1. Introduction

Surface plasmon (SP), excited by the interaction between light and metal surface [1–3] is known as an effective means to improve the emission efficiency of optoelectronic devices [4, 5]. Recently, numerous studies have been conducted to improve the band edge emission in ZnO based materials and different metallic materials (Ag, Au, Pt, graphene etc.) have been used as capping layers [6–14]. Enhancements in these types of ultraviolet (UV) emissions were mainly interpreted in terms of coupling between plasmons of the attached metallic materials and excitons in the ZnO material. However, Y. J. Fang et al. [15] have shown that the large enhancement of the UV emissions of metal coated ZnO nanowires originates from the electron transfer process (from metal to ZnO) due to the Ohmic contact formed between them. J. Song et al. [16] and A. Dev et al. [17] have shown that metal-deposition on ZnO can lead to strong enhancement of the UV emissions by passivating some defect states. Actually, surface modification by metal coatings may impact light extraction efficiency at metal/ZnO, which plays an important role in the quantum efficiency of the emitters due to the SP and scattering properties of metals [18]. However, there have been few reports to clearly explicate the contribution of light extraction caused by surface modification toward emission enhancement of SP-emitter, and SP coupling mechanism of excitons and light scattering mechanism are far from being fully understood. On the other hand, MgZnO, which can be achieved by alloying ZnO with MgO, has been regarded as a very promising material for short wavelength optoelectronic device applications, such as UV and deep UV light emitting and laser diodes [19, 20]. However, MgO has a rock-salt structure which is dissimilar to the wurtzite structure of ZnO. Thus, the incorporation of Mg in ZnO may deteriorate the crystalline quality and usually produces defects with the complication of lattice mismatch and Mg substitutional impurities, which limits the emission efficiency. It is therefore of high importance to study the influence of metals on the optical properties of MgZnO films in detail, which will be of great significance to their future applications in the field of MgZnO-based optoelectronics.

In this work, we have investigated the contribution of SP coupling and light extraction toward light enhancement of Pt nano-patterns (NPs) capped MgZnO films. Temperature dependence of the integrated photoluminescence (PL) intensity indicates that this improvement can be attributed to the surface modification and surface plasmonic coupling, while increased extraction efficiency caused by surface modification via Pt plays a dominant role as compared to SP coupling in increasing bandgap emission of MgZnO films. Here we give a detailed interpretation about the mechanism of this interesting phenomenon.

2. Experimental details

Mg_0.06Zn_0.94O films with thickness of ~200 nm were prepared on c-plane sapphire substrate by pulsed laser deposition in an ultrahigh vacuum chamber, and were then subsequently capped with Pt coatings. A monolayer of highly ordered polystyrene (PS) spheres (400 nm in diameter) was self-assembled on the surface of MgZnO sample. Pt was subsequently sputtered onto the monolayer PS/MgZnO sample. After sputtering, the PS spheres were removed by acetone dissolving for 2 min in an ultrasonic bath, then Pt NPs were formed on the top of MgZnO film. In addition, a Pt film was directly sputtered onto MgZnO film as the reference sample and the thickness of Pt (~8 nm) is the same as the maximum height of Pt NPs. All samples were cut from one MgZnO wafer and were identical. Atomic force microscopy (AFM) was utilized to characterize the morphology of the Pt NPs/MgZnO film. PL spectra were obtained with a 325 nm He–Cd laser excitation. The time resolved PL (TRPL) spectra were measured at room temperature to determine the decay dynamics by time-correlated single photon counting with a resolution of 10 ps. A 266 nm coherent femtosecond pulsed laser with a repetition rate of 76 MHz, a pulse width of 200 fs, and an excitation power of 10 mW was used as the excitation source.

3. Results and discussion

Figure 1(a) shows the room temperature PL spectra for bare MgZnO, MgZnO with Pt film, and MgZnO with Pt NPs. The inset Fig. 1(b) is AFM image (3D display) of MgZnO film with Pt NPs, and Fig. 1(c) is the schematic arrangement of the PL measurement configuration. Highly ordered Pt NPs with ~300 nm holes can be clearly observed from Fig. 1(b). For bare MgZnO film, a bandgap emission around 3.4 eV and a broad emission band around 2.3 eV were observed in the PL spectra. It has been recognized that in MgZnO films the broad PL band at 2.3 eV is related to oxygen vacancies or zinc interstitials [21]. As shown in Fig. 1(a), the integrated PL intensity of MgZnO coated with Pt film was four times stronger than that of bare MgZnO, while for Pt NPs/MgZnO, the integrated PL intensity is enhanced by as much as eight folds, which is remarkably larger than that of the Pt thin film/MgZnO. The inherent physical mechanism for enhanced bandgap emission is the coupling between excitons in MgZnO films and the SP of Pt patterns, passivation of surface states, as well as the enhanced light extraction from Pt NPs by the surface modification. It has been reported [5, 10] that the SP-exciton coupling results in a peak shift of the PL spectrum and emission enhancement. The lower PL peak energy and enhanced emission for Pt-coated samples shown in Fig. 1(a) suggests that the SP-exciton coupling process has made contribution. Since the SP coupling and light extracting phenomena are sensitive to metal surface morphology, it is reasonable to deduce that the strength of exciton-SP coupling and/or light extracting efficiency of the samples with Pt NPs and Pt film are different. Further discussion will be presented in detail for a deeper understanding this mechanism in later sections. For the bare MgZnO sample, defect emission is strong comparing with the bandgap emission. However, the defect emission decreases rapidly after Pt coating, and it is almost completely suppressed at the noise level for the MgZnO with Pt film, as illustrated in Fig. 1(a). This quenching of defect emission is possibly attributed to the reduction in surface defects after Pt coating.

 

Fig. 1 (a) Room temperature PL spectra for MgZnO with Pt NPs, Pt film, and bare MgZnO. The inset (b) is AFM image (3D display) of MgZnO with Pt NPs and the inset (c) is a schematic configuration for the PL measurement.

Download Full Size | PDF

Figure 2 shows the TRPL of bare and two MgZnO films with Pt coating. The decay curve can be well fitted by using the equation: $I_{(t)}=I_0{e}^{-t/\tau }$, where I and I₀ are the luminescence intensities at time t and t = 0, respectively, and τ is the decay time. It was observed that the bare MgZnO film has the fastest decay process with the decay time of 0.20 ns. The fitted decay times are increased to 1.21 ns for Pt film/MgZnO, and 0.70 ns for Pt NPs/MgZnO respectively. The PL decay time of bare MgZnO can be expressed as 1/${\tau }_{MgZnO}$ = 1/${\tau }_R$ + 1/${\tau }_{NR}$. While after Pt coating, the PL decay time can be expressed as 1/${\tau }_{Pt/MgZnO}$ = 1/${\tau }_R^{\ast }$ + 1/${\tau }_{SP}$ + 1/${\tau }_{NR}^{\ast }$, where ${\tau }_{NR}^{\ast }$ (${\tau }_{NR}$) and ${\tau }_R^{\ast }$ (${\tau }_R$) are the non-radiative and radiative decay times of MgZnO capped with (without) Pt, and 1/${\tau }_{SP}$ is the exciton-SP coupling rate. The incorporation of Mg in ZnO may deteriorate crystalline quality and usually produces defects. These surface states in the bare MgZnO film can act as non-radiative recombination or trapping centers. After capped with Pt, the surface states were remarkably reduced by passivation, in turn the non-radiative recombination or trapping centers decrease. Hence more incident photons can contribute to radiative recombination process, leading to the non-radiative recombination rate decreases and radiative recombination increases. Though the radiative recombination rate increased after Pt coating, surface defects passivation by Pt coating can significantly decrease the non-radiative recombination rate and play a dominant role in the decay channel. Hence the PL decay rates of MgZnO capped with Pt coatings are even smaller than bare MgZnO although the radiative recombination rate increases and there is an additional SP coupling decay channel. It is noted that the measured lifetime for the Pt film coated MgZnO is longer than that of the Pt NPs coated MgZnO, while the lifetime for a Pt film coated ZnO is shorter than that of a Pt NPs coated ZnO reported by K. W. Liu et al [7]. A possible reason for such a difference is that the surface defect states for MgZnO are different from those of ZnO. It was noted that the growth of the MgZnO alloys usually produces surface defects caused by Mg substitutional impurities, implying that more surface defects are possibly created in MgZnO film than in ZnO. Since the Pt coverage for the Pt film/MgZnO is much larger than the Pt NPs/MgZnO, it is reasonable to deduce that more surface defects are passivated for the Pt film/MgZnO sample, leading to a lower non-radiative recombination rate. Hence we observed a longer lifetime for Pt film coated MgZnO than Pt NPs coated sample. This can also explain the weaker defect-related emission from the Pt film/MgZnO sample than that from the Pt NPs/MgZnO as shown in Fig. 1(a).

 

Fig. 2 The TRPL decays of MgZnO, Pt film/MgZnO, and Pt NPs/MgZnO at 3.4 eV. The solid lines denote the fitting of each decay curve.

Download Full Size | PDF

As mentioned earlier, the SP coupling and light extracting are sensitive to metal surface morphology, therefore, in order to gain more insights of the enhanced mechanism and the contribution of SP coupling and light extraction toward light enhancement in the Pt coating/MgZnO samples, temperature dependent PL measurement was performed from 10 K to room temperature. Figure 3 shows the temperature dependence of the normalized integrated PL intensity for bare MgZnO, MgZnO with Pt NPs, and MgZnO with Pt film. It was found that the intensity of the MgZnO bandgap emission decreased remarkably with increasing temperature, which was partly due to the increased non-radiative recombination. As evidenced in Fig. 1(a), eight-fold and four-fold enhancements of bandgap emission from MgZnO film were observed by sputtering Pt NPs and Pt film onto MgZnO film, respectively. The increased bandgap emission PL intensity indicates an improvement in the external quantum efficiency (${\eta }_{ext}$) of MgZnO film with Pt coatings, which is related to the internal quantum efficiency (${\eta }_{int}$) and light extraction efficiency (${\eta }_{ext}{}_{raction}$) of lights from MgZnO to air by the following relation: ${\eta }_{ext}={\eta }_{int}\times {\eta }_{ext}{}_{raction}$. The ${\eta }_{int}$ of MgZnO film was estimated by comparing PL intensities assuming that the ${\eta }_{int}$ is 100% at 10 K regardless of excitation carrier density [5]. As Fig. 3 shows, the ${\eta }_{int}$ of MgZnO film, Pt film/MgZnO and Pt NPs/MgZnO was 4.3%, 8.5%, and 9.0%, respectively, indicating that the ${\eta }_{int}$ of MgZnO with Pt coating has increased doubly. The increase in ${\eta }_{int}$ after Pt coatings can be attributed to the reduction of surface defects as well as the coupling between excitons in MgZnO films and the SP of Pt patterns. The ${\eta }_{int}$ of bare MgZnO can be expressed as ${\eta }_{int(}{}_{MgZnO)}$ = (1/${\tau }_R$)/[(1/${\tau }_R$) + (1/${\tau }_{NR}$)]. While after Pt coating, the ${\eta }_{int}$ can be expressed as ${\eta }_{int(}{{}_{Pt/}}_{MgZnO)}$ = [(1/${\tau }_R^{\ast }$) + (1/${\tau }_{SP}$)] /[(1/${\tau }_R^{\ast }$) + (1/${\tau }_{SP}$) + (1/${\tau }_{NR}^{\ast }$)]. As mentioned previously, there are many surface states in the bare MgZnO film which act as non-radiative recombination centers. After capped with Pt layer, the surface states were passivated, in turn the non-radiative recombination rate decreases and radiative recombination increases. Moreover, the PL decay rates are enhanced through the exciton-SP coupling rate, this new recombination path increases the spontaneous recombination rate. Hence, the ${\eta }_{int(}{{}_{Pt/}}_{MgZnO)}$ was enhanced. We need to mention that the ${\eta }_{int}$of Pt film/MgZnO and Pt NPs/MgZnO were increased up to 2.0-fold and 2.1-fold respectively. Based on the observed increases of 4-fold and 8-fold for the ${\eta }_{ext}$ of Pt film/MgZnO and Pt NPs/MgZnO respectively, the increase in ${\eta }_{ext}{}_{raction}$ was estimated to be 2.0-fold and 3.8-fold from the equation ${\eta }_{ext}={\eta }_{int}\times {\eta }_{ext}{}_{raction}$ for Pt film/MgZnO and Pt NPs/MgZnO respectively. These results indicate that increased extraction efficiency caused by surface modification via Pt coating plays a dominant role in increasing bandgap emission of MgZnO film. This observation differs slightly from previous reports [5, 7]. In these earlier reports, it has been demonstrated that the enhancements in these types of UV emissions were mainly attributed to the coupling between plasmons of the attached metals and excitons.

 

Fig. 3 Temperature dependence of the integrated PL intensity for the three samples. PL integrated intensities at 10 K were normalized to 1.

Download Full Size | PDF

To get a clearer understanding of the giant PL intensity enhancement and the dominantly increased extraction efficiency from Pt/MgZnO system, we presented a hypothesis and illustrated it in Fig. 4 . Usually, exciton photo-generated in semiconductor materials decay through both radiative (k_rad) and non-radiative recombination (k_non) processes (Fig. 4(a)). Only a small fraction of the lights generated inside MgZnO film can escape because of total internal reflection at the interface of MgZnO and the outer medium air. A rough estimation for the MgZnO/air single interface leads to an extraction efficiency of about ${\eta }_{ext}$~1/4n²~6% [22]. For bare MgZnO film, the surface defect states are rather high according to the observation of a strong defect emission. After blanketing with Pt coating, the surface states were passivated, which in turn decreases the non-radiative recombination rate and increases radiative recombination. On the other hand, the Pt layer can increase the light ${\eta }_{ext}$ since the photons can be converted to free space radiation via scattering by rough surface of Pt layer (Fig. 4(b) and Fig. 4(c)). This implies that the increase of ${\eta }_{ext}{}_{raction}$ after Pt coating is attributed to the enhanced light extraction from Pt coating/MgZnO film by the surface modification. By proper engineering of the metal structures, light can be energy transferred from MgZnO film into a thin metal layer and scattered from the metal layer into free space, thereby enormously increase ${\eta }_{ext}{}_{raction}$of the light. While the distinction of ${\eta }_{ext}{}_{raction}$in the two metal patterns implies that, for Pt NPs, surface roughness of Pt film/MgZnO can enhance bandgap emission, but some of the energies can be thermally dissipated by non-radiative recombination (k_non) (Fig. 4(b)). As for the Pt NPs capping, the demonstrated eight-fold enhancement of bandgap emission is attributed to the fact that the periodic Pt NPs layer not only increase the ${\eta }_{int}$ but also offer a large amount of light transfer and scattering, which enormously increase the ${\eta }_{ext}{}_{raction}$ up to 3.8-fold (Fig. 4(c)).

 

Fig. 4 Schematic showing the enhancement mechanism of the bandgap emission for: (a) MgZnO, (b) Pt film/MgZnO, and (c) Pt NPs/MgZnO. K_rad: radiative recombination; k_non: non-radiative recombination.

Download Full Size | PDF

4. Conclusion

In summary, we present on a systematic study of the contribution of SP coupling and light extraction toward emission enhancement of Pt-capped MgZnO films. Temperature dependence of the integrated PL intensity indicates that this improvement can be attributed to the increased extraction efficiency as well as increased internal quantum efficiency due to the surface modification and surface plasmonic coupling, while increased light extraction efficiency caused by surface modification plays a dominant role in increasing bandgap emission of MgZnO films. The TRPL results indicate that the Pt coating can greatly reduce the non-radiative recombination rate by passivation of surface states, making the decay slow down.