For industrial high-dose dosimetry calibration services, the measurement quantity of interest is absorbed dose (to water, primarily), reported in units of gray. At NIST, absorbed dose is realized by a water calorimeter in a gamma-ray field produced by a Vertical Beam Co-60 source with an activity (as of April, 2007) of 48 TBq (1.3 kCi). Because the dose rate for this source is relatively low, a high-dose-rate source is needed to perform customer calibrations. The bulk of the services are provided through three Gammacell 220 Co-60 irradiators (Nordion, Canada). Their activities are: 37 TBq (1.0 kCi, serial number GC45); 137 TBq (3.7 kCi, serial number GC232); and 666 TBq (18 kCi,, serial number GC207), all as of April, 2007. Though also available for service operation, the Pool Co-60 source (0.15 kCi as of April, 2007) is rarely requested for calibration work by customers because its low absorbed-dose rate is no longer relevant to industrial needs; however, the Pool source continues to play an important role in the NIST source calibration scheme described here.

Calibration of the gamma sources within the NIST high-dose calibration facility are performed by measuring the ratio of the alanine dosimeter response for the source being calibrated to that of a reference source. The absorbed dose for these internal calibrations is 1 kGy or less. This approach simplifies the source comparison to a measurement of two quantities, dosimeter response and time. Absorbed dose is not computed for this calibration exercise because these added steps will introduce additional uncertainties inherent in the calibration curve, and it avoids any issues that might arise from non-linearity in the dosimetry system dose response.

The four Co-60 sources above are used for NIST high-dose calibration services. Prior to 2004, the Pool source and the Gammacells were each calibrated by a direct dosimeter response ratio to the Vertical Beam Co-60 source. However, the Vertical Beam 60Co source dose rate has decayed to a level that requires excessive periods of time (>24 h) to perform comparisons at the absorbed doses (≥1 kGy) routinely used. Since the Vertical Beam Co-60 source irradiations are performed under water, with the water surface in the vessel exposed to the room environment, there were concerns that a variation in the water level would contribute significantly to the uncertainty of the measurement, as it would be difficult to keep the water level constant for a prolonged period. To address the increasingly longer Vertical Beam irradiation times, modifications to the calibration scheme were developed. In 2004 the calibration scheme was modified so that the Vertical Beam Co-60 source would be compared only to the Pool Source. To improve several aspects of the measurement, the absorbed dose for this comparison was adjusted lower (140 Gy) for the Pool/Vertical-Beam source comparison. The three Gammacells are calibrated by comparison to the Pool Source. The Pool source serves well as an intermediary source in the calibration scheme as its dose rate is closer to that of the Gammacells; this permits longer exposure times, resulting in reduced timer uncertainties.

The calibration scheme begins with the known dose rate of the Vertical Beam Co-60 source, established in 1990. To transfer that dose rate, eight alanine pellets are irradiated in the calorimeter water tank. The pellets are stacked in a watertight polystyrene cylinder whose axis is fixed perpendicular to the Vertical Beam Co-60 source at a scale distance of 58.8 cm. The water surface is set at a scale distance of 53.8 cm. This design differs from the published scheme. In the published scheme, this irradiation was done in a polystyrene phantom and a scaling theorem was used to correct for differences in photon interaction cross sections between polystyrene and water. This direct underwater measurement was an improvement as it eliminated scaling theorem uncertainties. In the current calibration scheme the dosimeters are irradiated at the appropriate distance underwater, and no additional corrections are applied to the measured data.

Concurrent to the Vertical Beam irradiation, alanine dosimeters are irradiated to the same absorbed dose in the absolute center of the isodose region of the Pool-source gamma field. This comparison may be repeated as necessary to achieve an acceptable precision of 0.5 %. The dosimeters are measured using EPR, and the dosimeter response is divided by the irradiation time to convert to units of response/s. Once the measurements are converted to these common units, the established dose rate in the Vertical Beam source can be used to determine the dose rate in the Pool source.

Similarly, a series of comparisons are made between the dose rates at the center positions of the Pool source and the three Gammacells (GC45, GC232, and GC207) with the alanine transfer vial placed on a polystyrene pedestal set to position the dosimeters in the absolute center of the isodose region of the gamma field. For these comparisons a higher dose is used (e.g., 1 kGy) to reduce the contribution of uncertainty in the timer settings for the highest dose-rate Gammacell. In the GC232 and GC207, irradiations are performed on a pedestal either in a stainless-steel dewar or without a dewar; the dewar is used to improve temperature control at the extremes of the irradiation temperature ranges. The dosimeters are measured and the response/s is determined. The center-position dose rate for each Gammacell is determined by comparison to the Pool source center-position dose rate.

It should be noted that the equivalent transit time, the time subtracted from the timer setting that accounts for the absorbed dose received by the dosimeters during the delivery of the dosimeters to and from the irradiation position, is determined for each source. To measure the equivalent transit time, alanine dosimeters are irradiated for a series of very short times. Typically, these times are 5 s, 10 s, 20 s, 30 s, 40 s, and 50 s. The dosimeter response is measured and plotted versus irradiation time. A linear regression of these data is extrapolated to the x axis. The absolute value of the x intercept is the equivalent transit time.

Customer-supplied dosimeters for calibration are routinely irradiated in one of three calibration geometries: ampoule, Perspex, and film block. The rates for each of these positions in the Gammacells (though not all positions are used in each Gammacell), with and without a dewar present, are determined by comparison of the response/s for dosimeters irradiated in these positions to the response/s for dosimeters irradiated in the center position of the respective Gammacell. This final portion of the calibration scheme remains unchanged from that previously published. As a final check of the dose rates, all irradiation geometries are compared to confirm an equivalent measurement response for dosimeters irradiated to 1 kGy.

Approximately annually, dosimeter-response comparisons between the Gammacells, Pool and Vertical Beam Co-60 sources are performed. The ratios of source dose-rates are determined and ongoing control charts are maintained. These measurements have a standard uncertainty of approximately 0.5 % (k=1).

When possible, dosimetry comparisons are performed between NIST and the high-dose calibration facility of the National Physical Laboratory of the United Kingdom. Dosimeters from each facility are exchanged, measured, and the results compared. Participation in other NMI or multi-NMI international comparisons occurs as appropriate.