Introduction

Carbon and nitrogen isotope ratios have recently been used to study various kinds of research fields, such as biogeochemistry and ecology (Fry 2006; Michener and Lajtha 2007; Wada 2009; Ohkouchi et al. 2010). In limnology, stable isotope technique has been applied to both nutrient dynamics and food-web research in rivers (e.g., Yamada et al. 1996; Finlay 2004; Doi et al. 2007; Kohzu et al. 2009) and lakes (e.g., Cabana and Rasmussen 1996; Yamada et al. 1998; Vander Zanden et al. 1999; Post et al. 2000; Ogawa et al. 2001; Post 2002; Maki et al. 2010). The natural abundances of carbon-13 (13C) and nitrogen-15 (15N) are expressed as per mil (‰) deviations from international standards: δ13C (‰) or δ15N (‰) = (R sample/R standard − 1) × 103, where R is 13C/12C for δ13C or 15N/14N for δ15N, respectively. Vienna Peedee belemnite (VPDB) and atmospheric nitrogen (AIR) were used as international standards for carbon and nitrogen, respectively (Coplen 1994; Böhlke and Coplen 1995). A short history of selecting standards is described in Coplen et al. (2002). Recently, Coplen et al. (2006) recommended that δ13C values of both organic and inorganic materials measured by an isotope ratio mass spectrometer (IRMS) should be corrected and expressed relative to VPDB on a scale normalized by assigning consensus values of −46.6‰ to L-SVEC lithium carbonate and +1.95‰ to NBS19 calcium carbonate. For nitrogen, two-point calibration with IAEA-N-1 ammonium sulfate (+0.43‰) and IAEA-N-2 ammonium sulfate (+20.41‰) or IAEA-N-1 ammonium sulfate and USGS32 potassium nitrate (+180‰) should be applied (Böhlke and Coplen 1995).

The aim of this paper is to prepare certified reference materials of organic compounds in the Japanese community of stable isotope science based on the approved correction method of two-point anchoring. The multiple certified reference materials are useful in comparing carbon and nitrogen isotope ratios measured by different laboratories. The Center for Ecological Research (CER), Kyoto University, is approved as a Joint Usage/Research Center by the Ministry of Education, Culture, Sports, Science and Technology (MEXT), Japan, and the reference materials (CERKU-01 to CERKU-10) are available on request. BG-A, BG-P, and BG-T are also available from the Institute of Biogeosciences (BioGeos), Japan Agency for Marine-Earth Science and Technology (JAMSTEC), on request.

Materials and methods

In the CER, Kyoto University, special-grade reagents (Wako Co. Ltd.) of dl-Alanine (C3H7NO2), l-Alanine (C3H7NO2), glycine (C2H5NO2), l-Leucine (C6H13NO2), l-Threonine (C4H9NO3), starch-C3 ((C6H10O5) n ), starch-C4 ((C6H10O5) n ), dl-Alanine (C3H7NO2), and l-Tyrosine (C9H11NO3); and reagent for peptide (Kokusan Chemical Co. Ltd.) of l-Alanine (C3H7NO2) were purchased and named CERKU-01 to CERKU-10, respectively. CERKU-8 and CERKU-9 have previously been used as internal standards in the CER. They were purchased in 500-g bottles and manually well mixed before isotope measurement. CERKU-02 was prepared by dissolving 500 g of l-Alanine (special grade reagents, Wako) and 40 mg 15N-98%-l-Alanine (NLM-454, Cambridge Isotope Labs, USA) in 5 l deionized water. The solution was lyophilized and well mixed with a large bowl and pestle to obtain fine powder. In the BioGeos, JAMSTEC, reference materials named BG-A, BG-P, and BG-T also represent analytical grade reagents (Aldrich Chemical Co., Inc) of l-Alanine (C3H7NO2), l-Proline (C5H9NO2), and l-Tyrosine (C9H11NO3), respectively. Materials of BG-A and BG-P were carefully pulverized before use.

In this study, to calibrate δ13C values, international isotope standards NBS19 and L-SVEC were used, whereas IAEA-N-1, IAEA-N-2, and USGS32 were used for calibrating δ15N values. In addition, IAEA-CH-3 cellulose ((C6H10O5) n ), IAEA-CH-6 sucrose (C12H22O11), IAEA-CH-7 polyethylene ((CH2) n ), and USGS40, USGS41 l-Glutamic acid (C5H9NO4) (Qi et al. 2003) were prepared. Stable isotope measurements were preformed in three laboratories: CER, Kyoto University; BioGeos, JAMSTEC; and Tokyo Institute of Technology. In the CER, samples were analyzed using an isotope ratio mass spectrometer (Finnigan MAT Delta S, Germany) coupled with an elemental analyzer (Fisons EA1108, Italy) via Conflo II as an interface. The combustion temperature was set to 1,020°C, and the reduction furnace was kept at 650°C (Qi et al. 2003). One set of measurements was performed after introducing new combustion and reduction tubes until the exhaustion. Reference materials were folded into tin capsules containing carbon and nitrogen equivalent to 1 mg alanine (404 μgC for δ13C analysis and 157 μgN for δ15N analysis) within 10% variation of each amount. They were supplied to sequential dual isotope analysis with dilution mode for δ13C but undiluted mode for δ15N. Samples for δ13C and δ15N analysis were prepared separately, except for alanine.

Samples were analyzed sequentially in the order CERKU-01, CERKU-02, CERKU-03 for C, CERKU-03 for N, CERKU-04 for C, CERKU-04 for N, CERKU-05 for C, CERKU-05 for N, CERKU-06, CERKU-07, IAEA-N-1, USGS40 for C, USGS40 for N, USGS41 for C, USGS41 for N, IAEA-CH-3, IAEA-CH-6, IAEA-N-1, IAEA-N-2, L-SVEC, NBS19, USGS32, CERKU-08, CERKU-09 for C, CERKU-09 for N, CERKU-10, BG-A, BG-P for C, BG-P for N, BG-T for C and BG-T for N. This order was then repeated backward. Additional replicated analyses of the sequence were performed forward and backward in turn. In some sequences, several kinds of reference materials were not analyzed. A blank consisting of an empty 8 × 5-mm tin capsule was analyzed before the sequence; however, the amount of blank generally was so small compared with the sample peak that blank correction was not applied in the CER. All data were averaged, and two-point linear calibration of NBS19 (δ13C = 1.95‰) and L-SVEC (δ13C = −46.6‰) (#1–#5) was applied for δ13C (Coplen et al. 2006); and two-point linear calibration of IAEA-N-1 (δ15N = 0.43‰) and IAEA-N-2 (δ15N = 20.41‰) (#1–#3) or that of IAEA-N-1 (δ15N = 0.43‰) and USGS32 (δ15N = 180‰) (#4, 5) was applied for δ15N (Böhlke and Coplen 1995). The slope of linear regression of measured and trued-values of two certified reference standards were calculated in the sequences #1–#5.

At BioGeos JAMSTEC, we used a Delta-plus XP isotope ratio mass spectrometer (Thermo Finnigan, Germany) coupled to an elemental analyzer (Flash EA 1112, Thermo Finnigan, Germany) via ConFlo III interface with modifications to improve sensitivity (Ogawa et al. 2010). Temperatures of combustion and reduction furnaces were set to 1,050°C and 750°C, respectively. Samples were carefully prepared to contain 12 ± 1 μgC for δ13C or 3.1 ± 0.3 μgN for δ15N, which is equivalent to IRMS intensity of 15 ± 2 V (m/z 44) or 3.0 ± 0.5 V (m/z 28), respectively. Blank collection was performed to each raw data to offset isotope effect from background contamination of carbon and nitrogen (Ogawa et al. 2010) before further calculation and calibration. All data were averaged, and two-point linear calibration of NBS19 (δ13C = 1.95‰) and L-SVEC (δ13C = −46.6‰) was applied for δ13C (Coplen et al. 2006) and two-point linear calibration of IAEA-N-1 (δ15N = 0.43‰) and IAEA-N-2 (δ15N = 20.41‰) for δ15N.

In the Tokyo Institute of Technology (TokyoTech), samples were analyzed using a conventional off-line method consisting of sealed tube combustion followed by cryogenic separation of the resulting carbon dioxide (CO2) and nitrogen gas (N2) using a high-vacuum line system and IRMS measurement. Briefly, ~2.5 mg of sample was combusted in an evacuated and sealed quartz tube with 1.8 g copper(II) oxide (CuO) as the oxidant and 0.8 g copper (Cu) as the reductant at 850°C for 2 h. The sample tube was attached to the line system with a glass tube cracker and then cooled to −196°C for 5 min. The sample tube was then broken, and the released N2 was transferred to a 6-mm i.d. Pyrex® glass tube containing three pellets of molecular sieves 5A (1/16 Pellets, Wako) attached to the line system at −196°C. The Pyrex® glass tube was then flame sealed. Subsequently, the sample tube was heated to −78°C for 5 min. The released CO2 was transferred to a 6-mm i.d. Pyrex® glass tube attached to the line system at −196°C. The Pyrex® glass tube was then flame sealed. Measurements of δ13C and δ15N for the resulting CO2 and N2 gases in the Pyrex® glass tubes were made using an isotope ratio mass spectrometer (Finnigan MAT 252, Germany, and Delta-plus XP, USA) with a dual-inlet system. In the TokyoTech, isotopic standardization for δ13C against VPDB was accomplished by a two-point linear calibration of IAEA-CH-6 (δ13C = −10.45‰) and IAEA-CH-7 (δ13C = −32.15‰) instead of NBS19 and L-SVEC. For δ15N, isotopic standardization was accomplished by comparison with reference N2 gas (99.9999%; Taiyo Nippon Sanso Corp.), which was calibrated against AIR using the international nitrogen isotope standard N-SVEC (δ15N = −2.78‰).

Results and discussion

Table 1 shows the δ13C values of reference materials based on the two-point linear calibration. Two-point anchoring has repeatedly been recommended by many authors (e.g., Coplen et al. 2006; Paul et al. 2007). In this study, δ13C values of 13 analytical-grade reagents were carefully determined by the two-point linear calibration of NBS19 and L-SVEC. The calculated δ13C values of other international standards, including IAEA-CH-6, IAEA-CH-3, and USGS40, were well fitted to the literature values within 0.04‰, although USGS41 was slightly (0.19‰) lower than the accepted value. This is consistent with Skrzypek et al. (2010), who have shown that a higher mean error was observed outside the range bracketed by the δ values of the anchor standards (L-SVEC and NBS19).

Table 1 δ13CVPDB values corrected by two-point calibration indicated by asterisk
Full size table

Table 2 shows the δ15N values of reference materials based on the two-point linear calibration. δ15N values of 13 analytical-grade reagents were determined by the two-point linear calibration of IAEA-N-1 and IAEA-N-2 (#1–#3), or IAEA-N-1 and USGS32 (#4). Similar to the δ13C values, the calculated δ15N values of additional international standards [USGS41, USGS40 and IAEA-N-2 (#4)] were well fitted to the literature values within 0.03‰.

Table 2 δ15NAir values corrected by two-point calibration indicated by asterisk
Full size table

In the studies of limnology and ecology, stable isotope ratios are often used for investigating nutrient dynamics and food-web structure. In such cases, stable isotopic comparisons among study subjects are appropriate, whether they are within-ecosystem or between-ecosystem, if their calculated values are based on the analysis with high analytical precision. However, the isotopic composition is expressed as a relative value to the international standard, so that accuracy is important for the quality of interlaboratory comparisons, such as meta-analysis based on many papers. Coplen et al. (2006) recommended that δ13C values measured by an IRMS should be anchored at two-point linear calibration of NBS19 (δ13C = 1.95‰) and L-SVEC (δ13C = −46.6‰). They concluded that δ13C value of the international standard NBS22 oil should be replaced by −30.03‰ (Coplen et al. 2006). In this paper, we determined δ13C and δ15N values of reference materials by two-point anchoring and confirmed with several international standards. The slope of correlation between measured and true values of two-point anchors in carbon and nitrogen isotope ratio analysis of #1–#5 was 1.0061 ± 0.0009 (1.0047–1.0071) and 1.0006 ± 0.0014 (0.9988–1.0022) [mean ± standard deviation (SD) with the ranges in parenthesis], respectively. However, the range of slope values varied within 1% of 1:1 line in the delta S system in the CER, concluding that single-point anchoring standardization causes little error in the natural abundance level in this system.

In summary, we determined carbon and nitrogen isotope ratios of 13 reference materials based on the two-point anchoring method. The reference materials cover δ13CVPDB range of −34.92 to −9.45‰ and δ15NAir range of −5.22 to 22.71‰, thus can be useful for measuring isotope ratios of naturally occurring substances.