Influence of defects and traps in the scintillation process. Anna Vedda University of Milano Bicocca, Italy

Transcription

1 Influence of defects and traps in the scintillation process Anna Vedda University of Milano Bicocca, Italy LUMDETR 2009 July 15th, 2009

2 Outline Trapping processes during ionizing irradiation De trapping and e h recombinations Thermoluminescence (TSL) Wavelength resolved TSL measurements TSL vs RL spectra Examples of uncommon applications: amorphization, phase transitions TSL equations Data analysis: initial rise, variable heating rate method, curve fitting Role of traps in scintillation Prediction of RT trap lifetimes Slow tails in scintillation decay, afterglow, permanent trapping Examples: Lead tungstate, perovskites, crystalline silicates

3 Scintillation: Physical processes e High Energy photon absorption e e e Conduction Band E g Valence Band h h h Traps L exc L gnd Luminescent Centre Photon ( VIS-UV ) h CONVERSION TRANSPORT LUMINESCENCE

4 TRAPPING-DETRAPPING PROCESSES: case of electron traps e A direct process Conduction band B delayed process Ionizing radiation E T T L exc L ground hν Radio luminescence (X, γ ) h Valence band Mean time spent in the trap = τ = C exp(e T /kt) According to the E T value (typically from 10-2 up to 10 0 ev) a very wide range of τ can be measured (of even less than1 s up to thousands years)

5 TSL-TSC process simple scheme (electron traps) Conduction band TSC Temperature ( C) Heating TSL intensity (arb. units) E T T L exc L ground hν TSL Time (s) Temperature ( C) Valence band TSL signal measured by a photomultiplier, no information about the recombination centre Probability of escape from trap: P = K exp(-e T /kt) P increases very strongly with temperature A luminescence peak appears (thermally stimulated luminescence)

6 Case of hole traps Conduction band TSC TSL intensity (arb. units) hν TSL Temperature ( C) T E T Valence band It is not possible to determine from a simple glow curve whether electron or holes are detrapped during heating In a TSL process, the name trap is attributed to the defect from which carriers are freed by heating The name recombination centre is attributed to the defect in which carriers are stably trapped, and in which carriers of opposite sign recombine radiatively The same defects can act as traps or recombination centres in different temperature intervals

7 Different kinds of recombination paths TSC C.B. 1. Classical recombination through conduction band 2. Thermally assisted tunneling (trap-centre recombination) E TC E TT 1 T 2 T L e L g hν TSL V.B.

8 Wavelength resolved TSL measurements A wavelength resolved TSL measurement consists in a collection of emission spectra measured at constant temperature intervals (for example 1 K). Usually spectra are measured by a CCD. Study of both traps and recombination centers link to photo-, radio- luminescence, and scintillation

9 An example: low T Thermally Stimulated Luminescence (TSL) of PbWO 4 After temperature integration TSL NORMALIZE (D) 90 K 8 (C) 60 K (B) 40 K 4 (A) 10 K ENERGY (ev) TSL emission spectra 3D TSL measurement after x-ray irradiation at 10 K After wavelength integration RT decay times in the micro-milli second time scale TSL NORMALIZED INTENSITY (B) (A) (C) TEMPERATURE (K) TSL glow curves

10 Comparison between TSL and radio-luminescence (RL) spectra One would expect that they are the same but the nature of traps as well as specific trap-centre spatial correlations makes them often different Wavelength (nm) Normalized TSL and RL amplitude RL G - 6 H 5/2 x 3 P 0-3 H 4 5 D - 7 F 4 x 5d - 4f 1 1 D - 3 H P 0-3 H 5 5 D - 7 F 3 x TSL 3 P0-3 H 6 Sm Tm Tb Ce Case of Lu 2 SiO 5 :RE Significant differences between RL and TSL: RL spectra are governed by the emissions of principal dopant ions TSL spectra display mainly emissions from RE which capture holes during irradiation (Ce 3+,Tb 3+ Ce 4+,Tb 4+ ) even if they are present as trace impurities undoped Energy (ev) Evidence of the electronic nature of traps

11 Spatial correlation between Gd 3+ and traps in silica RL intensity (arb.units) RL 3 mol% Gd 0.1 mol% Ce 6 P 7/2-8 S 7/2 Gd 3+ 5d-4f Ce mol% Ce,3 mol% Gd Energy (ev) 0.1 mol% Ce,3 mol% Gd Gd 3+ emission TSL The difference between RL (where both Ce 3+ and Gd 3+ emissions are observed) and TSL spectra (featuring only Gd 3+ emission line) is due to the spatial correlation between Gd 3+ and oxygen-related electron traps.

12 TSL glow curves are sensitive to material amorphization Amorphous silica Crystalline quartz The amorphous structure of a material often induces broadening of TSL glow peaks due to the presence of continuous distributions of trap levels TSL INTENSITY ( Arb. Un. ) SiO 2 :500 ppmce TRAP ENERGY (ev) T ( C) Partial cleaning Temperature ( C)

13 TSL glow curves can also be sensitive to phase transitions Example of Ammonium Bromide (NH 4 Br) P.D. Townsend et al., Rad. Meas. 27, 31 (1997) Modifications of trap depths Increased effects in case of localized trap-centre recombinations Possibility to monitor temperature lags between heater and sample, which are a common error source in TSL measurements

14 TSL equations I ( T ) s T sn exp( E / kt )exp exp E / kt ' dt ' 0 T T T 0 First order recombination: no retrapping b / b1 T ( ) '' exp( / ) / 1 1 ''/ exp / ' ' I T s n E kt 0 b s E kt dt T T T 0 General order recombination: non negligible probability of retrapping

15 Data analysis: partial cleaning and initial rise method Case of YALO 3 RT X-RAY IRRADIATION Normalized TSL Amplitude GFE D C B 0.62 ev 0.53 ev 0.40 ev 0.26 ev 0.23 ev 0.15 ev /T (1/K) A 0.09 ev E I ( T ) nos exp kt PRE-HEATING T=T STOP RAPID COOLING TO RT TSL GLOW CURVE This procedure allows to evaluate the trap depth independently upon the kinetic order

16 10 6 EVALUATION OF TRAP DEPTHS partial cleaning of glow curves and initial rise (10 4 ppm Ce, 50 ppm Zr)-doped LuAG TSL intensity (arb. units) E 1 =1.05 ev τ RT 5.5 h E 2 =1.60 ev τ RT 10 6 y E 3 =1.92 ev τ RT y T stop =260 C E = 1.60 ev I=c exp(-e/k b T) /T (1/K) T ( C)

17 Evaluation of the frequency factor (s). Simple case of first order kinetics The presence of first order kinetics can be tested by verifying if TSL peak maxima do not vary with increasing dose. In this case, TSL intensity (arb. units) LYSO:Ce dose dependence Temperature ( o C) If E is already known, s can be evaluated (often only its order of magnitude due to several error sources) TSL intensity (arb. un.) 2, , E kt 2 m E s exp kt m Lu 3 Ga(x)Al(1-x) 5 O 12 :Ce 15s 20cm 10mA 150s 20cm10mA 300s 20cm 20mA 300s 10cm 20mA Temperature (K)

18 Variable heating rate method 1, LuGaG TSL intensity (arb.un.) C/s 1,5 C/s 2 C/s 2,5 C/s 3 C/s From: 11,8 E kt 2 m E s exp kt m Plotting lnt m2 /β vs 1/T m Temperature (K) 11,6 11,4 E = 0.92 ev s = 3x10 10 s -1 Slope = E/k ln(t m 2 /ß) 11,2 11 Intercept = ln (E/sk) 10,8 Possibility to extend to general order kinetics Need of good separation between different peaks 10,6 2, , , , , , , /T m (1/K)

19 Glow Curve Fit TSL Amplitude x 100 x T (K) TSL glow curve of YAP:0.1%Eu (full circles). The continuous red line represents the numerical fit of the glow curve in the framework of first order kinetics. Made in the framework of first or general order kinetics Difficult: E,s, and b (order of kinetic) are implicit parameters If several overlapping peaks are present, the number of parameters can be very high Need of preliminary information by other methods

20 Role of traps in scintillation If the RT decay time is of the order of micro- or milli- seconds Slow tails in the scintillation decay Longer (minutes, hours) Afterglow Very long (days, years) Permanent trapping Traps can be studied by heating at a constant rate after irradiation (slow scintillation tails correspond to TSL peaks at cryogenic temperatures, while afterglow and permanent trapping are related to peaks above RT) The determination of trap parameters by some method of analysis alows the evaluation of the order of magnitude of the RT lifetime (higher precision is commonly prevented by several error sources and by the temperature dependence of the frequency factor, which is hardly predictable)

21 The investigation of the role of traps in scintillation involves the following steps: Evidence of slow scintillation tails or afterglow Need to understand the nature of responsible defects Study of defects by an independent technique giving information about thermal stability of defects and their radiative recombination properties (TSL) Correlation of TSL data with other techniques which allow structural information about traps (e.g. EPR) Tuning of material synthesis to reduce the investigated defects

22 Role of traps in scintillation time decay Comparison between silicates, garnets, and perovskites lum. intensity [arb.units] ns Lu Al O :Ce (0.12%) I = 40% fast ns a - scintillation b - PL Photoluminescence decay time = 54 ns Lum. Intensity [arb. units] I(t)=459exp(-t/45ns) + 4.1exp(-t/714ns) I fast = 88% Lu 2-x Y x SiO 5 :Ce Lum. Intensity [arb. units] I(t) = 954exp(-t/24ns) + 110exp(-t/60ns) + 17exp(-t/329ns)+7.89 I fast = 65% YAlO 3 :Ce time [ns] time [ns] time [ns] The presence of fewer shallow traps in silicates with respect to other scintillators (very low TSL intensity below RT) is in agreement with the absence of slower components in the scintillation decay. TSL Intensity (arb. units) YALO :Ce 3 Lu Y SiO :Ce 2-x x 5 Lu Al O :Ce Temperature (K)

23 Example of lead tungstate (PbWO 4 PWO) Room temperature time decays TSL of undoped PWO Defect structures determined by EPR measurements 293 K time decay (s) Present only in Mo-doped samples WO 4 3- MO 4 3- Pb + - V 0 WO A Pb 10-7 WO La 3+ Scintillation time decay a 460 ppm La b 80 ppm La c undoped Intrinsic RT scintillation decay time Tmax (K) Suppressed by trivalent ion doping Doping with La 3+ reduces slow components in the scintillation decay, as well as TSL glow peaks at cryogenic temperatures La 3+ favours the reduction of intrinsic lattice defects

24 Effect of different ion dopings in very slow decay components in PWO Scintillation decays of a)undoped PWO and b)pwo(2750mo,500nb,50y) at RT. TSL glow curves of differently doped PWO samples Alpha coefficient: normalized difference between the signal level before the excitation pulse and the true background (It measures the level of very slow components) Good correlation between alpha coefficients of differently doped samples and TSL integral intensities below RT

25 YAlO 3 undoped crystals TSL Amplitude 154 K 190 K 232 K Wavelength integration 66 K 89 K 110 K 30 K Temperature (K) Temperature (K) intrinsic emissions 5d -4f Ce 3+ 1 Wavelength (nm) Temperature integration Normalized TSL Amplitude K K K defects Several emission centres for each glow peak 0.2 Recombinations involving valence/conduction bands Energy (ev)

26 Structural nature of traps Correlation with EPR signals EPR, TSL intensity (arb. units) O I O II TSL 50 ppm Ce 5000 ppm Ce O III O IV T (K) The TSL peaks at 154 K, 190 K, and 232 K are due to the de-trapping of holes from O - centres followed by recombination with electrons stored in oxygen vacancies

27 3 P 0-3 H 4 Doping of YAlO 3 with Yb 3+, Eu 3+, and Tm 3+ Trapping of electrons during irradiation 0.1% Tm TSL emissions 1 D 2-3 H 6 3 P 0-3 H 5 0.1% Tm K 3 P 0-3 H 6 Normalized TSL Amplitude x 10 x 100 x 2 2% Yb 0.1% Eu TSL Amplitude 2 F 5/2-2 F 7/2 CT emissions x 10 5D 0-7F x K 5d -4f Ce K K K 2% Yb 0.1% Eu K Undoped Undoped K K defects intrinsic emissions Temperature (K) Energy (ev) Hole traps are similar to those observed in undoped crystals. Rare earth ions compete with oxygen vacancies in electron trapping (becoming temporarily Yb 2+, Eu 2+, Tm 2+ ) and act as recombination centres in the TSL process

28 Detrapping and recombination: Case of YAlO 3 :Yb TSL peak E (ev) ν (1/s) at RT (s) 45 K K K K 190 K 232 K 110 K K (O - ) K (O - ) K (O - ) Due to their long RT decay times, traps responsible for the TSL peaks above 100 K (mostly O- centres) cause slow tails in the scintillation time decay

29 Doping with Ce 3+, Pr 3+, and Tb 3+ Trapping of holes during irradiation a TSL pattern different from that occurring with Yb 3+, Eu 3+, Tm 3+ doping is expected Tunneling emission 4 Tb 3+ emission Evidence of one or two dominant TSL peaks, similar to those observed in the cases of undoped, Yb, Eu, and Tm doped crystals TSL spectral emissions of rare- earth ions trapping holes Normalized TSL Intensity Ce 3+ emission O - O - O - 0.1% Tb Pr 3+ emission 0.1% Pr 0.5% Ce Undoped Temperature (K)

30 1 D 2-3 H 6 0.1% Tm 3 P 0-3 H 5 RE-doped YAP TSL spectra Example: 190 K peak 0.1% Pr 5d 1-3 H 4,5 1 D 2-3 H 4 3 P 0-3 H 4,5 3 P 0-3 H 4 3 P 0-3 H 6 2 F - 2 F 5/2 7/2 x 10 2% Yb TSL Amplitude 5 D - 7 F 4 x 5d -4f 1 5d -4f 1 Ce traces 5 D 3-7 F x 0.1% Tb 0.5% Undoped Ce TSL Amplitude CT emissions 5D 0-7F x 0.1% Eu defects intrinsic emissions Energy (ev) defects 5d 1-4f Ce traces intrinsic emissions Undoped Energy (ev) Emissions from electron and hole trapping RE ions observed in the same TSL peaks.

31 In RE-doped YAP, traps and recombination centres do not act like separate entities, but rather as parts of defect complexes in which carriers can be transferred from intrinsic levels (O - centres and oxygen vacancies) to the rare earth ion levels. The observation of a nearly temperature independent TSL emission suggests that tunnelling driven recombination processes occur in addition to the temperature driven recombination, supporting the existence of such complexes with close spatial correlation between traps and centres. Due to the formation of defect complexes, the kind of recombination centre (capturing holes or electrons, e.g. Ce or Yb) is not always a proof of the nature (electron- or hole-kind) of a trap The comparison of the TSL patterns of crystals doped with a variety of rare earth ions proved to be essential in order to highlight this phenomenology, which nevertheless could be more common in TSL experiments than expected.

32 Consequences of trap levels RT decay times on scintillation efficiency Sample Concentr. Melt (%) Growth method &Crucible N phels /MeV (phels per MeV) FWHM (%) at 662 kev N phels % of standard RL_intens. % of standard YAP:Ce Ce-0.05 CZ - Ir YAP:Ce Ce-0.5 CZ Mo Czech standard YAP:Pr Pr CZ - Ir YAP:Pr Pr 1.0 CZ - Ir YAP:Pr Pr 0.5 CZ - Mo Czech YAP:Pr Czech Pr 1.1 CZ - Mo n.m. As the glow curve peak occurs at a much higher temperature in the case of Pr doping, the RT lifetime of it main trap is about one order of magnitude larger with respect to that of Ced o p e d Y A P. Normalized TSL Amplitude (RT) ~ 10-3 s Pr-0.1% (RT) ~ 10-4 s Ce-0.1% x 10 Undoped Light yield ( 1 μs) Steady state RL This can be one of the reasons why delayed radiative recombination processes in YAP:Pr are more pronounced and its photoelectron yield is lower with respect to that of YAP:Ce in spite of its higher steady state RL efficiency Temperature [K]

33 Lu 2 SiO 5 :Ce and Lu 1.96 Y 0.04 SiO 5 :Ce identification of the nature of traps thanks to their localized recombination mechanism LSO:Ce TSL intensity (arb. units) 2 1 x 20 LSO:Ce LYSO:Ce Recombination mechanism: Ce 4+ + e (freed from the electron trap) Ce 3+* (excited state) Ce 3+ + hν (TSL emitted light) Evidence of the electronic nature of TSL traps Temperature ( o C) Trap depth (ev) LSO:Ce LYSO:Ce 78 C 135 C 181 C 236 C 300 C Tstop ( o C) Constant trap depth for all glow peaks except the 300 C one

34 Frequency factor (1/s) C 135 C LSO LYSO Lu1-O distance (A) 4.0 LSO structure oxygen site no. 181 C 236 C O-Lu distance (Angstrom) No dependence of maximum temperatures from dose No retrapping E kt E s exp 2 m kt m In the case of a thermally assisted tunneling process s exp r Where s is the frequency factor and r is the distance between the trap and the recombination centre Exponential dependence of the frequency factors from Lu1(Ce)-Oxygen distances Oxygen vacancies act as electron traps responsible for TSL peaks

35 Conclusions TSL is a powerful technique for the study of luminescent materials; especially when performed in wavelength resolved mode, it can really provide a deep insight in the trapping-recombination processes occurring in a scintillator. Several types of recombination mechanisms between traps and luminescent centres exist. Their comprehension needs very often the possibility to vary sample parameters like kind of doping, dopant concentration, synthesis and annealing conditions, and to probe their influence on TSL features. Anyway, in some cases only qualitative descriptions of the recombination processes remain possible. Data analyses are delicate; the handling of trap parameters needs a clear consciousness of the error sources linked to experimental data and numerical methods. This is important especially if a comparison with similar parameters obtained by other techniques has to be performed. Such consciousness of the limits of the technique doesn t lower, but increases its potential allowing to handle data in a critical and therefore constructive way.

36 References 1. M. Martini, F. Meinardi, G. Spinolo, A. Vedda, M. Nikl, Y. Usuki, Shallow traps in PbWO 4 by wavelength resolved thermally stimulated luminescence, Phys. Rev. B 60, 4653 (1999). 2. Vedda, M. Nikl, M. Fasoli, E. Mihokova, J. Pejchal, M. Dusek, G. Ren, C.R. Stanek, K. J. McClellan, D.D. Byler, Thermally stimulated tunneling in rare-earth doped Lu-Y oxyorthosilicates, Phys. Rev. B 78, (2008). 3. A. Vedda, N. Chiodini, D. Di Martino, M. Fasoli, L. Griguta, F. Moretti, E. Rosetta, Thermally stimulated luminescence of Ce and Tb doped SiO 2 sol-gel glasses, Journal of Non-Cryst. Solids 351, 3699 (2005). 4. Vedda, D. Di Martino, M. Martini, V.V. Laguta, M. Nikl, E. Mihokova, J. Rosa, K. Nejechleb, K. Blazek, Thermoluminescence of Lu 3 Al 5 O 12 :Ce crystals, Physica Status Solidi A 195, R1 (2003). 5. P.D. Townsend, A.P. Rowlands, G. Corradi, Thermoluminescence during a phase transition, Rad. Meas. 27, 31 (1997). 6. M. Nikl, K. Nitsch, P. Bohacek, M. Martini, E. Mihokova, A. Vedda, S. Croci, G.P. Pazzi, P. Fabeni, S. Baccaro, B. Borgia, I. Dafinei, M. Diemoz, G. Organtini, E. Auffray, P. Lecoq, M. Kobayashi, M. Ishii, Y. Usuki, "Decay kinetics and thermoluminescence of PbWO 4 :La 3+ ", Appl. Phys. Lett. 71, 3755, (1997). 7. V.V. Laguta, M. Martini, A. Vedda, E. Rosetta, M. Nikl, E. Mihokova, Y. Usuki, Electron traps related to oxygen vacancies in PbWO 4, Phys. Rev. B 67, (2003). 8. M. Nikl, P. Bohacek, E. Mihokova, N. Solovieva, A. Vedda, M. Martini, G.P.Pazzi, P. Fabeni, M. Kobayashi, M. Ishii, Enhanced efficiency of PbWO 4 :Mo,Nb scintillator, J. Appl. Phys. 91, 5041 (2002). 9. A. Vedda, M. Fasoli, M. Nikl, V.V. Laguta, E. Mihokova, J. Pejchal, A. Yoshikawa, M. Zhuravleva, Trap-centre recombination processes by rare earth activators in YAlO 3 single crystal host, Phys. Rev. B 80, (1)-(9) (2009). 10. V.V. Laguta, M. Nikl, A. Vedda, E. Mihokova, J. Rosa, K. Blazek, Hole and electron traps in the YAlO 3 single crystal scintillator, Phys. Rev. B 80, (1)-(10) (2009). 11. A. Vedda, M. Martini, F. Meinardi, J. Chval, M. Dusek, J.A. Mares, E. Mihokova, M. Nikl, Tunneling process in Thermally Stimulated Luminescence of mixed Lu x (Y 3+ ) 1-x AlO 3 :Ce crystals, Phys. Rev. B 61, 8081 (2000).