Development and dosimetric characterization of an in vitro benchtop Am-241 alpha irradiator platform

Abstract

Purpose:

To expand preclinical efforts aimed at understanding the unique radiobiology of high linear energy transfer (LET) alpha particles and their implications for cancer therapy, robust, accessible, and high-throughput in vitro alpha irradiation platforms are needed. Such platforms must be extensively characterized and validated to ensure rigor and reproducibility. Here, we describe the construction and dosimetric characterization of a benchtop in vitro alpha irradiation platform using an Americium-241 (²⁴¹Am) source.

Methods and Materials:

A square foil with embedded ²⁴¹Am was used as the alpha source (activity 3.3785 mCi). The source was secured in the roof of a custom housing incorporating a 3D-printed microcapillary array collimator to exclude large-angle alpha particles, thereby narrowing the energy spectrum, and improving homogeneity. A movable shelf accommodates cells grown on glass slides at variable distances from the source. Dosimetric characterization included direct measurement of dose homogeneity, source self-attenuation, energy spectra, absorbed dose, and dose rate.

Results:

The ²⁴¹Am source, when used in conjunction with the microcapillary array, provided a spatially uniform and geometrically stable irradiation field. The mylar entrance window resulted in predictable shifting and broadening of the alpha particle energy spectra. Absorbed dose and dose-rate measurements (5.815 – 7.721 mGy/s) demonstrated feasibility of controlled in vitro alpha particle irradiation studies.

Conclusions:

We established a turnkey, low-cost in vitro alpha irradiation platform suitable for high-throughput preclinical studies. The homogeneity, alpha particle energy spectra, absorbed dose and dose rate were characterized and validated through direct measurements. Given the approximately 432-year half-life of ²⁴¹Am, this platform will provide stable and reliable dose rates for decades, facilitating long-term radiobiological investigations.

Introduction

Ionizing radiation (IR) using high linear energy transfer (LET) alpha particles has the potential to offer distinct dosimetric and biological advantages over low-LET photon and particulate radiation. The indications and modalities for delivering alpha radiation clinically are growing, including Radium-223 dichloride ([²²³Ra]RaCl₂; which increases survival in prostate cancer patients with bone metastases)^1–4, targeted radioligands using alpha-emitters^5,6, Diffusing alpha-emitter radiation therapy (DaRT) implants^7–9, and external beam heavy ion administration using the helium nucleus (an alpha particle)^10,11. Alpha irradiation may become increasingly clinically relevant not only due to the potential for many of these modalities to deliver better organ at risk (OAR) sparing and improved local control – but also due to potential synergy with immune checkpoint inhibition (ICI) thought to be related to the biologically distinct double strand DNA breaks (DSB) caused by alpha irradiation^12–14.

Despite that, the degree of preclinical research on alpha irradiation is dwarfed by the body of literature concerning x-ray irradiation, at least in part due to the logistical difficulties associated with conducting in vitro alpha irradiation compared to x-rays, including cost and limited availability of alpha-emitting radiopharmaceuticals or synchrotrons for helium irradiation. To increase the validity, reproducibility, and rigor of preclinical alpha irradiation research, any system used to deliver alpha particles requires extensive dosimetric characterization and validation to establish, interpret, and apply the data being generated. With the increase in the options and ability to deliver alpha irradiation in the clinic (and the intense interest in combination therapy with ICI), there is a critical need to understand the unique radiobiology of alpha and high-LET irradiation in general.

Here we present the construction, and dosimetric characterization of an in vitro benchtop alpha irradiator using an ²⁴¹Am foil source. We describe the housing and shielding apparatus, collimator, and irradiation procedures, in addition to characterizing the energy, dose rates, and field homogeneity of the alpha particles being delivered. This system is relatively inexpensive, easy to use, and could be employed by any investigators interested in studying alpha particle radiobiology.

Materials and Methods

²⁴¹Am source initial design and characterization

The source selected (Model A-001, NRD, LLC, Grand Island, NY) was a 50 mm × 50 mm square foil (0.184 mm total thickness) consisting of layers of gold plating (0.508 μm), gold foil (10.16 μm), Americium-241 (²⁴¹Am; 5.08 μm,), gold (7.62 μm), silver (101.6 μm – 177.8 μm), and flash plating for identification of the active surface (Figure 1A). The ²⁴¹Am source has a half-life of approximately 432 years and had a total source activity of 3.3785 mCi, or 0.1250045 GBq, at the date of purchase. To remove large angle alpha particles emitted by the source that would otherwise reach the cells, a 52 × 52 × 3.12 mm collimator was 3D printed from MicroFine Green^™ (ABS-Like Micro Resolution) with a recess to fit the source foil and a 2 mm thick microcapillary “honeycomb” array for collimation (Figure 1B). Mylar sheets (3.0 μm thick) were placed on either side of the source to decrease the risk of contamination from scratches to the source or galvanic corrosion between the source and the housing (Chemplex, Cat. No. 3010). Housing and shielding for the source were constructed using stainless steel with the source housed in the roof of the unit with the active side of the source facing down. A shutter is used to shield the cells until irradiation commences. Cells are grown on glass slides and placed in a holder in a shelf which can be placed at various distances from the source (9 slots total, with Slot 1 having a source-to-target distance of 4 mm and subsequent slots at 5 mm intervals) (Figure 1C).

Figure 1. Source, housing, and collimator of in vitro ²⁴¹Am alpha irradiation platform.

A) Cross section of the 50 × 50 mm ²⁴¹Am foil (NRD Model A-001, 0.184 mm total thickness) consisting of layers of (1) gold plating (0.508 μm), (2) gold (10.16 μm), (3) ²⁴¹Am source (5.08 μm) (activity 3.3785 mCi at date of purchase), (4) gold (7.62 μm), (5) silver (101.6 μm – 177.8 μm), and (6) flash plating for identification of the active side. B) A collimator was 3D printed from MicroFine Green^™ (ABS-Like Micro Resolution) in a microcapillary “honeycomb” array to increase the alpha particle energy homogeneity at the target. C) Shielding and housing of the source was constructed from stainless steel. The source is placed in between 3.0 μm mylar sheets in either the collimator, or a non-collimated holder, and secured to the “roof” of the housing. Cells are grown on a glass slide which is held securely on a shelf which is inserted into slots at various distances from the source (Slots 1–9 with source to target distances of 4 – 44 mm at 5 mm intervals). A shutter blocks alpha particles from the target until the start of irradiation.

Initial measurements of the ²⁴¹Am source were performed to evaluate the homogeneity of the plated activity and assess its suitability for measurement in a larger-area parallel plate ionization chamber. For these studies, unlaminated Gafchromic EBT3 film was used, with measurements acquired at four source-to-film distances: directly on top of the “honeycomb” array and at the first three downstream slots corresponding to nominal separations of 4, 9, and 14 mm from the source (Slots 1–3, respectively). The film was cut with a laser cutter (Boss Laser, Sanford, FL), and all handling procedures followed recommendations from AAPM Task Group 235¹⁵. Prior to and following irradiation, the films were scanned on a flatbed scanner (Epson 10000XL), and image analysis was performed using a custom Python script.

A film-based analysis was used to assess the spatial uniformity and effective dimensions of the plated ²⁴¹Am source, as well as whether the apparent source size changed as a function of distance from the source (Figure 2). Scanned film images were converted to optical density (OD), and a square region of interest (ROI) centered on the irradiated area was defined (Figure 2A), with the horizontal lines indicating the locations of the band profiles used for analysis. The ROI encompassed the full irradiated region of the source, spanning approximately 500 pixels in each direction (≈42–43 mm at 300 dpi), and therefore covered nearly the entire active length of the plated source. Within this ROI, multiple horizontal band locations were selected, and band profiles were generated (Figure 2B) by averaging adjacent rows to reduce local noise while preserving spatial trends. These profiles were used to calculate effective source lengths in the horizontal and vertical directions at various distances from the source (Figure 2C), where edge transitions corresponding to the geometric boundaries of the plated activity were analyzed via full width at half maximum (FWHM) to determine effective source dimensions, using the known scanner pixel size for spatial calibration.

Figure 2. Dose homogeneity and effective source lengths.

The field homogeneity and effective dimensions of the ²⁴¹Am source foil with 3D printed collimator was assessed at multiple distances from the source corresponding to positions where cells will be irradiated (slots 1–3; 4, 9, and 14 mm from source to target, respectively) and on the array using Gafchromic unlaminated EBT3 film. A) Representative two-dimensional optical density (OD) distribution within the region of interest (ROI) selected for film analysis. B) Representative horizontal band profiles extracted from the ROIs. C) Extracted effective source lengths derived from horizontal and vertical band analyses on the array, and at slots 1–3. Error bars are standard deviations (SD). For each slot, the reported SD was calculated as the sample SD of repeated independent measurements acquired at that slot, using all replicate horizontal (or vertical) length measurements to quantify measurement variability.

Attenuation measurements of the ²⁴¹Am source

To quantify alpha-particle energy loss due to attenuation in the source assembly and intervening materials, a custom vacuum alpha-spectrometry chamber was designed. The chamber was required to (1) accommodate a large-area source and (2) support measurements in both air and vacuum. A schematic of the vacuum chamber (Figure 3A) which was used for all alpha spectrometry measurements of the ²⁴¹Am source is provided.

Figure 3. Alpha particle dosimeter.

A) Schematic of a large vacuum chamber used to contain the Passivated Implanted Planar Silicon (PIPS) detector and planar ²⁴¹Am source. B) Perspective drawing of printed circuit boards for the alpha ionization chamber. Window frame supports aluminized mylar roughly 1.6 mm away from collector. Epoxy is applied to the frame surrounding the rectangular opening with a stencil, and the mylar is attached such that the conductive side faces the collector and makes electrical contact with the window electrode.

A custom insert was fabricated to ensure compatibility with the source holder slots. The alpha spectrometer was energy-calibrated using a multi-nuclide calibration source (Eckert & Ziegler) containing ²³⁹Pu, ²⁴¹Am, and ²⁴⁴Cm, with primary alpha emissions of 5.157 MeV at 70.8(1)%, 5.486 MeV at 84.5(1)%, and 5.805 MeV at 76.7(4)% intensity, respectively as reported by the Decay Data Evaluation Project (DDEP)^16–18. To determine the thickness of the mylar window used in the construction of the ionization chamber, an additional spectrum of the calibration source was acquired under vacuum with the mylar window positioned between the source and the detector. Shifts in the measured alpha-particle peak energies were attributed to energy loss in the mylar. These measured energy losses, together with interpolated stopping power data from the NIST ASTAR database, were used to calculate the effective mylar thickness¹⁹. ASTAR data were used for this interpolation rather than ICRU 90, as the latter does not provide updated stopping power recommendations for alpha particles in mylar²⁰.

Measurements of the ²⁴¹Am source energy spectrum were subsequently performed with the alpha spectrometer mounted in the vacuum chamber. Spectra were acquired in air rather than under vacuum to characterize the alpha particle energy distribution incident at the entrance window of the ionization chamber, incorporating attenuation effects from both the source assemble and the intervening air path. Because the ionization chamber and alpha spectrometer were separated by a known distance, the measured energy distributions in air were used as input source models for Monte Carlo (MC) simulations.

For all ²⁴¹Am measurements reported in this work, spectra were acquired at Slots 3 and 4 with source-to-detector distances of 14 and 19 mm, respectively. These downstream slots were selected based on preliminary ionization chamber measurements performed across all positions that produced measurable signal, which demonstrated that Slots 3 and 4 yielded the highest collected charge and signal-to-noise ratio. These positions therefore provided the most robust conditions for subsequent alpha spectrometry and absorbed-dose measurements while maintaining sufficient alpha-particle fluence for spectral characterization.

Design of an ionization chamber for absorbed dose to water measurements

A custom parallel-plate ionization chamber was developed to enable direct physical measurements of absorbed dose from alpha particle ionization in air. The design was adapted from prior work on windowed extrapolation chambers for alpha emitting radionuclides, with the key distinction that a fixed air gap was maintained between the entrance window and the collecting electrode rather than a variable electrode separation^21,22.

Two printed circuit boards (PCBs) were fabricated for the chamber. The first PCB supported the collecting electrode, which was designed as a rounded-edge rectangle with nominal dimensions of 38.1(1) mm × 63.5(1) mm and corner radii of 2.54 mm. A 0.2 mm guard isolation gap surrounded the collecting electrode. Accounting for nominal PCB fabrication tolerances of 0.05 mm, the effective collecting area was calculated to be 24.34(7) cm².

The second PCB supported the entrance window, to which a thin aluminized mylar foil was affixed. When assembled, the air gap between the entrance window and the collecting electrode was nominally defined by the thickness of the entrance window PCB, measured as 1.455 mm. Electrostatic distortion of the mylar window was measured for 300-volt bias conditions by measuring the capacitive displacement current. Combined with the effective collecting area, this geometry yielded a sensitive air volume of 3.54 cm³. A schematic of the custom ionization chamber is shown in Figure 3B. In addition to manufacturer specifications, independent micrometer measurements were performed to verify the collecting electrode dimensions and entrance window PCB thickness, confirming agreement within stated tolerances.

A continuously variable, bipolar DC power supply capable of up to ±1000 V was used to bias the entrance window. The electrometer collector, guard, and chassis ground were maintained at the same electrical potential. The chassis grounds of both the electrometer and the bias supply were connected to the facility AC ground.

A second ionization chamber was constructed to serve as a quality assurance (QA) chamber for routine measurements of the ²⁴¹Am source. This chamber was fabricated using the same mechanical design and materials as the reference chamber but without detailed characterization of the sensitive volume. As such, this QA chamber is intended to be analogous to clinical ionization chambers that rely on an empirical calibration or beam quality correction factor rather than a fully characterized collecting volume.

Absorbed dose to water formalism for the ²⁴¹Am source

The absorbed dose to water at the surface of each source holder slot, ${\dot{D}}_{\mathit{water}}$, from the ²⁴¹Am source was determined using the following formalism²³:

$${\dot{D}}_{\mathit{water}}=\frac{{\left(\frac{\bar{W}}{e}\right)}_{\mathit{air}}\cdot {\bar{S}}_{\mathit{air}}^{\mathit{water}}}{{\rho }_0\cdot V_{\mathit{eff}}}\cdot I\cdot \left(k_{\mathit{elec}}\cdot k_{TP}\cdot k_{\mathit{recom}}\cdot k_{\mathit{pol}}\cdot k_{\mathit{attract}}\right)\cdot \left(k_{\mathit{wind}}\cdot k_{\mathit{vol}}\right)$$

where ${\left(\frac{\bar{w}}{e}\right)}_{\mathit{air}}$ is the average energy required to create an ion pair in air, which is given as 34.96 J/C ^[1,2]. ${\bar{S}}_{\mathit{air}}^{\mathit{water}}$ is the stopping power conversion from air to water. ${\rho }_0$ is the density of air at standard temperature and pressure. $V_{\mathit{eff}}$ accounts for the sensitive air volume of measurement. $I$ is the measured ionization current. For all measurements with the ionization chamber, for each slot of measurement three trials of three 60 second charge measurements were integrated.

The correction factors $k_{\mathit{elec}},k_{TP},k_{\mathit{recom}},k_{\mathit{pol}}$, and $k_{\mathit{attract}}$ account for the measurement related effects. Specifically, $k_{\mathit{elec}}$ is the electrometer correction factor given by calibration at an accredited dosimetry calibration laboratory (ADCL), $k_{TP}$ is the temperature and pressure correction factor at the time of measurement, $k_{\mathit{recom}}$ is the recombination correction factor which accounts for both initial and general recombination as defined in prior studies^22,23, $k_{\mathit{pol}}$ is the polarity correction factor, and $k_{\mathit{attract}}$ is the correction for the electrostatic attraction of the entrance window. Lastly, $k_{\mathit{wind}}$ and $k_{\mathit{vol}}$ are the MC-based correction factors, which will be discussed in the subsequent section.

To determine the electrostatic attraction correction factor, capacitance measurements were performed using the constructed ionization chamber. With the bias supply, voltage steps of 10 V were applied at nominal operating voltages of 25, 50, 100, 200, and 300 V. Because the chamber measurements were performed at 300 V, the ratio of the measured capacitance at the lowest and highest applied voltages was used to quantify changes in the effective air gap arising from electrostatic displacement of the mylar entrance window. This ratio was used to derive $k_{\mathit{attract}}$, accounting for the voltage-dependent deformation of the entrance window.

Monte Carlo Simulations

For all MC simulations, the Tool for Particle Simulation (TOPAS), version 3.7 was used²⁴. A custom physics list was used based off previously validated work, containing “g4em-standard_opt4,” “g4h-phy_QGSP_BIC_HP,” “g4ion-binarycascade,” “g4h-elastic_HP,” “g4_stopping,” “g4_decay,” and “tsradioactivedecay”²⁵. All simulations used a production threshold of 0.001 μm to account for the attenuation of the alpha particles within the plated ²⁴¹Am source.

In this study, the detailed microscopic geometry of the ²⁴¹Am source plating was not explicitly modeled in TOPAS. Instead, MC–derived correction factors were calculated as functions of alpha particle energy and subsequently combined with experimentally measured alpha particle spectra in air. The measured in-air alpha energy spectra at each measurement plan were used directly as the input energy distributions for the MC simulations, ensuring that the resulting correction factors reflect the experimentally observed energy degradation of the alpha particles. This approach was motivated by film measurements demonstrating that the effective source dimensions did not increase with increasing distance from the source, indicating that the microcapillary “honeycomb” array produces a directionally constrained, quasi-parallel alpha particle field rather than isotropic emission. Under this condition, the spatial evolution of the beam is dominated by energy loss rather than geometric divergence.

Accordingly, energy dependent correction factors were interpolated using the normalized measured spectral distributions to determine the appropriate values for each measurement plane. This method allows the experimentally observed energy degradation to be incorporated directly into the Monte Carlo framework without requiring explicit geometric modeling of the plated source structure.

The entrance window correction factor, $k_{\mathit{wind}}$, accounts for alpha-particle attenuation and scatter in the aluminized mylar entrance window and was defined as:

$$k_{\mathit{wind}}=\frac{D_{\mathit{air,vol,square}}}{D_{\mathit{mylar,vol,square}}}$$

where $D_{\mathit{air,vol,square}}$ and $D_{\mathit{mylar,vol,square}}$ are the MC mean dose scored in the effective measurement volume with the entrance window set to air and mylar, respectively.

The volume correction factor, $k_{\mathit{vol}}$, adjusts from a volume-averaged measurement to a surface dose prescription at each slot and was defined as:

$$k_{\mathit{vol}}=\frac{D_{\mathit{air,surf,square}}}{D_{\mathit{air,vol,square}}}$$

where $D_{\mathit{air,surf,square}}$ is the mean MC dose scored at the surface within a rectangular scoring region of thickness 10 μm, meant to simulate the size of a cell. These correction factors were applied multiplicatively in the absorbed-dose formalism previously described. Statistical uncertainties for the MC-derived quantities were obtained from the variance of the scored dose tallies.

Because the volume correction factor, $k_{\mathit{vol}}$, adjusts the ionization chamber response to a surface dose corresponding to a small scoring region directly downstream of the alpha spectrometry measurement location, the measured alpha particle energy spectra was also used in the determination of stopping power ratios at each measurement slot.

Energy dependent mass stopping powers for alpha particles in air and water were obtained from ICRU Report 90. The stopping power ratio, ${\bar{S}}_{\mathit{air}}^{\mathit{water}}$, was calculated by interpolating the tabulated stopping power data as a function of energy and weighting the values according to the normalized alpha particle energy distributions extracted from the spectrometry results. The resulting fluence-weighted stopping power ratios were used to convert measured dose to air to absorbed dose to water within the formalism previously discussed. Measured dimensions of the custom parallel plate ionization chamber were verified prior to dose measurements and were found to be consistent with the nominal values specified by the manufacturer and fabrication tolerances previously described. Independent micrometer measurements confirmed agreement for both the effective collecting area and the air gap thickness used to define the sensitive volume. Ionization chamber measurements of the ²⁴¹Am source were performed at slots 3 and 4. For each slot, three independent trials were acquired, with each trial consisting of three consecutive 60 second charge integrations. Polarity and recombination correction factors were evaluated in conjunction with these measurements.

Results

Field Homogeneity

Figure 2 summarizes the evaluation of spatial homogeneity and effective dimensions of the plated ²⁴¹Am source using Gafchromic unlaminated EBT3 film. Figure 2A shows the two-dimensional OD distribution within the selected ROI, along with the locations of the horizontal band profiles used for analysis. The ROI is uniformly irradiated across its full extent, with only localized reductions in OD observed near the lateral boundaries corresponding to the edges of the plated region.

Horizontal band profiles extracted from within the ROI are shown in Figure 2B. Across the central portion of the irradiated area, the profiles exhibit consistent OD values with no systematic spatial trends, indicating a uniform irradiation field. Sharp OD falloff is observed only at the periphery of the plated region and is consistent with geometric truncation rather than variations in plated activity. The close agreement among profiles sampled throughout the ROI demonstrates that the apparent source uniformity is maintained across the measurement region and is insensitive to small variations in sampling location or source to detector separation.

The effective source dimensions derived from horizontal and vertical band analyses are summarized in Figure 2C. The horizontal extent of the source is highly consistent across all measurement configurations, with an average length of approximately 45.4 mm and small associated uncertainties. The vertical extent exhibits modestly greater variability, though still submillimeter: however, no systematic increase in the effective source dimensions is observed with increasing distance from the source. This absence of geometric broadening indicates that the microcapillary “honeycomb” array produces a directionally constrained, quasi-parallel alpha-particle field rather than the lateral spreading expected from isotropic emission. The observed variability in the vertical dimension is therefore attributed to experimental positioning and irradiation geometry rather than non-uniformity in the plated activity.

Collectively, these results confirm that the plated ²⁴¹Am source, when used in conjunction with the microcapillary array, provides a spatially uniform and geometrically stable irradiation field suitable for subsequent absorbed dose measurements.

Alpha spectrometry

Figure 4 shows alpha spectra acquired using a multi-nuclide calibration source both with and without the mylar window positioned between the source and detector. Introduction of the mylar window resulted in systematic shifts of all three prominent alpha peaks toward lower energies, consistent with energy loss in the window material. These measured peak shifts were used to determine the effective mylar thickness by comparison with interpolated stopping power data from NIST ASTAR. This analysis yielded an effective mylar thickness of 2.98(8) μm compared to nominal thickness of 3.0 μm.

Figure 4. Alpha particle energy spectra with and without mylar.

Alpha spectra acquired using a multi-nuclide calibration source both with and without an aluminized mylar window positioned between the source and detector.

Measured alpha spectra for the plated ²⁴¹Am source acquired in air are shown in Figure 5 for the two measurements slots used in this study. A systematic shift in the peak energy between measurement slots is observed, reflecting differences in air attenuation between the source and detector. Aside from this expected energy degradation, the spectral shapes remain consistent, indicating stable alpha emission characteristics across configurations. These measured ²⁴¹Am spectra in air were discretized into energy bins and used to define the input energy distributions for MC simulations. The resulting spectra served as the basis for calculating ionization chamber correction factors and for generating phase space files at the entrance and exit of the sensitive air volume used in stopping-power ratio determinations.

Figure 5. Alpha particle energy spectra at various source-to-target distances.

Alpha spectra acquired in ²⁴¹Am housing with 3D printed microcapillary collimator and mylar sheet in place. Energy spectrum was measured at two distances from the source at which the cell targets will be located (3^rd and 4^th slots; 14 and 19 mm from the source, respectively).

MC correction factors and stopping power ratio

MC simulations were used to quantify correction factors for entrance window attenuation and for converting the chamber’s volume-averaged response to a surface dose at each measurement slot.

For Slot 3 (14 mm source to target), the fluence weighted transport power ratio at the entrance was 1.1572. The entrance window correction factor, $k_{\mathit{wind}}$, was 0.833 with a MC statistical uncertainty of 0.08%. The volume correction factor, $k_{\mathit{vol}}$, was 0.892 (0.1% statistical uncertainty), reflecting the expected adjustment from a volume averaged dose in the collecting air gap to the prescribed surface dose at the slot.

For Slot 4 (19 mm source to target), the stopping power ratio exhibited a larger energy-loss-driven change across the sensitive volume, where it was 1.1557. The entrance window correction factor increased to 0.899 (0.08%). The volume correction factor was 0.898 (0.1%), indicating a smaller correction from volume averaged to surface dose compared with Slot 3.

Collectively, these results show that both the window correction and the fluence weighted stopping power ratio are sensitive to the measured in air alpha energy distribution at each measurement plane, while the volume correction reflects the near-surface dose gradient and finite sensitive volume averaging in the parallel plate chamber geometry.

Absorbed dose rate

The polarity correction factor was measured once per plane and was identical for both planes $\left(k_{\mathit{pol}}=0.9983\right)$. The recombination correction factor was determined to be $k_{\mathit{recom}}=1.010$, and the electrometer calibration correction was $k_{\mathit{elec}}=0.999$. Temperature and pressure correction factors were applied on a per-trial basis and ranged from 1.035 to 1.037 over the course of the measurements. The electrostatic window attraction correction factor, evaluated at an applied bias of 300 V, was $k_{\mathit{attract}}=0.993$.

Corrected charge measurements were converted to absorbed dose to water using the formalism previously described, incorporating the effective sensitive air volume, air density at standard temperature and pressure, the mean energy per ion pair in air, and MC–derived correction factors for entrance window attenuation and volume averaging. The absorbed dose rate to water was determined to be 5.815 mGy/s at slot 3 and 7.721 mGy/s at slot 4 (14 and 19 mm from source, respectively).

Discussion

A major challenge in delivering alpha irradiation in vitro is the limited range of alpha particles in aqueous media. Prior researchers have addressed this by growing cells on mylar, Kapton or other materials thin enough to permit alpha particles to traverse the surface and irradiate cells from below^26–28. While effective, these methods present potential challenges related to cell adhesion, altered growth kinetics, and logistical complexity for subsequent assays. Our platform overcomes these limitations by transiently removing culture media from cells grown on standard glass slides, irradiating the monolayer from above, and immediately restoring media following irradiation, thereby enabling alpha-particle exposure while preserving conventional cell culture conditions and assay compatibility.

Current dosimetry methods for alpha particles rely predominantly on computational techniques, which necessarily incorporate several assumptions, including negligible self-attenuation and dependence on cross-section data such as stopping power. Prior studies demonstrated that uncertainties in these cross-section datasets can approach ~2.5% and that absorbed dose calculations are highly sensitive to the choice of physics models and transport settings, introducing additional systematic uncertainty^23,25. As a result, experimental dosimetry offers a more direct approach to absorbed dose determination, providing increased confidence in dosimetry without complete reliance on calculation algorithms or underlying data libraries. Assuming negligible fluence perturbation within the air cavity, the absorbed dose measurements reported here are representative of surface or shallow dose deposition. For in vitro experiments in which alpha particles traverse appreciable material thickness before reaching the cell nucleus, depth-dependent correction factors may be required. Future work will therefore investigate the implications of the steep dose gradients characteristic of alpha-particle irradiation and their impact on detector response, cell survival, and other biological readouts.

An increase in absorbed dose rate to water was observed at slot 4 (19 mm; 7.721 mGy/s) compared to slot 3 (14 mm; 5.815 mGy/s) despite the larger source-to-detector distance. While a decrease in fluence with increasing distance would be typically expected from an isotropically-emitting source, the Am-241 source is monodirectional due to the presence of the micro-capillary collimator occluding alphas traversing at an angle. Therefore, the energy deposition is dominated by stopping power, which increases as a function of decreasing energy (or increasing media thickness). As alpha particles traverse air, they lose energy and approach the Bragg peak, and alpha spectrometry measurements confirm a shift toward lower energies at more distal planes, resulting in increased energy deposition per unit path length. Under these conditions, the increased stopping power outweighs the reduction in particle fluence, leading to higher ionization within the sensitive volume, consistent with the MC–derived stopping-power ratios and volume correction factors.

For the alpha irradiation delivered clinically there can be a wide variety of mean absorbed dose-rates depending on a variety of factors. For example, external beam helium ions would typically deliver average dose rates of 0.15 – 0.20 Gy/s^29–31, but this can vary depending on the number of beams used, time in between beam set ups, respiratory motion management, the LET/target volume/energy/scanning beam path utilized. For DaRT implants dose rates can vary from 0.75 – 0.05 grays (Gy) per minute, but the dose rates vary widely over the course of the implant with the half-life of Ra²²⁴ being on the order of 3.63 days, and the implant delivering 90% of the dose over the course of 12–13 days. Dose rates with DaRT can also be influenced by tumor vasculature (due to diffusion and clearance of the daughter radionuclides), and the dose rate to a specific portion of the tumor varies based on distance from the DaRT seed(s)^32–35. Radiopharmaceuticals delivering alpha irradiation may require more advanced microdosimetric techniques to characterize the dose rates delivered clinically, making delivered dose-rates difficult to estimate - even prior to biologic factors such as tumor vasculature and clearance of the radiopharmaceutical^36–38. For example, in a study of Xofigo ([223Ra]RaCl) alpha irradiation in prostate cancer, the absorbed dose varied within the same lesion by orders of magnitude (0.0265 – 0.0001 Gy/min)³⁹. The addition of targeting ligands could add variation based of the biology of the patients tumor (i.e. variable expression of the target between tumor cells). Ultimately, dose rate may be less important for tumor efficacy with high-LET radiation than it is for X-rays^40–42.

Experimental dosimetry for alpha particles remains relatively underexplored in the context of radiopharmaceutical therapy. A key distinction between the present work and radiopharmaceutical alpha dosimetry lies in the directionality of particle incidence. In this study, alpha particles predominantly exhibit unidirectional transport toward the cell. Under such conditions, microdosimetric considerations may be less critical; however, this assumption will be rigorously evaluated in future computational microdosimetry studies.

Conclusions

We established a low-cost in vitro alpha irradiation platform suitable for high-throughput preclinical studies. Irradiation homogeneity, alpha particle energy spectra, absorbed dose, and dose rate were rigorously characterized and validated through direct measurement. Given the approximate 432-year half-life of ²⁴¹Am, this platform will provide a stable and usable dose rate for decades. This platform, and others developed using similar principles, may reduce logistical barriers to high-LET alpha-particle radiobiology research and facilitate broader adoption by investigators interested in preclinical and translational studies.

Data Availability:

All data relevant to this publication is presented in the manuscript or is stored in an institutional repository and will be shared upon request to the corresponding author(s).