Atmospheric Corrosion of Silver in Outdoor Environments and Modified Accelerated Corrosion Chambers

Atmospheric corrosion of bare metals has been extensively studied for failures of structures in environments as diverse as: industrial, marine, urban, rural, or in many cases a combination of these areas.^1-7  Industrial atmospheres are more corrosive than rural atmospheres because of sulfur contaminants generated by the burning of fossil fuel, where the sulfur contaminants can form very corrosive sulfurous and sulfuric acids.^1-5  In coastal marine atmospheres, chloride deposition can occur from aerosols derived from seawater, which contains 3.5 wt% sodium chloride (NaCl).^6-7  This is a good electrolyte and can induce various atmospheric corrosion phenomena such as pitting, galvanic corrosion, exfoliation, and crevice corrosion.^6-7 

Atmospheric corrosion is primarily a result of humidity and oxygen (O₂), but is also accentuated by contaminants such as sulfur compounds and chlorides. A fundamental requirement for the atmospheric corrosion process is the presence of an electrolyte. The thin layer of the electrolyte forms on metallic surfaces under various atmospheric exposure conditions, where atmospheric corrosion then proceeds by balancing anodic and cathodic reactions. The anodic oxidation reaction involves the dissolution of the metal, while the cathodic reaction is mainly assumed to be the oxygen reduction reaction and sometimes the production of hydrogen under a coating, which leads to blistering of the coating.⁸  While the atmospheric oxygen reduction is one of the most important reactions, other oxidizing species including ozone (O₃), sulfur, and nitrogen species are also involved in the reduction reactions in the presence of air pollutants.^1-2,8 

Atmospheric corrosion, however, is related not only to the chemical and physical interactions among corrodents (sulfur dioxide [SO₂], hydrogen sulfide [H₂S], oxides of nitrogen, and chlorides), but also to weathering factors (temperature, moisture, rainfall, solar radiation, and wind velocity). The combination of these parameters makes the process of atmospheric corrosion very complex and diverse. The atmospheric corrosion of silver at outdoor exposure sites has been extensively studied as a result of monitoring the corrosion of silver from environmental corrodents and weathering factors.^9-13  Chen, et al., observed that the rapid oxidation reaction of silver (Ag) to silver oxide (Ag₂O) required both ozone and ultraviolet (UV) radiation for the photodecomposition of ozone to facilitate the formation of Ag₂O, while the corrosion reaction of Ag was barely affected by the relative humidity (RH) in the laboratory tests to mimic Ag corrosion in outdoor environments.⁹  Liang, et al., also examined the Ag corrosion behavior in laboratory environments containing ozone, UV radiation, and RH with deposited NaCl particles and proposed four reaction pathways of Ag corrosion by reactive chlorine-containing species leading to the formation of the AgCl corrosion product and/or intermediate product of Ag₂O.¹⁰  Lin, et al., identified Ag₂SO₄ and AgO in outdoor locations including AgCl, Ag₂O, and AgS using x-ray diffraction and coulometric reduction techniques.¹¹  Sanders, et al., also identified Ag₂CO₃ by x-ray photoelectron spectroscopy, in addition to differentiating between Ag₂SO₃ and Ag₂SO₄ resulting from the presence of alkali cations.¹² 

In accelerated corrosion tests, the ASTM B117 salt spray test has been commonly used to evaluate the relative resistance to corrosion of coated and uncoated metallic specimens when exposed to a salt spray climate at an elevated temperature.¹³  However, the results of this test have been criticized for the discrepancy with the results of field exposure and failure to predict actual service performance.^10-11,14-17  Therefore, the development of an accelerated test protocol including environmental factors is needed to accurately predict the performance lifetime of materials (both bare and coated metals). Liang, et al., observed that there is no AgCl film on pure Ag after 4 months of ASTM B117 testing.⁹  Wan, et al., modified a standard salt spray corrosion test chamber to introduce ozone and UV radiation with same condition of ASTM B117 protocol.¹⁴  It was shown that the corrosion rate of silver was accelerated by 20 times in modified chamber with only ozone at 23 ppm that of Daytona Beach, FL.¹⁴  Lin, et al., suggested that the NaCl deposition rate and AgCl dissolution rate are quite high in ASTM B117 chamber, with the rate of formation of AgCl relatively low. Therefore, it is hard to detect AgCl formation after ASTM B117 exposure testing.¹¹ 

In this paper, the effect of exposure to O₃, along with UV radiation, was investigated with temperature and RH as weathering factors. This work presents the results for pure Ag in up to 2 y outdoor exposure in eight different locations. The corroded surfaces of the retrieved samples were investigated by scanning electron microscopy (SEM) and the chemical compositions of the corrosion products were determined by energy dispersive spectroscopy (EDS). Exposure experiments up to 1,000 h in the both a B117 and modified B117 chamber were also performed to better understand the corrosion behavior of pure Ag samples to compare to the corrosion results of the field exposures. The chamber exposure samples were also analyzed by SEM and EDS. The environmental and climatic factors in the modified chamber include RH, temperature, UV radiation, and O₃.

The bare, pure Ag from Battelle were mounted on exposure cards and deployed to the exposure sites.¹⁷  The coupons were 1/2 in × 3 in × 1/16 in (1.27 cm × 7.62 cm × 0.16 cm) and secured to the 4 in × 5 in (10.16 cm × 12.7 cm) mounting card with nylon standoffs, screws, and nuts. Each coupon was weighed to the nearest 0.01 mg prior to mounting and after retrieval (post-analysis) from the field following a standard cleaning method.¹⁸  Coulometric reduction on the Ag coupons after field and chamber exposures was performed to measure the film thickness of the silver corrosion product.^9-11  The electrolyte used for the experiment was 0.1 M sodium sulfate (Na₂SO₄) with the pH adjusted to 10.0 with sodium hydroxide (NaOH) and deaerated for 1 h with argon (Ar). A mercury/mercury sulfate (Hg/HgSO₄) reference electrode was used to avoid chloride contamination. The Ag samples were exposed to a constant cathodic current of 0.1 mA with surface area of 1 cm², and the potential was recorded as a function of time. The amount of time to reduce the Ag at each potential is related to total charge at a constant current, which can be used to measure the film thickness of silver corrosion product by using the Faraday constant (96,485 C/mol), the equivalent weight (AgCl: 143.32 g/mol, Ag₂O: 231.74 g/mol), and density of the compound being reduced (AgCl: 5.56 g/cm³, Ag₂O: 7.14 g/cm³).^9-11  There were six land-based exposure sites in the United States: Daytona Beach (FL), Pt. Judith (RI), Kirtland Air Force Base (AFB, NM), Hickam AFB (HI), Tyndall AFB (FL), and Wright-Patterson AFB (OH). There were two University National Oceanic Laboratory (UNOLS) ship-based exposure sites: the R/V Hugh R. Sharp, based at the University of Delaware in Lewes, DE (referred to as East Coast Ship) and the R/V Thomas G. Thompson, based at the University of Washington, Seattle, WA (referred to as West Coast Ship). The bare, pure Ag coupons were attached to the exposure racks, installed at each exposure location, and retrieved every 3 months over a 2 y period for a total of eight cumulative exposure intervals and eight sets of coupons per exposure site.

Weather monitoring stations containing temperature, RH, UV, and O₃ sensors were deployed at six exposure sites: Pt. Judith (RI), Kirtland AFB (NM), Tyndall AFB (FL), Wright-Patterson AFB (OH), East Coast Ship (DE), and West Coast Ship (WA). The weather monitoring station was not deployed at Daytona Beach (FL), where Battelle already has a weather monitoring station (except for UV radiation); a weather monitoring station was also not deployed at the Hickam AFB (HI) site, as there was not a reliable power supply in close proximity to the exposure site. The weather monitoring systems consisted of a HOBO U23 Pro v2^† temperature/RH data logger (Onset Computer Corporation), a Series 130^† O₃ transmitter (Aeroqual Ltd.), and an SU-100 UV Sensor^† (Apogee Instrument, Inc.). The weather data were recorded on an hourly basis and downloaded approximately every 3 months using a HOBO U12^† 4-channel external data logger and HOBO U-DT-1^† shuttle data transporter. The chemical composition and morphology of each coupon were analyzed using EDS (Genesis 2000^†) and SEM (Zeiss EVO-50XVP^†), respectively.

In order to investigate the role of environmental and climatic factor in the field exposure sites, a standard corrosion test chamber (Q-Lab Corporation) was modified for conducting the ASTM B117 test with the introduction of both ultraviolet A (UVA, 400 nm to 320 nm) light and O₃. The UVA lamps (Q-Lab Corporation) were installed on the chamber lid and illuminated the coupons in the chamber through a quartz window. UVA light intensity was inversely proportional to the distance from the UVA lamps within the chamber. A commercial O₃ generator (Pacific O₃ Model L11^†), O₃ monitor (Teledyne Model M450^†), and microprocessor (Love Controls 2600 Series^†) were put in line with the exposure chamber to provide and control O₃ levels. The effect of UVA and O₃ on the corrosion behavior of bare Ag coupons, which were identical to the samples deployed at the eight exposure sites, was investigated. The coupons were angled at 30° from the vertical. Therefore, there is a directly exposed front surface of the coupons to UV and NaCl deposition, but the reverse surface of the coupon could have provided a sheltering effect on Ag corrosion. For this paper, the concentrations of O₃ in the modified chamber were either 100 ppb or 800 ppb, while the UVA intensities were either 5.44 W/m² or 46.8 W/m².

The weather data monitored at the field exposure sites were: UV, O₃, temperature, and RH. O₃ levels measured at the land-based exposure sites demonstrated good agreement with the local Environmental Protection Agency (EPA) monitoring sites. Figure 1 shows the daily and yearly O₃ level between the monitoring stations at Kirtland AFB compared to the local EPA site, located 12 miles away. As can be seen in Figure 1(a), the daily pattern in winter was very similar, between 0 ppb and 40 ppb from December 1-7, 2009. The yearly pattern in Figure 1(b) shows a more visible seasonal variation over the 2 y monitoring. During the summer, the O₃ levels were observed to reach a maximum of 80 ppb.

Figure 2 shows a relationship between the following parameters: temperature, O₃, UV, and RH at hourly intervals from June 1-7, 2010 at Kirtland AFB. Despite the variability of the weather conditions, typical daily patterns were observed. The measured O₃ concentration, UV, and temperature values are correlated positively with each other, while %RH has a negative relationship with O₃ concentration, UV, and temperature.

Figure 3 shows the distribution of the cumulative frequency of O₃ concentration, UV, temperature, and RH at all of the monitored exposure locations. The boundary of the box closest to zero indicates the 25th percentile. The red line and black line within the box mark the median and 50th percentile, respectively. The boundary of the box farthest from zero indicates the 75th percentile. Whiskers above and below the box indicate the 90th and 10th percentile. The dots at the bottom and top represent the 5th and 95th percentile, respectively. O₃ concentration, temperature, and RH at Daytona Beach were taken from Battelle’s weather monitoring system. A weather monitoring station was not deployed at Hickam AFB because of a limitation of access to electrical power. The East Coast Ship and West Coast Ship weather data were excluded from comparison to the local EPA monitoring sites because of their off-shore cruising areas, even though there was a good agreement between coastal based EPA sites and ship locations when the ships were near the shore. O₃ levels ranged from 46.3 ppb to 66.5 ppb for 95% of the cumulative frequency from the West Coast Ship to the East Coast Ship; for UV levels, there was a range of 29.6 W/m² to 51.9 W/m² from the West Coast Ship to the Pt. Judith site for the 95% frequency. In terms of temperature, there was a fairly close range of 75.6°F to 92.6°F for the 95% cumulative frequency, and in regards to RH, for 95% of the cumulative frequency, all of the exposure sites exhibited 100% RH except for the Kirtland AFB exposure site, which was lower (approximately 81.5% RH).

The bare coupons of pure Ag were evaluated by coulometric reduction measurement after exposure to measure the film thickness of corrosion product. Figure 4 shows the coulometric reduction curves for Ag after exposure at Daytona Beach and Kirtland AFB for 9 months. AgCl is the most common corrosion product from all outdoor exposure sites except Kirtland AFB. The main corrosion products of Ag were found to be Ag₂O and AgO, along with AgCl from the Kirtland AFB site.^9-11,14  The film thickness of Ag corrosion product could be indicative of the cumulative chloride, oxide, and sulfate deposition levels at the exposure sites.¹¹  Ag corrosion product film thickness data for the pure Ag coupons at all exposure sites are given in Table 1. Figure 5 shows the thickness of corrosion product on pure Ag coupons at all locations during the 2 y exposure. The thickness is the average value, including both sides of the Ag sample. As can be seen, Ag coupons at Daytona Beach, East Coast Ship, and West Coast Ship sites showed high film thicknesses, while those at Kirtland AFB site showed the lowest film thickness.

Figure 6 shows the EDS and SEM image of the pure Ag coupons exposed at the Daytona Beach site over 24 months. Before exposure, only pure Ag was detected on the surface by coulometric reduction. After exposure, the Ag sample contained Cl and O elements at 3 month exposure and later O, Na, Mg, Al, Si, and Cl were detected (Table 2). As can be seen in Figure 6, the morphology of the corrosion products were not uniform across the surface. The particles were found to contain Ag and Cl, as well as a small amount of Na, as detected by EDS.

Similarly, Figure 7 shows the EDS and SEM image of a pure Ag coupon exposed at Kirtland AFB. Relatively small amounts of Cl were detected on the Kirtland AFB samples compared to Daytona Beach (Table 3). The SEM image in Figure 7 shows the scratch marks on the surface after exposure of 21 months, which indicates that the corrosion product had not fully covered the Ag surface.

Figure 8 shows the Cl/Ag ratios as a function of exposure time on the Ag coupons at eight different exposure locations. As can be seen for the Cl/Ag ratios on the samples, Daytona Beach, the West Coast Ship, and the East Coast Ship had high Cl/Ag ratios, while Wright-Patterson AFB and Kirtland AFB had low Cl/Ag ratios. The Cl/Ag ratios of Ag at Pt. Judith, Hickam AFB, and Tyndall AFB were intermediate between the other two groups. This trend of the Cl/Ag ratio on the Ag surface is similar to that observed for film thicknesses calculated from the coulometric reduction of Ag surface films as a function of exposure site location, as presented in Figure 5.

Along with continuous 5 wt% NaCl spray under B117 conditions, the modified chamber also included UVA light and O₃. The amounts of UVA irradiance and O₃ are shown in Table 4. Figure 9 shows the coulometric reduction curves for Ag after exposure with different UV and O₃ conditions for 500 h and B117 for 400 h in chamber. AgCl is the most common corrosion product of Ag from chamber exposure conditions except for B117, where no AgCl is formed. Under high UV and high O₃, the exposed side of Ag sample (i.e., the side facing the UV irradiation) exhibited a dramatic drop of reduction curve compared to the back side of the exposed sample, which was not facing the UV irradiation and is similar to the reduction curve of B117. This behavior happened for all other conditions, which indicates that the exposed side of the Ag sample experienced the continuous wet deposition of NaCl and AgCl dissolution,^9-11  which is also similar to the curve for the Ag coupon exposed at Daytona Beach for 9 months (Figure 4).

Figure 10 shows the average film thickness of pure Ag in the various chamber conditions including B117. Without UV radiation and O₃, the thickness of film was very low, while it significantly increased with UV and O₃. The effect of 5.44 V/m² to ~46.8 V/m² UV and 100 ppb to ~800 ppb O₃ on the thickness of film was complicated, which might be related to the experimental range of UV and O₃ in the chamber compared to the field conditions, shown in Figure 3. The concentrations of O₃ in the chamber are much higher than those in the field, while the level of UV irradiation in the chamber is similar to that in the field. However, the thickness of film in the high UV and high O₃ were higher than low UV and low O₃. Figure 11 shows the comparison between outdoor and chamber exposures for average film thickness on pure Ag. It is clear that film thickness under the modified chamber is much higher than those measured for the outdoor exposure locations, as well as the standard B117 exposures, which indicates that the corrosion rates of Ag in the chamber were accelerated in comparison to that of outdoor exposures, and also there is a significant effect of O₃ for the atmospheric corrosion of Ag.

The effects of environmental factors on the atmospheric corrosion of Ag during a 2 y outdoor exposure were analyzed through the use of coulometric reduction of film on Ag samples and SEM-EDS. Differences of film thickness data between Daytona Beach and Kirtland AFB sites were observed. The morphologies of corrosion products were not uniform across the surface of the Ag samples. The Ag exposed at Kirtland AFB showed relatively small amounts of Cl and O compared to Daytona Beach, and showed scratch marks on the surface after exposure periods up to 21 months.

An accelerated test protocol including UV radiation and O₃ was developed to more accurately reproduce the field exposure results. The corrosion rates of Ag in the accelerated chamber with the addition of UV and O₃ were much higher than those of the field exposure and the standard ASTM B117 salt spray test. The combination of O₃ exposure (100 ppb and 800 ppb) and UVA irradiation (5.44 W/m² and 46.8 W/m²) affected the corrosion rate of Ag in the following order: high UV and high O₃ > high UV and low O₃ > low UV and high O₃ > low UV and low O₃.

This work was sponsored by the SERDP-ESTCP Program under contract FA8601-06-D-0013. Special thanks to Mr. James Postel (Univ. Washington, R/V Thomas G. Thompson), Capt. William Byam (Univ. of Delaware, R/V Hugh R. Sharp), Lt. Col. P. Legendre (Kirtland AFB, NM), Mr. John Puu (Hickam AFB, HI), Mr. P. Mitchell (Tyndall AFB, FL), and Dr. J. Moran (ALCOA) for assistance in the field exposure placement and retrieval of the sample coupons and weather monitoring systems.