Trap centers in molybdates

Optical Materials 35 (2013) 2465–2472

Contents lists available at SciVerse ScienceDirect

Optical Materials journal homepage: www.elsevier.com/locate/optmat

Trap centers in molybdates D.A. Spassky a,b,⇑, V. Nagirnyi a, V.V. Mikhailin b, A.E. Savon b, A.N. Belsky c, V.V. Laguta d, M. Buryi d, E.N. Galashov e, V.N. Shlegel f, I.S. Voronina g, B.I. Zadneprovski h a

Institute of Physics, University of Tartu, Riia 142, Tartu 51014, Estonia Skobeltsyn Institute of Nuclear Physics, M.V. Lomonosov Moscow State University, Moscow 119991, Russia c Institute of Light and Matter, CNRS, University Lyon1, Villeurbanne 69622, France d Institute of Physics AS CR, Cukrovarnicka 10, Prague 162 00, Czech Republic e Novosibirsk State University, Novosibirsk 630090, Russia f Nikolaev Institute of Inorganic Chemistry SB RAS, Novosibirsk 630090, Russia g A.M. Prokhorov General Physics Institute of RAS, Vavilov str. 38, Moscow 119991, Russia h Central Research and Development Institute of Chemistry and Mechanics, Moscow 115487, Russia b

a r t i c l e

i n f o

Article history: Received 1 June 2013 Received in revised form 21 June 2013 Accepted 30 June 2013 Available online 20 July 2013 Keywords: Molybdate single crystals Thermostimulated luminescence Scintillating bolometers Self-trapping of charge carriers

a b s t r a c t Charge carrier trapping centers have been studied in molybdates CaMoO4, SrMoO4 and PbMoO4 with the scheelite crystal structure as well as in ZnMoO4, which crystallize in a-ZnMoO4 structural type. The trap parameters such as activation energies and frequency factors have been determined. It is shown for the first time that both electrons and holes are trapped by the elements of regular crystal structure in ZnMoO4. The effect of the charge carrier trapping on luminescence properties is demonstrated. Potential influence of the traps on the scintillation process is discussed. Ó 2013 Elsevier B.V. All rights reserved.

1. Introduction The scintillation properties of crystals are commonly dependent on the presence of traps in the material. Intermediate localization of charge carriers at shallow traps may result in the appearance of slow components in scintillation decay curves while the suppression of light yield may be expected in case of deep traps when the energy of thermal vibrations is not sufficient for the release of the trapped charge carriers. Molybdate single crystals are perspective for the cryogenic scintillating bolometers that operate in ultra-low temperature conditions (tens of mK) [1]. In view of such application, a set of promising molybdate crystals including CaMoO4 [2–5], SrMoO4 [1], PbMoO4 [1,6,7], Li2MoO4 [8,9], Li2Zn2(MoO4)3 [10] and ZnMoO4 [11–13] have been recently studied. Special attention has been paid to the following two representatives of the molybdate family. Calcium molybdate is a well-studied compound for cryogenic detectors [2–5], which has already been chosen as a cryogenic scintillator for the AMoRE experiment [14]. Zinc molybdate is a new crystal and only recently it has been successfully grown in a bulk form using the conventional ⇑ Corresponding author at: Institute of Physics, University of Tartu, 51014, Estonia. E-mail address: [email protected] (D.A. Spassky). 0925-3467/$ - see front matter Ó 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.optmat.2013.06.054

Czochralski technique [15,16] or the low temperature gradient Czochralski technique which allowed obtaining crystals with better optical quality [17,18]. ZnMoO4 attracts enhanced attention also as a material for cryogenic experiments due to its high energy resolution and efficient a background rejection [11–13], however the light yield of this crystal is rather low. A measurement of the absolute value of a scintillation light yield at the operating temperatures of cryogenic scintillators is a complicated task. The value of energy yield was estimated at 1.1 keV/MeV only for ZnMoO4 [13]. Taking into account the average energy of a single emission photon as 2.0 eV [18] the light yield of ZnMoO4 can be estimated to be 550 ph/MeV. The data on the light yield of molybdates are commonly presented with respect to the light yield of CaWO4 at temperatures around 10 K. The light yield of CaMoO4 is 95% of that in CaWO4 [19] while it is only 34% in PbMoO4 [6]. Taking into account the estimation of the absolute light yield of BGO as 23,700 ph/MeV which has been shown to be 150% of that in CaWO4 at 6 K [19], one can conclude that the absolute value is about 15000 ph/MeV for CaMoO4 and 5500 ph/MeV for PbMoO4. Therefore, at cryogenic temperatures, the light yield of ZnMoO4 is 30 times lower than the light yield of CaMoO4. Such low value of the light yield in ZnMoO4 may be related to the strong decrease of luminescence intensity upon temperature decrease from 100 to 50 K, which has been observed under VUV [18,20] and X-ray [11] irradiation

2466

D.A. Spassky et al. / Optical Materials 35 (2013) 2465–2472

of the crystal in good correlation with the position of a strong peak of thermostimulated luminescence at 70 K. Charge carrier trapping is a serious problem for cryogenic scintillators. Under the operating conditions, even shallow traps can have a strong impact on scintillation efficiency due to capturing separated charge carriers. Therefore, the study of the origin of traps in molybdates is of crucial importance for understanding their scintillation properties at low temperatures. By using the EPR and TSL techniques it has been shown for some molybdates that charge carriers are trapped by regular complexes [21–23]. Despite the quite low energy of the thermal release from intrinsic trap centers, the trapping itself will definitely worsen the scintillation properties at low temperatures. The influence may be expected even for the most perfect crystals with low content of uncontrollable impurities and crystal structure defects, which are commonly introduced during the crystal growth. Here we discuss the origin of trap centers for the scheelite type crystals CaMoO4, SrMoO4, PbMoO4 as well as for ZnMoO4, which are considered for application as cryogenic scintillators. For the first time the origin of the traps is determined for ZnMoO4 by using the EPR and TSL techniques. It is shown that the temperature dependence of the luminescence intensity under VUV and X-ray excitation is affected by the presence of traps. The influence of the traps on the scintillation process is discussed. 2. Experimental details CaMoO4, SrMoO4 and PbMoO4 single crystals were grown by the conventional Czochralski method in a Pt crucible in air. Samples of ZnMoO4 were grown using low temperature gradient Czochralski technique that allows growing crystals with improved optical properties [17]. The concentrations of contaminating impurities were determined by atomic-emission spectral analysis. This method allows determining the presence and concentration of 72 different elements from Li to U. The atoms of Nb and Ta belong to the structural materials of the ion source used in the analysis and their concentration in the samples was not detected. The data on the impurities detected in the investigated samples is summarized in Table 1. The concentrations of other impurities were below 3 ppm. It follows from the table that the total concentration of impurities remained between 100 and 400 ppm. Tungsten impurity which is an attendant of Mo was detected in all the samples studied, whereas it was a dominating impurity in PbMoO4 and ZnMoO4. X-ray excited luminescence, thermostimulated luminescence (TSL) curves and spectra were measured at the Laboratory PCML, Claude Bernard Lyon University. An X-ray source with a tungsten anode operating at U = 30 keV was used as an excitation source. Luminescence characteristics under excitation in UV and VUV spectral region have been measured at the SUPERLUMI station in the synchrotron radiation channel of the storage ring DORIS III (DESY, Hamburg) [24], and at the laboratory setup of the Institute of Physics, University of Tartu. All measurements have been carried out on the freshly cleaved vitreous surfaces of the samples. The EPR measurements for ZnMoO4 were performed at 9.23 GHz with the standard 3 cm wavelength of the EPR spectrometer in the temperature range 10–290 K using an Oxford Instrument cryostat. The

Table 1 The concentration of contaminating impurities in the studied crystals according to the data of atomic-emission spectral analysis. CaMoO4 SrMoO4 PbMoO4 ZnMoO4

Ba (100 ppm), Sr (60 ppm), Na (30 ppm), Ag (10 ppm), W (10 ppm) Si (70 ppm), Ca (20 ppm), Cl (15 ppm), W (10 ppm), Ba (10 ppm) W (300 ppm), Ca (40 ppm), S (10 ppm), Bi (4 ppm), K (4 ppm) W (200 ppm), Si (40 ppm), Cd (4 ppm)

sample was directly irradiated in the spectrometer cavity by X-rays at temperature about 30 K. 3. Experimental results 3.1. Characterization of traps in molybdates The TSL curves and spectral composition of the TSL peaks of molybdates are presented in Figs. 1–4 for the temperature region 10–300 K. The calculated trap parameters and spectral positions of the emission bands in TSL peaks are summarized in Table 2. Almost all TSL peaks were fitted using the first order decay approximation supposing that the probability for a charge carrier released from traps to be captured by traps again is much lower than the probability of radiative recombination. In this case, we used the following fitting expression:

EA x0 kB T 2 ðtÞ EA =kB TðtÞ Ilum ðtÞ ¼ nð0Þx0 exp  e  kB TðtÞ EA T 0 ðtÞ

! ð1Þ

where n(0) is the concentration of filled traps just after the end of sample excitation (here it is in arbitrary units), EA is activation energy of the trap, x0 is frequency factor of the trap, kB is Boltzmann constant, T is the temperature and T0 is heating rate [25]. However for a pronounced peak at 43 K in PbMoO4 and for a peak at 166 K in SrMoO4 the fitting with this formula gives unsatisfactory result (see the unsatisfactory approximation for 43 K peak in PbMoO4 in Fig. 3a). The symmetric shape of these peaks without distinctive bends does not allow to suggest their complex structure and to perform fitting with several peaks in the first order decay

Fig. 1. (a) TSL curve of CaMoO4. Thin red line represents the fitting of TSL curve with the trap parameters that are presented in Table 2. In the inset-emission under X-ray excitation and spectral composition of TSL recorded at T = 10–30 K (all-curve 1), T = 40–55 K (2) and T = 120–160 K (3) and (b) temperature dependence of the luminescence intensity measured at Eex = 4.1 eV (1), Eex = 11 eV (2), X-ray excitation (3) and of the scintillation output under excitation with a-particles (4) presented in [19]. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

D.A. Spassky et al. / Optical Materials 35 (2013) 2465–2472

Fig. 2. (a) TSL curve of SrMoO4. Thin red lines represent the fitting of TSL curve with the trap parameters that are presented in Table 2. In the inset – emission under Xray excitation and spectral composition of TSL recorded at T = 20–52 K, 95–120 K and 145–190 K (all-curve 1) and T = 190–240 K (2) and (b) temperature dependence of the luminescence intensity measured at Eex = 4.27 eV (1), Eex = 7.3 eV (2) and Xray excitation (3). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

approximation. We also used the second order decay approximation for the fitting of these peaks. In this approximation, the probability of a released charge carrier to be retrapped is supposed to be much higher than the probability to relax radiatively and the peak shape can be fitted with formula:

Ilum ðtÞ ¼ 

n2 ð0Þx0 T 0 ðtÞeEA =kB TðtÞ 1þ

x0 nð0ÞkB T 2 ðtÞ E =k TðtÞ e A B 0

2

ð2Þ

EA T ðtÞ

Here the parameters nomenclature is the same as for the first order decay approximation except x0 which is expressed as x0 ¼ x0 =cN0 , where N0 is the concentration of traps and c is the coefficient of proportionality [25]. As far as the absolute value of traps concentration is not known, we cannot determine the frequency factor x0 for the case of the second order approximation and it is not shown in Table 2. The spectral composition for relatively low-temperature TSL peaks in the molybdates is similar to that of the intrinsic emission observed under steady excitation with X-rays. The intrinsic luminescence of molybdates is characterized by a single broad emission band that is commonly ascribed to the radiative annihilation of the excitons self-trapped at the MoO4 complexes. In some cases (e.g., by using time-resolved spectroscopy) it is possible to observe a complex structure of luminescence spectra [18,26–28], which is attributed to the emission of singlet and triplet self-trapped excitons (STEs) [26–28]. A slight shift of the emission band observed for the TSL peaks at 35 K and 76 K in ZnMoO4 follows the shift of the STE emission band measured under steady X-ray excitation.

2467

Fig. 3. (a) TSL curve of PbMoO4. Thin red lines represent the fitting of TSL curve with the trap parameters that are presented in Table 2. Dashed green line represents the fitting of peak at 43 K with the first order kinetic model. In the insetemission under X-ray excitation and spectral composition of TSL recorded at T = 15– 22 K and 30–60 K (all-curve 1) and T = 75–90 K and 100–120 K (2) and (b) temperature dependence of the luminescence intensity measured at Eex = 3.5 eV (1), Eex = 11 eV (2), X-ray excitation (3) and of the scintillation output under excitation with a-particles (4) presented in [6]. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

This shift has been ascribed to the complex structure of the intrinsic emission band [18]. Relatively high-temperature TSL peaks of other molybdates demonstrating an emission band shifted to the low-energy region with respect to the STE emission can be attributed to defect-related emission centers. Charge carrier trapping affects the process of energy transfer to emission centers. The temperature dependence of luminescence intensity measured at different excitation energies is presented for molybdates in Figs. 1–4 (plot b). When the luminescence is excited in the region of direct creation of excitons the temperature dependence is mostly determined by the intracenter thermal quenching process. In such case, the dependence can be fitted with the Mott formula [29] allowing the estimation of activation energy of the process. Slight deviations from the simple Mott formula can appear due to an additional decay channel connected with the thermal disintegration of STE as it was proposed for PbWO4 in [30]. The deviation is significant for the temperature dependence of SrMoO4, in which the slope of the curve is very flat and cannot be fitted with the Mott formula with reasonable parameters. The origin of such behavior is not clear. For the excitation with VUV radiation and X-rays the temperature dependences become more complicated. The most distinctive features in the temperature dependences of emission intensities measured under such excitation are correlated to the positions of the intensive TSL peaks, which is connected with the partial trapping of separated electrons and holes. As a result, the emission intensity decreases in the region of the low-temperature slope of a corresponding TSL peak. The correlation is most pronounced for PbMoO4 and especially for ZnMoO4. We have performed additional experiments in order to determine which part of excited charge carriers are localized on traps

2468

D.A. Spassky et al. / Optical Materials 35 (2013) 2465–2472

Fig. 4. (a) TSL curve of ZnMoO4. Thin red lines represent the fitting of TSL curve with the trap parameters that are presented in Table 2. In the inset-emission under X-ray excitation at T = 60 K and spectral composition of TSL recorded at T = 50–85 K and 85–100 K (all-curve 1) and T = 28–38 K (2), (b) temperature dependence of the luminescence intensity measured at Eex = 4.1 eV (1), Eex = 11 eV (2) and X-ray excitation (3). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

(Fig. 5). Each sample was irradiated for 600 s with X-rays and then kept at low temperature in order to observe phosphorescence. The phosphorescence has been detected for PbMoO4 and ZnMoO4. 700 s after irradiation the samples were heated with the constant rate 10 K/min and TSL curves were recorded. The luminescence was registered with a CCD detector, integrating the signal over

Table 2 Data on the parameters of TSL peaks in molybdates. Crystal

Peak maximum (K)

Activation energy (meV)

Frequency factor (s1)

kmax (nm)a

CaMoO4

20 52 150 32 37 50 103 112 166 205 238 19 43 85 110 35 76 92

20 112 262 47 52 67 238 310 337 354 370 21 57 (50)b 229 260 20 62 75

8.2  103 7.8  1010 1.7  107 3.1  106 9.5  105 5  105 1.5  1010 2.5  1012 – 7.7  106 9.2  105 6.1  104 (343.2)b 1.8  1012 2.5  1011 10 2.4  102 2  102

537 555 590 525 525 525 530 530 535 565 565 530 530 567 567 563 588 588

SrMoO4

PbMoO4

ZnMoO4

a

The maximum of the emission band in the corresponding TSL peak. In the brackets the parameters of TSL peak at 43 K in PbMoO4 are presented which was approximated with the first order kinetics. b

the spectral range of the luminescence band(s) every 6 s during the whole time of the experiment, i.e. during the irradiation, phosphorescence measuring and heating. The ratio of the areas under the TSL curves to the areas under the curves obtained during 600 s under excitation allows estimating the part of the trapped charge carriers relatively to the part of charge carriers involved in luminescence under steady excitation. It should be noted that the non-radiative recombination of electron–hole pairs is not considered in such estimation, while its rate may differ in the crystals studied. However, the performed experiment allows doing certain conclusions. For CaMoO4 and SrMoO4 the part of trapped charge carriers is estimated as 3–5%. A considerably larger storage efficiency is observed for PbMoO4 (22%) and especially for ZnMoO4 (150%). Therefore in ZnMoO4, the probability for charge carriers to be captured by traps is even higher than the probability to emit light under X-ray excitation. The result cannot be explained by the enhanced concentration of uncontrollable impurities and structural defects in the given sample of ZnMoO4. The concentrations of the impurities are approximately at the same levels as in other studied molybdates (Table 1). The studied crystal was also of high optical quality without visible inclusions. Moreover, the storage efficiency measured for the sample of yellowish ZnMoO4 that was grown using the conventional Czochralski method and studied in [20], was even lower than that in the best ZnMoO4 available. Therefore, we suppose that the high storage efficiency of ZnMoO4 is not connected with the contamination of the crystal with impurities but it is due to fundamental reasons. 3.2. Study of trap centers origin in ZnMoO4 using TSL and EPR In order to obtain information about the origin of traps in ZnMoO4 and to clarify the reason of high trapping probability of charge carriers we have performed the EPR studies of irradiated samples. The EPR spectra recorded from a sample X-ray irradiated at 30 K for 0.5 h are shown in Fig. 6. The spectra of two centers with central lines at 3270 and 3470 G can be clearly distinguished. Let us label them as center #1 and #2, respectively. The spectral line of the center #1 varies with the crystal rotation in the g factor range of 2.001–2.028. One can thus conclude that this center is of a hole-type, most probably an O ion created as a result of a hole self-trapping at a lattice oxygen ion. The second center is characterized by the g factors in the range of 1.858–1.927. These g factors are typical of electron-type centers. One can assume that this center is created by the self-trapping of an electron by the (MoO4)2 complex, like this takes place in CaWO4 or PbWO4 [31–33]. This is confirmed by the observation of the well-resolved hyperfine lines from 95,97Mo isotopes which have the nuclear spins 5/2 and 7/2 with almost equal magnetic moments. The total natural abundance of these two isotopes is 25.2%. The O center also shows similar 95,97Mo hyperfine structure with a smaller splitting of the hyperfine lines. Both centers are not visibly perturbed by other defects and are thermally stable only to about 50 K. Detailed analysis of EPR spectra will be presented in a separate paper. The temperature dependence of the EPR signals of both centers along with the TSL curve and the temperature dependence of luminescence intensity under VUV excitation is presented in Fig. 7. The region of thermal destruction of EPR centers is perfectly correlated to that of the thermal release of the charge carriers observed in TSL. The self-trapped electrons are less stable than the self-trapped holes. Therefore, the following process of relaxation of trapped charge carriers may be proposed. The self-trapped electrons are thermally released at T > 50 K. They migrate partially to the selftrapped holes, forming STEs and giving rise to the TSL peak with a maximum at 76 K. Some part of the released self-trapped electrons is recaptured by deeper traps. The self-trapped hole centers

D.A. Spassky et al. / Optical Materials 35 (2013) 2465–2472

2469

Fig. 5. Integrated intensity of luminescence under excitation with X rays (50–650 s), phosphorescence (650–1300 s) and TSL (heating starts at t > 1300 s) of the molybdates. In the inset to ZnMoO4 plot: the integrated intensity of luminescence under excitation with X-rays at T = 10 K (1) and T = 100 K (2).

Fig. 6. (a) EPR spectra measured at 35 K in ZnMoO4 X-ray irradiated at 30 K and (b) simulated spectrum. The crystal was oriented with the axis a along magnetic field.

are stable up to 90 K. However, the number of self-trapped holes starts to decrease already at T > 50 K because of recombination with the thermally released electrons. When self-trapped holes become mobile they migrate to the deeper electron traps. The process is followed by the appearance of a TSL peak at 92 K. The luminescence spectrum in the TSL peak coincides with that of the intrinsic band implying that the corresponding recombination takes place at the MoO4 complex presumably perturbed by a defect nearby. A remarkably high phosphorescence intensity after the stopping crystal irradiation at T = 10 K also supports the conclusion about an

Fig. 7. Temperature dependencies of the EPR intensities of self-trapped electron (1) and hole (2) centers, TSL curve (3) and temperature dependence of the luminescence intensity under Eex = 11 eV (4) in ZnMoO4.

extra high concentration of occupied traps in ZnMoO4 (inset in Fig. 5c). The phosphorescence is connected with the STE creation from spatially correlated self-trapped electrons and holes. According to [34] the upper limit of stability caused by the spatial neighborhood of electrons and hole trapping centers is reached at the concentration of the centers about 1020/cm3. In a crystal where each oxyanionic complex may act as a trapping center, the concentration of trapping centers is obviously higher.

2470

D.A. Spassky et al. / Optical Materials 35 (2013) 2465–2472

When the temperature is increased up to 100 K, all charge carriers are released from the traps and the phosphorescence is not observed any more. The temporal dependence of the luminescence intensity under continuous X-ray excitation also remarkably differs for T = 10 and 100 K (see inset in Fig. 5c). The increase of the luminescence signal observed at T = 10 K in the first 300 s of irradiation is connected with the gradual saturation of intrinsic charge traps and it is not the case at T = 100 K when the traps are not filled.

Table 3 The release temperatures of self-trapped holes and electrons in the molybdates, determined by the maximum position of the corresponding TSL peak. Crystal

T (K)

Origin of the trapping center

References

CaMoO4 SrMoO4 CdMoO4 PbMoO4 ZnMoO4

150 200 69 140 or > 40 and >120 76 92

Self-trapped Self-trapped Self-trapped Self-trapped Self-trapped Self-trapped

[21] [22] [21] [23,44] This work

hole hole hole electron electron hole

4. Discussion The presented experimental data indicate a variety of crystal structure defects and impurities which are responsible for charge carrier trapping in each of the studied molybdates. However, charge carries can be also trapped by the regular structure elements of a crystal, and self-trapping of charge carriers is inherent to molybdates as well. So far the investigation of self-trapped holes and electrons has been performed in calcium, strontium, cadmium and lead molybdates. Existence of self-trapped holes has been detected by means of EPR and TSL technique in CaMoO4, CdMoO4 and SrMoO4 [21,22]. The holes have been shown to be self-trapped by MoO4 complexes due to lattice polarization. To the best of our knowledge, self-trapped electrons have not been experimentally detected in any molybdate, except PbMoO4. In PbMoO4, electrons are self-trapped in the form of axially distorted MoO43 complexes [23]. On the other hand, no self-trapped holes have been observed in PbMoO4. To explain this fact the presence of Pb 6s states at the very top of the valence band has been suggested and confirmed later by the recent calculations of the band structure of PbMoO4 [35]. The stabilization of oxygen-related holes is considered improbable in such case due to their possible migration over the lead states. The data on the decay temperatures of the self-trapped electrons and holes are summarized in Table 3. As it follows from the table the temperature of disintegration of self-trapped electrons in PbMoO4 is still under discussion. So far no data on coexistence of self-trapped electrons and holes was presented for molybdates. In the present paper, we obtained experimental evidence for existence of both, self-trapped electrons and self-trapped holes in ZnMoO4. This rather unusual property of ZnMoO4 may be due to the peculiarities of its crystal structure. The low-symmetry triclinic crystal structure of ZnMoO4 [36], in comparison to other studied molybdates, which crystalize in a tetragonal scheelite-type crystal structure, may be beneficial for the lattice polarization which is induced by the self-trapping of charge carriers. Previously the coexistence of self-trapped holes and self-trapped electrons has been suggested for Sc2O3 crystals [37,38]. The process of self-trapping was followed by the slowly damping phosphorescence in the STE emission band and the similar behavior is observed by us in ZnMoO4. Let us further consider the TSL curves obtained for CaMoO4, SrMoO4 and PbMoO4 taking into account the existence of selftrapped charge carriers.

observed in the long wavelength region relatively to the STE emission [39]. The position of the TSL peak at 52 K is perfectly correlated with the temperature region in which the anomalies of dielectric properties and electron spin–lattice relaxation times were observed in CaMoO4 [40]. These anomalies are connected with the second-order ferroelastic phase transition. Probably, the phase transition is accompanied by the partial delocalization of the trapped charge carriers. A high frequency factor obtained for the corresponding TSL peak indicates to the release of an electron from a trap. Because of the lack of the information on the possible origin of electron traps in molybdates, we may refer to the tungstates that commonly demonstrate similar luminescent properties. According to [41] there are generally two types of electron traps in CaWO4 – one is due to the presence of oxygen vacancies as described above, another is related to the Mo impurity. By analogy, we paid special attention to the tungsten impurity which was present in all the molybdates studied (Table 1). Tungsten impurity ions substitute molybdenum, thus forming WO4 complexes in molybdates. The lowest energy of electron transition from oxygen to the metal in WO4 complex is higher than the energy of corresponding transition in MoO4 complex [42]. Therefore, in contrast to the drastic role of Mo impurity in TSL processes in tungstates, where Mo forms the discrete energy levels below the bottom of the conduction band, the electronic states of W ions are expected to appear within the conduction band of molybdates. Delocalization of electrons from the traps formed by oxygen vacancies is the most probable explanation for the TSL peak at 52 K. As the electrons migrate to the holes self-trapped at oxyanionic complexes the emission spectrum of this TSL peak correspond to STE emission. The presence of another type of recombination centers might be suggested in calcium molydbate. The peak at 20 K is characterized by the extremely low value of the frequency factor, which may indicate to the recombination of a trapped hole with a closely situated electron trap. The situation is similar to that observed in some other oxide materials for the case of so-called center-to-center recombination [43]. The recombination of released holes with trapped electrons results in the TSL peak with the spectrum of the STE emission, which might indicate to the presence of electron trap centers near the regular MoO4 complex. 4.2. Strontium molybdate

4.1. Calcium molybdate Three distinct TSL peaks are observed upon the heating of an irradiated CaMoO4 crystal. The high-temperature peak at 150 K is connected with the thermal release of self-trapped holes (see Table 3). The emission band in this TSL peak is shifted to the longer wavelength region with respect to that of STE that implies hole recombination with electrons trapped at defects, presumably at oxygen vacancies, which similarly to tungstates can be considered as electron traps in molybdates. It is also known for molybdates that the emission from oxygen-deficient oxyanion complexes is

The intrinsic hole trap centers are inherent to SrMoO4 and their thermal decay has been previously detected at 200 K [22]. We observed rather weak TSL peak in the region of 205 K that may be connected with the thermal release of self-trapped holes. The low intensity of the hole-related peak may be due to the relatively lower activation energy of electron traps in SrMoO4. In this case, self-trapped holes will be almost exhausted by 200 K due to the recombination with electrons released from shallower traps. The emission spectrum observed in all TSL peaks below 190 K coincides with that of the STE emission, which may indicate to

D.A. Spassky et al. / Optical Materials 35 (2013) 2465–2472

recombination of electrons released from different traps with the self-trapped holes. The emission spectra of the TSL peaks at 205 and 236 K are shifted towards longer wavelengths. Such spectrum is expected for the released self-trapped holes (similarly to CaMoO4) demonstrating that the corresponding radiative recombination occurs at defect-related emission centers. The study of the origin of the low-temperature trap centers (T = 30–40 K) may be of special interest because only these are responsible for the slight changes observed in the luminescence intensity at low temperatures in SrMoO4. 4.3. Lead molybdate It has been reported earlier that the temperature of self-trapped electron disintegration in PbMoO4 is 40, 120 [44] or 140 K [23]. The results presented in Fig. 3 do not reveal the TSL peaks at 120 or 140 K. All TSL peaks are observed below these temperatures in our experiments with PbMoO4. A special attention should be paid to the most intensive TSL peak at 43 K. The peak is situated exactly in the temperature region where a significant decrease of the luminescence intensity is observed upon the cooling of a PbMoO4 sample under VUV and X-ray excitation (Fig. 3b, curves 2 and 3). The decrease is about 25–35% of the maximum intensity at the temperatures above the TSL peak. It implies the trap centers responsible for this peak to be very efficient in competition with the radiative recombination of charge carriers. Even the release of self-trapped holes in CaMoO4 does not result in such changes of the luminescence intensity (Fig. 1). We suppose that the peak at 43 K is connected with the thermal release of self-trapped electrons in PbMoO4. It is known that the thermal release of self-trapped electrons has been observed at 50 K in PbWO4 [41], which is close to the position of the peak at 43 K in PbMoO4. According to [45] the self-trapping of electrons in PbWO4 is also followed by the substantial decrease of the luminescence intensity (30–60%) under VUV excitation. The TSL peaks that were detected at 120 and 140 K that were previously attributed to an intrinsic electron trap center [23] might be connected with the trapping of an electron on a regular complex that is perturbed by a defect or impurity nearby. As it has been shown for PbWO4, such kind of trap centers are thermally destroyed at temperatures higher than that for the unperturbed regular trap centers [41]. Additional EPR and TSL experiments are necessary to verify the above suggested origin of the TSL peak at 43 K. 4.4. The influence of intrinsic trap centers on the scintillation process in molybdates The scintillation yield will be affected by traps at the migration stage of charge carrier relaxation. The self-trapping of charge carriers takes place at the regular oxyanionic complexes. Since these complexes act also as emission centers in molybdates, it should not substantially affect the scintillation yield if there were only one kind (electron or hole related) of the intrinsic trapping center in the crystal. Therefore, the number of self-trapped carriers (and the intensity of TSL peaks) is defined by the concentration of defect-related electron traps in CaMoO4 and SrMoO4 and hole traps in PbMoO4. It follows from the presented results that the concentration of the defects responsible for the hole trapping in PbMoO4 is considerably higher than the concentration of the defects responsible for the electron trapping in CaMoO4 and SrMoO4 (Fig. 5). The ratio of the light sum emitted in TSL peaks to that emitted under the steady excitation is higher in 4–7 times in PbMoO4 than in CaMoO4 or SrMoO4. Weak phosphorescence is also observed in PbMoO4 after the end of excitation indicating a high concentration of the filled traps. Hole traps may be created in oxides due to the substitution of a cation. For example, in the

2471

well-known model compound MgO, holes are trapped at the oxygen ions perturbed by the Ca2+ or Be2+ ions that substitute for neighboring cations Mg2+ [46,47]. The samples of PbMoO4 studied by us were contaminated with Ca ions (see Table 1), which could act as hole traps. The role of contamination with hole-trapping bivalent ions is not so essential in the recombination processes in CaMoO4 and SrMoO4 because of the efficient self-trapping of holes. The concentration of oxygen vacancies, which are suggested to be responsible for the trapping of electrons in these crystals may be rather low. Actually, we observed the emission from the oxygen deficient complexes only in TSL peaks, while this band was not observed in emission spectra measured under VUV and X-ray excitation. The existence of self-trapped electrons and holes in ZnMoO4 implies immobility of separated electrons and holes at low temperatures and should negatively affect the scintillation light yield at low temperatures even in case of perfect crystals, in which the defects of crystals structure are absent. As it is shown in Fig. 1–4b the decrease of the STE emission intensity under VUV and X-ray excitation is correlated with the position of the most intensive TSL peaks. Most striking results were obtained for ZnMoO4 right because of the intrinsic character of electrons and holes traps. Let us find whether such processes can be experimentally revealed in the scintillation response. Temperature dependencies of the scintillation light yield have been obtained previously for PbMoO4 and CaMoO4 under a-particles excitation in [6,19] and the results are plotted on Figs. 1b and 3b, curves 4. Although the experiments reported in [6] and [19] were performed on different samples of PbMoO4 and CaMoO4, the modification of the scintillation yield due to the trapping is generally the same as in the case of VUV (CaMoO4, PbMoO4) and X-ray (PbMoO4) excitation. Therefore, based on the analysis of the temperature dependencies of luminescence intensity under VUV or X-ray excitation it is possible to make suggestions concerning the scintillation yield in ZnMoO4 and SrMoO4. One may expect the lowest scintillation yield in ZnMoO4, in which a significant decrease of the yield at T < 80 K is connected with the immobility of charge carriers. The conclusion is consistent with the data on the extremely low light yield of 550 ph/MeV in ZnMoO4 at low temperatures [13]. A substantially higher light yield is expected for SrMoO4, which does not show self-trapping of electrons. The most efficient traps modulate only slightly the luminescence temperature dependence in this crystal. Thus, SrMoO4 may be considered as a candidate for the application in cryogenic scintillation detectors. 5. Conclusions The origin of charge carrier traps in molybdates that are potential materials for application in cryogenic scintillation detectors is discussed and the influence of trapping phenomena on luminescence properties is demonstrated. Self-trapped electrons or selftrapped holes have been reported in all the studied crystals. The present study supplies arguments in favor of the presence of selftrapped holes in CaMoO4 and SrMoO4 and self-trapped electrons in PbMoO4. It is shown for the first time that both self-trapped electrons and holes coexist in ZnMoO4. The immobility of charge carriers at T < 50 K results in the substantial decrease of the probability of STE creation with consequent worsening of the luminescent properties and in an unusually low scintillation light yield of ZnMoO4 at low temperatures. Acknowledgements The financial support from 7th FP INCO.2010-6.1 Grant Agreement No. 266531 (project SUCCESS), Mobilitas ESF program (Grant

2472

D.A. Spassky et al. / Optical Materials 35 (2013) 2465–2472

MTT83), Estonian Research Council – Institutional Research Funding IUT02-26, RFBR 11-02-01506-a and BMBF Project RUS 10/037 grants is gratefully acknowledged. We are grateful to Dr. A. Kotlov for his assistance during experiments at the SUPERLUMI station. References [1] S. Pirro, J.W. Beeman, S. Capelli, M. Pavan, E. Previtali, P. Gorla, Phys. Atom. Nucl. 69 (2006) 2109. [2] S. Belogurov, V. Kornoukhov, A. Annenkov, A. Borisevich, A. Fedorov, M. Korzhik, V. Ligoun, O. Missevitch, S.K. Kim, S.C. Kim, S.Y. Kim, J.W. Kwak, H.S. Lee, J. Lee, S.S. Myung, M.J. Lee, Y.D. Kim, J.Y. Lee, J.I. Lee, H.J. Kim, Y.J. Kwon, M.J. Hwang, J.J. Zhu, IEEE Trans. Nucl. Sci. 52 (2005) 1131. [3] V.B. Mikhailik, S. Henry, H. Kraus, I. Solskii, Nucl. Instrum. Method Phys. Res. A 583 (2007) 350. [4] A.N. Annenkov, O.A. Buzanov, F.A. Danevich, A.Sh. Georgadze, S.K. Kim, H.J. Kim, Y.D. Kim, V.V. Kobychev, V.N. Kornoukhov, M. Korzhik, J.I. Lee, O. Missevitch, V.M. Mokina, S.S. Nagorny, A.S. Nikolaiko, D.V. Poda, R.B. Podviyanuk, D.J. Sedlak, O.G. Shkulkova, J.H. So, I.M. Solsky, V.I. Tretyak, S.S. Yurchenko, Nucl. Instrum. Method. Phys. Res. A 584 (2008) 334. [5] S.J. Lee, J.H. Choi, F.A. Danevich, Y.S. Jang, W.G. Kang, N. Khanbekov, H.J. Kim, I.H. Kim, S.C. Kim, S.K. Kim, Y.D. Kim, Y.H. Kim, V.V. Kobychev, V.N. Kornoukhov, J.I. Lee, J.S. Lee, K.B. Lee, M.K. Lee, Y.H. Lee, S.S. Myung, J.H. So, V.I. Tretyak, Y. Yuryev, Astropart. Phys. 34 (2011) 732. [6] F.A. Danevich, B.V. Grinyov, S. Henry, M.B. Kosmyna, H. Kraus, N. Krutyak, V.M. Kudovbenko, V.B. Mikhailik, L.L. Nagornaya, B.P. Nazarenko, A.S. Nikolaiko, O.G. Polischuk, V.M. Puzikov, A.N. Shekhovtsov, V.I. Tretyak, Yu Ya Vostretsov, Nucl. Instrum. Method. Phys. Res. A 622 (2010) 608. [7] M. Minowa, K. Itakura, S. Moriyama, W. Ootani, Nucl. Instrum. Method. Phys. Res. A 320 (1992) 500. [8] O.P. Barinova, F.A. Danevich, V.Ya. Degoda, S.V. Kirsanova, V.M. Kudovbenko, S. Pirro, V.I. Tretyak, Nucl. Inst. Method Phys. Res. A 613 (2010) 54. [9] O.P. Barinova, F. Cappella, R. Cerulli, F.A. Danevich, S.V. Kirsanova, V.V. Kobychev, M. Laubenstein, S.S. Nagorny, F. Nozzoli, V.I. Tretyak, Nucl. Inst. Method Phys. Res. A 607 (2009) 573. [10] N.V. Bashmakova, F.A. Danevich, V.Ya. Degoda, I.M. Dmitruk, V.M. Kudovbenko, S.Yu. Kutovyi, V.V. Mikhailin, S.S. Nagorny, A.S. Nikolaiko, S. Nisi, A.A. Pavlyuk, S. Pirro, A.E. Savon, S.F. Solodovnikov, Z.A. Solodovnikova, D.A. Spassky, V.I. Tretyak, S.M. Vatnik, E.S. Zolotova, Funct. Mater. 16 (3) (2009) 266. [11] J.W. Beeman, F.A. Danevich, V.Y. Degoda, E.N. Galashov, A. Giuliani, I.M. Ivanov, M. Mancuso, S. Marnieros, C. Nones, G. Pessina, E. Olivieri, C. Rusconi, V.N. Shlegel, V.I. Tretyak, Y.V. Vasiliev, J. Low Temp. Phys. 167 (2012) 1021. [12] J.W. Beeman, F. Bellini, S. Capelli, L. Cardani, N. Casali, I. Dafinei, S. Di Domizio, F. Ferroni, E.N. Galashov, L. Gironi, F. Orio, L. Pattavina, G. Pessina, G. Piperno, S. Pirro, V.N. Shlegel, Ya.V. Vasilyev, C. Tomei, M. Vignati, Astropart. Phys. 35 (2012) 813. [13] L. Gironi, C. Arnaboldi, J.W. Beeman, O. Cremonesi, F.A. Danevich, V.Ya. Degoda, L.I. Ivleva, L.L. Nagornaya, M. Pavan, G. Pessina, S. Pirro, V.I. Tretyak, I.A. Tupitsyna, J. Instrum. (2010) 5P11007. [14] H. Bhang, R.S. Boiko, D.M. Chernyak, et al., J. Phys.: Conf. Ser. 375 (2012) 042023.

[15] L. Ivleva, I. Voronina, L. Berezovskaya, P. Lykov, V. Osiko, L. Ishakova, Crystallogr. Rep. 5 (6) (2008) 1122. [16] A.M. Dubovik, Yu Ya Vostretsov, B.V. Grinyov, F.A. Danevich, H. Kraus, L.L. Nagornaya, V.B. Mikhailik, I.A. Tupitsyna, Acta Phys. Pol. A 117 (2010) 15. [17] E.N. Galashev, T.M. Denisova, I.M. Ivanov, E.P. Makarov, V.N. Mamontov, V.N. Shlegel, Yu.G. Stenin, Ya.V. Vasiliev, V.N. Zhdankov, Func. Mater. 17 (4) (2010) 504. [18] D.A. Spassky, V.V. Mikhailin, A.E. Savon, E.N. Galashov, V.N. Shlegel, Ya V. Vasiliev, Opt. Mater. 34 (2012) 1804. [19] V.B. Mikhailik, H. Kraus, Phys. Status Solidi B 247 (2010) 1583. [20] D. Spassky, A. Vasil’ev, I. Kamenskikh, V. Kolobanov, V. Mikhailin, A. Savon, L. Ivleva, I. Voronina, L. Berezovskaya, Phys. Status Solidi A 206 (2009) 1579. [21] A. Hofstaetter, R. Oeder, A. Scharmann, D. Schwabe, Z. Phys. B 35 (1979) 1. [22] A. Hofstaetter, R. Oeder, A. Scharmann, D. Schwabe, J. Lumin. 22 (1981) 419. [23] A. Hofstaetter, R. Oeder, A. Scharmann, D. Schwabe, B. Vitt, Phys. Stat. Sol. (b) 89 (1978) 375. [24] G. Zimmerer, Rad. Meas. 42 (2007) 859. [25] V.V. Mikhailin, A.N. Vasil’ev, Introduction to solid-state spectroscopy, Izd, MGU, Moscow, 1987. [26] T. Kajitani, M. Itoh, Phys. Stat. Sol. C 8 (2011) 108. [27] M. Itoh, J. Lumin. 132 (2012) 645. [28] M. Itoh, T. Kajitani, Phys. Rev. B 87 (2013) 085201. [29] N.F. Mott, Proc. Roy. Soc. Lond. A 167 (1938) 384. [30] V. Murk, M. Nikl, E. Mihokova, K. Nitsch, J. Phys.: Condens. Matter 9 (1997) 249. [31] A. Hofstaetter, J. Planz, A. Scharmann, Z. Naturforsch. 32a (1977) 957. [32] A. Hofstaetter, A. Scharmann, D. Schwabe, B. Vitt, Z. Phys. B 30 (1978) 305. [33] V.V. Laguta, A. Vedda, D. Di Martino, M. Martini, M. Nikl, E. Mihokova, J. Rosa, Y. Usuki, Phys. Rev. B 71 (2005) 235108. [34] E. Beyer, R. Biederbick, A. Hofstaetter, R. Oeder, A. Scharmann, D. Schwabe, B. Vitt, Phys. Stat. Sol. (b) 87 (1978) K35. [35] Y. Zhang, N.A.W. Holzwarth, R.T. Williams, Phys. Rev. B 57 (1998) 12738. [36] W. Reichelt, T. Weber, T. Soehnel, S. Daebritz, Z. Fuer Anorg. Allg. Chem. 626 (2000) 2020. [37] A. Lushchik, M. Kirm, Ch. Lushchik, I. Martinson, G. Zimmerer, J. Lumin. 87–89 (2000) 232. [38] Ch. Lushchik, A. Lushchik, T. Kyarner, M. Kirm, S. Dolgov, Russ. Phys. J. 43 (2000) 171. [39] D.A. Spasskii, V.N. Kolobanov, V.V. Mikhailin, L.Yu. Berezovskaya, L.I. Ivleva, I.S. Voronina, Opt. Spectrosc. 106 (2009) 556. [40] J. Simon, J. Banys, J. Hoentsch, G. Volkel, R. Bottcher, A. Hofstaetter, A. Scharmann, J. Phys. Condens. Matter 8 (1996) L359. [41] M. Nikl, V.V. Laguta, A. Vedda Phys, Stat. Sol. (b) 245 (2008) 1701. [42] R. Grasser, E. Pitt, A. Scharmann, G. Zimmerer, Phys. Stat. Sol. (b) 69 (1975) 359. [43] A. Meijerink, G. Blasse, J. Phys. D: Appl. Phys. 24 (1991) 626. [44] M. Bohm, O. Erb, A. Scharmann, J. Lumin. 33 (1985) 315. [45] N.R. Krutyak, V.V. Mikhailin, D.A. Spassky, V.N. Kolobanov, M.B. Kosmyna, B.P. Nazarenko, V.M. Puzikov, A.N. Shekhovtsov, J. Appl. Spectrosc. 79 (2012) 211. [46] S.A. Dolgov, T. Kärner, A. Lushchik, A. Maaroos, S. Nakonechnyi, E. Shablonin, Phys. Sol. State 53 (2011) 1244. [47] S.A. Dolgov, V. Isakhanyan, T. Kärner, A. Maaroos, S. Nakonechnyi, J. Phys. Cond. Mater. 15 (2003) 6871.