- Email: [email protected]
Ion beam induced luminescence analysis of defect evolution in lithium fluoride under proton irradiation
Optical Materials 49 (2015) 1–5
Contents lists available at ScienceDirect
Optical Materials journal homepage: www.elsevier.com/locate/optmat
Ion beam induced luminescence analysis of defect evolution in lithium fluoride under proton irradiation A. Quaranta a,⇑, G. Valotto b, M. Piccinini c, R.M. Montereali c a
University of Trento, Department of Industrial Engineering, Via Sommarive 9, I-38132 Povo, Trento, Italy Department of Environmental Sciences, Informatics and Statistics, Università Ca’ Foscari Venezia, Dorsoduro 2137, I-30123 Venezia, Italy c ENEA C.R. Frascati, UTAPRAD-MNF, Photonics Micro and Nanostructures Laboratory, Via E. Fermi 45, 00044 Frascati, Rome, Italy b
a r t i c l e
i n f o
Article history: Received 24 May 2015 Received in revised form 15 August 2015 Accepted 16 August 2015
Keywords: Ion beam induced luminescence Lithium fluoride Multiple linear regression
a b s t r a c t Ion beam induced luminescence (IBIL) spectra of pure LiF under irradiation by a 2 MeV proton beam were analyzed as a function of the dose in order to deepen the kinetic mechanisms underlying the formation of luminescent point defects. The intensity evolution with dose at several emission wavelengths has been studied within a wide spectral interval, from ultraviolet (UV) to near infrared (NIR), and their different change rates have been correlated to the electronic defect formation processes. The intensity at few selected wavelengths was analyzed with a multiple linear regression (MLR) method in order to demonstrate that a linear calibration curve can be obtained and that an on-line optical dose monitor for ion beams can be realized. Ó 2015 Elsevier B.V. All rights reserved.
1. Introduction The exposure of pure LiF to different radiations, like ion beams, electrons, neutrons, gamma and X-rays or UV light, gives rise to the formation of color centers [1] characterized by fluorescence features whose light yield depends both on the radiation type and dose. The optical properties of such luminescent defects have been widely studied for several applications, like thermoluminescent dosimetry [2–4], miniaturized active channel waveguides [5], solid-state lasers [6] and neutron imaging detectors [7]. Moreover, the possibility to write luminescent submicrometricpatterns with X-rays, electron beams or ion beams and the stable formation of high contrast fluorescence images, has been demonstrated [8–11]. In spite of the great amount of work devoted to the analysis of LiF luminescent defects, some points are still to be fully understood, especially for what concerns the defects produced under ion beam irradiation. In this field, the most studied luminescence features are a structured peak in the UV and two broad bands, corresponding to optical transitions of F+3 and F2 centers, peaked around 540 nm and 670 nm, respectively. The UV peaks, studied mainly by Skuratov and co-workers with swift heavy ions [12–15], have been attributed to intrinsic or impurity/defect related self trapped exciton (STE) transitions, but at present a well
⇑ Corresponding author. E-mail address: [email protected] (A. Quaranta). http://dx.doi.org/10.1016/j.optmat.2015.08.015 0925-3467/Ó 2015 Elsevier B.V. All rights reserved.
defined origin is still to be established. Besides these main spectral features, other emission peaks are sometimes observed, like a band around 400 nm related to oxygen or Mg impurities [16,17] and the + NIR band of F 3 and F2 defects located around 900 nm [17,18]. A peculiar behaviour of ion beam induced luminescence (IBIL), also indicated as ionoluminescence (IL) by many authors, is that both intensity and shape of spectra change during irradiation owing to the point defect formation kinetic mechanisms. Typically, the yield decreases with the dose due to the formation of quenching or trapping centers produced by the irradiation. On the other hand, the intensity of features related to point defects increases during irradiation, due to the formation of luminescent centers induced by the ion beam itself, and then decreases, after reaching a maximum, when the center concentration is so high to trigger aggregation processes [19–21]. In lithium fluoride, it has been observed that the IBIL yield of F+3 and F2 bands follows the typical trend of point defects, while UV features, analyzed under swift heavy ion irradiation, monotonically decrease. The increasing rate depends on the defect formation mechanism, and it could be different for each defect. Up to now, the formation kinetics of F+3 and F2 centers has been deeply studied in samples irradiated with gamma-ray at low temperatures and then annealed at room temperature. In this case it was observed that the F+3 peak grows with a higher rate with respect to F2 owing to the diffusion of vacancies, allowing the reaction V+a + F2 ? F+3 [22]. On the contrary, under ion beam irradiation, the growth rates of the two bands are less studied. Only Russakova et al. made a
2
A. Quaranta et al. / Optical Materials 49 (2015) 1–5
comparison by means of optical analyses after heavy ion irradiation, finding similar increasing rates [23], but the conclusions came from the interpolation of few experimental points. In the same paper, the decreasing rate at higher doses has been examined. In particular, it has been observed that under irradiation with 14 MeV N2+ ions and 150 MeV Kr14+ ions the F+3 band yield decreases faster with respect to F2, and the effect has been attributed to optical self-absorption effects in the beam spot. Actually, at high doses the point defect luminescence bands decreases for several mechanisms which can be hardly distinguished, like the aggregation of the emitting centers into larger non-luminescent defects, concentration quenching or self-absorption. The aim of the present work is to study the behaviour of LiF point defects under proton beam irradiation by means of ion beam induced luminescence (IBIL) measurements. We focused on protons since they are widely used for applications ranging from material and device modification to radiobiology and radiotherapy. Very recently we started the investigation of the optical absorption and emission properties of CCs induced in LiF crystals and thin films by low energy protons. The stationary (after irradiation) visible photoluminescence (PL) spectra behaviour of radiationinduced color centers in LiF crystals was systematically studied as a function of irradiation fluence (interval of doses from 103 to 107 Gy), obtaining evidence of a linear optical response for the VIS peaks of F2 and F+3 color centers [24,25]. The detection of IBIL full spectra in real time during irradiation, presented here for the first time, allows to better understand the kinetics mechanisms involving the formation of luminescent point defects under ion irradiation. Moreover, the intensities as a function of the dose at selected wavelengths are interpolated by means of a multiple linear regression (MLR) method in order to demonstrate, at least from a conceptual point of view, that a linear relationship between the IBIL yield at some wavelengths and the dose can be obtained. This study can pave the way for the use of LiF crystals as on-line ion beam dose optical monitors, which could be particularly useful in hadrotherapy treatments [26].
allowing to follow the light intensity as a function of fluence within a wavelength interval ranging from 250 nm to 900 nm. This set-up allows on-line monitoring of the emitted intensity during proton irradiation, thus giving an exhaustive spectral analysis of the luminescence yield evolution. The spectra were collected with a resolution of 5 nm and they were not corrected for the response of the overall detection setup. 2.2. MLR analysis In order to analyze the possibility of extracting a calibration curve for the irradiation dose from IBIL spectra, we performed a multiple linear regression (MLR) analysis of the luminescence intensities at different wavelengths, with the aid of the code StatSoft 8.0. In particular, we fit the dose values D with a multiple linear function of the type:
D ¼ b0 þ
N X bi Ii
ð1Þ
i¼1
where bi are the fitting parameters and Ii are the intensities at the ki peak wavelengths taken as variables. For the fitting we used a forward stepwise method, by inserting one intensity variable at a time and by analyzing at every step the statistical significance of all the variables. The significance was evaluated with the parameters ‘‘tolerance”, ‘‘F to enter”, and ‘‘F to remove” set to 0.1, 100 and 10, respectively. ‘‘Tolerance” parameter is defined as 1 minus the squared multiple correlation of a selected variable with all the other independent variables used in the regression equation. Therefore, variables with small ‘‘tolerance” give a redundant contribution to the regression analysis. The ‘‘F to enter” (‘‘F to remove”) value determines how significant (insignificant) the contribution of a variable in the regression equation is, in order to be inserted (removed) into the equation. Finally, the R2 parameter was used as a performance indicator for the quality of the model. 3. Results and discussion
2. Materials and methods 3.1. IBIL spectra 2.1. IBIL measurements IBIL measurements were performed at the AN2000 accelerator of the INFN Laboratori Nazionali di Legnaro, by irradiating a nominally pure cleaved LiF crystal, of dimensions 10 10 1 mm3 at room temperature with a 2.0 MeV proton beam, perpendicular to the sample surface. The beam cross section was 1 mm2 and the current density was around 2.3 lA cm2, corresponding to an ion flux of 1.4 1013 H+ cm2 s1. During the ion irradiation, performed at 106 Torr, the chamber worked as a Faraday cup allowing to measure both the total charge impinging on the sample and the beam current. The maximum fluence value on the sample was 2.9 1015 H+ cm2. From the proton range (Rp = 37.9 lm in LiF calculated by SRIM2008 [27]) and taking into account the material density (q = 2.635 g/cm3) the fluence was converted into the average dose deposited along the ion path. In particular, the doses corresponding to the minimum and maximum fluence are 0.45 MGy and 92 MGy, respectively. The ionization energy of the impinging protons ranges from 33 eV ion1 nm1 at the sample surface, up to a maximum of 120 eV ion1 nm1 at the depth corresponding to the Bragg peak. The luminescence light was collected into the vacuum chamber by a silica fiber (600 lm diameter) put in front of the irradiated surface, at a distance of around 15 cm and at an angle of 20° with respect to the beam line direction. IBIL spectra were recorded by an Ocean Optics QE65000 spectrometer coupled to the chamber by means of a fiber vacuum connector. During the irradiation, 200 spectra were recorded, each one within 1 s of integration time,
In Fig. 1 are shown the IBIL spectra of a LiF crystal, collected at 1.4 MGy, 51 MGy and 93 MGy of irradiation dose, corresponding to
Fig. 1. IBIL spectra of the 2 MeV proton irradiated LiF crystal collected at 1.4 MGy, 51 MGy and 92 MGy of irradiation dose, corresponding to 4.4 1013 H+ cm2, 1.6 1015 H+ cm2 and 2.9 1015 H+ cm2, respectively. In the inset the UV peaks at higher doses are evidenced.
A. Quaranta et al. / Optical Materials 49 (2015) 1–5
fluences of 4.4 1013 H+ cm2, 1.6 1015 H+ cm2 and 2.9 1015 H+ cm2, respectively. In the UV range, the spectrum collected at the lower dose exhibits a peak at 320 nm, with a shoulder at 385 nm. By increasing the dose, all the peaks decreases leaving a faint peak at 330 nm, with a broad shoulder at 310 nm, and a peak at 395 nm, shown in the inset of Fig. 1. Skuratov and co-workers discussed the two main components in the UV range, reported at 296 and 330 nm [12,15], observed under swift heavy ion irradiation. The peak at 330 nm was univocally attributed to the STE recombination, while the lower wavelength feature has been subjected to different interpretations. In particular, it has been sometimes attributed to the STE recombination near an unstable defect [12], or near impurities [13]. The feature at higher wavelengths, reported in literature at 400 nm [13], has been identified as an impurity related band. The intensity of F+3 and F2 bands at 540 and 670 nm increases from 1.4 MGy to 51 MGy and then decreases at 92 MGy. Besides these features, a faint shoulder at 740 nm, attributed to F4-like + centers, and an intense band around 890 nm, due to F 3 and F2 centers [22,28], can be observed (the small valley at 875 nm is an instrumental artifact). A more complete representation of the IBIL spectra evolution can be observed in Fig. 2, where a 2-D plot is shown reporting the intensity at each wavelength normalized to its own maximum value. As it can be observed, two ranges with different trends can be identified: in the range from UV to 450 nm the intensity monotonically decreases with different rates, whereas at wavelengths higher than 500 nm all the luminescence yields increase, reaching a maximum at doses ranging from 10 to 30 MGy, and then monotonically decrease. In Fig. 3a and b are reported the effective (a) and normalized to the maximum (b) intensities as a function of the dose at five representative wavelengths, namely 315 nm, 385 nm, 545 nm, 670 nm and 890 nm. The yields at 315 and 385 nm monotonically decrease with different rates, since the features originate from different emitting centers, and definitely indicates that these features are not related to point defects, but to STE recombination processes in the lattice or near impurities. Skuratov et al. observed indeed a monotonic decrease [13,15], but the high ionization energy density deposited by swift heavy ions could induce to some misleading interpretation, since it should be possible that the dose rate release is so high to overcome the yield turning point just from the irradiation beginning. The same trend observed under proton irradiation, with a ionization energy orders of magnitude lower, is the definite evidence that STE giving the UV features decays in the lattice or near impurities. The luminescence yields at 545 nm, 670 nm and 890 nm increase with dose up to 11 MGy, 14 MGy and 37 MGy (3.4 1014, 4.4 1014 and 1.1 1015 H+ cm2), respectively. The yields of F+3 centers at 545 nm and of F2 centers at 670 nm increase with a slightly different rate. This trend is opposite to the behaviour observed during the room temperature thermal annealing of irradiated crystals [22]. It has to be taken into account that a 2 MeV proton can excite a around 1.5 105 electrons in the conduction band (value obtained by dividing the proton energy by the LiF band gap of 13.7 eV [29]). It gives an average number of 2 1018 electrons cm2 migrating through the lattice and reaching point defects. So, under beam irradiation, the large number electrons scattered by the ions through the crystal enhance the annihilation probability of F+3 centers through the reaction F+3 + e ? F3, thus slowing down their formation rate [23,30]. After reaching the maximum, the two features decrease during irradiation with a quite similar rate. Russakova et al. observed a
3
faster decreasing rate for F+3 centers, attributed mainly to selfabsorption effects induced by the heavy damage released by high energy heavy ions [23]. In the sample studied in this work, the colored spot left by the ion beam on the sample surface was very faint and optical self-absorption effects could be discarded in this work. The decrease of the two yields with the same rate, here indicates that the two kinds of defects under proton beam irradiation are subjected to the same aggregation processes giving rise to larger structures quenching the luminescence. The luminescence intensity at 890 nm, whose trend as a function of the dose is observed here for the first time, reaches its maximum at higher dose and decreases with a slower rate with respect to F+3 and F2 centers. The lower increasing rate can be explained by the need of a more complex reaction for the production of the centers which are responsible of this feature. For instance, F 3 centers can be formed either by the combination of a F2 center with a F defect, or by trapping two electrons into a F+3 center. The former mechanism can be realized when the proper local concentrations of F centers are achieved by the ion beam induced damage. At the same time, scattered electrons can promote the reactions e + F ? F and 2e + F+3 ? F 3 . The slower yield decrease of the NIR feature is less clear for the moment. Shiran et al. [17] observed that in cathodoluminescence spectra with high e-beam density, NIR peaks are practically unchanged, while visible features of F2 and F+3 centers disappears. This result points out that centers giving rise to high wavelengths emissions are more resistant under irradiation. Moreover, the presence of an energy transfer process from F2 centers to NIR emitting defects, contributing to the lower decreasing rate of this feature, cannot be neglected [28]. 3.2. MLR analysis IBIL spectra of materials like LiF are characterized by a strongly non-linear behaviour as a function of the dose or fluence. This behaviour changes for the different spectrum wavelengths due to the different nature of the luminescent centers and to their response to ion irradiation. In this work we want to study how the pattern of the yield variations as a function of the dose can be exploited for realizing dose or beam optical monitoring systems. For the MLR interpolation the yields at wavelengths near the main peaks, namely at 312 nm, 383 nm, 537 nm, 664 nm and 855 nm, were taken as main variables. Following the stepwise procedure previously described, we observed that 383 nm and 537 nm did not satisfy both ‘‘tolerance” and ‘‘F” values. In fact, the intensities at these wavelengths evolve in a way similar to the values at 312 and 664 nm, respectively, and they do not add significant information to the MLR analysis. By keeping only the intensities at 664 and 312 nm as variables, we obtained a R2 value of 0.92. This condition gives a poor prediction performance for higher and lower dose values used in this work. By including as a variable the intensity at 855 nm, the R2 value increased up to 0.99 with a remarkable improvement of the prediction performance. In Table 1 are reported the MLR coefficients calculated with the intensities at the three selected wavelengths. Due to the decrease of the luminescence yield over a certain dose value, all the coefficients are negative. In the same Table the standardized regression coefficients are shown. These parameters were obtained by fitting the standardized variables, that is the intensities normalized in order to give mean intensity equal to 0 and standard deviation equal to 1. Standardized regression coefficients allow to compare the relative contribution of each independent variable in the prediction of the dose value. In our case, the most influent variable is the yield at 312 nm, followed by 855 nm and 664 nm. In Fig. 4 are reported the experimental dose values (indicated as ‘‘observed” in the ordinate axis) as a function of the ‘‘predicted”
4
A. Quaranta et al. / Optical Materials 49 (2015) 1–5
Fig. 2. Normalized intensities of IBIL spectra as a function of both the dose and the wavelength. For each wavelength, the intensity was normalized to its own maximum value. Horizontal lines are due to fluctuations of the ion beam current.
Table 1 MLR regression coefficients (RC) and the corresponding standard errors (SE) for the intercept b0 and the wavelength intensities bk . The standardized regression coefficients (SRC) where calculated with the by fitting the standardized variables, where the mean intensity value and the respective standard deviation of each variable was set to 0 and 1, respectively.
b0 b312 b664 b855
RC
RC–SE
SRC
SRC–SE
20.0 58 4.1 23.9
0.3 1 0.2 0.8
– 1.22 0.34 0.71
– 0.02 0.01 0.02
Fig. 4. Observed vs. predicted dose values obtained from the MLR of the luminescence yield at 312 nm, 664 nm and 855 nm. Circles are the experimental points and the red line is the interpolation curve obtained from MLR. Dashed lines indicate the 95% confidence interval for the interpolation curve.
Fig. 3. IBIL yield as a function of the dose as collected (a) and normalized to the maximum (b) recorded at five wavelengths representative of the main spectral features.
values, obtained by substituting the corresponding IBIL intensities at the selected wavelengths in the MLR interpolation formula. As it can be observed, the points are well aligned along an interpolating line with slope 1, also reported in the figure. The dashed lines indicate the 95% confidence interval for the interpolation.
It is worth noting that a very good correlation was obtained with only three variables, allowing to realize a reliable calibration curve with a low waste of time and calculations. A standard error of around 3 MGy was evaluated in the estimation of the predicted dose from the MLR curve. This value is also related to the beam current fluctuations and could be improved by providing a set-up for recording in real time the charge corresponding to every single spectrum and by normalizing the spectrum to this value. Further improvements could be obtained by focusing the analysis on narrower dose intervals, suited for monitoring purposes.
A. Quaranta et al. / Optical Materials 49 (2015) 1–5
Besides these considerations, it is worth noting that this result is the first proof of concept demonstrating that IBIL spectra are suitable for the realization of on-line optical monitoring systems. 4. Conclusions In this work we have studied the detailed IBIL yield evolution of a pure LiF crystal under proton irradiation. The monotonic decrease of the luminescence intensity in the UV range with increase of dose allows to conclude that the emission features are related to STE transitions near lattice impurities rather than to point defects. + Concerning point defects, the formation of F2, F+3, and F 3 /F2 centers has been monitored by the intensity changes of the corresponding luminescence bands. The formation rate of F2 centers is slightly higher than F+3 centers, owing to the annealing of the latter by means of scattered electrons. On the other hand, the decreasing rate of the two yields is quite similar evidencing similar processes of aggregation and quenching for the two defects. + The evolution of the NIR F 3 /F2 band as a function of the dose was monitored here for the first time, evidencing that it increases and decreases with slower rates with respect to the other point defects, due to both more complex formation mechanisms and higher radiation hardness with respect to the other centers. Moreover energy transfer effects from F2 centers cannot be discarded. In order to study how the IBIL yield as a function of the dose can be used for the realization of beam or dose monitoring systems, a MLR analysis was performed by interpolating the luminescence intensity at few selected wavelengths. With the intensities at three wavelengths a very good regression was obtained, and a calibration curve for the interval from 1 to 90 MGy was found. Even better interpolations could be obtained by selecting narrower does ranges or by controlling the collected charge in real time. Further work is planned by analyzing several crystals of the same type in order to study the reproducibility of the approach for the production of reliable IBIL based dosimeters. References [1] W.B. Fowler, Physics of Color Centers, Academic Press, New York and London, 1968. [2] W.L. McLaughlin, A. Miller, S.C. Ellis, A.C. Lucas, B.M. Kapsar, Radiation-induced color centers in LiF for dosimetry at high absorbed dose rates, Nucl. Instr. Method 175 (1980) 17–18. [3] G. Baldacchini, A.T. Davidson, V.S. Kalinov, A.G. Kozakiewicz, R.M. Montereali, E. Nichelatti, A.P. Voitovich, Thermoluminescence of pure LiF crystals and color centers, J. Lumin. 122–123 (2007) 371–373. [4] D.R.S. Ribeiro, D.N. Souza, A.F. Maia, S.L. Baldochi, L.V.E. Caldas, Applicability of pure LiF in dosimetry, Rad. Meas. 43 (2–6) (2008) 1132–1134. [5] R.M. Montereali, M. Piccinini, E. Burattini, Amplified spontaneous emission in active channel waveguides produced by electron-beam lithography in LiF crystals, Appl. Phys. Lett. 78 (26) (2001) 4082–4084. [6] T.T. Basiev, P.G. Zverev, S.B. Mirov, in: C.E. Webb, J.D.C. Jones (Eds.), Handbook of Laser Technology and Applications, Taylor & Francis Group, CRC Press, Boca Raton, USA, 2003, pp. 499–522. [7] M. Matsubayashi, A. Faenov, T. Pikuz, Y. Fukuda, Y. Kato, Neutron imaging of micron-size structures by color center formation in LiF crystals, Nucl. Instr. Method A 622 (3) (2010) 637–641.
5
[8] S. Almaviva, F. Bonfigli, I. Franzini, A. Lai, R.M. Montereali, D. Pelliccia, A. Cedola, S. Lagomarsino, Hard X-ray contact microscopy with 250 nm spatial resolution using a LiF film detector and a tabletop microsource, Appl. Phys. Lett. 89 (5) (2006) 054102-1–054102-3. [9] A. Ustione, A. Cricenti, F. Bonfigli, F. Flora, A. Lai, T. Marolo, R.M. Montereali, G. Baldacchini, A. Faenov, T. Pikuz, L. Reale, Scanning near-field optical microscopy images of microradiographs stored in lithium fluoride films with an optical resolution of k/12, Appl. Phys. Lett. 88 (14) (2006) 141107-1– 141107-3. [10] P. Adam, S. Benrezzak, J. Bijeon, P. Royer, S. Guy, B. Jacquier, P. Moretti, R.M. Montereali, M. Piccinini, F. Menchini, F. Somma, C. Seassal, H. Rigneault, Fluorescence imaging of submicrometric lattices of colour centres in LiF by an apertureless scanning near-field optical microscope, Opt. Exp. 9 (7) (2001) 353–359. [11] J. Martin, L. Bischoff, R. Wannemacher, Microscopy of ion-beam generated fluorescent color-center patterns in LiF, Opt. Commun. 188 (1–4) (2001) 119– 128. [12] V.A. Skuratov, A.Yu. Didyk, S.M. Abu Al Azm, In-situ investigation of highenergy heavy ion irradiation effects: high-energy ionoluminescence of LiF, Rad. Phys. Chem. 50 (2) (1997) 183–188. [13] V.A. Skuratov, Luminescence of LiF and a-Al2O3 crystals under high energy irradiation, Nucl. Instr. Method B 146 (1–4) (1998) 385–392. [14] V.A. Skuratov, S.M. Abu Al Azm, V.A. Altynov, Luminescence of aggregate centers in lithium fluoride irradiated with high energy heavy ions, Nucl. Instr. Method B 191 (1–4) (2002) 251–255. [15] V.A. Skuratov, K.J. Gun, J. Stano, D.L. Zagorski, In situ luminescence as monitor of radiation damage under swift heavy ion irradiation, Nucl. Instr. Method B 245 (1) (2006) 194–200. [16] G. Baldacchini, O. Goncharova, V.S. Kalinov, R.M. Montereali, E. Nichelatti, A. Vincenti, A.P. Voitovich, Optical properties of coloured LiF crystals with given content of oxygen, hydroxyl and metal impurities, Phys. Status Sol. (c) 4 (3) (2007) 744–748. [17] N. Shiran, A. Belsky, A. Gektin, S. Gridin, I. Boiaryntseva, Radioluminescence of color centers in LiF crystals, Rad. Meas. 56 (2013) 23–26. [18] A. Perez, J. Davenas, C.H.S. Dupuy, Ionisation induced defects in alkali halide single crystals, Nucl. Instr. Method 132 (1976) 219–227. [19] C. Jardin, B. Canut, S.M.M. Ramos, The luminescence of sapphire subjected to the irradiation of energetic hydrogen and helium ions, J. Phys. D 29 (8) (1996) 2066–2070. [20] M. Fujiwara, T. Tanabe, H. Miyamaru, K. Miyazaki, Ion-induced luminescence in silica glasses, Nucl. Instr. Method B 116 (1–4) (1996) 536–541. [21] A. Quaranta, F. Gramegna, V. Kravchuk, C. Scian, Radiation damage mechanisms in CsI(Tl) studied by ion beam induced luminescence, Nucl. Instr. Method B 266 (12–13) (2008) 2723–2728. [22] A.P. Voitovich, V.S. Kalinov, L.P. Runets, A.P. Stupak, E.F. Martynovich, R.M. Montereali, G. Baldacchini, Color centers aggregation kinetics in lithium fluoride after gamma irradiation, J. Lumin. 143 (2013) 207–214. [23] A. Russakova, M.V. Sorokin, K. Schwartz, A. Dauletbekova, A. Akilbekov, M. Baizhumanov, M. Zdorovets, M. Koloberdin, Color center accumulation in LiF crystals under irradiation with MeV ions: optical spectroscopy and modeling, Nucl. Instr. Method B 313 (2013) 21–25. [24] M. Piccinini, F. Ambrosini, A. Ampollini, M. Carpanese, L. Picardi, C. Ronsivalle, F. Bonfigli, S. Libera, M.A. Vincenti, R.M. Montereali, Optical spectroscopy and imaging of colour centres in lithium fluoride crystals and thin films irradiated by 3 MeV proton beams, Nucl. Instr. Method B 326 (2014) 72–75. [25] M. Piccinini, F. Ambrosini, A. Ampollini, M. Carpanese, L. Picardi, C. Ronsivalle, F. Bonfigli, S. Libera, M.A. Vincenti, R.M. Montereali, Solid state detectors based on point defects in lithium fluoride for advanced proton beam diagnostics, J. Lumin. 156 (2014) 170–174. [26] L. Czopyk, G.A.P. Cirrone, G. Cuttone, M. Klosowski, P. Olko, U. Sroka, M.P.R. Waligórski, 2-D dosimetry of a proton radiotherapy beam using large-area LiF: Mg, Cu, P TL detectors, Rad. Meas. 43 (2–6) (2008) 977–980. [27] J.F. Ziegler, M.D. Ziegler, J.P. Biersac, SRIM – the stopping and range of ions in matter (2010), Nucl. Instrum. Method B 268 (11–12) (2010) 1818. [28] G. Baldacchini, Colored LiF: an optical material for all seasons, J. Lumin. 100 (1–4) (2002) 333–343. [29] M. Fox, Optical Properties of Solids, second ed., Oxford University Press, Oxford, 2010. [30] M.V. Sorokin, K. Schwartz, C. Trautmann, A. Dauletbekova, A.S. El-Said, Modeling of defect accumulation in lithium fluoride crystals under irradiation with swift ions, Nucl. Instr. Method B 326 (2014) 307–310.
Contents lists available at ScienceDirect
Optical Materials journal homepage: www.elsevier.com/locate/optmat
Ion beam induced luminescence analysis of defect evolution in lithium fluoride under proton irradiation A. Quaranta a,⇑, G. Valotto b, M. Piccinini c, R.M. Montereali c a
University of Trento, Department of Industrial Engineering, Via Sommarive 9, I-38132 Povo, Trento, Italy Department of Environmental Sciences, Informatics and Statistics, Università Ca’ Foscari Venezia, Dorsoduro 2137, I-30123 Venezia, Italy c ENEA C.R. Frascati, UTAPRAD-MNF, Photonics Micro and Nanostructures Laboratory, Via E. Fermi 45, 00044 Frascati, Rome, Italy b
a r t i c l e
i n f o
Article history: Received 24 May 2015 Received in revised form 15 August 2015 Accepted 16 August 2015
Keywords: Ion beam induced luminescence Lithium fluoride Multiple linear regression
a b s t r a c t Ion beam induced luminescence (IBIL) spectra of pure LiF under irradiation by a 2 MeV proton beam were analyzed as a function of the dose in order to deepen the kinetic mechanisms underlying the formation of luminescent point defects. The intensity evolution with dose at several emission wavelengths has been studied within a wide spectral interval, from ultraviolet (UV) to near infrared (NIR), and their different change rates have been correlated to the electronic defect formation processes. The intensity at few selected wavelengths was analyzed with a multiple linear regression (MLR) method in order to demonstrate that a linear calibration curve can be obtained and that an on-line optical dose monitor for ion beams can be realized. Ó 2015 Elsevier B.V. All rights reserved.
1. Introduction The exposure of pure LiF to different radiations, like ion beams, electrons, neutrons, gamma and X-rays or UV light, gives rise to the formation of color centers [1] characterized by fluorescence features whose light yield depends both on the radiation type and dose. The optical properties of such luminescent defects have been widely studied for several applications, like thermoluminescent dosimetry [2–4], miniaturized active channel waveguides [5], solid-state lasers [6] and neutron imaging detectors [7]. Moreover, the possibility to write luminescent submicrometricpatterns with X-rays, electron beams or ion beams and the stable formation of high contrast fluorescence images, has been demonstrated [8–11]. In spite of the great amount of work devoted to the analysis of LiF luminescent defects, some points are still to be fully understood, especially for what concerns the defects produced under ion beam irradiation. In this field, the most studied luminescence features are a structured peak in the UV and two broad bands, corresponding to optical transitions of F+3 and F2 centers, peaked around 540 nm and 670 nm, respectively. The UV peaks, studied mainly by Skuratov and co-workers with swift heavy ions [12–15], have been attributed to intrinsic or impurity/defect related self trapped exciton (STE) transitions, but at present a well
⇑ Corresponding author. E-mail address: [email protected] (A. Quaranta). http://dx.doi.org/10.1016/j.optmat.2015.08.015 0925-3467/Ó 2015 Elsevier B.V. All rights reserved.
defined origin is still to be established. Besides these main spectral features, other emission peaks are sometimes observed, like a band around 400 nm related to oxygen or Mg impurities [16,17] and the + NIR band of F 3 and F2 defects located around 900 nm [17,18]. A peculiar behaviour of ion beam induced luminescence (IBIL), also indicated as ionoluminescence (IL) by many authors, is that both intensity and shape of spectra change during irradiation owing to the point defect formation kinetic mechanisms. Typically, the yield decreases with the dose due to the formation of quenching or trapping centers produced by the irradiation. On the other hand, the intensity of features related to point defects increases during irradiation, due to the formation of luminescent centers induced by the ion beam itself, and then decreases, after reaching a maximum, when the center concentration is so high to trigger aggregation processes [19–21]. In lithium fluoride, it has been observed that the IBIL yield of F+3 and F2 bands follows the typical trend of point defects, while UV features, analyzed under swift heavy ion irradiation, monotonically decrease. The increasing rate depends on the defect formation mechanism, and it could be different for each defect. Up to now, the formation kinetics of F+3 and F2 centers has been deeply studied in samples irradiated with gamma-ray at low temperatures and then annealed at room temperature. In this case it was observed that the F+3 peak grows with a higher rate with respect to F2 owing to the diffusion of vacancies, allowing the reaction V+a + F2 ? F+3 [22]. On the contrary, under ion beam irradiation, the growth rates of the two bands are less studied. Only Russakova et al. made a
2
A. Quaranta et al. / Optical Materials 49 (2015) 1–5
comparison by means of optical analyses after heavy ion irradiation, finding similar increasing rates [23], but the conclusions came from the interpolation of few experimental points. In the same paper, the decreasing rate at higher doses has been examined. In particular, it has been observed that under irradiation with 14 MeV N2+ ions and 150 MeV Kr14+ ions the F+3 band yield decreases faster with respect to F2, and the effect has been attributed to optical self-absorption effects in the beam spot. Actually, at high doses the point defect luminescence bands decreases for several mechanisms which can be hardly distinguished, like the aggregation of the emitting centers into larger non-luminescent defects, concentration quenching or self-absorption. The aim of the present work is to study the behaviour of LiF point defects under proton beam irradiation by means of ion beam induced luminescence (IBIL) measurements. We focused on protons since they are widely used for applications ranging from material and device modification to radiobiology and radiotherapy. Very recently we started the investigation of the optical absorption and emission properties of CCs induced in LiF crystals and thin films by low energy protons. The stationary (after irradiation) visible photoluminescence (PL) spectra behaviour of radiationinduced color centers in LiF crystals was systematically studied as a function of irradiation fluence (interval of doses from 103 to 107 Gy), obtaining evidence of a linear optical response for the VIS peaks of F2 and F+3 color centers [24,25]. The detection of IBIL full spectra in real time during irradiation, presented here for the first time, allows to better understand the kinetics mechanisms involving the formation of luminescent point defects under ion irradiation. Moreover, the intensities as a function of the dose at selected wavelengths are interpolated by means of a multiple linear regression (MLR) method in order to demonstrate, at least from a conceptual point of view, that a linear relationship between the IBIL yield at some wavelengths and the dose can be obtained. This study can pave the way for the use of LiF crystals as on-line ion beam dose optical monitors, which could be particularly useful in hadrotherapy treatments [26].
allowing to follow the light intensity as a function of fluence within a wavelength interval ranging from 250 nm to 900 nm. This set-up allows on-line monitoring of the emitted intensity during proton irradiation, thus giving an exhaustive spectral analysis of the luminescence yield evolution. The spectra were collected with a resolution of 5 nm and they were not corrected for the response of the overall detection setup. 2.2. MLR analysis In order to analyze the possibility of extracting a calibration curve for the irradiation dose from IBIL spectra, we performed a multiple linear regression (MLR) analysis of the luminescence intensities at different wavelengths, with the aid of the code StatSoft 8.0. In particular, we fit the dose values D with a multiple linear function of the type:
D ¼ b0 þ
N X bi Ii
ð1Þ
i¼1
where bi are the fitting parameters and Ii are the intensities at the ki peak wavelengths taken as variables. For the fitting we used a forward stepwise method, by inserting one intensity variable at a time and by analyzing at every step the statistical significance of all the variables. The significance was evaluated with the parameters ‘‘tolerance”, ‘‘F to enter”, and ‘‘F to remove” set to 0.1, 100 and 10, respectively. ‘‘Tolerance” parameter is defined as 1 minus the squared multiple correlation of a selected variable with all the other independent variables used in the regression equation. Therefore, variables with small ‘‘tolerance” give a redundant contribution to the regression analysis. The ‘‘F to enter” (‘‘F to remove”) value determines how significant (insignificant) the contribution of a variable in the regression equation is, in order to be inserted (removed) into the equation. Finally, the R2 parameter was used as a performance indicator for the quality of the model. 3. Results and discussion
2. Materials and methods 3.1. IBIL spectra 2.1. IBIL measurements IBIL measurements were performed at the AN2000 accelerator of the INFN Laboratori Nazionali di Legnaro, by irradiating a nominally pure cleaved LiF crystal, of dimensions 10 10 1 mm3 at room temperature with a 2.0 MeV proton beam, perpendicular to the sample surface. The beam cross section was 1 mm2 and the current density was around 2.3 lA cm2, corresponding to an ion flux of 1.4 1013 H+ cm2 s1. During the ion irradiation, performed at 106 Torr, the chamber worked as a Faraday cup allowing to measure both the total charge impinging on the sample and the beam current. The maximum fluence value on the sample was 2.9 1015 H+ cm2. From the proton range (Rp = 37.9 lm in LiF calculated by SRIM2008 [27]) and taking into account the material density (q = 2.635 g/cm3) the fluence was converted into the average dose deposited along the ion path. In particular, the doses corresponding to the minimum and maximum fluence are 0.45 MGy and 92 MGy, respectively. The ionization energy of the impinging protons ranges from 33 eV ion1 nm1 at the sample surface, up to a maximum of 120 eV ion1 nm1 at the depth corresponding to the Bragg peak. The luminescence light was collected into the vacuum chamber by a silica fiber (600 lm diameter) put in front of the irradiated surface, at a distance of around 15 cm and at an angle of 20° with respect to the beam line direction. IBIL spectra were recorded by an Ocean Optics QE65000 spectrometer coupled to the chamber by means of a fiber vacuum connector. During the irradiation, 200 spectra were recorded, each one within 1 s of integration time,
In Fig. 1 are shown the IBIL spectra of a LiF crystal, collected at 1.4 MGy, 51 MGy and 93 MGy of irradiation dose, corresponding to
Fig. 1. IBIL spectra of the 2 MeV proton irradiated LiF crystal collected at 1.4 MGy, 51 MGy and 92 MGy of irradiation dose, corresponding to 4.4 1013 H+ cm2, 1.6 1015 H+ cm2 and 2.9 1015 H+ cm2, respectively. In the inset the UV peaks at higher doses are evidenced.
A. Quaranta et al. / Optical Materials 49 (2015) 1–5
fluences of 4.4 1013 H+ cm2, 1.6 1015 H+ cm2 and 2.9 1015 H+ cm2, respectively. In the UV range, the spectrum collected at the lower dose exhibits a peak at 320 nm, with a shoulder at 385 nm. By increasing the dose, all the peaks decreases leaving a faint peak at 330 nm, with a broad shoulder at 310 nm, and a peak at 395 nm, shown in the inset of Fig. 1. Skuratov and co-workers discussed the two main components in the UV range, reported at 296 and 330 nm [12,15], observed under swift heavy ion irradiation. The peak at 330 nm was univocally attributed to the STE recombination, while the lower wavelength feature has been subjected to different interpretations. In particular, it has been sometimes attributed to the STE recombination near an unstable defect [12], or near impurities [13]. The feature at higher wavelengths, reported in literature at 400 nm [13], has been identified as an impurity related band. The intensity of F+3 and F2 bands at 540 and 670 nm increases from 1.4 MGy to 51 MGy and then decreases at 92 MGy. Besides these features, a faint shoulder at 740 nm, attributed to F4-like + centers, and an intense band around 890 nm, due to F 3 and F2 centers [22,28], can be observed (the small valley at 875 nm is an instrumental artifact). A more complete representation of the IBIL spectra evolution can be observed in Fig. 2, where a 2-D plot is shown reporting the intensity at each wavelength normalized to its own maximum value. As it can be observed, two ranges with different trends can be identified: in the range from UV to 450 nm the intensity monotonically decreases with different rates, whereas at wavelengths higher than 500 nm all the luminescence yields increase, reaching a maximum at doses ranging from 10 to 30 MGy, and then monotonically decrease. In Fig. 3a and b are reported the effective (a) and normalized to the maximum (b) intensities as a function of the dose at five representative wavelengths, namely 315 nm, 385 nm, 545 nm, 670 nm and 890 nm. The yields at 315 and 385 nm monotonically decrease with different rates, since the features originate from different emitting centers, and definitely indicates that these features are not related to point defects, but to STE recombination processes in the lattice or near impurities. Skuratov et al. observed indeed a monotonic decrease [13,15], but the high ionization energy density deposited by swift heavy ions could induce to some misleading interpretation, since it should be possible that the dose rate release is so high to overcome the yield turning point just from the irradiation beginning. The same trend observed under proton irradiation, with a ionization energy orders of magnitude lower, is the definite evidence that STE giving the UV features decays in the lattice or near impurities. The luminescence yields at 545 nm, 670 nm and 890 nm increase with dose up to 11 MGy, 14 MGy and 37 MGy (3.4 1014, 4.4 1014 and 1.1 1015 H+ cm2), respectively. The yields of F+3 centers at 545 nm and of F2 centers at 670 nm increase with a slightly different rate. This trend is opposite to the behaviour observed during the room temperature thermal annealing of irradiated crystals [22]. It has to be taken into account that a 2 MeV proton can excite a around 1.5 105 electrons in the conduction band (value obtained by dividing the proton energy by the LiF band gap of 13.7 eV [29]). It gives an average number of 2 1018 electrons cm2 migrating through the lattice and reaching point defects. So, under beam irradiation, the large number electrons scattered by the ions through the crystal enhance the annihilation probability of F+3 centers through the reaction F+3 + e ? F3, thus slowing down their formation rate [23,30]. After reaching the maximum, the two features decrease during irradiation with a quite similar rate. Russakova et al. observed a
3
faster decreasing rate for F+3 centers, attributed mainly to selfabsorption effects induced by the heavy damage released by high energy heavy ions [23]. In the sample studied in this work, the colored spot left by the ion beam on the sample surface was very faint and optical self-absorption effects could be discarded in this work. The decrease of the two yields with the same rate, here indicates that the two kinds of defects under proton beam irradiation are subjected to the same aggregation processes giving rise to larger structures quenching the luminescence. The luminescence intensity at 890 nm, whose trend as a function of the dose is observed here for the first time, reaches its maximum at higher dose and decreases with a slower rate with respect to F+3 and F2 centers. The lower increasing rate can be explained by the need of a more complex reaction for the production of the centers which are responsible of this feature. For instance, F 3 centers can be formed either by the combination of a F2 center with a F defect, or by trapping two electrons into a F+3 center. The former mechanism can be realized when the proper local concentrations of F centers are achieved by the ion beam induced damage. At the same time, scattered electrons can promote the reactions e + F ? F and 2e + F+3 ? F 3 . The slower yield decrease of the NIR feature is less clear for the moment. Shiran et al. [17] observed that in cathodoluminescence spectra with high e-beam density, NIR peaks are practically unchanged, while visible features of F2 and F+3 centers disappears. This result points out that centers giving rise to high wavelengths emissions are more resistant under irradiation. Moreover, the presence of an energy transfer process from F2 centers to NIR emitting defects, contributing to the lower decreasing rate of this feature, cannot be neglected [28]. 3.2. MLR analysis IBIL spectra of materials like LiF are characterized by a strongly non-linear behaviour as a function of the dose or fluence. This behaviour changes for the different spectrum wavelengths due to the different nature of the luminescent centers and to their response to ion irradiation. In this work we want to study how the pattern of the yield variations as a function of the dose can be exploited for realizing dose or beam optical monitoring systems. For the MLR interpolation the yields at wavelengths near the main peaks, namely at 312 nm, 383 nm, 537 nm, 664 nm and 855 nm, were taken as main variables. Following the stepwise procedure previously described, we observed that 383 nm and 537 nm did not satisfy both ‘‘tolerance” and ‘‘F” values. In fact, the intensities at these wavelengths evolve in a way similar to the values at 312 and 664 nm, respectively, and they do not add significant information to the MLR analysis. By keeping only the intensities at 664 and 312 nm as variables, we obtained a R2 value of 0.92. This condition gives a poor prediction performance for higher and lower dose values used in this work. By including as a variable the intensity at 855 nm, the R2 value increased up to 0.99 with a remarkable improvement of the prediction performance. In Table 1 are reported the MLR coefficients calculated with the intensities at the three selected wavelengths. Due to the decrease of the luminescence yield over a certain dose value, all the coefficients are negative. In the same Table the standardized regression coefficients are shown. These parameters were obtained by fitting the standardized variables, that is the intensities normalized in order to give mean intensity equal to 0 and standard deviation equal to 1. Standardized regression coefficients allow to compare the relative contribution of each independent variable in the prediction of the dose value. In our case, the most influent variable is the yield at 312 nm, followed by 855 nm and 664 nm. In Fig. 4 are reported the experimental dose values (indicated as ‘‘observed” in the ordinate axis) as a function of the ‘‘predicted”
4
A. Quaranta et al. / Optical Materials 49 (2015) 1–5
Fig. 2. Normalized intensities of IBIL spectra as a function of both the dose and the wavelength. For each wavelength, the intensity was normalized to its own maximum value. Horizontal lines are due to fluctuations of the ion beam current.
Table 1 MLR regression coefficients (RC) and the corresponding standard errors (SE) for the intercept b0 and the wavelength intensities bk . The standardized regression coefficients (SRC) where calculated with the by fitting the standardized variables, where the mean intensity value and the respective standard deviation of each variable was set to 0 and 1, respectively.
b0 b312 b664 b855
RC
RC–SE
SRC
SRC–SE
20.0 58 4.1 23.9
0.3 1 0.2 0.8
– 1.22 0.34 0.71
– 0.02 0.01 0.02
Fig. 4. Observed vs. predicted dose values obtained from the MLR of the luminescence yield at 312 nm, 664 nm and 855 nm. Circles are the experimental points and the red line is the interpolation curve obtained from MLR. Dashed lines indicate the 95% confidence interval for the interpolation curve.
Fig. 3. IBIL yield as a function of the dose as collected (a) and normalized to the maximum (b) recorded at five wavelengths representative of the main spectral features.
values, obtained by substituting the corresponding IBIL intensities at the selected wavelengths in the MLR interpolation formula. As it can be observed, the points are well aligned along an interpolating line with slope 1, also reported in the figure. The dashed lines indicate the 95% confidence interval for the interpolation.
It is worth noting that a very good correlation was obtained with only three variables, allowing to realize a reliable calibration curve with a low waste of time and calculations. A standard error of around 3 MGy was evaluated in the estimation of the predicted dose from the MLR curve. This value is also related to the beam current fluctuations and could be improved by providing a set-up for recording in real time the charge corresponding to every single spectrum and by normalizing the spectrum to this value. Further improvements could be obtained by focusing the analysis on narrower dose intervals, suited for monitoring purposes.
A. Quaranta et al. / Optical Materials 49 (2015) 1–5
Besides these considerations, it is worth noting that this result is the first proof of concept demonstrating that IBIL spectra are suitable for the realization of on-line optical monitoring systems. 4. Conclusions In this work we have studied the detailed IBIL yield evolution of a pure LiF crystal under proton irradiation. The monotonic decrease of the luminescence intensity in the UV range with increase of dose allows to conclude that the emission features are related to STE transitions near lattice impurities rather than to point defects. + Concerning point defects, the formation of F2, F+3, and F 3 /F2 centers has been monitored by the intensity changes of the corresponding luminescence bands. The formation rate of F2 centers is slightly higher than F+3 centers, owing to the annealing of the latter by means of scattered electrons. On the other hand, the decreasing rate of the two yields is quite similar evidencing similar processes of aggregation and quenching for the two defects. + The evolution of the NIR F 3 /F2 band as a function of the dose was monitored here for the first time, evidencing that it increases and decreases with slower rates with respect to the other point defects, due to both more complex formation mechanisms and higher radiation hardness with respect to the other centers. Moreover energy transfer effects from F2 centers cannot be discarded. In order to study how the IBIL yield as a function of the dose can be used for the realization of beam or dose monitoring systems, a MLR analysis was performed by interpolating the luminescence intensity at few selected wavelengths. With the intensities at three wavelengths a very good regression was obtained, and a calibration curve for the interval from 1 to 90 MGy was found. Even better interpolations could be obtained by selecting narrower does ranges or by controlling the collected charge in real time. Further work is planned by analyzing several crystals of the same type in order to study the reproducibility of the approach for the production of reliable IBIL based dosimeters. References [1] W.B. Fowler, Physics of Color Centers, Academic Press, New York and London, 1968. [2] W.L. McLaughlin, A. Miller, S.C. Ellis, A.C. Lucas, B.M. Kapsar, Radiation-induced color centers in LiF for dosimetry at high absorbed dose rates, Nucl. Instr. Method 175 (1980) 17–18. [3] G. Baldacchini, A.T. Davidson, V.S. Kalinov, A.G. Kozakiewicz, R.M. Montereali, E. Nichelatti, A.P. Voitovich, Thermoluminescence of pure LiF crystals and color centers, J. Lumin. 122–123 (2007) 371–373. [4] D.R.S. Ribeiro, D.N. Souza, A.F. Maia, S.L. Baldochi, L.V.E. Caldas, Applicability of pure LiF in dosimetry, Rad. Meas. 43 (2–6) (2008) 1132–1134. [5] R.M. Montereali, M. Piccinini, E. Burattini, Amplified spontaneous emission in active channel waveguides produced by electron-beam lithography in LiF crystals, Appl. Phys. Lett. 78 (26) (2001) 4082–4084. [6] T.T. Basiev, P.G. Zverev, S.B. Mirov, in: C.E. Webb, J.D.C. Jones (Eds.), Handbook of Laser Technology and Applications, Taylor & Francis Group, CRC Press, Boca Raton, USA, 2003, pp. 499–522. [7] M. Matsubayashi, A. Faenov, T. Pikuz, Y. Fukuda, Y. Kato, Neutron imaging of micron-size structures by color center formation in LiF crystals, Nucl. Instr. Method A 622 (3) (2010) 637–641.
5
[8] S. Almaviva, F. Bonfigli, I. Franzini, A. Lai, R.M. Montereali, D. Pelliccia, A. Cedola, S. Lagomarsino, Hard X-ray contact microscopy with 250 nm spatial resolution using a LiF film detector and a tabletop microsource, Appl. Phys. Lett. 89 (5) (2006) 054102-1–054102-3. [9] A. Ustione, A. Cricenti, F. Bonfigli, F. Flora, A. Lai, T. Marolo, R.M. Montereali, G. Baldacchini, A. Faenov, T. Pikuz, L. Reale, Scanning near-field optical microscopy images of microradiographs stored in lithium fluoride films with an optical resolution of k/12, Appl. Phys. Lett. 88 (14) (2006) 141107-1– 141107-3. [10] P. Adam, S. Benrezzak, J. Bijeon, P. Royer, S. Guy, B. Jacquier, P. Moretti, R.M. Montereali, M. Piccinini, F. Menchini, F. Somma, C. Seassal, H. Rigneault, Fluorescence imaging of submicrometric lattices of colour centres in LiF by an apertureless scanning near-field optical microscope, Opt. Exp. 9 (7) (2001) 353–359. [11] J. Martin, L. Bischoff, R. Wannemacher, Microscopy of ion-beam generated fluorescent color-center patterns in LiF, Opt. Commun. 188 (1–4) (2001) 119– 128. [12] V.A. Skuratov, A.Yu. Didyk, S.M. Abu Al Azm, In-situ investigation of highenergy heavy ion irradiation effects: high-energy ionoluminescence of LiF, Rad. Phys. Chem. 50 (2) (1997) 183–188. [13] V.A. Skuratov, Luminescence of LiF and a-Al2O3 crystals under high energy irradiation, Nucl. Instr. Method B 146 (1–4) (1998) 385–392. [14] V.A. Skuratov, S.M. Abu Al Azm, V.A. Altynov, Luminescence of aggregate centers in lithium fluoride irradiated with high energy heavy ions, Nucl. Instr. Method B 191 (1–4) (2002) 251–255. [15] V.A. Skuratov, K.J. Gun, J. Stano, D.L. Zagorski, In situ luminescence as monitor of radiation damage under swift heavy ion irradiation, Nucl. Instr. Method B 245 (1) (2006) 194–200. [16] G. Baldacchini, O. Goncharova, V.S. Kalinov, R.M. Montereali, E. Nichelatti, A. Vincenti, A.P. Voitovich, Optical properties of coloured LiF crystals with given content of oxygen, hydroxyl and metal impurities, Phys. Status Sol. (c) 4 (3) (2007) 744–748. [17] N. Shiran, A. Belsky, A. Gektin, S. Gridin, I. Boiaryntseva, Radioluminescence of color centers in LiF crystals, Rad. Meas. 56 (2013) 23–26. [18] A. Perez, J. Davenas, C.H.S. Dupuy, Ionisation induced defects in alkali halide single crystals, Nucl. Instr. Method 132 (1976) 219–227. [19] C. Jardin, B. Canut, S.M.M. Ramos, The luminescence of sapphire subjected to the irradiation of energetic hydrogen and helium ions, J. Phys. D 29 (8) (1996) 2066–2070. [20] M. Fujiwara, T. Tanabe, H. Miyamaru, K. Miyazaki, Ion-induced luminescence in silica glasses, Nucl. Instr. Method B 116 (1–4) (1996) 536–541. [21] A. Quaranta, F. Gramegna, V. Kravchuk, C. Scian, Radiation damage mechanisms in CsI(Tl) studied by ion beam induced luminescence, Nucl. Instr. Method B 266 (12–13) (2008) 2723–2728. [22] A.P. Voitovich, V.S. Kalinov, L.P. Runets, A.P. Stupak, E.F. Martynovich, R.M. Montereali, G. Baldacchini, Color centers aggregation kinetics in lithium fluoride after gamma irradiation, J. Lumin. 143 (2013) 207–214. [23] A. Russakova, M.V. Sorokin, K. Schwartz, A. Dauletbekova, A. Akilbekov, M. Baizhumanov, M. Zdorovets, M. Koloberdin, Color center accumulation in LiF crystals under irradiation with MeV ions: optical spectroscopy and modeling, Nucl. Instr. Method B 313 (2013) 21–25. [24] M. Piccinini, F. Ambrosini, A. Ampollini, M. Carpanese, L. Picardi, C. Ronsivalle, F. Bonfigli, S. Libera, M.A. Vincenti, R.M. Montereali, Optical spectroscopy and imaging of colour centres in lithium fluoride crystals and thin films irradiated by 3 MeV proton beams, Nucl. Instr. Method B 326 (2014) 72–75. [25] M. Piccinini, F. Ambrosini, A. Ampollini, M. Carpanese, L. Picardi, C. Ronsivalle, F. Bonfigli, S. Libera, M.A. Vincenti, R.M. Montereali, Solid state detectors based on point defects in lithium fluoride for advanced proton beam diagnostics, J. Lumin. 156 (2014) 170–174. [26] L. Czopyk, G.A.P. Cirrone, G. Cuttone, M. Klosowski, P. Olko, U. Sroka, M.P.R. Waligórski, 2-D dosimetry of a proton radiotherapy beam using large-area LiF: Mg, Cu, P TL detectors, Rad. Meas. 43 (2–6) (2008) 977–980. [27] J.F. Ziegler, M.D. Ziegler, J.P. Biersac, SRIM – the stopping and range of ions in matter (2010), Nucl. Instrum. Method B 268 (11–12) (2010) 1818. [28] G. Baldacchini, Colored LiF: an optical material for all seasons, J. Lumin. 100 (1–4) (2002) 333–343. [29] M. Fox, Optical Properties of Solids, second ed., Oxford University Press, Oxford, 2010. [30] M.V. Sorokin, K. Schwartz, C. Trautmann, A. Dauletbekova, A.S. El-Said, Modeling of defect accumulation in lithium fluoride crystals under irradiation with swift ions, Nucl. Instr. Method B 326 (2014) 307–310.