Available online at www.sciencedirect.

com

ScienceDirect
Physics Procedia 90 (2017) 10 – 16

Conference on the Application of Accelerators in Research and Industry, CAARI 2016,

30 October – 4 November 2016, Ft. Worth, TX, USA

Acceptance Tests for AMS Radiocarbon Measurements at

iThemba LABS, Gauteng, South Africa
Vela L. Mbele*, Simon M. Mullins, Stephan R. Winkler, Stephan Woodborne
iThemba LABS, National Research Foundation, P. Bag X11, Wits 2050, South Africa

Abstract

The accelerator mass spectrometer was commissioned recently at the iThemba LABS 6 MV tandem accelerator.
Improvements in the vacuum system, requiring procurement of cryo-pumps and the reducing the tank pressure of the
N2 + CO2 insulation gas mixture below the level used for IBA measurements, were necessary. This resulted in the
reduction of the nitrogen background and improved the resolution of 14C from 14N background in the ionisation
chamber. The nitrogen was leaking to the stripping canal because of inadequate sealing. The analysing magnet was
scaled to detect C3+ ions, at 3 MV terminal potential. The first sensible spectra allowed for the pin-pointing of many
persistent issues. This resulted in measurements with a precision better than 1 pMC, and current blank levels
correspond to 12 half-lives of 14C or ~68000 years. The radiocarbon sample preparation laboratory has reached
production status. A brief outlook of the work towards the implementation of the measurement and chemical
preparation protocols for radionuclides 10Be and 26Al is also summarised in the conclusion
© 2017 The Authors. Published by Elsevier B.V.
© 2017 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
Peer-review under responsibility of the Scientific Committee of the Conference on the Application of Accelerators in Research
Peer-review
and Industry.under responsibility of the Scientific Committee of the Conference on the Application of Accelerators in Research and Industry

Keywords: Accelerator mass spectrometry; AMS; IBA; South Africa; radiocarbon; beryllium; aluminium, tandem; accelerator

* Corresponding author. Tel.: +27 11 351 7032; fax: +27 11 351 7053.

E-mail address: mbele@tlabs.ac.za

1875-3892 © 2017 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
Peer-review under responsibility of the Scientific Committee of the Conference on the Application of Accelerators in Research and Industry
doi:10.1016/j.phpro.2017.09.009
 Vela L. Mbele et al. / Physics Procedia 90 (2017) 10 – 16 11

1. Introduction

It has long been thought that research programmes at iThemba LABS should be extended to include Accelerator Mass
Spectrometry (AMS). The idea dates to the times the branch of iThemba LABS at the University of the Witwatersrand,
Johannesburg was part of the university, and known as WITS-CSIR Schonland Research Centre for Nuclear Sciences
(SRCNS). To this end, Prof J. P. F. Sellschop, who founded the SRCNS originally known as the Nuclear Physics
Research Unit as part of the department of physics in 1958, attended and gave the closing remarks at the Fourth
International Symposium on Accelerator Mass Spectrometry held at Niagara-on-the-Lake, Canada in 1987 (Sellschop
1987). After the SRCNS had been donated to (South African) National Research Foundation in 2004, the first head
of department of this new era went on an IAEA funded fact finding trip to ETH-Zurich. When SMM became the head
of department and, following repeated applications to funding agencies, the proposal was picked for funding, and
preparation for the procurement of the AMS detection systems began in earnest.

In anticipation of the inclusion of the AMS various changes were made to the accelerator, beamlines and the vault.
These changes were not only meant to benefit the AMS but also looked to improve detection limits in the Ion Beam
Techniques and Nuclear Physics detection beamlines. Summaries on the states of these particle experiments can be
found in various publications (Andeweg, et al., 1997; Msimanga, et al., 2013, 2016; Przybylowicz, et al., 1993). Some
of these articles demonstrate the thought, towards the expansion of the research themes of the SRCNS, and eventually
iThemba LABS, which was originally a nuclear physics research facility, to a more inclusive institute with research
interests extending to geology, archeology and possibly paleaontology. The AMS would then be at the centre of
operations.

The AMS development started with the preparation. Implementation of 14C, will, in the near future extend to 10Be and
26
Al, followed by the other radionuclides. These developments will be in response to research imperatives in the region.
The design of the system follows closely the Scottish Universities Environmental Research Centre (SUERC) 5 MV
AMS system. We hope in the long run, to measure a similar spectrum of isotopes to SUERC (Freeman, et al., 2004;
Maden, et al., 2007). The 14C for which capability has been established and is the subject of this report, will be
measured, together with the radionuclides 10Be and 26Al in chronological dating applications ranging from commercial
to academic research. Academic research will include assessments of landscape evolution in Southern Africa (Bierman
and Caffee 2001) and the continent (Brown, et al. 1994). We are looking to implement exposure (Bierman and Caffee
2001) and burial dating (Owen, et al. 2011) of cosmogenic radionuclides. We further hope to contribute to the
knowledge of species evolution by dating, in partnership with local and international colleagues, caves and fossil sites
in the region and the continent (Granger and Muzikar 2001). The context of southern African paleoscience places the
iThemba LABS AMS facility in a good position to contribute towards answering unique localized research
opportunities such as African human origins (Tattersal 2009).

2. The spectrometer

The improvements that preceded the installation of AMS spectrometer included better alignment of the injection and
extraction beam line optics. In the detection lines the changes included, but were not limited to, optical alignment of
the entire beamlines and detection lines, the upgrade of compressed air and cooling water systems, vacuum system
placed under computer control, the redesign and upgrade of the IBA injection source, refurbishment and installation
of larger beam line magnets, replacing the belt charging system by a Pelletron charging system, installing new high-
voltage grading resistors, installing new axial electric field and spiraled magnetic field accelerator tubes, installing a
new terminal potential stabiliser system (TPS), converting the gas stripper to a recirculating gas with HV terminal
pumping.

Changes were made to the building housing the accelerator. These included compartmentalising the accelerator
building into an accelerator vault, experiments hall for all detectors, electronics room, control room, and offices for
engineers and the technical staff. The installation of the AMS system included the low and high energy spectrometers.
The low energy spectrometer comprised the installation of the 64 cathode ion source, einzel lenses, electrostatic
12 Vela L. Mbele et al. / Physics Procedia 90 (2017) 10 – 16

steerers, various slits, beam profile monitors, electrostatic spherical analyser E/q, the installation of the 90 o injection
magnetic with bouncing capabilities, and, these were married into already existing IBA beamline, allowing for the
alternate use of the system operating the 860C negative sputter ion source, the Alphatross ion source or the AMS
source.

The high energy spectrometer, was installed behind the magnet directing the beam to the IBA experiments. The IBA
analysing magnetic is degaussed when the system is in the AMS mode. This spectrometer comprises a 90o 1270 mm
radius, ME/q2 = 176, M/ΔM = 725, double focusing magnet, three off axis movable Faraday cups, a 20o electrostatic
cylindrical analyser with a 3810 mm radius and E/ΔE=200, a quadrupole doublet and a switching magnet. There is
just one AMS detector connected through the 20o degree port of the switching magnet, at present. The five anode gas
ionisation detector is not connected directly to the switching magnet but is rather behind a quadrupole doublet (Maden,
et al. 2007). Although iThemba LABS is imposed the AMS system on a system already used for materials science and
nuclear physics research, the final configuration is similar to the SUERC 5 MV Pelletron AMS system (Freeman, et
al., 2004; Maden, et al., 2007). The most important components necessary for AMS operation in the iThemba LABS
tandem accelerator are shown in Fig. 1.

Figure 1: Schematic of the iThemba LABS AMS facility. The AMS facility has the low energy injection system comprising of the source, Einzel
lens, electrostatic cylindrical analyser, Einzel lens and the injection magnet. After the tank the high energy spectrometer is composed of the
electrostatic quadrupole triplet, X and Y steerers, IBA bending magnetic (degaussed during AMS operation), beam profile monitor, electrostatic
steerer, AMS bending magnet, offset Faraday cups, magnetic quadrupole, 20 o electrostatic analyser, switcher magnet with the magnetic
quadrupole, beam profile monitor and detector connected to the is 20 o degree port. Pumping stations along the beam lines and other important
accessory equipment are not labelled.

3. Experimental Challenges

The functionality of the components was tested at various stages. After the complete installation tests to check the
interfacing between the components and the marriage of the old High Voltage Engineering tank with the modern
National Electrostatics Corporation components, and the overall effect of the improvements to the AMS system, was
checked. Test runs provided inconsistent and unacceptable results. One of the biggest challenges in the 14C
measurements is the intermittent background of 14N, the source of which we have verified to be the accelerator
terminal. This interference could not originate from the ion source because nitrogen does not form negative ions.
Molecular interference is destroyed in the terminal stripper of the accelerator, or rather by selecting the 3+ or 4+
charge state through the Coulomb explosion of molecular isobars 13CH, and 12CH2 is assured.
 Vela L. Mbele et al. / Physics Procedia 90 (2017) 10 – 16 13

The background was from the tank insulation the N2 + CO2 gas mixture leaking into the gas stripper system. The leak
is because the components of a vacuum system were not coupled together in a high integrity leak-free fashion. Various
types of mating flanges were used incorporating the seals and gaskets, instead of using metal components exclusively.
The problem was further compounded by extremely poor vacuum at the accelerator base due to cryo-pump
malfunctions. The poor vacuum also led to conditioning bursts that created massive peaks of ionization from the gas
stripper, some with beam currents of up to 100 nA as measured in a Faraday cup after the tank. Under these conditions
the nitrogen background overwhelmed any signal in the detector after the mass spectrometer.

This challenge was overcome by remediating leaks on both the high-energy and low-energy base of the accelerator,
maintenance and replacement of the pumps, re-assembly of the gas stripper system, replacement of all O-ring seals,
lowering of the insulation gas pressure of the tank, and the procurement of metal O-rings designed for the high-
pressure to use in the stripper gas system. Not all these measures have been fully implemented, due to various
unforeseen delays, though the vacuum has been improved to the point that the nitrogen background and the
conditioning are manageable in the detector. Because of the vacuum issue the 6 MV tandem accelerator has only been
run at voltages about 4.5 MV or below, requiring conditioning up over several days, and hardly ever going beyond
5.1 MV. For 14C AMS there are 3+ and 4+ high energy charged states. At 5 MV the 4+ charge state would have been
used despite the possible 7Li2+ m/q interference, but under the conditions we selected the 3+ charge state with an
optimum theoretical charge state yield with the tandem terminal at about 3 MV (Suter 1990). The 14N4+ peak can be
clearly discerned in Fig. 2 and can be further resolved by improved to the system.

The low-energy mass spectrometer features a multi-beam switching facility, which allows sequential transmission of
beams of different mass to the accelerator in cycles of about 0.1 ms. With an extraction voltage of 40 kV and the
capability of slowing down the ions in the magnet with a 5 kV field it is possible to implement fast switching for 14C
and 13C. However, with the +/-5kV power supply we are limited to a 14C/13C measurement for radiocarbon without
the online δ13C capability. In order to measure the 14C/13C ratio the magnetic field is adjusted for the injection of mass
14 and mass 13 injection by applying positive voltage (acceleration-deceleration). This is because only one polarity
contributes to the effective range as the focusing is significantly different for the case of deceleration then acceleration,
versus the opposite case of acceleration/deceleration configuration. These different focusing conditions led to 12C and
13
C injection not being optimized simultaneously with regard to focusing, and that hampers the reproducibility of the
high-energy 13C versus 12C ratios. We mitigate this by measuring δ13C “off-line” using traditional mass spectrometry.
One of the big upsides of the low-energy injector is the current output achieved by the ion source.
12
C− beam currents of 100 μA and more could be extracted easily from the sputter targets in the low-energy offset
Faraday cup after the injection magnet. This offset cup is positioned so that 12C− is collected when 14C− is injected into
the accelerator. While this cup position will introduce complications for other isotopes, it is very convenient for
radiocarbon measurements. It also became necessary to reprogram the AccelNET gating and sequencing setup for the
data acquisition system, as this was designed for an injector where the cup is either movable in x, or placed in the 13C−
position.

4. Experimental Results

With the improved understanding of the system it was possible to test the reproducibility of our AMS-system for
14
C. In addition to several cathodes with standard graphite that had been used during the commissioning phase we
added three new OXI (NIST oxalic acid I) standard samples and four new machine blank samples (Merck UPC-1
unprocessed graphite powder 99.995%) for this test run.

We performed nine measurements on each sample over three minutes. The two dimensional spectra shown in Fig. 2
demonstrate the persistent 14N4+ peak in the OXI standard and while the blank (UPC) shows no peak other than the
nitrogen. The average absolute (14C/13C) ratio over the three samples was (1.008±0.001)∙10 -10 with standard deviation
of the mean agreeing with the uncertainty derived from counting statistics. This ratio, with the reported uncertainty,
is not in agreement with the absolute nominal (14C/13C) ratio of (1.068)∙10-10 however represents a step in the right
14 Vela L. Mbele et al. / Physics Procedia 90 (2017) 10 – 16

direction. AMS measurements use standards to normalize isotope specific differences in instrument efficiency. Some
of the 14C is lost due to the cuts in the spectra, there is a difference in charge state transmission (different velocities
for 13C compared 14C when interacting with the stripper gas), and potentially there are additional losses of 14C relative
to 13C through the electro-static analyser, the switching magnet, and the 8x8mm silicon nitride (Freeman, et al. 2004)
detector window.

The reduced Chi-squared for the twenty seven runs (hypothesis same value) is 1.22, just within one standard deviation
for the Poisson distribution over twenty-six degrees of freedom. The machine blank over four samples came to a
14 13
C/ C ratio of (2.15±0.15)∙10-14 – almost a factor of 5000 lower than the primary standard – which would correspond
to a conventional age of ~68000 years.

Figure 2. 2d-spectra (ΔE2 vs ΔE1; ΔE3 vs ΔE1) for standard sample material OXI and blank sample material UPC-1. While the nature of the
events in ΔE2 vs ΔE1 for the blank is unclear, ΔE3 vs ΔE1 suggest that some of them are real 14C counts matching the centre of mass of the
14
C-bin.

5. Conclusion and Outlook

14
Amid slow progress the iThemba LABS spectrometer is imminently ready for C AMS measurements. After it
 Vela L. Mbele et al. / Physics Procedia 90 (2017) 10 – 16 15

became clear that the 13C/12C ratio online cannot be measured online plans are underway to measure δ13C offline. This
will be achieved by using a standalone mass spectrometer. The issues relating to the background from the insulating
gas leaking into stripping canal and interfering in the 14C measured in the gas ionisation detector are being attended
to. We have attacked this problem in two fronts, by improving the seals and gaskets and also by lowering the insulating
gas mixture pressure. The 14C/13C ratio is getting closer to the expected nominal ratio. There should be even better
improvements when all seals and gaskets have been replaced by metal vaccuum vetted accessories.

The radiocarbon sample preparatory chemistry laboratory is now in full operation. We have also begun tests on 26Al
AMS measurements. We are looking to use the foil-stack to limit the interference from the stable isobar 10B in 10Be
AMS measurements reaching the detector, as opposed to the popularly used gas cell to limit the 10B interference. It is
also hoped that over time the second detector will be installed on one of the other ports of the switching magnet. This
detector when installed will be dedicated to.the detection of 10Be and perhaps even 36Cl. In the case of 36Cl the foilstack
will be used for limiting the 36S interference. A single polypopylene extraction fume cupboard for cosmogenic
radionuclide sample preparation laboratory with a capability for preparing fifty 10Be and fifty 26Al per month is being
tested for acceptance and will be reported on shortly. As one of the ways to test the integrity of this fume cupboard
and the laboratory will be by measuring process blank levels.

Even though the tandem is currently optimally operated at a relatively low achievable terminal potential for Cl AMS
the machine maybe capable to measure 36Cl, with some effort focussed on understanding the effect of the background.
Further experience will determine whether the routine lower than design potential operation is practical, but other labs
have demonstrated these measurements are possible at even lower terminal voltages (Martschini, et al. 2013)

Acknowledgements

We are grateful to colleagues in the accelerator department of the NRF iThemba LABS for having led efforts of the
installation of the accelerator mass spectrometer. We appreciate the assistance of colleagues from various laboratories
around the world, in particular, Peter Steier (VERA), Tom Brown (LLNL), and Albert Zondervan (GNS Science).
These efforts have been supported by the IAEA. This work is based on the research supported by the National Research
Foundation of South Africa (Grant Number 98920).

References

Andeweg, A. H., S. Ballestrero, J. U. M. Beer, J. E. Butler, M. B. H. Breese, S. H. Connell, L. Dini, et al. 1997. “The Schonland Micro-Scanning
Ion Beam Analysys Facility.” Nuclear Instruments and Methods in Physics Research B 37 - 44.

Bierman, P. R., and M. Caffee. 2001. “Slow Rates of Rock Surface Erosion and Sediment Production accross the Namib Desert and Escarpment,
Southern Africa.” American Journal of Science 326-356.

Brown, E. T., D. L. Bourlès, F. Colin, Z. Sanfo, G. M. Raisbeck, and F. Yiou. 1994. “The development of iron crust lateritic systems in Burkina
Faso, West Africa examined with in-situ-produced cosmogenic nuclides.” Earth and Planetary Science Letters 19-33.

Freeman, S., P. Bishop, C. Bryant, G. Cook, A. Fallick, D. Harkness, S. Metcalf, M. Scott, R. Scott, and M. Summerfield. 2004. “A new
environmental sciences AMS laboratory in Scotland.” Nuclear Instruments and Methods in Physics Research Section B 31-34.

Granger, D. E., and P. F. Muzikar. 2001. “Dating sediment burial with in situ-produced cosmogenic nuclides: theory, techniques, and
limitations.” Earth and Planetary Science Letters 269-281.

Maden, C., P. A. F. Anastasi, A. Dougans, S. P. H. T. Freeman, R. Kitchen, G. Klody, C. Schnabel, M. Sundquist, K. Vanner, and S. Xu. 2007.
“SUERC AMS ion detection.” Nuclear Instruments and Methods in Physics Research Section B 131-139.

Martschini, M., P. Andersson, O. Forstner, R. Golser, D. Hanstorp, A. O. Lindahl, W. Kutschera, et al. 2013. “AMS of 36Cl with the VERA 3 MV
tandem accelerator.” Nuclear Instruments and Methods in Physics Research Section B 115-120.

Msimanga, M., C. Pineda-Vargas, and M. Madhuku. 2016. “K-shell X-ray production cross sections in Ti by 0.3–1.0 MeV/u 12C and 28Si ions for
heavy ion PIXE.” Nuclear Instruments and Methods in Physics Research Section B 90-93.
16 Vela L. Mbele et al. / Physics Procedia 90 (2017) 10 – 16

Msimanga, M., D. Wamwangi, C. M. Comrie, C. A. Pineda-Vargas, M. Nkosi, and T. Hlatshwayo. 2013. “The new Heavy Ion ERDA set up at
iThemba LABS Gauteng: Multilayer thin film depth profiling using direct calculation and Monte Carlo simulation codes.” Nuclear
Instruments and Methods in Physics Research Section B 54-60.

Owen, L. A., K. Frankel, J. R. Knott, S. Reynhout, F. C. Finkel, J. F. Dolan, and J. Lee. 2011. “Beryllium-10 terrestrial cosmogenic nuclide
surface exposure dating of Quaternary landforms in Death Valley.” Geomorphology 541-557.

Partridge, T. C., D. E. Granger, M. W. Caffee, and R. J. Clarke. 2003. “Lower Pliocene Hominid Remains from Sterkfontein.” Science 607-612.

Przybylowicz, W. J., J. I. W. Watterson, H. J. Annegarn, S. H. Connell, R. W. Fearick, A. H. Andeweg, and J. P. F. Sellschop. 1993. “The new
Schonland PIXE microprobe and applications to geological and archaeological samples.” Nuclear Instruments and Methods in Physics
Research Section B 539-543.

Sellschop, J. P. F. 1987. “Progress in accelerator mass spectrometry.” Progress in accelerator mass spectrometry 439-445.

Suter, M. 1990. “Accelerator mass spectrometry: State of the art in 1990.” Nuclear Instruments and Methods in Physics Research Section B 221-
223.

Tattersal, I. 2009. “Human Origins: Out of Africa.” Proceedings of the National Academy of Sciences 16018-16021.