Department of Geology, Western Michigan University, Kalamazoo, Michigan 49008. A simple and inexpensive method for reacting carbonate samples with ...
Anal. Chem. 1997, 69, 4256-4258
Technical Notes
A Simple, Inexpensive Carbonate-Phosphoric Acid Reaction Method for the Analysis of Carbon and Oxygen Isotopes of Carbonates R. V. Krishnamurthy,* E. A. Atekwana, and Hillol Guha
Department of Geology, Western Michigan University, Kalamazoo, Michigan 49008
A simple and inexpensive method for reacting carbonate samples with phosphoric acid for the measurement of stable carbon and oxygen isotope ratios in carbonates is described. The technique is applied to samples taken in a sealed septum tube inside which a small tube is attached using ordinary household glue. The acid is loaded into this small tube. The whole assembly is evacuated and the vessel is tilted so that the acid makes contact with the carbonate. Isotope determinations were made using calcite at 25 °C and calcite and dolomite at 50 °C. With dolomite, the reaction was complete in 12 h. Measurements of yield and isotope ratios confirm that this method is as good as the traditional method, which uses potentially breakable Pyrex reaction vessels and costly grease or greaseless stopcocks. The δ13C and δ18O signatures of carbonates have been used extensively in paleoclimate research and in studies of the processes involved in their formation in a wide variety of environments. Isotope measurements are usually carried out using carbon dioxide gas produced in the reaction between carbonates and 100 or 95% phosphoric acids.1 In the traditional method, the carbonate sample is loaded in glass bottles with a side arm to hold the acid. The whole assembly is equilibrated at 25 °C, and the reaction vessel is “tilted” so as to bring the sample and the acid into contact with each other. This method requires cleaning of the vessels after each run and also use of grease, although it is possible to replace grease joints with O-ring fitted valves. In other variants of this method, the samples are reacted at 50 °C rather than at 25 °C to accelerate the reaction.2-5 Some laboratories are equipped with automated carbonate reaction systems in which carbon dioxide produced in the acid reaction is automatically introduced into the mass spectrometer. These systems are very expensive and obviously require elaborate maintenance. We have developed a simple and inexpensive method to carry out the acid reaction. This makes use of 16 × 100 mm glass septum tubes (Vacutainer Serum Tubes, Becton Dickson & Co., Franklin Lakes, NJ 07417) and a Cajon fitted needle assembly. (1) McCrea, J. M. J. Chem. Phys. 1950, 18, 849-857. (2) Rosenbaum, J.; Sheppard, S. M. F. Geochim. Cosmochim. Acta 1965, 50, 1147-1150. (3) Shackleton, N. J.; Opdyke, N. D. Quat. Res. 1973, 3, 39-45. (4) Al-Aasam, I. S.; Taylor, B. E.; South, B. Chem. Geol. 1990, 80, 119-125. (5) Swart, P. K.; Burns, S. J.; Leder, J. J. Chem. Geol. 1991, 86, 89-96.
4256 Analytical Chemistry, Vol. 69, No. 20, October 15, 1997
Figure 1. Schematic showing the reaction vessel. The septum tube is ∼10 ml in capacity, and the inside boat holding the acid is made of a 9 mm diameter Pyrex tube. Also shown is the needle assembly used to connect the septum tube with the vacuum extraction line.
The apparatus is similar to the one first suggested for the equilibration of water with carbon dioxide gas6 and that used by us for the study of dissolved inorganic carbon in water samples.7 EXPERIMENTAL SECTION Figure 1 shows the extraction apparatus, where the septum tube is 10 mL in capacity. After removal of the septum, a known amount of the sample (ranging from 10 to 40 mg) was transferred to the tube, and a Pyrex boat (9 mm diameter, ∼5 cm long) was glued to the side of the septum tube on the inside. For this, a drop of an ordinary household glue (Elmer’s Wonderbond Plus Super Glue, Borden Inc. HPPG, Columbus, OH 43215) was used. Approximately 1-1.5 mL of 100% phosphoric acid was then dropped into this Pyrex boat and the septum put back in place to close the sample tube. The whole assembly was evacuated using a needle assembly consisting of a 26 gauge needle (∼12 mm long) affixed to an inner Luer ground joint and attached to the vacuum line by way of a cajon union. After the required vacuum conditions (6) Socki, R. A.; Karlsson, H. R.; Gibson, E. K., Jr. Anal. Chem. 1992, 64, 829-831. (7) Atekwana, A. E.; Krishnamurthy, R. V. J. Hydrol., submitted for publication. S0003-2700(97)00204-7 CCC: $14.00
© 1997 American Chemical Society
Table 1. Extraction Efficiencies and Carbon and Oxygen Isotope Ratios of Various Carbonate Samples Prepared by Treatment with Phosphoric Acid in Sealed Septum Tubesa reacn time (h)
reacn temp (°C)
glue (n)3)
4-12
25
LAB-2 (n ) 5)
4
25
96.0 ( 1.5
LAB-3 (n ) 5)
4
25
LAB-3 + soil (1:1 mix) (n ) 5) Na2CO3‚H2O (n ) 8)
4
sample
estd CaCO3 (%)
δ13CPDB (‰)
δ18OPDB (‰)
BD -4.59 ( 0.06
-1.70 ( 0.07
94.0 ( 2
1.79 ( 0.05
0.30 ( 0.06
25
51.8 ( 3.3
1.82 ( 0.06
0.15 ( 0.02
4
25
96.2 ( 1.1
-7.32 ( 0.06
C-BJ085 (n ) 5) C-BJ015 (n ) 5) LAB-2 (n ) 3)
4 4 6
25 25 50
20.4 ( 1.1 2.3 ( 0.1 95.0 ( 2
-6.20 ( 0.1 -8.81 ( 0.5 -4.65 ( 0.02
-9.18 ( 0.2 -8.79 ( 0.8 -1.70 ( 0.02
dolomite-1 (n ) 4)
12
50
82.0 ( 5
0.90 ( 0.1
-9.50 ( 0.09
dolomite-2 (n ) 4)
12
50
77.0 ( 5
1.28 ( 0.1
-9.59 ( 0.09
-18.27 ( 0.3
comments the glue used to hold the acid produced no detectable CO2 internal carbonate standard values obtained by the normal method: δ13CPDB ) -4.65 ( 0.07, δ18OPDB ) -1.74 ( 0.1 internal carbonate standard values obtained by the normal method: δ13CPDB ) 1.87 ( 0.02; δ18OPDB ) 0.29 ( 0.1 indicates that the soil matrix has no effect in the extraction efficiency or δ measurements the δ18O has no significance in view of the water of hydration loess carbonate loess carbonate a fractionation factor of 1.010 25 at 25 °C was used8 thin sections indicated less than 90% dolomite; a fractionation factor of 1.010 at 50 °C was used2 thin sections indicated less than 90% dolomite; a fractionation factor of 1.010 at 50 °C was used2
a LAB-2 and LAB-3, our internal standards, represent Holocene limestone. These were calibrated using NBS-19. n denotes the number of analyses of each sample. BD ) below detection levels. Reproducibilities represent 1 standard deviation.
(10-3 Torr) were established, the whole assembly was immersed in a 25 °C bath for 30 min to allow the reaction vessel containing the sample and the acid to reach thermal equilibrium. Following this, the acid was allowed to slide outside the boat, flow down the inside wall, and cover the sample. This procedure was done inside the water bath itself to minimize effects arising from temperature differences between the bath and the laboratory air. The main precaution taken in this step was to ensure that the septum tube was tilted in a manner that allowed the acid to slide out of the boat and not flow back into the boat. The reacting mixture was kept in the constant-temperature bath for 6 h, although pure carbonates may react completely in much less time. Upon the completion of the reaction, the CO2 was extracted from the septum tube using the same needle assembly connected to the vacuum line with a Cajon union. Before extraction of the sample, it was necessary to pump away the air in the needle assembly. This was achieved by inserting the needle into the septum deep enough to cover its opening but not deep enough to puncture the septum completely. After evacuation of the needle assembly, the needle was pushed deep enough to puncture the septum so that reaction products, namely, CO2 and water, could be drawn out of the septum tube. The CO2 was then purified cryogenically and the yield measured manometrically in a calibrated cold finger. Isotope measurements are reported in the usual δ notation where
δ‰ )
Rsample - Rstandard - 1 (R ) 18O/16O, 13C/12C) Rstandard
The δ values are reported with respect to PDB using NBS-19 as the main calibration standard.
RESULTS The results of the present study are given in Table 1. Before this technique was implemented, the glue used for attaching the boat was reacted with phosphoric acid to ensure that it did not produce any carbon dioxide. For this, more than 10 drops (10 times that used to attach the boat to the inner wall of the septum tube) were treated with the acid for more than 12 h. As shown in Table 1, the reaction did not produce any gas that could be measured. We therefore consider that the glue did not contribute anything above the normal background. We do recommend, however, that before this method is adopted, verification be made that the glue to be used is indeed inert, as it was with our experience. From Table 1, it is clear that this technique provides δ measurements that are comparable to or even better than those obtained using the traditional glass reaction vessels with a side arm. The results from LAB-2 and LAB-3, our internal carbonate standards, confirm this. Sodium carbonate also gave very good results although δ18O of this sample had a larger standard deviation. This arises from the fact that the solid sample used by us was in the hydrated form (Na2CO3‚H2O), and so the δ18O value actually has little meaning. In order to simulate a natural carbonate soil, we mixed an equal proportion of a sandy loam soil (previously decarbonated with hydrochloric acid) with LAB-3. Since the δ values of LAB-3 are known, analysis of this mixture would provide a further test regarding the efficiency of this method. As shown in Table 1, analysis of this mixture gave an extraction efficiency of 97% and excellent reproducibility of the δ values. Table 1 also shows the analysis of two soil carbonates as well as the analysis of calcite and dolomites at 50 °C. The results from the soil carbonates are interesting in that while the percentage calcium carbonate estimated using this method gave very small errors, the δ values of one sample, namely, C-BJ015, gave considerably large errors (e.g., (0.5 for δ13C and (0.8 for δ18O). (8) Friedman, I.; O’Neil, J. R. U.S. Geol. Surv. Prof. Pap. 1977, No. 440-KK.
Analytical Chemistry, Vol. 69, No. 20, October 15, 1997
4257
It should be noted that this sample also had a very low carbonate content (2.3 ( 0.1%). With such a low carbonate content, it is possible that only part of the sample may represent an authigenic precipitate and may in fact be a mixture of in situ and windblown carbonates of heterogenous isotope ratios. We argue that the larger spread in the δ values is a direct result of the carbonate composition rather than the low carbonate content of the sample. The isotope measurements of calcite and dolomite at 50 °C show that carbonates can be reacted with this technique at this temperature. Previous work has shown that acceptable δ values of dolomites can be obtained by carrying out the reaction at this temperature for 4 h and that up to 24 h may be required for 100% completion of the reaction.4 In our case, the reaction after 12 h gave a yield of ∼82%. Since thin-section studies revealed these samples to be composed of 80% dolomite (Heidi Wines, personal communication), we believe that the reaction was complete in 12 h. In a previous study dealing with the dissolved inorganic carbon in water samples,7 we have seen that samples may be stored in these septum tubes for several weeks without compromising the vacuum inside or the isotopic ratios. These experiments suggest that the rubber septum used to close the tubes do not absorb or react with the carbon dioxide gas chemically. An additional advantage of this technique is that reaction of “tough” carbonates such as siderite and dolomite with phosphoric acid can be carried out at temperatures lower than 50 °C for an extended period of time. Finally, we wish to indicate that this technique is suitable for handling small amounts of samples. To date, sample amounts as small as 0.2-0.3 mg have been processed without difficulty. These were natural shell samples processed for stable oxygen and carbon isotope ratios (Norman Lovan, personal communication). We believe even smaller samples can be processed in the same system or by using smaller septum tubes. Although most
4258
Analytical Chemistry, Vol. 69, No. 20, October 15, 1997
laboratories use the conventional carbonate extraction techniques, it will be interesting to couple the modification described here to automated carbonate extraction systems. If that is successful, the advantages in terms of speed and cost savings are obvious. CONCLUSIONS We have developed a simple method to carry out the phosphoric acid-carbonate reaction for the determination of δ13C and δ18O values in carbonates. The method is a modification of the original technique proposed by McCrea1 but makes use of a septum tube. Since septum tubes are inexpensive, they can be discarded after each run and there is no need to recycle. Also, use of grease is totally eliminated. Analyses of laboratory carbonate standards and several field samples show that this method is comparable to or even better than the traditional method. Results from field samples showed that a sample with low carbonate content had a large standard deviation in the isotopic value, although the reproducibility of the carbonate content was excellent. We suggest that, at low carbonate content, the isotopic heterogeneity may be related to the carbonate composition of the sediment rather than the efficiency of this technique. ACKNOWLEDGMENT This work was supported by NSF Grant EAR9632034. We thank Dr. Bill Harrison for providing us with the dolomite samples and H. R. Karlsson for a constructive review of the manuscript. Received for review February 19, 1997. Accepted August 5, 1997.X AC9702047 X
Abstract published in Advance ACS Abstracts, September 15, 1997.
