1. Introduction
The graphitization process, required for the preparation of solid carbon samples for 14C AMS, has two disadvantages: it is laborious and contaminates the samples, compromising the measurement background and precision at masses below 150 µgC. Avoiding graphitization, the AMS analysis of CO2 samples is attractive for biomedical applications with high sample throughput and for the dating of small sample masses. The AMS measurement of CO2 samples requires a gas handling device that transfers the CO2 from the gas source to the AMS system with a specific flow and timing, and an ion source capable of producing a C– beam from CO2 (Bronk Ramsey and Hedges Reference Bronk Ramsey and Hedges1994; Wacker et al Reference Wacker, Fahrni, Hajdas, Molnar, Synal, Szidat and Zhang2013).
The HVE AMS ion sources support the operation with gas samples and are used with gas interfaces independently developed by several laboratories (Braione et al Reference Braione, Maruccio, Quarta, D’Elia and Calcagnile2015; Duijn et al Reference Duijn, Sandman, Grossouw, Mocking, Coulier and Vaes2014; Klein et al Reference Klein, Mous and Scognamiglio2023; Stolz et al Reference Stolz, Dewald, Altenkirch, Herb, Heinze, Schiffer, Feuerstein, Müller-Gatermann, Wotte, Rethemeyer and Dunai2017, Reference Stolz, Dewald, Heinze, Altenkirch, Hackenberg, Herb, Müller-Gatermann, Schiffer, Zitzer, Wotte, Rethemeyer and Dunai2019, Reference Stolz2020). Recently, HVE designed and manufactured its own gas interface and performed the first tests to determine its contributions to the AMS measurement in terms of background, precision, and efficiency.
2. The gas interface system
The HVE gas interface system is based on the design developed at the University of Cologne (Stolz et al Reference Stolz, Dewald, Altenkirch, Herb, Heinze, Schiffer, Feuerstein, Müller-Gatermann, Wotte, Rethemeyer and Dunai2017, Reference Stolz, Dewald, Heinze, Altenkirch, Hackenberg, Herb, Müller-Gatermann, Schiffer, Zitzer, Wotte, Rethemeyer and Dunai2019, Stolz Reference Stolz2020). It connects the AMS system to different peripherals (Figure 1), accepting gas from lecture bottles, typically containing a blank or standard gas for calibration purposes, and other devices. The external gas source used for the tests is an elemental analyzer (EA), whose operation is controlled by the gas interface software. The EA combusts solid or liquid samples and separates in time the gases resulting from the combustion, which reach the EA exit port in a He stream. A thermal conductivity detector generates a signal for each combustion product and allows the identification of the CO2 peak, enabling the transfer to the gas interface system. The mixture of CO2 and He passes through a zeolite trap at room temperature, which adsorbs the CO2 only, while the He is directed to waste. By heating the trap to 500ºC, the CO2 is released from the zeolite and transferred to a motor-driven syringe. CO2 from lecture bottles, instead, flows directly to the syringe. CO2 in the syringe is then diluted with He to the desired mixing ratio. The syringe drives the mixture of CO2 and He via a capillary to the AMS ion source with a controlled gas flow rate. Between processing two samples, the gas interface system runs a cleaning procedure, consisting of purging the trap, syringe, and capillaries with He. The trap is heated to 500ºC during purging and set back to room temperature before accepting the next sample.
The gas interface system connects the AMS machine with different peripherals.
The gas interface system features two syringes and two zeolite traps, which are used in parallel to minimize the dead time required for purging, as shown in Figure 2. The resulting sample throughput of more than 10 samples per hour is limited only by the combustion time needed by the EA.
Schematics of the HVE gas interface. Having two zeolite traps and two syringes allows parallel operation and fast sample throughput. Switching valves direct the CO2 to the desired trap or syringe. While one trap is collecting CO2 from the EA combustion (yellow arrows), the other trap releases the CO2 to the syringe (red arrows). Similarly, when one syringe is loaded (red arrows), the other injects CO2 into the AMS ion source (green arrows). He can flow through the trap to collect the residual CO2, to the syringe to dilute pure CO2 or to the AMS ion source to purge the transfer capillary while pre-sputtering the targets (blue arrows).
3. Test setup
The described measurements are performed using the test setup shown in Figure 3. The HVE gas interface system is coupled to a ThermoFisher FlashSmart EA and a HVE 210 kV 14C–dedicated AMS machine, whose overall performance is described in (Klein et al Reference Klein, Podaru and Mous2019; Scognamiglio et al Reference Scognamiglio, Klein, van der Hoef, Podaru and Mous2021).
The test setup used for the measurements, with the HVE gas interface accepting CO2 samples from a ThermoFisher FlashSmart EA and transferring CO2 to the HVE 210 kV AMS machine.
The CO2 from the gas interface is injected into the AMS ion source head through a capillary with a flow rate of about 1.5 µgC/min following the results at CologneAMS (Stolz et al Reference Stolz, Dewald, Altenkirch, Herb, Heinze, Schiffer, Feuerstein, Müller-Gatermann, Wotte, Rethemeyer and Dunai2017). The ion source is a hybrid SO-110C Cs sputter ion source, which accepts both solid and CO2 samples (Klein and Mous Reference Klein and Mous2017). The CO2 is guided onto the Ti insert mounted at the tip of the gas target piece and is adsorbed on the Ti surface from where it is sputtered, as shown in Figure 4. To minimize cross-talk effects between one sample and the next, the target piece is replaced and the capillary between the gas interface and the AMS is flushed with He.
Left: The sample CO2 reaches the source head via the target holder and is guided onto a Ti insert placed in the target piece. Right: Section of a gas target piece.
The samples used for the tests have known 14C/12C isotopic ratios and are listed in Table 1.
Samples used for the tests
Blank and OxII CO2 are admitted to the interface from gas bottles. These samples go directly from the bottle to the syringe. Solid samples supplied by the IAEA with 14C/12C isotopic ratios ranging between 3 · 10–15 and 1.5 · 10–12 are combusted in the EA. To make sure that the desired number of 14C counts is reached and that the AMS analysis doesn’t stop due to lack of sample material, a high sample mass is used for combustion (> 200 µgC per sample) and the excess is discarded at the end of the AMS analysis.
All non-prediluted samples are mixed with He in the syringe to obtain a final sample with a CO2 content of about 5% for optimal source efficiency.
4. Measurement results and discussion
The experiments with CO2 samples have been done to characterize the interface contribution to the efficiency, precision, and background of the AMS measurements. The measurement time has been 15 minutes for OxII samples and 20 minutes for blank samples.
4.1. Source output
During the tests, the ion source is operated with a target voltage of 10 kV, an extraction voltage of 30 kV and a Cs reservoir temperature of 110ºC to obtain 12C– currents of a few µA. Figure 5 shows the typical 12C– beam current during the AMS measurement of the OxII reference gas.
12 C– beam current of an OxII bottle gas sample measured in the Faraday cup at the accelerator entrance. The plot shows a peak in the current due to the target insertion (A), the admission of sample CO2 into the ion source (B) and the purging of the feed capillary with He (C).
Before admitting the sample CO2 into the ion source and starting an AMS analysis, a new target piece with Ti insert is loaded and pre-sputtered, while pure He is injected into the ion source. Environmental CO2 that is adsorbed on the surface of the Ti insert generates a small pulse of 12C– output current, that quickly drops to a background current of about 50 nA. When this low current is stabilized, the target is considered to be cleaned by sputtering.
Upon the CO2 injection into the ion source at a flow rate of 1.5 µgC/min, the 12C– beam current rises within seconds to a value between 2 and 2.5 µA, corresponding to an ion source efficiency of about 1%. At the end of the AMS measurement, the gas feed system is purged with He, and the target piece is replaced.
Higher currents are expected to be obtained with a higher Cs reservoir temperature. Furthermore, in a study realized by CologneAMS on their SO110B ion source, it has been shown that the Cs beam is focused on the Ti insert on a small spot of less than 1 mm in diameter and that the ion source efficiency can increase to 3% by enlarging the sputtering area (Stolz et al Reference Stolz, Dewald, Altenkirch, Herb, Heinze, Schiffer, Feuerstein, Müller-Gatermann, Wotte, Rethemeyer and Dunai2017). This was achieved by changing the target piece’s shape and the ion source geometry. Further investigations are required to confirm that similar changes beneficial to the sputtering efficiency can be applied to the SO110C.
4.2. Background
The 14C/12C background ratio is measured as 2.3 · 10–14 ± 10% from the blank gas bottle and 2.8 · 10–14 ± 9% from the IAEA-C-4 wood, with the measurement error determined by the counting statistics. The small discrepancy can be explained by the use of the EA and the trap for the IAEA-C-4 samples, which are not used for gas bottle samples. The contribution of the AMS machine is in the mid 10–16 level for solid samples and therefore insignificant for gas measurements.
The measured background is about 2 PMC, which is consistent with the background current being about 2% of the 12C– measured current (see Section 4.1). This indicates that contamination of the target piece represents the main contribution to the gas sample background and that the gas interface contribution is therefore in the 10–15 level. Improving the pre-sputtering procedure, e.g. extending its duration, is expected to provide a reduction of the measured 14C/12C background to the 10–15 level, as measured at CologneAMS (Stolz et al Reference Stolz, Dewald, Heinze, Altenkirch, Hackenberg, Herb, Müller-Gatermann, Schiffer, Zitzer, Wotte, Rethemeyer and Dunai2019).
4.3. Precision and linearity
The precision was determined by measuring 10 targets that were fed with OxII gas samples from the bottle and 10 targets that were fed with IAEA-C-3 samples combusted in the EA in two different batches. The results are plotted in Figure 6 and listed in Table 2.
Results of the AMS measurement of 10 OxII and 10 IAEA-C3 standards. Each data point is reported with the statistical error. The orange lines indicate the calculated average value ± its standard deviation. The normalization factor, defined as the ratio between the nominal and the measured ratio, is about 1.18.
AMS results of two batches of reference samples. Each batch measured 10 samples of the same kind, whose values are shown in Figure 6
Since the 13C/12C ratio is not influenced by the counting statistics, it is a good indication of the AMS machine’s stability and achievable precision. The 13C/12C ratios of the combusted samples and gas bottle samples have a standard deviation of 0.4% and 0.23%, respectively. Since the AMS system precision measured for solid samples is 0.2%, the gas interface contribution to the 13C/12C ratio error of bottle samples is minor. The EA and the traps are instead the main contribution when combusted samples are measured.
Concerning the 14C/12C precision, a statistical error of 1.3% and a standard deviation of 1.8% are measured for both batches. The main contribution to the 14C/12C achievable precision is therefore the counting statistics in this case. Further investigations with a higher number of counts are required to determine the contribution of the gas interface to the 14C/12C precision. However, the measured 13C/12C precision suggests that a 14C/12C precision at or below 1% is achievable.
IAEA standards with different 14C/12C isotopic ratios are used to determine the linearity of the measurement with the setup based on EA, gas interface and AMS system. The measured ratios are displayed in Figure 7. Since the analysis time is the same for all target pieces, the uncertainties are larger for samples with a lower 14C content and consequently poorer counting statistics. The slope of the linear fit suggests an average measured/nominal ratio of 77%, which is lower than the typically measured 85% (Table 2). The differences are explained by the applied AMS system settings (especially the einzel lens), which were not adjusted for gas measurements after preceding solid sample tests using 30-fold higher beam currents.
Measured (non-normalized) 14C/12C ratio of different IAEA samples as a function of their nominal value.
5. Conclusions and outlook
A gas interface between an EA and an AMS system has been developed by HVE in collaboration with the University of Cologne. The prototype has been manufactured and coupled to the HVE 210 kV AMS system for preliminary functional and performance tests.
The ion source settings have been kept moderate resulting in a 12C– current of 2–3 µA. The measured background level in the low 10–14 is consistent with the 50 nA current measured without CO2 admission and is related to contamination of the target pieces. Consequently, the background contribution from the gas interface system is in the 10–15 regime. The 14C/12C precision measurement was limited to 2% by counting statistics. However, a gas interface contribution at or below 1% is anticipated, since the measured 13C/12C precision is below 0.5%.
