Journal of Non-Crystalline Solids 377 (2013) 137–141

Contents lists available at ScienceDirect

Journal of Non-Crystalline Solids

j ourna l homepage: www.e lsev ie r .com/ locate / jnoncryso l

Samarium-doped oxyfluoride borophosphate glasses for x-ray dosimetry inMicrobeam Radiation Therapy

Vincent Martin a,⁎, Go Okada a, Dancho Tonchev a, George Belev b, Tomasz Wysokinski b,Dean Chapman a,b, Safa Kasap a

a Department of Electrical and Computer Engineering, University of Saskatchewan, Saskatoon, SK, Canada S7N 5A9b Canadian Light Source, University of Saskatchewan, Saskatoon, SK, Canada S7N 0X4

⁎ Corresponding author. Tel.: +1 3062026759; fax: +E-mail address: Vincent.Martin@dal.ca (V. Martin).

0022-3093/$ – see front matter © 2012 Elsevier B.V. Allhttp://dx.doi.org/10.1016/j.jnoncrysol.2012.12.015

a b s t r a c t

a r t i c l e i n f o

Article history:Received 4 October 2012Received in revised form 5 December 2012Available online 8 January 2013

Keywords:Samarium;Alkaline-earth;Oxyfluoride;Photoluminescence;Microbeam Radiation Therapy

One of the key parameters and most difficult challenges related to Microbeam Radiation Therapy (MRT) is theexact measurement of radiation dose delivered. The approach presented here is based on the conversion ofthe oxidation state of samarium ions (embedded in a suitable glass host) upon exposure to high energy ra-diation. The conversion from Sm3+ to Sm2+ under the effect of high x-ray irradiation doses has been studiedin four oxyfluoride glasses based on Ca, Mg, Sr and Ba. These glasses can easily be prepared in air, contrary topure fluoride glasses. They provide good transparency in the visible range, essential for analysis by confocalmicroscopy. Only the Mg-based glasses exhibited a clear conversion of Sm3+ to Sm2+ under x-ray exposure;and hence show potential for use as MRT dosimetry. The conversion did not saturate even up to ~20 kGy(dose in air at the sample surface) under a 50 keV synchrotron beam. Further, we were able to determinethe spatial dose profile of the x-ray beam through confocal fluoroscopic microscopy; and the dose profilematched closely that from the Monte Carlo simulation.

© 2012 Elsevier B.V. All rights reserved.

1. Introduction

Microbeam Radiation Therapy (MRT) is an experimental radiationtherapy for the treatment of many types of cancer. This synchrotron-based technique has the potential to improve the efficiency of thetreatment compared to broad-beam radiation treatment [1,2]. Itinvolves the irradiation of the tumor by parallel x-ray planes withthickness ranging from 20–80 μm, spaced uniformly in the range of100–400 μm [3] as illustrated in Fig. 1. One of the challenges relatedto this technique is the accurate irradiation-dose measurement. Thisincludes a measurement of the total x-ray dose as well as the peak-to-valley ratio. X-ray doses required by MRT can reach 1000 Gy andmore. Therefore, the detector has to be able to measure high x-raydoses and also to be sensitive enough to low x-ray doses (valleydoses). Furthermore the glass should allow a high spatial resolutionto provide an accurate dose profile of the irradiation used to calculatethe x-ray doses. The reduction of Sm3+ to Sm2+ under the effect ofx-ray irradiation has been considered in different Sm-doped fluorideglasses [4]. So far, Sm-doped fluorophosphate and fluoroaluminateglasses have shown good potential for use up to the kiloGray regime[5]. Sm3+ and Sm2+ ions provide intense photoluminescence emis-sion bands in the red region when suitably excited with a blue orUV light [6]. The Sm2+ photoluminescence emission spectrum has

1 306 6494916.

rights reserved.

only small overlap with that from Sm3+. This provides an easy detec-tion and quantification of each ion, which can then be related to theincident x-ray dose. Our goal is to find a host material that allowsthe conversion from Sm3+ to Sm2+ and the stabilization of the oxida-tion state (+II) for high irradiation doses. To minimize its cost interm of production, it has to be prepared under normal atmosphere,i.e. without any need for an O2-free or reducing atmosphere. Oxideglasses do not seem to provide an ideal environment for the conver-sion of Sm3+ to Sm2+ under MRT's experimental conditions. Howev-er, samarium in its oxidation state (+II) can be found in some oxidecrystals. The most interesting case is the strontium tetraborate crystalSrB4O7 which hosts only Sm2+ even when it is synthesized fromSm2O3 in air [7,8]. In this crystal, the samarium is surrounded by 15oxygen atoms located at distances ranging from 2.52 to 3.20 Å [9].At low samarium content, the Sm2+ ion is also found in MBPO5:Smcrystals (M=Ca, Sr, and Ba) and in M1−xSmxSO4 (M=Ba, Sr) in addi-tion to Sm3+ [7,10]. Borophosphate glasses can provide a highamount of non-bridging oxygens that are apparently necessary to sta-bilize Sm2+, brought by tetrahedral boron and phosphorus species.Furthermore, borophosphate glasses are able to provide good chemi-cal durability [11–16]. Given the above observations, we have decidedto synthesize four types of Sm-doped oxyfluoride glasses that canpotentially allow the conversion of Sm3+ to Sm2+ upon irradiationwith x-rays. These glasses are Mg-, Ca-, Sr- and Ba-based oxyfluorideborophosphate glasses, doped by the addition of SmF3. Following syn-thesis, the above glasses were exposed to a high dose radiation at the

Fig. 1. Illustration of the principle of Micribeam Radiation Therapy (MRT). Synchrotron generated x-rays are sent through a multi-slit collimator to create a micro-layers pattern,then these planes irradiate the tumor. The total x-ray dose as well as the valley and peak doses have to be accurately determine for a successful therapy.

Fig. 2. Example of the photoluminescence peaks of Sm3+ and Sm2+ before and afterirradiation for the Mg-based glass.

138 V. Martin et al. / Journal of Non-Crystalline Solids 377 (2013) 137–141

Canadian Synchrotron (CLS) up to several thousand Grays. Thenphotoluminescence experiments were carried out to examine theSm3+ to Sm2+ conversion. Further, a microbeamwas also used to ex-amine the spatial resolution that is achievable with these Sm-dopedglasses under confocal fluoroscopic readout of the Sm2+ lumines-cence. The results indicate that Sm-doped Mg-based glasses showgood potential for use in high-dose high-resolution dosimetry.

2. Experiment

The glasses were all prepared from the same nominal composition:

18:6 MF2−18:6 MO−37:4 P2O5−24:9 B2O3−0:5 SmF3

with M=Mg, Ca, Sr, and Ba.The glasses were prepared from (NH4)2HPO4 (Alfa Aesar, 98%),

B2O3 (Alfa Aesar, 99%), SmF3 (Alfa Aesar, 99.99%), MgF2 (Balzers),MgO (Puratronic, grade 1), CaF2 (ESPI), CaO (Alfa Aesar, 99.95%),SrF2 (Alfa Aesar, 99%) and SrCO3 (Alfa Aesar, 99%, 1% Ba), BaF2 (AlfaAesar, 99%), and BaCO3 (Alfa Aesar, tech.). All syntheses were pre-pared following the same procedure. First, the oxides and carbonateswere mixed in an alumina crucible, heated for 4 h at 600 °C then at1300 °C for 1 h. Each sample was annealed at a temperature 15 °Cbelow its respective Tg for 4 h. Afterward, the melt was quenchedat room temperature on a brass plate. The obtained oxide glass wasgrounded in thin powder into a ceramic mortar and placed backinto the same crucible. The fluorides were added and the mixturewas heated at 1300°C for 1 h and quenched at room temperature.Each sample was annealed at a temperature 15 °C below its respec-tive Tg for 4 h. The prepared samples were cut and polished, andthen irradiated with x-rays at the CLS on the Biomedical ImagingTherapy beamline. The peak energy of the x-ray beam was 50 keV(used in MRT). The dose rate was experimentally estimated to be ap-proximately 0.0064 Gy s−1mA−1 at 250 mA of the storage ring cur-rent. The reported dose values correspond to dose in air at thesample surface. Several pieces of each sample were irradiated for dif-ferent doses from 0.5 to 20,000 Gy, and then they were wrappedwith aluminum foil and stored until the photoluminescence re-sponse measurements were carried out. A piece of Mg-based glasswas also irradiated with a microbeam at a total dose of 10,000 Gy.The size of the slit in the multi-slit collimator was 50 μm wide andthe slits were separated by a distance of 400 μm center-to-center.The distance between the collimator and the sample was 1 m. Thedepth of measurement from the surface of the sample was 20 μm.The response of a sample upon x-ray irradiation was readout usinga confocal microscopic technique. We used a 473 nm diode-pumped solid-state laser as the excitation source, and the con-sequent photoluminescence signals from Sm3+ and Sm2+ were

measured by two separate photomultiplier tubes (PMT) in order tomeasure the response as:

R ¼PL Sm2þð ÞPL Sm3þð Þ

¼ IPMT 2þð ÞIPMT 3þð Þ

!irradiated

−IPMT 2þð ÞIPMT 3þð Þ

!non−irradiated

ð1Þ

where IPMT(2+) and IPMT(3+) are measured PMT signals. These are in-tegrated photoluminescent signals over the ranges of 570 to 650 nmand 660 to 800 nm, respectively. The first term takes the signalvalues for irradiated samples; while, the second term is those fornon-irradiated samples, which subtracts the contribution of typical4G5/2→

6H11/2 emission of Sm3+ to the IPMT(2+) signal, hence the re-sponse value without irradiation is zero. Fig. 2 shows the emissionpeak from the photo-excitation of Sm3+ and Sm2+ before and afterirradiation for the Mg-based glass. The value given by Eq. (1) is therelative intensity of the photoluminescence generated by the Sm2+

created by irradiation. It is an indicator of the Sm3+→Sm2+

conversion.The thermal analysis of the samples were performed using a

Seratam TG-DSC 111 from room temperature to 800 °C. Prior to anal-ysis, the samples were ground to powder in a ceramic mortar.

Table 1Glass transition temperature Tg and temperatures of crystallization Tc1 and Tc2 fromroom temperature to 800 °C. The uncertainty on the temperatures was estimated at±1 °C.

Sample Tg Tc1 Tc2

Mg-based 589 °C 751 °C NoneCa-based 648 °C 705 °C 762 °CSr-based 640 °C None NoneBa-based 607 °C None None

139V. Martin et al. / Journal of Non-Crystalline Solids 377 (2013) 137–141

The optical transmission was measured from 200 nm to 2000 nmusing a Perkin Elmer Lambda 900 spectrometer. The measurementswere performed on cut and polished samples before being irradiatedwith synchrotron generated x-rays. The absorption coefficient wascalculated from these measurements and the samples' thicknessusing the Beer-Lambert law. The optical absorption spectra for theMg-sample were also obtained before and after x-ray irradiation.

3. Results

All the peaks present on the absorption coefficient spectra in thenear-UV region and in the infrared region (see Fig. 3) are typical ofSm3+-containingmaterials [17–20]. The absorption jump and noise lo-cated at 890 nm and below are due to the light bulb change in the spec-trometer during the analysis and were not removed from the spectra(This jump does not interfere with the peak at 920 nm). The UV-absorption edge in borophosphate glasses is generally observed around200–350 nm depending on the amount of B2O3, and the nature andamount of additives [16,21]. This is the case for all four samples withan absorption edge in the range from 300 to 350 nm for the Ba-, Sr-and Ca-containing glasses. In the Mg-based glass, the absorption edgeoccurs at a much lower wavelength (~250 nm). The UV-absorptionedge is shifted towards shorter wavelengths as the size of thealkaline-earth decreases. All non-irradiated samples present a low ab-sorption coefficient in the visible range. The dark coloration after irradi-ation greatly degraded the optical transmission in the visible for theirradiated Mg-based sample. The values of the glass transition temper-atures and temperatures of crystallization are listed in Table 1. In thisrange, the magnesium sample shows a crystallization peak at 751 °C,and the calcium sample shows two crystallization peaks at 705 °C and762 °C. No crystallization occurs for the strontium and barium con-taining samples under 800 °C. Nevertheless, a sudden rise of the heatflow above 775 °C for these two samples suggest the possible presenceof crystallization close to, but higher than 800 °C. A typical thermogramfor the Mg-based sample is presented in Fig. 4, where Tg and the peakcrystallization temperature Tc1 have been marked.

The response of all four types of Sm3+-doped glasses, as deter-mined from Eq. (1), after x-ray irradiation at different doses, isshown in Fig. 5. The Ca-, Sr- and Ba-containing glasses present apoor photoluminescence response (Rb0.02) even at high irradiationdoses under blue excitation (473 nm). Put differently, we do notdetect any conversion of Sm3+ to Sm2+ within errors involved inthe measurement of R. However, it has to be noted that a small

Fig. 3. Absorption coefficient of the four different non-irradiated samples from near-UV to IRradation of the transparency in the visible.

conversion was detectable under green excitation (535 nm). On theother hand, the Mg-based glass shows a good response, ~0.35 fordoses of ~20 kGy. This sample does not show signs of saturation athigh x-ray doses and could probably measure even larger doses,which is a distinct advantage. The response measured under lowdoses (b10 Gy) is limited by the noise floor of the confocal micro-scopic technique used, and also the noise associated with the subtrac-tion of the two spectra. The dose profile of the x-ray microbeam alongthe surface of the Mg-based glass is shown in Fig. 6(a). The peaks cor-respond to the layers of x-rays passing through the collimator. Thepeak dose is measured at 10,000 Gy, corresponding to the total dosedelivered. The valley dose is estimated at 100 Gy. This value doesnot go down to 0 Gy due to dispersion/scattering of the x-rays.

4. Discussion

The transparency of the sample in the visible range makes possiblefor the measurements of the photoluminescence of Sm cations by con-focal microscopy. The absorption peak at 400 nm does not interferewith the quantification of Sm2+ and Sm3+ by photoluminescence;their emission peaks being located in the red region. The Sr-basedsample does not provide any encouraging results. As mentioned pre-viously, Sm2+ is present in the SrB4O7 crystal even when it is pre-pared in air. But its oxidation state becomes (+III) in the glassmatrix of the same composition (SrO.2B2O3) even when preparedunder a reducing atmosphere [7]. In polycrystalline SrBPO5:Sm3+,an excellent samarium conversion has been observed after only55 s irradiation [4]. From these observations, it appears that thelocal environment of the samarium is a critical aspect for its abilityto convert, and a crystalline environment seems to be more suitable.The Mg-based sample is the only one that is able to show good con-version under high x-ray doses well above the kiloGray regime. Adose profile along the surface is presented in Fig. 6 for this sample.

-B. The spectrum of the irradiated Mg-based sample is also displayed, showing the deg-

Fig. 4. Thermogram of the Mg-based glass sample from 500 °C to 800 °C. The glass transition temperature has been determined at 589±1 °C for this sample. One crystallizationpeak is visible at 751±1 °C.

Fig. 5. Response R generated by Sm2+ under blue excitation (473 nm) created after x-rayirradiation. The values of R are obtained from Eq. (1). The variations of R observed for theCa-, Sr-, and Ba-based samples at high X-ray doses are due to the lack of sensitivity of thephotomultipliers at low intensities.

Fig. 6. a) Dose profile along the surface of the Mg-based sample after irradiation by a x-ray msured microbeam in the Mg-based glass with predicted data calculated by a Monte Carlo si

140 V. Martin et al. / Journal of Non-Crystalline Solids 377 (2013) 137–141

The profile of the two peaks, corresponding to the irradiated areas, iscompared to predicted data calculated by Monte Carlo simulationfromNettelbeck et al. [22] in Fig. 6(b). A good agreement is found be-tween the experimentally determined profile and that predicted byMonte Carlo simulation, confirming the high-resolution capabilityof the Mg-based glass. Since a crystalline environment seems to in-crease the samarium conversion, pieces of the magnesium-basedglass have been annealed at 740 °C (above Tg and below TC1 for thisglass) for 6 and 12 h respectively to generate the nucleation ofnano-crystals in the glass. The samples subsequently have been irra-diated and the response has been recorded according to the sameprocedures as described above. The results of the photoluminescenceresponse of the converted Sm2+ are shown in Fig. 7 for the three mag-nesium samples. Surprisingly, the response (conversion) is degradedcompared to the original glass material. The maximum observed re-sponse is decreased and it reaches a saturation at lower doses as theannealing time increases. It has to be noted that at this point, no x-raydiffraction analysis has been performed to detect and characterize thenano-crystals generated by the annealing treatment.

icrobeam. The total delivered dose was 10,000 Gy. b) Comparison of the shape of mea-mulation (Nettelbeck et al. [22]).

Fig. 7. Response R generated by Sm2+ under blue excitation (473 nm) created after x-rayirradiation for the Mg-based samples with no additional thermal treatment, and with anannealing treatment at 740 °C for 6 h and 12 h. The values of R are obtained from Eq. (1).

141V. Martin et al. / Journal of Non-Crystalline Solids 377 (2013) 137–141

5. Conclusion

Four oxyfluoride borophosphate glasses were prepared with fourdifferent alkaline-earths. All samples exhibited typical optical proper-ties of Sm-containing materials. From the literature, strontium seemedto be the best element able to provide a good Sm3+→Sm2+ conver-sion, but it did not meet the expected results in this case. However,the Mg-based sample presented an excellent ability to measurehigh x-ray doses up to 20 kGy under a 50 keV x-ray beam. Further,we were able to determine the spatial dose profile of the x-raybeam through confocal fluoroscopic microscopy; and the dose pro-file matched closely the Monte Carlo simulation. Given that theMg-based glass did not exhibit any saturation at 20 kGy, it is quitelikely that we can consider it as a potential dosimetric material forMRT. The nucleation of crystals at a temperature close to the firstcrystallization peak degraded the quality of the samarium conver-sion. This new glass provides a new starting point for further studiesto explain the conversion of samarium by x-ray irradiation. In futurework, the structure of the glass, and eventually glass-ceramics, could

be studied by 19F NMR before and after irradiation to reveal structur-al changes around the samarium for example.

Acknowledgments

We would like to thank the NSERC for financial support andTeledyne-DALSA for sponsoring this project.

References

[1] D. Slatkin, F. Dimanian, P. Spanne, M. Sandborg, Med. Phys. 19 (1992) 1395–1400.[2] F.A. Dilmanian, P. Romanelli, Z. Zhonge, R. Wang, M. Wagshul, J. Kalef-Ezra,

M. Maryanski, E. Rosen, D. Anshel, Eur. J. Radiol. 38 (2008) S129–S136.[3] M. Ptaszkiewicz, E. Braurer-Kirsch, M. Klosowski, L. Czopyk, P. Olko, Radiat. Meas.

43 (2008) 990–993.[4] G. Belev, G. Okada, D. Tonchev, C. Koughia, C. Varoy, A. Edgar, T. Wysokinski,

D. Chapman, S. Kasap, Phys. Status Solidi C 8 (2011) 2822–2825.[5] S. Vahedi, G. Okada, B. Morrell, E. Muzar, C. Koughia, A. Edgar, C. Varoy, G. Belev,

T. Wysokinski, D. Chapman, S. Kasap, J. Appl. Phys. 112 (2012) 073108.[6] G. Okada, B. Morrell, C. Koughia, A. Edgar, C. Varoy, G. Belev, T. Wysokinski,

D. Chapman, S. Kasap, Appl. Phys. Lett. 99 (121105) (2011) 1–3.[7] P. Mikhail, J. Hulliger, Comments Inorg. Chem. 21 (1999) 263–283.[8] H.B. Liang, Q. Su, S. Lu, Y. Tao, T.D. Hu, T. Liu, Mater. Res. Bull. 37 (2002) 1981–1989.[9] A. Perloff, S. Block, Acta Crystallogr. 20 (1966) 274–279.

[10] P. Mikhail, A. Sieber, T. Samtleben, B. Trusch, T. Lüthi, J. Hulliger, J. Solid StateChem. 154 (2000) 535–541.

[11] R. Brow, J. Non-Cryst. Solids 194 (1996) 267–273.[12] R. Brow, D.R. Tallant, J. Non-Cryst. Solids 222 (1997) 396–406.[13] L. Koudelka, P. Mosner, J. Pospísil, L. Montagne, G. Palavit, J. Solid State Chem. 178

(2005) 1837–1843.[14] L. Koudelka, J. Šubčík, P. Mosner, L. Montagne, L. Delevoye, J. Non-Cryst. Solids

353 (2007) 1828–1833.[15] K.V. Shah, M. Goswami, S. Manikandan, V.K. Shrikhande, G.P. Kothiyal, Bull.

Mater. Sci. 32 (2009) 329–336.[16] J.W. Lim, M.L. Schmitt, R.K. Brow, S.W. Yung, J. Non-Cryst. Solids 356 (2010)

1379–1384.[17] A. Edgar, C.R. Varoy, C. Koughia, D. Tonchev, G. Belev, G. Okada, S.O. Kasap, H. Von

Seggern, M. Ryan, Opt. Mater. 31 (2009) 1459–1466.[18] N. Yaru, L. Chunhua, Z. Yan, Z. Qitu, X. Zhongzi, J. Rare Earths 25 (2007) 94–98.[19] M. Jayasimhadri, E.-J. Cho, K.-W. Jang, H.S. Lee, S.I. Kim, J. Phys. D: Appl. Phys. 41

(2008) 175101.[20] Y.K. Sharma, S.S.L. Surana, R.K. Singh, J. Rare Earths 27 (2009) 773–780.[21] D. Kim, C. Hwang, D. Gwoo, T. Kim, Y. Kim, N. Kim, B.K. Ryu, Electron. Mater. Lett.

7 (2011) 309–312.[22] H. Nettelbeck, G.J. Takacs, M.L.F. Lerch, A. Rozenfeld, Med. Phys. 36 (2009) 447–456.