The MZTFs were prepared by a method combining the procedures reported by Zhao^{[ 18]} and Sanchez^{[ 10, 11, 12]}. Firstly, the surfactant F127 were dissolved in anhydrous ethanol (EtOH) solution stirring for 30 min. Secondly, the zirconium (IV) chloride (98% anhydrous) were added into the above surfactant-ethanol solution with stirring, quickly, the solution turned into colorless and transparent. Thirdly, zircon (IV) propoxide (70wt% solution in 1-propanol) were added into the second-step solution with stirring further for 2 h at 26℃ in a closed glass vessel. The composition ratio of the ultimate prepared solution was: n(ZrCl₄): n(Zr(OC₃H₇)₄): n(F127): n(EtOH)=0.7: 0.3: 0.01: (30-40). Then, the mesoporous zirconia thin films were prepared by dip-coating in a chamber with relative humidity (RH) of 55% and temperature of 26℃. The clean glass substrate was dipped into the ultimate solution (colorless and transparent) and then withdrew at a rate of 1-5 mm/s. The thin films turned into opaque as soon as the substrate left the zirconia precursor solution. Then, the films turned into transparent again by increasing the RH of the surroundings. Repeat this process twice. The ultimate transparent surfactant-zirconia thin films were aged in the chamber at 26℃ and 55% RH for at least 24 h. The surfactant templates were removed by calcination at 350℃ for 3 h in air.

The small-angle X-ray diffraction (SAXRD) patterns of the films were conducted in Rigaku Rotaflex diffractometer equipped with a rotating anode and using CuKα radiation. The transmission electron microscope (TEM) images, selective area electron diffraction (ED) pattern and simultaneous energy-dispersive spectroscope (EDS) spectrum were executed in a JEM-2100F field emission TEM with an accelerating voltage of 200 kV. UV-Vis absorption spectra were measured on Shimadzu UV-3101. The thicknesses and linear refractive indexes of the films were measured to be (233±10) nm and 1.92 at 1064 nm by SGC-10 film thickness measurement apparatus. The transmittances of the films were measured to be 81.1% by dividing the transmitted laser energy with the incident laser energy. A mode-locked Nd:YAG laser was used in the Z-scan experiment. The laser focused to a beam waist of 35 μm at the focus then went through an aperture with an aperture factor of 0.0192. The power of the incident laser was 1.26 × 10⁶ (±5%) W.

Figure 1 shows the SAXRD patterns of the air-dried and templates removed MZTF. The two diffraction peaks at 2 θ = 0.84^o and 1.35^oare indexed as (110) and (200) reflections of the cubic-phased mesoporous structure^{[ 19]} which indicates an ordered mesoporous structure of the prepared zirconia thin films, as seen in Fig. 1. But the periodic ordering of the thin films is partly lost after the surfactant templates are removed by calcination at 350℃ for 3 h in the air, as the inset in Fig. 1 shown. From the inset SAXRD pattern, we can see that the (110) reflection is not distinct as that of the air-dried one and moved towards a higher 2 θ region of 0.85^o. This may be due to the further contraction and copolymerization of the mesoporous zirconia framework by calcination.

Fig. 1

	Figure Option ViewDownloadNew Window
	Fig. 1 SAXRD pattern of the air-dried MZTF, and the inset is the template-removed MZTF

Figure 2 shows the TEM images of the air-dried MZTF and the gold NPs incorporated MZTF. From Fig. 2(a), the ordered arrangement of the cubic-phased mesopores is observed in the air-dried MZTF, as indicated by the two white circles in the image. The circled areas can be indexed as (111) and (100) planes according to the geometry of the cubic-phased mesoporous structure. The result is consistent with the above SAXRD analysis. From Fig. 2(b), we can see the gold NPs (the black dots in the image) highly disperse within the MZTF with an average diameter of 6 nm mainly due to the confinement effect of the mesoporous channels^{[ 7, 9]}. The existence of the gold NPs and a semi-quantitative gold content (1.2at%) of the composite thin films can be confirmed and measured by simultaneous EDS spectrum, as seen in Fig. 3, which shows the distinctive signal peaks of the elements Au and Zr.

Fig. 2

	Figure Option ViewDownloadNew Window
	Fig. 2 TEM images of (a) air-dried and (b) gold NPs incorporated MZTF

Fig. 3

	Figure Option ViewDownloadNew Window
	Fig. 3 EDS spectra of the gold NPs incorporated MZTF

Figure 4 shows the UV-Vis absorption spectra of the blank MZTF and the gold NPs incorporated MZTF. The blank MZTF did not appear any absorption in the detection region from 350 to 800 nm in the UV-Vis absorption spectra, as seen curve (a) in Fig. 4. While the gold NPs incorporated MZTF presented a distinct absorption peak at 560 nm in the UV-Vis absorption spectra, which was corresponded to the characteristic surface plasmon resonance (SPR) absorption of the gold NPs, as curve (b) in Fig. 4 shown. The SPR absorption peak was apparently red-shifted compared with that (at around 541 nm) of gold NPs incorporated mesoporous silica thin films^{[ 7]} as the dielectric constant of the zirconia matrix is higher than that of the silica matrix^{[ 20]}. The increased background absorption of the gold NPs incorporated MZTF originated from the inter-band transition of valence electrons of the gold NPs from dband to the Fermi surface^{[ 21]}.

Fig. 4

	Figure Option ViewDownloadNew Window
	Fig. 4 UV-Vis absorption spectra of (a) blank and (b) gold NPs incorporated MZTF

The third-order optical nonlinearity of the gold NPs incorporated MZTF was measured by Z-scan technique using a 1064 nm incident laser. Both the nonlinear optical (NLO) refraction and absorption of the samples can be measured by Z-scan technique^{[ 22]}.

Figure 5 shows the closed aperture (CA) and open aperture (OA) Z-scan outputs of the gold NPs incorporated MZTF. The CA Z-scan output presented a symmetrical peak followed by a valley in the normalized transmittance trace which indicates a negative NLO refractive index and a self-defocusing NLO process. Neither saturation nor anti-saturation absorption appeared in the OA Z-scan out- put, indicating that the composite thin films did not have distinctive NLO absorptions, as seen Fig. 5(b). The experimental phenomenon is consistent with that of our previously investigated gold-silica composite thin films mainly due to the off-resonant measurement at 1064 nm^{[ 23]}.

Fig. 5

	Figure Option ViewDownloadNew Window
	Fig. 5 (a) CA and (b) OA Z-scan outputs of the gold NPs incorporated MZTF

The absolute values of the NLO refractive index ( n₂) and the χ⁽³⁾ of the gold NPs incorporated MZTF were calculated to be 3.47×10^-9 and 7.08×10^-10esu according the method reported in literature^{[ 23]}. The χ⁽³⁾ of the present material is significantly higher than that ( χ⁽³⁾ ≈ 5.62×10^-11 esu) of the highly dispersed gold NPs incorporated mesoporous silica thin films measured by a time-resolved optical Kerr effect at 800 nm^{[ 7]} and comparable with that ( χ⁽³⁾ ≈ 8.23×10^-10 esu) of the gold NPs incorporated mesoporous silica thin films with an ultrahigh gold content of 30wt% measured by Z-scan experiment at 1064 nm^{[ 23]}. The relative high χ⁽³⁾ of the gold NPs incorporated MZTF could be attributed to the higher linear refractive index of zirconia matrix (2.1) than that of silica (1.4). So the enhanced off-resonant χ⁽³⁾ of the gold NPs incorporated MZTF may be assigned to the higher linear refractive index of zirconia matrix and the high dispersion of the gold NPs. And the optical nonlinearity of the gold-zirconia composite thin films was believed to originate from the intra-band transition of electrons near the Fermi surface in the incorporated gold NPs^{[ 7]}.

Highly dispersed gold NPs incorporated mesoporous zirconia thin films were successfully synthesized by an easy deposition-precipitation method. The third-order optical nonlinearities of the composite thin films were measured by Z-scan technique at 1064 nm. The absolute values of optical nonlinear refractive index and third-order optical nonlinear susceptibility of the composite thin films were calculated to be 3.47×10^-9 and 7.08×10^-10esu. The enhanced third-order optical nonlinearity of the material could be attributed to the high linear refractive index of the zirconia matrix and the high dispersion of the gold NPs.