Size dependence of exciton-phonon coupling in sol-gel ZnMgO powders
C. H. Chia,1,a)J. N. Chen,1T. C. Han,1J. W. Chiou,1Y. C. Lin,2W. L. Hsu,2and W. C. Chou2 1
Department of Applied Physics, National University of Kaohsiung, Kaohsiung 81148, Taiwan
2
Department of Electrophysics, National Chiao Tung University, Hsin-Chu 30010, Taiwan (Received 28 January 2011; accepted 6 February 2011; published online 23 March 2011)
We found that the exciton-phonon coupling in the Zn1xMgxO powders (0.01 5x 5 0.07) is greatly influenced by the crystalline-size. Two well-resolved photoluminescence (PL) bands due to recombination of free exciton and its longitudinal optical (LO)-phonon replicas enable us to analyze the relative intensities among free excitons, one-LO-phonon replicas, and two-LO-phonon replicas. As crystalline size increases, a larger enhancement of the PL-intensity ratio of a free exciton to its LO-phonon replicas was found compared to that of two LO-phonon replica to one-LO-phonon replica.VC 2011 American Institute of Physics. [doi:10.1063/1.3563574]
I. INTRODUCTION
There is a large potential for the development of light-emitting devices operating in the UV-to-deep-UV spectral range using Zn1xMgxO ternary alloy semiconductors as an active layer in optoelectronic devices.1,2In our previous let-ter, we reported the biexcitonic emissions of sol-gel Zn1xMgxO powders.
3
It was found that the emission of a longitudinal-optical (LO) phonon sideband is obvious in the sol-gel Zn1xMgxO powders at high temperature (T). As ex-citation density increases, the exciton-phonon interaction competed with the inelastic exciton-exciton scattering pro-cess, which is an important transition for stimulated emission in ZnO-based nanostructures. It is well known that the exci-ton-phonon coupling (EPC) effect is significant in ZnO-based materials, which possess large ionicity and polarity. It affects the stability of excitons at highT. Therefore the EPC is an important issue to ZnO-based materials.
The photoluminescence (PL) properties of ZnO-based nanostructures have been intensively studied due to their potential application in the nanoscale optoelectronic devices.4–9 It was found that the EPC can differ significantly from the bulk counterpart to the nanostructure systems. Among the liter-ature, several interpretations were given for the enhancement of EPC in ZnO nanostructures, such as free exciton reabsorp-tion,4imperfections in the crystal,5–8and lattice heating.9In particular, the room-temperature (RT) PL of ZnO-based nano-structures was characterized by a single emission band due to free excitons (FX) and LO phonon-assisted excitonic recombi-nations.4–8 No well-resolved emission band was detected because the studies only limited to samples with small size.
We systematically studied the EPC of Zn1xMgxO pow-ders with a size from nano- to micrometer. The separation of FX-emission band and its LO-phonon sideband allows us to investigate the crystalline-size (dsize) dependence of EPC. We found that thedsizeof the powders greatly influences the emission peak energy and integrated intensity of phonon-assisted excitonic transitions.
II. EXPERIMENT
The Zn1–xMgxO powders were grown from aqueous solu-tion prepared using zinc nitrate hexahydrate [Zn(NO3)26H2O] and magnesium nitrate hexahydrate [Mg(NO3)26H2O] as the starting materials, de-ionized water as the solvent, and citric acid (C6H8O7) as the stabilizer. The precursor solutions were mixed thoroughly until the formation of sols. The sols were preheated in a furnace at 120C for 12 h to evaporate the sol-vent and successively at 600C for 2 h to remove the organic residuals. The powders obtained from the dried sols were then post-annealed at calcination-temperature (TC)¼ 600, 800, and 1000C for 2 h in ambient air. The Mg contentx in the sam-ples is 0.01 5 x 5 0.07, determined by the exciton energy obtained from low-T reflection spectra, according to Ref.10. From the results of x-ray diffraction measurement (x-ray beam CuKa¼ 0.154 nm), no signal of MgO phase was detected from the samples. Thedsizeof the Zn1xMgxO pow-ders was confirmed by field emission scanning electron micro-scope (FESEM).
The PL spectra were measured by a 32-cm-long mono-chromator and a charge-coupled device camera. The samples were excited by 266-nm line of neodymium-doped yttrium aluminum garnet (Nd:YAG) laser. The pulsed laser has a pulse width of 10 ns and a repetition rate of 20 Hz. The exci-tation-power was kept as low as 20 W/cm2to avoid high-ex-citation effect of excitons. A closed cycle refrigerator was used to perform theT-dependent measurement.
III. RESULTS AND DISCUSSION
The FESEM images of the samples (0.01 5x 5 0.07) withTC¼ 600, 800, and 1000C are shown in Figs.1(a) to 1(f), respectively. Regardless of Mg concentrations, thedsize of the powders depends crucially on TC. As TC increases from 600 to 1000C, thedsizeof the powders increases. The average dsize of the samples annealed at 600, 800, and 1000C are 0.2, 0.5, and 0.9 lm, respectively, as shown in Fig.2(a). The corresponding RT-PL spectra of these samples are shown in Fig.2(b). A single emission band was observed for the samples with dsize 0.2 lm, whereas two well-resolved emission bands were observed in the samples with
a)Author to whom all correspondence should be addressed. Electronic mail:
0021-8979/2011/109(6)/063526/4/$30.00 109, 063526-1 VC2011 American Institute of Physics JOURNAL OF APPLIED PHYSICS 109, 063526 (2011)
dsize 0.5 and 0.9 lm. With comparison to the RT-reflec-tance spectra (not shown), the high-energy emission bands can be attributable to the radiative recombination of FX, denoted as open squares. The energetic separations of the low-energy bands and theFX bands are 82 and 110 meV for the samples with dsize 0.5 and 0.9 lm, respectively. The FX emissions of the samples with dsize 0.2 lm are weak, and the PL are dominated by emission bands with peak ener-gies about 65 meV lower than theFX energies determined from RT-reflectance measurement.
To clarify the origin of low-energy PL bands of the samples, we performed the T-dependent PL measurement. In Figs. 3(a)–3(c), we show the typical T-dependent PL spectra of the Zn1xMgxO powders (0.04 5x 5 0.05) with dsize 0.2, 0.5, and 0.9 lm, respectively. The PL spectra are dominated by radiative recombination of excitons local-ized at alloy potential fluctuation at lowT, judging from the Stoke’s shift between the PL peak energy and the FX energy obtained from low-T reflectance measurement. At 100 K, the PL peak energies blueshift because of delocali-zation of excitons. After 100 K, as excitons become mobile, new emissions attributable to LO phonon-sidebands ofFX (denote asFX-1LO) emerge in the PL spectra of all of the samples. At 100 K, The energetic separations between these FX-1LO lines and the FX lines are about 56 meV. The value matches well with the theoretical energetic separation cal-culated from DE¼ ELO-1.5kBT (Ref.11) withELO¼ 72 meV.12 The two-LO-phonon assisted excitonic transitions (FX-2LO) appear in the PL bands after 200 K. This could be inter-preted as being due to the thermalization effect of FX. At highT, FX’s with large wave vector k increase, and the si-multaneous emission of photons and LO phonons becomes a most efficient FX annihilation process.13 Therefore we can assign the low-energy PL bands to a combinedFX-1LO andFX-2LO transition in the samples. It is well-known that the Fro¨hlich interaction plays a major role in EPC in highly polar ZnO materials.14However, the Fro¨hlich interaction is
FIG. 1. The field emission scanning electron microscope image of Zn1xMgxO powders (a)
x 0.01, TC¼ 600C; (b) x 0.03,
TC¼ 800C; (c)x 0.07, TC¼ 1000C; (d)
x 0.04, TC¼ 600C; (e) x 0.05,
TC¼ 800C; (f)x 0.05, TC¼ 1000C.
FIG. 2. (a) Thedsizeof Zn1xMgxO powders (0.01 5x 5 0.07) as a function
ofTC. (b) The corresponding RT-PL of the Zn1xMgxO powders. The open
squares and solid circles represent the peak energy positions ofFX and LO-phonon sidebands, respectively. The averagedsize’s and the Mg
concentra-tions are also given at the left-hand side.
063526-2 Chia et al. J. Appl. Phys. 109, 063526 (2011)
extremely small in a perfect crystal due to the parity conser-vation, which results in a weak exciton-LO phonon cou-pling.5,9At low T, because the lateral extension of exciton wave function is limited by the localization,15 the EPC is expected to be small according to the Huang–Rhys model,16 which states that the EPC decreases as the separation between electrons and holes decreases. Therefore phonon sidebands are negligible at low T. As an exciton becomes free, a strong EPC is still not expected because of parity conservation mentioned in the preceding text. The lattice heating9 effect is unlikely under the low-excitation power (20 W/cm2). Also, the FX reabsorption effect4 can not explain the increase in FX-emission intensity in the large-dsize powder because a larger reabsorption is expected for the samples with larger dsize. Therefore we interpret the strong LO-phonon-assisted FX emission observed in the
samples as being due to the crystal-imperfection in the pow-ders.5–8In Fig. 4(a), we show the low-T PL spectra for all of the samples. The deep level emission is negligible, indi-cating crystal imperfection inside the powders could be excluded for consideration. Therefore we emphasize that the most decisive factor resulting in a strong LO-phonon-assistedFX emissions mainly stem from the surface defects. We note that the PL-intensity ratios of FX, FX-1LO, and FX-2LO are significantly different among the samples with dsize 0.2, 0.5, and 0.9 lm. To quantify the relative contributions of the FX, FX-1LO, and FX-2LO processes to the RT-PL spectra of the Zn1-xMgxO powders (0.04 5 x 5 0.05), the line shape of each PL band was fitted by three Gaussians that give a good approximation to the line shapes of both the FX and its LO phonon replica (bold solid lines in Fig. 3). The spectral positions of FX-1LO and FX-2LO are about 55 and 120 meV lower than that of FX. These values are smaller than the energies of one-LO phonon (72 meV) and two-LO phonons (144 meV) because of thermal distribution at RT.11 The inte-grated intensity ratio of the FX (IFX) to its LO phonon
FIG. 3. (a)T-dependent PL spectra for Zn0.95Mg0.05O powders (dsize 0.9
lm). (b) T-dependent PL spectra for Zn0.95Mg0.05O powders (dsize 0.5 lm).
(c)T-dependent PL spectra for Zn0.96Mg0.04O powders (dsize 0.2 lm). Lines
are given to guide the evolution of the PL band ofFX, FX-LO, and FX-2LO as a function ofT. The solid circles indicate the peak position of the combined 1LO- and 2LO-phonon replicas. Bolded solid curves on the PL spectra of 300 K are the fitting results. The dotted, dot-dashed, and dashed curves represent the contributions due toFX, FX-1LO, and FX-2LO, respectively.
FIG. 4. (a) The low-T PL spectra of Zn1-xMgxO powders. The PL spectra
are dominated by radiative recombination of localized excitons and the deep level emission is negligible. (b) The PL-intensity ratio I2LO/I1LO (solid
squares) andIFX/I1LOþ2LO(solid circles).
063526-3 Chia et al. J. Appl. Phys. 109, 063526 (2011)
replica (I1LOþ2LO) is shown in Fig. 4(b). TheIFX/I1LOþ2LO ratio increases from 0.01 to 0.47 as dsize increases from 0.2 to 0.9 lm because there is a larger portion ofFX that radiatively recombines inside the interior region than that at the proximity of surface region in the powder with larger dsize. We also show the integrated PL intensity-ra-tio I2LO/I1LO in Fig. 4(b). The ratio increases from 0.1 to 0.43 asdsize increases from 0.2 to 0.9 lm, leading to the redshift of LO-phonon-assisted FX emissions. The I2LO/ I1LO ratio, which is proportional to the Huang–Rhys fac-tor S,16 could be an indicator for the strength of EPC. The increase in EPC with increasing dsize was also reported for ZnO quantum dots17 and nanowires,18 using Raman spectroscopy. As the dsize increases, the extent of excitonic wave function can spread over a larger region with longer electron-hole separation. These excitons strongly coupled with the polarization field of LO pho-nons at the surface region, leading to largerS.17
IV. CONCLUSION
In conclusion, by investigating the RT-PL of a series of Zn1xMgxO powders (0.01 5x 5 0.07), we found that the EPC effect in the powders is greatly influenced by thedsize but irrelevant to the Mg concentrations. Two well-resolved PL bands due to recombination of FX and its LO-phonon replicas enable us to study the relative intensities amongFX, FX-1LO, and FX-2LO. As dsizeincreases, a larger enhance-ment ofIFX/I1LOþ2LO ratio was found compared to that of I2LO/I1LO ratio. This implies that both ratios have to be considered when discussing the origin of RT-PL from ZnO-based nanostructures having various sizes.
ACKNOWLEDGMENTS
This research was supported by National Science Council of Taiwan under Grant No. NSC-99–2112-M-390–001-MY3.
1H. Tampo, H. Shibata, K. Maejima, A. Yamada, K. Matsubara, P. Fons, S.
Niki, T. Tainaka, Y. Chiba, and H. Kanie,Appl. Phys. Lett.91, 261907 (2007).
2H. Zhu, C. X. Shan, B. H. Li, Z. Z. Zhang, J. Y. Zhang, B. Yao, D. Z.
Shen, and X. W. Fan,J. Appl. Phys.105, 103508 (2009).
3
C. H. Chia, Y. J. Lai, W. L. Hsu, T. C. Han, J. W. Chiou, Y. M. Hu, Y. C. Lin, W. C. Fan, and W. C. Chou,Appl. Phys. Lett.96, 191902 (2010).
4S. L. Chen, S. K. Lee, W. M. Chen, H. X. Dong, L. Sun, Z. H. Chen, and I.
A. Buyanova,Appl. Phys. Lett.96, 033108 (2010).
5
T. Voss, C. Bekeny, L. Wischmeier, H. Gafsi, S. Bo¨rner, W. Schade, A. C. Mofor, A. Bakin, and A. Waag,Appl. Phys. Lett.89, 182107 (2006).
6C. H. Ahn, S. K. Mohanta, N. E. Lee, and H. K. Cho,Appl. Phys. Lett.94,
261904 (2009).
7
W. K. Hong, G. Jo, M. Choe, T. Lee, J. I. Sohn, and M. E. Welland,Appl. Phys. Lett.94, 043103 (2009).
8X. Gu, K. Huo, G. Qian, J. Fu, and P. K. Chu, Appl. Phys. Lett. 93,
203117 (2008).
9
Y. J. Xing, Z. H. Xi, Z. Q. Xue, X. D. Zhang, J. H. Song, R. M. Wang, J. Xu, Y. Song, S. L. Zhang, and D. P. Yu,Appl. Phys. Lett.83, 1689 (2003).
10M. Lorenz, E. M. Kaidashav, A. Rahm, Th. Nobis, J. Lenzner, G. Wagner,
D. Spemann, H. Hochmuth, and M. Grudmann, Appl. Phys. Lett. 86, 143113 (2005).
11
C. Klingshirn,Phys. Status Solidi B71, 547 (1975).
12U¨ . O¨zgu¨r, Ya. I. Alivov, C. Liu, A. Teke, M. A. Reshchikov, S. Dog˘an,
A. Avrutin, S.-J. Cho, and H. Morkoc,J. Appl. Phys.98, 041301 (2005).
13
W. Shan, W. Walukiewicz, J. W. Ager III, K. M. Yu, H. B. Yuan, H. P. Xin, G. Cantwell, and J. J. Song,Appl. Phys. Lett.86, 191911 (2005).
14R. P. Wang, G. Xu, and P. Jin,Phys. Rev. B69, 113303 (2004).
15T. Makino, K. Tamura, C. H. Chia, Y. Segawa, M. Kawasaki, A. Ohtomo,
and H. Koinuma,Phys. Rev. B66, 233305 (2002).
16
K. Huang, and A. Rhys,Proc. R. Soc. London, Ser. A204, 406 (1950).
17R. C. Ray, Y. Low, H. M. Tsai, C. W. Pao, J. W. Chiou, S. C. Yang, F. Z.
Chien, W. F. Pong, M.-H. Tsai, K. F. Lin, H. M. Cheng, W. F. Hsieh, J. F. Lee,Appl. Phys. Lett91, 262101 (2007).
18
S. Nomura, and T. Kobayashi,Phys. Rev. B45, 1305 (1992).
063526-4 Chia et al. J. Appl. Phys. 109, 063526 (2011)
