UNIVERSITY OF CINCINNATI
Date:_February 07, 2006__
I, _ Rajasundaram Rajasekaran______, hereby submit this work as part of the requirements for the degree of: Master of Science in:
Electrical & Computer Engineering and Computer Science It is entitled: Dependence of LASER Performance on Number of Quantum Wells in InAlGaAs Semiconductor LASERS
This work and its defense approved by:
Chair: _Dr.David J.Klotzkin______Dr.Kenneth P. Roenker ______Dr.Fred Beyette______
Dependence of LASER Performance on Number of Quantum Wells in InAlGaAs Semiconductor LASERS
A thesis submitted to the Division of Graduate Studies and Research of the
University of Cincinnati
in partial fulfillment of the requirements for the degree of
MASTER OF SCIENCE
in the Department of Electrical and Computer Engineering and Computer Science of the College of Engineering
February 2006 By
Rajasundaram Rajasekaran Bachelor of Engineering, Bharathidasan University, India, 2001
Thesis Advisor and Committee Chair: Dr. David J. Klotzkin
Abstract
This thesis is undertaken to understand the trade-off in the performance of semiconductor lasers with the increase in well number. The five and seven quantum well InAlGaAs lasers operating at an optical communication wavelength of 1300nm were used in this investigation. The dependence of laser performance on the waveguide structure and the geometry of the laser device were also researched. Experimental setup were designed and implemented to perform DC and dynamic characterization. Numerical simulations were carried out using Mode Solver to compare the optical confinement in five and seven quantum wells. The seven quantum well InAlGaAs lasers showed better optical confinement and characteristic temperature compared to the five quantum well lasers.
The increase in well number aids in lowering threshold current and improving the optical confinement factor, characteristic temperature, net gain, material gain and d-factor with trade off in slope efficiency, differential quantum efficiency and cavity loss. The design curves were plotted to aid the laser designer to understand the performance trade off in increasing the well number. The comparison plots would aid the designer to optimize the number of quantum wells when designing uncooled semiconductor lasers for the application in optical communication. Seven quantum well InAlGaAs lasers exhibited low threshold current of 13.8mA and high characteristic temperature of 71°c. The five quantum well lasers showed high slope efficiency of 0.198W/A and a low cavity loss of
18/cm. The better modulation responses of the seven quantum well lasers were shown by determining the D-factor. This study also helps in optimizing the number of quantum wells for the given cavity length.
To mom, dad and Rachana ACKNOWLEDGEMENTS
I would like to express my heartfelt gratitude to my advisor Dr. David Klotzkin for his valuable guidance and constant encouragement throughout my graduate studies. His patience, generosity and wit have all made working under him a very memorable and thoroughly enjoyable experience. I would also like to thank him especially for creating a
good learning environment in the research group. He has been a true mentor,
academically as well as in real life. I would also like to thank Dr. Kenneth P. Roenker
and Dr. Fred Beyette along for reviewing my thesis and obliging to be in my defense
committee
I would also like to express my deep appreciation to my colleagues for helping me out in
times of need, and for making the Optoelectronic Devices Lab such a wonderful place to
work in. Many thanks go to Hua Tan, Ajit Balagopal, Ashwin Chincholi, Siddartha
Banerjee, Hui Shen, Haichuan Mu, Yaling Zhou and Ansuman Banerjee. I would like to
especially mention the help and guidance Hua Tan provided during the experimental
setup and measurements.
Finally, I am grateful to my family in India and to all my friends for their unending love,
support and encouragement throughout my academic career.
Table of Contents
1. Introduction…………………………………………………...... 8
1.1 Thesis Overview…………………………………………………………………..8
1.2 Quantum well Lasers ……………………………………………………………..8
1.3 Advantage of introducing quantum wells in semiconductor lasers……………...10
1.4 Multi quantum well lasers……………………………………………………...... 13
1.5 Advantage of graded index separate confinement heterostructure………………14
1.6 Effect of strain in quantum well lasers…………………………………………..15
1.7 Long wavelength quantum well lasers in optical communication……………….17
1.8 InAlGaAs/InP and Conventional InGaAsP/InP 1.3 µm Quantum well lasers…..18
2. Fabery- Perot In1-x-yAlxGayAs RWG MQW laser structure and determination of optical confinement factor……………………..23
2.1 Introduction………………………………………………………………………23
2.1.2 Design of Fabery-Perot 1.3µm InAlGaAs-InP MQW laser…………………...24
2.1.3 Device Structure of the fabricated InAlGaAs RM-RWG MQW laser 1………25
2.1.3.1 InAlGaAs-InP reverse mesa (RM) RWG lasers……………………………..26
2.1.3.2 Effect of the structural parameters of the active layers………………………29
2.1.3.3 In1-x-yAlxGayAs and In1-xGaxAsyP1-y energy band gap……………………….30
2.1.3.4 In1-xGaxAs energy bandgap…………………………………………………..31
2.1.4 Determination of Refractive index of each layer………………………………32
2.1.5 Optical Confinement Factor Theory…………………………………………...33
2.1.6 Simulation using MODE solutions for InAlGaAs MQW lasers……………….34
1 2.1.7 Effect of well number on optical confinement factor………………………….36
2.1.8 Conclusions…………………………………………………………………….38
3. Effects of well number and the waveguide structure on the performance of InAlGaAs MQW lasers…………………………39
3.1 Introduction………………………………………………………………………39
3.2.1 L-I characteristics plot of 1.3µm RM–RWG InAlGaAs MQW laser………….40
3.2.2 Threshold current dependence on well number for InAlGaAs MQW
Laser………………………………………………………………………………….41
3.3 Determination of Slope Efficiency, External Differential Quantum Efficiency an
Internal Quantum Efficiency of InAlGaAs MQW laser……………………………..44
3.4 Threshold current dependence on temperature for InAlGaAs MQW Laser……..45
3.5 Conclusion……………………………………………………………………….49
4. Effect of well number in Gain and Intrinsic modulation responses in InAlGaAs MQW lasers ……………………………62
4.1 Introduction………………………………………………………………………62
4.1.1 Experimental details – Gain measurement for InAlGaAs MQW lasers……….63
4.2 Gain measurement of InAlGaAs MQW lasers…………………………………..64
4.3 Gain measurement at various temperatures for InAlGaAs MQW lasers………..66
4.4 Measurement of Optical loss…………………………………………………….71
4.5 Material gain determination for InAlGaAs MQW lasers………………………..73
4.6 Experimental setup for modulation response of InAlGaAs MQW lasers………..74
4.7 Determination of Frequency response for InAlGaAs MQW lasers……………...76
4.8 Conclusion……………………………………………………………………….80
2
5. Results and Future Work……………………………………...81
5.1 Effect of well number and waveguide structure in the optical confinement
factor…………………………………………………………………………………81
5.2 Effect of well number and waveguide structure on the performance of the
InAlGaAs MQW lasers………………………………………………………………81
5.3 Effect of well number on Gain and Intrinsic modulation response in InAlGaAs
MQW lasers………………………………………………………………………….82
5.4 Scope for future work……………………………………………………………82
3 List of Figures
Fig. 1.1 Schematic diagram of Multi Quantum well lasers
Fig. 2.1 Schematic diagram of 1.3µm Fabery-Perot RM RWG InAlGaAs MQW
Fig. 2.2 SEM of 1.3µm InAlGaAs seven QW RM RWG laser
Fig. 2.3 SEM of a RWG MQW laser
Fig. 2.4 Fabery-Perot 1.3µm InAlGaAs RM RWG Lasers
Fig. 2.5 Modal confinement in Fabery–Perot InAlGaAs RM RWG MQW lasers
Fig. 3.0 Schematic Setup for Characterization of InAlGaAs MQW Lasers
Fig. 3.1 L-I plot for InAlGaAs MQW lasers at room temperature
Fig. 3.2 (a) L-I plot for 7 QW InAlGaAs- Device 1 at various temperatures
Fig. 3.2 (b) L-I plot for 7 QW InAlGaAs- Device 2 at various temperatures
Fig. 3.3 (a) L-I plot for 5 QW InAlGaAs- Device 1 at various temperatures
Fig. 3.3 (b) L-I plot for 5 QW InAlGaAs- Device 2 at various temperatures
Fig. 3.4 (a) Characteristic Temperature for device-1 7QW InAlGaAs
Fig. 3.4 (b) Characteristic Temperature for device-2 7QW InAlGaAs
Fig. 3.5 (a) Characteristic Temperature for device-1 5QW InAlGaAs
Fig. 3.5 (b) Characteristic Temperature for device-2 5QW InAlGaAs
Fig. 3.6 Slope efficiency dependence on temperature for InAlGaAs MQW lasers
Fig. 4.1 Schematic representation of the experimental setup for gain measurement
Fig. 4.2 Sample plot of modulation response at ASE
Fig 4.3 Net Gain at various current below threshold at room temperature (a) 7Qw (b) 5Qw
Fig. 4.4 Net Peak gain at various currents below threshold at room temperature
4 Fig. 4.5 (a) Net peak gain (G) versus current – 7 Qw InAlGaAs lasers
Fig. 4.5 (b) Net peak gain (G) versus current – 5 Qw InAlGaAs lasers
Fig. 4.6 Material Gain of InAlGaAs MQW lasers
Fig. 4.7 Schematic diagram of the experimental setup for frequency response of
InAlGaAs
Fig. 4.8 Frequency response for InAlGaAs MQW lasers
Fig. 4.9 (a) Frequency responses at varying temperatures – 5Qw InAlGaAs Lasers
Fig. 4.9 (b) Frequency responses at varying temperatures – 7Qw InAlGaAs Lasers
Fig. 4.10 D-factor versus temperature for InAlGaAs Lasers
5 List of Tables
Table 2.1 Refractive index of the layers of the InAlGaAs MQW laser
Table 2.2 Optical Confinement factor for five and seven InAlGaAs quantum well laser
Table 3.1 Threshold current and threshold current density of InAlGaAs MQW lasers
Table 3.2 Slope efficiency and differential quantum efficiency of InAlGaAs MQW lasers
Table 3.3 Characteristic temperature for InAlGaAs MQW lasers
Table 3.4 Summary of the DC characteristic measurement of InAlGaAs lasers
Table 4.1 Material property of InAlGaAs MQW laser dependence on Well number
Table 4.2 Parameters as function of well numbers for InAlGaAs MQW lasers
6
7 Introduction
1.1 Thesis Overview
This research is on the study of the effects of the well number in the laser function and
intrinsic modulation in both five and seven fabry-perot 1.3µm In1-x-yAlxGayAs multi
quantum well reverse mesa ridge and tapered waveguide lasers. This introduction chapter
introduces the concept of quantum well lasers, graded index separate confinement
structure, strained quantum well lasers and the application of long wavelength uncooled
semiconductor lasers in optical communication transmitters. The subsequent sections
deals with the of the 1.3µm In1-x-yAlxGayAs- InP MQW laser in terms of characteristic temperature and other parameter with respect to the conventional In1-xGaxAsyP1-y-InP
MQW lasers. It also talks about the previous study in the effect of well number on characteristic temperature in InGaAsP-InP conventional lasers. The second chapter introduce to the laser material structure and waveguide and determination of the optical confinement factor for each laser type under study. The subsequent chapters talks about the experimental setup, characterization work carried out and analysis of the dependence of laser function and intrinsic modulation on the well number. The results and discussion is done in detail in the conclusion chapter.
1.2 Quantum well Lasers
A quantum well device is one that is formed by sandwiching a narrow bandgap semiconductor between two wider bandgap semiconductor materials. The two semiconductor materials are lattice matched so that the interface defects due to mismatch of crystal dimensions between them are minimal. As it is formed by the sandwich of two different semiconductor materials there is going to be a change in the bandgap Eg at the
8 1 interface . This causes discontinuities in conduction band (Ec) and valence band (Ev) at the interface. And this change in ∆Ec and ∆Ev depend on the doping of the semiconductor materials1. The schematic diagram of a quantum well is given below in the figure 1.1.
The wavelength emission of the quantum well semiconductor device formed is different
from the respective bandgap energy of the two individual semiconductor materials1.
Figure 1.1 Schematic diagram of a single quantum well laser
Thus it could be summed up as a new material that is only remotely related to the bulk
material it is formed from. Advantage of this device is the effective carrier confinement
and the nature of the electronic density of states that results in operation at lower
threshold currents in comparison with the "bulk" active layers. In addition, the use of a quantum well, with discrete transition energy levels also provides a mean to “tune” to the desired wavelength of emission.
The gain region in quantum well in contrast with a bulk laser is a thin region and is embedded inside a larger optical confinement region1. The carriers in the quantum wells
are confined in one dimension and have discrete energy bands. To model the complete
1 P. Bhattacharya, “semiconductor optoelectronic devices,” Prentice hall, second edition, 1997
9 carrier injection one need to examine both the nature of its confinement and QW regions
and their mutual coupling2. On the other hand in bulk region both the carrier and the optical mode are confined in the same region of operation. The carrier and gain nonlinearities in QW lasers are determined, to a large extent, by the properties of the confinement region3,4,5,6. Hence a good understanding in the carrier confinement in quantum well helps to design better quantum well structure.
1.3 Advantage of introducing quantum wells in semiconductor lasers
The density of states of electrons and holes in bulk material varies as є1/2 and from Fermi-
Dirac static’s it is known that the probability of occupation of the levels in the band decreases rapidly as є increases. In a three dimensional bulk semiconductor lasers the
carriers are distributed over a wide range in the band and with relatively smaller density
at the band edges1. In the case of a two dimensional quantum well lasers the electrons are
2 N. Tessler, and G. Eisenstein “On Carrier Injection and Gain Dynamics in Quantum Well
lasers,” IEEE J. Quantum Electron. vol. 29, pp. 1586-1596. Jun. 1993
3 N. Tessler, R. Nagar, and G. Eisenstein. “Structure dependent modulation responses in quantum
well laser,” IEEE J. Quantum Electron.. vol. 28, pp. 2242-2250. Oct. 1992.
4 W. Rideout, W. f. Sharfin, E. S. Koteles, M. 0. Vassell, and B. Elman. “Well barrier hole
burning in quantum well lasers.” IEEE Photon. Technol. Lett., vol. 3, pp. 784-786, Sept. 1991
5 S. C. Kan, D. Vassilovski. T. C. Wu, and K. Y. Lau. “On the effects of carrier diffusion and
quantum capture in high speed modulation of quantum well lasers,” Appl. Phgs. Lett., vol. 61, pp.
752-754, 1992
6 P. W. M. Blom, R. F. Mols, J. E. M. Haverkort, M. R. Leys, and J. H. Wolter, “Picosecond
carrier capture by a separate confinement laser structure,” European Conf. Opt. Commun.
Amsterdam. 1990, paper MoB3.4.
10 spread over a smaller energy range with a higher density at the band edge. Thus giving
rise to population inversion at low injected carrier density than compared to a
conventional laser diode. That’s the reason why lowest threshold current are measured in
the quantum well lasers.
In general lasers with low threshold current are important for minimizing the power
consumption and cross talk of parallel optical interconnects in switching and
supercomputers as they are needed in array configuration. Quantum well lasers are used
for this purpose. The combination of biaxial strain and quantum confinement has been
proposed to reduce the in-plane hole effective mass and in turn reducing the threshold
current7.
In order to achieve low threshold current and better characteristic temperature the
confinement energy should be large to prevent the carrier overflow. And added to it these low threshold quantum well lasers require a well designed optical waveguide for high modal gain and low cavity loss8. This is achieved by opting for either separate confinement structure and graded index separate confinement structure (GRIN-SCH).
The biaxial strain in the quantum well structure helps in maximizing the carrier confinement while the GRIN-SCH helps in better optical confinement.
7 A. R. Adams, “Band-structure engineering for low threshold high efficiency semiconductor
lasers,” Electron. Lett., vol. 22, no. 5, pp. 249-250, 1986.
8 S.R. Selmic, T.M. Chou; J. Sih, J.B. Kirk, A. Mantle, J.K. Butler, D. Bour, G.A. Evans,”
Design and characterization of 1.3-µm AlGaInAs-InP multiple-quantum-well lasers”, IEEE
Journal of selected topics in Quantum Electronics, Vol. 7, No. 2, pp.340 – 349, March-April
2001
11 In a single quantum well laser the modal confinement is less but shows high carrier
confinement with compressive strain8. The modal confinement is increased in single
quantum well by the introduction of separate confinement structure. Thus the Quantum
well semiconductor laser diodes are widely used as their inherent nature of multimode operation could be exploited to produce spectrally broad gain band width lasers. And they also show better line width and temperature stability.
The application of quantum well structures to semiconductor laser diodes has received considerable attention because of its physical interest and as well as its superior laser characteristics8. By controlling the width of the quantum wells one can modify the electron and hole wave function which leads to modification of laser characteristics as well as introduction of new concepts to optical devices.
To achieve better confinement and temperature stability quantum well lasers with more than one well were introduced. These lasers were termed as multiple quantum well lasers.
In general a multi quantum lasers were found to have higher amplification per injected carrier than bulk lasers. This results in lasers with high differential quantum efficiency, a property for which multiquantum well lasers (MQW) lasers are known to excel9. Detailed characteristics of the MQW lasers would be discussed in the coming section. The gain in the quantum well laser is embedded inside a large optical confinement region. The
Carriers in the quantum well care confined in one dimension and have discrete energy
9D. Gershoni, C. H. Henry, and G. A. Baraff, “Calculating the optical properties of multi
dimensional heterostructure: application to modeling of quaternary quantum well lasers,” IEEE J.
Quantum Electron., vol. 29, pp. 2433-2450, 1993
12 bands. The property of the confinement factor gives the nonlinearity in the gain and carriers in the quantum well.
1.4 Multi quantum well lasers
In general a MQW well laser preserves the advantage of a single quantum well (SQW) laser in addition to higher light powers. The internal and differential quantum efficiencies are greater in MQW lasers for all cavity lengths. But the SQW lasers exhibit lower threshold current than MQW lasers. The lowering of threshold current in SQW due to quantum size effect could largely be offset by the small width of the gain region in a
SQW lasers and which causes the optical confinement to be poor. This problem is overcome by opting MQW structure for the gain region and separate confinement heterostructure (SCH) to enhance the optical confinement factor. And also the introduction of Graded Refractive Index – SCH results in lower threshold current.
The high amplification per carrier in a MQW laser is a consequence of the smaller active volume which requires fewer carriers to reach transparency and also due to the confinement in one dimension9. This is particularly true for compressively strained MQW lasers which have the best performance and the description on strain in quantum well lasers are seen in the subsequent sections in this chapter. In creasing the number of wells increases the modal confinement and doesn’t increase the gain which is a function of the carrier density. Previous study by Gershoni et al shows that the MQW lasers tend to have higher amplification per injected carrier in both 1.5v and 1.3µm10. Theoretically
10 Juan. A. Mart and M. Sanchez, “Comparison between a graded and step-index optical cavity in
InGaN MQW laser diodes”, Semicond. Sci. Technol., vol.20, pp. 290–295, 2005
13 predicted by Arakawa et al.11, MQW lasers have a number of superior high-speed
properties due to the quantum size effect. But previous works in MQW lasers also shows
that the well number is inversely proportional to the cavity length of the laser. Thus just
increasing the well number also improves the internal loss. Thus well number is
optimized for each corresponding cavity length of the laser.
1.5 Advantage of graded index separate confinement heterostructure
In a single quantum well (SQW) laser the carriers would have to scatter in order to lose
enough energy to be collected by the well. This would in turn increase the threshold
current and would not be advantageous. Thus the well is separated from the cladding by a
composition such that the barrier height is less than cladding. And such a structure is
called separate confinement heterostructure. In addition the bandgap of cladding region
could be altered and designed using the popular method of graded index separate
confinement structure. That would aid in the increase carrier confinement. Thus in GRIN-
SCH lasers the optical confinement is no longer dependent on the carrier
confinement12,13.
The GRIN region not only enhances carrier confinement but also assists in the thermalization of carriers into the quantum well. It acts as a funnel and enhances the collection of electrons in the active layer. It also helps in guiding the electromagnetic
11 Y. Arakawa and A. Yariv, “Theory of gain, modulation response, and spectral line width in
AlGaAs quantum well lasers,” ZEEE J. Quantum Electron., vol. QE-21, pp. 1666-1674, 1985
12 G. Agrawal and N. Dutta, Semiconductor Lasers, New York: Van Nostrand Reinhold, 1993
13 J. O’Gorman, A. F. J. Levi, T. Tanbun-Ek, D. L. Coblentz, and R. A. Logan, “Temperature
dependence of long wavelength semiconductor lasers”, Oct. 1992
14 wave in an optical fiber of graded index and thus maintaining a significant overlap of the optical and electrical distributions10.
On comparing the step index to graded index structure it is to be noted that in SQW the
former gives the highest confinement factor. For quantum well structure with well number more than six, the confinement factor is always higher in the structure with a
GRIN cavity, independent of the cladding layer width14.
1.6 Effect of strain in quantum well lasers
A method for optimizing the MQW laser is to construct a strained MQW laser. Strain in a hetero junction occurs when the lattices at the junction are different. Thin films of a lattice mismatched semiconductor can be grown without serious defects. There are two types of strain that can be applied, compressive and tensile. Strain effectively changes the band structure of the heterostructure. Due to the slow growth techniques such as MBE and MOVPE, thin films can be grown on a substrate such that they do not relax.
Relaxation is what causes undesired effects such as dangling bonds and defect states.
Tsang et al.8. found a substantial reduction in Jth for strained layer MQW (SL-MQW) lasers
The effects of strain on the band structure affects the optical properties of the lasers and are manipulated for the better designing of lasers. Two kinds of strain could be imposed in a quantum well structure and they are tensile and compressive strain. In the conduction
band the strain shifts the position of the band edge and has a small effect on the effective
14 D. Gershoni, C. H. Henry and G. A. Baraff, “Calculating the Optical Properties of
Multidimensional Heterostructure: Application to the Modeling of Quatemary Quantum Well
Lasers,” IEEE J. Quantum Electron. vol. 29, pp. 2433-2450. Sep. 1993
15 mass. On the other hand it affects the valence band structure. In addition to the change in the band gap the energy difference between the light and heavy hole band also changes.
The tensile strain tends to push the bands closer to degeneracy or will serve to moves them apart in the case of compressive strain. The compressive strain tends to help in more efficient population inversion. Resulting in the increase of differential gain and the single mode lasing occurs at lower injection currents. On the other hand in the case of tensile strain the differential gains increases1.
The other important effect due to strain is the change in the auger rates. The strain changes the separation between the band edge and the split off band. It also changes the threshold current density and carrier concentration .Since auger rate is proportional to the cube of carrier concentration it changes auger rates too. This effect is used to suppress the auger recombination channels in strained quantum well structure which in turn reduces the temperature sensitivity and aids in the design of long wavelength lasers15.
The performance of compressive strain reduces at higher temperature (to = 100k) but on the other hand the tensile strained quantum well lasers are able to operate at a much higher temperature (to 140 k) 15. Advantage of compressive strain is the hole band associated with the lasing transition has a relatively low effective mass. The other lesser advantage of compressive strain is that the well depth for confining the electrons is deeper in compressive strain. MQW lasers with a compressive strain have significantly higher differential gains both at 1.55µm and 1.3 µm. Thus it results in higher modulation
15 E. P. O’Reilly and M. Silver, “Temperature sensitivity and high temperature operation of long wavelength semiconductor lasers”, Sep.1993
16 speed than the bulk lasers. The main disadvantage of MQW lasers is that they operate at
higher carrier densities than bulk lasers16.
1.7 Long wavelength quantum well lasers in optical communication
The continuing development in providing high bandwidth fiber optical system and the
increase in demand for greater data transmission at faster time over the fiber optic cables
is dependent on the qualities offered by the semiconductor laser optical transmitters. The
semiconductor laser used for such a practical communication system should require a
robust temperature performance for application in ambient, uncooled environments where
the device may be required to operate between 40ºC and 85ºC and also should be
emitting wavelength at the spectral range preferred by the optic fibers with high gain13.
The dispersion and loss are minimal at the wavelength near 1.3µm and 1.5µm in optical fibers. The continuing development of high bandwidth fiber optical communication systems and the unceasing demand for greater data transmission capacity over fiber-optic cables is dependent on the superior qualities offered by semiconductor laser optical transmitters. At 1.3µm a standard silica-based optical fiber has zero dispersion, while minimum loss may be obtained at 1.55 µm. Highly efficient uncooled semiconductor lasers operating in the above spectral range are preferred and they are termed as long wavelength lasers1. The practical use of these lasers is impaired by an extreme sensitivity of threshold current to temperature. At room temperature and above, non-radiative Auger recombination is the dominant physical mechanism responsible for such sensitivity. On this account it was suggested that strained quantum well lasers should show reduced
16 Hersee, S. D, de Cremoux, B.; Duchemin, J. P. Some characteristics of the GaAs/GaAlAs
graded-index separate-confinement heterostructure quantum well laser structure
17 temperature sensitivity due to suppression of Auger recombination channels in these
structures17.
To improve these devices, it is important to analyze the effects that are responsible for
modulation bandwidth and also to study the temperature dependent response for these
devices.
In contrast the gain in the quantum well laser is embedded inside a large optical
confinement region. Carriers in the quantum well care confined in one dimension and
have discrete energy bands. The carrier and gain nonlinearities in quantum well are
determined to a large extent by the properties of the confinement region. The Continuous
Wavelength properties of multiple quantum well long wavelength lasers has focused
heavily on the properties of the threshold current, specifically its reduction with changes
in device structure and geometry and its temperature sensitivity. Increase in the above
threshold efficiency is also advantageous in applications where CW operation with a
minimum of current consumption is desired such a high power externally modulated
sources for use a transmitters in optical fiber communication systems18.
1.8 InAlGaAs/InP and Conventional InGaAsP/InP 1.3 µm Quantum well lasers
Uncooled semiconductor quantum well lasers operating with wavelengths of 1.3µm have
attracted much attention in optical fiber communication systems. Especially, in their
17 E. Yablonovitch and E. O. Kane, “Band structure engineering of semiconductor lasers for optical communications,” J. Lightwave Technol., vol. 6, no. 8, pp. 1292-1299, 1988
18 K. Prosyk, J. G. Simmons, and J. D. Evans, “Well number, length, and temperature dependence
of efficiency and loss in InGaAsP-InP compressively strained MQW ridge waveguide lasers at
1.3 µm,” IEEEJ. Quantum Electron., vol. 33, pp. 1360–1368, 1997
18 application to fiber-in-the-loop (FITL) and fiber-to-the-home (FTTH), they are required
to operate at high temperatures. This section implies on how the InAlGaAs lasers shows
better temperature reliability and its use in optical communication on comparison with
the conventional InGaAsP-InP.
The uncooled conventional InGaAsP-InP laser emitting at the long wavelength (optical
fiber window) shows temperature characteristics too poor to meet Bellcore’s Generic
Requirement19 for the fiber in loop applications20. Thus in present laser transmitter, thermoelectric cooler are required to keep the laser operating temperature constant. This is one of the most troublesome components in a present laser module which not only makes the module expensive and complicated but also may degrade its long term reliability. This increases the cost due to the need for set up consisting of temperature sensor current supply and a controller in the communication setup to maintain the temperature21. The reason for the poor performance of these uncooled lasers are partly
due to auger recombination in the low bandgap material and partly due to poor electron
confinement resulting due to small conduction band offset (∆Ec= 0.4 ∆Eg) of the
19 G. P. Agrawal and N. K. Dutta, “Long-Wavelength Semiconductor Lasers”, New York Van
Nostrand Reinhold, 1986
20 C. E. Zah, R. Bhat, B. N. Pathak, F. Favire, W. Lin, M. C. Wang, N. C. Andreadakis, D. M.
Hwang, M. A. Koza, T. P. Lee, Z. Wang, D. Darby, D. Flanders, and J. J. Hsieh, “High
performance uncooled 1.3µm AlxGayIn1_x_yAs–InP strained-layer quantum-well lasers for subscriber loop applications,” IEEE J. Quantum Electron., vol. 30, pp. 511–523, Feb. 1994
21 H. Wada, K. Takemasa, T. Munakata, M. Kobayashi, T. Kamijoh, “Effects of well number on
temperature characteristics in 1.3µm AlGaInAs/InP quantum well lasers,” Semiconductor Laser
Conference, vol.5, no.3,pp. 29 – 30, 1998
19 conventional lasers22. Thus even the long term reliability is also a concern. Thus effect of strain on the conventional laser structure was implemented to reduce the threshold current, auger recombination and intervalence band absorption5,6,7,8. Strained quantum well InGaAsP shows low threshold current23,24,25,26,27and maximum operating temperature23,24,27,28 but with a decrease in differential quantum efficiency of more than 1
22 W. Rideout, W. f. Sharfin, E. S. Koteles, M. 0. Vassell, and B. Elman. “Well barrier hole burning in quantum well lasers.” IEEE Photon. Technol. Lett., vol. 3, pp. 784-786, Sept. 1991
23 C. E. Zah, R. Bhat, F. J. Favire, Jr., S. G. Menocal, N. C. Andreadakis, K. W. Cheung, D. M.
Hwang, M. A. Koza, and T. P. Lee, “Low- threshold 1.5 µm compressive-strained multiple and single quantum well lasers,” IEEE J. Quantum Elecrron., vol. 27, pp. 1440-1450, June 1991
24 C. E. Zah, R. Bhat, B. Pathak, C. Caneau, F. J. Favire, N. C. Andreadakis, D. h4. Hwang, M. A.
Koza, C. Y. Chen, and T. P. Lee, “Low threshold 1.5 µm tensile strained single quantum well lasers,” Electron. Lett., vol. 27, pp. 1414-1415, 1991
25 P. J. A. Thijs, L. F. Tiemeijer, P. I. Kuindersma, J. J. M. Binsma, and T. van Dongen, “High performance 1.5 µm wavelength InGaAs-InGaAsP strained quantum well lasers and amplifiers,”
IEEE J Quantum Electron., vol. 27, pp. 1426-1439, June 1991
26 P. J. A. Thijs, T. van Dongen, L. F. Tiemeijer, R. W. M. Slootweg, and J. J. M. Binsma,“High output power (380 mW) low threshold current (1.3 mA), low linewidth enhancement factor (5
2)X 1.3 µm strained quantum well lasers,” Opt Commun. vol. 2, Paris, France, Sept. 1991, 48-51
27 P. J. A. Thijs, J. J. M. Binsma, L. F. Tiemeijer, and T. van Dongen, “Submilliamp threshold current (0.62 mA at 0ºC) and high output power (220 mW) 1.5 µm tensile strained InGaAs single quantum well lasers,” Electron. Lett., vol. 28, no. 9, pp. 829-830, 1992
28 T. Namegay, A. Kasukawa, N. Iwai, and T. Kikuta, “High temperature operation of 1.3 pm
GaInAsP/InP GRINSCH strained-layer quantum well lasers,” Electron. Lett., vol. 29, no. 4, pp.
392-393, 1993
20
db for the temperature range of 25ºC to 80ºC26,27. This factor affects both the system link power budget and modulation bandwidth. Thus the choice for InAlGaAs-InP with large conduction band offset (∆Ec= 0.72 ∆Eg) with stronger electron confinement in the wells
is opted29,30. Thus these uncooled InAlGaAs lasers are more useful in the optical
communication system because of the better carrier confinement gain and better
characteristic temperature compared to the conventional InGaAsP. Superior temperature
characteristics have been reported by Zah et al24. In general in quantum well structure the
carrier leakage through the barrier layers is one important factor causing the laser performance degradation at high temperature and high carrier density. Due to large conduction band offset we have better confinement. By introducing biaxial strains the compressive strain and tensile strain better performance of the device are obtained. It is also easier to form graded index structure by just changing the values of x in InAlGaAs compared with InGaAsP where we need to change both x and y24. The 1.3µm InAlGaAs-
InP lasers have been studied by several groups and a characteristic temperature of
29 Y. Sugiyama, T. Inata, T. Fujii, Y. Nakata, S. Muto, and S. Hiyamizu, “Conduction band edge
discontinuity of InoszGao r In, 52(GalAL)o 48AS (0 5 2 5 1) heterostructure,”Japan. J. Appl.
Phys., vol. 25, no. 8, pp. L648-L650, 1986
30 C. G. Van de Walle, “Band lineups and deformation potentials in the model-solid theory,”
Phys. Rev B, vol. 39, no. 3, pp, 1871-181, 1989
21 threshold current higher than 100K and lasing operation up to 185ºC have been demonstrated31,32,33.
31 H. Temkin, D. Coblentz, R. A. Logan, J. P. van der ZieL T. Tanbun-Ek,R. D. Yadvish, and A.
M. Sergent, “High temperature characteristics of InGaAsP/InP laser structures,” Appl. Phys. Lett., vol. 62, pp. 2402-2404, May 1993
32 T. R. Chen, P. C. Chen, J. Unger, M. A. Newkirk, S. Oh, and N. Bar Chaim, “Low threshold and high temperature operation of InGaAlAs-InP lasers,” IEEE Photon. Technol. Lett., vol. 9, pp.
17–18, 1997
33T. Ishikawa, T. Higashi, T. Uchida, T. Fujii, T. Yamamoto, H. Shoji, and M. Kobayashi,
“Evaluation of differential gain of 1.3µm AlGaInAs/InP strained MQW lasers,” in Proc. 10th Int.
Conf. InP and Related Materials, 1998, pp. 729–732, paper ThP-55
22 Chapter 2
Fabery- Perot In1-x-yAlxGayAs RWG MQW laser structure and
determination of optical confinement factor
2.1 Introduction
The devices studied in this investigation are InAlGaAs five and seven multi quantum well Fabery-Perot reverse mesa ridge waveguide operating at 1.3µm. The aim of this chapter is to study the structure of the fabricated lasers and determine their optical confinement factor. In the section 2.1.2-2.1.7 the structure of the fabricated device and determination of refractive index of each layer is presented. In the subsequent sections the layout of the fabricated quantum well lasers are created using Lumerical- MODE solution. Then the layout is simulated and the optical confinement factors for the quantum well lasers are obtained. From the simulated results the optical confinement in each well is also determined for both seven and five quantum well lasers. Effect of the well number on the optical confinement factor is then studied from the determined values. Determination of device rules for picking the number in an InGaAlAs system could then be determined and dependence of device properties on number of quantum wells could be studied. Optical mode confinement comparison between the seven and five InAlGaAs quantum well lasers is carried out.
23
2.1.2 Design of Fabery-Perot 1.3µm InAlGaAs-InP MQW laser
The 1.3µm InAlGaAs MQW laser structure is grown by molecular beam epitaxy on an
InP-substrate. The active region is composed of (~X Å thick) compressively-strained
InAlGaAs wells separated by (~Y Å thick) tensile-strained InAlGaAs barriers. The number of wells and barriers vary for five and seven QW laser as 5 wells/4 barriers and 7 wells/6 barriers respectively. The undoped active region (~1µm thick) is sandwiched by the InAlGaAs confinement layers. The confinement layers are undoped and linearly graded from the active region to the cladding layer. On both sides of these separate confinement layers p-cladding and n-cladding regions are grown. The n-cladding regions are lattice matched to the InP substrate. A p-cap layer is grown over the p-cladding region for the ohmic contact. The epitaxial layers of the Fabery- Perot devices are then processed into ridged waveguides to confine optical waves in the lateral direction. The ridge width of the reverse mesa ridge waveguide under discussion is of size 2µm. These
InAlGaAs RWG used for this investigation are fabricated with constant cavity length of
300µm. A general schematic diagram of the Fabery-Perot RM RWG InAlGaAs MQW laser is shown below in the fig 2.1.
24 p-contact
p-cladding
etch stop p-spacer p-transition p-cladding p-GRIN
undoped active region
n-GRIN n-cladding n-transition GRIN
n-substrate
Fig 2. 1 Schematic diagram of 1.3µm Fabery-Perot RM RWG InAlGaAs MQW
A detail study on the QW layers and waveguide structure of MQW laser is done in the
following sections.
2.1.3 Device Structure of the fabricated InAlGaAs RM-RWG MQW laser
One of the aims of this work is to understand the effect of well number on the optical confinement factor of the Fabery-Perot 1.3 InAlGaAs RWG lasers. In order to study this effect it necessary to get an insight to the ridge waveguide structure and QW layers incorporated in the investigated InAlGaAs MQW lasers. To start with the SEM image of the investigated 1.3µm seven quantum well InAlGaAs RM RWG laser fabricated is
25 shown below in the fig 2.2. The RM RWG and the QW region of the fabricated device
under investigation are indicated in the SEM image. It also gives the dimension of the
ridge neck and the cap.
4 µm
RM RWG
2 µm
Active Layers
Fig 2. 2 SEM of 1.3µm InAlGaAs seven QW RM RWG laser
2.1.3.1 InAlGaAs-InP reverse mesa (RM) RWG lasers
A waveguide is defined as a region of dielectric through which light is propagated, surrounded by dielectric regions having a smaller dielectric constant34. In order to form a
waveguide various techniques are employed that would effectively and selectively create regions of varying refractive index. In order to obtain single-mode propagation it is necessary to delineate the guiding region in the lateral direction by causing a refractive index change. This is achieved in a Ridge Waveguide (RWG) structure. It has been reported that when RWG is combined with MQW active regions they show wide
34 P. Bhattacharya, “semiconductor optoelectronic devices,” Prentice hall, second edition, 1997
26 temperature range (WTR) performance35,36,37,38,39, simpler fabrication with high process uniformities3, low parasitics6 and high reliability3,5. Since these structures don’t employ any buried semiconductors it relaxes the leakage current and long term reliability limitations to a great extent. These reported ridge wave guide employs vertical side walls and are termed as vertical mesa ridge waveguide (VM RWG). Disadvantages of these
RWG are limitation to WTR and high power performance due to their high electrical and
35 K. Y. Liou, W. T. Tsang, F. S. Choa, E. C. Burrows, G. Raybon and C. A. Burrus, “Low threshold and high-temperature operation of 1.55µm self-aligned ridge-waveguide multiple- quantum-well lasers grown by chemical-beam epitaxy,” Appl. Phys. Lett., vol. 59, no. 26,pp.
3381–3383, 1991
36 N. Matsumoto, T. Fukushima, H. Nakayama, Y. Ikegami, T. Namegaya, A. Kasukawa, and M.
Shibata, “High-reliability and high-temperature operation of GaInAsP/InP multiple-quantum-well ridge-waveguide lasers emitting at 1.3µm with an excellent process-uniformity,” in Tech. Dig.
ECOC ’93, Montreux, Switzerland, 1993, ThP11.5
37 B. Stegmuller, E. Vuhoff, J. Rieger, and H. Hedrich, “High-temperature (130°C) CW operation of 1.53µm InGaAsP ridge-waveguide lasers using strained quaternary quantum wells,” Electron.
Lett., vol. 29, no.19, pp. 1691–1693, 1993
38 C. E. Zah, R. Bhat, B. Pathak, F. Favire, M. C. Wang, W. Lin, N. C. Andreadakis, D. M.
Hwang, M. A. Koza, T. P. Lee, Z. Wang, D. Darby, D. Flanders, and J. J. Hsieh, “High- performance uncooled 1.3µm AlGaInAs/InP strained-layer quantum-well lasers for fiber-in-the- loop applications,” in Tech. Dig. OFC ’94, San Jose, CA, 1994, ThG1
39 A. P. Wright, A. T. R. Briggs, A. D. Smith, R. S. Baulcomb, K. J. Warbrick, “22 GHz- bandwidth 1.5µm compressively strained InGaAsP MQW ridge-waveguide DFB lasers,”
Electron. Lett., vol. 29, no. 21, pp. 1848–1849, 1993
27 thermal resistances and they also have large threshold current. This lead to the proposal of a novel InP- based RWG structure by Mashiro Aoki et al, to improve the performance of the long wavelength RWG lasers. The proposed RWG structure depicted a reverse trapezoidal shape and coined as the reverse mesa ridge waveguide (RM RWG). From the investigation carried by mashiro et al, it is noted that the advantage of the RM over the conventional VM RWG is that they offer lower threshold current smaller electrical and thermal resistance and smaller waveguide losses. This novel structure is quite suitable for
WTR and high power operations and can be used for practical purposes. This RWG structure when employed in Fabery-Perot long wavelength lasers shows better characteristic temperature and decrease in slope efficiency at high temperatures. From the
SEM image shown in the previous section it could be noted that that shape of the ridge waveguide is a reverse trapezoid. The Fabery-Perot laser under study with the RM RWG and strained quantum well should show better temperature characteristics due to small amount of lateral current leakage which would be seen in the next chapter
(a) Reverse Mesa (RM) (b) Vertical Mesa (VM)
Fig 2.3 SEM of a RWG MQW laser13
28
2.1.3.2 Effect of the structural parameters of the active layers
In quantum well structures, the carrier leakage through the barrier layer is one important
factor causing the laser performance degradation at high temperature and high carrier density. To improve the temperature characteristics the InAlGaAs material system has been chosen to replace the conventional InGaAsP system due to higher conduction band offset and better carrier confinement as seen in the introduction chapter. The energy difference between the well and the barriers helps to suppress the leakage of electron at high temperature and provides better electron confinement An overview of the layers of the active region on the MQW laser is given in the section 2.1.2. The carrier over flow is one of the factors that influence the temperature characteristics of QW lasers. This is overcome by increasing the doping of the cladding regions. Previous study shows that increasing the number of wells helps in reducing the threshold current of the MQW lasers. This could be optimized by optimizing the cavity length which is inversely proportional the number of wells. Effect of strain on the band structure gives rise to changes in optical properties, which could be used to the advantage in designing better lasers which is seen in the section of strained quantum well lasers in the introduction chapter. Thus the compressive strained well in this laser helps in improving the suppression of career leakage. While the tensile strain of the barriers avoids the crystal degradation due to strain accumulation of the wells in the MQW laser. To optimize optical confinement factor in the fabricated laser graded index separate confinement
29 structure is implemented40. Another practical method by which electron confinement is improved is by the introduction of electron stop layer. Electron stop layers do not impede the carrier injection to the active region at the same time reduces the thermionic emission of carriers out of the active region and increases the value of internal quantum efficiency41.
2.1.3.3 In1-x-yAlxGayAs and In1-xGaxAsyP1-y energy band gap
To calculate the physical parameter of the In1-x-yAlxGayAs a linear interpolation between different binary semiconductors InAs , AlAs and GaAs are used to formulate the equation
P(In1−x−y AlxGay As) = P(InAs)(1− x − y) + P(AlAs)x + P(GaAs)y (eV) (2.1)
42,43 The band gap of the lattice matched In1-x-yAlxGayAs s given by
40 T. Namegaya, N. Matsumoto, N. Yamanaka, N. Iwai, H. Nakayama, and A. Kasukawa,
“Effects of well number in 1.3µm GaInAsP/InP GRIN-SCH strained-layer quantum-well lasers,”
IEEE J. Quantum Electron., Vol. 30, pp. 578–584, 1994
41 K. Takemasa, T. Munakata, M. Kobayashi, H. Wada, and T. Kamijoh, “1.3µm AlGaInAs-
AlGaInAs strained multiple quantum well lasers with a p-AlInAs electron stopper layer,” IEEE
Photonics Technology Letters., Vol.10, pp.495–497, April 1998
42 J. Minch, S. H. Park, T. Keating, and S. L. Chuang, “Theory and experiment of In1-xGaxAsyP1-y and In1-x-yAlxGayAs long wavelength strained quantum well lasers,” IEEE Journal of Quantum
Electron., Vol. 35, pp.771–782, May 1999
43 M. Jain, C. Bryce, S. Pinches and C. N. Ironside, “Broadening of gain spectrum using
InGaAs/InAlGaAs multiple width quantum wells at 1550nm,” Conference on Lasers and Electro-
Optics, paper CTh053, May 2002
30 2 2 Eg (In1−x−y AlxGay As) = 0.36 + 2.093x + 0.629y + 0.557x + 0.436y +1.013xy − 2.0xy(1− x − y) (eV) (2.2)
The band gap of the quaternary device could be determined given their operating wavelength 1.3 µm and using the formula.
hc E = (ev) (2.3) g λ
Where c is the speed of light and h is the planks constant. Since the well and the barrier region in the fabricated device are strained. It is necessary to include the effect of strain.
In the quantum well laser the lattice constant of the quaternary epitaxial layer a and ao the lattice constant of the substrate are used44
a − a ε = ε = ε = 0 (2.4) xx yy a
And
⎛ c ⎞ ⎜ 12 ⎟ ε zz = −2⎜ ⎟ε (2.5) ⎝ c11 ⎠ where c12 and c11 are elastic constants.
2.1.3.4 In1-xGaxAs energy bandgap
The linear interpolation of the material parameter M is given using the binary semiconductors InAs & GaAs is given by
44 S.R. Selmic, T.M. Chou; J. Sih, J.B. Kirk, A. Mantle, J.K. Butler, D. Bour, G.A. Evans,”
Design and characterization of 1.3µm AlGaInAs-InP multiple-quantum-well lasers”, IEEE
Journal of selected topics in Quantum Electronics, Vol. 7, No. 2, pp.340 – 349, March-April
2001
31 M (In1− x Ga x As ) = M (InAs )(1 − x) + M (GaAs )x (eV) (2.6)
1,11 Energy bandgap ( Eg ) for In1-xGaxAs is given
2 Eg (In1−xGax As) = 0.36 + 0.505x + 0.555x (eV) (2.7)
2.1.4 Determination of Refractive index of each layer
To determine the optical confinement factor of the InAlGaAs MQW laser it is necessary to calculate the refractive index of the each layer. This could be used to estimate the effective refractive index and thus the optical confinement in the lateral direction in the quantum well laser. The laser is fabricated such that each layer is lattice matched to the layer on which it is grown. From the lattice constants and with the equations obtained in section 3.15 and 3.16 refractive indexes of each layer are calculated and tabulated in table
2.2. From the tabulated value it could be inferred the fabricated device as a lateral step index. Lateral step index is an important criterion as only for an appropriate step index value, the laser operates in single lateral mode4. To obtain the desired step index value the thickness of the p-spacer and GRIN layer are adjusted4.
Layer Strain Refractive Index
InP-cap lattice match 3.1
InGaAs-cladding lattice match 3.1
InP-cladding lattice match 3.2
InGaAsP-etch stop lattice match 3.34
InP-cladding lattice match 3.3
InAlAs-cladding lattice match 3.23
InAlGaAs-GINSCH lattice match 3.23
32 InAlGaAs-barr/2 tensile strain 3.34
InAlGaAs-Wells compression strain 3.49
InAlGaAs- barriers tensile strain 3.34
InAlGaAs- barr/2 tensile strain 3.34
InAlGaAs-GRNSCH lattice match 3.23
InP-buffer 3.2
InP substrate
Table 2.1 Refractive index of the layers of the InAlGaAs MQW laser4
The determination of the optical confinement factor by estimating the effective refractive index is shown in the coming section.
2.1.5 Optical Confinement Factor Theory
The optical confinement factor is defined as the fraction of the optical power of the mode contained in the active quantum-well layer1. Thus the optical confinement factor is denoted by Γ is given by the equation45,46,47
Γ = N wγLz (2.8)
45 S.H. Park, “High temperature characteristics of strained InGaAs/InGaAlAs quantum well lasers” Japanese Journal of Applied Physics., Vol.36, pp.3528-3530, June 1997
46 M. Aoki, M. Komori, T. Tsuchiya, H. Sato, K. Nakahara, and K. Uomi, “InP-based reversed- mesa ridge-waveguide structure for high performance long-wavelength laser diodes,” IEEE
Journal of selected topics Quantum Electron., Vol. 3, pp. 672–683, April 1997
47 C. Weisbuch, B. Vinter, “Quantum Semiconductor Structures (Fundamentals and
Applications)”, Academic Press, Inc. Harcourt Brace Jovanovich, Publishers, Chapter 5, 1991
33 Where the number of wells is Nw , Lz is the well thickness and γ is the optical confinement factor per unit well.
The laser threshold condition in terms of optical confinement factor, modal gain (G ) and internal loss (α )48 is given by
ΓG = α (2.9)
The optical confinement plays a major role in the determined the threshold current of the device. In the next chapter the dependence of well number on the gain of the InAlGaAs quantum well lasers are determined. Since optical confinement factor is an important parameter that affects the gain of the laser. In the upcoming section the optical confinement factor of the fabricated InAlGaAs five and seven quantum well lasers are determined using the Lumeical – MODE Solutions, simulation software. Its dependence on the number of wells is also determined.
2.1.6 Simulation using MODE solutions for InAlGaAs MQW lasers
The first step to determine the optical confinement factor is to calculate the effective refractive index of the laser structure. Then using the layout editor the layout of the
InAlGaAs quantum well laser structure is created for both seven and five quantum well lasers. Refractive index of the value determined from the previous section for each layer is included in the layout parameter. The ridge waveguide helps in determining the mode confinement. Thus it is important to reproduce these shapes in layout structure. But the shape of the RM and tapered ridge waveguide structure of the QW laser couldn’t be reproduced using the layout editor. As trapezoidal structure couldn’t be drawn using the
48 Y. Peng, B. Wang, H. Sun, W. Chen, S. Liu, “Design of quantum structure stripe lasers for low threshold current”, Optical and Quantum Electronics., Vol. 31 , pp23-28, January 1999
34 layout editor. Thus an equivalent vertical mesa ridge waveguide structure is drawn for both types of InAlGaAs MQW lasers. In the equivalent layout of the reverse mesa ridge waveguide the width of the vertical mesa ridge neck is around 2 µm. On the other hand in the case of tapered waveguide it is 0.5µm wide. The layout of each QW device is shown below. The fig 2.5 (a) and (b) shows the layout for five and seven InAlGaAs RM RWG
QW lasers
MQW region
(a) Five QW (b) Seven QW
Fig2.4 Fabery-Perot 1.3µm InAlGaAs RM RWG Lasers
The area selected as shown in the layout diagram by the orange rectangular box is simulated to determine if it mode propagation in the ridge waveguide is single mode or not and also to find the optical confinement factor. Setting the calculation parameters for the layout created the confinement factor and the effective refractive index of both the
35 lasers are calculated. The mode confinement simulation images of the quantum well lasers are given below in Fig 2.6(a) and (b).
(a) Five QW laser (b) Seven QW laser
Fig 2.5 Modal confinement in Fabery–Perot InAlGaAs RM RWG MQW lasers
2.1.7 Effect of well number on optical confinement factor
The layout of the InAlGaAs five and seven QW lasers are simulated to check whether a single mode confines in the ridge waveguide region. From the simulated result it is understood there is single mode confinement in the case of narrow tapered waveguide but not perfect single mode confinement in the case of reverse mesa ridge waveguide. Thus better optical confinement in the case of tapered waveguide is predicted. Once the layout is properly constructed and simulated the optical confinement factor is determined for the both lasers by specifying the active region. From the literature survey it is inferred that for InGaAsP/InP MQW lasers the optical confinement factor will increase with the well number and it is effective at low threshold current10. The tabulation results give the
36 optical confinement factor for five and seven InAlGaAs QW lasers for reverse mesa and tapered ridge waveguide.
% Five QW RM Seven QW RM
Γtotal
( Active Region) 10.902 17.8646
Γind 0.507/0.519/1.058/1.04 0.560/0.569/0.578/0.587/1.
( per QW) 1/1.022 187/1.152/1.115
Γavg
0.829 0.821
Table 2.2 Optical Confinement factor for five and seven InAlGaAs quantum well laser
In the table 2.2 Γtotal denotes the overall % confinement factor of the active region Γavg denotes average % confinement factor andΓind denotes % individual confinement factor per unit well. In the literature survey it states that the optical confinement factor per well is independent of the number of wells in the case of GaAs/AlGaAs MQW lasers49. From the table it is inferred that well number does have effect on the total optical confinement factor.
49 A. Kurobe, H. Furuyama, S. Naritsuka, N. Sugiyama, Y. Kokubun and M. Nakamura, “Effects of well number, cavity length, and facet reflectivity on the reduction of threshold current of
GaAs/AlGaAs multi quantum well lasers,” IEEE Journal of Quantum Electron., Vol. 24, pp. 635-
640, April 1988
37 2.1.8 Conclusions
The structure of the fabricated InAlGaAs five and seven quantum well lasers are studied in detail. The refractive indexes of the various layers are calculated with the given parameters of the fabricated device. Using the MODE Solution software the single mode confinement is determined in these quantum well lasers. The optical confinement factors for two different devices were found using the simulation tool. From the obtained data it could be concluded that the % optical confinement factor of the structure increases with the increase in the number of well. On the other hand the % average confinement per individual well remains the same for the two devices.
38 Chapter 3
Effects of well number and the waveguide structure on the performance
of InAlGaAs MQW lasers
3.1 Introduction
The previous chapters gave an overview on the advantage of 1.3µm InAlGaAs MQW lasers over the conventional InGaAsP MQW laser and its enormous application in the optical communication devices. Furthermore a detailed discussion was carried on the structural design of the InAlGaAs MQW lasers, and the confinement factors were calculated. This chapter takes the next step of analyzing and examining the quality, performance and characteristics of the lasers being researched. As a result a methodical approach is pursued to extract the principal laser parameters and analyze its dependence on well number and waveguide structure. To facilitate the extraction of the geometrical and material properties of the lasers an experimental setup is designed and constructed.
The first section of this chapter gives a detailed description on the arrangement of the experimental setup utilized for the characterization of the lasers. Then the subsequent sections deal in the determination of geometrical and structural parameters of the lasers such as threshold current, threshold current density, characteristic temperature, slope efficiency, differential quantum efficiency, internal quantum efficiency and dynamic series resistance. From this analysis a detail study is carried on to understand the dependence of these parameters on well number. It also discusses the previous studies that were carried out in determining the dependence of well number on the performance of the laser.
39
3.2 Experimental arrangement to test and characterize InAlGaAs MQW laser
This section gives a detail description on the experimental setup that is designed and implemented to test and characterize the InAlGaAs MQW lasers. The setup composed of laser diode controller, laser stage, optical spectrum analyzer, power detector, lensed fiber, fiber coupler, GPIB bus and a computer for data acquisition. The schematic diagram of the experimental setup is shown below in fig 3.0
Computer - Laser Diode Lab View Controller
Fiber GPIB Power Detector Coupler Laser Stage
Fig 3.0 Schematic Setup for Characterization of InAlGaAs MQW Lasers
The laser stage is an arrangement of heat sink stage, probes and a microscope. The heat sink stage is composed of vertical stand, copper place and thermo electric cooler. The thermo epoxy resin was used to bond the copper plate and the TE cooler to the vertical stand. The stage is built such that when the temperature is varied there is a uniform distribution of heat on the copper plates on which eventually the laser would be placed during the investigation. The temperature of the stage could be varied from 20ºc or lower to 100ºc with the aid of a laser diode controller. The temperature fluctuation in the laser stage is determined to be ± 0.5ºc for the setup implemented. The power detector is placed in alignment with the quantum well laser facet to detect the output optical power lased.
Once the laser is checked for lasing then a lensed fiber is used to couple to optical power
40 from the laser facet to the power detector. A lensed fiber is used instead of directly aligning the power detector as the rest of the experiments require the need for the fiber splitter. The amount of power loss incurred due to the usage of lensed fiber is detected to be -4dBm. An alignment setup is also arranged with the laser stage to align the lensed fiber to the InAlGaAs laser cavity stage. This lensed fiber is placed in a fiber holder in the alignment arrangement. The lensed fiber is then connected to the connected the power detector. The lensed fiber alignment with laser cavity is checked with the help of power detector. The lensed fiber is used to optimize maximum coupling of light, as loss due to coupling of light through the fiber is a major drawback in the setup. As shown in the diagram the laser diode controller and the power detector are connected through a GPIB bus to a computer. A specifically built in VI (Virtual Instrument) in LabVIEW 6.1 is programmed to acquire data from the power meter and optical spectrum analyzer. The drawbacks of the experimental setup are that the laser is not isolated from the surrounding temperature and vibrations, and there is a quite bit of loss due to coupling of light through the lensed fiber. Care should be taken to optimize the alignment of the setup. Each experiment is carried out for four sets of each type of InAlGaAs MQW lasers.
3.2.1 L-I characteristics plot of 1.3µm RM–RWG InAlGaAs MQW laser
An important characteristic of a laser to be measured is the amount of light it emits as current is injected into the device. This measurement generates the L-I plot. From the L-I plot the threshold current (Ith) of a laser is determined. The threshold current is the value at which the straight line hits the horizontal axis drawn by extending the lasing portion of
41 the L-I plot50. In general a low threshold current value is desirable as it would result in more efficient devices. It is a one measure used to quantify the performance of a laser. It depends upon the quality of the semiconductor material from which the device is fabricated and the geometrical design of the waveguide structure. It is also dependent on the geometry of the laser device. Thus when comparing the threshold current values of different devices, it is more appropriate to refer to threshold current density 1,51.
Threshold current density is denoted by the symbol (Jth) and is determined by dividing the experimentally obtained threshold current value Ith by the area of the laser. Threshold current density is one of the parameters that give a direct indication of the quality of the semiconductor material from which the device is fabricated. The experimental setup described in the above section 3.2 is used to plot the L-I curve of the InAlGaAs MQW lasers being researched. The L-I plot for the two sets of devices measured at room temperature are shown below in the figure 3.1. From the obtained characteristic L-I curve the parameters the threshold current and threshold current density are determined and tabulated in for the four sets of devices in the tabular column 3.1. The figure 3.1 displays the plot of just one device of seven and five quantum well InAlGaAs lasers.
While the tabular column gives the values measured for the rest of the tested devices whose plots are not shown here. The current is varied from (0 mA-30 mA) in steps of 1 mA using the laser diode controller.
50 P. Bhattacharya, “semiconductor optoelectronic devices,” Prentice hall, second edition, 1997
51 Peter S.Zory, “Quantum well lasers,” Academic Press, Inc 1993
42
Fig 3.1 L-I plot for InAlGaAs MQW lasers at room temperature
Average
Average threshold InAlGaAs threshold current MQW Threshold current in (mA) Threshold current density in (A/cm2) current density lasers in (mA) in
(A/cm2)
7 QW 14.17 13.88 13.77 14.32 14.03 1889 1850.6 1835.81 1909.28 1871.17
5 QW 17.18 16.84 17.52 16.90 17.11 2290.6 2245.10 2335.94 2253.70 2281.33
3.1 Threshold current and threshold current density of InAlGaAs MQW lasers
From the plot shown in figure 3.1 and from the table the variation in the threshold current value is evident. This variation in Ith and Jth are due to their dependence in the well number and shape of the ridge waveguide structure. The obvious conclusions that could be drawn are (a) threshold current and threshold current density decrease with increase in the number of wells. This is due to the increase in optical confinement factor to the gain region. This dependence of threshold current on well number and waveguide structure are
43 discussed in detail in the following sections. The average threshold current densities of five and seven quantum well InAlGaAs lasers are determined to be 2281.33 A/cm2 and
1871.17 A/cm2. The L-I plot is also used to determine the slope efficiency which is also dealt in the forthcoming sections.
3.2.2 Threshold current dependence on well number for InAlGaAs MQW Laser
Previous study states that threshold current decreases with increase in the well number for an optimum cavity length in quantum well lasers52,53,54,55,56,57. The cavity length is an
52 A. Kurobe, H. Furuyama, S. Naritsuka, N. Sugiyama, Y. Kokubun, and M. Nakamura, “Effects of well number, cavity length, and facet reflectivity on the reduction of threshold current of
GaAs/AlGaAs multiquantum well lasers,” IEEE J. Quantum Electron., vol. 24, pp.635–640, 1988
53 C. P. Seltzer, S. D. Perrin and P. C. Spurdens, “Low threshold current, high output power buried heterostructure MQW lasers with strained InGaAsP wells,” Electron. Letter., vol. 28, pp.
1819–1821, 1992
54 T. Namegaya, N. Matsumoto, N. Yamanaka, N. Iwai, H. Nakayama, and A. Kasukawa,
“Effects of well number in 1.3µm GaInAsP/InP GRIN-SCH strained-layer quantum-well lasers,”
IEEE J. Quantum Electron., vol. 30, pp. 578–584, 1994
55 K. Prosyk, J. G. Simmons, and J. D. Evans, “Well number, length, and temperature dependence of efficiency and loss in InGaAsP-InP compressively strained MQW ridge waveguide lasers at
1.3 µm,” IEEE J. Quantum Electron., vol. 33, pp. 1360–1368, 1997
56 M. Silver and E. P. O’Reilly, “Optimization of long wavelength InGaAsP strained quantum- well lasers,” IEEE J. Quantum Electron., vol. 31, pp. 1193–1200, 1995
57 P. Blood, E. D. Fletcher, and K. Woodbridge, “Dependence of threshold current on the number of wells in AIGaAs-GaAs quantum well lasers,” Applied Physics Letter., vol. 47, pp. 193-195,
1985
44 important geometrical parameter which plays a major role in the laser performance. The reason for the decrease in threshold current density with the increase in the number of well is due to the fact that the optical confinement factor in the well increases with well number5. The increase in the optical confinement factor with the increase in the well number was discussed and shown using the Lumerical Simulation in the previous chapter. At the same time increase in well number also gives rise to larger internal loss and large beam divergence5. Hence an optimum number of wells are designed for the desired cavity length of the laser. Threshold current is also depends on the thickness of the well. Both the cavity length (300µm) and the thickness of the well are kept a constant in these InAlGaAs MQW lasers being studied. The low threshold current density of seven quantum well indicates that it need less amount of injection current to start lasing compared with the five quantum well lasers.
3.3 Determination of Slope Efficiency, External Differential Quantum Efficiency an
Internal Quantum Efficiency of InAlGaAs MQW laser
As desired to have a low threshold current as possible it is also important to have rapid increase in output light emission for small change in the input current above the threshold current. This measurement of the change in the output power versus injected current
(∆P/∆I) above the threshold current is coined as the term slope efficiency1. The slope efficiency of each device is determined from the L-I characteristic curve plotted in the fig
3.1. Some of other important parameters that are used to gauge the performance of a laser are determined from the calculated slope efficiency value. The external quantum efficiency (ηd) which is a figure of merit, measured in percentage indicates the efficiency of laser device in converting the injected electron hole pairs to photons emitted from the
45 device. In a theoretical laser the recombination of each electron- hole pair results in the generation of one photon and additionally the photon survives its travel through the laser waveguide structure and is emitted from the device2. In a practical laser, some electron- hole pair recombination results in the generation of photons, while the rest results in the generation of other undesirable forms of energy, such as heat. In addition not all the photons are emitted from the device some of them are reabsorbed by the structure2.
Hence the value of ηd is not 100% for a fabricated laser device. The value of the external quantum efficiency is calculated from the slope efficiency value using the below formula
⎡ ∇ P ⎤ ⎡ λ q ⎤ η d = ⋅ (3.1) ⎣⎢ ∇ I ⎦⎥ ⎣⎢ hc ⎦⎥
λ- wavelength of the laser, h-planks constant, c-speed of light
Both slope efficiency and differential quantum efficiency are dependent on the cavity length of the laser devices58. In general these parameters are dependent on the laser geometry. The internal quantum efficiency (ηi) is a measure of the efficiency of a laser in converting electron-hole pairs (injected current) into photons within the laser structure.
Unlike the external quantum efficiency the internal quantum efficiency is independent of the geometry of the laser device such as the cavity length or the strip width. This parameter is extracted for the comparison of material properties of the laser devices. The carrier losses could be caused by lateral spreading of carriers (ηs), carrier escape from the active region (ηe), and by recombination losses within the active layers (ηr). All the three loss mechanisms contribute to the internal efficiency given by the equation below.
58 K. Tanaka, K. Wakao, T. Yamamoto, H. Nobuhara, and T. Fujii, “Dependence of differential quantum efficiency on the confinement structure in InGaAs/InGaAsP strained-layer multiple quantum-well lasers,” IEEE Photon Technology Letter., vol. 5, pp. 602–605, 1993
46 η i = η s ⋅η e ⋅η r (3.2)
The internal quantum efficiency is determined from the plot of inverse of differential efficiency versus cavity length. In this investigation the cavity length of the lasers being characterized are kept as constant. A different approach is followed to deduce the internal quantum efficiency of the laser devices. The relation between differential quantum efficiency and the internal quantum efficiency is given by the equation
1 ⎡ αi ⎤ ηd = ⋅ ⎢1+ L⎥ (3.3) ηi ⎣ ln(1/ R) ⎦
Where αi is the internal loss, L is the length of the cavity and R is the reflectivity of the facets. The value of ηd is less than the value of ηi. And the ratio of ηd/ ηi gives the number of photons emitted from the laser to the number of photons generated within the laser2.
The light that propagates through the laser diode cavity suffers from losses as is the case of light propagation in any optical waveguide. The internal loss is the parameter which corresponds to the loss of the optical wave. In the chapter to follow a detail description is given on the measurement of internal loss and quantum efficiency.
From the characteristic L-I curve show in the fig 3.1 and with the help of the equation given in 3.1 the differential efficiency is calculated for the four MQW laser devices and tabulated. The table 3.2 below shows the slope efficiency and the differential quantum efficiency of the devices being characterized.
47 Average Average differential InAlGaAs slope Differential Quantum Slope efficiency in (W/A) quantum MQW lasers efficiency in Efficiency (DQE) in (%) efficiency in (W/A) (%)
7 QW 0.171 0.168 0.176 0.169 0.171 35.9 35.3 37.0 35.6 35.95
5 QW 0.196 0.196 0.204 0.198 0.198 40.9 40.9 42.6 413 41.43
3.2 Slope efficiency and differential quantum efficiency of InAlGaAs MQW lasers
The tabulation indicates that the slope efficiency is better for five quantum well lasers than seven quantum well lasers due to more absorption in the quantum well region in the later. This is due to the less loss in the active region of the five quantum well region in comparison with the seven quantum well lasers. Even though the threshold current density is greater incase of the five quantum well lasers the slope efficiency is lesser than the seven quantum well lasers. The average differential quantum efficiency is greater in case of the five quantum well lasers. Thus the slope efficiency and differential quantum efficiency are dependent on the well number and they increase with the decrease in well number. The cavity length on the other hand is inversely proportional to the differential quantum efficiency. The external quantum efficiency as discussed above is dependent on the cavity length of the laser. The next sets of analysis are done to understand the effect of temperature on these parameters. Furthermore to understand how the device performance of the lasers are affected on subjection to temperature.
48 3.4 Threshold current dependence on temperature for InAlGaAs MQW Laser
The ability of the laser to perform well at elevated temperatures is of great interest. It is a great concern especially in the case of lasers used in the application of optical communication controllers, as the amount of heated generated in theses environment causes the device temperature to rise significantly. Thus it is of utmost importance for the semiconductor crystal to be robust enough so as not to degrade due to device operation at high temperatures. In order to understand the stability of laser performance when subjected to temperature it is necessary to determine the characteristic temperature of laser. The characteristic temperature (T0) is a measure of the temperature sensitivity of the laser device1 and is given below by the equation 3.4.
⎡ T ⎤ Ith = I0 exp⎢ ⎥ (3.4) ⎣T0 ⎦
The equation indicates that the Ith increases exponentially with temperature. I0 is the proportionality constant. Higher values of T0 imply that the threshold current density and the differential quantum efficiency of the device increase less rapidly with increasing temperatures. This translates into the laser being more thermally stable. This reduction in the temperature sensitivity of the threshold current is the major advantage in a 2 dimensional quantum well laser in comparison with a 3 dimensional heterostructure laser59. It is because of the temperature dependence of the effective density of states60.
59 K. Hess, B. A. Vojak, N. Holonyak, Jr. R. Chin, and P. D. Dapkus, “Temperature dependence of threshold current for a quantum well heterostructure laser,” Solid State Electron., vol. 23, pp.
585-589, 1980
49 For commercial purposes the optical sources in communication network are required to be temperature independent to avoid the need for temperature controllers. The InAlGaAs-
InP 1.3µm quantum well lasers are more suitable for this purpose on comparison with the conventional InGaAsP-InP quantum well lasers. This is due to their large conduction band offset which helps in better electron confinement in the quantum wells61,62,63,64,65,66.
60Y. Arakawa and H. Sakaki, “Multidimensional quantum well laser and temperature dependence of its threshold current,” Applied Physic Letter, vol. 40, pp. 939-941, 1982
61 T. R. Chen, P. C. Chen, J. Unger, M. A. Newkirk, S. Oh, and N. BarChaim, “Low-threshold and high-temperature operation of InGaAlAs-InP lasers,” IEEE Photon Technology Letter., vol.
9, pp. 17–18, 1997
62 M. C. Wang, W. Lin, T. T. Shi, and Y. K. Tu, “Ultrahigh temperature and ultrahigh speed operation of 1.3µm strain-compensated AlGaInAs-InP uncooled laser diodes,” Electron Letter., vol. 31, no. 18, pp. 1584–1585, Aug. 1995
63 K. Takemasa, T. Munakata, M. Kobayashi, H. Wada, and T. Kamijoh, “High temperature operation of 1.3µm AlGaInAs strained multiple quantum well lasers,” Electron Letter., vol. 34, no. 12, pp. 1231–1233, June 1998
64 T. Ishikawa, T. Higashi, T. Uchida, T. Fujii, T. Yamamoto, H. Shoji, and M. Kobayashi,
“Evaluation of differential gain of 1.3 µm AlGaInAs/InP strained MQW lasers,” in Proc. 10th
Int. Conf. Indium Phosphide and Related Materials, Tsukuba, Japan, pp. 729–732, paper ThP-55,
1998
65 C. E. Zah, M. C. Wang, R. Bhat, T. P. Lee, S. L. Chuang, Z. Wang, D. Darby, D. Flanders, and
J. J. Hseih, “High-temperature modulation dynamics of 1.3µm AlxGayIn1_x_yAs–InP compressive strained multiple-quantum-well lasers,” in Proc. Int. Semiconductor Laser Conf., 1994, pp. 215–
216
50 Recently to improve the characteristic temperature of InAlGaAs multiple quantum well lasers wide band gap barriers are formed for the suppression of thermal leakage15,16,17,67
68,69. Zah et al.16 proposed improved temperature characteristic in InAlGaAs MQW lasers by the employing strain in the quantum well region. A better characteristic temperature could also be achieved by the suppression of thermal leakage carriers into barrier and separate confinement heterostructure (SCH) layers70.
In order to measure the characteristic temperature of a laser it is necessary to experimentally measure the characteristic L-I. curve of a laser at various temperatures.
The same experimental setup described in the section 3.1 is employed. The temperature is
66 C. E. Zah, R. Bhat, B. N. Pathak, F. Favire, W. Lin, M. C. Wang, N. C. Andreadakis, D. M.
Hwang, M. A. Koza, T. P. Lee, Z. Wang, D. Darby, D. Flanders, and J. J. Hsieh, “High- performance uncooled 1.3µm AlxGayIn1_x_yAs–InP strained-layer quantum-well lasers for subscriber loop applications,” IEEE J. Quantum Electron., vol. 30, pp. 511–523, Feb. 1994
67 Z.Wang, D. B. Darby, R. Panock, P. Whitney, and D. C. Flanders, “High speed, ultra low noise operation from 40ºc –100ºc tensile strained InGaAlAs MQW lasers emitting at 1300 nm,” in
Proc. Int. Semiconductor Laser Conf., 1994, pp. 23–24
68 C. E. Zah, R. Bhat, and T. P. Lee, “High temperature operation of AlGaInAs/InP lasers,” in
Proc. 17th Int. Conf. InP and Related Materials, 1995, pp. 14–17
69 M. C. Wang, W. Lin, C. Y. Chang, H. H. Liao, and Y. K. Tu, “Highly-reliable, high- performance 1.3µm m low-cost laser diodes for fiber-to-the-home applications,” in Proc. IEEE-
LEOS Annu. Meeting, 1996, pp. 415–416
70 T. Higashi, J. Sweeney, A. F. Phillips, A. R. Adams, E. P. O’Reilly, T. Uchida, and T. Fujii,
“Experimental analysis of temperature dependence in 1.3µm AlGaInAs–InP strained MQW lasers,” IEEE J. Selected Topics Quantum Electron., vol. 5, pp. 413–419, Mar. 1999
51 varied and monitored using the laser diode controlled. The copper plate heat sink stage assembled detects even a small increment in temperature applied to the stage by the laser diode controller. It is also made sure that there is a uniform distribution of temperature on the surface of the stage. The temperature is varied from (20ºc-70ºc) and the corresponding characteristic L-I curve for each laser device is plotted. From these experimentally measured L.I. curves the characteristic temperature of the device is determined by plotting the natural logarithm of threshold current Ith versus the temperature and then measuring the slope of the linear curve fit line. This is explained numerically by the following equations.
⎡ T ⎤ Ln(Ith ) − Ln(I0 ) = ⎢ ⎥ (3.5) ⎣T0 ⎦
1 ∆ln()ITh = ⋅ (∆T ) (3.6) T0
The L-I plots ate various temperature for two sets of each InAlGaAs MQW laser is shown below. The fig 3.2 (a) and (b) device 1 and device 2 of 7 QW InAlGaAs lasers for temperature varied between 20ºc-75ºc is shown.
52
Fig 3.2 (a) L-I plot for 7 QW InAlGaAs- Device 1 at various temperatures
Fig 3.2 (b) L-I plot for 7 QW InAlGaAs- Device 2 at various temperatures
The L-I plot for device 1 and device 2 of 5Qw InAlGaAs lasers for temperature between
20ºc-60ºc are shown below in fig 3.3 (a) and (b).
53
Fig 3.3 (a) L-I plot for 5 QW InAlGaAs- Device 1 at various temperatures
Fig 3.3 (b) L-I plot for 5 QW InAlGaAs- Device 2 at various temperatures
The next step is to determine the characteristic plot for these devices from the above characteristic L-I curve at various temperatures. The above plots clearly show the increase in threshold current with the increase in temperature. Thus to determine the
54 characteristic temperature the natural logarithm of threshold current versus temperature needed to be plotted first. The characteristic temperature values for each device are calculated from the plot and are shown below. Thus the characteristic temperatures are determined and compared for both seven and five InAlGaAs ridge waveguide quantum well lasers.
Fig 3.4 (a) Characteristic Temperature for device-1 7QW InAlGaAs
55
Fig 3.4 (b) Characteristic Temperature for device-2 7QW InAlGaAs
Fig 3.5 (a) Characteristic Temperature for device-1 5QW InAlGaAs
56
Fig 3.5 (b) Characteristic Temperature for device-2 5QW InAlGaAs
From curve fitting the characteristic temperature plot the value T0 is determined and it shows better characteristic temperature for seven quantum well lasers compared to five quantum well lasers. Higher To indicates reduced threshold current dependence on temperature for seven quantum well lasers. The tabular column 3.3 below gives the characteristic temperature for all the four devices for which the measurement are taken.
InAlGaAs MQW laser T0 (°c) Average T0 (°c)
5 QW 57.0 58.0 57.5 59.0 57.9
7 QW 70.9 71 72 70 70.1
Table 3.3 Characteristic temperature for InAlGaAs MQW lasers
The measured values present a comprehensible fact that the characteristic temperature increases with well number. From the previous chapter it is evident that increase in the number of wells increases the optical confinement factor. The threshold gain also
57 decreases with increase in the number of wells thus causing the characteristic temperature to increase with well number. In the later section we would also see the increase in the gain with well number and the dependence of gain with temperature and current. Since the cavity lengths of both these devices are constant its effects on temperature are not discussed. The previous study reveals that the laser structure also plays an important role in characteristic temperature dependence in InAlGaAs MQW lasers71,72,73,74,75. From the laser device structure tabulated in the previous chapter it could be inferred that that barrier width of the fabricated InAlGaAs quantum well laser was twice that of the width
71 K. Takemasa, T. Munakata, M. Kobayashi, H. Wada, and T. Kamijoh, “1.3µm AlGaInAs-
AlGaInAs strained multiquantum-well lasers with a p-AlInAs electron stopper layer,” IEEE
Photon. Technology Letter., vol. 10, pp. 495–497, 1998
72 R. F. Kazarinov and G. L. Belenky, “Novel design of AlGaInAs-InP lasers operating at
1.3µm,” IEEE J. Quantum Electron., vol. 31, pp. 423–426, 1995
73 KOCH, T.L., KOREN, U,, EISENSTEIN, G., YOUNG, M.G., ORON, M., GILES, C. R., and
MILLER, B.I.: ‘Tapered waveguide InGaAs/InGaAsP multiple-quantum-well lasers’, IEEE
Photonics Technology Letter.
74 P. W. A. McIlroy, A. Kurobe, and Y. Uematsu, “Analysis and application of theoretical gain curves to the design of multi-quantum-well lasers,” IEEE J. Quantum Electron., vol. QE-21, pp.
1958–1963, 1985
75 LEALMAN, L.F., SELTZER, C.P., RIVERS, L.J., HARLOW, M.J., and PERRIN, S .D, ‘Low threshold current 1.6µm InGaAsP-InP tapered active layer multiquantum well laser with improved coupling to cleaved single mode fiber’, Electron. Letter., 1994, 30, (12), pp. 973- 975
58 of the wells. In the study by Blood et al76 show the influence of the barriers on the temperature dependence of threshold current in AlGaAs quantum well lasers. They have also introducing electron stop layers to get better characteristic temperature in quantum well lasers. This high characteristic temperature value of InAlGaAs MQW lasers is one of the key parameters of its high use in the optical communication devices in recent years. The other deduction that could be derived from the L-I plot at various temperatures is the change of slope efficiency with temperature. Given below is figure 3.6 which shows the change in slope efficiency of five and seven InAlGaAs MQW lasers at various temperatures.
Fig 3.6 Slope efficiency dependence on temperature for InAlGaAs MQW lasers
76 M. Aoki, M. Komori, T. Tsuchiya, H. Sato, K. Nakahara, and K. Uomi, “InP-based reversed- mesa ridge-waveguide structure for high performance long-wavelength laser diodes,” IEEE
Journal of selected topics Quantum Electron., Vol. 3, pp. 672–683, April 1997
59 From the above plot two things are understood that (a) the slope efficiency decreases with increase in temperature for both seven and five quantum well lasers and (b) the change in slope efficiency with increase in temperature is less in the case of five quantum well laser on comparison with the seven quantum well. The rapid change in the slope efficiency at high temperature is due to the increase in the optical loss in the quantum well laser with higher number of wells. The rapid decrease in the slope efficiency at high temperature in
InAlGaAs quantum well with temperature is caused by the hole leakage due to the valence band offset in InAlGaAs is smaller11. The next chapter gives a detail description on the gain measurement and its dependence with the well number and waveguide structure.
60 3.5 Conclusion
The material and the geometrical parameters of the InAlGaAs MQW lasers are extracted and analyzed. The dependence of these parameters on the various geometry of the laser is understood. The dependence of these parameters on well number and the waveguide shape was researched. The laser is subjected to temperature and the dependence of threshold current and slope efficiency with temperature are calculated. The performance of the laser is tested as desired. The advantage of increasing the well number and its effects on laser performance is studied. The high temperature characteristics of InAlGaAs
MQW lasers are determined and the reasons analyzed. From the so far analyzed DC characteristics it is concluded that the threshold current value and T0 of seven quantum well InAlGaAs lasers are less than the five quantum well lasers due to the better optical confinement in the quantum well region of the former. On the other hand better slope efficiency and less dependence of slope efficiency with increase in temperature is exhibited by five quantum well lasers in comparison with the seven quantum well lasers due to more absorption loss in the quantum well region of the later. The tabular column
3.4 below gives you the design trade of faced in the DC characteristics of the InAlGaAs
MQW lasers in increasing the number of wells.
Threshold Well Characteristic Slope efficiency Differential quantum current density Number efficiency in (%) temperature (°c) (W/A) (A/cm2)
5 41.43 2281 57.9 0.198
7 38.95 1871 70.1 0.171
Table 3.4 Summary of the DC characteristic measurement of InAlGaAs lasers
61 Chapter 4
Effect of well number in Gain and Intrinsic modulation responses in
InAlGaAs MQW lasers
4.1 Introduction
In this chapter further investigation is done on the InAlGaAs MQW lasers to determine the effect of well number on gain and intrinsic modulation response. The determination of net gain, optical loss and material gain in the multi quantum well lasers are important parameters needed to further understand the performance of the lasers. The effects of well number and temperature on these parameters are measured and useful information is inferred from the obtained data. The intrinsic modulation responses of these lasers are measured to estimate the operating bandwidth of these lasers. The determination of these parameters gives the laser operating point needed in order to optimize the laser performance. The measurement are carried out only for the five and seven quantum well lasers as the measurement on tapered waveguide lasers gave low output power and thus making it difficult to characterize. Different experimental setup were designed and constructed to extract there parameters. The experimental setup and procedure are explained in detail in the corresponding subsections.
62 4.1.1 Experimental details – Gain measurement for InAlGaAs MQW lasers
Semiconductor lasers are considered to be key components in fiber optic system as their cost represents a substantial part of the total cost of an optical transmitter. The temperature dependence of the lasers makes it necessary to use expensive and unreliable thermo-electric coolers with high power consumption. This is one reason for the interest, both experimentally and theoretically, in the temperature dependence of these devices.
The optical gain in a semiconductor lasers is an important parameter in determining many of its operating characteristics including threshold current, spectral width and modulation behavior. As a result, there has been much effort to calculate and measure the gain in the
InP based material system77. The experimental setup used to measure the net gain is given below in figure 4.1. Almost the same experimental setup is used for the measurement of L-I characteristics plot of InAlGaAs MQW lasers in chapter 3.
Computer - Optical Spectrum Laser Diode Lab View Analyzer Controller
GPIB Fiber Splitter Laser Stage
Power Detector
4.1 Schematic representation of the experimental setup for gain measurement
77 A. R. Adams, M. Asada, Y. Suematsu, and S. Arai, “The temperature dependence of the efficiency and threshold current of InGaAsP lasers related to intervalence band absorption,” Jap.
Journal Applied Physics, vol. 19, pp. L621-L624, 1980
63 In this setup the optical power output from the laser is coupled using a 90/10 fiber splitter whose ends are connected to the optical spectrum analyzer and power detector. The 90% output of the fiber end is connected to the power detector and the 10% output end is connected to the spectrum analyzer. The connecting ends of the fiber splitter are switched as per the need of the experiment. In the above experimental setup the coupling loss incurred due to the usage single mode fiber instead of lensed fiber is measured to be around 7db. Care is taken to optimize the coupling of light into the fiber split so as to minimize the coupling loss.
4.2 Gain measurement of InAlGaAs MQW lasers
One of the most important parameters relating the physical properties of the semiconductor structure to the output characteristics of the laser diode is the optical gain.
Optical gain and its dependence on the operating conditions determine not only the basic output characteristics, such as threshold current, but also the temperature dependence of the output characteristics, as well high-speed performance of the laser. A common technique used to measure modal gain was introduced by Hakki and Paoli and involves estimating the gain from the sub threshold emission spectrum of the laser78. .The carrier distribution in the active layer strongly depends on drive current below threshold, but not above threshold. Physical processes below threshold are critical in determining the operating point of the laser. Therefore studying the optical emission below threshold is often more informative in the process of understanding the device performance. At currents higher than threshold it is predominantly stimulated emission, below threshold it
78 B. W. Hakki and T. L. Paoli, “Gain spectra in GaAs double heterostructure injection lasers,”
Journal Applied Phys., vol. 46, pp.1299-1306, Mar. 1975
64 is amplified spontaneous emission (ASE)2. The gain spectra are obtained by measuring the depth of modulation caused by the Fabry-Perot resonance in the emission spectrum.
The data acquisition system is programmed to reject all the data input except the maxima and minima of the Fabry-Perot resonance in this region. The maxima and minima are stored in a matrix array until the whole spectrum is scanned. The system then computes
1 the gain at any wavelength by first averaging each two consecutive peaks ⋅[]P + P 2 i i+1
and dividing by the intermediate valley Vi thereby obtaining the depth of modulation Ri in the form of equation 4.1
Pi + Pi+1 Ri = (4.1) 2Vi
A sample plot of (a) power versus wavelength is given in below in the figure 4.2
P i+1
Pi
) mW (
Power
Vi
Wavelength (nm) 4.2 Sample plot of modulation response at ASE [2]
The net gain measured in this setup is given by the equation 4.2 below
G = Γg −α i −α m (4.2) where Г is the active layer optical confinement factor, g is the material gain, αi, is the optical loss and αm, is the mirror loss. Mirror loss is given by equation 4.3
65 2 α m = −[ln(R )]/ L (4.3) where L is the cavity length and R is the mirror reflectivity. The net gain measured using the Hakki Paoli method is given by equation 4.4
1/ 2 1/ 2 G = {ln[(Ri −1)/(Ri +1)]}/ L (4.4)
There are several other methods that are used to determine the material gain in semiconductor lasers. The difficulties in using this Hakki Paoli technique is that the devices must be selected with vary good transverse mode behavior and the substrate emission can obscure the measurement and cause incorrect values of gain to be inferred especially for energies near the bandgap. Thus the gain measures using the experimental setup described in figure 4.1 gives the net gain. The mirror loss is calculated with the given length L= 300µm and R=0.28 as αm= 42/cm. Thus the next step is to measure the cavity loss. Before measuring the optical loss and then determining the material gain g.
The net gain G is measured experimentally. The optical loss is a sum of the cavity loss and the mirror loss. The mirror loss could be calculated from the equation 4.3 knowing the value of R and cavity length L. The optical loss is deuced by extrapolating net gain at threshold current (Ith) = 0. From the obtained optical loss the cavity loss is measured and shown in the section 4.4. The effect of temperature on the net gain for both five and seven InAlGaAs quantum well lasers is also measured in the next section.
4.3 Gain measurement at various temperatures for InAlGaAs MQW lasers
A high volume market exists for 1.3 µm semiconductor lasers operating over a wide temperature range with reduced temperature sensitivity. This market drives the search for understanding of T0, a parameter describing the temperature dependence of threshold.
66 Temperature dependent optical loss79,80,81 as well as heterobarrier leakage82 has been cited as the main cause of observed temperature sensitivity of threshold. Many prior investigations have indicated nonradiative recombination due to various Auger processes as the primary cause of strong threshold temperature dependence in long wavelength lasers83,84,85,86,87,88,89,90,91,92. Most of the studies of Auger recombination have focused on
79 M. Asada, A. R. Adams, K. E. Stubkjaer, Y. Suematsu, Y. Itaya, and S. Arai, “The temperature dependence of the threshold current of GaInAsP DH lasers,’’ IEEE J. Quantum Electron., vol.
QE-17,pp. 611-619, 1981
80 M. Asada and Y. Suematsu, “The effects of loss and nonradiative recombination of the temperature dependence of threshold current in 1.5-1.6µm GaInAsP/InP lasers,” IEEE J.
Quantum Electron., vol. QE-19, pp.917-923, 1983
81 M. Yano, H. Nishi, and M. Takusagawa., “Temperature characteristics of threshold current in
InGaAsP double heterostructure lasers,” Journal Applied Physics., vol. SI, pp. 40224028, 1980
82 Y. Horikoshi and Y. Furukawa, “Temperature sensitive threshold current of InGaAsP-InP double heterostructure lasers,” Jap. J. Appl. Phys.,vol. 18, pp. 809-815, 1979
83 G. B. H. Thompson and G. D. Henshall, “Nonradiative carrier loss and temperature sensitivity of threshold in 1.27 µm (GaIn)(AsP)/InP D.H.Lasers,’’ Elecrron. Lett., vol. 16, pp. 42-44, 1980
84 A. Sugimura, “Band-to-band Auger recombination in InGaAsP lasers,” Appl. Phys. Lett., vol.
39, pp. 21-23, 1981
85 T. Uji, K. Iwamoto and R. Lang, “Nonradiative recombination in InGaAsP light sources causing light emitting diode output saturation and strong laser-threshold-current temperature sensitivity,” Appl. Phys
86J. I. Pankove, “Temperature dependence of emission efficiency and lasing threshold in laser diodes.” IEEE J. Quantum Electron., vol. 4, pp. 119-122, 1968
67 either the temperature dependence of the Auger processes9,12,14,16 or the role of Auger processes in consuming current 7,8,10,12,13,15. While there is no consensus on the temperature dependence, it is generally agreed that the non-linear dependence of Auger recombination on carrier density tends to boost long wavelength laser thresholds with temperature. Recently however, in contrast to previous findings 93,94,95, gain has been
87 N. K. Dutta and R. J. Nelson, “Temperature dependence of the lasing characteristics of the 1.3 pm InGaAsP-InP and Ga-As-Al0.36Ga0.64As DH lasers,” IEEE J. Quantum Electron., vol. 18, pp. 871-878, 1982
88 N. K. Dutta and R. J. Nelson, “The case for Auger recombination in InGaAsP”Journal Applied
Physics. vol. 53, pp. 74-92, 1982
89A. Mozer, K. M. Romanek, 0. Hildebrand, W. Schmid, and M. H. Pilkuhn, “Losses in
GaInAsP/InP and GaAlSbAs/GaSb lasers-The influence of the split-off valence band.” IEEE J.
Quantum Electron., vol. QE-19, pp. 913-916, 1983
90 A. P. Mozer, S. Hausser, and M. H. Pilkuhn, “Quantitative evaluation of gain and losses in quaternary lasers,” IEEE J. Quantum Electron., vol. QE-21, 719-725, 1985
91 E. P. O’Reilly and M. Silver, “Temperature sensitivity and high temperature operation of long wavelength semiconductor lasers,” Appl. Phys
92 M. C. Wang, K. Kash, C. E. Zah, R. Bhat, and S. L. Chuang, “Measurement of nonradiative and radiative recombination rates in strained-layer quantum-well systems,” Appl. Phys. Lett., vol.
62, pp. 166-168, 1993
93 Y. Zou, J. S. Osinski, P. Grodzinski, P. D. Dapkus, W. C. Rideout, W. F. Sharlin, J. Schlafer, and F. D. Crawford, “Experimental study of Auger recombination, gain, and temperature sensitivity of 1.5 µm compressively strained semiconductor lasers,” IEEE J. Quantum
Electron.,vol. 29, pp. 1565-1575, 1993
68 identified as a dominant cause of low T0 1.3µm strained multiquantum well (MQW) and bulk lasers96. The net gain (G) is measured for both five and seven quantum well lasers at various currents below threshold current. The plot at various currents below threshold at room temperature for both seven and five quantum well InAlGaAs lasers is given below in the figure 4.3 (a) &(b)
4.3 Net Gain at various current below threshold at room temperature (a) 7Qw (b)
5Qw
94N. K. Dutta and N. A. Olsson, “Temperature dependence of threshold current of injection lasers for short pulse excitation,” Appl. Phys. Lett
95 P.-L. Liu, J. P. Heritage, and 0. E. Martinez, “Temperature dependence of the threshold current of an InGaAsP laser under 130-ps electrical pulse pumping,” Appl. Phys. Lett., vol. 44, pp. 370-
372, 1983
96 D. A. Ackerman, P. A. Morton, G. E. Shtengel, M. S. Hybertsen, R. F. Kazarinov, T. Tanbun-
Ek, and R. A. Logan, “Analysis of T0 in1.3µm multi quantumwell and bulk active lasers,” Appl.
Phys. Lett.,accepted for publication
69 From the above figures some of the obvious conclusions that could be derived for both the seven and five InAlGaAs quantum well lasers are that the net gain increases with increase in well number and net gain increases with current. The better gain exhibited by the seven quantum well lasers is due to higher optical confinement in comparison with the five quantum well lasers. From the plots the net peak gain (G) for each current is calculated for both seven and five quantum well lasers. The following graphs will throw much better light to it. They also show how the how peak gain varies linearly increases with current. The plot of peak gain versus current at room temperature is shown below for InAlGaAs quantum well lasers figure 4.4.
4.4 Net Peak gain at various currents below threshold at room temperature
The change in the peak gain with current (dGpeak/dI) is greater in the case of seven quantum well lasers when compared to five quantum well lasers. Thus the effect of well number on net gain is shown. The better gain characteristics exhibited in the seven quantum well lasers is due to the better optical confinement and less leakage with increase in temperature. It could ask be noted from the above plots that the peak gain
70 decreases with temperature. This could be better understood by the determination of the cavity loss and the material gains of the seven and five quantum well lasers.
4.4 Measurement of Optical loss
To determine the material gain the cavity loss should be measured. The cavity loss αi, of semiconductor lasers is an important parameter that influences many aspects of device operation, such as external quantum efficiency, threshold carrier density, resonance frequency and differential gain. An early and widely used technique for measuring internal optical loss requires a set of lasers, varying in length but otherwise equivalent, to estimate the average value of at threshold. Several new techniques have recently been proposed for optical loss measurements. They are all based on the simple relationship connecting modal gain, material gain and total optical loss in the laser given by the equation 4.2 in the section 4.2. Thus to calculate the cavity loss the value of the net gain should be measured at the point where the modal gain turn to zero. Using the following assumption the cavity loss at varying current at room temperature is calculated below and plotted against current for both five and seven quantum well lasers in the figure 4.5 (a) and (b).
71
4.5 (a) Cavity loss below threshold at room temperature- 5QW InAlGaAs
4.5 (b) Cavity loss below threshold current at room temperature- 7QW InAlGaAs
The cavity loss decreases with the increase in current. It is linearly dependent with current at room temperature in the case of seven quantum well lasers. The cavity loss of seven and five quantum well lasers is determined to be 32/cm and 18/cm respectively.
The material loss is calculated to be 40/cm.
72 4.5 Material gain determination for InAlGaAs MQW lasers
From the value of the net gain measured above experimentally and using the equation 4.2 the material gain for both the seven and five quantum well lasers are calculated and plotted in the figure 4.7 shown below. The value of the optical loss is determined from the sum of the cavity loss and the mirror loss. Then the optical confinement factor for both the quantum well lasers are used form the Lumerical simulation results determined in the chapter 2. The optical confinement factor for five and seven InAlGaAs quantum well lasers determined in the chapter 2 are 10.902 % and 17.8646 % respectively. Using the equation 4.2-4.4 the material gain at 1mA below threshold and at room temperature for the laser is determined.
4.6 Material Gain of InAlGaAs MQW lasers
The material gain of the laser determined is higher for the seven than the five quantum well lasers. This is due the increase in the number of well aiding in better carrier confinement5,7,8,11. Even though the optical losses are more for seven quantum well lasers the higher net gain and optical confinement factor helps to boost up its material gain.
73 Given below is the tabular column for the comparison of material property of InAlGaAs
MQW lasers (table 4.1).
Differential Quantum Internal Quantum
QW Cavity Loss Efficiency (ηd ) Efficiency (ηi )
(%) (%)
5 -18 41.43 75
7 -32 35.95 56
Table 4.1 Material property of InAlGaAs MQW laser dependence on Well number
The internal quantum efficiency is calculated using the formula below.
⎡ α i ⎤ η i = η d ⋅ ⎢1 + L ⎥ ⎣ ln( 1 / R ) ⎦
4.6 Experimental setup for modulation response of InAlGaAs MQW lasers
In addition to their CW properties, there is great current interest in the dynamical properties of semiconductor lasers. For example, the modulation characteristics of these lasers provide a fundamental limit on the bandwidth of optical communication systems.
The direct modulation bandwidth of a semiconductor laser is limited by its resonance frequency. The experimental setup used to measure the resonance frequency of the
InAlGaAs MQW lasers is shown below in the diagram
74
Frequency Spectrum Analyzer
Attenuator Computer - Laser Diode Lab View Controller Photo Detector
GPIB Fiber Splitter Laser Stage
Power Detector
4.7 Schematic diagram of the experimental setup for frequency response of
InAlGaAs
The experimental setup shown above is almost identical to the one described in the figure
4.1 except few more components are added to measure the modulation response. The optical signal obtained in the previous setup should effectively convert into electrical signal for measurement. Thus instead of one end of the fiber splitter being connected to the optical spectrum analyzer it is connected to the photo detector and then to a frequency spectrum analyzer via an attenuator. The attenuator is used to attenuate the noises caused by the reflection of ac signal. The photo detector is to converts the output optical signal into equivalent electrical signals and sends it to the frequency spectrum analyzer for measurement. The lightwave converts were used as the photo detectors in this setup.
They are fast accurate DC-coupled optical to electrical converters packaged as small optical probes. They also ensure low very low signal distortion and hence improved output signal. The experiment is monitored and data is obtained using the LabView.
75 4.7 Determination of Frequency response for InAlGaAs MQW lasers
The use of ultrahigh speed semiconductor lasers is expected to enable gigabit optical digital communication systems and optical subcarrier transmission systems. A desirable feature of a laser is the constant amplitude it shows right after the laser turns on, the amplitude varies for a while and then gets stabilized to a constant. The frequency before the laser gets stabilized is known as the relaxation oscillation frequency which is measured using the experimental setup described in section 4.6. The reason to obtain the value of the relaxation oscillation frequency is that physically, it is easy to get resonance when added a frequency similar to that of the system; in resonance large responses could be obtained. The maximum modulation bandwidth given fmax, in sa emiconductor lasers is limited by the K factor, defined as the ratio of the laser’s damping factor (γ) to the square of the laser’s relaxation oscillation frequency. The K factor is expressed by the differential gain (dg /dN) and the nonlinear gain coefficient as where Гp is photon lifetime. The second term, related to nonlinear gain, plays a significant role in the modulation behavior of semiconductor lasers. Hence determination of these parameters and their effect on gain and quantum efficiency is indispensable. The first step in the process is to determine the frequency response versus current for both the seven and five quantum well lasers. The square of the frequency response is directly proportional to the damping factor. The effect of well on the high speed performance of the lasers is seen from the plot shown below in the figure 4.9.
76
4.8 Frequency response for InAlGaAs MQW lasers
In the above plot the curve fit is drawn and then extended to meet the x-axis. The line cuts the x-axis at the threshold current of the laser device. The D factor is calculated from the slope of the line of square of the frequency measured with respect to current. The D factor is calculated to be 1.681 GHz/mA1/2 and 1.336 GHz/mA1/2 for seven and five quantum well lasers. Thus it could be concluded that the D factor increases with number of wells. This proves better performance of these devices with the increase in the number of wells. The dependence of the frequency response on temperature is also an important measure as it gives the operating band width of the lasers. Thus the frequency response versus current at varying temperature is measured and plotted in the figure shown below
5.10 (a) and (b).
77
4.9 (a) Frequency responses at varying temperatures – 5Qw InAlGaAs Lasers
4.9 (b) Frequency responses at varying temperatures – 7Qw InAlGaAs Lasers
The seven quantum well lasers shows better temperature stability in comparison with the five quantum well lasers. The measurements are carried out above the threshold current.
The modulation bandwidth wasn’t measured directly as the samples did not have
78 microwave contacts. From the plot 5.10 (a) and (b) the D factor is calculated for various temperatures for InAlGaAs MQW lasers. From the calculated D factor values a plot of D facto variation with temperature is plotted. The plot shows a linear variation of damping factor with temperature. The damping factor decreases with increase in temperature.
4.10 D-factor versus temperature for InAlGaAs Lasers
79 4.8 Conclusion
The effect of well number on the gain, optical losses, leakage and frequency response of the InAlGaAs lasers were measured and studied. From the obtained measurements the factors that affect the performance of these lasers were understood. The drawback in the experimental setup for each measurement was analyzed and the losses incurred due to that were added in the calculations. The effect of well number on gain, optical confinement, material gain, cavity losses and frequency response were measured and analyzed. The effect of temperature on gain was less in the case of seven quantum well lasers on comparison with the five quantum well lasers. The determination of the material gain and its dependence on the well number was also calculated and plotted. Though the cavity loss in the five quantum well lasers is more the net gain and the higher value of optical gain helps it to achieve better material gain than the five quantum well lasers. The measurement of cavity losses and its effect on the net gain was also calculated and plotted. The modulation efficiency of seven quantum well lasers are better than the five quantum well lasers. The damping factor was calculated from the relaxation oscillation frequency and it effect on temperature were determined and plotted.
InAlGaAs Net Gain Cavity loss Material Gain D Factor MQW lasers (1/cm) (1/cm) (1/cm) (GHz/mA1/2)
5 -11.4 -18 -6.578 1.336
7 -6.24 -32 -4.435 1.681
Table 4.2 Parameters as function of well numbers for InAlGaAs MQW lasers
80 Chapter 5
Results and Future Work
5.1 Effect of well number and waveguide structure in the optical confinement factor
The 1.3µm InAlGaAs reverse mesa ridge waveguide multi quantum well lasers device structure and its design parameters were analyzed and studied in detail. From the given device structure the refractive index of each layer was determined and the effective refractive index of the active region was calculated. Using the Mode solution software tool the structural layout of both the reverse mesa ridge waveguide and tapered waveguide structure for five and seven quantum well lasers were drawn. The Lumerical-
Simulation tool was then used to determine the optical confinement factor of each device.
From the obtained measurement the effect of well number and waveguide structure on optical confinement factor was determined.
5.2 Effect of well number and waveguide structure on the performance of the
InAlGaAs MQW lasers
Experimentally the L-I characteristic plots were plotted for all the four devices. From the plot the material and geometry parameters of the devices were determined. The effect of well number and waveguide structure on both the material and the geometrical parameters of the InAlGaAs MQW lasers were determined and researched. The effect of temperature on these parameters were also measured and studied due to their extensive use of these devices in optical communication modules. The drawback in the experimental setup was also discussed and suggestions were made to improve the setup for obtaining accurate data’s in the future. The high characteristic temperature and
81 performance of seven and five quantum well lasers for both reverse mesa ridge was discussed.
5.3 Effect of well number on Gain and Intrinsic modulation response in InAlGaAs
MQW lasers
The performance of the 1.3µm InAlGaAs reverse mesa ridge waveguide multi quantum well lasers were studied by the measurement of gain and intrinsic modulation response.
These measurements provide critical experimental feedback in the process of laser diode optimization. Thus the net gain, optical loss, material gain and internal quantum efficiency were measured and its dependence on well number and temperature were also researched. The measurement of intrinsic modulation helped to understand the performance of these laser devices in high speed communication devices. The net gain was measured to increase linearly with current and decrease with increase in temperature.
The wavelength of the peak gain shifted towards right with the increase in the injection current. Once again the drawbacks of the experimental setup designed were also researched and ways to improve the setup was also suggested.
5.4 Scope for future work
The experimental arrangement could be made to put the devices on a sub mount for microwave measurements. To better understand the effect of well number on the performance of the InAlGaAs MQW lasers experiments could be carried out over a wide range of cavity lengths and higher well numbers. Also further experiments could be carried out for understanding the effect of other structural parameters on the performance of the InAlGaAs MQW lasers.
82