- Email: [email protected]
The defect passivation effect of hydrogen on the optical properties of solution-grown ZnO nanorods
Physica B 480 (2016) 48–52
Contents lists available at ScienceDirect
Physica B journal homepage: www.elsevier.com/locate/physb
The defect passivation effect of hydrogen on the optical properties of solution-grown ZnO nanorods Z.N. Urgessa a,n, C.M. Mbulanga a, S.R. Tankio Djiokap a, J.R. Botha a, M.M. Duvenhage b, H.C. Swart b a b
Department of Physics, Nelson Mandela Metropolitan University, P.O. Box 77000, Port Elizabeth 6031, South Africa Department of Physics, University of the Free State, P.O. Box 339, Bloemfontein ZA9300, South Africa
art ic l e i nf o
a b s t r a c t
Article history: Received 14 May 2015 Accepted 24 June 2015 Available online 6 July 2015
In this study the effect of annealing environment on both low temperature and room temperature photoluminescence (PL) characteristics of ZnO nanorods, grown in solution, is presented. Particular attention is given to the effect of hydrogen defect passivation and its PL related line. It is shown that, irrespective of annealing ambient, an optimum annealing temperature of 300 °C suppresses the defect related emission and significantly improves the UV emission. By considering the stability of hydrogen impurities, the observed results in the PL spectra are analyzed. There is an observed asymmetric broadening on the low energy side of the bound exciton luminescence in the low temperature annealed samples which is explained by a high concentration of ionized impurities related to hydrogen. This has been attributed primarily to the conversion of hydrogen molecule to substitutional hydrogen on the oxygen site (HO) as a result of annealing. & 2015 Elsevier B.V. All rights reserved.
Keywords: ZnO PL Annealing Hydrogen Defect Passivation
1. Introduction Studies of the behavior of hydrogen in semiconducting material are not new, as it incorporates readily into most semiconductors and often behaves as an amphoteric impurity [1–3]. The knowledge of how hydrogen interacts with itself or other defects is not only of fundamental scientific interest, but also of crucial importance for the understanding of device processing and performance [2]. Early studies performed in the 1950s by D.G. Thomas and co-workers [4] detected n-type conductivity after introducing hydrogen into ZnO at elevated temperatures. Since then, the influence of hydrogen on doping and defect passivation in ZnO has been an active research area. In ZnO atomic hydrogen is exclusively a donor [1–4] and is found to incorporate on two stable sites: interstitially (Hi) and substitutionally at an oxygen vacancy (HO) [1–3]. In addition to these well-known hydrogen species, it appears that there are also a “hidden” reservoir of hydrogen in hydrothermally grown ZnO that can release Hi by annealing at 400 °C, even in the absence of external hydrogen. Interstitial H2 molecules (H2;int), which are neutral and invisible by infrared (IR) spectroscopy [5], were suggested to be a candidate for the hidden hydrogen [5]. The possibility of H2 in oxygen vacancies (VO2+–H2) n
Corresponding author. E-mail address: [email protected] (Z.N. Urgessa).
http://dx.doi.org/10.1016/j.physb.2015.06.028 0921-4526/& 2015 Elsevier B.V. All rights reserved.
rather than H2;int has also been suggested by first principle calculations [6]. Both of these defects can release Hi when the sample is annealed at around 400 °C [5,6]. Theoretical calculations predict the stability of Hi up to 125 °C and HO up to 475 °C [7]. Additionally, annealing experiments have been performed to understand the thermodynamics and kinetics of H in ZnO. In intentionally hydrogenated ZnO samples, using either a hydrogen plasma [1,8–10] or by annealing in H2 gas [11], hydrogen was found to passivate defects and increase the mobility and carrier concentration [1,8,11,12]. The increase in mobility and carrier concentration after implantation, and the subsequent decrease in these quantities after annealing, was ascribed to defect passivation by hydrogen [8,11]. From these reports it became accepted that while Hi anneals out at approximately 150 °C [11], HO is stable up to 500 °C [8,11]. In solution-grown ZnO (including nanorods investigated in this study) hydrogen is a constituent of the precursors and is the main donor impurity; thus, the very nature of the growth environment leads to hydrogen-related surface adsorbed and bulk impurities [10,12–14]. It is commonly found that low temperature ( 300 °C) annealing significantly increases the intensity of the hydrogen related PL lines, as well as improves the optical and electrical properties of the material [10,15–17]. Annealing primarily reduces the physically absorbed hydroxyl layer from the surface of ZnO rods and also removes organic and residual surface impurities. As
Z.N. Urgessa et al. / Physica B 480 (2016) 48–52
a result, the effect of annealing is mostly interpreted in terms of the removal of these surface adsorbed impurities [10,15–17]. However, annealing also causes hydrogen diffusion, altering the optical and electrical properties. Hence, the defect passivation effect of hydrogen in solution grown ZnO had not been studied in detail. Hydrogen in ZnO is often analyzed by low temperature PL [2,4,8,13], X-ray Photoelectron spectroscopy (XPS) [14,15,18], Raman and FTIR [1,11] experiments. In this study, the defect passivation effects of hydrogen on the optical properties of solutiongrown ZnO nanorods are characterized via PL and XPS. In the first part of this study, low temperature PL is used to systematically identify the temperature at which hydrogen out-diffuses from the sample. Then systematic isochronal and isothermal low temperature annealing was conducted and its effect on the RT PL characteristics is presented. Possible mechanisms that are responsible for the annealing effects are investigated and discussed.
2. Experiment
49
were chosen as it has been previously shown that hydrogen out diffuses from the nanorods in this range [13]. The PL spectra were measured at 10 K. Fig. 1(a) shows PL spectra obtained at 10 K for an as-grown sample, and samples annealed at 300 °C, 325 °C, 350 °C and 450 °C, respectively. A full study on the effect of annealing (from 200 to 900 °C) on the low temperature PL spectra is reported in our previous paper [13]. As can be seen from Fig. 1(a) hydrogen HO (I4) is the main donor impurity for samples annealed below 300 °C and it is annealed out after annealing at 450 °C. The concentration of this impurity, which causes an asymmetric broadening in the low energy tail [20], was calculated based on the Stark effect caused by ionized impurities [20,21]. According to theory, outlined in references [20,21], the presence of both donor and acceptor impurities at high concentrations cause some donors and compensating acceptors to be ionized at low temperature in n-type material. Thus, the emission energy of a bound exciton will decrease by an amount ΔE and its intensity is given by Ref. [20]
I ( ΔE ) = CA ΔE −1.75 Exp
Samples were grown from 50 mM hexamine and 50 mM zinc nitrate hexahydrate using de-ionized water on a seed coated substrate. Details of the growth technique are presented in Ref. [19]. The growth time and temperature were fixed to 90 min and 85 °C, respectively. Various characterization techniques namely Scanning Electron Microcopy (SEM), XRD, PL, XPS were used to investigate the effects of the annealing. For the low temperature PL experiments, the samples were excited with the 325 nm line of a He–Cd laser at 11 K and the emission was dispersed with a high resolution Horiba monochromator (FHR 1000) and detected by a photomultiplier tube (R666S) with a GaAs photocathode. Roomtemperature (RT) PL measurements were performed using a miniPL UV Laser System 5.0 (Photon systems, USA), which employs a NeCu laser with excitation wavelength of 248.6 nm.
3. Results and discussion
{ − 0.75A ΔE
−0.75
}
A ≡ πNci a o3 (4Eio )0.75 where C is constant, Nci is the concentration of charged impurities/ defects (donors plus acceptors), ao and Eio are the effective Bohr radius and binding energy of the intrinsic (free) exciton, respectively. The shape of the spectrum is related to the concentration of charged impurities Nci. From the equation above, the maximum of the curve lies at the 4
point ΔE = 8.727Nci 3 ao Eio [20]. Below this maximum point, the tail of the curve is characterized by ΔE −1.75, extending to lower energies. Above the maximum there appears a sharp cut off which reflects the exponential energy dependence originating from the assumption of a Poisson distribution of the nearest charged impurities during theoretical calculations [20]. Using ao ¼2 nm [22] and Eio ¼60 meV for ZnO, the above equation can be simplified as
{
I ( ΔE ) ≡ (1.4896−18) Nci ΔE −1.75 exp −(1.12−18) Nci ΔE −0.75 For the low temperature PL study, segments were cleaved from an as-grown sample, one of which was subsequently used as a reference. The others were annealed in oxygen for 30 min at different temperatures between 200 and 450 °C. These temperatures
} (1)
Fig. 1(b) shows the PL results for the sample annealed at 300 °C. The spectrum has been fitted on the low energy side using Eq. (1).
3.328
l
350 oC 400 oC 450 oC As-grown
y erg
low
tai
PL Intensity (arb. units)
PL Intesnity (arb. units)
300 oC 325 oC
en
I4
3.363 eV
3.336
3.344
3.352
3.360
Energy (eV)
3.368
3.376
As-grown
300 oC Theory
Nci = 2.4 x 1018 cm-3
3.33
3.34
3.35
3.36
3.37
3.38
Energy(eV)
Fig. 1. (a) 10 K PL spectra of as-grown rods and samples annealed 300 °C, 325 °C, 350 °C and 450 °C [13]; (b) theoretical fit of the low energy tail for the sample annealed at 300 °C, using the theory outlined in Refs. [20,21]. The dots represent the measured spectrum, while the solid line was calculated according to Eq. (1). The inset in (b) shows Nci as a function of annealing temperature.
50
Z.N. Urgessa et al. / Physica B 480 (2016) 48–52
Fig. 2. Comparison of the RT PL spectra of as-grown material to those of samples annealed in different environments for 30 min: (a) oxygen, (b) argon, (c) vacuum and (d) hydrogen. The effects of annealing time and temperature are shown in (e) and (f), respectively. (g, h) Effect of annealing temperature on the UV and deep level emission (DLE) intensity as a function of annealing temperature, respectively.
The value of Nci was found to be 2 1018 cm 3 and is comparable to the hydrogen related carrier concentration reported after hydrogen plasma treatment [8,11]. The insert in Fig.1(b) shows that the as-grown sample and samples annealed at low temperature are characterized by high concentrations of ionized impurities, which decrease significantly with annealing temperature. 3.1. Effect of annealing environment on the RT PL characteristics Once again, segments were cleaved from an as-grown sample and used to determine the effect of annealing conditions on the RT PL characteristics. Since the highest PL intensity was observed in the case of a sample annealed at 300 °C, the effect of annealing ambient and annealing time was studied at this temperature. The
effects of annealing temperature and environment were studied by annealing for a fixed time of 30 min. Fig. 2 compares the RT PL spectra of as-grown material to those of samples annealed in different environment for 30 min: (a) oxygen, (b) argon, (c) vacuum and (d) hydrogen. The effects of annealing time and temperature are shown in (e) and (f), respectively. Fig. 2(g) and (h) shows the effect of annealing temperature on the UV and deep level emission (DLE) intensity as a function of annealing temperature, respectively. As can be seen from Fig. 2(a)–(f) the RT PL spectra of nanorods display a UV emission at around 2.28 eV and broad DLE from 2 eV to 2.6 eV. It is well known that the RT PL spectra for ZnO usually display two major peaks: UV emission around 3.3 eV and broad visible emission (DLE). The UV peak is attributed to band-edge emission
Z.N. Urgessa et al. / Physica B 480 (2016) 48–52
related to free exciton, while the broad visible bands are generally attributed to intrinsic defects in ZnO [9,12]. The two most often reported intrinsic defect includes the oxygen vacancy VO at 2.5 eV and the zinc vacancy VZn at 2.3 eV [9,12]. All the spectra clearly indicate that, irrespective of the annealing environment, heat treatment enhances the UV emission significantly, while the DLE is reduced. The ratios of the intensity of the UV emission after and before annealing are found to be 4, 13, 14 and 14 when samples are annealed in oxygen, argon, vacuum and hydrogen, respectively. It is somewhat unexpected that annealing in an oxygen ambient leads to a weaker enhancement in UV emission compared to annealing the other environment. However, it has previously been reported by Liu et al. [23] that the exposure of ZnO rods to air caused the adsorption of O2 molecules on the surface. The author indicated that the adsorbed O2 molecule can capture an electron to form a superoxide anion O−2 , which causes band bending near the surface. This idea of oxygen molecule adsorption on (101¯ 1) surfaces has also been suggested using first-principles total-energy calculations by Yanfa et al. [24]. It is thus suggested that while annealing removes surface adsorbed species, an oxygen environment leads to the adsorbed of O2 molecules on the surfaces of the rods, resulting in a “residual” electric field near the surface that will reduce the intensity of the UV emission. In Fig. 2(e), the effect of annealing time at 300 °C in oxygen is presented. It is evident that 30 min is the optimum time to enhance the UV emission, and no significant improvement was observed as the annealing time was increased. The same study was conducted in hydrogen and nitrogen (not shown here) with the same result. Fig. 2(f) shows the effect of annealing temperature on the spectra. The change in the intensities of the UV and DLE as a function of annealing temperature is displayed in Fig. 2(g) and (h), respectively. For annealing temperatures below 300 °C a strong increase in the UV and a concomitant decrease in the DLE intensity can be seen. Between 300 °C and 500 °C the UV emission decreases rapidly, whereas the DLE increases significantly. Between 500 °C and 700 °C no significant change is observed in the UV intensity; however, the DLE now decreases. Above 700 °C, both the UV and DLE increase significantly. The increase in UV emission and suppression of the defectrelated emission after low temperature annealing ( 300 °C) has been observed in solution grown ZnO nanostructures and has been mostly attributed to the removal of surface adsorbed impurities [10,14–17]. It has also been shown that the most likely
51
surface adsorbed impurities are H, OH and different oxygen-related defects like O2, and CO [14,23,25,26]. These defects on the surface will introduce band bending near the surface and therefore strongly influence the PL properties of ZnO [15,25,27,28]. As most of these surface defects can be removed by low temperature annealing, it is natural to associate the improvement of the PL intensity after annealing with their removal. Iza et al. [10] studied the effect of annealing conditions (temperature and environment) on electrochemically deposited ZnO nanorods and reported a significant increase in the UV emission after annealing at 300 °C. Hang et al. [17] has also reported similar observations following annealing of chemical bath deposition (CBD) grown ZnO nanorods. Similar to what is observed in this study, these authors observed an increase in the UV emission after annealing between 200 and 300 °C, with a sample annealed at 300 °C yielding a high intensity of UV emission, a low resistivity and the lowest DLE intensity [10,17]. Further annealing between 300 °C and 500 °C, decreased the UV emission [10,17]; with the highest resistivity and strongest DLE observed for samples annealed at 400 °C [10]. Iza et al. [10] ascribe the increase in UV emission after annealing between 100 to 200 °C to decomposition of Zn(OH)2 and the further increase after annealing up to 300 °C to an enhancement in crystal quality. They supported the decomposition of the Zn(OH)2 with the observed quenching of OH related lines in the infrared (IR) spectra after annealing at 200 °C [10]. The decomposition temperature of Zn(OH)2 is 160 °C [10]. The decrease in the UV emission after annealing beyond 300 °C was ascribed to the formation of point defects as a result of annealing [10,17]. However, the attribution of the significant increase in PL intensity by annealing at 300 °C to an enhancement in crystal quality is unrealistic. Crystal quality is unlikely to be effected by such low annealing temperatures. Additionally, higher temperature annealing would further improve the crystal quality and thus UV emission; however, this is not seen in their results. In our previous studies, it has been shown that sequential annealing (from 200 to 850 °C) did not induce any significant change in the crystallinity of the nanorods as determined by XRD [13]. Based on the abundance and stability of hydrogen in solution grown ZnO, the enhancement in the UV emission irrespective of the annealing ambient, of low temperature annealed material, is primarily ascribed to the defect passivation effect of hydrogen. The removal of some non-radiative recombination channels (induced by band bending) as a result of annealing can also not be excluded. XPS was used to investigate the surface composition of asgrown and annealed samples. Fig. 3 presents the effect of
~ 530.5 eV
Count (/s)
Count (/s)
As-grown o 300 C o 400 C o 600 C o 850 C
532±3 eV 850 oC As-grown
526
528
530
532
534
Binding energy (eV)
536
538
526
528
530
532
534
536
538
Binding energy (eV)
Fig. 3. XPS spectra after sputter for core level O 1s from the as-grown and samples annealed in oxygen at different temperature. The annealing time was 30 min. (a) O 1s and (b) Gaussian curve fitting to the O 1s spectra. The scattered points are experimental data and the solid lines are Gaussian fit for the as-grown and sample annealed at 850 °C.
52
Z.N. Urgessa et al. / Physica B 480 (2016) 48–52
annealing on the O 1s XPS spectra of samples annealed in oxygen at different temperatures. The binding energies were calibrated by taking the carbon C 1s peak (284.7 eV) as reference. As can be seen from Fig. 3(b), for both the as-grown and annealed sample, the O 1s XPS line can be fitted to two distinct peaks centered, around 530.5±0.3 eV and 532.0±0.3 eV and commonly called O1 and O2, respectively [14,15]. The FWHM of these two lines are found to be around 1.1 eV and 1.5 eV, respectively. The O 1s signal in our case is much narrower than previously reported for similar material, in which up to 4 peaks were fitted to the overall signal [14]. The O1 peak is associated with O atoms in the wurtzite structure, surrounded by four Zn atoms with their full complement of nearestneighbor O atoms [14,15], whereas O2 is related to OH on the surface [14,15,18]. The presence of OH surface defects even after high temperature annealing is due to the fact that the samples were exposed to ambient during transfer from the annealing furnace to the XPS system. The same procedure also applies to the PL measurements subsequent to annealing. Thus, even though the presence of surface adsorbed impurities on the nanorods is undeniable, the effect of these impurities on the RT PL results is inconsequential. A possible alternate explanation for the observed RT PL after annealing at 300 °C is the activation of “hidden” interstitial [5] and substitutional hydrogen molecules [6] that can be converted into atomic hydrogen. By combining local mode and free carrier infrared (IR) absorption measurements, Shi et al. [5] studied the effect of annealing on as-received hydrothermally grown ZnO substrate. They reported an IR line at 3326.3 cm 1 (at 4 K) which was activated only after annealing near 400 °C. The intensity of this line increased with annealing temperature between 400 °C and 500 °C and then vanished upon annealing at 600 °C [5]. The same line could be produced after hydrogenation at 725 °C and subsequent quenching to room temperature. The donor was unstable, annealing out at 150 °C, but could be reactivated (presumably from the dissociation of “hidden” molecular hydrogen) by annealing near 450 °C. In the present work, the observed increase in the RT PL intensity after annealing at 300 °C, irrespective of annealing conditions, seems to agree with these findings. For instance, the slight increase in the UV emission and decrease in DLE after annealing at 200 °C can be related to diffusion of atomic hydrogen, causing the passivation of intrinsic defects like VZn and VO, which are the dominant contributors to the DLE [9]. The strong increase in the UV emission and concurrent decrease in the DLE after annealing at 300 °C (see Fig. 2), irrespective of annealing environment, however, should be primarily related to the dissociation of the hidden hydrogen (H2;int [5] or VO2+ –H2 [6]) into atomic hydrogen, which in turn will produce the thermally more stable HO shallow donors and the passivation of vacancies. As shown in Fig. 2(h) the highest UV intensity is observed in the case of a sample annealed at 300 °C, indicating that this temperature should be optimum for the dissociation of the hidden hydrogen in these nanorods. This can explain the increase in the conductivity after annealing at 300 °C as reported in the literature [8]. As the annealing temperature increases further from 300 °C to 500 °C, HO will dissociate and hydrogen will out-diffuse, resulting in a decrease in the I4 emission in the low temperature PL (Fig. 1). The out-diffusion of hydrogen will cause the reappearance of the DLE (Fig. 2(h)) and a decrease in the RT UV emission (Fig. 2(g)). The overall increase in emission intensity at high annealing temperatures (above 600 °C) is suggested to result from point defect annealing in the bulk, which enhances the free exciton emission, and the simultaneous degradation of the near-surface region, which could activate the DLE there.
4. Conclusion In summary, the effect annealing conditions (ambient, time and temperature) on the optical and surface properties of solution grown ZnO nanorods has been presented. Low temperature PL indicated that in samples annealed at temperature below 450 °C, the main donor impurity is hydrogen and its concentration is estimated to be 1018 cm 3. Irrespective of the annealing environment, 300 °C is shown to be the optimum temperature for enhancing the UV emission and suppressing the DLE. Based on XPS results, the defect passivation effect of hydrogen is deduced to be the possible cause for the observed enhancement of the UV emission after annealing at such lower temperatures.
Acknowledgments This work is based upon research supported by the South Africa Research Chairs Initiative of the Department of Science and Technology and the National Research Foundation (NRF), South Africa. The financial support from post-doctoral fellowship at Nelson Mandela Metropolitan University (NMMU) is also gratefully acknowledged.
References [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23] [24] [25] [26] [27] [28]
E.V. Lavrov, F. Herklotz, J. Weber, Phys. Rev. B 79 (2009) 165210. J. Weber, Phys. Status Solidi: C 5 (2008) 535. C.G. Van de Walle, J. Neugebauer, Nature 423 (2003) 626. D.G. Thomas, J.J. Lander, J. Chem. Phys. 25 (1956) 1136–1142. G. Alvin Shi, Marjan Saboktakin, Michael Stavola, S.J. Pearton, Appl. Phys. Lett. 85 (2004) 5601. M.-H. Du, K. Biswas, Phys. Rev. Lett. 106 (2011) 115502. J. Bang, K.J. Chang, Appl. Phys. Lett. 92 (2008) 132109. J.J. Dong, X.W. Zhang, J.B. You, P.F. Cai, Z.G. Yin, Q. An, X.B. Ma, P. Jin, Z.G. Wang, P.K. Chu, ACS Appl. Mater. Interfaces 2 (2010) 1780–1784. C. Ton-That, L. Weston, M.R. Phillips, Phys. Rev. B 86 (2012) 115205. D.C. Iza, D.M. Rojas, Q. Jia, B. Swartzentruber, J.L. MacManus-Driscoll, Nanoscale Res. Lett. 7 (2012) 655. G. Alvin Shi, M. Stavola, S.J. Pearton, M. Thieme, E.V. Lavrov, J. Weber, Phys. Rev. B 72 (2005) 195211. C. Guillen, J. Herrero, Vacuum 84 (2010) 924–929. Z.N. Urgessa, J.R. Botha, M.O. Eriksson, C.M. Mbulanga, S.R. Dobson, S.R. Tankio Djiokap, K.F. Karlsson, J. Appl. Phys. 116 (2014) 123506. L.L. Yang, Q.X. Zhao, M. Willander, X.J. Liu, M. Fahlman, J.H. Yang, Appl. Surf. Sci. 256 (2010) 3592–3597. L.L. Yang, Q.X. Zhao, M.Q. Israr, J.R. Sadaf, M. Willander, G. Pozina, J.H. Yang, J. Appl. Phys. 108 (2010) 103513. M.Y. Cho, M. Su Kim, H.Y. Choi, K.G. Yim, Jae-Y. Leem, Bull. Korean Chem. Soc. 32 (2011) 180. D.R. Hang, S.E. Islam, K.H. Sharma, S.W. Kuo, C.Z. Zhang, J.J. Wang, Nanoscale Res. Lett. 9 (2014) 632. A. Kushwaha, M. Aslam, J. Phys. D: Appl. Phys. 46 (2013) 485104. Z.N. Urgessa, O.S. Oluwafemi, E.J. Olivier, J.H. Neethling, J.R. Botha, J. Alloy. Compd. 580 (2013) 120–124. K. Hiroshi, S. Shigeo, T. Seizo, M. Kazuo, J. Phys. Soc. Jpn. 28 (1970) 110. H. Hanamura, J. Phys. Soc. Jpn. 28 (1970) 120. Y. Gu, I.L. Kuskovsky, M. Yin, S. O'Brien, G.F. Neumark, Appl. Phys. Lett. 85 (2004) 3833. W.Z. Liu, H.Y. Xu, J.G. Ma, C.Y. Liu, Y.X. Liu, Appl. Phys. Lett. 100 (2012) 203101. Yanfa Yan, M.M. Al-Jassim, Su-H. Wei, Phys. Rev. B 72 (2005) 161307. Z.M. Liao, H.Z. Zhang, Y.B. Zhou, J. Xua, J.M. Zhang, D.P. Yu, Phys. Lett. A 372 (2008) 4505–4509. I. Shalish, H. Temkin, V. Narayanamurti, Phys. Rev. B 69 (2004) 245401. W.C. Sun, Y.C. Yeh, C.T. Ko, H.H. Jr, Mi.J. Chen, Nanoscale Res. Lett. 6 (2011) 556. J.P. Richters, T. Voss, D. SKim, R. Scholz, M. Zacharias, Nanotechnology 19 (2008) 305202.
Contents lists available at ScienceDirect
Physica B journal homepage: www.elsevier.com/locate/physb
The defect passivation effect of hydrogen on the optical properties of solution-grown ZnO nanorods Z.N. Urgessa a,n, C.M. Mbulanga a, S.R. Tankio Djiokap a, J.R. Botha a, M.M. Duvenhage b, H.C. Swart b a b
Department of Physics, Nelson Mandela Metropolitan University, P.O. Box 77000, Port Elizabeth 6031, South Africa Department of Physics, University of the Free State, P.O. Box 339, Bloemfontein ZA9300, South Africa
art ic l e i nf o
a b s t r a c t
Article history: Received 14 May 2015 Accepted 24 June 2015 Available online 6 July 2015
In this study the effect of annealing environment on both low temperature and room temperature photoluminescence (PL) characteristics of ZnO nanorods, grown in solution, is presented. Particular attention is given to the effect of hydrogen defect passivation and its PL related line. It is shown that, irrespective of annealing ambient, an optimum annealing temperature of 300 °C suppresses the defect related emission and significantly improves the UV emission. By considering the stability of hydrogen impurities, the observed results in the PL spectra are analyzed. There is an observed asymmetric broadening on the low energy side of the bound exciton luminescence in the low temperature annealed samples which is explained by a high concentration of ionized impurities related to hydrogen. This has been attributed primarily to the conversion of hydrogen molecule to substitutional hydrogen on the oxygen site (HO) as a result of annealing. & 2015 Elsevier B.V. All rights reserved.
Keywords: ZnO PL Annealing Hydrogen Defect Passivation
1. Introduction Studies of the behavior of hydrogen in semiconducting material are not new, as it incorporates readily into most semiconductors and often behaves as an amphoteric impurity [1–3]. The knowledge of how hydrogen interacts with itself or other defects is not only of fundamental scientific interest, but also of crucial importance for the understanding of device processing and performance [2]. Early studies performed in the 1950s by D.G. Thomas and co-workers [4] detected n-type conductivity after introducing hydrogen into ZnO at elevated temperatures. Since then, the influence of hydrogen on doping and defect passivation in ZnO has been an active research area. In ZnO atomic hydrogen is exclusively a donor [1–4] and is found to incorporate on two stable sites: interstitially (Hi) and substitutionally at an oxygen vacancy (HO) [1–3]. In addition to these well-known hydrogen species, it appears that there are also a “hidden” reservoir of hydrogen in hydrothermally grown ZnO that can release Hi by annealing at 400 °C, even in the absence of external hydrogen. Interstitial H2 molecules (H2;int), which are neutral and invisible by infrared (IR) spectroscopy [5], were suggested to be a candidate for the hidden hydrogen [5]. The possibility of H2 in oxygen vacancies (VO2+–H2) n
Corresponding author. E-mail address: [email protected] (Z.N. Urgessa).
http://dx.doi.org/10.1016/j.physb.2015.06.028 0921-4526/& 2015 Elsevier B.V. All rights reserved.
rather than H2;int has also been suggested by first principle calculations [6]. Both of these defects can release Hi when the sample is annealed at around 400 °C [5,6]. Theoretical calculations predict the stability of Hi up to 125 °C and HO up to 475 °C [7]. Additionally, annealing experiments have been performed to understand the thermodynamics and kinetics of H in ZnO. In intentionally hydrogenated ZnO samples, using either a hydrogen plasma [1,8–10] or by annealing in H2 gas [11], hydrogen was found to passivate defects and increase the mobility and carrier concentration [1,8,11,12]. The increase in mobility and carrier concentration after implantation, and the subsequent decrease in these quantities after annealing, was ascribed to defect passivation by hydrogen [8,11]. From these reports it became accepted that while Hi anneals out at approximately 150 °C [11], HO is stable up to 500 °C [8,11]. In solution-grown ZnO (including nanorods investigated in this study) hydrogen is a constituent of the precursors and is the main donor impurity; thus, the very nature of the growth environment leads to hydrogen-related surface adsorbed and bulk impurities [10,12–14]. It is commonly found that low temperature ( 300 °C) annealing significantly increases the intensity of the hydrogen related PL lines, as well as improves the optical and electrical properties of the material [10,15–17]. Annealing primarily reduces the physically absorbed hydroxyl layer from the surface of ZnO rods and also removes organic and residual surface impurities. As
Z.N. Urgessa et al. / Physica B 480 (2016) 48–52
a result, the effect of annealing is mostly interpreted in terms of the removal of these surface adsorbed impurities [10,15–17]. However, annealing also causes hydrogen diffusion, altering the optical and electrical properties. Hence, the defect passivation effect of hydrogen in solution grown ZnO had not been studied in detail. Hydrogen in ZnO is often analyzed by low temperature PL [2,4,8,13], X-ray Photoelectron spectroscopy (XPS) [14,15,18], Raman and FTIR [1,11] experiments. In this study, the defect passivation effects of hydrogen on the optical properties of solutiongrown ZnO nanorods are characterized via PL and XPS. In the first part of this study, low temperature PL is used to systematically identify the temperature at which hydrogen out-diffuses from the sample. Then systematic isochronal and isothermal low temperature annealing was conducted and its effect on the RT PL characteristics is presented. Possible mechanisms that are responsible for the annealing effects are investigated and discussed.
2. Experiment
49
were chosen as it has been previously shown that hydrogen out diffuses from the nanorods in this range [13]. The PL spectra were measured at 10 K. Fig. 1(a) shows PL spectra obtained at 10 K for an as-grown sample, and samples annealed at 300 °C, 325 °C, 350 °C and 450 °C, respectively. A full study on the effect of annealing (from 200 to 900 °C) on the low temperature PL spectra is reported in our previous paper [13]. As can be seen from Fig. 1(a) hydrogen HO (I4) is the main donor impurity for samples annealed below 300 °C and it is annealed out after annealing at 450 °C. The concentration of this impurity, which causes an asymmetric broadening in the low energy tail [20], was calculated based on the Stark effect caused by ionized impurities [20,21]. According to theory, outlined in references [20,21], the presence of both donor and acceptor impurities at high concentrations cause some donors and compensating acceptors to be ionized at low temperature in n-type material. Thus, the emission energy of a bound exciton will decrease by an amount ΔE and its intensity is given by Ref. [20]
I ( ΔE ) = CA ΔE −1.75 Exp
Samples were grown from 50 mM hexamine and 50 mM zinc nitrate hexahydrate using de-ionized water on a seed coated substrate. Details of the growth technique are presented in Ref. [19]. The growth time and temperature were fixed to 90 min and 85 °C, respectively. Various characterization techniques namely Scanning Electron Microcopy (SEM), XRD, PL, XPS were used to investigate the effects of the annealing. For the low temperature PL experiments, the samples were excited with the 325 nm line of a He–Cd laser at 11 K and the emission was dispersed with a high resolution Horiba monochromator (FHR 1000) and detected by a photomultiplier tube (R666S) with a GaAs photocathode. Roomtemperature (RT) PL measurements were performed using a miniPL UV Laser System 5.0 (Photon systems, USA), which employs a NeCu laser with excitation wavelength of 248.6 nm.
3. Results and discussion
{ − 0.75A ΔE
−0.75
}
A ≡ πNci a o3 (4Eio )0.75 where C is constant, Nci is the concentration of charged impurities/ defects (donors plus acceptors), ao and Eio are the effective Bohr radius and binding energy of the intrinsic (free) exciton, respectively. The shape of the spectrum is related to the concentration of charged impurities Nci. From the equation above, the maximum of the curve lies at the 4
point ΔE = 8.727Nci 3 ao Eio [20]. Below this maximum point, the tail of the curve is characterized by ΔE −1.75, extending to lower energies. Above the maximum there appears a sharp cut off which reflects the exponential energy dependence originating from the assumption of a Poisson distribution of the nearest charged impurities during theoretical calculations [20]. Using ao ¼2 nm [22] and Eio ¼60 meV for ZnO, the above equation can be simplified as
{
I ( ΔE ) ≡ (1.4896−18) Nci ΔE −1.75 exp −(1.12−18) Nci ΔE −0.75 For the low temperature PL study, segments were cleaved from an as-grown sample, one of which was subsequently used as a reference. The others were annealed in oxygen for 30 min at different temperatures between 200 and 450 °C. These temperatures
} (1)
Fig. 1(b) shows the PL results for the sample annealed at 300 °C. The spectrum has been fitted on the low energy side using Eq. (1).
3.328
l
350 oC 400 oC 450 oC As-grown
y erg
low
tai
PL Intensity (arb. units)
PL Intesnity (arb. units)
300 oC 325 oC
en
I4
3.363 eV
3.336
3.344
3.352
3.360
Energy (eV)
3.368
3.376
As-grown
300 oC Theory
Nci = 2.4 x 1018 cm-3
3.33
3.34
3.35
3.36
3.37
3.38
Energy(eV)
Fig. 1. (a) 10 K PL spectra of as-grown rods and samples annealed 300 °C, 325 °C, 350 °C and 450 °C [13]; (b) theoretical fit of the low energy tail for the sample annealed at 300 °C, using the theory outlined in Refs. [20,21]. The dots represent the measured spectrum, while the solid line was calculated according to Eq. (1). The inset in (b) shows Nci as a function of annealing temperature.
50
Z.N. Urgessa et al. / Physica B 480 (2016) 48–52
Fig. 2. Comparison of the RT PL spectra of as-grown material to those of samples annealed in different environments for 30 min: (a) oxygen, (b) argon, (c) vacuum and (d) hydrogen. The effects of annealing time and temperature are shown in (e) and (f), respectively. (g, h) Effect of annealing temperature on the UV and deep level emission (DLE) intensity as a function of annealing temperature, respectively.
The value of Nci was found to be 2 1018 cm 3 and is comparable to the hydrogen related carrier concentration reported after hydrogen plasma treatment [8,11]. The insert in Fig.1(b) shows that the as-grown sample and samples annealed at low temperature are characterized by high concentrations of ionized impurities, which decrease significantly with annealing temperature. 3.1. Effect of annealing environment on the RT PL characteristics Once again, segments were cleaved from an as-grown sample and used to determine the effect of annealing conditions on the RT PL characteristics. Since the highest PL intensity was observed in the case of a sample annealed at 300 °C, the effect of annealing ambient and annealing time was studied at this temperature. The
effects of annealing temperature and environment were studied by annealing for a fixed time of 30 min. Fig. 2 compares the RT PL spectra of as-grown material to those of samples annealed in different environment for 30 min: (a) oxygen, (b) argon, (c) vacuum and (d) hydrogen. The effects of annealing time and temperature are shown in (e) and (f), respectively. Fig. 2(g) and (h) shows the effect of annealing temperature on the UV and deep level emission (DLE) intensity as a function of annealing temperature, respectively. As can be seen from Fig. 2(a)–(f) the RT PL spectra of nanorods display a UV emission at around 2.28 eV and broad DLE from 2 eV to 2.6 eV. It is well known that the RT PL spectra for ZnO usually display two major peaks: UV emission around 3.3 eV and broad visible emission (DLE). The UV peak is attributed to band-edge emission
Z.N. Urgessa et al. / Physica B 480 (2016) 48–52
related to free exciton, while the broad visible bands are generally attributed to intrinsic defects in ZnO [9,12]. The two most often reported intrinsic defect includes the oxygen vacancy VO at 2.5 eV and the zinc vacancy VZn at 2.3 eV [9,12]. All the spectra clearly indicate that, irrespective of the annealing environment, heat treatment enhances the UV emission significantly, while the DLE is reduced. The ratios of the intensity of the UV emission after and before annealing are found to be 4, 13, 14 and 14 when samples are annealed in oxygen, argon, vacuum and hydrogen, respectively. It is somewhat unexpected that annealing in an oxygen ambient leads to a weaker enhancement in UV emission compared to annealing the other environment. However, it has previously been reported by Liu et al. [23] that the exposure of ZnO rods to air caused the adsorption of O2 molecules on the surface. The author indicated that the adsorbed O2 molecule can capture an electron to form a superoxide anion O−2 , which causes band bending near the surface. This idea of oxygen molecule adsorption on (101¯ 1) surfaces has also been suggested using first-principles total-energy calculations by Yanfa et al. [24]. It is thus suggested that while annealing removes surface adsorbed species, an oxygen environment leads to the adsorbed of O2 molecules on the surfaces of the rods, resulting in a “residual” electric field near the surface that will reduce the intensity of the UV emission. In Fig. 2(e), the effect of annealing time at 300 °C in oxygen is presented. It is evident that 30 min is the optimum time to enhance the UV emission, and no significant improvement was observed as the annealing time was increased. The same study was conducted in hydrogen and nitrogen (not shown here) with the same result. Fig. 2(f) shows the effect of annealing temperature on the spectra. The change in the intensities of the UV and DLE as a function of annealing temperature is displayed in Fig. 2(g) and (h), respectively. For annealing temperatures below 300 °C a strong increase in the UV and a concomitant decrease in the DLE intensity can be seen. Between 300 °C and 500 °C the UV emission decreases rapidly, whereas the DLE increases significantly. Between 500 °C and 700 °C no significant change is observed in the UV intensity; however, the DLE now decreases. Above 700 °C, both the UV and DLE increase significantly. The increase in UV emission and suppression of the defectrelated emission after low temperature annealing ( 300 °C) has been observed in solution grown ZnO nanostructures and has been mostly attributed to the removal of surface adsorbed impurities [10,14–17]. It has also been shown that the most likely
51
surface adsorbed impurities are H, OH and different oxygen-related defects like O2, and CO [14,23,25,26]. These defects on the surface will introduce band bending near the surface and therefore strongly influence the PL properties of ZnO [15,25,27,28]. As most of these surface defects can be removed by low temperature annealing, it is natural to associate the improvement of the PL intensity after annealing with their removal. Iza et al. [10] studied the effect of annealing conditions (temperature and environment) on electrochemically deposited ZnO nanorods and reported a significant increase in the UV emission after annealing at 300 °C. Hang et al. [17] has also reported similar observations following annealing of chemical bath deposition (CBD) grown ZnO nanorods. Similar to what is observed in this study, these authors observed an increase in the UV emission after annealing between 200 and 300 °C, with a sample annealed at 300 °C yielding a high intensity of UV emission, a low resistivity and the lowest DLE intensity [10,17]. Further annealing between 300 °C and 500 °C, decreased the UV emission [10,17]; with the highest resistivity and strongest DLE observed for samples annealed at 400 °C [10]. Iza et al. [10] ascribe the increase in UV emission after annealing between 100 to 200 °C to decomposition of Zn(OH)2 and the further increase after annealing up to 300 °C to an enhancement in crystal quality. They supported the decomposition of the Zn(OH)2 with the observed quenching of OH related lines in the infrared (IR) spectra after annealing at 200 °C [10]. The decomposition temperature of Zn(OH)2 is 160 °C [10]. The decrease in the UV emission after annealing beyond 300 °C was ascribed to the formation of point defects as a result of annealing [10,17]. However, the attribution of the significant increase in PL intensity by annealing at 300 °C to an enhancement in crystal quality is unrealistic. Crystal quality is unlikely to be effected by such low annealing temperatures. Additionally, higher temperature annealing would further improve the crystal quality and thus UV emission; however, this is not seen in their results. In our previous studies, it has been shown that sequential annealing (from 200 to 850 °C) did not induce any significant change in the crystallinity of the nanorods as determined by XRD [13]. Based on the abundance and stability of hydrogen in solution grown ZnO, the enhancement in the UV emission irrespective of the annealing ambient, of low temperature annealed material, is primarily ascribed to the defect passivation effect of hydrogen. The removal of some non-radiative recombination channels (induced by band bending) as a result of annealing can also not be excluded. XPS was used to investigate the surface composition of asgrown and annealed samples. Fig. 3 presents the effect of
~ 530.5 eV
Count (/s)
Count (/s)
As-grown o 300 C o 400 C o 600 C o 850 C
532±3 eV 850 oC As-grown
526
528
530
532
534
Binding energy (eV)
536
538
526
528
530
532
534
536
538
Binding energy (eV)
Fig. 3. XPS spectra after sputter for core level O 1s from the as-grown and samples annealed in oxygen at different temperature. The annealing time was 30 min. (a) O 1s and (b) Gaussian curve fitting to the O 1s spectra. The scattered points are experimental data and the solid lines are Gaussian fit for the as-grown and sample annealed at 850 °C.
52
Z.N. Urgessa et al. / Physica B 480 (2016) 48–52
annealing on the O 1s XPS spectra of samples annealed in oxygen at different temperatures. The binding energies were calibrated by taking the carbon C 1s peak (284.7 eV) as reference. As can be seen from Fig. 3(b), for both the as-grown and annealed sample, the O 1s XPS line can be fitted to two distinct peaks centered, around 530.5±0.3 eV and 532.0±0.3 eV and commonly called O1 and O2, respectively [14,15]. The FWHM of these two lines are found to be around 1.1 eV and 1.5 eV, respectively. The O 1s signal in our case is much narrower than previously reported for similar material, in which up to 4 peaks were fitted to the overall signal [14]. The O1 peak is associated with O atoms in the wurtzite structure, surrounded by four Zn atoms with their full complement of nearestneighbor O atoms [14,15], whereas O2 is related to OH on the surface [14,15,18]. The presence of OH surface defects even after high temperature annealing is due to the fact that the samples were exposed to ambient during transfer from the annealing furnace to the XPS system. The same procedure also applies to the PL measurements subsequent to annealing. Thus, even though the presence of surface adsorbed impurities on the nanorods is undeniable, the effect of these impurities on the RT PL results is inconsequential. A possible alternate explanation for the observed RT PL after annealing at 300 °C is the activation of “hidden” interstitial [5] and substitutional hydrogen molecules [6] that can be converted into atomic hydrogen. By combining local mode and free carrier infrared (IR) absorption measurements, Shi et al. [5] studied the effect of annealing on as-received hydrothermally grown ZnO substrate. They reported an IR line at 3326.3 cm 1 (at 4 K) which was activated only after annealing near 400 °C. The intensity of this line increased with annealing temperature between 400 °C and 500 °C and then vanished upon annealing at 600 °C [5]. The same line could be produced after hydrogenation at 725 °C and subsequent quenching to room temperature. The donor was unstable, annealing out at 150 °C, but could be reactivated (presumably from the dissociation of “hidden” molecular hydrogen) by annealing near 450 °C. In the present work, the observed increase in the RT PL intensity after annealing at 300 °C, irrespective of annealing conditions, seems to agree with these findings. For instance, the slight increase in the UV emission and decrease in DLE after annealing at 200 °C can be related to diffusion of atomic hydrogen, causing the passivation of intrinsic defects like VZn and VO, which are the dominant contributors to the DLE [9]. The strong increase in the UV emission and concurrent decrease in the DLE after annealing at 300 °C (see Fig. 2), irrespective of annealing environment, however, should be primarily related to the dissociation of the hidden hydrogen (H2;int [5] or VO2+ –H2 [6]) into atomic hydrogen, which in turn will produce the thermally more stable HO shallow donors and the passivation of vacancies. As shown in Fig. 2(h) the highest UV intensity is observed in the case of a sample annealed at 300 °C, indicating that this temperature should be optimum for the dissociation of the hidden hydrogen in these nanorods. This can explain the increase in the conductivity after annealing at 300 °C as reported in the literature [8]. As the annealing temperature increases further from 300 °C to 500 °C, HO will dissociate and hydrogen will out-diffuse, resulting in a decrease in the I4 emission in the low temperature PL (Fig. 1). The out-diffusion of hydrogen will cause the reappearance of the DLE (Fig. 2(h)) and a decrease in the RT UV emission (Fig. 2(g)). The overall increase in emission intensity at high annealing temperatures (above 600 °C) is suggested to result from point defect annealing in the bulk, which enhances the free exciton emission, and the simultaneous degradation of the near-surface region, which could activate the DLE there.
4. Conclusion In summary, the effect annealing conditions (ambient, time and temperature) on the optical and surface properties of solution grown ZnO nanorods has been presented. Low temperature PL indicated that in samples annealed at temperature below 450 °C, the main donor impurity is hydrogen and its concentration is estimated to be 1018 cm 3. Irrespective of the annealing environment, 300 °C is shown to be the optimum temperature for enhancing the UV emission and suppressing the DLE. Based on XPS results, the defect passivation effect of hydrogen is deduced to be the possible cause for the observed enhancement of the UV emission after annealing at such lower temperatures.
Acknowledgments This work is based upon research supported by the South Africa Research Chairs Initiative of the Department of Science and Technology and the National Research Foundation (NRF), South Africa. The financial support from post-doctoral fellowship at Nelson Mandela Metropolitan University (NMMU) is also gratefully acknowledged.
References [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23] [24] [25] [26] [27] [28]
E.V. Lavrov, F. Herklotz, J. Weber, Phys. Rev. B 79 (2009) 165210. J. Weber, Phys. Status Solidi: C 5 (2008) 535. C.G. Van de Walle, J. Neugebauer, Nature 423 (2003) 626. D.G. Thomas, J.J. Lander, J. Chem. Phys. 25 (1956) 1136–1142. G. Alvin Shi, Marjan Saboktakin, Michael Stavola, S.J. Pearton, Appl. Phys. Lett. 85 (2004) 5601. M.-H. Du, K. Biswas, Phys. Rev. Lett. 106 (2011) 115502. J. Bang, K.J. Chang, Appl. Phys. Lett. 92 (2008) 132109. J.J. Dong, X.W. Zhang, J.B. You, P.F. Cai, Z.G. Yin, Q. An, X.B. Ma, P. Jin, Z.G. Wang, P.K. Chu, ACS Appl. Mater. Interfaces 2 (2010) 1780–1784. C. Ton-That, L. Weston, M.R. Phillips, Phys. Rev. B 86 (2012) 115205. D.C. Iza, D.M. Rojas, Q. Jia, B. Swartzentruber, J.L. MacManus-Driscoll, Nanoscale Res. Lett. 7 (2012) 655. G. Alvin Shi, M. Stavola, S.J. Pearton, M. Thieme, E.V. Lavrov, J. Weber, Phys. Rev. B 72 (2005) 195211. C. Guillen, J. Herrero, Vacuum 84 (2010) 924–929. Z.N. Urgessa, J.R. Botha, M.O. Eriksson, C.M. Mbulanga, S.R. Dobson, S.R. Tankio Djiokap, K.F. Karlsson, J. Appl. Phys. 116 (2014) 123506. L.L. Yang, Q.X. Zhao, M. Willander, X.J. Liu, M. Fahlman, J.H. Yang, Appl. Surf. Sci. 256 (2010) 3592–3597. L.L. Yang, Q.X. Zhao, M.Q. Israr, J.R. Sadaf, M. Willander, G. Pozina, J.H. Yang, J. Appl. Phys. 108 (2010) 103513. M.Y. Cho, M. Su Kim, H.Y. Choi, K.G. Yim, Jae-Y. Leem, Bull. Korean Chem. Soc. 32 (2011) 180. D.R. Hang, S.E. Islam, K.H. Sharma, S.W. Kuo, C.Z. Zhang, J.J. Wang, Nanoscale Res. Lett. 9 (2014) 632. A. Kushwaha, M. Aslam, J. Phys. D: Appl. Phys. 46 (2013) 485104. Z.N. Urgessa, O.S. Oluwafemi, E.J. Olivier, J.H. Neethling, J.R. Botha, J. Alloy. Compd. 580 (2013) 120–124. K. Hiroshi, S. Shigeo, T. Seizo, M. Kazuo, J. Phys. Soc. Jpn. 28 (1970) 110. H. Hanamura, J. Phys. Soc. Jpn. 28 (1970) 120. Y. Gu, I.L. Kuskovsky, M. Yin, S. O'Brien, G.F. Neumark, Appl. Phys. Lett. 85 (2004) 3833. W.Z. Liu, H.Y. Xu, J.G. Ma, C.Y. Liu, Y.X. Liu, Appl. Phys. Lett. 100 (2012) 203101. Yanfa Yan, M.M. Al-Jassim, Su-H. Wei, Phys. Rev. B 72 (2005) 161307. Z.M. Liao, H.Z. Zhang, Y.B. Zhou, J. Xua, J.M. Zhang, D.P. Yu, Phys. Lett. A 372 (2008) 4505–4509. I. Shalish, H. Temkin, V. Narayanamurti, Phys. Rev. B 69 (2004) 245401. W.C. Sun, Y.C. Yeh, C.T. Ko, H.H. Jr, Mi.J. Chen, Nanoscale Res. Lett. 6 (2011) 556. J.P. Richters, T. Voss, D. SKim, R. Scholz, M. Zacharias, Nanotechnology 19 (2008) 305202.