- Email: [email protected]
oxine solvent extraction method for the preconcentration of trace metals from seawater using metal exchange back-extraction
ELSEVIER
Marine Chemistry
55 (1996) 381-388
Improved dithiocarbamate/oxine solvent extraction method for the preconcentration of trace metals from seawater using metal exchange back-extraction Grant J. Batterham School of Mathematical
and Physical
Sciences, Northern
Received 6 October
*,
David L. Parry Territory
1995; accepted
Unicvrsity
Darwin,
0909 NT, Australia
10 July 1996
Abstract A rapid single extraction procedure using dithiocarbamate complexing agent, a di-isobutyl ketone (DIBK) organic phase and Hg exchange back-extraction is described for the simultaneous quantitative preconcentration of Cd, Co, Cu, Fe, Ni, Pb and Zn in seawater. An 8-hydroxyquinoline (oxine) extraction technique is also presented for the subsequent determination of Mn using the same reagents and operating system. The proposed method gave quantitative spike recoveries for all metals and was in excellent agreement with the certified results for CASS-3 marine reference seawater. Blanks and detection limits are in the low ppt range. The method gives a preconcentration factor of 72 for an SO-ml sample, facilitating the analysis of pristine seawater samples. Kqvwords:
heavy metals; seawater;
extraction;
preconcentration;
dithiocarbamate
1. Introduction Solvent extraction procedures are widely used for the determination of trace metals in seawater. The earliest standard methods were based upon metal complexation with ammonium pyrrolidine dithiocarbamate (APDC) and extraction into methyl isobutyl ketone (MIBK) (Brooks et al., 1967; Jan and Young, 1978). Kinrade and Van Loon (1974) examined the APDC-MIBK system in some detail and added diethylammonium diethyldithiocarbamate in combination with APDC to improve complex stability and broaden the effective working pH range.
_ Corresponding 0304-4203/96/$15.00 P/I
author. Copyright
SO304-4203(96)00068-O
The direct analysis of MIBK extracts by flame atomic absorption spectrometry (FAAS) was limited by the poor stability of dithiocarbamate (DTC) complexes in MIBK (Brooks et al., 1967; Kinrade and Van Loon, 1974). Danielsson et al. (1978) and Jan and Young (1978) overcame this by developing acid back-extractions to transfer DTC complexed metals into a stable acidified aqueous phase. Magnusson and Westerlund (198 1) used acid back-extraction efficiencies to show DTC complex stability was greatest in chloroform, followed by MIBK and freon TF. Graphite furnace atomic absorption spectrometry (GFAAS), with its improved sensitivity, replaced FAA& and MIBK was no longer used as a solvent due to its high water miscibility which resulted in
0 1996 Elsevier Science B.V. All rights reserved.
382
G.J. Batterham,
D.L. Parry/Marine
salt carry-over and matrix interference (Magnusson and Westerlund, 1981). Freon TF was subsequently used as the solvent of choice, but dual extractions were required for the quantitative recovery of most common trace metals (Danielsson et al., 1982; Statham, 1985). Metal exchange back-extraction was developed by Lo et al. (1982), using Hg to overcome the slow kinetics of acid back-extraction. This method required a single 20-min extraction into chloroform to preconcentrate (> 90%) Cd, Co, Cu, Fe, Ni, Pb and Zn. Improvements to this method were recently made by Sachsenberg et al. (1992) who only required a 90-s single extraction into di-isobutyl ketone (DIBK), with Pd back-extraction, to quantitatively determine Pb, Cu, Ni and Cd in seawater. This technique produced high preconcentration factors from relatively small sample volumes, but was limited by the range of metals determined. Manganese, an important element in geochemical and speciation studies, has a low affinity for DTC ligands (Shen et al., 1980), necessitating high DTC concentrations and dual extractions to achieve quantitaive preconcentration from seawater (Statham, 1985). Previous studies have also used &hydroxyquinoline (oxine) as a complexing agent to extract Mn into chloroform (Klinkhammer, 1980) or MIBK (Sturgeon et al., 1980). Sachsenberg et al. (1992) demonstrated their rapid DTC/DIBK solvent extraction method for Cd, Cu, Ni and Pb and did not evaluate Co, Fe, Mn and Zn which have much poorer recoveries with DTC solvent extraction systems (Stary and Kratzer, 1968; Wyttenbach and Bajo, 1975; Bajo and Wyttenbach, 1979; Shen et al., 1980). Di-isobutyl ketone has been found to be an efficient solvent for the extraction of a wide range of DTC-metal chelates from non-saline waters (Bone and Hibbert, 1979). The purpose of our study was to develop a rapid and reliable single extraction method with metal exchange back-extraction that allows the quantitative determination of Cd, Co, Cu, Fe, Mn, Ni, Pb and Zn in seawater. Due to the low extractability of Mn with DTC, we also examined the use of oxine, independently and in combination with DTC, for extraction into DIBK and its compatibility with metal exchange back-extraction.
Chemistry
55
C1996)
381-388
2. Experimental
2.1. Chemicals Acids and ammonia (25%) were of Suprapur/Aristar/Ultrex quality. Complexing and back-extracting agents were of analytical grade. Technical grade 2,6-dimethyl-heptan-4-one (DIBK), Merck (75% DIBK, 25% 2,4-dimethyl-heptan-6-one) or Ajax (95%), was used as purchased, with the latter preferred due to lower metal impurities. Ultrapure water used to make up all solutions was obtained from a Permutit HI-PURE water system fed with water from a reverse osmosis system. Dithiocarbamate complexing solution, 0.5% each APDC/sodium diethyldithiocarbamate (NaDDC), was made up immediately prior to use in IO-ml volumetric flasks with pH 2.5 (HNO,) ultrapure water. The solutions were purified by extraction with DIBK (1 ml, I-min shake). Stock oxine solution (2%) was prepared by dissolving 10 g of oxine in 500 ml of 0.5% HCl (v/v> and purified with DIBK (2 X 5 ml). Stock ammonia (3 M)/acetic acid (2 M) buffer was made up weekly and did not require purification. Stock 1000 ppm Hg back extracting solution was prepared from AR grade Hg(NO,), in 0.02 M HNO, and diluted to 100 ppm with 0.02 M HNO, as required. 2.2. Seawater Seawater was collected directly from Stokes Hill Wharf in Darwin Harbour, Australia, using a battery-operated polystyrene, submersible, in-line pump (Niagara LVM 114) fitted to a weighted polyvinylchloride (PVC) pole. The pump was run for several minutes prior to the attachment of a Gelman groundwater sampling capsule for in-line filtration (0.45 pm). The filter capsule was flushed with seawater for several minutes prior to sample collection. Filtered seawater was collected in an acid-washed (10% HNO,) 20-l polyethylene carboy and acidified to 0.02 M HNO, (Ultrex). The seawater had a salinity of 35%0 and pH of 8.2. CASS-3 Nearshore Standard Reference Material was obtained from the National Research Council of Canada to validate methods.
G.J. Batterham, D.L. Parry/Marine
Low-density polypropylene bottles (Nalgene) reserved for regular seawater sampling were prepared by soaking in detergent (1% Decon 90) for 24 h, followed by 3 rinses in ultrapure water and soaking in 10% HNO,/ultrapure water for at least 48 h. The bottles were again rinsed 3 times in ultrapure water, dried in a clean oven, and sealed in polyethylene plastic bags. The bottles were thoroughly rinsed with filtered seawater immediately prior to sample collection. Samples were placed on ice and acidified to 0.02 M HNO, with ultrapure acid in a clean environment. 2.3. Extraction
and preconcentration
procedure
The extraction procedure was performed in a Class-100 laminar flow cabinet. To limit exposure and potential contamination, caps were only removed from separating funnels/centrifuge tubes for the introduction or removal of sample/reagents. All labware was acid washed in the manner described for the sampling bottles. In addition, all acid-washed polyethylene pipette tips were again doubly rinsed in a nitric acid bath (5%) and an ultrapure water bath inside the laminar flow cabinet immediately prior to use. These baths and glassware used for dispensing sample/reagents were kept under plastic wrap (Gladwrap) whilst not in use. Acidified seawater (80 ml) was placed in a screwcap polypropylene separating funnel (Nalgene 125 ml) using a lOO-ml measuring cylinder. The separating funnel was mounted in a Lab-Line Multi-Wrist Shaker. The pH was adjusted to N 4.5 with buffer (N 0.73 ml), as determined on a separate representa-
Table I Furnace parameters
Cd
Element Drying and ashing
Tube clean
3x3
55 CIY96) 383-388
tive sample that was not extracted. Finally, 100 pl of 0.5% each APDC/NaDDC complexing agent and 5 ml of DIBK were added. The samples were shaken at 1000 rpm with a l-cm amplitude for 10 min and allowed to stand for 5 min for phase separation. The lower seawater phase was drained and retained for Mn analysis (if required). Using an adjustable pipette (Gilson 5 ml), fitted with a long tip, 4.5 ml of the DIBK phase was withdrawn and placed in a lo-ml screw-cap polypropylene, tapered, centrifuge tube. One ml of Hg solution (100 ppml was added and the tube was shaken by hand for 2 min for back-extraction. After phase separation, 3 ml of the upper DIBK was removed by adjustable pipette, and N 0.9 ml of the lower aqueous phase was transferred to an autosampler cup for GFAAS analysis. A preconcentration factor of 72 was obtained by this method. Blanks were determined by extraction of 80 ml of acidified ultrapure water. To determine Mn, 50 ml of the retained seawater (or a new aliquot) was returned to the separating funnel and the pH was adjusted to N 9.2 with ammonia solution (25%). Finally, 250 pl of 2% oxine solution and 5 ml of DIBK were added. The samples were shaken for 10 min and following phase separation, 4 ml of the DIBK was withdrawn and back-extracted as per the other trace metals. 2.4. Optimisation
of extraction procedure
To optimise the extraction procedure, eight replicate acidified seawater or high-purity water samples were extracted. Four of these were spiked: Cd 0.1 pg 1-l; Co, Ni and Pb 0.5 pg 1-l; Cu, Fe and Mn
for GFAAS analysis Temperature/ramp
Atomisation
Chemist?
(gas stop)
Co/Ni
75/5 I 80/25 600/20 2400/ 2400/2
2400/3
time (“C)/(s)
I
CU
Fe
Mn
Pb
75/s 350,‘25 1000/15
75/5 350/25 1000/15
2500/ 1 2500/3.5
1ooo/ 1 2500/l 2500/3.5
2500/I 2500/4
2500/ 1 2500/2
2400/ 2400/2
2500/3
2500/3
2500/3
2400/3
2500/3
75/5 350/25 500/20
75/5 350/25 800/20
Zn
15/5 350/25 800/ 15
I
75/5 350/25 500/ I5 2000/l 2000/2
2400/3
384
G.J. Batterham, D.L. Party/Marine
l-2 pg 1-l and Zn 1-4 p,g I-‘. Extraction efficiencies were calculated on the basis of spike recoveries. 2.5. Instrument parameters The GFAAS was a Varian SpectrAA equipped with a graphite tube atomiser GTA-95 and autosampler. Pyrolytic coated platform graphite tubes and deuterium background correction were used. Ammonium dihydrogen phosphate (0.5%) purified by DTC/DIBK extraction was used for matrix modification. The sample injection volume was 10 p,l with 2 pl of modifier. Standard addition calibration was necessary for Cd and Pb determination. Standards for all other elements were matrix matched with the back-extracting solution. The general operating parameters were as specified by Varian and the furnace parameters are listed in Table 1.
3. Results and discussion
3.1. Method optimisation In order to attain quantitative recoveries of Co, Fe and Zn with a DTC/DIBK solvent extraction technique it was essential that we improve the extractability of the system proposed by Sachsenberg et al. (1992). We began by increasing the solvent:sam-
Table 2 Extraction
efficiency
Run Extracting
agent
Hg (ppm) Extraction pH Cd co CU Fe Mn Ni Pb Zn
(%) as a function
Chemistry 55 (1996) 381-388
ple ratio from 1: 100 to 1: 16 and replaced hand-shaken volumetric flasks with a mechanical wrist shaker and polypropylene separating funnels. We then investigated the optimum conditions for this system, including reagent type, quantity, extraction pH and shaking times. 3. I .I. The effect of pH and oxine on extraction ejjkiency The pH dependence of DTC extraction into DIBK has not been examined, and as extraction is solvent dependent (Magnusson and Westerlund, 1981), we examined the extraction efficiency over the critical pH range (Table 2). The recovery of Cd, Co, Cu and Pb were quantitative over the entire pH range examined. Manganese was only recovered in the presence of oxine (run 5), but the addition of oxine interfered with Zn recovery, and only Mn and Zn could not be extracted simultaneously in a mixed oxine/NaDDC/APDC extraction system. Iron recovery diminished with increasing pH due to the formation of Fe hydroxy complexes (Danielsson et al., 1978, 1982). The quantitative recovery of Fe at pH 9.2 (Table 2, run 5) was due to complexation with oxine rather than DTC. Experiments at this pH using oxine as the sole complexing agent (no DTC) produced good recoveries of Mn (105 & lo%), Fe (98 f 2%) and Cu (85 + 7%), but poor recoveries ( < 20%) of Cd, Co, Ni, Pb
of pH (n = 4 for each run)
I
2
3
4
5
1% APDC a
0.5% APDC/NaDDC
0.5% APDC/NaDDC
0.5% APDC/NaDDC
500 3.1 99 + 4 lOI+_ 105 *2 91*5 <5 96+ 12 92 + 2 16 f 5
100 4.5 101+4 91+ 7 101 f 4 104+2 <5 102*8 94 * 4 I0056
100 5.6 101 *4 98 f 4 99 f 4 79 f 4 <5 91+7 96 f 3 97 Yh5
100 4.5/8.6 104+2 101 + 3 100+7 65 + 5 <5 87 k 9 98 + 6 104i2
0.5% APDC/NaDDC, 2% oxine 100 4.5/9.2 ’ 97 + 5 99 + 2 91+2 93 f 3 94* I 90 c 3 96 f 2 12 * 4
’
’ Mixed APDC/NaDDC could not be used at pH 3.7, due to the instability of NaDDC at low pH and I % APDC was used. b The DTC was added at pH 4.5 and mixed for 15 s prior to pH adjustment and extraction at pH 8.6. ’ A sequential extraction was carried out: extraction with DTC for IO min at pH 4.5, followed by the addition of oxine and a further 10 min extraction at pH 9.2.
G.J. Batterham, D.L. Party/Marine
and Zn. At low pH (< 4), Zn recoveries decreased due to the acid decomposition of DTC and competitive extraction of protonated DTC (Danielsson et al., 1978). An extraction pH of 4.5 was chosen as the extractability of Fe with DTC was quantitative and optimal at this pH. Sachsenberg et al. (1992) used an extraction pH of 4, but this would be too low for the quantitative extraction of Zn. Attempts to reduce the extraction time showed that a 5-min extraction was generally sufficient for all metals except Fe, which required a IO-min extraction. 3.1.2. The effect of HCl and HNO, sample acidification and APDC
and NaDDC/APDC
on extraction
eflciency
Sachsenberg et al. (1992) achieved good recoveries of Cd, Cu, Ni and Pb from spiked seawater using APDC alone, with mixed NaDDC/APDC only having a slightly better within-the-batch standard deviation. We trialled APDC as the sole complexing agent with CASS-2 Standard Reference Seawater, but obtained Fe concentrations slightly above the certified levels, which we attributed to insufficient extraction of Fe in the blank. To investigate this we examined spike extraction efficiency using APDC and mixed NaDDC/APDC complexing agent for ultrapure water acidified (0.02 M) with either HCI or HNO, (Table 3). This trial showed that mixed NaDDC/APDC complexing agent was essential for quantitative recoveries of all trace metals (runs 3 and 4). Using APDC alone, both Fe and Zn were not quantitatively recovered (runs 1 and 21, Zn having a Table 3 Acidified high-purity water (0.02 M HNO, or HCl) extraction efficiency (%I at pH 4.5 with I% APDC and 0.5% each APDC/NaDDC (n = 4 for each run) 0.5% each APDC/NaDDC
Dithiocarbamate:
1% APDC
Acidification:
HNO,
HCI
I
2
HNO, 3
HCI 4
96+2 98+2 100*2 95i-2 100*2 93+_2 83*1
101*3 101+2 100*2 66+5 107* 1 93*5 96*4
99*3 99*1 101+4 101*2 104*4 96+_2 99+3
96i-I 101* 1 101 f 1 100+6 98i-6 97k2 98*3
Cd co CU Fe Ni Pb Zn
Chemistry 55 (1996) 381-388
385
lower extraction constant with APDC in comparison to NaDDC (Shen et al., 1980). Whilst sample acidification did not affect recoveries with the mixed complexing agent, low Fe recoveries were obtained with APDC for HCI acidified samples (run 2) compared to HNO, acidified samples (run 1). This may be due to the participation of carbon disulphide, an acid decomposition breakdown product of DTC (Hulanicki, 1967), in the formation of mixed Fe complexes, such as Fe(DTC),(CS,). The decomposition reaction of DTC with HCI is slow in comparison to HNO, (Bode and Neumann, 1959; Magnusson and Westerlund, 198 I). Steric factors favour the formation of mixed complexes for Fe, which unlike the other trace metals examined in this study, forms a tris complex. Using the mixed complexing agent, sample acidification was probably not important due to the relative instability of NaDDC with respect to acid decomposition (Cheng et al., 1982 and references therein) and subsequent mixed complex formation. 3.1.3. Factors effecting Hg back-extraction During the development of the metal exchange back-extraction system, we noted several factors affecting the efficiency of the system. At low pH, extraction of protonated DTC reduced the Hg available for metal exchange, lowering the back-extraction efficiency. At pH 3.7 it was necessary to increase the Hg concentration from 100 to 500 ppm to overcome this (Table 2). Carry-over seawater also affected Hg availability for metal exchange, due to the formation of stable mercury(R) chloride species, such as HgCli- (Lo et al., 1982). In this regard, care was necessary to avoid seawater droplets in the transferral of DIBK for back-extractions. Di-isobutyl ketone meets an important criterion for Hg back-extraction: it has a low water miscibility (0.06 k 0.02 ml/l00 ml; Bone and Hibbert, 1979). Seawater carry-over also caused a matrix interference in GFAAS determinations for Pb and Cd, which required standard additions calibration. Danielsson et al. (1978) reported a similar salt interference using freon TF as the solvent. In our trials with low Hg concentrations, incomplete back-extraction primarily affected Co recovery, as Co is primarily extracted as Co(DTC), and undergoes a slow transformation to kinetically stable
386
G.J. Batterham,
D.L. Partyv/Marine
Co(DTC), (Stary and Kratzer, 1968). In some instances where there was no Co recovery, Cu recovery was also reduced: Cu having the highest extraction constant out of the metals examined in this study after Hg (Cheng et al., 1982 and references therein). Another crucial factor that we noted whilst trying to improve extraction efficiencies was that increasing NaDDC and Hg concentrations affected solution stability. Excessive carry-over of DTC into back-extractions resulted in Hg(DTC), exceeding the solubility limit. Some GFAAS samples also became unstable, slowly turning yellow as a precipitate containing elemental sulfur formed (EDAX analysis). This resulted in GFAAS pipetting problems. The precipitate was essentially a function of redox reactions involving CS, formed from the acid decomposition of NaDDC, which is more unstable than APDC (Cheng et al., 1982 and references therein). We began our trials using Pd for back extractions, as used by Sachsenberg et al. (19921, but found this promoted the redox reaction and precipitation in comparison to Hg. With the recommended procedure, back-extraction solutions have been stored prior to analysis in the centrifuge tubes at 4°C for 24 h without precipitation problems or loss of trace metals. Graphite furnace samples are stable with respect to solution discolouration for at least 7 days, but should be analysed as soon as practical. 3.2. Recommended
procedure
The procedure described here requires a single IO-min extraction and 2-min back extraction to quanTable 4 Analysis of CASS-3 nearshore
Bottle I Bottle 2 Certified
”
Blank Detection limit (3~)
standard
reference
Chemists
55 (19961381-388
titatively preconcentrate 7 elements. This compares favourably with existing dual extraction procedures (Magnusson and Westerlund, 1981; Danielsson et al., 1982; Statham, 1985) and with the single 20-min chloroform extraction method of Lo et al. (1982). Single extraction procedures have advantages over dual extraction procedures in terms of contamination potential due to fewer sample manipulations. Another advantage of this technique is that extraction is quantitative. Whilst our DTC procedure could not extract Mn, a method, replacing DTC with oxine, was developed that uses the same general extraction procedure and reagents. Mercury exchange back-extraction was found to be applicable to the oxine solvent extraction system. The procedure we developed can even re-use the same seawater aliquot previously used for the DTC extraction, but extreme care must be taken to avoid contamination and we generally use a fresh aliquot of the sample. Statham (198.5) required over 150 times the level of DTC used here to simultaneously extract Mn with other trace metals. Such high DTC levels are not amenable to metal exchange back-extraction. Whilst nitric acid back-extraction works with our technique, Hg exchange back-extraction has considerable preconcentration advantages due to the small back-extraction volume that can be used. Our technique gives a preconcentration factor of 72 for an 80-ml sample. The analytical data obtained for CASS-3 Nearshore Standard Reference Seawater with the recommended procedure were in excellent agreement with the certified values (Table 4). The standard deviations, blanks and detection limits quoted here are at the low end of reported values for solvent
material (n = 4)
Cd
co
CU
Fe
Mn
Ni
Pb
Zn
(PPt)
(PPt)
(PPt)
(ppb)
(PPb)
(PPt)
(PPt)
(ppb)
3I .7 f 0.4 33.2 i 0.8 30 f 5
45 * 3 47 + 6 41 *9 b.d.1. 15
502 f 4 532 + I 517i62 20-30 6
I .38 i 0.02 1.21 * 0.02 1.26+0.17 b.d.1. 0.03
2.48 + 0.02 2.64 + 0.03 2.5 I + 0.36
375 + 1I 390 + 14 386 + 62 b.d.1. 18
< 15 11 f4h 1214 b.d.1. 8
I .29 + 0.03 I .20 f 0.04 I .25 i 0.24
b.d.1.
I .o
b.d.1. = below detection limit. a Two different bottles of CASS-3 were analysed. ’ A greater sample injection volume was used in the GFAAS determination to 8 ppt.
b.d.1.
0.03
0.03-0.05 0.0 I
of Pb in Bottle-2 samples, lowering the detection limit from 15
C.J. Batterham,
D.L. Pan-v/Marine
extraction. Lower detection limits than those presented in Table 4 have been achieved in our laboratory by reducing the Hg back-extraction volume by up to a quarter and increasing the Hg concentration by the same factor, giving a preconcentration factor of up to 288. The use of small reagent volumes and dilute reagent concentrations reduces blank levels and allows very low detection limits to be obtained. The vast majority of our blanks were contributed by technical grade DIBK. The use of 95% DIBK (Ajax> in comparison to 75% (Merck) lowered blanks by nearly an order of magnitude for most trace metals and purification by distillation was not necessary. Of the reagents used in our procedure, only minimal precleaning of the extracting agents was required. Researchers undertaking our procedure should conduct an initial spike recovery investigation to ensure that their shaking time is adequate to overcome problems associated with Fe hydroxy complex formation. Regular spike recovery trials should also be conducted as part of a laboratory quality control program to ensure continued reagent viability, particularly with regard to NaDDC decomposition. Sodium diethyldithiocarbamate is critical for quantitative Zn recovery and it has been our experience that aged NaDDC results in declining Zn recoveries. In this case, a mean zinc extraction efficiency can be used to correct incomplete extraction in a particular batch of samples.
4. Conclusions The proposed procedure is applicable to the determination of trace metals from pristine coastal and open-ocean levels to polluted harbour concentrations. A dozen samples can be readily preconcentrated simultaneously on a wrist shaker within 1 h. The procedure has been successfully trialled in our associated consultancy laboratory on Darwin Harbour seawater samples collected weekly over an g-month period. The major advantage of this procedure is that large preconcentration factors can be rapidly achieved for a wide range of metals from relatively small samples. Modified versions of this technique have successfully been applied to 40-ml samples and trials
Chemistry
55 (1996) 381-388
387
with smaller volumes are being undertaken to allow the determination of these trace metals in pore waters.
Acknowledgements This research was generously sponsored by McArthur River Mining Pty. Ltd. The authors also wish to thank Dr. N.C. Munksgaard and the technical staff of the Northern Territory University for their assistance.
References Bajo, S. and Wyttenbach, A., 1979. Lead. cadmium and zinc bis(diethyldithiocarbamate1 and diethyldithiocarbamic acid as reagents for liquid-liquid extraction. Anal. Chem., 5 l(3): 376-378. Bode, H. and Neumann, F., 1959. Untersuchungen iiber disubstituierte Dithiocarbamidate. Fresenius Z. Anal. Chem., 169: 1410-1416. Bone, K.M. and Hibbert, W.D., 1979. Solvent extraction with ammonium pyrrolidinedithiocarbamate and 2.6.dimethyl-4heptanone for the determination of trace metals in effluents and natural waters. Anal. Chim. Acta, 107: 219-229. Brooks, R.R., Presley, B.J. and Kaplan, I.R.. 1967. APDC-MIBK extraction system for the determination of trace elements in saline waters by atomic-absorption spectrometry. Talanta. l-1(71: 809-816. Cheng, K.L., Ueno, K. and Imamura, T., 1982. CRC handbook of organic analytical reagents. CRC Press, Boca Raton, FL. pp. 389-398. Danielsson, L.-G., Magnusson. B. and Westerlund, S., 1978. An improved metal extraction procedure for the determination of trace metals in sea water by atomic absorption spectrometry with electrothermal atomization. Anal. Chim. Acta, 98: 47-57. Danielsson, L.-G., Magnusson, B., Westerlund, S. and Zhang, K.. 1982. Trace metal determination in estuarine waters by electrothermal atomic absorption spectrometry after extraction 01 dithiocarbamate complexes into freon. Anal. Chim. Acta, 144: 183-188. Hulanicki, A.. 1967. Complexation reactions of dithiocarbamates. Talanta, 14: 1371-1392. Jan, T.K. and Young, D.R., 1978. Determination of microgram amounts of some transition metals in seawater by methyl isobutyl ketone-nitric acid successive extraction and tlameless atomic absorption spectrophotometry. Anal. Chem.. So(S): I250- 1253. Kinrade, J.D. and Van Loon, J.C., 1974. Solvent extraction for use with tlame atomic absorption spectrometry. Anal. Chem.. 46(131: 1894-1898.
388
G.J. Batterham, D.L. Parry/Marine
Klinkhammer, G.P., 1980. Determination of manganese in seawater by flameless atomic adsorption spectrometry after pre-concentration with R-hydroxyquinoline in chloroform. Anal. Chem., 52: 117-120. Lo, J.M., Yu, J.C., Hutchison, F.I. and Wal, CM., 1982. Solvent extraction of dithiocarbamate complexes and back-extraction with mecury(II) for determination of trace metals by atomic absorption spectrometry. Anal. Chem., 54: 2536-2539. Magnusson, B. and Westerlund, S., 1981. Solvent extraction procedures combined with back-extraction for trace metal determinations by atomic absorption spectrometry. Anal. Chim. Acta, 131: 63-72. Sachsenberg, S., Klenke, Th., Krumbein, W.E. and Zeeck, E., 1992. A back-extraction procedure for the dithiocarbamate solvent extraction method. Rapid determination of metals in seawater matrices, Fresenius J. Anal. Chem., 342: 163-166. She”, L.H., Yeh, S.J. and Lo, J.M., 1980. Determination of
Chemistry 55 (1996) 381-388 extraction constants for lead(B), zinc(B), thallium(I), and manganese(B) dithiocarbamates by a two-step extraction method. Anal. Chem., 52(12): 1882-1885. Stary, J. and Kramer, K., 1968. Determination of extraction constants of metal diethyldithiocarbamates. Anal. Chim. Acta, 40: 93-100. Statham, P.J., 1985. The determination of dissolved manganese and cadmium in sea water at low nmol L- ’ concentrations by chelation and extraction followed by electrothermal atomic absorption spectrometry. Anal. Chim. Acta, 169: 149-159. Sturgeon, R.E., Berman, S.S., Desauliniers, A. and Russell, D.S., 1980. Pre-concentration of trace metals from sea-water for determination by graphite furnace atomic-absorption spectrometry. Talanta, 27: 85-94. Wyttenbach, A. and Bajo, S., 1975. Extractions with metal-dithiocarbamates as reagents. Anal. Chem., 47(11): 1813-1817.
Marine Chemistry
55 (1996) 381-388
Improved dithiocarbamate/oxine solvent extraction method for the preconcentration of trace metals from seawater using metal exchange back-extraction Grant J. Batterham School of Mathematical
and Physical
Sciences, Northern
Received 6 October
*,
David L. Parry Territory
1995; accepted
Unicvrsity
Darwin,
0909 NT, Australia
10 July 1996
Abstract A rapid single extraction procedure using dithiocarbamate complexing agent, a di-isobutyl ketone (DIBK) organic phase and Hg exchange back-extraction is described for the simultaneous quantitative preconcentration of Cd, Co, Cu, Fe, Ni, Pb and Zn in seawater. An 8-hydroxyquinoline (oxine) extraction technique is also presented for the subsequent determination of Mn using the same reagents and operating system. The proposed method gave quantitative spike recoveries for all metals and was in excellent agreement with the certified results for CASS-3 marine reference seawater. Blanks and detection limits are in the low ppt range. The method gives a preconcentration factor of 72 for an SO-ml sample, facilitating the analysis of pristine seawater samples. Kqvwords:
heavy metals; seawater;
extraction;
preconcentration;
dithiocarbamate
1. Introduction Solvent extraction procedures are widely used for the determination of trace metals in seawater. The earliest standard methods were based upon metal complexation with ammonium pyrrolidine dithiocarbamate (APDC) and extraction into methyl isobutyl ketone (MIBK) (Brooks et al., 1967; Jan and Young, 1978). Kinrade and Van Loon (1974) examined the APDC-MIBK system in some detail and added diethylammonium diethyldithiocarbamate in combination with APDC to improve complex stability and broaden the effective working pH range.
_ Corresponding 0304-4203/96/$15.00 P/I
author. Copyright
SO304-4203(96)00068-O
The direct analysis of MIBK extracts by flame atomic absorption spectrometry (FAAS) was limited by the poor stability of dithiocarbamate (DTC) complexes in MIBK (Brooks et al., 1967; Kinrade and Van Loon, 1974). Danielsson et al. (1978) and Jan and Young (1978) overcame this by developing acid back-extractions to transfer DTC complexed metals into a stable acidified aqueous phase. Magnusson and Westerlund (198 1) used acid back-extraction efficiencies to show DTC complex stability was greatest in chloroform, followed by MIBK and freon TF. Graphite furnace atomic absorption spectrometry (GFAAS), with its improved sensitivity, replaced FAA& and MIBK was no longer used as a solvent due to its high water miscibility which resulted in
0 1996 Elsevier Science B.V. All rights reserved.
382
G.J. Batterham,
D.L. Parry/Marine
salt carry-over and matrix interference (Magnusson and Westerlund, 1981). Freon TF was subsequently used as the solvent of choice, but dual extractions were required for the quantitative recovery of most common trace metals (Danielsson et al., 1982; Statham, 1985). Metal exchange back-extraction was developed by Lo et al. (1982), using Hg to overcome the slow kinetics of acid back-extraction. This method required a single 20-min extraction into chloroform to preconcentrate (> 90%) Cd, Co, Cu, Fe, Ni, Pb and Zn. Improvements to this method were recently made by Sachsenberg et al. (1992) who only required a 90-s single extraction into di-isobutyl ketone (DIBK), with Pd back-extraction, to quantitatively determine Pb, Cu, Ni and Cd in seawater. This technique produced high preconcentration factors from relatively small sample volumes, but was limited by the range of metals determined. Manganese, an important element in geochemical and speciation studies, has a low affinity for DTC ligands (Shen et al., 1980), necessitating high DTC concentrations and dual extractions to achieve quantitaive preconcentration from seawater (Statham, 1985). Previous studies have also used &hydroxyquinoline (oxine) as a complexing agent to extract Mn into chloroform (Klinkhammer, 1980) or MIBK (Sturgeon et al., 1980). Sachsenberg et al. (1992) demonstrated their rapid DTC/DIBK solvent extraction method for Cd, Cu, Ni and Pb and did not evaluate Co, Fe, Mn and Zn which have much poorer recoveries with DTC solvent extraction systems (Stary and Kratzer, 1968; Wyttenbach and Bajo, 1975; Bajo and Wyttenbach, 1979; Shen et al., 1980). Di-isobutyl ketone has been found to be an efficient solvent for the extraction of a wide range of DTC-metal chelates from non-saline waters (Bone and Hibbert, 1979). The purpose of our study was to develop a rapid and reliable single extraction method with metal exchange back-extraction that allows the quantitative determination of Cd, Co, Cu, Fe, Mn, Ni, Pb and Zn in seawater. Due to the low extractability of Mn with DTC, we also examined the use of oxine, independently and in combination with DTC, for extraction into DIBK and its compatibility with metal exchange back-extraction.
Chemistry
55
C1996)
381-388
2. Experimental
2.1. Chemicals Acids and ammonia (25%) were of Suprapur/Aristar/Ultrex quality. Complexing and back-extracting agents were of analytical grade. Technical grade 2,6-dimethyl-heptan-4-one (DIBK), Merck (75% DIBK, 25% 2,4-dimethyl-heptan-6-one) or Ajax (95%), was used as purchased, with the latter preferred due to lower metal impurities. Ultrapure water used to make up all solutions was obtained from a Permutit HI-PURE water system fed with water from a reverse osmosis system. Dithiocarbamate complexing solution, 0.5% each APDC/sodium diethyldithiocarbamate (NaDDC), was made up immediately prior to use in IO-ml volumetric flasks with pH 2.5 (HNO,) ultrapure water. The solutions were purified by extraction with DIBK (1 ml, I-min shake). Stock oxine solution (2%) was prepared by dissolving 10 g of oxine in 500 ml of 0.5% HCl (v/v> and purified with DIBK (2 X 5 ml). Stock ammonia (3 M)/acetic acid (2 M) buffer was made up weekly and did not require purification. Stock 1000 ppm Hg back extracting solution was prepared from AR grade Hg(NO,), in 0.02 M HNO, and diluted to 100 ppm with 0.02 M HNO, as required. 2.2. Seawater Seawater was collected directly from Stokes Hill Wharf in Darwin Harbour, Australia, using a battery-operated polystyrene, submersible, in-line pump (Niagara LVM 114) fitted to a weighted polyvinylchloride (PVC) pole. The pump was run for several minutes prior to the attachment of a Gelman groundwater sampling capsule for in-line filtration (0.45 pm). The filter capsule was flushed with seawater for several minutes prior to sample collection. Filtered seawater was collected in an acid-washed (10% HNO,) 20-l polyethylene carboy and acidified to 0.02 M HNO, (Ultrex). The seawater had a salinity of 35%0 and pH of 8.2. CASS-3 Nearshore Standard Reference Material was obtained from the National Research Council of Canada to validate methods.
G.J. Batterham, D.L. Parry/Marine
Low-density polypropylene bottles (Nalgene) reserved for regular seawater sampling were prepared by soaking in detergent (1% Decon 90) for 24 h, followed by 3 rinses in ultrapure water and soaking in 10% HNO,/ultrapure water for at least 48 h. The bottles were again rinsed 3 times in ultrapure water, dried in a clean oven, and sealed in polyethylene plastic bags. The bottles were thoroughly rinsed with filtered seawater immediately prior to sample collection. Samples were placed on ice and acidified to 0.02 M HNO, with ultrapure acid in a clean environment. 2.3. Extraction
and preconcentration
procedure
The extraction procedure was performed in a Class-100 laminar flow cabinet. To limit exposure and potential contamination, caps were only removed from separating funnels/centrifuge tubes for the introduction or removal of sample/reagents. All labware was acid washed in the manner described for the sampling bottles. In addition, all acid-washed polyethylene pipette tips were again doubly rinsed in a nitric acid bath (5%) and an ultrapure water bath inside the laminar flow cabinet immediately prior to use. These baths and glassware used for dispensing sample/reagents were kept under plastic wrap (Gladwrap) whilst not in use. Acidified seawater (80 ml) was placed in a screwcap polypropylene separating funnel (Nalgene 125 ml) using a lOO-ml measuring cylinder. The separating funnel was mounted in a Lab-Line Multi-Wrist Shaker. The pH was adjusted to N 4.5 with buffer (N 0.73 ml), as determined on a separate representa-
Table I Furnace parameters
Cd
Element Drying and ashing
Tube clean
3x3
55 CIY96) 383-388
tive sample that was not extracted. Finally, 100 pl of 0.5% each APDC/NaDDC complexing agent and 5 ml of DIBK were added. The samples were shaken at 1000 rpm with a l-cm amplitude for 10 min and allowed to stand for 5 min for phase separation. The lower seawater phase was drained and retained for Mn analysis (if required). Using an adjustable pipette (Gilson 5 ml), fitted with a long tip, 4.5 ml of the DIBK phase was withdrawn and placed in a lo-ml screw-cap polypropylene, tapered, centrifuge tube. One ml of Hg solution (100 ppml was added and the tube was shaken by hand for 2 min for back-extraction. After phase separation, 3 ml of the upper DIBK was removed by adjustable pipette, and N 0.9 ml of the lower aqueous phase was transferred to an autosampler cup for GFAAS analysis. A preconcentration factor of 72 was obtained by this method. Blanks were determined by extraction of 80 ml of acidified ultrapure water. To determine Mn, 50 ml of the retained seawater (or a new aliquot) was returned to the separating funnel and the pH was adjusted to N 9.2 with ammonia solution (25%). Finally, 250 pl of 2% oxine solution and 5 ml of DIBK were added. The samples were shaken for 10 min and following phase separation, 4 ml of the DIBK was withdrawn and back-extracted as per the other trace metals. 2.4. Optimisation
of extraction procedure
To optimise the extraction procedure, eight replicate acidified seawater or high-purity water samples were extracted. Four of these were spiked: Cd 0.1 pg 1-l; Co, Ni and Pb 0.5 pg 1-l; Cu, Fe and Mn
for GFAAS analysis Temperature/ramp
Atomisation
Chemist?
(gas stop)
Co/Ni
75/5 I 80/25 600/20 2400/ 2400/2
2400/3
time (“C)/(s)
I
CU
Fe
Mn
Pb
75/s 350,‘25 1000/15
75/5 350/25 1000/15
2500/ 1 2500/3.5
1ooo/ 1 2500/l 2500/3.5
2500/I 2500/4
2500/ 1 2500/2
2400/ 2400/2
2500/3
2500/3
2500/3
2400/3
2500/3
75/5 350/25 500/20
75/5 350/25 800/20
Zn
15/5 350/25 800/ 15
I
75/5 350/25 500/ I5 2000/l 2000/2
2400/3
384
G.J. Batterham, D.L. Party/Marine
l-2 pg 1-l and Zn 1-4 p,g I-‘. Extraction efficiencies were calculated on the basis of spike recoveries. 2.5. Instrument parameters The GFAAS was a Varian SpectrAA equipped with a graphite tube atomiser GTA-95 and autosampler. Pyrolytic coated platform graphite tubes and deuterium background correction were used. Ammonium dihydrogen phosphate (0.5%) purified by DTC/DIBK extraction was used for matrix modification. The sample injection volume was 10 p,l with 2 pl of modifier. Standard addition calibration was necessary for Cd and Pb determination. Standards for all other elements were matrix matched with the back-extracting solution. The general operating parameters were as specified by Varian and the furnace parameters are listed in Table 1.
3. Results and discussion
3.1. Method optimisation In order to attain quantitative recoveries of Co, Fe and Zn with a DTC/DIBK solvent extraction technique it was essential that we improve the extractability of the system proposed by Sachsenberg et al. (1992). We began by increasing the solvent:sam-
Table 2 Extraction
efficiency
Run Extracting
agent
Hg (ppm) Extraction pH Cd co CU Fe Mn Ni Pb Zn
(%) as a function
Chemistry 55 (1996) 381-388
ple ratio from 1: 100 to 1: 16 and replaced hand-shaken volumetric flasks with a mechanical wrist shaker and polypropylene separating funnels. We then investigated the optimum conditions for this system, including reagent type, quantity, extraction pH and shaking times. 3. I .I. The effect of pH and oxine on extraction ejjkiency The pH dependence of DTC extraction into DIBK has not been examined, and as extraction is solvent dependent (Magnusson and Westerlund, 1981), we examined the extraction efficiency over the critical pH range (Table 2). The recovery of Cd, Co, Cu and Pb were quantitative over the entire pH range examined. Manganese was only recovered in the presence of oxine (run 5), but the addition of oxine interfered with Zn recovery, and only Mn and Zn could not be extracted simultaneously in a mixed oxine/NaDDC/APDC extraction system. Iron recovery diminished with increasing pH due to the formation of Fe hydroxy complexes (Danielsson et al., 1978, 1982). The quantitative recovery of Fe at pH 9.2 (Table 2, run 5) was due to complexation with oxine rather than DTC. Experiments at this pH using oxine as the sole complexing agent (no DTC) produced good recoveries of Mn (105 & lo%), Fe (98 f 2%) and Cu (85 + 7%), but poor recoveries ( < 20%) of Cd, Co, Ni, Pb
of pH (n = 4 for each run)
I
2
3
4
5
1% APDC a
0.5% APDC/NaDDC
0.5% APDC/NaDDC
0.5% APDC/NaDDC
500 3.1 99 + 4 lOI+_ 105 *2 91*5 <5 96+ 12 92 + 2 16 f 5
100 4.5 101+4 91+ 7 101 f 4 104+2 <5 102*8 94 * 4 I0056
100 5.6 101 *4 98 f 4 99 f 4 79 f 4 <5 91+7 96 f 3 97 Yh5
100 4.5/8.6 104+2 101 + 3 100+7 65 + 5 <5 87 k 9 98 + 6 104i2
0.5% APDC/NaDDC, 2% oxine 100 4.5/9.2 ’ 97 + 5 99 + 2 91+2 93 f 3 94* I 90 c 3 96 f 2 12 * 4
’
’ Mixed APDC/NaDDC could not be used at pH 3.7, due to the instability of NaDDC at low pH and I % APDC was used. b The DTC was added at pH 4.5 and mixed for 15 s prior to pH adjustment and extraction at pH 8.6. ’ A sequential extraction was carried out: extraction with DTC for IO min at pH 4.5, followed by the addition of oxine and a further 10 min extraction at pH 9.2.
G.J. Batterham, D.L. Party/Marine
and Zn. At low pH (< 4), Zn recoveries decreased due to the acid decomposition of DTC and competitive extraction of protonated DTC (Danielsson et al., 1978). An extraction pH of 4.5 was chosen as the extractability of Fe with DTC was quantitative and optimal at this pH. Sachsenberg et al. (1992) used an extraction pH of 4, but this would be too low for the quantitative extraction of Zn. Attempts to reduce the extraction time showed that a 5-min extraction was generally sufficient for all metals except Fe, which required a IO-min extraction. 3.1.2. The effect of HCl and HNO, sample acidification and APDC
and NaDDC/APDC
on extraction
eflciency
Sachsenberg et al. (1992) achieved good recoveries of Cd, Cu, Ni and Pb from spiked seawater using APDC alone, with mixed NaDDC/APDC only having a slightly better within-the-batch standard deviation. We trialled APDC as the sole complexing agent with CASS-2 Standard Reference Seawater, but obtained Fe concentrations slightly above the certified levels, which we attributed to insufficient extraction of Fe in the blank. To investigate this we examined spike extraction efficiency using APDC and mixed NaDDC/APDC complexing agent for ultrapure water acidified (0.02 M) with either HCI or HNO, (Table 3). This trial showed that mixed NaDDC/APDC complexing agent was essential for quantitative recoveries of all trace metals (runs 3 and 4). Using APDC alone, both Fe and Zn were not quantitatively recovered (runs 1 and 21, Zn having a Table 3 Acidified high-purity water (0.02 M HNO, or HCl) extraction efficiency (%I at pH 4.5 with I% APDC and 0.5% each APDC/NaDDC (n = 4 for each run) 0.5% each APDC/NaDDC
Dithiocarbamate:
1% APDC
Acidification:
HNO,
HCI
I
2
HNO, 3
HCI 4
96+2 98+2 100*2 95i-2 100*2 93+_2 83*1
101*3 101+2 100*2 66+5 107* 1 93*5 96*4
99*3 99*1 101+4 101*2 104*4 96+_2 99+3
96i-I 101* 1 101 f 1 100+6 98i-6 97k2 98*3
Cd co CU Fe Ni Pb Zn
Chemistry 55 (1996) 381-388
385
lower extraction constant with APDC in comparison to NaDDC (Shen et al., 1980). Whilst sample acidification did not affect recoveries with the mixed complexing agent, low Fe recoveries were obtained with APDC for HCI acidified samples (run 2) compared to HNO, acidified samples (run 1). This may be due to the participation of carbon disulphide, an acid decomposition breakdown product of DTC (Hulanicki, 1967), in the formation of mixed Fe complexes, such as Fe(DTC),(CS,). The decomposition reaction of DTC with HCI is slow in comparison to HNO, (Bode and Neumann, 1959; Magnusson and Westerlund, 198 I). Steric factors favour the formation of mixed complexes for Fe, which unlike the other trace metals examined in this study, forms a tris complex. Using the mixed complexing agent, sample acidification was probably not important due to the relative instability of NaDDC with respect to acid decomposition (Cheng et al., 1982 and references therein) and subsequent mixed complex formation. 3.1.3. Factors effecting Hg back-extraction During the development of the metal exchange back-extraction system, we noted several factors affecting the efficiency of the system. At low pH, extraction of protonated DTC reduced the Hg available for metal exchange, lowering the back-extraction efficiency. At pH 3.7 it was necessary to increase the Hg concentration from 100 to 500 ppm to overcome this (Table 2). Carry-over seawater also affected Hg availability for metal exchange, due to the formation of stable mercury(R) chloride species, such as HgCli- (Lo et al., 1982). In this regard, care was necessary to avoid seawater droplets in the transferral of DIBK for back-extractions. Di-isobutyl ketone meets an important criterion for Hg back-extraction: it has a low water miscibility (0.06 k 0.02 ml/l00 ml; Bone and Hibbert, 1979). Seawater carry-over also caused a matrix interference in GFAAS determinations for Pb and Cd, which required standard additions calibration. Danielsson et al. (1978) reported a similar salt interference using freon TF as the solvent. In our trials with low Hg concentrations, incomplete back-extraction primarily affected Co recovery, as Co is primarily extracted as Co(DTC), and undergoes a slow transformation to kinetically stable
386
G.J. Batterham,
D.L. Partyv/Marine
Co(DTC), (Stary and Kratzer, 1968). In some instances where there was no Co recovery, Cu recovery was also reduced: Cu having the highest extraction constant out of the metals examined in this study after Hg (Cheng et al., 1982 and references therein). Another crucial factor that we noted whilst trying to improve extraction efficiencies was that increasing NaDDC and Hg concentrations affected solution stability. Excessive carry-over of DTC into back-extractions resulted in Hg(DTC), exceeding the solubility limit. Some GFAAS samples also became unstable, slowly turning yellow as a precipitate containing elemental sulfur formed (EDAX analysis). This resulted in GFAAS pipetting problems. The precipitate was essentially a function of redox reactions involving CS, formed from the acid decomposition of NaDDC, which is more unstable than APDC (Cheng et al., 1982 and references therein). We began our trials using Pd for back extractions, as used by Sachsenberg et al. (19921, but found this promoted the redox reaction and precipitation in comparison to Hg. With the recommended procedure, back-extraction solutions have been stored prior to analysis in the centrifuge tubes at 4°C for 24 h without precipitation problems or loss of trace metals. Graphite furnace samples are stable with respect to solution discolouration for at least 7 days, but should be analysed as soon as practical. 3.2. Recommended
procedure
The procedure described here requires a single IO-min extraction and 2-min back extraction to quanTable 4 Analysis of CASS-3 nearshore
Bottle I Bottle 2 Certified
”
Blank Detection limit (3~)
standard
reference
Chemists
55 (19961381-388
titatively preconcentrate 7 elements. This compares favourably with existing dual extraction procedures (Magnusson and Westerlund, 1981; Danielsson et al., 1982; Statham, 1985) and with the single 20-min chloroform extraction method of Lo et al. (1982). Single extraction procedures have advantages over dual extraction procedures in terms of contamination potential due to fewer sample manipulations. Another advantage of this technique is that extraction is quantitative. Whilst our DTC procedure could not extract Mn, a method, replacing DTC with oxine, was developed that uses the same general extraction procedure and reagents. Mercury exchange back-extraction was found to be applicable to the oxine solvent extraction system. The procedure we developed can even re-use the same seawater aliquot previously used for the DTC extraction, but extreme care must be taken to avoid contamination and we generally use a fresh aliquot of the sample. Statham (198.5) required over 150 times the level of DTC used here to simultaneously extract Mn with other trace metals. Such high DTC levels are not amenable to metal exchange back-extraction. Whilst nitric acid back-extraction works with our technique, Hg exchange back-extraction has considerable preconcentration advantages due to the small back-extraction volume that can be used. Our technique gives a preconcentration factor of 72 for an 80-ml sample. The analytical data obtained for CASS-3 Nearshore Standard Reference Seawater with the recommended procedure were in excellent agreement with the certified values (Table 4). The standard deviations, blanks and detection limits quoted here are at the low end of reported values for solvent
material (n = 4)
Cd
co
CU
Fe
Mn
Ni
Pb
Zn
(PPt)
(PPt)
(PPt)
(ppb)
(PPb)
(PPt)
(PPt)
(ppb)
3I .7 f 0.4 33.2 i 0.8 30 f 5
45 * 3 47 + 6 41 *9 b.d.1. 15
502 f 4 532 + I 517i62 20-30 6
I .38 i 0.02 1.21 * 0.02 1.26+0.17 b.d.1. 0.03
2.48 + 0.02 2.64 + 0.03 2.5 I + 0.36
375 + 1I 390 + 14 386 + 62 b.d.1. 18
< 15 11 f4h 1214 b.d.1. 8
I .29 + 0.03 I .20 f 0.04 I .25 i 0.24
b.d.1.
I .o
b.d.1. = below detection limit. a Two different bottles of CASS-3 were analysed. ’ A greater sample injection volume was used in the GFAAS determination to 8 ppt.
b.d.1.
0.03
0.03-0.05 0.0 I
of Pb in Bottle-2 samples, lowering the detection limit from 15
C.J. Batterham,
D.L. Pan-v/Marine
extraction. Lower detection limits than those presented in Table 4 have been achieved in our laboratory by reducing the Hg back-extraction volume by up to a quarter and increasing the Hg concentration by the same factor, giving a preconcentration factor of up to 288. The use of small reagent volumes and dilute reagent concentrations reduces blank levels and allows very low detection limits to be obtained. The vast majority of our blanks were contributed by technical grade DIBK. The use of 95% DIBK (Ajax> in comparison to 75% (Merck) lowered blanks by nearly an order of magnitude for most trace metals and purification by distillation was not necessary. Of the reagents used in our procedure, only minimal precleaning of the extracting agents was required. Researchers undertaking our procedure should conduct an initial spike recovery investigation to ensure that their shaking time is adequate to overcome problems associated with Fe hydroxy complex formation. Regular spike recovery trials should also be conducted as part of a laboratory quality control program to ensure continued reagent viability, particularly with regard to NaDDC decomposition. Sodium diethyldithiocarbamate is critical for quantitative Zn recovery and it has been our experience that aged NaDDC results in declining Zn recoveries. In this case, a mean zinc extraction efficiency can be used to correct incomplete extraction in a particular batch of samples.
4. Conclusions The proposed procedure is applicable to the determination of trace metals from pristine coastal and open-ocean levels to polluted harbour concentrations. A dozen samples can be readily preconcentrated simultaneously on a wrist shaker within 1 h. The procedure has been successfully trialled in our associated consultancy laboratory on Darwin Harbour seawater samples collected weekly over an g-month period. The major advantage of this procedure is that large preconcentration factors can be rapidly achieved for a wide range of metals from relatively small samples. Modified versions of this technique have successfully been applied to 40-ml samples and trials
Chemistry
55 (1996) 381-388
387
with smaller volumes are being undertaken to allow the determination of these trace metals in pore waters.
Acknowledgements This research was generously sponsored by McArthur River Mining Pty. Ltd. The authors also wish to thank Dr. N.C. Munksgaard and the technical staff of the Northern Territory University for their assistance.
References Bajo, S. and Wyttenbach, A., 1979. Lead. cadmium and zinc bis(diethyldithiocarbamate1 and diethyldithiocarbamic acid as reagents for liquid-liquid extraction. Anal. Chem., 5 l(3): 376-378. Bode, H. and Neumann, F., 1959. Untersuchungen iiber disubstituierte Dithiocarbamidate. Fresenius Z. Anal. Chem., 169: 1410-1416. Bone, K.M. and Hibbert, W.D., 1979. Solvent extraction with ammonium pyrrolidinedithiocarbamate and 2.6.dimethyl-4heptanone for the determination of trace metals in effluents and natural waters. Anal. Chim. Acta, 107: 219-229. Brooks, R.R., Presley, B.J. and Kaplan, I.R.. 1967. APDC-MIBK extraction system for the determination of trace elements in saline waters by atomic-absorption spectrometry. Talanta. l-1(71: 809-816. Cheng, K.L., Ueno, K. and Imamura, T., 1982. CRC handbook of organic analytical reagents. CRC Press, Boca Raton, FL. pp. 389-398. Danielsson, L.-G., Magnusson. B. and Westerlund, S., 1978. An improved metal extraction procedure for the determination of trace metals in sea water by atomic absorption spectrometry with electrothermal atomization. Anal. Chim. Acta, 98: 47-57. Danielsson, L.-G., Magnusson, B., Westerlund, S. and Zhang, K.. 1982. Trace metal determination in estuarine waters by electrothermal atomic absorption spectrometry after extraction 01 dithiocarbamate complexes into freon. Anal. Chim. Acta, 144: 183-188. Hulanicki, A.. 1967. Complexation reactions of dithiocarbamates. Talanta, 14: 1371-1392. Jan, T.K. and Young, D.R., 1978. Determination of microgram amounts of some transition metals in seawater by methyl isobutyl ketone-nitric acid successive extraction and tlameless atomic absorption spectrophotometry. Anal. Chem.. So(S): I250- 1253. Kinrade, J.D. and Van Loon, J.C., 1974. Solvent extraction for use with tlame atomic absorption spectrometry. Anal. Chem.. 46(131: 1894-1898.
388
G.J. Batterham, D.L. Parry/Marine
Klinkhammer, G.P., 1980. Determination of manganese in seawater by flameless atomic adsorption spectrometry after pre-concentration with R-hydroxyquinoline in chloroform. Anal. Chem., 52: 117-120. Lo, J.M., Yu, J.C., Hutchison, F.I. and Wal, CM., 1982. Solvent extraction of dithiocarbamate complexes and back-extraction with mecury(II) for determination of trace metals by atomic absorption spectrometry. Anal. Chem., 54: 2536-2539. Magnusson, B. and Westerlund, S., 1981. Solvent extraction procedures combined with back-extraction for trace metal determinations by atomic absorption spectrometry. Anal. Chim. Acta, 131: 63-72. Sachsenberg, S., Klenke, Th., Krumbein, W.E. and Zeeck, E., 1992. A back-extraction procedure for the dithiocarbamate solvent extraction method. Rapid determination of metals in seawater matrices, Fresenius J. Anal. Chem., 342: 163-166. She”, L.H., Yeh, S.J. and Lo, J.M., 1980. Determination of
Chemistry 55 (1996) 381-388 extraction constants for lead(B), zinc(B), thallium(I), and manganese(B) dithiocarbamates by a two-step extraction method. Anal. Chem., 52(12): 1882-1885. Stary, J. and Kramer, K., 1968. Determination of extraction constants of metal diethyldithiocarbamates. Anal. Chim. Acta, 40: 93-100. Statham, P.J., 1985. The determination of dissolved manganese and cadmium in sea water at low nmol L- ’ concentrations by chelation and extraction followed by electrothermal atomic absorption spectrometry. Anal. Chim. Acta, 169: 149-159. Sturgeon, R.E., Berman, S.S., Desauliniers, A. and Russell, D.S., 1980. Pre-concentration of trace metals from sea-water for determination by graphite furnace atomic-absorption spectrometry. Talanta, 27: 85-94. Wyttenbach, A. and Bajo, S., 1975. Extractions with metal-dithiocarbamates as reagents. Anal. Chem., 47(11): 1813-1817.