ZnO nanoparticles embedded in polymeric matrices

NanoSti-uctured Materials. Vol. I. No. 6. pp. 659466.1996 Elsevia Science Ltd Copyright @ 1996 Acta Metallurgica Inc. Printed in the USA. All rights reserved 0965-9773,96 $15.00 + .OO

Pergamon

PI1 So965-9773(96)00043-8

ZnO NANOPARTICLES

EMBEDDED

IN POLYMERIC

MATRICES

S. Mahamuni’, B.S. Bendre’, VJ. Leppert2, C.A. Smith2, D. Cooke3, S.H. Risbud2 and H.W.H. Lee3 ‘Department of Physics, University of Pune, Pune 411 007, India 2Department of Chemical Engineering and Materials Science, University of California, Davis, CA 95616 3Lawrence Livermore National Laboratory, Livermore, CA 94550 (Accepted May 1996) Abstract- Highly stable, wurtzite quantum sized ZnO colloids encapsulated in polymers have been synthesized. The particles can be obtained in a powder form and are partially redissolvable in organic media. A shift in the optical absorption spectrum confirms quantum size effects. Stationaryfluorescence measurements exhibit excitonic as well as trappedfluorescence. The intensity of trapped fluorescence changes with capping. The time resolved fluorescence measurements indicate considerably short decay times.

INTRODUCTION The synthesis of nanophase structures of electronic and optical materials is at the heart of many fundamental issues in chemical and materials science, for this size regime is where dramatic deviations from bulk electronic and structural properties have recently been observed. In the range of particle sizes between 1 and 20 nm, strong spatial delocalization of valence electrons creates properties characteristic of neither the bulk state nor the molecular state. This unique intermediate hybrid property range is of much current interest in exciting new research that has resulted in the processing and synthesis of novel materials. For example, nanosize effects have been observed and are being studied in several types of quantum confined crystallites (CdS, CdSe, metallic clusters possessing “magic numbers”, and most recently, superconducting Cm and C6& (A= an alkali metal or halogen)). Applications of nanoclusters include uses in luminescent devices, photoelectrochemical cells, photocatalysis and nonlinear optics (l-3). Although significant changes in properties are observed due to spatial confinement of charge carriers in three dimensions, it is also established that surface properties are important factors in determining the properties of these materials (4-7). Since the number of surface atoms in nanoclusters is comparable to the number of bulk atoms, surface defects and states can have a major effect on properties. Zinc oxide has many applications in luminescent devices. photocatalysis and photoelectrochemistry (8-12). Previous work on the synthesis of ZnO nanocrystals has been 659

660

S

MAHAMUNI ET AL.

restricted so far to unprotected ZnO clusters, which are quasi-stable due to excess acetate/ hydroxyl ions in solution. Such clusters (12) have been used to form films which exhibit novel electrochemical properties; although the clusters grown were somewhat larger than the particle sizes of interest in this study. Inorganic semiconductors such as ZnS and CdS have also been used as a cappant (4), which can lead to emulsion quenching. The surface chemistry, in addition to a narrow size distribution, was found to govern the photophysical properties. A promising approach in nanocluster synthesis from the practical viewpoint is to use stabilizing agents or to grow the particles in structured media. In the present work, we report the synthesis and stabilizing effects of polymers on the photophysical properties of wurtzite ZnO nanoclusters. The study of semiconductor nanocluster polymer matrix composite materials is especially useful because these doped polymeric systems have been reported to be promising materials for fast non-linear optical applications where third-order non-linearity is sought. Insightful chemical modifications and processing methods need to be developed to advance the science of synthesis and processing of semiconductor doped polymer composites. Further, sophisticated characterization of the interfaces between the nanosize semiconductor and the polymer matrix are essential because surface and interface states are very important contributors to the observed enhancement in the third order optical non-linearity and luminescence behavior.

EXPERIMENTAL

PROCEDURE

We initially carried out synthesis of ZnO nanoclusters by various routes proposed by Spanhel and Anderson (1 l), Bahneman et al. (8) and Koch Edal. (9) in the presence of polymeric media. The underlying reaction chemistry governed the choice of various reagents in the synthesis and ZnO nanocrystals were prepared by the protocol described below. A solution of 0.05 M Zn(ClO&, 6H2O in methanol was mixed with the polymeric capping agent. Various cappants such as poly-vinyl pyrrolidone (PVP), poly-vinyl alcohol (PVA) and sodium hexa-meta phosphate @IMP) were tried. Free standing, re-dissolvable nanocrystalline powder could only by obtained by using PVP. The use of PVAdid not allow aredissolvable powder to be obtained, while HMP resulted in amorphous or short range ordered ZnO clusters. The mixture of Zn(ClO& and PVP was then mixed with 0.07 M LiOH of the same quantity. It was stored at room temperature for a few hours, followed by separation of the colloids by centrifugation. The resultant powder could then be dispersed in methanol for optical absorption as well as for steady state luminescence studies. A Hitachi 303 spectrophotometer was used to obtain optical absorption spectra. Steady state luminescence measurements were obtained with a Perkin Elmer FS-50 fluorimeter with Xenon lamp and monochromator. The phase of the nanocrystallites was determined from powder X-ray diffraction experiments using a Philips PW-1840 X-ray powder diffractometer with Cu Ka incident radiation. Samples for transmission electron microscopy (TBM) were prepared by sonicating particles in solution 10 minutes immediately before transfer to aFormvar/carbon coated copper grid. An Hitachi 600 scanning transmission electron microscope (STEM) operated at 100 kV with a point-to-point resolution of 0.45 nm was used to obtain low magnification images. High resolution images were obtained with aTopcon - 002B microscope with an interpretable resolution limit of 0.18 nm operated at 200 kV.

2~0 Nmowmcus

EMEEDDED IN POLYMERIC MATRICES

661

Time resolved photoluminescence spectroscopy was performed using two sources: (i) the -100 fs output from a self-modelocked Ti:sapphire laser frequency-doubled with an appropriate KDP crystal to give excitation pulses varying from 355400 nm and at a 82 MHz repetition rate, (ii) the -150 fs output from a Ti:sapphire regenerative amplifier frequency-doubled to give 400 nm excitation pulses at a 1 kHz repetition rate. Photoluminescence spectra were recorded with a 0.25 m monochromator and an intensified optical multichannel analyzer. The photoluminescence spectrum is not corrected. A photomultiplier tube (1.5 ns resolution) was used to detect the photoluminescence decay, which was recorded at 450,500,550 and 600 nm using appropriate bandpass filters. All experiments were performed at rOOmtemperature.

RESULTS AND DISCUSSION Figure 1 shows optical absorption curves for (a) uncapped, (b) uncapped aged and (c) PVP protected clusters. Freshly preparedclustersexhibited a blue shift in absorption, with the excitonic shoulder appearing at about 300 nm compared to the band edge for macrocrystalline ZnO at 365 nm. Some aging effects of these nanoclusters were observed, but after about 2 days the excitonic absorption hump for uncapped clusters stabilized at 322 nm. PVP protected clusters exhibited an excitonic hump at about 310 nm. The approximate size of the nanoclusters was estimated from X-ray diffraction patterns using the Scherrer formula (10,13,14). Particle sizes were 2.8 nm for

300

400

Wavelength (nm) Figure 1. Optical absorption spectra of (a) freshly prepared uncapped, (b) aged and uncapped, and (c) PVP protected ZnO nanocrystals.

662

s

!hAMUNl

ET AL.

+::::::::::::::::::::::I:::::::::::

I

64

60

56

52

48

44

40

36

32

Two Theta (degrees) Figure 2. XRD pattern of (a) freshly prepared uncapped, (b) aged and uncapped, and (c) PVP protected ZnO nanoparticles.

the uncapped sample, 5.1 nm for the uncapped and aged sample, and 3.5 nm for the PVP protected sample. It is obvious that freshly prepared ZnO and PVPcapped ZnO clusters will exhibit quantum size effects as the Bohr exciton radius of ZnO is about 4.0 nm (11). Figure 2 shows x-ray diffraction patterns for the (a) uncapped, (b) uncapped and aged, and (c) PVP protected ZnO particles. The diffraction peaks are rather broader for freshly prepared ZnO nanoclusters, but become more intense and narrower after aging. PVP capped ZnO particles do not exhibit change in either size or crystallinity after several months. The interplanar spacing values are in good agreement with JCPD data (10) and hence the wurtzite ZnO phase is well established. TEM results for the uncapped and aged sample are shown in Figure 3, while those for the PVP protected sample are shown in Figure 4. The PVP protected sample had an average particle diameter of about 5.0 nm, while the uncapped and aged sample had an average particle diameter of about 12.0 nm. The size trend is consistent with the x-ray diffraction results, with TEM measurements indicating a slightly larger particle diameter for the PVPprotected sample (5.0 nm vs. 3.5 nm), and a considerably larger particle diameter for the uncapped and aged sample (12.0 nm vs. 5.1 nm). The larger particle diameter measured for the uncapped sample may be due to aging during the several weeks that elapsed between X-ray diffraction measurements and TEM analysis. Selected area diffraction (SAD) patterns for all samples were consistent with d-spacing values for wurtzite ZnO. High resolution electron microscopy (HREM) analysis of the uncapped sample showed ZnO particles to be single crystal in nature (Figure 3).

ZNO NANOPARTICLES EMEIEDOED IN POLYMERIC MATRICES

663

Figure 3. Micrographs showing (a) TEM of uncapped and aged ZnO particles (black dots 12 nm in diameter) with SAD pattern showing rings that match d-spacings for the wurtzite structure of ZnO, and (b) HREM of a smaller particle (- 6 nm) of uncapped and aged ZnO.

Figure 4. TEM of PVP protected ZnO nanoparticles (black dots - 5 nm in diameter) with SAD pattern showing rings that match d-spacings for the wurtzite structure of ZnO.

S MAHAMUNI ET AL.

400

600

500

Wavelength

(nm)

Figure 5. Stationary luminescence spectra of (a) uncapped and (b) PVP protected ZnO nanoparticles &xc: 325 nm). Figure 5 shows the steady state luminescence curves for uncapped and PVP protected ZnO nanocrystals at room temperature. Excitation wavelength was kept constant at 325 nm. Luminescence spectra exhibit two features at about 360 nm and 5 10 nm. The UV luminescence peak at 360 nm is due to HOMO - LUMO gap excitonic transition (5,8). Visible fluorescence at 5 10 nm is characteristically observed in ZnO. It is believed (8,l l- 12) that an anion vacancy trap level within the forbidden gap is responsible for this feature. Moreover, in the case of nanocrystals, trapped luminescence has been prominent, whereas the gap luminescence was extremely weak. It is interesting to note that the proportion of UV band gap luminescence observed in the case of our samples is much higher than that reported by Bahnemann et al. (8) and Koch et al. (9). Also, it should be noted that PVP capping further causes an increase in the intensity of UV emission compared with that of visible emission. Quenching of visible luminescence along with the higher stability offered by PVP capping may have possible technological implications. We are further investigating fluorescence quenching experiments of differently sized ZnO particles. Due to the hole transfer to an adsorbed species, Kamat et al. (4) also observed quenching of visible fluorescence. We tentatively ascribe the quenching of luminescence to adsorbed species and/or reduction in surface defects as a result of the capping agent. Figure 6 shows a time resolved luminescence decay curve for a representative capped ZnO sample in the visible luminescence regime. The decay curves are biexponential and can be fit by a nonlinear least squares procedure using the formula F(t) = al exp(-t/ tl) + a2 exp(-t/t2). The two

ZNO NANOPARTIC~ES EMBEDDED IN POLYMERIC MATRICES

665

0.20A .* g 2 CI 8 9 2 .3 6 ‘i3 ;s e

O.lS-

O.lO-

I

Wavelength (PL) = 550 nm Bi-exponential Fit: .8 ns, 7.3 ns

0.05 -

Time (ns) Figure 6. Time resolved luminescence spectrum of a representative capped ZnO nanoparticle sample showing a biexponential fit with time constants of 0.8 and 7.3 ns.

decay components have time constants of roughly l-2 and 8- 10 ns. The decay times observed are much shorter than reported earlier by Bahnemann ef al. (8) and Kamat et al. (4) (about 14 ns and 190 ns, respectively). Moreover, use of different capping agents also indicated similar shorter decay time constants. This provides additional evidence for the earlier conclusion that visible luminescence is due tothe surface vacancies responsible for shallow traps in the gap. Further work on determination of the nature and number of carrier traps is underway. SUMMARY

AND CONCLUSIONS

Stable suspensions of capped wurtzite phase ZnO nanoclusters (- 5 nm in diameter) encapsulated in polymeric media were synthesized. A blue shift in the absorption spectra compared to the band edge for macrocrystalline ZnO was observed. Steady state luminescence exhibited band edge as well as trapped luminescence, with surface capping resulting in a substantial increase in band edge UV luminescence yield relative to trap induced visible luminescence yield. Encapsulation is also believed to be responsible for shorter decay time constants than have been previously reported in the literature for nanocrystalline ZnO.

666

s

b.hiA~~Nl

ET AL.

ACKNOWLEDGMENTS

SM would like to thank the Department of Science and Technology, New Delhi for supporting this research. VJL, CAS, and SHR acknowledge support from the Electronic Materials Program at NSF through grant DMR94- 11179. The I-REM work was performed at the National Center for Electron Microscopy at the Lawrence Berkeley National Laboratory. The authors wish to thank Dr. Mark Fendorf for his assistance in interpreting HREM images. REFERENCES 1.

L. Brus, Act. Chem. Res. 2, 183 (1993).

2. 3. 4. 5. 6. 7. 8. 9.

A. Henglein, Chem. Rev. Se, 1861(1989). L. Brus, Appl. Phys. A Z$, 465 (1991). P.V. Kamat and B. Patrick, J. Phys. Chem. $& 6829 (1992). S. Hotchandani and P.V. Kamat, J. Phys. Chem. %, 6835 (1992). A. Hasselbarth, A. EychmulIer and H. Weller, Chem. Phys. Lett. m, 27 1 (1993). R.R. Chandler, J.L. Coffer, S.J. Atherton and PT. Snowden, J. Phys. Chem. %, 2713 (1992). D.W. Bahnemann, C. Kormann andM.R. Hoffmann, J. Phys. Chem. %,3789 (1987). U. Koch, A. Fojtik, H. Weller and A. Henglein, Chem. Phys. Lett. m, 507 (1985). M. Haase, H. Weller and A. Henglein, J. Phys. Chem. z, 482 (1988). L. Spanhel and M.A. Anderson, J. Am. Chem. Sot. u, 2826 (1991). P. Hoyer, R. Eichberger and H. Weller, Ber Bunsenges Phys. Chem. Z, 630 (1993). Y. Wang and N. Herron, Phys. Rev. B &,7253 (1990). S. Mahamuni, A. Khosravi, M. Kundu, A. Kshirsagar, A. Bedekar, D.B. Avasare, P. Singh and S.K. Kulkarni, J. Appl. Phys. B, 5237 (1993).

10. 11.

12. 13. 14.