Examining electrometer performance checks with direct‐current generator in a clinic: Assessment of generated charges and implementation of electrometer checks

Abstract Purpose Medical physicists use a suitable detector connected to an electrometer to measure radiotherapy beams. Each detector and electrometer has a lifetime (due to physical deterioration of detector components and electrical characteristic deterioration in electronic electrometer components), long‐term stability [according to IEC 60731:2011, ≤0.5% (reference‐class dosimeter)], and calibration frequency [according to Muir et al. (J Appl Clin Med Phys. 2017; 18:182‐190), generally 2 years]; thus, physicists should check the electrometer and detector separately. However, to the best of our knowledge, only one study (Blad et al., Phys Med Biol. 1998; 43:2385–2391) has reported checking the electrometer independently from the detector. The present study conducts performance checks on electrometers separately from the detector in clinical settings, using an electrometer equipped with a direct current (DC) generator (EMF 521R) capable of injecting DC (effective range: ±20 pA to ±20 nA) into itself or another electrometer. Methods First, to check the nonlinearity of the generated currents from ±20 pA to ±20 nA, charges generated from the DC generator were measured with the EMF 521R electrometer. Next, six reference‐class electrometers classified according to IEC 60731:2011 were checked for repeatability at a current of ±20 pA or a minimum effective indicated value meeting IEC 60731:2011, as well as for nonlinearity within the current range from ±20 pA to ±20 nA. Results The nonlinearities for the measured currents were less than ±0.05%. The repeatability for the six electrometers was < 0.1%. While the nonlinearity of one electrometer reached up to 0.22% at a current of –20 pA, all six electrometers displayed nonlinearities of less than ±0.1% at currents of ±100 pA or higher. Conclusions This work suggests that it is possible to check the nonlinearity and repeatability of clinical electrometers with DCs above the ±30 pA level using a DC generator in a clinic.

Radiotherapy dosimeters, specified in IEC 60731:2011, 1 are measurement systems usually consisting of an ionization chamber, extension cable, and electrometer. According to an addendum to AAPM TG-51, 2 the uncertainty in the raw reading measured with these systems, excluding any variability in beam delivery, is due to the electrometer as well as the ionization chamber. Therefore, it is important to develop methods to test clinical electrometers.
All measurement systems have a limited lifetime; therefore, it is necessary to make routine checks of systems used in clinical settings. 2,3 The addendum to AAPM TG-51 showed the measurement of polarity correction for a simple QA check of the measurement systems. 2 This addendum also described a method to monitor the long-term stability of the systems with a check source or a linac-beam. 2 These checks are meant for monitoring the entire system.
Various techniques, such as stereotactic radiosurgery, stereotactic body radiotherapy, and intensity-modulated radiotherapy, have been introduced in many modern radiotherapy clinics. Consequently, clinical physicists often select the most suitable detector for the measurement object (e.g., small-field-, reference-, relative-dosimetry, and pre-treatment dose verification) with various radiotherapy beams; namely, an electrometer is connected to various detectors in these clinical measurements. Each detector and electrometer has a lifetime (due to physical deterioration of the detector's components and electrical characteristic deterioration in the electrometer's electronic parts), long-term stability [according to IEC 60731:2011, ≤0.5% (reference-class dosimeter)], and calibration frequency (according to recent surveys on reference dosimetry practices, 4 generally 2 years). Therefore, to ensure accurate measurements, each component of the measurement systems should be checked individually.
While an electrometer plays an important role in clinical dosimetry, to the best of our knowledge, performance checks on an electrometer separated from the detector have been poorly documented. 5 Therefore, this work examined the feasibility of conducting performance checks on electrometers with a direct-current (DC) generator in radiotherapy clinics. In this work, charges generated from the DC generator in a clinical setting were first verified. Then, electrometer performance checks were implemented using the DC generator to determine whether it can be used by physicists for electrometer performance checks.

| MATERIALS AND METHODS
Section 2.A summarizes the characteristics of the DC generator used in this work. Section 2.B describes the verification of charges generated from the DC generator. Section 2.C describes the implementation of electrometer checks.

2.A.1 | Range of generated DCs and injection time
The DC generator used in this work was equipped with an electrometer [EMF 521R (EMF Japan Co., Ltd., Hyogo, Japan)] that can inject DC into itself or another electrometer for performance checks ( Fig. 1). This DC generator had two terminals for injecting DCs: output-1 and −2 [ Fig. 1(a)]. Effective currents from the DC generator ranged from AE20 pA to AE2 nA for output-1 and from AE200 pA to AE20 nA for output-2. The injection time of the DC generator can be set in the range of 0.1-1000 s. Because charge is the product of current and time, the DC generator can be utilized as a charge generator as well.

2.A.2 | Traceability of the DCs to primary standards
The generated constant currents can be traced to primary standards as follows: 3. The voltage source and standard capacitor for the standard electrometer were traceable to the primary standards for DC voltage and capacitance in Japan, respectively. The frequency standard was traceable to the primary standard for frequency in Japan.

2.A.3 | Uncertainty in the DCs
The accuracies (k = 1) of the DCs from the DC generator were 0.15% and 0.10% for output-1 and output-2, respectively. The uncertainty (k = 1) in the injection time of the DCs was 0.003%. Table 1 summarizes the uncertainties in charges generated from the DC generator, which were estimated following the Guide to the Expression of Uncertainty in Measurements. 6 The principal sources of these uncertainties were as follows: 1. In-house standard device (i.e., electrometer and a universal counter) calibrations: These were obtained from their calibration certificates.   The raw electrometer readings were multiplied by the electrometer calibration coefficient (P elec ) supplied by ANTM, and the products were assumed to be the measured charges.
The nonlinearities checked here were expressed as percentage deviations [σ (%)] from linearity for the ratio of the measured charges to set charges between the RP and the other points, which were obtained by the following equation: where M (i.e., measured charge) is produced by Q (i.e., set charge at RP) and m (i.e., measured charges) is derived from q (i.e., set charges at each measurement point except for the RP). These observed nonlinearities would consist of two elements: generated current and measurement instrument nonlinearities, which were within AE0.1% each, according to the specifications. Assuming each of the two uncertainties (each of their probability distributions was assumed to be a rectangular distribution within AE0.1%) were AE0.06% (k = 1), the relative combined standard uncertainty in the observed nonlinearities would be 0.09%. Then, whether the observed nonlinearities were within 0.09% was verified. If the observed nonlinearities were 0.09% or less, it was assumed that the generated charge nonlinearities met the performance specified in the specification.
In addition to the nonlinearity checks, differences between the measured charges and set charges were assessed. The uncertainty in the measured charges, which consisted of the raw electrometer reading, P elec , and the charges generated at the radiation facility, was estimated in accordance with the Guide to the Expression of Uncertainty in Measurements. 6

2.C | Electrometer check with the DC generator
Six electrometers, which met the requirements for a reference-class electrometer described in IEC 60731:2011, were checked. The characteristics of the six electrometers are listed in Table 2.
In this work, the electrometer performance checks were carried out for repeatability and nonlinearity, which were generally performed according to IEC 60731:2011. 1 The DCs were injected into the electrometers via the following procedure:   Table 2, the same process was performed because it was difficult to inject the minimum detectable charge (i.e., 2 pC) into the electrometer.
To evaluate the repeatability for the electrometers, the relative standard deviation of 10 successive readings, expressed as a percentage of the mean indicated value, was calculated.

2.C.2 | Nonlinearity
The ranges of current values for nonlinearity checks performed here are listed in Table 3. According to IEC 60731:2011, to assess nonlinearity for an electrometer, the currents injected into it should range as follows: from the minimum effective reading to the maximum effective reading for a single-range electrometer; from the minimum effective reading on the most sensitive dose-rate range to the maximum effective reading on the least sensitive dose-rate range for a multiple-range electrometer. It was impossible to carry out nonlinearity checks for the ranges described in IEC 60731:2011 because effective currents from the DC generator were in the range of 20 pA to 20 nA. Hence, this study checked nonlinearity from AE20 pA to AE20 nA (or the maximum effective reading, when that value was less than AE20 nA) for all six electrometers.
Different currents in the range of AE20 pA to AE20 nA (or the maximum effective reading, less than AE20 nA), which were spaced at approximately the same interval of the current range expressed as a logarithmic scale, were transferred for 50 s to the six electrometers. Half of the maximum effective reading was taken as the RP for each electrometer (with the exception of electrometer 4, for which AE20 nA was taken as the RP).
These nonlinearities were expressed as percentage deviations from linearity for the ratios of the electrometer readings to set charges between the RP and the other points, which were evaluated using the equation specified in Section 6.3.3 in IEC 60731:2011.
T A B L E 2 Characteristics of the six electrometers checked in this work.  Figure 2 shows nonlinearities for measured currents from the DC generator. In all situations investigated here, the nonlinearities were below AE0.05%; the nonlinearities for the measured currents from output-1 were -0.04% (at 20 pA) or less (well below AE0.01% at each current of above 20 pA), whereas those from output-2 were well below AE0.01%.

3.A | Verification of charges from the DC generator
The discrepancies between the measured charges and set charges were within AE0.03%. The uncertainties in these measured charges (k = 1) were 0.18% for output-1 and 0.14% for output-2 (Table 4).

3.B.2 | Nonlinearity
The nonlinearities for the six electrometers were usually AE0.05% or less in the range of AE0.1 to AE20 nA (Fig. 3). When constant currents of AE20 and AE30 pA were transferred to the electrometers, the non- were well below AE0.09%. These findings suggest that the generatedcurrent nonlinearities met the performance in the specification.
Differences between the measured charges and set charges were within the uncertainties of the measured charges. These findings suggest that the DC generator generated accurate charges.

4.B | Electrometer check with the DC generator
The findings described in Section 3.B were not intended to be an indicator of the performance of the six electrometers because the F I G . 2. Nonlinearities for measured current from the DC generator: (a) from output-1 and (b) from output-2. The y-axis shows the nonlinearity (%); the x-axis indicates the injected currents on a logarithmic scale.
T A B L E 4 Uncertainty budget (k = 1) for measured charges.

4.B.1 | Repeatability
As can be seen from Table 5, the repeatability for the six electrometers was within AE0.1%. Because these values were less than AE0.1% despite the variations in the ambient environment, no significant variation would occur in their repeatability.

4.B.2 | Nonlinearity
As noted in section 3.B.b, while the nonlinearities from AE0.1 to AE20 nA were usually AE0.05% or less for the six electrometers, those at AE20 and AE30 pA occasionally ranged from AE0.05% to AE0.2% for   In this work, although the nonlinearities at 20 and 30 pA were typically somewhat larger than those for the currents above AE100 pA, our values were within the specifications described in the catalogs for the six electrometers.

| CONCLUSION
At currents of AE20-30 pA, nonlinearity and repeatability may not depend solely on the performance of the electrometer itself; variations in the electrometer's zero reading and injected current can also play a role. At current levels of 100 pA or more, however, such variations have no significant effect.

CONFLI CT OF INTEREST
We borrowed the DC generator used in this work from EMF Japan Co., Ltd. We borrowed the electrometers from Chiyoda Technol Corporation, EMF Japan Co., Ltd., and TOYO MEDIC Co., Ltd.

D A T A A V A I L A B I L I T Y S T A T E M E N T
The data that support the findings of this study are available from the corresponding author upon reasonable request.