Dose Reconstruction After a WD‒XRF Spectrometer Accident Using Minute-Rate TLD Measurements Authors: Yiying Gao, Haoyu Wang, Ling He, Liangguo He, Wantian Wang, Linyu Li, Yurou Zhang, Yuxing Xue, Shuyu Zhang, Deming Liu Published: 2026 ABSTRACT Accurate dose estimation following radiation accidents remains challenging, particularly when personal dosimeter data are unavailable. This study presents a practical post-accident dose reconstruction approach based on minute-rate thermoluminescent dosimeter (TLD) measurements combined with exposure scenario reconstruction, demonstrated through a wavelength-dispersive X-ray fluorescence (WD‒XRF) spectrometer accident. The exposure scenario was reconstructed by placing LiF(Mg,Cu,P) TLDs at seven representative positions corresponding to head and hand locations, with the spectrometer operated under the recorded technical parameters (40 kV, 7 mA). Minute personal dose equivalent rates, Hp′(10), were measured and combined with reconstructed retention times to calculate cumulative doses, from which skin absorbed doses were derived. The measured dose rates ranged from 0.13 to 358.38 mSv/min for the head and from 0.07 to 620.24 mSv/min for the hands. The resulting cumulative skin absorbed doses were 7.58 Gy for the head and 12.52 Gy for the hands, consistent with reported dose ranges associated with localized radiation injury. This approach provides a feasible supplementary strategy for retrospective dose reconstruction under emergency or resource-limited conditions when conventional dosimetric information is unavailable. Introduction Accurate dose estimation following radiation accidents is widely recognized as a critical but challenging task. 1 IAEA accident investigation reports consistently indicate that direct personal dosimeter data are often unavailable at the time of exposure, particularly in industrial, medical, and occupational accident scenarios.2,3 Consequently, post-accident dose assessment frequently relies on indirect reconstruction approaches rather than real-time measurements. 4 According to IAEA investigations and peer-reviewed studies, absorbed dose estimation after accidents is most commonly performed using Monte Carlo–based physical reconstruction, biological dosimetry, or combinations of physical measurements and clinical indicators.5,6 Monte Carlo simulation has been widely applied when sufficient information on radiation source characteristics, geometry, and exposure duration can be obtained, and reasonable dose estimates have been reported in well-documented cases. 7 However, both IAEA reports and journal literature emphasize that the accuracy of Monte Carlo reconstruction strongly depends on the availability and reliability of exposure information, which is often incomplete or uncertain after an accident. Biological dosimetry represents another important pillar of post-accident dose assessment. Cytogenetic methods, particularly dicentric chromosome analysis, are recommended by international guidance and have been applied in several reported accidents when physical dosimetry is lacking.8,9 Nevertheless, the use of biological dosimetry is constrained by practical factors, including the need for specialized laboratory capacity, trained personnel, and appropriate time windows for sample collection and analysis. IAEA guidance indicates that these requirements may not be met in many real accident settings, especially during the early response phase or in resource-limited environments.5,6 When neither detailed physical reconstruction nor biological dosimetry is readily feasible, dose estimation may rely on indirect approaches such as clinical manifestations or retrospective physical measurements. 10 Both IAEA reports and journal publications highlight that these approaches introduce substantial uncertainty, particularly for localized exposures and non-uniform dose distributions.11-13 Taken together, these limitations indicate the need for practical supplementary strategies in post-accident dose estimation when conventional methods are constrained. 14 In this context, measurement-based approaches, such as TLD-based reconstruction combined with exposure scenario analysis, may provide a feasible option for retrospective dose assessment. Building on this perspective, the present study uses a WD‒XRF spectrometer accident as a representative case to apply a practical post-accident dose reconstruction approach based on minute-rate TLD measurements and exposure scenario reconstruction. Rather than introducing a new dosimetric model, this study focuses on the practical implementation of this approach under conditions where conventional dose assessment methods are constrained. By focusing on localized skin dose estimation, this study aims to provide a feasible supplementary strategy for such accident scenarios. Materials and Methods All measurement instruments and detection devices used in this study complied with the national verification and calibration standards. Prior to data collection, each device was subjected to routine performance checks and was confirmed to meet the required technical specifications to ensure measurement accuracy and reliability. Exposure Scenario and Case Context On November 28, 2020, a malfunction occurred in a WD‒XRF spectrometer at an industrial facility, providing the exposure scenario analyzed in this study. During an extended maintenance operation, the safety interlock system had been removed or was not in place, and the equipment remained energized, resulting in prolonged proximity of the head and hands to the X-ray source. The exposure involved repeated manual access to the interior of the instrument and lasted approximately six hours, leading to non-uniform and localized irradiation of the head and hands. Clinical symptoms consistent with radiation-induced skin injury developed shortly after the exposure and progressed over the following weeks. The affected individual was ultimately diagnosed with acute radiation dermatitis at a specialized medical center, supporting the relevance of localized skin dose assessment in this case. Because no personal dosimeter was worn at the time of exposure, direct dose records were unavailable. Therefore, dose estimation was performed using a retrospective reconstruction strategy based on exposure scenario analysis, operational simulation, and onsite environmental dose measurements. Geometric configuration, exposure duration, and exposure mode were integrated to estimate absorbed skin doses to the head and hands. Exposure Parameters for Dose Reconstruction Field investigation identified the involved equipment as a WD‒XRF spectrometer manufactured by Jiangsu Skyray Instrument Co., Ltd. The instrument was produced in January 2016 and installed in June 2016, with a rated operating capacity of 40 kV and 7 mA. To support retrospective dose reconstruction, the exposure scenario was characterized using a set of representative operator positions during maintenance activities. Seven key positions (labeled A–G) were identified based on onsite inspection, operator recollection, and reconstruction of the maintenance procedure. For each position, the distance to the radiation source and the corresponding retention time (Ts) were determined. Distances were measured directly at the site, while Ts values were estimated through reconstruction of the maintenance procedure. The reconstructed exposure geometry and representative positions are shown in Figure 1 and summarized in Table 1. These positions were intended to represent the major postures and locations associated with head and hand exposure during the repair process.Figure 1.Operator positions A–G used for dose reconstruction in relation to the WD‒XRF spectrometer. (A), in front of the spectrometer; (B), left side of the spectrometer; (C), in front of the computer desk; (D), right side, near the cable line; (E), left side, near the cable line; (F), right side, sample-loading area; (G), in front of the front panel. The corresponding source-to-head distances, source-to-hands distances, and retention times are summarized in Table 1Table 1.Positions (A–G): Distance to Source and Retention Time L Distance to source (cm)T (min)HeadHandsA (In front of the spectrometer)301060B (Left side of the spectrometer)20560C (In front of computer desk)12083180D (Right side, near cable line)873040E (Left side, near cable line)402010F (Right side, sample-loading area)681030G (In front of the front panel)301220Note. Distances were measured on site; retention times were estimated through reconstruction of the maintenance procedure. Operator positions A–G used for dose reconstruction in relation to the WD‒XRF spectrometer. (A), in front of the spectrometer; (B), left side of the spectrometer; (C), in front of the computer desk; (D), right side, near the cable line; (E), left side, near the cable line; (F), right side, sample-loading area; (G), in front of the front panel. The corresponding source-to-head distances, source-to-hands distances, and retention times are summarized in Table 1 Table 1.Positions (A–G): Distance to Source and Retention Time L Distance to source (cm)T (min)HeadHandsA (In front of the spectrometer)301060B (Left side of the spectrometer)20560C (In front of computer desk)12083180D (Right side, near cable line)873040E (Left side, near cable line)402010F (Right side, sample-loading area)681030G (In front of the front panel)301220Note. Distances were measured on site; retention times were estimated through reconstruction of the maintenance procedure. Positions (A–G): Distance to Source and Retention Time Note. Distances were measured on site; retention times were estimated through reconstruction of the maintenance procedure. Dose Collection Scheme By simulating the onsite environment and maintenance procedures corresponding to the exposure scenario, personal dose equivalents at representative head and hand locations were obtained and subsequently used to estimate localized absorbed doses. During the measurements, the irradiation angle was fixed, and only the effect of source-to-detector distance associated with different exposure positions was considered. At each of the seven representative exposure positions, five TLDs were placed separately near the head and hand locations. The WD‒XRF spectrometer was operated under the same technical parameters as during the incident (40 kV, 7 mA), with an exposure duration of 1 min per measurement. The TLDs were then read using a TLD reader (T360A) to obtain personal dose equivalents for each position. TLD Calibration and Quality Control The TLD system used in this study was calibrated according to standard procedures before measurement. Background signals were measured and subtracted from the readings. For each representative position, five TLDs were performed to improve reliability and to evaluate variability among readings. Because the WD‒XRF exposure involved low-energy X-rays and non-uniform irradiation geometry, the potential influence of energy dependence and angular response was considered in the interpretation of the reconstructed doses. Dose Estimation Method For each position L (A–G), five TLDs readings Hp,i(10) L (i = 1,…,n; n = 5) were acquired with a fixed irradiation angle and an exposure time of t = 1 min per reading. The position-specific minute dose rate is as follows: (1) Hp′(10)L=1n∑i=1nHp,i(10)Lt(mSv/min) The personal dose equivalent contributed by position L during the incident is obtained by multiplying the minute dose rate by the retention time at that position: (2) Hp(10)L=Hp′(10)L T(mSv) Summing the per-position contributions over positions A–G yields the cumulative personal dose equivalent for the body site (head or hands): (3) Hp(10)=∑L=AGHp(10)L(mSv) The skin absorbed dose is obtained from Hp(10) using a conversion coefficient fpr (GBZ/T 244-2017): (4) Ds=fpr Hp(10)(mGy) (5) fpr=Cks/CkP(mGy/mSv) Here, Cks is the conversion coefficient from the air kerma to the skin absorbed dose (mGy/mSv), and CkP is the conversion coefficient from the air kerma to the personal dose equivalent (mSv/mGy). Coefficients were taken for a photon energy of 0.04 MeV with the appropriate irradiation geometry: AP for Cks and 0° for the hands/45° for the head for CkP. Because the exposure was localized and non-uniform, the selected conversion coefficients were applied as an approximate approach for retrospective skin dose estimation rather than as a fully individualized organ dose conversion. Results Measured Minute Personal Dose Equivalent Rates Hp’ (10) at Positions A–G Table 2 summarizes the minute personal dose equivalent rates Hp′(10) (mSv/min) for the head and hands across positions A–G, derived from the measurement scheme above. For the head, Hp′(10) ranged from 0.13 mSv/min (D) to 358.38 mSv/min (G); for the hands, the range was 0.07 mSv/min (D) to 620.24 mSv/min (G). To better visualize the large variation in Hp′(10) across positions, the position-specific values for the head and hands are presented in Figure 2 using a logarithmic y-axis.Table 2.Measured Personal Dose Equivalent Rates Hp′(10) (mSv/min) for the Head and Hands at Seven Retention Positions (A–G) n Head Hp′(10) (mSv/min)Hands Hp′(10) (mSv/min)ABCDEFGABCDEFG14.013.320.210.090.701.10320.83157.575.950.490.070.564.06612.7124.083.920.390.080.610.50313.15134.336.370.490.070.565.04618.1033.824.770.210.080.580.72317.58119.776.370.490.070.635.32614.7444.833.040.200.160.580.57665.42136.435.460.490.070.564.83636.2355.102.370.220.260.570.65174.91138.955.810.490.070.634.83619.43Mean4.373.480.250.130.610.71358.38137.415.990.490.070.594.82620.24Note. Hp′(10) denotes the minute personal dose equivalent rate (mSv/min). Each value is the mean of five TLDs measurements for the respective position and body site.Figure 2.Bar chart showing the distribution of minute personal dose equivalent rates Hp′(10) (mSv/min) at positions A–G for the head and hands. Values represent the mean of five TLD measurements at each position. A logarithmic y-axis was used to better visualize the large variation across positions Table 2.Measured Personal Dose Equivalent Rates Hp′(10) (mSv/min) for the Head and Hands at Seven Retention Positions (A–G) n Head Hp′(10) (mSv/min)Hands Hp′(10) (mSv/min)ABCDEFGABCDEFG14.013.320.210.090.701.10320.83157.575.950.490.070.564.06612.7124.083.920.390.080.610.50313.15134.336.370.490.070.565.04618.1033.824.770.210.080.580.72317.58119.776.370.490.070.635.32614.7444.833.040.200.160.580.57665.42136.435.460.490.070.564.83636.2355.102.370.220.260.570.65174.91138.955.810.490.070.634.83619.43Mean4.373.480.250.130.610.71358.38137.415.990.490.070.594.82620.24Note. Hp′(10) denotes the minute personal dose equivalent rate (mSv/min). Each value is the mean of five TLDs measurements for the respective position and body site. Measured Personal Dose Equivalent Rates Hp′(10) (mSv/min) for the Head and Hands at Seven Retention Positions (A–G) Note. Hp′(10) denotes the minute personal dose equivalent rate (mSv/min). Each value is the mean of five TLDs measurements for the respective position and body site. Bar chart showing the distribution of minute personal dose equivalent rates Hp′(10) (mSv/min) at positions A–G for the head and hands. Values represent the mean of five TLD measurements at each position. A logarithmic y-axis was used to better visualize the large variation across positions Overall, the dose rates followed the order G > A > B > F > E > C > D for both the head and hands, with positions A and G exhibiting the highest values. The observed differences among positions mainly reflect variations in exposure geometry and shielding conditions, rather than being interpreted as an independent comparison of position-specific risk. Skin Absorbed Dose (Ds) for Head and Hands Using the measured Hp′(10) at positions A–G together with the corresponding retention times Ts, we computed cumulative skin absorbed doses Ds for the head (Table 3) and hands (Table 4) by aggregating contributions over all seven positions and their total exposure durations via Ds=fprHp(10). This calculation incorporated all seven positions and their total exposure durations, yielding cumulative absorbed dose estimates that represent the total radiation received by each body part over the course of the accident. The results revealed that the cumulative skin absorbed dose for the head was 7.58 Gy, whereas that for the hands was markedly higher (12.52 Gy).Table 3.Calculated Skin Absorbed Dose (Ds) for the Head at Seven Retention Positions (A–G) L Hp′(10)L (mSv/min)T (min)Hp(10) L (mSv)Hp′(10) (mSv)Cks (mGy/mSv)CkP (mSv/mGy)fpr (mGy/mSv)Ds (Gy)A4.3760262.27716.200.8780.8940.987.58B3.4860208.8C0.2518045D0.13405.2E0.61106.1F0.713021.3G358.38207167.6Table 4.Calculated Skin Absorbed Dose (Ds) for the Hands at Seven Retention Positions (A–G) L Hp′(10)L (mSv/min)T (min)Hp(10) L (mSv)Hp′(10) (mSv)Cks (mGy/mSv)CkP (mSv/mGy)fpr (mGy/mSv)Ds (Gy)A137.41608244.621250.300.8781.490.5912.52B5.9960359.4C0.4918088.2D0.07402.8E0.59105.9F4.8230144.6G620.242012404.8 Table 3.Calculated Skin Absorbed Dose (Ds) for the Head at Seven Retention Positions (A–G) L Hp′(10)L (mSv/min)T (min)Hp(10) L (mSv)Hp′(10) (mSv)Cks (mGy/mSv)CkP (mSv/mGy)fpr (mGy/mSv)Ds (Gy)A4.3760262.27716.200.8780.8940.987.58B3.4860208.8C0.2518045D0.13405.2E0.61106.1F0.713021.3G358.38207167.6 Calculated Skin Absorbed Dose (Ds) for the Head at Seven Retention Positions (A–G) Table 4.Calculated Skin Absorbed Dose (Ds) for the Hands at Seven Retention Positions (A–G) L Hp′(10)L (mSv/min)T (min)Hp(10) L (mSv)Hp′(10) (mSv)Cks (mGy/mSv)CkP (mSv/mGy)fpr (mGy/mSv)Ds (Gy)A137.41608244.621250.300.8781.490.5912.52B5.9960359.4C0.4918088.2D0.07402.8E0.59105.9F4.8230144.6G620.242012404.8 Calculated Skin Absorbed Dose (Ds) for the Hands at Seven Retention Positions (A–G) Discussion This study demonstrates the feasibility of dose reconstruction for localized skin exposure from a WD–XRF spectrometer accident, using a method that integrates minute-rate TLD measurements with scenario reconstruction. When neither personal dosimeter data nor biological dosimetry was available, our approach still yielded quantitative dose estimates for the head and hands and captured the non-uniform exposure pattern. The reconstructed skin absorbed doses were 7.58 Gy for the head and 12.52 Gy for the hands, which are consistent in magnitude with reported dose ranges associated with localized radiation-induced skin injury. This agreement supports the plausibility of the reconstruction and suggests that the estimated doses are clinically reasonable in the context of localized, non-uniform skin exposure. In addition, the present findings provide practical reference data for dose reconstruction in similar radiation accidents. Analysis of the relationship between reconstructed doses and clinical outcomes may also improve understanding of dose–response relationships and support the refinement of exposure limits and injury prevention strategies. Current approaches to absorbed dose estimation rely mainly on personal dosimeter measurements, physical simulations, or combinations of both. For example, TLDs or electronic personal dosimeters are often used for real-time exposure monitoring, followed by dose reconstruction based on recorded data. 15 Alternatively, Monte Carlo simulations and computational human phantoms are applied to model radiation transport and energy deposition for organ-specific dose estimation, particularly when direct measurements are unavailable. 16 Compared with these methods, the approach presented here combines post-incident TLD measurements with exposure scenario reconstruction, providing a practical basis for retrospective dose estimation in the absence of real-time monitoring. This measurement-based strategy integrates empirical measurements with exposure scenario reconstruction and may improve practical applicability while providing plausible dose estimates under constrained post-accident conditions. Beyond methodological feasibility, the proposed dose reconstruction approach provides practical value for post-accident dose assessment when personal dosimeter data are unavailable. By combining post-incident TLD measurements with exposure scenario reconstruction, individual radiation doses can be estimated retrospectively under constrained conditions.17,18 By providing quantitative dose estimates, this approach may provide useful reference information for post-accident dose assessment and clinical interpretation. Although further validation and additional data would improve accuracy, the method can be readily applied in practice. Model parameters should be updated periodically to reflect current standards and guidelines.19-21 Wider application of this approach may contribute to improved radiation protection and occupational safety in industrial settings.22-24 Several limitations should be considered. First, the reconstruction relied on retrospectively estimated retention times and reconstructed exposure positions, which may introduce uncertainty in the dose estimation. Second, the irradiation geometry was simplified, and individualized anatomical modeling was not performed, which may affect the precision of localized dose estimation.25,26 Third, due to the localized and non-uniform nature of the exposure, the conversion from Hp(10) to skin absorbed dose should be interpreted as an approximate retrospective estimate. Finally, biological dosimetry data were not available, which limited independent validation of the reconstructed doses. In accordance with the Chinese regulatory system Classification of Radiation Apparatus (MEE Order No. 66, 2017), 27 WD–XRF spectrometers are classified as Class III radiation devices and are generally considered to pose a relatively low risk. However, despite this classification, insufficient routine supervision and inadequate operational control can still lead to severe radiation accidents. These findings indicate that WD–XRF spectrometers should be incorporated into radiation device management systems with more effective oversight and control measures. Effective management of radiation devices is essential for ensuring radiation safety and must be implemented in accordance with regulatory requirements.28,29 Regulatory authorities should strengthen safety training and establish clear operational guidelines to ensure the proper use and maintenance of radiation equipment.30,31 Standardized management and supervision play key roles in reducing radiation risk and preventing accidents. 32 Conclusion This study presents a practical measurement-based approach for dose reconstruction after a WD–XRF spectrometer accident using minute-rate TLD measurements and exposure scenario reconstruction. The estimated skin absorbed doses were 7.58 Gy for the head and 12.52 Gy for the hands, which were consistent in magnitude with reported dose ranges associated with localized radiation-induced skin injury. This approach may serve as a supplementary strategy for retrospective dose assessment when conventional dosimetric data are unavailable. Further refinement of exposure scenario reconstruction and uncertainty analysis would help strengthen similar applications in future radiation accident assessments.