- Email: [email protected]
Effects induced by UV laser radiation on the blue luminescence of silica nanoparticles
Journal of Luminescence 138 (2013) 39–43
Contents lists available at SciVerse ScienceDirect
Journal of Luminescence journal homepage: www.elsevier.com/locate/jlumin
Effects induced by UV laser radiation on the blue luminescence of silica nanoparticles L. Spallino n, L. Vaccaro, S. Agnello, M. Cannas Dipartimento di Fisica e Chimica, Universita di Palermo, Via Archirafi 36, I-90123 Palermo, Italy
a r t i c l e i n f o
abstract
Article history: Received 25 September 2012 Received in revised form 15 January 2013 Accepted 24 January 2013 Available online 4 February 2013
The effects induced on the blue luminescence centered around 2.8 eV, characteristic of silica nanoparticles, were investigated by monitoring its intensity during and after exposure to the third and the fourth harmonic of a Nd:YAG pulsed laser. The luminescence trend is found to be dependent on the UV photon energy: 3.50 eV photons induce a partial bleaching followed by a recovery in the postirradiation stage; 4.66 eV photons cause a total bleaching permanent after the irradiation. These results are interpreted as the conversion of luminescent defects towards stable and metastable configurations. & 2013 Elsevier B.V. All rights reserved.
Keywords: Silica nanoparticles Surface defects Time-resolved photoluminescence UV laser irradiation Conversion processes
1. Introduction The high photon emissivity in a wide range, from IR to visible, is one of the most relevant phenomena emerging from the reduction of silica down to nanoscale [1–4]; hence it is promising for the development of optical nanotechnologies (down converters, probes, displays) [5–7]. It is well accepted that the origin of this luminescence is related to the high specific surface of silica nanoparticles ( 100 m2 =g) [1,2,5,6,8], however, the role of specific optically active centers remains poorly understood. An obstacle to the interpretation of data is due to the huge variety of defects arising from the reactions, at the surface sites, of molecular and atomic species of surrounding atmosphere [9–13]. Nevertheless, a bright photoluminescence (PL) centered around 2.8 eV (blue band), stable in the ambient atmosphere, is usually observed in silica nanoparticles. It can be excited in a wide range, extending from visible to UV, and it is characterized by a fast decay time, in the nanoseconds timescale [8,9,14]. The monotonic increase of its intensity on increasing the specific surface area indicates that it originates from defects localized in the nanoparticles surface shell [15]. The origin of this PL band remains an open question, actually only a model has been proposed: a defect pair consisting of a dioxasilirane, ¼ SiðO2 Þ, and a silylene, ¼ Si [8]. Recently, it has been demonstrated that this band decreases and red-shifts by addition of H2 O2 or mineral acids (HCl
n
Corresponding author. Tel.: þ39 091238 91788. E-mail address: [email protected] (L. Spallino).
0022-2313/$ - see front matter & 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.jlumin.2013.01.027
or H2 SO4 ) [16,17]; whatever the model, these results suggest a possible quenching of the luminescent centers. In order to gain further information on the structure of the blue luminescent defects, in the present work we have monitored the PL both during and after UV laser irradiation. This approach is useful to investigate the stability of these centers under UV exposure and the induced chemical reactions with atomic or molecular species of the ambient atmosphere.
2. Experimental methods We used samples obtained by commercial fumed silica nanoparticles (Aerosil) supplied in powder form by Evonik-Industries [18]. These materials differ for the specific surface (S) and the average particle diameter (d). In particular, we investigated two typologies of materials: one, named AE300, with S¼300730 m2/g and d¼771 nm; the other, named AE150, has S¼150715 m2/g and d¼1472 nm. In a previous study performed by atomic force microscopy technique [19], it is shown that the diameters of these fumed silica nanoparticles have a statistical distribution with average values in agreement with that declared by the producer. Silica nanoparticles have been pressed at 300 MPa into tablet of 1 mm thickness. UV-irradiations were performed at room temperature using the third (3.50 eV) and the fourth harmonic (4.66 eV) of a pulsed Quanta System SYL 201 Nd:YAG laser (pulse width 5 ns, spot size 0:4 cm2 ). The 3.50 eV exposure was done with pulse repetition rates of 10 Hz, 2 Hz, and 1 Hz and energy density of
40
L. Spallino et al. / Journal of Luminescence 138 (2013) 39–43
F1 ¼ 7:5 mJ=cm2 , F2 ¼ 37:5 mJ=cm2 , and F3 ¼ 75 mJ=cm2 , respectively, so as to deposit on the sample the same energy density in the same time interval. The 4.66 eV exposure was carried out at 2 Hz repetition rate and 37.5 mJ/cm2 pulse energy density. The mean power of the laser beam was monitored by a pyroelectric detector and kept constant during each irradiation. Visible luminescent centers were investigated by ex situ timeresolved PL technique. The PL spectra were performed at room temperature in a standard front-scattering geometry; pulsed light at excitation energy Eexc ¼ 4:96 eV (pulse width 5 ns, repetition rate 10 Hz) was provided by a VIBRANT OPOTEK optical parametric oscillator laser system, pumped by the third harmonic (3.50 eV) of a Nd:YAG laser. The pulse energy was monitored with a pyroelectric detector and the fluence/pulse was maintained at F ¼ 0:8 mJ=cm2 , low enough to avoid the generation of any defect. The emitted light was spectrally resolved by a monochromator (SpectraPro 2300i, PI/Acton) equipped with a 150 grooves/ mm grating blazed at 300 nm and acquired by an intensified charge coupled device (CCD) camera driven by a delay generator (PI-MAX Princeton Instruments) that sets the time acquisition of the emitted light (gate width DT ¼ 200 ns and time delay TD ¼ 10 ns after the laser pulse). All the emission spectra were detected with a bandwidth of 20 nm and were corrected for the monochromator dispersion.
3. Results As a representative view, the spectroscopic characteristics of one of the as grown samples (AE150) are reported in Fig. 1. The blue emission is centered around 2.8 eV with full width at half maximum (FWHM) 0:7 eV, its excitation shows two peaks around 3.9 eV and 5.0 eV, its decay occurs in the nanoseconds timescale. These characteristics are in good agreement with previous works [8,9,14]. We also observe a contribution at lower intensity, around 3.5 eV. UV luminescence of high surface silica has been attributed to hydrocarbons contamination [9] or surface silanols (Si–OH) [20], however, in this work we will not deal with it. Prior to study the effects induced by high power laser radiation, we have exposed for 1 h the samples to UV photons coming from the laser used for the ex situ PL measurements (Eexc ¼ 4:96 eV, fluence=pulse 0:8 mJ=cm2 ), up to store a total
Fig. 1. PL spectrum of the as grown AE150 sample excited at Eexc ¼ 4:96 eV and acquired with TD ¼ 10 ns and DT ¼ 200 ns. The insets show, at the top, the excitation spectrum acquired at emission energy Eem ¼ 2:74 eV; at the bottom, the decay time of the blue band peak obtained acquiring the PL spectra with DT ¼ 1 ns and TD ranging from 0 to 20 ns.
fluence 30 000 mJ=cm2 . We observed that this does not cause any variation of the blue band intensity. Then, we have followed the evolution of the spectral and lifetime features related to the blue band, both during and after irradiation with the third harmonic of the Nd:YAG laser (hn1 ¼ 3:50 eV). We have exposed our samples to an increasing number of pulses, from 1 to 20 000, with fluence/pulse F2 ¼ 37:5 mJ=cm2 so as to deposit a total fluence of 750 J/cm2. During the irradiation stage, we performed ex situ time resolved PL spectra under Eexc ¼ 4:96 eV. In Fig. 2a we report some representative curves observed in the AE150 sample; as indicated by the arrow the blue band decreases on increasing the number of deposited pulses. The post irradiation trend of the PL spectra, monitored for a time of about 5 h after the last irradiation, is depicted in Fig. 2b. The blue band intensity increases back to its value before irradiation. We note that its spectral features (peak position and FWHM) remain the same both during and after irradiation. In Fig. 3 we report the decay time recorded before, during and after the irradiation. All the curves are described by a stretched exponential, characteristic of luminescent defects embedded in a disordered matrix [12]; we estimate a common lifetime, measured as the time necessary to reduce the intensity by a factor 1/e, t ¼ 2:0 70:3 ns. We appreciate some differences at longer delay time: the decay rate is faster after a fluence of 150 J=cm2 and after the recovery. The same spectroscopic phenomenology is observed in the AE300 sample. To quantitatively characterize the bleaching and the recovery of the blue band, its PL intensity is reported in Fig. 4. During the irradiation stage, the PL kinetics is plotted as a function of the fluence in a semilogarithmic scale, while it is plotted versus time in the post irradiation stage, where t ¼0 corresponds to the end of the irradiation session. In the same graph we show the PL intensities measured in analogous experiments under exposure
Fig. 2. Evolution of PL spectra measured ex-situ in the sample AE150 during (a) and after (b) irradiation with hn1 ¼ 3:50 eV and fluence/pulse ¼37.5 mJ/cm2. All spectra are excited at Eexc ¼ 4:96 eV and acquired with TD ¼ 10 ns and DT ¼ 200 ns. The arrows indicate the trend of the blue band.
L. Spallino et al. / Journal of Luminescence 138 (2013) 39–43
to F1 ¼ 7:5 mJ=cm2 , F2 ¼ 37:5 mJ=cm2 and F3 ¼ 75 mJ=cm2 . The PL intensity decreases to a value that depends on Fi : it reduces by 50% for F1 ¼ 7:5 mJ=cm2 ; by 80% for F2 ¼ 37:5 mJ=cm2 and almost totally for F3 ¼ 75 mJ=cm2 . The recovery also depends on the fluence/pulse and keeps memory of irradiation stage: in about 5 h, the PL intensity gets back to its as-grown value for F1 ¼ 7:5 mJ=cm2 , while it recovers 70% and 20% of its original value for F2 ¼ 37:5 mJ=cm2 and F3 ¼ 75 mJ=cm2 , respectively. In Fig. 5 the comparison between the effects induced on blue PL intensity by radiation (fluence/pulse fixed at 37.5 mJ/cm2) coming from the third (hn1 ¼ 3:50 eV) and the fourth
41
(hn2 ¼ 4:66 eV) harmonic of the Nd:YAG laser is reported. The comparison suggests that, higher the radiation energy, higher the bleaching effect on the blue band during irradiation is. Indeed we observe that, at a fluence 200 mJ=cm2 , the radiation energy hn1 ¼ 3:50 eV reduces the PL intensity by 75%, while hn2 ¼ 4:66 eV radiation totally removes the PL band. Moreover, the photon laser energy also influences the post-irradiation: after hn1 ¼ 3:50 eV exposure the PL intensity recovers 75% of its original value within 5 h, whereas the bleaching induced by the fourth harmonic of the Nd:YAG laser (hn2 ¼ 4:66 eV) is long lasting.
4. Discussion
Fig. 3. Decay curves of the blue band recorded in the AE150 sample before irradiation (filled circles), during irradiation at hn1 ¼ 3:50 eV after depositing a total fluence of 15 J=cm2 (triangles) and 150 J=cm2 (squares) and 5 h after the end of irradiation (crossed circles). All the curves are obtained acquiring the PL spectra under Eexc ¼ 4:96 eV with DT ¼ 1 ns and TD ranging from 0 to 15 ns. The curves are arbitrarily scaled for viewing purposes.
The luminescence properties of as received silica nanoparticles are mainly characterized by a blue luminescence (peak emission around 2.8 eV), that is excited in the UV and exhibits a fast decay time ( 2 ns) in accordance with a singlet–singlet allowed transition. These spectroscopic features do not allow to address the defect structure; from this viewpoint, the attribution of the blue band to the pair of a dioxasilirane, ¼ SiðO2 Þ, and a silylene ¼ Si , [8] remains an open question. For this purpose, the comparison with the visible luminescence properties associated with the dioxasilirane and silylene could be useful. Table 1 reports the main spectroscopic features of these luminescent defects. It is evident the disagreement among the decay properties: the silylene and dioxasilirane have a much longer lifetime, characteristic of a forbidden transition. This finding would lead to hypothesize that the formation of the defects pair introduces an energy levels configuration with a very fast (ns) transition rate from the excited state. The structural properties of the blue luminescent defect can be discussed on the basis of its response to UV radiation. Our results point out that UV laser photons govern two conversion processes
Fig. 4. PL intensity detected in the AE150 sample at Eem ¼ 2:60 eV as a function of hn1 ¼ 3:50 eV radiation fluence (left) and time (right). The origin of the time scale corresponds to the end of the irradiation session. The three curves refer to: F1 ¼ 7:5 mJ=cm2 (circles), F2 ¼ 37:5 mJ=cm2 (squares), and F3 ¼ 75 mJ=cm2 (crossed squares). The intensity is normalized to that measured in the as-grown sample; its uncertainty is 10%.
42
L. Spallino et al. / Journal of Luminescence 138 (2013) 39–43
Fig. 5. PL intensity detected in the AE150 sample at Eem ¼ 2:60 eV as a function of radiation fluence (left) and of the time elapsed from irradiation (right); the curves refer to hn1 ¼ 3:50 eV (squares) and hn2 ¼ 4:66 eV (circles). The intensity is normalized to that measured in the as-grown sample; its uncertainty is 10%.
Table 1 Main spectroscopic features of visible luminescent defects of silica nanoparticles: blue luminescent defect studied in this work, silylene and dioxasilirane. Defect
Emission peak (eV) Excitation peak (eV) Lifetime
Blue luminescent defect 2.8 2.75 Silylene ¼ Si [23] dioxasilirane ¼ SiðO2 Þ [12] 2.4
3.9 and 5.0 5.0 3.2 and 5.2
2 ns 18 ms 7:5 ms
of the blue luminescent defects at the surface of silica nanoparticles. These consist of a bleaching and a back-conversion observed during and after the UV exposure. It is also evidenced that the luminescence lifetime is weakly influenced by laser radiation. Since the lifetime is governed by radiative and non-radiative decay rates, t ¼ ðkR þkNR Þ 1 is an exponent then we infer that in the excited state the interactions with the surrounding are poorly affected by laser radiation, the quantum yield Z ¼ kR t does not change dramatically and the luminescence bleaching and recovery are mainly due to a concentration variation of the defects. As concerns the bleaching process, it is not observed at fluence/pulse as low as 1 mJ=cm2 , whereas it is effective if the fluence/pulse increases of an order of magnitude. In particular, as evidenced in Fig. 4 (left panel), the bleaching rate increases on increasing the fluence/pulse, in agreement with a non-linear effect, likely based on a multi-photon absorption. The backconversion process is evidenced by the luminescence recovery in the post irradiation stage, nevertheless we argue that it is also effective in the laser interpulse time span. Under this hypothesis, the kinetics of the luminescent centers during the irradiation is governed by the competition between bleaching and back-conversion; we expect that such competition is influenced by the irradiation laser parameters, such as the repetition rate and the fluence/pulse. The use of the Nd:YAG third and fourth harmonic as irradiation beam is crucial for the fate of the luminescent defect (DPL):
(1) Under exposure to hn1 ¼ 3:50 eV photons, DPL converts towards a different configuration. Since the PL recovery is
observed at room temperature, we deduce that this configuration (CMS) is metastable thus allowing a back-conversion. (2) Under exposure to hn1 ¼ 4:66 eV photons, DPL also converts to a new configuration (CS), but in this case the PL recovery is not observed, thus indicating its stability, at least at room temperature. The two conversion pathways are accounted for by the following equation: kc
k0c
C MS ! DPL ,C S
ð1Þ
kbc
Both conversion rates, kc and k0c , depend non-linearly on the photon fluence. As concerns the back-conversion process, it is active from CMS at room temperature and we suppose its rate is of the form kbc ¼ k0 eDE=KT , the activation energy DE being comparable with the thermal energy at room temperature ( 25 meV). On the basis of the comparison with literature data dealing with post-irradiation processes [21], we can only tentatively suggest a k
bi-molecular reaction of the form: A þB-C, in which the two species A and B react to form a species C with a conversion rate k. In the general case of solids [22], these processes are governed by an activation energy due to the diffusion of the atomic or molecular species in the network and, for endothermic processes, the reaction itself requires an amount of activation energy. In our case, the samples are dominated by a high specific surface so that the defects are easily accessible from species of the environment atmosphere. This makes negligible any diffusion and the process could be only reaction limited. We note that these results provide interesting information about the stability of the blue luminescent defect, but it is not enough to unambiguously and ultimately establish its nature. On the basis of the structural model that ascribes the blue band to the defect pair ¼ SiðO2 Þ and ¼ Si [8], the PL bleaching could be interpreted as the effect of O2 detrapping, induced by UV laser photons, consequently the O2 trapping leads to the PL recovery. It is worth noting that previous experiments have pointed out that the reactions of O2 with ¼ Si requires an activation energy of 0:3 eV [11], much larger than ( 25 meV),
L. Spallino et al. / Journal of Luminescence 138 (2013) 39–43
roughly estimated from our data. However, we cannot rule out the hypothesis that two faced ¼ Si make more likely the O2 trapping. To clarify this issue, we aim to deepen our experiments in a thermochemical controlled atmosphere, so as to have clear and useful indications about its origin.
5. Conclusion High power UV radiation, provided by the third (3.50 eV) and the fourth (4.66 eV) harmonic of a Nd:YAG pulsed laser, causes the bleaching of blue luminescence characteristic of silica nanoparticles. Under the higher radiation energy the effect is permanent, otherwise the band follows a post-irradiation recovery kinetics, likely due to the reactions, at the surface sites, with atomic or molecular species of the atmosphere.
Acknowledgments The work was partially supported by FAE project, PO FESR Sicilia 2007/2013 4.1.1.1. We thank the group of the Laboratory of Amorphous Materials Physics (Palermo University) (http://www. fisica.unipa.it//amorphous) for their support and stimulating discussions and G. Napoli and G. Tricomi for their technical assistance. References [1] Y.D. Glinka, S.H. Lin, Y.T. Chen, Appl. Phys. Lett. 75 (1999) 778. [2] Y.D. Glinka, S.H. Lin, Y.T. Chen, Phys. Rev. B 62 (2000) 4373.
43
[3] S. Bonacchi, D. Genovese, R. Juris, M. Montalti, L. Prodi, E. Rampazzo, N. Zaccheroni, Angew. Chem. Int. Ed. 50 (2011) 4056. [4] C. Zhang, J. Lin, Chem. Soc. Rev. 41 (2012) 7938. [5] B. Hamilton, Semicond. Sci. Technol. 10 (1995) 1187. [6] A.G. Cullis, L.T. Canham, P.D.J. Calcott, J. Appl. Phys. 82 (1997) 909. [7] G.L. Davies, J.E. McCarthy, A. Rakovich, Y.K. Gun’ko, J. Mater. Chem. 22 (2012) 7358. [8] A. Anjiki, T. Uchino, J. Phys. Chem. C 116 (2012) 15747. [9] A. Aboshi, N. Kurumoto, T. Yamada, T. Uchino, J. Phys. Chem. C 111 (2007) 8483. [10] V.A. Radzig, Colloids Surf. A 74 (1993) 91. [11] V.A. Radzig, Chem. Phys. Rep. 14 (1995) 1206. [12] L. Vaccaro, A. Morana, V.A. Radzig, M. Cannas, J. Phys. Chem. C 115 (2011) 19476. [13] C.M. Carbonaro, P.C. Ricci, R. Corpino, M. Marceddu, A. Anedda, J. Non-Crys. Solids 357 (2011) 1904. [14] L. Vaccaro, G. Vaccaro, S. Agnello, G. Buscarino, M. Cannas, Solid State Commun. 150 (2010) 2278. [15] G. Vaccaro, S. Agnello, G. Buscarino, M. Cannas, L. Vaccaro, J. Non Cryst. Solids 351 (2011) 1941. [16] S. Banerjee, S. Honkote, A. Datta, J. Phys. Chem. C 115 (2011) 1576. [17] S. Banerjee, S. Maity, A. Datta, J. Phys. Chem. C 115 (2011) 22804. [18] Basic Characteristics of Aerosil, 4th ed., Degussa, Frankfurt (2001). [19] G. Buscarino, V. Ardizzone, G. Vaccaro, S. Agnello, F.M. Gelardi, J. Appl. Phys. 108 (2001) 074314. [20] C.M. Carbonaro, F. Clemente, R. Corpino, P.C. Ricci, A. Anedda, J. Phys. Chem. B 109 (2005) 14441. [21] F. Messina, M. Cannas, J. Non-Cryst. Solids 353 (2007) 522. [22] R.H. Doremus, Diffusion of Reactive Molecules in Solids and Melts, John Wiley and Sons, 2002. [23] V.N. Bagratashivli, S.I. Tsypina, V.A. Radtsig, A.O. Rybaltovskii, P.V. Chernov, S.S. Alimpiev, Y.O. Simanovskii, J. Non-Cryst. Solids 180 (1995) 221.
Contents lists available at SciVerse ScienceDirect
Journal of Luminescence journal homepage: www.elsevier.com/locate/jlumin
Effects induced by UV laser radiation on the blue luminescence of silica nanoparticles L. Spallino n, L. Vaccaro, S. Agnello, M. Cannas Dipartimento di Fisica e Chimica, Universita di Palermo, Via Archirafi 36, I-90123 Palermo, Italy
a r t i c l e i n f o
abstract
Article history: Received 25 September 2012 Received in revised form 15 January 2013 Accepted 24 January 2013 Available online 4 February 2013
The effects induced on the blue luminescence centered around 2.8 eV, characteristic of silica nanoparticles, were investigated by monitoring its intensity during and after exposure to the third and the fourth harmonic of a Nd:YAG pulsed laser. The luminescence trend is found to be dependent on the UV photon energy: 3.50 eV photons induce a partial bleaching followed by a recovery in the postirradiation stage; 4.66 eV photons cause a total bleaching permanent after the irradiation. These results are interpreted as the conversion of luminescent defects towards stable and metastable configurations. & 2013 Elsevier B.V. All rights reserved.
Keywords: Silica nanoparticles Surface defects Time-resolved photoluminescence UV laser irradiation Conversion processes
1. Introduction The high photon emissivity in a wide range, from IR to visible, is one of the most relevant phenomena emerging from the reduction of silica down to nanoscale [1–4]; hence it is promising for the development of optical nanotechnologies (down converters, probes, displays) [5–7]. It is well accepted that the origin of this luminescence is related to the high specific surface of silica nanoparticles ( 100 m2 =g) [1,2,5,6,8], however, the role of specific optically active centers remains poorly understood. An obstacle to the interpretation of data is due to the huge variety of defects arising from the reactions, at the surface sites, of molecular and atomic species of surrounding atmosphere [9–13]. Nevertheless, a bright photoluminescence (PL) centered around 2.8 eV (blue band), stable in the ambient atmosphere, is usually observed in silica nanoparticles. It can be excited in a wide range, extending from visible to UV, and it is characterized by a fast decay time, in the nanoseconds timescale [8,9,14]. The monotonic increase of its intensity on increasing the specific surface area indicates that it originates from defects localized in the nanoparticles surface shell [15]. The origin of this PL band remains an open question, actually only a model has been proposed: a defect pair consisting of a dioxasilirane, ¼ SiðO2 Þ, and a silylene, ¼ Si [8]. Recently, it has been demonstrated that this band decreases and red-shifts by addition of H2 O2 or mineral acids (HCl
n
Corresponding author. Tel.: þ39 091238 91788. E-mail address: [email protected] (L. Spallino).
0022-2313/$ - see front matter & 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.jlumin.2013.01.027
or H2 SO4 ) [16,17]; whatever the model, these results suggest a possible quenching of the luminescent centers. In order to gain further information on the structure of the blue luminescent defects, in the present work we have monitored the PL both during and after UV laser irradiation. This approach is useful to investigate the stability of these centers under UV exposure and the induced chemical reactions with atomic or molecular species of the ambient atmosphere.
2. Experimental methods We used samples obtained by commercial fumed silica nanoparticles (Aerosil) supplied in powder form by Evonik-Industries [18]. These materials differ for the specific surface (S) and the average particle diameter (d). In particular, we investigated two typologies of materials: one, named AE300, with S¼300730 m2/g and d¼771 nm; the other, named AE150, has S¼150715 m2/g and d¼1472 nm. In a previous study performed by atomic force microscopy technique [19], it is shown that the diameters of these fumed silica nanoparticles have a statistical distribution with average values in agreement with that declared by the producer. Silica nanoparticles have been pressed at 300 MPa into tablet of 1 mm thickness. UV-irradiations were performed at room temperature using the third (3.50 eV) and the fourth harmonic (4.66 eV) of a pulsed Quanta System SYL 201 Nd:YAG laser (pulse width 5 ns, spot size 0:4 cm2 ). The 3.50 eV exposure was done with pulse repetition rates of 10 Hz, 2 Hz, and 1 Hz and energy density of
40
L. Spallino et al. / Journal of Luminescence 138 (2013) 39–43
F1 ¼ 7:5 mJ=cm2 , F2 ¼ 37:5 mJ=cm2 , and F3 ¼ 75 mJ=cm2 , respectively, so as to deposit on the sample the same energy density in the same time interval. The 4.66 eV exposure was carried out at 2 Hz repetition rate and 37.5 mJ/cm2 pulse energy density. The mean power of the laser beam was monitored by a pyroelectric detector and kept constant during each irradiation. Visible luminescent centers were investigated by ex situ timeresolved PL technique. The PL spectra were performed at room temperature in a standard front-scattering geometry; pulsed light at excitation energy Eexc ¼ 4:96 eV (pulse width 5 ns, repetition rate 10 Hz) was provided by a VIBRANT OPOTEK optical parametric oscillator laser system, pumped by the third harmonic (3.50 eV) of a Nd:YAG laser. The pulse energy was monitored with a pyroelectric detector and the fluence/pulse was maintained at F ¼ 0:8 mJ=cm2 , low enough to avoid the generation of any defect. The emitted light was spectrally resolved by a monochromator (SpectraPro 2300i, PI/Acton) equipped with a 150 grooves/ mm grating blazed at 300 nm and acquired by an intensified charge coupled device (CCD) camera driven by a delay generator (PI-MAX Princeton Instruments) that sets the time acquisition of the emitted light (gate width DT ¼ 200 ns and time delay TD ¼ 10 ns after the laser pulse). All the emission spectra were detected with a bandwidth of 20 nm and were corrected for the monochromator dispersion.
3. Results As a representative view, the spectroscopic characteristics of one of the as grown samples (AE150) are reported in Fig. 1. The blue emission is centered around 2.8 eV with full width at half maximum (FWHM) 0:7 eV, its excitation shows two peaks around 3.9 eV and 5.0 eV, its decay occurs in the nanoseconds timescale. These characteristics are in good agreement with previous works [8,9,14]. We also observe a contribution at lower intensity, around 3.5 eV. UV luminescence of high surface silica has been attributed to hydrocarbons contamination [9] or surface silanols (Si–OH) [20], however, in this work we will not deal with it. Prior to study the effects induced by high power laser radiation, we have exposed for 1 h the samples to UV photons coming from the laser used for the ex situ PL measurements (Eexc ¼ 4:96 eV, fluence=pulse 0:8 mJ=cm2 ), up to store a total
Fig. 1. PL spectrum of the as grown AE150 sample excited at Eexc ¼ 4:96 eV and acquired with TD ¼ 10 ns and DT ¼ 200 ns. The insets show, at the top, the excitation spectrum acquired at emission energy Eem ¼ 2:74 eV; at the bottom, the decay time of the blue band peak obtained acquiring the PL spectra with DT ¼ 1 ns and TD ranging from 0 to 20 ns.
fluence 30 000 mJ=cm2 . We observed that this does not cause any variation of the blue band intensity. Then, we have followed the evolution of the spectral and lifetime features related to the blue band, both during and after irradiation with the third harmonic of the Nd:YAG laser (hn1 ¼ 3:50 eV). We have exposed our samples to an increasing number of pulses, from 1 to 20 000, with fluence/pulse F2 ¼ 37:5 mJ=cm2 so as to deposit a total fluence of 750 J/cm2. During the irradiation stage, we performed ex situ time resolved PL spectra under Eexc ¼ 4:96 eV. In Fig. 2a we report some representative curves observed in the AE150 sample; as indicated by the arrow the blue band decreases on increasing the number of deposited pulses. The post irradiation trend of the PL spectra, monitored for a time of about 5 h after the last irradiation, is depicted in Fig. 2b. The blue band intensity increases back to its value before irradiation. We note that its spectral features (peak position and FWHM) remain the same both during and after irradiation. In Fig. 3 we report the decay time recorded before, during and after the irradiation. All the curves are described by a stretched exponential, characteristic of luminescent defects embedded in a disordered matrix [12]; we estimate a common lifetime, measured as the time necessary to reduce the intensity by a factor 1/e, t ¼ 2:0 70:3 ns. We appreciate some differences at longer delay time: the decay rate is faster after a fluence of 150 J=cm2 and after the recovery. The same spectroscopic phenomenology is observed in the AE300 sample. To quantitatively characterize the bleaching and the recovery of the blue band, its PL intensity is reported in Fig. 4. During the irradiation stage, the PL kinetics is plotted as a function of the fluence in a semilogarithmic scale, while it is plotted versus time in the post irradiation stage, where t ¼0 corresponds to the end of the irradiation session. In the same graph we show the PL intensities measured in analogous experiments under exposure
Fig. 2. Evolution of PL spectra measured ex-situ in the sample AE150 during (a) and after (b) irradiation with hn1 ¼ 3:50 eV and fluence/pulse ¼37.5 mJ/cm2. All spectra are excited at Eexc ¼ 4:96 eV and acquired with TD ¼ 10 ns and DT ¼ 200 ns. The arrows indicate the trend of the blue band.
L. Spallino et al. / Journal of Luminescence 138 (2013) 39–43
to F1 ¼ 7:5 mJ=cm2 , F2 ¼ 37:5 mJ=cm2 and F3 ¼ 75 mJ=cm2 . The PL intensity decreases to a value that depends on Fi : it reduces by 50% for F1 ¼ 7:5 mJ=cm2 ; by 80% for F2 ¼ 37:5 mJ=cm2 and almost totally for F3 ¼ 75 mJ=cm2 . The recovery also depends on the fluence/pulse and keeps memory of irradiation stage: in about 5 h, the PL intensity gets back to its as-grown value for F1 ¼ 7:5 mJ=cm2 , while it recovers 70% and 20% of its original value for F2 ¼ 37:5 mJ=cm2 and F3 ¼ 75 mJ=cm2 , respectively. In Fig. 5 the comparison between the effects induced on blue PL intensity by radiation (fluence/pulse fixed at 37.5 mJ/cm2) coming from the third (hn1 ¼ 3:50 eV) and the fourth
41
(hn2 ¼ 4:66 eV) harmonic of the Nd:YAG laser is reported. The comparison suggests that, higher the radiation energy, higher the bleaching effect on the blue band during irradiation is. Indeed we observe that, at a fluence 200 mJ=cm2 , the radiation energy hn1 ¼ 3:50 eV reduces the PL intensity by 75%, while hn2 ¼ 4:66 eV radiation totally removes the PL band. Moreover, the photon laser energy also influences the post-irradiation: after hn1 ¼ 3:50 eV exposure the PL intensity recovers 75% of its original value within 5 h, whereas the bleaching induced by the fourth harmonic of the Nd:YAG laser (hn2 ¼ 4:66 eV) is long lasting.
4. Discussion
Fig. 3. Decay curves of the blue band recorded in the AE150 sample before irradiation (filled circles), during irradiation at hn1 ¼ 3:50 eV after depositing a total fluence of 15 J=cm2 (triangles) and 150 J=cm2 (squares) and 5 h after the end of irradiation (crossed circles). All the curves are obtained acquiring the PL spectra under Eexc ¼ 4:96 eV with DT ¼ 1 ns and TD ranging from 0 to 15 ns. The curves are arbitrarily scaled for viewing purposes.
The luminescence properties of as received silica nanoparticles are mainly characterized by a blue luminescence (peak emission around 2.8 eV), that is excited in the UV and exhibits a fast decay time ( 2 ns) in accordance with a singlet–singlet allowed transition. These spectroscopic features do not allow to address the defect structure; from this viewpoint, the attribution of the blue band to the pair of a dioxasilirane, ¼ SiðO2 Þ, and a silylene ¼ Si , [8] remains an open question. For this purpose, the comparison with the visible luminescence properties associated with the dioxasilirane and silylene could be useful. Table 1 reports the main spectroscopic features of these luminescent defects. It is evident the disagreement among the decay properties: the silylene and dioxasilirane have a much longer lifetime, characteristic of a forbidden transition. This finding would lead to hypothesize that the formation of the defects pair introduces an energy levels configuration with a very fast (ns) transition rate from the excited state. The structural properties of the blue luminescent defect can be discussed on the basis of its response to UV radiation. Our results point out that UV laser photons govern two conversion processes
Fig. 4. PL intensity detected in the AE150 sample at Eem ¼ 2:60 eV as a function of hn1 ¼ 3:50 eV radiation fluence (left) and time (right). The origin of the time scale corresponds to the end of the irradiation session. The three curves refer to: F1 ¼ 7:5 mJ=cm2 (circles), F2 ¼ 37:5 mJ=cm2 (squares), and F3 ¼ 75 mJ=cm2 (crossed squares). The intensity is normalized to that measured in the as-grown sample; its uncertainty is 10%.
42
L. Spallino et al. / Journal of Luminescence 138 (2013) 39–43
Fig. 5. PL intensity detected in the AE150 sample at Eem ¼ 2:60 eV as a function of radiation fluence (left) and of the time elapsed from irradiation (right); the curves refer to hn1 ¼ 3:50 eV (squares) and hn2 ¼ 4:66 eV (circles). The intensity is normalized to that measured in the as-grown sample; its uncertainty is 10%.
Table 1 Main spectroscopic features of visible luminescent defects of silica nanoparticles: blue luminescent defect studied in this work, silylene and dioxasilirane. Defect
Emission peak (eV) Excitation peak (eV) Lifetime
Blue luminescent defect 2.8 2.75 Silylene ¼ Si [23] dioxasilirane ¼ SiðO2 Þ [12] 2.4
3.9 and 5.0 5.0 3.2 and 5.2
2 ns 18 ms 7:5 ms
of the blue luminescent defects at the surface of silica nanoparticles. These consist of a bleaching and a back-conversion observed during and after the UV exposure. It is also evidenced that the luminescence lifetime is weakly influenced by laser radiation. Since the lifetime is governed by radiative and non-radiative decay rates, t ¼ ðkR þkNR Þ 1 is an exponent then we infer that in the excited state the interactions with the surrounding are poorly affected by laser radiation, the quantum yield Z ¼ kR t does not change dramatically and the luminescence bleaching and recovery are mainly due to a concentration variation of the defects. As concerns the bleaching process, it is not observed at fluence/pulse as low as 1 mJ=cm2 , whereas it is effective if the fluence/pulse increases of an order of magnitude. In particular, as evidenced in Fig. 4 (left panel), the bleaching rate increases on increasing the fluence/pulse, in agreement with a non-linear effect, likely based on a multi-photon absorption. The backconversion process is evidenced by the luminescence recovery in the post irradiation stage, nevertheless we argue that it is also effective in the laser interpulse time span. Under this hypothesis, the kinetics of the luminescent centers during the irradiation is governed by the competition between bleaching and back-conversion; we expect that such competition is influenced by the irradiation laser parameters, such as the repetition rate and the fluence/pulse. The use of the Nd:YAG third and fourth harmonic as irradiation beam is crucial for the fate of the luminescent defect (DPL):
(1) Under exposure to hn1 ¼ 3:50 eV photons, DPL converts towards a different configuration. Since the PL recovery is
observed at room temperature, we deduce that this configuration (CMS) is metastable thus allowing a back-conversion. (2) Under exposure to hn1 ¼ 4:66 eV photons, DPL also converts to a new configuration (CS), but in this case the PL recovery is not observed, thus indicating its stability, at least at room temperature. The two conversion pathways are accounted for by the following equation: kc
k0c
C MS ! DPL ,C S
ð1Þ
kbc
Both conversion rates, kc and k0c , depend non-linearly on the photon fluence. As concerns the back-conversion process, it is active from CMS at room temperature and we suppose its rate is of the form kbc ¼ k0 eDE=KT , the activation energy DE being comparable with the thermal energy at room temperature ( 25 meV). On the basis of the comparison with literature data dealing with post-irradiation processes [21], we can only tentatively suggest a k
bi-molecular reaction of the form: A þB-C, in which the two species A and B react to form a species C with a conversion rate k. In the general case of solids [22], these processes are governed by an activation energy due to the diffusion of the atomic or molecular species in the network and, for endothermic processes, the reaction itself requires an amount of activation energy. In our case, the samples are dominated by a high specific surface so that the defects are easily accessible from species of the environment atmosphere. This makes negligible any diffusion and the process could be only reaction limited. We note that these results provide interesting information about the stability of the blue luminescent defect, but it is not enough to unambiguously and ultimately establish its nature. On the basis of the structural model that ascribes the blue band to the defect pair ¼ SiðO2 Þ and ¼ Si [8], the PL bleaching could be interpreted as the effect of O2 detrapping, induced by UV laser photons, consequently the O2 trapping leads to the PL recovery. It is worth noting that previous experiments have pointed out that the reactions of O2 with ¼ Si requires an activation energy of 0:3 eV [11], much larger than ( 25 meV),
L. Spallino et al. / Journal of Luminescence 138 (2013) 39–43
roughly estimated from our data. However, we cannot rule out the hypothesis that two faced ¼ Si make more likely the O2 trapping. To clarify this issue, we aim to deepen our experiments in a thermochemical controlled atmosphere, so as to have clear and useful indications about its origin.
5. Conclusion High power UV radiation, provided by the third (3.50 eV) and the fourth (4.66 eV) harmonic of a Nd:YAG pulsed laser, causes the bleaching of blue luminescence characteristic of silica nanoparticles. Under the higher radiation energy the effect is permanent, otherwise the band follows a post-irradiation recovery kinetics, likely due to the reactions, at the surface sites, with atomic or molecular species of the atmosphere.
Acknowledgments The work was partially supported by FAE project, PO FESR Sicilia 2007/2013 4.1.1.1. We thank the group of the Laboratory of Amorphous Materials Physics (Palermo University) (http://www. fisica.unipa.it//amorphous) for their support and stimulating discussions and G. Napoli and G. Tricomi for their technical assistance. References [1] Y.D. Glinka, S.H. Lin, Y.T. Chen, Appl. Phys. Lett. 75 (1999) 778. [2] Y.D. Glinka, S.H. Lin, Y.T. Chen, Phys. Rev. B 62 (2000) 4373.
43
[3] S. Bonacchi, D. Genovese, R. Juris, M. Montalti, L. Prodi, E. Rampazzo, N. Zaccheroni, Angew. Chem. Int. Ed. 50 (2011) 4056. [4] C. Zhang, J. Lin, Chem. Soc. Rev. 41 (2012) 7938. [5] B. Hamilton, Semicond. Sci. Technol. 10 (1995) 1187. [6] A.G. Cullis, L.T. Canham, P.D.J. Calcott, J. Appl. Phys. 82 (1997) 909. [7] G.L. Davies, J.E. McCarthy, A. Rakovich, Y.K. Gun’ko, J. Mater. Chem. 22 (2012) 7358. [8] A. Anjiki, T. Uchino, J. Phys. Chem. C 116 (2012) 15747. [9] A. Aboshi, N. Kurumoto, T. Yamada, T. Uchino, J. Phys. Chem. C 111 (2007) 8483. [10] V.A. Radzig, Colloids Surf. A 74 (1993) 91. [11] V.A. Radzig, Chem. Phys. Rep. 14 (1995) 1206. [12] L. Vaccaro, A. Morana, V.A. Radzig, M. Cannas, J. Phys. Chem. C 115 (2011) 19476. [13] C.M. Carbonaro, P.C. Ricci, R. Corpino, M. Marceddu, A. Anedda, J. Non-Crys. Solids 357 (2011) 1904. [14] L. Vaccaro, G. Vaccaro, S. Agnello, G. Buscarino, M. Cannas, Solid State Commun. 150 (2010) 2278. [15] G. Vaccaro, S. Agnello, G. Buscarino, M. Cannas, L. Vaccaro, J. Non Cryst. Solids 351 (2011) 1941. [16] S. Banerjee, S. Honkote, A. Datta, J. Phys. Chem. C 115 (2011) 1576. [17] S. Banerjee, S. Maity, A. Datta, J. Phys. Chem. C 115 (2011) 22804. [18] Basic Characteristics of Aerosil, 4th ed., Degussa, Frankfurt (2001). [19] G. Buscarino, V. Ardizzone, G. Vaccaro, S. Agnello, F.M. Gelardi, J. Appl. Phys. 108 (2001) 074314. [20] C.M. Carbonaro, F. Clemente, R. Corpino, P.C. Ricci, A. Anedda, J. Phys. Chem. B 109 (2005) 14441. [21] F. Messina, M. Cannas, J. Non-Cryst. Solids 353 (2007) 522. [22] R.H. Doremus, Diffusion of Reactive Molecules in Solids and Melts, John Wiley and Sons, 2002. [23] V.N. Bagratashivli, S.I. Tsypina, V.A. Radtsig, A.O. Rybaltovskii, P.V. Chernov, S.S. Alimpiev, Y.O. Simanovskii, J. Non-Cryst. Solids 180 (1995) 221.