Femtosecond third-order nonlinear spectra of lead-germanium oxide glasses containing silver nanoparticles

Femtosecond third-order nonlinear spectra of lead-germanium oxide glasses containing silver nanoparticles Leonardo De Boni,1,* Emerson C. Barbano,1 Thiago A. de Assumpção,2 Lino Misoguti,1 Luciana R. P. Kassab,3 and Sergio C. Zilio1 1

Instituto de Física de São Carlos, Universidade de São Paulo, CP 369, 13560-970 São Carlos, SP, Brazil 2 Escola Politécnica,Universidade de São Paulo, 05508-900, São Paulo, SP, Brazil 3 Faculdade de Tecnologia de São Paulo - CEETEPS, 01124060, São Paulo, SP, Brazil * [email protected]

Abstract: This work reports on the spectral dependence of both nonlinear refraction and absorption in lead-germanium oxide glasses (PbO-GeO2) containing silver nanoparticles. We have found that this material is suitable for all-optical switching at telecom wavelengths but at the visible range it behaves either as a saturable absorber or as an optical limiter. ©2012 Optical Society of America OCIS codes: (190.0190) Nonlinear optics; (160.2750) Glass and other amorphous materials.

References and links 1. 2.

P. N. Prasad, Nanophotonics (Wiley, (2004)). D. M. da Silva, L. R. P. Kassab, S. R. Lüthi, C. B. de Araújo, A. S. L. Gomes, and M. J. V. Bell, “Frequency upconversion in Er3+ doped PbO–GeO2 glasses containing metallic nanoparticles,” Appl. Phys. Lett. 90(8), 081913 (2007). 3. L. R. P. Kassab, F. A. Bomfim, J. R. Martinelli, N. U. Wetter, J. J. Neto, and C. B. Araújo, “Energy transfer and frequency upconversion in Yb3+–Er3+-doped PbO-GeO2 glass containing silver nanoparticles,” Appl. Phys. B 94(2), 239–242 (2009). 4. L. A. Gomez, F. E. P. dos Santos, A. S. L. Gómes, C. B. de Araujo, L. R. P. Kassab, and W. G. Hora, “Nearinfrared third-order nonlinearity of PbO–GeO2 films containing Cu and Cu2O nanoparticles,” Appl. Phys. Lett. 92(14), 141916 (2008). 5. C. B. de Araújo, T. R. Oliveira, E. L. Falcão-Filho, D. M. Silva, and L. R. P. Kassab, “Nonlinear optical properties of PbO–GeO2 films containing gold nanoparticles,” J. Lumin. In press. 6. D. Faccio, P. D. Trapani, E. Borsella, F. Gonella, P. Mazzoldi, and A. M. Malvezzi, “Measurement of the thirdorder nonlinear susceptibility of Ag nanoparticles in glass in a wide spectral range,” Europhys. Lett. 43(2), 213– 218 (1998). 7. M. Sheik-Bahae, A. A. Said, and E. W. Van Stryland, “High-sensitivity, single-beam n(2) measurements,” Opt. Lett. 14(17), 955–957 (1989). 8. L. De Boni, A. A. Andrade, L. Misoguti, C. R. Mendonça, and S. C. Zilio, “Z-scan measurements using femtosecond continuum generation,” Opt. Express 12(17), 3921–3927 (2004). 9. T. Hayakawa, S. T. Selvan, and M. Nogami, “Field enhancement effect of small Ag particles on the fluorescence from Eu3+-doped SiO2 glass,” Appl. Phys. Lett. 74(11), 1513–1515 (1999). 10. S. Qu, Y. Zhang, H. Li, J. Qiu, and C. Zhu, “Nanosecond nonlinear absorption in Au and Ag nanoparticles precipitated glasses induced by a femtosecond laser,” Opt. Mater. 28(3), 259–265 (2006). 11.P. P. Kiran, B. N. S. Bhaktha, D. N. Rao, and G. De “Nonlinear optical properties and surface-plasmon enhanced optical limiting in Ag-Cu nanoclusters co-doped SiO2 Sol-Gel,” J. Appl. Phys. 96(11), 6717–6723 (2004). 12.U. Gurudas, E. Brooks, D. M. Bubb, S. Heiroth, T. Lippert, and A. Wokaun, “Saturable and reverse saturable absorption in silver nanodots at 532nm using picoseconds laser pulse,” J. Appl. Phys. 104(7), 073107 (2008). 13. L. Boni, E. L. Wood, C. Toro, and F. E. Hernandez, “Optical saturable absorption in gold nanoparticles,” Plasmonics 3(4), 171–176 (2008). 14. X. H. Wang, D. P. West, N. B. McKeown, and T. A. King, “Determining the cubic susceptibility (3) of films or glasses by the maker fringe method: a representative study of spin-coated films of copper phthalocyanine derivation,” J. Opt. Soc. Am. B 15(7), 1895–1902 (1998). 15. V. Mamidala, G. Xing, and W. Ji, “Surface plasmon enhanced third-order nonlinear optical effects in Ag-Fe3O4 nanocomposites,” J. Phys. Chem. C 114(51), 22466–22471 (2010). 16. E. Falcão-Filho, C. Bosco, G. Maciel, L. Acioli, C. de Araújo, A. Lipovskii, and D. Tagantsev, “Third-order optical nonlinearity of a transparent glass ceramic containing sodium niobate nanocrystals,” Phys. Rev. B 69(13), 134204 (2004).

#161877 - $15.00 USD

(C) 2012 OSA

Received 20 Jan 2012; revised 29 Feb 2012; accepted 4 Mar 2012; published 9 Mar 2012

12 March 2012 / Vol. 20, No. 6 / OPTICS EXPRESS 6844

17.D. S. Corrêa, L. De Boni, L. Misoguti, I. Cohanoschi, F. E. Hernandez, and C. R. Mendonca, “Z-scan theoretical analysis for three-, four- and five-photon absorption,” Opt. Commun. 277(2), 440–445 (2007). 18. G. I. Stegeman, Nonlinear Optics of Organic Molecules and Polymers (CRC, (1997)). 19. V. Mizrahi, K. W. Delong, G. I. Stegeman, M. A. Saifi, and M. J. Andrejco, “Two-photon absorption as a limitation to all-optical switching,” Opt. Lett. 14(20), 1140–1142 (1989). 20. J. M. P. Almeida, L. De Boni, A. C. Hernandes, and C. R. Mendonça, “Third-order nonlinear spectra and optical limiting of lead oxifluoroborate glasses,” Opt. Express 19(18), 17220–17225 (2011). 21. G. Lin, F. Luo, H. Pan, M. M. Smedskjaer, Y. Teng, D. Chen, J. Qiu, and Q. Zhao, “Universal preparation of novel metal and semiconductor nanoparticle-glass composites with excellent nonlinear optical properties,” J. Phys. Chem. C 115(50), 24598–24604 (2011). 22. Q.-Q. Wang, J.-B. Han, H.-M. Gong, D.-J. Chen, X.-J. Zhao, J.-Y. Feng, and J.-J. Ren, “Linear and nonlinear optical properties of Ag nanowire polarizing glass,” Adv. Funct. Mater. 16(18), 2405–2408 (2006).

1. Introduction Optical materials containing metallic nanoparticles (NPs) are attracting a great deal of attention because these may contribute to the enhancement of the optical nonlinearities [1]. Such improvement is important for the development of all-optical photonic devices such as optical switches and limiters. Among several materials that can be employed for photonic applications, glasses are particularly interesting due to their chemical and mechanical stability, wide compositional range, fast response times and high optical nonlinearities. The latter can be achieved in glasses containing heavy-metal atoms as they have easily deformable electron clouds that may lead to high hyperpolarizabilities. In particular, lead-germanium oxide (PGO) glasses constitute a class of materials that fulfills the above requirements. In addition, they exhibit large transmittance window from the visible to the infrared region, low cutoff phonon energy (~700 cm−1) and high refractive index (~2.0). Nucleation of metallic NPs in PGO glasses was recently reported, showing the enhancement of rare earth ions luminescence in the presence of silver NPs [2, 3]. The growth of near-infrared nonlinearities in PGO films containing Cu NPs was also demonstrated [4]. Furthermore, studies of the third-order susceptibility in pico- and sub-picosecond regimes of glassy PbO-GeO2 films containing gold NPs revealed ultrafast response and enhanced values of the nonlinear refractive index at 532 and 800 nm due to the presence of nanoparticles [5]. Another interesting investigation measured the third-order nonlinear susceptibility as 6 x10−9 esu for Ag nanoparticles embedded in BK7 glasses between 380 nm and 470 nm (at the Ag surface plasmon band) [6]. In all these previous studies, the nonlinearities were investigated at some specific wavelengths. However, in order to fully evaluate the potential of a given material for photonics applications, the knowledge of the nonlinear response in a broad spectral range is required. A material presenting excellent nonlinear refractive index may not perform well as an all-optical switch if it exhibits deleterious strong two-photon absorption (2PA). Therefore, the spectral behavior of both nonlinear refraction and absorption has to be characterized in order to select the optimum operational wavelength for an optical device using a given sample. Within this context, the present work investigates the spectral dependence of the femtosecond third-order nonlinearity in PbO-GeO2 glasses with and without Ag NPs. We have found that this material is suitable for all-optical switching at telecom wavelengths (around 1.5 μm) independently of the presence of the NPs. However, the NPs play an important role at the visible range and this material can behave either as a saturable absorber (below 450 nm) or as an optical limiter (from 480 to 550 nm). 2. Materials and methods The glass samples were prepared by the melting-quenching technique using the following composition: 59PbO-41GeO2 (in wt. %). Initially, a sample without AgNO3 was prepared to be used as reference (sample A). The nucleation of Ag NPs followed the procedure described in previous report [3]. The doping specie used was AgNO3 (5.0 wt. %). The reagents were melted at 1200 °C in an alumina crucible for 1 h, quenched in a preheated brass mold and

#161877 - $15.00 USD

(C) 2012 OSA

Received 20 Jan 2012; revised 29 Feb 2012; accepted 4 Mar 2012; published 9 Mar 2012

12 March 2012 / Vol. 20, No. 6 / OPTICS EXPRESS 6845

annealed at 420 °C for 1 h (sample B). Then, they are further annealed at 480°C for 3 h (sample C). These heat treatments were performed to thermally reduce Ag+ ions to Ag0 and to nucleate silver nanoparticles. After the heat treatment, the samples were cooled to room temperature inside the furnace. Polished flat samples with a thickness of approximately 1.3 mm were used for both linear and nonlinear measurements. The linear refractive indices of the samples studied here were measured to be 1.9 ± 0.1 at 632 nm. Linear absorption spectra were measured with a Shimadzu UV-1800 spectrometer, while nonlinear absorption and refraction were obtained by means of the Z-scan [7] and white-light continuum (WLC) Z-scan [8] techniques. These techniques are able to provide the nonlinear optical response of a given material by moving the sample through the focal plane of a Gaussian laser beam and monitoring intensity changes in the far field. The nonlinear absorption was measured with the open-aperture WLC Z-scan because it provides a better spectral resolution. However, the nonlinear refraction (n2) was obtained with the traditional closed-aperture Z-scan technique. In this case, we used an optical parametric amplifier (OPA) as the excitation light source. It provides pulses with approximately 120 fs, covering the spectral range from 460 to 2000 nm, with a 1 kHz repetition rate. The OPA was pumped by a Ti:sapphire chirped pulse amplified system delivering 150 fs pulses at 775 nm. In the case of the WLC Z-scan measurement, light at 1100 nm coming from the OPA was focused into a 3 cm-thick homemade fused silica cuvette containing distillated water to generate the WLC. After the cuvette, an IR filter (KG3) was used to remove the infrared part of the pumping light and of the WLC beam. After passing through the filter, the WLC beam was collimated by an achromatic lens and used in the Zscan setup. For the WLC Z-scan technique, the output signal was measured by a portable spectrometer, and for the traditional closed aperture Z-scan, the output signal was obtained using two silicon photo-detectors, one in the sample channel and the other as a reference. Both detectors were coupled to a lock-in amplifier that integrates 1000 shots per Z-scan point. The OPA pulse energies and beam waist size used in this experiment range from 6 to 62 nJ and 13 to 20 μm, respectively, depending to the excitation wavelength. In the case of a material presenting 2PA, the total absorption coefficient is intensity dependent and can be written as α = α0 + βI, where I is the laser beam irradiance, α0 is the linear absorption coefficient, and β is the 2PA coefficient, which is determined by monitoring changes in the transmittance while the sample is scanned through the focal position in the open-aperture configuration. By integrating the transmitted power over time, and assuming a temporal Gaussian pulse, one can obtain the normalized transmittance (NT) as [7]:

T =

1

π q0

∞

ln  1 + q ( z , 0 ) e τ ∫  z , 0 ( ) −

2

0

−∞

 dτ , 

(1)

where = q0 ( z , t ) β I 0 (t ) Leff (1 + z 2 / z02 ) −1 , Leff = [1-exp(-α0L)]/α0, L is the sample thickness, z0 = kw02/2 is the Rayleigh length, w0 is the beam waist, k = 2π/λ is the wave vector, λ is the laser wavelength, z the sample position, and I0 is the on-axis irradiance at the focus (z = 0). Since the linear absorption is very small, we can safely assume Leff ≈L. The 2PA coefficient β can be determined by fitting the experimental Z-scan curves using Eq. (1). The nonlinear refraction coefficient, n2, is obtained in the closed-aperture configuration by fitting the experimental curve with [7]:

T ( z , ∆φ0 ) =+ 1

#161877 - $15.00 USD

(C) 2012 OSA

4∆φ0 x , ( x 2 + 9)( x 2 + 1)

(2)

Received 20 Jan 2012; revised 29 Feb 2012; accepted 4 Mar 2012; published 9 Mar 2012

12 March 2012 / Vol. 20, No. 6 / OPTICS EXPRESS 6846

where x = z/z0 and ∆φ0 is the on-axis phase shift defined as ∆φ0 = kn2 I 0 L. In order to calibrate the irradiance and beam parameter values, we measured closed-aperture Z-scans in a 1 mm-thick fused silica sample, which has a well-known value of n2. Optical limiting measurements were performed with the OPA tuned to 520 nm, and using two polarizers to control the irradiance. To obtain a reference signal for the input intensity, the incident laser light was separated by a beam splitter located after the second polarizer. Then, the laser beam was focused into the sample with a f = 15 cm lens (w0 = 15 μm). Fused silica was also used as a reference material for the optical limiting experiments. 3. Results and discussion Figure 1(a) shows linear absorption spectra of the PGO samples studied here. It also depicts nonlinear absorption spectra, but these will be discussed later. It is clearly seen that the sample without NPs (sample A) is transparent for wavelengths longer than 400 nm (solid line). The sample containing Ag NPs and annealed at 420 °C for 1 h (sample B) presents an emerging absorption band located around 460 nm (dashed line), which is related to the surface plasmon resonance (SPR) of the Ag NPs.1 After the sample is further heat-treated at 480 °C for 3 h (sample C), it is possible to observe that the surface plasmon band is amplified by a factor of nearly 10 (dotted line).

Fig. 1. (a) Linear absorption (lines) and normalized transmittance (symbols) spectra for glass samples with and without Ag NPs. (b) Size distribution histogram of Ag NPs for sample C. The inset shows the image of Ag NPs investigated with a high resolution TEM.

By considering the classical theory of Mie-Drude [9] which explains the characteristics of surface plasmon resonance (SPR) band, and using the approximation of Doyle and Kreibig, one can determine the NP size from the linear absorption data. By using the SPR band width at half maximum and the Fermi velocity of silver (νf = 1.397 x108 cm/s), the dependence with the nanoparticle size is 2= R ν f ∆ω1 2 , where R is the radius of the Ag NP, and ∆ω1/2 is the angular frequency variation that is a function of the SPR band width at the half maximum. The average radius obtained are R = 6.7 nm and 2.7 nm for samples B and C respectively. A high-resolution transmission electron microscope was used to confirm the result for sample C, see Fig. 1(b). The inset of Fig. 1(b) shows the TEM image for this case, while the size distribution histogram corroborates the value obtained with the analysis of the SPR band. A filling factor of about 2% was determined for sample C. Figure 2 shows (a) open- and (b) closed-aperture Z-scan curves for different wavelengths measured in sample C. The solid lines represent the fits provided respectively by Eqs. (1) and (2). Similar results were obtained for the other two samples at this spectral range (data not show here).

#161877 - $15.00 USD

(C) 2012 OSA

Received 20 Jan 2012; revised 29 Feb 2012; accepted 4 Mar 2012; published 9 Mar 2012

12 March 2012 / Vol. 20, No. 6 / OPTICS EXPRESS 6847

Fig. 2. Typical (a) open- and (b) closed-aperture Z-scan curves obtained at several wavelengths in sample C. The solid lines represent fittings achieved with Eq. (1) and (2) respectively. In (b) the curves are vertically shifted by 0.5. The pulse energy and the beam waist used in the open aperture Z-scan are 34 nJ and 16 μm for 500 nm, 29 nJ and 17 μm for 550 nm, 27 nJ and 18 μm for 600 nm, 23 nJ and 19 μm for 650 nm, 19 nJ and 20 μm for 700. The pulse energy and the beam waist used in the closed aperture Z-scan are 8 nJ and 16 μm for 560 nm, 16 nJ and 17μm for 650 nm, 16 nJ and 18 μm for 750, 17 nJ and 18.5 μm for 800 nm and 62 nJ and 20 μm for 1300 nm.

In Fig. 2(a), it is clearly seen that the nonlinear absorption increases as the excitation wavelength approaches the band gap energy. These results are summarized in Fig. 1(a), that shows the minimum value of the NT curves (at z = 0) obtained in the open-aperture WLC Zscan technique. The decrease in the NT for wavelengths between 550 and 700 nm is the same for all samples and is attributed to a 2PA process intrinsic to the glass host. On the other hand, the 2PA process between 480 and 550 nm is enhanced in the sample that presents the strongest SPR band (sample C). This increment was also observed by Qu et al. [10] in Ag NPs precipitated in glasses and was attributed to an interband transition, also mediated by a 2PA excitation. A similar result was also observed for Ag-Cu nanoclusters [11]. Another feature seen in Fig. 1(a) is the saturable absorption (SA) observed for samples with Ag NPs, occurring near the region of the SPR band. The SA effect is a nonlinear optical process, caused by the saturation of the linear absorption in the plasmon band that has already been reported for different NPs [12, 13]. According to Ref. 12, the decrease in the absorption on the SPR band (bleaching) for Ag nanoparticles is result of a change in the oscillation frequency of the surface electrons when they are excited by light. In the case of the nonlinear refraction measurements, all curves depicted in Fig. 2(b) present the same profile, a minimum followed by a maximum. No inversion is observed at this wavelength interval, meaning that the nonlinearity induced on the sample is always positive. The magnitude of n2, obtained by means of Eq. (2), ranges from 10 to 25 x10−20 m2/W. These values are about one order of magnitude higher than that of fused silica [14]. Figure 3 shows the 2PA coefficient (a) and nonlinear refraction (b) spectra of the three samples studied. The SA region was excluded and the linear absorption is also plotted in order to associate linear and nonlinear effects. There is an enhancement of nonlinear effects as the excitation wavelength approaches the absorbing region for all three samples. In the case of 2PA (Fig. 3(a)), the results show a clear dependence on the presence of NPs. For sample C (circles), the enhancement of β is about 2.5 at 500 nm when compared to samples A and B. In the case of the nonlinear refraction at approximately 560 nm (Fig. 3(b)), n2 increases nearly 2.5 times when compared to the value observed at 1300 nm. However, it does not shown any clear dependence with the NPs presence, meaning that this nonlinearity originates primarily from the glass host.

#161877 - $15.00 USD

(C) 2012 OSA

Received 20 Jan 2012; revised 29 Feb 2012; accepted 4 Mar 2012; published 9 Mar 2012

12 March 2012 / Vol. 20, No. 6 / OPTICS EXPRESS 6848

Fig. 3. (a) Two-photon absorption coefficient for samples A (open squares), B (open circles) and C (open triangles), together with absorbance spectra (lines); (b) nonlinear refraction spectra for samples A (open squares), B (open circles) and C (open triangles), together with absorbance spectra (lines). In (b), the black dashed line is just to guide the eyes and the error bars come from the average of three measurements.

Similar results were observed in Ag-Fe3O4 nanocomposites by Mamidala et al. [15]. In that paper, the author measured an increase of about 1.6 in the 2PA coefficient when Ag NPs (~7 nm size) are present in a solution of Fe3O4/toluene compared to a solution without NPs. On the other hand, the nonlinear refraction in these nanocomposites is not clearly enhanced when Ag nanoparticles are added, analogous to the results found in our case. Experimentally, they obtained nearly the same value for n2 (inside of the error bar) for different NPs sizes and without Ag NPs. According to their theoretical model, n2 was expected to have an increase in the magnitude in the same way β was found to have. They explained the discrepancy in the case of the nonlinear refraction between the theoretical and the experimental results as a consequence of the interaction between the metal NPs and Fe3O4 cubes. Another interesting work demonstrated that changes in n2 are observed when the filling factor is larger than 5% in transparent glass-ceramic containing sodium niobate nanocrystals [16]. For a filling factor of about 8%, the authors observed an improvement in the nonlinear refractive index of about 20% at 800 nm, meaning that the filling factor is an important parameter that needs to be taken into account in order to improve the nonlinear effects. In our case, the filling factor of just 2% could explain the negligible dependence with the NPs. As shown in Fig. 1(a), the 2PA effect in PGO glasses containing Ag NPs could be successfully used to build an optical limiting device between 480 and 550 nm. Indeed, as depicted in Fig. 4(a), there is an optical limiting effect at 520 nm for all samples studied here, which presents a trend similar to the theoretical calculations [17]. For the sake of comparison, we also plotted the transmission of silica (stars), which shows no limiting effect. Moreover, Fig. 4(a) shows that samples A and B present approximately the same output intensity curve, as it was expected due to the same value of β at 520 nm. On the other hand, sample C has an enhancement in the optical limiting behavior of about twice (open triangles), following the increment in the magnitude of β at 520 nm. For a material to perform well as an all-optical switch, it should exhibit low linear and nonlinear losses, large Kerr nonlinearity, and ultrafast response [18]. For our samples, one of the figures of merit, W = ∆nmax λα 0 , is always greater than 2, meaning that the Kerr nonlinearity is much more pronounced than the small linear absorption. Another important figure of merit is expressed by FOM = 2βλ/n2 [19]. A good candidate for photonic switching applications must have FOM < 1 at telecommunication wavelengths. Figure 4(b) displays the FOM of the PGO samples as a function of the wavelength. One can see that FOM < 1 is observed for wavelengths greater than 600 nm for all samples and, practically, for wavelengths greater than 750 nm the value of FOM is nearly constant. Around 1.3 μm, samples present n2 ≈12 x10−20 m2/W, β