- Email: [email protected]
Luminescence study of alumina nanopowders prepared by various methods
Accepted Manuscript Luminescence Study of Alumina Nanopowders Prepared by Various Methods Marek Oja, Eliko Tõldsepp, Eduard Feldbach, Henri Mägi, Sergey Omelkov, Marco Kirm PII:
S1350-4487(15)30093-7
DOI:
10.1016/j.radmeas.2015.12.014
Reference:
RM 5495
To appear in:
Radiation Measurements
Received Date: 25 October 2015 Revised Date:
15 December 2015
Accepted Date: 19 December 2015
Please cite this article as: Oja, M., Tõldsepp, E., Feldbach, E., Mägi, H., Omelkov, S., Kirm, M., Luminescence Study of Alumina Nanopowders Prepared by Various Methods, Radiation Measurements (2016), doi: 10.1016/j.radmeas.2015.12.014. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT
Luminescence Study of Alumina Nanopowders Prepared by Various Methods Marek Oja*, Eliko Tõldsepp, Eduard Feldbach, Henri Mägi, Sergey Omelkov, and Marco Kirm
RI PT
Institute of Physics, University of Tartu, Ravila 14C, 50 411, Tartu, Estonia * Corresponding author. Tel.: +372 737 4762, email: [email protected]
SC
Abstract
TE D
M AN U
Mixed nanopowders of transition alumina prepared by combustion synthesis and phase pure ultra-porous α-alumina by oxidation method were investigated using low temperature time-resolved cathodoluminescence and photoluminescence spectroscopy under VUV-XUV excitation. In all samples along with the 7.6 eV emission of self-trapped excitons of α-alumina, luminescence bands due to F, F+ centres with maxima at 3 and 3.8 eV and other UV-visible luminescence bands of intrinsic and extrinsic origin with varying intensity depending on sample preparation method and thermal treatment were studied. In alumina nanopowders the excitonic excitation peak at ~9.1 eV near fundamental absorption edge is shifted to the higher energies by 0.15 eV in comparison with the same feature in single crystals. The nanostructure of alumina is responsible for this shift.
EP
Keywords: transition and α-alumina, relaxation of electronic excitations, intrinsic and extrinsic luminescence, nanopowders
1. Introduction
AC C
Aluminium oxide (Al2O3) is an important material for different optical and technological applications in the form of ceramics or single crystals, but there is a growing interest in nanostructured alumina, because of its large surface area being favourable for catalyst support and filtration. Significant research has been carried out on single crystals of α-Al2O3 using luminescence and other methods (see Kirm et al., (1999); Valbis and Itoh, (1991) and references therein). Recently, various modification (incl. α-, θ-, γ-Al2O3) of phase pure ultra-porous alumina nanopowder (hereafter called as UPA) with the low impurity content was investigated using low temperature time-resolved luminescence spectroscopy (Museur et al., 2013).
ACCEPTED MANUSCRIPT
M AN U
SC
RI PT
Metastable aluminium oxide phases (e.g. δ-, θ-, κ- and γ-Al2O3) with the common name of transition alumina also exist in the nature. The latter phases have different crystal structure, electronic properties and have been significantly less studied, but such activities based on luminescence spectroscopy are rising (e.g. Kirm et al., (2010); Oja et al., (2013); Pustovarov et al., (2012); Trinkler et al., (2012)). Due to their thermochemical properties the preparation of pure phase samples is a real challenge and therefore metastable alumina samples typically contain multiple Al2O3-phases complicating identification of elementary processes in such mixtures. Depending on the preparation method often unintended impurities are introduced into alumina samples. In order to reduce influence of surface related effects (luminescence quenching etc.) the XUV and electron beam excitation is applied, methods which penetrate deeper into the samples prohibiting near-surface effects at fundamental absorption regime, which has to be considered in investigation of wide gap crystals (Lushchik et al., 1994). The main goal of this work is to enlighten electronic properties, relaxation of electronic excitations, identify phase and crystallite size specific features in luminescence of alumina samples prepared with different synthesis methods.
TE D
2. Sample preparation, structural properties and experimental techniques
AC C
EP
Nanostructured UPA was produced using oxidation of high purity aluminum (99.999 %) through a liquid mercury and silver film, which leads to formation of highly porous (99%) macroscopic (cm dimensions), but nanostructured (tangled alumina fibers of 5 nm diameter) monoliths as described by Vignes et al., (2008). The further post-treatment at elevated temperatures (1200-1250 °C) results in nanostructured α-alumina with characteristic crystallite size of 200-300 nm as described in the paper by Museur et al., (2013). This method has advantage over combustion and solid state synthesis as amount of impurities is well controlled by using high purity aluminium and its impurity content is smaller in comparison with samples prepared by combustion synthesis. Two samples were prepared using the combustion synthesis method analogous to Oja et al., (2013). The chemicals used in the synthesis of these samples were Al(NO3)3·9H2O (for the ALO 1600 - Alfa Aesar, 99.999% and for the ALO 5050 - REAHIM, 99.99%) and urea (Alfa Aesar) as a fuel. Sample ALO 1600 was produced using stoichiometric mixture with fuel/nitrate molar ratio 5:2 whereas
ACCEPTED MANUSCRIPT
RI PT
sample ALO 5050 was synthesised using fuel rich mixture which corresponds to a molar ratio of 25:4. According to Ianoş et al., (2009) fuel-rich synthesis gives lower combustion temperature and therefore different phase composition and higher specific surface area compared to stoichiometric synthesis. Burning out of residual carbon and phase composition adjustment is achieved by an annealing procedure resulting in an increase of crystallite size. Samples were annealed in air at 1600˚C for 5 hours (ALO 1600) and at 1000˚C for 1 hour (ALO 5050). After annealing procedure the sample ALO 1600 was white powder but sample ALO 5050 was still greyish powder indicating either presence of carbon or absorption of oxygen vacancies.
TE D
M AN U
SC
The phase composition and crystallite size of the samples were determined by powder X-ray diffraction (XRD) analysis and the obtained values are presented in Table 1. Sample ALO 5050 consists of several Al2O3 phases with different crystallite size, which is a general complication in studies of transition alumina whereas other studied samples were composed of α-Al2O3 phase. Analysis of the quantitative phase composition (using the Rietveld method, program FULLPROF) is not straight forward because of the overlap of XRD patterns from various phases are responsible for relatively large a few percent error bars on the obtained values. The γ-Al2O3 phase has only very broad diffraction lines referring to small crystallite size below 10 nm. Although there can be an a few nm error in absolute size, the crystallite size of α-Al2O3 is practically an order of magnitude larger (see Table 1). Crystallite size estimations were performed using the Scherrer equation.
AC C
EP
Low temperature cathodoluminescence (CL) studies were performed in Tartu using a home-built setup equipped with VUV and UV-visible monochromators, pulsed electron gun (EGPS-3101, Kimball Physics) and multiscaler photon counter (MSA-300, Becker&Hickl GmbH) for accumulation of luminescence decay curves. It was possible to tune the energy of electrons (1 – 10 keV), pulse width of exciting electrons (δt) and interval between successive pulses (∆t). In the present study energy of 10 keV, δt = 100 ns and ∆t = 200 µs were used. Time-resolved low temperature luminescence studies in VUV region up to 35 eV were carried out at the SUPERLUMI station (Zimmerer, 2007) and under XUV excitation (starting from 40 eV) at the beamline BW3 (Kirm et al., 2003) of HASYLAB at DESY (Hamburg, Germany). It is important to point out that emission spectra are not corrected to the transmission of detection system because of the extremely wide (1.5-8.5 eV) energy range used in studies. The shifts of the VUV luminescence peaks recorded at the CL-setup and at the BW3 are due to differences in the analysing channel (monochromator + detector). However, there are clearly sample specific features observable in the spectra (see Fig. 1).
ACCEPTED MANUSCRIPT
3. Results and discussion
AC C
EP
TE D
M AN U
SC
RI PT
Figure 1 demonstrates the time-integrated luminescence spectra recorded at 10 K and 78 K (LNT) at the BW3 (130 eV photons) and CL-setup (10 keV electrons), respectively. In both cases penetration depth (~50 nm for 130 eV photons and ~500 nm for 10 keV e-beam) is higher than near intrinsic absorption where corresponding penetration depth is 10 nm (absorption 10-6 cm-1) in wide gap crystals. 130 eV photons are able to excite electrons from Al 2s (~118 eV) and 2p (~72 eV) core shells (see luminescence study by Kirm et al., (2003)). The selftrapped exciton (STE) emission at 7.6 eV is well pronounced for dominantly αalumina containing samples (ALO 1600 and UPA) as expected. Ratio of UV/VUV emission intensities is in favour the former ones in ALO 5050 sample, where according to the XRD up to 50 % of the sample contains small < 10 nm γAl2O3 crystallites, which may act as perfect quenching centres for mobile electronic excitations. The UPA sample has well pronounced luminescence peaks of the F+ and F centres (Fig 1a), which is not the case for the combustion synthesized samples. The latter ones have rather broad emission all over UVvisible range and inhomogeneous broadening effects play here a role in smoothing luminescence spectra as pointed out by Oja et al., (2013). A specific feature in the emission spectra of combustion synthesized alumina is the luminescence band with the maxima at 2.4 eV. In the single crystalline samples there have been two suggestions of the nature of this emission band. It has been previously assigned to Al+i centre - interstitial Al ions by Valbis and Itoh, (1991) and Surdo et al., (2005). Later the 2.4 eV emission band has been assigned to F2 centre (oxygen divacancy with 4 electrons) by Evans et al., (1994), Itou et al., (2009), Kortov et al., (2011) and Tale et al., (1996). The emission band 2.4 eV has a decay time of around 68 ms (Itou et al., 2009) or 56 µs (Surdo et al., 2005) assigned for the F2 and Al+i centre, respectively. According to Tale et al., (1996) the 2.4 eV emission is stronger in crystal with heavily doped with heterovalent impurities this means the crystals have less order than crystals with less impurities but the spectroscopic parameters do not depend on the presence or absence of impurities in the crystal thus showing the defect is intrinsic. In Kortov et al., (2011) they have proposed in alumina nanopowders with different phase composition the 2.4 eV emission is due to F2 centre because of disorder in anionic and cationic sublattices but in their work the 2.4 eV disappears after annealing at 1400 ˚C as the order is restored. In our work ALO 1600 sample has strong 2.4 eV emission after annealing at 1600 ˚C. The decays for 2.4 eV under electron-beam excitation are presented on Figure 3 and the fit results in Table 2. For the samples ALO 1600 and ALO 5050 we have fast component in one µs time range and the second component is 25 and 47
ACCEPTED MANUSCRIPT
RI PT
µs, respectively. This time constant are definitely shorter than previously measured 50 ms by Evans and Stapelbroek, (1980) and 68 ms by Itou et al., (2009) but in the same range as 56 µs by Surdo et al., (2005). Also the formation of F2 centres assumes the presence of F centres in a higher concentration, which is not supported by CL data on Figure 1. Hence, the time-resolved luminescence spectroscopy data available supports that very likely we deal with the Al+i centre in our samples. Due to complexity of emission centres it cannot be excluded that other colour and impurity centres contribute as well to recorded luminescence spectra (see Kortov et al., (2014)).
M AN U
SC
The preparation of the UPA and nanostructured alumina by combustion synthesis occurs at rather different conditions. The first one is taking place at low temperatures and moderate thermal post-treatment allows receiving desired phase composition, which for obtaining α-phase is only 1250 °C (see Museur et al., (2013)). ALO 1600 and ALO 5050 samples were annealed at 1600 °C and 1000 °C, which is still comparable with that for the UPA. Obviously, much higher temperatures during the combustion process may facilitate formation of interstitial cations along formation with other colour centres. It has been shown (Tale et al., 1996) that slow cooling facilitates aggregation of pre-existing anion vacancies into complex colour centres.
AC C
EP
TE D
In the UV range of 4 to 6 eV there is broad non-elementary luminescence band with maxima positioned from 4.8 to 4.4 eV in a row ALO 5050, ALO1600 and UPA under selective XUV excitation. In such conditions intrinsic electronic excitations are preferentially created like free excitons in MgO and STE-s in Al2O3 as shown earlier for oxides by Kirm et al., (2003). This 4.5 eV emission in combustion synthesized alumina have been tentatively assigned to the recombination of self-trapped excitons because its excitation onset coincides with intrinsic absorption of Al2O3 as shown in the paper by Oja et al., (2013). Similar hypothesis was proposed by Museur et al., (2013) for UPA where additionally two UV bands at 5.2 eV and 4.1 eV were ascribed to singlet (life time in ns range) and triplet (life time in µs range) self-trapped excitons. The CL spectra do differ from those of recorded under XUV because exciting electrons generate wide range of secondary electrons able to excite along intrinsic excitations also extrinsic ones like impurities and defects by impact mechanism investigated in details for doped alkali halides under photoexcitation (Feldbach et al., 1997) and by energy transfer processes by electrons and holes. It results in narrow emission bands typical for transition metals impurities (Kortov et al., 2014) observed between broad luminescence bands in ALO 5050 sample (Fig. 1 c).
ACCEPTED MANUSCRIPT
SC
RI PT
In Figure 2 the excitation spectra for 7.6 eV STE emission near intrinsic absorption region are shown for all samples at 10 K. In agreement with earlier studies (Kirm et al., 1999) the corresponding peak for single crystal has a maximum at 8.95 eV and is significantly narrower width than that of UPA and ALO 1600 samples. Al2O3 is an optically anisotropic material, where a small shift can be introduced if samples are at different orientation in respect to the polarised synchrotron radiation. Such studies have been performed for other low symmetry materials like tungstate crystals (see Feldbach et al., (2000) and references therein) and BeO (Pustovarov et al., 2001). The studied single crystal had the caxis in the direction normal to the incident synchrotron radiation and therefore no orientation effects are expected. Nanopowders naturally do not have a preferential orientation. Therefore, the observed shifts by 0.15 eV are not due such anisotropy effects.
TE D
M AN U
The nanosize structure of samples can cause quantum confinement effects responsible, which can modulate the energy states in the system. This has been shown to occur in semiconductors with large radius excitons (Wang and Herron, 1991). In alumina one deals with strongly localised intrinsic excitations so called self-shrunk excitons (Kirm et al., 1999), which spatial extent is in order of a lattice constant. Therefore, this effect for crystallites with size of a few hundred nm for α-alumina studied cannot cause significant changes of electronic states. It has been considered (Song et al., 1989) that the residual stress effects due treatment of sample surface can cause such shifts in the reflection spectra. This is not excluded, but no special procedures were applied to the nanopowders studied, which leaves open the possibility that internal stress is induced during synthesis. Obviously, the nanostructure of samples in α-alumina is causing this 0.15 eV shift of the excitation peak, but its origin needs further studies.
AC C
EP
Figure 3 shows decays of several emission bands in UV and visible spectral range excited by pulsed electrons at LNT. The decay curve fitting results together with literature results at RT (from Surdo et al., (2005)) are collected in the Table 2. By selecting the excitation pulse length 100 ns, the intentional discrimination of ultrafast F+ luminescence at 3.8 eV with τ - 2 ns (Surdo et al., 2001, 2005). Our study show that decay times of tens of µs are characteristic for recombination processes populating states of F+ and F centres, respectively. Kortov et al., (2008) recorded decays for nanostructured alumina powders. According to their studies the bands in region 3.0 – 4.5 eV have microsecond decay time constants. In our work for 4.5 eV emission we identified a complex decay with sub-µs decay constant, time constant about 10 µs and for samples ALO 5050 and ALO 1600 time constant at least 0.1 ms. Such decay time support our hypothesis of triplet self-trapped excitons with emission peak in UV range below 5 eV (Oja et.al.,
ACCEPTED MANUSCRIPT
RI PT
2013). It is known from the literature that α-alumina has several decay components extending from ns to µs range (Mürk and Ismailov, 1993) due to the more complex exciton structure and relaxation processes than alkali halides. The most probable cause of observed differences in various decay components of studied samples is non-radiative relaxation channels becoming more dominant with decreasing crystallite size.
4. Conclusions
TE D
Acknowledgements
M AN U
SC
We have studied nanostructured alumina samples and showed that the excitonic excitation peak is shifted to 9.1 eV, which is by 0.15 eV higher energies in comparison with single crystals. The shift is arising due to the nanostructure of the samples. The origin and spectral-kinetic properties of complex luminescence bands were investigated and discussed. The UV emission in the range of 4.4 -4.8 eV with microsecond lifetime is assigned to radiative recombination of triplet self-trapped excitons. In combustion synthesised alumina the well pronounced 2.4 eV emission with microsecond lifetime is due to intrinsic defects, which are known for single crystals i.e. more likely Al+i centre, but contribution for F2 centres is not excluded.
AC C
EP
We are grateful for Dr. Kanaev who kindly provided UPA alumina samples and for challenging discussions. Our research has received funding from the European Community's Seventh Framework Programme (FP7/2007-2013) under grant agreement n° 312284 supporting activities al large scale facilities. This work has been also supported by Estonian Materials Technology program project 3.2.1101.12-0014 „Synthesis, characterization and application of activated rareearth compounds in synthetic fuel synthesis reactors and other high-tech devices“ at University of Tartu, Estonia and from the Estonian Research Council grant IUT-2-26. The support from the French Government (the BGF scholarship) is gratefully acknowledged by E. Tõldsepp and from the Estonian-French cooperation program „G. F. PARROT” is kindly appreciated.
ACCEPTED MANUSCRIPT
References Evans, B.D., Pogatshnik, G.J., Chen, Y., 1994. Optical properties of lattice defects in α-Al2O3. Nucl. Instruments Methods Phys. Res. Sect. B Beam Interact. with Mater. Atoms 91, 258–262. doi:10.1016/0168-583X(94)96227-8
RI PT
Evans, B.D., Stapelbroek, M., 1980. Optical vibronic absorption spectra in 14.8 MeV neutron damaged sapphire. Solid State Commun. 33, 765–770. doi:10.1016/0038-1098(80)90827-3
SC
Feldbach, E., Jönsson, L., Kirm, M., Kotlov, A., Lushchik, A., Nagirnyi, V., Svensson, G., Åsberg-Dahlborg, M., 2000. Polarized emission of PbWO4 and CdWO4 crystals. J. Lumin. 87-89, 1213–1216. doi:10.1016/S00222313(99)00516-5
M AN U
Feldbach, E., Kamada, M., Kirm, M., Lushchik, A., Lushchik, Ch., Martinson, I., 1997. Direct excitation of Tl+ impurity ions by hot photoelectrons in wide-gap crystals. Phys. Rev. B 56, 13908–13915. doi:10.1103/PhysRevB.56.13908 Ianoş, R., Lazău, I., Păcurariu, C., 2009. The influence of combustion synthesis conditions on the α-Al2O3 powder preparation. J. Mater. Sci. 44, 1016–1023. doi:10.1007/s10853-008-3226-5
TE D
Itou, M., Fujiwara, A., Uchino, T., 2009. Reversible photoinduced interconversion of color centers in α-Al2O3 prepared under vacuum. J. Phys. Chem. C 113, 20949–20957. doi:10.1021/jp908417m
EP
Kirm, M., Feldbach, E., Kotlov, A., Liblik, P., Lushchik, A., Oja, M., Palcevskis, E., 2010. VUV spectroscopy and electronic excitations in nano-size alumina. Radiat. Meas. 45, 618–620. doi:10.1016/j.radmeas.2009.12.006
AC C
Kirm, M., Lushchik, A., Lushchik, Ch., Vielhauer, S., Zimmerer, G., 2003. Luminescence of pure and doped Al2O3 and MgO single crystals under inner-shell excitation. J. Lumin. 102-103, 307–312. doi:10.1016/S0022-2313(02)00521-5 Kirm, M., Zimmerer, G., Feldbach, E., Lushchik, A., Lushchik, Ch., Savikhin, F., 1999. Self-trapping and multiplication of electronic excitations in Al2O3 and Al2O3:Sc crystals. Phys. Rev. B 60, 502–510. doi:10.1103/PhysRevB.60.502 Kortov, V.S., Ermakov, A. E., Zatsepin, A. F., Uĭmin, M. A., Nikiforov, S. V., Mysik, A. A., Gaviko, V.S., 2008. Specific features of luminescence properties of nanostructured aluminum oxide. Phys. Solid State 50, 957–961. doi:10.1134/S1063783408050259
ACCEPTED MANUSCRIPT Kortov, V.S., Pustovarov, V.A., Spiridonova, T.V., and Zvonarev, S.V., 2014. Photoluminescence of ultradisperse alumina ceramics under VUV excitation J. Appl. Spectroscopy 80, 835-840. doi:10.1007/s10812-014-9852-6
RI PT
Kortov, V.S., Zvonarev S. V., and Medvedev, A. I., 2011. Pulsed cathodoluminescence of nanoscale aluminium oxide with different phase composition J. lumen. 131, 1904-1907. doi: 10.1016/j.jlumin.2011.05.006 Lushchik, A., Feldbach, E., Lushchik, Ch., Kirm, M., Martinson, I., 1994. Multiplication mechanisms of electronic excitations in KBr and KBr:Tl crystals. Phys. Rev. B 50, 6500–6503. doi:10.1103/PhysRevB.50.6500
SC
Mürk, V. V., Ismailov, K. M., 1993. Decay kinetics of the luminescence of selftrapped excitons in Al2O3 and Y3Al5O12. Phys. solid state 35, 259–260.
M AN U
Museur, L., Bouslama, M., Amamra, M., Kanaev, A., 2013. Electronic transitions in α, θ and γ polymorphs of ultraporous monolithic alumina. Phys. status solidi Rapid Res. Lett. 7, 1026–1029. doi:10.1002/pssr.201308040 Oja, M., Feldbach, E., Kotlov, A., Mändar, H., Vielhauer, S., Kirm, M., 2013. Intrinsic and extrinsic luminescence of nanosize transition alumina powders. Radiat. Meas. 56, 411–414. doi:10.1016/j.radmeas.2013.01.067
TE D
Pustovarov, V. A., Ivanov, V.Y., Kirm, M., Kruzhalov, A. V., Korotaev, A. V., Zimmerer, G., 2001. Relaxation of electronic excitations in beryllium oxide: A time-resolved vacuum-UV spectroscopy study. Phys. Solid State 43, 1233–1240. doi:10.1134/1.1386456
EP
Pustovarov, V. A., Kortov, V. S., Zvonarev, S. V., Medvedev, A. I., 2012. Luminescent vacuum ultraviolet spectroscopy of Cr3+ ions in nanostructured aluminum oxide. J. Lumin. 132, 2868–2873. doi:10.1016/j.jlumin.2012.06.001
AC C
Song H., French R.H., Coble R.L., 1989. Effect of Residual Strain on the Electronic Structure of Alumina and Magnesia. J. Am. Ceram. Soc., 72, 990-994. doi:10.1111/j.1151-2916.1989.tb06257.x Surdo, A. I., Pustovarov, V. A., Kortov, V. S., Kishka, A. S., Zinin, E. I., 2005. Luminescence in anion-defective α-Al2O3 crystals over the nano-, micro- and millisecond intervals. Nucl. Instruments Methods Phys. Res. Sect. A Accel. Spectrometers, Detect. Assoc. Equip. 543, 234–238. doi:10.1016/j.nima.2005.01.189
ACCEPTED MANUSCRIPT Surdo, A. I., Kortov, V. S., Pustovarov, V. A., 2001. Luminescence of F and F+ centers in corundum upon excitation in the interval from 4 to 40 eV. Radiat. Meas. 33, 587–591. doi:10.1016/S1350-4487(01)00064-6
RI PT
Tale, I., Piters, T. M., Barboza-Flores, M., Perez-Salas, R., Aceves, R., Springis, M., 1996. Optical Properties of Complex Anion Vacancy Centres and PhotoExcited Electronic Processes in Anion Defective Alpha-Al2O3. Radiat. Prot. Dosimetry. 65, 235-238. Trinkler, L., Berzina, B., Jevsjutina, Z., Grabis, J., Steins, I., Baily, C. J., 2012. Photoluminescence of Al2O3 nanopowders of different phases. Opt. Mater. 34, 1553–1557. doi:10.1016/j.optmat.2012.03.029
M AN U
SC
Valbis, J., Itoh, N., 1991. Electronic excitations, luminescence and lattice defect formation in α-Al2O3 crystals. Radiat. Eff. Defects Solids 116, 171–189. doi:10.1080/10420159108221357
Vignes, J.-L., Frappart, C., Di Costanzo, T., Rouchaud, J.-C., Mazerolles, L., Michel, D., 2008. Ultraporous monoliths of alumina prepared at room temperature by aluminium oxidation. J. Mater. Sci. 43, 1234–1240. doi:10.1007/s10853-0072260-z
TE D
Wang, Y., Herron, N., 1991. Nanometer-sized semiconductor clusters: materials synthesis, quantum size effects, and photophysical properties. J. Phys. Chem. 95, 525–532. doi:10.1021/j100155a009
AC C
EP
Zimmerer, G., 2007. SUPERLUMI: A unique setup for luminescence spectroscopy with synchrotron radiation. Radiat. Meas. 42, 859–864. doi:10.1016/j.radmeas.2007.02.050
ACCEPTED MANUSCRIPT Figure captions
RI PT
Fig. 1. TI emission spectra excited by 10keV electrons (solid lines) at 78 K and 130 eV photons (dashed lines) at 10K. a) Blue – Al2O3 UPA; b) Green – Al2O3 1600; c) Magenta – Al2O3 5050. CL spectra were recorded with two monochromators. The range above 6 eV with the VUV instrument and below with the UV-visible spectrometer. Fig. 2. TI excitation spectra for 7.6 eV luminescence at 10K. Black (*) – Al2O3 crystal; Green (+) – Al2O3 1600 and blue (o) – Al2O3 UPA.
AC C
EP
TE D
M AN U
SC
Fig. 3 Decay curves for alumina samples excited by pulsed (100 ns) 10 keV electrons at 78 K. blue (1) – UPA, green (2) – 1600, magenta (3) – 5050. Decays for emissions a) 2.6 eV, b) 3.0 eV, c) 3.8 eV and d) 4.5 eV.
ACCEPTED MANUSCRIPT
Table 1 Phase composition of studied Al2O3 samples and estimated crystallite size based on the XRD analysis Phase composition
Al2O3 1600 Al2O3 5050
α-phase α-phase (~40-50%) γ-phase (~40-50%) α -phase
Al2O3 UPA
Volume weighted crystallite size α-phase ~180nm α-phase ~80nm γ-phase <10nm α -phase ~200-300 nm
RI PT
Sample name
2.4 eV decay
3.8 eV decay
4.5 eV decay
Al2O3 1600
1.5 µs 25.1 µs
0.5 µs 22.9 µs
0.7 µs 28.8 µs
0.6 µs 7.8 µs 148.3 µs
Al2O3 5050
0.7 µs 47.6 µs
0.7 µs 14.2 µs
0.6 µs 19.4 µs
0.6 µs 7.9 µs >200 µs
0.6 µs 7.8 µs
0.4 µs 4.7 µs
0.5 µs 16.3 µs
0.8 µs 10.9 µs
34 ms
2 ns
TE D
3.0 eV decay
EP
Sample name
M AN U
SC
Table 2 Fitted decay parameters for Al2O3 1600, Al2O3 5050 and Al2O3 UPA UVvisible luminescence bands compared with literature (Surdo et al., 2005) values at room temperature (RT).
Al2O3 UPA
AC C
Decay times from literature at RT
56 µs
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIPT Highlights: Combustion synthesized alumina is compared with phase pure one using luminescence; Excitonic excitation peak at 9.1 eV is shifted towards higher energies in nanopowders;
AC C
EP
TE D
M AN U
SC
RI PT
Time resolved luminescence was used to identify nature of emission centres
S1350-4487(15)30093-7
DOI:
10.1016/j.radmeas.2015.12.014
Reference:
RM 5495
To appear in:
Radiation Measurements
Received Date: 25 October 2015 Revised Date:
15 December 2015
Accepted Date: 19 December 2015
Please cite this article as: Oja, M., Tõldsepp, E., Feldbach, E., Mägi, H., Omelkov, S., Kirm, M., Luminescence Study of Alumina Nanopowders Prepared by Various Methods, Radiation Measurements (2016), doi: 10.1016/j.radmeas.2015.12.014. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT
Luminescence Study of Alumina Nanopowders Prepared by Various Methods Marek Oja*, Eliko Tõldsepp, Eduard Feldbach, Henri Mägi, Sergey Omelkov, and Marco Kirm
RI PT
Institute of Physics, University of Tartu, Ravila 14C, 50 411, Tartu, Estonia * Corresponding author. Tel.: +372 737 4762, email: [email protected]
SC
Abstract
TE D
M AN U
Mixed nanopowders of transition alumina prepared by combustion synthesis and phase pure ultra-porous α-alumina by oxidation method were investigated using low temperature time-resolved cathodoluminescence and photoluminescence spectroscopy under VUV-XUV excitation. In all samples along with the 7.6 eV emission of self-trapped excitons of α-alumina, luminescence bands due to F, F+ centres with maxima at 3 and 3.8 eV and other UV-visible luminescence bands of intrinsic and extrinsic origin with varying intensity depending on sample preparation method and thermal treatment were studied. In alumina nanopowders the excitonic excitation peak at ~9.1 eV near fundamental absorption edge is shifted to the higher energies by 0.15 eV in comparison with the same feature in single crystals. The nanostructure of alumina is responsible for this shift.
EP
Keywords: transition and α-alumina, relaxation of electronic excitations, intrinsic and extrinsic luminescence, nanopowders
1. Introduction
AC C
Aluminium oxide (Al2O3) is an important material for different optical and technological applications in the form of ceramics or single crystals, but there is a growing interest in nanostructured alumina, because of its large surface area being favourable for catalyst support and filtration. Significant research has been carried out on single crystals of α-Al2O3 using luminescence and other methods (see Kirm et al., (1999); Valbis and Itoh, (1991) and references therein). Recently, various modification (incl. α-, θ-, γ-Al2O3) of phase pure ultra-porous alumina nanopowder (hereafter called as UPA) with the low impurity content was investigated using low temperature time-resolved luminescence spectroscopy (Museur et al., 2013).
ACCEPTED MANUSCRIPT
M AN U
SC
RI PT
Metastable aluminium oxide phases (e.g. δ-, θ-, κ- and γ-Al2O3) with the common name of transition alumina also exist in the nature. The latter phases have different crystal structure, electronic properties and have been significantly less studied, but such activities based on luminescence spectroscopy are rising (e.g. Kirm et al., (2010); Oja et al., (2013); Pustovarov et al., (2012); Trinkler et al., (2012)). Due to their thermochemical properties the preparation of pure phase samples is a real challenge and therefore metastable alumina samples typically contain multiple Al2O3-phases complicating identification of elementary processes in such mixtures. Depending on the preparation method often unintended impurities are introduced into alumina samples. In order to reduce influence of surface related effects (luminescence quenching etc.) the XUV and electron beam excitation is applied, methods which penetrate deeper into the samples prohibiting near-surface effects at fundamental absorption regime, which has to be considered in investigation of wide gap crystals (Lushchik et al., 1994). The main goal of this work is to enlighten electronic properties, relaxation of electronic excitations, identify phase and crystallite size specific features in luminescence of alumina samples prepared with different synthesis methods.
TE D
2. Sample preparation, structural properties and experimental techniques
AC C
EP
Nanostructured UPA was produced using oxidation of high purity aluminum (99.999 %) through a liquid mercury and silver film, which leads to formation of highly porous (99%) macroscopic (cm dimensions), but nanostructured (tangled alumina fibers of 5 nm diameter) monoliths as described by Vignes et al., (2008). The further post-treatment at elevated temperatures (1200-1250 °C) results in nanostructured α-alumina with characteristic crystallite size of 200-300 nm as described in the paper by Museur et al., (2013). This method has advantage over combustion and solid state synthesis as amount of impurities is well controlled by using high purity aluminium and its impurity content is smaller in comparison with samples prepared by combustion synthesis. Two samples were prepared using the combustion synthesis method analogous to Oja et al., (2013). The chemicals used in the synthesis of these samples were Al(NO3)3·9H2O (for the ALO 1600 - Alfa Aesar, 99.999% and for the ALO 5050 - REAHIM, 99.99%) and urea (Alfa Aesar) as a fuel. Sample ALO 1600 was produced using stoichiometric mixture with fuel/nitrate molar ratio 5:2 whereas
ACCEPTED MANUSCRIPT
RI PT
sample ALO 5050 was synthesised using fuel rich mixture which corresponds to a molar ratio of 25:4. According to Ianoş et al., (2009) fuel-rich synthesis gives lower combustion temperature and therefore different phase composition and higher specific surface area compared to stoichiometric synthesis. Burning out of residual carbon and phase composition adjustment is achieved by an annealing procedure resulting in an increase of crystallite size. Samples were annealed in air at 1600˚C for 5 hours (ALO 1600) and at 1000˚C for 1 hour (ALO 5050). After annealing procedure the sample ALO 1600 was white powder but sample ALO 5050 was still greyish powder indicating either presence of carbon or absorption of oxygen vacancies.
TE D
M AN U
SC
The phase composition and crystallite size of the samples were determined by powder X-ray diffraction (XRD) analysis and the obtained values are presented in Table 1. Sample ALO 5050 consists of several Al2O3 phases with different crystallite size, which is a general complication in studies of transition alumina whereas other studied samples were composed of α-Al2O3 phase. Analysis of the quantitative phase composition (using the Rietveld method, program FULLPROF) is not straight forward because of the overlap of XRD patterns from various phases are responsible for relatively large a few percent error bars on the obtained values. The γ-Al2O3 phase has only very broad diffraction lines referring to small crystallite size below 10 nm. Although there can be an a few nm error in absolute size, the crystallite size of α-Al2O3 is practically an order of magnitude larger (see Table 1). Crystallite size estimations were performed using the Scherrer equation.
AC C
EP
Low temperature cathodoluminescence (CL) studies were performed in Tartu using a home-built setup equipped with VUV and UV-visible monochromators, pulsed electron gun (EGPS-3101, Kimball Physics) and multiscaler photon counter (MSA-300, Becker&Hickl GmbH) for accumulation of luminescence decay curves. It was possible to tune the energy of electrons (1 – 10 keV), pulse width of exciting electrons (δt) and interval between successive pulses (∆t). In the present study energy of 10 keV, δt = 100 ns and ∆t = 200 µs were used. Time-resolved low temperature luminescence studies in VUV region up to 35 eV were carried out at the SUPERLUMI station (Zimmerer, 2007) and under XUV excitation (starting from 40 eV) at the beamline BW3 (Kirm et al., 2003) of HASYLAB at DESY (Hamburg, Germany). It is important to point out that emission spectra are not corrected to the transmission of detection system because of the extremely wide (1.5-8.5 eV) energy range used in studies. The shifts of the VUV luminescence peaks recorded at the CL-setup and at the BW3 are due to differences in the analysing channel (monochromator + detector). However, there are clearly sample specific features observable in the spectra (see Fig. 1).
ACCEPTED MANUSCRIPT
3. Results and discussion
AC C
EP
TE D
M AN U
SC
RI PT
Figure 1 demonstrates the time-integrated luminescence spectra recorded at 10 K and 78 K (LNT) at the BW3 (130 eV photons) and CL-setup (10 keV electrons), respectively. In both cases penetration depth (~50 nm for 130 eV photons and ~500 nm for 10 keV e-beam) is higher than near intrinsic absorption where corresponding penetration depth is 10 nm (absorption 10-6 cm-1) in wide gap crystals. 130 eV photons are able to excite electrons from Al 2s (~118 eV) and 2p (~72 eV) core shells (see luminescence study by Kirm et al., (2003)). The selftrapped exciton (STE) emission at 7.6 eV is well pronounced for dominantly αalumina containing samples (ALO 1600 and UPA) as expected. Ratio of UV/VUV emission intensities is in favour the former ones in ALO 5050 sample, where according to the XRD up to 50 % of the sample contains small < 10 nm γAl2O3 crystallites, which may act as perfect quenching centres for mobile electronic excitations. The UPA sample has well pronounced luminescence peaks of the F+ and F centres (Fig 1a), which is not the case for the combustion synthesized samples. The latter ones have rather broad emission all over UVvisible range and inhomogeneous broadening effects play here a role in smoothing luminescence spectra as pointed out by Oja et al., (2013). A specific feature in the emission spectra of combustion synthesized alumina is the luminescence band with the maxima at 2.4 eV. In the single crystalline samples there have been two suggestions of the nature of this emission band. It has been previously assigned to Al+i centre - interstitial Al ions by Valbis and Itoh, (1991) and Surdo et al., (2005). Later the 2.4 eV emission band has been assigned to F2 centre (oxygen divacancy with 4 electrons) by Evans et al., (1994), Itou et al., (2009), Kortov et al., (2011) and Tale et al., (1996). The emission band 2.4 eV has a decay time of around 68 ms (Itou et al., 2009) or 56 µs (Surdo et al., 2005) assigned for the F2 and Al+i centre, respectively. According to Tale et al., (1996) the 2.4 eV emission is stronger in crystal with heavily doped with heterovalent impurities this means the crystals have less order than crystals with less impurities but the spectroscopic parameters do not depend on the presence or absence of impurities in the crystal thus showing the defect is intrinsic. In Kortov et al., (2011) they have proposed in alumina nanopowders with different phase composition the 2.4 eV emission is due to F2 centre because of disorder in anionic and cationic sublattices but in their work the 2.4 eV disappears after annealing at 1400 ˚C as the order is restored. In our work ALO 1600 sample has strong 2.4 eV emission after annealing at 1600 ˚C. The decays for 2.4 eV under electron-beam excitation are presented on Figure 3 and the fit results in Table 2. For the samples ALO 1600 and ALO 5050 we have fast component in one µs time range and the second component is 25 and 47
ACCEPTED MANUSCRIPT
RI PT
µs, respectively. This time constant are definitely shorter than previously measured 50 ms by Evans and Stapelbroek, (1980) and 68 ms by Itou et al., (2009) but in the same range as 56 µs by Surdo et al., (2005). Also the formation of F2 centres assumes the presence of F centres in a higher concentration, which is not supported by CL data on Figure 1. Hence, the time-resolved luminescence spectroscopy data available supports that very likely we deal with the Al+i centre in our samples. Due to complexity of emission centres it cannot be excluded that other colour and impurity centres contribute as well to recorded luminescence spectra (see Kortov et al., (2014)).
M AN U
SC
The preparation of the UPA and nanostructured alumina by combustion synthesis occurs at rather different conditions. The first one is taking place at low temperatures and moderate thermal post-treatment allows receiving desired phase composition, which for obtaining α-phase is only 1250 °C (see Museur et al., (2013)). ALO 1600 and ALO 5050 samples were annealed at 1600 °C and 1000 °C, which is still comparable with that for the UPA. Obviously, much higher temperatures during the combustion process may facilitate formation of interstitial cations along formation with other colour centres. It has been shown (Tale et al., 1996) that slow cooling facilitates aggregation of pre-existing anion vacancies into complex colour centres.
AC C
EP
TE D
In the UV range of 4 to 6 eV there is broad non-elementary luminescence band with maxima positioned from 4.8 to 4.4 eV in a row ALO 5050, ALO1600 and UPA under selective XUV excitation. In such conditions intrinsic electronic excitations are preferentially created like free excitons in MgO and STE-s in Al2O3 as shown earlier for oxides by Kirm et al., (2003). This 4.5 eV emission in combustion synthesized alumina have been tentatively assigned to the recombination of self-trapped excitons because its excitation onset coincides with intrinsic absorption of Al2O3 as shown in the paper by Oja et al., (2013). Similar hypothesis was proposed by Museur et al., (2013) for UPA where additionally two UV bands at 5.2 eV and 4.1 eV were ascribed to singlet (life time in ns range) and triplet (life time in µs range) self-trapped excitons. The CL spectra do differ from those of recorded under XUV because exciting electrons generate wide range of secondary electrons able to excite along intrinsic excitations also extrinsic ones like impurities and defects by impact mechanism investigated in details for doped alkali halides under photoexcitation (Feldbach et al., 1997) and by energy transfer processes by electrons and holes. It results in narrow emission bands typical for transition metals impurities (Kortov et al., 2014) observed between broad luminescence bands in ALO 5050 sample (Fig. 1 c).
ACCEPTED MANUSCRIPT
SC
RI PT
In Figure 2 the excitation spectra for 7.6 eV STE emission near intrinsic absorption region are shown for all samples at 10 K. In agreement with earlier studies (Kirm et al., 1999) the corresponding peak for single crystal has a maximum at 8.95 eV and is significantly narrower width than that of UPA and ALO 1600 samples. Al2O3 is an optically anisotropic material, where a small shift can be introduced if samples are at different orientation in respect to the polarised synchrotron radiation. Such studies have been performed for other low symmetry materials like tungstate crystals (see Feldbach et al., (2000) and references therein) and BeO (Pustovarov et al., 2001). The studied single crystal had the caxis in the direction normal to the incident synchrotron radiation and therefore no orientation effects are expected. Nanopowders naturally do not have a preferential orientation. Therefore, the observed shifts by 0.15 eV are not due such anisotropy effects.
TE D
M AN U
The nanosize structure of samples can cause quantum confinement effects responsible, which can modulate the energy states in the system. This has been shown to occur in semiconductors with large radius excitons (Wang and Herron, 1991). In alumina one deals with strongly localised intrinsic excitations so called self-shrunk excitons (Kirm et al., 1999), which spatial extent is in order of a lattice constant. Therefore, this effect for crystallites with size of a few hundred nm for α-alumina studied cannot cause significant changes of electronic states. It has been considered (Song et al., 1989) that the residual stress effects due treatment of sample surface can cause such shifts in the reflection spectra. This is not excluded, but no special procedures were applied to the nanopowders studied, which leaves open the possibility that internal stress is induced during synthesis. Obviously, the nanostructure of samples in α-alumina is causing this 0.15 eV shift of the excitation peak, but its origin needs further studies.
AC C
EP
Figure 3 shows decays of several emission bands in UV and visible spectral range excited by pulsed electrons at LNT. The decay curve fitting results together with literature results at RT (from Surdo et al., (2005)) are collected in the Table 2. By selecting the excitation pulse length 100 ns, the intentional discrimination of ultrafast F+ luminescence at 3.8 eV with τ - 2 ns (Surdo et al., 2001, 2005). Our study show that decay times of tens of µs are characteristic for recombination processes populating states of F+ and F centres, respectively. Kortov et al., (2008) recorded decays for nanostructured alumina powders. According to their studies the bands in region 3.0 – 4.5 eV have microsecond decay time constants. In our work for 4.5 eV emission we identified a complex decay with sub-µs decay constant, time constant about 10 µs and for samples ALO 5050 and ALO 1600 time constant at least 0.1 ms. Such decay time support our hypothesis of triplet self-trapped excitons with emission peak in UV range below 5 eV (Oja et.al.,
ACCEPTED MANUSCRIPT
RI PT
2013). It is known from the literature that α-alumina has several decay components extending from ns to µs range (Mürk and Ismailov, 1993) due to the more complex exciton structure and relaxation processes than alkali halides. The most probable cause of observed differences in various decay components of studied samples is non-radiative relaxation channels becoming more dominant with decreasing crystallite size.
4. Conclusions
TE D
Acknowledgements
M AN U
SC
We have studied nanostructured alumina samples and showed that the excitonic excitation peak is shifted to 9.1 eV, which is by 0.15 eV higher energies in comparison with single crystals. The shift is arising due to the nanostructure of the samples. The origin and spectral-kinetic properties of complex luminescence bands were investigated and discussed. The UV emission in the range of 4.4 -4.8 eV with microsecond lifetime is assigned to radiative recombination of triplet self-trapped excitons. In combustion synthesised alumina the well pronounced 2.4 eV emission with microsecond lifetime is due to intrinsic defects, which are known for single crystals i.e. more likely Al+i centre, but contribution for F2 centres is not excluded.
AC C
EP
We are grateful for Dr. Kanaev who kindly provided UPA alumina samples and for challenging discussions. Our research has received funding from the European Community's Seventh Framework Programme (FP7/2007-2013) under grant agreement n° 312284 supporting activities al large scale facilities. This work has been also supported by Estonian Materials Technology program project 3.2.1101.12-0014 „Synthesis, characterization and application of activated rareearth compounds in synthetic fuel synthesis reactors and other high-tech devices“ at University of Tartu, Estonia and from the Estonian Research Council grant IUT-2-26. The support from the French Government (the BGF scholarship) is gratefully acknowledged by E. Tõldsepp and from the Estonian-French cooperation program „G. F. PARROT” is kindly appreciated.
ACCEPTED MANUSCRIPT
References Evans, B.D., Pogatshnik, G.J., Chen, Y., 1994. Optical properties of lattice defects in α-Al2O3. Nucl. Instruments Methods Phys. Res. Sect. B Beam Interact. with Mater. Atoms 91, 258–262. doi:10.1016/0168-583X(94)96227-8
RI PT
Evans, B.D., Stapelbroek, M., 1980. Optical vibronic absorption spectra in 14.8 MeV neutron damaged sapphire. Solid State Commun. 33, 765–770. doi:10.1016/0038-1098(80)90827-3
SC
Feldbach, E., Jönsson, L., Kirm, M., Kotlov, A., Lushchik, A., Nagirnyi, V., Svensson, G., Åsberg-Dahlborg, M., 2000. Polarized emission of PbWO4 and CdWO4 crystals. J. Lumin. 87-89, 1213–1216. doi:10.1016/S00222313(99)00516-5
M AN U
Feldbach, E., Kamada, M., Kirm, M., Lushchik, A., Lushchik, Ch., Martinson, I., 1997. Direct excitation of Tl+ impurity ions by hot photoelectrons in wide-gap crystals. Phys. Rev. B 56, 13908–13915. doi:10.1103/PhysRevB.56.13908 Ianoş, R., Lazău, I., Păcurariu, C., 2009. The influence of combustion synthesis conditions on the α-Al2O3 powder preparation. J. Mater. Sci. 44, 1016–1023. doi:10.1007/s10853-008-3226-5
TE D
Itou, M., Fujiwara, A., Uchino, T., 2009. Reversible photoinduced interconversion of color centers in α-Al2O3 prepared under vacuum. J. Phys. Chem. C 113, 20949–20957. doi:10.1021/jp908417m
EP
Kirm, M., Feldbach, E., Kotlov, A., Liblik, P., Lushchik, A., Oja, M., Palcevskis, E., 2010. VUV spectroscopy and electronic excitations in nano-size alumina. Radiat. Meas. 45, 618–620. doi:10.1016/j.radmeas.2009.12.006
AC C
Kirm, M., Lushchik, A., Lushchik, Ch., Vielhauer, S., Zimmerer, G., 2003. Luminescence of pure and doped Al2O3 and MgO single crystals under inner-shell excitation. J. Lumin. 102-103, 307–312. doi:10.1016/S0022-2313(02)00521-5 Kirm, M., Zimmerer, G., Feldbach, E., Lushchik, A., Lushchik, Ch., Savikhin, F., 1999. Self-trapping and multiplication of electronic excitations in Al2O3 and Al2O3:Sc crystals. Phys. Rev. B 60, 502–510. doi:10.1103/PhysRevB.60.502 Kortov, V.S., Ermakov, A. E., Zatsepin, A. F., Uĭmin, M. A., Nikiforov, S. V., Mysik, A. A., Gaviko, V.S., 2008. Specific features of luminescence properties of nanostructured aluminum oxide. Phys. Solid State 50, 957–961. doi:10.1134/S1063783408050259
ACCEPTED MANUSCRIPT Kortov, V.S., Pustovarov, V.A., Spiridonova, T.V., and Zvonarev, S.V., 2014. Photoluminescence of ultradisperse alumina ceramics under VUV excitation J. Appl. Spectroscopy 80, 835-840. doi:10.1007/s10812-014-9852-6
RI PT
Kortov, V.S., Zvonarev S. V., and Medvedev, A. I., 2011. Pulsed cathodoluminescence of nanoscale aluminium oxide with different phase composition J. lumen. 131, 1904-1907. doi: 10.1016/j.jlumin.2011.05.006 Lushchik, A., Feldbach, E., Lushchik, Ch., Kirm, M., Martinson, I., 1994. Multiplication mechanisms of electronic excitations in KBr and KBr:Tl crystals. Phys. Rev. B 50, 6500–6503. doi:10.1103/PhysRevB.50.6500
SC
Mürk, V. V., Ismailov, K. M., 1993. Decay kinetics of the luminescence of selftrapped excitons in Al2O3 and Y3Al5O12. Phys. solid state 35, 259–260.
M AN U
Museur, L., Bouslama, M., Amamra, M., Kanaev, A., 2013. Electronic transitions in α, θ and γ polymorphs of ultraporous monolithic alumina. Phys. status solidi Rapid Res. Lett. 7, 1026–1029. doi:10.1002/pssr.201308040 Oja, M., Feldbach, E., Kotlov, A., Mändar, H., Vielhauer, S., Kirm, M., 2013. Intrinsic and extrinsic luminescence of nanosize transition alumina powders. Radiat. Meas. 56, 411–414. doi:10.1016/j.radmeas.2013.01.067
TE D
Pustovarov, V. A., Ivanov, V.Y., Kirm, M., Kruzhalov, A. V., Korotaev, A. V., Zimmerer, G., 2001. Relaxation of electronic excitations in beryllium oxide: A time-resolved vacuum-UV spectroscopy study. Phys. Solid State 43, 1233–1240. doi:10.1134/1.1386456
EP
Pustovarov, V. A., Kortov, V. S., Zvonarev, S. V., Medvedev, A. I., 2012. Luminescent vacuum ultraviolet spectroscopy of Cr3+ ions in nanostructured aluminum oxide. J. Lumin. 132, 2868–2873. doi:10.1016/j.jlumin.2012.06.001
AC C
Song H., French R.H., Coble R.L., 1989. Effect of Residual Strain on the Electronic Structure of Alumina and Magnesia. J. Am. Ceram. Soc., 72, 990-994. doi:10.1111/j.1151-2916.1989.tb06257.x Surdo, A. I., Pustovarov, V. A., Kortov, V. S., Kishka, A. S., Zinin, E. I., 2005. Luminescence in anion-defective α-Al2O3 crystals over the nano-, micro- and millisecond intervals. Nucl. Instruments Methods Phys. Res. Sect. A Accel. Spectrometers, Detect. Assoc. Equip. 543, 234–238. doi:10.1016/j.nima.2005.01.189
ACCEPTED MANUSCRIPT Surdo, A. I., Kortov, V. S., Pustovarov, V. A., 2001. Luminescence of F and F+ centers in corundum upon excitation in the interval from 4 to 40 eV. Radiat. Meas. 33, 587–591. doi:10.1016/S1350-4487(01)00064-6
RI PT
Tale, I., Piters, T. M., Barboza-Flores, M., Perez-Salas, R., Aceves, R., Springis, M., 1996. Optical Properties of Complex Anion Vacancy Centres and PhotoExcited Electronic Processes in Anion Defective Alpha-Al2O3. Radiat. Prot. Dosimetry. 65, 235-238. Trinkler, L., Berzina, B., Jevsjutina, Z., Grabis, J., Steins, I., Baily, C. J., 2012. Photoluminescence of Al2O3 nanopowders of different phases. Opt. Mater. 34, 1553–1557. doi:10.1016/j.optmat.2012.03.029
M AN U
SC
Valbis, J., Itoh, N., 1991. Electronic excitations, luminescence and lattice defect formation in α-Al2O3 crystals. Radiat. Eff. Defects Solids 116, 171–189. doi:10.1080/10420159108221357
Vignes, J.-L., Frappart, C., Di Costanzo, T., Rouchaud, J.-C., Mazerolles, L., Michel, D., 2008. Ultraporous monoliths of alumina prepared at room temperature by aluminium oxidation. J. Mater. Sci. 43, 1234–1240. doi:10.1007/s10853-0072260-z
TE D
Wang, Y., Herron, N., 1991. Nanometer-sized semiconductor clusters: materials synthesis, quantum size effects, and photophysical properties. J. Phys. Chem. 95, 525–532. doi:10.1021/j100155a009
AC C
EP
Zimmerer, G., 2007. SUPERLUMI: A unique setup for luminescence spectroscopy with synchrotron radiation. Radiat. Meas. 42, 859–864. doi:10.1016/j.radmeas.2007.02.050
ACCEPTED MANUSCRIPT Figure captions
RI PT
Fig. 1. TI emission spectra excited by 10keV electrons (solid lines) at 78 K and 130 eV photons (dashed lines) at 10K. a) Blue – Al2O3 UPA; b) Green – Al2O3 1600; c) Magenta – Al2O3 5050. CL spectra were recorded with two monochromators. The range above 6 eV with the VUV instrument and below with the UV-visible spectrometer. Fig. 2. TI excitation spectra for 7.6 eV luminescence at 10K. Black (*) – Al2O3 crystal; Green (+) – Al2O3 1600 and blue (o) – Al2O3 UPA.
AC C
EP
TE D
M AN U
SC
Fig. 3 Decay curves for alumina samples excited by pulsed (100 ns) 10 keV electrons at 78 K. blue (1) – UPA, green (2) – 1600, magenta (3) – 5050. Decays for emissions a) 2.6 eV, b) 3.0 eV, c) 3.8 eV and d) 4.5 eV.
ACCEPTED MANUSCRIPT
Table 1 Phase composition of studied Al2O3 samples and estimated crystallite size based on the XRD analysis Phase composition
Al2O3 1600 Al2O3 5050
α-phase α-phase (~40-50%) γ-phase (~40-50%) α -phase
Al2O3 UPA
Volume weighted crystallite size α-phase ~180nm α-phase ~80nm γ-phase <10nm α -phase ~200-300 nm
RI PT
Sample name
2.4 eV decay
3.8 eV decay
4.5 eV decay
Al2O3 1600
1.5 µs 25.1 µs
0.5 µs 22.9 µs
0.7 µs 28.8 µs
0.6 µs 7.8 µs 148.3 µs
Al2O3 5050
0.7 µs 47.6 µs
0.7 µs 14.2 µs
0.6 µs 19.4 µs
0.6 µs 7.9 µs >200 µs
0.6 µs 7.8 µs
0.4 µs 4.7 µs
0.5 µs 16.3 µs
0.8 µs 10.9 µs
34 ms
2 ns
TE D
3.0 eV decay
EP
Sample name
M AN U
SC
Table 2 Fitted decay parameters for Al2O3 1600, Al2O3 5050 and Al2O3 UPA UVvisible luminescence bands compared with literature (Surdo et al., 2005) values at room temperature (RT).
Al2O3 UPA
AC C
Decay times from literature at RT
56 µs
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIPT Highlights: Combustion synthesized alumina is compared with phase pure one using luminescence; Excitonic excitation peak at 9.1 eV is shifted towards higher energies in nanopowders;
AC C
EP
TE D
M AN U
SC
RI PT
Time resolved luminescence was used to identify nature of emission centres