High-efficiency InGaAs/InGaAsP compressively strained multiple quantum-well laser diode

September 10, 2017 | Autor: Haruhisa Soda | Categoría: Optical physics, Electrical And Electronic Engineering
Share Embed


Descripción

ACKNOWLEDGMENTS

The authors wish to express their sincere thanks to Dr. Y. Imamura and Dr. T. Izawa of N I T for their valuable suggestions and constant encouragement. They also wish to thank Dr. M. Fukuda and Dr. T. Sasaki of NTT for their useful discussions. REFERENCES 1. J . P. van der Ziel, R. D. Dupuis, R. A. Logan, and C. J. Pinzone, “Degradation of GaAs Lasers Grown by Metalorganic Chemical Vapor Deposition on Si Substrates,” Appl. Phys. Lett., Vol. 51, pp. 89-91, 1987. 2. M. Sugo, H. Mori, Y. Sakai, and Y. Itoh, “Stable cw Operation at Room Temperature of a 1.5-prn Wavelength Multiple Quantum Well Laser on a Si Substrate,” Appl. Phys. Lett., Vol. 60, pp. 462-473. 1992. 3. M. Sugo, H. Mori, Y. Itoh, and Y. Sakai, “InP-Based Optical Devices on Si Substrates,” in Extended Abstracts of International Conference on Solid State Devices and Materials, Japan Society of Applied Physics, 1992. pp. 656-658. 4. L. A. Coldren, K. Iga, B. I. Miller, and J . A. Rentschler, “GaInAsP1InP Stripe-Geometry Laser with A Reactive-IonEtched Facet,” Appl. Phys. Lett., Vol. 37, 1980, pp. 681-683. 5 . J. L. Merz and R. A. Logan, “Integrated GaAs-AI,Ga,.,As Injection Lasers and Detectors with Etched Reflectors,” Appl. Phys. Lett., Vol. 30, 1977, pp. 530-533. 6. K. Iga and B. I. Miller, “GaInAsPIInP Laser with Monolithically Integrated Monitoring Detector,” Electron. Lett., Vol. 16, 1980, pp. 342-343. 7. E A. Blum, K. L. Lawly, and W. C. Holton, “Monolithic Ga,~,In,AsMesa Lasers with Grown Optical Facets,” J . Appl. Phys., Vol. 46, 1975, pp. 2605-2611. 8. K. Utaka, K. Kobayashi, F. Koyama, Y. Abe, and Y. Suemune, “Single-Wavelength Operation of 1.53 pm GaInAsP1InP BuriedHeterostructure Integrated Twin-Guide Laser with Distributed Bragg Reflector under Direct Modulation up to 1 GHz,” Electron. Lett., Vol. 17, 1981, pp. 368-369. 9. H. Blauvelt. N. Bar-Chaim, D. Fekete, S. Margalit, and A. Yariv. “AIGaAs Lasers with Micro-Cleaved Mirrors Suitable for Monolithic Integration,” Appl. Phys. Lett., Vol. 40, 1982, pp. 289-290. 10. 0. Wada, S. Yamakoshi, T. Fujii, S. Hiyarnizu, and T. Sakurai, “AIGaAsiGaAs Microcleaved Facet (MCF) Laser Monolithically Integrated with Photodiode,” Electron. Lett., Vol. 18, 1982, pp. 189-190. 11. U. Koren. A. Hasson, K. L. Yu, T. R. Chen, S. Margalit, and A. Yariv. “Low Threshold InGaAsPllnP Lasers with Microcleaved Mirrors Suitable for Monolithic Integration,” Appl. Phys. Lett., Vol. 41, 1982, pp. 791-793. Received 5-6-93; revised 7-7-93 Microwave and Optical Technology Letters, 713, 143-145 0 1994 John Wiley & Sons, Inc. CCC 0895-2477194

HIGH-EFFICIENCY InGaAs/lnGaAsP COMPRESSIVELY STRAINED MULTIPLEQUANTUM-WELL LASER DIODE Shouichi Ogita, Hirohiko Kobayashi, Toshio Higashi, Niro Okazaki, Osamu Aoki, and Haruhisa Soda Fujitsu Laboratories Ltd., Atsugi 10-1 Morinosato-Wakamiya Atsugi 243-01, Japan

KEY TERMS Multiple-quanium-well laser diode, compressively strained quantum wells, InCaAsllnCaAsP, Optical-fiber amplifer, High-power laser diode ABSTRACT The characteristics of the InCaAsilnGaAsP compressively strained multiple-quantum-well laser diodes with thin wells were studied experimentally to realize the pumping laser diode for ihe optical-fiber amplifer. High external quantum efficiency and comparatively low threshold current density were obtained when the thickness of the well was reduced to less than 2 nm. It was also found that the characteristic temperature depends only on the optical confinement to the wells when [he optical confinement factor is smaller than 1.5%. Output power as high us 210 mW was demonstrated in the coated sample with the optimized well sfructure. 0 1994 John Wiley & Sons. Inc. 1. INTRODUCTION

Compressively strained multiple-quantum-well (MQW) structure is one of the most suitable structures for the high-power pumping laser diode of the optical fiber amplifier [1-7]. In the high-power laser diode, the mirror loss is comparatively large, so the MQW structure should be optimized under this condition. To achieve high-power operation at a moderately low drive current, we should consider the following two requirements. 1. High slope efficiency. 2. Low threshold carrier density.

First, the high slope efficiency is indispensable for reducing the drive current at high output power. Sufficiently low threshold carrier density is also required, because the high threshold carrier density induces the increase of the loss in the active layer caused by the increase of the carrier density at the barrier layer, and also accelerates the decrease of the slope efficiency at a high drive current caused by the heating of the active layer. To obtain a high slope efficiency, the internal loss should be reduced as much as possible. Generally, the internal loss is mainly determined by the loss in the active layer, so in the MQW structure, a decrease in the number of wells is very effective for this purpose. However, the threshold carrier density drastically increased when the number of wells decreased. A high gain is required for lasing in the small optical confinement to the wells; however, the differential gain of the MQW laser tends to saturation at a high carrier density [8]. The saturation of the differential gain strongly depends on the thickness of the wells. The thin well has a large density of state and so its differential gain does not saturate at a high carrier density. Therefore, the reduction of the thickness of wells in the MQW structure is an effective method to achieve high-power operation. In this article we studied the dependence of the external differential quantum efficiency on the thickness and the number of wells whose thickness of less than 2 nm. We also studied the temperature dependence of the lasing characteristics, and optimized the MQW structure based on the obtained results. 2. STRUCTURE

The schematic structure of our 1.48-pm high power diode is shown in Figure 1. After the growth of the MQW structure using low-pressure MOVPE on the n-type InP substrate, the FBH structure was fabricated by the conventional LPE 191.

MICROWAVE AND OPTICAL TECHNOLOGY LETTERS / Vol. 7, No. 3, February 20, 1994

145

p-lnP n-lnP

p-lnP AR (SIN)

' SL-MQW

Figure 1 The schematic structure of the 1.48-pm high-power laser diode for the Er-doped optical fiber amplifier

The front facet was coated with SIN film to reduce the facet reflectivity. The rear facet had a high-reflectivity coating which consists of three pairs of a-Si film and SiOz film. The reflectivity at each facet was optimized based on the theory taking account of the spatial hole burning effect [lo]. The cavity is 900 pm long. The band structure of tested samples is shown in Figure 2. The InGaAs wells were used, because the P-source is not used in the MOVPE growth of the InGaAs layer, so the uniformity and controllability of the InGaAs layer is good enough to grow the very thin wells. The compressive strain was introduced in the wells to reduce the threshold current density. The thickness d, of the well layer was changed from 0.9 nm (3 monolayers) to 3.3 nm (11 monolayers). For the content x of In in the In,Ga,,As-well layer, we tested two cases of 0.62 and 0.7 to set the lasing wavelength around 1.48 pm. The measured lasing wavelength of the above samples was 1.42-1.52 Fm.The InGaAsP SCH layers with the bandgap wavelength of hsCHwere introduced besides the MQW layer for improving the optical confinement to the well layer. The samples used for optimizing the MQW structure had 600pm-long cavities and the as-cleaved facets, whose mirror loss is almost the same values as that of the 900-pm-long coated samples. For the evaluation of the threshold current density, we also fabricated the oxide stripe laser with the 60-pm-wide stripe.

where r is the optical confinement factor to the wells, (Rz,) is the optical matrix element, qn is the intraband relaxation time, g ( k ) dk is the density of state, fc is the Fermi-Dirac distribution function of the conduction band, f v is that of the valence band, E,, is the transition energy. A is the differential gain; N is the injected carrier density; NG is the transparent carrier density. In the MQW structure, g ( k ) dk is roughly proportional to the reciprocal of the thickness of a well d,, therefore in the MQW structure with the thin wells, the differential gain does not saturate even at a high injection carrier density. N is proportional to the density of state, so NG also increases for the thin wells. This means that the threshold current density Jfh at a low mirror loss condition could be comparatively large for the thin wells. We fabricated the oxide stripe lasers with the different thicknesses and the different numbers of wells, while keeping the optical confinement to the wells constant. Figure 3 shows the measured Jthby varying the cavity length. In this measurement, the hscHis 1.3 pm to obtain the large optical confinement factor of 2.2%. The vertical axis shows the mirror loss, which is proportional to the reciprocal of the cavity length. From Figure 3, the MQW structure with a large number of thin wells shows the sufficiently low J t h even at a high mirror loss condition, which could improve the temperature characteristics. We also measured the temperature dependence of Jthusing the same samples, which is shown in Figure 4. The cavity length is 600 pm. As can be seen in Figure 4, the increase of the J t h of the 1.5-nm-thick MQW is smaller than that of the 3.3-nm-thick MQW. From Figures 3 and 4, we confirmed that the MQW structure with the thin wells is very effective for a high-power laser operating at high injection levels. If the internal loss is determined only by the loss in the well layer, a MQW with a large number of thin wells is the best structure to realize the high-power laser diode.

100

I

I

I

3. EFFECT OF THE THIN WELLS

First, we confirmed the effect of the thin well in the MQW structure. From the semiclassical theory taking account of the effect of the intraband relaxation of electrons [ l l ] , the modal gain G , is given by the following, which is roughly approxi-

1

r

-

InP

InGaAsP (1.3 pm)

-0

J,,

lnxGal.,As 100nm 10nm dw l 0 0 n m Figure 2 The band structure of the InGaAsiInGaAsP compressively strained MQW laser

146

2

1

3 4 (kA/cm2)

5

Figure 3 The dependence of the threshold current density Jthon the mirror loss. The samples were the oxide stripe laser with the same optical confinement factor to the wells. The bandgap wavelength hscr, of the SCH layer was 1.3 pm. The x of In,Ga,.,As wells was kept at 0.62

MICROWAVE AND OPTICAL TECHNOLOGY LETTERS / Vol. 7, No. 3, February 20, 1994

10 600 p n

As-cleaved

cu^ E 0

h 2 5

1

5

10

7

1.8 2.1

0.1

0

20 40 60 80 Temperature ("C)

100

However, the loss in the barrier layer and the SCH layer is not negligible because the optical confinement factor to the wells is less than a few percent. Therefore, the thickness and the number of wells should be optimized to achieve high slope efficiency and sufficiently good temperature characteristics. 4. OPTIMIZATION OF THE MQW STRUCTURE

Next, we studied the dependence of the external differential quantum efficiency r)d at a low drive current and Jth on the number N , of wells by varying the d, from 0.9 to 2.1 nm. Figure 5 shows the output power versus injection current characteristics of the four-QW laser with the 0.9-nm-thick wells. The large 71,) of 27%/facet (0.25 mW/mA/facet) and the sufficiently low threshold current of 20 mA were obtained even for the d , = 0.9 nm. Figure 6 shows the dependence of r)(( on N,. From Figure 6, vd increases by reducing N , when N,,, 2 5. This is due to the reduction of internal loss. The maximum value of the qdwas 3l%/facet (0.27 nW/mA/facet) for the four-pair 1.2-nm-thick wells. When N , 5 3, the q d decreases by reducing the N,. Figure 7 shows the threshold current density Jth.The Jth was estimated from the measured threshold current, cavity length, and the width of the active

9

7

Figure 6 The dependence of the external differential quantum efficiency qd on the number N , of wells. The parameters were the thickness d , of wells and the In contents x of the wells. The tested laser had a 600-pm-long cavity and as-cleaved facets

region obtained from SEM photograph. The Jthdrastically increases when N , 5 3, which corresponds to a decrease of Td. Therefore, a decrease of T~ for N , 5 3 could be caused by an increase in the threshold carrier density, which induces an increase in the internal loss. For N , 2 5 , the Jth takes an almost constant value of 1.5 kA/cm2. From these two figures, we confirmed that the MQW structure with thin wells of thickness less than 2.1 nm are very effective in achieving high Td. These two figures also showed that r)d and Jth are almost independent of d,, especially when N , 2 5. This means the loss in the barrier layer could not be negligible for the very thin wells less than 2.1 nm. The difference of the lasing characteristics between x = 0.62 and x = 0.7 was not clear for d, 5 2.1 nm from our results. Figures 8 and 9 show the internal loss a, and internal quantum efficiency vi, respectively. Figure 8 shows that the ai decreases when the N , decreases for N , 2 4, and drastically increases for N , = 3. Figure 8 also shows that the ai is almost independent of the d , when N , 2 5. These results support the above discussion about r)d. r), takes a large value of more

8 20

5 NW

Figure 4 The temperature dependence of the threshold current density Jth.The cavity is 600 pm long. Each facet was as cleaved

n

3

1

0.62 0.62

2 1 6-;OO pm 1 As-cleaved

n e l

E

0

h Y

!

W

0

Q

7

'

,

+

I

6 -

I

I

d w (nm) 0.9 ' I 1.2 A 1.5 0 1.8 0 1.8

x 0.7 0.7 0.7 0.7 0.62

4

5

7

2

0 ' 0

' 50 '

8

I

I

100 150 Current (mA)

As-cleaved I

I

200

Figure 5 The output power versus drive current characteristics of the MQW-FBH laser diode with the very thin wells of 0.9-nm thickness. The tested laser had a 600-pm-long cavity and as-cleaved facets

MICROWAVE AND OPTICAL TECHNOLOGY LETTERS / Vol. 7, No. 3, February 20, 1994

147

-

dw (nm) 1.2 A 1.5

V

I

r;

201

6-

10

0

W

x 0.7 0.7

I

2 W

I-0

1

6o

40

1

i

600 pm As-cleaved

i

60-g00c

A 0 0 X

0' 1

I

3

I

5 Nw

1

I

0

9

7

Figure 8 The dependence of the estimated internal loss a,on the number N,+ of wells. The parameters were the thickness d, of wells and the In contents x of the wells

than 90%. which is independent of both the d, and the N,, shown in Figure 9. This means the carrier injection to the well layers is sufficiently large for the d, 5 2.1 nm. Next, we measured the characteristic temperature T,, for estimation of the decrease of the slope efficiency at a high drive current. Figure 10 shows the To values measured in the high-temperature range from 60 to 90 "C. The samples had the FBH structures, so the measured Tovalues showed comparatively small values due to the effect of the current leak through the BH region. The horizontal axis is the calculated optical confinement factor to the wells using the simple slabwaveguide analysis. From this figure, the samples with the same optical confinement factor had almost the same To when the optical confinement factor is less than 1.5%. This means the TI depends only on the volume of the well layers when the optical confinement factor is less than 1.5%. Therefore, a large optical confinement to the well layers is indispensable in improving the high-temperature characteristics, which correspond to the high-power characteristics at a high drive current. To as high as 45 K, which is equal to the values for the bulk cases, is obtained for an optical confinement factor of

Figure 10 The dependence of the T,, measured at the temperature range from 60 to 90 "C on the calculated optical confinement factor to the wells. The tested laser had a 600-pm-long cavity and as-cleaved facets

more than 1.4%. From the discussion about the slope efficiency in this chapter, the internal loss slightly increases as the number of the wells increases, and is almost independent of the thickness of wells. Therefore, the number of the wells should be selected as small as possible even when the optical confinement factor is larger than 1.4%. Based on these results, we can obtain an optimum MQW structure; N, = 5 , d , = 1.8 nm. a n d x = 0.62. 5. HIGH-POWER CHARACTERISTICS

Based on the above results, we fabricated the high-power laser diode emitting at 1.48 pm. The output power versus injection current characteristics of the optimized MQW structure is shown in Figure 11. The high slope efficiency of 0.43 mW/ mA and the high-power operation exceeding 210 mW were obtained at room temperature. The drive current at the output power of 150 mW is moderately low, around 500 mA. The lasing wavelength is 1.485 pm at 150-mW output power. At the submount temperature of 70 "C, an output power of 100

250 1

I

0'

1

I

3

I

I

5

7

/

50

90

x 0.7 0.7 0.62

n

9

NW Figure 9 The dependence of the estimated internal quantum efficiency 7,on the number N,, of wells. The parameters were the thickness d,, of wells and the In contents x of the wells

148

I

70

1 25

I

0

c d w (nm) V 1.2 A 1.5 0 1.8

I

0.7 0.62 0.62

0.5 1.0 1.5 2.0 Optical confinement factor (%)

n

75

I

1.5 1.8 1.8 2.1

"0

400 800 Current (mA)

1200

Figure 11 The high-power characteristics of the optimized MQW laser diode with a 900-pm-long cavity and AR-HR coated facets. The parameter was the submount temperature. The laser diode was mounted as the junction down

MICROWAVE AND OPTICAL TECHNOLOGY LETTERS / Vol 7, No. 3, February 20, 1994

sity. The To value, which gives the limitation of the highpower operation, drastically decreased for the small optical confinement to the well layer. Based on these experiments, we obtained an optimum MQW structure with the five-pair 1.8-nm-thick wells. We also demonstrated high-power operation exceeding 150 m W with a moderately low drive current of less than 500 mA. ACKNOWLEDGMENT

The authors wish t o thank Mr. S. Hirose for the device fabrication and Dr. S. Yamakoshi for useful discussions.

-45

0

45

Angle (degree)

-45

0

45

Angle (degree)

Figure 12 The measured far-field pattern of the optimized MOW laser diode. The figure on the left shows the horizontal pattern, and that on the right shows the vertical pattern

mW was obtained at a drive current of 500 mA. These data are sufficient values to realize the pumping laser diode for the optical-fiber amplifier. In our design, the well is very thin, and the SCH layer consists of a wide-gap InGaAsP (ASCH = 1.1 pm), so the narrow optical beam could be expected. Figure 12 shows the measured far-field pattern. A sufficiently narrow beam was obtained, which enables us to make the coupling efficiency to the single-mode fiber very high. The characteristics of the laser diode module with the optical-fiber pigtail using this device was shown in Figure 13. The high output power of 100 mW coupled to the single-mode fiber was achieved at the sufficiently low drive current of 550 mA. The coupling efficiency in our module exceeded 70%. 6. CONCLUSION

We studied the dependence of the external differential efficiency and the characteristic temperature on the thickness and the number of wells in the compressively strained InGaAsi InGaAsP MQW laser for designing a high-power pumping laser diode. High external quantum efficiency of more than 30%/facet was obtained for thin wells with a thickness of less than 2 nm without the increase of the threshold current den-

c

3 CL 3 0

c

50 -

L

Q,

P

25 -

i i 0

0

200 400 600 Current (mA)

Figure 13 The output power versus drive current characteristics of the 1.48-pm laser diode module with the optical-fiber pigtail

REFERENCES 1. P. J. A. Thijs, L. F. Tiemeijer, P. I. Kuindersma. J . J. M. Binsma, and T. V. Dongen, “High-Performance 1.5 pm Wavelength InGaAs-InGaAsP Strained Quantum Well Lasers and Amplifiers,” IEEEJ. Quantum Electron., Vol. 27, June 1991, pp. 14261439. 2. T. Tanbun-Ek, R. A. Logan, N. A. Olsson. H . Temkin. A. M. Sergent, and K. W. Wecht. “High Power Output 1.48-1.51 prn Continuously Graded Index Separate Confinement Strained Quantum Well Lasers.” Appl. Phys. Lett., Vol. 57. July 1990. pp. 224-226. 3. H. Asano. S. Takano. M. Kawaradani, M. Kitamura. and I . Mito. “1.48 pm High-Power InGaAs/InGaAsP MQW LD’s for ErDoped Fiber Amplifiers,” IEEE Photon. Technol. Lett., Vol. PTL-3. May 1991, pp. 415-417. 4. M. Joma, H . Horikawa. Y. Matsui. and T. Kamijo. “High-Power 1.48 pm Multiple Quantum Well Lasers with Strained Quaternary Wells Entirely Grown by Metalorganic Vapor Phase Epitaxy.” Appl. Phys. Lett., Vol. 58, May 1991. pp. 2220-2222. 5 . T. Namegaya. R . Katsumi. Y. Imajo. N. Iwai. S . Namiki. A. Kasukawa. Y. Hiratani, and T. Kikuta, “1.48 pm High Power GaInAsP GRIN-SCH MQW LD with Circular Beam.” in Technical Digest of Fourth Optoelectronics Conference (OEC’Y2). Makuhari Messe, Japan. July 1992. Paper 17B2-1. 6 . R . Suzuki, S. Tsuji. T. Tuchiya, T. Taniwatari. T. Toyonaka. and Y. Ono. “Comparison of the Compressively and Tensile Strained MQW-BH Lasers Operating at 1.48 pm.” in Technical Digest of Fourth Optoelectronics Conference (OEC’Y2),Makuhari Messe. Japan. July 1992. Paper 17B2-2. 7. H . Kamei, N. Tatoh, J . Shinkai, H . Hayashi. and M. Yoshimura. “Ultrahigh Output Power of 1.48 pm GaInAsPiGaInAsP Strained-Layer MQW Laser Diodes.” Technical Digest of OFC’YZ, San Jose, CA, Feb. 1992, Paper TuH6. 8. M. Mittelstein. Y. Arakawa. A. Larsson. and A. Yariv. “Second Quantized State Lasing of a Current Pumped Single Quantum Well Laser,” Appl. Phys. Lett., Vol. 49. Dec. 1986. pp. 16891691. 9. H . Ishikawa, H . Soda, K. Wakao. K. Kihara. K. Kamite, Y. Kotaki, M. Matsuda, H . Sudo, S. Yamakoshi, S. Isozumi, and H . Imai, “Distributed Feedback Laser Emitting at 1.3 pm for Gigabit Communication Systems,” J . Lightwave Technol., Vol. LT-5, June 1987. pp. 848-855. 10. T. Higashi, S. Ogita, H . Soda, H. Kobayashi. H . Kurakake. 0. Aoki, and N. Okazaki, “Optimum Asymmetric Mirror Facet Structure for High Efficiency Semiconductor Lasers.” IEEE J . Quantum Electron., Vol. QE-29, June 1993 (to be published). 11. M. Asada and Y. Suematsu, “Density-Matrix Theory of Semiconductor Lasers with Relaxation Broadening Model-Gain and Gain-Suppression in Semiconductor Lasers,” IEEE 1.Quantum Electron., Vol. QE-21, May 1985. pp. 434-442. Received 5-7-93; revised 8-9-93 Microwave and Optical Technology Letters, 713, 145-149 0 1994 John Wiley & Sons. Inc. CCC 0895-2477194

MICROWAVE AND OPTICAL TECHNOLOGY LETTERS I Vol. 7, No. 3, February 20, 1994

149

Lihat lebih banyak...

Comentarios

Copyright © 2017 DATOSPDF Inc.