The method of 14C counting used by Libby and his co-workers involved measuring radioactivity using modified Geiger counters. The next development in counting technology was the conversion of sample carbon to CO2 gas for measurement in Gas Proportional counters. In the early 1950's, the first attempts were made to detect 14C by the Liquid Scintillation (LS) counting method. In the 1940's, Broser and Kallman (1947) discovered that certain organic compounds (scintillators) fluoresced when exposed to ionising radiation. Each fluorescence event is proportional to a radioactive decay event, and the frequency of these events is directly proportional to the number of 14C atoms present in the sample. A brief summary of the LS method and some of the problems associated with the technique are given here. A more detailed report can be found in Polach (1987). Discussion on the LS method can be broken down into a number of topics : counting liquid, counting vials, and the LSC method, including counting instrumentation and potential problems associated with measurement of beta (ß)-decay by LS counting.
In the majority of LSC facilities, the scintillation solvent is benzene (C6H6) or a mixture of benzene and toluene (C6H6CH3). Benzene has been chosen because of its excellent light transmission properties and the high chemical conversion yield of sample C to benzene. The sample is first converted to CO2, then reacted with molten lithium to form lithium carbide (Li2C2), before being catalytically trimerised to benzene. Benzene synthesis procedures vary considerably in different laboratories. The details outlined below describe the procedures used in the Waikato laboratory to give the reader some idea of how the chemical conversions may be carried out.
The carbon is first oxidised to CO2, either by acid hydrolysis (for carbonates), or combustion in an oxygen stream or combustion bomb (for organic materials). The combustion gases are passed over heated CuO to complete the oxidation of CO, NO, N2O and NO2 and also of tar substances. The CO2 may then be then purified using a chain of wet chemical reagents; for example, AgNO3 and Hg(NO3)2 to precipitate halogens, and the oxidising agents KI/I2 and K2Cr2O7 to remove nitrogen and sulphur compounds. Silica gel and dry ice traps (-80oC) remove any water remaining in the gas (see figure 1).
This reaction was first described by Barker (1953), who developed the reaction using lithium instead of barium, and later improved by Polach and Stipp (1967). The CO2 is bled slowly onto the molten Li where it is converted to Li2C2. The carbide is heated to ca. 800oC (furnace temperature) and placed under active vacuum for 30 minutes to remove any unreacted gases and complete the carbide synthesis (Gupta and Polach, 1985). The lithium carbide is cooled and then hydrolysed to acetylene gas (Li2C2):
The acetylene is purified by passing through a phosphoric acid trap to remove ammonia compounds, and again, dry ice traps to remove water vapour. Finally the acetylene is trimerised to benzene using a suitable catalyst. There are a variety of vanadium or chromium activated catalysts available, including a silica-alumina vanadium activated catalyst developed by John Noakes (CAIS, Univ. of Georgia), a vanadium-alumina-silica catalyst produced at the Institute of Geography, Univ. of Petersburg by Dr Kh. A. Arslanov and another at the British Museum by Dr J. Ambers. The catalytic trimerisation for some catalysts may be more efficient at reduced temperatures (e.g. 5oC).
Benzene is then driven off the catalyst at ca. 100oC and collected under vacuum at ca. -65oC. The benzene is then stored in a vial under refrigeration to await counting.
This sequence of reactions requires a high degree of operator skill because of the complexity of the equipment and the nature of the reactions. It is important that a standardised routine is followed carefully and consistently, so that yields remain high and there is little cross-contamination between samples.
The synthesised benzene is transferred into counting vials of a variety of types. The counting vials contain the sample solvent and the scintillator. Commercially available LS vials used for benzene counting are commonly composed of either Teflon, quartz or low-K glass (e.g. Polach et al.1983; Haas 1979; Devine & Haas 1987; Noakes & Valenta 1989; Hogg & Noakes, 1992; Hogg 1993). Polyethylene and polypropylene vials are very useful for non-aromatic solvents, but their permeability makes them unsuitable for repeatable long-term benzene use. In a comparison of different vial types, Hogg et al . (1991) concluded that, for spectrometers using true anticoincidence detection and extensive passive shielding (as in the Perkin Elmer 1220 Quantulus), Teflon or synthetic silica vials produced the highest performance in terms of 14C detection efficiency and background. The authors also noted that synthetic silica had superior physical properties to Teflon. The silica vials vary in size according to the weight of the sample benzene being counted, including 0.3 ml and 1 ml minivials, 3 ml standard vials and 12 ml vials, suitable for high precision applications. Low-K glass vials are supplied in either 7 ml or 20 ml sizes.
The most commonly used scintillator in 14C dating is PPO + POPOP, either dissolved in toluene or directly in sample benzene. Polach et al (1983) made a comparative study of various scintillators and concluded that dry powder butyl-PBD dissolved in sample benzene (15g/l) gave superior and stable performance even under extreme quench conditions. Some Packard spectrometers utilise an active plastic holder (Pico adaptor) which helps to reduce background radiation. These instruments require a secondary wave shifter (usually bis-MSB, see Cook, Harkness & Anderson 1989) to be added to the butyl-PBD. In the Waikato laboratory, the scintillator (butyl-PBD) is dissolved directly into the sample benzene at a concentration of 15g/l. The vial is then transferred into a Quantulus spectrometer and allowed to cool and dark adapt for a minimum of 8 hours prior to the commencement of counting.
The essential electronic components of the Liquid Scintillation (LS) Spectrometer according to Gupta and Polach (1985), are photo multiplier tubes (PMTs), high voltage supply, signal preamplifiers, pulse and summing amplifiers, coincidence logic, timer and scaler (Gupta and Polach, 1985:50). The LS spectrometer measures electronically the pulses of light generated from photon emissions emitted by a scintillator in response to a radioactive decay event. The PMTs register an electronic pulse proportional to the energy of a particular beta decay when a photon or light particle is emitted within the benzene cocktail. The anode current generated from a PMT, then, is a function of the level of radioactivity (Horrocks, 1974).
Optical cross-talk occurs between opposed PMTs. An event occurring in one tube initiates a pulse in the other (Noakes, 1977). Ionising events may occur in either PMT due to interactions between the molecules of the tube material and surrounding radiation. Butterfield and Polach (1983) and Gupta and Polach (1985) have described methods used to reduce optical cross talk in LS counters, both by optimising and refining electronic circuitry to enable differentiation between sample and non-sample events and by masking areas in the counting chamber to reduce reflection. Carefully designed vials and vial holders can also help to reduce cross-talk, by minimising the view of opposing PMTs to each other.
Line transmitted switching noises and variation in the high voltage supply can influence reproducibility in LS counting. Power supplies are therefore frequently filtered and regulated, either with line noise suppression electronic transformers or a suitable uninterruptable power supply (UPS) unit. A clean and stable voltage supply is crucial because of the relationship between electronic particle detection and the energy of the decay event.
Radiofrequency interference is generated by lights, motors and switches in the vicinity of the counter. RF detection circuitry can be fitted which will identify RF signals and ensure they are removed from the sample 14C spectrum.
Static induced noise results from the build up of static electricity through friction generated from the movement of vials throughout the counting system. Earthing of vials and holders and an ionising unit which the vials move past, can be used to remove static electricity prior to counting.
222Rn emits alpha and beta particles, the latter with a decay energy of 5.587 meV. This decay interferes with the 14C spectrum, resulting in higher count rates and radiocarbon ages that are too young. Radon's short half-life of 3.82 days, however, means that a delay in counting eliminates the problem. Sample benzene is therefore often left for 3 - 4 weeks to allow any radon that may be present to decay. Potential 222Rn contamination is more significant in older samples because of their lower count rates.
Natural atmospheric radiation influences sample count rates. This is reduced through shielding of the counting environment. This may be passive or active shielding. The passive shield consists of lead which shields the counting chamber from external radiation. The term 'active shielding' describes electronic means of recognising and eliminating background causing events from the sample beta spectrum. Active shielding occurs in two forms in modern LS counters. The Perkin Elmer Quantulus, utilises an electronic anti-coincidence guard, comprising a liquid scintillator guard containing it's own phototubes, surrounding the counting chamber and sample PMTs. Ionising radiation from non-sample external sources passing through the guard scintillator leaves excited atoms and molecules in it which are seen by the guard PMTs. If an ionising event is recorded simultaneously in the guard and sample tubes, it is rejected as a background count. Some Packard counters utilise 'time resolved counting' to differentiate between, sample events with a low pulse index (see Roessler et al, 1991 for details), and background events with a higher pulse index.
Most higher performance counters use multi-channel analysers (MCAs) to store and evaluate counting data. The MCAs are software-controlled to provide complete flexibility in data acquisition.
LSC Dating by the Perkin Elmer "Quantulus"
One of the most widely known applications of LSC of 14C is in high precision dating of the radiocarbon timescale for the purposes of calibration. This was conducted primarily at the Belfast Radiocarbon Dating Laboratory (an LSC facility) and the Quaternary Dating Centre at the University of Washington (Gas Proportional Counting). The following LSC laboratories present information on the technique and its application:
For a review of LSC see: Polach, H.A. (1992) Four Decades of LS Counting and spectrometry. in, Radiocarbon after four Decades. An interdisciplinary Perspective. (eds) Taylor, R.E, Long, A and Kra, R. Springer-Verlag, NY.
Nevada Technical Associates, Inc. has some web pages devoted to radiation safety, radiochemistry and related areas. Equipment including Liquid Scintillation Counters is available here.