Lectura y comprensión

The3H activity present in a sample can be measured using several methods, depending on the type of samples and the tritium activities present (Table 2.3). Tritium activities are expressed in tritium units (TUs), Bequerels (Bq) (the SI unit of radioactivity) or Curies (Ci)(a pre-SI unit of radioactivity), where 1 TU = 1 3H atom per 1018 H atoms (Okada and Momoshima, 1993), which is equivalent to 0.118 Bq HTO per litre of H2O, and 1 Ci = 3.7 x 1010 Bq (NCRP, 1979). Liquid scintillation counting is a common method for the quantitative analysis of radioactivity, includingβ particles (Table 2.3). It has appropriate limits of detection for the samples in the present study, and it is rapid and cost-effective. The 3He ingrowth method can also be used to determine3H activities in extremely low level samples (<0.1 TU; Clarkeet al, 1976); however, the lengthy ingrowth period of∼6 months is a major disadvantage to this method.

2 .2 . T ri tiu m a n a ly s is 4 9

Table 2.3: Methods of tritium measurement

Method Type of

sample

Limit of detec- tion

Brief details of method Advantages Disadvantages References

Liquid scintilla- tion counting

Colourless liq- uids

∼_{0.005 Bq/g Sample and solvent are mixed with}

scintillators, which release energy as light in response to an α or β decay event. This light is measured by photomultiplier tubes.

Low background (∼₁

cpm). Relatively effi- cient detection of low energy β-emitters such as tritium. Interference by other radionu- clides if present. Chemilumines- cence. Quenching (self-absorption) can prevent tritium detection. Kallman, 1950; Reynolds et al, 1950; Dyer, 1974; L’Annunziata, 2003a 3

He ingrowth Water sam-

ples

0.1 TU (for water samples (∼_40g)

sealed for 6 months to 1 year)

Degassed water sample is sealed in a (low He permeability) glass bulb for a length of time. The3

He/4 He ratio of the water is measured us- ing a noble gas mass spectrometer. This allows the amount of3

He at- tributed to tritium decay to be cal- culated.

Very low detection limit. Simple analyt- ical procedure.

Length of time for ingrowth of 3

He (∼_{6 months).}

Liquid Scintillation counting

Theory and development The development of liquid scintillation counting is at- tributed to Kallman (1950) and Reynolds et al (1950), who reported that homogeneous solutions of samples with organic scintillators or ‘fluors’ emit light photons, which can be measured, on interaction withα and β-particles (Figure 2.9; Dyer, 1974). Improvements to the technique have focused on:

• Increasing the efficiency and miscibility with water of scintillation cocktails, and reducing their impact on health and the environment (Dyer, 1974).

• Reducing noise in the electronics by using two photomultiplier (PM) tubes and the coincidence counting technique, which reduces background counts significantly by eliminating thermal and electrical noise (L’Annunziata, 2003a).

Radioactive decay

in sample Solvent Scintillator

Photo-multiplier Tube Multi-channel Analyser β- CH3 e.g. toluene O N

e.g. PPO (2,5-diphenyloxazole) Chemical quench Physical/colour quench light electronics n p+ _β- + + ν

Figure 2.9: The basic mechanism of liquid scintillation counting (after Dyer, 1974).

The transfer of energy from a β particle (electron or positron) to the solvent, the fluor and finally detection by the photomultiplier tubes can be substantially reduced by quenching. Quenching most strongly affects weakβ-emitters such as3H and14C because it can result in photons from a particular decay event not being detected by the PM tubes (L’Annunziata, 2003a). There are three main types of quenching; physical, chemical and colour quench, which occur at different stages in this energy transfer (Figure 2.9). Physical quenching occurs when energy is not transferred from the radioactive particle to the scintillator solution, or when photons of light are physically absorbed within the vial. For example, a heterogeneous counting mixture may prevent the maximum interaction between the β particles and the solvent/scintillant mixture. Chemical quench occurs when non-fluorescent molecules absorb the energy of the solvent molecules instead of the scintillant molecules, and the energy is lost as heat instead of light. Any compound without an aromatic structure will produce quench, but the strongest quenching agents are halogenated compounds; salts, bases, acids, alcohols and water have less quenching effect and are described as diluters (L’Annunziata, 2003a). Chemical quench cannot generally be avoided, so it must be accurately corrected for. There are a wide range of methods for doing this but the most accurate methods use either an internal or external standard of known activity to allow the efficiency of the energy transfer process to be calculated for

2.2. Tritium analysis 51 each individual sample:

Ef f iciency = amount out(CP M)/amount in (DP M)

, where CPM = counts per minute and DPM = disintegrations per minute.

The Quantulus 1220 Liquid Scintillation Counter, used in the present study, applies the Spectral Quench Parameter of the External Standard (SQP(E)) method to correct for quenching effects. Aγ-ray source (37 000 Bq of152Eu), placed next to the vial, irradiates the sample, producing Compton electrons with a constant spectrum of energies. This allows the effect of quench (measured as the position of 99.5 % of the endpoint of the external standard spectrum) to be quantified (L’Annunziata, 2003a).

Colour quenching, when coloured compounds absorb light, occurs after the fluores- cence stage, reducing the number of photons leaving the vial. Colour can be produced either by reactions between the scintillator solution and the sample, or by the introduction of a coloured sample (Dyer, 1974). The SQP(E) method is not as accurate at correct- ing for colour quench as for chemical quench (L’Annunziata, 2003a), so the analysis of coloured samples has been avoided in the present study.

Measurement and calibration To measure the tritium activity of an aqueous sample, an aliquot (normally 8 ml) was pipetted into a 22 ml polythene scintillation vial, to which 12 ml of Goldstar scintillation cocktail was added. The vial was labelled, shaken to homogenise the phases, and dark-adapted overnight. It was then counted for 2 hours (10 x 12 minutes) on the Wallac 1220 ultra-low-level liquid scintillation counter. Only 5 ml of supernatant from the water-extractable tritium procedure was usually analysed. The total tritium activity (Bq/g) was then calculated using the equation:

A= CP M 60 · 100 E · Vt Vm ·1000 M · 100 Rs

, where CPM = counts per minute (output of LSC); E = efficiency; Vt = Total volume of bubbler; Vm = Measured volume; M = Sample mass; and Rs = Standard recovery.

MilliQ pure water blanks were also used to determine the instrument background. This was subtracted from the measured cpm to produce the actual count used in the calculation (CPM). The efficiency was calculated using an empirically-defined calibration curve, which corrected for the amount of quench in the samples. The standard calibration curve used was for 4 to 10 ml of MilliQ pure water spiked with a certified tritiated water source (supplied by Amersham plc, Cardiff), mixed with the appropriate volume of Goldstar Scintillation cocktail to make up 20 ml in a 22 ml polythene scintillation vial and counted for 1 hour (Figure 2.10). The limit of detection for this technique was 0.02 Bq/g of fresh sediment, based on 1.3 cpm background and 120 minutes count time.

Interferences The combustion furnace method was used because most matrix elements and radionuclides remain in the sample residue after combustion. However, 14_{C is a} potential interference because it can also dissolve in water and the tail of the14C energy

SQPE 700 720 740 760 780 800 Counting ef fi c ie nc y ( % ) 0 5 10 15 20 25 30 35 y = -0.000572193x2 _{+ 1.027845868x - 423.224826113} r2 _{= 0.969475311}

Figure 2.10: The standard calibration curve for 3

H measurement on the Quantulus 1220 liquid scintillation counter (instrument Q1), calculated by Dr J. Oh on 23/06/04. Aqueous samples, ranging in volume from 3 to 10 ml, were spiked with a tritiated water standard (from Amersham plc, Cardiff), then made up to 20 ml in a 22 ml polythene scintillation vial, dark adapted and counted for 1 hour.

spectrum overlaps with the 3H energy spectrum (Figure 2.11). To counteract this, a solution of 0.1 M HNO3 was used instead of water because the acid reduced the solubility of CO2.

Chemiluminescence, or the emission of light from chemical reactions, is a major in- terference for tritium because it produces peaks with similar energies to the 3H energy spectrum (Figure 2.11). It is often produced by exposure of the scintillation cocktail to UV light, which is short-lived, or by reaction with basic samples (pH 8 to 14), which is usually more persistent. To prevent chemiluminescence, samples can either be acidified or stored in a cool dark place overnight (L’Annunziata, 2003a); in the present study samples were dark adapted overnight before counting.

C o u n ts x 1 0 3

Relative pulse height

2 4 6 8 52 54 chemiluminescence chemiluminescence 3 H scintillation 3 H scintillation 14_{C scintillation}

Figure 2.11: Schematic diagram showing the potential interferences with the measurement

of tritium activities by liquid scintillation counting, including chemiluminescence and the presence of14

2.3. X-Ray fluorescence (XRF) analysis 53