Annealing effect on the optical properties and laser ... - OSA Publishing

47 downloads 0 Views 266KB Size Report
State Key Laboratory of Coal Conversion, Institute of Coal Chemistry, Chinese Academy of Sciences,. Taiyuan 030001, China, and College of Material Science ...
1066

J. Opt. Soc. Am. B / Vol. 24, No. 5 / May 2007

Liang et al.

Annealing effect on the optical properties and laser-induced damage resistance of solgel-derived ZrO2 films Liping Liang State Key Laboratory of Coal Conversion, Institute of Coal Chemistry, Chinese Academy of Sciences, Taiyuan 030001, China, and College of Material Science and Engineering, Taiyuan University of Science and Technology, Taiyuan 030024, China, and Graduate School of Chinese Academy of Sciences, Beijing 100049, China

Yao Xu State Key Laboratory of Coal Conversion, Institute of Coal Chemistry, Chinese Academy of Sciences, Taiyuan 030001, China

Lei Zhang State Key Laboratory of Coal Conversion, Institute of Coal Chemistry, Chinese Academy of Sciences, Taiyuan 030001, China and Graduate School of Chinese Academy of Sciences, Beijing 100049, China

Yonggang Sheng State Key Laboratory of Coal Conversion, Institute of Coal Chemistry, Chinese Academy of Sciences, Taiyuan 030001, China and Graduate School of Chinese Academy of Sciences, Beijing 100049, China

Dong Wu and Yuhan Sun State Key Laboratory of Coal Conversion, Institute of Coal Chemistry, Chinese Academy of Sciences, Taiyuan 030001, China Received July 7, 2006; revised December 18, 2006; accepted January 10, 2007; posted January 18, 2007 (Doc. ID 72767); published April 17, 2007 By modifying some structural characteristics, the annealing process can have considerable effects on the optical performance of the solgel-derived ZrO2 xerogel films. Annealing at increasing temperature from 150° C to 750° C gives rise to first an increase of refractive index from 1.63 (at 633 nm) to 1.93 and then a decrease to 1.86 with the watershed temperature of 550° C. This can be associated with the evolutions in both packing density and structure order of the films due to the removal of organic segments, material crystallization, and phase transformation. The optical bandgap is found to decrease from 5.63 to 4.97 eV over the entire temperature range, suggesting an increasing nonlinear absorption in the case of high-power laser irradiation. Moreover, annealing completely destroys the network structure of the xerogel films that is suspected to facilitate the energy relaxation. Thus, the combined effect of the greatly weakened endurance and possible enhanced absorption to irradiation laser leads to a monotonous decrease of the laser-induced damage threshold from 55 to 10 J / cm2 (at 1053 nm, 10 ns pulse duration, and R/1 testing mode). © 2007 Optical Society of America OCIS codes: 160.6060, 160.4670, 310.3840, 140.3330, 240.6490.

1. INTRODUCTION ZrO2, with its unique properties such as the high refractive index, wide optical bandgap and the corresponding good laser damage resistance, low absorption and dispersion in the visible and near-infrared (NIR) spectral regions, as well as the high chemical and thermal stabilities, has proved to be a potential candidate for the highrefractive index film materials in depositing optical mirrors, especially the highly reflective (HR) mirrors, for application in high-power lasers.1–5 Among the various physical and chemical techniques developed to deposit 0740-3224/07/051066-9/$15.00

ZrO2 films,4–9 the soft solgel route exhibits a great number of advantages. The most attractive is its excellent control of the purity and homogeneity of the materials, which greatly diminishes the probability of the defect formation and then defect-induced laser damage of films.1–5,8–11 Besides, the solgel route usually provides films with porous network structures that tend to be more favorable to energy relaxation, thus endowing films with low internal stresses and high laser-induced damage resistance.1–5,10,11 Many excellent studies have been reported on the sol-

© 2007 Optical Society of America

Liang et al.

gel deposition of ZrO2 films for application in high-power lasers.1–3,5,12 Compared with the physical deposition methods,6,13 the solgel route endows ZrO2 films with a much higher laser-induced damage threshold (LIDT) of 45– 50 J / cm2 (at 1064 nm, 5 – 7 ns pulse duration and R/1 testing mode).1–3,5,12 It is a pity that, however, the porous network structure, which is appreciated in improving the LIDT, will lead to a low packing density and then a low refractive index of the films. Furthermore, the solgel ZrO2 xerogel films usually exhibit an organic–inorganic (O/I) hybrid structure that is derived from the chemically controlled hydrolysis and condensation of zirconium necessary to obtain stable sols for film deposition.14–17 Such an O/I hybrid structure will result in another decrease in the refractive index of films. Thus, to prepare ZrO2 films for the common optical application, the xerogel films usually experience a high-temperature densification. Floch et al.18 once predicted, however, that such a heat-treating process would be impracticable in depositing coatings for highpower laser application because it would introduce high internal stresses and greatly diminish the LIDT of the films. Unfortunately, a thorough understanding has not been achieved on how far and how the annealing process will affect the laser-induced damage behavior of the solgel–derived ZrO2 films. In addition, although it is possible to find in current literature a large number of studies about the annealing effect on the structural properties15,19–21 and several papers dealing with the annealing dependence of the refractive index15,21 for solgelderived ZrO2 films, the studies on their optical properties are far from systematic, and the reported structural data show a large scatter probably due to the different synthesis routes. Therefore, the present work is devoted to investigating systematically the annealing effect on the optical properties and laser-induced damage resistance of the solgelderived ZrO2 films and seeking a reasonable correlation between the observed evolutions in structure, optical properties, and laser-induced damage resistance.

2. EXPERIMENT A. Sample Preparation All chemicals, zirconium (IV) n-propoxide [Zr共OPr兲4, Strem chemicals], diethanolamine (DEA, Tianjin chemicals), and anhydrous ethanol (Tianjin chemicals), were of reagent grade and used without further purification. The entire process, including the sol preparation and film deposition, was carried out in a clean and controlled atmosphere with temperature 20 ° C and relative humidity 45%. Stable ZrO2 sol was prepared by the controlled hydrolysis and condensation of Zr共OPr兲4, with DEA as the chelating agent and anhydrous ethanol as solvent. During synthesis, two different but equal parts of alcohol solutions were prepared. In the first part, Zr共OPr兲4 was dissolved into anhydrous ethanol containing DEA. The solution was sealed and kept stirring for 30 min to achieve a complete chelation between the alkoxide and DEA. The second part of the solution was then prepared by mixing the deionized water with anhydrous ethanol. Hydrolysis was carried out under the atmospheric condition by rapidly mixing

Vol. 24, No. 5 / May 2007 / J. Opt. Soc. Am. B

1067

these two solutions at vigorous stirring, which led to the formation of transparent sol. In the total mixture, the initial concentration of Zr共OPr兲4 was 0.28 mol/ L, the molar ratios of DEA to Zr共OPr兲4 and H2O to Zr共OPr兲4 were fixed at 0.7:1 and 2:1, respectively. The fresh sol was hermetically stirred for 60 min and then held under static conditions to ensure the completion of the hydrolysis and condensation. After 20 day aging, the precursor sol was deposited onto the well-cleaned fused-silica substrates by spin coating at 2000– 3000 rpm for 30 min. The as-deposited films were dried at 20° C for 10 h and then underwent a conventional furnace annealing process at different temperatures ranging from 150° C to 750° C for 60 min. In our study, two series of ZrO2 films have been prepared. One (named Series I) was used for determining the structural and optical properties. The other (named Series II) was specially designed for the laser irradiation test, with an optical thickness equal to about one quarter of the irradiation laser wavelength, ␭0 共␭0 = 1053 nm兲. For the sake of clarity, the unannealed and annealed films were named hereafter FX020 and FXY, where X is the series number, I or II, and Y indicates the annealing temperature in °C. In deposition of the films in Series I, two spinning rates of the substrate were used. For samples FI020 and FI150, the spinning rate was 2000 rpm and two layers were deposited. For samples FI350–FI750, the spinning rate was 3000 rpm and eight layers were deposited, which were also the thickest films that could be deposited without cracking. The thickness of every monolayer in samples FI350–FI750 was about 150 nm before annealing and 40– 50 nm after annealing. In deposition of the films in Series II, special attention has been paid to the control of their optical thickness, which was realized by adjusting the spinning rate between 2000 and 3000 rpm and making the first harmonic peak of the transmission spectrum occur at 1053 nm. The samples FII020 and FII150 were single-layer films, whereas samples FII350–FII750 were three-layer films. B. Sample Characterization The transmission spectra of films in Series I were measured using an UV/Vis/NIR spectrometer (Shimadzu, UV3150) over the spectral range of 190– 1500 nm. The optical constants including refractive index n and absorption coefficient ␣ were determined from the transmission spectra based on Swanepoel’s method.22 The dispersion of the calculated refractive index was analyzed on the basis of the Wemple-DiDomenico (WDD) model.23–25 The optical bandgap, Eg was derived from the spectral dependence of the absorption coefficient. The laser irradiation experiments were carried out on the films in Series II with the optics shown in Fig. 1. A Q-tuned Nd:YLF laser (Beijing Jiepu Trend Technology Company) was used to provide a nearly Gaussian-type pulse beam (spatially and temporally) at 1053 nm wavelength. The maximum output energy was 1000 mJ, the pulse duration was 10 ns, and the repeat frequency was 1 Hz. A fixed energy attenuator was installed in the beam path to provide energy control. Two wedges were used to pick off a portion of the beam for a standard set of beam diagnostics that included a calorimeter (energy measure-

1068

J. Opt. Soc. Am. B / Vol. 24, No. 5 / May 2007

Fig. 1. Schematic of the testing optics of LIDT:B, calorimeter; C, photodiode; M1, M2, mirrors.

ment), a photodiode (temporal properties), and a CCD camera (spatial profile). The laser spot on the tested film was adjusted to 0.65 mm2. The R/1 testing procedure was carried out at 30 different locations that were arranged into a 6 ⫻ 5 array. By increasing the irradiation energy from 0 in a step of 0.5 mJ, the films were shot until damage occurred. The damage was estimated from the visual inspection of plasma flash and detected in situ with a Normarski interferential contrast microscope (Nikon, E600W). The crystalline phases of films was examined on an x-ray diffractometer (XRD) (Bruker AXS, D8 Advance). Line traces were collected over 2␪ values ranging from 20° to 70°. Narrow scan analysis was conducted in 2␪ range of 26° to 33°, and the average size of the tetragonal crystallites was calculated using Scherrer’s equation.26 The chemical changes accompanying the annealing process were probed by the Fourier transform infrared (FTIR) spectrometer (Nicolet, Magna-II 550) in the range of 400 to 4000 cm−1. The thermal behavior of the unannealed sample was studied by the thermograviemtric (TG) and differential scanning calorimetry (DSC) analysis (NETZSCH, STA 449C) in air atmosphere with a heating rate of 10° C / min, and the starting weight of the sample was ⬃15 mg. For FTIR and TG analyses, the films were peeled from the substrates with a blade.

Liang et al.

ness; and (5) separate calculation of absorption coefficient in the weak-medium and strong absorption regions. In weak and medium absorption spectral regions, the refractive indices of ZrO2 films were calculated based on the envelopes, TM共␭兲 and Tm共␭兲, and the transmission spectrum of bare substrate, Ts.22,27 For the fused-silica substrate, Ts is almost a constant of 93.1% in the spectral region of 400– 1500 nm, giving a refractive index, s = 1.47. The calculated refractive indices of all the film samples are shown by the scattered dots in Fig. 3. The refractive indices at the He–Ne laser wavelength of 633 nm are collected in Table 1. At 633 nm, the refractive index of ZrO2 xerogel film, FI020, is 1.63, larger than the values of 1.57–1.59 reported in current literature for solgel-derived ZrO2 films.4,29 This could be attributed to the comparatively sufficient hydrolysis and condensation of Zr共OPr兲4 due to the proper selection of reaction system. The sample annealed at 550° C, FI550, shows a refractive index of

Fig. 2. A typical transmission spectrum of ZrO2 films, FI550. TM and Tm denote the upper and lower envelope curves, respectively.

3. RESULTS A. Optical Characterization A typical transmission spectrum of the annealed ZrO2 film is shown in Fig. 2. The presence of the interference fringes is an indication of the thickness uniformity of films.22,27 According to the important simplification proposed by Manifacier et al.28 and adopted by Swanepoel,22 the envelopes that trace the extremes of the interference fringes, TM and Tm can be considered as continuous functions of wavelength, ␭, as shown by the dotted curves in Fig. 2. Following Swanepoel’s method,22 the derivation of the optical constants in our case involves the following procedures: (1) the determination of the envelope curves, TM共␭兲 and Tm共␭兲, in weak and medium absorption regions 共␭ 艌 400 nm兲 by the spline interpolation to the experimental data; (2) calculation of the refractive index in weak and medium absorption regions; (3) dispersion analysis of the calculated refractive index and estimation of refractive index in the strong absorption region 共␭ 艋 400 nm兲 by the extrapolation method; (4) estimation of the film thick-

Fig. 3. Spectral dependence of the refractive indices and a typical plot of the refractive-index factor 共n2 − 1兲−1 versus 共hv兲2 (inset) of ZrO2 films. The scattered dots are determined from the weak and medium absorption regions of the transmission spectra according to the envelope method. The broken curves are obtained by extrapolation based on the WDD dispersion model.

Liang et al.

Vol. 24, No. 5 / May 2007 / J. Opt. Soc. Am. B

1069

Table 1. Structural and Optical Parameters of ZrO2 Filmsa Dispersion Parameters

Sample

n (at 633 nm)

FI020 FI150 FI350 FI450 FI550 FI650 FI750

1.63± 0.01 1.68± 0.01 1.85± 0.01 1.91± 0.01 1.93± 0.01 1.88± 0.01 1.86± 0.01

E0 (eV) 10.30± 0.09 10.15± 0.11 9.84± 0.08 9.62± 0.09 9.71± 0.11 9.51± 0.12 9.47± 0.11

Ed (eV)

d (nm)

Eg (eV)

DT (nm)

16.46± 0.51 17.81± 0.55 22.92± 0.47 24.48± 0.52 25.48± 0.54 23.04± 0.59 22.25± 0.53

358± 6 326± 7 392± 6 356± 6 330± 6 327± 7 332± 6

5.63± 0.05 5.61± 0.05 5.57± 0.05 5.36± 0.05 5.28± 0.05 5.13± 0.05 4.97± 0.05

— — — 18.2± 0.5 19.5± 0.5 25.3± 0.5 25.6± 0.5

a n is the refractive index, E0 is the single-oscillator energy, and Ed is the dispersion energy. d is the physical thickness of the films, and Eg is the optical bandgap. DT denotes the average size of the tetragonal crystallites. The uncertainties of the calculated values for the optical parameters were determined by the least-squares fits on five separate measurements of the transmission spectrum.

1.93, matching the reported value of 1.92 for ZrO2 films annealed at the same conditions.21 In addition, with the increase of annealing temperature, the refractive index first increases and then decreases with the watershed temperature of 550° C. Such a change behavior is very similar to that reported by Liu et al.21 The dispersion of the calculated refractive indices has been analyzed on the basis of the WDD model,23–25 which is derived from the Sellemeier-type dispersion equation and based on the single-oscillator formula n2共h␯兲 = 1 +

E 0E d E02

− 共h␯兲2

,

共1兲

where E0 is the single-oscillator energy and directly related with the optical bandgap, Eg, Ed is the dispersion energy and closely associated with the structural order of the materials, and h is the Planck constant. By plotting 共n2 − 1兲−1 against 共h␯兲2 and fitting a straight line, E0 and Ed can be determined. A typical WDD plot for the sample, FI550, is presented as the inset of Fig. 3. A good linear relationship in the visible spectral region confirms the validity of the WDD model for our films. The calculated dispersion parameters are summarized in Table 1. With the increase in annealing temperature, a decrease in E0 is observed. Ed is found to first increase and then decrease with the watershed temperature of 550° C, indicating first an increase and then a decrease in the structural order of films.23,25 By extrapolating Eq. (1) toward the short wavelength, the refractive indices in the strong absorption region can be estimated.22,24,27 The extrapolated data are also plotted in Fig. 3 as broken curves. To estimate the film thickness, d, a number of thickness, d⬘, were calculated based on the refractive indices at two adjacent extremes of the same type (maximum– maximum or minimum–minimum) in weak and absorption region. As was observed by Swanepoel22 and Bhaskar et al.,30 there was some dispersion in the calculated values of d⬘. In our case, the differences between the values of d⬘ for a certain sample were within 2%. Thus, we estimated the film thickness, d, as the average value of d⬘. The calculated data of d are summarized in Table 1. The absorption coefficient, ␣, was then determined based on the refractive index, n, the upper envelope, TM, and the transmission spectra of films in strong absorption

Fig. 4. Spectral dependence of the absorption coefficients and the determination of optical bandgap from the plots of 共␣hv兲2 versus photon energy, hv for ZrO2 films.

region.22,24,27 The typical wavelength dependences of the absorption coefficient for our films are given in Fig. 4. All curves show a steep rise in absorbance at the absorption edge due to the optical transition.24,27 To determine the nature of the involved optical transition and deduce the gap width, the strong absorption regions in Fig. 4 were analyzed based on the following equation24,27:

␣h␯ = A共h␯ − Eg兲p ,

共2兲

where h␯ is the photon energy, A is a constant about the transition probability, Eg is the optical bandgap, and exponent p depends on the type of the optical transition between the valence and conduction bands (p usually takes the value of 1 / 2 or 2, for the direct or indirect optical transition, respectively). The common method for determining Eg is to plot a curve of 共␣h␯兲1/p versus h␯, which tends asymptotically toward a linear section in the high-photon energy region. We plotted 共␣h␯兲1/p versus h␯ for the investigated samples and found the best fit was obtained for p = 1 / 2 as shown in the inset of Fig. 4, implying that the direct transition is the most probable mechanism responsible for the optical absorption in our case. The optical bandgap, Eg, was estimated by extrapolating the linear portion of the plot 共␣h␯兲2 versus h␯ to 共␣h␯兲2 = 0, as shown in the inset of Fig. 4. The obtained Eg values are in the

1070

J. Opt. Soc. Am. B / Vol. 24, No. 5 / May 2007

Liang et al.

range of 4.97– 5.63 eV (as collected in Table 1), in agreement with the literature values of 5.1– 5.4 eV (Ref. 15) for solgel-derived ZrO2 films. B. Laser Irradiation Test The laser-induced damage resistance of all the films in Series II was measured in R/1 mode. The representative morphologies of the damaged areas are shown in Fig. 5. For the unannealed film, FII020, a large circular damaged spot with a diameter of 700– 800 ␮m can be observed [Fig. 5(a)], which gradually merges into the undamaged area with a blurry boundary. In the damaged area there are several annular regions, displaying a ra-

Fig. 6. LIDT values of ZrO2 films (at 1053 nm wavelength, 10 ns pulse duration, and R/1 mode). (a) FII020, (b) FII150, (c) FII350, (d) FII450, (e) FII550, (f) FII650, (g) FII750. (Uncertainty of the measurement of LIDT is estimated to be ±1 J / cm2.)

dial expansion of laser energy and the different damaged extent along the film surface. To clearly exhibit the annular feature, the damaged spot was covered with a layer of vapor when the photograph was taken, as shown by the left side of Fig. 5(a). Different from the damage morphology of sample FII020, the damaged regions on the annealed samples are confined to a much smaller area [Figs. 5(b) and 5(c)]. Moreover, the damage appears more catastrophic and the boundary between damaged and undamaged areas becomes much sharper, suggesting either a different damage mechanism was occurring or a similar one was causing damage over films with different microstructures. Figure 6 shows the LIDT values of all the films, which correspond to the highest energies at which no damage occurred. Among all the investigated samples, the unannealed sample exhibits the highest LIDT of 55 J / cm2 (at 1053 nm, 10 ns pulse duration), matching the literature values of 45– 50 J / cm2 (at 1064 nm, 5 – 7 ns pulse duration) for the solgel ZrO2 films.1–3,5,12 With increasing annealing temperature from 20° C to 750° C, a monotonic decrease of LIDT from 55 to 10 J / cm2 is found, and a steep decrease occurs after the film is annealed at 350° C.

Fig. 5. Representative morphologies of the damaged region obtained by a Normarski microscope. (a) FII020, (b) FII350, (c) FII550.

C. Structural Characterization Figure 7 displays the XRD patterns of the films in Series I. All samples show a broad hump at ⬃22° due to the diffraction from the amorphous silica substrates. Samples FI020 and FI350 show no obvious diffraction peaks, indicating an amorphous nature of the films. After being annealed at 450° C – 650° C, the films have only tetragonal 共t兲 phase, whereas annealing above 750° C results in a mixture of the monoclinic 共m兲 and tetragonal phases. The average size of the tetragonal crystallites, DT, was estimated to be 18.2– 25.6 nm as collected in Table 1. Our XRD analysis indicated that amorphous ZrO2 began to crystallize into the metastable tetragonal phase at 450° C, similar to the values reported by Ehrhart et al.15 and Liu et al.21 However, the temperature at which the t → m phase transformation occurs, is higher than the literature

Liang et al.

Fig. 7. XRD patterns of ZrO2 films annealed at various temperatures.

Vol. 24, No. 5 / May 2007 / J. Opt. Soc. Am. B

1071

due to C–N, Zr–O–C, and Zr–OC vibrations disappear [Fig. 8(b)], implying that annealing at 350° C results in a nearly complete removal of the organic residues and decomposition of the O/I hybrid materials. Annealing at temperatures higher than 350° C gives rise to only a gradual loss of the hydroxyl groups. Figure 9 presents the thermal analysis results of the xerogel sample that has been scraped from the unannealed films. On the TG curve, four weight loss stages can be distinguished at ⬃30° C – 150° C, 250° C – 350° C, 350° C – 500° C, and 650° C – 730° C, corresponding to one endothermic and three exothermal peaks on the DSC curve. The peak at 30° C – 150° C should correspond to the desorption of physically adsorbed water and solvent, the peak at 250° C – 350° C could be caused by the oxidative decomposition of the organic segments in hybrid film materials, the peak at 350° C – 500° C must be caused by both the Zr–OH decomposition and crystallization, and the peak at 650° C – 730° C might be attributed to the combustion of the accumulated carbonaceous residues on the surface of ZrO2 particles.17 Besides, combined with XRD analysis results, we conjectured that the t → m phase transformation of the film materials must be partially responsible for the broad exothermal peak at ⬃700° C on the DSC curve, though it may have no contribution on the weight loss.

4. DISCUSSION

Fig. 8. FTIR spectra of ZrO2 powders scraped from the corresponding films. (a) FII020, (b) FII350, (c) FII450, (d) FII550, (e) FII750.

values. This deviation might be due to the different amorphous structures derived from the different synthesis routes. Figure 8 comparatively displays the FTIR spectra of powder samples that have been scraped from the films. FTIR spectra indicate the O–H stretching vibration at ⬃3440 cm−1, C–H and N–H stretching vibrations at 2929 and 2867 cm−1, O–H bending vibration in the adsorbed and/or coordinated water at 1640 cm−1, bending vibrations of the bridging and nonbridging OH groups at 1547 and 1400 cm−1, C–H and N–H deformation vibrations in the region of 1400– 1200 cm−1, C–N, Zr–O–C and Zr–OC stretching vibrations at 1103, 1072, and 926 cm−1, as well as Zr–OH and Zr–O–Zr stretching vibrations at ⬃618 and 470 cm−1.16,17,31–34 For the unannealed sample [Fig. 8(a)], the strong absorptions of O–H, C–H, N–H, and C–O vibrations suggest the presence of the organic residues, and absorptions of Zr–O–C and Zr–OC vibrations show evidence of the chelating agents anchored in the inorganic networks. After annealing at 350° C, the absorption lines

It is generally acknowledged that the chemically controlled hydrolysis and condensation of zirconium alkoxides usually involve three principal reaction steps: the nucleophilic substitution of the alkoxide groups by chelating ligands, hydrolysis of the remaining alkoxide groups, and condensation to form Zr–O and Zr–O–Zr backbones.33 The controlled hydrolysis and condensation produces a large number of colloidal particles, which connect to each other to form the network structure of the precursor sols and then xerogel films. Compared with the alkoxide groups, the chelating ligands are more stable against hydrolysis. Thus, most of them might become anchored to a zirconium oxopolymeric backbone and form O/I networks.33 The presence of DEA segments in the unannealed films

Fig. 9. TG/DSC curves of the xerogel powder scraped from the unannealed ZrO2 film.

1072

J. Opt. Soc. Am. B / Vol. 24, No. 5 / May 2007

has been confirmed by the strong FTIR absorption lines at 1103, 1072, and 926 cm−1 [Fig. 8(a)] and the peak at 250° C – 350° C on the TG/DSC curves (Fig. 9). It is the porous O/I hybrid structure that makes the unannealed film exhibiting a refractive index of 1.63, much lower than that of bulk ZrO2 [⬃2.17 (Ref. 15)]. Annealing at 150° C leads to the evaporation of physically adsorbed water and solvent, resulting in a slight increase of refractive index from 1.63 to 1.68. After annealing at 350° C, a nearly complete removal of the organic segments is achieved. The purification and densification of the films give rise to a rapid increase of refractive index, n, from 1.68 to 1.85. With a further increase of the annealing temperature, n increases from 1.85 to 1.93 between 350° C and 550° C, but decreases from 1.93 to 1.86 between 550° C and 750° C. According to the XRD analysis results (Fig. 7), the continuous increase of the refractive index between 350° C and 550° C can be ascribed to the crystallization of amorphous ZrO2, which results in a further densification of the films. After being annealed at 650° C, the film is still composed of t-ZrO2 from the XRD analysis, and our experiment showed that the obvious peaks of m-ZrO2 could be observed only after film was annealed at 750° C. However, Li et al.35 found that the t → m phase transformation of ZrO2 usually initiated at the surface region of the particles or crystallites. They also demonstrated that such a process was only detectable for the surface-sensitive UV Raman spectroscopy, but was unperceivable for the XRD analysis that mainly reflected the bulk properties of materials. Thus, we conjectured that in sample FI650, there might be a thin layer of m-ZrO2 at the surface of the t-ZrO2 particles that composed of the film structure, but they were undetectable for XRD analysis. In other words, we considered that the t → m phase transformation of ZrO2 might take place at temperature lower than 650° C. Based on this conjecture, we assigned the decrease in refractive index between 550° C and 750° C to the decrease in packing density of the films, which might occur as the consequence of both volume expansion accompanying the phase transformation and increase of porosity due to the crystallites’ growth (as shown in Table 1). The above-mentioned annealing dependence of the refractive index can also be explained from the viewpoint of the structural order of films. As mentioned earlier, the dispersion energy, Ed, can be directly associated with the structural order of the investigated material. By comparing the evolutions of refractive index, n, and structural ordering parameter, Ed, as shown in Table 1, one can find that both variables exhibit the similar changing tendency, indicating a one-to-one correspondence between the evolutions in refractive index and microstructure order of the films. Apart from refractive index, n, and dispersion parameters, E0 and Ed, we have deduced another two important optical constants, namely, the absorption coefficient, ␣, and optical bandgap, Eg. Here, we emphasize the discussion particularly on Eg. It is clear from Table 1 that Eg of the studied films decreases from 5.63 to 4.97 eV with increasing annealing temperature from 20° C to 750° C. Such changing behavior could be related to evolutions in both the microstructure and chemical composition of

Liang et al.

films. Pamu et al.36 reported that in the case of TiO2 films, Eg tends to increase for films composed of smaller crystallite size and/or containing amorphous phase. Similar results have been observed by Tigau et al.27 and Bhaskar et al.30 in their separate investigations on Sb2O3 and Pb1−xLaxTi1−x/4O3 films. Therefore, we partially ascribed the decrease of Eg to the gradually improved crystallinity and increased crystallites size of the films (as shown in Fig. 7 and Table 1). Moreover, Jana et al.37 have detected, in their investigation on the air-annealed solgel-derived ZrO2 films, the presence of a small amount of oxygen vacancy and trivalent zirconium that led to a decrease of Eg. In our case, the films were also annealed in air atmosphere. Hence, it is possible that in our annealed films, an amount of oxygen vacancy and trivalent zirconium was also present, which becomes another factor that causes the decrease of Eg. However, it is noticeable that the amount of the nonstoichiometry defects in our films must be very small, since the combined effect of increased crystallinity and nonstoichiometry defects only results in a small difference of about 0.6 eV between the Eg values of amorphous film and film annealed at 750° C. The significance of the derivation of ␣ and Eg lies in the fact that these two parameters can reflect the absorption characteristics of materials in the case of light irradiation, especially high-power laser irradiation. Therefore, they are closely related to the laser-induced damage behavior of the materials. Figure 4 clearly shows that at 1053 nm, the absorption coefficient, ␣, is approximately zero, indicating a negligible linear absorption for all the investigated films. Thus, in the case of high-power laser irradiation, the damage to films must be initiated by the massive nonlinear absorption of film material such as the multiphoton absorption and the absorption resulted by the localized absorbing defects. A large number of researchers38–41 have demonstrated that the nonlinear absorption and LIDT of the materials strongly depended on its apparent optical bandgap, Eg. Although the complexity of the laser-induced damage process makes it too difficult to establish a quantitative correlation between them, there is still a generally accepted viewpoint that material with larger Eg usually exhibits weaker nonlinear absorption and then a relatively higher LIDT value. Such a viewpoint has been demonstrated again in our present work. It is obvious from Table 1 and Fig. 6 that the LIDT of films decreases following the decrease of its optical bandgap, Eg. However, in our case, the variation in Eg is too small to be responsible for the comparatively large decrease in LIDT on its own, especially the sharp decrease in the low-temperature region. Combining with the morphology evolution of the damaged spots (Fig. 5), we conjectured that the structure properties, especially mechanical property, must have a considerable effect on the film endurance to the irradiated laser energy, and thus might be another crucial factor determining the laser-induced damage behavior of films. Based on the combined effect of the optical bandgap and structural properties, a possible explanation of the annealing dependence of laser-induced damage resistance of our studied films can be described as follows. Figures 5 and 6 clearly show that the unannealed sample, FII020, exhibits much better laser-induced damage resistance with a LIDT of

Liang et al.

55 J / cm2 and a larger but less catastrophic damaged spot than the annealed films. This must benefit from its large optical bandgap and network structure that permits a sufficiently fast energy relaxation within the films. Annealing at 150° C can induce tensile stress in the film,42,43 which diminishes, to a certain extent, the strength of the film and then accounts for the large decrease of LIDT from 55 to 44 J / cm2. After annealing at 350° C, the porous network structure of the xerogel film is wholly destroyed, and simultaneously a rapid densification and a further increase in the tensile stress of films occurs. The more dense structure is suspected to exhibit a longer energy relaxation time.2,4,5 Therefore, after such a densification, the absorbed laser energy must be restricted within a much smaller region on the tested films. It is the combined effect of the concentrated laser energy and weakened film mechanical strength that results in the sharp decrease of LIDT from 44 to 27 J / cm2. The convincing evidence can be found in the morphology evolution of the damaged spots, because although the lower irradiation energy corresponding to the lower LIDT may also give rise to a smaller damage spot, the observed more catastrophic damage and distinct boundary of the damaged spot for the annealed films can only be explained be the more concentrated energy and less durable film structure. With a further increase in the annealing temperature to 750° C, the monotonous decrease in Eg and the gradually increased internal stress lead to a further decrease in LIDT of the films. In addition, the increased amount of absorbing and structural defects will be partially responsible for the monotonous decrease in LIDT during the entire temperature region. It is worth noting, however, that the films annealed at a temperature below 550° C still show a high LIDT comparable to that of the films prepared by physical deposition. Hereto, we have given a systematic investigation on the annealing dependence of the optical performance of solgel-derived ZrO2 films. Among all the examined films, the xerogel film, FII020, exhibits the best laser damage resistance but a comparatively low refractive index, whereas the annealed film, FII550, displays the highest refractive index but only an acceptable laser-induced damage resistance. Based on these two typical high-index layers, two HR mirrors, substrate/ FII020/ 共SiO2 / FII020兲n (HR1) and substrate/ FII550/ 共SiO2 / FII550兲n (HR2), have been deposited. It is found that, benefiting from the lowtemperature deposition process of HR1, a perfect stress match between the high- and low-index layers enables the successful deposition of a nearly full-reflection mirror (n equal to 10), with the minimum transmittance of 1.6% (at 1053 nm) and high LIDT of ⬃39 J / cm2, whereas, for HR2, only a three-layer mirror (n equal to 1) can be obtained due to stress incompatibility between the high- and lowindex layers. Unfortunately, when the porous ZrO2 xerogel film is really applied in the high-power lasers, its O/I hybrid composition and porous structure can be troublesome problems, since they may greatly diminish the performance stability of the HR mirrors due to the possible decomposition of the hybrid backbones and inevitable adsorption effect under the service conditions. The ill effects of the porous structure on optical performance have been found in our practical application of the solgel SiO2 anti-

Vol. 24, No. 5 / May 2007 / J. Opt. Soc. Am. B

1073

reflective mirrors. Therefore, to make full use of the advantages of the solgel route in depositing the multilayer HR mirrors for high-power laser applications, a special annealing process for the high-index ZrO2 layers should be developed, and in some sense, an instantaneous posttreatment may be a practicable way.

5. CONCLUSION Based on a detailed characterization on the optical properties, laser-induced damage resistance, as well as structural characteristics, the annealing dependence of the optical performance for solgel-derived ZrO2 films has been studied. It is found that, due to the porous O/I network structure, the unannealed film exhibits the best laserinduced damage resistance but a low refractive index, whereas the film annealed at 550° C shows the highest refractive index but only acceptable laser-induced damage resistance derived from its rigid structure. The present study will help to establish a better understanding of how and how far the annealing process will affect the optical performance of the solgel-derived ZrO2. However, when the solgel-derived ZrO2 film is really applied in the HR mirrors for high-power-laser application, a special annealing process must be developed. And this will be the focus of our further investigation.

ACKNOWLEDGMENTS The financial support from the National Key Native Science Foundation, China (grant 20133040) was gratefully acknowledged. Y. Xu is the corresponding author and can be reached at [email protected].

REFERENCES 1.

2.

3.

4.

5.

6.

7.

I. Thomas, “Preparation of dielectric HR mirrors from colloid oxide suspensions containing organic polymer binders,” in Sol-Gel Optics III, J. D. Mackenzie, ed., Proc. SPIE 2288, 50–55 (1994). Ph. Belleville, C. Bonnin, and J. J. Priotton, “Roomtemperature mirror preparation using sol-gel chemistry and laminar-flow coating technique,” J. Sol-Gel Sci. Technol. 19, 223–226 (2000). Q. Y. Zhang, J. Shen, J. Wang, G. M. Wu, and L. Y. Chen, “Sol-gel derived ZrO2 – SiO2 highly reflective mirrors,” Int. J. Inorg. Mater. 2, 319–323 (2000). D. Grosso and P. A. Seromon, “Scandia optical mirrors for application at 351 nm,” Thin Solid Films 368, 116–124 (2000). Ph. Belleville, Ph. Prené, C. Bonnin, L. Beaurain, Y. Montouillout, and É. Lavastre, “How smooth chemistry allows high power laser optical coating preparation,” in Advances in Optical Thin Films, C. Amra, N. Kaiser, and H. A. Macleod, eds. Proc. SPIE 5250, 196–202 (2004). G. L. Tian, J. B. Huang, T. Wang, H. B. He, and J. D. Shao, “Microstructure and laser-damage threshold of ZrO2 mirrors dependence on annealing temperature,” Appl. Surf. Sci. 239, 201–208 (2005). D. W. Zhang, S. H. Fan, Y. A. Zhao, W. D. Gao, J. D. Shao, R. Y. Fan, Y. J. Wang, and Z. X. Fan, “High laser-induced damage threshold HfO2 films prepared by ion-assisted electron beam evaporation,” Appl. Surf. Sci. 243, 232–237 (2005).

1074 8.

9.

10.

11.

12.

13.

14. 15.

16.

17.

18. 19. 20.

21.

22. 23. 24.

25.

26.

J. Opt. Soc. Am. B / Vol. 24, No. 5 / May 2007 M. Reichling, A. Bodemann, and N. Kaiser, “Defect induced laser damage in oxide multilayer mirrors for 248 nm,” Thin Solid Films 320, 264–279 (1998). Y. A. Zhao, W. D. Gao, J. D. Shao, and Z. X. Fan, “Roles of absorbing defects and structural defects in multilayer under single-shot and multi-shot laser radiation,” Appl. Surf. Sci. 227, 275–281 (2004). Y. Xu, B. Zhang, W. H. Fan, D. Wu, and Y. H. Sun, “Sol-gel broadband anti-reflective single-layer silica films with high laser damage threshold,” Thin Solid Films 440, 180–183 (2003). Y. Xu, L. Zhang, D. Wu, Y. H. Sun, Z. X. Huang, X. D. Jiang, X. F. Wei, Z. H. Li, B. Z. Dong, and Z. H. Wu, “Durable sol-gel antireflective films with high laser-induced damage thresholds for inertial confinement fusion,” J. Opt. Soc. Am. B 22, 905–912 (2005). M. C. Ferrara, M. R. Perrone, M. L. Protopapa, J. SanchoParramon, S. Bosch, and S. Mazzarelli, “High mechanical damage resistant sol-gel mirror for high power lasers,” in Advances in Optical Thin Films, C. Amra, N. Kaiser, and H. A. Macleod, eds., Proc. SPIE 5250, 537–545 (2004). Y. A. Zhao, T. Wang, D. P. Zhang, S. H. Fan, J. D. Shao, and Z. X. Fan, “Laser conditioning of ZrO2 : Y2O3 / SiO2 mirror coatings prepared by E-beam evaporation,” Appl. Surf. Sci. 239, 171–175 (2005). G. S. Kathryn and S. L. Jeffrey, “Characterization of titanium and zirconium valerate sol-gel films,” Chem. Mater. 6, 890–898 (1994). G. Ehrhart, B. Capoen, O. Robbe, Ph. Boy, S. Turrell, and M. Bouazaoui, “Structural and optical properties of n-propoxide sol-gel derived ZrO2 thin films,” Thin Solid Films 496, 227–233 (2006). J. Méndez-Vivar, R. Mendoze-Serna, and L. Valdez-Castro, “Control of the polymerization process of multicomponent (Si, Ti, Zr) sols using chelating agents,” J. Non-Cryst. Solids 288, 200–209 (2001). J. P. Zhao, W. H. Fan, D. Wu, and Y. H. Sun, “Synthesis of highly stabilized zirconia sols from zirconium n-propoxide-diglycol system,” J. Non-Cryst. Solids 261, 15–20 (2000). H. G. Floch, J. P. Priotton, and I. M. Thomas, “Optical coatings prepared from colloidal media,” Thin Solid Films 175, 173–178 (1989). W. M. Liu, Y. X. Chen, C. F. Ye, and P. Y. Zhang, “Preparation and characterization of doped sol-gel zirconia films,” Ceram. Int. 28, 349–354 (2002). A. Hartridge, M. Ghanashyam Krishna, and A. K. Bhattacharya, “Temperature and ionic size dependence on the structure and optical properties of nanocrystalline lanthanide doped zirconia thin films,” Thin Solid Films 384, 254–260 (2001). W. C. Liu, D. Wu, A. D. Li, H. Q. Ling, Y. F. Tang, and N. B. Ming, “Annealing and doping effects on structure and optical properties of sol-gel derived ZrO2 thin films,” Appl. Surf. Sci. 191, 181–187 (2002). R. Swanepoel, “Determination of the thickness and optical constants of amorphous silicon,” J. Phys. E 16, 1214–1222 (1983). S. H. Wemple and M. DiDomenico, Jr., “Behavior of the electronic dielectric constant in covalent and ionic materials,” Phys. Rev. B 3, 1338–1351 (1971). J. M. González-Leal, R. Prieto-Alcón, J. A. Angel, and E. Márquez, “Optical properties of thermally evaporated amorphous As40S60−xSex films,” J. Non-Cryst. Solids 315, 134–143 (2003). N. K. Sahoo, S. Thakur, and R. B. Tokas, “Superior refractive index tailoring properties in composite ZrO2 / SiO2 thin film systems achieved through reactive electron beam codeposition process,” Appl. Surf. Sci. 253, 618–626 (2006). S. Shukla and S. Seal, “Thermodynamic tetragonal phase

Liang et al.

27.

28.

29.

30.

31. 32.

33. 34.

35.

36.

37. 38.

39.

40.

41.

42.

43.

stability in sol-gel derived nanodomains of pure zirconia,” J. Phys. Chem. B 108, 3395–3399 (2004). N. Tigau, V. Ciupina, and G. Prodan, “The effect of substrate temperature on the optical properties of polycrystalline Sb2O3 thin films,” J. Cryst. Growth 277, 529–535 (2005). J. C. Manifacier, J. Gasiot, and J. P. Fillard, “A simple method for the determination of the optical constants n , k and the thickness of a weakly absorbing thin film,” J. Phys. E 9, 1002–1004 (1976). L. F. Cueto, E. Sánchez, L. M. Torres-Martínez, and G. A. Hirata, “On the optical, structural, and morphological properties of ZrO2 and TiO2 dip-coated thin films supported on glass substrates,” Mater. Charact. 55, 263–271 (2005). S. Bhaskar, S. B. Majumder, M. Jain, P. S. Dobal, and R. S. Katiyar, “Studies on the structural, microstructural and optical properties of sol-gel derived lead lanthanum titanate thin films,” Mater. Sci. Eng., B 87, 178–190 (2001). Y. Sorek, M. Zevin, and R. Reisfeld, “Zirconia and zirconiaORMOSIL planar waveguides prepared at room temperature,” Chem. Mater. 9, 670–676 (1997). M. K. Mishra, B. Tyagi, and R. V. Jasra, “Effect of synthetic parameters on structure, textural, and catalytic properties of nanocrystalline sulfated zirconia prepared by sol-gel technique,” Ind. Eng. Chem. Res. 42, 5727–5736 (2003). K. G. Severin, J. S. Ledford, B. A. Torgerson, and K. A. Berglund, “Characterization of titanium and zirconium valerate sol-gel films,” Chem. Mater. 6, 890–898 (1994). Y. X. Hao, J. S. Li, X. J. Yang, X. Wang, and L. D. Lu, “Preparation of ZrO2 – Al2O3 composite membrane by solgel process and their characterization,” Mater. Sci. Eng., A 367, 243–247 (2004). M. J. Li, Z. C. Feng, G. Xiong, P. L. Ying, Q. Xin, and C. Li, “Phase transformation in the surface region of zirconia detected by UV Raman spectroscopy,” J. Phys. Chem. B 105, 8107–8111 (2001). D. Pamu, M. Ghanashyam Krishna, K. C. James Raju, and A. K. Bhatnagar, “Ambient temperature growth of nanocrystalline titanium dioxide thin films,” Solid State Commun. 135, 7–10 (2005). S. Jana and P. K. Biswas, “Characterization of oxygen deficiency and trivalent zirconium in sol-gel derived zirconia films,” Mater. Lett. 30, 53–58 (1997). V. N. Strekalov, “Absorption of laser light by ions as a mechanism of optical damage in solids,” in Laser-Induced Damage in Optical Materials, G. Exarhos, A. Guenther, M. Kozlowski, K. L. Lewis, and M. J. Soileau, eds., Proc. SPIE 3244, 26–30 (1997). V. N. Strekalov, “The path integrals method and description of valence band impact ionization under multiphoton heating of electrons,” in Laser-Induced Damage in Optical Materials, G. Exarhos, A. Guenther, M. Kozlowski, K. L. Lewis, and M. J. Soileau, eds., Proc. SPIE 3578, 256–263 (1998). C. W. Carr, H. B. Radousky, and S. G. Demos, “Experimental study of wavelength dependent damage threshold in DKDP,” in Laser-Induced Damage in Optical Materials, G. Exarhos, A. Guenther, N. Kaiser, K. L. Lewis, M. J. Soileau, C. J. Stolz, A. Giesen, and H. Weber, eds., Proc. SPIE 4932, 385–389 (2003). D. Li, Z. Q. Zhu, X. Y. Fu, and F. M. Qiu, “Relations between the laser induced damage to optical thin film and the material band-gap,” Opto-Electron. Eng. 26, 58–62 (1999). R. Brenier and A. Gagnaire, “Densification and aging of ZrO2 films prepared by sol-gel,” Thin Solid Films 392, 142–148 (2001). H. K. Pulker and S. Schlichtherle, “Density related properties of metal oxide films,” in Advances in Optical Thin Films, C. Amra and N. Kaiser, eds., Proc. SPIE 5250, 1–11 (2004).