Journal of Radioanalytical and Nuclear Chemistry, Vol. 254, No. 1 (2002) 193–200

Removal of cobalt- and mercury-EDTA chelates from aqueous solutions with a macroporous char T. W. Marrero,1 S. E. Manahan,2* J. S. Morris3 1 2

University of Missouri, Department of Chemistry, 125 Chemistry, Columbia, Missouri 65211, USA University of Missouri, Department of Chemistry, 125 Chemistry, Columbia, Missouri 65211, USA 3 University of Missouri Research Reactor, Research Park, Columbia, Missouri 65211, USA (Received March 26, 2002)

A bench-scale method was developed to remove cobalt- and mercury-EDTA chelates from water onto macroporous char. Experimental parameters included variations on solution pH, char pre-treatment, cobalt oxidation state, and apparatus configuration. The use of 60Co and 203Hg radiotracers allowed for total accountability of the metals in the char and effluents. Overall experimental results demonstrate the applicability of treated macroporous chars for the effective removal of both cobalt and mercury EDTA chelates from aqueous solutions. The char system was optimized to sequester 97.5% and 99.8% of the cobalt- and mercury-EDTA from 20 ppm solutions, respectively.

Introduction Radioactive and toxic heavy metals pose substantial problems in the nuclear and manufacturing industries. A particularly difficult type of such waste consists of hazardous heavy metals or radioactive metals, for example, cobalt and mercury, bound by strong, biodegradation resistant chelating agents, such as ethylenediaminetetraacetic acid, EDTA. EDTA chelates are stable and degradation resistant. They hold metals as anionic species that are not strongly bound by cationexchanging soil and are, therefore, mobile in groundwater. These waste streams are not easily and efficiently decontaminated by traditional treatment methods. Previous research has shown that the ChemChar hazardous waste gasification process can be used to thermally treat mixed hazardous chemical and radioactive wastes on a matrix of porous granular carbon prepared by three successive partial gasifications of subbituminous coal under conditions that enable complete retention of cobalt and mercury in the gasification system.1,2 The research reported in this paper was undertaken with the objective of developing means of removing low levels of EDTA-chelated cobalt and mercury from water onto a solid surface that can be subjected to gasification to destroy the EDTA and immobilize the metals in a poorly leachable form. The solid material upon which the chelated metals were immobilized and treated in this study consisted of an acid-washed granular char called Triple Reverse Burn char (TRB char) prepared by three successive gasifications of subbituminous coal. TRB char is a porous, stable, mechanically strong material that has been widely used for the removal, and subsequent destruction by gasification, of a variety of radioactive and hazardous chemical wastes.1–6

An advantage of TRB char for waste treatment is that it may be used as a support material for other substances that have an affinity for wastes. An example of such a substance is freshly precipitated iron(III) hydroxide, Fe(OH)3, which has the ability to coprecipitate heavy metal ions from water and to remove chelated metals from water.7 Aqueous wastes cannot be filtered over iron(III) hydroxide because its gelatinous nature will not allow liquid to flow through it at a significant rate. However, by depositing precipitated Fe(OH)3 directly into the pores of TRB char (macropores about 10 micrometers in size), a granular material is obtained that allows for free flow of liquids over it. Figure 1 shows the structure of a non-metal-laden TRB char exhibiting the macropores that make it a very effective support for the immobilization of materials used to treat wastewater. Experimental Radiotisotopes of 59Fe, 60Co, 64Cu, and 203Hg were prepared by neutron irradiation of the appropriate targets in the 10-MW reactor at the University of MissouriColumbia (MURR). The gamma-ray energies observed during analysis of these radiotracers by high-resolution gamma-ray spectroscopy are listed in Table 1.

Fig. 1. Scanning electron micrograph of triple reverse burn char * E-mail: [email protected] 0236–5731/2002/USD 17.00 © 2002 Akadémiai Kiadó, Budapest

Akadémiai Kiadó, Budapest Kluwer Academic Publishers, Dordrecht

T. W. MARRERO et al.: REMOVAL OF COBALT- AND MERCURY-EDTA CHELATES FROM AQUEOUS SOLUTIONS

Table 1. Nuclear data for target irradiations Element Fe Co Cu Hg

Target isotope

Reaction

Product isotope

(n, γ) (n, γ) (n, γ) (n, γ)

58

Fe Co 63 Cu 202 Hg 59

Half-life

59

Fe Co 64 Cu 203 Hg

Observed γ-energy, keV

44.50 d 5.27 y 12.70 h 46.61 d

60

1099 1173 511 279

Table 2. Experimental parameters

Trial

Co 1 Co 2 Co 3 Co 4 Co 5 Co 6 Co 7 Co 8 Co 9 Co 10 Co 11 Co 12 Co 13 Co 14 Co 15 Co 16 Co 17 Hg 1 Hg 2 Hg 3 Hg 4 Hg 5 Hg 6 Hg 7 Hg 8 EDTA 1 EDTA 2

Initial pH TRB Stock char solution 7.00 4.09 3.84 3.84 3.88 4.17 3.97 4.01 4.13 3.95 3.92 4.12 4.18 7.92 8.37 4.10 4.27 3.50 3.00 4.15 4.10 4.05 4.18 4.03 4.01 3.99 4.22

5.60 4.34 4.07 4.08 4.08 4.06 4.06 3.90 4.04 4.36 4.20 4.36 4.08 8.00 7.91 3.89 3.90 5.30 5.30 4.35 4.35 4.08 4.11 4.43 4.13 3.90 4.05

Segment 1 (125 ml) – 7.20 4.96 4.51 6.85 9.28 2.96 2.62 7.40 4.10 4.10 4.20 4.20 8.00 8.40 4.15 4.10 – – 4.80 4.64 4.44 4.70 4.1 4.2 3.59 6.34

Final pH Segment 2 (125 ml)

Segment 3 Column (C) (100 ml) or batch (B)

– 7.60 4.47 4.73 7.87 9.37 2.85 2.49 7.13 – – – – – – – – – – 4.50 4.60 4.18 4.37 – – 3.17 5.95

Various solutions were contacted with char material on a laboratory scale to study removal of solutes onto the char. Radioactivity was counted in (1) the TRB char through which the solutions of radionuclides were filtered, (2) the effluent of filtration, (3) the solid residue isolated from the filtrate, and (4) glass wool contained in a plug at the bottom of the filter. The seven parameters observed during the experimental trials were (1) initial pH of the TRB char, (2) initial pH of the stock solutions, (3) final pH of the effluent, (4) column or batch study, (5) loading or absence of iron(III) hydroxide on the TRB char, (6) chelated or unchelated metals, and (7) cobalt oxidation state. The various conditions employed are

194

9.00 7.80 4.29 4.77 8.13 9.36 2.80 2.43 6.86 – – – – – – – – 5.60 3.60 4.40 4.52 4.13 4.00 – – 3.06 5.52

C C C C C C C C C B B B B B B B B C C C C C C B B C C

pH adjusted

Iron loaded

EDTA

Cobalt oxidation state

No No No No No No No No No Yes Yes Yes Yes Yes Yes Yes Yes No No No No No No Yes Yes No No

Yes Yes No No Yes Yes No No Yes No No Yes Yes Yes Yes Yes Yes Yes Yes Yes No No Yes No Yes No Yes

Yes Yes Yes Yes Yes No No No No Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes No No Yes Yes Yes Yes

2 2 2 3 3 2 2 3 3 2 3 2 3 2 3 2 3

summarized in Table 2. For column studies the char solid was contained in a vertically positioned 46 cm long, 2.2 cm inner diameter Vycor (silica glass) tube over which the water containing radionuclides was passed. The batch methods employed gentle agitation to ensure contact of the stock solutions with the macroporous char. The initial pH ranged from 3.0 to 8.4 on the char surface and in the stock solutions. The pH was monitored without correction during preliminary trials; in later runs the hydrogen ion concentration was regulated throughout the experiment to maintain a constant pH in the system.

T. W. MARRERO et al.: REMOVAL OF COBALT- AND MERCURY-EDTA CHELATES FROM AQUEOUS SOLUTIONS

Preparation of iron-free TRB char adsorption columns Granular 14–20 mesh TRB char was soaked in a concentrated hydrochloric acid solution for three hours to remove iron oxides and other acid-soluble minerals. The acid-treated char was rinsed with large volumes of deionized (DI) water until the pH of the effluent was almost neutral. It was then placed in an oven to dry at 110 to 136 °C for approximately 12 hours and immediately stored in a desiccator. Preliminary experiments indicated that an initial pH 4 for the TRB chars and the stock solutions was optimal for the removal of aqueous cobalt and mercury. Therefore, the TRB char and the stock solutions were conditioned to a pH 4 prior to their mixing. Conditioning of the acid washed TRB char was conducted in the 46 cm long, 2.2 cm inner diameter Vycor column. Deionized water spiked with small aliquots of dilute nitric acid (HNO3) or dilute sodium hydroxide (NaOH) were run through the column and the pH of aliquots of the effluent were monitored potentiometrically until the desired pH was established, after which excess water was removed by blowing air through the column. Preparation of metal-EDTA stock solutions All water used for sample preparation was a high purity 18 MΩ·cm water. Stock 20-ppm chelated and unchelated cobalt(II), cobalt(III), and mercury(II) solutions were prepared in 1 liter volumetric flasks. A 1 liter EDTA solution free of chelatable metals was also prepared with a concentration of unchelated EDTA the same as that of the chelated EDTA in the 20-ppm cobaltEDTA solution. The 60Co radiotracer used was prepared from an existing radiolabeled cobalt(II) nitrate solution. A 0.05 ml aliquot of the cobalt solution containing 12 µCi 60Co was diluted to 8.0 ml with a 2.5% nitric acid solution. The 60Co solution was mixed with 10 ml of dissolved Na-EDTA. The total volume was increased to 25 ml with 5% nitric acid and excess cobalt(II) nitrate solid was added. The pH of the solution was increased to 12.8 with 1M NaOH. A blue precipitate of unchelated cobalt(II) hydroxide formed in the tube and was removed by filtration through a 0.5 µm filter. The filtrate was diluted to 500 ml. The pH of the chelated cobalt solution was adjusted to approximately 5.6 with the 5% HNO3 solution, transferred to a 1 l volumetric flask, and diluted to one liter. The cobalt(III) solutions were prepared prior to diluting the solutions to one liter by adding hydrogen peroxide oxidizer to the cobalt(II) radiotracer and carrier and gently heating. The cobalt(III) activities generated were identical to those of the cobalt(II) solutions.

The 203Hg radiotracer was prepared from a mercury(II) nitrate solid. The solid was placed in a highdensity polyethylene (HDPE) vial encased in a transport vessel and subjected to a thermal neutron flux of 8.1013 n.cm–2.s–1. A pneumatic tube system was used to deliver the solid Hg(NO3)2 to the reactor core for a predetermined irradiation period. Multiple targets were irradiated for preparation of the various mercury stock solutions required in the trials. Mercury radiotracer activity of 5 to 20 µCi was attained. The irradiated mercury solid was dissolved in a 5% HNO3 solution and transferred to a 50 ml plastic tube that contained 0.04 g EDTA solid and 5 ml of 5% HNO3. The EDTA and Hg(NO3)2 were swirled together for one minute. The pH of the Hg-EDTA solution was raised to above twelve by addition of 10 to 20 ml of a 1M NaOH. A slight excess of freshly precipitated mercury(II) oxide was added to ensure that all the EDTA present was chelated to mercury in a 1:1 ratio and the excess solid was removed by filtration. This stock solution of radiotracer mercury was mixed with one of 1:1 nonradioactive Hg:EDTA prepared by precipitating excess mercury(II) at an elevated pH from a mixture of EDTA and Hg(II) nitrate and diluted quantitatively to 1 l. Within a pH range of 3 to 12 only a negligible fraction of the mercury(II) is present as the unchelated metal in a 1:1 Hg : EDTA mixture. Preparation of iron(III) hydroxide loaded TRB char adsorption columns To measure the loading of iron on the char, radiochemical analysis was performed with a radiotracer of 59Fe, (T1/2 = 44.44 d). An iron wire was irradiated to produce 9 µCi of 59Fe. The irradiated iron wire was dissolved with 2:1 conc. HCl and conc. HNO3 in a 500 ml volumetric flask. A quantity of 143 g of iron(III) nitrate, Fe(NO3)3.9H2O, was dissolved with 5% HNO3 solution in the flask to generate upon dilution a 1 g Fe/25 ml of stock solution. For each trial, a total of 25.00±0.02 ml of the iron stock solution was adsorbed onto 39.6±0.1 g of the TRB char to produce a maximum loading of 2.6 wt.% iron on the char matrix. For precipitation of iron(III) hydroxide in the char matrix, the iron nitrate-laden TRB char was placed in a horizontally positioned Vycor tube loosely plugged with glass wool at each end. Ammonia gas was introduced from a lecture bottle or entrained in a stream of nitrogen gas bubbled through a solution of concentrated ammonia. The exothermic reaction produced by formation of iron(III) hydroxide could be followed by the visible transformation of the iron to a “rusty” precipitate and by heat generated as the conversion progressed down the char column. Once the heated area reached the end of the tube the pH of the outlet gas was

195

T. W. MARRERO et al.: REMOVAL OF COBALT- AND MERCURY-EDTA CHELATES FROM AQUEOUS SOLUTIONS

measured with moist pH indicator paper until a pH of 12 was observed in the outlet gas. At that time the ammonia gas flow was stopped and the iron-loaded char was washed exhaustively with deionized water until the effluent pH was between 3 and 7. The final washing removed any of the iron(III) hydroxide that was not sequestered on the surface of the char. Air was blown through the column of freshly made iron-loaded TRB char to remove excess water. Adsorption procedure For column studies, a column containing 39.6 g of appropriately treated char held in place in the Vycor column by a glass wool plug and 1-hole stopper was used and 350 ml of stock solution was run slowly through the column over a period of 30 minutes. The column effluent was collected in three aliquots of 125, 125, and 100 ml. Once collected, the effluent was vacuum filtered through a 0.45 µm filter paper to ensure that no residual solid remained with the solution. The filter paper was saved and counted by high-resolution gamma-ray spectroscopy. For the batch trials the solution was added to the TRB char in a 1 l beaker and gently agitated for 30 minutes. At the end of the batch trials the solutions were separated from the TRB char by filtering through a column with glass wool at the end of an empty Vycor tube. In both the column and batch trials containing ironloaded TRB char, the Fe(OH)3 would partially separate from the char during the mixing process and would pass through the glass wool filter into the collection beakers. The solid Fe(OH)3 in the collection beakers was isolated by centrifugation. The centrifugate was decanted and filtered through a 0.45 µm filter paper. The filter paper and the collected iron(III) hydroxide were soaked in dilute nitric acid and filtered again through a 5.0 µm filter paper to remove any char dust. The filtrate was placed in a 120 ml disposable plastic specimen cup and the residue was placed with the char for measurement of activity by gamma-ray spectroscopy. At the end of each trial the TRB char, effluent, glass wool, and separated Fe(OH)3 were collected for radioassay using a high-resolution gamma-ray spectrometer. Addition of free iron(III) nitrate Iron(III) nitrate solution was added to two batch trials with Co-EDTA (Co-16 and Co-17, Table 2) to determine if precipitation of dissolved iron during the adsorption process resulted in higher removals of CoEDTA. One milliliter of the iron(III) nitrate stock solution was mixed with the contents of the beaker at the beginning of the trial. The acidic iron(III) nitrate solution aided in the maintenance of the pH at 4. 196

Use of copper-64 to monitor removal of unchelated EDTA Two trials were conducted to determine the affinity of acid-washed TRB char and Fe(OH)3-laden TRB char for unchelated EDTA. A volume of 350 ml of a solution containing unchelated EDTA of a concentration that would form a 1:1 chelate with 20 ppm Co was introduced to (1) iron-free or (2) iron-loaded TRB char and the effluent was collected in a beaker. The pH of the effluent was measured but not adjusted during the trial. The effluent was collected at the outlet in a beaker and analyzed for non-retained EDTA. The EDTA in the effluent was measured by mixing the effluent with a stock solution of copper(II) radiolabeled with 64Cu, (T1/2 = 12.7 h), precipitating excess copper at a pH adjusted to above 10 and measuring the copper remaining in solution as a 1:1 chelate with EDTA (KCu(EDTA) = 6.3.1018).8 A known volume of the filtrate from precipitation of excess copper was pipetted into a 250 ml polyethylene bottle and the copper activity remaining in the filtrate was measured by high-resolution gamma-ray spectroscopy. The amount of EDTA retained on the macroporous char was calculated by difference. Preparation of samples and standards Four types of geometries were used for counting the radiotracers. (1) All of the solid char was counted in a specimen cup. The sample masses were relatively the same; consequently, the height of the char in the specimen cup was constant, ±1 mm, for all experiments. Both the char samples and char standards were mixed with a stir rod in an attempt to homogenize the metal and radiotracer in the vessel. (2) The effluents from the column and batch trials, and the dissolved iron(III) hydroxide, were counted in a specimen cup. Using gravimetric analysis, a correction factor for the total sample mass of the solutions and the fraction of the mass used for counting were calculated in order to account for the total activity of each sample. Each liquid sample and standard was brought to a constant height to achieve equivalent geometries for comparative purposes. The column effluents were counted in three aliquots of 125, 125, and 100 ml. The batch trials used one fraction of the whole sample for counting. (3) The glass wool was placed at the bottom of a specimen cup. To ensure reproducibility, the glass wool was pressed to the bottom with plastic filler. (4) In the column trials the residue collected on the filter paper was folded neatly into a 0.5 ml polyvial. All of the samples were compared to standards made from the same stock solution used in their trial. There

T. W. MARRERO et al.: REMOVAL OF COBALT- AND MERCURY-EDTA CHELATES FROM AQUEOUS SOLUTIONS

were two types of char standards: (1) a TRB char that was loaded with 25.00±0.02 ml of the stock solution and thoroughly mixed to make it as homogeneous as possible and (2) a liquid standard (25.00±0.02 ml of stock solution) made in the same geometry as the char and comprising a perfectly homogeneous standard. A correction factor was used for the char standard made with water because of the difference in self-absorption of the gamma rays relative to the char matrix.9 For the liquids counted in the specimen cup, a known volume of the stock solution was pipetted into the specimen cup and diluted to the correct height with water. The glass wool standard was made by pipetting a known volume of the stock solution onto the filter paper at the same height as the glass wool. The solid residues collected in the 0.5 ml vials were compared to a standard made from a 0.4 ml aliquot of stock solution. Char homogeneity One possible source of significant experimental error is the counting of the char samples. It is not obvious by visual inspection whether or not the char is homogeneous. If it is not mixed well in the counting vessel, a significant experimental error can potentially occur. This is an especially important consideration for the column trials. For example, more of the radiotracer may be absorbed at the top of the column compared to the bottom. If the char is poured directly into the specimen cup, it will have a higher concentration of radiotracer at the top and less at the bottom. This stratification problem will drastically affect the observed count rate. Potential errors are exacerbated if the char standard is also inhomogeneous. To account for these problems, the standards made from TRB char and water mixed with 25.00±0.02 ml of stock solution were compared to the char samples. The char samples and both types of standards were counted in the normal and inverted positions. A comparison was made between the ratios of the decay-corrected counts of the two standards to the sample. The standard with the normal:inverted ratio closest to the sample value was used for comparative analysis. Gamma-ray spectroscopy The samples and standards were analyzed by a high purity germanium (HPGe) detector, that is part of the MURR Canberra VMS high-resolution gamma-ray spectroscopy network. The relatively long half-lives of the 59Fe, 60Co and 203Hg made these radioisotopes excellent candidates for the HPGe analysis allowing for the use of long count times. These three radionuclides

are completely resolved via high-resolution gamma-ray spectroscopy. Counting position and measurement time were manipulated to obtain a counting statistical error of 0.5% or less in most cases. Although 64Cu has a relatively shorter half-life of 12.7 hours, the procedure was completed in less than one 64Cu half-life. Consequently, similarly low (approximately 0.5%) counting statistical errors were also achieved for 64Cu. Accuracy and precision in the determination of the activity balance Every attempt was made to account for the movement of the radiotracers within the system. Considering the cumulative experimental error, the accountability was 100% for all samples. The use of the most representative standards for each individual sample produced counting geometry errors of less than one percent for the solutions and 0.04% to 7.7% for the char samples (Table 3). Results The results of the column and batch equilibration studies are summarized in Table 3. The trial designations referred to in this table are given in Table 2. Char homogeneity The range of percent difference between the ratios of normal:inverted counting of the char samples to the char standards was 3.1–6.4%. Neither the char nor water standard was clearly superior in more closely representing the cobalt or mercury distribution in the char samples. Similar observations of the difference between the char standards made from water and TRB char were evident with the iron analysis (0.7–7.5%). The ironloaded char did seem to be better represented by the char standard made from water more frequently than the ones made from TRB char. Accuracy and precision in the determination of the activity balance All of the cobalt challenged to the char system was accounted for by radiotracer analysis. Table 3 presents the geometrical error associated with the measurement of cobalt in the char samples. The mean total accountability for the 60Co and 203Hg in the char system was 102.7±5.5% and 106.2±7.5%, respectively.

197

T. W. MARRERO et al.: REMOVAL OF COBALT- AND MERCURY-EDTA CHELATES FROM AQUEOUS SOLUTIONS Table 3. Accountability of 60Co and 203Hg radioisotopes and EDTA 60

Trial Co 1 Co 2 Co 3 Co 4 Co 5 Co 6 Co 7 Co 8 Co 9 Co 10 Co 11 Hg 1 Hg 2 Hg 3 Hg 4 Hg 5 Hg 6 Hg 7

Char 42.0 92.5 28.2 47.9 69.0 108.1 4.6 4.7 77.1 57.9 85.1 89.8 101.1 111.2 74.3 103.2 97.9 118.8

Trial

Char

Co 12 Co 13 Co 14 Co 15 Co 16 Co 17 Hg 8

40.2 47.9 41.8 32.3 51.6 61.4 94.8

EDTA 1c EDTA 2c

77.7 96.7

Error of char to standard, %

Co or 203Hg activity per section, % Effluent or Glass wool filtrate

Residue

Totala,b

1.9 0.5 0.9 2.2 5.8 6.4 2.6 0.1 1.1 1.1 0.7 3.5 4.0 4.8 2.7 6.4 7.7 1.7

56.0 2.5 74.6 53.4 27.8 0.02 93.6 87.8 16.6 43.9 18.1 2.6 0.1 4.4 25.6 0.1 0.4 0.2

0.66 0.29 0.61 0.59 1.56 0.23 0.47 0.56 0.70 N/A N/A 0.41 0.41 0.05 0.29 0.01 0.02 N/A

0.18 0.02 0.03 0.01 0.05 0.16 0.00 0.00 2.9 N/A N/A 10.64 0.01 0.00 0.00 0.00 0.00 N/A

Error of char to standard, %

Effluent or filtrate

Metal in dissolved “freed” iron

Residue

61.7 48.8 56.5 56.8 33.6 22.4 6.4

2.8 8.6 12.6 14.5 28.7 23.8 7.14

N/A N/A N/A N/A N/A N/A N/A

104.7 105.3 110.9 103.6 113.9 107.6 108.3

22.2 3.3

N/A N/A

N/A N/A

100 100

3.1 2.9 0.04 1.6 0.6 1.1 3.1 N/A N/A

98.9 95.3 103.5 101.9 98.5 108.5 98.6 93.1 97.4 101.8 103.2 103.5 101.6 115.6 100.2 103.3 98.4 119.0 Total,ab

a

Mean = 102.7±5.5%. Mean = 106.2±7.5%. c EDTA 1 and EDTA 2 accounts for the unchelated EDTA using the 64Cu analysis. The 64Cu analysis assumes that if the EDTA is not in the filtrate, it is in the char. b

Possible errors include the replication of the geometry of the glass wool and the effluents. Because they are homogeneous solutions, the effluents match their comparison standards better than char; hence the errors in the effluent measurements are lower. The solid iron hydroxide was centrifuged, filtered and dissolved in nitric acid minimizing the error in quantifying 59Fe. As seen in Table 3, a significant portion of the cobalt was captured in the freed iron hydroxide solid. This cobalt can be considered removed from the aqueous solution since it was collected by a solid. Radiotracer analysis Table 2 is a comprehensive summary of the conditions for all of the trials and Table 3 is the calculated data for the final results. The 60Co and 203Hg removed by the system varied over a range of 6.4–100.0% and 74.4–99.9%, respectively. The

198

optimum conditions for the removal of EDTA-chelates are coincidentally in Trial 2 for each metal (Table 3). Trial “EDTA 2” represents the removal of unchelated EDTA on a char column loaded with iron(III) hydroxide. Analysis of the effluent by the measurement of copper(II) remaining in solution spiked with copper after adjustment to pH 10 indicated that 97% of the EDTA was absorbed onto the column. Discussion Unchelated and chelated metals were removed to varying degrees. In a basic environment the EDTA will strongly chelate with iron(III), cobalt(II) and mercury(II). In an acidic environment the complexation of the metal-EDTA chelate is weaker, but still significant. The cobalt(III)-EDTA chelate is considered strong in both acidic and basic solutions.10 Iron(III) is strongly chelated by the EDTA at acidic pH values.

T. W. MARRERO et al.: REMOVAL OF COBALT- AND MERCURY-EDTA CHELATES FROM AQUEOUS SOLUTIONS

Removals of unchelated cobalt(II), cobalt(III), and mercury(II) by char and by iron-loaded char are shown in Table 3 (Co 6-9, Hg 5, Hg 6). Unchelated Co(II) and Co(III) are removed much more effectively by ironloaded char than by acid-washed char. Both acid-washed TRB char and the iron-loaded TRB char used removed essentially all unchelated mercury as shown in Hg trials 5 and 6. The highest removal efficiency of chelated Hg was 99.8%, using iron-loaded TRB char in a column. The highest removal efficiency of the cobalt(II)-EDTA was 97.5 %. The cobalt(III) EDTA chelates showed a lower affinity for the solid sorbent with a maximum of 81.9% removed from solution. Cobalt(III)-EDTA can be distinguished from the cobalt(II)-EDTA chelate by its more optically dense pink color. The intensification of the pink color made it apparent that an oxidation did occur with the chelated cobalt(II) species. Additional qualitative experiments were conducted to determine if the char, iron, or pH change were primarily responsible for the oxidation of the cobalt. The final analysis determined that the acidwashed TRB char provided the environment necessary for cobalt to be oxidized. The presence of iron and/or the change of pH did not increase the intensity of color due to dissolved cobalt chelate that would indicate the transition of the cobalt(II)- to the cobalt(III)-EDTA chelate. It was also observed that the systems that started at a low pH and were allowed to change without intervention showed approximately the same metal removals as those in which the pH was held constant during equilibration. The batch trials with an adjusted pH had a mean cobalt and mercury removal of 57.3±16.4% and 96.7±4.4%, respectively. Similarly, the column trials where the pH was not adjusted to maintain a constant pH had a mean cobalt and mercury removal of 55.4±37.9% and 94.5± 10.0%, respectively. A trial was conducted to determine the interaction of unchelated EDTA with the TRB char without influence of chelated metals in solution. Although the acid-washed TRB char did show a reasonably high affinity for the EDTA (77.8%), the ironloaded char was even more effective for unchelated EDTA removal (96.7%). The iron-loaded char did cause measurable increases in pH over the equilibration period of 30 minutes. This could be due to ammonia used to precipitate the Fe(OH)3 being released from the solids or to binding of H+ ion from solution by the freshly precipitated Fe(OH)3 surface. Another possible cause for the increase in pH in the cobalt systems is the oxidation of cobalt from the +2 to the +3 oxidation states to produce hydroxide ions (OH–): Co2+ → Co3+ + e– (1) 4e– + 2H+ + O2 → 2OH–

The iron-loaded char had a more pronounced pH increase over time than the acid-washed TRB char. This is plausibly due to residual basicity of the freshly made iron hydroxide char. In cases where cobalt(II) was radioanalytically proven to be substantially removed from the aqueous solution, the end color was a more intense pink than the initial stock solution of cobalt(II). This result is indicative of oxidation of cobalt(II) to cobalt(III). In all of the trials where iron was present on the char, the iron did not dissolve into the aqueous phase. In all cases iron-laden char removed more cobalt than iron-free char. Table 3 shows the importance of the iron on the char surface. The “dissolved freed iron” refers to the iron(III) hydroxide that was originally on the char at the initiation of the trial but which was sloughed off the char surface during contact with the cobalt solutions. The percent of cobalt found in the freed iron ranged from 2.8% to 28.7% for similar loadings of iron. The percent of mercury found in this freed iron was 7.1%. Conclusions The use of a macroporous char for the removal of cobalt- or mercury-EDTA chelates is a potentially useful technique. Under the parameters of this study and through the analytical use of the radiotracers, the optimal conditions for the removal of cobalt-EDTA chelates by TRB char occurs when the cobalt is in the 2 oxidation state and a 2 wt.% iron-loading utilizing a column apparatus with the pH of the char and chelate solution equal to 4.0 prior to mixing. The optimal conditions for the removal of Hg-EDTA chelates (99.8%) is by using iron-loaded TRB char in a column apparatus with the pH of the char and chelate solution equal to 4.0 prior to mixing. Adjustment of the pH over time is not required for removal of either metal. An added benefit to the method is that the metal absorbed on the char can be subjected to a reductive thermal gasification treatment process that will retain the cobalt and mercury in a non-leachable form suitable for volume reduction while converting the organics to environmentally friendly gases.3,9,11 * Professor R. Kent MURMANN, University of Missouri-Columbia, Department of Chemistry, is appreciated for his help in the review of cobalt oxidation states and EDTA chelation and Louis M. ROSS Jr., University of Missouri-Columbia, Department of Geological Sciences, for the use of the scanning electron microscope. The Department of Energy Radiochemistry Education Advancement Program, DE-FC0993SR18262 (Coop #135), supported the research.

(2)

199

T. W. MARRERO et al.: REMOVAL OF COBALT- AND MERCURY-EDTA CHELATES FROM AQUEOUS SOLUTIONS

References 1. R. S. MARTIN, S. E. MANAHAN, J. S. MORRIS, Chemosphere, 37 (1998) 531. 2. R. S. MARTIN, K. E. GARRISON, S. E. MANAHAN, J. S. MORRIS, D. W. LARSEN, J. Environ. Sci. Health, A34 (1999) 767. 3. D. W. LARSEN, S. E. MANAHAN, Process for Treatment of Hazardous Waste by Reverse Burn Gasification, 1990, U.S. Patent No. 4,987,477. 4. J. W. COLBERT, R. S. MARTIN, S. E. MANAHAN, J. S. MORRIS, J. Radioanal. Nucl. Chem., 230 (1998) 61. 5. R. S. MARTIN, S. E. MANAHAN, J. S. MORRIS, D. W. LARSEN, J. Environ. Sci. Health, Part A: Toxic/Hazard. Subst. Environ. Eng., A34 (1999) 241. 6. B. D. MEDCALF, D. W. LARSEN, S. E. MANAHAN, Waste Managem., 18 (1998) 197.

200

7. A. Y. NIKIFOROV, N. V. KADNIKOVA, I. A. NIKIFOROV, L. A. IL’INA, Colloid J., 62 (2000) 457. 8. D. A. SKOOG, D. M. WEST, F. J. HOLLER, Complex-Formation Titrations in Fundamentals of Analytical Chemistry, 7th ed., Chapter 14, Saunders College Publishing, 1997, p. 278. 9. T. W. MARRERO, Dissertation, University of Missouri, Columbia, MO, 2002. 10. R. J. SERNE, A. R. FELMY, K. J. CANTRELL, K. M. KRUPKA, J. A. CAMPBELL, H. BOLTON, J. K. FREDRICKSON, Characterization of Radionuclide-Chelating Agent Complexes Found in Low-Level Radioactive Decontamination Waste; Division of Regulatory Applications, Office of Nuclear Regulatory Research, U.S. Nuclear Regulatory Commission, Washington D.C., 1996, p. 31; 109. 11. R. S. MARTIN, Dissertation, University of Missouri, Columbia, MO, 1999.