Expand this Topic clickable element to expand a topic
Skip to content
Optica Publishing Group

Internal strain induced significant enhancement of deep ultraviolet light extraction efficiency for AlGaN multiple quantum wells grown by MOCVD

Open Access Open Access

Abstract

In this work, combined analysis of internal strain effects on optical polarization and internal quantum efficiency (IQE) were conducted for the first time. Deep ultraviolet light extraction efficiency of AlGaN multiple quantum wells (MQWs) have been investigated by means of polarization-dependent photoluminescence (PD-PL) and temperature-dependent photoluminescence (TD-PL). With the increase of compressive internal strain applied to the MQWs by an underlying n-AlGaN layer, the degree of polarization (DOP) of the sample was improved from −0.26 to −0.06 leading to significant enhancement of light extraction efficiency (LEE) as the PL intensity increased by 29.2% even though the internal quantum efficiency declined by 7.7%. The results indicated that proper management of the internal compressive strain in AlGaN MQWs can facilitate the transverse electric (TE) mode and suppress the transverse magnetic (TM) mode which could effectively reduce the total internal reflection (TIR) and absorption. This work threw light upon the promising application of compressively strained MQWs to reduce the wave-guide effect and improve the LEE of deep ultraviolet light emitting diodes (DUV LEDs).

© 2018 Optical Society of America under the terms of the OSA Open Access Publishing Agreement

1. Introduction

The AlGaN based DUV LEDs are energy-saving, environment-friendly and of long life span compared to conventional UV lamp. They are promising for a variety of applications including water purification, phototherapy and sterilization [1, 2]. But their further applications still held back due to the poor efficiency compared to the near UV or InGaN based visible LEDs owing to several issues such as high threading dislocation density, doping efficiency, quantum-confined Stark effect (QCSE) and anisotropic optical polarization property [1–4]. The three former issues were the common challenges that shared with InGaN based LEDs and numerous efforts have been made to relevant researches. With respect to anisotropic optical polarization, it is a unique property of AlGaN alloys owing to different structures of the valence subbands between AlN and GaN [4–6]. Specifically, due to the spin–orbit and crystal-field splitting effects, the topmost valence subband near the Γ point are heavy-hole (HH) for GaN and crystal-field split-off hole (CH) band for AlN. Such effects lead to the predominant optical polarization of light emissions are TE (E⊥c) and TM (E∥c) mode for GaN and AlN, respectively. The TM-polarized light propagates mainly in the lateral direction other than TE mode to the vertical direction. Therefore, TM mode undergoes strong effects of TIR due to the large incident angle and results in much lower light extraction efficiency [7]. The AlGaN based DUV LEDs intrinsically incorporated high Al content and emitted with predominant TM mode emission, suffering from the wave-guide effect a lot. It should be considered as an obstacle to enhance the LEE of DUV LEDs grown on c-plane substrates. Some previous researches have been done concerning the optical polarization dependence on strain, temperature, injection current and quantum well structure, respectively [8–13]. It is worthy to be noted that the strain issue in the epitaxy wafer is inevitable due to lattice mismatch between adjacent epitaxy layers with different Al contents. In typical flip-chip DUV LEDs, the underlying Si-doped AlGaN contact layer for the MQWs is of higher Al-content than the MQWs to avoid light absorption, meanwhile compressive internal strain is introduced to the MQWs. Therefore, it is of great sense to take advantage of this strain issue to benefit device performance during the epitaxy process. Several studies have revealed the relation between the strain and the optical polarization of AlGaN layers and MQWs [8–10, 14–17]. Our previous study has demonstrated that the internal compressive strain on the AlGaN epilayer would improve TE mode emission by modulating the band structure [15]. However, the previous works were limited to simulations or reporting the varying optical polarization. Lack of research has experimentally investigated the internal strain effect on the optical polarization combined with the IQE, thus demonstrated the enhancement of LEE in compressively strained AlGaN MQWs for DUV LEDs.

In this paper, to take advantage of this strain-related issue, tunable compressive internal strain were introduced to AlGaN MQWs by increasing Al-content of the underlying n-AlGaN layer. For the first time, combined analysis of optical polarization and IQE of the AlGaN MQWs was conducted under the internal strain effect. It showed that with the increase of compressive internal strain applied to the MQWs by the underlying n-AlGaN contact layer, the DOP of the sample was improved from −0.26 to −0.06, leading to significant enhancement of PL intensity even despite a lower IQE which directly demonstrated the improvement of LEE. The results indicated that proper increase of the internal compressive strain can facilitate the transverse electric (TE) mode and suppress the transverse magnetic (TM) mode which could effectively reduce the total internal reflection (TIR) and absorption in AlGaN MQWs.

2. Experiments

A series of AlGaN based MQWs were grown on c-plane sapphire substrates by metalorganic chemical vapor deposition (MOCVD). The wafers were simplified typical DUV LED structure except for electron blocking layer and p-GaN contact layer. Trimethylaluminum (TMA), triethylgallium (TEG) and ammonia (NH3) were used as precursors of Al, Ga and N, respectively. Hydrogen was the carrier gas. Firstly, 20 nm AlN was grown on c-plane sapphire substrate at 560 °C as the nucleation layer. Subsequently, a 800 nm AlN layer was grown at 1100 °C, followed by a Si-doped AlxGa1-xN layer with the thickness of 2.5 μm where x refers to 0.55 0.64 and 0.73 for sample A to sample C, respectively. Then, 5 periods of Si-doped Al0.35Ga0.65N/Al0.5Ga0.5N MQWs were deposited on the n-AlGaN layer followed by the last barrier of 12 nm-thick Si-doped Al0.5Ga0.5N. The nominal thickness of the wells and barriers were 2.5 nm and 12.5 nm. The carrier concentrations were estimated to be 2 × 1018 cm−3 in the n-AlGaN layer, 1 × 1018 cm−3 in the MQWs and 3 × 1017 cm−3 in the last barrier by Van-der-Pauw Hall measurement. It should be noted that the sample A with 0.55 Al-content of the n-AlGaN layer was the reference sample deduced from the flip-chip DUV LED structure.

HR-XRD 2θ-ω and ω rocking curve scans were performed by PANalytical X’pert PRO MRD Holland to investigate the structure and crystal quality of the samples. RSM of asymmetric (105) reflection was conducted to determine the strain condition of the MQWs. PL measurements were performed by using the 4th harmonic of a Q-switched Cr:YAG laser with the wavelength of 266 nm as an excitation source. The PD-PL setup was shown in our previous study [15]. To characterize the optical polarization, DOP was defined as [4],

DOP=IEc-IE//cIEc+IE//c
where IEc and IEc represented integrated PL intensity of TE mode and TM mode in this study, respectively. The TD-PL measurement was conducted by the same 266 nm excitation laser system joined by a closed-cycle Helium cryostat that was capable of varying the temperature from 15 K to 300 K.

3. Results and discussion

To qualitatively illustrate the sample structure, the XRD 2θ-ω scans of all the samples were shown in Fig. 1. The peaks of MQWs were slightly left shifted as 35.262°, 35.251° and 35.239° with the increase of Al content of n-AlGaN layer which should be attributed to strain-induced lattice deformation as discussed in other work [15]. The Al contents of n-AlGaN layer were determined as 0.55, 0.64 and 0.73 for sample A to sample C by XRD measurement. At least 3 satellite peaks can be resolved from the curves. The intensity of 2θ-ω scans for satellite peaks was relatively low due to only 5 periods of quantum wells in the samples. The full width at half maximum (FWHM) of ω scan rocking curve at (002) / (102) reflection for n-AlGaN layers were measured as 177/631, 186/713 and 181/838 for sample A to sample C, respectively, which were summarized in Table 1 together with other crucial experimental results. As the template of the MQWs, the crystal quality of n-AlGaN layer greatly affect the crystal quality of the MQWs.

 figure: Fig. 1

Fig. 1 HR-XRD 2θ-ω scan for all the samples. At least 3 satellite peaks can be resolved and the peak of MQWs slightly left shifted.

Download Full Size | PDF

Tables Icon

Table 1. Summary of the parameters for the samples.

RSM results were shown in Fig. 2 to illustrate the strain condition in the samples. As the AlGaN layer was 2.5 μm it should be considered fully relaxed. With respect to sample A as shown in Fig. 2(a), the peak of MQWs was totally merged in the peak of AlGaN so it was reasonable to assume the MQWs were fully strained to n-AlGaN. In Fig. 2(b) and 2(c), the MQWs and the n-AlGaN layer shared the same Qx coordinate that demonstrated pseudomorphical growth of MQWs. The in-plane strain εxx in the MQWs should be defined as [18],

εxx=aMQWsa0MQWsa0MQWs
where aMQWs was the lattice constant of MQWs measured by RSM and a0MQWs was the relaxed lattice constant which can be derived according to Vegard’s law. The strain conditions in sample A to sample C were calculated as −0.49%, −0.71% and −0.96%, respectively. The negative values indicated compressive strain. The TD-PL data from 15K to 300K were plotted in Fig. 3(a)-3(c) for sample A to sample C, respectively. No ‘S-shaped’ behavior of the peak energy was observed denoted weak exciton localization in the MQWs [19]. Therefore, the temperature induced band shrinkage was the dominant mechanism for the red-shift of the peaks. The IQE was estimated as the ratio between the integral PL intensity at 300 K to that at low temperature which were calculated as 29.9%, 35.5% and 27.6% for sample A to sample C, respectively. It should be noted that the nonradiative recombination cannot totally freeze out at 15 K so the calculated IQE should be in some degree overestimated [20]. However, they were still valid for comparison as they were acquired by steady state PL under the same measure condition of low excitation density. It was demonstrated that the IQE were mainly determined by the dislocation density which could be represented by the FWHM of rocking curves in this study. Meanwhile, the compressive strain induced piezoelectric polarization with opposite direction compared to the spontaneous polarization could relieve the tilt of energy band and suppress the QCSE, which would improve the IQE [21]. As the case of sample B, since the crystal quality did not decrease much compared to sample A, the strain effect in favor of the improvement of IQE. However, excessive strain would deteriorate the crystal quality owing to large lattice mismatch and introduce more nonradiative recombination centers that would lead to the decline of IQE as the case of sample C. Therefore, the increase of Al content in our experiment for the n-AlGaN layer should be well considered to avoid further deterioration of the crystal quality and further efforts could be made to optimize the epitaxy process. However, this declined IQE did not result in lower PL intensity which would be discussed in the following paragraph.

 figure: Fig. 2

Fig. 2 RSM of all the sample indicated pseudomorphical growth of MQWs on the n-AlGaN layer so that they shared the same in-plane lattice a.

Download Full Size | PDF

 figure: Fig. 3

Fig. 3 (a)-(c) TD-PL spectra plot against photon energy for sample A to sample C at temperature range from 15 K to 300 K.

Download Full Size | PDF

The optical polarization was investigated by the PD-PL setup to demonstrated the compressive strain effect. Owing to the symmetry of rotation angles, all the points tested through this setup should have a corresponding point in the range from 0 to 90 degrees. Integral PL intensities normalized by the value of TM mode from the lateral facet were plotted as a function of rotation angle of the Glan-Taylor prism shown in Fig. 4(a). To better exhibit the improvement of DOP, normalized integral intensity was plotted against the rotation angle in Fig. 4(b). The symbols indicated experimental data and the solid lines were the fitted curves of sine function. The TE mode was corresponded to 0 and 180 degrees in the polar coordinates while TM mode to 90 and 270 degrees. As analyzed from the data, the DOP of sample A to sample C was calculated by Eq. (1) to be −0.26, −0.13 and −0.06. The scale of DOP acquired in our work accorded well with the trend demonstrated in previous study [9]. The increase of TE mode was attributed to the escalated compressive strain induced valence band modification. The discrepancy between the topmost CH band and the subjacent HH band at the Brillouin zone center shrank owing to the compressive strain. Therefore the emission of TE mode which predominantly derived from the HH band would be relatively enhanced. The increase of TE mode can facilitate the light extraction pathway to the surface, as shown in the inset of Fig. 4(a), so that greatly suppress the wave-guide effect. Figure 4(c) depicted the dependence of DOP on the in-plane strain which exhibited approximate linear relation between them.

 figure: Fig. 4

Fig. 4 (a) Normalized integral intentisy of PD-PL plotted as a function of rotation angle range from 0 to 90. The inset showed the schematic propagation of TE and TM mode. (b) Normalized PD-PL intensity fitted with sine function showed in polar coordinates. (c) Plot of DOP as a function of in-plane strain.

Download Full Size | PDF

PL intensity from sample surface was depicted in Fig. 5 as a function of the wavelength.

 figure: Fig. 5

Fig. 5 PL intensity probed from the surface of each sample. The inset showed the IQE as well as the normalized PL and light extraction coefficient η of the samples for comparison.

Download Full Size | PDF

Compared to sample A, sample B and sample C showed 32.9% and 29.2% enhancement of the PL intensity. To better illustrate the improvement of LEE, a light extraction coefficient η was defined as,

η=cIPLIQE
where IPL is the PL intensity from the sample surface at room temperature, c is a parameter used to normalize the coefficient for better comparison which was calculated to be 0.3. The light extraction coefficient η was introduced as a modified expression of the relations between the external quantum efficiency (EQE), IQE and LEE in DUV LEDs [22]. As the assuption that IQE was 100% at low temperature, the LEE was comparable to EQE which could be expressed as the PL intensity at low temperature. The coefficient η combined the PL intensity and IQE where a large value of it represented a high LEE. By solving Eq. (3), η were achieved as 1, 1.12 and 1.40 for sample A to sample C, respectively. It demonstrated that with the escalation of in-plane strain, η improved 12% and 40% for sample B and sample C compared to sample A. We plotted the IQE, PL intensity and η in the inset of Fig. 5 for clear comparison in which the PL intensity and η were normalized by the value of that for sample A. For all the three samples, the value of η increase with the compressive strain other than vary with IQE or PL intensity which demonstrate the dependence of light extraction on compressive strain and DOP was significant. The worth noted point was that with respect to sample C, although the IQE was declined by 7.7% than that of sample A, the PL intensity was remarkably enhanced by 29.2% and the η was as high as 1.4. It demonstrated that the strain induced improvement of DOP was able to enhance the LEE dramatically because TE mode was more easily to be extracted from the surface than TM mode [7]. Which would potentially give rise to DUV LEDs with high light output power.

4. Conclusion

In conclusion, the tunable compressive internal strain in AlGaN MQWs was introduced by properly increasing the Al content of underlying n-AlGaN contact layer in order to investigate its impact on the light extraction efficiency. The sample structure, strain condition, IQE and DOP of each sample was investigated by means of HR-XRD, RSM, TD-PL and PD-PL. The increase of compressive strain induced the improvement of TE mode which would suppress the wave-guide effect that was to blame for the loss of light emission in high Al content MQWs with predominant TM mode emission. The LEE exhibited significant enhancement as the increase of PL intensity even despite the decline of IQE due to the compressive strain. The results provided a promising application of compressively strained MQWs with high light extraction efficiency which could potentially lead to the fabrication of DUV LEDs with high light output power.

Funding

Key Project of Chinese National Development Programs (Grant No. 2016YFB0400901, 2016YFB0400804); Key Laboratory of Infrared Imaging Materials and Detectors, Shanghai Institute of Technical Physics, Chinese Academy of Sciences (Grant No. IIMDKFJJ-15-07); National Natural Science Foundation of China (Grant No. 61675079, 11574166, 61377034, 61774065); Director Fund of WNLO.

References and links

1. A. Khan, K. Balakrishnan, and T. Katona, “Ultraviolet light-emitting diodes based on group three nitrides,” Nat. Photonics 2(2), 77–84 (2008). [CrossRef]  

2. H. Hirayama, S. Fujikawa, N. Noguchi, J. Norimatsu, T. Takano, K. Tsubaki, and N. Kamata, “222-282 nm AlGaN and InAlGaN-based deep-UV LEDs fabricated on high-quality AlN on sapphire,” Nat. Photonics 2(2), 77–84 (2009).

3. Z.-H. Zhang, Y. Zhang, W. Bi, H. V. Demir, and X. W. Sun, “On the internal quantum efficiency for InGaN/GaN light-emitting diodes grown on insulating substrates,” Optik -physica status solidi (a) 206(6), 1176–1182 (2016).

4. K. B. Nam, J. Li, M. L. Nakarmi, J. Y. Lin, and H. X. Jiang, “Unique optical properties of AlGaN alloys and related ultraviolet emitters,” physica status solidi (a) 213(12), 3078-3102 (2004).

5. I. Vurgaftman and J. R. Meyer, “Band parameters for nitrogen-containing semiconductors,” J. Appl. Phys. 94(6), 3675–3696 (2003). [CrossRef]  

6. H. Lu, T. Yu, G. Yuan, C. Jia, G. Chen, and G. Zhang, “Valence subband coupling effect on polarization of spontaneous emissions from Al-rich AlGaN/AlN Quantum Wells,” Opt. Express 20(25), 27384–27392 (2012). [CrossRef]   [PubMed]  

7. H.-Y. Ryu, I.-G. Choi, H.-S. Choi, and J.-I. Shim, “Investigation of Light Extraction Efficiency in AlGaN Deep-Ultraviolet Light-Emitting Diodes,” Appl. Phys. Express 6(6), 062101 (2013). [CrossRef]  

8. S. Fan, Z. Qin, C. He, X. Wang, B. Shen, and G. Zhang, “Strain effect on the optical polarization properties of c-plane Al0.26Ga0.74N/GaN superlattices,” Opt. Express 22(6), 6322–6328 (2014). [CrossRef]   [PubMed]  

9. J. E. Northrup, C. L. Chua, Z. Yang, T. Wunderer, M. Kneissl, N. M. Johnson, and T. Kolbe, “Effect of strain and barrier composition on the polarization of light emission from AlGaN/AlN quantum wells,” Appl. Phys. Lett. 100(2), 021101 (2012). [CrossRef]  

10. T. Kolbe, A. Knauer, C. Chua, Z. Yang, V. Kueller, S. Einfeldt, P. Vogt, N. M. Johnson, M. Weyers, and M. Kneissl, “Effect of temperature and strain on the optical polarization of (In)(Al)GaN ultraviolet light emitting diodes,” Appl. Phys. Lett. 99(26), 261105 (2011). [CrossRef]  

11. M. Hou, Z. Qin, C. He, J. Cai, X. Wang, and B. Shen, “Effect of injection current on the optical polarization of AlGaN-based ultraviolet light-emitting diodes,” Opt. Express 22(16), 19589–19594 (2014). [CrossRef]   [PubMed]  

12. W. Wang, H. Lu, L. Fu, C. He, M. Wang, N. Tang, F. Xu, T. Yu, W. Ge, and B. Shen, “Enhancement of optical polarization degree of AlGaN quantum wells by using staggered structure,” Opt. Express 24(16), 18176–18183 (2016). [CrossRef]   [PubMed]  

13. T. Kolbe, A. Knauer, C. Chua, Z. Yang, S. Einfeldt, P. Vogt, N. M. Johnson, M. Weyers, and M. Kneissl, “Optical polarization characteristics of ultraviolet (In)(Al)GaN multiple quantum well light emitting diodes,” Appl. Phys. Lett. 97(17), 171105 (2010). [CrossRef]  

14. T. K. Sharma, D. Naveh, and E. Towe, “Strain-driven light-polarization switching in deep ultraviolet nitride emitters,” Phys. Rev. B 84(3), 035305 (2011). [CrossRef]  

15. H. Long, F. Wu, J. Zhang, S. Wang, J. Chen, C. Zhao, Z. C. Feng, J. Xu, X. Li, J. Dai, and C. Chen, “Anisotropic optical polarization dependence on internal strain in AlGaN epilayer grown on AlxGa1−xN templates,” J. Phys. D Appl. Phys. 49(41), 415103 (2016). [CrossRef]  

16. C. Reich, M. Guttmann, M. Feneberg, T. Wernicke, F. Mehnke, C. Kuhn, J. Rass, M. Lapeyrade, S. Einfeldt, A. Knauer, V. Kueller, M. Weyers, R. Goldhahn, and M. Kneissl, “Strongly transverse-electric-polarized emission from deep ultraviolet AlGaN quantum well light emitting diodes,” Appl. Phys. Lett. 107(14), 142101 (2015). [CrossRef]  

17. D. Fu, R. Zhang, B. G. Wang, B. Liu, Z. L. Xie, X. Q. Xiu, H. Lu, Y. D. Zheng, and G. Edwards, “Ultraviolet emission efficiencies of AlxGa1−xN films pseudomorphically grown on AlyGa1−yN template (x<y) with various Al-content combinations,” Thin Solid Films 519(22), 8013–8017 (2011). [CrossRef]  

18. M. A. Moram and M. E. Vickers, “X-ray diffraction of III-nitrides,” Rep. Prog. Phys. 72(3), 036502 (2009). [CrossRef]  

19. F. Wu, H. Sun, I. A. Ajia, I. S. Roqan, D. Zhang, J. Dai, C. Chen, Z. C. Feng, and X. Li, “Significant internal quantum efficiency enhancement of GaN/AlGaN multiple quantum wells emitting at ~350 nm via step quantum well structure design,” J. Phys. D Appl. Phys. 50(24), 245101 (2017). [CrossRef]  

20. M. Shatalov, J. Yang, W. Sun, R. Kennedy, R. Gaska, K. Liu, M. Shur, and G. Tamulaitis, “Efficiency of light emission in high aluminum content AlGaN quantum wells,” J. Appl. Phys. 105(7), 073103 (2009). [CrossRef]  

21. V. Fiorentini, F. Bernardini, and O. Ambacher, “Evidence for nonlinear macroscopic polarization in III–V nitride alloy heterostructures,” Appl. Phys. Lett. 80(7), 1204–1206 (2002). [CrossRef]  

22. M. Kneissl, T. Kolbe, C. Chua, V. Kueller, N. Lobo, J. Stellmach, A. Knauer, H. Rodriguez, S. Einfeldt, Z. Yang, N. M. Johnson, and M. Weyers, “Advances in group III-nitride-based deep UV light-emitting diode technology,” Semicond. Sci. Technol. 26(1), 014036 (2011). [CrossRef]  

Cited By

Optica participates in Crossref's Cited-By Linking service. Citing articles from Optica Publishing Group journals and other participating publishers are listed here.

Alert me when this article is cited.


Figures (5)

Fig. 1
Fig. 1 HR-XRD 2θ-ω scan for all the samples. At least 3 satellite peaks can be resolved and the peak of MQWs slightly left shifted.
Fig. 2
Fig. 2 RSM of all the sample indicated pseudomorphical growth of MQWs on the n-AlGaN layer so that they shared the same in-plane lattice a.
Fig. 3
Fig. 3 (a)-(c) TD-PL spectra plot against photon energy for sample A to sample C at temperature range from 15 K to 300 K.
Fig. 4
Fig. 4 (a) Normalized integral intentisy of PD-PL plotted as a function of rotation angle range from 0 to 90. The inset showed the schematic propagation of TE and TM mode. (b) Normalized PD-PL intensity fitted with sine function showed in polar coordinates. (c) Plot of DOP as a function of in-plane strain.
Fig. 5
Fig. 5 PL intensity probed from the surface of each sample. The inset showed the IQE as well as the normalized PL and light extraction coefficient η of the samples for comparison.

Tables (1)

Tables Icon

Table 1 Summary of the parameters for the samples.

Equations (3)

Equations on this page are rendered with MathJax. Learn more.

DOP= I Ec - I E//c I Ec + I E//c
ε xx = a MQWs a 0MQWs a 0MQWs
η=c I PL IQE
Select as filters


Select Topics Cancel
© Copyright 2024 | Optica Publishing Group. All rights reserved, including rights for text and data mining and training of artificial technologies or similar technologies.