Transition mechanism of InAs/GaAs quantum-dot infrared photodetectors with
different InAs coverages
Chi-Che Tseng, Tung-Hsun Chung, Shu-Cheng Mai, Kuang-Ping Chao, Wei-Hsun Lin, Shih-Yen Lin, and Meng-Chyi Wu
Citation: Journal of Vacuum Science & Technology B 28, C3G28 (2010); doi: 10.1116/1.3368607 View online: http://dx.doi.org/10.1116/1.3368607
View Table of Contents: http://scitation.aip.org/content/avs/journal/jvstb/28/3?ver=pdfcov Published by the AVS: Science & Technology of Materials, Interfaces, and Processing
Articles you may be interested in
The influence of In composition on InGaAs-capped InAs/GaAs quantum-dot infrared photodetectors J. Appl. Phys. 106, 054512 (2009); 10.1063/1.3212983
Growth of high quality InAs quantum-dot multilayer structures on InP for infrared photodetector applications J. Vac. Sci. Technol. B 24, 1656 (2006); 10.1116/1.2201452
Influence of rapid thermal annealing on a 30 stack InAs/GaAs quantum dot infrared photodetector J. Appl. Phys. 94, 5283 (2003); 10.1063/1.1609634
Transport characteristics of InAs/GaAs quantum-dot infrared photodetectors Appl. Phys. Lett. 83, 752 (2003); 10.1063/1.1594285
Near-room-temperature operation of an InAs/GaAs quantum-dot infrared photodetector Appl. Phys. Lett. 78, 2428 (2001); 10.1063/1.1362201
with different InAs coverages
Chi-Che Tseng
Institute of Photonics Technologies, National Tsing Hua University, Hsinchu 300, Taiwan
Tung-Hsun Chung, Shu-Cheng Mai, Kuang-Ping Chao, and Wei-Hsun Lin
Institute of Electronics Engineering, National Tsing Hua University, Hsinchu 300, Taiwan
Shih-Yen Lina兲
Research Center for Applied Sciences, Academia Sinica, Nankang, Taipei 11529, Taiwan, Department of Photonics, National Chiao-Tung University, Hsinchu 300, Taiwan,
and Institute of Optoelectronic Sciences, National Taiwan Ocean University, Keelung 20224, Taiwan
Meng-Chyi Wu
Institute of Photonics Technologies, National Tsing Hua University, Hsinchu 300, Taiwan and Institute of Electronics Engineering, National Tsing Hua University, Hsinchu 300, Taiwan
共Received 22 September 2009; accepted 1 March 2010; published 25 March 2010兲
In this article, the authors investigate the influences of different InAs coverages on the photoluminescence excitation 共PLE兲 spectra and spectral responses of InAs/GaAs quantum-dot infrared photodetectors共QDIPs兲. An increase in InAs coverage would lead to an increase in energy separation between heavy-hole state and light-hole state in the wetting layer共WL兲 region in the QD PLE spectra. The results suggest that most of the strain resulted from the InAs/GaAs lattice mismatch may be accumulated in the WL instead of the QD region. Also observed are the similar energy separations of energy levels responsible for the intraband absorption in the PLE spectra of the QDIPs such that similar detection wavelengths are observed for the devices. © 2010 American
Vacuum Society.关DOI: 10.1116/1.3368607兴
I. INTRODUCTION
The growth of self-assembled 共InGa兲As quantum dots 共QDs兲 on GaAs substrates for practical applications and fun-damental research have been widely investigated in recent years.1–3The influence of strain accumulation on energy lev-els has also been investigated.4 For practical applications, long-wavelength laser diodes based on QD structures5 and quantum-dot infrared photodetectors 共QDIPs兲 have been developed.1,2,6The performance of the devices suggests that strain reduction in QDs plays a key role in obtaining good optical characteristics of the QDIP structure. Beside the en-ergy levels in QD structures, the enen-ergy level that resulted from the two-dimensional wetting layer 共WL兲 is also fre-quently observed in related reports.7,8The heavy-hole 共HH兲 and light-hole 共LH兲 splittings that resulted from the strain accumulation have been reported elsewhere.9,10 Theoretical calculation has been proposed to derive the wetting layer thickness according to the HH-LH energy splitting.10 How-ever, until now, there is no report regarding the influence of strain in QD structures on the performance of practical de-vices such as QDIPs.
In this article, the photoluminescence excitation 共PLE兲 spectra and spectral response of InAs/GaAs QDIPs with dif-ferent coverages are investigated. Increasing the energy sepa-ration between HH state and LH state in the WL region is observed in the PLE spectra with increasing InAs coverage of the QD structures. The results imply that most of the strain
that resulted to the InAs/GaAs lattice mismatch may be ac-cumulated in the WL instead of the QD region. Further in-vestigation is still required to confirm this conclusion. The multilongitudinal optical共multi-LO兲 phonon peaks observed in the PLE spectra are similar to the phonon energies of the InAs bulk material, which further confirms the attribution. A model is also proposed for the transition mechanisms respon-sible for the peaks observed in the PLE spectra. The similar energy separations of energy levels responsible for the intra-band absorption for QDIPs with different InAs coverages indicate in part why most QDIPs could only operate in the midwavelength infrared range共MWIR兲 共3–5 m兲.
II. EXPERIMENT
The QDIP samples investigated here were grown on 共100兲-oriented semi-insulating GaAs substrates in a Riber Compact 21 solid-source molecular beam epitaxy system. Sandwiched between 0.3 and 0.6 m thick GaAs contact layers with n = 1⫻1018 cm−3, ten-period InAs/GaAs QDs
with 1.5, 2.0, 2.5, and 3.0 ML 共monolayer兲 InAs coverages and 30 nm GaAs barriers are prepared. The sample structures are shown in TableI. The PL and PLE spectra of the QDIP samples were measured at 10 K by using a Jobin Yvon’s NanoLog3 system with a tungsten-halogen lamp as the light source. For the measurements of spectral responses, the QDIP samples were fabricated into 100⫻100 m2 devices
using standard photolithographic techniques, contact metal evaporation, and wet chemical etching. The measurement
system for spectral response consists of a Spectral 100 Fou-rier transformation infrared spectrometer with a cryostat and a current preamplifier.11
III. RESULTS AND DISCUSSION
Normalized 10 K PL spectra of the QDIP samples with different InAs coverages are shown in Fig.1. The decrease in PL peak energy at 1.135, 1.116, 1.108, and 1.092 eV is ob-served with increasing InAs coverage. For the sample with a 1.5 ML InAs coverage, the luminescence at 1.24 eV is attrib-uted to higher order excited state of 1.5 ML InAs QDs. Since relatively weaker QD luminescence intensity is observed for the sample, the luminescence of the n-type GaAs contact layer is observed. To further investigate the PL peak shift with different InAs coverages, there are two major mecha-nisms proposed for the energy level transitions in the QD structures. One mechanism is related to the height of QDs, where the lower energy levels are expected for taller QDs. The other mechanism is related to strain accumulation in the QD structures, where a band gap broadening is expected with larger compressive strain accumulation. The larger redshift of PL peak wavelength with a higher InAs coverage for the QDIP samples is attributed to the former mechanism. The results may also indicate that strain accumulation does not play a significant role in the QD structures with increasing InAs coverage. If the strain does accumulate in the QD struc-tures with increasing InAs coverage, the influence of both mechanisms may cancel out each other such that the mono-tonically decrease in PL peak energy could not be observed. To further investigate the optical characteristics of the QDIP samples, the 10 K PLE spectra with the PL peak en-ergy as the detection enen-ergy are shown in Fig. 2共a兲. It has been reported elsewhere that the PLE peak wavelength of
QDs will change with different detection energies.3,12 There-fore, although not shown here, the PLE peaks of QD1 chang-ing under different detection energies are related to the ab-sorptions in the QD structures while the peaks of WL1 and WL2 with unchanged location resulted from the QW-like InAs WLs, as shown in Fig.2共a兲. The GaAs band-gap-edge absorption is also observed in the spectra, which is labeled as
EGaAs. By setting the detection energies as zero, the shifted
PLE spectra of the QDIP samples are shown in Fig.2共b兲. An inspection of Fig. 2共b兲 reveals that the peaks with the same energy separation of 32 meV are observed in the spectra for the samples with different InAs coverages. The peaks are attributed to the involvement of 1LO and 2LO phonons in the QD ground-state luminescence. Considering the LO pho-ton energy of bulk InAs as 28 meV, similar phonon energies of the samples indicate that strain accumulation in the QDs does not increase with increasing the InAs coverage. The results are consistent with the previous assumption observed from the PL spectra.
To explain the transition mechanisms of the QD PLE spectra, a simplified band structure and the different transi-tion mechanisms in PL/PLE spectra of the QDIP samples are shown in Fig. 3. Assuming that most of the compressive strain resulted from the InAs/GaAs lattice mismatch is accu-mulated at the WL, HH-LH splitting can be observed in the valence band of WL. For the InAs QDs, since no pronounced strain accumulation is observed in the investigated InAs cov-erage range of 1.5–3.0 ML, HH-LH degeneracy should still TABLEI. Sample structure of the QDIPs with different InAs coverages.
Top contact 300 nm GaAs n = 2⫻1018 cm−3
30 nm GaAs共ten times兲 Undoped
InAs QDs共ML兲 共ten times兲 1.5 2.0 2.5 3.0
30 nm GaAs Undoped
Bottom contact 600 nm GaAs n = 2⫻1018 cm−3
Substrate 350 m共100兲 semi-insulating GaAs
1.05 1.10 1.15 1.20 1.25 1.30 1.5 ML 2.0 ML 2.5 ML 3.0 ML
Photon Energy (eV)
PL
Intensity
(a
.u.)
10 K
FIG. 1. Normalized 10 K spectra of the QDIP samples with different InAs coverages. 1.1 1.2 1.3 1.4 1.5 1.6 WL Region QD Region WL2 WL1 QD1 3.0 ML 2.5 ML 2.0 ML 1.5 ML 10 K
Photon Energy (eV)
PLE In te ns it y (a.u .) EGaAs (a) 0.06 0.12 0.18 0.24 10 K 3.0 ML 2.5 ML 2.0 ML 1.5 ML 2LO Energy (eV) PLE In te ns it y (a.u .) 1LO (b)
FIG. 2. 共a兲 10 K PLE spectra with the PL peak energy as the detection energy and共b兲 shifted PLE spectra by setting the detection energies as zero of the QDIP samples.
C3G29 Tseng et al.: Transition mechanism of InAs/GaAs quantum-dot infrared photodetectors C3G29
be observed. In this case, three states EHH-LH,QD, EHH,WL, and
ELH,WL, which correspond to the QD ground state, WL HH
state, and WL LH state, respectively, are obtained in the valence band. In the conduction band, the three states EQD,0,
EQD,1, and EWL correspond to QD ground state, QD first
excited state, and the WL state, respectively, as depicted in this figure. In addition, the optical recombination process of 共a兲 EQD,0− EHH-LH,QDwould result in the PL peaks observed
in Fig. 1. For the absorption processes, there are three al-lowed transitions of 共b兲 EHH,WL− EQD,1, 共c兲 EHH,WL− EWL,
and 共d兲 ELH,WL− EWL in the structure. For absorption 共b兲,
although the wave functions of the two states are of different parities, the transition should still be allowed. The reasons are 共a兲 the initial state EHH,WL is at the WL region and the
destination state EQD,1 is at the QD region and 共b兲 QD and WL structures are adjacent to each other. In this case, the wave function overlapping is not negligible such that the absorption process共b兲 would still be observed.
Absorptions共b兲, 共c兲, and 共d兲 correspond to the three PLE peaks of QD1, WL1, and WL2, respectively, as shown in Fig.2共a兲. In this case, the energy difference between absorp-tions 共c兲 and 共d兲 would be equal to the HH-LH splitting in the WL. Therefore, as shown in Fig. 2共a兲, the increase in HH-LH splitting 48, 62, 63, and 69 meV with increasing the InAs coverage is observed for the QDIP samples. It implies that an increase in InAs coverage would lead to pronounced compressive strain accumulation in the WL region. Com-bined with the previous assumptions, it could be concluded that with increasing the InAs coverage, most of the compres-sive strains resulted from the InAs/GaAs lattice mismatch will be accumulated in the WL instead of the QD region. This phenomenon would result in the increase in HH-LH
splitting in the WL, similar phonon energy with the bulk InAs material in the QDs, and the redshift of PL peak wave-length with increasing InAs coverage, as discussed above. To confirm this conclusion, more detailed investigation is still required in the future.
The 10 K spectral responses of the four samples biased at 0.6 V are shown in Fig.4. As shown in this figure, similar responses in the range of 3 – 9 m with peak detection wavelengths of ⬃6 m are observed for the four samples. The broad detection wavelengths are frequently observed for the InAs/GaAs QDIPs. The phenomenon indicates that al-though the QDIPs used different InAs coverages, the energy difference between the confinement states responsible for the intraband transitions is almost the same for the four samples. Going back to the 10 K PLE spectra shown in Fig. 2共a兲, another interesting phenomenon observed in the spectra is the similar energy differences of 0.22, 0.21, 0.2, and 0.2 eV between PLE peaks of QD1 and WL1 for the samples with 1.5, 2.0, 2.5, and 3.0 ML InAs coverages. The values are quite close to the peak responses of the devices at ⬃6 m 共0.21 eV兲. As described in the previous paragraph, the energy difference between PLE peaks of QD1 and WL1 would cor-respond to the energy difference in EQD,1− EWL. In this case, it is reasonable to conclude that for the InAs/GaAs QDIPs,
EQD,1− EWLtransitions are responsible for the observed
spec-tral responses. Since most of the strain would be accumu-lated in the WL instead of the QD region, energy level low-ering with increasing dot height will be observed with increasing InAs coverage. The broadening of the HH-LH splitting would compensate the lowering of the QD energy levels such that similar energy differences are observed for the QDIP samples with different InAs coverages. In this case, similar detection wavelengths are observed for QDIPs with different InAs coverages. The results explain in part why most QDIPs would only operate in the MWIR range 共3–5 m兲.
IV. CONCLUSIONS
In conclusion, we have demonstrated the effects of cover-age on the PL/PLE and spectral responses of InAs/GaAs QDIPs. An increase in InAs coverage would lead to an in-crease in HH-LH splitting in the WL, similar phonon energy ELH,WL EHH,WL EHH-LH,QD EQD,0 EQD,1 EWL EGaAs (a) (b) (c) (d)
(a) EQD,0-EHH-LH,QDobserved in the PL (b) EHH,WL-EQD,1observed in the PLE (c) EHH,WL-EWLobserved in the PLE (d) ELH,WL-EWLobserved in the PLE
FIG. 3. Simplified band structures and the different transition mechanisms in PL/PLE spectra of the QDIP samples.
2 3 4 5 6 7 8 9 10 1.5 ML 2.0 ML 2.5 ML 3.0 ML Wavelength (μμm) 10 K 0.6 V Respons iv ity (a.u. )
with the bulk InAs material in the QDs, and a redshift of PL peak wavelength. Most of the compressive strains are thus accumulated in the WL instead of the QD region. A correla-tion between the PLE spectra and the spectral responses of the devices is also established. The transition between the QD first excited state and the WL state is responsible for the responses of the InAs/GaAs QDIPs. The similar detection wavelengths of the devices suggest that if longer detection wavelengths are expected to be achieved for QDIPs, it is necessary to insert an addition energy level between the QD first excited state and the WL state.
ACKNOWLEDGMENT
This work was supported in part by the National Science Council, Taiwan under Grant No. NSC 98-2221-E-001-001.
1C. C. Tseng et al., IEEE Photonics Technol. Lett. 20, 1240共2008兲. 2S. T. Chou, M. C. Wu, S. Y. Lin, and J. Y. Chi, Appl. Phys. Lett. 88,
173511共2006兲.
3D. Bimberg, M. Grundmann, and N. N. Ledentsov, Quantum Dot
Hetero-structures共Wiley, Chichester, 1999兲.
4D. M. T. Kuo and Y. C. Chang, Phys. Rev. Lett. 99, 086803共2007兲. 5W. S. Liu, D. M. T. Kuo, J. I. Chyi, W. Y. Chen, H. S. Chang, and T. M.
Hsu, Appl. Phys. Lett. 89, 243103共2006兲.
6S. F. Tang, S. Y. Lin, and S. C. Lee, Appl. Phys. Lett. 78, 2428共2001兲. 7H. C. Lin, M. Gao, J. McCaffrey, Z. R. Wasilewski, and S. Fafard, Appl.
Phys. Lett. 78, 79共2001兲.
8C. Y. Huang, T. M. Ou, S. T. Chou, C. S. Tsai, M. C. Wu, S. Y. Lin, and
J. Y. Chi, J. Vac. Sci. Technol. B 23, 1909共2005兲.
9
G. SJk, P. Poloczek, K. Ryczko, J. Misiewicz, A. Löffler, J. P. Reithmaier, and A. Forchel, J. Appl. Phys. 100, 103529共2006兲.
10A. S. Bhatti, M. Grassi Alessi, M. Capizzi, P. Frigeri, and S. Franchi,
Phys. Rev. B 60, 2592共1999兲.
11S. T. Chou, C. C. Tseng, C. N. Chen, W. H. Lin, S. Y. Lin, and M. C. Wu,
Appl. Phys. Lett. 92, 253510共2008兲.
12R. Heitz, M. Veit, N. N. Ledentsov, A. Hoffmann, D. Bimberg, V. M.
Ustinov, P. S. Kop’ev, and Zh. I. Alferov, Phys. Rev. B 56, 10435共1997兲.
C3G31 Tseng et al.: Transition mechanism of InAs/GaAs quantum-dot infrared photodetectors C3G31
