- Email: [email protected]
Photoluminescence and surface photovoltage spectroscopy studies of hydroxyapatite nano-Bio-ceramics
ARTICLE IN PRESS
Journal of Luminescence 122–123 (2007) 936–938 www.elsevier.com/locate/jlumin
Photoluminescence and surface photovoltage spectroscopy studies of hydroxyapatite nano-Bio-ceramics G Rosenmana, D. Aronova, L. Osterb,, J. Haddadb, G. Mezinskisc, I. Pavlovskac, M. Chaikinad, A. Karlovd a School of Electrical Engineering-Physical Electronics, Tel Aviv University, 69978, Israel Sami Shamoon Academic College of Engineering, Bialik/Basel Sts. Beer-Sheva 84100, Israel c Institute of Silicate Materials of the Riga Technical University, Riga, Latvia d Center for Orthopedic and Medical Material Sciences of the Siberian Branch of the Russian Academy of Medical Sciences, 634029 Tomsk, Russia b
Available online 20 March 2006
Abstract Photoluminescence (PL) and surface photovoltage spectroscopy applied to nanostructural bioceramics hydroxyapatite (HAp) allowed to study electron (hole) energy states spectra of HAP and distinguish bulk and surface localized levels. Studied PL excitation spectra allowed obtaining an exact value of the energy band gap in HAP: Eg ¼ 3.95 eV.This result is consistent with Eg value determined by the contact potential difference (DCPD) curves treatment method as Eg ¼ 3.94 eV. Comparison between DCPD and PL spectra indicates that the energy spectra of electron – hole levels studied by two different experimental spectroscopy techniques are very similar. This comparison enables to conclude that all HAp samples have identical electron – hole states structures consisting of five bulk states and one surface state. It is assumed that the deep electron (hole) charged states may be responsible for high bioactivity of the HAp nanoceramics. r 2006 Elsevier B.V. All rights reserved. Keywords: Hydroxyapatite; Photoluminescence; Photovoltage spectroscopy.
1. Introduction Recent advances in biomaterial research have discovered that electrically polarized hydroxyapatite (HAp) ceramics produces significant biological response [1,2]. It has been demonstrated that the enhanced bone formation is observed at the negatively polarized HAp ceramic surface [3]. Important biological effects are ascribed to highdensity stored charges reaching hundreds of mC/cm2 [2]. The proposed method of polarization is based on bulk electrical polarization of the ceramics by the application of external electric field. According to these studies, the tailored electric charge is ascribed to ionic polarization and partly related to migration of protons in the columnar (OH) channels of HAp [1,2]. However, long-term charge storage in dielectric solids may also occur due to electron (hole) trapping at local electronic states in the energy gap and formation of electric charge [4]. Studies of the trap Corresponding author. Tel.: +972 8 6475759, Fax: +972 8 6475758.
E-mail address: [email protected] (L. Oster). 0022-2313/$ - see front matter r 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.jlumin.2006.01.331
spectroscopy in HAp may shed light on microscopic mechanism of this bioceramics interaction with the bone tissue. In this work we present the results of electron states spectroscopy of nanostructural HAp bioceramics. Combination of photoluminescence and surface photovoltage spectroscopy methods allowed obtaining detailed electron (hole) trap spectrum of HAp and distinguishing bulk and surface states.
2. Experimental details HAp nanopowder, synthesized by mechanochemical reactions was used as a raw material for preparation of ceramic platelets. It was fabricated in a planetary mill by mechanochemical synthesis in activator in 1 min by steel drums with steel balls. The size of powder particles was about 20–30 nm. Two type of HAp nano-powder ‘‘A’’ and ‘‘P’’ were used for ceramic samples fabrication. HAp powder ‘‘P’’ was
ARTICLE IN PRESS G. Rosenman et al. / Journal of Luminescence 122–123 (2007) 936–938
3. Results and discussion The measurements of absorption spectra from HAp were very problematic since the samples are non-transparent and represent strongly dispersed media. However significant dispersion increasing (relative to the air) observed at the level of photon energy higher than 3.7 eV might be related to the region of the beginning of fundamental absorption edge. The basic optical data were measured by means of excitation spectrum of PL. First of all, we evaluated the spectral emission region of PL. Excitation of HAp ceramics by photon energy of 3.44 eV led to a very wide, continuous optical emission PL spectrum with a wide plateau in the range 540–680 nm. The excitation spectra (Fig. 1) were measured in the region (2.5–6.2) eV using emission band 64075 nm determined from the plateau of the emission spectrum. It should be noted that both sorts of samples ‘‘A’’ and ‘‘P’’ show very similar spectra but the PL intensities are strongly different. The continuous increase of PL intensity was observed starting from the exciting photon energy of 3.8 eV and PL intensity reaches the maximum value at 5.6 eV. This excitation spectrum behavior in this spectrum region is a firm evidence of the
6 A PL Intensity [a.u.]
annealed at 900 1C in 2 h and then dispersed in alcohol in 2 min whilst the powder ‘‘A’’ was not subjected to any thermal treatment. Such a preliminary high-temperature treatment of the powder ‘‘P’’ should lead to a strong dehydration of HAp which was confirmed by subsequent XPS analysis of the HAp ceramics samples. Optical absorption spectra were measured with a Genesis-5 spectrophotometer (Milton Roy, USA) equipped with PC-IBM. Photoluminescence (PL) excitation and emission spectra were measured with a FP-6200 (Jasco, Japan) spectrofluorometer supported by a Pentium-4 computer. The PL excitation bands were resolved into a minimum number of individual Gaussian components using the ‘‘Peak-Fit’’ deconvolution program. We defined the minimum number of the components by second derivative method: in many cases, the hidden bands which evidence no local maxima in the original spectrum do appear as local minima in a smoothed second derivative. SPS measurements were performed in air using commercial Kelvin probe arrangement (Besocke Delta Phi, Ju¨lich, Germany) with sensitivity of 1 meV. The vibrating metallic probe consisted of a 2.5 mm diameter semitransparent gold grid mounted at a piezoelectric actuator. The distance between the probe and the powder was kept to a fraction of a millimeter. The piezoelectric crystal was moved by an external oscillator at a frequency of 170 Hz. The sample was illuminated by a 250 W tungsten-halogen lamp using a grating monochromator (Jarrell Ash). A value of the contact potential difference (CPD) and its changes with photon energy were measured using lock-in amplifier (LIA) and were processed by a Pentium-3 computer.
937
P
4
2
0 2.5
3.0
3.5 4.0 Photon Energy [eV]
4.5
5.0
Fig. 1. PL excitation spectrum from HAp (270 – 500 nm). Emission wavelength 640 nm.
Table 1 Energy structure of electron (hole) states in HAp obtained from Photoluminescence Sample type
E1 (eV)
E2 (eV)
E3 (eV)
E4 (eV)
E6 (eV)
Eg (eV)
A P
2.63 2.61
2.84 2.91
3.03 3.02
3.17 3.17
3.41 3.34
3.95 3.97
fundamental absorption (interband transitions). The fitting of the edge of fundamental optical absorption allowed to evaluate the width of the forbidden band Eg in HAp ceramics for both ‘‘A’’ and ‘‘P’’ samples between Eg ¼ 3.8–4.0 eV. The measured spectra (Fig. 1) represents a wide nonsymmetric nonmonotonic optical band indicating that number of localized energy levels of electron/hole origin are responsible for the spectrum. They were resolved into individual Gaussian components. The energies of these components are shown in Table 1. The deconvolution treatment of the experimental data allowed to obtain an exact value of the energy band gap Eg ¼ 3.95 eV. Several individual energy states were also found. They are located in the energy gap in the range (2.6–3.9) eV. Fig. 2 shows a light-induced variation of contact potential difference, DCPD. The DCPD spectra of both investigated Hap ceramics samples (‘‘A’’ and ‘‘P’’) are identical. Since light illumination typically tends to decrease the surface band-bending, this would result in a positive DCPD in P-type samples and a negative DCPD in N-type samples. The obtained DCPD spectra demonstrate a positive sign of the DCPD ‘‘knee’’ which allows relating both HAp samples to P-type [5]. Despite of very similar structure of DCPD spectra a pronounced difference is found for absolute values of DCPD which is by 10 times higher for the ‘‘P’’ sample (Fig. 2). Another basic application of SPS is measurements of a sample band gap Eg and energy position of localized states. Strong monotonic variation of DCPD (Fig. 2) occurs due to increase of light absorption coefficient near the band
ARTICLE IN PRESS G. Rosenman et al. / Journal of Luminescence 122–123 (2007) 936–938
938
12
1 P
0.5
6
0 1.5
2.5
∆CPD [mV]
∆CPD [mV]
A
0 5.5
3.5 4.5 Photon Energy [eV]
Fig. 2. Surface photovoltage spectra of HAp ceramic samples without (‘‘A’’) and after (‘‘P’’) thermal treatment.
Table 2 Energy structure of electrons (hole) states in HAp measured by Surface Photovoltage Spectroscopy method Sample E1 (eV) E2 (eV) E3 (eV) E4 (eV) E5 (eV) E6 (eV) Eg (eV) type A P
2.64 2.61
2.82 2.87
2.99 3.00
3.17 3.17
3.30 3.24
3.43 3.35
3.94 3.86
ques are very similar (Tables 1, 2). However the electron state E5 ¼ (3.24–3.30) eV found by SPS method is not observed in PL spectrum. Contact potential difference generated between Kelvin probe and illuminated sample surface affected both by surface and near surface-bulk states. However PL intensity totally depends on number of states participating in recombination process resulting in photon emission. PL is mainly contributed by bulk states. It allows relating the electron state E5 to the surface state which does not contribute sufficiently to PL. Comparison of the PL and SPS experimental spectroscopy data enables to conclude that both ‘‘P’’ and ‘‘A’’ HAp samples have identical electron-hole states structures in five bulk states E1, E2, E3, E4 and E6 and one surface state E5. However, the observed one order of magnitude variation of DCPD between two ‘‘A’’ and ‘‘P’’ HAp samples (Fig. 2) indicates a strong difference in density of elementary defects responsible for the found electron–hole energy state structure. Specifically, preliminary performed thermal treatment of ‘‘P’’ powder led to a higher density of defect concentration compared to ‘‘A’’. They might be (OH) vacancies where high concentration was found in ‘‘P’’ samples. 4. Conclusions
gap energy edge which is observed around 3.6–4.0 eV. According to the developed technique of DCPD curves treatment [5] the sharpest change in the slope of DCPD (Fig. 2) is related to the region of the fundamental light absorption. As a result the value of the energy gap in HAp was determined as Eg ¼ 3.94 eV (Table 2) which is consistent with Eg value obtained from the PL data. Identical approach [5] was applied for estimation of energy positions of bulk and surface electron (hole) states. Excitation of electrons from bulk or surface states to the conduction band contributes to a positive change in the surface charge and hence a negative DCPD is expected. Conversely, excitation of holes to the valence band makes the surface charge more negative and positive DCPD should be observed. The combination of the DCPD threshold energy and the slope sign allows finding the absolute energy positions of bulk and surface states. They are determined as tangents intersection of a slope change points at DCPD curves [5]. Table 2 concentrates estimated bulk and surface states energies for both HAp samples that were obtained from the DCPD data. Energy of six localized states are in the range (2.6–3.3) eV. Comparison between the DCPD and the PL spectra indicates that the energy spectra of electron–hole levels studied by two different experimental spectroscopy techni-
Studies of optical absorption, photoluminescence, surface photovoltage phenomena of nanostructural hydroxyapatite ceramics combining with some high-resolution characterization methods allowed to find basic electronic parameters of HAp such as energy band and fine electron–hole energy structure. Deep energy-localized states maybe responsible for high bioactivity of HAp ceramics. Acknowledgements The authors appreciate support of European Commission PROJECT NMP3-CT-504937 ‘‘PERCERAMICS’’. References [1] K. Yamashita, K. Kitagaki, T. Umegaki, J. Am. Ceram. Soc. 78 (1995) 1191. [2] S. Nakamura, H. Takeda, K. Yamashita, J. Appl. Phys. 89 (2001) 5386. [3] T. Kobayashi, M. Ohgaki, S. Nakamura, K. Yamashita, in: H. Ohgushi, G.W. Hastings, T. Yoshikawa (Eds.), Bioceramics, 12, World Scientific, Singapore, 1999, pp. 291–294. [4] G. Sessler (Ed.), Electrets, Springer, Berlin, 1980. [5] L. Kronik, Y. Shapira, Surf. Interface Anal. 31 (2001) 954.
Journal of Luminescence 122–123 (2007) 936–938 www.elsevier.com/locate/jlumin
Photoluminescence and surface photovoltage spectroscopy studies of hydroxyapatite nano-Bio-ceramics G Rosenmana, D. Aronova, L. Osterb,, J. Haddadb, G. Mezinskisc, I. Pavlovskac, M. Chaikinad, A. Karlovd a School of Electrical Engineering-Physical Electronics, Tel Aviv University, 69978, Israel Sami Shamoon Academic College of Engineering, Bialik/Basel Sts. Beer-Sheva 84100, Israel c Institute of Silicate Materials of the Riga Technical University, Riga, Latvia d Center for Orthopedic and Medical Material Sciences of the Siberian Branch of the Russian Academy of Medical Sciences, 634029 Tomsk, Russia b
Available online 20 March 2006
Abstract Photoluminescence (PL) and surface photovoltage spectroscopy applied to nanostructural bioceramics hydroxyapatite (HAp) allowed to study electron (hole) energy states spectra of HAP and distinguish bulk and surface localized levels. Studied PL excitation spectra allowed obtaining an exact value of the energy band gap in HAP: Eg ¼ 3.95 eV.This result is consistent with Eg value determined by the contact potential difference (DCPD) curves treatment method as Eg ¼ 3.94 eV. Comparison between DCPD and PL spectra indicates that the energy spectra of electron – hole levels studied by two different experimental spectroscopy techniques are very similar. This comparison enables to conclude that all HAp samples have identical electron – hole states structures consisting of five bulk states and one surface state. It is assumed that the deep electron (hole) charged states may be responsible for high bioactivity of the HAp nanoceramics. r 2006 Elsevier B.V. All rights reserved. Keywords: Hydroxyapatite; Photoluminescence; Photovoltage spectroscopy.
1. Introduction Recent advances in biomaterial research have discovered that electrically polarized hydroxyapatite (HAp) ceramics produces significant biological response [1,2]. It has been demonstrated that the enhanced bone formation is observed at the negatively polarized HAp ceramic surface [3]. Important biological effects are ascribed to highdensity stored charges reaching hundreds of mC/cm2 [2]. The proposed method of polarization is based on bulk electrical polarization of the ceramics by the application of external electric field. According to these studies, the tailored electric charge is ascribed to ionic polarization and partly related to migration of protons in the columnar (OH) channels of HAp [1,2]. However, long-term charge storage in dielectric solids may also occur due to electron (hole) trapping at local electronic states in the energy gap and formation of electric charge [4]. Studies of the trap Corresponding author. Tel.: +972 8 6475759, Fax: +972 8 6475758.
E-mail address: [email protected] (L. Oster). 0022-2313/$ - see front matter r 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.jlumin.2006.01.331
spectroscopy in HAp may shed light on microscopic mechanism of this bioceramics interaction with the bone tissue. In this work we present the results of electron states spectroscopy of nanostructural HAp bioceramics. Combination of photoluminescence and surface photovoltage spectroscopy methods allowed obtaining detailed electron (hole) trap spectrum of HAp and distinguishing bulk and surface states.
2. Experimental details HAp nanopowder, synthesized by mechanochemical reactions was used as a raw material for preparation of ceramic platelets. It was fabricated in a planetary mill by mechanochemical synthesis in activator in 1 min by steel drums with steel balls. The size of powder particles was about 20–30 nm. Two type of HAp nano-powder ‘‘A’’ and ‘‘P’’ were used for ceramic samples fabrication. HAp powder ‘‘P’’ was
ARTICLE IN PRESS G. Rosenman et al. / Journal of Luminescence 122–123 (2007) 936–938
3. Results and discussion The measurements of absorption spectra from HAp were very problematic since the samples are non-transparent and represent strongly dispersed media. However significant dispersion increasing (relative to the air) observed at the level of photon energy higher than 3.7 eV might be related to the region of the beginning of fundamental absorption edge. The basic optical data were measured by means of excitation spectrum of PL. First of all, we evaluated the spectral emission region of PL. Excitation of HAp ceramics by photon energy of 3.44 eV led to a very wide, continuous optical emission PL spectrum with a wide plateau in the range 540–680 nm. The excitation spectra (Fig. 1) were measured in the region (2.5–6.2) eV using emission band 64075 nm determined from the plateau of the emission spectrum. It should be noted that both sorts of samples ‘‘A’’ and ‘‘P’’ show very similar spectra but the PL intensities are strongly different. The continuous increase of PL intensity was observed starting from the exciting photon energy of 3.8 eV and PL intensity reaches the maximum value at 5.6 eV. This excitation spectrum behavior in this spectrum region is a firm evidence of the
6 A PL Intensity [a.u.]
annealed at 900 1C in 2 h and then dispersed in alcohol in 2 min whilst the powder ‘‘A’’ was not subjected to any thermal treatment. Such a preliminary high-temperature treatment of the powder ‘‘P’’ should lead to a strong dehydration of HAp which was confirmed by subsequent XPS analysis of the HAp ceramics samples. Optical absorption spectra were measured with a Genesis-5 spectrophotometer (Milton Roy, USA) equipped with PC-IBM. Photoluminescence (PL) excitation and emission spectra were measured with a FP-6200 (Jasco, Japan) spectrofluorometer supported by a Pentium-4 computer. The PL excitation bands were resolved into a minimum number of individual Gaussian components using the ‘‘Peak-Fit’’ deconvolution program. We defined the minimum number of the components by second derivative method: in many cases, the hidden bands which evidence no local maxima in the original spectrum do appear as local minima in a smoothed second derivative. SPS measurements were performed in air using commercial Kelvin probe arrangement (Besocke Delta Phi, Ju¨lich, Germany) with sensitivity of 1 meV. The vibrating metallic probe consisted of a 2.5 mm diameter semitransparent gold grid mounted at a piezoelectric actuator. The distance between the probe and the powder was kept to a fraction of a millimeter. The piezoelectric crystal was moved by an external oscillator at a frequency of 170 Hz. The sample was illuminated by a 250 W tungsten-halogen lamp using a grating monochromator (Jarrell Ash). A value of the contact potential difference (CPD) and its changes with photon energy were measured using lock-in amplifier (LIA) and were processed by a Pentium-3 computer.
937
P
4
2
0 2.5
3.0
3.5 4.0 Photon Energy [eV]
4.5
5.0
Fig. 1. PL excitation spectrum from HAp (270 – 500 nm). Emission wavelength 640 nm.
Table 1 Energy structure of electron (hole) states in HAp obtained from Photoluminescence Sample type
E1 (eV)
E2 (eV)
E3 (eV)
E4 (eV)
E6 (eV)
Eg (eV)
A P
2.63 2.61
2.84 2.91
3.03 3.02
3.17 3.17
3.41 3.34
3.95 3.97
fundamental absorption (interband transitions). The fitting of the edge of fundamental optical absorption allowed to evaluate the width of the forbidden band Eg in HAp ceramics for both ‘‘A’’ and ‘‘P’’ samples between Eg ¼ 3.8–4.0 eV. The measured spectra (Fig. 1) represents a wide nonsymmetric nonmonotonic optical band indicating that number of localized energy levels of electron/hole origin are responsible for the spectrum. They were resolved into individual Gaussian components. The energies of these components are shown in Table 1. The deconvolution treatment of the experimental data allowed to obtain an exact value of the energy band gap Eg ¼ 3.95 eV. Several individual energy states were also found. They are located in the energy gap in the range (2.6–3.9) eV. Fig. 2 shows a light-induced variation of contact potential difference, DCPD. The DCPD spectra of both investigated Hap ceramics samples (‘‘A’’ and ‘‘P’’) are identical. Since light illumination typically tends to decrease the surface band-bending, this would result in a positive DCPD in P-type samples and a negative DCPD in N-type samples. The obtained DCPD spectra demonstrate a positive sign of the DCPD ‘‘knee’’ which allows relating both HAp samples to P-type [5]. Despite of very similar structure of DCPD spectra a pronounced difference is found for absolute values of DCPD which is by 10 times higher for the ‘‘P’’ sample (Fig. 2). Another basic application of SPS is measurements of a sample band gap Eg and energy position of localized states. Strong monotonic variation of DCPD (Fig. 2) occurs due to increase of light absorption coefficient near the band
ARTICLE IN PRESS G. Rosenman et al. / Journal of Luminescence 122–123 (2007) 936–938
938
12
1 P
0.5
6
0 1.5
2.5
∆CPD [mV]
∆CPD [mV]
A
0 5.5
3.5 4.5 Photon Energy [eV]
Fig. 2. Surface photovoltage spectra of HAp ceramic samples without (‘‘A’’) and after (‘‘P’’) thermal treatment.
Table 2 Energy structure of electrons (hole) states in HAp measured by Surface Photovoltage Spectroscopy method Sample E1 (eV) E2 (eV) E3 (eV) E4 (eV) E5 (eV) E6 (eV) Eg (eV) type A P
2.64 2.61
2.82 2.87
2.99 3.00
3.17 3.17
3.30 3.24
3.43 3.35
3.94 3.86
ques are very similar (Tables 1, 2). However the electron state E5 ¼ (3.24–3.30) eV found by SPS method is not observed in PL spectrum. Contact potential difference generated between Kelvin probe and illuminated sample surface affected both by surface and near surface-bulk states. However PL intensity totally depends on number of states participating in recombination process resulting in photon emission. PL is mainly contributed by bulk states. It allows relating the electron state E5 to the surface state which does not contribute sufficiently to PL. Comparison of the PL and SPS experimental spectroscopy data enables to conclude that both ‘‘P’’ and ‘‘A’’ HAp samples have identical electron-hole states structures in five bulk states E1, E2, E3, E4 and E6 and one surface state E5. However, the observed one order of magnitude variation of DCPD between two ‘‘A’’ and ‘‘P’’ HAp samples (Fig. 2) indicates a strong difference in density of elementary defects responsible for the found electron–hole energy state structure. Specifically, preliminary performed thermal treatment of ‘‘P’’ powder led to a higher density of defect concentration compared to ‘‘A’’. They might be (OH) vacancies where high concentration was found in ‘‘P’’ samples. 4. Conclusions
gap energy edge which is observed around 3.6–4.0 eV. According to the developed technique of DCPD curves treatment [5] the sharpest change in the slope of DCPD (Fig. 2) is related to the region of the fundamental light absorption. As a result the value of the energy gap in HAp was determined as Eg ¼ 3.94 eV (Table 2) which is consistent with Eg value obtained from the PL data. Identical approach [5] was applied for estimation of energy positions of bulk and surface electron (hole) states. Excitation of electrons from bulk or surface states to the conduction band contributes to a positive change in the surface charge and hence a negative DCPD is expected. Conversely, excitation of holes to the valence band makes the surface charge more negative and positive DCPD should be observed. The combination of the DCPD threshold energy and the slope sign allows finding the absolute energy positions of bulk and surface states. They are determined as tangents intersection of a slope change points at DCPD curves [5]. Table 2 concentrates estimated bulk and surface states energies for both HAp samples that were obtained from the DCPD data. Energy of six localized states are in the range (2.6–3.3) eV. Comparison between the DCPD and the PL spectra indicates that the energy spectra of electron–hole levels studied by two different experimental spectroscopy techni-
Studies of optical absorption, photoluminescence, surface photovoltage phenomena of nanostructural hydroxyapatite ceramics combining with some high-resolution characterization methods allowed to find basic electronic parameters of HAp such as energy band and fine electron–hole energy structure. Deep energy-localized states maybe responsible for high bioactivity of HAp ceramics. Acknowledgements The authors appreciate support of European Commission PROJECT NMP3-CT-504937 ‘‘PERCERAMICS’’. References [1] K. Yamashita, K. Kitagaki, T. Umegaki, J. Am. Ceram. Soc. 78 (1995) 1191. [2] S. Nakamura, H. Takeda, K. Yamashita, J. Appl. Phys. 89 (2001) 5386. [3] T. Kobayashi, M. Ohgaki, S. Nakamura, K. Yamashita, in: H. Ohgushi, G.W. Hastings, T. Yoshikawa (Eds.), Bioceramics, 12, World Scientific, Singapore, 1999, pp. 291–294. [4] G. Sessler (Ed.), Electrets, Springer, Berlin, 1980. [5] L. Kronik, Y. Shapira, Surf. Interface Anal. 31 (2001) 954.