Environ. Sci. Technol. 2002, 36, 2273-2278
Investigation into the Rejuvenation of Spent Electroless Nickel Baths by Electrodialysis P A U L T . B O L G E R * ,† A N D D A V I D C . S Z L A G Clean Processes Branch, National Risk Management Research Laboratory, U.S. Environmental Protection Agency, 26 West Martin Luther King Drive, Cincinnati, Ohio 45268
Electroless nickel plating generates substantially more waste than other metal-finishing processes due to the inherent limited bath life and the need for regular bath disposal. Electrodialysis can be used to regenerate electroless nickel baths, but poor membrane permselectivity, leading to high losses of valuable bath components, continues to be a weakness of the technology. This research has investigated improving electrodialysis permselectivity for removing contaminants (sodium, orthophosphite, and sulfate) in a spent electroless nickel bath while minimizing the losses of valuable bath ions (nickel, hypophosphite, and organic acids). Ion permselectivity was explored with respect to electrodialysis operating conditions, membrane type, and cell configuration. Excellent permselectivity for sodium over nickel was attained irrespective of operating condition, membrane, or cell configuration. Studies on the effects of four different operating conditions (current density, pH, flow rate, and temperature) on anion permselectivity revealed bath pH and current density to be critical operating parameters. The type of anion exchange membrane used had a crucial effect on selectivity; one membrane (Ionac MA-3475) was identified as having superior selectivity for bath contaminants particularly for sulfate. The improvements in electrodialysis permselectivity established by this research will decrease waste generation within the electroless nickel process and increase resource productivity by minimizing the loss of valuable plating chemicals.
Introduction The electroless nickel (EN) process is the most important industrial catalytic plating process currently in use with a growth rate of 5% per year (1-3). The EN process produces metal deposits with better physical properties than electroplated nickel and is used to deposit metal coatings on small parts or in situations in which electrodeposition is not possible (4, 5). The general equation for the chemical deposition of nickel by hypophosphite is shown in eq 1
Ni2+ + 2H2PO2- + 2H2O f Ni + 2H2PO3- + 2H+ + H2 hypophosphite orthophosphite (1) One of the major shortcomings of the EN process is the inherent limited bath life due to the accumulation of harmful anions (sulfate and orthophosphite) and cations (sodium) * Corresponding author phone: +353-45-439510; fax: +353-45434207; e-mail:
[email protected]. † Current address: Bord na Mo ´ na Environmental Ltd., Main Street, Newbridge, Co. Kildare, Ireland. 10.1021/es015610t CCC: $22.00 Published on Web 04/19/2002
2002 American Chemical Society
in the bath, which results in rough deposits, poor corrosion resistance, and decreased plating rates (5-7). The orthophosphite is the product of hypophosphite oxidation during nickel reduction, and the sodium and sulfate are counterions to hypophosphite and nickel, respectively, which are continuously replenished in the bath. Without an effective means to remove these harmful components, the EN bath must eventually be discarded resulting in high disposal and raw material costs (8). Over the past decade, there has been significant progress in extending EN bath life via selective chemical precipitation, ion exchange, electrodialysis (ED), replacing sodium hypophosphite with nickel hypophosphite, and improved in-process bath maintenance (9). ED has been at the forefront of these EN bath regeneration technologies, and the technique has been patented for this purpose (10). The quality of an EN deposit from an EN bath rejuvenated by ED has been reported to be the same as a fresh EN bath deposit (11-13). Generally, the contaminants in a spent EN bath are removed only to a concentration where they no longer affect the plating rate or quality; various ED trials indicate that between 15 and 80% of orthophosphite, sodium, and sulfate can be removed (9). However, in trials where high levels of the contaminants are removed, there are invariably equally high losses of the useful constituents, i.e., hypophosphite and organic acids. Conceivably, due to proprietary protection concerns, most of the reports on spent EN bath rejuvenation by ED contain few details on the membranes used, the ED cell configuration, the operating conditions, or the loss of organic acids (14-16). To fill this information gap and to explore the possibility of improving the selectivity of ED for contaminants over useful components, a study of the ED of a spent EN solution was undertaken with the following objectives: (i) establish a baseline for the removal of the various bath components under four ED operating conditions, i.e., current density, bath pH, bath temperature, and solution flow rate; (ii) develop a complete ion analysis system for an EN solution to monitor the fate of all bath components (including the organic acid stabilizing agents) as previous studies on EN bath rejuvenation have assessed ion removal using crude criteria such as total dissolved solids or solution conductivity; (iii) evaluate a range of commercial anion and cation exchange membranes to determine their removal efficiency and selectivity in treating spent EN baths; and (iv) explore different ED cell configurations to determine if one particular configuration has an improved permselectivity for bath contaminants over useful components. The ED of a spent EN solution was evaluated under three separate criteria: the percent removal of each ion, the permselectivity of the membrane, and the current efficiency. The percent ion removal was calculated using eq 2
% ion removal ) [(CFC - CIC)/CID] × 100%
(2)
where CFC ) final ion concentration in concentrating stream, CIC ) initial ion concentration in concentrating stream, and CID ) initial ion concentration in depleting stream (EN solution). The total anion permselectivity (TAP) of the membrane was evaluated by comparing the beneficial removal of bath contaminants (orthophosphite, sulfate) with the loss of valuable components (hypophosphite, propanoic, and succinic acid). The TAP was calculated (see eq 3) using a modified version of the equations used by Sata (17) to calculate ion exchange equilibrium constants and permselectivities. VOL. 36, NO. 10, 2002 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
2273
TAP ) (CRcont./CRval.)/(CBcont./CBval.)
(3)
where CRcont. ) moles of contaminants removed; CBcont. ) moles of contaminants in EN bath; CRval. ) moles of valuable components removed; and CBval. ) moles of valuable components in EN bath. The orthophosphite/hypophosphite permselectivity (OHP) of a membrane was calculated by an equation similar to eq 3, replacing moles of contaminants with moles of orthophosphite and moles of valuable components with moles of hypophosphite. Both TAP and OHP include a concentration factor that accounts for the concentration difference between contaminants and the valuable components in an EN bath. The current efficiency (fraction of electrical charge involved in the transport of ions across a membrane) was calculated using eq 4
current efficiency ) (Σ(CR × Z) × F)/I × T
(4)
where CR ) moles of ion removed; Z ) ion charge; T ) time (sec); F ) Faraday’s number (96 500 A s mol-1); and I ) current (amps).
Experimental Methods A simulated spent EN bath was utilized in order to control the composition and pH of the EN bath and to use organic acids that can be separated from the hypophosphite and orthophosphite anions by ion chromatography. The composition of the simulated solution was based on a typical spent EN bath (4, 5, 18) and is shown in Table 1 along with the concentration of each component and the method of analysis for that component. The following chemicals were used in bath preparation: nickel sulfate hexahydrate (Aldrich, 98+%), sodium phosphite pentahydrate (Riedel-de-Hae¨n, 98%), phosphorus acid (Riedel-de-Hae¨n, 98%), sodium sulfate (Aldrich, 99+%), sodium hypophosphite monohydrate (Fluka, 99%), propanoic acid (Aldrich, 99%), and succinic acid (Aldrich, 99%). The pH of a simulated bath with the ion concentrations given in Table 1 was 4.5. In the trials where pH was a variable parameter, the pH was adjusted below 4.5 using phosphorus acid (H3PO3) and above 4.5 using sodium orthophosphite (Na2HPO3). In the aforementioned trials, the concentration of sodium in the bath was not constant due to the different amounts of H3PO3 and Na2HPO3 used but the amount of phosphite remained the same. To determine the effect that a relatively high bath pH (pH > 5) had on the removal of the weakly dissociating organic acids by ED, a second spent EN bath was prepared without nickel sulfate. The nickel was not added to a bath at a pH > 5 as nickel orthophosphite slowly precipitates from solution at this pH. After each ED trial, the ions remaining in the bath were quantified and the bath was rebuilt to the original composition (see Table 1) and reused. The ED cell was a bench-scale ElectroMPcell (Electrocell AB) with a nickel plate cathode and a platinized titanium anode. The compartment frames and gaskets had a membrane/membrane and membrane/electrode gap of 12 mm and the surface area of exposed membrane (or electrode) was 10 cm2. A range of cation and anion exchange membranes were obtained from different companies (via The Electrosynthesis Co.) and were selected on the basis of their chemical and physical characteristics. An inexpensive, general purpose membrane was chosen as the benchmark for comparison; the different chemical and physical features of the other membranes were monovalent ion selectivity, low resistivity, and chemical stability. The current was controlled by a 20 amp variable current power supply (TecNu Durapulse), and the potential drop across the cell was measured using a multimeter. The flow rate was measured by a flow meter and controlled by a needle valve, and all trials were 2274
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 36, NO. 10, 2002
TABLE 1. Concentration and Analysis Technique for the Ions in a Simulated Spent EN Bath component
g L-1
molarity
analysis
nickel sodium orthophosphite sulfate hypophosphite propanoic acid succinic acid
5 75 135 60 15 10 10
0.086 3.260 1.687 0.625 0.230 0.135 0.085
ICPa ICP ICb IC IC IC IC
a
Inductively coupled plasma spectroscopy.
b
Ion chromatography.
carried out at room temperature with the exception of trials where temperature was the varied parameter. In these trials, the temperature was controlled by circulating water heated in a water bath around a silicone tube placed in the electrolyte solution tanks. A 0.1 M sodium sulfate solution was used as the anolyte/ catholyte except in ED trials where a phosphate buffer 8 solution was required; the buffer solution was prepared according to the CRC Handbook (19). The anolyte and catholyte were kept in separate reservoirs for recirculation around the electrode compartments in operating condition trials, but in membrane testing trials, the anolyte and catholyte were mixed together as a single electrolyte. All ED experiments were conducted over a 24 h period, and samples were taken from each feed tank at regular intervals for ion and pH analysis. The anions were quantified on a Dionex 500 Ion Chromatograph using a 4 mm Dionex AS11-HC anion exchange column (mixed isocratic-hydroxide gradient elution) and an ED50 conductivity detector. The ion chromatography procedure for separating the anions was developed in-house and is the subject of a separate publication (20). Sodium and nickel ions were quantified by inductively coupled plama spectroscopy on a Perkin-Elmer 3300 DV instrument.
Results Effect of ED Operating Conditions on Ion Removal. Cation Removal and Permselectivity. The metal cations in the EN solution were sodium and nickel, and the cation exchange membrane used in the operating condition trials was a monovalent ion selective Tokuyama Neosepta CMS membrane. The ED trials on the EN solution at the four different operating conditions showed that the permselectivity of the Tokuyama Neosepta CMS cation selective membrane was very much in favor of sodium over nickel; the nickel loss from the EN solution was 200-400 mA cm-2, the percent removal of the lower concentration anions
FIGURE 1. Percentage of each anion removed from an EN bath and the TAP as a function of current density on ED at 2 L min-1, pH 4.5, and 23 °C.
FIGURE 2. Percentage of each anion removed from an EN bath and the TAP as a function of pH on ED at 2 L min-1, 400 mA cm-2, and 23 °C. (hypophosphite, propanate, and succinate) remained constant but the percent removal of the higher concentration anions (orthophosphite and sulfate) continued to increase at the same rate (see Figure 1). It is probable that the limiting current density for the lower concentration anions was in the region of 200-400 mA cm-2 and that above this current density the percent removal for the lower concentration anions was mass-transfer-controlled. The pH of the spent EN solution has a very significant effect on the percent anion removal and ED permselectivity (see Figure 2). The percent anion removal was highest for all anions at the normal bath pH of 4.5. However, in direct contrast to the percent anion removal, the TAP was highest at pH 2. On proceeding from a high (pH 4.5) to low pH (pH 2), the decrease in percent removal for hypophosphite, propanate, and succinate was relatively greater than that for orthophosphite and sulfate and consequently the TAP increased almost linearly from pH 5 to pH 2. The range of values for solution flow for the flow rate trials were chosen to encompass the optimum flow velocity of a sheet-flow ED frame, which is estimated to be 0.05 m s-1 (22). The horizontal cross-sectional area of an upright ED frame in the ED cell was 0.0012 m2 so flow rates of 0.5, 2, 4, and 10 L min-1 corresponded to velocities of 0.0069, 0.027, 0.054, and 0.139 m s-1, respectively. The percent removal of all the anions in the EN solution increased as the flow rate was increased (see Figure 3). Although higher flow velocities increased the percent removal of each anion uniformly, the higher fluid velocities did not have a corresponding impact on the TAP. The effect of EN bath temperature on the percent anion removal by ED is shown in Figure 4. A temperature increase in an electrochemical system usually enhances mass transport rates via a combination of decreased solution viscosity and increased diffusion coefficients. The percent removal for all anions (except sulfate) increased on raising the bath temperature from 20 to 35 °C.
FIGURE 3. Percentage of each anion removed from an EN bath and the TAP as a function of flow rate on ED at 400 mA cm-2, pH 4.5, and 23 °C.
FIGURE 4. Percentage of each anion removed from an EN bath and the TAP as a function of temperature on ED at 400 mA cm-2, pH 4.5, and 2 L min-1. However, the percent anion removal remained constant on increasing the bath temperature further from 35 to 50 °C. The TAP decreased slightly from 0.57 to 0.50 on raising the temperature from 20 to 50 °C. It is apparent that for all of the operating conditions explored, the percent sulfate removal was quite low relative to the other bath anions. This low removal rate for sulfate was also a feature of most of the anion exchange membranes tested with the exception of the Ionac MA-3475 (see Membrane Testing). Membrane Testing. Cation Exchange Membranes. The percent removal of sodium and nickel ions from a spent EN solution by ED using different commercial cation exchange membranes is shown in Table 2.
TABLE 2. Comparison of Percent Cation Removal from a Simulated Spent EN Solution on ED Using Four Different Commercial Cation Exchange Membranesa membrane name/source
membrane features
Na+ % rem
Ni2+ % rem
ESC-7001 electrosynthesis CMS Tokuyama
general purpose
11.2
0.0
monovalent permselective low resistance high stability
13.5
0.0
13.1 13.0
0.0 0.0
CM-1 Tokuyama Nafion 350 DuPont
a At a current density of 400 mA cm-2, pH of 4.5, solution flow rate of 2 L min-1, and temperature of 23 °C.
The same anion exchange membrane (Tokuyama AM-1) was used for the assorted trials. Although the four cation exchange membranes tested had quite different properties (monovalent cation selectivity, low resistance, or high stability), there was not much difference between percent VOL. 36, NO. 10, 2002 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
2275
TABLE 3. Comparison of the Percent Anion Removal, TAP, OHP, and Total Moles of Anions Removed on ED of a Simulated Spent EN Solution for a Range of Commercial Anion Exchange Membranes at a Current Density of 400 MA cm-2, pH of 4.5, Solution Flow Rate of 2 L min-1, and Temperature of 23 °C (CEM Was Tokuyama cm-1 for All Trials) H2PO2- propanate H2PO3- succinate SO42moles anion % rem % rem % rem % rem % rem TAP OHP rem'd
membrane name/source
membrane features
ACM Tokuyama ESC-7001a electrosynthesis ACS Tokuyama ASV Asahi glass AAV Asahi glass AM-1 Tokuyama MA-3475b Ionac AMT Asahi glass AHA Tokuyama
low H+ transport general purpose monovalent permselective monovalent permselective low H+ transport low resistance general purpose concentration stable to oxidation and high pH
a High membrane resistance; current density ) 200 mA cm-2. too high across membrane; no current passed.
b
9
23.9 7.3 10.3 13.7 R 11.2 5.2 13.3 14.5
13.8 7.8 13.2 16.5 R 13.5 7.2 12.7 12.3
21.8 7.3 12.1 15 R 13.4 6.2 13.7 13.0
7.0 0.59 9.5 1.01 7.2 0.87 9.8 0.85 R R 13.7 1.14 12.6 1.30 12.9 0.95 6.4 0.81
0.58 0.89 0.81 0.79 R 1.02 0.95 0.95 0.90
0.33 0.17 0.28 0.35 R 0.28 0.15 0.27 0.37
High membrane resistance; current density ) 170 mA cm-2. c R ) resistance
removals of sodium and nickel for the different membranes. There was only a 17% difference between the highest and the lowest values for sodium removal. There was no detectable loss of nickel from the spent EN bath for any of the cation exchange membranes tested, and there appeared to be no selectivity advantage to using a monovalent CEM such as Tokuyama CMS. The ED trials were carried out over a 24 h period, and it is possible that over a longer time period a monovalent CEM might emerge as a superior membrane for preventing nickel loss. Anion Exchange Membranes. The results of the anion exchange membrane investigations are presented in Table 3. The percent anion removal for the Asahi AAV membrane could not be determined as the membrane resistance was too high to pass current with a 20 V power supply. Similarly, the strikingly low levels of anion removal for the Electrosynthesis ESC-7001 and the Ionac MA-3475 AEMs were due to high membrane resistance and it was only possible to pass current densities of 200 and 170 mA cm-2, respectively, for these membranes. The percent anion removal for the AEMs varied significantly from membrane to membrane, and consequently, the TAP and OHP of the various membranes were very different. The Tokuyama ACM membrane had a comparatively low TAP (0.59) and OHP (0.58) that was attributable to high losses of valuable hypophosphite and organic acids. The five membranes ESC-7001, ACS, ASV, AMT, and AHA had a similar range of values for TAP (0.81-1.01) and OHP (0.79-0.95) even though these membranes had quite different individual properties, i.e., ion concentration, monovalent selective, oxidation stability, etc. The only membrane with an OHP > 1 was the Tokuyama AM-1 with an OHP of 1.02 and a TAP of 1.14. However, it was the Ionac-3475 that had the highest TAP (1.30) even though it had only a slightly above average OHP value (0.95). The high TAP for the Ionac-3475 membrane was due to the exceptionally high removal of SO42- in comparison with the other anions. The percent SO42- removal for the Ionac-3475 was almost twice that of any other bath anion, which was in direct contrast to other AEMs where the SO42- removal rate was significantly less than the other anions. The TAP and OHP of the Ionac-3475 membrane were increased to 1.4 and 1.0, respectively, by decreasing the pH of the EN bath to pH 3 and running the cell with the anolyte and catholyte in separate reservoirs. ED Cell Configuration. The four different ED cell configurations explored in this work are illustrated in Figures 5-8. The term “cell configuration” refers to the number and arrangement of the membranes and the number of electrolyte and EN solution streams within an ED system. For example, ED cell configurations 1 and 2 (Figures 5 and 6, respectively) differ because in ED cell configuration 1 the anolyte and catholyte streams are separate but in ED cell configuration 2276
19.5 8.7 16.4 20.5 Rc 13.2 7.5 13.2 13.6
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 36, NO. 10, 2002
FIGURE 5. ED cell configuration 1: three compartment, three stream, and two membrane.
FIGURE 6. ED cell configuration 2: three compartment, two stream, and two membrane. 2 the streams are mixed giving an overall two stream system. The essential objective of the cell configuration studies was to manipulate the pH conditions within the system to selectively recover valuable monovalent acids. The TAP and OHP results for ED trials with ED cell configurations 1-4 are presented in Table 4. ED cell configuration 2 was investigated to determine if a mixed anolyte/catholyte scenario (anolyte pH 5.5) had a different permselectivity to a separate anolyte/catholyte configuration (steady state anolyte pH ∼ 1). However, despite the anolyte pH difference, ED cell configurations 1 and 2 displayed a similar TAP and OHP. Phosphate buffer (pH 8) was used as an electrolyte in ED cell configuration 2 (expt 3) to attempt to control the electrolyte pH between 7 and 8. After 1-2 h of ED, the buffer pH had decreased to pH 5.5 and over a period of 24 h the TAP and OHP was found to be the same as with Na2SO4 as electrolyte. In one trial using ED cell configuration 2, the mixed electrolyte pH was controlled at pH 7-7.2 by continuously adding concentrated KOH to the electrolyte (expt 4) but this did not appear to significantly increase the value of the TAP or OHP. In ED cell configuration 3 (Figure 7), the anions in the spent EN bath were transported across AEM 2 from com-
FIGURE 7. ED cell configuration 3: five compartment, two stream, and four membrane.
ions. However, even under these conditions where the pH was artificially maintained at pH 7 by KOH addition, there was no improvement in selectivity for hypophosphite and propanate over the other anions. ED cell configuration 4 was designed to allow the electrolyte solution in compartments C and E to contact the cathode and thereby maintain a high pH during ED. This eliminated the need to add concentrated KOH to the electrolyte to keep the pH between pH 7-8. When ED was initiated, the pH in compartment C and E increased to pH 12, the pH of compartment A decreased to pH 2, and the pH of the EN solution stream remained steady. After 6 h of ED, the anolyte volume (compartment A) had decreased by 80% due to migration of the H3O+ ion across CEM 1 with consequent dilution of the EN solution. Although this configuration can produce the high pH necessary to maintain the orthophosphite and succinate in divalent form, there is a water balance issue due to the low pH anolyte being in contact with the cation exchange membrane.
Discussion
FIGURE 8. ED cell configuration 4: five compartment, three stream, and four membrane.
TABLE 4. TAP and OHP for Different ED Cell Configurations, Electrolytes, and pH Values at a Current Density of 400 MA cm-2, Solution Flow Rate of 2 L min-1, and Temperature of 23 °C (CEM, Tokuyama CMS; AEM, Tokuyama ACM; Monovalent AEM, Tokuyama ACS) expt no.
configuration/electrolyte/pH
TAP
OHP
1 2 3 4 5 6 7 8
1/Na2SO4 2/Na2SO4 2/buffer 8 2/buffer 8/pH controla 3/Na2SO4 3/buffer 8 3/Na2SO4/pH control 4/Na2SO4
0.57 0.52 0.52 0.51 STb 0.57 0.48 WBc
0.58 0.57 0.60 0.59 0.59 0.55 0.57 WB
a pH held at 7.0-7.2 by KOH addition. transport. c WB ) water balance issue.
b
ST ) electrolyte sulfate
partment D to compartment C. The inclusion of a monovalent AEM in ED cell configuration 3 offered the opportunity to recover monovalent acids such as hypophosphite and propanate back to the EN solution (in compartment B). ED cell configuration 3 was operated with and without electrolyte pH control. With no pH control (expt. 5 and 6), the electrolyte sulfate ions migrated across the monovalent AEM and accumulated in the EN solution. The transport of the electrolyte anions appeared to be a distinctive problem for this type of cell configuration; for example, where phosphate buffer was used as an electrolyte, phosphate ions migrated across the monovalent AEM to accumulate in the EN solution. Under pH control (expt. 7), the pH of the mixed electrolyte in compartment C was neutral/alkaline maintaining orthophosphite and succinate in a divalent state and thus monovalent hypophosphite and propanate may have been more selectively recovered. The transport of electrolyte anions did not occur with electrolyte pH control (expt 7) presumably due to preferential transfer of excess OH- ions over electrolyte
The parameter of primary concern in this study was the permselectivity of the ED membranes for sodium, orthophosphite, and sulfate over other components in the EN solution. The selective removal of sodium over nickel from an EN bath was relatively straightforward to achieve, and the only operating condition variable that had an effect on sodium permselectivity was current density. The high permselectivity for sodium over nickel can be attributed not only to the monovalent cation selectivity of the membrane but also to the complexation of nickel ions by the organic acids and the larger ionic size of nickel(II). In fact, the use of a monovalent cation exchange membrane for selective sodium removal may not be necessary as there was no detectable loss of nickel for either a monovalent membrane or a nonmonovalent membrane and there were only minor differences in the percent sodium removal between all of the cation exchange membranes. This indicates that the choice of a cation exchange membrane for rejuvenating spent EN baths by ED should be based on a membrane feature such as cost, low resistance, or high stability rather than on monovalent selectivity. The selective removal of orthophosphite and sulfate anions from an EN bath was more problematic and complex than the selective removal of the sodium cation. The current density investigations showed that the percent removal of the lower concentration anions in the EN bath was restricted by their lower limiting current density relative to the high concentration anions. In a spent EN bath, the lower concentration anions are the components that are desirable to retain in the solution. It is then possible to optimize ED permselectivity for contaminant ions over valuable bath ions, i.e., increase the TAP by operating at higher current densities. Although increasing the current density from 100 to 600 mA cm-2 tripled the moles of anion removed from the EN solution, the current efficiency decreased from 32% at 100 mA cm-2 to 15% at 600 mA cm-2. The pH of the EN bath exerts considerable influence on the percent anion removal and the TAP. A reduction in EN bath pH from 4.5 to 2.0 resulted in a doubling of the TAP value. This result can be explained by an examination of the dissociation constants of the acids in the bath. The first acid dissociation constants of hypophosphorus, phosphorus, propanoic, and succinic acid are 1.1, 1.3, 4.9, and 4.2, respectively (19, 23). The organic acids will be mostly protonated at pH 2-3 and will be unavailable for ED transport resulting in a higher TAP. The percent anion removal in the EN solution increased almost linearly as the flow rate was increased. This is in accordance with electrochemical theory that an increase in flow velocity will reduce the boundary layer thickness at the membrane and decrease concentration VOL. 36, NO. 10, 2002 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
2277
polarization. However, there is no benefit to operating at a high flow rate from an anion permselectivity perspective. A temperature increase from 23 to 50 °C did increase the percent anion removal, but the temperature change did not greatly affect the TAP. The relative stability of the TAP on increasing bath temperature is important as many commercial ED systems used for rejuvenating EN baths are frequently operated above room temperature due the high temperatures of the plating bath. In contrast to the cation exchange membranes, the various anion exchange membranes provide very different values for percent anion removals, OHP, and TAP. There is no apparent correlation between specific AEM properties (such as low resistance, monovalent selectivity, or oxidation stability) and the TAP and OHP. The Ionac-3475 AEM has a clear selectivity advantage over the other AEMs investigated with a TAP (1.30) that is almost 40% higher than the average TAP (0.94) for the membranes. The high TAP value is due to an above normal membrane selectivity for sulfate. It is not possible to relate this high selectivity to a specific membrane property except that the Ionac-3475 AEM is thicker and has a higher resistance than most of the other membranes. Given the importance of the EN bath pH to membrane anion permselectivity, the ED cell configuration research was carried out with a view to controlling the pH of the electrolyte during ED. Both hypophosphorus and propanoic acid are monobasic acids, but dibasic phosphorus acid has a pKa2 ) 6.7 so at a neutral pH the majority of orthophosphite will exist as divalent H2PO32- (along with SO42-). It was thought that an alternate ED cell configuration might be able to exploit this difference in acid valence chemistries and provide improved permselectivity for contaminants over valuable components. The concept was to remove the anions by normal ED through a non-monovalent selective membrane into an electrolyte at pH 7 and then to selectively recover the monobasic acids (hypophosphite and propanate) through a monovalent membrane. Clearly, at the pH that phosphorus acid is divalent, succinic acid will also be in a divalent form (pKa2 ) 5.6) and thus cannot be recovered by operating with higher pH electrolyte. The recovery of monovalent anions in this manner proved impractical due to the undesirable transport of the high concentration electrolyte anion across the monovalent AEM into the EN solution, i.e., the EN solution becomes contaminated with electrolyte anions. The ED cell configuration studies indicate that a simple three compartment cell provides the same permselectivity results as the five compartment cells. In addition, there was not a significant permselectivity difference between using a mixed anolyte/ catholyte and a separate anolyte and catholyte. A mixed anolyte/catholyte should then be used in EN bath rejuvenation operations as the extreme electrolyte pH values generated under a separate anolyte and catholyte configuration can hasten membrane degradation. This work has shown that selective removal of sodium over nickel from an EN bath can be achieved using a cheap, general purpose cation exchange membrane. It also identified that pH and current density are the critical operating parameters in selecting contaminant anions over the valuable anions. By simply decreasing the pH of a spent EN bath by one pH unit, it may be possible to substantially reduce losses of valuable EN bath components during ED. This decrease in pH may be carried out as part of normal bath operation. The pH of an EN bath will tend to decrease naturally with use due to the production of hydrogen ions during plating (see eq 1), and typically, a base (such as ammonia) has to be added to raise the pH. However, it would be possible to allow the pH to decrease to approximately pH 3-3.5 toward the end of bath life to optimize the permselectivity for the contaminants over the valuable bath components during ED rejuvenation. An increase in the current density across 2278
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 36, NO. 10, 2002
the ED cell will provide higher anion removal and better permselectivity, but an electroplater would have to weigh this advantage against a lower overall cell current efficiency and an increased energy requirement. Finally, it is worthwhile for an EN plater to give some consideration to the anion exchange membrane used for rejuvenating a spent EN bath. This work demonstrates that the Ionac-MA3475 membrane provides much better selectivity for contaminants over valuable components than the other anion exchange membranes.
Acknowledgments This research was supported by an appointment to the Postgraduate Research Participation Program at the National Risk Management Research Laboratory administered at the Oak Ridge Institute for Science and Education through an interagency agreement between the U.S. Department of Energy and the U.S. Environmental Protection Agency. The discussion of certain technologies and mentioning of trade names in this paper does not in any way represent an endorsement of those technologies/products by the U.S. Environmental Protection Agency.
Supporting Information Available A table containing data for the percent ion removal, TAP, and total moles of anions removed on ED of a simulated spent EN bath at different operating conditions. This material is available free of charge via the Internet at http:// pubs.acs.org.
Literature Cited (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13) (14) (15) (16) (17) (18) (19) (20) (21) (22) (23)
Bates, J. Plat. Surf. Finish. 1998, 85 (May), 14. Niederpru ¨ m, H. Angew. Chem. Int. Ed. 1975, 14, 614. Parker, K. Plat. Surf. Finish. 1992, 79 (March), 29. Mallory, G. O. Electroless Plating: Fundamentals and Applications; Mallory, G. O., Hajdu, J. B., Eds.; AESF: Orlando, FL, 1990; pp 1-56. Riedel, W. Electroless Nickel Plating; Finishing Publications Ltd.: Herts, U.K., 1991. Duncan, R. N. Plat. Surf. Finish. 1996, 83 (October), 64. Mallory, G. O.; Parker, K. Plat. Surf. Finish. 1994, 81 (December), 55. Richmond, J.; Top, P. Proceedings of the 11th AESF/EPA Conference on Environmental Control for the Surface Finishing Industry; AESF: Orlando, FL, 1990; pp 1-16. Bolger, P. T.; Szlag, D. C. Clean Prod. Proc. 2001, 2, 209. Vaughan, D. J. Process and Equipment for Reforming and Maintaining Electroless Nickel Baths. U.S. Patent, US5419821, 1993. Steffani, C.; Meltzer, M. Electroless Nickel Recycling via Electrodialysis; NTIS document, DE95011808, Lawrence Livermore National Laboratory: 1995; pp 1-4. Li, C. L.; Zhou, D.; Yu, X. J. Harbin Inst. Technol. (New Series) 1998, 5, 13. Stencel, N.; O’Donnell, J. Proceedings of the 4th Annual Air Force Worldwide Pollution Prevention Conference and Exhibition, San Antonio, TX, 1995; pp 433-438. Crotty, D. E. Proceedings of the 1999 Electroless Nickel Conference, Cincinnati, OH, 1999; pp 22-1-22-12. Kuboi, Y.; Takeshita, R. Proceedings of the 1989 Electroless Nickel Conference, Cincinnati, OH, 1989; pp 16-1-16-15. Williams, M.; Vogel, D. Proceedings of the 1999 AESF/EPA Conference for Environmental Excellence, Orlando, FL, 1999; pp 149-163. Sata, T. J. Membr. Sci. 1994, 93, 117. Li, C. L.; Zhao, H. X.; Tsuru, T.; Zhou, D.; Matsumura, M. J. Membr. Sci. 1999, 157, 241. Lide, D. R. CRC Handbook of Chemistry and Physics, 48th ed.; CRC: London, 1997; pp 8-42. Bolger, P. T.; Szlag, D. C. Plat. Surf. Finish. 2002, 89 (March), 52. Cowan, D. A.; Brown, J. H. Ind. Chem. Eng. 1959, 51, 1445. Davis, T. A.; Genders, J. D.; Pletcher, D. A First Course in IonPermeable Membranes; The Electrochemical Consultancy: Hants, England, 1997; p 155. Van Wazer, J. R. Phosphorus and Its Compounds; Chemistry, Interscience Publishers Inc.; New York, 1958; Vol. 1, p 360.
Received for review July 14, 2001. Revised manuscript received February 21, 2002. Accepted March 4, 2002. ES015610T
