Nonlinear refraction properties of nickel oxide thin

JOURNAL OF APPLIED PHYSICS 106, 093517 共2009兲

Nonlinear refraction properties of nickel oxide thin films at 800 nm Ronaldo P. de Melo, Jr.,1 Blenio J. P. da Silva,2 Francisco Eroni P. dos Santos,2 A. Azevedo,2 and Cid B. de Araújo2,a兲 1

Programa de Pós-Graduação em Ciência de Materiais, Universidade Federal de Pernambuco, Recife 50670901, PE, Brazil and Colégio Militar do Recife, Exército Brasileiro, Recife 50730-120, PE, Brazil 2 Departamento de Física, Universidade Federal de Pernambuco, Recife 50670-901, PE, Brazil

共Received 22 July 2009; accepted 28 September 2009; published online 10 November 2009兲 Measurements of the nonlinear refractive index, n2, of nickel oxide films prepared by controlled oxidation of nickel films deposited on substrates of soda-lime glass are reported. The structure and morphology of the samples were characterized by scanning electron microscopy, atomic force microscopy, and x-ray diffractometry. Samples of excellent optical quality were prepared. The nonlinear measurements were performed using the thermally managed eclipse Z-scan technique at 800 nm. A large value of n2 ⬇ 10−12 cm2 / W and negligible nonlinear absorption were obtained. © 2009 American Institute of Physics. 关doi:10.1063/1.3254233兴 I. INTRODUCTION

Advances in photonic devices are motivating the development of materials that present high nonlinear 共NL兲 optical response, low optical absorption for the spectral range of interest, small moisture sensitivity, high mechanical resistance and thermal stability, large linear refractive index, and simple fabrication processing. Although a wide range of inorganic and organic photonic materials have been already discovered the search for new materials is still very active. Among the materials being studied, transition metal oxides deserve special attention because they present remarkable chemical stability as well as peculiar optical, electrical, and magnetic properties. In particular, among the binary 3d transition metals compounds, nickel oxide 共NiO兲 has been considered as an attractive material for many technological applications. It has been used in spin-valve devices,1 as p-type transparent conducting material,2 in electrochromic devices,3 as gas sensors,4 as well as for resistive random access memory.5 NiO is an antiferromagnetic material with Neel temperature of 523 K. At room temperature the point symmetry is 2/m that presents inversion as one symmetry operation.6 The linear optical absorption spectrum of NiO was deeply studied and the absorption bands in the visible and in the nearinfrared were associated with transitions within the 共3d兲8 levels of Ni2+.6 The optical band gap is in the near-ultraviolet range. NL optical techniques based on the second-order susceptibility, ␹共2兲, have been applied by several authors to study basic phenomena associated with interface magnetism of NiO films. For instance, second harmonic generation 共SHG兲 was applied to investigate electronic and magnetic properties of NiO. The SHG signal was attributed to combined contributions from magnetic-dipole and electric-dipole transitions between Ni2+ levels.6 Also resonance-enhanced two-photon sum-frequency generation was exploited to investigate NiO a兲

Author to whom correspondence should be addressed. Electronic mail: [email protected].

0021-8979/2009/106共9兲/093517/6/$25.00

and revealed weak magneto-optical transitions difficult to observe using conventional techniques.7 Moreover, ultrafast reorientation of Ni2+ spins due to changes in the magnetic anisotropy was studied8 and coherent oscillations of the magnetic-order parameter due to photoinduced phasetransition were observed. In spite of NiO be considered for many applications there have been few works to investigate its third-order NL optical properties. Since NiO is a centrosymmetric material, the knowledge of ␹共3兲, the third-order susceptibility, provides important information to analyze bulk experiments. On the other hand it may help the interpretation of ␹共2兲 experiments to allow the disentanglement of bulk and surface NL contributions. This is particularly useful for thermally poled glasses as demonstrated in Ref. 9. Two-photon absorption 共TPA兲 of NiO single crystal films was reported in Ref. 10 where the TPA spectrum and its dynamics were discussed. More recently the third-order susceptibility of NiO nanoparticles 共NPs兲 suspended in toluene was studied at 532 nm using the Z-scan technique with 80 ps laser pulses at 7 Hz.11 Colloids with NPs volume fraction of 10−8 – 10−7 present large values of the NL refractive index, n2 ⬀ Re ␹共3兲, in the range of 10−13 – 10−12 cm2 / W and TPA coefficients, ␣2 ⬀ Im ␹共3兲, smaller than 0.2 cm/GW. The electronic contribution for n2 is four orders of magnitude larger than for silver NPs. The large n2 and small ␣2 determined in Ref. 11 motivated the present work with films of NiO where we report on the preparation and characterization of NiO thin films obtained by oxidation of nickel 共Ni兲 films grown by thermal evaporation onto soda-lime glass substrates. The samples prepared were characterized by x-ray diffraction 共XRD兲, scanning electron microscopy 共SEM兲, atomic force microscopy 共AFM兲, optical absorption, m-line technique, and the thermally managed eclipse Z-scan 共TM-EZ scan兲 technique.12 Films of NiO with excellent optical quality and large stability, produced using a method based on the controlled oxidation of Ni films, were used in the experiments. Negligible NL absorption and large n2 ⬇ 10−12 cm2 / W were determined using a laser operating at 800 nm.

106, 093517-1

© 2009 American Institute of Physics

Author complimentary copy. Redistribution subject to AIP license or copyright, see http://jap.aip.org/jap/copyright.jsp

093517-2

de Melo, Jr., et al.

FIG. 1. NiO film thickness dependence as a function of the oxidation time.

II. SAMPLES PREPARATION

Nickel films were deposited by thermal evaporation of Ni chips 共99.99%兲 onto substrates of soda-lime glass. During deposition the substrate was kept at room temperature in a vacuum chamber with base pressure of ⬇3.5⫻ 10−6 torr. Films with thicknesses from 30 to 350 nm were prepared. In order to produce NiO films the Ni films were inserted into a previously heated tube furnace for oxidation. The oxidation process was carried out at 400 ° C under action of a high purity oxygen flux 共50 cm3 / min兲 during various time intervals, from few minutes up to 20 h. Using this method we were able to investigate the dynamics of the oxidation process. After each oxidation run the thickness of the NiO film was monitored by measuring the step height at an edge of the film using a profilometer. The surface morphology development was investigated by SEM and AFM. The film composition and preferential crystal plane formation were confirmed by the XRD technique. The NiO films obtained are very stable with respect to deterioration. However, in order to compare the surface stability against degradation, we also prepared NiO films by two other techniques: 共i兲 by rf sputtering from a NiO target and 共ii兲 by thermal evaporation of Ni followed by oxidation in a muffle furnace. In both cases the NiO films presented a much higher deterioration than the films prepared by the oxidation technique described above. III. RESULTS AND DISCUSSION A. Film thickness evolution

The time evolution of the Ni oxidation process clearly occurs in two stages, as shown in Fig. 1 for a 160 nm thick film. In the first stage the oxidation process takes place during the first 6 h and the NiO film thickness exhibits a linear dependence on time. In the second stage the Ni film is fully oxidized to NiO and the final thickness saturates at ⬇300 nm. Therefore as a result of the oxidation process the film thickness increased up to 88% in comparison with the asdeposited Ni films. As the nominal ratio between the molar volume of NiO 共11.14 cm3 / mol兲 and Ni 共6.59 cm3 / mol兲 is ⬇1.7, the NiO film exhibits small degree of porosity of ⬇5.6%. By using the time dependence diffusion length obtained from the data the diffusion coefficient of 6.2 nm2 / min was estimated. Hence, by adjusting the thickness of initial Ni

J. Appl. Phys. 106, 093517 共2009兲

FIG. 2. 共Color online兲 X-ray diffraction spectra showing the evolution of the oxidation process for different processing times. The curves were shifted in the vertical axis to prevent overlap among them.

film, carefully monitoring the processing conditions and assuming that the oxidation process is uniform, the final thickness of the NiO layer may be controlled. B. XRD, AFM, and SEM analyses

In order to investigate more details of the dynamics of the oxidation process, the sample was removed from the oxidation furnace after successive time intervals for XRD and thickness measurements. Figure 2 shows XRD patterns obtained using Cu K␣ radiation for the same film investigated in Fig. 1. The face-centered cubic 共fcc兲 structure of Ni and the rock salt structure of NiO were observed. Other nickel oxide phases, such as Ni2O3, previously detected in thermally oxidized Ni films,13 were not present. The height of the diffraction peaks corresponding to the 共111兲 and 共200兲 crystal planes increases with the processing time. Consequently, the XRD pattern associated with Ni gradually decreases and practically disappears after 20 h of processing. To determine the average crystallite size we analyzed the spectra obtained after 1.5 h of heat treatment for the Ni film with 300 nm thickness. The average NiO crystallite size can be estimated using Scherrer´s formula L = 0.89␭ / B cos ␪, where L is the crystallite size, ␭ = 1.5405 Å is the wavelength of the x-ray radiation, and B is the full width half maximum 共FWHM兲 of the diffraction peak measured at 2␪. The average size determined from the FWHM of the 共111兲 and 共200兲 peaks of NiO is ⬇19 nm while the average size of the fcc Ni crystallites is ⬇31 nm. Figure 3共a兲 shows the AFM surface topography of the as-deposited Ni film 共2.5⫻ 2.5 ␮m2 sweep兲 with a thickness of 38 nm and rms of 6.29 nm while Fig. 3共b兲 shows the NiO film 共5.0⫻ 5.0 ␮m2 sweep兲 after 1.5 h of oxidation with 47 nm thickness and rms of 8.89 nm. Crystallite sizes for asdeposited Ni films and NiO films were measured by AFM and the size distributions are shown in Figs. 4共a兲 and 4共b兲. In both cases the average crystallite sizes show excellent agreement with the values calculated using the XRD data shown in Fig. 2. No defects such as cracks or pores are observed by the AFM topography analysis. The other films prepared show the same good quality. Figure 5 shows a SEM micrograph of the NiO film with 300 nm thickness obtained by oxidation of a Ni film of 160


093517-3


J. Appl. Phys. 106, 093517 共2009兲

FIG. 4. 共Color online兲 Histogram of crystallite size distribution obtained by AFM: 共a兲 Ni film as-evaporated and 共b兲 NiO film after 1.5 h of time oxidation.

FIG. 3. 共Color online兲 AFM surface topography of the samples. 共a兲 Asdeposited Ni film with 380 nm and rms of 6.29 nm. 共b兲 NiO film after 1.5 h of oxidation with 470 nm and rms of 8.89 nm.

FIG. 5. The SEM micrograph for the NiO film of 300 nm thickness after 90 min of oxidation. It can be seen that the film is dense and continuous, free of cracks, and has low porosity.

nm thickness. Surface morphology reveals a dense, continuous film with very fine grains. The other films prepared show similar optical quality.

C. Linear and nonlinear optical properties

The NiO films obtained exhibit optically clear surfaces with light green color by reflection and a brown color by transmission. The linear optical absorption measurements were performed using a commercial spectrophotometer from 300 to 800 nm. Figure 6 shows the spectrum obtained for a 300 nm NiO thick film, assumed as totally oxidized. The inset shows a plot of the linear absorption coefficient squared, ␣20, versus the photon energy. The optical band gap 共Eg ⬇ 3.8 eV兲 was evaluated by extrapolation of the linear

FIG. 6. 共Color online兲 Optical absorption spectrum 共NiO film thickness: 300 nm兲. The inset shows the behavior of the linear absorption coefficient squared, ␣20, as a function of photon energy.


093517-4


J. Appl. Phys. 106, 093517 共2009兲

FIG. 7. 共Color online兲 Transmission spectra for samples prepared by oxidation of a Ni film with thickness of 38 nm. The spectrum of the substrate is also shown for comparison.

portion of the data toward the energy axis. The result is in agreement with previous measurements.14,15 The band gap of the substrate was 4.0 eV. Transmission spectra were obtained as a function of the annealing time and Fig. 7 presents results for three different oxidation degrees showing that the optical transmission increases for longer annealing times due to the NiO film development. We also measured the linear refractive index by using the m-line technique and found 2.4380 at 632.8 nm, in good agreement with the literature.16 Such large linear refractive index is a strong indication of large third-order NL response according to Miller⬘s law.17 The NL measurements were performed using the TM-EZ scan technique, recently introduced in Ref. 12. This technique allows measurements of the electronic nonlinearity of materials using high repetition pulses and has been exploited to investigate different materials.12,18–21 The experimental setup used is illustrated in Fig. 8. A mode-locked Ti-sapphire laser 共800 nm, ប␻L = 1.56 eV, pulse duration of 150 fs, and pulse repetition rate of 76 MHz兲 was used. To measure n2 the sample is scanned in the focal region of lens L3 along the beam direction. The analysis of the light intensity transmitted through the sample is made by positioning an opaque disk in front of the lens L4 and the eclipsed beam is directed toward the Pd1 detector. The temporal evolution of the EZ-scan signal is obtained by delaying the Pd1 signal acquisition time with respect to the chopper opening time 共t = 0兲. The measurements are made by monitoring the time evolution of the Pd1 signal with the sample placed in prefocal and postfocal positions with respect to lens L3. From these measurements the TM-EZ scan curves representing the sample transmittance as a function of time 共when the sample is in the peak and in the valley positions of the EZ-scan profiles兲 are con-

FIG. 8. 共Color online兲 Experimental TM-EZ scan setup. The labels refer to lenses 共L1-L5兲, beam splitter 共BS兲, photodiodes 共Pd1, Pd2兲, and chopper 共Ch兲.

FIG. 9. 共Color online兲 TM-EZ scan results for two NiO films with thickness of 47 nm 共a兲 and 65 nm 共b兲.

structed. Crossing of the two curves 共corresponding to the pre- and postfocal signals兲 indicates the presence of thermal and electronic nonlinearities with contributions of opposite signs. When the signs of both nonlinearities are the same the curves do not cross but the transmittance change with time. Parallel 共horizontal兲 curves indicate the absence of thermal contribution. The NL absorption coefficient can be measured monitoring the light intensity transmitted by the sample analyzing the signal recorded by the Pd2 detector. Experiments for measuring the NL parameters were carried out for NiO films having different thicknesses. Figure 9 shows TM-EZ scan results 共transmittance profile and temporal evolution兲 obtained for two NiO films with 47 nm and 65 nm thicknesses, using an opaque disk with diameter 1.7 cm in front of lens L4. Notice that the TM-EZ scan profile with the peak 共valley兲 for z ⬍ 0 共z ⬎ 0兲 indicates that n2 is positive. The light induced refractive index change is determined from measurements of ⌬T PV, the peak-to-valley difference in the samples transmittance. The data were analyzed following the procedure of Ref. 12 where ⌬T PV = 15.2␲Leffn2I0␭−1, with Leff = 关1 − exp共−␣0L兲兴 / ␣0, where L is the sample thickness and I0 = 2.1 GW/ cm2 is the on-axis peak intensity at the focus of lens L3. The solid lines describing the temporal evolution of the transmittance are numerical fits that allow to determine the electronic contribution for n2 from the ⌬T PV value at t = 0, assuming that no other fast mechanisms contribute during the chopper opening rise time.


093517-5

J. Appl. Phys. 106, 093517 共2009兲


FIG. 10. 共Color online兲 Energy levels diagram of 3d8 states of Ni2+ and NiO energy bands.

To verify if the sample changes its characteristics due to the laser excitation, sequential measurements were made in two regimes of intensity 共high and low after the NL measurements兲 and the results did not change. As demonstrated in Ref. 22 NL measurements with films may be affected by optical damage but the Z-scan technique cannot provide clear evidence. In the present experiments we assure, with basis on the results of the sequential measurements, that optical damage is not occurring. The measured ⌬T PV allows to obtain a value for the effective NL refractive index, neff 2 , that depends on the nonlinearity of film and substrate. From the intersection of the solid lines with the vertical axis 共t = 0兲 in Fig. 9 we obtain neff 2 of 共3.61⫾ 0.20兲 ⫻ 10−16 cm2 / W 共film with 47 nm thickness兲 and 共4.21⫾ 0.20兲 ⫻ 10−16 cm2 / W 共film with 65 nm thickness兲. Taking into account the relative ratio between the thickness of the films and substrate and the small NL refractive index of the soda-lime glass 共3 ⫻ 10−16 cm2 / W, Ref. 23兲 we obtain n2 ⬇ 共1.15⫾ 0.30兲 ⫻ 10−12 cm2 / W. This value is three orders of magnitude larger than the value obtained for liquid CS2, used as reference standard 关n2 = 2.50

⫻ 10−15 cm2 / W 共Ref. 12兲兴. The NL absorption coefficient ␣2 was smaller than the minimum value that can be detected by our setup 共⬍660 cm/ GW兲 and the results for all films are similar. In order to understand the NL results we recall that the optical transitions of NiO for photon energy smaller than Eg are determined by local 3d-3d transitions of the Ni2+ 共3d兲8 electrons. Valence-to-conduction band transitions occur for photon energies larger than ⬇3.8 eV. The electronic structure of NiO was studied in Refs. 24 and 25 and Fig. 10 共based on Ref. 10兲 indicates the electronic energy levels. Notice that ប␻L 共1.56 eV兲 and 2ប␻L 共3.12 eV兲 are off-resonance with real transitions of Ni2+ ions and the energy detuning 兵2ប␻L − 关E共 3⌫+4 兲 − E共 3⌫+2 兲兴其 ⬇ 0.2 eV is large enough to make the contribution of the inside gap states negligible. Therefore, considering a two-band model to describe the NL response of the NiO crystal,26,27 we understand that ␣2 should be small because of the large energy detuning 兩2ប␻ − Eg兩 ⬇ 0.82 eV. Moreover, according to Ref. 27, it can be seen that the positive value obtained for n2 is expected because ប␻L / Eg ⬇ 0.4. Table I presents a comparison between the value of n2 for the NiO films prepared in this work and other highly NL materials. Because ␣2 could not be determined it is not possible to provide a precise figure-of-merit F = n2 / ␭␣2 from the present measurements. The value of ␣2, indicated in Table I, is the minimum value that our setup allows to measure but the actual value of ␣2 for our samples may be much smaller than it is indicated. According to Ref. 10 the TPA coefficient for excitation with two-photon energy of 3.12 eV is ⬇20 cm/GW. Considering this value we obtain F = 12.5 which is very appropriate for all-optical switching.

IV. SUMMARY

NiO films were grown by thermal oxidation of evaporated films of Ni deposited on glass substrates. The oxidation was carried out in a furnace under action of a high purity oxygen flow and the NiO film microstructure was investigated as a function of temperature and annealing time. Surface morphology analysis showed that the films are continuous and fine-grained with no defects such as cracks and pores. The measured NL refractive index obtained is almost four orders of magnitude larger than CS2 at 800 nm. The linear and NL optical properties show that the NiO films reported here are strong candidates for photonic applications.

TABLE I. NL parameters of the NiO films and other highly NL films.

Material NiO film NiO film PbO– GeO2 film with Cu NPs Bi2Nd2Ti3O12 film Bi3.35La0.75Ti3O12 film CuO film

␭ 共nm兲 800 800 532 532 800

Pulse duration

n2 ⫻ 1010 共cm2 / W兲

␣2 共cm/GW兲

Reference

150 fs 10 ns 150 fs 35 ps 35 ps 50 fs

0.2 ¯ 0.063 7 3.1 0.0039

⬍660 20 ⬍660 3.1⫻ 104 3 ⫻ 104 ⫺16.9

This work 10 19 28 29 30


093517-6

ACKNOWLEDGMENTS

We acknowledge the financial support of the Brazilian agencies: Conselho Nacional de Desenvolvimento Científico e Tecnológico 共CNPq兲 and Fundação de Amparo à Ciência e Tecnologia do Estado de Pernambuco 共FACEPE兲. We also thank Jorlândio F. Felix, Clécio G. dos Santos, and João Carlos Albuquerque for their help with the measurements of film thickness, AFM, XRD, and SEM. This work was supported by the National Institute of Photonics 共INCT Photonics兲 project. 1

E. Fujii, A. Tomozawa, H. Torii, and R. Takayama, Jpn. J. Appl. Phys., Part 2 35, L328 共1996兲. H. Sato, T. Minami, S. Takata, and T. Yamada, Thin Solid Films 236, 27 共1993兲. 3 C. M. Lampert, Mater. Today 7, 28 共2004兲. 4 H. Kumagai, M. Matsumoto, K. Toyoda, and M. Obara, J. Mater. Sci. Lett. 15, 1081 共1996兲. 5 L. Courtade, Ch. Turquat, Ch. Muller, J. G. Lisoni, L. Goux, D. J. Wouters, D. Goguenheim, P. Roussel, and L. Ortega, Thin Solid Films 516, 4083 共2008兲. 6 M. Fiebig, D. Frölich, Th. Lottermoser, V. V. Pavlov, R. V. Pisarev, and H. J. Weber, Phys. Rev. Lett. 87, 137202 共2001兲共and references therein兲. 7 T. Satoh, Th. Lottermoser, and M. Fiebig, Appl. Phys. B: Lasers Opt. 79, 701 共2004兲. 8 N. P. Duong, T. Sato, and M. Fiebig, Phys. Rev. Lett. 93, 117402 共2004兲. 9 O. Deparis, F. P. Mezzapesa, C. Corbari, P. G. Kazansky, and K. Sakaguchi, J. Non-Cryst. Solids 351, 2166 共2005兲. 10 S. I. Shablev and R. V. Pisarev, Phys. Solid State 45, 1742 共2003兲. 11 L. A. Gómez, C. B. de Araújo, L. M. Rossi, S. H. Masunaga, and R. F. Jardim, Opt. Lett. 32, 1435 共2007兲. 2

J. Appl. Phys. 106, 093517 共2009兲

de Melo, Jr., et al. 12

A. S. L. Gomes, E. L. Falcão-Filho, C. B. de Araújo, D. Rativa, and R. E. de Araujo, Opt. Express 15, 1712 共2007兲. 13 B. Sasi, K. G. Gopchandran, P. K. Manoj, P. Koshy, P. P. Rao, and V. K. Vaidyan, Vacuum 68, 149 共2002兲. 14 R. Newman and R. M. Cerenko, Phys. Rev. 114, 1507 共1959兲. 15 L. Ai, G. Fang, L. Yuan, N. Liu, M. Wang, C. Li, Q. Zhang, J. Li, and X. Zhao, Appl. Surf. Sci. 254, 2401 共2008兲. 16 R. J. Powell and W. E. Spicer, Phys. Rev. B 2, 2182 共1970兲. 17 R. W. Boyd, Nonlinear Optics 共Academic, San Diego, 2003兲. 18 A. S. L. Gomes, E. L. Falcão Filho, C. B. de Araújo, D. Rativa, R. E. de Araujo, K. Sakaguchi, F. P. Mezzapesa, I. C. S. Carvalho, and P. G. Kazansky, J. Appl. Phys. 101, 033115 共2007兲. 19 D. Rativa, R. E. de Araujo, C. B. de Araújo, A. S. L. Gomes, and L. R. P. Kassab, Appl. Phys. Lett. 90, 231906 共2007兲. 20 L. A. Gómez, F. E. P. dos Santos, A. S. L. Gomes, C. B. de Araújo, L. R. P. Kassab, and W. G. Hora, Appl. Phys. Lett. 92, 141916 共2008兲. 21 F. E. P. dos Santos, F. C. Fávero, A. S. L. Gomes, J. Xing, Q. Chen, M. Fokine, and I. C. S. Carvalho, J. Appl. Phys. 105, 024512 共2009兲. 22 K. Fedus, G. Boudebs, C. B. de Araújo, M. Cathelinaud, F. Charpentier, and V. Nazabal, Appl. Phys. Lett. 94, 061122 共2009兲. 23 K. Jamshidi-Ghaleh, N. Mansour, and A. Namdar, Laser Phys. 15, 1714 共2005兲. 24 M. Takahashi and J. Igarashi, Phys. Rev. B 54, 13566 共1996兲. 25 T. Bredow and A. R. Gerson, Phys. Rev. B 61, 5194 共2000兲. 26 H. S. Brandi and C. B. de Araújo, J. Phys. C 16, 5929 共1983兲. 27 M. Sheik-Bahae, D. J. Hagan, and E. W. van Stryland, Phys. Rev. Lett. 65, 96 共1990兲. 28 B. Gu, H. Wang, X. C. Peng, J. P. Ding, J. L. He, and H. T. Wang, Appl. Phys. Lett. 85, 3687 共2004兲. 29 H. Shin, H. J. Chang, R. W. Boyd, M. R. Choi, and W. Jo, Opt. Lett. 32, 2453 共2007兲. 30 A. Chen, G. Yang, H. Long, F. Li, Y. Li, and P. Lu, Thin Solid Films 517, 4277 共2009兲.
