Modern bone contains ca 25% protein material, most of which is collagen. Amino acids separated from collagen isolated from bone are suitable for 14C dating of fossil bone, but attempts to carry out this procedure on bones seriously depleted in protein can yield erroneous 14C dates. Amino-acid analysis of fossil bone gives quantitative information on the degree of preservation of its organic component. Also, the relative abundance of the amino-acid components reveal the degree to which the collagen-like pattern has been altered. Alteration may be caused by addition of extraneous material. A 1mg sample of bone material is sufficient for this preliminary analysis. We have developed a series of acceptance criteria for whether a particular specimen is likely to yield the correct 14C age. 14C dating of fossil bones not seriously depleted in protein is a straightforward procedure and yields reliable dates.

Collagen-based pretreatment methods for bone yield inconsistent results for those samples where protein preservation is low, as frequently found in bones from the semi-arid zones of Australia and North America. New methods for dealing with low collagen bones are needed, and this paper suggests that the non-collagenous proteins, particularly the blood proteins, may offer advantages for AMS dating because of their better preservation. Amino-acid profiles of collagen and non-collagenous proteins suggest that such differential preservation may be due to physico-chemical differences, and help to explain the poor results from dating low-collagen bones.

A semi-automated continuous-flow system used to process archaeological bone to purified gelatin or amino acids for 14C dating is described. Powdered bone is retained in flow cells specifically designed to permit the sequential leaching of the bone with acid, alkali and water. Crude collagen obtained by this process is gelatinized, and than either purified directly using a macroporous cation exchange resin (BioRad AGMP-50), or hydrolyzed and the amino acids desalted on BioRad 50W-X8 resin. When compared with previous methods used by the laboratory, the new method allows more samples to be treated to a higher degree of purification. Examples of dates obtained on “standard” bones are presented, and confirm that no contamination is introduced from the components used in the new process.

Radon gas is a serious contaminant in radiocarbon dating by radiometry. The low specific ionizations associated with the α-particle emitting radon and its β-particle emitting daughters overlap within the 14C counting window. Elimination of radon is therefore imperative for precise 14C age determinations. This paper deals with the sources and mechanism of incorporation of radon affecting 14C dating by liquid scintillation (LS) counting, and reviews conventional radon elimination practices in 14C laboratories. It demonstrates, based on rigorous multichannel and multiparameter α- and β-particle spectral analyses of some 1000 benzene samples, that parent radium is not present and that its daughter radon is quantitatively eliminated during dynamic vacuum recovery of benzene at −78°C. However, the radon-free benzene can be recontaminated by exposure to air containing traces of radon, such as is common in concrete or low-lying laboratories. The use of radon-free air, when exposing the benzene to the atmosphere, and the monitoring of radon counts from the environment and sample benzene in a fixed ‘radon window', are essential prerequisites to the quality control of 14C age determinations in very low background systems.

Gas chromatographic analysis of the benzene obtained for radiocarbon dating revealed only very low levels (<200 ppm in total) of impurities, most of which were alkylbenzenes. Traces of acetone and phenol were also found. These impurities, at such low levels are thought unlikely to affect counting efficiency.

A fast 14C sample preparation technique for organic material in conventional 14C counting is described. The basic difference from conventional preparation is oxidation of the organic substances in a closed system under an oxygen pressure of ca 10 bars. After the explosion-like combustion, the reaction products SO2 and NO2 are oxidized immediately to SO42-and NO3− and precipitated on the wall of the reaction vessel. The residual gas mixture is passed first through a cold trap at −78° to remove water vapor and then through an activated charcoal column at 0° for purification. CO2 is removed from the remaining O2−CO2 gas mixture in a LN2 trap at a pressure of ≃ 100mbar. With this method sample preparation time is reduced from ca 10 hr to ca 1 hr. The efficiency and reproducibility of the procedure is shown with 14C and 13C measurements of a sugar-carbon standard. The results obtained by the new technique agree well with conventionally prepared wood and bone samples.

Preparation techniques used in dating freshwater peat were evaluated and compared using a homogeneous sample. This paper details various approaches to pretreatment of this material and presents the resultant 14C dates. General trends in the results are discussed.

Origins of carbon in potsherds are studied through field experiments which involve firing pots containing selected components and laboratory analyses. As temper seems to be the main source of carbon, radiocarbon dates may be too old if calcium carbonate is used. We can avoid the carbon from the clay by low-temperature burning of samples.

The standard procedure for measuring 14C at the Gif-sur-Yvette Tandetron AMS facility is described. A new sample manipulator and automated measurements are being used and have been operational for six months. Evidence of isotopic fractionation in the sputter ion source is provided. We take this into account by measuring the 13C/12C ratio of the sample in the accelerator.

Determination of the more common light elements such as nitrogen and chlorine at trace levels is difficult because of their high abundance on sample surfaces, in materials used to build analysis instruments, and in the residual gas of the instrument vacuum. We present here a new approach to analysis of these elements: accelerator mass spectrometry (AMS) combined with neutron activation. The problem of contamination is overcome by using neutron activation to produce long-lived radioisotopes which generally have low concentrations in the environment. For measurement of 14N and 35Cl, AMS can provide sensitive background-free measurements of their neutron activation products 14C and 36Cl and, in addition, can provide depth profiles. These are the first results of this new method: depth profiles of nitrogen and chlorine implanted in semiconductor grade silicon.

Since Middleton showed the potential for a sputter source working from CO2, we have been building a source which can automatically change samples. The source produces a maximum beam of 25μA of C−, with typical operation between 10 and 20μA. Although beam generation from the source is very reliable, the mechanics of sample changing have given considerable problems. The changing of samples also involves considerable care in gas handling, and a computer control system has been written which ensures the correct sequence of the 16 operations required for sample change.

We have made the first measurements without pre-enrichment of 41Ca in terrestrial rock and bone samples using accelerator mass spectrometry. Although the results in tufa deposits from Egypt are in good agreement with the saturation value of 8×10-15 predicted by Raisbeck and Yiou (1979), the average 41Ca:40Ca ratio of 2×10-15 (range: 0.6 to 4.2×10-15) that we measure in modern bone is an order of magnitude lower than that obtained previously by Henning, et al (1987) on a cow bone that was measured using AMS following isotope enrichment. The low value and the variability (more than a factor of seven) of the 41Ca:40Ca ratio in modern bone make the possibility of dating bones using 41Ca unlikely.

We made preliminary AMS measurements of 41Ca/Ca ratios in bone and limestone specimens with the Argonne Tandem-Linac Accelerator System (ATLAS). We were able to avoid pre-enrichment of 41Ca used in previous experiments due to a substantial increase in Ca-beam intensity. Most of the measured ratios lie in the 10-14 range, with a few values below 10-14. In general, these values are higher than the ones observed by the AMS group at the University of Pennsylvania. We discuss possible implications of these results. We also present the current status of half-life measurements of 41Ca and discuss 41Ca production processes on earth.

High-intensity ion sources are now capable of producing negative molecular ion beams of Ca of >2μA (Middleton et al, 1989). This and the advanced status of 41Ca AMS at several laboratories (Pennsylvania, Argonne National Laboratory, Rochester, Munich, Weizmann Institute) have revived interest in 41Ca (T1/2 = 1.0×105 yr (Mabuchi et al, 1974)) as a possible dating tool (Yamaguchi, 1963; Raisbeck & Yiou, 1979; Kubik et al, 1986).

Liquid scintillation (LS) 14C spectrometry, using ultra low-level, high-efficiency and resolution α and β particle detection equipment is seen as an excellent tool supporting future 14C radiometry. Modern technology ensures electronic excellence: high-precision, small-sample (100mg) dating ability, flexibility of use for the detection of α and β particle emitting radioisotopes at and below natural abundance levels. Multiple applications to environmental research, supported by elegant software and commercial availability, are its attributes. However, mastering the new technology takes time as lessons learned in gas-proportional 14C radiometry and high-background counting are not directly applicable to ultra low-level 14C LS spectrometry using benzene as the counting medium. That discordant results can be obtained by any technique based on physical measurement is well established. This is demonstrated for 14C age determination by reference to an international cross-check organized by the University of Glasgow. Reasons for aberrations are explored and the question is raised if it is not timely to introduce the type of quality assurance practiced by other analytical disciplines. Better results will not only serve the daters and users but will generally enhance the value of worldwide 14C research.

A Packard low-level liquid scintillation counting system is described which provides superior low-background beta-counting capability for 3H and 14C. The design is based on a novel pulse-discrimination circuit that separates background from valid scintillation pulses. Background discrimination is further enhanced by adding a plastic vial holder that acts as an anticoincidence guard. When excited by background radiation, the scintillation properties of the plastic provide an increased burst of photons to the detection electronics, which discriminate based on the number of component pulses in the burst. Experimental data demonstrate the low-level counting capabilities of this counter.

Important advances have been made in reducing the background counting rate of gas proportional counters for 14C dating through detailed and systematic study of the background components. Until recently, limited work has been reported on the study of the background of liquid scintillation counters (LSC). During the last few years, commercial systems with greatly reduced background have been introduced. It is shown that the best gas proportional counters and LSC have similar backgrounds for the same amount of sample material. Similar results with less effort may be expected with more detailed and fundamental knowledge of the components of the background of LSC. We report the results of a study of one photomultiplier LSC system where we research all parameters of importance: light collection efficiency, absorbed energy per photo-electron, pulse height spectrum and background counting rate.