Characterization of selected additive manufacturing materials for synchrotron monochromatic imaging and broad-beam radiotherapy at the Australian synchrotron-imaging and medical beamline

Objective. This study aims to characterize radiological properties of selected additive manufacturing (AM) materials utilizing both material extrusion and vat photopolymerization technologies. Monochromatic synchrotron x-ray images and synchrotron treatment beam dosimetry were acquired at the hutch 3B and 2B of the Australian Synchrotron-Imaging and Medical Beamline. Approach. Eight energies from 30 keV up to 65 keV were used to acquire the attenuation coefficients of the AM materials. Comparison of theoretical, and experimental attenuation data of AM materials and standard solid water for MV linac was performed. Broad-beam dosimetry experiment through attenuated dose measurement and a Geant4 Monte Carlo simulation were done for the studied materials to investigate its attenuation properties specific for a 4 tesla wiggler field with varying synchrotron radiation beam qualities. Main results. Polylactic acid (PLA) plus matches attenuation coefficients of both soft tissue and brain tissue, while acrylonitrile butadiene styrene, Acrylonitrile styrene acrylate, and Draft resin have close equivalence to adipose tissue. Lastly, PLA, co-polyester plus, thermoplastic polyurethane, and White resins are promising substitute materials for breast tissue. For broad-beam experiment and simulation, many of the studied materials were able to simulate RMI457 Solid Water and bolus within ±10% for the three synchrotron beam qualities. These results are useful in fabricating phantoms for synchrotron and other related medical radiation applications such as orthovoltage treatments. Significance and conclusion. These 3D printing materials were studied as potential substitutes for selected tissues such as breast tissue, adipose tissue, soft-tissue, and brain tissue useful in fabricating 3D printed phantoms for synchrotron imaging, therapy, and orthovoltage applications. Fabricating customizable heterogeneous anthropomorphic phantoms (e.g. breast, head, thorax) and pre-clinical animal phantoms (e.g. rodents, canine) for synchrotron imaging and radiotherapy using AM can be done based on the results of this study.


Introduction
Synchrotron can produce ultra-high dose rate x-rays that can be used in cancer treatment with small beam fields.It has the capability as well of producing high resolution images considering its phase contrast effects in x-ray imaging.The Australian Synchrotron (AS)-Imaging and Medical Beamline (IMBL) provides researchers the capability to image at micron-scale to see small difference in the anatomy or object structure.
In addition, material specific quantities such as attenuation coefficients are possible to measure using its monochromatic computed tomography (CT) capability.Due to its ability to deliver small field of high radiation doses relative to conventional Linac radiotherapy, it can treat small tumours using synchrotron x-rays through a broad-beam mode or through spatial fractionation technique called microbeam radiation therapy (MRT) configuration.Synchrotron radiotherapy is useful in treating small tumour targets like brain tumour given its small beam divergence, broad spectrum (range of 50 keV and 600 keV), and high intensity.In addition, it has a radiobiological advantage for MRT called the dose volume effect as the normal tissue tolerance increases with small radiation fields (Bräuer-Krisch et al 2010, Cameron et al 2017).
The success of radiation therapy highly depends on the quality assurance (QA) done to characterize the beam condition before doing any treatment procedures.Water phantom is a standard equipment utilized for QA as the human body is mainly composed of water.However, it is difficult to set up for daily QA procedure.Thus, solid phantoms are commonly preferred because they are easier to use.As of writing, there is still no standard QA equipment used in synchrotron imaging and radiotherapy.Plastic phantoms commonly used in conventional imaging and radiotherapy (MV photon and MeV electron) are utilized in beam measurements and research.These commercial phantom materials are generally specified for high energy (MV and MeV) radiotherapy and may not be always applicable for low energy treatment beams specifically for synchrotron x-ray beam quality (Cameron et al 2017).In addition, these standard phantoms are manufactured in simple shapes, and do not conform with the actual patient or animal anatomy for pre-clinical experiments.This can result in wrong assumptions and can affect the whole QA process particularly with specialized treatment modalities such as MRT.Thus, there is a need to characterize for the specific beam qualities of MRT with energies around keV.
Aside from radiotherapy capability, the AS-IMBL is used for high-resolution x-ray imaging that makes it possible to assess internal structure of samples (Suortti andThomlinson 2003, Arhatari et al 2021).The role of imaging in treatment guidance is also possible to implement in improving the accuracy of pre-clinical treatment of rats (Paino et al 2021).CT is a 3D imaging technique that has been utilized in clinical setups for radiotherapy planning and diagnostic procedure (Khan and Gibbons 2014).However, the resolution of medical CT machines might not be enough to resolve the porosity and internal defects inside an object (Sinico et al 2021).Moreover, clinical CT number in Hounsfield Unit (HU) is specific for the x-ray spectrum of the machine.Hence, it is challenging to acquire accurate material specific quantities like the attenuation coefficient without using dual-energy CT technique (Fonseca et al 2023).The AS-IMBL is aimed to acquire superior resolution of CT images using monochromatic and polychromatic x-ray beam useful in breast imaging where tissue imaging contrast is highly important.
Three-dimensional (3D) printing or additive manufacturing (AM) is an emerging technology in fabricating objects based on 3D computer-aided designs.There are several research done for AM applications in conventional radiotherapy applications such as phantoms, immobilization device, and bolus (Bustillo et al 2019, Tino et al 2019, Asfia et al 2020).However, limited studies are only available for applications of AM in synchrotron-based imaging and therapy.
There are published studies on the investigation of experimental attenuation coefficient of AM materials using monochromatic synchrotron x-rays to fabricate a physical breast imaging phantom (Ivanov et al 2018, Mettivier et al 2022).However, these data are still limited to selected AM materials and focused on selected energies for breast imaging application.Also, it is interesting to compare the results got from the AS with the data got from other synchrotrons by other studies.
The design and fabrication of synchrotron x-ray dosimetry phantoms remain a research prospect for this advanced radiation modality.Hence, AM or 3D printing can be utilized to create basic to complex geometries of phantoms and other dosimetry equipment.This study is a significant step in prototyping 3D printed synchrotron dosimetry equipment such as phantoms, boluses, and treatment support.Furthermore, the data presented in this study will be useful in planning the use of biofabricated materials for medical radiation physics applications that would be needing 3D printing support.
This study aims to characterize selected AM materials for synchrotron imaging and radiotherapy.Broad-beam accumulated dose measurements and monochromatic CT scanning were utilized to characterize all the 3D printing materials using synchrotron radiation.International Commission on Radiation Units and Measurements (ICRU) published attenuation databases and by the National Institute of Standards and Technology (NIST) were used to compare with experimental attenuation based on monochromatic CT.Experimental attenuation of AM materials and reference materials were also compared as both RMI457 Solid Water and bolus are used in experiments done at the IMBL.Finally, the measured broad-beam dosimetry data were compared with the simulated dose profiles acquired using an established Geant4 Monte Carlo simulation for AS-IMBL.

NIST attenuation coefficients of studied and reference materials
The materials investigated in this study are divided into four categories: (a) polylactic acid (PLA)-based filaments (4 filaments), (b) co-polyester filaments (3 filaments), (c) other material extrusion (MEX) filaments (5 filaments), and (d) photopolymer resins (3 resins).In addition, two reference materials used in conventional radiotherapy were used: bolus, and RMI457 Solid Water.A total of 15 AM and 2 reference materials were used.Technical details about the AM materials' main polymer composition, and manufacturer are provided in the appendix (table A1).
Table 1 shows the assumed elemental composition by weight fraction of all the studied materials.Materials used in radiotherapy QA (water-equivalent plastic, and bolus) are utilized as reference.RMI457 Solid Water was used as it was reported to have best match with water for synchrotron radiation dosimetry at the AS-IMBL (Cameron et al 2017).
Mass attenuation coefficient ( µ / ρ ) of a material for a given energy of photons or uncharged particles is defined by equation ( 1) where dN/N refers to the average fraction of particles interacting in a material thickness of dl with density ρ, and µ is the linear attenuation coefficient (International Commission On Radiation Units And Measurements (ICRU) 2011).While the transmitted intensity, I, in monochromatic irradiation is expressed as the integral of linear attenuation coefficient through the material thickness (x) in the direction of the x-ray beam as shown in equation ( 2), where I o is the initial intensity.Theoretically, it is possible to determine µ using this equation, In this study, the theoretical mass attenuation coefficient of each material was calculated using the online database of the NIST (www.nist.gov).This may underestimate the actual mass attenuation coefficient given that it is a composite material, but it would be sufficient for this study (Cameron et al 2017).Elemental composition of each material was determined based on the manufacturer's datasheet, published articles, and available databases (Alssabbagh et al 2017).Data for reference materials such as soft tissue, brain tissue, adipose tissue, and breast tissue were referenced from NIST database (Berger et al 2010).This is needed to calculate the linear attenuation coefficient for a specific x-ray beam energy.Note that the elemental composition of some materials is based on its major polymer component due to the unavailability of the exact compositional details.It is worth noting that exact material characteristics may differ depending on many manufacturing aspects.
Attenuation coefficient is the focus of this study as the CT number (in Hounsfield Units, HUs) has been reported by previous studies related to AM technologies (Bustillo et al 2019, Tino et al 2019).In addition, CT number is a CT machine dependent quantity.Synchrotron monochromatic CT imaging is capable of directly measuring attenuation coefficients which can be compared to theoretical values.
Twelve materials are based on the material extrusion printing technology (also called fused deposition modeling, FDM), while three materials are used with vat-photopolymerization technology (also known as stereolithography, SLA).All FDM materials were printed using Ultimaker S5 and Creality CR-10 printers with 100% slicer infill setting.Note that this setting will still produce microscopic air gaps due to the printing resolution limitations of this technology such as the nozzle size.While SLA produces more homogeneous printouts using Formlabs Form 2 with 0.1 mm of layer height slicer setting.
These printing technologies are selected as they have wide applications for example in prototyping, dental medicine, robotics, and aerospace.In addition, these materials are considered in fabricating phantoms, and dosimetry support structures in both synchrotron imaging and radiotherapy pre-clinical experiments.These include experiments involving animals (e.g.rodents, canine) requiring customizable dosimetry support that are commonly not available commercially.

Monochromatic CT imaging
The AS, as shown in figure 1, has an electron storage ring controlled at the nominal 3 GeV maintaining a ring current of 200 mA and operating in top-up mode (regularly injecting current).AS-IMBL has a superconducting multipole wiggler (SCMPW) source capable of producing radiation energies of more than 100 keV.The x-ray will then pass through a set of in-vacuum filters (i.e.graphite, aluminium, copper, molybdenum).Selecting monochromatic x-ray energy is done at the Hutch 1 A using a dual-crystal Laue monochromator (DCLM) by mechanically bending the Si crystal relative to the second crystal following the Bragg equation expressed in equation ( 3) where n is a whole number, λ is the wavelength, d is the distance between Si crystal planes, and θ refers to the angle between incident x-ray beam and crystal plane (Midgley et al 2019),  The experimentally acquired attenuation values at the beamline have an excellent agreement with NIST values above 70 keV.However, discrepancies have been reported in the lower-energy data due to harmonic contamination (Stevenson et al 2017).It was estimated experimentally that the third harmonic (n = 3) contamination for 35 keV and 60 keV are 1.0% and 1.5% relative to principal beam energy (n = 1), respectively (Midgley and Schleich 2015).Further investigations were done in a previous copper attenuation measurement experiment wherein the contribution of third harmonic using a RUBY image receptor is 0.4% for both 30 keV and 35 keV, and 0.6% for 40 keV (Midgley et al 2019).Thus, these third-harmonic contamination measurements are small systematic error in the chosen energy range and will be sufficient for this study.The harmonic contribution decreases with higher beam energies given the wider gap between principal and harmonic energies.
The monochromatic CT imaging was done in hutch 3B located around 140 m from the source.Four 3D printed rectangular test objects were placed in stack on a motorized sample table as shown in figure 2 RMI457 Solid Water and a radiotherapy bolus served as standard materials for comparison.A portion of each homogeneous test object was scanned using a monochromatic x-ray source.The acquired CT images were used to assess the materials' attenuation coefficient as the printed test object is uniform in structure.The test objects were approximately 40 cm from the RUBY detector while approximately 137 m away from the x-ray source.Before the image acquisition, the rocking angle ∆θ of the monochromator was adjusted based on the detected incident x-ray flux of the ion chamber to get the optimal incident photon fluence.
Before taking any projections of the phantom materials, 99 frames of dark field images (I dark ) were acquired without radiation by closing the beam shutter then averaging the signal of the detector.In addition, 99 frames of flat field background images (I flat ) were acquired without the sample material along the beam direction by averaging the detector transmitted intensity signals.Acquisition of both dark and flat fields were also done after each CT scan because the beam profile may vary as a function of time.Their weighted average was used to correct the material intensity readings (I) considering the non-uniform signal of the detector system as shown in equation ( 4) (Ivanov et al 2018).The detector was used under its saturation condition (65 535 for RUBY detector) by operating on the shoulder of the Bragg peak.Post image processing was done to reduce noise contribution, Acquired projection data were preprocessed using dark-field and flat-field images.Geometrical and ring artefact corrections were also done using the X-TRACT software package (Gureyev et al 2011) and the recommended image processing setup by beamline scientists.The reconstructed image produced the linear attenuation coefficient values (µ) in cm −1 based on the transmitted synchrotron x-ray readings.Analysis of the attenuation coefficient was accomplished using ImageJ (National Institute of Health, Bethesda, Maryland, USA) by obtaining mean attenuation coefficient value from the ROIs placed on the homogeneous part of the image slices.The upper and lower part of the image were avoided as it may contain reduced vertical beam intensity as suggested by Midgley (Midgley et al 2019).In addition, 3D printed objects fabricated using MEX or FDM are known to have pores or air gaps due to the limitation on its printing resolution (Bustillo et al 2019, Ma et al 2021).These air gaps were avoided in acquiring ROIs as these can be easily resolved due to the micron resolution of the synchrotron CT imaging.These AM experimental attenuation values (sample, s) were compared relative to the NIST attenuation coefficients and reference materials (r) by calculating the per cent differences expressed in equation ( 5), (5)

Broad-beam dosimetry experiment and simulation
The broad-beam dosimetry experiment was done at the hutch 2B of the AS-IMBL to acquire attenuated point dose data behind the printed sample.The synchrotron x-ray spectrum has an energy range around 40 keV to 250 keV.The spectrum peak and mean energies can be varied using a set of filter paddles to harden the beam.While the intensity can be modified by varying the SCMPW magnetic field strength (Paino et al 2021).For this experiment, the SCMPW was set to operate in 4.0 T and three set of filters were used: AlAl, MoMo, and CuCu.This is to characterize the materials using synchrotron x-ray spectra used in radiotherapy dosimetry experiments.Three-dimensional (3D) printed rectangular prism test objects with a size of (40 mm × 20 mm × 15 mm) were fabricated using material extrusion and vat photopolymerization printing techniques.Commonly used AM materials were considered in this study.This will help synchrotron scientists in deciding which material to use for their purpose as these materials are readily available commercially.
The Dynamic MRT (DynMRT) stage was used to place a set of 10 cm × 10 cm RMI457 Solid Water slabs with a total thickness of 8.5 cm.A slab dosimeter insert for a microdiamond detector (PTW-Freiburg Ref TN60019) was used to place the centre of the detector at 5 mm behind the 3D printed sample.The 3D printed test object was inserted at the middle of a 15 mm thick bolus material facing the x-ray beam direction as shown in figure 3 to hold the test object along the beam direction.Reference radiotherapy materials (RMI457 Solid Water, and bolus) were used in the beam transmittance measurement.A beam defining aperture of 0.532 mm and a mask of size of 10 mm (horizontal) × 20 mm (vertical) were used to have enough beam field size relative to the size of test object.
A Geant4 Monte Carlo radiation transport model of the AS-IMBL developed by previous beamline researchers was implemented using the simulated energy spectrum of 4T wiggler magnetic field for a given set of filters used, material specifications, and experimental setup (Dipuglia et al 2019).This code has been validated experimentally for both broad-beam and microbeam irradiation setup.A broad-beam irradiation configuration was implemented for this study following the setup used in the actual experiment.Phase space of particles from the three sets of filters were generated and stored to use as incident radiation field in studying different materials.Per cent dose difference in attenuated dose (D), shown in equation ( 6), acquired in simulation and in experiment for AM sample (s) and reference material (r) were compared.Uncertainties of about 10% has been considered acceptable during the preclinical stage of synchrotron radiation due to the intrinsic challenges of synchrotron dosimetry of small radiation fields as discussed by previous studies (Cameron et al 2017).These data will give additional insights regarding the difference between the actual experimental implementation and the ideal scenario in a Monte Carlo simulation,

Monochromatic CT imaging
Monochromatic imaging capabilities of the AS allows the possibility of comparing theoretical attenuation coefficients with its experimental counterparts.This is a good method in characterizing radiological properties as a function of synchrotron x-ray energy.This section shows the comparison of NIST calculated and experimental attenuation coefficients of various materials as a function of energy.All the numerical attenuation data are summarized in the appendix (table A2).The differences between the experimental and the NIST calculated attenuation coefficients are summarized in figure 4. The experimental measurement standard deviations are small and not reflected in the figure.Note that NIST attenuation coefficient values are based on assumed compositions as these details are proprietary.
The differences in attenuation coefficients of PLA-based polymers acquired from monochromatic CT images relative to NIST values are presented in figure 4(A).Not all PLA filaments have similar attenuation properties as seen in the acquired attenuation values consistent with the differences in their material specifications.This has been reported as well in previous studies using dual energy CT imaging (Ma et al 2021).PLA filaments were observed to have a maximum per cent difference of 16.30% for PLA (Creality) under 40 keV.This specific PLA filament has above 10% difference except for 50 keV.One reason for this is the variation in the PLA synthesis and presence of polymer additives that affects its actual composition.In addition, a pure PLA composition was assumed by using the chemical formula of the repeating monomer of PLA for NIST calculation.PLA (PolyLite) and eSilk PLA (eSun) seem to have similar trend where they have above 10% difference for 30 keV to 40 keV while below 10% for higher energies.PLA+ (eSun) appears to have a very good accuracy relative to the NIST calculated attenuations for this energy range.
Figure 4(B) illustrates that CPE has good similarity for 30 keV to 35 keV, and 50 keV to 65 keV.While CPE+ matches NIST for 50 keV to 65 keV.All the studied materials under this group have per cent differences below 13% relative to the NIST attenuations.PETG has a good match (<10% difference) with the NIST attenuation except for 45 keV.
Other MEX based filaments were grouped together which include materials with different composition compared to PLA and CPE filaments.The differences in the attenuation coefficients of calculated and experimental are presented in figure 4(C).This illustrates that ABS filament has a very good match with its NIST calculated coefficient with differences below 5%.While the rest of the filaments seem to have varying response per energy.ASA matches the NIST for the 30 keV, 50 keV, 55 keV, and 65 keV, PP and TPU for 50 keV to 65 keV, and nylon for 30 keV and 45 keV to 65 keV.This shows that radiological response varies as a function of x-ray energy.
Lastly, all the resins used have good match with their NIST counterparts below 10% difference for all the energies.Figure 4(D) shows the differences between the calculated and experimental coefficients.The more homogeneous printing characteristic of these resins using SLA technique may have been a factor in having accurate match (<6% difference) between experimental attenuation values and NIST calculated.
To have a better illustration of the comparison of AM materials' experimental attenuations with the reference materials' attenuation, figure 5 shows the per cent difference of materials that can simulate RMI457 Solid Water within <10% difference.
Figure 5(A) shows that all PLA-based filaments are good RMI457 substitutes for the energies 30 keV to 50 keV.This energy range is useful in fabricating water-equivalent phantoms for breast CT imaging.Interestingly, PLA (Creality) shows to be able to simulate RMI457 for all the energies studied within ±10%.All the studied resins, utilizing SLA printing, have similar linear attenuation coefficients compared with RMI457 for 45 keV to 65 keV.While White resin has attenuation properties similar to RMI457 for 35 keV and 40 keV as well.
Likewise, figure 5(B) shows that White resin has good equivalence with RMI457 for 35 keV to 65 keV.Both Flexible and Draft resins have good match for 45 keV to 65 keV.Thus, PLA-based materials are generally solid water equivalent for low monochromatic energies investigated, while resin materials are useful for higher monochromatic energies studied (>35 keV).These results are important in deciding which material and 3D printing technique is the best for a specific monochromatic imaging application.
Comparing the experimental monochromatic attenuation coefficients with selected ICRU44 attenuation coefficients showed that many of the studied AM materials can be a good tissue-equivalent substitute.As shown in figure 6(A), PLA+ has similar attenuation coefficient with soft tissue (ICRU44) for all the studied x-ray energies.CPE has a good match with the attenuation of soft tissue except for 30 keV.All other PLA-based filaments are good substitute for soft tissue for 45 keV to 65 keV.
Brain-equivalence, as shown in figure 6(B), is another tissue of interest due to one of the applications of the beamline in brain radiotherapy.PLA+ has the best match with attenuation coefficient of brain (ICRU44).Similar for soft tissue, all other PLA-based filaments are good substitute for brain tissue for 35 keV to 65 keV.Nylon, CPE, CPE+, TPU, White, and Flex can also be used as brain tissue substitute for 50 keV to 65 keV.It is evident in figures 6(A) and (B) that differences flip from negative to positive starting 50 keV except for PLA and Nylon.
Adipose and breast tissues, shown in figures 6(C) and (D), are also considered for the medical imaging application of the synchrotron beamline.ABS, ASA, and Draft resin have similar linear attenuation coefficient with adipose tissue within ±10% for all the x-ray energies studied.PLA (Creality), CPE+, TPU, and White resins have similar attenuation coefficient with breast tissue (ICRU44) within ±10% for all the x-ray energies studied.Lastly, other PLA-based filaments and PETG (not shown in the figure) can also be used as breast tissue for low energies such as 30 keV to 45 keV within ±10%.These data are quite consistent with the results reported using the monochromatic CT capability of the European synchrotron (Ivanov et al 2018).Combining these specified AM materials can be useful in fabricating a breast or water-equivalent phantom for synchrotron monochromatic medical CT imaging QA.

Experimental and simulated synchrotron broad-beam dosimetry
Simulated synchrotron radiation spectra, shown in figure 7, were based on the generated phase space file using a benchmarked Geant4 Monte Carlo simulation code (Dipuglia et al 2019).All the experimental parameters utilized in the synchrotron broad-beam experiment are shown in table 2. Note that each synchrotron radiation spectra will be referred based on the variable 4th and 5th in-vacuo filters used.Three filtration modes were used for the broad-beam experiments to produce variable synchrotron x-ray spectra: 4T-AlAl, 4T-CuCu, and 4T-MoMo.Figure 8 illustrates the per cent difference of the simulated and experimental attenuation properties of AM materials relative to both RMI457 Solid Water and bolus.
As shown in figure 8 (4T AlAl), experimental attenuated doses of AM materials are lower (negative per cent difference) relative to RMI457 dose measurements.This signifies that the studied materials generally attenuate more 4T-AlAl x-rays than the RMI457.This attributed to the variations between the assumed and 7. Hutch 2B synchrotron x-ray spectra calculated using a benchmarked Geant4 Monte Carlo simulation code following the experimental parameters used in the synchrotron beamline.The legend refers to the wiggler magnetic field (4 T) and the 4th and 5th in-vacuo filters used to change the beam quality.a Dose rates at 20 mm depth in a water phantom.More details of the filters are available in this paper (Stevenson et al 2017).actual composition of AM and reference materials.The 4T AlAl synchrotron spectrum has a low mean energy (57.35 keV), thus a high dependence on the effective atomic number is expected due to a dominant photoelectric interaction (Fonseca et al 2023).On the other hand, experimental attenuated doses of AM materials relative to bolus appear to be higher given the positive per cent differences.Accordingly, the studied materials are attenuating less x-rays compared to attenuated dose measurements at the back of bolus.Although this experimental result is the opposite of the negative per cent differences for simulation, the trend is quite comparable.These dose difference relative to water were also observed for MV x-rays in previous studies due to the variation in mass and electron densities (Burleson et al 2015).The simulated and the actual bolus composition is the reason for these differences.Presence of pores inside the MEX-based objects will also influence the interaction of synchrotron x-rays.These air gaps are observable in the synchrotron micro-CT imaging.Overall, Geant4 simulated results for 4T-AlAl show that the studied AM materials can simulate RMI457 within ±5%, except for PP (material 10) and bolus within ±10%.While experimental data show all materials are good RMI457 Solid Water substitute except for eSilk PLA, CPE, TPU, and White.In addition, PP is not a good bolus equivalent material based on experimental results.The RMI457 experimental data presented in figure 8 (4T CuCu) show that the studied materials attenuate less 4T-CuCu x-rays relative to RMI457 given the positive per cent differences.All experimental differences are below 10% except for PP (material 10) On the other hand, experimental data relative to bolus have differences below ±5% except for PP.Geant4 simulation shows that the differences in the attenuated dose readings relative to both RMI457 and bolus are <5% except for CPE (material 6).The experimental and simulated dose data relative to bolus seems to be almost consistent.Majority of the studied materials' experimental dose data have negative per cent difference (<5% difference, except for PP) relative to bolus which means it attenuates x-rays more compared to bolus.Due to the higher mean energy of this spectrum (101.40 keV), less dependence on photoelectric interactions is expected.Thus, lower magnitudes of differences were observed.
Figure 8 (4T MoMo) shows a good match between the simulated and experimental attenuated dose readings which can be explained by the higher mean x-ray energy (157.91 keV) of the spectrum.Similar to 4T-AlAl case, experimental data shows that the AM materials attenuate more x-rays relative to RMI457 (all differences < −10%) while they attenuate less x-rays relative to bolus (all differences <+5%, except for PP).The simulation results show that all the studied AM materials can mimic both RMI457 and bolus with <5% difference.All the data for 4T-MoMo show that the studied AM materials can simulate both RMI457 and bolus within ±10% for this case.
Overall, the broad-beam simulated and experimental data show that synchrotron x-ray spectrum should always be considered when using a specific material for QA and irradiation experiments.4T-AlAl shows to be more material dependent which is expected given its lower x-ray energy compared with both 4T-CuCu and 4T-MoMo with higher x-ray energies (<100 keV).Photoelectric interaction is more probable under this x-ray energy range.Thus, energy dependence based on material composition and detector response is an important consideration (Bushberg et al 2011).The results suggest that the studied AM materials can be good tissue-equivalent materials in fabricating phantoms for synchrotron applications.PP material has been observed to produce the largest difference which can be explained by its low mass density (0.89 g cm −3 ) compared with the other studied AM material with mass density above 1 g cm −3 .It is also worth noting that characterizing AM materials is very important before using it for radiation dosimetry application as the composition may vary depending on different manufacturing factors such as reproducibility in batch production, and polymer additives.

Conclusion and recommendations
This study investigated the suitability of 15 AM materials to simulate both RMI457 Solid Water and bolus which are commonly used in synchrotron pre-clinical imaging and radiotherapy experiments.In addition, these materials were studied as potential substitutes for selected tissues such as breast tissue, adipose tissue, soft-tissue, and brain tissue useful in fabricating 3D printed phantoms for synchrotron imaging and therapy applications.Results confirmed that most AM materials can simulate both Solid Water and bolus depending on the synchrotron monochromatic energy and x-ray spectrum to be used.PLA+ matches the linear attenuation coefficient of both soft tissue and brain tissue for the energies used in this study.ABS, ASA, and Draft resin have good equivalence for adipose tissue with ABS as the best material substitute (<5% difference).PLA (Creality), CPE+, TPU, and White resins are promising materials to substitute breast tissue.Lastly, most of the studied materials have good radiological tissue-equivalence (<10% difference) and can be considered in manufacturing synchrotron x-ray dosimetry phantoms.
It is recommended to study other available AM materials and technologies which might be useful in this advanced medical radiation technology.Repeating this experiment using other synchrotrons and energies is also an interesting extension of this study.AM material properties important for phase contrast imaging can be investigated as well.Finally, fabricating heterogeneous anthropomorphic phantoms (e.g.breast, head, thorax) and pre-clinical animal phantoms (e.g.rodents, canine) for synchrotron imaging and radiotherapy using AM can be done based on these results.

Figure 1 .
Figure 1.Simplified schematic diagram of the AS-IMBL following the direction of the synchrotron x-ray beam from left to right highlighting selected components of the experimental setup of this study: superconducting multipole wiggler (SCMPW), dual-crystal Laue monochromator (DCLM), beam defining aperture (BDA), air ionization chamber (AIC), additive manufacturing (AM) object.Both pink-beam and monochromatic beam (in yellow) are presented.Broad-beam dosimetry and monochromatic CT imaging were done in hutch 2B and 3B, respectively.

Figure 2 .
Figure 2. (A) Experimental setup for monochromatic CT imaging (40 cm from the RUBY detector) acquiring one CT slice per sample, (B) actual setup at the IMBL hutch 3B (red arrow shows the direction of the x-ray beam).
. The RUBY image detector (Monash University, Dynamic Imaging laboratory) composed of an optical camera (pco.edgeCMOS camera), lens (Nikon Micro-Nikkor 105 mm f /2.8 G Extra low dispersion) coupled with scintillator, and a mirror system was utilized in acquiring CT images of all samples (Hall et al 2013).Monochromatic x-ray beams with 8 different energies around the range of 30 keV ⩽ E ⩽ 65 keV were used in 5 keV increment.A lens setting of 1:3, pixel size of 17.50 µm with horizontal field of view of 44.8 mm were utilized.The wiggler magnetic field is set to 3 Tesla for a set of in-vacuum filters of C(0.45), C[hd](5), C[hd](10), and Al(1)-Al(1).

Figure 4 .
Figure 4. Difference between experimental and NIST calculated attenuation coefficients as a function of x-ray energy in keV for (A) Polylactic acid (PLA)-based filaments, (B) Co-polyester-based filaments, (C) Other MEX filaments, (D) Photopolymer resins.

Table 1 .
Estimated elemental composition of AM polymers and reference materials used in NIST and in Geant4 Monte Carlo calculations.

Table 2 .
Synchrotron experimental parameters utilized in the broad-beam dosimetry.

Table A2 .
NIST calculated and experimental linear attenuation coefficients (cm −1 For the experimental data, the mean attenuation coefficient and one standard deviation of the measured attenuation for different CT slices are presented). (