SPIE Proceedings [SPIE Symposium on Integrated Optics - San Jose, CA (Saturday 20 January 2001)]...

Optimization of the barrier height in 1.3 µm InGaAsPmultiple-quantum-well active regions for high temperature operation

Sebastian Mogg∗ a, and Joachim Piprek

b

aLaboratory of Semiconductor Materials, Department of Electronics,

Royal Institute of Technology, Stockholm, S-164 40 Kista

bDepartment of Electrical and Computer Engineering,

University of California at Santa Barbara, Santa Barbara, CA 93106

ABSTRACT

We present a study of barrier height effects on the high-temperature performance of 1.3 µm strained layer InGaAsP/InPquantum well lasers. Broad-area Fabry-Perot lasers were fabricated and their light-current characteristics were measured attemperatures between 20 °C and 80 °C. Based on our experimental results we analyze the effect of the barrier bandgapusing the commercial laser simulation software LASTIP. The simulator calculates all relevant physical mechanisms,including their dependence on temperature and local carrier density, self-consistently. The strained quantum-well opticalgain computation is based on the 4 x 4 kp method considering valence-band mixing effects. A drift-diffusion modelincluding thermionic emission at hetero-interfaces is used for the calculation of the carrier transport. Careful adjustments ofmaterial parameters, in agreement with data reported in the literature, are performed in order to reproduce themeasurements. Lowering the barrier height in the active region leads to an improved performance of our laser with respectto threshold current and slope efficiency. An optimum barrier bandgap range of 1.21 - 1.24 eV is identified for our laser.This is partially attributed to the non-uniform carrier-distribution across the quantum-wells.

Keywords: InGaAsP/InP, laser characterization, semiconductor lasers, quantum well devices, numerical analysis,semiconductor device modeling

1. INTRODUCTION

Long-wavelength (1.3 µm and 1.55 µm) InGaAsP/InP laser diodes have gained a great deal of interest in the last twodecades due to telecommunication applications. Much effort has been devoted to the fabrication, analysis and optimizationof these devices considering a wide range of adjustable design parameters. The performance at elevated temperatures is ofspecial interest with respect to low threshold currents and high slope efficiencies [1], [2]. The laser performance isinfluenced by several physical effects such as Auger recombination [3], carrier leakage out of the active region [4],intervalence band absorption (IVBA) [5], lateral carrier spreading [6], absorption and spontaneous recombination withinpassive layers [7], and optical gain reductions [8]. The self-consistent combination of all these mechanisms in one model isessential to correctly describe high-temperature effects [2]. In recent years, several authors have studied the influence of thebarrier height in multi-quantum well (MQW) structures on the operational characteristics of InGaAsP lasers [9], [10]. It wasshown that the overall performance of laser structures improves as the barrier bandgap is decreased. This is partiallyattributed to a non-uniform carrier distribution within the active region [11], [12]. Direct hole transport measurements instrained InGaAsP MQW structures using time-resolved photoluminescence measurements support this theory [13]. In thispaper, we analyze measurements on InGaAsP/InP laser structures considering all of the above physical mechanisms and

∗ Correspondence: Email: [email protected]; Telephone: +46 8 752 1162; Fax: +46 8 750 5173

Physics and Simulation of Optoelectronic Devices IX, Yasuhiko Arakawa, Peter Blood, Marek Osinski,Editors, Proceedings of SPIE Vol. 4283 (2001) © 2001 SPIE · 0277-786X/01/$15.00 227

Downloaded From: http://proceedings.spiedigitallibrary.org/ on 09/22/2013 Terms of Use: http://spiedl.org/terms

their interaction self-consistently. Based on this analysis we investigate theoretically the effects of varying barrier heights onthe high-temperature performance of InGaAsP/InP laser diodes.

The accurate modeling of III-V semiconductor lasers requires the specification of many material parameters which are oftennot well known [14]. Furthermore, the actual structure may differ from the original laser design due to insufficient growthcontrol. For instance, the actual composition of the active region may deviate from the intended composition [15]. For thesereasons and owing to the complexity of semiconductor lasers, measurements need to be used for parameter calibration in themodel. More confidence in the choice of parameters and their relative importance can be obtained by simultaneousreproduction of several experimental results, e.g., with variation of temperature or laser length. The simulated situationsimplifies if pulsed laser operation at different stage temperatures is employed. This way, self-heating of the device can beneglected and its temperature is equal to the stage temperature. This excludes the thermal conductivity parameter from thesimulation which is not exactly known in semiconductor heterostructures [16] and may lead to inaccurate active regiontemperatures in CW simulations.

In this paper, we first measure and analyze the temperature dependence of 1.3 µm InGaAsP/InP laser diodes with a barrierbandgap of 1.29 eV. We then stepwise reduce the barrier bandgap in the simulation to find an optimum design for high-temperature operation. For this analysis we employ the commercial laser diode simulator LASTIP [19]. Section 2 gives adescription of the device structure and the experimental results. Physical models and material parameters are discussed inSection 3. The results of our numerical analysis are presented in Section 4.

2. DEVICE STRUCTURE AND EXPERIMENTAL RESULTS

The laser structures are grown by low-pressure metal organic vapor phase epitaxy (LP-MOCVD) in a commercialAIXTRON AIX 200 planar horizontal reactor at a temperature of 710 °C and a pressure of 50 mbar on <100> sulfur doped,n-type InP substrates. Details of growth and material quality are described elsewhere [17]. The active region of the lasersconsists of five 6.5 nm thick In0.90Ga0.10As0.51P0.49 quantum wells with 0.94 % compressive strain, emitting near 1.31 µm.The wells are separated by 7.3 nm thick tensile strained (0.36 %) In0.90Ga0.10As0.085P0.915 barriers (λ = 0.96-µm,hν = 1.291 eV). The multi quantum well (MQW) stack is embedded in 100 nm thick separate confinement layers (SCL) of1.02-µm InGaAsP followed by InP cladding layers. The MQW and the SCL on the p-side of the structure are notintentionally doped in order to keep the absorption losses low and to prevent back diffusion of the zinc (Zn) dopants into theactive region. The residual n-type doping is about 1 × 1015 cm-3. A highly doped p-InGaAs contact layer is used.

3.4 3.5 3.6 3.7 3.8 3.9-0.5

0.0

0.5

1.0

1.5

MQW p-InPSCLSCLn-InP

heavy holes

light holes

hole Fermi level

electron Fermi level

valence band

conduction band

En

erg

y [e

V]

Vertical Position [µm]

FIG. 1: Energy band diagram of the laser structure at lasing threshold (T = 20 °C).

Proc. SPIE Vol. 4283228


Table 1 summarizes composition, layer thickness and doping concentration in all layers. A calculation of the active regionenergy diagram at lasing threshold is depicted in Fig. 1. Broad-area oxide-stripe lasers are processed with a stripe width of50 µm and cleaved into four different lengths of L = 300 µm, 600 µm, 900 µm and 1200 µm. No coatings are applied to thefacets so that the laser mirrors remain as-cleaved. Up to seven devices for each length are mounted p-side up on siliconsub carriers. Several gold wires are bonded to each laser to maintain homogenous current injection and to prevent damagingof the devices due to direct probing.

TABLE 1: Layer parameters as used in the simulation.

LAYER MATERIAL THICKNESS [nm] DOPING [cm-3]

contact InGaAs 150 p - 3 × 1019 (Zn)

cladding InP 2000 p - 7 × 1017 (Zn)

cladding InP 200 p - 1 × 1017 (Zn)

separate confinement In0.54Ga0.46As0.19P0.81 100 n - 1 × 1015

barrier In0.90Ga0.10As0.085P0.915 7.3 n - 1 × 1015

well In0.90Ga0.10As0.51P0.49 6.5 n - 1 × 1015


well In0.90Ga0.10As0.51P0.49 6.5 n - 1 × 1015


well In0.90Ga0.10As0.51P0.49 6.5 n - 1 × 1015


well In0.90Ga0.10As0.51P0.49 6.5 n - 1 × 1015


well In0.90Ga0.10As0.51P0.49 6.5 n - 1 × 1015


separate confinement In0.54Ga0.46As0.19P0.81 100 n - 5 × 1016 (Si)

buffer InP 500 n - 1 × 1018 (Si)

substrate InP 3000 n - 6 × 1018 (Si)

The sub carriers with the bonded lasers are mounted to a copper heat sink which temperature can be controlled in a rangebetween –25 °C and +150 °C using an ILX LDT-5190B temperature controller in conjunction with an AD590 temperaturesensor. The light-current (L-I) characteristics of the lasers are measured under pulsed operation using an ILX LDP-3840/3Apulsed current source. The pulses are 2.5 µs short and repeated with a duty-cycle of 0.1 % to keep self-heating negligible.This is confirmed by observing the laser emission wavelength being red-shifted if self-heating occurs. Calibrated L-Imeasurements at 20 °C are initially performed for all laser lengths and shown in Fig. 2. For a device with L = 617 µm wemeasure a threshold current of Ith = 196 mA, a threshold current density of jth = 636 A/cm2, a total slope efficiency ofηd = 0.47 and a lasing wavelength of λ = 1312 nm. The measured L-I curves scatter although different sets of laser diodesare theoretically identical. This scattering is attributed to intrinsic microstructural as well as to processing related differencesbetween the lasers. As a result of these unavoidable variations laser parameter extraction becomes more difficult. Thethreshold current Ith and slope efficiency η of each device is determined by the I-axis intercept of a least mean square fitthrough the linear part of the measured L-I data. The inverse slope efficiency ηd

-1 obtained from the L-I curves is thenplotted versus cavity length L in order to extract the internal optical loss αi and the differential internal efficiency ηi (Fig. 3).Owing to the scattering the data points do not fall on a straight line as it is expected. The average facet reflectivity isassumed to be R = 0.28. Taking all data points into account linear regression gives αi = 9.6 /cm ± 0.5 /cm andηi = 67.3 % ± 1.4 %. Additionally to the uncertainty in the measured data the ηd

-1 (L) method itself contains an inherentinaccuracy since it neglects the length dependence of both parameters, αi and ηi [18].

Proc. SPIE Vol. 4283 229


Header for SPIE use

0 100 200 300 400 5000

5

10

15

20

25symbols: measurementsolid lines: simulation

square: 300umcircle: 617umup triangle: 892umdown triangle: 1200um

T = 293K

Lig

ht

Po

wer

/ F

acet

[m

W]

Current [mA]

FIG. 2: Pulsed laser power versus current (L-I) characteristics atdifferent laser length (symbols: measurement, lines:simulation).

200 400 600 800 1000 1200 14001.6

1.8

2.0

2.2

2.4

2.6

2.8

3.0

Inve

rse

Slo

pe

Eff

icie

ncy

ηd-1

Laser Cavity Length L [µm]

FIG. 3: Extraction of internal differential efficiency η i and internaloptical loss α i from the measured slope efficiencies (squares).

For our study of temperature effects on Ith and ηd we chose a typical laser from the set having cavity lengths of about600 µm. The device temperature is varied from 20 °C (room temperature) to 80 °C by heating the copper stage. Thethermally induced red-shift of the laser emission wavelength is measured to dλ/dT = 0.5 nm/K which is typical for thismaterial system. Temperature effects on our pulsed L-I characteristics are shown in Fig. 4.

0 100 200 300 400 500 600 700 800 9000.0

2.5

5.0

7.5

10.0

12.5

15.0

17.5

20.0

22.5

25.0symbols: measurementsolid lines: simulation

square: 293 Kcircle: 313 Kup triangle: 333 Kdown triangle: 353 K

L = 617um

Lig

ht

Po

wer

/ F

acet

[m

W]

Current [mA]

FIG. 4: Pulsed laser power versus current (L-I) characteristics at various temperatures (symbols: measurement, lines: simulation).

ηi = 67.3 % ± 1.4 %α i = 9.6 /cm ± 0.5 /cm



3. PHYSICAL MODELS AND MATERIAL PARAMETERS

We have employed LASTIP [19], an advanced commercial laser simulation software, to analyze our measurements and tostudy the effects of varying barrier heights on the temperature performance of 1.3 µm strained MQW lasers. The softwareself-consistently combines two-dimensional (2-D) carrier transport, heat flux, optical gain computation, and wave guidingwithin the transversal plane (x, y). Our case does not necessitate including heat flux calculations since the lasers areoperated under pulsed conditions. In our investigation, the temperature is homogenous throughout the device and is equal tothe stage temperature. Longitudinal variations are also not considered and believed to be of minor importance. Furtherdetails of the laser model can be found elsewhere [20], [21]. In the following we discuss those aspects of the model whichare most important for our analysis.

p-InP

n-InP

SCL

SCL

MQW

metal oxide

verticalleakage

lateral leakage

x

y

FIG. 5: Illustration of leakage currents from the waveguide region in the lateral (x) and vertical (y) directions.

Fermi-statistics and thermionic emission at heterointerfaces is included in the drift-diffusion model of carrier transport [22].Thermionic emission is mainly controlled by the offset of the conduction band (∆EC ) and the valence band (∆EV ) at theheterojunction. There have been several studies on the band offset in the InGaAsP/InP material system [23]. However,considerable disagreement in the reported data is found. We assume in our simulations a band offset ratio of∆EC / (∆EC+∆EV) = 0.4 independent of temperature. Fig. 5 illustrates leakage currents in our device. Vertical leakagecurrent is identified as part of the total current which passes through the active layer in vertical y-direction. For electronsleaking out of the MQW stack into the p-cladding we draw the border along the SCL-InP interface. Vertical hole leakage isnegligible. All carriers leaving the active region in the lateral x-direction at the position of the metal-oxide interfaceconstitute lateral leakage. Within passive layers, a temperature- and carrier density-independent radiative recombinationcoefficient of B = 10-10 cm3/s is assumed. The spontaneous recombination rate within the quantum wells is calculated self-consistently by numerical integration of the spontaneous emission spectrum. Defect and surface recombination is lumped inthe Shockley-Read-Hall (SRH) recombination lifetime of electrons and holes, we assume τn = τp = 10 ns in the active regionand 100 ns elsewhere. For long-wavelength semiconductor lasers, the Auger recombination process is often the predominantnonradiative mechanism. In LASTIP, the local Auger recombination rate is calculated by Rau = (Cn + Cp )(np – ni

2)(n - electron density, p - hole density, ni - intrinsic carrier density) where the theoretical Auger coefficients, Cn and Cp, aresomewhat different from the experimental parameter C. This difference has been investigated by Seki et al. [24]. We followtheir approach and assume that the conduction-hole-hole-split-off (CHHS) Auger process dominates (Cn = 0). From our fitto measurements (Fig. 4) we extract an Auger coefficient of Cp = 9.5 × 10-29 cm6/s at room temperature (20 °C) which risesto Cp = 2.3 × 10-28 cm6/s at 80 °C corresponding to an activation energy of Ea = 132 meV. Common experimentally deduced



room temperature values for the Auger coefficient at λ = 1.3 µm are between C = 2.0 × 10-29 cm6/s and 5.0 × 10-29 cm6/swhereas theoretical investigations suggest values as high as C = 2.0 × 10-28 cm6/s [25]. Our fit result may include Augerinduced carrier leakage [26]. Hot carriers with energies in excess of the heterobarrier height are produced in the Augerprocess and can diffuse out of the active region into the cladding layers contributing to the leakage current. This Augerenhanced leakage is not included in the LASTIP model. The total injected current I into the device is composed of both theleakage currents and all radiative as well as nonradiative currents within the active region.

For the efficient computation of the optical gain in our compressively strained QW’s we fit a parabolic subband model to avalence mixing subband model which is calculated by the 4 × 4 kp method [27]. Lorentzian gain broadening and intrabandrelaxation due to carrier-carrier and carrier-LO-phonon scattering is considered following Asada [28]. We assume alongitudinal optical phonon energy of hνLO = 34 meV and an optical dielectric constant of ε∞ = 11.4. Usually, the scatteringtime is only known to an order of magnitude (10-13 s). The intraband scattering model used in the simulations calculates thetemperature dependent scattering time self-consistently. Fig. 6 depicts the calculated scattering time versus the active regiontemperature.

280 300 320 340 360 380 4000.44

0.46

0.48

0.50

0.52

0.54

0.56

0.58

Sca

tter

ing

Lif

etim

e τ

[10-1

3 s]

Temperature [K]

FIG. 6: Calculated intraband relaxation time versus active region temperature.

Bandgap shrinkage owing to exchange effects is considered as ∆Eg = -ξN1/3 where we assume an exchange coefficient ofξ = 3.2 × 10-8 eV cm [29]. The thermal bandgap reduction parameter dEg/dT = -0.35 meV/K is extracted from the measuredthermal shift of the MQW photoluminescence peak wavelength and it is used in all layers except the barriers. We obtain thebest agreement between experiment and simulation in the temperature range from 20 °C to 80 °C by assumingthat the barrierlayer bandgap (BLBG) depends on temperature as BLBG(T) = BLBG(293 K) + dEg/dT ⋅ T with dEg/dT =-0.43 meV/K.

LASTIP includes several loss mechanisms other than interband absorption. These other mechanisms include intraband freecarrier absorption, intervalence band absorption (IVBA), and light scattering by material inhomogeneity. The correspondinglocal loss depends on the density of electrons and holes and is considered as α = knn + kpp + αb. The constant backgroundloss αb accounts for carrier density independent mechanisms such as light scattering at defects. Free carrier absorptioncaused by electrons is small in 1.3 - 1.55 µm InGaAsP/InP lasers (kn = 10-21 cm2). Absorption due to holes in the valenceband can be attributed to intraband transitions (free carrier absorption) or to interband transitions (IVBA). Both mechanismsdepend nearly linearly on the hole density and are lumped in the parameter kp. However, their individual contribution to thetotal loss is difficult to assess and values of kp are difficult to find in the literature. In our simulation we assume kn and kp



uniform throughout the device and independent of temperature. From our fit to L-I measurements of lasers with differentlengths (Fig. 2), vanishing background loss (αb = 0) and R = 0.28 power reflectance per facet, we extract kp = 13 × 10-18 cm2

at room temperature. This corresponds to an absorption loss of 13 cm-1 for a hole concentration of p = 1 × 1018 cm-3 inperfect agreement with results of other researchers [30]. The mirror loss coefficient is defined as am = 1/L ln(1/R). Ourcavity length of L = 617 µm gives am = 20.6 cm-1.

4. SIMULATION AND ANALYSIS

The simulated L-I characteristics are shown in Fig. 2 and Fig. 4 for different cavity lengths at room temperature and fordevice temperatures ranging from 20 °C to 80 °C (L = 617 µm), respectively. We obtain reasonable agreement with themeasurements by careful adjustments of key material parameters as discussed in the previous section. This agreementconfirms the validity of the laser model. The same model and the same parameters are now employed to study the effect ofthe barrier bandgap theoretically. We repeat the calculations by lowering the BLBG energy from 1.291 eV to 1.166 eV insteps of 25 meV. Since we keep all other parameters constant, the potential depth of the wells will decrease with decreasingBLBG. Beside the decrease in the spatial overlap between the electron and hole wavefunctions in the wells with decreasingBLBG we also expect the amount of vertical carrier leakage out of the QW’s into the cladding layers to vary. The structurewith the smallest BLBG exhibits the largest total energy barrier for electron leakage from the quantum well level into the p-side InP-cladding suggesting a minimum leakage current. However, other effects like the non-uniform MQW carrierdistribution may also be important [11].

1.16 1.18 1.20 1.22 1.24 1.26 1.28 1.30100

150

200

250

300

350

400

450

500

550

Th

resh

old

Cu

rren

t I th

[m

A]

BLBG [eV]

1.16 1.18 1.20 1.22 1.24 1.26 1.28 1.30

0.30

0.35

0.40

0.45

0.50

0.55

0.60

0.65

0.70

0.75

293 K 313 K 333 K 353 K

Slo

pe

Eff

icie

ncy

η [

W/A

]

BLBG [eV]

FIG. 7: Variation of threshold current Ith (left) and slope efficiency η (right) with device structure for several different temperatures(617 µm cavity length). The highest barrierlayer bandgap (BLBG = 1.291 eV) corresponds to the measured structure.

Fig. 7 shows the calculated variation of the threshold current Ith and slope efficiency η with BLBG energy for differenttemperatures. The threshold current is minimum for barrier bandgaps in the range BLBG = 1.21 - 1.24 eV which is close tothe bandgap of the separate confinement layer (1.22 eV). The minimum is more pronounced at higher temperatures. Theslope efficiency η increases and finally saturates with decreasing BLBG energy. At high temperatures, about 1.22 eV seemto be an overall optimum for the barrier bandgap.



3.60 3.65 3.700.0

0.5

1.0

1.5

2.0

2.5

BLBG 1.166 eV

BLBG 1.291 eV

T = 293 KI = 309 mA

L = 617 um

p-side

BLBG 1.166 BLBG 1.191 BLBG 1.216 BLBG 1.241 BLBG 1.266 BLBG 1.291

Ele

ctro

n d

ensi

ty [

101

8 cm

-3]


3.60 3.65 3.700.0

0.5

1.0

1.5

2.0

2.5

BLBG 1.166 eV

BLBG 1.291 eVT = 353 KI = 699 mA

L = 617 um

p-side

Ele

ctro

n d

ensi

ty [

101

8 cm

-3]


FIG. 8: MQW electron density n(y) along the laser axis for different barrierlayer bandgaps (left: T = 20 °C, right: T = 80 °C).

To study the physical mechanisms behind this behavior, we first compare the MQW carrier density distributions at identicalinjection currents (Fig. 8). The higher the barrier bandgap the more non-uniform is the distribution of the carriers due to thelarger energy barrier for thermionic emission between quantum wells. This affects mainly the hole transport across theMQW. Higher barriers make it more difficult for the holes to penetrate from the p-side to the n-side. From this it followsthat both the holes and electrons pile up at the p-side of the active region due to coulomb-interaction. The carrier non-uniformity is less pronounced at higher temperature due to stronger thermionic emission.

However, the non-uniform MQW carrier distribution is known to affect the laser performance substantially [11]. The non-uniformity increases with rising injection current leading to enhanced Auger recombination and to enhanced electronleakage from the p-side quantum well into the cladding. The high-temperature Auger recombination rate and spontaneousrecombination rate within the five QWs is shown in Fig. 9 and Fig. 10, respectively. It is strongly non-uniform for structureswith high BLBG energies and it becomes more equalized for structures with lower barrier bandgaps. By careful inspectionof Fig. 9 one observes that the integration of the Auger recombination rate yields a minimum at an intermediate BLBGenergy.

3.60 3.65 3.700.00

0.05

0.10

0.15

0.20

0.25

0.30

0.35

0.40

BLBG 1.166 eV

BLBG 1.291 eV

p-side


T = 353 KI = 699 mA

L = 617 um

Au

ger

rec

om

bin

atio

n r

ate

[1028

cm

-3/s

]


FIG. 9: Auger recombination as calculated for differentbarrierlayer bandgaps at high current injection (T = 80 °C).

3.60 3.65 3.700.00

0.05

0.10

BLBG 1.166 eV

BLBG 1.291 eV

p-side


T = 353 KI = 699 mA

L = 617 um

Sp

on

tan

eou

s re

com

bin

atio

n r

ate

[10

28 c

m-3/s

]


FIG. 10: Calculated spontaneous emission for differentbarrierlayer bandgaps at high current injection (T = 80 °C).



Header for SPIE use

The minimum of the Auger recombination is better visible in Fig. 11 which also plots the other current loss mechanisms atboth temperatures. At room temperature, spontaneous emission contributes most to the current loss except for the structureexhibiting the highest barrier height where Auger recombination dominates. Vertical leakage is of little importance at roomtemperature. As expected, lateral leakage and SRH recombination hardly change with BLBG variation. At hightemperatures, Auger recombination is the dominant current loss mechanism for all BLBGs. It can also be seen that verticalleakage is more important at elevated temperatures. It increases with higher BLBG due to the more non-uniform MQWcarrier distribution.

1.16 1.18 1.20 1.22 1.24 1.26 1.28 1.30

20

40

60

80

100

I = 309 mAT = 293 K

L = 617 um

lateral leakage vertical leakage Auger recombination spontaneous emission SRH recombination

Cu

rren

t lo

sses

ab

ov

e th

resh

old

[m

A]

BLBG [eV]

1.16 1.18 1.20 1.22 1.24 1.26 1.28 1.30

50

100

150

200

250

300

350

400

I = 699 mAT = 353 K

L = 617 um

Cu

rren

t lo

sses

ab

ov

e th

resh

old

[m

A]

BLBG [eV]

FIG. 11: Computed current loss versus barrier bandgap at high injection (left: T = 293 K, right: T = 353 K)

5. SUMMARY

We have performed various light versus current measurements to study the temperature dependence of 1.3 µm InGaAsP/InPlaser diodes with a barrier bandgap of 1.29 eV. Based on this analysis we have investigated theoretically the effects ofvarying barrier heights on the high-temperature performance of InGaAsP/InP laser diodes employing the commercial laserdiode simulator LASTIP. In the simulation we have stepwise reduced the barrier bandgap to find an optimum design forhigh-temperature operation. Lowering the barrier height in the active region leads to an improved performance of our laserwith respect to threshold current and slope efficiency. An optimum barrier bandgap range of 1.21 - 1.24 eV is identified forour laser. This is partially attributed to the non-uniform carrier-distribution across the quantum-wells.

ACKNOWLEDGMENT

The authors would like to gratefully acknowledge C. Silfvenius for the laser structure design and growth, N. Chitica forvaluable discussions with respect to device analysis, and M. Hammar for supporting this work.



REFERENCES

[1] G. P. Agrawal and N. K. Dutta, Semiconductor Lasers. New York: Van Nostrand Reinhold, 1993.[2] J. Piprek, P. Abraham, and J. E. Bowers, “Self-consistent analysis of high-temperature effects on strained-layer

multiquantum-well InGaAsP-InP lasers”, IEEE J. Quantum Electron., vol. 36, pp. 366-374, 2000.[3] J. Braithwaite, M. Silver, V. A. Wilkinson, E. P. O’Reilly, and A. R. Adams, “Role of radiative and nonradiative

processes on the temperature sensitivity of strained and unstrained 1.5 µm InGaAs(P) quantum well lasers”, Appl.Phys. Lett., vol. 67, pp. 3546-3548, 1995.

[4] L. J. P. Ketelsen and R. F. Kazarinov, “Carrier loss in InGaAsP-InP lasers grown by hydride CVD”, IEEE J.Quantum Electron., vol. 34, pp. 811-813, 1995.

[5] J. Piprek, D. Babic, and J. E. Bowers, “Simulation and analysis of double-fused 1.55 µm vertical-cavity lasers”, J.Appl. Phys., vol. 81, pp. 3382-3390, 1997.

[6] Y. Yoshida, H. Watanabe, K. Shibata, A. Takemoto, and H. Higuchi, “Analysis of characteristic temperature forInGaAsP BH lasers with p-n-p-n blocking layers using two-dimensional device simulator”, IEEE J. QuantumElectron., vol. 34, pp. 1257-1262, 1998.

[7] A. A. Bernussi, H. Temkin, D. L. Coblentz, and R. A. Logan, “Effect of barrier recombination on the hightemperature performance of quaternary multiquantum well lasers”, Appl. Phys. Lett., vol. 66, pp. 67-69, 1995.

[8] D. A. Ackermann, G. E. Shtengel, M. S. Hybertsen, P. A. Morgan, R. F. Kazarinov, T. Tanbun-Ek, and R. A.Logan, “Analysis of gain in determining T0 in 1.3 µm semiconductor lasers”, IEEE J. Select. Topics QuantumElectron., vol. 1, pp. 250-262, 1995.

[9] J. F. Hazell, J. G. Simmons, J. D. Evans, and C. Blaauw, “The effect of varying barrier height on the operationalcharacteristics of 1.3 µm strained-layer MQW lasers”, IEEE J. Quantum Electron., vol. 34, pp. 2358-2363, 1998.

[10] M. J. Hamp, D. T. Cassidy, B. J. Robinson, Q. C. Zhao, D. A. Thompson, and M. Davies, “Effect of barrier heighton the uneven carrier distribution in asymmetric multiple-quantum-well InGaAsP lasers”, IEEE Photon. Technol.Lett., vol. 10, pp. 1380-1382, 1998.

[11] J. Piprek, P. Abraham, and J. E. Bowers, “Carrier non-uniformity effects on the internal efficiency of multi-quantum-well lasers”, Appl. Phys. Lett., vol. 74, pp. 489-491, 1999.

[12] M. J. Hamp, D. T. Cassidy, B. J. Robinson, Q. C. Zhao, and D. A. Thompson, “Nonuniform carrier distribution inasymmetric multiple-quantum-well InGaAsP laser structures with different numbers of quantum wells”, Appl.Phys. Lett., vol. 74, pp. 744-746, 1999.

[13] C. Silfvenius, G. Landgren, and S. Marcinkevicius, “Hole distribution in InGaAsP 1.3 µm multiple quantum welllaser structures with different hole confinement energies”, IEEE J. Quantum Electron., vol. 35, pp. 603-607, 1999.

[14] S. Adachi, Physical Properties of III-V Semiconductor Compounds. New York: Whiley, 1992.[15] S. Rapp, J. Piprek, K. Streubel, J. Andre, and J. Wallin, “Temperature sensitivity of 1.54 µm vertical-cavity lasers

with an InP-based Bragg reflector”, IEEE J. Quantum Electron., vol. 33, pp. 1839-1845, 1997.[16] J. Piprek, T. Troeger, B. Schroeter, J. Kolodzey, and C. S. Ih, “Thermal conductivity reduction in GaAs-AlAs

distributed Bragg reflectors”, IEEE Photon. Technol. Lett., vol. 10, pp. 81-83, 1998.[17] C. Silfvenius and G. Landgren, “Design, growth and performance of different QW structures for improved 1300

nm InGaAsP lasers”, Conference on Metal Organic Vapor Phase Epitaxy, La Jolla, California, USA, 1998.[18] J. Piprek, P. Abraham, and J. E. Bowers, “Cavity length effects on internal loss and quantum efficiency of multi-

quantum-well lasers”, IEEE J. Select. Topics Quantum Electron., vol. 5, pp. 643-647, 1999.[19] LASTIP 5.4.2 by Crosslight Software Inc., 2000.[20] Z.-M. Li, “Physical models and numerical simulation of modern semiconductor lasers”, Proc. SPIE, vol. 2994,

pp. 698-708, 1997.[21] LASTIP user manual. [Online]. Available HTTP: http://www.crosslight.com.[22] W. R. Frensley, “Heterostructure and quantum well physics”, in Heterostructures and Quantum Devices, N. G.

Einspruch and W. R. Frensley, Eds. San Diego, CA: Academic, 1994.[23] G. Margaritondo and P. Perfetti, “The problem of heterojunction band discontinuities”, in Heterojunction Band

Discontinuities, Physics and Device Applications, ed. F. Capasso, G. Margaritondo, pp. 61-114, North Holland,Amsterdam, 1987.

[24] S. Seki, W. W. Lui, and K. Yokoyama, “Explanation for the temperature insensitivity of the Auger recombinationrates in 1.55 µm InP-based strained-layer quantum-well lasers”, Appl. Phys. Lett., vol. 66, pp. 3093-3095, 1995.

[25] G. P. Agrawal and N. K. Dutta, Long-Wavelength Semiconductor Lasers, pp. 107, 122, New York: Van NostrandReinhold, 1986.



[26] L. C. Chiu, A. Yariv, “Comments on the nature of electron leakage in InGaAsP/InP double heterostructure”, Jpn. J.Appl. Phys., vol. 21, pp. L305-6, 1982.

[27] S. L. Chuang, “Efficient band structure calculations of strained quantum wells”, Phys. Rev. B, vol. 43, pp. 9649-9661, 1991.

[28] M. Asada, “Intraband Relaxation Effect on Optical Spectra”, in Quantum Well Lasers, ed. P.S. Zory, pp. 97-130,Academic Press, San Diego, 1993.

[29] L. A. Coldren and S. W. Corzine, Diode Lasers and Photonic Integrated Circuits, p. 138, New York: Wiley, 1995.[30] C. H. Henry, R. A. Logan, F. R. Merritt, and J. P. Luongo, “The effect of intervalence band absorption on the

thermal behavior of InGaAsP lasers”, IEEE J. Quantum Electron., vol. QE-19, pp. 947-952, 1983.



SPIE Proceedings [SPIE Symposium on Integrated Optics - San Jose, CA (Saturday 20 January 2001)]...

Documents

Transcript of SPIE Proceedings [SPIE Symposium on Integrated Optics - San Jose, CA (Saturday 20 January 2001)]...