Ir544 Barium sulphate method for consecutive determination of radium-226 and radium-228 on the same source

2 Methodology

The BaSO₄ precipitation method utilises micro-filtration, rather than electro-deposition to deposit radium as a thin source for alpha counting. Poorer resolution as compared to electro-deposition techniques, and unquantifiable radon retention means that only ²²⁶Ra can be effectively resolved from the spectrum on these sources (Sill 1987, Medley et al 2005).

A method has been developed for counting the sources prepared for ²²⁶Ra determination by the BaSO₄ precipitation method, with a High Purity Germanium (HPGe) detector for ²²⁸Ra determination. The gamma decay lines of the fast ingrowing ²²⁸Ac daughter of ²²⁸Ra are utilised for this. The low efficiency, high background counts and low peak decay probabilities of the high energy lines of ²²⁸Ac, gives relatively high detection limits for this method. However, the detection limits are lower than the standard gamma method using a pressed geometry due to the thin, flat geometry of the source which improves the efficiency for counting.

To achieve very low-level detection limits for ²²⁸Ra determination a complementary method for digestion of the filter paper after a sufficient time for ingrowth of the ²²⁸Th daughter and measurement of ²²⁸Th via alpha spectrometry has also been developed.

2.1 Gamma spectrometry method development

This section describes the preparation of samples and calibration of the spectrometry systems for determination of ²²⁸Ra by the BaSO₄ method, utilising high resolution gamma spectrometry. Instrument Detection Limits (IDLs), Method Detection Limits (MDLs), Efficiency, Energy, Recovery and Region of Interest (ROI) calibrations are included.

2.1.1 Source preparation

Source discs for alpha and gamma spectrometry are prepared by methods described in Medley et al (2005). The system set-up and configuration for the alpha and gamma spectrometry systems are described in Martin & Hancock (2004b) and Marten (1992) respectively. Prepared sources are initially counted via alpha spectrometry for ²²⁶Ra determination, then for ²²⁸Ra via gamma spectrometry.

To accurately analyse prepared sources from the BaSO₄ method via gamma spectrometry, the sources must be mounted in a standard geometry. To enable this, sample holders were custom made from PVDF. This material was deemed the most suitable material for the following reasons:

• 
  Easy to work to necessary dimensions with a high degree of accuracy
  
• 
  High chemical resistance, including to H₂SO₄ and caustic agents – this helped in preparation of standard sources (see method for sealed standards)
  
• 
  Resistance to temperatures up to at least 80C
  

Necessary for preparation of sealed standards

Essential for cleaning of holders which uses an alkaline (pH 10.6) 0.2 M DTPA wash at 80C for removing Ba/RaSO₄

• 
  Relatively low cost; PTFE has a higher chemical resistance and melting point than PVDF but was ruled out due to significantly higher cost
  

The inner diameter of the PVDF holders is 24 mm to ensure the source discs are not bent while in the holder. The outer diameter is not a factor that will influence the gamma counting efficiency. The thickness of the PVDF holder must be consistent for standards and source discs. Source discs are placed in the base of the holder then the top is placed on to hold the disc centrally and pressed flat to maintain the same geometry.

Figure 1 Illustration of mounting radium source discs in PVDF holders for gamma spectrometry. Polypropylene filters have a 17.5 mm active diameter, and a total diameter of 22 mm.

²²⁸Ra is a beta emitter, and cannot be directly measured via alpha or gamma spectrometry. ²²⁸Ac, the direct daughter of ²²⁸Ra, is very short lived with a half-life of only 6.15 hours (Martin & Hancock 2004b). After a short ingrowth period ²²⁸Ac will have sufficiently ingrown to have reached secular equilibrium with ²²⁸Ra (Figure 2).

This is calculated using the radioactive decay and ingrowth equation (Equation 1) adapted from Friedlander (1981).

Ingrowth of ²²⁸Ac from the parent isotope ²²⁸Ra. Equation 1

Where –

A_Ac – Activity of ²²⁸Ac

A_Ra – Activity of ²²⁸Ra

λ_Ra – Decay constant of ²²⁸Ra

λ_Ac – Decay constant of ²²⁸Ac

As λ_Ra << λ_Ac equation 1 can be written as:

Figure 2 Time taken for ²²⁸Ac to reach equilibrium with ²²⁶Ra parent after radium separation.

Calculated using Equation 2

2.1.2 Gamma counting

The 911 keV and 969 keV gamma lines of ²²⁸Ac are used as they are sufficiently separated from interfering lines from the ¹³³Ba tracer, ²²⁶Ra and daughters, and have a high decay probability (Table 3). Figure 3 shows a typical spectrum with major lines of ¹³³Ba, ²²⁶Ra and daughters and shows that there are no lines interfering in the region of interest of the 911 keV and 969 keV peaks.

Figure 3 Spectrum of sample with ¹³³Ba and ²²⁶Ra, counted on an HPGe spectrometer. Major peaks of these 2 isotopes can be seen to be distinctly separate from ²²⁸Ac decay lines at 911 and 969 keV. Spectrum produced at eriss.

Table 3 Gamma decay energies and emission probabilities for relevant nuclides and decay lines in samples and standards prepared for ²²⁶Ra and ²²⁸Ra analysis via the BaSO₄ method. All gamma energies are taken from Canet & Jacquemin (1990), except for ²²⁸Ac which are taken from Marten (1992).

Isotope	Main gamma emissions (keV)	Probability (%)	Isotope	Main gamma emissions (keV)	Probability (%)
^232Th	59	1.9	^226Ra	186	3.28
^228Ac	338 911 969 & 964 (doublet peak)	12.3 26.6 20.4	^214Pb	295 352	18.7 35.8
			^228Th	84	1.2
			^224Ra	241	3.9
^214Bi*	609 934 964 1120 1238	46.1 3.16 0.38 15.0 5.9	^133Ba**	53 81 223 303 356	2.2 33.8 0.5 18.4 62.1

2.1.3 Detection system calibration

Calibration of the gamma spectroscopy system is necessary before any results can be calculated. The parameters that need to be determined are:

• 
  Channel/energy calibration
  
• 
  Detector efficiency
  
• 
  Background
  
• 
  Detection limits.
  

2.1.3.1 Energy calibration

Energy calibrations are performed with the sealed standards used for efficiency determination and the ¹³³Ba standard discs. Quality control checks with a standard of known activity, with a broad energy range of gamma emitting nuclides, are counted after each sample to ensure energy calibration is maintained.

2.1.3.2 Efficiency

The efficiency of an HPGe gamma spectrometer varies with source geometry, sample matrix and the energy of the gamma decay line, thus a gamma spectrometry system requires accurate efficiency determination for the geometries used and for each peak of interest to be calculated. To do so a source of known activity with the same geometry and sample matrix (as close as possible) is used. Typical efficiency curves for a series of spectrometers calibrated with a source containing isotopes with a range of gamma photon energies are shown in Figure 4.

Figure 4 Typical efficiency curves for a series (A, D, F, G & K) of gamma spectrometer configurations. Reproduced from Marten (1992).

For this method the chemical recoveries of the sources are determined by comparison with a standard of known activity so determination of the efficiencies of the decay lines for the ¹³³Ba peaks is not necessary. However, the system needs calibration for the ²²⁸Ac.

The preparation of standards is often the most critical step in the use of radiation detection systems. For gamma spectrometric measurement of ²²⁸Ra via the gamma decay of ²²⁸Ac, a pure source of ²²⁸Ra is ideally used to calibrate the system. Due to the short half of ²²⁸Ra, it is not easily available and quite expensive to obtain as a pure source. Pure thorium salts however are easy to obtain and relatively inexpensive. An isotopically pure ²³²Th (t½ = 1.41 x 10¹⁰ years, Joshi et al 1983) salt will reach secular equilibrium with the daughter ²²⁸Ra (t½ = 6.7 years, Joshi et al 1983) after approximately 5 half lives, and thus can be used as an essentially pure ²²⁸Ra standard. It is important that ²³⁰Th contamination of the thorium salt is negligible, as the daughter ²²⁶Ra is unwanted in calibration standards. A review of the main gamma emissions from ²³²Th and daughters other than ²²⁸Ra showed no significant interferences with the high energy ²²⁸Ac gamma emissions, though the lower energy emissions could not be used (338 keV, Table 3).

2.1.3.3 Background

There are essentially 2 types of background that need to be considered in gamma spectrometry systems. The background introduced by the source falls in two main categories, either Compton scattering or Bremsstrahlung. Compton scattering is caused by gamma photons imparting energy to weakly bound electrons which in turn release photons. Bremsstrahlung creates a broad spectrum and is caused by beta particles (electrons) interacting with the semiconductor detector and the absorbing material of the sample and holder (Canet & Jacquemin 1990). The type and magnitude of background interference will vary with each source (Marten 1992).

There may also be systematic background as a result of contamination of the counter or counting well and natural system (ie – electronic interference in old counting systems, cosmic and other natural radiation). Lead and other types of shielding are routinely used to minimise natural background. Interferences in the production of a final spectrum will vary with each detector system (Marten 1992).

2.1.3.4 Detection limits

Detection limits represent the point at which a detection system can no longer be considered to have detected a peak.

There are several types of detection limits described for radiation detection systems – instrument detection limits, IDLs, method detection limits, MDLs, and overall detection limits, DDLs. There are also a number of ways to determine the detection limits.

Detection limits are essentially the measure of the relative effectiveness of:

• 
  IDL – the detector system configuration only
  
• 
  MDL – IDL and the method components – any processing that affects each sample
  
• 
  DDL – MDL and the preparation and counting conditions that are unique for each prepared source.
  

Instrument detection limit

The IDL for ²²⁸Ra for this method can be determined by measuring:

• 
  the efficiency of the detector for the ²²⁸Ac peaks to be used
  
• 
  the background spectrum as a result of the detector and sample holder configuration only.
  

Detector efficiencies for the method were determined from a series of sealed standards prepared at varying activities from an isotopically pure, calibrated thorium nitrate (Th(NO₃)₂) solution in secular equilibrium with ²²⁸Ra. The solution was pipetted onto a small disc of absorbent paper (17.5 mm diameter – the same size as polypropylene filter papers) in the PVDF sample holders for gamma measurement. A small volume of 60% ethanol was added. Th(NO₃)₂ is highly hygroscopic so it is essential the standards are dried when weighed to ensure accurate calibration (this was performed on a hotplate at 70°C), once dried, it is also essential that the standards are sealed. The sealed standards were counted under identical conditions to those as for samples, and the efficiency of each ²²⁸Ac line was determined (assuming 100% recovery). The IDL can be calculated after measuring a number of instrument background spectra, using methods derived by Currie (1968).

Method detection limit

The method detection limit, uses the same factors as for the IDL, but also takes into account the processes samples undergo for sample preparation. For this method these are:

• 
  chemical separation and associated typical recovery of the analyte of interest
  
• 
  additional background counts resulting from trace contaminants in chemical solutions and laboratory equipment
  
• 
  the accuracy of other preparation steps – eg weighing of the sample (only has a small influence on uncertainty in radiation measurement techniques, De Regge & Fajgelj 1999)
  
• 
  Appropriate sample size and type may also be influencing factors – for example there are limits on the amount of organic material that can be tested before the chemical separation and deposition of the analyte is adversely affected.
  

Unsealed standards were prepared to determine the method detection limit and to assess whether it is significantly affected by the ¹³³Ba activity which is used as a tracer to determine the chemical recovery of radium from the procedure. These standards were made from the prepared and calibrated Th(NO₃)₂ tracer solutions in secular equilibrium, and an isotopically pure ¹³³Ba tracer solution.

Unsealed standards were also prepared with varying amounts of the ¹³³Ba tracer solution only in order to assess the correlation of ¹³³Ba activity to background counts in the ²²⁸Ac peaks.

After counting, these standards were given a period of time (approximately 6 ½ months) to allow ingrowth of the ²²⁸Th daughter from ²²⁸Ra, so they could be used for calibration of the ingrowth part of this method.

Overall detection limit

Overall detection limits, are affected by all of the above considerations, but vary with each sample, this is due to several factors:

• 
  Counting time – the most significant source of uncertainty in radiological measurements are counting statistics (De Regge & Fajgelj 1999), with longer count times uncertainty is reduced, and therefore the detection limits are reduced
  
• 
  Individual sample composition
  

Some samples contain much higher levels of trace contaminants which may affect the chemistry (such as barium) than others

Samples with higher levels of ²²⁶Ra can affect the overall Compton background due to the 934 keV gamma line of ²¹⁴Bi (Table 3), and in extreme cases (very high ²²⁶Ra: ²²⁸Ra activity ratios, eg >1000) another very low probability gamma emission of ²¹⁴Bi at 964 keV can also affect the spectrum due to its proximity to the ²²⁸Ac peak at 969 keV

• 
  Chemical recovery of the sample.
  

2.1.4 Chemical recovery determination

A barium standard disc is prepared, and the chemical recovery of the disc determined using a NaI detector according to methods described by Medley et al (2005).

This disc is then counted in an HPGe detector and the count rate compared with that of the sample is used for chemical recovery determination. Actual tracer activities are not needed as the relative recovery of each sample is determined by comparison with the count rates obtained for the standard disc, using the same amount of tracer.

Thus the chemical recovery is determined according to the following basic principle:

Principle of chemical recovery determination. Equation 2

Where –

R_sample – Sample recovery

N_sample – Net count rate for the sample

N_standard – Net count rate for the standard disc

The 302 keV and 356 keV gamma decay lines of ¹³³Ba are used, other lines with a high decay probability are interfered with by lines from daughters of ²²⁶Ra, and so cannot be used for accurate recovery determination. To account for the varying peak decay probabilities of the two gamma decay lines, and the varying count times and tracer masses of the sample and standard, the following set of Equations 3–5 are used for chemical recovery determination:

Chemical recovery determination. Equation 3

Where –

R₃₀₂ – Sample Recovery calculated using the 302 keV gamma line

R₃₅₆ – Recovery calculated using the 356 keV gamma line

Recovery at each ¹³³Ba decay line is calculated using Equation 4.

Chemical recovery determination for each peak. Equation 4

Where –

R_Peak – The sample recovery at a given peak

C_Peak – The gross counts in the peak

B_L – The background counts in a region to the left of the gamma line spanning half the width of the main region of interest

B_R – The background counts in a region to the right of the gamma line spanning half the width of the main region of interest

t – The count time in kiloseconds (ks)

PDP – The Peak Decay Probability of the peak. For the 302 keV peak this is 0.1833, and the 356 keV peak is 0.6205

M_sample – The mass of the tracer used in the sample

CF_Std – The ¹³³Ba standard correction factor

Equation 5 is used to calculate the ¹³³Ba standard correction factor.

¹³³Ba standard correction factor determination. Equation 5

Where –

CF_Std – Gross counts of a given peak of the standard disc

R_Std – The recovery of the standard disc, this is calculated from methods described in Medley et al (2005)

M_Std – The mass of the tracer used in the sample.

All other parameters are the same as described above, though for the standard disc.

2.1.5 ²²⁸Ra activity determination

Based on the above calibrations of the spectroscopy system ²²⁸Ra activity is determined using Equation 6.

Back calculation of ²²⁸Ra activity from ²²⁸Ac activity measured. Equation 6

Where –

Net count rate determination. Equation 7

and –

N_c – Combined net counts of the 911 and 969 keV regions of interest

C₉₁₁ – Counts in the 911 keV region of interest

C₉₆₉ – Counts in the 969 keV region of interest

BL₉₁₁, BL₉₆₉, BR₉₁₁, BR₉₆₉ – The background counts in a region to the left and right of the 911 keV and 969 keV gamma lines spanning half the width of the main region of interest

PDP_c – The Combined Peak Decay Probability of the 911 and 969 keV which is 0.4705

ε – The calculated combined efficiency of the 911 and 969 keV peaks which is 2.70%.

2.1.6 Quality control and quality assurance

Quality assurance and quality control (QA/QC) is an important part of implementing a developed method to ensure calibrated parameters remain within acceptable limits.

Counting of a high activity energy calibration source for 10 minutes after each sample count provides a means of ensuring energy calibration of each spectrum can be adjusted for minor variations after counting.

Regular background checks are conducted for a standard pressed geometry, though this does provide relevant information regarding the background spectrum of the detector system, these shall be implemented for unused and chemical blank polypropylene discs at bi-monthly intervals. It will take some time to fill control charts with sufficient data points to provide meaningful QC data (>20 data points, Currie 1968).

Due to the discs being used for determination of ²²⁶Ra by alpha spectrometry as well as ²²⁸Ra by gamma spectrometry, all samples are counted with a NaI detector for chemical recovery determination prior to alpha counting. Alpha counting should always precede gamma counting to prevent losses on the PVDF containers used for gamma counting from affecting final results.

Comparison of the chemical recovery obtained through NaI and HPGe spectrometers provides an additional QA process. The calibration of high resolution gamma spectrometers allows for accurate determination of chemical recovery in samples for ²²⁶Ra analysis where ²²⁶Ra activity levels are very high. The gamma emissions of radon daughter ²¹⁴Pb, and increased Compton background from higher energy gamma emissions of ²¹⁰Bi and others (Table 3) can increase the number of counts in the broad region of interest used for chemical recovery determination through the NaI spectrometer, increasing the measured recovery.

2.2 Alpha spectrometry method development

This section describes the preparation of samples and calibration of the spectrometry systems for determination of ²²⁸Ra by the BaSO₄ method and subsequent separation and measurement of the ²²⁸Th daughter utilising high resolution alpha spectrometry. Instrument Detection Limits (IDLs), Method Detection Limits (MDLs) and Recovery calibrations are included. Efficiency, Energy and Region of Interest (ROI) calibrations are detailed in (Martin & Hancock 2004b).

2.2.1 Methodology

Prepared sources are screened for ²²⁸Ra activity via HPGe gamma spectrometry using the ²²⁸Ac daughter. If the activity is below the detection limit for this method, the samples are stored until the ²²⁸Th daughter has ingrown sufficiently to be measured via alpha spectrometry (Figure 5). The maximum activity of ²²⁸Th is reached after approximately 1650 days, or just over 4 years. Suitable activities for analysis of many environmental can be reached after 12 months ingrowth.

Figure 5 Ingrowth of ²²⁸Th from the parent isotope ²²⁸Ra, based on Equation 1. Note that this ingrowth curve is valid regardless of the initial activity of ²²⁸Ra.

²²⁸Th activity was determined by methods described in Martin & Hancock (2004b), and ²²⁸Ra activity was back calculated using Equation 8 given below.

Back calculation of ²²⁸Ra activity from measured ²²⁸Th activity. Equation 8

Where –

A_Ra - The activity of a ²²⁸Ra at time

A_Th - The initial activity of ²²⁸Th

λ_Ra – The decay constant of ²²⁸Ra

λ_Ra – The decay constant of ²²⁸Th

t - Time since separation of ²²⁸Ra from the sample

2.2.2 Source preparation for ²²⁸Th measurement

A standard method for separation of thorium isotopes and analysis via alpha spectrometry used is described in Martin & Hancock (2004b), a flow diagram summary is given in Figure 6. Based on a number of assumptions, this method has been modified for the current application.

Figure 6 Flow diagram of standard method for thorium isotope analysis (Martin & Hancock 2004b). Steps highlighted have been omitted for this project.

The source disc filter paper used in radium determination is 100% polypropylene. To break this down into a suitable chemical form and release for thorium analysis, a digestion process is required (Figure 7). Concentrated sulphuric acid (98% H₂SO₄), is used first to char the filter, and then concentrated nitric acid/hydrogen peroxide (69% HNO₃/35% H₂O₂) treatment is used to remove excess carbon from the sample matrix. Once the digestion is complete, the sample can be taken up in appropriate acid medium, and undergo tributyl phosphate (TBP) extraction for thorium, eliminating the iron hydroxide precipitation step.

Figure 7 Flow diagram of digestion procedure for ²²⁸Th ingrowth method

Due to the very similar chemical properties of thorium and uranium, an anion exchange step is usually employed to ensure complete separation of these two elements (Martin & Hancock 2004b). For the current application it can be assumed that uranium has already been effectively removed from the sample in the initial radium separation, and thus the anion exchange step can be eliminated. To prevent from allowing organic residue to carry through from the TBP extraction step (usually removed in the anion exchange step), after evaporation of the extract, concentrated HNO₃ is added, and evaporated at moderate heat (approximately 80°C).

The solution is then passed through the electrodeposition process, whereby thorium is electroplated onto a stainless steel planchet, and counted via alpha spectrometry according to methods described Martin & Hancock (2004b).There are several assumptions to this approach for ²²⁸Ra determination, and a number of calibration and performance assessment procedures were undertaken to ensure the validity of this method.

Th(SO₄)₂ is very soluble in water and acid media (Hyde 1960), and there is an assumption that there is a 100% recovery of ²²⁸Th from the digestion procedure. This was confirmed by the analysis of unsealed standards prepared initially for calibration of both the gamma and alpha spectrometric techniques.

2.2.3 Alpha counting

Alpha counting for ²²⁸Th calibration of the detection system, and QA/QC for the alpha spectrometry determinations is performed according to procedures detailed in Martin & Hancock (2004b) and others detailed below:

• 
  Bi-monthly blanks are counted to monitor background levels
  
• 
  Efficiency calibration of the spectrometry system is performed by calibration with a known activity alpha-emitting standard
  
• 
  Peak tailing calibrations are performed with prepared standards of isotopically pure ²²⁸Th, ²²⁹Th and ²³⁰Th
  
• 
  Energy calibrations are performed with a prepared standard of ²³²Th in secular equilibrium with its decay series
  
• 
  Chemical blanks are run with each batch of samples to monitor procedural contamination
  
• 
  All equipment used for chemical separation and handling is washed according to procedures described in Appendix 1, section A1.4.1 of IR501 (Medley et al 2005).
  

2.2.4 ²²⁸Ra activity determination

Back calculation of ²²⁸Ra activity from ²²⁸Th determination is performed using Equation 8.