Determine the mechanism for copper plating and methods for Its Elimination from HVAC Systems.
COPPER PLATING LITERATURE SEARCH
The literature search, summarized in the project final report (Kauffman 2007), included the Materials Compatibility & Lubricants Research Program final report (Calm 2000). The literature search focused on papers and reports describing the occurrences of copper surface corrosion or copper plating on steel coupons during sealed tube tests or on steel compressor parts during accelerated laboratory testing and commercial use. In many of the papers and reports, the copper plating was a secondary observation.
As the starting point for developing the copper plating mechanism, the basic copper plating process (Lilje 2000 and ASHRAE 2002) occurs in the following four separate steps:
1. Corrosion of copper containing surface
2. Dissolution/suspension of copper species by circulating refrigerant/oil
3. Transportation of copper species to steel surface
4. Plating of copper metal onto steel surface
Inhibiting one or more steps of the above mechanism will inhibit copper plating. Refrigeration systems using sulfur dioxide did not suffer from copper plating even though the removed mineral oils were dark green in color due to dissolved copper (McGovern 1939). So in sulfur dioxide systems, steps 1-3 of the above mechanism occurred but copper plating did not follow due to the inhibition of step 4. Consequently, any additive or process capable of inhibiting one or more of the individual steps listed above would be capable of inhibiting the copper plating on steel surfaces. Conversely, contaminants that promote one or more of the individual steps would be capable of accelerating the copper plating of steel surfaces.
Corrosion of Copper Containing Surface
The first step in the copper plating mechanism listed above is the corrosion of a copper metal or copper alloy surface. Based on the position of copper with respect to hydrogen on the electrochemical series of metals, copper is a noble metal and is not reactive with hydrochloric (HCl) acid produced by the thermal degradation of the CFC and HCFC refrigerant/lubricating oil systems (Spauchus 1963 and Steinle 1964). In the absence of oxygen, carboxylic acids present due to hydrolysis of the polyolester (POE) oils or oxidation of any type oil will not react with copper even in the presence of water (Field and Henderson 1998).
In contrast to acids, lubricating oils may contain contaminants or other type degradation products capable of copper corrosion. Copper corrosion occurs in the presence of lubricating oils containing hydroperoxides (oxidizing agent) (Scott 1965) or "oleoresins" (Steinle 1964) removed by current refinement processes. In addition to naturally occurring base-stock contaminants, manufacturing fluid contamination of refrigeration systems has been shown to cause copper corrosion (Cavestri and Schooley 1996 and Rohatgi 2007).
High quality lubricating oils and processes to ensure system cleanliness could be used to totally inhibit the direct chemical corrosion of the copper metal surfaces in refrigeration systems. However, the presence of air in the refrigeration system originating during manufacture, storage, operation and/or repair is one contaminant that cannot be totally excluded. Even very low (ppm) levels of air in flowing fluorinated liquids resulted in the formation of copper (I) oxides on the surfaces of copper tubing (Lee 1970). The thickness of the copper (I) oxide layer increased with air concentration until it reached a maximum (approximately 1 micron) at which point the copper (I) oxide layer protected the underlying copper metal from further oxidation.
In contrast to copper metal, copper (I) oxide as well as copper (II) oxide will react with both carboxylic and hydrochloric acids. Formicary (also called "ant nest") corrosion of copper tubing is well known and occurs when a corrosive agent such as carboxylic acid (manufacturing residue) reacts with the copper (I) oxide according to the reaction mechanism (Corbett 2000):
Cu(I) Oxide + Carboxylic Acid [right arrow] Cu(I) Carboxylate [unstable] + Water
Cu(I) Carboxylate + Oxygen [right arrow] Cu(II) Carboxylate + Cu(I) Oxide
Cu(II) Carboxylate + Cu(0) metal [right arrow] 2 Cu(I) Carboxylate
Once initiated, the formicary corrosion will continue as long as oxygen is available to produce the Cu (II) carboxylate. In the presence of water, galvanic corrosion between the pits in the copper surface and the copper (I) oxide increases the rate of the overall corrosion process.
Dissolution of Corroded Copper Species
Once the copper surface has been corroded, the dissolution of the copper species will rely on both the solvency of the refrigerant/mineral oil system as well as the chemical composition of the copper species.
Chlorinated Copper Compounds. In the case of CFC/mineral oil systems, rolling sealed tube tests with R-12 and oils with different "oleoresin" contents showed that dissolved copper increased with "oleoresin" concentration and stressing temperature (Steinle and Bosch 1964). The increasing amount of dissolved copper was accompanied by increased iron and chloride concentrations. It has also been reported (Spauchus 1963) that copper corrosion did not result from direct attack by hydrochloric (HCl) acid but rather by oil previously chlorinated by reaction with HCl acid.
Copper dissolution occurred in sealed tubes containing only heated lubricant and the results indicated that the same compounds present in oils that produced sludge and carboxylic acids during oxidation tests were also responsible for copper dissolution by the oil (McGovern 1939).
Non-Chlorinated Copper Species. Carboxylic acids are produced by the oxidation of all lubricating oils and by the hydrolysis of POE oils during storage and operation in refrigerant systems. The dissolution of the copper species will be totally dependent upon the capability of the refrigerant/oil system to dissolve the copper carboxylate produced by the carboxylic acid corrosion of the copper oxides. The long chain carboxylates produced by the oxidation of oils will be highly oil soluble while the shorter chain carboxylates produced by POE hydrolysis will have reduced oil solubility (oil solubility increases with the length of the carboxylic acid).
Although the majority of the copper carboxylates produced by the lubricating oil corrosion of copper oxide surfaces will be oil soluble, the solubility of the copper carboxylates will be reduced by the presence of the refrigerant, especially HFC refrigerants. Reddish to reddish brown precipitates formed on the copper coupon surface when water or hexanoic acid was added to sealed tube tests of HFC-134a/POE oil although the dissolved copper concentration never exceeded 3ppm (Field and Henderson 1998). Also, research on the thermal reactions between iron carboxylate and POE basestocks determined that increasing the HFC:lubricant ratio greatly decreased the solubilities of different carboxylates (Kauffman 2003).
Transportation of Copper Species
Insoluble copper containing particles or soluble copper compounds need to be transported to the compressor for copper plating to occur on the steel surfaces. Very little information on copper transportation was identified by the literature search.
With regard to CFC/mineral oil systems, sealed tube tests indicated that the copper corrosion compounds are soluble making the transportation of the copper species to the steel surfaces continuous as the refrigerant circulates throughout the system. However, the copper corrosion compounds in HFC/POE systems will precipitate due to their low solubility which was observed (DeVos 1997) during the development of an accelerated compressor test to evaluate the capillary plugging tendencies of different contaminated R-134a/POE systems. The deposits that formed at the inlet of the capillary tube of the R-134a/POE systems (least soluble) contained copper and iron carboxylates (carboxylic acids from brazing flux, POE hydrolysis, ester manufacturing oils, etc.) while mineral oil, silicone oil and other light oils deposited toward the exit of the tube (more soluble). Capillary plugging of refrigeration systems is reported to be worse for HFC/POE systems, due to the low solubility of the metal carboxylates and various manufacturing liquids, than for CFC/mineral oil systems (Cartlidge and Schellhase 2003).
Copper Plating Mechanisms
Once the copper containing compounds reach the steel surfaces, two basic mechanisms can occur to produce metallic copper plating on the steel surfaces: 1) thermal decomposition of dissolved or suspended copper compounds or 2) electrochemical reactions between the iron containing surface (are oxidized then dissolved) and soluble copper compounds (are reduced then plate iron surface).
Thermal Decomposition of Copper Compounds. Commercial inks based on copper carboxylates undergo thermal decomposition above 120[degrees]C (248[degrees]F) in the absence of air to produce metallic copper. The copper films from thermal decomposition would form on the hottest surfaces of the compressor, regardless of surface metallurgy.
Electrochemical Reaction at Steel Surface with Dissolved Copper Compounds. As opposed to the thermal decomposition copper plating mechanisms, the electrochemical reaction mechanism will only occur at steel surfaces and other metal surfaces which are anodic to copper in the electrochemical series (e.g., unoxidized aluminum, zinc, etc.). In the electrochemical plating mechanism, the iron is oxidized/ dissolves into the surrounding liquid as the dissolved copper compound is reduced to insoluble metallic copper which plates onto the steel surface. In refrigerant/oil systems, electrochemical reactions are inhibited by the low conductivity of the organic liquid (poor electron transfer) limiting the rate of electrical flow, and consequently, electrochemical reaction between the dissimilar metals. However, steel coupons were completely copper plated in heated [175[degrees]C (347[degrees]F)] R-134a/ POE oil systems when a 9Vdc battery was attached to a copper/steel coupon pair (Kauffman 1993).
In lubricating systems, additive studies evaluating film formation on a rotating steel ball inside a chamber filled with mineral or POE oil heated to 100[degrees]C (212[degrees]F) determined that copper metal oleates (cathodic/noble to iron) formed films (>100 nanometers) containing metallic copper and iron (II) oleate (Ratoi 2003). For mineral oils, the iron (II) oleate built-up as part of the film (insoluble) and for POE oils the iron oleate dissolved into the bulk fluid (soluble). When the test was repeated without rubbing, no copper species were detected on the steel surface. Based on these results, the following electrochemical reaction was postulated to explain the formation of the copper metal on the steel surfaces:
Fe (0) + Cu (II) oleate (soluble) [right arrow] Fe (II) oleate (insoluble/soluble) + Cu (0)
Whether the rubbing was necessary to produce a clean, unoxidized metal surface (air present) or to increase the temperature of the surface further promoting the above reaction was not determined.
Contaminants that Promote Copper Plating
Accelerated Stability Tests of System Contaminants. The majority of the identified references studied the effects of air, water, different carboxylic acids, hydrochloric acid and/or chlorinated refrigerant system residues on the occurrence of copper plating during sealed tube tests and accelerated compressor testing. The majority of reported sealed tube tests indicated that copper plating did not increase with water concentration for CFC/mineral oil or HFC/POE oil systems. Water did increase copper plating when the water became insoluble (Walker 1962) or if HC1 was present (Factor and Miranda 1991). Mineral oils obtained from compressor burnouts [both strong and weak acids with up to 50 ppm water (Kauffman 1992)] or previously exposed to HCl acid (Spauchus 1963) greatly increased copper plating during CFC/HCFC sealed tube tests. During accelerated compressor tests (Cavestri 2000), copper plating increased with water (200 ppm) or other contaminants (carboxylic acids and chlorinated refrigerants in HFC systems) only in the presence of air (4%v/v).
Field Reports of Copper Plating. As opposed to the accelerated sealed tube and compressor tests, the field reports of copper plating indicated that the presence of water levels as low as 200 ppm increased copper plating for R-134a/POE oil systems (amount of air not measured) and that the iron concentrations increased with copper plating (Herbe and Lundqvist 1997). Air conditioning systems being converted from CFC-12/mineral oil systems to HFC-134a/polyalkylene glycol (PAG) systems suffered heavy copper plating and heavy filter plugging (insoluble copper compounds) when shipboard refrigeration systems were converted to HFC-134a/POE system (Nickens 1992). Early work with PAG oils (1600 ppm water in PAG) reported heavy copper plating for compressor test stands and refrigerators running on R-134a (Reyes-Gavilan 1993). In contrast to other reports of copper plating, refrigeration systems running for 5 years with R-12/mineral oils or R-134a/POE oils showed no signs of copper plating (Reimer and Hansen 1996).
Using a 3/4-hp reciprocating, semi-hermetic compressor with a bronze main bearing to evaluate different R-134a/oil combinations, light to heavy copper plating occurred on the valve plate for 11 out of 17 PAG oils tested (Sundaresan and Finkenstadt 1992). Even though many of the 22 POE oils tested had initially high water or carboxylic acid contents and the majority of the compressor tests were terminated due to severe oil discoloration observed in the sight glass, no copper plating was observed.
System Driers/Filters. Sealed tube tests were run to determine the effects of molecular sieves (4A and 3A) and alumina on the stability of R-22/mineral oil, R-32/POE and R-134a/POE oil systems (Rohatgi 1998). Copper and steel coupon corrosion and copper plating were reported for all of the R-22/mineral oil systems run at 120[degrees]C (248[degrees]F) with the steel corrosion and copper plating being heavier in the presence of the alumina systems than for the molecular sieve systems. No copper plating was observed for the R-22/mineral oil/desiccant tests run at 100[degrees]C (212[degrees]F) even though elevated chloride ion contents were detected. Although the R-32/POE and R-134a/POE oil systems with alumina had organic acid and fluoride ion contents as high as 3% and 0.5%, respectively, no copper plating was observed for the non-chlorine refrigerant/oil systems.
Manufacturing Process Contaminants. A comprehensive list of manufacturing process fluids that may be present as contaminants in refrigeration systems (Cavestri and Schooley 1996) and sealed tube tests 150[degrees]C (302[degrees]F) and 175[degrees]C (347[degrees]F) to determine the effects of process fluids at 0.1 and 1.0% concentrations on the thermal stability of R-134a/mixed acid POE oil systems (Rohatgi 2006) have been reported. Of the chemicals tested, zinc chloride (used in brazing fluxes) was the only chemical to produce copper plating at 0.1 and 1% at both temperatures. Ferric phosphate dihydrate at 1% was the only other chemical that caused copper plating at 150C (302F). Sodium paratoluene sulfonate, potassium tetraborate tetrahy-drate, ferric phosphate dihydrate and sodium trisilicate hydrate used in cleaners, metal working fluids, and rust inhibitors produced copper plating at 175[degrees]C (347[degrees]F) at 0.1 and 1.0%.
INITIAL LABORATORY TESTS AND COPPER PLATING ANALYSES
Simple laboratory experiments in closed containers were performed in the absence of refrigerant to evaluate the literature search results. The refrigerant was not used because it was expected to inhibit the dissolution of the copper salts/compounds, and consequently, copper plating. The tests were run in small, closed containers to promote copper plating by minimizing the copper species transportation to the steel surface.
Copper Compound Thermal Decomposition Tests
Different copper compounds were heated on iron surfaces (steel foil, iron powder, etc.) in a nitrogen atmosphere from 25 to 400[degrees]C (77 to 752[degrees]F) to determine if their thermal decomposition formed copper metal films or was catalyzed by the iron surface. The tests determined that copper (II) carboxylates undergo minimal weight loss up to 250[degrees]C (482[degrees]F). Of the carboxylates tested, only copper (I) acetate was unstable and lost weight (decomposing) at 150[degrees]C (302[degrees]F). Copper (I) butanethiolate ("oleoresins" corrosion) underwent exothermic weight loss before 100[degrees]C (212[degrees]F) and produced an organic residue containing copper particles. Copper chlorides and copper oxides (gained weight) in the tested temperature range. Copper plating or catalysis by the iron surface was not observed for any of the tested copper compounds, i.e., thermal degradation appears unimportant in the copper plating mechanism.
Glass Vial Tests
To simulate the chemistries of acidic (carboxylic and HCl acids) oil drops adhering to copper tubing and/or compressor steel surfaces in the absence and presence of air contamination, several tests were run in 10-milliliter (mL) glass vials. Half of the prepared vials were allowed to sit at room temperature for 3 days while the second set was heated at 150[degrees]C (302[degrees]F) for 3 days. Due to their larger surface areas (accelerate the corrosion process), copper metal beads and copper (I and II) oxide powders were used in the tests instead of copper coupons. Copper (II) caprylate (octanoic acid), copper (I) acetate and copper (II) acetate were the carboxylates used in the tests. Chlorine containing compounds were supplied as chlorinated hydrocarbons, dry HCl gas in organic solvents, Cu (I and II) chlorides and iron (III) chlorides. The studied carboxylic acids included linear pentanoic/octanoic and branched 3,5,5 trimethylhexanoic acids to simulate the hydrolysis of linear and branched acid POE oils, respectively. To simulate the oxidation of POE or mineral oils, oleic (linear, unsaturated) acid was also used. Valve steel coupons were used in the tests to allow observance and analysis of the produced copper metal films.
Carboxylic Acid/Copper Compounds. To perform the tests, a carboxylic acid, a copper compound and a steel coupon were put in different glass vials. Regardless of the temperature or carboxylic acid, after 3 days the acids were deep blue in color for the vials containing copper (I or II) oxide or carboxylate and colorless for the vials containing copper beads. Regardless of the dissolved copper content, temperature or atmosphere, no copper plating was observed.
Repeat tests were then performed with the copper beads, oxides and carboxylates in which a water drop was added to the different vials during preparation. After 3 days at room temperature or after 4 hours of heating at 90[degrees]C (194[degrees]F), copper plating was observed on the liquid portion of the steel coupons for all of the tests with copper oxide or carboxylate (acid deep green) and the copper metal tests run in air (acid light green). In nitrogen or under vacuum, the copper beads produced no plating (reacted acid colorless). Surface corrosion was visible on the vapor portion of all the steel coupons and increased with copper plating and acid volatility.
The carboxylic acids from the copper plating tests were then analyzed for metal content using X-Ray fluorescence (XRF) spectrometry and the stressed carboxylic acids fell into three main groups. Only iron was detected by XRF (steel coupon corrosion) in the reacted carboxylic acids removed from the copper bead tests heated in vacuum and water. For the copper oxide tests that did not produce copper plating, the blue colored carboxylic acids contained high levels of copper (>100 ppm) and very low levels of iron (below 5 ppm). For the tests in which copper plating occurred, the green colored carboxylic acids contained varying levels of both iron and copper (in agreement with literature search).
To better characterize the composition/structure of the copper plating and steel corrosion, all of the steel coupons with copper plating were analyzed with Scanning Electron Microscope/Energy Dispersive Spectrometry (SEM/EDS). SEM microphotographs representative of the produced corroded steel surfaces and copper platings are shown in Figure 1. The microphotographs in Figure 1 show that the steel surface is pitted (up to 10 micron wide) and that the copper plating is heterogeneous consisting of copper nodules (copper also detected in corrosion pits).
[FIGURE 1 OMITTED]
Chlorinated Compounds. To evaluate the effects of the chlorinated metal salts or solvents on copper plating, iron (III) chloride (with copper beads), copper (I) chloride or copper (II) chloride was added to 1-chlorodecane (simulate chlorinated oil) and a steel coupon in different glass vials and sealed under air or vacuum. To evaluate the effects of HCl on copper plating, dry HCl gas in acetic acid, a steel coupon and copper metal/copper (I) oxide mixture were added to the 1-chlorodecane in different glass vials and sealed under an air atmosphere or under vacuum. Regardless of the temperature, atmosphere (air or vacuum), chlorinated metal salts or dry HCl gas used in the test, no copper plating or steel corrosion was detected by SEM/EDS analyses on the steel coupons and no dissolved metal species were detected by XRF analyses of the reacted 1-chlorodecanes removed from the test vials.
In contrast to the dry tests with chlorinated species, the addition of 1 drop of DI water to each prepared vial containing copper or iron chloride caused heavy copper plating to occur within one hour at room temperature. The copper plating was so heavy that the glass wall where the steel coupon was touching also was copper plated. As opposed to the copper platings formed by the organic acids and copper carboxylates, the copper platings produced by the metal chlorides contained a high level (40%) of chlorine.
Summary. The initial tests performed in the absence of refrigerant (expected to inhibit the dissolution of the copper salts/compounds) showed that carboxylic acids containing a conductive species (water, copper carboxylate, HCl, etc.) were sufficient to produce copper plating if steel surface corrosion also occurred. The copper platings were primarily copper metal for the carboxylates and copper chlorides for the chlorinated compounds. Air is only required for copper plating when copper metal surfaces are involved (forms acid reactive oxides on the inert copper surface).
ANALYSES OF COPPER PLATING REMOVED FROM FIELD REFRIGERATION SYSTEMS
To determine if the copper platings produced by the initial steel coupon tests performed in air and vacuum were similar in composition/morphology to platings produced on steel parts in field refrigeration compressors, the copper platings present on compressor crankshafts obtained from field refrigeration systems were analyzed: three R-22/mineral oil produced platings and one R-410/POE oil produced plating. The parts were washed with acetone and hexane prior to analysis in an attempt to remove residual oils/organic residues. All of the copper platings cut from the crankshaft surfaces for analysis were thin and discontinuous similar to those shown in Figure 2.
[FIGURE 2 OMITTED]
The EDS analyses of the copper platings did not detect elements such as chlorine, sulfur or phosphorus that would originate from chlorinated species or oil additives. The SEM microphotographs of the crankshaft surfaces/copper platings shown in Figure 3 were selected to illustrate the copper platings at the initial stages of development in wear and non-wear surfaces. To better understand the features in the microphotographs, a linear EDS scan was performed across the right crankshaft surface and copper plating in Figure 3 to determine their elemental compositions. The elemental spectra shown in right microphotograph in Figure 3 are the elements detected by the EDS analyses.
[FIGURE 3 OMITTED]
The left crankshaft surface in Figure 3 is covered by copper nodules (initial sites of copper formation) next to the wear scars and accumulated wear debris. The copper nodules are smaller, less numerous than those observed in the copper platings from the laboratory tests (Figure 1). The right crankshaft surface in Figure 3 contains pits and is covered by two separated layers of copper that appear to be made from numerous copper nodules similar in size to those of the laboratory tests (Figure 1). The elemental scan shows that the iron signal decreases only partially in the areas of copper plating (plating below 1 micron in thickness) while the weak silicon and oxygen signals are greatest for the unplated steel surfaces (Si and O in steel surface not copper plating). No chlorine was detected.
To study the thickness and elemental composition of the copper platings, Auger Emission Spectrometry (AES) with ion sputtering capability was used to depth profile the R-22 and R-410 produced platings as shown in Figure 4. The elemental profiles in Figure 4 indicate that the selected areas of the R-22 and R-410 copper platings are approximately 70 and 40 nanometers (nm) in thickness, respectively, (interface between copper plating and steel surface considered to be point where copper and iron elemental plots cross in Figure 4). The other two R-22 copper platings were approximately 45 and 15 nm in thickness.
[FIGURE 4 OMITTED]
Also, the AES analyses determined that the platings are primarily copper (80% for the thicker copper platings in Figure 4), so that the strong iron signals for the copper plating in Figure 3 are coming from the underlying iron surfaces (indicates Cu film less than 1 micron) and not from iron contamination of the copper platings. The elemental plot in Figure 4 indicates that once a thin (approximately 2-5 nm) layer of oil/varnish residue is penetrated, the copper plating contains less than 10% carbon and less than 5% total concentration of chlorine (regardless of the refrigerant), nitrogen, oxygen, or sulfur. Some of the non-copper content is due to detection of surrounding iron surfaces (concentrations of the non-copper elements do not trend with the copper concentration during the depth profiling).
Therefore, the elemental analyses in Figures 1- 4 of the lab and field samples determined that the submitted copper platings were copper nodules at initiation on the steel surface and were primarily copper as the platings increased in thickness (up to 100 nm or 0.1 micron) regardless of the refrigerant/ oil system. Two of the R-22 copper platings were produced with mineral oils containing triaryl phosphates indicating that steel surface passivation would not inhibit copper plating. The chemical and physical characteristics of the laboratory and compressor produced copper platings appear to be very similar, i.e., produced by similar electrochemical processes between steel surfaces and oil soluble copper compounds.
ANALYSES OF POWDERS REMOVED FROM REFRIGERANT COPPER TUBING SUPPLIES
In addition to the copper platings, two powders removed from commercial supplies of refrigerant copper tubing were also submitted for analysis. The SEM microphotographs showed that the powders were actually flakes (2-5 microns thick) as large as 100 microns in width. The EDS analyses of the flakes detected a wide range of copper and oxygen ratios ranging from 4:1 (copper and [Cu.sub.2]O) to 1:1 (CuO).
Since the EDS analyses were unable to obtain a consistent copper to oxide ratio (required to determine if powders are cuprous oxide, cupric oxide, oxidized copper particles, etc.) bulk analytical tests were performed. When copper, cuprous oxide and cupric oxide particles were heated in a thermogravimetric analyzer under nitrogen containing trace amounts of oxygen, the particles increased 25, 13 and 0% in weight, respectively, due to complete oxidation to cupric oxide. Both the cuprous oxide ([Cu.sub.2]O) and copper tubing powders reached an 11% (13% expected) weight gain. These results indicate that the copper tubing powder is primarily cuprous oxide. Both the copper tubing and cuprous oxide powders contain some cupric oxides on their surfaces accounting for the slightly lower than expected weight gains.
Therefore, the analyses determined that the copper tubing powder is cuprous oxide platelets/flakes with small amounts of cupric oxide on their surfaces. These results are in agreement the cited research in which cuprous oxides were produced on copper tubing surfaces by trace amounts of air in flowing fluorinated liquids (Lee 1970)) and by formicary corrosion during tubing storage (Corbett 2000). The cuprous oxide sample gave similar analytical results to the copper tubing powder, and consequently, was used as the copper source in the subsequent sealed tube and flowing copper plating tests.
SUMMARY OF LITERATURE SEARCH AND INITIAL LABORATORY TESTS AND ANALYSES
The basic conclusions determined from the literature search and initial research and analyses were:
* copper plating process is most likely electrochemical involving copper carboxylates
* the steel surface must undergo corrosion for copper plating to take place
* water promotes plating by supporting steel surface corrosion/providing conductive path
* air only has effect when copper metal surfaces need to undergo corrosion
* passivation of steel surfaces will not inhibit copper plating
* the copper platings created in the lab and field are similar in morphology and composition
* copper plating occurs on stationary and non-wearing rotating steel surfaces
HCFC AND HFC REFRIGERANT AND SOLVENT SCREENING TESTS
As opposed to the initial testing performed in both air and vacuum in tubes with sealable caps, the next sets of tests were performed using evacuated, sealed glass tubes/ampules. The tests were run with the copper species and steel surfaces in the same glass vial/tube to minimize the transportation requirements of the dissolved metal species between the copper/steel surfaces. A wide range of HCFC/HFC refrigerants and solvents containing different concentrations of lubricating oils [CGS commercial mineral oil and tetravaleric acid POE base-stock] and conductive species (carboxylic acids, water, HCl and copper carboxylates) were heated at temperatures up to 150[degrees]C (302[degrees]F) to produce copper platings on different metal surfaces. Chlorinated hydrocarbons were tested in this study to represent the chlorinated hydrocarbons produced by chlorination of the lubricating oil by degrading CFC/HCFC and thought to responsible for copper plating (Spauchus 1963).
Sealed Tube Tests--HCFC and HFC Refrigerants/Oils
The sealed tube tests were initiated using HCFC-22 refrigerant/mineral oil/oleic acid and HFC-134a refrigerant/POE oil/octanoic acid oil systems with valve steel strips and [Cu.sub.2]O/Cu metal powder heated to 150[degrees]C (302[degrees]F). The refrigerant:oil ratios were 50:50 by weight and the oils contained organic acids at concentrations of 0 - 10% (by weight). Even though the copper concentrations reached as high as 400 ppm (proportional to organic acid concentration), none of the sealed tubes had created visible copper plating on the steel coupons after 2 weeks at 150[degrees]C (302[degrees]F). The iron concentrations were below 5 ppm for the heated refrigerant: oil: acid solutions.
A second set of sealed tube tests [2 weeks at 150[degrees]C (302[degrees]F)] were then performed with water concentrations ranging from 50 ppm (dried oil) up to 2000 ppm water and organic acids at concentrations of 10%. Even with copper concentrations as high as 600 ppm and iron concentrations up to 50 ppm (coupon darkened/particles increased with water), no copper plating was visible on the steel coupons during the 2 week heating period.
A third set of sealed tube tests were then performed in which the refrigerant:oil ratio was varied from 10:90 to 90:10. The organic acid concentrations were held at 10% (acid to refrigerant:oil solution) in all of the sealed tubes. Since the previous sealed tube tests indicated that the oil:refrigerant reactions were minimal (HFC-32 and chloride ions were not detected in heated HCFC-22 tubes), the third set of sealed tube tests were performed at 90[degrees]C (194[degrees]F) for 4 weeks to minimize the tube pressures for the refrigerant:oil solutions with higher refrigerant:oil ratios. Again the dry tube tests produced no visible plating on the steel coupons even though the refrigerant: oil solutions were deep blue in color (copper concentration approximately 600 ppm in each tube). However, when 1000 ppm water was added to the sealed tubes with the high refrigerant:oil ratio prior to testing, copper plating occurred on the steel coupons during cooling.
Sealed Tube Tests-Solvents/Oils
Since the initial tests run in air and under vacuum were run without oil present and the refrigerant sealed tube tests only produced a homogeneous copper plating when the oil was at a minimum, the rest of the studies were run with oil concentrations of 5%. Since the HCFC and HFC refrigerants did not appear to be reactive with respect to the copper plating reaction and a copper plating proclivity test (Divers 1958) employed tetrachloromethane instead of refrigerants to promote plating, a series of chlorinated solvents were tested in the sealed tubes. The chlorinated solvents were mixed with 5% CGS mineral oil, 10% oleic acid and 500 ppm water and dispensed into a tube containing [Cu.sub.2]O/Cu metal powders and a steel coupon to test the effects of the chlorinated molecule structures on copper plating formation. The tubes were cooled, evacuated, sealed and allowed to sit at room temperature for 16 hours (over night). If copper plating was not observed then the sealed tubes were heated for 4 hours at 35[degrees]C (95[degrees]F). The sealed tube temperature was increased in steps of 10[degrees]C (18[degrees]F) up to 105[degrees]C (221[degrees]F) with four hours of heating at each step until copper plating was visible on the steel coupon. As the results listed in Table 1 demonstrate, the molecular structure has a strong effect on the copper plating tendencies of the chlorinated solvents.
Table 1. Temperatures Required to Produce Copper Plating with Solvents Solvent Chemical Formula 2,2-Dichloro-1,1,1-Trifluroethane [C.sub.2][Cl.sub.2][HF.sub.3] 1,1,2-Trichloro-2-Fluoroethane [C.sub.3][Cl.sub.3][H.sub.2]F 1,2,2-Trichloro-1-1,1-Difluoroethane [C.sub.2][Cl.sub.3][HF.sub.2] 1,1,2,2-Tetrachlorofluoroethane [C.sub.2][Cl.sub.4]HF Dichloromethane [CCl.sub.2][H.sub.2] Trichloromethane, Tetrachloromethane [CCl.sub.3]H, [CCl.sub.4] Tetrachloromethane [C.sub.2][Cl.sub.4] 1-Chlorodecane [C.sub.10][ClH.sub.21] 1-Chlorohexadecane [C.sub.16][ClH.sub.33] 3,3-Dichloro- [C.sub.3][Cl.sub.2][HF.sub.5] 1,1,1,2,2-Pentafluoropropane (HCFC-225ca) * 1,3-Dichloro- 1,1,2,2,3-Pentafluoropropane (HCFC-225cb) * 1,1,1,2,3,4,4,5,5,5-Decafluoropentane [C.sub.5][H.sub.2][F.sub.10] (HFC 43-10mee) [dagger] Solvent Plating Temperature ([degrees]C/[degrees]F) 2,2-Dichloro-1,1,1-Trifluroethane >105/221 1,1,2-Trichloro-2-Fluoroethane 85/185 1,2,2-Trichloro-1-1,1-Difluoroethane 105/221 1,1,2,2-Tetrachlorofluoroethane 65/149 Dichloromethane 45/113 Trichloromethane, Tetrachloromethane Room temperature Tetrachloromethane Room temperature 1-Chlorodecane 55/131 1-Chlorohexadecane 75/167 3,3-Dichloro-1,1,1,2,2-Pentafluoropropane >105/221 HCFC-225ca) * 1,3-Dichloro-1,1,2,2,3-Pentafluoropropane (HCFC-225cb) * 1,1,1,2,3,4,4,5,5,5-Decafluoropentane (HFC >105/221 43-10mee) [dagger] * 50/50 [dagger] Combined with octanoic acid and POE oil
All of the chlorinated hydrocarbon solvents in Table 1 reacted with the steel surface at temperatures below 75[degrees]C (167[degrees]F) to promote copper plating and the initial copper plating temperature decreased with increasing chlorine content in the molecule. For the tetrachloromethane, tetrachloroethylene and trichloromethane solvents, after several days at room temperature, the copper plating was no longer visible on the steel coupon surface, the submerged portion of the steel coupon became weak and the chlorinated solvent/oil solution turned into a gray, cloudy suspension (steel corrosion independent of copper plating). Several commercial remediation techniques use iron filings to dechlorinate tetrachloro -and trichloromethane contaminants in ground water. Ionic liquid chromatographic analyses were run on the water extracts of the stressed fluids produced by the sealed tube tests of the chlorinated hydrocarbons. As expected, the chloride concentrations of the stressed fluids decreased with the chlorine content of the hydrocarbon, i.e., tetrachloromethane, tetrachloroethylene and trichloromethane test liquids contained 30 - 45 ppm chloride ions while the 1-chlorodecane and 1-chlorohexadecane test liquids contained 4 - 8 ppm chloride ions.
The tertachloro- and trichloro -HCFC solvents in Table 1 required heating to produce copper plating. The plating temperatures of the HCFC solvents increased with decreasing chlorine content and were consistently higher than those of the chlorinated hydrocarbon solvents with similar chlorine contents. The chloride contents of the stressed HCFC test fluids also decreased with the chlorine content of the molecule. The chloride contents of the stressed dichloro-HCFC test fluids were below the blank.
The dichloro-HCFC and HFC solvents in Table 1 did not produce copper plating below 105[degrees]C (221[degrees]F). The tests were not increased past 105[degrees]C/221[degrees]F based on the fact that extended testing with R-22 and R-134a at 150[degrees]C (302[degrees]F) did not produce copper plating without the presence of excessive water and the temperature limits of the magnetic pump of the planned flowing tests were around 105[degrees]C (221[degrees]F). Therefore, various tests were performed in which the water and acid concentrations of the HCFC and HFC test systems in Table 1 were varied (5% oil concentration) to determine the minimum contaminant concentrations required to promote copper plating at temperatures below 105[degrees]C (221[degrees]F).
In the first set of sealed tube tests, the water was held constant at 2000 ppm and the acid concentration was decreased from 10% down to 0.1% as listed in Table 2. Since no chloride ions were detected in the test solutions' (below blank), the corrosion of the steel coupon (required for copper plating to take place) was being performed by the carboxylic acid in the presence of water. The main differences between the copper platings produced in Table 2 were the locations of the platings: the oleic acid (nonvolatile) platings occurred in the liquid phase and the octanoic acid (volatile) platings occurred in the vapor phase (no plating observed below liquid line).
Table 2. Temperatures Required to Produce Copper Plating with Varying Carboxylic Acid Concentrations Solvent Carboxylic Acid % Acid Plating * Temperature ([degrees]C/[degrees]F) HCFC-225ca/b Oleic 10 45/113 5 55/131 2.5 75/167 1.0 >105/221 0.1 >105/221 HFC 43-10mee Octanoic 5 45/113 2.5 75/167 1 105/221 0.1 >105/221 * Plating location: oleic acid--liquid and octanoic acid--vapor
In the second set of tests, the carboxylic acid was held constant at 5% to minimize the plating temperature and the water concentration was increased from 200 ppm up to 2000 ppm. As seen previously with the HCFC and HFC refrigerants, copper plating did not occur until 1000 ppm of water was added to the system. The copper plating occurred at 55[degrees]C (131[degrees]F) for both 1000 and 2000 ppm water regardless of the solvent/carboxylic acid system (position of plating again dependent on acid, oleic acid: liquid phase and octanoic acid: vapor phase).
Effects of Plated Metal Composition
In addition to the flowing test rig concerns, the plating surface metallurgy was also studied to better understand the plating mechanism. To test the compositional effects of the surface undergoing plating on the overall copper plating mechanism, metal wires of different compositions were sealed in separate glass ampules containing either 1-chlorodecane solution (5% CGS mineral oil, 5% oleic acid and 500 ppm water) or HFC 43-10mee solution (5% POE ester oil, 5% octanoic acid and 2000 ppm water) and [Cu.sub.2]O/Cu metal powders. The sealed ampules were heated at 55[degrees]C (131[degrees]F) for 24 hours before inspection for copper plating. For both solvent/oil/acid systems, the wires composed of stainless steel and Teflon (fluid wetted surfaces of pump, filter holders and plating chamber components) did not show any signs of corrosion or copper plating. However, the aluminum and galvanized (zinc coating) low carbon steel wires were plated with copper. Consequently, copper plating will only occur on surfaces which can be corroded/oxidized (aluminum, zinc, low carbon steel) and not on inert surfaces such as the stainless steel/ Teflon surfaces of the flowing rig. (Iron nails became plated, while carbon rods remained clean with dc voltage applied [Kauffman 2007)].
Summary of Lab Tests
The research with the sealed tubes and ampules indicated there are three basic types of low temperature [below 150[degrees]C (302[degrees]F)] copper plating mechanisms that can occur based on the type of refrigerant, oil and acid.
1. Chlorinated hydrocarbons [chlorinated oil postulated by Spauchus (1963)] undergo a dechlorination reaction with the iron surface and water causing iron surface to undergo corrosive oxidation promoting copper plating by dissolved copper species. Increasing the number chlorines in the hydrocarbon molecule increases the rate of dechlorination in the presence of water and steel surfaces without affecting the copper plating rate.
2. HCFC refrigerants/solvents are not directly involved in the copper plating reaction. Indirectly, HCFC degradation can produce chlorinated oil molecules to promote copper plating by the first mechanism. The copper plating occurs in the liquid phase involving reactions between long chain carboxylic acids (e.g., oleic acid), corroded iron surface, water and dissolved copper species. The presence of fluorines in the HCFC molecule greatly reduces the dechlorination susceptibility of the molecule.
3. HFC refrigerants/solvents are not involved in the copper plating mechanism to any degree. In fact, the HFC environment appears to inhibit copper plating so that plating occurs in the vapor phase by reactions between thin layers of volatile, short chain acids (e.g., pentanoic), corroded iron surface, condensing water and dissolved copper species (splashing, thermal gradient, etc. transport liquid containing dissolved copper species to reactive iron surface).
Copper plating will only occur on metal surfaces anodic to copper which can be corroded/oxidized (e.g., aluminum, zinc, low carbon steel) and not on inert surfaces such as the stainless steel, Teflon surfaces or graphite.
FLOWING TEST RIG AND COPPER PLATING ELIMINATION
Although the sealed tubes/ampules are useful for studying the individual steps of the copper plating mechanism, they are not as suitable for the copper plating elimination studies. Since the copper dissolution occurs at the copper tubing walls and the copper plating occurs on the steel parts of the compressor, the copper dissolution and plating processes should be separated in the laboratory studies used to study techniques for eliminating copper plating. Consequently, the flowing test rig shown in Figure 5 was designed to pump a refrigerant/oil/acid solution through 3 successive chambers containing:
[FIGURE 5 OMITTED]
1. Wetted [Cu.sub.2]O/Cu powder mixture (copper dissolution step/refrigerant hydration)
2. Copper plating inhibiting materials
3. Spinning stainless steel bar (window in chamber for viewing copper plating)
a. rectangular valve steel coupon attached to its upper or side surface (no rubbing)
b. circular valve steel coupon attached to its lower surface (rubbing against stainless steel)
Flowing Test Rig Studies--Promoting Copper Plating
The upper filter paper in Figure 5 was wetted with 0.5 mL of water (water concentration of flowing solvent dependent on water solubility) to ensure enough water was present to saturate the flowing solution and test rig's hydroscopic surfaces. Cu metal/[Cu.sub.2]O powders were added to the upper, wetted filter paper. The rig was evacuated and filled with 150 mL of 1-chlorodecane (representing chlorinated oil produced during HCFC-22 degradation) containing 5% CGS mineral oil and 10% oleic acid. The heater was turned on and when the steel chamber reached 55[degrees]C (131[degrees]F) the pump was turned on to start the solvent flow. Within 30 minutes of starting the test, the solution observed through the window of the stainless steel chamber was blue in color (copper dissolved). After 2 hours of starting the test, the valve steel coupon on the upper surface of the stainless steel coupon appeared red, i.e., copper plated (seen through the window in the stainless steel chamber). The test was allowed to proceed for another hour at which time the system plumbing was drained and the stainless chamber was opened (total test time = 3 hour) as shown in Figure 6.
[FIGURE 6 OMITTED]
As shown in Figure 6, both the upper surface (rectangular) and bottom surface (circular) valve steel coupons were copper plated. The center of the bottom valve steel coupon, which makes contact with the upraised portion of the stainless steel chamber floor (rubbing surface), had minimal copper plating apparent. No copper plating was apparent on any of the stainless steel chamber surfaces. The copper and iron concentrations of the stressed fluid were 150 and 40 ppm, respectively. Although enough water was added to the filter to increase the water concentration of the flowing fluid to 0.5%, Karl Fisher analysis determined the water concentration of the stressed fluid was 430 ppm.
In every test with at least 0.2 mL of water added to the filter paper, the stressed solution turned blue and copper plating occurred on both valve steel coupons. When no water was added to the filter paper, the flowing solution still turned blue but no copper plating was evident on the rectangular or circular valve steel coupon. The stressed fluid did not contain detectable iron (below 10 ppm) and the water content was 140 ppm.
Flowing Test Rig Studies--Copper Plating Elimination
Previous research during this study (Kauffman 2007) determined that iron surface passivation and electrochemical copper extraction with iron surfaces were unsuitable for inhibiting copper plating on compressor components. Another approach to inhibiting copper plating is to remove any dissolved copper from the circulating refrigerant by modified driers/adsorbents. Research to extract dissolved copper species from stored Navy jet fuels by solid supported chelators such as Silica Supported Triamine Tetraacetate (TATA), Diethylenetriamine Polymer Bound (DETA) and Silica Bound After Phosphoric Acid Treatment (DETA-P) have been reported (Puranik 2001). Based on preliminary testing tests, the flowing test rig studies to eliminate copper plating were performed with an alumina (25 grams)/TATA or DETA-P (9 grams) mixture placed on a paper filter in the lower filter cartridge in Figure 5.
Chlorinated Hydrocarbon Studies. The first set of flowing test rig studies to eliminate copper plating were performed with 1-chlorodecane (10% oleic acid/5% oil) with the stainless steel chamber heated to 55[degrees]C (131[degrees]F). The first test was performed without the addition of water to the upper filter paper in Figure 5 to confirm the fact that water had to be added to the flowing test rig to produce copper plating on the valve steel coupons in Figure 5. After 3 hours, the flowing solution was blue in color but no copper plating was apparent on the spinning valve steel coupon observed through the stainless steel chamber. The test was continued for another three hours and the stressed fluid was drained into a glass jar (Figure 7). As listed in Table 3, the drained fluid had below 5 ppm iron indicating iron corrosion and copper plating had not occurred (confirmed by steel coupons on the spinning bar). The blue color/copper concentration of 130 ppm in Table 3 confirm dissolved copper species were available for the plating mechanism. Consequently, the addition of water was required to initiate the corrosion of the steel coupon involved in the plating process.
[FIGURE 7 OMITTED]
Table 3. Flowing Test Rig Data for Stressed Fluids and Valve Steel Coupons Solvent Adsorbent Fe Cu (ppm) [H.sub.2]O Copper (ppm) (ppm) plating 1-Chlorodecane None * <5 130 139 No None 25 167 430 Yes Alumina/TATA <10 <10 175 No HCFC- None 11 110 545 Yes 225ca/b[dagger] Alumina/TATA <10 25 276 No HFC 43-10mee None 104/216 190/482 ** 1513 Yes [double dagger] ** Alumina/TATA 29 36 608 No * No water added [dagger] 3,3-Dichloro-1,1,1,2,2-Pentafluoropropane/1,3-Dichloro- 1,1,2,2,3-Pentafluoropropane [double dagger] 1,1,1,2,3,4,5,5,5,5-Decafluoropentane ** Unshaken/Shaken
Next, the 1-chlorodecane flowing test was repeated with the addition of 1 mL of water to the upper filter paper in Figure 5. The flowing liquid was again blue in color within three hours of starting the test and copper plating was already apparent on the spinning valve steel coupon observed through the stainless steel chamber. The test was terminated at three hours and the drained fluid contained 25 ppm iron in addition to the 167 ppm copper (Table 3). The dissolved iron content and copper plating on the steel coupons (similar to those in Figure 6) again indicate iron corrosion/dissolution is required for copper plating to occur. Although, enough water was added to the system to obtain a water concentration in the flowing fluid of 6,000 ppm, the water analysis of the stressed fluid was 430 ppm (remainder of water distributed among [Cu.sub.2]O/Cu powder mixture which looked like mud at end of test, filter paper in the lower filter cartridge, and other surfaces capable of supporting the insoluble, suspended water droplets).
To evaluate the copper plating elimination capabilities of the alumina/TATA mixture, the 1-chlorodecane flowing test was repeated with the addition of 1 mL of water to the upper filter paper and the alumina/TATA mixture to the lower filter paper in Figure 5. After six hours the stressed fluid was drained and as shown in Figure 7, the fluid was clear, colorless (TATA in lower filter cartridge was light blue indicative of chelated copper). Both the iron and copper contents of the stressed fluid were well below 10 ppm (Table 3) and no copper plating was observed on the valve steel coupons (Figure 7). A duplicate test with DETA-P in place of TATA had similar results, i.e., stressed fluid colorless, chelator light blue in color and no copper plating observed.
HCFC Solvent Studies. The second set of flowing test rig studies to eliminate copper plating were performed with HCFC-225ca/b (10% oleic acid/5% oil) with the stainless steel chamber heated to 55[degrees]C (131[degrees]F). As opposed to the 1-chlorodecane with a boiling point of 223[degrees]C (433F[degrees]), i.e., minimal vapor pressure, HCFC-225ca/b has a boiling point of 54[degrees]C (129[degrees]F) so that the stainless steel chamber was mainly vapor. Consequently, the upper, rectangular steel coupon was placed on the side of the stirring bar to interact with both the liquid and vapor phases of the heated, flowing fluid.
The HCFC-225ca/b flowing test was performed with the addition of 1 mL of water to the upper filter paper in Figure 5. The flowing liquid was again blue in color within three hours of starting the test but copper plating could not be confirmed since the steel coupon was on the side of the spinning bar. The test was terminated at six hours and the drained fluid contained 11 ppm iron and 110 ppm copper (Table 3). Although the trace metal contents were less than the 1-chlorodecane, both valve steel coupons (side and bottom) on the stirring bar were copper plated. The circular steel coupon on the bottom of the steel bar was partially plated (no plating in rubbing zone) similar to the circular steel coupon in Figure 6 and the water analysis of the stressed fluid was 545 ppm.
To further evaluate the copper plating elimination capabilities of the alumina/TATA mixture, the HCFC-225ca/b flowing test was repeated with the addition of the alumina/ TATA mixture to the lower filter paper in Figure 5. After six hours, the stressed fluid was drained and was clear, colorless (similar to 1-chlorodecane chelator test in Figure 7). The valve steel coupons on the stainless stirring bar were clean and free of copper plating. The iron and copper contents of the stressed fluid were below 10 and 25 ppm, respectively (Table 3) and the water content of the fluid was 276 ppm. A duplicate test with DETA-P in place of TATA had similar results, i.e., stressed fluid colorless and no copper plating observed.
HFC Solvent Studies. The third set of flowing test rig studies to eliminate copper plating were performed with HFC 43-10mee with the stainless steel chamber heated to 55[degrees]C (131[degrees]F). HFC 43-10mee has a boiling point of 55[degrees]C (131[degrees]F) so that the stainless steel chamber was mainly vapor. Consequently, the upper, rectangular steel coupon was again placed on the side of the stirring bar to interact with both the liquid and vapor phases of the heated, flowing fluid. In contrast to the chlorinated and HCFC tests which used one acid, oleic acid, the HFC 43-10mee tests were run with equal concentrations (2.5%) of hexanoic and octanoic acids (some pentanoic acid from hydrolysis and transesterification of POE ester) for a total of 5% carboxylic acid in the HFC 43-10mee.
The HFC 43-10mee flowing test was performed with the addition of 1 mL of water to the upper filter paper in Figure 5. The flowing liquid was again blue in color within three hours of starting the test but copper plating could not be confirmed since the steel coupon was on the side of the spinning bar. The test was terminated at six hours and the drained fluid was blue in color but hazy. So the liquid was allowed to settle/clear prior to the elemental test. The clear liquid contained 104 ppm iron and 190 ppm copper while the reshaken fluid contained 216 and 482 ppm of copper and iron, respectively (Table 3), indicating that the stressed HFC 43-10mee fluid contained both insoluble and soluble iron and copper species (hexanoic/octanoic acid carboxylates have low solubilities in flowing fluid). As to be expected from the elevated iron concentration, both steel coupons were copper plated. Again, neither coupon was completely covered by copper plating being restricted by either vapor/liquid patterns (side coupon) or rubbing (bottom coupon). The hydroscopic POE ester and carboxylic acids increased the water solubility of the solution up to 1513 ppm also helping to explain the high amounts of corrosion products (haziness) in the stressed fluid.
As a final evaluation of the copper plating elimination capabilities of the alumina/TATA mixture, the HFC 43-10mee flowing test was repeated with the addition of the alumina/TATA mixture to the lower filter paper in Figure 5. After six hours the stressed fluid was drained and was clear and colorless (similar to 1-chlorodecane chelator test in Figure 7). Although the iron and copper contents of the stressed fluid were elevated compared to the other chelator tests, 29 ppm iron and 36 ppm copper in Table 3, the steel coupons were clean and showed no signs of plating. The water content of the stressed fluid was also elevated compared to the other tests, 608 ppm. Again, a duplicate test with DETA-P in place of TATA had similar results, i.e., stressed fluid colorless and no copper plating observed.
Flowing Test Rig Summary
The research with the flowing rig test shown in Figures 6 and 7 agreed with previous results of the sealed tube/ampule tests (Kauffman 2007). The copper plating elimination capabilities of the adsorbent/chelator mixture were better in the flowing rig (copper corrosion/dissolution occurs in compartment separate from iron corrosion/copper plating) than in the sealed tube/ampules.
The water measurements (Table 3) indicated that the presence of alumina/chelator reduced the water contents of the stressed fluids. Follow-up acid number measurements of the stressed fluids also found decreases in the presence of alumina. Since the water and acid levels were reduced in the presence of alumina, the results indicate that the alumina was aiding in the copper plating inhibition by inhibiting the corrosion of the steel surface. In contrast to the alumina, both the silica supported TATA and DETA-P chelators were successful in eliminating the copper plating by removing the copper species from the stressed fluids, especially in the flowing test rig. Follow-up flowing rig tests showed that the TATA and DETA-P were capable of eliminating copper plating in the absence of alumina (higher levels of water and acid). As with any chelation system, the chelators have a finite capacity for adsorbing copper (copper oxide to chelator ratio of 1:12 eventually failed in sealed tubes while the 1:20 ratio used in the flowing test rig appeared sufficient for extended use).
FINALIZED COPPER PLATING MECHANISM
1. Reaction of Copper Tubing Surface with Oxygen to Produce Copper (I) Oxide ([Cu.sub.2]O) [Copper (II) Oxide (CuO) on surface of [Cu.sub.2]O particles]
Cu (0) Metal Surface + [O.sub.2] [right arrow] 2 [Cu.sub.2]O
2 [Cu.sub.2]O + [O.sub.2] [right arrow] 4 CuO (particle surface)
2. Reaction of Copper Oxides with Carboxylic Acids (R[CO.sub.2]H) to Produce Copper (II) Carboxylate [Cu[([O.sub.2]CR).sub.2]] [Copper (I) Carboxylates (CuO2CR) formed by interaction with metal: Formicary Corrosion]
[Cu.sub.2]O + 2 R[CO.sub.2]H [right arrow] 2 Cu[O.sub.2]CR (unstable) + [H.sub.2]O
8 Cu[O.sub.2]CR + [O.sub.2] [right arrow] 2 [Cu.sub.2]O + 4 Cu[([O.sub.2]CR).sub.2]
Cu[([O.sub.2]CR).sub.2] + Copper (0) Metal Surface [right arrow] 2 Cu[O.sub.2]CR
3. Dissolution of Copper Carboxylates by Refrigerant/Lubricating Oil (or Direct Corrosion of Copper Surfaces by Sulfur Containing Oils)/Excess Carboxylic Acids and Transportation of Copper Containing Oil Droplets by Circulating Liquid to Steel Surface
4. Oxidative Corrosion of Metal Surface (Fe) by Chlorinated Hydrocarbon (CFC/HCFC refrigerants) or Carbox-ylic Acids in Presence of Water. Other Corrosive Compounds (Oil Additives, Brazing Compounds)/Wear Can Also Contribute to Ferrous (II) Species
Fe (0) Metal Surface + Cl-Hydrocarbon + [H.sub.2]O [right arrow] Fe(II)Cl(OH) + Hydrocarbon
Fe (0) Metal Surface + [2 [(R[CO.sub.2]).sup.-] + [H.sup.+] in [H.sub.2]O] [right arrow] Fe(II)[([O.sub.2]CR).sub.2] + [H.sub.2]
5. Electrochemically Favored Reaction in Conductive Medium (e.g., Water, Carboxylic Acid) or Wear Region Between Copper Carboxylates and Reducing Agent [Fe (II) Species/Fe(0) Surface] to Produce Copper Metal Plating
Cu[([O.sub.2]CR).sub.2] + 2 Fe (II) Species [right arrow] 2 Fe(III)([O.sub.2]CR)Species + Cu (0) Metal Plating
Cu[O.sub.2]CR + Fe (II) Species [right arrow] Fe(III)([O.sub.2]CR)Species + Cu (0) Metal Plating
Cu[O.sub.2]CR + Fe (0) Surface (Wear) [right arrow] Fe(II)[([O.sub.2]CR).sub.2] + Cu (0) Metal Plating
The presence of water supports/accelerates many of the above reactions by hydrolyzing the POE esters to produce carboxylic acids, by increasing the solubility of the copper carboxylates, by promoting the corrosion of the steel surface (catalyzing dechlorination of the chlorohydrocarbons), by ionizing the carboxylic acid/other corrosive agents (zinc chloride used in brazing)] and by completing the electrical path of the electrochemical reaction in Step 5. The carboxylic acid in Step 2 could originate from tube forming oils, hydrolysis of POE esters or oxidation of mineral oil/POE ester lubricating oils. The presence of air or other oxidizing agent promotes the corrosion of the copper surface in Steps 1 and 2 but inhibits the copper plating of the steel surface in Step 5 (reducing iron (II) species oxidized by oxygen to inert iron (III) species). The iron (II) species can be produced by oxidative corrosion of non-wearing steel surfaces or in the wear region of the compressor (copper plating on edges of wear region from field samples and contact point from flowing rig test).
RECOMMENDATIONS TO PREVENT COPPER PLATING
Based on the results of the research and the proposed copper plating mechanism, copper plating can be prevented by excluding air and/or water from the refrigeration system as follows:
1. Exclusion of air from stored copper tubing and operating refrigeration system is the first priority because copper plating can not occur without copper oxidation.
a. Inhibits copper oxide formation in Reaction 1 required for copper corrosion and dissolution in Reactions 2 and 3. Copper oxides react with carboxylic acids to produce water promoting subsequent copper plating reactions.
b. Inhibits oil oxidation to produce carboxylic acids required for metal surface corrosion in Reactions 2 and 4.
2. Exclusion of water from stored lubricating oil and operating system is the next priority because copper plating can not occur without steel surface corrosion [Thin plating (blush) can still occur in wear regions (Reaction 5) in presence of copper carboxylates].
a. Inhibits corrosion of steel surface by carboxylic acid or chlorinated oil in Reaction 4.
b. Inhibits dissolution/transportation of ionic species to steel surface.
c. Inhibits electrochemical plating process in Reaction 5 by removing conductive/ionizing matrix.
d. Inhibits hydrolysis of POE lubricating oil or antiwear additives (triaryl phosphate) to produce corrosive carboxylic and phosphoric acids, respectively.
3. As a general rule, system cleanliness is important: carboxylic acids from lubricating oil, brazing chemicals such as zinc chloride, process chemicals such as carboxylates in drawing fluids from copper tubing, etc. should be excluded from the refrigeration system. The absence of air and water will decrease the effects of system contamination but copper plating can still occur if the contaminants promote the corrosion of the oxidized copper tubing and steel surfaces.
RECOMMENDATIONS TO ELIMINATE COPPER PLATING
If the air and water concentrations in the refrigeration system exceed the levels required for copper plating to occur, then copper plating can be eliminated by using an adsorbent/supported chelator combination. Alumina is especially useful as the adsorbent material for CFC and HCFC refrigerant/mineral oil systems in which chlorinated oil has been produced by refrigerant/oil reactions (thermal degradation/wear region). Also, any process that reduces the overall water and/or acid content of the refrigeration system will aid in eliminating continuing copper plating from the refrigeration system.
However, the most complete process for eliminating copper plating involves incorporation of a supported chelator(s) specifically designed for extracting copper and iron species from circulating systems, i.e., dissolved copper is removed from the flowing system prior to contact with compressor steel surfaces and reduction by dissolved iron (II) species. The silica supported Triamine Tetraacetate (TATA) and Diethylenetriamine after Phosphoric Acid Treatment (DETA-P) chelators tested during this research program are designed to remove dissolved Cu (II) species and were very effective in extracting dissolved copper species from both flowing HCFC/mineral oil and HFC/POE oil systems. If required, supported chelators to extract Cu (I) or other metal species of interest could be designed into the copper removal system to improve its effectiveness. The chelators change color as they absorb copper: light yellow - no copper present to blue - chelator saturated with copper. Therefore, the color change could be used to indicate the presence of dissolved copper species in the refrigeration system as well as a change out indicator (similar to indicating driers) to improve the reliability of the refrigeration system operation through both prevention and elimination of copper plating.
ASHRAE. 2002. 2002 ASHRAE handbook - Refrigerant System Chemistry, Chapter 5. Atlanta: American Society of Heating, Refrigerating and Air-Conditioning Engineers, Inc.
Calm, J.M. 2000. Consolidated project final reports: Materials Compatibility & Lubricants Research (MCLR) Program, The Air-Conditioning and Refrigeration Technology Institute, DOE-FG02-91CE23810.
Cartlidge, D. and Schellhase, H. 2003. Using acid number as a leading indicator of refrigeration and air conditioning performance. The Air-Conditioning and Refrigeration Technology Institute, ARTI-21CR/611-50060-01.
Cavestri, R.C. and Schooley, D.L. 1996. Compatibility of manufacturing process fluids with R-134a and polyolester lubricant. The Air-Conditioning and Refrigeration Technology Institute, ARTI MCLR Project No. 660 - 52000, DOE/CE/23810-55.
Cavestri, R.C., et al. 2000. Effect of selected contaminants in air conditioning and refrigeration equipment. The Air-Conditioning and Refrigeration Technology Institute, ARTI MCLR Project No. 655-53000, DOE/CE/23810-111, December 2000.
Corbett, R. 2000. Formicary corrosion [ant-nest corrosion] from CORROSION 2000, http://www.corrosion-lab.com/papers/formicary2000/formicary_corrosion-2000paper.htm.
DeVos, R. 1997. Evaluation of AHAM 134a stress test data. ASHRAE Trans., Vol. 103(2), pp. 640-648.
Divers, R.T. 1958. Better standards are needed for refrigeration lubricants. Refrigerating Engineering, pp. 40-45&59.
Factor, A. and Miranda, P.M. 1991. An investigation of the mechanism of the R-12/oil/steel reaction. Wear, Vol. 150, pp. 41-58.
Field, J.E., and Henderson, P.E. 1998. Corrosion of metals in contact with new refrigerants/lubricants at various moisture and organic acid levels. ASHRAE Transactions Vol.104(1A), pp. 210-220.
Herbe, L. and Lundqvist. 1997. CFC and HCFC refrigerants retrofit - experiences and results. International Journal Of Refrigeration, Vol.2 (1), pp 49-54.
Kauffman, R.E. 1992. Sealed tube tests of refrigerants from field systems before and after recycling. ASHRAE Research Project 683-RP, April 1992.
Kauffman, R.E. 1993. Accelerated screening methods for determining chemical and thermal stability of refrigerant-lubricant mixtures - Part II: Method assessment. Final Report. The Air-Conditioning and Refrigeration Technology Institute, ARTI MCLR Project No. 655-51500.
Kauffman, R.E. 2003. Mechanism for reaction between polyolester lubricant and ferrous metals. ASHRAE Research Project 1211-RP, Final report, Atlanta: American Society of Heating, Refrigerating and Air-Conditioning Engineers, Inc.
Kauffman, R.E., 2007. Determine the Mechanism for Copper Plating and Methods for Its Elimination from HVAC Systems, ASHRAE Research Project 1249 - RP, Final report, Atlanta: American Society of Heating, Refrigerating and Air-Conditioning Engineers, Inc.
Lee, L.H. 1970. Corrosion of copper in fluorochemical liquid coolants. Corrosion, Vol. 26(12), pp.529-532.
Lilje, K.C. 2000. The impact of chemistry on the use of polyolester lubricants in refrigeration. ASHRAE Transactions, Vol. 106, Part 2, pp. 661-667.
McGovern, E.W. 1939. Copper plating in refrigerant compressors," Refrigerating Engineering, Vol. 39, pp. 31-34.
Nickens, A.D., Brunner, G.F. and Hamilton, D.L. 1992. Navy investigations of HFC-134a as a replacement for CFC-12 in shipboard applications," Naval Engineers Journal, pp. 98-103.
Puranik, D., Morris, R.E. and Chang; E.L. 2001. Metal Complexing. US Patent No. 6,297,191.
Ratoi, M., Bovington, C. and Spikes, H. 2003. In situ study of metal oleate friction modifier additives. Tribology Letters, Vol. 14(1), pp. 33-40.
Reimer, A. and Hansen, P.E. 1996. Analysis of R-134a cabinets from the first series production in 1990. Proceedings of the 1996 International Refrigeration Conference at Purdue, pp. 501-505.
Reyes-Gavilan, J.L. 1993. Performance evaluation of naphthenic and synthetic oils in reciprocating compressors employing R-134a as the refrigerant," ASHRAE Trans., Vol. 99 (1), pp. 349-360.
Rohatgi, N.D. 2001.Effects of water in synthetic lubricant systems and clathrate formation: a literature search and review. The Air-Conditioning and Refrigeration Technology Institute, ARTI-21CR/610-50035-01.
Rohatgi, N.D. 1998. Effects of temperature on desiccant catalysis of refrigerant and lubricant decompositions. The Air-Conditioning and Refrigeration Technology Institute, ARTI MCLR Project No. 670 - 54200, DOE/CE/23810-95.
Rohatgi, N.D. 2006. Effects of chemicals in process fluids on the breakdown of HFC/POE systems. ASHRAE Trans., Vol. 112 (1) in press, Paper CH-06-14-3 at 2006 ASHRAE Winter Meeting.
Scott, G. 1965. Atmospheric Oxidation and Antioxidants. Chapter 7: Oxidative deterioration of saturated oil and polymers. Elsevier Publishing Company, New York.
Spauchus, H.O. 1963. Copper transfer in refrigeration oil studies. ASHRAE Journal, pp. 89-92, 140-142.
Steinle, H. and Bosch, R., 1964. Development and testing of lubricants for refrigerating machines. ASHRAE Transactions, Vol.70, pp. 195-202.
Sundaresan, S.G. and Finkenstadt, W.R. 1992. Polyalkylene glycol and polyolester lubricant candidates for use with HFC-134a in refrigeration compressors. ASHRAE Trans., Vol. 98(1), pp. 796-80.
Walker, W.O., Rosen, S. and Levy, S.L. 1962. Stability of mixtures of refrigerants and refrigerating oils. ASHRAE Journal, pp. 59-71.
This paper is based on findings resulting from ASHRAE Research Project RP-1249.
Robert E. Kauffman is a distinguished research chemist in the Nonstructural Materials Division, UDRI, Dayton, OH.