The Monte Carlo method has become a mainstream tool for studying induced radioactivity [7, 8]. FLUktuierende KAskade (FLUKA) [9, 10] is a well-established and user-friendly tool that has been utilized for decades in designing radiation shielding and calculating radiation doses for various large particle accelerators globally. Moreover, FLUKA is an open-source software, continuously updated, and equipped with a convenient card-type input structure. FLUKA presents evident advantages for the calculation of activation. It has its own built-in burn-up calculation model, eliminating the need for any external software [11, 12]. Additionally, this study leveraged Flair, an advanced user interface with fast geometric visualization [13]. In particular, this study used FLUKA version 4–2.2 and Flair version 3.1–15.1, respectively. The batch is set to five cycles, 30 spawns, and 2×10⁷ particles per cycle.

The activation calculation involves setting the ‘PRECISIOn’ model with ‘DEFAULT’ and compiling the Relativistic Quantum Molecular Dynamics Code [14] and Dual Parton model using ldpmqmd [15]. The ‘PHYSICS,’ ‘REDDECAY,’ ‘IRRPROFI,’ and ‘DCYTIMES’ cards are used to describe the physical process, irradiation profile, and cooling time. Combining ‘RESNUCIE’ and ‘USRBIN’ with their corresponding ‘DCYSCORE’ and ‘AUXSCORE’ generates the activity of radionuclides in a region and residual dose rate. ‘LOW-NEUT’ is the card that sets up the transport of low-energy neutrons. The contribution of low-energy neutrons to the production of induced radionuclides warrants attention. The radionuclide generation cross sections in the databases Evaluated Nuclear Data File (ENDF)/B-VIIR0 [16], ENDF/B-VIIR1 [17], Japanese Evaluated Nuclear Data Library (JENDL)-3.3 [18], JENDL-4.0 [19], and TALYS Evaluated Nuclear Data Library (TENDL)-13 [20] are selected from thermal neutrons to the energy range of 20 MeV.

A simplified estimation model is established to conservatively simulate the induced radioactivity of the patient and other materials around (Fig. 1). The overall size of the treatment room is 15.95 m×23.15 m×9.45 m, and the main shielding wall is made of concrete with the same thickness as that of real. Besides, some proton treatment rooms require additional local shielding to address inadequate protection in certain actual situations. Therefore, a local shielding wall is added to the concrete, which is a rectangular iron block measuring 20 cm×440 cm×445 cm, to investigate local shielding activation and assess its potential effect on the overall induced radioactivity in the treatment room (Fig. 1). The primary culprits for beam loss are frequently the patient and the treatment head in a treatment room. All magnets and other components, excluding the nozzle are disregarded, to streamline the process and factor in component activation of the treatment head. The patient and nozzle are considered predominant hot spots. Here, the nozzle is utilized to scan the beam, so it is simplified as a sphere with an 8 cm radius made of copper. A rectangular phantom of 30 cm×30 cm×70 cm (the same size as a human chest) [21] is used to represent the patient. The International Commission on Radiological Protection (ICRP) soft tissue material from the material library of FLUKA is introduced for more accurate results as it resembles the human body composition. The couch has the same size profile as the patient but a different thickness and composition, with a 10 cm polyethylene layer and a 2 cm stainless steel support. Materials mainly discussed in this paper include tissue (patient), copper (nozzle), polyethylene (patient couch), stainless steel (patient couch support), iron (local shielding), and air.

The materials air and ICRP soft tissue are all derived from the FLUKA MATERIAL database, except for the copper, iron, polyethylene, and stainless steel, which need to be manually added using ‘COMPOUND’ and ‘MATERIAL’ cards combined. Table 1 lists the density and elemental compositions of the materials [5, 22–26].

This study recorded the induced radioactivity exclusively after 2 minutes irradiation and selected the cooling intervals downtime, 1 minute, 3 minutes, 5 minutes, 10 minutes, 30 minutes, 1 hour, 2 hours, and 5 hours. Table 2 shows two hot spots (patient and nozzle) and three energy scenarios. We set 2.50 μSv/hr as the dose limit according to China’s national standard GB18871-2002 [27].

Fig. 2 shows the residual dose distribution at the downtime under three energy scenarios. The three energy scenarios are named scenarios A, B, and C, corresponding to 220 MeV at 3.06 nA, 150 MeV at 2.64 nA, and 70 MeV at 0.192 nA, respectively. The 2.50 μSv/hr range is emphasized. It indicates that residual dose distribution varied based on the irradiation type, with higher energy causing a greater residual dose. The residual dose rate will exceed 2.50 μSv/hr within a 100 cm distance at downtime (the distance between the point where the beam bombards the surface of the phantom and the point selected in the opposite direction of the beam incident, the same as below) under the conditions of scenarios A and B. In contrast, the residual dose rate is significantly lower in scenario C. The information indicates that caution must be taken when working under high energy and intensity conditions.

Fig. 3 displays the residual dose distribution for different cooling times under scenario A. Noteworthily, the residual dose rate significantly diminishes as time elapses. Figs. 2A and 3A show that the residual dose rate in the treatment room rapidly decreases multiple times within 1 minute of the end of treatment. It declines to approximately 1–2 orders of magnitude from 1 to 10 minutes. Additionally, combined with Fig. 2 above, the dose rates near the patient, nozzle, and couch are higher than in other treatment room areas. Hence, these regions are the major contributors to the residual dose rates.

Fig. 4 shows the one-dimensional residual dose rate distribution under three energy scenarios at various cooling times, whereas Table 3 lists the corresponding values at 30 cm and 100 cm from the phantom surface. A quantitative analysis of Table 3 and Fig. 4 reveals that different energies cause distinct residual radiation fields. In particular, the residual dose rates in scenarios B and C are approximately 0.73 and 0.03 times that of scenario A, respectively.

The residual dose rate in the treatment room will reach roughly 56.16 μSv/hr within a 30 cm distance at downtime after the 2 minutes treatment is finished under scenario A. The durations required to lower the dose rates at 30 cm to 2.50 μSv/hr are 10, 8, and 0 minute for scenarios A, B, and C, respectively. Similarly, the corresponding times are 5, 3, and 0 minute at 100 cm. This information enables us to regulate when employees can enter the premises and create customized safety protocols for various tumor treatment strategies.

Meanwhile, we analyzed the levels of induced radioactivity in various materials based on their specific activity. Tables 4–9 list the primary activation products found in different materials after shutdown and their corresponding specific activities. Only those products with a ratio to the total activity of >0.5% are included. Meanwhile, Figs. 5 and 6 illustrate the changes in specific activities of major contributing radionuclides (scenario A) in materials after shutdown, and the total activities under three energy scenarios with time.

The comparison of “total” curves illustrated in Fig. 5 confirms that the total activity of the patient exceeds that of the nozzle by a factor of 3–4. This result indicates that the patient is the primary source of residual radiation in the short term, but the significance of its total activity diminishes over time. In particular, the total specific activity of the patient exhibited a 50% reduction within 1 minute and a 75% reduction within 3 minutes after being cooled. This can be attributed to the presence of short half-life radionuclides, such as ¹⁵O (T_1/2= 2.03 minutes), ⁸Be (T_1/2=8.19×10⁻¹⁷ seconds), and ¹¹C (T_1/2= 20.39 minutes), which are the dominant ones generated in the patient. Trace amounts of metal and long half-life radionuclides can be detected in the patient, but their proportion contribution to the overall specific activity is minimal, only accounting for approximately 10% in total, including ²⁴Na (T_1/2=15 hours), ³⁹Ca (T_1/2=8.79 minutes), and ⁵⁵Fe (T_1/2= 2.74 a) only contribute a negligible fraction to the total specific activity, specifically 1.86×10⁻⁴, 3.28×10⁻², and 5.94× 10⁻⁹, respectively. However, the majority of the specific activity is responsible for ¹⁵O, accounting for 60% of the total.

Table 5 lists the primary radionuclides in the nozzle and their corresponding specific activities. Copper typically contains a diverse range of radionuclides, including isotopes of V, Cr, Sc, Mn, Fe, Co, Ni, Cu, and Zn, with an average distribution. Notably, the long half-life nuclides, such as ⁴⁹V (T_1/2=330 days), ⁵⁵Fe (T_1/2=2.74 a), and ⁶⁰Co (T_1/2=5.27 a), constitute <0.5% of the total radionuclides present in copper. Drawing insights from Table 5 and Fig. 5B, it becomes evident that the nozzle houses are predominantly short to medium half-life radionuclides, notably ⁶²Cu (T_1/2=9.74 minutes), ⁶⁶Cu (T_1/2= 5.12 minutes), ⁵⁹Cu (T_1/2=1.36 minutes), etc. Among these, ⁶²Cu appears as the primary contributor, constituting 68.95% of the total specific activity. Notably, the decline in total specific activity of the nozzle over time is not as rapid as observed in the patient. Approximately 8.31 minutes are required for the total specific activity of the nozzle to decrease by half after shutdown.

Fig. 6 depicts the activation levels of materials around the patient and nozzle, whereas Tables 6–9 detail the main activation products along with their specific activities. Dominant nuclides in the couch include ¹²B (T_1/2=20.2 millisecond), ⁸Li (T_1/2=838 millisecond), ⁸Be (T_1/2=8.19×10⁻¹⁷ seconds), and ¹¹C (T_1/2=20.39 minutes), and all except ¹¹C have extremely short half-lives. The total specific activity of the couch decreases quickly, dropping by approximately 87.29% within a minute, following the cessation of operation. Subsequently, the specific activity of ¹¹C becomes nearly equivalent to the total specific activity of the couch. The total specific activity of the couch is 1–2 orders of magnitude lower than that of the patient and nozzle for 1–5 minutes after shutdown. Hence, the radiation effects of the couch are deemed relatively negligible.

The couch support and local shielding wall, constructed from stainless steel and iron, respectively, yield multiple metal radionuclides, including ²⁸Al (T_1/2=2.25 minutes), ⁵²V (T_1/2=3.74 minutes), ⁵⁵Cr (T_1/2=3.5 minutes), ⁵¹Mn (T_1/2=46.2 minutes), ^52mMn (T_1/2=21.10 minutes), ⁵⁶Mn (T_1/2=2.508 hours), ⁵³Fe (T_1/2=8.51 minutes), among others. These radionuclides demonstrate medium to long half-lives, accompanied by low specific activity as detailed in Tables 7 and 8. The total specific activities of the couch support and local shielding wall exhibit a reduction of 48.66% and 40.91% within 1 minute cooling during downtime, respectively. Furthermore, these activities are 2–4 orders of magnitude lower than that of the patient, indicating less contribution to the total induced radiation field in the treatment room in the short term.

However, metal materials, such as copper nozzle, stainless steel couch support, and locally shielded iron wall, will produce nuclides with long half-lives postirradiation, and their ratio may seem minor initially, but their long-term effects will be significant. Consequently, implementing suitable actions is imperative to guarantee its secure management and prolonged oversight. Medical facilities or centers must alert healthcare workers, patients, and their staff to steer clear of the treatment heads and metallic components, whereas maintenance staff are advised to use professional gloves during operations. Additionally, metal materials in the treatment room are recommended to be minimized, substituting them with substances, such as polyethylene, known for their short half-life nuclides upon activation.

Table 9 and Fig. 6D indicate that the primary contributors to the total specific active of air at downtime include ¹⁶N, ¹⁵O, ¹⁴O, ¹³B, ¹³N, ¹³O, ¹²B, ¹²N, ¹¹C, ¹⁰C, ⁹Li, ⁹B, ⁹C, ⁸Li, ⁸Be, ⁸B, and ⁶He. Among them, ¹⁵O (T_1/2=2.03 minutes) stands out as the most significant contributor, followed by ⁸Be (T_1/2=8.19×10⁻¹⁷ seconds), which has an exceedingly short half-life, decaying to zero in <1 minute. Furthermore, other nuclides with similar short half-lives in the air show quick specific activity decline to zero in a matter of minutes. The total specific activity of the air decreases to one-third of its original value after 1 minute of cooling, after which ¹⁵O, ¹³N, and ¹¹C contribute 90% of the total specific activity. Furthermore, as the accelerator operates with negative pressure ventilation in this dynamic scenario, the actual activity of the air is expected to be lower than the simulated results. Consequently, the effect of the air, after a brief irradiation period, is comparatively minor compared to other materials.

The preceding analysis reveals that materials nearer to the source term demonstrate higher activity, and those with a greater atomic number generate a greater variety of radionuclides and durable nuclides. Nuclides with brief half-lives postshort irradiation are predominant, but those with extended half-lives must be considered over time, thus the use of high-Z materials in the treatment area should be reduced and kept away from the source term. Furthermore, the patient is a significant contributor to the induced radiation dose in the treatment room after a short period of exposure, but its importance decreases significantly over time due to the short half-life of the dominant radionuclides in the patient. The variation in the residual dose rate under the three energy scenarios is congruent with the total specific activity of the various materials. The higher the energy, the higher the residual dose rate and total specific activity of the material, thereby enabling us to limit the time of staff entering the treatment room based on different treatment scenarios, such as 10 minutes, 8 minutes, and no waiting for scenarios A, B, and C, respectively. This approach not only improves work efficiency but also safeguards medical personnel.