- Email: [email protected]
Ion chromatographic determination of selected ions in antarctic ice
Analytica Chimica Acta, 194 (1987) 281-286 Elsevier Science Publishers B.V., Amsterdam -Printed
in The Netherlands
Short Communication
ION CHROMATOGRAPHIC IN ANTARCTIC ICE
DETERMINATION OF SELECTED IONS
J. P. IVEY*’ Australian Government (A ustmlia)
Analytical
Labomtories,
P.O. Box 84, Kingston,
Tasmania
7150
D. M. DAVIES CSIRO Division (Australia)
of Atmospheric
Research,
Private Bag 1, Mordialloc,
Victoria 3195
(Received 3rd January 1986)
Summary. Ion chromatography is used to measure the concentrations of chloride, nitrate, sulphate, ammonium and sodium ions at the pg 1-l level in Antarctic ice and to investigate the occurrence of methanesulphonate, fluoride, formate, acetate and nitrite. Of the latter group of ions, only methanesulphonate was found in measurable concentrations.
The composition of Antarctic ice, in particular its soluble impurities, has been used to give insight into such questions as paleoatmospheric composition [ 11, and the natural cycling of nitrogen and sulphur compounds [l, 21. Recently, the presence of biogenic methanesulphonic ,and other organic acids in the maritime environment [ 3-51 has raised questions as to their presence or otherwise in Antarctic ice. If they are present, they can be expected to contribute to ice acidity and in the case of methanesulphonic acid to the sulphur budget [6]. Methanesulphonate measurements may also aid in the definition of the relative sulphate contributions from marine biogenic productivity and volcanic activity [l] . The sample preparation protocol is of fundamental importance in the eventual reliability of the analytical results. The problems associated with gaseous species in the preparation of ice for analysis have been noted in earlier work [ 71. In this communication, the manner in which the preparation problem was overcome is discussed as well as the optimization of the instrumentation for the determination of each ion, or group of ions. Experimental Sample preparation. Ice cores were collected as part of the Australian National Research Expedition (ANARE) glaciology research program at Law Dome (66 30’S, 111 OO’E). The core was cut into 15 equal (depth) sections *Present address: Department of Analytical Chemistry, University of New South Wales, P.O. Box 1, Kensington, N.S.W. 2033, Australia. 0003-2670/87/$03.50
o 1987 Elsevier Science Publishers B.V.
for each annual accumulation layer as shown by the oxygen isotope record. Each section was then trimmed to leave ca. 200 g of the inner part of the core. The method of washing and melting the 200-g sections was developed in three stages, only results obtained from the final two methods were utilized elsewhere [8] . All manipulations were done in a clean air laminar flow cabinet. All plastic ware was washed copiously with distilled/deionized water and then filled with distilled/deionized water, sealed, and allowed to stand for at least 24 h prior to use. The three stages were as follows. (1) Initially about 25% of the 200-g block was removed by washing in distilled/deionized water (resistivity > 18 M ohm cme2) after which a further 25% was removed by placing the washed ice in a polyethylene funnel and allowing the ice to melt as described by Legrand et al. [7]. Chloroform (multiply extracted with distilled/deionized water) was added to a polyethylene receiver vessel. The purpose of this was to preserve any organic acids which have been shown to be labile in the absence of a biocide [5]. Frozen distilled/deionized water taken through the procedure, however, showed disconcerting levels of acetate, formate and traces of nitrite by ion chromatography. The clean air cabinet was obviously allowing acidic (and presumably basic) gases through its filter, with the potential further to contaminate the melting sample with nitric acid and sulphurous/sulphuric acids and ammonia. (2) Air from a compressed air source was cleaned by passage through an in-line gas absorption tube containing Ascarite (8-20 mesh). The scrubbed air was then piped to an inverted funnel over the polyethylene funnel used to contain the melting ice (Fig. 1). A gas absorption tube packed with oxalic acid was included in the same line for removing ammonia. A 0.45-pm filter was last in line. The effectiveness of this can be seen in Fig. 2, for which two different samples were allowed to melt in the laminar flow cabinet (A) and the scrubbed-air (B) melt systems. (3) When the above system was used, any formate and acetate present were below the detection limit of the present method and thus no chloroform was required. The ice was washed with distilled/deionized water until 50% remained which was then simply melted in a sealed polyethylene bag (Whirlpak). The Whirlpaks were checked for possible contaminants by running blank determinations with distilled/deionized water. Equipment. The ion-chromatographic system for the determination of anions consisted of a Waters Associates M45 pump, a Rheodyne fixedvolume injector (model 7000), a Dionex 030985 fast-run anion separator column, a Dionex 035350 hollow-fibre or Dionex 030829 suppressor column and a Waters model 430 conductivity detector. The same instrumentation was used for the determination of sodium and ammonium except that a Wescan 269004 cation separator column and a Dionex cation suppressor column were used. As indicated by Legrand et al. [7] a large (5.5-ml) sample loop was found to be preferable to a concentrator column for the introduction of samples with low concentrations of analyte species. Samples
283
Fig. 1. Air-purification system used in the melting of ice cores: (1) compressed air source; (2) acid gas trap of ascarite (8-20 mesh); (3) alkali gas trap of oxalic acid; (4) self-indicating soda lime; (5) particle filter; (6) laminar flow clean cabinet; (7) polyethylene receiving bottle.
C
NLDRIDE
0
Fig. 2. Chromatograms of sample ice. (A) Ice melted in a clean air cabinet without contaminating gas removal showing the separation of acetate/fluoride, formate, methanesulphonate and chloride at respective concentrations of about 105, 5.8 and 50 ng ml-‘. (B) Ice melted after contaminating gas removal; methanesulphonate and chloride at respective concentrations of 16.7 and 310 ng ml-‘. (C) Sample ice with nitrate and sulphate concentrations of 17 and 31 ng ml-‘, respectively. Conditions: injection volume of 5.5 ml, flow rate 2.0 ml min-’ , chart speed 0.25 cm min-’ , eluents as in Table 1.
234
were loaded into the loop with a cleaned lo-ml polyethylene syringe. Eluent concentrations and conditions are shown in Table 1. Comparative-measurements of sodium and determinations of magnesium were done on a Varian atomic absorption spectrometer model 1475 with an air/acetylene flame. Results and discussion Legrand et al. [7] used an eluent for nitrate and sulphate which did not allow chloride to be quantified. They chose to use an alternative column specifically for chloride determinations. Here, the same column was used for all determinations but with a more dilute eluent for chloride. In the method of Legrand et al. [ 71, chloride elutes at a retention time of less than 2 min after passage of the water dip. Acetate, formate, methanesulphonate and fluoride all elute prior to chloride. Fluoride has been reported previously as being present in Antarctic ice but accurate quantitation was not possible because of interference from the water dip [l] . To examine the possible existence of these poorly retained anions, a very dilute eluent (0.0007 M sodium hydrogen carbonate) was used. Figure 2B is a typical sample chromatogram and Fig. 2A shows the degree of separation achieved. Fluoride was resolved from acetate though not adequately; but because no peak was observed at the acetate/fluoride retention time in any sample, this was not a problem. Chloride was also quantified with this system by using very high attenuation. Methanesulphonate ion was found in all samples examined. The concentration was between 9 and 0.2 pg 1-l as sulphur (S); the latter is the detection limit at twice the baseline noise. Standards with concentrations of 1.0, 2.5, 5.0, 7.5 and 10.0 fig 1-l (S) typically gave a linear peak height vs. TABLE 1 Summary of conditions for quantifying selected ions in Antarctic icea Ion
Eluent
Detection limit (ng)
Range of standards ( ng ml-’ )
Sodium Ammonium Fluoride/acetate Formate Methane sulphonate Chloride Nitrite Nitrate Sulphate
0.004 M HNO,
1 2 5120 20 1 2 4 2 3
O-500 O-50 -
0.0007
M NaHCO,
0.002 M Na,CO,
O-100 O-30 o-So0 o-5 O-25 O-l 50
‘Injection volume 5.5 ml, flow rate 2.0 ml min-’ ; columns etc. as discussed in text; sensitivity 5 ~1s f.s.d. for cations, 1 US and 5 $!l for chloride, with both 0.0007 M hydrogen carbonate and 0.002 M carbonate eluents.
285
mass calibration with a correlation coefficient of 0.9998, a standard error of 0.14 and an intercept of 0.02 pg 1-l (S). Nitrate and sulphate were both quantified by using 0.002 M sodium carbonate as eluent (Fig. 2C). Nitrite elutes between chloride and nitrate but was not observed in samples melted in scrubbed air or samples melted in polyethylene bags. Detection limits are given in Table 1. Sodium and ammonia were quantified by ion chromatography. Potassium was not measured in the present study. Comparison of the sodium results from ion chromatography and those obtained by a.a.s. on 30 samples is shown in Fig. 3. In general, the agreement between the two methods is reasonable (r = 0.988) and certainly comparable to a similar comparison between ion chromatographic and neutron activation measurements of sodium content [ 71. Accurate sodium and magnesium results are particularly important as they are used to define the relative contributions made to the sulphate budget from sea salt and gas to particle conversion [9]. The a.a.s. measurements of sodium and magnesium were done without the addition of an ionization suppressant to either standards or samples in order to avoid contamination of the samples. Ammonium concentrations were in the range O-l pg 1-l ; most samples were at, or about, the detection limit of 0.5 pg 1-l for the equipment used. The relative absence of ammonia from our samples indicates that the methodology for sample preparation is sound. Presence of ammonia has been found to be indicative of sample contamination [7] . Conclusion Whilst it is always preferable to be able to quantify all the species of interest from a single chromatogram, this is not practicable in the determination of anions in Antarctic ice. Valuable information may be lost if the concen-
Fig. 3. Comparison of sodium concentration obtained by a.a.s. (x axis) and ion chromatography (y axis) of 30 ice-core samples (ng ml-‘). Con-. coeff. 0.988, slope 1.034 and x intercept 8.8.
286
trations of early eluting species are not measured. The methodology presented above allows such measurement and forms the basis of work regarding the presence of methanesulphonate in Antarctic ice [8]. Experience leads to the conclusion that it is necessary not only to filter the air to remove particulate matter prior to contact with melting ice but also to exclude potentially contaminating gases. REFERENCES 1 J. M. Palais and M. Legrand, J. Geophys. Res., 90 (1985) 1143. 2 M. R. Legrand and R. J. Delmas, Atmos. Environ., 18 (1984) 1867. 3 E. S. Saltzman, D. L. Savoie, R. G. Zika and 3. M. Prospero, J. Geophys. Res., 88 (1983) 10897. 4 G. P. Ayers, J. P. Ivey and H. S. Goodman, J. Atmos. Chem., 4 (1986) 173. 5 W. C. Keene, J. N. Galloway and J. D. Holden, J. Geophys. Res., 88 (1983) 5122. 6 R. J. Delmas, Nature, 299 (1983) 677. 7 M. Legrand, D. DeAngelis and R. J. Delmas, Anal. Chim. Acta, 156 (1984) 181. 8 J. P. Ivey, D. M. Davies, V. Morgan and G. P. Ayers, Tellus, 5 (1986) 88. 9 W. C. Keene, A. A. P. Pszenny, J. N. Galloway and M. E. Hawley, J. Geophys. Res., 9 (1986) 6647.
in The Netherlands
Short Communication
ION CHROMATOGRAPHIC IN ANTARCTIC ICE
DETERMINATION OF SELECTED IONS
J. P. IVEY*’ Australian Government (A ustmlia)
Analytical
Labomtories,
P.O. Box 84, Kingston,
Tasmania
7150
D. M. DAVIES CSIRO Division (Australia)
of Atmospheric
Research,
Private Bag 1, Mordialloc,
Victoria 3195
(Received 3rd January 1986)
Summary. Ion chromatography is used to measure the concentrations of chloride, nitrate, sulphate, ammonium and sodium ions at the pg 1-l level in Antarctic ice and to investigate the occurrence of methanesulphonate, fluoride, formate, acetate and nitrite. Of the latter group of ions, only methanesulphonate was found in measurable concentrations.
The composition of Antarctic ice, in particular its soluble impurities, has been used to give insight into such questions as paleoatmospheric composition [ 11, and the natural cycling of nitrogen and sulphur compounds [l, 21. Recently, the presence of biogenic methanesulphonic ,and other organic acids in the maritime environment [ 3-51 has raised questions as to their presence or otherwise in Antarctic ice. If they are present, they can be expected to contribute to ice acidity and in the case of methanesulphonic acid to the sulphur budget [6]. Methanesulphonate measurements may also aid in the definition of the relative sulphate contributions from marine biogenic productivity and volcanic activity [l] . The sample preparation protocol is of fundamental importance in the eventual reliability of the analytical results. The problems associated with gaseous species in the preparation of ice for analysis have been noted in earlier work [ 71. In this communication, the manner in which the preparation problem was overcome is discussed as well as the optimization of the instrumentation for the determination of each ion, or group of ions. Experimental Sample preparation. Ice cores were collected as part of the Australian National Research Expedition (ANARE) glaciology research program at Law Dome (66 30’S, 111 OO’E). The core was cut into 15 equal (depth) sections *Present address: Department of Analytical Chemistry, University of New South Wales, P.O. Box 1, Kensington, N.S.W. 2033, Australia. 0003-2670/87/$03.50
o 1987 Elsevier Science Publishers B.V.
for each annual accumulation layer as shown by the oxygen isotope record. Each section was then trimmed to leave ca. 200 g of the inner part of the core. The method of washing and melting the 200-g sections was developed in three stages, only results obtained from the final two methods were utilized elsewhere [8] . All manipulations were done in a clean air laminar flow cabinet. All plastic ware was washed copiously with distilled/deionized water and then filled with distilled/deionized water, sealed, and allowed to stand for at least 24 h prior to use. The three stages were as follows. (1) Initially about 25% of the 200-g block was removed by washing in distilled/deionized water (resistivity > 18 M ohm cme2) after which a further 25% was removed by placing the washed ice in a polyethylene funnel and allowing the ice to melt as described by Legrand et al. [7]. Chloroform (multiply extracted with distilled/deionized water) was added to a polyethylene receiver vessel. The purpose of this was to preserve any organic acids which have been shown to be labile in the absence of a biocide [5]. Frozen distilled/deionized water taken through the procedure, however, showed disconcerting levels of acetate, formate and traces of nitrite by ion chromatography. The clean air cabinet was obviously allowing acidic (and presumably basic) gases through its filter, with the potential further to contaminate the melting sample with nitric acid and sulphurous/sulphuric acids and ammonia. (2) Air from a compressed air source was cleaned by passage through an in-line gas absorption tube containing Ascarite (8-20 mesh). The scrubbed air was then piped to an inverted funnel over the polyethylene funnel used to contain the melting ice (Fig. 1). A gas absorption tube packed with oxalic acid was included in the same line for removing ammonia. A 0.45-pm filter was last in line. The effectiveness of this can be seen in Fig. 2, for which two different samples were allowed to melt in the laminar flow cabinet (A) and the scrubbed-air (B) melt systems. (3) When the above system was used, any formate and acetate present were below the detection limit of the present method and thus no chloroform was required. The ice was washed with distilled/deionized water until 50% remained which was then simply melted in a sealed polyethylene bag (Whirlpak). The Whirlpaks were checked for possible contaminants by running blank determinations with distilled/deionized water. Equipment. The ion-chromatographic system for the determination of anions consisted of a Waters Associates M45 pump, a Rheodyne fixedvolume injector (model 7000), a Dionex 030985 fast-run anion separator column, a Dionex 035350 hollow-fibre or Dionex 030829 suppressor column and a Waters model 430 conductivity detector. The same instrumentation was used for the determination of sodium and ammonium except that a Wescan 269004 cation separator column and a Dionex cation suppressor column were used. As indicated by Legrand et al. [7] a large (5.5-ml) sample loop was found to be preferable to a concentrator column for the introduction of samples with low concentrations of analyte species. Samples
283
Fig. 1. Air-purification system used in the melting of ice cores: (1) compressed air source; (2) acid gas trap of ascarite (8-20 mesh); (3) alkali gas trap of oxalic acid; (4) self-indicating soda lime; (5) particle filter; (6) laminar flow clean cabinet; (7) polyethylene receiving bottle.
C
NLDRIDE
0
Fig. 2. Chromatograms of sample ice. (A) Ice melted in a clean air cabinet without contaminating gas removal showing the separation of acetate/fluoride, formate, methanesulphonate and chloride at respective concentrations of about 105, 5.8 and 50 ng ml-‘. (B) Ice melted after contaminating gas removal; methanesulphonate and chloride at respective concentrations of 16.7 and 310 ng ml-‘. (C) Sample ice with nitrate and sulphate concentrations of 17 and 31 ng ml-‘, respectively. Conditions: injection volume of 5.5 ml, flow rate 2.0 ml min-’ , chart speed 0.25 cm min-’ , eluents as in Table 1.
234
were loaded into the loop with a cleaned lo-ml polyethylene syringe. Eluent concentrations and conditions are shown in Table 1. Comparative-measurements of sodium and determinations of magnesium were done on a Varian atomic absorption spectrometer model 1475 with an air/acetylene flame. Results and discussion Legrand et al. [7] used an eluent for nitrate and sulphate which did not allow chloride to be quantified. They chose to use an alternative column specifically for chloride determinations. Here, the same column was used for all determinations but with a more dilute eluent for chloride. In the method of Legrand et al. [ 71, chloride elutes at a retention time of less than 2 min after passage of the water dip. Acetate, formate, methanesulphonate and fluoride all elute prior to chloride. Fluoride has been reported previously as being present in Antarctic ice but accurate quantitation was not possible because of interference from the water dip [l] . To examine the possible existence of these poorly retained anions, a very dilute eluent (0.0007 M sodium hydrogen carbonate) was used. Figure 2B is a typical sample chromatogram and Fig. 2A shows the degree of separation achieved. Fluoride was resolved from acetate though not adequately; but because no peak was observed at the acetate/fluoride retention time in any sample, this was not a problem. Chloride was also quantified with this system by using very high attenuation. Methanesulphonate ion was found in all samples examined. The concentration was between 9 and 0.2 pg 1-l as sulphur (S); the latter is the detection limit at twice the baseline noise. Standards with concentrations of 1.0, 2.5, 5.0, 7.5 and 10.0 fig 1-l (S) typically gave a linear peak height vs. TABLE 1 Summary of conditions for quantifying selected ions in Antarctic icea Ion
Eluent
Detection limit (ng)
Range of standards ( ng ml-’ )
Sodium Ammonium Fluoride/acetate Formate Methane sulphonate Chloride Nitrite Nitrate Sulphate
0.004 M HNO,
1 2 5120 20 1 2 4 2 3
O-500 O-50 -
0.0007
M NaHCO,
0.002 M Na,CO,
O-100 O-30 o-So0 o-5 O-25 O-l 50
‘Injection volume 5.5 ml, flow rate 2.0 ml min-’ ; columns etc. as discussed in text; sensitivity 5 ~1s f.s.d. for cations, 1 US and 5 $!l for chloride, with both 0.0007 M hydrogen carbonate and 0.002 M carbonate eluents.
285
mass calibration with a correlation coefficient of 0.9998, a standard error of 0.14 and an intercept of 0.02 pg 1-l (S). Nitrate and sulphate were both quantified by using 0.002 M sodium carbonate as eluent (Fig. 2C). Nitrite elutes between chloride and nitrate but was not observed in samples melted in scrubbed air or samples melted in polyethylene bags. Detection limits are given in Table 1. Sodium and ammonia were quantified by ion chromatography. Potassium was not measured in the present study. Comparison of the sodium results from ion chromatography and those obtained by a.a.s. on 30 samples is shown in Fig. 3. In general, the agreement between the two methods is reasonable (r = 0.988) and certainly comparable to a similar comparison between ion chromatographic and neutron activation measurements of sodium content [ 71. Accurate sodium and magnesium results are particularly important as they are used to define the relative contributions made to the sulphate budget from sea salt and gas to particle conversion [9]. The a.a.s. measurements of sodium and magnesium were done without the addition of an ionization suppressant to either standards or samples in order to avoid contamination of the samples. Ammonium concentrations were in the range O-l pg 1-l ; most samples were at, or about, the detection limit of 0.5 pg 1-l for the equipment used. The relative absence of ammonia from our samples indicates that the methodology for sample preparation is sound. Presence of ammonia has been found to be indicative of sample contamination [7] . Conclusion Whilst it is always preferable to be able to quantify all the species of interest from a single chromatogram, this is not practicable in the determination of anions in Antarctic ice. Valuable information may be lost if the concen-
Fig. 3. Comparison of sodium concentration obtained by a.a.s. (x axis) and ion chromatography (y axis) of 30 ice-core samples (ng ml-‘). Con-. coeff. 0.988, slope 1.034 and x intercept 8.8.
286
trations of early eluting species are not measured. The methodology presented above allows such measurement and forms the basis of work regarding the presence of methanesulphonate in Antarctic ice [8]. Experience leads to the conclusion that it is necessary not only to filter the air to remove particulate matter prior to contact with melting ice but also to exclude potentially contaminating gases. REFERENCES 1 J. M. Palais and M. Legrand, J. Geophys. Res., 90 (1985) 1143. 2 M. R. Legrand and R. J. Delmas, Atmos. Environ., 18 (1984) 1867. 3 E. S. Saltzman, D. L. Savoie, R. G. Zika and 3. M. Prospero, J. Geophys. Res., 88 (1983) 10897. 4 G. P. Ayers, J. P. Ivey and H. S. Goodman, J. Atmos. Chem., 4 (1986) 173. 5 W. C. Keene, J. N. Galloway and J. D. Holden, J. Geophys. Res., 88 (1983) 5122. 6 R. J. Delmas, Nature, 299 (1983) 677. 7 M. Legrand, D. DeAngelis and R. J. Delmas, Anal. Chim. Acta, 156 (1984) 181. 8 J. P. Ivey, D. M. Davies, V. Morgan and G. P. Ayers, Tellus, 5 (1986) 88. 9 W. C. Keene, A. A. P. Pszenny, J. N. Galloway and M. E. Hawley, J. Geophys. Res., 9 (1986) 6647.