- Email: [email protected]
Sb surfactant-mediated growth of strained InGaAs multiple-quantum wells by metalorganic vapor phase epitaxy at low growth temperatures
ARTICLE IN PRESS Journal of Crystal Growth 312 (2010) 359–362
Contents lists available at ScienceDirect
Journal of Crystal Growth journal homepage: www.elsevier.com/locate/jcrysgro
Sb surfactant-mediated growth of strained InGaAs multiple-quantum wells by metalorganic vapor phase epitaxy at low growth temperatures Tomonari Sato n, Manabu Mitsuhara, Yasuhiro Kondo NTT Photonics Laboratories, NTT Corporation, 3-1 Morinosato-Wakamiya, Atsugi, Kanagawa 243-0198, Japan
a r t i c l e in fo
abstract
Article history: Received 6 August 2009 Received in revised form 22 October 2009 Accepted 2 November 2009 Communicated by R. Bhat Available online 10 November 2009
We report on the influence of growth temperature on the Sb surfactant-mediated growth of strained InGaAs multiple-quantum wells by metalorganic vapor phase epitaxy and propose an effective method for obtaining the surfactant effect at low growth temperatures. When reducing the growth temperature from 620 to 540 1C, the Sb supply, which is needed to improve the surface morphology and the photoluminescence intensity, decreases to one tenth because of the surface segregation of the Sb atoms. With the help of Sb segregation, the surfactant effect at a growth temperature of 540 1C is obtained simply by supplying Sb prior to well growth. & 2009 Elsevier B.V. All rights reserved.
Keywords: A1. Segregation A3. Metalorganic vapor phase epitaxy B2. Semiconducting III–V materials B2. Semiconducting ternary compounds
1. Introduction Strained multiple-quantum-well (MQW) lasers have been widely used to increase emission wavelengths by increasing the strain of wells in the active regions. For example, strained InGaAs(N)/GaAs MQW lasers on GaAs can emit light in the 1.3–1.55 mm range [1,2]. Moreover, emission wavelengths longer than 2 mm have been achieved in strained In(Ga)As/InGaAs(P) MQW lasers on InP [3,4]. However, excessive strain causes threedimensional (3D) growth, and crystalline defects occur in the MQW structure. This 3D growth has been effectively suppressed by reducing the growth temperature or using a surfactant. Surfactant-mediated growth was investigated by Copel et al. [5] for a Si/Ge system using molecular beam epitaxy (MBE), and then applied to III–V heterostructures such as an InGaAs(N) layer on a GaAs substrate [6,7]. The combination of a surfactant and lowtemperature growth has been actively investigated with respect to MBE growth to further suppress 3D growth [8]. However, with metalorganic vapor phase epitaxy (MOVPE), reducing the growth temperature was the only sure way to suppress 3D growth. In contrast, we have recently reported that 3D growth can be suppressed by supplying a small amount of antimony (Sb) as a surfactant for strained InGaAs MQWs on InP substrates grown by MOVPE at a temperature as high as 620 1C [9]. When combining the use of a surfactant and low-temperature growth in MOVPE, it
n
Corresponding author. Tel.: + 81 46 240 2479; fax: + 81 46 240 4301. E-mail address: [email protected] (T. Sato).
0022-0248/$ - see front matter & 2009 Elsevier B.V. All rights reserved. doi:10.1016/j.jcrysgro.2009.11.003
is necessary to clarify the surfactant effect of Sb when the growth temperature is reduced. In this article, we report the effect of the growth temperature on the crystalline quality of strained InGaAs MQWs with an Sb surfactant. Specifically, we investigated the dependence of the morphologies and the optical properties of strained MQWs grown at 540 and 620 1C on the amount of Sb surfactant supplied. In addition, we examined the Sb depth profiles of MQWs grown at different temperatures by secondary ion mass spectroscopy (SIMS), and studied the behavior of Sb on the growing surface. Based on our consideration of the Sb behavior, we also investigated an effective way of supplying the surfactant at a low growth temperature.
2. Experimental details The samples were grown on n-type (1 0 0)-InP substrates in a horizontal MOVPE reactor at 50 Torr. We grew a 200-nm-thick InP buffer layer at a temperature (Tg) of 620 1C, followed by an MQW structure and a 100-nm-thick InP capping layer at 540 or 620 1C. The MQW structure contained three 11-nm-thick In0.79 Ga0.21As(Sb) wells with a compressive strain of 1.75% and four 18.5-nm-thick In0.48Ga0.52As barriers with a tensile strain of 0.4% with respect to InP. The growth rate and the V/III ratio of the wells were 0.75 nm/s and 40, respectively. The precursors for In, Ga, As, and P were trimethyl-indium (TMIn), triethyl-gallium (TEGa), arsine (AsH3), and phosphine (PH3), respectively. The Sb precursor was tris-dimethyl-amino-antimony (TDMASb). The TDMASb flow
ARTICLE IN PRESS T. Sato et al. / Journal of Crystal Growth 312 (2010) 359–362
rate was varied from 0 to 191 mmol/min. The solid compositions and thicknesses of the wells and the barriers were determined by X-ray diffraction (XRD) analysis. The surface morphology was studied using atomic force microscopy (AFM) in air. Room temperature photoluminescence (PL) measurements were performed using the 532 nm line of a frequency-doubled Nd:YVO4 laser. SIMS depth profiles of the samples were measured with an ATOMIKA SIMS-4000.
3. Results and discussion The 3D growth induced by increasing the well strain has been observed as a surface undulation for an InGaAs/GaAs MQW on a GaAs substrate [10]. In accordance with this observation, we first investigated the influence of supplying TDMASb on the surface morphologies of strained InGaAs MQWs on InP substrates. Fig. 1 shows AFM images of samples grown at 540 1C (a) without and (b) with a TDMASb flow rate of 15 mmol/min. The TDMASb was supplied only during well growth. In Fig. 1(a), hillocks can be seen on the surface of the sample grown without the TDMASb supply. On the other hand, these hillocks disappeared when TDMASb was supplied as shown in Fig. 1(b). The improvement in the surface morphology reflects the suppression of the 3D growth of the strained InGaAs wells owing to the surfactant effect of Sb. Supplying the TDMASb, the 3D growth could be suppressed
both for the strained InGaAs MQWs grown at 620 1C in an earlier study [9] and for those grown at 540 1C in this study. Next, we performed AFM and PL measurements for the samples grown at 540 and 620 1C with the TDMASb flow rate in the 0–191 mmol/min range to investigate the effect of the TDMASb supply on the structural and optical properties for each growth temperature. Figs. 2(a) and (b), respectively, show the dependences of the root mean square (RMS) surface roughness and the PL peak intensities of the samples on the TDMASb flow rate. For all the samples, the PL peak wavelength was approximately 2.1 mm. For the sample grown at 620 1C, the RMS roughness was 0.9 nm without the TDMASb supply. As the TDMASb flow rate increased, the RMS roughness gradually decreased until it reached a value of about 0.4 nm. When the growth temperature was 540 1C, the RMS roughness abruptly
Tg = 540°C
1.2
Tg = 620°C
RMS roughness (nm)
360
0.8
10 nm
0.4
0.0 0
50
100
150
200
TDMASb flow rate (μmol/min) 25
2 4 6 20
8
10 nm
PL peak intensity (arb.units)
μm
15
10 λPL ≈ 2.1μm
5
Tg = 540°C Tg = 620°C
2 0
4 μm
0
6 8
Fig. 1. AFM images of strained MQWs grown at 540 1C (a) without and (b) with the TDMASb supply. The TDMASb flow rate is 15 mmol/min.
50
100
150
200
TDMASb flow rate (μmol/min) Fig. 2. Dependence of (a) the RMS roughness and (b) the PL peak intensity of strained MQWs on the TDMASb flow rate.
ARTICLE IN PRESS T. Sato et al. / Journal of Crystal Growth 312 (2010) 359–362
decreased from 1.2 to 0.4 nm with the lower TDMASb supply. On the other hand, the PL peak intensities at each growth temperature increased as the TDMASb supply increased. The maximum intensities for 540 and 620 1C were, respectively, 3 and 7 times greater than the intensities without the TDMASb supply. When we compare the results in Figs. 2(a) and (b), we find that decrease in the surface roughness caused by supplying the TDMASb is directly related to the improvement in the PL intensity. This means that the surface roughness induced by the defect formation that leads to nonradiative recombination is suppressed by supplying TDMASb to the growing surface. As seen in Fig. 2, the TDMASb flow rates required to obtain the surfactant effect were very different for 540 and 620 1C. Specifically, the surfactant effect became marked at TDMASb flow rates of 15 and 122 mmol/min for growth temperatures of 540 and 620 1C, respectively. That is to say, the required flow rate for obtaining the surfactant effect at 540 1C was about a tenth of that at 620 1C. This appears to be because the Sb atoms supplied to the growing surface do not easily reevaporate at a low growth temperature. Sb used as a surfactant is incorporated slightly into films [9,11]. Therefore, to clarify the influence of the growth temperature on the behavior of Sb on the growing surface, we compared the SIMS depth profiles of Sb in MQWs grown at 540 and 620 1C. Fig. 3 shows SIMS depth profiles of Sb in samples grown at 540 and 620 1C. The samples were composed of two InGaAs wells with TDMASb flow rates of 15 and 122 mmol/min. The regions indicated by arrows in Fig. 3 correspond to the wells with the TDMASb supply. The right-hand side peak positions in each profile were misaligned because the barriers grown at 620 1C were thicker than those grown at 540 1C. For the sample grown at 620 1C, the Sb ion peak intensities in wells with TDMASb flow rates of 15 and 122 mmol/min were about 100 and 1000 counts/s, respectively, that is, the Sb concentration in the wells was
361
approximately proportional to the TDMASb flow rate. The Sb ion intensity of the sample grown at 540 1C also increased in proportion to the TDMASb flow rate, however it was 1.5 times larger than that of the sample grown at 620 1C. This means that the Sb was easier to incorporate at 540 1C than at 620 1C. However, increasing the Sb concentration by reducing the growth temperature does not adequately explain the TDMASb flow rate needed to obtain the notable surfactant effect at 540 1C, which was an order of magnitude lower than that at 620 1C as shown in Fig. 2. As can be seen in Fig. 3, the Sb ion intensity in the barrier layers for the sample grown at 620 1C decreased abruptly when the TDMASb supply was interrupted, but that for the sample grown at 540 1C decreased slowly. This indicates that the Sb atoms remained at the growing surface at a growth temperature of 540 1C. Namely, the Sb atoms segregated to the surface when the growth temperature was low. This surface segregation could increase the effective number of Sb atoms on the growing surface, thus providing a surfactant effect with the lower TDMASb flow rates at 540 1C shown in Fig. 2. Using the Sb segregation that occurs at low growth temperatures, the surfactant effect will be obtained by supplying TDMASb prior to well growth. To verify this, we grew an MQW at 540 1C with a TDMASb flow rate of 15 mmol/min only before the well growth. Fig. 4 shows PL spectra of the samples grown without and with the TDMASb supply. The PL peak intensity of the sample grown at 540 1C increased when the TDMASb was supplied before well growth. In addition, AFM revealed an improvement in the surface morphology. In contrast, for the sample grown at 620 1C, there was no increase in PL intensity or improvement in the surface morphology even when a TDMASb flow rate of 122 mmol/min was employed prior to well growth. As a result, we confirmed that the surfactant effect could be obtained by reducing the growth temperature and supplying the TDMASb only before growing the wells.
105 Tg = 540°C
with Sb (Tg = 540°C)
Tg = 620°C TDMASb flow rate: 122 μmol/min
103
PL intensity (arb.units)
Sb ion intensity (counts/s)
104
15 μmol/min
102
without Sb (Tg = 540°C)
101 with Sb (Tg = 620°C)
Growth direction
100 0
20
40
60 Depth (nm)
80
100
120
Fig. 3. SIMS depth profiles of Sb in MQWs grown with TDMASb flow rates of 15 and 122 mmol/min.
1.8
1.9
2.0 Wavelength (μm)
2.1
2.2
Fig. 4. PL spectra of MQWs grown with and without a TDMASb supply before the well growth.
ARTICLE IN PRESS 362
T. Sato et al. / Journal of Crystal Growth 312 (2010) 359–362
4. Conclusions
References
In conclusion, we investigated the effect of growth temperature on the structural and optical properties of strained InGaAs MQWs grown on InP substrates using MOVPE with Sb as a surfactant. For growth temperatures of 540 and 620 1C, these properties were improved by supplying TDMASb at flow rates of 15 and 122 mmol/min, respectively. We used SIMS analysis to confirm that the large difference in the TDMASb supply between 540 and 620 1C was due to the surface segregation of Sb atoms at 540 1C. Employing this surface segregation, we also obtained a surfactant effect with an Sb supply before (but not during) the well growth at 540 1C.
[1] K. Nakahara, M. Kondow, T. Kitatani, M.C. Larson, K. Uomi, IEEE Photon. Technol. Lett. 10 (1998) 487. [2] M. Fischer, M. Reinhardt, A. Forchel, Electron. Lett. 36 (2000) 1208. [3] M. Mitsuhara, M. Ogasawara, M. Oishi, H. Sugiura, Appl. Phys. Lett. 72 (1998) 3106. [4] T. Sato, M. Mitsuhara, T. Kakitsuka, T. Fujisawa, Y. Kondo, IEEE J. Sel. Top. Quantum Electron. 14 (2008) 992. [5] M. Copel, M.C. Reuter, E. Kaxiras, R.M. Tromp, Phys. Rev. Lett. 63 (1989) 632. [6] N. Grandjean, J. Massies, V.H. Etgens, Phys. Rev. Lett. 69 (1992) 796. [7] W.C. Chen, Y.K. Su, R.W. Chuang, M.C. Tsai, K.Y. Cheng, Y.S. Wang, J. Cryst. Growth 298 (2007) 145. [8] J.C. Harmand, L.H. Li, G. Patriarche, L. Travers, Appl. Phys. Lett. 84 (2004) 3981. [9] T. Sato, M. Mitsuhara, T. Watanabe, Y. Kondo, Appl. Phys. Lett. 87 (2005) 211903. [10] M. Sato, U. Zeimer, F. Bugge, S. Gramlich, M. Weyers, Appl. Phys. Lett. 70 (1997) 1134. [11] T. Suzuki, T. Ichihashi, K. Kurihara, K. Nishi, J. Cryst. Growth 221 (2000) 31.
Acknowledgement We thank A. Takano for technical assistance with the SIMS analysis. We also thank Y. Tohmori for his encouragement throughout this work.
Contents lists available at ScienceDirect
Journal of Crystal Growth journal homepage: www.elsevier.com/locate/jcrysgro
Sb surfactant-mediated growth of strained InGaAs multiple-quantum wells by metalorganic vapor phase epitaxy at low growth temperatures Tomonari Sato n, Manabu Mitsuhara, Yasuhiro Kondo NTT Photonics Laboratories, NTT Corporation, 3-1 Morinosato-Wakamiya, Atsugi, Kanagawa 243-0198, Japan
a r t i c l e in fo
abstract
Article history: Received 6 August 2009 Received in revised form 22 October 2009 Accepted 2 November 2009 Communicated by R. Bhat Available online 10 November 2009
We report on the influence of growth temperature on the Sb surfactant-mediated growth of strained InGaAs multiple-quantum wells by metalorganic vapor phase epitaxy and propose an effective method for obtaining the surfactant effect at low growth temperatures. When reducing the growth temperature from 620 to 540 1C, the Sb supply, which is needed to improve the surface morphology and the photoluminescence intensity, decreases to one tenth because of the surface segregation of the Sb atoms. With the help of Sb segregation, the surfactant effect at a growth temperature of 540 1C is obtained simply by supplying Sb prior to well growth. & 2009 Elsevier B.V. All rights reserved.
Keywords: A1. Segregation A3. Metalorganic vapor phase epitaxy B2. Semiconducting III–V materials B2. Semiconducting ternary compounds
1. Introduction Strained multiple-quantum-well (MQW) lasers have been widely used to increase emission wavelengths by increasing the strain of wells in the active regions. For example, strained InGaAs(N)/GaAs MQW lasers on GaAs can emit light in the 1.3–1.55 mm range [1,2]. Moreover, emission wavelengths longer than 2 mm have been achieved in strained In(Ga)As/InGaAs(P) MQW lasers on InP [3,4]. However, excessive strain causes threedimensional (3D) growth, and crystalline defects occur in the MQW structure. This 3D growth has been effectively suppressed by reducing the growth temperature or using a surfactant. Surfactant-mediated growth was investigated by Copel et al. [5] for a Si/Ge system using molecular beam epitaxy (MBE), and then applied to III–V heterostructures such as an InGaAs(N) layer on a GaAs substrate [6,7]. The combination of a surfactant and lowtemperature growth has been actively investigated with respect to MBE growth to further suppress 3D growth [8]. However, with metalorganic vapor phase epitaxy (MOVPE), reducing the growth temperature was the only sure way to suppress 3D growth. In contrast, we have recently reported that 3D growth can be suppressed by supplying a small amount of antimony (Sb) as a surfactant for strained InGaAs MQWs on InP substrates grown by MOVPE at a temperature as high as 620 1C [9]. When combining the use of a surfactant and low-temperature growth in MOVPE, it
n
Corresponding author. Tel.: + 81 46 240 2479; fax: + 81 46 240 4301. E-mail address: [email protected] (T. Sato).
0022-0248/$ - see front matter & 2009 Elsevier B.V. All rights reserved. doi:10.1016/j.jcrysgro.2009.11.003
is necessary to clarify the surfactant effect of Sb when the growth temperature is reduced. In this article, we report the effect of the growth temperature on the crystalline quality of strained InGaAs MQWs with an Sb surfactant. Specifically, we investigated the dependence of the morphologies and the optical properties of strained MQWs grown at 540 and 620 1C on the amount of Sb surfactant supplied. In addition, we examined the Sb depth profiles of MQWs grown at different temperatures by secondary ion mass spectroscopy (SIMS), and studied the behavior of Sb on the growing surface. Based on our consideration of the Sb behavior, we also investigated an effective way of supplying the surfactant at a low growth temperature.
2. Experimental details The samples were grown on n-type (1 0 0)-InP substrates in a horizontal MOVPE reactor at 50 Torr. We grew a 200-nm-thick InP buffer layer at a temperature (Tg) of 620 1C, followed by an MQW structure and a 100-nm-thick InP capping layer at 540 or 620 1C. The MQW structure contained three 11-nm-thick In0.79 Ga0.21As(Sb) wells with a compressive strain of 1.75% and four 18.5-nm-thick In0.48Ga0.52As barriers with a tensile strain of 0.4% with respect to InP. The growth rate and the V/III ratio of the wells were 0.75 nm/s and 40, respectively. The precursors for In, Ga, As, and P were trimethyl-indium (TMIn), triethyl-gallium (TEGa), arsine (AsH3), and phosphine (PH3), respectively. The Sb precursor was tris-dimethyl-amino-antimony (TDMASb). The TDMASb flow
ARTICLE IN PRESS T. Sato et al. / Journal of Crystal Growth 312 (2010) 359–362
rate was varied from 0 to 191 mmol/min. The solid compositions and thicknesses of the wells and the barriers were determined by X-ray diffraction (XRD) analysis. The surface morphology was studied using atomic force microscopy (AFM) in air. Room temperature photoluminescence (PL) measurements were performed using the 532 nm line of a frequency-doubled Nd:YVO4 laser. SIMS depth profiles of the samples were measured with an ATOMIKA SIMS-4000.
3. Results and discussion The 3D growth induced by increasing the well strain has been observed as a surface undulation for an InGaAs/GaAs MQW on a GaAs substrate [10]. In accordance with this observation, we first investigated the influence of supplying TDMASb on the surface morphologies of strained InGaAs MQWs on InP substrates. Fig. 1 shows AFM images of samples grown at 540 1C (a) without and (b) with a TDMASb flow rate of 15 mmol/min. The TDMASb was supplied only during well growth. In Fig. 1(a), hillocks can be seen on the surface of the sample grown without the TDMASb supply. On the other hand, these hillocks disappeared when TDMASb was supplied as shown in Fig. 1(b). The improvement in the surface morphology reflects the suppression of the 3D growth of the strained InGaAs wells owing to the surfactant effect of Sb. Supplying the TDMASb, the 3D growth could be suppressed
both for the strained InGaAs MQWs grown at 620 1C in an earlier study [9] and for those grown at 540 1C in this study. Next, we performed AFM and PL measurements for the samples grown at 540 and 620 1C with the TDMASb flow rate in the 0–191 mmol/min range to investigate the effect of the TDMASb supply on the structural and optical properties for each growth temperature. Figs. 2(a) and (b), respectively, show the dependences of the root mean square (RMS) surface roughness and the PL peak intensities of the samples on the TDMASb flow rate. For all the samples, the PL peak wavelength was approximately 2.1 mm. For the sample grown at 620 1C, the RMS roughness was 0.9 nm without the TDMASb supply. As the TDMASb flow rate increased, the RMS roughness gradually decreased until it reached a value of about 0.4 nm. When the growth temperature was 540 1C, the RMS roughness abruptly
Tg = 540°C
1.2
Tg = 620°C
RMS roughness (nm)
360
0.8
10 nm
0.4
0.0 0
50
100
150
200
TDMASb flow rate (μmol/min) 25
2 4 6 20
8
10 nm
PL peak intensity (arb.units)
μm
15
10 λPL ≈ 2.1μm
5
Tg = 540°C Tg = 620°C
2 0
4 μm
0
6 8
Fig. 1. AFM images of strained MQWs grown at 540 1C (a) without and (b) with the TDMASb supply. The TDMASb flow rate is 15 mmol/min.
50
100
150
200
TDMASb flow rate (μmol/min) Fig. 2. Dependence of (a) the RMS roughness and (b) the PL peak intensity of strained MQWs on the TDMASb flow rate.
ARTICLE IN PRESS T. Sato et al. / Journal of Crystal Growth 312 (2010) 359–362
decreased from 1.2 to 0.4 nm with the lower TDMASb supply. On the other hand, the PL peak intensities at each growth temperature increased as the TDMASb supply increased. The maximum intensities for 540 and 620 1C were, respectively, 3 and 7 times greater than the intensities without the TDMASb supply. When we compare the results in Figs. 2(a) and (b), we find that decrease in the surface roughness caused by supplying the TDMASb is directly related to the improvement in the PL intensity. This means that the surface roughness induced by the defect formation that leads to nonradiative recombination is suppressed by supplying TDMASb to the growing surface. As seen in Fig. 2, the TDMASb flow rates required to obtain the surfactant effect were very different for 540 and 620 1C. Specifically, the surfactant effect became marked at TDMASb flow rates of 15 and 122 mmol/min for growth temperatures of 540 and 620 1C, respectively. That is to say, the required flow rate for obtaining the surfactant effect at 540 1C was about a tenth of that at 620 1C. This appears to be because the Sb atoms supplied to the growing surface do not easily reevaporate at a low growth temperature. Sb used as a surfactant is incorporated slightly into films [9,11]. Therefore, to clarify the influence of the growth temperature on the behavior of Sb on the growing surface, we compared the SIMS depth profiles of Sb in MQWs grown at 540 and 620 1C. Fig. 3 shows SIMS depth profiles of Sb in samples grown at 540 and 620 1C. The samples were composed of two InGaAs wells with TDMASb flow rates of 15 and 122 mmol/min. The regions indicated by arrows in Fig. 3 correspond to the wells with the TDMASb supply. The right-hand side peak positions in each profile were misaligned because the barriers grown at 620 1C were thicker than those grown at 540 1C. For the sample grown at 620 1C, the Sb ion peak intensities in wells with TDMASb flow rates of 15 and 122 mmol/min were about 100 and 1000 counts/s, respectively, that is, the Sb concentration in the wells was
361
approximately proportional to the TDMASb flow rate. The Sb ion intensity of the sample grown at 540 1C also increased in proportion to the TDMASb flow rate, however it was 1.5 times larger than that of the sample grown at 620 1C. This means that the Sb was easier to incorporate at 540 1C than at 620 1C. However, increasing the Sb concentration by reducing the growth temperature does not adequately explain the TDMASb flow rate needed to obtain the notable surfactant effect at 540 1C, which was an order of magnitude lower than that at 620 1C as shown in Fig. 2. As can be seen in Fig. 3, the Sb ion intensity in the barrier layers for the sample grown at 620 1C decreased abruptly when the TDMASb supply was interrupted, but that for the sample grown at 540 1C decreased slowly. This indicates that the Sb atoms remained at the growing surface at a growth temperature of 540 1C. Namely, the Sb atoms segregated to the surface when the growth temperature was low. This surface segregation could increase the effective number of Sb atoms on the growing surface, thus providing a surfactant effect with the lower TDMASb flow rates at 540 1C shown in Fig. 2. Using the Sb segregation that occurs at low growth temperatures, the surfactant effect will be obtained by supplying TDMASb prior to well growth. To verify this, we grew an MQW at 540 1C with a TDMASb flow rate of 15 mmol/min only before the well growth. Fig. 4 shows PL spectra of the samples grown without and with the TDMASb supply. The PL peak intensity of the sample grown at 540 1C increased when the TDMASb was supplied before well growth. In addition, AFM revealed an improvement in the surface morphology. In contrast, for the sample grown at 620 1C, there was no increase in PL intensity or improvement in the surface morphology even when a TDMASb flow rate of 122 mmol/min was employed prior to well growth. As a result, we confirmed that the surfactant effect could be obtained by reducing the growth temperature and supplying the TDMASb only before growing the wells.
105 Tg = 540°C
with Sb (Tg = 540°C)
Tg = 620°C TDMASb flow rate: 122 μmol/min
103
PL intensity (arb.units)
Sb ion intensity (counts/s)
104
15 μmol/min
102
without Sb (Tg = 540°C)
101 with Sb (Tg = 620°C)
Growth direction
100 0
20
40
60 Depth (nm)
80
100
120
Fig. 3. SIMS depth profiles of Sb in MQWs grown with TDMASb flow rates of 15 and 122 mmol/min.
1.8
1.9
2.0 Wavelength (μm)
2.1
2.2
Fig. 4. PL spectra of MQWs grown with and without a TDMASb supply before the well growth.
ARTICLE IN PRESS 362
T. Sato et al. / Journal of Crystal Growth 312 (2010) 359–362
4. Conclusions
References
In conclusion, we investigated the effect of growth temperature on the structural and optical properties of strained InGaAs MQWs grown on InP substrates using MOVPE with Sb as a surfactant. For growth temperatures of 540 and 620 1C, these properties were improved by supplying TDMASb at flow rates of 15 and 122 mmol/min, respectively. We used SIMS analysis to confirm that the large difference in the TDMASb supply between 540 and 620 1C was due to the surface segregation of Sb atoms at 540 1C. Employing this surface segregation, we also obtained a surfactant effect with an Sb supply before (but not during) the well growth at 540 1C.
[1] K. Nakahara, M. Kondow, T. Kitatani, M.C. Larson, K. Uomi, IEEE Photon. Technol. Lett. 10 (1998) 487. [2] M. Fischer, M. Reinhardt, A. Forchel, Electron. Lett. 36 (2000) 1208. [3] M. Mitsuhara, M. Ogasawara, M. Oishi, H. Sugiura, Appl. Phys. Lett. 72 (1998) 3106. [4] T. Sato, M. Mitsuhara, T. Kakitsuka, T. Fujisawa, Y. Kondo, IEEE J. Sel. Top. Quantum Electron. 14 (2008) 992. [5] M. Copel, M.C. Reuter, E. Kaxiras, R.M. Tromp, Phys. Rev. Lett. 63 (1989) 632. [6] N. Grandjean, J. Massies, V.H. Etgens, Phys. Rev. Lett. 69 (1992) 796. [7] W.C. Chen, Y.K. Su, R.W. Chuang, M.C. Tsai, K.Y. Cheng, Y.S. Wang, J. Cryst. Growth 298 (2007) 145. [8] J.C. Harmand, L.H. Li, G. Patriarche, L. Travers, Appl. Phys. Lett. 84 (2004) 3981. [9] T. Sato, M. Mitsuhara, T. Watanabe, Y. Kondo, Appl. Phys. Lett. 87 (2005) 211903. [10] M. Sato, U. Zeimer, F. Bugge, S. Gramlich, M. Weyers, Appl. Phys. Lett. 70 (1997) 1134. [11] T. Suzuki, T. Ichihashi, K. Kurihara, K. Nishi, J. Cryst. Growth 221 (2000) 31.
Acknowledgement We thank A. Takano for technical assistance with the SIMS analysis. We also thank Y. Tohmori for his encouragement throughout this work.