ARTICLE IN PRESS

Applied Radiation and Isotopes 68 (2010) 1507–1509

Contents lists available at ScienceDirect

Applied Radiation and Isotopes

0969-80

doi:10.1

� Tel.

E-m1 Th

estimat

uncerta

(1995).

journal homepage: www.elsevier.com/locate/apradiso

An automated ionization chamber for secondary radioactivity standards

R. Fitzgerald �

Physics Laboratory, National Institute of Standards and Technology, 100 Bureau Drive, Gaithersburg, MD 20899, USA

a r t i c l e i n f o

Keywords:

Ionization chamber

Radioactivity

43/$ - see front matter Published by Elsevier

016/j.apradiso.2009.12.008

: +1 301 975 5597; fax: +1 301 926 7416.

ail address: ryan.fitzgerald@nist.gov

e combined standard uncertainty of a measu

ed standard deviation of that result. For m

inty analysis used in this paper, see Taylor and

a b s t r a c t

I report on the operation and characterization of a new ionization chamber system, ‘‘AUTOIC’’, featuring

a commercial digital electrometer and a commercial robotic sample changer. The relative accuracy of

the electrometer was improved significantly beyond the manufacturer’s specifications through an in-

house calibration of the various ranges, applied via software. The measurement precision and

repeatability of the system have been determined by measuring multiple samples of the same

radionuclide over the span of two or three years. The linearity of the system was examined by following

the decay of 99mTc, 99Mo and 133Xe sources for up to 19 half-lives and determining half-life values. All of

these values agree with the accepted literature values, within their combined uncertainties.

Published by Elsevier Ltd.

1. Introduction

Calibration factors for our reentrant ionization chamber ‘‘A’’are maintained as secondary standards for g-ray emitting radio-nuclei at NIST (Calhoun, 1986). These calibration factors areusually obtained from primary standardizations. We report on theoperation and characterization of a second ionization chamber,‘‘AUTOIC’’. This new system makes use of a commercial digitalelectrometer to measure the ionization current and a custom-designed commercial robotic sample changer. This new instru-ment, by remotely and automatically handling the radioactivesources, is safer for the user by minimizing radiation exposureand can increase the frequency in which samples can bemeasured. This second advantage can be important for half-lifedeterminations using a single sample as well as quality assurancemeasurements for large batches of sources.

For background information on reentrant ionization chambersand the uncertainties involved in their use as secondarystandards, see the report by Rytz (1978) and the recent reviewby Schrader (2007). For information about half-life measurementsusing ionization chambers and about half-life measurementuncertainties in general see the papers by Schrader (2004) andPomme (2007), respectively.

All uncertainties reported here on experimental results arecombined standard uncertainties, unless noted otherwise.1

Ltd.

rement result represents the

ore information about the

Kuyatt (1994) and ISO et al.

2. Instruments and methods

The AUTOIC consists of a Centronic2 IG11 reentrant ionizationchamber shielded by lead on all sides. The detector bias is negative1.10 kV. The sources are typically NBS-style flame-sealed ampoulescontaining 5 mL of solution (Calhoun, 1986; Rytz, 1978). Integral tothe system are a set of radium reference sources, each consisting of asealed 226Ra needle encapsulated in an acrylic right-circular cylinderhaving a diameter almost identical to that of the glass ampoules.Calibration factors for a particular combination of radionuclide andsource-type (container, volume, density, etc.) are determined relativeto a radium reference source, producing similar ionization current tothe test source. Each source to be measured is placed inside anacrylonitrile butadiene styrene (ABS plastic) holder that is closed witha screw top. The sample holders were made from a single batch ofplastic and machined by programmable lathe as a batch. The wallthickness of each sample holder is (2.4070.05) mm. The variability inthe ionization current due to variability in both holder and ampouledimensions was tested by measuring five ampoules of an 125Isolution, each in its own holder. Most of the photons emitted by 125Iare below 30 keV, so the ionization chamber response is highlysensitive to attenuation from the sample holder. The standarddeviation of the distribution of measured ionization currents was(0.1770.13) percent, where the uncertainty is the average standarddeviation of the mean for five repeated measurements of the sameampoule. As many as 100 such sample holders can be placed in asample tray, which is surrounded by lead shielding. The samples areloaded into the chamber by a custom-designed automatic sample

2 Certain commercial equipment, instruments, and materials are identified in

this paper to foster understanding. Such identification does not imply recom-

mendation or endorsement by NIST, nor does it imply that the materials or

equipment are the best available for the purpose.

ARTICLE IN PRESS

Table 1Corrections to manufacturer’s range settings on electrometer.

Range m b (pA)

20 pA 0.9995 0

200 pA 1.0000 0

2 nA 1.0010 0.140

20 nA 1.0010 0.140

200 nA 1.0010 0.140

2 mA 1.0007 126.0

The slope, m, and intercept, b, of the correction are listed. See text for equation.

Table 2Measured half-life values, T, and number of half-lives followed, nT, for each

radionuclide, where Tl is the literature value recommended by Be et al. (2007).

133Xe 99Mo 99mTc

T (5.246870.0011) d (2.748870.0012) d (6.007670.0014) h

nT 19 10 6

Tl (5.247470.0006) d (2.747970.0006) d (6.006770.0010) h

Fig. 1. Plot showing percent residuals for 99Mo decay over 10 half-lives. Vertical

lines show electrometer range changes. The uncertainty intervals represent the

combination of statistical precision and positioning variability, as described in the

text.

R. Fitzgerald / Applied Radiation and Isotopes 68 (2010) 1507–15091508

changer, (Changer Labs, Knoxville, TN, USA). When loaded into thechamber, the source holders sit upon an ABS-plastic stilt and are heldvertical by the robotic hand.

The ionization current is measured by a Keithley 6517A digitalelectrometer, (Keithley Instruments, Inc, Cleveland, OH, USA). Theprecision and reproducibility of the electrometer have beenenhanced significantly beyond the manufacturer’s ‘‘accuracy’’specifications by averaging current readings over measurementtimes of (100 to 300) s. This enhancement is most pronounced forsmall currents, in which the 1% accuracy for the short measure-ment times (166 ms) quoted by the manufacturer appears to belimited by shot-noise (statistical precision). For a low ionizationcurrent of 2 pA, a measurement time of 1000 s results in 0.1%precision. This precision is defined as the standard deviation ofthe mean for numerous current readings averaged during themeasurement time. This 2 pA current corresponds to 20 kBq of60Co, which is about as low an activity as would be measured inthis chamber. The background current is about 0.07 pA. Theprecision is statistics-limited, and thus improves in the usual way,n�1/2, as either the measurement duration or the ionizationcurrent is increased. This trend continues to a precision of at least0.005%. As demonstrated below, these measurements are repro-ducible over time to accuracy better than 0.03%.

Since measurements of a particular radionuclide are carriedout relative to a radium reference source, it is important that therelative currents are linear over a wide current range and that theratio of two currents is reproducible over time. The linearityamong the various electrometer ranges has been established bymeasuring various ionization currents on multiple ranges. Fromthis data, correction factors to the manufacturer’s calibrationwere derived and are listed in Table 1. The 200 pA range wasarbitrarily chosen to be the ‘‘correct’’ range. The measuredcurrent, Im, is related to the ‘‘correct’’ current, I, as I=m � Im+b.These correction factors are applied via software.

3. Linearity and stability tests

The linearity of the ionization-current measuring system wasexamined by following the decay of short-lived radionuclei for atleast six half-lives and comparing the observed half-lives toliterature values (Table 2 and Fig. 1). The measurements wereundertaken to verify the linearity of the system, not to provide newnuclear data. The uncertainty for each data point in a decay curve isthe combination of the standard deviation of the mean for repeatedcurrent determinations during a single measurement and theuncertainty due to positioning the source. The latter componentamounts to 0.15% for 133Xe and 0.02% for 99Mo and 99mTc. Thebetween-measurement variability for 133Xe was larger than thatfor other radionuclei due to the fact that the gas ampoule does notseat as securely in the sample holder, as seen in the plots of theresiduals from the data fits. The analysis method is describedfurther by Pibida et al. (2009), who plot residuals for fits to an 82Sr

decay curve. For 99Mo, the uncertainty analysis of Pomme (2007)was used to account for the possible trends in the residuals. Thisadditional uncertainty component was about equal to the standarddeviation of the half-life fit parameter, and the two componentswere added in quadrature. The full uncertainty evaluation for thesehalf-life measurements is described in Table 3.

All three measured half-life values are in agreement with therecommended literature values of Be et al. (2007), within theircombined uncertainties. Also, the half-life of 82Sr was determined tobe (25.4070.02) d by following the decay of a source for 7.5 half-lives. This value agreed with a determination by HPGe g-rayspectrometry and details of both determinations are given by Pibidaet al. (2009).

To quantify the reproducibility of measured activity ratios overtime, calibration-factors for various radionuclei were determinedover a span of two to three years. These calibration factors werederived from our well-established ionization chamber ‘‘A’’, whichhas been in use for over 30 years. The difference between thecalibration factor for 111In measured in August 2007 and August2008 was (0.0370.03)%, where the uncertainty listed is merely thestatistical precision (based on combined standard deviations of themeans). Similarly, the calibration factor for 60Co was measured fivetimes between June 2007 and January 2009 and the standarddeviation of the distribution for the five values was (0.0170.04)%.

4. Summary

A new ionization chamber (AUTOIC) system featuring automaticsample changing and a digital electrometer has been optimized for

ARTICLE IN PRESS

Table 3Uncertainty analysis for reported half-life determinations, where uc is the combined standard uncertainty.

Uncertainty component Typea ui (%)

133Xe 99Mo 99mTc

Standard deviation of the least-squares fit parameter A 0.020 0.036 0.021

Long-term non-linearity estimated by analyzing small apparent trends in the fit residuals using the method of Pomme (2007) B 0.005 0.027 0.010

Background B 0.004 0.005 0.004

uc (%) 0.021 0.045 0.024

The uncertainties due to radioactive impurities, counting statistics and positioning variability were included implicitly in the uncertainty from the fit parameter.

a The term type refers to the type of uncertainty evaluation, whether by statistical methods (type A) or by other methods (type B), as explained in the Guide to the

Expression of Uncertainty in Measurement, published by ISO et al. (1995).

R. Fitzgerald / Applied Radiation and Isotopes 68 (2010) 1507–1509 1509

secondary radioactivity determinations. The linearity of the systemhas been tested by measuring well-known half-lives. The long-termstability of the system is being determined by annual comparisonsof the instrument to our well-established chamber ‘‘A’’.

References

Be, M. et al., 2007. Decay data evaluation project. /http://www.nucleide.org/DDEP.htmS, Laboratoire National Henri Becquerel.

Calhoun, J.M., 1986. Radioactivity calibrations with the NBS ‘‘4p’’g ionizationchamber, and other NBS radioactivity calibration capabilities. NBS SP 250-10.

ISO, et al., 1995. Guide to the expression of uncertainty in measurement, (GUM),International Organization for Standardization, Switzerland.

Pibida, L., Fitzgerald, R., Unterweger, M., Hammond, M.M., Golas, D., 2009.Measurements of the Sr-82 half-life. Applied Radiation Isotopes 67, 636–640.

Pomme, S., 2007. Problems with the uncertainty budget of half-life measurements.A.C.S. Symposium Series 945, 282–292.

Rytz, A., 1978. International coherence of activity measurements. EnvironmentalInternational 1, 15–18.

Schrader, H., 2004. Half-life measurements with ionization chambers—a study ofsystematic effects and results. Applied Radiation Isotopes 60, 317–323.

Schrader, H., 2007. Ionization chambers. Metrologia 44, S53–S66.Taylor, B.N., Kuyatt, C.E., 1994. Guidelines for Evaluating and Expressing the

Uncertainty of NIST Measurements. NIST, Gaithersburg.