film is thought to form on the silver surface uniformly.

Phase Prediction

Fig. 6 shows the E-pH diagrams in Ag-H2O-100ppb H2S system obtained from the method suggested by Zappire20,21. In solid phase, silver metal, silver sulfide and silver sulfate were predicted as stable species. In liquid phase, silver ion and sulfite were predicted the stable species. In order to explain reaction mechanism, the concentrations of the silver ion and hydrogen ion were calculated using the following charge balance equation.

(1)

According to the XPS results that only silver sulfide was detected in tarnish film after the exposure of 75% R.H. environments, the concentration of silver ion was expressed using solubility product of silver sulfide. Each term was expressed as the concentration of hydrogen ion using the equilibrium constants listed in table 1,

(2)

,where PH2S and PCO2 is 100ppb and 1.8ppm, respectively.

From equation (2), the concentration of silver ion (1.9 10-16) was calculated and the equilibrium potential between Ag and Ag+ was mapped in Fig. 6. The potential where hydrogen gas is evolved is lower than the equilibrium potential between silver and silver ion, indicating that silver was not oxidized by hydrogen evolution. This calculated results agreed with the QCM results that silver was not oxidized without oxygen.

Discussion

In 100ppb H2S, silver was oxidized by oxygen, as was proved by the in situ measurements of mass showing that the tarnish rate of silver was negligible in the environments without oxygen. This corresponded to the results, obtained from thermodynamic calculation, that the equilibrium potential between Ag and Ag+ being higher than that for hydrogen gas evolution as shown in Fig. 6-(b). In addition to oxygen, the amount of adsorbed water affected the tarnish rate. This effect was explained by the change of ionic path on surface8 or the change of activity of other gas9. It could also be explained by the change of properties of adsorbed water with humidity6, 10. The formation of a few monolayers makes the behavior of adsorbed water be that of the bulk water. In the humid environments used in this tests (75% R.H.), the water was adsorbed as much as 2.5 monolayers. Furthermore, the presence of H2S increased the amount of water from 2.5 monolayers to 5.3 monolayers. Hence, the tarnish process in humid environments is supposed to be regarded as the reaction in aqueous solution. The following reactions in humid environments are suggested:

Anodic reaction: f Ag/Ag+ = - 0.443V (3)

Cathodic reaction: f O/O-2 = = 0.804V (4)

Overall reaction: (5)

where f Ag/Ag+ is calculated from [Ag+] 10-16 and the activity of water = 1.

In addition, the detection of linear region in humid environment in Fig. 5 could be also explained by the high oxidizing power in humid environment and the high electrical conductivity of silver sulfide film region (i.e., kAg2O 10-8W -1-cm-1, k Ag2S 10-8 W -1-cm-1)11, 12. In 100ppb H2S with air containing 15% R.H., the amount of adsorbed water was 14ng/cm2 ( 0.5 monolayer), whose physical properties are different from that of bulk water. Unlike the humid environment, the tarnish process in dry environments is supposed to be regarded as the reaction between metal and gases. Compared to dry environment, the dissolved oxygen in humid environments increased the oxidizing rate of silver, and sequentially enhanced tarnish rate. The presence of sulfur with high chemical state detected on tarnish film after longer than 40 hour exposure in low humidified (15% R.H.) air containing 100ppb H2S could be explained by low diffusion rate of silver. Low oxidizing power in dry environment and the growth of the film reduce the flux of silver, making sulfide have more chance to react with oxygen, to form stable phase such as sulfur with high chemical state.

In addition, the detection of linear region in humid environment in Fig. 5 could be also explained by the high oxidizing power in humid environment and the high electrical conductivity of silver sulfide film region (i.e., k Ag2O 10-8 W -1-cm-1, k Ag2S 10-8 W -1-cm-1)11, 12. Low electrical conductivity of tarnish film could not provide the detection of linear though the oxidizing power was enhanced.

figure6.gif (7572 bytes)

FIG. 6 The prediction of solid stable phase in Ag-100ppb H2S-H2O: (a) Liquid phase, (b) Solid phase, (dot line: total activity of dissolved silver (-log[Ag+]T), line a: f H+/H, line b: f O/O-2)

Conclusions

The aim of this investigation was the study of the effects of relative humidity and oxygen on the tarnish process of silver in various environments containing 100ppb H2S. The mass change was monitored during exposure and the tarnish film was analyzed by XPS. From the experimental results, the conclusions are as follows:

Tarnish film formed consisted of only silver sulfide covered by adventitious carbon, adsorbed water and possibly adsorbed oxygen and (OH)ad.

Silver was not tarnished without oxygen, which proved by QCM results and thermodynamic calculation.

The amount of adsorbed water in 75% R.H. environment was as much as 2.5 monolayers and the presence of H2S increased the amount of adsorbed water to 5.3 monolayers. It is enough to act as bulk water. Hence, the high tarnish rate in humid environment is due to the presence of bulk water layer in which oxygen is dissolved to oxidize silver.

In humid environment, the kinetic behavior was due to the high electrical conductivity of the tarnish film and the high oxidizing power due to adsorbed water.

References

1.    J. Guinement and C. Fiaud, in 13th Inter. Conf. on Electric Contacts, p. 383, Lausanne (1986).

2.    D. Simon, C. Perrin, D. Mollimard, M.T. Bajard, and J. Bardolle, in 13th Inter. Conf. on Electric Contacts, p. 333, Lausanne (1986).

3.    T. Kuhn and G.H. Kelsall, in Corrosion of Electrical Contacts, p. 51, Institute of Metals, Great Britain (1989).

4.    W.H. Abbott, in Proc. 4th Int. Res. Symp. Electrical Contact Phemonena, p. 35, Swansea (1968).

5.    T.E. Graedel, J.P. Franey, G.J. Gualtieri, G.W. Kammlott and D.L. Malm, Corr. Sci., 25, 1163 (1985).

6.    P. B. P. Phipps and D. W. Rice, in Corrosion Chemistry, p. 235, G.R. Brubaker and P.B.P. Phipps, Editors, Washington (1979).

7.    S. Bruckenstein and S. Swathirajann, Elecrochemica Acta, 30, 851 (1985).

8.    J-E Svensson and L-G Johansson, J. Electrochem. Sci., 143, 51 (1996).

9.    P. Ericksson and L-G Johansson, J. Electrochem. Sci., 138, 1227 (1991).

10.    E. McCafferty and A.C. Zettlemoyer, Disc. Faraday Soc., 52, 239 (1971).

11.    T. Nomura, T. Nagamunne, K. Izutsu and T.S. West, Bunseki KagaKu, 30, 494 (1981).

12.    Yu.G. Vlasov and Yu. E. Ermolenko, Elektrokhemiya, 17, 1301 (1981).

13.    D. Grientsching and W. Sitte, J. Phys. Chem. Solids. 52, 805 (1991).

14.    R.L. Allen and W.J. Moore, J. Chem. Phys., 63, 223 (1985).

15.    L. Volpe and P.J. Peterson, Corr. Sci., 29, 1179 (1989).

16.    M.P. Seah and W.A. Debch, Surf. and Interface Anal., 1, 2 (1979).

17.    D.A. Buttery, in Electroanalytical Chemistry vol. 17, A.J. Bard, Editor, p1, Marcel Dekker Inc., New York (1989).

18.    D.W. Rice, E.B. Rigby, P.B.P. Phipps, R.J. Cappell and R. Tremoureux, J. Electrochem. Soc., 128, 275 (1989).

19.    R. Suhrmann, J.M. Heras, L.V. de Heras, G. Welder, Ber. Bunsenges, Phisik. Chem. 72, 854 (1968).

20.    J.C. Angus, B. Lu and M.J. Zappia, J. Appl. Electrochem., 17, 1 (1987).

21.    J.C. Angus and M.J. Zappia, J. Electrochem. Soc., 134, 1374 (1987).

Table 1 Equilibrium Constants

Equilibrium constants

Value

Chemical equations

Henry’s constants

KH for CO2

3.4 10-2

[CO2]/PCO2

 

KH for H2S

10-0.99

[H2S]/PH2S

Dissociation constants

Kd1 for CO2

10-6.35

[H+] [HCO3-]/[CO2]

 

Kd2 for CO2

10-10.33

[H+] [CO3-2]/[HCO3-]

 

Kd1 for H2S

10-7

[H+] [HS-]/[H2S]

 

Kd2 for H2S

10-14

[H+] [S-2]/[HS-]

 

Kw for H2O

10-14

[H+] [OH-]

Solubility product

Ks.p. for A2S

10-50

[Ag+]2 [S-2]