INTRODUCTION
Accelerator mass spectrometry (AMS) is a detection technology with the advantages of less sample usage, short measurement time, and high sensitivity (Bennett et al. Reference Bennett, Beukens and Clover1977; Nelson et al. Reference Nelson, Korteling and Stott1977), making it relevant to various general cases. As the method with the highest sensitivity to measure 14C, it is widely used in archaeology, geology, oceanography, biomedicine, and environment science (Nielsen Reference Nielsen1952; Lubritto et al. Reference Lubritto, Rogalla and Rubino2004; Marzaioli et al. Reference Marzaioli, Lubritto and Battipaglia2005; Salehpour et al. Reference Salehpour, Hakansson and Possnert2015; Shen et al. Reference Shen, Pang, Jiang, He, Dong and Dou2015, Reference Shen, Sasa, Meng, Matsumura and Matsunak2019, Reference Shen, Zhang and Tang2022a, Reference Shen, Shi and Tang2022b; Cheng et al. Reference Cheng, Burr and Zhou2020). The high-precision and low-background 14C measurements rely on the appropriate graphite preparation conditions and reliable graphite preparation system. Therefore, establishing a high-quality 14C graphite preparation process and improving 14C graphite experimental conditions are significant work for every radiocarbon laboratory worldwide (Kitagawa et al. Reference Kitagawa, Masuzawa and Makamura1993; Meng et al. Reference Meng, Zhang and Chen2002; Santos et al. Reference Santos, Southon and Griffin2007a; Khosh et al. Reference Khosh, Xu and Trumbore2010; Hua et al. Reference Hua, Jacobsen and Zoppi2016; Barile et al. Reference Barile, Barone and Fedi2019; Shen et al. Reference Shen, Tang and Wang2022c).
As a vital method to synthesize 14C graphite, the Zn-Fe reduction method was first proposed by Jull et al. (Reference Jull, Donahue and Hatheway1986) and Slota et al. (Reference Slota, Jull and Linick1987). Xu et al. (Reference Xu, Trumbore and Zheng2007) and Walker et al. (Reference Walker and Xu2019) introduced a method for the synthesis of carbon samples by the Zn-Fe flame sealing reduction method, which not only overcomes the problem of atmospheric leakage in the reduction unit caused by the traditional online Zn-Fe reduction method (Bronic et al. Reference Bronic, Horvatincic and Sironic2010) but also avoids the problems of water vapor and CH4 production during the reaction, effectively reducing the background and improving the accuracy of the results. In addition, further studies by Macario et al. (Reference Macario, Alves and Oliveira2016) and Dee et al. (Reference Dee and Ramsey2000) showed that the ratios of the reducing agent and catalyst to sample, and their treatment method were significantly related to the quality and performance of the graphite target.
In this study, 14C graphite samples (≈1 mg C) were synthesized by the Zn-Fe flame sealing reduction method, and the experimental conditions were optimized based on a series of experiments, such as the selection of a reduction furnace, reaction temperature, time, reagent dose, and pretreatment of Fe. By analyzing the recovery rate, 12C– beam current, measurement sensitivity, and modern carbon contamination, the optimized experimental conditions for synthesizing trace 14C graphite samples by the Zn-Fe flame sealing reduction method were obtained. The background value of the experiment was evaluated using processed commercial graphite, and international standard samples (OXII, CSC, IAEA-C8) were used to evaluate the stability and reliability of the method, aiming to obtain higher performance 14C graphite for GXNU-AMS measurement.
MATERIAL AND METHODS
Reagents and Materials
Zn powder (Sigma-Aldrich #324930, <150 µm, 99.995%) was used for the reduction of CO2 to CO, Fe powder (Sigma-Aldrich #209309, 325 mesh, 97%) was used as a catalyst for reducing CO to C, and CuO powder (Sinopharm Chemical Reagent Co., Ltd, analytical grade, 99.7%) was used for combustion. The reference material OXII (SRM 4990C) was from the National Institute of Standards and Technology (NIST). C1 and C8 were from the International Atomic Energy Agency (IAEA). Chinese sugar carbon (CSC) was obtained from Xi’an AMS Center. The carbon powder samples were commercial graphite (CAS#1333-86-4) obtained from Alfa Aesar Co., Ltd.
A muffle furnace (SX2-2.5-12N, Shanghai Yiheng Scientific Instrument Co., Ltd., China) and a customized graphite reduction furnace (GXU-1, Chery Glass Products Co., Ltd., China) were used in the reduction and oxidation reactions. A high-temperature oven (GW-150B, Bangsi Instrument Technology (Shanghai) Co., Ltd.) was used in material baking. Two analytical balances (Mettler Toledo XP2 and ME155D, United States) with reading precisions of 1 μg and 10 μg, respectively, were used for sample weighing.
14C Sample Preparation Vacuum Line
The layout of the 14C sample vacuum line is shown in Figure 1. The main components include a vacuum pump, CO2 trap, quartz tube, water trap, valve, and vacuum gauge (Shen et al. Reference Shen, Tang and Wang2022c). According to its functions, the device is divided into a vacuum maintenance unit, a CO2 purification unit, and a CO2 reduction unit. The entire device adopts quartz glass as the main structural material, which has good vacuum performance and allows the entire experimental process to be observed.
Figure 1 Layout of the 14C sample vacuum line.
Before sample preparation, the quartz glass tubes were placed in a high-temperature furnace at 500°C for 5 hr to remove carbon contamination. The samples and CuO were mixed in the oxidation tube under vacuum conditions and then subjected to flame seal treatment. After combustion in a muffle furnace, the carbon in the sample was oxidized into CO2 and then introduced to the vacuum system. The CO2 gas first passed through alcohol liquid nitrogen cold trap 5 at –90°C to thoroughly remove the water vapor and then entered liquid nitrogen cold trap 6 at –196°C, where it was frozen. Next, any noncondensable gases, such as N2 and O2 were pumped away. The purified CO2 was heated and transferred to a gasreservoir 7 (with a fixed volume of approximately 23 mL), quantified by measuring the CO2 pressure (1 mg carbon corresponds to approximately 80 mbar), transferred to a reduction tube containing a catalyst Fe and reducing agent Zn using a liquid nitrogen cold trap, and sealed with a torch. Then, the reduction tube was placed inside a muffle furnace or a customized graphite reduction furnace, whereby graphite formed at the surface of the iron powder in the reduction tube. Finally, the graphite and iron powder were pressed into the AMS cathodes for measurement.
Type of Reduction Furnace
Most 14C laboratories use a muffle furnace for oxidation combustion and CO2 reduction treatment (i.e., the entire sealed Zn tubes are held at a constant temperature). To explore better synthesis conditions of graphite for the Zn-Fe method, in addition to the conventional muffle furnace method, we also adopted a 10-hole Condensing Graphite Reduction Furnace (CGRF) (Shen et al. Reference Shen, Tang and Wang2022c) for CO2 reduction, which was initially designed for the H2-Fe method and could apply a thermal gradient to Zn tube reactors during graphitization. With the 10-hole condensing graphite reduction furnace, the Fe and Zn catalysts are held at high reaction temperatures (∼600°C), whereas the tops of the Zn tube reactors are held at ambient temperatures (20–25°C). The principle of the method is very similar to that described by Xu et al. (Reference Xu, Trumbore and Zheng2007) and Orsovszki et al. (Reference Orsovszki and Rinyu2015) for the graphitization of 1–100 µg C samples. However, our study focuses primarily on improving Zn graphitization efficiency for 1 mg C samples.
Pictures of our reduction furnace and two Zn reactor tubes after different furnace graphitization processes are shown in Figure 2. At 600°C, zinc evaporates from the bottom of the outer tube, and the zinc oxide resulting from the reduction of CO2 forms a white band on the cooler part of the inside of the outer tube close to the heat source. In addition, excess zinc forms another fluffy band close to the cooler part of the reaction cell (Figure 2b), increasing the efficiency of filamentous graphite and/or Fe-carbide formation (Orsovszki et al. Reference Orsovszki and Rinyu2015).
Figure 2 A customized 10-hole graphite reduction furnace (a) and the reduction tube after graphitization using the customized furnace (b) and muffle furnace (c).
The performance of the graphite, such as 12C– beam current, 14C background, and recovery, synthesized by a condensate graphite reduction furnace and a muffle furnace, were compared, as shown in Table 1. There was no significant difference in 12C– beam current and 14C blank values between the two types of furnaces. However, the recovery of graphite synthesized by a condensate graphite reduction furnace is much higher than that of a muffle furnace, which further supports the experimental views of Orsovszki et al. (Reference Orsovszki and Rinyu2015) and Santos et al. (Reference Santos, Mazon and Southon2007b).
Table 1 Experimental results of reduction furnaces.
Pretreatment and Dose of CuO
Pretreatment of CuO
The pretreatment was meant to clean up the potentially absorbed atmospheric CO2 on the surface of CuO at high temperatures. In our study, ∼1 mg of commercial graphite was used as the blank, and ∼20 mg of CuO was precleaned in three groups: 900°C for 3 hr, 600°C for 3 hr, and an untreated group. The AMS measurement results are shown in Table 2.
Table 2 AMS results of different CuO treatments.
The data show that the trace amount of modern carbon in CuO will slightly affect the experimental background. The treatment of CuO at 900°C for 3 hr and 600°C for 3 hr is better. Considering the possible decomposition reaction of CuO at high temperatures and the pollution caused by reoxidation after cooling, we adopted 600°C for CuO pretreatment.
Dose of CuO
Considering that excess CuO may introduce modern carbon (Zhou et al. Reference Zhou and Zhang2001), we checked the relationship between the CO2 recovery rate in the combustion reaction and the amount of CuO used. Commercial graphite (∼1 mg) mixed with CuO sample sizes from 20 mg to 140 mg was placed in a muffle furnace at 900°C for 2 hr to fully react and generate CO2, quantified by measuring the CO2 pressure in our vacuum line. The results are shown in Table 3. As seen from the table, the CO2 recovery rate is stable at approximately 98%, indicating that the 20–30 mg CuO dose is sufficient for ∼1 mg of graphite to be fully oxidized to CO2 in our experimental conditions.
Table 3 Experimental results of CuO dose and CO2 recovery rate in the combustion.
Pretreatment and Dose of Fe powder
Pretreatment of Fe powder
Pretreatment of Fe powder is a key step in the 14C graphite synthesis process to clean up the potentially absorbed atmospheric CO2. Therefore, the sample contaminations from different Fe treatment processes were investigated. Approximately 3 mg of Fe powder was pretreated in five groups, i.e., heated under vacuum at 600°C for 0.5 hr, 1 hr, and 2 hr, heated at 400°C in open air for 3 hr, and an untreated group. The AMS measurement results are shown in Table 4.
Table 4 Results of the Fe powder pretreatment experiment.
The data show that the 14C backgrounds before and after the treatment of Fe powder are significantly different. The average background value heating at 400°C in open air is 0.54± 0.02 pMC, and the fluctuation is lower than in other conditions. Under vacuum conditions, the average background values of heating at 600°C for 0.5 hr, 1 hr, and 2 hr are 0.58± 0.17 pMC, 0.57 ± 0.05pMC, and 0.59± 0.07 pMC, respectively. There were no significant differences in the 14C background between the treatment groups under open air and under vacuum for different times. The average 12C– current is greater than 30 μA for the Fe powder heated under vacuum at 600°C for 0.5 hr, 1 hr, and 2 hr, and the average 12C– current is greater than 40 μA for the Fe powder heated at 400°C in open air for 3 hr, which is higher than that under other treatment conditions. The graphite recovery rates are all above 80% for Fe powder heating at 400°C in open air and heating at 600°C under vacuum for 0.5 hr and 1 hr, and the fluctuation is low. Additionally, we did not find significant differences in the graphite recovery rates between the treatment groups under open air and under vacuum for different times. Therefore, air heating at 400°C for 3 hr is an ideal pretreatment method for us considering the 12C– current, background, recovery rate, and convenience factor.
Ratio of Fe/C
The Fe/C ratio affects the values of the 12C– beam current and background of 14C. With increasing Fe powder dose, sample carbon contamination will increase by 0.35 μgC per mg Fe (Dee et al. Reference Dee and Ramsey2000). However, a low Fe/C ratio will cause beam current instability (Jull et al. Reference Jull, Donahue and Hatheway1986). Therefore, we investigated the relationship between the background level and the Fe/C ratio. For the commercial graphite of ∼1 mg, the sample sizes of catalyst Fe powder ranged from 0.1 mg to 5 mg (after heating in open air at 400°C for 3 hr), as shown in Figure 3. With the decrease in the Fe/C ratio, the background of 14C/12C decreases slowly before finally reaching a balance value of approximately 6.5×10–15, which is a considerably better value when the mass ratio of Fe/C is lower than 3.
Figure 3 Relationship between Fe/C and 14C/12C.
To explore the lower limit ratio of Fe/C, we also checked the relationship between the 12C– beam current value and the Fe/C ratio, as shown in Figure 4(a). With increasing Fe/C, the 12C– beam current decreases gradually because the decrease in the amount of carbon attached to the unit Fe powder affects the 12C– beam current. Meanwhile, we explored the 12C– beam current when Fe/C<1, as shown in Figure 4(b). With the decrease in Fe/C, the 12C– beam current first increased and then decreased. The maximum value (≈43 μA) is obtained when Fe/C is approximately 0.5.
Figure 4 Relationship between 12C– beam current and Fe/C ratio. (a) Fe/C >1, and (b) Fe/C <1.
However, the average graphite recovery rate under this condition is only approximately 50%, and the durability and stability of the 12C– beam current with an Fe/C ratio of 2.5 were better than those with an Fe/C ratio of 1.5, which is consistent with the conclusion of Orsovszki et al. (Reference Orsovszki and Rinyu2015). Therefore, the ratio of Fe/C is determined to be 2.5–3.0 in our laboratory, considering both the background value and the beam current durability.
Ratio of Zn/C
According to the chemical reaction formula, the mass ratio Zn/C required for the complete reaction of the equation is 10.88, and with the increase in Zn/C, the graphite yield and 12C– beam current increase rapidly (Macario Reference Macario, Alves and Oliveira2016). However, excess Zn powder may also introduce modern carbon contamination. Therefore, we checked the relationship between the 14C/12C backgrounds and the amount of Zn powder used. For the CO2 containing ∼1 mg carbon, the reducant Zn powder sample sizes ranged from 10 mg to 60 mg, as shown in Figure 5. With the increase in the Zn/C ratio, the background of 14C/12C decreases slightly and then gradually increases, indicating that the Zn powder contains a small amount of modern carbon, which considerably influences the 14C background.
Figure 5 Relationship between Zn/C and 14C/12C of blank samples.
As seen in Figure 5, the 14C/12C background value is stable at approximately 6.5×10–15 when the sample size of Zn is 15–25 mg, which is the most suitable range for the reduction of 1 mg of graphite. When the amount of Zn powder is more than 30 mg, the background increases with the increased sample size of Zn. Another possible reason is that when the amount of Zn is excessive, the carbides formed by Zn and C exist in the mixture of Fe powder and graphite, which affects the synthesis of graphite and increases the experimental background (Santos et al. Reference Santos, Southon and Griffin2007b). Therefore, the ratio of Zn/C was determined to be 18–22 mg Zn/mg C in our laboratory.
Temperature and Time of Reduction
The recovery rate of graphite is important and sensitive to the temperature and time of reduction. Therefore, we investigated the relationship between temperature or time of reduction and recovery rate when the other conditions were the same. The influence of the reduction temperature on the recovery rate is shown in Figure S1(a). The recovery rate increases with the reduction temperature and eventually reaches a balance value of approximately 90% at temperatures ranging from 650°C to 700°C. Figure S1(b) shows the influence of reduction time on the recovery rate, which increases with the reduction time and reaches an equilibrium value of 8–9 hr, with a recovery rate of over 98%. For these reasons, our lab’s reduction temperature and time are set to 650°C–700°C and 8–9 hr, respectively.
AMS MEASUREMENTS
According to the established experimental conditions for preparing 14C graphite using the Zn-Fe method, the samples were oxidized, purified, graphitized, and finally measured with AMS. Commercial blank graphite was directly measured to check the machine background, and the results are shown in Figure S2(a). The background value of unprocessed commercial graphite was 0.27 ± 0.02 pMC, and the 14C/12C value was 3.14 ± 0.27 × 10–15, equivalent to a 14C age of approximately 47,000 years. The experimental results for the process blank of the Zn-Fe method are shown in Figure S2(b). The background value ranged between 0.49 pMC and 0.62 pMC, the mean value was 0.55 ± 0.04 pMC, and the 14C/12C ratio was 6.47 ± 0.48 × 10–15, equivalent to a 14C age of approximately 40,500 years. In addition, three international standard samples, OXII, Chinese sugar carbon (CSC), and IAEA-C8, were prepared and measured using AMS. The experimental results for the oxalic acid standard are shown in Figure S2(c). The average pMC of oxalic acid was 134.11 ± 0.41 pMC, which is consistent with the recognized standard value of 134.07 pMC within the allowable error (the precision of the system ∼ 0.6%). The pMC value of CSC was 136.32 ± 0.48 pMC, which was consistent with the recognized standard value of 136.2 pMC. Moreover, the pMC value of IAEA-C8 was 15.17 ± 0.22 pMC, as shown in Figure S2(d), which was also consistent with the recognized standard value of 15.03 pMC.
CONCLUSION AND DISCUSSION
In this study, 14C samples of 1 mg of carbon were prepared by the Zn-Fe reduction method and verified with AMS under the following optimal conditions: 20–30 mg CuO powder was pretreated at 600°C for 3 hr, 2.5–3.0 mg Fe powder and 18–22 mg Zn powder were pretreated at 400°C in open air for 3 hr, and reduction treatments were performed in a condensed graphite reduction furnace at 650°C for 8 hr. The AMS measurement results show that the average value of OX-II is 134.11 ± 0.41 pMC, the CSC average value is 136.32 ± 0.48 pMC, and the IAEA-C8 average value is 15.17 ± 0.22 pMC, all of which are consistent with the recognized standard value. The 14C/12C ratio of the process blank was (6.47 ± 0.48) × 10–15, equivalent to a 14C age of approximately 40,500 years. After deducting the machine background of (3.14 ± 0.27) × 10–15, the background induced by the sample preparation process with the above method was (2.06 ± 0.55) × 10–15. The preliminary results show that this sample preparation method is reliable, can provide technical support for the development of 14C dating and tracing at GXNU, and contributes to biology and environmental science.
ACKNOWLEDGMENTS
This work was supported by the Central Government Guidance Funds for Local Scientific and Technological Development, China (No. Guike ZY22096024), the Guangxi Natural Science Foundation of China (Nos. 2019GXNSFDA185011 and 2017GXNSFFA198016), the National Natural Science Foundation of China (Nos. 11775057, 11765004, 12065003, and 12164006), and JSPS KAKENHI under Grant No. 23K23269.
SUPPLEMENTARY MATERIAL
To view supplementary material for this article, please visit https://doi.org/10.1017/RDC.2024.73
