Materials Science and Engineering B107 (2004) 66–69
Comparisons of InP/InGaAlAs and InAlAs/InGaAlAs
distributed Bragg reflectors grown by metalorganic
chemical vapor deposition
T.C. Lu, J.Y. Tsai, H.C. Kuo, S.C. Wang
∗Institute of Electro-Optical Engineering, National Chiao Tung University, 1001 Ta Hsueh Road, Hsinchu 30050, Taiwan, ROC Received 3 September 2003; accepted 13 October 2003
Abstract
Long wavelength vertical cavity surface emitting lasers (VCSELs) are considered the best candidates for the low cost reliable light emitters in fiber communications. The low refractive index contrast in the conventional InP-based lattice-matched distributed Bragg reflectors (DBRs), InP/InGaAsP, impeded the development of 1.3–1.5m VCSELs. However, the monolithic InP-based lattice-matched DBRs are still most attractive and desirable. The InP/InGaAlAs and InAlAs/InGaAlAs DBRs with larger refractive index contrast than the conventional InP/InGaAsP DBRs have been demonstrated recently. In this report, we compare these two material systems in terms of optical and electrical properties of DBRs. We found the InP/InGaAlAs DBRs have better electrical and optical properties, while the InAlAs/InGaAlAs DBRs have much lower growth complexity.
© 2003 Elsevier B.V. All rights reserved.
Keywords: Metalorganic chemical vapor deposition; Distributed Bragg reflectors
1. Introduction
Long wavelength (1.3–1.5m) vertical cavity surface emitting lasers (VCSELs) are very attractive light sources for fiber communications due to the advantages of sin-gle longitudinal mode, small divergence circular emission beam profile, low power consumption and low cost re-liable productions. However, the lack of high refractive index contrast materials for the distributed Bragg reflectors (DBRs) makes the development of long wavelength VC-SELs lagging behind the short wavelength (0.78–0.98m) VCSELs. In the past few years, high performance long wavelength VCSELs[1,2]were demonstrated using wafer fusion technique, which integrated the InP-based active lay-ers and GaAs-based DBRs together. But the capability of mass production could be an issue. Recently, the InGaNAs 1.3m VCSELs grown on GaAs substrates have been demonstrated with excellent characteristics[3,4]. However, to extend the InGaNAs gain to beyond 1.5m is still rather difficult.
∗Corresponding author. Tel.:+886-3-5712121x56320; fax:+886-3-5716631.
E-mail address: [email protected] (S.C. Wang).
Monolithically grown DBRs lattice-matched to InP sub-strates continues to attract interests due to the existing highly efficient InGaAsP and InGaAlAs gain materials with the wavelength window covering from 1.3 to 1.8m. The Sb-based DBRs with large refractive index contrasts
of n ranging from 0.43 to 0.44 have been successfully
applied in the VCSEL structures [5,6]. However, these DBRs have drawbacks such as low thermal conductivity and relatively high growth complexity. The conventional InP-based lattice-matched InP/InGaAsP DBR has the prob-lem of the small refractive index contrast (n = 0.27 for InP/InGaAsP) that requires a larger number of DBR pairs to obtain high reflectivity. In addition, the conventional DBRs not only increase the penetration depth causing more absorption, but also create the heat dissipation problem.
Recently, the DBRs based on relatively large refrac-tive index contrast material combinations of InP/InGaAlAs
(n = 0.34) and InAlAs/InGaAlAs (n = 0.3) have been
demonstrated[7–10]. However, the systematic comparisons of these two material combinations have not been re-ported. In this study, we used the low pressure metalorganic chemical vapor deposition (MOCVD) system to grow the InP/InGaAlAs and InAlAs/InGaAlAs DBRs. We compared these two material systems in terms of electrical and
op-0921-5107/$ – see front matter © 2003 Elsevier B.V. All rights reserved. doi:10.1016/j.mseb.2003.10.067
T.C. Lu et al. / Materials Science and Engineering B107 (2004) 66–69 67
tical properties of DBRs. In addition, due to the different group V-based materials of InP and InGaAlAs, the epitaxial growth technique and the complexity will also be discussed. We found that the InP/InGaAlAs DBRs have better optical and electrical properties, while the InAlAs/InGaAlAs DBRs have much lower growth complexity.
2. Experimental procedure
All DBR structures were grown in a vertical-type low pressure MOCVD system with a rotating disk. The disk ro-tated at 900 rounds per minute to maintain the laminar gas flow. The growth pressure was 9.33 kPa. The growth tem-perature was 625◦C. V/III ratio was 150 for InP and 200 for InGaAlAs. The growth rate was 36.5 and 35 nm/min for InP and InGaAlAs, respectively. The alkyl sources were trimethylindium, trimethylgallium, and trimethylaluminum, and the group V gases were AsH3 and PH3. The carrier
gas was hydrogen. Si2H6was used as the precursor of the
n-type dopant. The epitaxial layers were all grown on n-type (1 0 0)InP substrates. For the growth of InP/InGaAlAs DBRs, we adapted the growth interruption technique we established earlier[11]and chose the interruption time tpof
0.3 min to make compromises between the total growth time, the amount of source usage and the required high reflectiv-ity. As for the growth of the InAlAs/InGaAlAs DBRs, the epitaxial layers of InAlAs/InGaAlAs were directly switched since both materials have the same group V source.
The thickness of each layer and the vertical compo-sitional profiles were investigated by the field emission scanning electron microscope (SEM) and the secondary ion mass spectrometry (SIMS). The high-resolution transmis-sion electron microscope (TEM) was used to investigate the abruptness of the interface. The current–voltage (I–V) mea-surement was used to determine the resistance of the n-type doped DBRs. The spectrometer was used to determine the reflectivity of the DBRs. The reflectivity of the Au film was used as the reference.
3. Results and discussion
The DBRs with seven pairs n-type InP/InGaAlAs and InAlAs/InGaAlAs designed for 1.55m VCSELs were first grown and investigated. The quarter-wavelength thickness of InP, InAlAs and InGaAlAs was 122, 121 and 110 nm, respectively. To avoid the absorption of DBRs while the operating wavelength was 1.55m, the lattice-matched In0.53Ga0.39Al0.08As was used with a band gap emission
wavelength of 1.42m. The as-grown wafers were dry etched roughly 1.8m down to the n-type InP substrate to form a round mesa with the diameter of 50m. The periphery of the mesa was then passivated by SiNx. After the thinning of the n-type InP substrate, the AuGe/Ni/Au contacts were deposited on the both sides of the wafers.
Fig. 1 (a) and (b) show the interface conditions of the InP/InGaAlAs DBRs examined by TEM for two different
Fig. 1. The interface conditions of InP/InGaAlAs DBRs examined by TEM with different growth interruption time: (a)tp= 0.3 min; (b) tp= 0 min.
0.4 0.6 0.8 1.0 1 00 1 01 1 02 1 03 1 04 1 05 1 06 1 07 P In As
SECONDARY ION INTENSITY(cts/sec)
Depth(um) 0.4 0.6 0.8 1.0 1 00 1 01 1 02 1 03 1 04 1 05 1 06 1 07 Al Ga In As
SECONDARY ION INTENSITY(cts/sec)
Depth(um)
(a)
(b)
Fig. 2. The SIMS results of: (a) InP/InGaAlAs DBRs; (b) In-AlAs/InGaAlAs DBRs.
68 T.C. Lu et al. / Materials Science and Engineering B107 (2004) 66–69
growth interruption time. The growth sequence is indicated as the arrow direction.Fig. 1(a)is for the optimized growth condition with interruption time of 0.3 min. As can be seen, the interfaces between the InP and InGaAlAs are clear and abrupt.Fig. 1(b)shows the interfaces when the interruption time was 0 min. Some dark clusters can be seen at the inter-face between the InGaAlAs and InP. However, the interinter-face between InP and InGaAlAs did not contain these dark clus-ters. We attribute the dark clusters to the As carry-over ef-fect when grown under the non-optimized growth condition. The lattice-mismatched InAsP formed at the interface and became defects. These defects elongated upward to form the pits while the growth continues and reduce the reflectivity of the DBRs. On the contrary, there is no need to consider the interface switching problems to grow the InAlAs/InGaAlAs DBRs. The abrupt interface can be obtained without the in-terruption time between consecutive layers.
The SIMS results of InP/InGaAlAs and InAlAs/InGaAlAs DBRs are shown inFig. 2 (a) and (b), respectively.Fig. 2(a)
clearly indicates very low As carry-over for growth of
-150 -100 -50 0 50 100 150 -0.6 -0.4 -0.2 0.0 0.2 0.4 0.6 C u rr ent( m A ) Voltage drop(mV) InAlA s/InGaAl As DBRs InP/InGaAl As DBRs 1.1 1.2 1.3 1.4 1.5 -1.5 -1.0 -0.5 0.0 0.5 InGaAlAs InP En e rg y (e V) Distance (um) 1.1 1.2 1.3 1.4 1.5 InGaAlAs InAlAs Distance (um) (a) (b)
Fig. 3. (a) The I–V curves of InP/InGaAlAs and InAlAs/InGaAlAs DBRs with round mesas of 50m in diameter. (b) Simulation of the equilibrium band diagrams of the InP/InGaAlAs and InAlAs/InGaAlAs DBRs when the n-type concentration was chosen to be 1× 1018cm−3. The dashed line is the Fermi level.
InP/InGaAlAs DBRs under the optimized growth condi-tion.Fig. 2(b)also shows the abrupt InAlAs and InGaAlAs transition at the interface for reference. Fig. 3(a) shows the I–V curves of the InP/InGaAlAs and InAlAs/InGaAlAs DBRs with round mesas of 50m in diameter. The re-sistance per DBR pair is calculated to be 1.2 × 10−5 and
2.2 × 10−5 cm2for InP/InGaAlAs and InAlAs/InGaAlAs
DBRs, respectively. The lower value of the resistance of the InP/InGaAlAs is due to the smaller conduction band discontinuity (Ec= 0.15 for InP/InGaAlAs; Ec= 0.47
for InAlAs/InGaAlAs [9]) between two layers. Fig. 3(b)
shows the simulated equilibrium band diagrams of the InP/InGaAlAs and InAlAs/InGaAlAs DBRs for the n-type concentration of 1× 1018cm−3. The results suggest that the voltage drop is mainly located at the interface to overcome the potential barrier. Further reduction in the resistance value can be achieved by modulation doping of the inter-faces of the DBR structure to lower the potential barriers.
Two DBR structures with 35 pairs of InP/In0.53Ga0.39
T.C. Lu et al. / Materials Science and Engineering B107 (2004) 66–69 69 1400 1500 1600 1700 0 20 40 60 80 100 120
In AlAs/InG aAlAs DBRs InP/InG aAlAs DBRs
Rs ign al /R Au (% ) Wavelength (nm)
Fig. 4. The reflectivity curves of 35 pairs of InP/InGaAlAs and In-AlAs/InGaAlAs DBRs measured by the spectrometer.
were then grown for comparisons using the same interrup-tion time tpof 0.3 min for growth of InP/In0.53Ga0.39Al0.08As
DBRs.Fig. 4shows the reflectivity curves of these two sam-ples measured by the spectrometer. The measured reflectiv-ity of samples was normalized to the reflectivreflectiv-ity of the Au film. The maximum reflectivity of both DBRs exceeds 99% at center wavelength at 1.57 and 1.56m with the stopband width of 110 nm and 100 nm for the InP/InGaAlAs and InAlAs/InGaAlAs DBRs, respectively. The results of the reflectivity and the stopband width for the InP/InGaAlAs and InAlAs/InGaAlAs DBRs grown by MOCVD are com-parable with the previous reports[9,12]. The InP/InGaAlAs DBR has wider stopband because of its larger refractive in-dex contrast (n = 0.34 for InP/InGaAlAs; n = 0.3 for InAlAs/InGaAlAs[9]). However, the maximum reflectivity of the InP/InGaAlAs DBRs is susceptible to variation of the growth conditions.
4. Summary
In summary, we have grown the InP/InGaAlAs and the InAlAs/InGaAlAs DBRs with excellent electrical and optical properties using MOCVD and the growth inter-ruption technique. The DBRs show low resistance with an estimated resistance per DBR pair of 1.2 × 10−5 and
2.2 × 10−5 cm2for InP/InGaAlAs and InAlAs/InGaAlAs
DBRs, respectively. The maximum reflectivity of both
DBRs exceeds 99% with a stopband width of 110 nm for InP/InGaAlAs DBR and 100 nm InAlAs/InGaAlAs DBR. Although the InP/InGaAlAs DBRs have better optical and electrical properties, the InAlAs/InGaAlAs DBRs has much lower growth complexity. Both DBR structures should be applicable for fabrication of long wavelength VCSELs in 1.5–1.6m range.
Acknowledgements
The authors wish to thank the considerable technical port from Union Optronics Corporation. This work was sup-ported by the National Science Council of Republic of China (ROC) under contract No. NSC 91-2215-E009-030 and by the Academic Excellence Program of the Ministry of Edu-cation of ROC under the contract No. 88-FA06-AB.
References
[1] D.I. Babic, J. Piprek, K. Streubel, R.P. Mirin, N.M. Margalit, D.E. Mars, J.E. Bowers, E.L. Hu, IEEE J. Quantum Electron. 33 (8) (1997) 1369–1383.
[2] Y. Ohiso, C. Amano, Y. Itoh, H. Takenouchi, T. Kurokawa, IEEE J. Quantum Electron. 34 (10) (1998) 1904–1913.
[3] S. Sato, N. Nishiyama, T. Miyamoto, T. Takahashi, N. Jikutani, M. Arai, A. Matsutani, F. Koyama, K. Iga, Electron. Lett. 36 (24) (2000) 2018–2019.
[4] G. Steinle, F. Mederer, M. Kicherer, R. Michalzik, G. Kristen, A.Y. Egorov, H. Riechert, H.D. Wolf, K.J. Ebeling, Electron. Lett. 37 (10) (2001) 632–634.
[5] E. Hall, H. Kroemer, L.A. Coldren, Electron. Lett. 35 (5) (1999) 425–427.
[6] E. Hall, S. Nakagawa, G. Almuneau, J.K. Kim, L.A. Coldren, Elec-tron. Lett. 36 (17) (2000) 1465–1467.
[7] I. Sagnes, G. Le Roux, C. Meriadec, A. Mereuta, G. Saint-Girons, M. Bensoussan, Electron. Lett. 37 (37) (2001) 500–501.
[8] M. Linnik, A. Christou, IEEE Trans. Electron. Devices 48 (10) (2001) 2228–2237.
[9] M.H.M. Reddy, T. Asano, R. Koda, D.A. Buell, L.A. Coldren, Elec-tron. Lett. 38 (20) (2002) 1181–1182.
[10] J.P. Debray, I. Sagnes, G. Le Roux, P. Legay, M. Quillec, C. Kazmier-ski, R. Madani, J.F. Palmier, IEEE Photonics Technol. Lett. 11 (7) (1999) 770–772.
[11] T.C. Lu, J.Y. Tsai, J.T. Chu, Y.S. Chang, S.C. Wang, J. Cryst. Growth 250 (2003) 305–312.
[12] J.-H. Baek, I.H. Choi, B. Lee, W.S. Han, H.K. Cho, Appl. Phys. Lett. 75 (11) (1999) 1500–1502.
