Environ. Sci. Technol. 1993,27, 350-356
Eitzer, B. D.; Hites, R. A. Environ. Sci. Technol. 1989,23, 1389-95. Nakano, T.; Tsuji, M.; Okuno, T. Atmos. Environ. 1990, 24A, 1361-8. Eitzer, B. D.; Hites, R. A. Environ. Sci. Technol. 1989,23, 1396-401. Bidleman, T. Environ. Sci. Technol. 1988, 22, 361-7. Travis, C. C.; Hattemer-Frey, H. A. Sci. Total. Environ. 1991,104, 97-127. Crosby, D. G.; Wong, A. S. Science 1977, 195, 1337-8. Bacci, E.; Cerejeira, M. J.; Gaggi, C.; Chemello, G.; Calamari, D.; Vighi, M. Bull. Environ. Contam. Toxicol. 1992,48,8-17.
(19) S k u , W. Y.; Doucette, W.; Gobas, F.; Andren, A.; Mackay, D. Environ. Sci. Technol. 1988, 22, 651-8. Received for review J u l y 13, 1992. Revised manuscript received October 15,1992. Accepted October 26,1992. T h e information in this document has been funded wholly (or in part) by the U S , Environmental Protection Agency. I t has been subjected to the Agency’s peer and administrative review, and it has been approved for publication as a n E P A document. Mention of trade names or commercial products does not constitute endorsement or recommendation for use.
Reversible Photoreductive Deposition and Oxidative Dissolution of Copper Ions in Titanium Dioxide Aqueous Suspensions Nancy S. Foster, Rlchard D. Noble,+and Carl A. Koval*
Department of Chemistry and Biochemistry and the Cooperative Institute for Research in Environmental Science (CIRES), University of Colorado, Campus Box 215, Boulder, Colorado 80309-0215. Particulate titanium dioxide was used to remove and concentrate Cu(I1) ions in aqueous solutions through a cyclic process of photodeposition, separation, and oxidation. Illuminated, nitrogen-purged solutions containing copper sulfate, excess sodium formate (pH 3.6), and titanium dioxide formed a purple Cu-TiOz species. Cu(I1) concentrations in the supernatant were driven from 51 to 10.018 pg/mL. Upon purging with oxygen, this purple Cu-Ti02 system reverted back to white along with a corresponding increase in the Cu(I1) supernatant concentration. The photodeposition step and the air oxidation step were utilized to demonstrate a volume reduction process. Eighty-sixpercent of the Cu(I1) in a synthetic waste stream was concentrated to an organic-free solution having 7 % of the initial volume. The remaining waste solution contained only 1% of the initial Cu(II), and in a subsequent step, the remaining formate ion was destroyed using conventional TiOz photocatalytic oxidation. The overall process demonstrated the ability to separate copper ions from organics using only light and air. The reversible photoreduction deposition of Cu(I1) from solution was observed in the pH range 1.84-6.60. The reversible photoreduction deposition of copper(I1)was dependent on the organic used to scavenge holes and independent of the copper salt used.
Introduction Overview. Metals are discharged into the air, water, and soil from natural sources such as continental dust, sea spray, biological activity, forest fires, and volcanic eruptions. In the last century, however, increased industrialization through mining, smelting, metal refining, production of metallic products, and burning of fossil fuels has unlocked vast quantities of metals such as mercury, cadmium, lead, zinc, silver, and tin from minerals, ores, and bedrock into the environment. Metals are nondegradable, having infinite lifetimes, building up concentrations in food chains to toxic levels, and putting a heavy burden on ecosystems. Besides the metal pollution toxicity problem, there is a limited resource of metals in minerals and ore reserves. Through metal disposal and recovery, environmental metal pollution can be prevented and metal resources can be preserved. Copper is used in the electrical industry, construction, industrial machinery, transportation equipment, military Department of Chemical Engineering, Campus Box 424. 350
Envlron. Sci. Technol., Vol. 27, No. 2, 1993
supplies, and electroplating (1, 2). In 1987, the world production of primary (new) copper through mining was approximately 8.4 Mt (megaton) (1-3). The world reserve base (resourcesthat are economicallyviable) was estimated to be 570 Mt (1). The Minerals Handbook 1988-89 (3) estimated the static reserve life of copper to be 40 years. Obviously, recycling copper is important. The United States used 1.2 Mt of copper scrap in 1987; 41% of the totalU.S. copper consumption (1). Recycling scrap copper only requires 3-40% of the energy that is required to extract pure copper from ore (1). Besides saving energy when copper scrap is recycled, mining waste streams from the smelting and refining processes, which can pollute the environment (2),are also prevented. Current methods of removal and disposal of metals, including copper, include diafiltration, adsorption on activated carbon, precipitation, ion exchange, and encapsulation ( 4 ) . These methods are used as pretreatments to concentrate and fix metals before solidification. Before metals can be recycled from waste streams, the process must become economicallyfeasible. One way to lower costs is to use a preconcentration step before, for example, electrochemical reduction. Use of illuminated semiconductor particles in waste streams to remove and/or coricentrate copper offers advantages such as lower cost and higher efficiency than current methods. The oxidative recovery method described within this paper is limited to copper due to the photodeposition of most other metals such as gold and platinum involving photoreduction to the stable metallic state. Semiconductor Background. Since 1972, when Fujishima and Honda (5) discovered that water could be decomposed into oxygen and hydrogen by illuminating titanium dioxide, semiconductor electrochemistry has been studied for a variety of processes related to solar energy utilization (6-9). Semiconductors have been used in the photodeposition of metals (10-13) such as gold, silver, platinum, palladium, rhodium, mercury, lead, manganese, thallium, and cobalt from aqueous solutions. Although numerous examples of metal photodeposition on semiconductors exist, the extent of removal and potential use under various solution conditions has not been thoroughly studied. Particulate semiconductors are often used in photoelectrochemical experiments due to their low cost and large surface areas (7-50 m2/g for TiOJ. Illumination of titanium dioxide with a photon energy greater than the band
0013-936X/93/0927-0350$04.00/0
0 1993 American Chemical Society
gap (3.23 eV = 384 nm for pure anatase) (14) excites an electron from the valence band to the conduction band leaving a positively charged hole in the valence band. The conduction band edge (--0.20 eV vs NHE at pH 0) represents the reducing power of the photogenerated electron, and the valence band edge (- 3.0 eV vs NHE) represents the oxidizing power of the photogenerated hole. Depending on solution conditions, the electron may reduce protons, water, dioxygen, or metal ions. In a semiconductor particle, both the hole and the electron must be consumed to maintain neutrality. The hole may oxidize water to oxygen, oxidize water to hydroxyl radicals that in turn oxidize organics, or oxidize organics directly (15). Theoretically, a solution species with a standard reduction potential positive of the conduction band edge and negative of the valence band edge can be reduced or oxidized by the electron or hole, respectively. The photodeposition of most metals involves the reduction of metal ions by the conduction band electrons and the oxidation of water to molecular oxygen by the valence band holes (11). Since many waste streams contain organics in addition to metals, the valence band holes could be used to oxidize organics instead. There are numerous examples of illuminated Ti02used as a photocatalyst in the decomposition of a variety of organic compounds (10, 15, 16). Reiche et al. (17)studied the photoreduction of Cu(I1) at TiO2 They observed the full reduction of Cu(I1) to copper metal in aqueous solutions containing either acetate or no organic. Bideau et al. (18) studied the kinetics of the oxidation of formate in the presence of Cu(I1) ions. They observed the formation of a red Cu-Ti02 species when solutions containing Cu(I1) and formate were illuminated in the presence of Ti02 In this paper, we report the rapid reoxidation of photoreduced Cu(I1) ion with dioxygen and demonstrate how a cyclic photoredox process can be used to treat mixed copper ion/organic waste streams, such as electroless copper plating bath solutions. The effects of pH and organic hole scavenger on this process are discussed in detail. Experimental Section Materials. The titanium dioxide used was Degussa P-25, which was mostly anatase and had a BET (Brunauer-Emmett-Teller) surface area of 50 m2/g and an average particle diameter of 30 nm (19). All other chemicals were reagent grade. Photoreactors. Two types of photoreactors were used. Photoreactor I consisted of a sealed 150-mL Erlenmeyer flask illuminated from the side by two black lights. The reactor solution was stirred using a magnetic stir plate. This reactor was not purged with nitrogen and became oxygen starved after all of the oxygen was photoreduced. Photoreactor I1 was an annular photoreactor (Figure 1). The photoreactor consisted of two concentric borosilicate glass tubes 39 cm long and separated by -0.4 cm. A coiled glass rod was inserted into the photoreactor to provide turbulent flow. The total volume of the reactor including the connecting pieces was 150 mL. The extra reaction solution was contained in a separatory funnel. A nitrogen purge line was placed into the reservoir. The solution was recirculated through the photoreactor by a Masterflex Model peristaltic pump. The peristaltic pump had a variable capacity up to 8 cm3/s which provided a linear velocity inside the photoreactor of up to 2.5 cm/s. A 15-W (41 cm X 2.54 cm) Sylvania black light (F15T8/BLB) (mostly 365 nm) was placed in the center of the annular reactor. The photon flux of the photoreactor was determined by potassium ferrioxalate chemical actinometry (20-22). The
Black Light Bulb (15W) 365 nm
Purge Line
Photoreactor
Flgure 1. Schematic of photoreactor 11. Solution was recirculated through an annular reactor by a peristattlc pump. A black light (mostly 365 nm) was placed in the center of the annular reactor.
photon flux was 1.67 X 1Ol8 photons 300-388 nm/s (2.9 mW /cm2). Sampling, In the copper photodeposition experiments, all samples were taken from the reservoir with a syringe while exposure to oxygen was minimized. The samples were then immediately filtered using 0.2-pm nylon syringe filters. Samples were stored in plastic with a few drops of HC1 to minimize adsorption of copper ions to the container. In the formate oxidation experiment, samples were diluted with a saturated calcium carbonate solution to precipitate any carbon dioxide that may have formed that would have contributed to the total organic carbon reading. The samples were then filtered using 0.2-pm nylon syringe filters. Volume Reduction Experiment. The 2500 mL of Cu(I1) waste was processed using 500-mL portions at a time, reusing the Ti02 each time. Portions of the waste solution were placed in photoreactor 11. The photoreactor solution reservoir was a 500-mL flask which was stirred magnetically. Fresh TiO, (3 g) was placed in the reactor with the first portion. The solution was continuously purged with nitrogen and illuminated for 10 min. The reactor was drained under a flow of nitrogen into the reservoir flask. The flask was placed into a nitrogenpurged glovebag. A sample was taken and filtered. The purple TiO, was allowed to settle, and as much of the supernatant as possible was decanted through filter paper. The remaining solution was centrifuged and the supernatant decanted. To establish a concentrate solution for the fnst portion, 55 mL of 0.01 M H,S04 solution was used. The concentrate was added to the purple Ti02, and the solution was stirred and purged with compressed air to oxidize the Cu-TiO, species to Cu(I1) (white TiO,). The concentrate containing the white TiOzwas centrifuged, and the supernatant became the concentrate solution used for the next four portions. The Ti02was recovered from the centrifuge tubes using the next waste portion. The procedure was repeated four more times each using 500 mL of the waste copper solution and reusing the same TiOz each time. After the fifth portion, the TiO, was washed with 1% (v/v) nitric acid solution and dried under vacuum. At the end of the volume reduction experiment, the photoreactor and the TiO, were each washed with 100 mL of 1% (v/v) nitric acid solution to retrieve any copper or TiO,. Analyses. Copper concentrations were determined by atomic absorbance (AA) using a Perkin-Elmer 360 atomic absorbance spectrophotometer at 324.7 nm. The total Envlron. Scl. Technol., Vol. 27, No. 2, 1993 351
Conduction B a p c Z H + + Z ~ ;Zr . H~
g5
2oL
Ol
'0:
0
-
H-~OH e
2 b20
3 .
0
o
o
light, nitrogen purge light off, air purge
Valence Band
0
2
4
6
8
1 0 1 2
14
PH
Figure 2. Conduction band and valence band edges for TIO, and standard reductlon potentlals for water and copper species as a function of pH.
organic concentrations (TOC) were determined using an 0.1. analytical total organic analyzer. The pH of the solutions was determined using an Orion Research Model 701Aldigital ionalyzer. X-ray photoelectron (XPS) spectra were recorded using a Leybold-Heraeus Series 10 ESCA module with a Mg K a source at 13 keV and an emission current of 20 mA (power 260 W). The turbomolecular-pumped UHV chamber had a base pressure of 7 X Pa. The binding energies of the measured spectra were charge referenced using the C 1s binding energy of 284.6 eV. Electron paramagnetic resonance (EPR) spectra were measured on a Bruker ESP 300 at 9.45 GHz. Microwave frequencies were measured with a Hewlett-Packard frequency counter. Under a nitrogen atmosphere, vacuumdried samples were placed into 4-mm-diameter quartz EPR tubes. Results and Discussion Thermodynamic Considerations. Figure 2 shows the relative standard reduction potentials of relevant solution species (23) with respect to the conduction band and valence band edges of TiOz as a function of pH (24). This figure can be used to estimate solution conditions amenable to photoreduction of copper ions. At all pHs, reduction of oxygen is highly favorable. Gerischer and Heller (25) have predicted that O2 reduction will limit observed photooxidation of organics, but this conclusion has been questioned by Kormann et al. (26). If oxygen is present, it can be reduced preferentially with respect to copper ions; therefore, solutions in this study were purged with nitrogen or sealed to eliminate oxygen reduction as the conduction band process. Photoreduction of protons or water is barely favorable and was found to prevent reduction of copper ions only at pH