BACKGROUND Shale, as an important component of sedimentary rocks, is also a valuable source rock and “reservoir” of shale gas. The study of nitrogen isotope distribution characteristics can provide geochemical indexes for judging the sedimentary environment and oil-source correlation of crude oil. It is helpful to understand the organic matter enrichment mechanism of black shale, shale oil, and gas exploration^14-17. Nitrogen isotope ratio analysis in sedimentary rocks is faced with problems such as low content, and the influence of pretreatment on analysis and the reference materials, which affect the accurate analysis of nitrogen isotope ratio in shale and restrict the development of this technology in related studies. First, the low content in rocks will affect the analysis results by elemental analyzer-isotope ratio mass spectrometry (EA-IRMS)^16,18,20. Second, the method of direct sample analysis is often used to analyze nitrogen isotope ratio of whole rock, to improve the relative content of nitrogen and remove carbonate to meet the requirement of organic carbon isotope analysis; samples treated with acid can also be used for analysis. There are differences in treatment methods in different laboratories, which can also increase the variation in content and isotope ratios, as well as the difficulty of data comparison^9-13,21,25. Finally, there are relatively few international standard materials based on rocks and minerals for nitrogen isotope and there are still some problems and limitations in these standards^7,21,24,30. Therefore, the accurate analysis of nitrogen isotope ratio in shale is affected, thus restricting the development of this technology in related studies.

OBJECTIVES To improve the precision and accuracy of nitrogen isotope analysis in shales.

METHODS (1) The analysis was carried out by EA-IRMS, with the use of MAT253 Plus gas stable isotope mass spectrometer, Flash 2000 HT element analyzer and ConFlo Ⅳ interface. The furnace tube was filled with the recommended scheme of the instrument, and the water removal trap was adjusted according to the measured object, that is, only magnesium perchlorate was used or carbon adsorbent was added on this basis (the main component was sodium hydroxide), which accounted for 50% each. Carbon and nitrogen isotope ratios were measured separately.　　(2) Carbon and nitrogen isotope standard materials included: USGS40, USGS41a, IAEA-600, Urea, GBW04701, GBW04702, GBW04703, GBW07424 and GBW07107 as calibration, monitoring and experimental study (Table 1). Typical shale samples of the upper Yangtze Longmaxi Formation were selected as the research object. Both samples are rich in organic carbon. Combined with hand samples and XRD analysis, their mineral composition is mainly as follows: Quartz, potassium feldspar, illite and pyrite in SAMPLE-1, SAMPLE-2 are similar in mineral composition in general, but there is obvious dolomite in SAMPLE-2.　　(3) The experiment consisted of three parts. First, the impact of increasing oxygen injection time and using double tin cup encapsulation on nitrogen isotope ratio analysis was studied. Urea was applied to determine δ¹⁵N under the same conditions but with different oxygen injection time (3s or 5s) and the differences were compared. Urea and GBW07424 were used to encapsulate the sample with double tin cup (two layers of tin cup for the same sample) and conventional single tin cup for comparison under the same conditions. The composition of total nitrogen and total carbon isotope ratios in GBW04701, GBW04702 and GBW04703 was determined by single-tin cup encapsulation. Second, the influence of acid rinse method on the analysis was conducted. Standards were weighed and gradually added with an excess of 6mol/L HCl to remove carbonate minerals. During the reaction, the centrifuge tube was placed in an ultrasonic cleaner for ultrasonic treatment for 3 to 5 times, 10min each time. Centrifuge was used, acid was discarded and deionized water was added. After ultrasonic treatment in ultrasonic cleaner, centrifuge treatment was continued. This process was repeated until it reached neutral. The samples were then compared and analyzed with unprocessed samples and recommended values. Finally, the shale samples SAMPLE-1 and SAMPLE-2 from Longmaxi Formation were analyzed. The isotopic ratios of total nitrogen and total carbon were determined respectively. The samples were treated with the above acid rinse method to analyze the ratio of organic carbon and nitrogen isotopes, and the mineral composition before and after treatment was analyzed by XRD. Nitrogen isotope difference and total carbon and organic carbon isotope ratio were analyzed before and after treatment.

RESULTS The results of single/double tin cup method showed that the accuracy of δ¹⁵N value determined by double tin cup method were obviously worse than that by single tin cup method at the same oxygen injection time (Table 2). GBW07424 analysis showed that nitrogen amounts had more obvious influence on the analysis. The results of total nitrogen analysis in GBW04701, GBW04702 and GBW04703 measured under the condition of oxygen injection time 5s were consistent with the recommended values (Table 3), and the standard deviation of the samples was improved (SD≈0.1‰). After acid rinse, the nitrogen values of GBW04701 and GBW04702 were consistent with the recommended values (Table 3), and the relative deviation was better (SD<0.1‰). The overall carbon isotope ratio analysis values were consistent with the recommended values (Table 4). The difference of δ¹³C_VPDB between total carbon and organic carbon determined by GBW07107 was small and consistent with its carbon composition. The δ¹⁵N values of shale samples before and after acid treatment were basically the same in SAMPLE-1, while the δ¹⁵N values of shale samples after acid treatment were obviously changed in SAMPLE-2, which was related to the composition loss in the wash acid process of SAMPLE-2.

DISCUSSION (1) The optimized measurement conditions of EA-IRMS. Attempts were made to improve combustion efficiency by using double tin cup wrapping and increasing oxygen injection time. The relative deviation of double tin cup method was greater than 0.2‰ and obviously higher than that of single tin cup method, indicating that this method is not suitable. The reason may be related to the difference in combustion efficiency caused by the gap between double tin cups and the introduction of air when the sample is wrapped¹⁹. Appropriate increase of oxygen injection time in the analysis of GBW04701, GBW04702 and GBW04703 obtained satisfactory results, indicating that this method can effectively improve the accuracy of the technique, while this method did not significantly improve the standard deviation of δ¹⁵N value of Urea, which may be related to the matrix of the sample²⁰. It may be the abundant carbon in shale. Therefore, it is necessary for nitrogen isotope ratio analysis in shale to be determined by carbon and nitrogen separately and adding carbon adsorbent.　　(2) The acid rinse method. The acid rinse method adopted has no significant effect on GBW04701 and GBW04702 on δ¹⁵N analysis and can effectively remove carbonate components. Ultrasonic was used to promote the reaction between acid/water and sample in the process of acid reaction and acid washing in the experiment, and the centrifugation time was appropriately increased in the centrifugation stage, which was conducive to the preservation of fine particle components. Compared with static reaction, the addition of ultrasonic reaction could improve the washing effect. The less washing times were thus contributing to the accuracy of δ¹⁵N analysis⁴⁷.　　(3) The shale sample analysis. Although the acid treatment method adopted in the experiment has no obvious effect on the δ¹⁵N analysis of standards, significant composition loss and δ¹⁵N change were still observed in the analysis of shale samples. Moreover, the nitrogen content of shale samples of Longmaxi Formation was relatively high. Therefore, in order to accurately analyze the nitrogen content and δ¹⁵N analysis, direct analysis of samples is recommended for the whole-rock nitrogen isotope analysis of such samples.　　(4) The standard and reference materials. Carbon and nitrogen isotope reference materials with different characters and compositions were selected as calibration, monitoring and experimental study. USGS40, USGS41a and IAEA-600 standard samples were used for calibration and quality monitoring, and no obvious influence of matrix difference was found. Although they are Marine sediments, GBW04701, GBW04702, GBW04703 provide isotopic compositions of total carbon, organic carbon, total nitrogen and organic nitrogen. They are suitable for the analysis of carbon and nitrogen isotope ratios and the monitoring of acid treatment processes, especially the analysis of EA-IRMS technology⁵⁰. In shale analysis, GBW04701, GBW04702 and GBW04703 were selected as monitoring to meet the requirements of nitrogen and total carbon and organic carbon isotope ratio analysis.