May 1, 2010 / Vol. 35, No. 9 / OPTICS LETTERS 1449

Modified optical properties in a samarium dopedtitania inverse opal

Wei Wang, Hongwei Song,* Qiong Liu, Xue Bai, Yu Wang, and Biao DongState Key Laboratory on Integrated Optoelectronics, College of Electronic Science and Engineering, Jilin University,

Changchun 130012, China*Corresponding author: hwsong2005@yahoo.com.cn

Received February 1, 2010; accepted March 10, 2010;posted March 19, 2010 (Doc. ID 123537); published April 29, 2010

The TiO2:Sm3+ inverse opal has been prepared by a template-assisted method, and its modification onphotoluminescent (PL) properties of Sm3+ is studied. Transmittance spectra show a dip within the photonicbandgap, whose location shifts a little to red and then to blue with an increasing detection angle between theincident light and the surface normal; such an irregular shift is attributed to uneven shrinkage of lattices.Steady-state PL spectra indicate a spatial redistribution of emission lines. Decay kinetics demonstrate thatthe spontaneous emission rates of Sm3+ ions in the inverse opal decrease to varied extent depending on theemission wavelength and detection angle. All the facts are examined to be consistent with the photonicbandgap effect. © 2010 Optical Society of America

OCIS codes: 260.3800, 160.3918.

Ever since the birth of the concept of photonic crystal(PHC) [1], considerable amount of investigation, boththeoretical and experimental, has been carried outtoward its application [2]. The basic property of thePHC is that the density of the photon states (DOS)could be manipulated by changing the PHC struc-tural parameters. The rate of spontaneous emission,being proportional to the propagating DOS, thus un-dergoes a modification when the emitter molecule isincorporated in a PHC structure [3]. Present utiliza-tion of the PHC structure is mainly the manipulationof the traveling direction of light, namely, PHC fibers[4]. Other possible fields, in which the PHC could findan immediate use, include photosensitization [5] andcolor purification [6,7]. Previous studies mainly fo-cused on luminescent species with broadband emis-sion including semiconductor quantum dots, dye mol-ecules, and rare earth (RE) elements with 4f–5dtransitions [2,6]. Photonic effect on narrow-emissioned 4f–4f RE elements is to be more thor-oughly studied, despite a few works reported [7–10].

Titania nanocrystals (NCs) have certain enticingproperties, which include low phonon energy��700 cm−1�, wide bandgap, high refractive index,and biocompatibility [11]. It is widely investigated invarious fields, such as photolysis [12], photocatalysis[13], solar cells [14], sensors [15], and so on. The ca-pability to manipulate titania NCs into a desiredPHC structure and the possibility to modify the spon-taneous transition rate of the RE dopants in the tita-nia hosts might be interesting. In this work, wepresent our experimental approach to prepareTiO2:Sm3+ inverse opal and study the PHC effect onnarrowband emission of Sm3+.

The TiO2:Sm3+ inverse opal was prepared by atemplat-assisted method. Polystrene (PS) monodis-persed particles were purchased, and opal templateswere grown through vertical deposition [16].TiO2:Sm3+ (0.75% in molar ratio) precursor solutionswere prepared according to one of our previous works

[17]. Two PS templates were used in our experi-

0146-9592/10/091449-3/$15.00 ©

ments: a sample PC grown from PS spheres diam-etered at 450 nm and a sample REF from a mixtureof 200, 380, and 450 nm balls. After being heated inan oven at 100°C for 1 h for stronger physicalstrength, PS templates were submerged into the pre-cursor for maximum infiltration. Annealing was car-ried out with slowly elevated temperature �40°C/h�up to 500°C for 2 h.

X-ray diffraction shows that both samples werepure anatase in phase. The PC sample displays a red-dish color, while the REF gives no color at all. Whenwetted with ethanol, the reddish color of the PCfaded, as illustrated in Fig. 1(a). Figure 1(b) showsthe topview (along the substrate surface normal)scanning electron microscope (SEM) picture of thePC sample, in which the inverse opal structure of thePC can be clearly observed. Figure 1(c) shows a side-view (along the substrate surface) SEM image of thePC sample. It should be noted that along the normalthe air hole distance, d2=262 nm, is smaller than

Fig. 1. (Color online) (a) Photograph of the inverse opalsample (PC) wet with ethanol; (b) top-view SEM image;

(c) side-view SEM image.

2010 Optical Society of America

1450 OPTICS LETTERS / Vol. 35, No. 9 / May 1, 2010

that along the surface parallel, d1=300 nm. Since thediameter of PS balls is 450 nm, such a discrepency in-dicates a more severe shrinkage along the normalthan along the parallel. In a larger view the PCsample cracks into typically tens of micrometerpieces. Such a weakness is believed to be caused bythe different levels of shrinkage between the sub-strate and the material during calcination [3]. Thesample shrinkage along the parallel is to some extentcounteracted by the less shrinking substrate; yetalone the normal no such counteracting force exists,giving rise to the uneven shrinkage observed.

Figure 2 shows the transmittance spectra of the PCsample measured at different angles with respect tothe surface normal. Note that as measured from thenormal ��=0� the 450 nm PS template displays adeep and narrow bandgap centered at around 1100nm and as � is raised the bandgap shifts to blue. InFig. 2, the PC sample gives a wide bandgap centeredat 675 nm from the normal ��=0�. The bandgap shiftsa little to blue with the increasing angle, and after10° of rotation it starts to redshift and becomesbroader. Theoretically, the bandgap location of opalcan be calculated by Bragg’s law of diffraction com-bined with Snell’s law [3],

� =2dhkl

m�neff

2 − sin2 �, �1�

where � is the center of bandgap, m is the order ofBragg diffraction, dhkl is the hkl plane distance, neffis the average refractive index, and � is the anglefrom the incident light to the normal of the substratesurface. Assuming in our opal templates air takes26% of the space, the calculated bandgap fits the ex-perimental data, except the redshift of the PC band-gap with increasing detection angle, which might beattributed to the uneven shrinkage mentioned above.The (111) plane distance along the normal, which be-fore calcination equals to the �1̄11��11̄1��111̄� planedistance, now shrinks more severely, causing the red-shift of the bandgap.

The photoluminescent (PL) spectra of the PCsample taken at various angles with respect to the

Fig. 2. (Color online) Transmittance spectra taken atvarious angles with respect to surface normal in the PC

sample.

surface normal are given in Fig. 3(a). In all the spec-tra, the 4G5/2–6Hj�j=5/2,7/2,9/2,11/2� transitionsare observed, located at 585, 615, 664, and 728 nm,respectively. Note that the spectra are normalizedto the 585 nm peak, which is situated furthestaway from the normal bandgap observed.Figure 3(b) shows the dependence of the intensityratios of R1=I�4G5/2–6H9/2� /I�4G5/2–6H5/2� and R2=I�4G5/2–6H11/2� /I�4G5/2–6H5/2� on the angle of detec-tion. It is interesting to observe that R1 and R2 bothgradually increase with the increasing angle, inwhich the bandgap gradually shifts to red, becomesbroader and deeper, and such a trend in R1 is moreobvious than that in R2. This implies that the sup-pression of the bandgap on the peaks at 664 and 728nm becomes less obvious with the increasing angle,which could be attributed to the fact that not only thelocation, but also the depth and width of the bandgaplargely varies with the growing incident angle. Suchan angle-dependent behavior, which is absent inREF, is believed to arise from the PHC structure.

For both samples, the decay kinetic curves of vari-ous emission peaks were taken from different angles.Figure 4 gives the typical decay curves of the PC andREF samples (at 664 nm) measured at the normal��=0�. As is shown, the kinetic decay curves of bothTiO2:Sm3+ samples can be well fitted with a bi-exponential decay function [17],

I1e−a1t + I2e−a2t, �2�

where I1 and I2 represent contributions of the fasterand slower components �I1+I2=1�, respectively, whilea1 and a2 represent the faster and slower spontane-ous emission rates (SERs) of �4G5/2–6HJ. The bi-exponential dynamics could be attributed to Sm3+

ions located at different sites, surface and inner [17].It should be noted that the value of I1 (or I2) in Eq. (2)fluctuates between 0.5 and 0.57; we therefore decidethat it is not a good subject to study under the con-text of the PHC. Figure 5 shows the wavelength-dependence of the SERs of �4G5/2–6HJ in the PC andREF samples. First of all, for the sample REF with-out the PHC structure, both the faster and slower

Fig. 3. (Color online) (a) Steady-state emission spectrameasured at different angles for the PC sample;(b) intensity ratios of I�4G5/2–6H9/2� /I�4G5/2–6H5/2� andI�4G5/2–6H11/2� /I�4G5/2–6H5/2� as functions of the angle

of incident light.

May 1, 2010 / Vol. 35, No. 9 / OPTICS LETTERS 1451

SERs barely change when measured at differentangles and wavelengths. Second, for all the measuredwavelengths and angles, the SERs of �4G5/2–6HJ inthe sample PC are smaller than those in the sampleREF due to the modification of the PHC bandgap ex-cept to varying extent. Third, the decrease in thevalue for the slower SER is more obvious than thatfor the faster SER. Fourth, the SERs in the PC de-pend on both the measured wavelength and angle. Ingeneral, for the 585 and 615 nm peaks, which sitrelatively further away from the center of the band-gap, their SERs are larger than those for the 664 and728 nm peaks and have not only a smaller changewith increasing angle but also a smaller discrepancyfrom the sample REF. As for the 664 and 728 nm

Fig. 4. (Color online) Typical decay curves of the PC andREF samples. The dots are experimental data, and thesolid curves are fitting functions.

Fig. 5. (Color online) Dependence of SERs in the PC (mea-sured at different angles) and REF (measured at the nor-mal angle) samples on emission wavelength. (a) Fast com-

ponent. (b) Slow component.

peaks in the PC, their SERs are smaller than thosefor the 585 and 615 nm peaks and have more obviouschange with the varied angles. Overall, as the inci-dent angle is raised, SERs increase, which is under-standable when correlated with the transmittancespectra: the suppression on emission is weakened asthe angle is raised and bandgap center shifts awayfrom the emission peaks.

To sum up, we have fabricated TiO2:Sm3+ inverseopal and the modification of the photonic bandgap ef-fect on the emissions of Sm3+ was observed in termsof transmittance spectra, PL spectra, and decay dy-namics. The PL spectra show that the emissions ofSm3+ are direction-dependent, which might be usefulin color purification. The luminescence dynamicsshows decreased SERs within the bandgap and theextent of such a decrease depends on both the emis-sion wavelength and detection angle, which might beutilized in energy transfer processes such as photo-sensitization.

The authors are thankful for the financial supportof the National Talent Youth Science Foundation(grant no. 60925018), the National Natural ScienceFoundation of China (NSFC) (grants no. 50772042and no. 10704073), and the High-Tech Research andDevelopment Program of China (grant no.2007AA03Z314).

References

1. E. Yablonovitch, Phys. Rev. Lett. 58, 2059 (1987).2. S. V. Gaponenko, V. N. Bogomolov, E. P. Petrov, A. M.

Kapitonov, D. A. Yarotsky, I. I. Kalosha, A. A. Eych-mueller, A. L. Rogach, J. McGilp, U. Woggon, and F. J.Gindele, J. Lightwave Technol. 17, 2128 (1999).

3. D. W. McComb, B. M. Treble, C. J. Smith, R. M. De LaRue, and N. P. Johnson, J. Mater. Chem. 11, 142 (2001).

4. M. Fujita, S. Takahashi, Y. Tanaka, T. Asano, and S.Noda, Science 308, 1296 (2005).

5. D. Kovalev and M. Fujii, Adv. Mater. 17, 2531 (2005).6. E. Bovero and F. C. J. M. V. Veggel, J. Am. Chem. Soc.

130, 15374 (2008).7. Z. X. Li, L. L. Li, H. P. Zhou, Q. Yuan, C. Chen, L. D.

Sun, and C. H. Yan, Chem. Commun. 43, 6616 (2009).8. X. S. Qu, H. W. Song, G. H. Pan, X. Bai, B. Dong, H. F.

Zhao, Q. L. Dai, H. Zhang, R. F. Qin, and S. Z. Lu, J.Phys. Chem. C 113, 5906 (2009).

9. H. S. Yoo, J. Y. Han, S. W. Kim, D. Y. Jeon, and B. S.Bae, Opt. Express 17, 3732 (2009).

10. X. G. Huang, Z. W. Yang, L. Sun, Q. Xie, B. Li, J. Zhou,and L. T. Li, Mater. Chem. Phys. 115, 23 (2009).

11. P. A. Ramires, A. Romito, F. Cosentino, and E. Milella,Biomaterials 22, 1467 (2001).

12. A. Fujishima and K. Honda, Nature 238, 37 (1972).13. A. L. Linsebigler, G. Q. Lu, and J. T. Yates, Jr., Chem.

Rev. 95, 735 (1995).14. A. Mihi, M. E. Calvo, J. A. Anta, and H. Miguez, J.

Phys. Chem. C 112, 13 (2007).15. G. K. Mor, O. K. Varghese, M. Paulose, K. Shankar,

and C. A. Grimes, Sol. Energy Mater. Sol. Cells 90,2011 (2006).

16. L. M. Goldenberg, J. Wagner, J. Stumpe, B. R. Paulke,and E. Gornitz, Mater. Sci. Eng. C 22, 405 (2002).

17. L. Y. Hu, H. W. Song, G. H. Pan, B. Yan, R. F. Qin, Q. L.Dai, L. B. Fan, S. W. Li, and X. Bai, J. Lumin. 127, 371

(2007).