- Email: [email protected]
Concentration of trace metals from sea-water by complexation with 8-hydroxyquinoline and adsorption on C18-bonded silica gel
Talontq Printed
Vol. 29, pp. 167 to 171, 1982 in Great Britain. All rights reserved
Copyright
~39-9140/82/030167-051o3.oo/0 0 1982 Pergamon Press Ltd
CONCENTRATION OF TRACE METALS FROM SEA-WATER BY COMPLEXATION WITH 8-HYDROXYQUINOLINE AND ADSORPTION ON Cl ,-BONDED SILICA GEL R. E. STURGEON*,S. S. BERMANand S. N. WILLIE Division of Chemistry, National Research Council of Canada, Montreal Road, Ottawa, Ontario, Canada (Received 29 June 1981. Accepted 15 September 1981) Snmmary-A reversed-phase liquid chromatographic technique based on a combination of multielement chelation by 8-hydroxyquinoline with subsequent adsorption on C I.s-bonded silica gel is described for the concentration of Cd, Zn, Cu, Ni, Co, Mn and Fe from sea-water. Enrichment factors of 5(rlOO are readily obtained following elution of the absorbate with methanol to provide a matrix-free concentrate suitable for graphite-furnace atomic-absorption analysis. Quantitative recovery of these elements from near-shore samples of sea-water is demonstrated and the accuracy and precision of the technique are
Analysis of sea-water for heavy metals currently requires concentration and/or matrix separation prior to determination by most instrumental methods of analysis. A number of techniques have been successfully used for this purpose, including co-precipitation
and co-crystallization,” ’ chelation and solvent extraction,3-6 chelating ion-exchange resins,5* ‘-’ and electrolytic concentration. lo By far the most widely used procedure is based on the formation of metal dithiocarbamate complexes and their extraction into an organic solvent.” Recent studies have indicated that the true concentrations of heavy metals in open ocean water are much lower than previously thought, e.g., Cd 0.2 mg/l., Cu 80 ng/l., Zn 4 rig/l.... Pb 3.3 ng/I.,13 necessitating substantial preconcentration before analysis. Because extractions from large volumes are impractical, both from theoretical and physical considerations, there is an upper limit to the preconcentration factor which can be achieved with such single-stage separations. Again, although the application of chelating resins, such as Chelex-100,5S8 to the analysis of sea-water permits higher concentration factors to be attained, the method is time-consuming (sample flowrates l-2 ml/min) and removal of significant amounts of calcium and magnesium from the resin before elution of the trace metals requires careful washing procedures.5* l4 Hence other methods must be sought. In a preliminary report from this laboratory,15 a method based on the formation of metal-8-hydroxyquinoline complexes and their subsequent adsorption onto Cl,-bonded silica gel was described for the preconcentration of Cd, Zn, Cu, Pb, Ni, Mn and Fe from * To whom requests for reprints should be addressed. Crown copyrights reserved.
sea-water. Encouraged by the initial success of this approach, we report here a detailed investigation of the applicability of this method to the determination of Cd, Pb, Cu, Mn, Fe, Ni, Cr and Co in samples of near-shore sea-water. EXPERIMENTAL Apparatus
All atomic-absorption analyses were carried out with a Perkin-Elmer HdA-2200 heated graphite atomizer. Sample aliquots of 10 or 20 ul were delivered to the furnace with‘ a Peikin-Elmer AS-l autosampler and absorbance peak heights were recorded on a fast-response Speed Servo II strip-chart recorder (Esterline Corp., Indiannapolis, ,Ind., U.S.A.). Continuum-source background correction was used. A detailed description of the instrumentation has already been given.5 Figure 1 shows the preconcentration apparatus, consisting of a I-litre polypropylene reservoir connected by a short segment of Bev-A-line IV tubing (Cole Parmer Co., Chicago, Ill.) to a borosilicate glass column in which approximately 600mg of C,s-bonded silica gel (Bondapak, Porasil B, 37-?5 pm, Waters Associates, Milford, MA) was supported by a coarse sintered-glass frit. Reagents
The C,s-bonded silica gel was used without further purification. All acids, solvents and other reagents were purified as previously reported.’ Subsurface samples of nearshore Atlantic sea-water (salinity 29.5%) were collected during late autumn off the coast of Nova Scotia, filtered through a 0.45~pm membrane filter, acidified to pH 1.6 with nitric acid and stored in precleaned polypropylene bottles. Procedure
Sample preparation was done in a clean-laboratory equipped with laminar-flow benches and fume cupboards providing a class 100 working environment. The glass column was slurry-loaded with 600mg of the C,, gel suspended in redistilled methanol. Three lO-ml 167
168
R. E.
STURGEONet al.
of appropriate metal standards in 1% v/v nitric acid as well as by the addition of metal spikes to aliquots of the concentrate in order to effect a matrix match.
RESULTS AND DISCUSSION
Bev-A-line
C,,-Bonded Silica (Icm
gel bed1
Coarse slntered glass frlt
Teflon
-
stopcock _
Fig. 1. Preconcentration
apparatus.
portions of methanol were then passed through the bed under gravity flow to effect clean-up. Before passage of the last 20&300 ~1 of the methanol, distilled, demineralized water (DMW) was carefully added in small (100-200~1) portions, thereby changing the solvent from methanol to water, in the manner of a gradient elution, without disturbing the resin bed. Following complete switch to the aqueous system, the column-was washed with 5 bedvolumes of DMW and left covered with 4-5 ml of DMW. A SN-ml sea-water sample was loaded into the reservoir and I 0.5 ml of 5% purified’ Shydroxyquinoline solution was’added. The pH was adjusted to 8.9 with ammonia solution and the sample was drawn through the C,, bed with a water-aspirator, at a typical flow-rate of about 20ml/min. Visual observation confirmed that the metal chelates were adsorbed on the top layer of the bed to form a thin green band. Following passage of the sample, the column was washed free from sea-water with 10ml of pH-8.9 DMW containing 50 pegof oxine per ml. The adsorbed metal chelates of 8-hydroxyquinoline were then eluted from the column with 5 ml of methanol. This eluate was either collected in a lo-ml standard flask and made up to volume with 2% v/v nitric acid or collected in a “Vycor” crucible and evaporated to dryness in the presence of 200~1 of concentrated nitric acid and 500 ~1 of 60% perchloric acid (this being repeated until a clear, colourless solution was obtained) and thereafter diluted to lO.Oml with 1% v/v nitric acid. These procedures provide a theoretical concentration factor of 50. Blanks were determined by applying the procedure to 40ml of DMW (containing the same absolute amount of 8-hydroxyquinoline as the sample) instead of the sea-water sample. Between runs, the column was reslurried to ensure proper packing of the bed, thoroughly flushed with methanol and changed over to the aqueous system. The concentrates were analysed by graphite-furnace atomic-absorption, with calibration by means of solutions
The method described for the preconcentration of trace metals from sea-water is essentially reversedphase liquid column adsorption chromatography (RPLC). The application of this technique to the separation and concentration of trace amounts of heavy metals from a variety of matrices has recently been reviewed by Schwedt. l6 In RPLC, the stationary phase (in this case the Crs silica gel) is, by definition, less polar than the mobile phase (here, sea-water). Its effectiveness as a trace enrichment method in the present case lies in the fact that manipulation of secondary chemical equilibria in the aqueous mobile phase (formation of metal chelates) permits concentration of these hydrophobic, less polar complexes, at the head of the column. Furthermore, the weak surface attraction energies of the non-polar stationary phase promote rapid mobile-phase equilibration on the column during sample enrichment and elution, as well as gradient regeneration.” Reversed-phase techniques are versatile methods for trace enrichment and have been used on several occasions to separate and/or preconcentrate heavy metals from a variety of matrices.‘s-21 Applications of these techniques to sea-water analysis have been undertaken in both this laboratoryi and, more recently, by Mills and Quinn.22 The latter authors utilized prepacked Ci,, SEP-PAK cartridges (Waters Associates, Milford, MA) to collect both dissolved organic material and copper organometallic complexes from samples of near-shore sea-water. The use of SEP-PAK cartridges may prove very convenient for this application. As 8-hydroxyquinoline is capable of forming insoluble metal complexes with more than 25 elements, advantage may be taken of its non-specific reactivity to complex simultaneously a number of metals present in sea-water. With suitable adjustment of pH and oxine concentration, chelation of the matrix elements (Ca and Mg) can be minimized relative to that of the transition metals.23 Of the elements selected for study, manganese forms the weakest complex with 8-hydroxyquinoline (log p2 = 17.5).24 Initial studies15 indicated that manganese could be quantitatively recovered from sea-water at pH 8.9 with negligible recovery of calcium and magnesium. The recommended pH range for operation of bonded-phase silica gels is 1-9.25 Siloxane bonds are attacked by hydroxide in aqueous solution and the silica substrate must be “protected” by a high density of bonded organic phase in order to avoid dissolution. Although no deleterious effects on the performance of the columns were noted even after extended use at pH 8.9, care was taken to reslurry and repack
Concentration
Table 3. Analysis of sea-water sample 1
Table 1. Absolute blanks Blank, ng*
Cd Zn CU Mn Fe Ni co Pb
Concentration, ng/ml
Acid decomposed concentrate
Methanolic concentrate
Element
169
of trace metals from sea-water
Element Cd Zn cu Mn Fe Ni co
C concen&ion* 0.030 f 0.30 * 0.97 f 0.78 + 0.9 f 0.26 k 0.0199
0.002 0.07 0.02 0.02 0.1 0.01
Accepted valuet 0.027 f 0.41 + 0.96 + 0.68 f 1.03 * 0.31 f 0.015 *
0.003 0.05 0.04 0.05 0.04 0.04 0.007
10
*Mean and standard deviation for 12 determinations. tSkew distribution.
the bed after elution of each sample, because of the possibility of the silica dissolving under these conditions which might cause shifting and channelling of the bed. Analytical blanks. Absolute blanks are presented in Table 1. Data are given for both the direct procedure, wherein the methanolic concentrates were directly analysed, and also following total acid decomposition of the blank eluate with dissolution of the residue in 1% v/v nitric acid. Blank values for the latter procedure are significantly higher for lead and iron, reflecting the consequences of increased sample manipulation. Furthermore, the precision of the blank determinations for acid decomposition procedures is seriously poorer than that for direct analysis of the methanolic concentrates. Whereas reagent impurities can account for the blank values obtained for Cd, Mn, Ni, Co and Pb,5 the Cl,-bonded silica gel is probably the source of the high levels of Zn, Cu and Fe. Semiquantitative d.c. arc spectrographic analysis of this material revealed the presence of substantial concentrations sf iron (0.06% w/w); lower concentrations of copper (0.01% w/w) and zinc (0.01% w/w) were also present. Even slight leaching of the silica substrate at high pH during each
* Mean and standard deviation for triplicate analyses. t Values accepted by this laboratory fol-
lowing extensive analysis methods6 5 Single determination.
by
several
sample run could easily account for the high blank from these elements. The low analytical blanks obtained in this study make the technique attractive for sea-water analysis. This is a significant advantage over use of activated charcoal23 as the adsorption agent for oxine complexes and organometallic species.22 Recovery of spikes. The efficiency of recovery of spikes added to both DMW and sea-water samples is shown in Table 2. With the exception of cobalt, all elements of interest could be quantitatively recovered from DMW. From sea-water media, however, recovery of lead was erratic, varying from 17 to 88%. As a consequence, the technique could not be used for the determination of lead in sea-water. The reproducibility of the technique was slightly poorer for all elements when sea-water was analysed and an isolated check for chromium indicated that recovery of this element was less than 10%. Recovery of spikes from sea-water did not vary over the range of sample flow-rates 8-35 ml/min. Analytical results. Results for the analysis of two near-shore sea-water samples are given in Tables 3 and 4. The samples were filtered during collection and
Table 2. Recovery of metal spikes Recovery, %* Table 4. Analysis of sea-water sample 2 Element Cd Zn cu Mn Fe Ni co Pb
Demineralized water 94 * 102 f 104 + 95 * 98 + 105 f 65 + 94 f
10 13 4 5 7 11 5 5
Concentration, ng/ml
Sea-water 108 & 18 97k 12 91 + 12 99 + 7 104 f 8 99 f 20 67 * 9 17-88
* Mean and standard deviation for 4 DMW samples and 10 sea-water samples (sample preconcentration S&100-fold).
Element Cd Zn cu Mn Fe Ni co
C concent%ion* 0.031 f 0.002 0.4 & 0.1 0.11 + 0.02 0.98 + 0.05 7.7 + 0.5 0.46 & 0.06 0.017 * 0.001
Accepted valuet 0.025 + 0.001 0.28 f 0.01 0.13 + 0.01 1.06 f 0.02 6.9 k 0.2 0.39 * 0.01 0.017 &-0,002
*t See footnotes to Table 3.
170
R. E. STURGEON et al.
acidified to pH 1.6 with nitric acid for storage purposes. Each sample was analysed in triplicate following concentration (N-fold) from 500-ml portions of seawater. As a result of the lower blanks obtained by direct analysis of the methanolic concentrates (Table 1), all analyses were completed in this manner. Calibration against metal standards in 1% nitric acid or in 1% v/v nitric acid in 50% aqueous methanol was not successful, as the presence of oxine in the solution (0.3 mg/ml in the methanolic concentrates) produced significant positive interferences (signal enhancements of up to lOOo/,).Calibration was achieved by spiking a sample of the concentrate with the element of interest, thereby obtaining an exact matrix match (standard addition method). Apart frog the oxine, no other matrix constituents were present. Rejection of calcium and magnesium was excellent, there being less than 1 pg/ml (Ca) and approximately 2Opg/ml (Mg) in the concentrates. In calculation of the results account was taken of the 65% recovery of cobalt. For all other elements, quantitative recovery was assumed (Table 2). Data obtained by using the C1s concentration procedure are compared with “accepted” values for these samples (Tables 3 and 4). The “accepted” values were obtained by exhaustive analysis by several different techniques,‘j including isotope-dilution spark-source mass spectrometryz6 and are believed to be quite reliable. Application of the t-test (at the 95% confidence level) shows no significant difference between the results obtained by the C1 s adsorption procedrire and the accepted values. The precision of analysis averages 10% RSD (range of RSD being 4-25%) for the two samples and is generally worse for those elements for which blanks were relatively large (zinc and copper). CONCLUSIONS
The ability of 8-hydroxyquinoline to form chelates with a large number of elements, coupled with their non-selective adsorption onto C,s-bonded silica gel, provides a versatile, multielement concentration technique for trace metals in sea-water. The C,s-bonded gel may be reused repeatedly under the conditions cited in these experiments, thereby maintaining low operating costs. Additionally, all reagents are easily purified. The relatively fast sample processing (up to 35 ml/min) and minimum of sample manipulation permit the rapid attainment of high concentration factor%’ 5 contributing to its potential use as an on-site preconcentration procedure. Blank levels ultimately determine practical detection limits, and these can potentially be lowered by synthesis of a C,,-bonded phase with high-purity silica gel as the substrate, and perhaps by operating the column at lower pH (to minimize hydrolysis of the silica).
RPLC with C,s-bonded silica gel may also prove very useful for preconcentrating trace organics from sea-water, including that portion binding trace metals, thereby permitting speciation studies. A recent study by Mills and Quinn attests to the feasibility of this application.22 Future application of liquid column chromatography to the concentration of trace elements from sea-water could prove more rewarding if the chelation were performed within a column bed containing immobilized I-hydroxyquinoline. Work is in progress in this direction. AcknowledgementbThe authors thank R. Guevremont of the Atlantic Research Laboratory, National Research Council of Canada, for the sea-water samples, and V. P. Clancy for spectrographic analyses of the Cl8 gel.
REFERENCES
1. E. A. Boyle and J. M. Edmond, in Analytical Methods in Oceanoaraohv. T. R. P. Gibb Jr. ted.). D. 44. ACS Advances -in ‘Chemistry No. 147, W‘ash&ton, DC., 1975. 2. V. Hudnik, S. GomigEek and B. Gorenc, Anal. Chim. Acta, 1978,98,
39.
3. T. K. Jan and D. R. Young, Anal. Chem., 1978, 50, 1250. 4. L. Danielson, B. Magnusson and S. Westerlund, Anal. Chim. Acta, 1978, 98, 47.
5. R. E. Sturgeon, S. S. Berman, A. Desaulniers and D. S. Russell, Talanta, 1980, 27, 85. 6. R. E. Sturgeon, S. S. Berman, J. A. H. Desaulniers, A. P. Mykytiuk, J. W. McLaren and D. S. Russell, Anal. Chem., 1980,52, 1585. 7. J. P. Riley and D. Taylor, Anal. Chim. Acta, 1968, 40, 479. 8. T. M. Florence and G. E. Batley, Talanta, 1976, 23, 179. 9. P. Pakalns, G. E. Batley and A. J. Cameron, Anal. Chim. Acta, 1978,99, 333. 10. yjg Batley and J. P. Matousek, Anal. Chem., 1980, 52, 11, J. M. Bewers and H. L. Windom, Preliminary Scientific Renort of the IOCIWMONNEP Workshon on the It&calibration of’ Sampling Procedures, Bermuda, 1980; to be published by the Intergovernmental Oceanographic Commission. 12. K. W. Bruland and R. P. Franks, paper presented at the conference on “Trace Metals in Sea-water”, NATO Advanced Research Institute, Erice, Sicily, 30 March-l April 1981. 13. B. Schaule and C. C. Paterson, paper presented at “Trace Metals in Sea-water” conference (reference 12 above). 14. H. M. Kingston, I. L. Barnes, T. J. Brady, T. C. Rains and A. M. Champ, Anal. Chem., 1978, SO, 2064. 15. H. Watanabe, K. Goto, S. Taguchi, J. W. McLaren, S. S. Berman and D. S. Russell, ibid., 1981,53, 738. 16. G. Schwedt, in Topics in Current Chemistry, Instrumental Inorganic Chemistry, Vol. 85, p. 159. SpringerVerlag, New York, 1979. 17. N. H. C. Cooke and K. Olsen, J. Chromatog. Sci., 1980, 18, 1. 18. I. P. Alimarin and T. A. Bolshova, Pure Appl. Chem. 1972, 31, 493. 19. I. P. Alimarin, T. A. Bolshova, N. I. Ershova and M. B. Polinskaya, Zh. Analit. Khim., 1969, 24, 26. 20. G. Schwedt, Chromatographia, 1977. 11, 145.
Concentration
of trace metals from sea-water
21. P. C. Uden, D. M. Parees and F. H. Walters, Anal. Lett., 1975, 8, 795. 22. G. L. Mills and J. G. Quinn, Mar. Chem., 1981, 10.93. 23. B. M. Vanderborght and R. E. Van Grieken, Anal. Chem., 1977,49,311.
24. W. D. Johnston 1952,74, 5239. 25. R. E. Majors, J. 26. A. P. Mykytiuk, Chem., 1980,52,
171 and H. Freiser, .I. Am. Chem. Sot., Chromtog. Sci., 1980, 18, 488.. D. S. Russell and R. E. Sturgeon, Anal. 1281.
Vol. 29, pp. 167 to 171, 1982 in Great Britain. All rights reserved
Copyright
~39-9140/82/030167-051o3.oo/0 0 1982 Pergamon Press Ltd
CONCENTRATION OF TRACE METALS FROM SEA-WATER BY COMPLEXATION WITH 8-HYDROXYQUINOLINE AND ADSORPTION ON Cl ,-BONDED SILICA GEL R. E. STURGEON*,S. S. BERMANand S. N. WILLIE Division of Chemistry, National Research Council of Canada, Montreal Road, Ottawa, Ontario, Canada (Received 29 June 1981. Accepted 15 September 1981) Snmmary-A reversed-phase liquid chromatographic technique based on a combination of multielement chelation by 8-hydroxyquinoline with subsequent adsorption on C I.s-bonded silica gel is described for the concentration of Cd, Zn, Cu, Ni, Co, Mn and Fe from sea-water. Enrichment factors of 5(rlOO are readily obtained following elution of the absorbate with methanol to provide a matrix-free concentrate suitable for graphite-furnace atomic-absorption analysis. Quantitative recovery of these elements from near-shore samples of sea-water is demonstrated and the accuracy and precision of the technique are
Analysis of sea-water for heavy metals currently requires concentration and/or matrix separation prior to determination by most instrumental methods of analysis. A number of techniques have been successfully used for this purpose, including co-precipitation
and co-crystallization,” ’ chelation and solvent extraction,3-6 chelating ion-exchange resins,5* ‘-’ and electrolytic concentration. lo By far the most widely used procedure is based on the formation of metal dithiocarbamate complexes and their extraction into an organic solvent.” Recent studies have indicated that the true concentrations of heavy metals in open ocean water are much lower than previously thought, e.g., Cd 0.2 mg/l., Cu 80 ng/l., Zn 4 rig/l.... Pb 3.3 ng/I.,13 necessitating substantial preconcentration before analysis. Because extractions from large volumes are impractical, both from theoretical and physical considerations, there is an upper limit to the preconcentration factor which can be achieved with such single-stage separations. Again, although the application of chelating resins, such as Chelex-100,5S8 to the analysis of sea-water permits higher concentration factors to be attained, the method is time-consuming (sample flowrates l-2 ml/min) and removal of significant amounts of calcium and magnesium from the resin before elution of the trace metals requires careful washing procedures.5* l4 Hence other methods must be sought. In a preliminary report from this laboratory,15 a method based on the formation of metal-8-hydroxyquinoline complexes and their subsequent adsorption onto Cl,-bonded silica gel was described for the preconcentration of Cd, Zn, Cu, Pb, Ni, Mn and Fe from * To whom requests for reprints should be addressed. Crown copyrights reserved.
sea-water. Encouraged by the initial success of this approach, we report here a detailed investigation of the applicability of this method to the determination of Cd, Pb, Cu, Mn, Fe, Ni, Cr and Co in samples of near-shore sea-water. EXPERIMENTAL Apparatus
All atomic-absorption analyses were carried out with a Perkin-Elmer HdA-2200 heated graphite atomizer. Sample aliquots of 10 or 20 ul were delivered to the furnace with‘ a Peikin-Elmer AS-l autosampler and absorbance peak heights were recorded on a fast-response Speed Servo II strip-chart recorder (Esterline Corp., Indiannapolis, ,Ind., U.S.A.). Continuum-source background correction was used. A detailed description of the instrumentation has already been given.5 Figure 1 shows the preconcentration apparatus, consisting of a I-litre polypropylene reservoir connected by a short segment of Bev-A-line IV tubing (Cole Parmer Co., Chicago, Ill.) to a borosilicate glass column in which approximately 600mg of C,s-bonded silica gel (Bondapak, Porasil B, 37-?5 pm, Waters Associates, Milford, MA) was supported by a coarse sintered-glass frit. Reagents
The C,s-bonded silica gel was used without further purification. All acids, solvents and other reagents were purified as previously reported.’ Subsurface samples of nearshore Atlantic sea-water (salinity 29.5%) were collected during late autumn off the coast of Nova Scotia, filtered through a 0.45~pm membrane filter, acidified to pH 1.6 with nitric acid and stored in precleaned polypropylene bottles. Procedure
Sample preparation was done in a clean-laboratory equipped with laminar-flow benches and fume cupboards providing a class 100 working environment. The glass column was slurry-loaded with 600mg of the C,, gel suspended in redistilled methanol. Three lO-ml 167
168
R. E.
STURGEONet al.
of appropriate metal standards in 1% v/v nitric acid as well as by the addition of metal spikes to aliquots of the concentrate in order to effect a matrix match.
RESULTS AND DISCUSSION
Bev-A-line
C,,-Bonded Silica (Icm
gel bed1
Coarse slntered glass frlt
Teflon
-
stopcock _
Fig. 1. Preconcentration
apparatus.
portions of methanol were then passed through the bed under gravity flow to effect clean-up. Before passage of the last 20&300 ~1 of the methanol, distilled, demineralized water (DMW) was carefully added in small (100-200~1) portions, thereby changing the solvent from methanol to water, in the manner of a gradient elution, without disturbing the resin bed. Following complete switch to the aqueous system, the column-was washed with 5 bedvolumes of DMW and left covered with 4-5 ml of DMW. A SN-ml sea-water sample was loaded into the reservoir and I 0.5 ml of 5% purified’ Shydroxyquinoline solution was’added. The pH was adjusted to 8.9 with ammonia solution and the sample was drawn through the C,, bed with a water-aspirator, at a typical flow-rate of about 20ml/min. Visual observation confirmed that the metal chelates were adsorbed on the top layer of the bed to form a thin green band. Following passage of the sample, the column was washed free from sea-water with 10ml of pH-8.9 DMW containing 50 pegof oxine per ml. The adsorbed metal chelates of 8-hydroxyquinoline were then eluted from the column with 5 ml of methanol. This eluate was either collected in a lo-ml standard flask and made up to volume with 2% v/v nitric acid or collected in a “Vycor” crucible and evaporated to dryness in the presence of 200~1 of concentrated nitric acid and 500 ~1 of 60% perchloric acid (this being repeated until a clear, colourless solution was obtained) and thereafter diluted to lO.Oml with 1% v/v nitric acid. These procedures provide a theoretical concentration factor of 50. Blanks were determined by applying the procedure to 40ml of DMW (containing the same absolute amount of 8-hydroxyquinoline as the sample) instead of the sea-water sample. Between runs, the column was reslurried to ensure proper packing of the bed, thoroughly flushed with methanol and changed over to the aqueous system. The concentrates were analysed by graphite-furnace atomic-absorption, with calibration by means of solutions
The method described for the preconcentration of trace metals from sea-water is essentially reversedphase liquid column adsorption chromatography (RPLC). The application of this technique to the separation and concentration of trace amounts of heavy metals from a variety of matrices has recently been reviewed by Schwedt. l6 In RPLC, the stationary phase (in this case the Crs silica gel) is, by definition, less polar than the mobile phase (here, sea-water). Its effectiveness as a trace enrichment method in the present case lies in the fact that manipulation of secondary chemical equilibria in the aqueous mobile phase (formation of metal chelates) permits concentration of these hydrophobic, less polar complexes, at the head of the column. Furthermore, the weak surface attraction energies of the non-polar stationary phase promote rapid mobile-phase equilibration on the column during sample enrichment and elution, as well as gradient regeneration.” Reversed-phase techniques are versatile methods for trace enrichment and have been used on several occasions to separate and/or preconcentrate heavy metals from a variety of matrices.‘s-21 Applications of these techniques to sea-water analysis have been undertaken in both this laboratoryi and, more recently, by Mills and Quinn.22 The latter authors utilized prepacked Ci,, SEP-PAK cartridges (Waters Associates, Milford, MA) to collect both dissolved organic material and copper organometallic complexes from samples of near-shore sea-water. The use of SEP-PAK cartridges may prove very convenient for this application. As 8-hydroxyquinoline is capable of forming insoluble metal complexes with more than 25 elements, advantage may be taken of its non-specific reactivity to complex simultaneously a number of metals present in sea-water. With suitable adjustment of pH and oxine concentration, chelation of the matrix elements (Ca and Mg) can be minimized relative to that of the transition metals.23 Of the elements selected for study, manganese forms the weakest complex with 8-hydroxyquinoline (log p2 = 17.5).24 Initial studies15 indicated that manganese could be quantitatively recovered from sea-water at pH 8.9 with negligible recovery of calcium and magnesium. The recommended pH range for operation of bonded-phase silica gels is 1-9.25 Siloxane bonds are attacked by hydroxide in aqueous solution and the silica substrate must be “protected” by a high density of bonded organic phase in order to avoid dissolution. Although no deleterious effects on the performance of the columns were noted even after extended use at pH 8.9, care was taken to reslurry and repack
Concentration
Table 3. Analysis of sea-water sample 1
Table 1. Absolute blanks Blank, ng*
Cd Zn CU Mn Fe Ni co Pb
Concentration, ng/ml
Acid decomposed concentrate
Methanolic concentrate
Element
169
of trace metals from sea-water
Element Cd Zn cu Mn Fe Ni co
C concen&ion* 0.030 f 0.30 * 0.97 f 0.78 + 0.9 f 0.26 k 0.0199
0.002 0.07 0.02 0.02 0.1 0.01
Accepted valuet 0.027 f 0.41 + 0.96 + 0.68 f 1.03 * 0.31 f 0.015 *
0.003 0.05 0.04 0.05 0.04 0.04 0.007
10
*Mean and standard deviation for 12 determinations. tSkew distribution.
the bed after elution of each sample, because of the possibility of the silica dissolving under these conditions which might cause shifting and channelling of the bed. Analytical blanks. Absolute blanks are presented in Table 1. Data are given for both the direct procedure, wherein the methanolic concentrates were directly analysed, and also following total acid decomposition of the blank eluate with dissolution of the residue in 1% v/v nitric acid. Blank values for the latter procedure are significantly higher for lead and iron, reflecting the consequences of increased sample manipulation. Furthermore, the precision of the blank determinations for acid decomposition procedures is seriously poorer than that for direct analysis of the methanolic concentrates. Whereas reagent impurities can account for the blank values obtained for Cd, Mn, Ni, Co and Pb,5 the Cl,-bonded silica gel is probably the source of the high levels of Zn, Cu and Fe. Semiquantitative d.c. arc spectrographic analysis of this material revealed the presence of substantial concentrations sf iron (0.06% w/w); lower concentrations of copper (0.01% w/w) and zinc (0.01% w/w) were also present. Even slight leaching of the silica substrate at high pH during each
* Mean and standard deviation for triplicate analyses. t Values accepted by this laboratory fol-
lowing extensive analysis methods6 5 Single determination.
by
several
sample run could easily account for the high blank from these elements. The low analytical blanks obtained in this study make the technique attractive for sea-water analysis. This is a significant advantage over use of activated charcoal23 as the adsorption agent for oxine complexes and organometallic species.22 Recovery of spikes. The efficiency of recovery of spikes added to both DMW and sea-water samples is shown in Table 2. With the exception of cobalt, all elements of interest could be quantitatively recovered from DMW. From sea-water media, however, recovery of lead was erratic, varying from 17 to 88%. As a consequence, the technique could not be used for the determination of lead in sea-water. The reproducibility of the technique was slightly poorer for all elements when sea-water was analysed and an isolated check for chromium indicated that recovery of this element was less than 10%. Recovery of spikes from sea-water did not vary over the range of sample flow-rates 8-35 ml/min. Analytical results. Results for the analysis of two near-shore sea-water samples are given in Tables 3 and 4. The samples were filtered during collection and
Table 2. Recovery of metal spikes Recovery, %* Table 4. Analysis of sea-water sample 2 Element Cd Zn cu Mn Fe Ni co Pb
Demineralized water 94 * 102 f 104 + 95 * 98 + 105 f 65 + 94 f
10 13 4 5 7 11 5 5
Concentration, ng/ml
Sea-water 108 & 18 97k 12 91 + 12 99 + 7 104 f 8 99 f 20 67 * 9 17-88
* Mean and standard deviation for 4 DMW samples and 10 sea-water samples (sample preconcentration S&100-fold).
Element Cd Zn cu Mn Fe Ni co
C concent%ion* 0.031 f 0.002 0.4 & 0.1 0.11 + 0.02 0.98 + 0.05 7.7 + 0.5 0.46 & 0.06 0.017 * 0.001
Accepted valuet 0.025 + 0.001 0.28 f 0.01 0.13 + 0.01 1.06 f 0.02 6.9 k 0.2 0.39 * 0.01 0.017 &-0,002
*t See footnotes to Table 3.
170
R. E. STURGEON et al.
acidified to pH 1.6 with nitric acid for storage purposes. Each sample was analysed in triplicate following concentration (N-fold) from 500-ml portions of seawater. As a result of the lower blanks obtained by direct analysis of the methanolic concentrates (Table 1), all analyses were completed in this manner. Calibration against metal standards in 1% nitric acid or in 1% v/v nitric acid in 50% aqueous methanol was not successful, as the presence of oxine in the solution (0.3 mg/ml in the methanolic concentrates) produced significant positive interferences (signal enhancements of up to lOOo/,).Calibration was achieved by spiking a sample of the concentrate with the element of interest, thereby obtaining an exact matrix match (standard addition method). Apart frog the oxine, no other matrix constituents were present. Rejection of calcium and magnesium was excellent, there being less than 1 pg/ml (Ca) and approximately 2Opg/ml (Mg) in the concentrates. In calculation of the results account was taken of the 65% recovery of cobalt. For all other elements, quantitative recovery was assumed (Table 2). Data obtained by using the C1s concentration procedure are compared with “accepted” values for these samples (Tables 3 and 4). The “accepted” values were obtained by exhaustive analysis by several different techniques,‘j including isotope-dilution spark-source mass spectrometryz6 and are believed to be quite reliable. Application of the t-test (at the 95% confidence level) shows no significant difference between the results obtained by the C1 s adsorption procedrire and the accepted values. The precision of analysis averages 10% RSD (range of RSD being 4-25%) for the two samples and is generally worse for those elements for which blanks were relatively large (zinc and copper). CONCLUSIONS
The ability of 8-hydroxyquinoline to form chelates with a large number of elements, coupled with their non-selective adsorption onto C,s-bonded silica gel, provides a versatile, multielement concentration technique for trace metals in sea-water. The C,s-bonded gel may be reused repeatedly under the conditions cited in these experiments, thereby maintaining low operating costs. Additionally, all reagents are easily purified. The relatively fast sample processing (up to 35 ml/min) and minimum of sample manipulation permit the rapid attainment of high concentration factor%’ 5 contributing to its potential use as an on-site preconcentration procedure. Blank levels ultimately determine practical detection limits, and these can potentially be lowered by synthesis of a C,,-bonded phase with high-purity silica gel as the substrate, and perhaps by operating the column at lower pH (to minimize hydrolysis of the silica).
RPLC with C,s-bonded silica gel may also prove very useful for preconcentrating trace organics from sea-water, including that portion binding trace metals, thereby permitting speciation studies. A recent study by Mills and Quinn attests to the feasibility of this application.22 Future application of liquid column chromatography to the concentration of trace elements from sea-water could prove more rewarding if the chelation were performed within a column bed containing immobilized I-hydroxyquinoline. Work is in progress in this direction. AcknowledgementbThe authors thank R. Guevremont of the Atlantic Research Laboratory, National Research Council of Canada, for the sea-water samples, and V. P. Clancy for spectrographic analyses of the Cl8 gel.
REFERENCES
1. E. A. Boyle and J. M. Edmond, in Analytical Methods in Oceanoaraohv. T. R. P. Gibb Jr. ted.). D. 44. ACS Advances -in ‘Chemistry No. 147, W‘ash&ton, DC., 1975. 2. V. Hudnik, S. GomigEek and B. Gorenc, Anal. Chim. Acta, 1978,98,
39.
3. T. K. Jan and D. R. Young, Anal. Chem., 1978, 50, 1250. 4. L. Danielson, B. Magnusson and S. Westerlund, Anal. Chim. Acta, 1978, 98, 47.
5. R. E. Sturgeon, S. S. Berman, A. Desaulniers and D. S. Russell, Talanta, 1980, 27, 85. 6. R. E. Sturgeon, S. S. Berman, J. A. H. Desaulniers, A. P. Mykytiuk, J. W. McLaren and D. S. Russell, Anal. Chem., 1980,52, 1585. 7. J. P. Riley and D. Taylor, Anal. Chim. Acta, 1968, 40, 479. 8. T. M. Florence and G. E. Batley, Talanta, 1976, 23, 179. 9. P. Pakalns, G. E. Batley and A. J. Cameron, Anal. Chim. Acta, 1978,99, 333. 10. yjg Batley and J. P. Matousek, Anal. Chem., 1980, 52, 11, J. M. Bewers and H. L. Windom, Preliminary Scientific Renort of the IOCIWMONNEP Workshon on the It&calibration of’ Sampling Procedures, Bermuda, 1980; to be published by the Intergovernmental Oceanographic Commission. 12. K. W. Bruland and R. P. Franks, paper presented at the conference on “Trace Metals in Sea-water”, NATO Advanced Research Institute, Erice, Sicily, 30 March-l April 1981. 13. B. Schaule and C. C. Paterson, paper presented at “Trace Metals in Sea-water” conference (reference 12 above). 14. H. M. Kingston, I. L. Barnes, T. J. Brady, T. C. Rains and A. M. Champ, Anal. Chem., 1978, SO, 2064. 15. H. Watanabe, K. Goto, S. Taguchi, J. W. McLaren, S. S. Berman and D. S. Russell, ibid., 1981,53, 738. 16. G. Schwedt, in Topics in Current Chemistry, Instrumental Inorganic Chemistry, Vol. 85, p. 159. SpringerVerlag, New York, 1979. 17. N. H. C. Cooke and K. Olsen, J. Chromatog. Sci., 1980, 18, 1. 18. I. P. Alimarin and T. A. Bolshova, Pure Appl. Chem. 1972, 31, 493. 19. I. P. Alimarin, T. A. Bolshova, N. I. Ershova and M. B. Polinskaya, Zh. Analit. Khim., 1969, 24, 26. 20. G. Schwedt, Chromatographia, 1977. 11, 145.
Concentration
of trace metals from sea-water
21. P. C. Uden, D. M. Parees and F. H. Walters, Anal. Lett., 1975, 8, 795. 22. G. L. Mills and J. G. Quinn, Mar. Chem., 1981, 10.93. 23. B. M. Vanderborght and R. E. Van Grieken, Anal. Chem., 1977,49,311.
24. W. D. Johnston 1952,74, 5239. 25. R. E. Majors, J. 26. A. P. Mykytiuk, Chem., 1980,52,
171 and H. Freiser, .I. Am. Chem. Sot., Chromtog. Sci., 1980, 18, 488.. D. S. Russell and R. E. Sturgeon, Anal. 1281.