Distributed Feedback Quantum Cascade Laser ...

Distributed Feedback Quantum Cascade Laser Arrays for Chemical Sensing

A dissertation presented by Benjamin Guocian Lee to the School of Engineering and Applied Sciences in partial fulfillment of the requirements for the degree of Doctor of Philosophy in the subject of Applied Physics Harvard University Cambridge, Massachusetts March 2009

c 2009 - Benjamin Guocian Lee

All rights reserved.

Thesis advisor

Author

Federico Capasso

Benjamin Guocian Lee

Distributed Feedback Quantum Cascade Laser Arrays for Chemical Sensing

Abstract Quantum cascade lasers (QCLs) are unipolar semiconductor lasers based on intersubband transitions in heterostructures. The emission wavelengths of mid-infrared QCLs span from 3 to 24 µm and cover the “fingerprint” region of molecular absorption. This makes QCLs particularly interesting for spectroscopic applications. Single-mode emission is required for most spectroscopic applications. To achieve single-mode emission, QCLs can be made as distributed feedback (DFB) lasers or integrated with an external cavity (EC). EC-QCLs are widely tunable but are cumbersome and complex to build; they require high quality anti-reflection coatings, wellaligned external optical components including a grating for tuning, and piezoelectric controllers. DFB-QCLs are very compact and can be readily micro-fabricated, but a single DFB-QCL has limited tunability of ⇠ 10 cm 1 . In this thesis, I developed arrays of DFB-QCLs as widely-tunable, single-mode laser sources, and I demonstrated their applications to chemical sensing. I demonstrated a DFB-QCL array with 32 single-mode lasers on a single chip, emitting in a range over 85 cm

1

near 9 µm wavelength, operated pulsed at room

temperature. The DFB-QCL array can be continuously tuned, since the separation in

iii

Abstract

iv

nominal emission frequencies is small enough that we can use temperature tuning to span the frequency gaps between adjacent lasers in the array. To show the applications for chemical sensing, absorption spectroscopy was performed using the DFB-QCL array; the absorption spectra of several fluids were obtained, with results that were comparable to conventional Fourier transform infrared spectrometers. Achieving overlapped beams at extended distances can be important for a number of applications envisioned for DFB-QCL arrays, particularly remote sensing. Using the technique of spectral beam combining, the total angular divergence of the DFBQCL array was reduced to less than 2 milliradians, which is 40 times better than without beam combining. Using the beam-combined array, absorption spectroscopy was performed at a distance of 6 m from the laser chip. An ultra-broadband DFB-QCL array was developed to further increase the coverage and tuning range. The array emitted in a range over 220 cm wavelength, operated pulsed at room temperature.

1

near 9 µm

Contents Title Page . . . . Abstract . . . . . Table of Contents List of Figures . . Acknowledgments

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

i iii v vii xiii

. . . . . . . . spectroscopy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

1 1 4 7 9 10 12 13 14 16 18 18 20

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

23 24 27 30 31 35 39

3 Distributed feedback QCL arrays 3.1 Device fabrication . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2 Array results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

41 42 45

1 Introduction 1.1 General introduction . . . . . . . . . . . . 1.2 State-of-the-art: Fourier transform infrared 1.3 Mid-infrared laser sources . . . . . . . . . 1.4 Quantum cascade lasers . . . . . . . . . . 1.4.1 Invention of the QCL . . . . . . . . 1.4.2 Intrinsic advantages of QCLs . . . 1.4.3 Development of QCLs . . . . . . . 1.4.4 Bound-to-continuum design . . . . 1.4.5 State-of-the-art . . . . . . . . . . . 1.5 Frequency control of QCLs . . . . . . . . . 1.5.1 External cavity QCLs . . . . . . . 1.5.2 Distributed feedback QCLs . . . .

. . . . .

. . . . .

2 Distributed feedback QCL design 2.1 Introduction: distributed feedback . . . . . . . 2.2 DFB-QCL grating design . . . . . . . . . . . . 2.3 DFB coupling strength and end facet mirrors . 2.3.1 Intensity profile and power output . . . 2.3.2 Single-mode selection . . . . . . . . . . 2.4 Summary . . . . . . . . . . . . . . . . . . . .

v

. . . . .

. . . . . .

. . . . .

. . . . . .

. . . . .

. . . . . .

. . . . .

. . . . . .

. . . . .

. . . . . .

. . . . . .

Contents

vi . . . . .

45 48 50 52 54

4 DFB-QCL arrays for mid-infrared spectroscopy 4.1 DFB-QCL array source . . . . . . . . . . . . . . . . . . . . . . . . . . 4.2 Absorption spectroscopy . . . . . . . . . . . . . . . . . . . . . . . . . 4.3 Discussion and summary . . . . . . . . . . . . . . . . . . . . . . . . .

56 57 57 61

5 Beam combining of QCL arrays 5.1 Introduction . . . . . . . . . . . 5.2 DFB-QCL array . . . . . . . . . 5.3 Near and far-field beam profiles 5.4 Beam overlap . . . . . . . . . . 5.5 Remote sensing demonstration . 5.6 Summary . . . . . . . . . . . .

. . . . . .

65 66 68 70 73 75 77

3.3 3.4

3.2.1 Short array . . . . . . . . . . . . 3.2.2 Array with anti-reflection coating 3.2.3 Longer array . . . . . . . . . . . Threshold and slope efficiencies . . . . . Summary . . . . . . . . . . . . . . . . .

6 Ultra-broadband DFB-QCL 6.1 Heterogeneous cascade . . 6.2 DFB-QCL array . . . . . . 6.3 Device performance . . . . 6.4 Summary . . . . . . . . .

. . . . . .

array . . . . . . . . . . . . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . .

. . . . . .

. . . . .

. . . . . .

. . . . .

. . . . . .

. . . . .

. . . . . .

. . . . .

. . . . . .

. . . . .

. . . . . .

. . . . .

. . . . . .

. . . . .

. . . . . .

. . . . .

. . . . . .

. . . . .

. . . . . .

. . . . .

. . . . . .

. . . . .

. . . . . .

. . . . .

. . . . . .

. . . . .

. . . . . .

. . . . .

. . . . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

78 78 81 82 86

7 Microfluidic tuning of DFB-QCLs 7.1 Introduction . . . . . . . . . . . . . . 7.2 Device fabrication and encapsulation 7.3 Experimental results . . . . . . . . . 7.4 Summary and future directions . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

87 87 88 92 97

. . . .

8 Conclusions

100

Bibliography

104

A List of Publications

110

B List of Patents

112

List of Figures 1.1 1.2

1.3 1.4

1.5

1.6

2.1

Schematic diagram of a Fourier transform infrared spectrometer. Adapted from a diagram by Nicolet Instruments. . . . . . . . . . . . . . . . . Here we compare the infrared absorption spectra of two isomers of propanol. We note that they have substantatively di↵erent spectra in the fingerprint region of the mid-infrared. . . . . . . . . . . . . . . . Principal characteristics of (a) an interband transition and (b) an intersubband transition in a quantum well. . . . . . . . . . . . . . . . (a) Schematic of a QCL’s conduction band diagram. Each stage of the structure has an active region and a relaxation/injection region. Electrons can emit up to one photon per stage. (b) Outline of the design principle. There is a three-level system in the active region. The lifetime of the transition from level 3 to 2 has to be longer than the lifetime of level 2 to have population inversion. . . . . . . . . . . Schematic conduction band diagram of a bound-to-continuum design. The wells are tilted due to the applied electric field (35 kV/cm) and the moduli squared of the computed wavefunctions is plotted. Courtesy of R. Maulini. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (a) SEM image of the facet of a distributed feedback quantum cascade laser, with a top grating. This image is taken from Faist et al. [1997]. The red arrow drawn onto the image denotes the direction of light emission from the laser ridge. (b) SEM image of a cross-sectional slice of a buried DFB grating, showing the grating grooves. . . . . . . . . Plot of the spatial intensity distribution of the fundamental mode of the cavity, at di↵erent levels of the coupling strength L ⇠ ⇡ n/⇤. This figure taken from Kogelnik and Shank [1972]. . . . . . . . . . .

vii

5

6 10

11

15

21

26

List of Figures 2.2

2.3

2.4

(a) Schematic of the DFB-QCL waveguide structure. The grating is etched in an InGaAs layer (blue) in the top waveguide cladding, just above the active region (red). There is InP overgrown on top of the grating (yellow in diagram). (b) Mode simulation of the mode on the low-frequency side of the photonic gap of the DFB grating. The plot displays the magnitude of the electric field in the laser structure. The low-frequency mode has more of the electric field concentrated in the high-index part of the grating as expected. (c) Mode simulation of the mode on the high-frequency side of the photonic gap of the DFB grating. Again, we display the magnitude of the electric field. The high-frequency mode has more of the electric field concentrated in the low-index part of the grating as expected. (d) The dependence of the grating coupling strength per unit length  on the etch depth of the grating. The real part of  is shown as squares and the imaginary part as circles. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The intensity profile of the laser light along the length of the cavity. (a) The intensity profile for coupling constants L = 1, 4, 11 in the absence of mirrors. (b) The intensity profile for L = 4 and a very asymmetric mirror configuration — one mirror is coincident with the grating peaks and the other with the grating troughs. (c) The intensity profile for L = 11 and a very asymmetric mirror configuration — one mirror is coincident with the grating peaks and the other with the grating troughs. (d) A plot of the variability of the power emitted from the front facet of the laser as a fraction of the total power from both facets, for arbitrary positions of the end facet mirrors relative to the grating. The grating coupling strength here is L = 4.6 + 0.025i. . . . . . . . These three plots show the di↵erence in total loss between the highand low-frequency modes, for di↵erent coupling strengths L of the grating and di↵erent configurations of the end mirrors. The possible positions of the mirrors with respect to the grating are designated by a phase, which can range between 0 and ⇡ — corresponding to the case where the mirror is exactly in line with a grating groove and exactly anti-aligned respectively. When the surface of the plot of total loss is above zero, it means that the high-frequency mode has greater loss than the low-frequency mode. When the plot is below zero, then the high-frequency mode has the lower loss. (a) A plot for the case where L = 4.6 + 0.025i and both mirrors have 30% reflectivity. (b) A plot for the case where L = 4.6 + 0.025i, the first mirror has an antireflection coating with 1% residual reflectivity and the second mirror has 30% reflectivity. (c) A plot for the case where L = 11 + 0.058i and both mirrors have 30% reflectivity. . . . . . . . . . . . . . . . . . . . . . .

viii

29

32

38

List of Figures 3.1 3.2

3.3

3.4

3.5

Fabrication process flow. In step 7, the red arrow denotes the direction of light emitted from the facet of the laser ridge. . . . . . . . . . . . (a) DFB grating etched into the InGaAs layer, just above the active region. This SEM image is taken before regrowth of InP on top, which buries the grating. (b) A single ridge laser device, out of the 32 in the array, as viewed from its front facet. (c) SEM image of a cross-sectional slice of the laser ridge, taken along the direction of the ridge, showing the grating grooves. (d) The finished array, shown on top of a dime for size comparison. . . . . . . . . . . . . . . . . . . . . . . . . . . . (a) Spectra of 32 lasers from an array with L = 4.6 + 0.025i. All the lasers are single-mode, but 18 of the lasers are lasing on the highfrequency side of the DFB gratings photonic gap, while 14 are lasing on the low-frequency side. The inset shows the two modes on either side of the photonic bandgap (b) Spectrum of a representative DFB from the array, plotted on a log scale to show >20 dB side-mode suppression. (c) Plot of the voltage (left axis) and light output intensity (right axis) from several lasers in the same array, as a function of pump current. There is a small amount of variability in the threshold current, and a larger variation in the slope efficiencies, which leads to a significant variation in the peak output power. . . . . . . . . . . . . . . . . . . (a) Spectra of 32 lasers from the array with L = 4.6 + 0.025i, after an anti-reflection coating is applied on the front facet, with 1% residual reflectivity. All but one of the lasers are single-mode, with 25 of the lasers are lasing on the high-frequency side of the DFB gratings photonic gap, and 6 lasing on the low-frequency side. One of the lasers has both modes lasing. (b) Plot of the voltage (left axis) and light output intensity (right axis) from several lasers in the same array, as a function of pump current. . . . . . . . . . . . . . . . . . . . . . . . . (a) Spectra of 32 lasers from an array with L = 11 + 0.058i. All the lasers are single-mode, and they all lase on the high-frequency side of the DFB gratings photonic gap. (b) Plot of the voltage (left axis) and light output intensity (right axis) from several lasers in the same array, as a function of pump current. . . . . . . . . . . . . . . . . . . . . .

ix

43

44

46

49

51

List of Figures 3.6

4.1

4.2

4.3

4.4

5.1 5.2

(a) Histogram of the threshold current densities for the three di↵erent DFB-QCL arrays. The array with L = 4.6 + 0.025i is denoted in red with square markers, and after coating the front facet with an antireflection coating the new thresholds are denoted in blue with triangles, finally the array with L = 11 + 0.058i is denoted in green with circles. (b) Histogram of the slope efficiencies for the three di↵erent DFB-QCL arrays. The array with L = 4.6 + 0.025i is denoted in red with squares, and after coating the front facet with an anti-reflection coating the new slope efficiencies are denoted in blue with triangles, finally the array with L = 11 + 0.058i is denoted in green with circles. . . . . . . . .

x

53

(a) Schematic of the widely tunable quantum cascade laser source with a distributed feedback laser array driven by a custom microelectronic controller. The dotted line denotes the routing of the current pulse and the DC bias to a specific laser (second from top) to fire it at a specified wavelength. (b) The microelectronic controller, as built, with the DFB-QCL array chip located in the bottom centre of the image. 58 Absorption spectra of isopropanol (squares), acetone (circles) and methanol (triangles) obtained with the laser array and with a Bruker Vertex 80v Fourier-Transform Infrared spectrometer (continuous lines). (inset) Experimental setup for mid-infrared spectroscopy of liquids with the quantum cascade laser source. . . . . . . . . . . . . . . . . . . . . . 60 (a) Temperature tuning achieved using a DC bias current to heat an individual DFB-QCL in the array. 5 cm 1 of tuning range is achieved by using 300 mA of bias current. (b) Temperature tuning achieved by heating the entire substrate of the DFB-QCL array by varying its heatsink temperature using heaters and a thermoelectric cooler. 5 cm 1 of tuning range is achieved by using 70 K di↵erence in temperature. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61 Absorption spectrum of isopropanol taken using the DFB-QCL array source operated at di↵erent temperatures, in order to have a continuous measure of the spectrum (points). Data taken using a Bruker Vertex 80v FTIR shown for comparison (solid line). . . . . . . . . . . . . . 62 Beam combining can be a powerful technique, as evidenced by Star Wars! . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Schematic diagram of spectral beam combining, where an external cavity containing a grating both combines the beams and also provides the optical feedback that selects the emission wavelengths of the lasers in the array. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

66

67

List of Figures 5.3

5.4

5.5

5.6

6.1

6.2

6.3

6.4 6.5

Schematic diagram of spectral beam combining with an array of distributed feedback quantum cascade lasers (DFB-QCLs). The emission wavelengths of the lasers are selected by the individual DFBs on each laser ridge in the array, and beam combining is accomplished by a suitably placed grating that compensates for the angular dispersion of the beams in the array. . . . . . . . . . . . . . . . . . . . . . . . . . . . (a) Image of the beam of a representative laser, just after it has been reflected from the grating. The white bar is 1 cm. (b) Image of the farfield spot of a representative laser. The white bar is 1 milliradian. (c) Linescan in the horizontal (solid) and vertical (dotted line) directions of the far-field image of a representative laser. . . . . . . . . . . . . . (a) Image of several lasers showing the extent of the residual angular dispersion in the beams. From the left, we have laser elements #18, 24, 28, and 31 in the array. Lasers 18 and 31 have the largest relative angular spread in the entire array. The white bar is 1 milliradian. (b) A plot of the angular deviation of the laser beams, as a function of the laser frequency. Squares represent the angular spread of laser beam positions from the entire array, as measured relative to the position of laser 31 (rightmost point in the plot). The line is a calculation of the angular spread using the grating equation (Eq. 5.1) with the wavelengths of the DFB-QCL array and a grating angle of 54.65 degrees. Absorption spectrum of isopropanol measured using the spectrally beamcombined DFB-QCL array at a distance of 6 m (squares). Fourier transform infrared spectrometer measurement of the same sample using a Bruker Vertex 80v FTIR instrument (solid line). . . . . . . . . Electroluminescence of a mesa structure fabricated from the QCL wafer, showing the gain spectrum of the material. The mesa was pumped pulsed at 80kHz at room temperature, with a 5% duty cycle. Voltages of 9V, 12.8V, and 14.8V were applied (ascending order of traces). . . SEM micrograph showing a cross-section of the device, which has been cut along the laser ridge. The grating corrugation can be seen as the rectangular wave near the top of the image, and the two active regions are below, with a thin InGaAs spacer between them. . . . . . . . . . Spectra of 24 single-mode distributed feedback lasers in the array. Laser frequencies are spaced ⇠ 9.5 cm 1 apart and span a range of ⇠ 220 cm 1 . (inset) Spectrum of a representative laser in the array on a log scale, showing side-mode suppression > 20 dB. . . . . . . . . . Plot of the threshold current required for lasing, as a function of the laser frequency. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Plot of the voltage (left axis) and light output (right axis) characteristics of several representative lasers in the array as functions of current.

xi

68

72

74

76

80

81

84 85 86

List of Figures 7.1

7.2

7.3 7.4

7.5

7.6

(a) Schematic cross section of the processed DFB-laser. The current is injected laterally from two Ti/Au contacts. The Bragg grating is etched in the top layer composing the waveguide and is exposed to air/liquid. Diagrams showing (b) the laser as bonded and mounted on a Cu-heatsink, (c) the di↵erent parts entering into the fabrication of the liquid chamber and (d) a device after encapsulation. . . . . . . . An image of an encapsulated laser device. The laser facet is at the front/bottom. The laser ridge sits inside an empty micro-chamber where the fluid can be introduced, via one of the two attached tubes. Optical spectra obtained with an encapsulated device without liquid at room temperature and with di↵erent current levels. . . . . . . . . (a) Optical spectra obtained at a fixed current (4.1 A) at room temperature with di↵erent liquids in the fluid chamber. The refractive index of the fluids varied from 1.3 to 1.735. (b) Optical spectra shown on a log scale measured without fluid, with the liquid with n = 1.53 and methanol. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Comparison between the peak position obtained experimentally from the data shown in Fig. 7.4 (a) and the results of FDTD simulations. The result obtained with methanol in the fluid chamber is also displayed. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (a) Voltage and output power vs. current curves obtained at room temperature with di↵erent liquids in the fluid chamber. Note that these data were obtained with an encapsulated DFB QCL di↵erent from the one used to produce the data shown in Fig. 7.3, 7.4 and 7.5. (b) Mode hopping observed at a constant current by simply immersing the Bragg grating in fluid (n = 1.735). The wavelength shift is reproducible when the procedure is repeated after evaporation of the liquid. . . . . . . .

xii

90

92 93

95

96

98

Acknowledgments I want to thank the many people who’ve helped me — family, friends, colleagues and mentors — for all your encouragement, advice and support. I’ll start by thanking two of my earliest academic mentors — my high-school physics teacher Ping Lai, and my classics and Latin teacher Eugene DiSante — one started me on my journey of research and science, and the other broadened my horizons so that my curiosity might wander more freely. My undergraduate research advisor, Axel Scherer, provided an excellent environment where I got to try out random new ideas; moreover, as a teaching assistant for his lab course, I benefitted from the opportunity to excite and interact with students. I began my graduate studies with Bob Westervelt, who I thank for his kindness, his flashes of keen physical insight, and the opportunity I was given to mentor several wonderful undergrads — Susannah Dickerson, Li Wang, Will Shanks and Arthur Reynolds. I remain tremendously thankful to Bob and to Federico Capasso for allowing me to switch research groups. It has been an honour to have Federico as my advisor; I greatly appreciate his style of asking questions, his breath of scientific interest, his keen physical intuition, and his advising — I would not have completed my graduate studies without his interest, guidance, and sometimes criticism that helped me in becoming a better scientist and a more serious researcher. I would also like to acknowledge the other members of my thesis committee — David Weitz, Peter Pershan and Ken Crozier, and to thank the School of Engineering and Applied Sciences for an atmosphere of interdisciplinary research and excellent facilities. It has been a great pleasure to work with the members of the Capasso group. I learned a lot from working with the postdocs Marko Lonˇcar, Misha Belkin, Laurent

xiii

Acknowledgments

xiv

Diehl and Christian Pfl¨ ugl; thank you all for your e↵orts and patience. My research was also helped by two very talented undergraduates — Peter Behroozi and Ross Audet, and by new graduate student Haifei Zhang, who I’m sure will take the project in exciting new directions. Thanks additionally to Chris Mullaney for keeping the group running smoothly. Outside of the group, I have been fortunate to also interact with a number of excellent scientists in collaborations, including Anish Goyal, Jan Kansky, Antonio Sanchez and George Turner at MIT Lincoln Labs, Tom Tague of Bruker Optics, and Jerome Faist of ETH Zurich. I have been blessed with many friends during my time at Harvard and in Boston. I thank you all for your love and caring, and I hope I will be able to keep a connection with you into the future. In particular, thank you Derek, Annie, Crystal, Kaisey, Zsuzsa, Dev, Lela, Azamat, Tom, Lukas, Jillian, Giro, Christina, Wesley, Heidi, Raymond, Sam, Paul and Jo. My warm thanks also go to my friends and colleagues on the Graduate Student Council, the Canadian Club (particularly Karen, Mike, Jonathan and Rob), the Harvard Energy Journal Club (thanks especially to Kurt for organizing), the Dudley dragonboat team (Sarah, Audrey, Grace, Mark, Sheila, Brett and many others), Dudley House (Susan and Chad), the resident advisor and housing team (all my co-RAs, Ellen, Sheila, Patty, Jill, Bob and Garth), and last but not least the Jazz 5:30 church community (Dan, Nancy, Lolita, Claudette, David, Sister Jayne, Carolyn and others). I am more than grateful to my parents William and Corinna — for all their love and support, their encouragement in difficult times, and their firmness and criticism when I have needed it. My accomplishments would not have been possible without

Acknowledgments

xv

you. Finally I thank God, the Creator of our universe, who loves us, and who as the living Word, is still speaking.

Chapter 1 Introduction 1.1

General introduction

The rainbow in nature is a testament to the spectral nature of light, although this remained unrecognized until fairly recently in human history. Spectroscopy as a science began with Issac Newton, who first demonstrated in 1666 that the sun’s white light can be dispersed into a continuous spectrum of colours using a prism. Subsequently, infrared radiation was discovered by William Herschel in the early 1800s, when he placed a thermometer just beyond the edge of the visible spectrum of sunlight that was dispersed by a prism. This was a serendipitous discovery, since Herschel wanted to measure the temperature of the visible radiation and only placed the thermometer outside the path of the visible rays as a control, but was surprised to discover a higher temperature there. Spectroscopy with electromagnetic radiation has evolved into an important tool for both basic science and applications. Techniques include absorption spectroscopy, flu-

1

Chapter 1: Introduction

2

orescence spectroscopy, x-ray spectroscopy and crystallography, atomic emission spectroscopy, spark emission spectroscopy, and Raman spectroscopy. For spectroscopy, the infrared region is typically divided into the near-infrared (0.7 to 3 µm wavelength), the mid-infrared (3 to 25 µm) and the far-infrared (25 µm to 1 mm). The mid-infrared region is particularly interesting for chemical sensing applications. Many gas- and liquid-phase molecules have fundamental rotational and vibrational energies corresponding to mid-infrared wavelengths, giving rise to characteristic absorption features that can be used to identify them. These absorption features are very specific to the structure of each di↵erent molecule, allowing highly selective detection. Hence the mid-infrared spectral region is frequently referred to as the “molecular fingerprint region” of electromagnetic radiation. Applications of chemical sensing in the molecular fingerprint region include: medical diagnostics, such as breath analysis, pollution monitoring, environmental sensing of the greenhouse gases responsible for global warming, and remote detection of toxic chemicals and explosives.[Kosterev and Tittel, 2002] In this thesis, I will focus on the development of quantum cascade laser (QCL) devices for chemical sensing. The invention of the QCL in the research group of Federico Capasso at Bell Labs in 1994 [Faist et al., 1994] has opened up new prospects for chemical sensing in the mid-infrared. QCLs are unipolar semiconductor lasers that utilize resonant tunneling and optical transitions between electronic states in the conduction band of a multi-quantum well heterostructure. The emitted photon energy of QCLs can be chosen by design. Specifically, the photon energy is determined by the thicknesses of the wells and barriers in the heterostructure, and can be designed


3

by “band-gap engineering.” Since their invention, QCLs has been been shown to operate reliably throughout the mid-infrared from 3-25 µm wavelength[Capasso et al., 2002]. Moreover, QCLs have been demonstrated with watt-level high output power, operating continuouswave at room temperature[Lyakh et al., 2008; Bai et al., 2008]. Having a laser source opens up a lot of applications where a bright and single-frequency source are desirable. These include sensitive detection at the parts-per-billion level, remote sensing, and spectroscopy of dense media with strong absorption. [Kosterev and Tittel, 2002; Nelson et al., 2002; Pushkarsky et al., 2006] Controlling the emission frequency of a laser source is critical for spectroscopic applications. For chemical sensing with absorption spectroscopy, one clearly must be able to produce radiation at the frequency of the absorption feature. Moreover, it is important to achieve sufficient spectral resolution to resolve di↵erent absorption features the spectral resolution should be comparable or better than the linewidth of the targeted absorption features. For laser sources, there are several ways to control the emission frequency, in particular by using distributed-feedback in the laser cavity, or by incorporating the laser in an external cavity that provides frequency-selective optical feedback. In this thesis, I will detail the development of tunable, single-frequency mid-infrared sources, based on single and arrayed distributed-feedback quantum cascade lasers.


1.2

4

State-of-the-art: Fourier transform infrared spectroscopy

Currently, Fourier Transform Infrared (FTIR) spectrometers are the preferred method used for infrared spectroscopy. Most infrared spectrometers have a broadband infrared source, typically in the form of a thermal source of blackbody radiation. A common source is called a “globar” and consists of a silicon carbide rod of 5 to 10 mm width and 20 to 50 mm length which is electrically heated up to 1000 to 1650 C. For absorption spectroscopy, a sample (a cell containing the analyte of interest) is placed in the path of the infrared radiation. The original infrared spectrometers used dispersive elements such as prisms or gratings to separate the individual frequencies of energy in the transmitted infrared beam. A detector is used to measure the amount of energy at each frequency, in order to generate the spectrum. We show a diagram of a commercial FTIR from Nicolet in Fig. 1.1. FTIR spectrometers employ a Michelson interferometer with one movable mirror. Infrared radiation from the source is directed towards a beamsplitter. One beam reflects o↵ a flat mirror which is fixed in place. The other beam reflects o↵ a second flat mirror which can be moved a short distance, typically a few millimetres. The two beams reflect o↵ of their respective mirrors and are recombined when they meet back at the beamsplitter. The signal exiting the interferometer is thus the interference signal between the beams for a specific path length di↵erence. By moving the mirror and changing the path length di↵erence, the temporal coherence of the source can be mapped as an interferogram.


5

Movable Mirror

Figure 1.1: Schematic diagram of a Fourier transform infrared spectrometer. Adapted from a diagram by Nicolet Instruments.


6

For a signal with a long temporal coherence, such as a laser signal, the interferogram will exhibit interference fringes for all measured path length di↵erences between the beams. By contrast, for an incoherent signal, such as that from a globar infrared source, the interferogram will only have a few fringes corresponding to small path length di↵erences. We can recover a frequency spectrum of the source by performing a Fourier transform of the time-domain data in the interferogram. To perform absorption spectroscopy, a sample can be placed in the path of the beam between the interferometer and the detector. The spectrum obtained with the sample can be compared to a measurement of the background spectrum without the sample, to obtain the desired absorption data, as in Fig. 1.2. Representa)ve IR spectra of liquids (ﬁngerprint region in yellow)  Propan‐2‐ol, CH3‐CH‐CH3 

Propan‐1‐ol, CH3‐CH2‐CH2‐OH 

Transmission, % 

Transmission, % 

OH 

4000    3000    2000               1000   1/cm 

4000    3000    2000              1000   1/cm 

Note that even isomers of the same chemical have  large spectral diﬀerences 

Figure 1.2: Here we compare the infrared absorption spectra of two isomers of propanol. We note that they have substantatively di↵erent spectra in the fingerprint region of the mid-infrared.


1.3

7

Mid-infrared laser sources

While FTIR spectroscopy is an excellent tool for chemical sensing in the infrared, there are important advantages to laser spectroscopy. Chief among these are the brightness of laser sources and their narrow linewidth. Laser sources are orders of magnitude brighter than the thermal sources used in FTIR, allowing for higher signalto-noise in detection. This is important for applications involving sensitive partsper-billion level of detection (such as for trace gases), for remote sensing, and for spectroscopy of highly absorptive, optically dense materials (such as fluids or analytes in water). Laser linewidth from quantum cascade lasers can be smaller than 0.001 nm [Williams et al., 1999], which is considerably better than what is available from a typical “bench-top” commercial FTIR (⇠1 nm). Thus, a laser source can achieve significantly higher spectral resolution over its tuning range than an FTIR. Several classes of mid-infrared lasers exist. The CO2 laser was one of the earliest lasers to be developed [Patel, 1964] and is still one of the most useful. The principal emission wavelengths of the CO2 laser are near 9.6 and 10.6 µm, and they can provide high power of up to several hundred milliwatts running continuous-wave (CW). CO2 lasers have found wide use in industrial applications like cutting and welding metals, plus they have been used in surgical procedures. In spectroscopy, CO2 lasers can provide relatively compact, high power sources. However, they are limited by the fact that they can only be tuned on the vibrational-rotational transition energies of the CO2 molecule. These are at discrete wavelengths between ⇠ 9.2 and 11.5 µm separated by ⇠1 cm

1

gaps [Webber et al., 2005].

Lead-salt diode lasers were also pioneered in the 1960s, and can operate in the


8

mid-infrared from 3-25 µm wavelength. However, lead-salt diode lasers typically require cryogenic cooling for operation, particularly for CW operation [Feit et al., 1996]. Lead-salt lasers have been useful in a number of research settings, but haven’t become widely adopted commercially. Coherent mid-infrared light can also be generated using non-linear crystals with a large

(2)

susceptibility, by di↵erence frequency generation (DFG) and optical para-

metric oscillation (OPO). DFG mixes two single-frequency sources, called the pump and signal, in the non-linear material to generate tunable narrowband radiation at a third frequency, equal to the di↵erence between the two input frequencies. DFG can achieve large tuning ranges, as the tuning range of the pump is transferred to the generated radiation at the much lower di↵erence frequency, resulting in a much larger relative tuning. For example in Sanders et al. [1996], tuning a signal at 980 nm by 32 nm (3% relative tuning) corresponds to tuning the DFG output from 3.6 to 4.3 µm, which is a 17% change. However, DFG does have a limitation due to the extremely low conversion efficiency of the non-linear process, resulting in maximum output power levels of hundreds of microwatts to a few milliwatts. OPO consists essentially of an optical resonator and a nonlinear optical crystal. The optical resonator serves to resonate at least one of signal and idler waves. In the non-linear optical crystal, the pump, signal and idler waves overlap. The interaction between these three waves leads to amplitude gain for signal and idler waves (parametric amplification) and a corresponding deamplification of the pump wave. Power conversion efficiencies of OPO can be high, in the range of 10% and can deliver several watts of CW output power through a wide tuning range [van Herpen et al., 2004].


9

OPOs have found spectroscopic applications in photoacoustic and cavity ring-down spectroscopy in the mid-infrared [Ngai et al., 2006]. However, both OPO and DFG require fairly complex optical components and pump lasers (i.e. Nd-YAG) that can be expensive. Quantum cascade lasers, first invented in 1994, have become an attractive source for coherent mid-infrared radiation. They are compact, powerful (several hundred milliwatts CW), room-temperature mid-IR laser sources, which can be readily microfabricated in large quantities using the same techniques as the diode lasers used in the telecom industry. In the following section, I will detail the development of QCLs and their advantages.

1.4

Quantum cascade lasers

Quantum cascade lasers (QCLs) are unipolar semiconductor lasers based on intersubband transitions in heterostructures. Laser light is generated from a QCL due to a radiative transition resulting from confined states of the electron in the conduction band. This is in contrast to standard diode lasers, which the radiative transition is interband, across the material’s bandgap, and involves both electrons and holes. The material’s bandgap determines the frequency of the emitted light for diode lasers. QCLs have the advantage that by a designed change of the thicknesses of the quantum wells and barriers in the heterostructure, we can change the frequency of the emitted light. For interband transitions, the joint density of states for the transition is constant for energies larger than the transition energy E21 . But for intersubband transitions,


10

the joint density of states is atomic-like and peaked at E21 , resulting in a narrow gain linewidth. Moreover, because the initial and final subbands have the same curvature (neglecting non-parabolicity), this linewidth depends only indirectly on the subband populations through collision processes. Another fundamental property of intersubband transitions is their short lifetimes. For subbands whose energy is separated by more than an optical phonon energy h ¯ !LO , the principal scattering process emits these phonons, resulting in lifetimes of the order of a picosecond.

Figure 1.3: Principal characteristics of (a) an interband transition and (b) an intersubband transition in a quantum well.

1.4.1

Invention of the QCL

The rst proposal to use intersubband transitions in a semiconductor heterostructure for light amplication goes back to the early seventies with the foundational work of Kazarinov and Suris [1971]. However, it wasn’t until 1994 that the first intersubband laser, called the quantum cascade laser by its inventors, was realized in Federico


11

Capasso’s group at AT&T Bell Laboratories, Murray Hill, NJ, USA [Faist et al., 1994]. This breakthrough was achieved using a structure with many cascades, where each stage consists of an active region and a relaxation/injection region, shown in Fig. 1.4.

Figure 1.4: (a) Schematic of a QCL’s conduction band diagram. Each stage of the structure has an active region and a relaxation/injection region. Electrons can emit up to one photon per stage. (b) Outline of the design principle. There is a three-level system in the active region. The lifetime of the transition from level 3 to 2 has to be longer than the lifetime of level 2 to have population inversion. The active region, which in this case consisted of three coupled quantum wells, is a three-level system in which population inversion between levels 3 and 2 is achieved by engineering the lifetimes. The nonradiative relaxation time between levels 3 and 2 can be increased by employing a transition with a reduced spatial overlap of the


12

wavefunctions — for example, a diagonal transition between the states. Also, the lifetime of state 2 is minimized by making the spacing with level 1 resonant with the optical phonon energy, so that there will be a rapid depopulation of state 2. The active region is left undoped because the presence of dopants signicantly broadens the lasing transition by introducing a tail of impurity states. The relaxation/injection region has several functions. It is in this region that the electrons relax after the optical transition and are injected in the next period by resonant tunnelling. It also introduces an additional energy drop between the lower state of the laser transition and the ground state of the period, thus reducing the thermal backlling of the former. Finally, it is doped to act as an electron reservoir, insuring that the integrated negative charge in the structure is compensated by the positive donors, even in the situation of strong injection to prevent the formation of space-charge domains.

1.4.2

Intrinsic advantages of QCLs

A key advantage of the quantum cascade laser is that the energy of the lasing transition is determined by the confinement energy of the electrons in the quantum wells, rather than the intrinsic bandgap energy of the material. This means that the emission frequency can be chosen by design, and frees us from “bandgap slavery.” Moreover, using the same material system, a significant frequency range can be covered by adjusting the design, with the transition energy being adjustable over a range of up to 50% of the conduction band o↵set. The cascaded nature of the QCL gives a second advantage — if there are Nc


13

repeated cascades in the active region of the device (typically 20 to 50), then the electrons will be recycled Nc times. This results in a large di↵erential quantum efficiency ⌘d which is proportional to Nc . This means that high output power from the laser can be achieved, even with devices that have a small pumped area, such as narrow-ridge devices that have the benefit of lasing in a single spatial mode. This is preferable to using broad ridge devices which would lase in multiple transverse modes. Finally, QCLs employ only electrons and not holes, so they can be called “unipolar” lasers. Thus there are no problems with surface recombination of carriers that would degrade the performance of QCLs. This makes QCLs relatively robust.

1.4.3

Development of QCLs

Here I will quickly review the development of QCLs since their invention. The initial design involved an active region with three quantum wells. Subsequently, the injector region of the design was modified to act as a Bragg reflector for the electrons in the upper laser state [Faist et al., 1995], which reduces the probability that these electrons will escape to the continuum (states not bound in the quantum wells). A significantly improved design was dubbed the “two-phonon” design; it uses a four quantum well structure. It has a four-level active region where the lasing transition occurs between levels 4 and 3, and the three lower levels are each separated by an optical phonon energy. This ensures a rapid depopulation of the lower lasing level, state 3, and ensures that there is a smaller thermal population of that level than in the three quantum well design. Room temperature, continuous-wave operation of QCLs was first achieved by Beck et al. [2002] using this method.


14

A dramatically di↵erent design was pursued by Scamarcio et al. [1997], where the lasing transition was not between discrete energy states, but rather between di↵erent “minibands” of a superlattice structure. These minibands are a series of energetically closely-spaced levels with overlapping, delocalized electronic wavefunctions, bearing a resemblance to the electronic bands of crystal lattices. Population inversion with this design relies on the long relaxation time between minibands as compared to within the same miniband.

1.4.4

Bound-to-continuum design

The innovation of the superlattice structure with lasing between minibands led to the development of the so-called bound-to-continuum design by Faist et al. [2001]. The basic idea of this design is that it combines the advantages of the original three quantum well design with the superlattice idea. Electrons are injected into the “bound” (localized) upper state by resonant tunnelling as in the three quantum well design, and they make a lasing transition to a lower miniband of states. As previously mentioned for the superlattice design, the miniband of states rapidly scatters the electron through the fast intraminiband relaxation time. An important and useful feature of the bound-to-continuum design (Fig. 1.5) is that the oscillator strength is not just in a single transition, but instead is spread out over several di↵erent transitions going from the bound upper state to the highest energy states of the lower miniband. Since these states are typically spaced by an energy of ⇠ 20 meV, the result is a broader gain curve than for other active region designs. This makes the bound-to-continuum design particularly well-suited


15

for making a broadband laser source in the mid-infrared. In this thesis, we will detail the development of distributed-feedback quantum cascade laser arrays utilizing the bound-to-continuum active region design, to pioneer broadband single-frequency QCL sources for mid-IR spectroscopy and chemical sensing.

Figure 1.5: Schematic conduction band diagram of a bound-to-continuum design. The wells are tilted due to the applied electric field (35 kV/cm) and the moduli squared of the computed wavefunctions is plotted. Courtesy of R. Maulini.

The bound-to-continuum design we used was first presented by Maulini et al. [2004]. Its calculated band diagram is presented in Fig. 1.5. The layer sequence of one period, in nanometres, starting from the injection barrier is 3.9/ 2.2/ 0.8/ 6/ 0.9/ 5.9/ 1/ 5.2/ 1.3/ 4.3/ 1.4/ 3.8/ 1.5/ 3.6/ 1.6/ 3.4/ 1.9/ 3.3/ 2.3/ 3.2/ 2.5/ 3.2/ 2.9/ 3.1 where In0.52 Al0.48 As layers are in bold, In0.53 Ga0.47 As in roman and the doped layers (Si 2.3⇥1017 cm?3 ) are underlined.


1.4.5

16

State-of-the-art

At the present time, mid-infrared QCLs with emission wavelengths from 3 to 24 µm have been demonstrated [Capasso et al., 2002]. Additionally, QCLs have been developed that emit in the terahertz range of frequencies [Kohler et al., 2002], which up until recently had lacked solid-state sources of radiation. Standard terahertz QCLs require cryogenic cooling, with the current record operating temperature being 178K [Belkin et al., 2008b]. In an exciting development, a novel design that produces terahertz radiation from two mid-infrared QCL sources via di↵erence-frequency generation operates readily at room temperature [Belkin et al., 2008a]. An important milestone for the commercialization of QCLs occurred when metalorganic vapor phase epitaxy (MOVPE) growth of QCLs was demonstrated [Roberts et al., 2003]. MOVPE is the same technology that is used to make telecom laser diodes. This method has several advantages over its competitor, molecular beam epitaxy (MBE), for commercial production of QCLs. MOVPE reactors can deposit multiple wafers simultaneously, and can be more easily maintained since they do not require the same level of ultra-high vacuum conditions needed by MBE. Also, with MOVPE, thick phosphide layers can be grown rapidly; thus the entire structure including top InP cladding layers can be completed quickly in a single step process. Continuous-wave (CW) operation at room temperature, first demonstrated by Beck et al. [2002] has become routine for mid-infrared QCLs. High-power CW QCLs operating at room temperature was first achieved by Yu et al. [2003]. In continuing developments, high-power CW QCLs made using MOVPE techniques have been demonstrated [Diehl et al., 2006a], and recently QCLs with watt-level power have


17

been reported [Bai et al., 2008; Lyakh et al., 2008]. Broadband wavelength coverage is very useful for QCL applications, and so the development of QCLs with broadband gain has been an important goal. As previously mentioned, the bound-to-continuum design with inherently broad gain over 200 cm

1

was pioneered by Faist et al. [2001]. Gmachl et al. [2002a] developed an

ultra-broadband laser with broad gain by having a heterogeneous cascade, with many di↵erent cascades in the active region of the laser, designed to emit at di↵erent frequencies. Combining these two ideas, Maulini et al. [2006] demonstrated a QCL with gain over 350 cm

1

near 9 µm wavelength, by having a heterogeneous cascade of two

bound-to-continuum structures designed to emit at 8.4 and 9.6 µm respectively. Most QCLs have been made in the InGaAs/AlGaAs material system on InP substrates (this includes all the literature cited thus far), but devices made with GaAs/AlGaAs were demonstrated by Sirtori et al. [1998], ones with InAs/AlSb were done by Ohtani and Ohno [2003], and ones with InGaAs/AlAsSb by Revin et al. [2004]. The latter two material systems are particularly interesting for short-wavelength QCLs below 4 µm, as the conduction band o↵set of those material systems is large (2.1 and 1.6 eV respectively). All-in-all, quantum cascade lasers have made great strides as coherent sources of mid-infrared and terahertz radiation. Particularly in the important mid-infrared “molecular fingerprint” region of interest for chemical sensing, QCLs have matured as a high-power, continuous-wave, room-temperature laser source, with significant potential for commercialization and mass-production.


1.5

18

Frequency control of QCLs

For most applications, it is necessary to have QCL devices operating at a single desired frequency. In particular for spectroscopic applications, it is important to achieve sufficient spectral resolution to resolve di↵erent absorption features — the spectral resolution should be comparable or better than the linewidth of the targeted absorption features. For laser sources, there are several ways to control the emission frequency, in particular by using distributed-feedback in the laser cavity, or by incorporating the laser in an external cavity that provides frequency-selective optical feedback.

1.5.1

External cavity QCLs

Incorporating the QCL into an grating-coupled external cavity can be an e↵ective way of controlling its frequency. For example, in what is called the Littrow external cavity setup, the first-order di↵racted beam from the grating is fed directly back into the laser chip. This enforces single-frequency lasing, as the di↵racted beam only provides feedback for one selected wavelength, set by the angle of the grating. To tune the emission frequency, one can vary the angle of the grating relative to the laser chip. For the simplest Littrow setup, only a single lens (to collimate the laser light) and a suitable grating on a rotation mount are required. Certainly this is sufficient to obtain coarse frequency tuning; however, achieving good fine-tuning of the laser emission frequency is much more difficult. Firstly, a high-quality anti-reflection coating (less than 1% residual reflectivity) is needed for the front facet of the laser device. Otherwise the back-reflection from


19

this facet will contribute significantly to the optical feedback, producing lasing on the Fabry-Perot modes of the laser chip, rather than on the modes selected by the external cavity. Moreover, tuning the frequency by rotating the grating will result in mode-hopping rather than continuous tuning, unless the (external) cavity length is kept constant at the same time. This occurs because the QCL lases on the FabryPerot modes of the full cavity, so there will be mode-hopping if the cavity length changes. The same dilemma a↵ects fine-tuning of the emission frequency by changing the temperature of the laser devices, since the e↵ective cavity length changes with the refractive index shift. A solution is to use two piezoelectric controllers to simultaneously control the grating angle and adjust the length of the cavity. However, this adds significantly to the cost and complexity of the system. Nonetheless, significant strides have been made in the development of externalcavity QCLs (EC-QCLs). The first realization of the EC-QCL was by Luo et al. [2001], who demonstrated a tuning range of 32 cm cm

1

1

for a 4.5 µm laser and a 34

range for a 5.1 µm laser, operating pulsed at 80K. This first result did not

achieve continuous tuning as the chip was not anti-reflection coated and operated on the Fabry-Perot modes of the chip. Since then, progress has been rapid, with continuous tuning demonstrated (though still with mode-hopping) [Luo et al., 2002], pulsed room temperature operation [Totschnig et al., 2002], broadband tuning of 150 cm

1

around 10 µm wavelength (also pulsed at room temperature) using a bound-to-

continuum active region design [Maulini et al., 2004], and continuous-wave operation of a thermoelectrically-cooled EC-QCL [Maulini et al., 2005]. The state-of-the-art in tuning range is now over 265 cm

1

[Maulini et al., 2006],


20

done by external-cavity tuning of a QCL with a heterogeneous cascade of two boundto-continuum structures designed to emit at 8.4 and 9.6 µm respectively. EC-QCLs are beginning to be used in applications to chemical sensing and trace-gas analysis, for example in Wysocki et al. [2005]; Pushkarsky et al. [2006]; Kosterev et al. [2008]; Wysocki et al. [2008].

1.5.2

Distributed feedback QCLs

A distributed feedback (DFB) grating has a periodic variation in refractive index (and/or the gain/loss) along its length; this provides optical feedback via backwards Bragg scattering from the periodic perturbations [Kogelnik and Shank, 1972]. DFB lasers are a well-known technique for achieving single-mode operation. Their use is standard in telecom diode lasers. The first distributed-feedback quantum cascade laser (DFB-QCL) was realized by Faist et al. [1997]. The grating was etched into the top cladding of the QCL and covered with the metal top contact. As such, the grating was both index- and losscoupled; the metal introduced losses for laser modes that were spatially overlapped with it. Gmachl et al. [1997] developed buried DFB gratings, where the grating corrugation is an etched corrugation right next to the QCL active region, subsequently buried when the top cladding of the laser is grown above it. The proximity of the grating to the active region results in a significant spatial overlap between the laser mode and the grating, giving a large coupling strength for the DFB. Another advantage of buried DFBs is that they are made with semiconductor layers that give a negligible contribution to waveguide losses.


21

(a)

(b)

1 um

Figure 1.6: (a) SEM image of the facet of a distributed feedback quantum cascade laser, with a top grating. This image is taken from Faist et al. [1997]. The red arrow drawn onto the image denotes the direction of light emission from the laser ridge. (b) SEM image of a cross-sectional slice of a buried DFB grating, showing the grating grooves. Several other kinds of DFB-QCLs have been developed, including devices with second-order gratings for surface emission [Hofstetter et al., 1999; Pflugl et al., 2005] and devices with laterally etched DFBs [Golka et al., 2005; Kennedy et al., 2006]. The best performance is still with buried gratings, as they have the largest coupling strength and minimal contributions to waveguide losses. DFB-QCLs operating continuous-wave at room temperature with hundreds of milliwatts of output power is the current state-of-the-art [Wittmann et al., 2006]. One of the main drawbacks of DFB lasers is that the tuning range is limited for individual DFBs. DFB-QCLs can be tuned by changing their operating temperature, by heating the laser with a DC bias current or by changing the heatsink temperature. However, the tuning coefficient is typically only ⇠0.07 cm 1 /K [Gmachl et al., 2002b]. Thus, the tuning range of a single DFB-QCL is limited to ⇠ 10 cm 1 . This limitation can be circumvented by making arrays of DFB-QCLs, so that a much larger range of frequencies can be covered; this is the subject of the work presented in this thesis.


22

DFB-QCL arrays can be used to generate a broadband mid-infrared laser source. In this thesis, I will discuss the design of DFB-QCL arrays, their performance, applications to chemical sensing, and potential improvements to their performance. In chapter 2, I will describe the design considerations for DFB-QCL arrays. In chapter 3, I will review the fabrication of these arrays, and their performance characteristics. Chapter 4 will show an application of a DFB-QCL array to infrared absorption spectroscopy. Chapter 5 will demonstrate how all the beams from di↵erent lasers in the array can be combined together, with important implications for applications such as remote sensing. In chapter 6, I will exhibit an ultra-broadband DFB-QCL array that covers a range of > 200 cm 1 . Chapter 7 will discuss some exploratory work towards the use of DFB-QCLs for intra-cavity sensing. Finally, I will present my conclusions and ideas for future work in chapter 8.

Chapter 2 Distributed feedback QCL design In this chapter, we describe the theoretical framework developed for generic distributed feedback gratings, and we apply these methods to design DFB-QCL arrays. We calculate how the performance characteristics of DFB-QCLs are a↵ected by the coupling strength L of the grating, and the relative position of the mirror facets at the ends of the laser cavity. Performance characteristics we examine include singlemode selection, threshold, slope efficiency and output power. Single-mode selection refers to the ability to choose a desired mode/frequency of laser emission with a DFB grating. We discuss how single-mode selection can be improved by design. This chapter serves as a prelude to the next chapter, where we obtain experimental results from fabricated DFB-QCL arrays, and compare those results to the predictions made here.

23

Chapter 2: Distributed feedback QCL design

2.1

24

Introduction: distributed feedback

To inform our discussion of the waveguide structure and grating design, we briefly review the properties of DFB gratings. In general, a DFB grating has a periodic variation in refractive index (and/or the gain/loss) along its length; this provides optical feedback via backwards Bragg scattering from the periodic perturbations. In other words, the corrugations of the grating cause light to be reflected, and the net e↵ect of all the in-phase reflections is optical feedback for light that satisfies the Bragg condition:

n = 2⇤. Here n is an integer, denoting the di↵raction order,

(2.1) is the wavelength of the

light, and ⇤ is the period of the grating. We are interested in first-order di↵raction from the grating (n=1), as first-order DFBs are typically preferred to achieve stronger optical feedback. The Bragg frequency is defined as ⌫0 = c/2n⇤, where n is the average index of refraction of the grating structure. A periodic variation in the refractive index gives rise to a photonic gap around the Bragg frequency; the DFB supports lasing for longitudinal modes on either side of this gap. There are two modes that are directly at either edge of the photonic gap — these two modes have the lowest lasing threshold of all the possible modes supported by the DFB. We will call these the low- and high-frequency modes — the modes have frequencies ⌫1 and ⌫2 respectively. The low-frequency mode is more concentrated in the higher-index part of the grating, with a modal e↵ective refractive index n1 = c/2⌫1 ⇤. The high-frequency mode is


25

more concentrated in the lower-index part of the grating, so it has a lower modal e↵ective refractive index n2 = c/2⌫2 ⇤. In the coupled-wave formalism for DFBs developed by Kogelnik and Shank [1972], the gratings coupling strength (per unit length) is a complex number , whose real part is proportional to the photonic gap, and whose imaginary part corresponds to the di↵erence in loss (or gain) of the two DFB modes across the gap. In terms of the e↵ective refractive indices of the two modes directly across the photonic gap:

 = ⇡ (n + ik)/ 0 . Here

(2.2)

(n + ik) is the di↵erence in the complex e↵ective refractive indices of the

two modes, and

0

is the Bragg wavelength.

We typically speak of the coupling strength as L, which is a dimensionless quantity and includes the length of the grating L. We can see the e↵ect of the coupling strength L on the spatial distribution of mode intensity along the cavity in Fig. 2.1. For a critically-coupled DFB (L ⇠ 1), the intensity is evenly distributed along the cavity, while the intensity is towards the ends of the cavity in the under-coupled case and towards the centre in the over-coupled case. This is a very visual way of seeing how the DFB provides optical feedback in the cavity, and how this changes for di↵erent coupling strengths. The larger the coupling strength, the more light is reflected by the DFB grating, causing the light intensity to be more concentrated towards the centre of the cavity. A larger coupling strength gives us more optical feedback for the supported DFB modes, which improves the mode selectivity and decreases the threshold current re-


26

quired for lasing. However, a large coupling strength also results in less light escaping from the cavity, meaning that less output power from the laser. Thus in choosing the coupling strength L for a DFB laser, one needs to balance the competing goals of having strong optical feedback and obtaining good output power. Previous theoretical investigations of DFB lasers [Buus, 1985] and experimental studies of DFB-QCLs [Gmachl et al., 2002b] have shown that it is desirable to have slightly over-coupled DFBs where L ⇠ 4.

Figure 2.1: Plot of the spatial intensity distribution of the fundamental mode of the cavity, at di↵erent levels of the coupling strength L ⇠ ⇡ n/⇤. This figure taken from Kogelnik and Shank [1972].

In our designs, we will target large values of  in order to have relatively short lasers. Shorter lasers are desirable since less current is required to pump them.


2.2

27

DFB-QCL grating design

To achieve our target, we need to consider various aspects of the grating geometry, including the location of the grating, the depth of the corrugation, the duty cycle, and the choice of material for the grating. These choices influence the coupling strength of the grating, and also the optical losses of the waveguide. In terms of the location of the grating, one choice is to use top gratings, where the corrugation is an air-dielectric interface etched into the surface of the top cladding of the device, as in Faist et al. [1997]. However, this type of structure su↵ers from a relatively small coupling strength (per unit length) since the grating is far from the active region of the laser, so that the overlap of the grating with the laser mode is small. Another choice is to employ buried DFB gratings, where the grating corrugation is an etched corrugation right next to the QCL active region, subsequently buried when the top cladding of the laser is grown above it. Buried DFBs were first demonstrated for QCLs by Gmachl et al. [1997], and they have the benefit of having both a large coupling strength and negligible contribution to waveguide losses. Given the advantages of buried gratings, we choose to make a first-order DFB that is formed in the upper waveguide cladding of the QCL, as an etched corrugation in a buried InGaAs layer just above the active region of the laser. This is shown in Fig. 2.2a. Referring to Fig. 2.2a, the DFB-QCL structure consists of: a bottom waveguide cladding of 4 µm of InP doped 1⇥1017 cm doped 3⇥1016 cm

3

3

(yellow), followed by 580 nm of InGaAs

(green), a 2.4 µm-thick lattice-matched active region (red), 580

nm of InGaAs doped 3⇥1016 cm

3

where the grating is etched 500 nm deep (blue),

and a top waveguide cladding (yellow), consisting of 4 µm of InP doped 1⇥1017 cm

3


28

and 0.5 µm of InP doped 5⇥1018 cm 3 . The active region consists of 35 stages based on a bound-to-continuum design emitting at ⇠ 9 µm — this active region design is shown in chapter 1.4.4 of this thesis, and also in Maulini et al. [2004]. The grating provides distributed feedback by having a refractive index contrast between the corrugated InGaAs layer and the InP material that is overgrown above it. Locating the grating just above the active region means that there is excellent overlap between the laser mode and the grating, resulting in a larger grating coupling strength as compared to surface gratings. The duty cycle of the grating is chosen to be 50%, which is typical of first-order DFBs, as this maximizes the refractive index contrast. Finally, we performed simulations to determine the appropriate etch depth of the grating for our desired grating coupling strength. In a 2D simulation done using the commercial software COMSOL 3.2, we find numerically the longitudinal modes of the structure shown in Fig. 2.2a, for a fixed transverse mode (ie. TM00 ). Figs. 2.2b and c shows the mode profile calculated for the low- and high-frequency DFB modes, for a single period of the grating. In the grating structure, the part which is indented down (“grating troughs”) has a lower e↵ective refractive index than the part which is raised up (“grating peaks”). As previously mentioned, the low-frequency mode has higher electric field magnitude in the grating peaks which is the higher-index part of the grating, and also that the high-frequency mode is more concentrated in the lower-index part of the grating. We perform the simulation for di↵erent grating depths. From the simulation, we determined the emission frequencies of the two DFB modes and also the e↵ective refractive index experienced by each mode. For a 500 nm grating depth, the complex


29

(a)

(b)

(c)

(d)

Figure 2.2: (a) Schematic of the DFB-QCL waveguide structure. The grating is etched in an InGaAs layer (blue) in the top waveguide cladding, just above the active region (red). There is InP overgrown on top of the grating (yellow in diagram). (b) Mode simulation of the mode on the low-frequency side of the photonic gap of the DFB grating. The plot displays the magnitude of the electric field in the laser structure. The low-frequency mode has more of the electric field concentrated in the high-index part of the grating as expected. (c) Mode simulation of the mode on the high-frequency side of the photonic gap of the DFB grating. Again, we display the magnitude of the electric field. The high-frequency mode has more of the electric field concentrated in the low-index part of the grating as expected. (d) The dependence of the grating coupling strength per unit length  on the etch depth of the grating. The real part of  is shown as squares and the imaginary part as circles.


30

e↵ective refractive indices (n + ik) of the low- and high-frequency modes were 3.191 + 5.85⇥10 4 i and 3.182 + 5.34⇥10 4 i, respectively. The real part of these numbers corresponds to the mode e↵ective refractive index, while the imaginary part corresponds to waveguide loss/absorption ↵w = 4⇡⌫k/c. These waveguide losses are 8.1 and 7.4 cm cm

1

1

for the low- and high-frequency modes, which is comparable to the 8.3

found experimentally for buried-heterostructure QCLs with a similar waveguide

[Diehl et al., 2006a]. The origin of the waveguide losses is free carrier absorption due to the doping of the semiconductor layers. Using the values of refractive index from the simulation, we calculated that the coupling strength per unit length of the 500 nm deep grating is  = (31 + 0.17i) cm 1 . The grating coupling strength per unit length was also found for shallower gratings of 100 nm and 300 nm depth (Fig. 2.2d). With shallower gratings there is less index contrast, so  is smaller. Since we desire short devices for lower pump power and L ⇠ 4, we chose to etch the grating 500 nm deep, so that devices can be only ⇠ 2 mm long.

2.3

DFB coupling strength and end facet mirrors

The coupling strength of the DFB grating and the presence of mirrors have a significant impact on the performance of the device, including single-mode selection and power output. In order to gain a greater understanding of these e↵ects, we present a theoretical analysis of the DFB coupling strength. We first show the intensity profile of the laser light along the length of the DFB for an idealized DFB without end mirrors and then proceed to discuss the impact of reflections from end facet mirrors.


31

It is known that reflectivity and position of the end facet mirrors are critical factors a↵ecting power output and single-mode selection. We investigate ways to decrease the e↵ect of the mirrors and improve single-mode selection for DFB-QCLs.

2.3.1

Intensity profile and power output

We illustrate in Fig. 2.3a the intensity profile along the length of the laser ridge for di↵erent magnitudes of the coupling constant |L| in an approximation by Kogelnik and Shank [1972] where we assume no mirror reflections. This figure shows that the light is more confined to the central part along the length of the cavity when L is large, versus being more evenly distributed throughout the cavity when L is smaller. Taking into account the end mirror reflections in a real device, we will show that the relative position of the end facets can make a large di↵erence in the intensity profile along the length of the cavity. Here we build upon work by Streifer et al. [1975] who treated the case of a finite-length grating with end reflectors outside the region where the grating exists. They allowed the relative position of the end reflectors with respect to the grating to vary. This matches our case, where the position of our end facet mirrors relative to the grating is basically random for each of the lasers in the array, as we have no precise control over position of the facet with our present fabrication methods. Following the work of Streifer et al. [1975], we consider coupled-wave solutions of the electric field along the grating. The grating couples right- and left-traveling waves R(z) = r1 e

z

+ r2 e

be determined and

z

and S(z) = s1 e

z

+ s2 e

z

where r1 , r2 , s1 , s2 are constants to

obeys the eigenvalue equation:


32

(c) 

(b) 

2nd mirror position (units of π)

(a) 

(d) 

1.0 0.5

0.8 0.6

0.0

0.4 0.2

-0.5 -1.0 -1.0 -0.5 0.0 0.5 1.0 1st mirror position (units of π)

Figure 2.3: The intensity profile of the laser light along the length of the cavity. (a) The intensity profile for coupling constants L = 1, 4, 11 in the absence of mirrors. (b) The intensity profile for L = 4 and a very asymmetric mirror configuration — one mirror is coincident with the grating peaks and the other with the grating troughs. (c) The intensity profile for L = 11 and a very asymmetric mirror configuration — one mirror is coincident with the grating peaks and the other with the grating troughs. (d) A plot of the variability of the power emitted from the front facet of the laser as a fraction of the total power from both facets, for arbitrary positions of the end facet mirrors relative to the grating. The grating coupling strength here is L = 4.6 + 0.025i.


33

( L)2 +(L)2 sinh2 ( L)(1 ⇢2l )(1 ⇢2r )+2iL(⇢l +⇢r )(1 ⇢l ⇢r ) Lsinh( L)cosh( L) = 0. (2.3) Also

satisfies the dispersion relation

2

= (↵

i )2 + 2 and

= (1 + ⇢l ⇢r )2

4⇢l ⇢r cosh2 ( L). Here ⇢l and ⇢r are the reflectivities of the left and right mirrors respectively, including both the amplitude and phase of the reflection. ↵ is the total loss of the mode, which can also be seen as the threshold gain required for that mode to lase. is the frequency of the mode relative to the Bragg frequency; in particular the lowand high-frequency modes directly on either side of the bandgap have

< 0 and

>0

respectively. If we set ⇢l = ⇢r = 0, we recover the specific case of no mirror reflections in Kogelnik and Shank [1972]:

( L)2 + (L)2 sinh2 ( L) = 0.

(2.4)

By solving the general eigenvalue equation, we can find the allowed values for , corresponding to di↵erent longitudinal modes supported by the grating. The solutions also yield the total loss ↵ and frequencies

(relative to the Bragg frequency) of these

modes. Moreover, the intensity profile of each mode along the length of the cavity can be obtained from the sum of the right- and left-traveling waves R(z) and S(z). We investigate a few illustrative cases to predict the performance of DFB-QCL arrays. If the mirrors are arranged symmetrically (i.e. the relative position of both end facets with respect to the grating grooves is the same) then clearly the intensity profile


34

remains symmetric. However, in general the mirrors are not arranged symmetrically. In Figs. 2.3b and c we illustrate particularly asymmetric cases, where one mirror is coincident with a “trough” of the grating while the other is coincident with a “peak” of the grating. In Fig. 2.3b, we see an asymmetric intensity profile for a grating where L = 4 and the mirror reflectivity is 30%. The curves in the graph are the total intensity, and the intensities of the right-traveling and left-traveling waves R(z) and S(z). The power output from the right facet is given by the intensity of the right-traveling wave at the facet multiplied by the transmission coefficient of that facet (70%). Similarly the power output from the left facet is the intensity of the left-traveling wave at the facet multiplied by the transmission coefficient of that facet. From the large asymmetry of the intensity profile, we can see that the light intensity output from one facet can be an order of magnitude greater than from the other facet. Fig. 2.3c shows that the asymmetry of the light output remains for larger L — here we show the case where L = 11. The overall intensity profile varies less, since most of the mode is concentrated in the center of the laser cavity. However, the ratio between the intensities of light output from the left and right facets remains large. We repeat this calculation for arbitrary mirror positions and a fixed L (= 4.6 + 0.025i). Fig. 2.3d shows the power output from the front facet of the laser, as a fraction of the total power emitted from both end facets. While a slight majority of plotted points have between 40% and 60% of the light being emitted from the front facet, there are mirror configurations which give 5% or 95% of the power from one facet. Thus there can be a large variability in the output light intensity measured


35

from the front facet of lasers with di↵erent mirror configurations, as great as an order of magnitude in range. With an array of DFB-QCLs where the end mirrors all have arbitrary positions relative to the grating, this variability will be seen in large di↵erences in the output power measured from di↵erent lasers in the array. This variability will also appear in the slope efficiencies dP/dI of di↵erent lasers in our DFB-QCL arrays. Our calculations demonstrate the inherently large variability in the light intensity output of DFB-QCLs in an array when we have significant reflectivity from the end mirrors and arbitrary mirror positions. In order to remove this variability one could either find a way to reliably terminate the laser cavity at the same point relative to the local grating for each of the lasers in the array, or reduce the reflectivity of both mirror facets with anti-reflection coatings.

2.3.2

Single-mode selection

The end facet mirrors can also have a significant impact on the single-mode selection, which we defined earlier as the ability to choose a specific desired mode/frequency of laser emission with a distributed feedback grating. More specifically, we will have good single-mode selection if we can reliably cause only the low-frequency (or only the high-frequency) mode to lase for every DFB-QCL in the array. A mode lases at a specific frequency when the gain at this frequency exactly compensates for the total optical losses, which include the waveguide losses and the mirror losses over one round-trip in the cavity. For our lasers, we can calculate from the gain spectrum (FWHM of ⇠ 300 cm 1 ) that the amount of modal gain for lasing

Chapter 2: Distributed feedback QCL design at closely-spaced frequencies (ie. 3.1 cm

1

36

apart) di↵ers by 1% or less. This means

that the di↵erence in total loss between the two DFB modes primarily determines which one of them actually lases. Specifically, the DFB mode with the lower optical loss will lase. The two DFB modes have di↵erent waveguide losses. From the previous calculations of waveguide losses, these are 8.1 and 7.4 cm

1

for the low- and high-frequency

DFB modes respectively. Now if the e↵ect of end facet mirrors could be ignored, then the mode with the smaller waveguide losses will always lase. However, the presence of end facet mirrors gives reflections that constructively or destructively interfere with the DFB modes in the laser cavity. This interference a↵ects the mirror loss of each mode, and can determine which mode lases. We note that the e↵ect of the mirrors is largest when the position of both mirrors coincide with a peak in electric-field amplitude of one DFB mode, which is also when the mirrors are at a node for the other DFB mode. When both mirrors coincide with the peaks, then the reflections from the end mirrors maximally constructively interfere with the mode present in the laser cavity. This results in a lower total loss, due to the constructive contribution of the mirrors. When both mirrors coincide with the nodes, then the reflections from the end mirrors destructively interfere with the mode present in the laser cavity. This results in a higher total loss, due to the contribution of the mirrors. Using the method developed by Streifer et al. [1975], we calculated the total losses for the two DFB modes for arbitrary end mirror positions and plotted the di↵erence between these losses in Fig.2.4. The total losses include both the waveguide losses and

Chapter 2: Distributed feedback QCL design the mirror losses, which can vary by several cm

37 1

and thus dominate the di↵erence

in waveguide losses. Here the possible positions of the mirrors with respect to the grating are designated by a phase, which can range between 0 and ⇡ — corresponding to the case where the mirror is either exactly in line with a grating trough or with a grating peak. A total loss di↵erence above zero means that the high-frequency mode has greater loss than the low-frequency mode. When the plot is below zero, then the high-frequency mode has the lower loss. Fig. 2.4a shows the case for a grating with coupling strength L = 4.6 + 0.025i and both mirrors having 30% reflectivity. This corresponds to a laser array that is 1.5 mm long, fabricated with the grating structure we previously simulated. The end facets of the laser ridge give a 30% reflectivity from the semiconductor/air interface. Fig. 2.4a predicts that, on average with random mirror positions, the high-frequency mode will have lower loss 57% of the time, while the low-frequency mode will have lower loss 43% of the time. This means that, for a sufficiently large array of lasers where the statistics hold, 57% of the time the high-frequency mode will lase and 43% of the time the low-frequency one will. One way of decreasing the e↵ect of the end mirror facets is to put an anti-reflection coating on the output facet, which will decrease the reflectivity of that facet and hence its e↵ect on the total losses of the DFB modes. In Fig. 2.4b, we show the case where we have a grating with coupling strength L = 4.6 + 0.025i, the first mirror having an antireflection coating that reduces its reflectivity to 1% and the second mirror having 30% reflectivity. We see that the first mirror now barely has an e↵ect on the losses, which increases the probability that the high-frequency mode lases to 75%, while now


(a) 

38

(b) 

(c)  Figure 2.4: These three plots show the di↵erence in total loss between the high- and low-frequency modes, for di↵erent coupling strengths L of the grating and di↵erent configurations of the end mirrors. The possible positions of the mirrors with respect to the grating are designated by a phase, which can range between 0 and ⇡ — corresponding to the case where the mirror is exactly in line with a grating groove and exactly anti-aligned respectively. When the surface of the plot of total loss is above zero, it means that the high-frequency mode has greater loss than the low-frequency mode. When the plot is below zero, then the high-frequency mode has the lower loss. (a) A plot for the case where L = 4.6 + 0.025i and both mirrors have 30% reflectivity. (b) A plot for the case where L = 4.6 + 0.025i, the first mirror has an antireflection coating with 1% residual reflectivity and the second mirror has 30% reflectivity. (c) A plot for the case where L = 11 + 0.058i and both mirrors have 30% reflectivity.


39

the low-frequency mode only lases 25% of the time. This is due to the smaller e↵ect of the mirrors and the lower waveguide losses of the high frequency mode. Another way to decrease the e↵ect of the end mirrors is to have a stronger DFB grating. Having a stronger DFB grating means that more of the laser light will be reflected by the grating grooves and the laser mode will be more confined towards the center part of the laser cavity, and have much lower intensity towards the ends of the cavity where the mirrors are. This situation was illustrated in Fig. 2.3, and described in our previous discussion of the intensity profile of a DFB with L = 11. In Fig. 2.4c we show the case where L = 11 + 0.058i and both mirrors have 30% reflectivity. This corresponds to a laser array that is 3.5 mm long, fabricated with the grating structure we previously simulated. Now the loss di↵erence due to the end mirrors for the two modes is almost always less than the di↵erence in their waveguide losses. Hence we are sufficiently insensitive to the end mirrors that the mode with the smaller waveguide losses, the high-frequency mode, will lase 95% of the time.

2.4

Summary

We investigated DFB-QCL arrays to predict their performance characteristics — single-mode selection, threshold, slope efficiency and output power. We expect that the single-mode selection of the DFB gratings will be a↵ected by both the coupling strength L of the grating, and by the position of the end mirror facets. The end mirror facets, which are randomly positioned relative to the grating, strongly a↵ect which DFB mode lases. Better single-mode selection can be achieved by either depositing anti-reflection coatings on the end facets or by using a strongly over-coupled

Chapter 2: Distributed feedback QCL design grating (|L|

40

1). We expect large variability in the slope efficiency and output

power among di↵erent lasers in the array, also caused by the end mirrors. In the next chapter we will show the results of fabricated DFB-QCL arrays and make a comparison with our predictions.

Chapter 3 Distributed feedback QCL arrays Distributed feedback quantum cascade laser (DFB-QCL) arrays are fabricated as a proof-of-concept. We investigated their performance characteristics — singlemode selection of the DFB grating, and variability in threshold, slope efficiency and output power of di↵erent lasers in the array. Single-mode selection refers to the ability to choose a desired mode/frequency of laser emission with a DFB grating. In the previous chapter, we applied a theoretical framework developed for general DFB gratings to analyze DFB-QCL arrays. Here, we look at the results from several DFBQCL arrays that are fabricated and then tested. We achieve desired improvements in single-mode selection, and we observe the predicted variability in the threshold, slope efficiency and output power of the DFB-QCLs. These results are also reported in the literature as Lee et al. [2007] and Lee et al. [2008].

41

Chapter 3: Distributed feedback QCL arrays

3.1

42

Device fabrication

We fabricated several DFB-QCL arrays to experimentally determine their characteristics. The QCL material used to fabricate the laser arrays was grown by MOVPE. The layer structure grown is as described in the previous chapter, section 2.2. Device processing started with the fabrication of arrays of 32 buried DFB gratings in the QCL material. The fabrication process is described in Fig 3.1. Grating periods ranged between 1.365 and 1.484 µm, satisfying the Bragg condition for lasing wavelengths between 8.71 and 9.47 µm, assuming an e↵ective refractive index of 3.19 as calculated in our mode simulations. To fabricate the buried gratings, the top waveguide cladding was removed down to the first InGaAs layer, using concentrated HCl as a selective wet etch. Then, a 200 nm thick layer of Si3 N4 was deposited on top of the InGaAs by chemical vapor deposition. First-order Bragg gratings were exposed onto AZ-5214 image-reversal photoresist by optical lithography, using a photomask where the grating patterns had been defined by electron-beam writing. This pattern was transferred into the Si3 N4 by using a CF4 -based dry-etch. The gratings were then etched 500 nm deep into the top InGaAs layer with a HBr/BCl3 /Ar/CH4 plasma in an inductively coupled plasma reactive ion etching machine (ICP-RIE). The InP top cladding was re-grown over the gratings using MOVPE. Laser ridges, 15 µm wide and spaced 75 µm apart, were defined on top of the buried gratings by dry-etching the surrounding areas 9 µm deep with a HBr/BCl3 /Ar/CH4 plasma using ICP-RIE. During this step, the back facet of the lasers was also defined. The bottom and the sidewalls of the laser ridges were insulated by Si3 N4 and a 400nm-thick gold top contact was deposited. The samples were then thinned to 200 µm


43

Figure 3.1: Fabrication process flow. In step 7, the red arrow denotes the direction of light emitted from the facet of the laser ridge.


44

and a metal bottom contact was deposited. Finally, the front facets of the lasers were defined by cleaving. The device in various stages of completion is shown in Fig. 3.2. In this paper we discuss results from two di↵erent arrays; one of these was cleaved to obtain 1.5-mm-long lasers, and the other was cleaved to obtain 3.5-mm-long lasers. Each laser array was indium-soldered onto a copper block for testing. The entirety of each DFB laser array chip is only 4 mm by 5 mm in size (Fig. 3.2d).

(a)

(c)

(b)

1 um

(d)

Figure 3.2: (a) DFB grating etched into the InGaAs layer, just above the active region. This SEM image is taken before regrowth of InP on top, which buries the grating. (b) A single ridge laser device, out of the 32 in the array, as viewed from its front facet. (c) SEM image of a cross-sectional slice of the laser ridge, taken along the direction of the ridge, showing the grating grooves. (d) The finished array, shown on top of a dime for size comparison.


3.2

45

Array results

To test an array of DFB-QCLs, we applied electrical current pulses to individual laser ridges at room temperature. We also determined the coupling strength of the grating, by measuring the luminescence of the device below the lasing threshold. The luminescence spectrum was measured with a sub-threshold DC current at 77 K. The spectrum shows a photonic bandgap of 3.1 cm

1

(Fig. 3.3a, inset), from which the

real part of  is obtained ( = 30 cm 1 ) in excellent agreement with the theoretical prediction (31 cm 1 ).

3.2.1

Short array

We first present the results of testing the laser array that is 1.5 mm long, which corresponds to a coupling strength L = 4.6 + 0.025i. The spectra for the 32 lasers of this array are presented in Fig. 3.3a. All the lasers operate single-mode with greater than 20dB suppression of side-modes (Fig. 3.3b). However, the emission frequencies are not spaced regularly apart (by ⇠2.7 cm 1 ) as desired. Instead, there is an uneven pattern. As discussed in the previous chapter (section 2.1), there are two possible lasing modes, which exist on either side of the photonic bandgap — the low- and highfrequency DFB modes. The uneven pattern of spectra results because some lasers are lasing in the low-frequency DFB mode while others are lasing in the high-frequency DFB mode. The frequency spacing between adjacent lasers falls under 3 categories: the spacing is the desired spacing (2.7 cm 1 ), the spacing is the sum of the desired spacing and the bandgap (2.7 + 3.1 = 5.8 cm 1 ), or the spacing is the di↵erence


46

(a)

(b)

(c)

Figure 3.3: (a) Spectra of 32 lasers from an array with L = 4.6 + 0.025i. All the lasers are single-mode, but 18 of the lasers are lasing on the high-frequency side of the DFB gratings photonic gap, while 14 are lasing on the low-frequency side. The inset shows the two modes on either side of the photonic bandgap (b) Spectrum of a representative DFB from the array, plotted on a log scale to show >20 dB side-mode suppression. (c) Plot of the voltage (left axis) and light output intensity (right axis) from several lasers in the same array, as a function of pump current. There is a small amount of variability in the threshold current, and a larger variation in the slope efficiencies, which leads to a significant variation in the peak output power.


47

of the desired spacing and the bandgap (2.7 - 3.1 = -0.4 cm 1 ). These correspond respectively to the cases where: two adjacent lasers lase on the same side of the bandgap, the first laser lases in the low-frequency mode and the second in the highfrequency mode, the first laser lases in the high-frequency mode and the second in the low-frequency mode. This e↵ect arises because of the impact of the end facets at either end of the laser ridges — these end facets are partially reflecting (⇠ 30% reflectivity) mirrors. Moreover, the position of these end facets with respect to the grating can vary in a basically random manner, from the random position of the cleaved facet relative to the grating grooves. These mirror positions a↵ect which of the two modes on either side of the photonic gap has lower total loss, and hence which mode lases. In the previous discussion of the impact of end facet mirrors in the previous chapter (section 2.3.2) and Fig. 2.4a, we predicted that 57% of the time the high-frequency mode will lase, while 43% of the time the low-frequency one will. The results agree, with 18 out of 32 lasers lasing on the high-frequency mode and 14 lasers lasing on the low-frequency mode — this is 56% and 44% respectively. We note that the mode selection is stable the same mode lases if the laser is turned o↵ and on again, or if the pump current is varied. This is additional evidence that the variation in mode selection is due to a fixed factor such as the mirror positions. We also observe that there is significant variability in the slope efficiencies and peak output power of the lasers. This can be seen from the light output data in Fig. 3.3c. The I-V characteristic of di↵erent lasers in the array is basically identical, but the light output varies by an order of magnitude, which is consistent with our


48

prediction discussed in the previous chapter (section 2.3.2).

3.2.2

Array with anti-reflection coating

We coated the front facet of the array with 1.5 mm long lasers, to try to improve the single-mode selection — specifically to increase the probability of observing the high-frequency DFB mode for each laser in the array. For the anti-reflection coating, we evaporated a two-layer stack of YF3 and ZnSe, a techique also employed in Maulini et al. [2006]. YF3 has a refractive index of 1.335 at 9 microns wavelength and ZnSe has a refractive index of 2.412. We deposited 610 nm of YF3 and 340 nm of ZnSe films on the front facet of the lasers — this decreases the residual reflectivity of the facet to below 1%. After applying the anti-reflection coating to the array, the laser characteristics were measured again. Fig. 3.4a shows the spectra of the 32 lasers in the array. Now 25 out of 32 lasers are lasing on the high-frequency DFB mode and 6 are lasing on the low-frequency mode, with 1 laser lasing on both modes. Disregarding the one multi-mode laser, this means that 81% of the time the high-frequency mode lases and 19% of the time the low-frequency mode does. This is in reasonable agreement with the prediction of a 75% / 25% split, shown in Fig. 2.4b in the previous chapter. Since the anti-reflection coating increases the mirror losses, the threshold current for lasing is slightly increased. This can be seen in Fig. 3.4b. However, the slope efficiency of the lasers is also higher, since more light is coupled out of the laser cavity from the coated end. Overall the peak output power from the lasers in the array is similar to the situation before the anti-reflection coating was applied. This is simply a


49

(a)

(b)

Figure 3.4: (a) Spectra of 32 lasers from the array with L = 4.6 + 0.025i, after an anti-reflection coating is applied on the front facet, with 1% residual reflectivity. All but one of the lasers are single-mode, with 25 of the lasers are lasing on the highfrequency side of the DFB gratings photonic gap, and 6 lasing on the low-frequency side. One of the lasers has both modes lasing. (b) Plot of the voltage (left axis) and light output intensity (right axis) from several lasers in the same array, as a function of pump current.


50

coincidence of the fact that the e↵ects of the increased threshold and slope efficiency a↵ect the peak output power in opposing ways and roughly cancel each other. Having anti-reflection coatings on the back facet as well would further improve the single-mode selection. However, this is difficult to realize with the present DFB-QCL array geometry where the back facets are etched.

3.2.3

Longer array

We recall that an even more e↵ective way of getting more of the lasers in an array to emit in the same DFB mode (specifically the high-frequency DFB mode) is to have an array with a stronger coupling L. We can achieve this by fabricating an array with longer lasers. Here we present the results of a longer array where the lasers are 3.5 mm long, corresponding to L = 11 + 0.058i. Looking at the spectra of the array in Fig. 3.5a, we see that all the lasers are single-mode and they all lase on the high-frequency side of the DFB gratings photonic gap. This is even better than the theoretical prediction of a 95%/5% split between high- and low-frequency modes in Fig. 2.4c in the previous chapter. So having a strongly over-coupled grating is a very e↵ective way of suppressing the e↵ect of the end facet mirrors in mode selection. However, the light output intensity of the lasers su↵er. Fig. 3.5b shows that the slope efficiency and the peak output power of the lasers is lower than with the shorter array. This is because the light is more highly confined in the central part along the length of the laser ridge by a larger number of reflections from the stronger grating. Less of the light makes it out of the laser cavity so the light output is smaller. One


51

(a)

(b)

Figure 3.5: (a) Spectra of 32 lasers from an array with L = 11 + 0.058i. All the lasers are single-mode, and they all lase on the high-frequency side of the DFB gratings photonic gap. (b) Plot of the voltage (left axis) and light output intensity (right axis) from several lasers in the same array, as a function of pump current.


52

benefit is that the threshold current density required for lasing is also smaller, due to the strong optical feedback in the grating. We see that a more strongly coupled grating can help to reliably select one of the two DFB modes to lase. However, this comes at the cost of decreasing the output power of the lasers.

3.3

Threshold and slope efficiencies

While we have already examined the thresholds and slope efficiencies of the laser arrays, it is worthwhile to summarize those results and compare them. One way of getting a more quantitative view, given the variability inherent in each array, is to view the thresholds and slope efficiencies as histograms. A histogram of threshold currents (or slope efficiencies) shows how many lasers in each array have thresholds (slope efficiencies) within a given range of values. Fig. 3.6a shows a histogram of the threshold current densities for the three cases (short array, array with anti-reflection coating, longer array). The histogram confirms that the threshold current density is lowest for the longer array and highest for the array with anti-reflection coating. As previously discussed, this is expected because the longer array has greater optical feedback from its grating so that the threshold should be lower; also the anti-reflection coating reduces the optical feedback, giving a higher lasing threshold. Fig. 3.6b shows a histogram of the slope efficiencies for the three cases. The histogram displays a clear trend in the magnitude of the slope efficiencies — the longer array has the lowest slope efficiencies, while the short array has higher slope


53

(a)

16 14 12 10 8 6 4 2 0 2.0

2.5

3.0

2

threshold current density (kA/cm )

(b)

20

15

10

5

0 100

200 300 400 slope efficiency (mW/A)

500

600

Figure 3.6: (a) Histogram of the threshold current densities for the three di↵erent DFB-QCL arrays. The array with L = 4.6 + 0.025i is denoted in red with square markers, and after coating the front facet with an anti-reflection coating the new thresholds are denoted in blue with triangles, finally the array with L = 11 + 0.058i is denoted in green with circles. (b) Histogram of the slope efficiencies for the three di↵erent DFB-QCL arrays. The array with L = 4.6 + 0.025i is denoted in red with squares, and after coating the front facet with an anti-reflection coating the new slope efficiencies are denoted in blue with triangles, finally the array with L = 11 + 0.058i is denoted in green with circles.


54

efficiencies, and the anti-reflection coated array has the highest. Again, this is in line with our previous discussion (section 2.3.1). The longer array has stronger coupling |L| ⇠ 11. Thus the light is more highly confined in the central part along the length of the laser ridge by a larger number of reflections from the stronger grating. Less of the light makes it out of the laser cavity so the light output is smaller. The array with anti-reflection coating allows more light to come out of the coated end, so its slope efficiency is higher. The histogram also shows that there is almost an order of magnitude variation in the slope efficiency for each of the three cases. This is due to the e↵ect of the end facet mirrors. The reflections from the end facet mirrors cause highly asymmetric intensity profiles along the length of the laser cavity. This was predicted from theory in the previous chapter (section 2.3.1). Here we reiterate that a highly asymmetric intensity profile along the laser cavity means that the amount of light emitted from one facet of the laser does not have to be equal to the amount emitted from the other facet. For an array of lasers, this results in a large variability in the measured output power from the front facet between di↵erent lasers in the array. As the slope efficiency is dP/dI, it also shows up as a large variability in the slope efficiency.

3.4

Summary

We investigated DFB-QCL arrays for their performance characteristics — singlemode selection, threshold, slope efficiency and output power. We observed that the single-mode selection of the DFB gratings is a↵ected by both the coupling strength L of the grating, and by the position of the end mirror facets. The end mirror


55

facets, which are randomly positioned relative to the grating, strongly a↵ect which DFB mode lases. Better single-mode selection can be achieved by either depositing anti-reflection coatings on the end facets or by using a strongly over-coupled grating (|L|

1). The variability of the threshold, slope efficiency and output power among

di↵erent lasers in the array are also caused by the end mirrors.

Chapter 4 DFB-QCL arrays for mid-infrared spectroscopy Our DFB-QCL array is a mid-infrared source that can emit any frequency within a designed range. It covers that range by being continuously tunable, since the separation in nominal emission frequencies is small enough that we can use temperature tuning to span the spacing. With such an array, we can perform infrared spectroscopy, with many potential applications in chemical sensing, including medical diagnostics such as breath analysis, pollution monitoring, and remote detection of toxic chemicals and explosives. In this chapter, we demonstrate the usefulness of arrays of DFB-QCLs for absorption spectroscopy with fluids. Condensed phase materials such as fluids can have significant absorption, and typically have broadened absorption features that are well-suited to measurements using a wide-coverage DFB-QCL array source. Our results here were published in Lee et al. [2007].

56

Chapter 4: DFB-QCL arrays for mid-infrared spectroscopy

4.1

57

DFB-QCL array source

We connect our DFB-QCL array up to a custom-built microelectronic controller. The custom-designed controller consists of pulse generators to power the lasers, direct current bias circuitry to heat individual lasers in the array for temperature tuning, and a serial port interface for computer control of laser firing. This is shown in Fig. 4.1a. In Fig. 4.1b, we have a photo of the controller, with the DFB-QCL array chip attached.

4.2

Absorption spectroscopy

We perform absorption spectroscopy by firing the array of lasers one-by-one through an analyte, and looking at the transmitted signal intensity, as compared to a reference case without the analyte. The absorption at any frequency can then be calculated by comparing the two measurements. The 3.5 mm long DFB-QCL array described in the previous chapter was used, since all of its lasers operated single-mode at the designed frequencies. Our setup consisted of the QCL array source, a transparent BaF2 fluid cell (23.6 µm chamber thickness) containing the analyte, and a mercury cadmium telluride (MCT) liquid-nitrogen cooled detector (Fig. 4.2, inset). A single lens, 12 mm in diameter and with a 12 mm focal length, was used to image the 2.5-mm-wide QCL array onto the 0.25 mm by 0.25 mm active area of the detector. To take a spectrum, the lasers were fired sequentially, and the intensities of the transmitted beams were recovered from the detector using a gated integrator. One of the lasers was damaged


58

Controller Board 

(a)

Pulse Generators &  Biasing Circuitry  Laser Array 

Computer/  Laptop 

USB 

DSP 

(b) 20 cm

Figure 4.1: (a) Schematic of the widely tunable quantum cascade laser source with a distributed feedback laser array driven by a custom microelectronic controller. The dotted line denotes the routing of the current pulse and the DC bias to a specific laser (second from top) to fire it at a specified wavelength. (b) The microelectronic controller, as built, with the DFB-QCL array chip located in the bottom centre of the image.


59

during handling, so only 31 out of the 32 lasers were used. After taking the background and sample spectra, we obtained the absorption spectrum using a frequency table with data for each laser in the array (Fig. 4.2). The spectra took less than 10 s to obtain using the DFB-QCL array. The present limitation on speed is due to the fastest repetition rate (100 kHz) achievable using electronics we custom-built for the array, and also due to the delay in transmitting both control instructions and data over a slow serial connection between the electronics and our lab computer. With faster repetition rates and a higher data-rate connection, sampling can occur much quicker and the measurement time could be reduced to milliseconds. Our results compare favorably with spectra obtained using a conventional Fourier-Transform Infrared (FTIR) spectrometer, also shown in Fig. 4.2. In order to obtain continuous spectral coverage between the nominal emission frequencies of the individual lasers in the array, one can tune the lasers in a small range. This is done by temperature tuning the lasers can either be heated locally by applying a sub-threshold DC current to tune an individual laser, or the lasers can be heated globally by changing the temperature of the heatsink on which the laser array chip sits. Temperature tuning using both of these techniques is shown in Fig. 4.3. The local heating using DC current can achieve the desired temperature in milliseconds, while heatsink temperature changes typically take tens of seconds. Absorption spectroscopy using the DFB-QCL array with continuous coverage between the nominal emission frequencies of the individual lasers can then be performed. This is demonstrated in Fig. 4.4. Here the temperature of the laser array chip was adjusted globally by controlling the heatsink temperature using a thermoelectric cooler.


60

Figure 4.2: Absorption spectra of isopropanol (squares), acetone (circles) and methanol (triangles) obtained with the laser array and with a Bruker Vertex 80v Fourier-Transform Infrared spectrometer (continuous lines). (inset) Experimental setup for mid-infrared spectroscopy of liquids with the quantum cascade laser source.


(a)

61

(b)

Figure 4.3: (a) Temperature tuning achieved using a DC bias current to heat an individual DFB-QCL in the array. 5 cm 1 of tuning range is achieved by using 300 mA of bias current. (b) Temperature tuning achieved by heating the entire substrate of the DFB-QCL array by varying its heatsink temperature using heaters and a thermoelectric cooler. 5 cm 1 of tuning range is achieved by using 70 K di↵erence in temperature. The absorption spectra obtained with the DFB-QCL array in each case agree well with the results found using FTIR.

4.3

Discussion and summary

We note that the frequency resolution of our QCL source is determined by the lasers linewidths, which were measured to be ⇠0.1 nm (0.01 cm 1 ) in pulsed operation. This resolution is an order of magnitude higher than that provided by a standard ”bench top” commercial Fourier Transform Infrared Spectrometer (FTIR) and is roughly comparable to the linewidth of gas absorption features at atmospheric pressure. We note that, with careful stabilization of the current source and the device temperature, a QCL operating in the continuous-wave regime can achieve linewidth


62

Figure 4.4: Absorption spectrum of isopropanol taken using the DFB-QCL array source operated at di↵erent temperatures, in order to have a continuous measure of the spectrum (points). Data taken using a Bruker Vertex 80v FTIR shown for comparison (solid line).


63

smaller than 0.001 nm [Williams et al., 1999]. We intend to pursue this in our future work. In comparing QCL-based sensors to FTIR spectrometers for applications, it is useful to consider di↵erent factors such as resolution, spectral range, brightness, size, portability, and cost. We have already pointed out the significantly improved resolution over a FTIR. Although the spectral bandwidth provided by a broadband QCL is smaller than that of a FTIR spectrometer, this isnt an impediment for most sensing applications. Most of the important absorption lines occur in the 8-12 micron spectral range and QCLs with broadband gain that covers most this spectral range have already been demonstrated. In addition, we typically know what chemicals we want to monitor and we can design a QCL active region to have laser gain at the wavelengths where those chemicals absorb light. The much higher brightness of QCLs as compared to the thermal sources (glowbars) used in FTIR spectrometers should lead to substantial improvements in signal to noise. This is particularly true when using condensed-phase analytes such as fluids with high absorption. In terms of size and portability, there is no question that QCL-based sensors would be much more compact and portable than FTIR spectrometers. Because QCLs can be produced in the same foundries that produce diode lasers for the telecom industry, the cost of QCLs can potentially go down to the levels of laser diodes. Thus, QCL-based sensors can be quite inexpensive, particularly when compared to FTIR spectrometers which contain sophisticated mechanical and optomechanical components.


64

With all the benefits of QCL-based sensors, including their higher resolution, brightness, lower cost, size and portability, we hope that they will find applications as an alternative to FTIR spectrometers.

Chapter 5 Beam combining of QCL arrays For a number of applications envisioned for QCL arrays, it is important to have the beams from the individual lasers in the array propagate together, so that the beams overlap in the far-field. For example, for remote sensing applications, if the beams can be collimated and propagated a long distance where they all overlap, then only a single detector is required at the end of the beam path to measure the resulting signal. However, due to the spacing gap between lasers in the array, the beams will point at slightly di↵erent angles in the simple case where their light is collected and collimated by a lens. In that case, the angular dispersion will result in the beams from the array diverging from each other as they propagate any significant distance, and so the beams will be spatially separated in the far-field. In this chapter, we use spectral beam combining to overlap the beams from the lasers in the array, in the far-field. We then perform absorption spectroscopy at a distance separated from the laser source, to demonstrate a proof-of-principle application to remote sensing.

65

Chapter 5: Beam combining of QCL arrays

5.1

66

Introduction

The overall principle of spectral beam combining is to take spatially separated beams with distinct optical spectra, and combine them using a wavelength-sensitive beam combiner. Examples of wavelength-sensitive beam combiners are prisms and di↵raction gratings, which can deflect incident beams according to their wavelength, so that they propagate in the same direction after the combiner. In e↵ect, this is the reverse of the classic experiment with a prism, which takes a single beam of white light containing many wavelengths, and splits this light into angularly resolved monochromatic beams.

Figure 5.1: Beam combining can be a powerful technique, as evidenced by Star Wars!

Spectral beam combining of laser sources has been advanced by researchers at MIT Lincoln Laboratory, Lexington MA, USA, including beam-combining of diode laser arrays [Daneu et al., 2000; Huang et al., 2007], and fibre lasers [Augst et al., 2003]. In one form of spectral beam combining, the laser array elements are incorporated in an external cavity containing a di↵raction grating, and the cavity with grating both combines the beams and also provides optical feedback to the laser elements


67

that selects their emission wavelengths, as in the original experiment of Daneu et al. [2000]. This configuration is shown in Fig. 5.2.

Figure 5.2: Schematic diagram of spectral beam combining, where an external cavity containing a grating both combines the beams and also provides the optical feedback that selects the emission wavelengths of the lasers in the array. Here the front facet of the diode laser array is antireflection coated, and the laser resonator is formed for each array element between the high-reflectivity coated back facet of the laser and the partially reflecting (e.g. 10%) output coupler. The transform lens, located one focal distance away from the array, acts to transform the position of the laser element in the array into an angle of incidence on the grating. The flat output coupler forces the beams to co-propagate, since the beams must all be normal to this mirror in order to satisfy the feedback condition for the external cavity. The angles of incidence on the grating for the beams from all the array elements are di↵erent, so the external cavity will select di↵erent wavelengths for all the array elements as needed to enforce the co-propagation of the beams. In another form of spectral beam combining, the laser array elements have their


68

emission wavelengths selected separately from the grating that compensates for the angular dispersion of the beams and combines them. For example, a volume Bragg grating [Chann et al., 2006] or distributed feedback grating in the laser (i.e. the work presented here) can be used for wavelength selection.

5.2

DFB-QCL array

We present an array of distributed feedback quantum cascade lasers (DFB-QCLs) where we use spectral beam combining to co-propagate the beams, as shown in Fig. 5.3.

Figure 5.3: Schematic diagram of spectral beam combining with an array of distributed feedback quantum cascade lasers (DFB-QCLs). The emission wavelengths of the lasers are selected by the individual DFBs on each laser ridge in the array, and beam combining is accomplished by a suitably placed grating that compensates for the angular dispersion of the beams in the array.

Our DFB-QCL array is composed of 32 ridge lasers, emitting frequencies from 1061 to 1148 cm 1 , with the emission frequency of adjacent lasers separated by ⇠ 2.74 cm 1 . The laser ridges are each 15 µm wide and separated by a centre-to-centre


69

distance of 75 µm. The QCL active region for this array is a bound-to-continuum design for emission around 9 µm wavelength, as reported in Maulini et al. [2004], and the fabrication of the array is detailed in Lee et al. [2007]. The details of the design and fabrication of the DFB-QCL array can also be found in this thesis, in sections 2.2 and 3.1. The DFB-QCL array was connected to a custom-built electronics controller, which allows us to individually address and power each of the laser devices in the array. Out of the 32 laser ridges, only 28 were operational, with lasers #1, 21, 22 and 32 not emitting. The DFB-QCL array is oriented so that the lasers emit in a horizontal line. A 2.54 cm diameter ZnSe lens (f = 2.54 cm) was placed one focal distance away from the front facets of the DFB-QCL array, as a transform lens. To enable fine adjustment of the lens position, it was attached to an x-y-z stage. The lens position was adjusted to ensure that the individual laser beams were collimated, and that beams near the centre of the array were propagating on-axis; this was verified using a thermal IR camera to image the beam spots. A reflection grating with 750 lines/cm was inserted in the beam path after the transform lens, around 3 cm away from the lens. The grating is attached to a rotation stage, allowing it to be rotated in a plane parallel to the laser array. Ideally, the grating would be located one focal distance away from the lens, which would ensure that the beams would overlap at the grating. However, the size of the components and the need to ensure that the beam path remain unobstructed constrained the placement of the grating. The required angle for the grating to co-propagate all the beams can be deduced from the grating equation, 5.1:


d(sin✓m + sin✓n ) = m

70

n.

(5.1)

Here d is the groove spacing of the grating, ✓m is the output angle of the m-th di↵raction order, ✓n is the incident angle of the n-th laser beam on the grating, and n

is the wavelength of that laser. We have m=1, as our grating is blazed for high

efficiency in first di↵raction order. The incident angles ✓n of the lasers in the array are all di↵erent, with ✓n = ✓grating + atan(xn /flens ), where xn is the position of the n-th laser in the array and flens is the focal length of the transform lens. For all the beams to co-propagate, we require that all the lasers in the array have the same output angle ✓m from the grating. From equation 5.1 we see that this entails that the angle of the grating ✓grating should be ⇠ 55 degrees. We set the grating at this angle and then adjusted while monitoring the far-field beam spot, rotating the grating until the beams from the extreme ends of the array (lasers 2 and 31, since 1 and 32 were not working) were overlapped.

5.3

Near and far-field beam profiles

We used a thermal IR camera to get images of the beams coming o↵ the grating. A flat mirror was placed in the beam path, just after the grating. This mirror was used to steer the beams and also provided a convenient location to image the beam profile near the grating. We placed the IR camera in the path of the laser beams, and focussed the camera on the mirror surface to view the beam profile at that location. A representative image of one of the laser beams is shown in Fig. 5.4a. The beam is clipped by the edges of the transform lens, which is 2.54 cm in diameter. The


71

resulting circular beam becomes oval-shaped after di↵racting from the grating; the major axis is ⇠ 3.5 cm long and the minor axis is still 2.5 cm long. The laser array chip was indium-soldered to a copper submount, which acts as a heatsink and also as the electrical ground of the device. Unfortunately the mounting of the array chip was done in such a way that the chip is located slightly back from the edge of the submount, so that some of the light emitted from the lasers reflects o↵ the submount. This reflected light interferes with the direct emission from the laser facet, resulting in the horizontal interference fringes visible in the beam profile in Fig. 5.4a. The spacing of these fringes is consistent with interference from a reflection ⇠ 200 µm away from the laser facet, which is the thickness of the array chip substrate. In order to image the far-field beam profiles, we placed a 2.88 m diameter spherical mirror in the path of the beams. The spherical mirror was angled slightly so that the reflected, converging beam could be focussed onto the IR camera’s CCD array (without the camera’s imaging lens), which was placed in the focal plane (f = 1.44 m) of the mirror. In the focal plane of the mirror, the angular extent of the beams is resolved as horizontal and vertical shifts in position. By individually imaging all of the beams from the laser array, we can find the spot size of the beams (in angular units) and we can also quantify the overlap between the beams in the far-field. The far-field beam profile of a representative laser is shown in Fig. 5.4b. There is a characteristic Airy ring pattern, which is a result of the “top hat” profile of the beam emerging from the 2.54 cm diameter transform lens. Taking a linescan of the far-field beam profile (Fig. 5.4c), we can quantify the angular extent of the main lobe of the Airy pattern, from null to null, as 0.93 milliradians in the horizontal direction


72

(a)

(b)

intensity (arb. units)

(c)

-2

-1

0

1

2

angular spread (milliradians)

Figure 5.4: (a) Image of the beam of a representative laser, just after it has been reflected from the grating. The white bar is 1 cm. (b) Image of the far-field spot of a representative laser. The white bar is 1 milliradian. (c) Linescan in the horizontal (solid) and vertical (dotted line) directions of the far-field image of a representative laser.


73

and 1.3 milliradians in the vertical direction. For comparison, the di↵raction limited spot size at 9 µm wavelength for a beam collimated with a 2.54 cm diameter f/1 lens is ✓ ⇠ 2.44 /D, which is 0.86 milliradians. The likely reason that the beam diameter is larger than the di↵raction limit is that the beams are not perfectly collimated by the transform lens.

5.4

Beam overlap

The overlap of the beams in the far-field can be determined by imaging all the beams and overlaying those images to measure any shifts in position, which will show the residual angular dispersion. Fig. 5.5a shows a composite image of the beam spots from 4 di↵erent laser elements in the array (#18, 24, 28, 31). We measure the centre-to-centre distance between the beam spots. Lasers 18 and 31 have the largest relative angular spread in the entire array — 2 milliradians. All of the beam spots are collinear in the horizontal direction. Originally, there was also a vertical shift in the beam spots, which was due to a small tilt of the laser array relative to the plane of rotation of the grating. However, this e↵ect was compensated with fine adjustment of the tilt of the grating. Fig. 5.5b shows the angular spread of laser beam positions from the entire array. The angular deviation is measured relative to laser 31. The figure compares the experimental results to a calculation of the angular spread using the grating equation (Eq. 5.1) where we input the wavelengths of the DFB-QCL array and a grating angle of 54.65 degrees. There is good agreement between the results and the calculation. The residual angular divergence is due to the fact that the grating’s angular dispersion


74

angular deviation (milliradians)

(a)

1.5

(b)

1.0

0.5

0.0 1080

1100 1120 wavenumber (1/cm)

1140

Figure 5.5: (a) Image of several lasers showing the extent of the residual angular dispersion in the beams. From the left, we have laser elements #18, 24, 28, and 31 in the array. Lasers 18 and 31 have the largest relative angular spread in the entire array. The white bar is 1 milliradian. (b) A plot of the angular deviation of the laser beams, as a function of the laser frequency. Squares represent the angular spread of laser beam positions from the entire array, as measured relative to the position of laser 31 (rightmost point in the plot). The line is a calculation of the angular spread using the grating equation (Eq. 5.1) with the wavelengths of the DFB-QCL array and a grating angle of 54.65 degrees.


75

is not a linear function of frequency. We achieved beam combining with a residual angular divergence of 2 milliradians in the worst case for the beams. Without spectral beam combining, the angular divergence would have been 86 milliradians, so we have better than a factor of 40 improvement. In order to reduce the residual angular divergence even further, we could use a laser array where either the spacing of frequencies in the array or the physical spacing of the laser elements is not linear. In particular, the required spacing of the laser frequencies or laser element positions can be calculated using Eq. 5.1. The grating efficiency was measured to be 55%. This was gauged by measuring the power intensity of the di↵racted beam, and comparing that to the power of the laser beam, collimated with the 2.54 cm ZnSe lens, but without the di↵raction grating present. The measurement was done by focussing the beam with a BaF2 lens (f = 19 cm) onto an integrating sphere connected to a thermoelectrically-cooled Vigo MCT detector. The fact that QCLs are TM polarized limits the grating efficiency, as the electric field polarization is parallel to the grating lines, resulting in higher losses.

5.5

Remote sensing demonstration

To demonstrate the potential of spectrally beam-combined DFB-QCL arrays for remote sensing, we performed a simple absorption spectroscopy measurement at a distance of 6 m from the laser array. At a distance of 6 m, a BaF2 lens (f = 19 cm) was placed in front of a thermoelectrically-cooled Vigo MCT detector, to collect the laser light from the array onto the detector area. Since the beams were not perfectly overlapped in the far-eld, it was not possible to focus them all onto the


76

detector. Instead, we placed the lens slightly closer than its focal distance away from the detector, so that part of every beam from the array would fall on the detector area, giving sufficient signal for a measurement. A BaF2 fluid cell (chamber thickness 27.2 µm) was placed in the path of the beams between the lens and the detector. The fluid cell was filled with isopropanol for sample measurements, or left empty to measure the background. 3.0

absorbance

2.5 2.0 1.5 1.0 0.5

1060

1080

1100

1120

1140

wavenumber (1/cm)

Figure 5.6: Absorption spectrum of isopropanol measured using the spectrally beamcombined DFB-QCL array at a distance of 6 m (squares). Fourier transform infrared spectrometer measurement of the same sample using a Bruker Vertex 80v FTIR instrument (solid line).

To take a spectrum, the lasers were fired sequentially, and the intensities of the transmitted beams were recovered from the detector using a gated integrator. After taking the background and sample spectra, we obtained the absorption spectrum using a frequency table with data for each laser in the array (Fig. 5.6). The spectra took less than 10 s to obtain using the DFB-QCL array. The present limitation on speed is due to the fastest repetition rate (100 kHz) achievable using electronics


77

we custom-built for the array, and also due to the delay in transmitting both control instructions and data over a slow serial connection between the electronics and our lab computer. With faster repetition rates and a higher data-rate connection, sampling can occur much quicker and the measurement time could be reduced to milliseconds. Our results compare favorably with spectra obtained using a conventional Fouriertransform infrared (FTIR) spectrometer, also shown in Fig. 5.6. We note that the noise in the measurement is dominated by electromagnetic pickup in our custombuilt electronics, rather than by any fundamental sources of noise in the lasers or the amount of laser signal available.

5.6

Summary

Spectral beam-combining was used to co-propagate beams from a 32-element DFB-QCL array. The residual angular divergence was less than 2 milliradians, which is a factor of 40 better than without spectral beam combining. To demonstrate the applicability of spectrally beam-combined DFB-QCL arrays for remote sensing, we obtained the absorption spectrum of isopropanol at a distance of 6 m from the laser array. In the future, we envision using spectrally beam-combined QCL arrays for remote sensing at distances of tens and hundreds of meters. We will also achieve fully overlapped beams in the far-eld, by choosing appropriate laser frequencies so that we eliminate the residual angular divergence observed in this paper. Finally, we hope to make laser arrays with higher power output, for instance watt-level peak power from each laser in the array.

Chapter 6 Ultra-broadband DFB-QCL array In this chapter, we investigate DFB-QCL arrays using a heterogeneous cascade structure to further increase the coverage and tuning range. A heterogeneous cascade based on two bound-to-continuum designs centered at 8.4 and 9.6 µm was used to fabricate an ultra-broadband DFB-QCL array. This array emitted in a range over 220 cm

1

near 9 µm wavelength, operated pulsed at room temperature. The output power

of the array varied between 100 and 1100 mW peak intensity. In both the extent of frequency coverage and the level of output power, this is a significant improvement from our first arrays.

6.1

Heterogeneous cascade

In our previous work, we demonstrated a proof-of-concept DFB-QCL array, which achieved single-mode lasing coverage of 85 cm

1

near 9 µm wavelength [Lee et al.,

2007]. We used a bound-to-continuum active region design, which has inherently

78

Chapter 6: Ultra-broadband DFB-QCL array

79

broad gain with full-width half-max of nearly 300 cm 1 , as demonstrated by Maulini et al. [2004]. We did not attempt to cover the entire range of the gain spectrum with DFB-QCLs in our first attempt. Since it is advantageous for chemical sensing to have broadband single-mode laser sources, we are interested in increasing the range of frequencies that can be covered by a single DFB-QCL array. Thus we now choose to employ an active region design with an even broader gain spectrum, and we aim to space the design frequencies of our DFB-QCLs so that they span that gain spectrum. An ultra-broadband laser was first demonstrated by Gmachl et al. [2002a], accomplished by using a heterogeneous cascade to support lasing over a broad range of frequencies. Maulini et al. [2006] employed a heterogenous cascade of two boundto-continuum designs (with individual gain maxima at 8.4 and 9.6 µm) to achieve a gain spectrum with a record FWHM of 350 cm 1 . Using this “two-stack” boundto-continuum QCL, they obtained a single-mode tuning range of 265 cm

1

in an

external-cavity setup. Here, we will employ the same two-stack active region design to achieve an ultra-broadband DFB-QCL array. The QCL material used to fabricate the laser array was grown by metalorganic vapor phase epitaxy (MOVPE). The structure grown consists of a bottom waveguide cladding of 2 µm of InP doped 1⇥1017 cm 3 , followed by 200 nm of InGaAs doped 3⇥1016 cm 3 , a first active region which is 1.4 µm thick, 100 nm of InGaAs doped 3⇥1016 cm 3 , a second active region which is 1.3 µm thick, and 600 nm of InGaAs doped 3⇥1016 cm 3 . The first active region consists of 20 stages based on a boundto-continuum design emitting at ⇠ 9.6 µm. The second active region is 20 stages


80

of a bound-to-continuum design at 8.4 µm. The design of the two active regions is outlined in Maulini et al. [2006]. The structure was grown without an upper cladding, since it was anticipated that DFB gratings would be fabricated in the top 600 nm thick InGaAs layer. We processed mesa structures to measure the luminescence of the QCL material and to determine the gain spectrum. Circular mesa structures with 200 µm diameter were fabricated and then the samples were cleaved to split the mesas into semi-circles, with an exposed facet where the luminescence could be measured. The luminescence was measured pulsed at 80 kHz at room temperature, with long pulses corresponding to a duty cycle of 5% (Fig. 6.1). From the luminescence we can see that the gain spectrum has a full-width half-maximum of over 350 cm 1 , centered at around 1150 cm 1 .

Figure 6.1: Electroluminescence of a mesa structure fabricated from the QCL wafer, showing the gain spectrum of the material. The mesa was pumped pulsed at 80kHz at room temperature, with a 5% duty cycle. Voltages of 9V, 12.8V, and 14.8V were applied (ascending order of traces).


6.2

81

DFB-QCL array

We chose to design an array of DFB-QCLs to cover the spectral range from 1000 to 1310 cm 1 , with laser emission frequencies spaced regularly by 10 cm 1 . This was done to cover the central portion of the measured luminescence spectrum. By simulating the waveguide structure of the QCL, we obtained an e↵ective refractive index of 3.19 for the TM00 mode. Assuming this e↵ective refractive index, we chose the grating periods to range between 1.197 and 1.567 µm, corresponding to the desired emission frequencies between 1000 and 1310 cm 1 .

Figure 6.2: SEM micrograph showing a cross-section of the device, which has been cut along the laser ridge. The grating corrugation can be seen as the rectangular wave near the top of the image, and the two active regions are below, with a thin InGaAs spacer between them.

Device processing started with the fabrication of an array of 32 buried DFB gratings in the QCL material. To fabricate the buried gratings, a 200 nm thick layer of


82

Si3 N4 was deposited on top of the exposed top InGaAs by chemical vapor deposition. First-order Bragg gratings (with periodicities as mentioned above) were exposed onto PMMA using an Elionix ELS-7000 100-keV electron beam lithography system. This pattern was transferred into the Si3 N4 by using a CF4 -based dry-etch. The gratings were then etched 590 nm deep into the InGaAs layer with a HBr/BCl3 /Ar/CH4 plasma in an inductively coupled plasma reactive ion etching machine (ICP-RIE) (Fig. 6.2). A top cladding consisting of 2 µm of InP doped 5⇥1016 cm nm of InP doped 5⇥1018 cm

3

3

and 500

was re-grown over the gratings using MOVPE. The

grating coupling strength per unit length was calculated using coupled wave theory [Kogelnik and Shank, 1972] to be  ⇠ 42 cm 1 . Laser ridges, 15 µm wide and spaced 75 µm apart, were defined on top of the buried gratings by dry-etching the surrounding areas 9 µm deep with a HBr/BCl3 /Ar/CH4 plasma using ICP-RIE. During this step, the back facet of the lasers was also defined. The bottom and the sidewalls of the laser ridges were insulated by Si3 N4 and a 600nm-thick gold top contact was deposited. The samples were then thinned to 200 µm and a metal bottom contact was deposited. Finally, the front facets of the lasers were defined by cleaving to obtain 2-mm-long lasers and the laser array was indiumsoldered onto a copper block for testing. The facets were left uncoated. The entire DFB laser array chip is only 4 mm by 5 mm in size.

6.3

Device performance

All 32 lasers in the array were individually tested using a probe station at room temperature. The lasers were tested in pulsed mode with 50 ns pulses at a repetition


83

rate of 80 kHz. Lasing was observed from 24 out of the 32 lasers, with all operating devices being single-mode (Fig. 6.3). The lasing frequencies of the QCLs in the array were spaced ⇠ 9.5 cm

1

apart and spanned from 1020 to 1240 cm

1

(8.1 to

9.8 µm wavelength). All the DFB lasers remained single-mode with > 20 dB sidemode suppression up to at least 2.0 A in current (Fig. 6.3 inset). Sub-threshold measurements revealed the bandgap of the DFB grating to be 4.5 cm 1 , giving a coupling strength per unit length of  ⇡ 45 cm 1 , in agreement with our calculations. With the lasers being 2 mm long, the DFBs are strongly over-coupled with L ⇠ 9. All the lasers were lasing on the same side of the DFB’s photonic gap, specifically in the DFB mode that is on the high-frequency side of the gap. We attribute the good single-mode selection — which refers to the selection of the same DFB mode for all the lasers in the array — to the fact that the lasers are strongly over-coupled and contain a small amount of loss-coupling (⇠ 0.05 cm 1 ). The topic of single-mode selection for DFB-QCLs is discussed in detail in section 2.3.2. Lasing was not observed from lasers at the extreme ends of the array, likely because there was insufficient gain at those frequencies, which are towards the edges of the gain spectrum. We can see this by examining the threshold currents for the DFB lasers in the array, which ranged from 1.3 to 1.6 A, corresponding to current densities of 4.3 to 5.3 kA/cm2 . Plotting the threshold current required for lasing, we see that there is a marked increase in the thresholds for lasers near the edges of the array (Fig. 6.4). Next we examine the output power and I-V characteristics of devices in the array, shown in Fig. 6.5. There were small di↵erences in the I-V curves of the lasers resulting


84

Figure 6.3: Spectra of 24 single-mode distributed feedback lasers in the array. Laser frequencies are spaced ⇠ 9.5 cm 1 apart and span a range of ⇠ 220 cm 1 . (inset) Spectrum of a representative laser in the array on a log scale, showing side-mode suppression > 20 dB.


85

threshold current (A)

1.60 1.55 1.50 1.45 1.40 1.35 1.30 1050

1100

1150

1200

wavenumber (1/cm)

Figure 6.4: Plot of the threshold current required for lasing, as a function of the laser frequency. from the varying series resistances of di↵erent length gold contacts running between the laser ridges and the wirebond pads on the QCL array chip. We also observed a large variation in the slope efficiency of the lasers between 250 mW/A and 2500 mW/A. The variation in slope efficiency is not correlated with the scatter in threshold current density, and is likely due to the random variation in the position of the end mirror facets relative to the laser ridge gratings. As shown in Streifer et al. [1975], the variation in the position of the laser facet alters the distribution of light intensity within the laser cavity, which results in a variation in the amount of light emitted from a facet. This e↵ect for arrays of distributed feedback quantum cascade lasers is discussed in detail in section 2.3.1.


86

Figure 6.5: Plot of the voltage (left axis) and light output (right axis) characteristics of several representative lasers in the array as functions of current.

6.4

Summary

In conclusion, we have demonstrated a broadband DFB-QCL array covering 220 cm

1

around 9 µm wavelength; this range is nearly 20% of the center frequency.

The broad coverage was made possible by using a gain element with a heterogeneous cascade providing a wide gain spectrum with FWHM > 350 cm 1 . We achieved a significant increase in coverage from our previous results [Lee et al., 2007] where we demonstrated an array spanning ⇠ 85 cm 1 . We envision that broadband DFBQCL arrays will be used in spectroscopic applications where it is desirable to monitor di↵erent widely-spaced spectroscopic features. In the future, we hope to integrate DFB-QCL arrays into a variety of spectroscopic devices, including portable midinfrared spectrometers.

Chapter 7 Microfluidic tuning of DFB-QCLs In this chapter, we report the tuning of the emission wavelength of a single mode distributed feedback quantum cascade laser by controlled changes of the mode effective refractive index using fluids. A robust fabrication procedure to encapsulate the devices in polymers for microfluidic delivery is presented. The integration of microfluidics with semiconductor laser (optofluidics) is promising for new compact and portable lab-on-a-chip applications. These results were originally reported by the author and his collaborators in [Diehl et al., 2006b].

7.1

Introduction

So far, QCLs have been used essentially for trace gas sensing in combination with multi-pass or photo-acoustic cells [Kosterev and Tittel, 2002], allowing the detection of numerous gases at parts per million to parts per billions by volume levels. Alternative detection techniques include for example sensors based on hollow fibers

87

Chapter 7: Microfluidic tuning of DFB-QCLs

88

[Charlton et al., 2003]. Although it is very promising, only little work has been done to develop QCL-based techniques enabling the quantification of specific compound levels in liquids. All experiments reported in the literature are so far direct absorption measurements, in which the high power of QCLs is used to increase the useful optical path length and hence the sensitivity of the detection method [Chen et al., 2005b; Lendl et al., 2000]. Note that liquids have recently been used to help dissipating the heat from QCLs operated in continuous wave at room temperature [Chen et al., 2005a]. In this initial work, we report a novel way to integrate electrically pumped QCLs with microfluidic channels and demonstrate wavelength tuning by varying the refractive index of liquid in direct contact with the laser.

7.2

Device fabrication and encapsulation

The devices used in this work are DFB-QCLs first reported in Hofstetter et al. [2000]. The core of the laser waveguide is composed of a 1.75 µm thick active region based on a so-called diagonal transition design. The top cladding layer consists of a low-doped In0.52 Ga0.47 As layer (Si, 1⇥1017 cm3 , thickness 2.2 µm) followed by a highly doped In0.52 Ga0.47 As cap layer (Si, 1⇥1019 cm3 , thickness 0.7 µm). The Bragg grating was first defined holographically and etched in the cap layer before 34 µm wide ridges were wet etched. A thin layer of insulating Si3 N4 was deposited on the sidewalls of the ridges. Electrical contacts were provided by narrow Ti/Au stripes evaporated at the edges of the ridges. The conductivity of the highly doped cap layer is high enough to ensure an e↵ective and homogeneous current injection in the gain medium. As shown on the sketch presented in Fig. 7.1(a), the top of the ridge is


89

exposed to air, which allows a fraction of the mode to leak out in the air. Although the mode overlaps mostly with the semiconductor, replacing air by a liquid with a high refractive index n results in a relatively large tuning of the emission wavelength as described later in the text. In order to deliver liquids to the sensitive part of the QCLs in a convenient and controlled manner, we developed a reliable and robust encapsulation procedure for building microcavities around lasers or other photonic devices. Ordinarily, polydimethyl-siloxane (PDMS) is the standard material for use in microfluidic applications due to the comparative ease of its molding process [Whistesides et al., 2001]. However, using PDMS for the fluid channel led to serious problems: as the wire bonds delivering power to the laser were also embedded in the PDMS, they would detach from the laser immediately if any external stress were applied. Hence, we developed a new technique to build the laser cavity almost entirely out of SU-8 negative photoresist, so as to encase the wire bonds in a tough protective layer while still providing a leak proof device. The main steps of the fabrication process are illustrated in Figure 7.1. First the laser chip without encapsulation is mounted on a Cu-submount, wirebonded and tested (Fig. 7.1(b)). A layer of Microchem SU-8 50 is then spun on (2500 rpm, 30 s) and soft-baked (65 C, 5 min, then 95 C, 20 min). Pre-coating with Microchem Omnicoat is optional (3000 rpm, 30 s, followed by baking at 200 C for 1min). The SU-8 is exposed using a mask that protects the desired cavity region (shown in the figure as a T-shape), a very short region at the front of the laser (as SU8 is absorbing at the wavelength output of our lasers), and finally the gold electrical contacts on either side of the laser.


90

Figure 7.1: (a) Schematic cross section of the processed DFB-laser. The current is injected laterally from two Ti/Au contacts. The Bragg grating is etched in the top layer composing the waveguide and is exposed to air/liquid. Diagrams showing (b) the laser as bonded and mounted on a Cu-heatsink, (c) the di↵erent parts entering into the fabrication of the liquid chamber and (d) a device after encapsulation.


91

The SU-8 is then cross-linked (65 C for 1 min, 95 C for 5 min) and developed with PGMEA to wash away the unexposed regions. Using a microsyringe, Shipley S1818 positive photoresist is deposited in the cavity region. Surface tension keeps the photoresist from spilling over beyond the cavity edges. In addition, the same procedure is used to protect the front of the laser device. The S1818 need not be fully baked: only a thin protective skin has to form at its surface, and therefore the standard procedure of baking may be followed (105 C, 5 min). In the next step, another layer of SU-8 50 is deposited, exposed with the exception that the mask no longer protects the cavity region and developed. This results in a hard covering forming over the photoresist-protected area. Degassed PDMS is then poured over the entire laser to form a layer several millimeters thick which is baked to a hard cure (65 C, 3 hours). PDMS actually sticks weakly to SU-8 and is therefore peeled o↵ the device before applying silicone potting cement (e.g., Loctite 5140) on the SU8. The PDMS layer is then placed back on the laser and allowed to sit for 24 hours as the cement dries. Afterwards, the PDMS may be trimmed with a razor to reveal the electrodes and the front laser facet. Holes are drilled in the PDMS over the fluid cavity all the way down to the copper substrate. PGMEA is subsequently used to dissolve entirely the protective photoresist regions. Finally, tubing may be inserted into the holes thus created, and fluid may be pumped into the fluid cavity for device testing. A finished device is shown in Fig. 7.2.


92

Figure 7.2: An image of an encapsulated laser device. The laser facet is at the front/bottom. The laser ridge sits inside an empty micro-chamber where the fluid can be introduced, via one of the two attached tubes.

7.3

Experimental results

After encapsulation, the lasers were tested in pulsed mode (5 kHz repetition rate, 25 ns pulses) at 300 K. The light output from the devices was sent either onto a calibrated thermopile detector for power measurements or into a Fourier-transform infrared spectrometer (Nicolet 860) equipped with a deuterated triglycine sulphate detector for spectral characterization. Figure 7.3 shows the optical spectra obtained with one of several encapsulated devices. The laser is clearly single mode close to threshold (Ithreshold = 3.1 A), and as the current increases, a second mode appears. The separation between the two modes is 7.1 cm

1

and corresponds to a higher order

lateral mode due to the large width of the laser ridges. The current was kept at a constant value close to threshold (4.1 A) during the measurements with fluids, in


93

order to ensure single mode operation of the laser.

Figure 7.3: Optical spectra obtained with an encapsulated device without liquid at room temperature and with di↵erent current levels.

Fig. 7.4 (a) shows the optical spectra obtained with fluids with di↵erent refractive index (immersion oils made by the company Cargilles) in the SU-8 chamber. The measurements were performed in pulsed mode (repetition rate 5 kHz, pulse length 25 ns) at 300 K at a constant current of 4.1 A. For simplicity, the liquids were delivered to the device using microsyringes attached to the tubing inserted in the PDMS. Before replacing a fluid with another one, the SU-8 chamber was cleaned thoroughly by flushing methanol and then dry air through it several times. A continuous and reproducible red shift of the wavelength was observed as the refractive index of the fluids was increased. Note that the spectral linewidth is of the order of 0.28 cm 1 , which corresponds to the resolution limit of our experimental set-up. Even in the presence of liquid, the optical spectrum remained clearly single mode, with a side-


94

mode suppression ratio close to or better than 20 dB as shown in Fig. 7.4 (b). The maximum wavelength tuning was obtained with the fluid having the largest refractive index and corresponds to 1.15 cm 1 . It is close to the value found before encapsulation of the device (0.8 cm 1 ), i.e. when a drop of the same liquid was directly deposited onto the top of the laser with a micro-pipette. The results are not a↵ect by the fact that the SU-8 chamber does not cover the entire DFB grating. The graph in Fig. 7.5 displays the experimentally observed wavelength shift as function of the refractive index and compares it with the one obtained using a two-dimensional finite di↵erence time domain (FDTD) program. A good agreement is found between the results of our simulations (

/ n=10.7 nm) and the experiment (linear fit with

/ n=11 nm) when the data point corresponding to the liquid n=1.735 is not taken into account. The latter is clearly an outlier for reasons that are not fully understood yet. The refractive index of methanol was deduced using the above experimental linear coefficient. The value found corresponds to 1.29, in good agreement with the value found in Kim and Su [2004] (n=1.32). Figure 7.6 (a) shows the voltage and light intensity vs. current curves (V-I and LI) obtained with a second encapsulated laser. Its spectral properties are quantitavely and qualitatively similar to those described in Fig. 7.3, 7.4 and 7.5. These V-I and L-I curves were taken with and without liquid in the SU-8 chamber and under the same conditions as used for the spectral characterization, i.e. in pulsed mode at 300 K. The optical power was measured with a calibrated thermopile detector. The V-I curves are not a↵ected by the liquids since they are not conductive. The same conclusions were reached when the devices were tested electrically with deionized


95

(a)

(b)

Figure 7.4: (a) Optical spectra obtained at a fixed current (4.1 A) at room temperature with di↵erent liquids in the fluid chamber. The refractive index of the fluids varied from 1.3 to 1.735. (b) Optical spectra shown on a log scale measured without fluid, with the liquid with n = 1.53 and methanol.


96

Figure 7.5: Comparison between the peak position obtained experimentally from the data shown in Fig. 7.4 (a) and the results of FDTD simulations. The result obtained with methanol in the fluid chamber is also displayed. water and methanol. On the contrary, the L-I curves are clearly modified as seen on Fig. 7.6 (a). The threshold current density increases consistently when liquids with increasing refractive index were pumped into the SU-8 chamber. This change cannot be explained by an absorption in the liquid as the most transparent fluid used actually had the highest refractive index. The increase in threshold current density can be explained by the fact that the laser backfacet is immersed in the liquid, resulting in the lowering of the reflectivity of the mirror facet and thus an increase of the mirror losses. A change of threshold current density obtained by varying the mirror losses can be used to determine the waveguide losses of the laser [Sirtori et al., 1995]. The waveguide losses deduced from the data presented in Fig. 7.6(a) are equal to 18.3 cm 1 , which is comparable to the calculated value (24.6 cm 1 ).


97

Mode hopping due to the presence of liquid was also observed, as shown in Fig. 7.6 (b). In that particular case, the emission wavelength shifted by 34 nm (3.3 cm 1 ) in a reproducible manner when the liquid with n=1.735 was covering the laser. Because of the reproducibility and the magnitude of the tuning measured, the control of mode hopping using fluids is potentially very interesting for future applications. The temperature tuning coefficient for the DFB QCL used is equal to

⌫/ T=

-0.063 cm 1 /K [Hofstetter et al., 2000]. The amount of tuning that we report in the present letter can be obtained simply by varying the heatsink temperature by less than 15 K. However, the tuning caused by the fluids is essentially limited by the very small (approximately 0.01%) fraction of the optical mode propagating in the liquid. The overlap can be improved greatly by changing the geometry of the laser, e.g. by reducing the width of a ridge waveguide QC laser and leaving the sidewalls uncovered with metal. A QCL structure with a special cavity based on a photonic crystal has been proposed recently. The novelty of this design is the fact that the holes are placed in the plane perpendicular to the growth axis. The tuning range of this device is expected to be close to

7.4

⌫/ n=100 cm

1

[Loncar et al., 2007].

Summary and future directions

In this initial work we have concentrated on demonstrating the on-chip integration of microfluidics with quantum cascade lasers by tuning single mode devices via changes of the e↵ective refractive index of the mode induced by fluids. Our future research however will concentrate on lasers with integrated microfluidic delivery of analytes in-situ to the laser cavity for new compact and portable lab-on-a chip appli-


98

(a)

(b)

Figure 7.6: (a) Voltage and output power vs. current curves obtained at room temperature with di↵erent liquids in the fluid chamber. Note that these data were obtained with an encapsulated DFB QCL di↵erent from the one used to produce the data shown in Fig. 7.3, 7.4 and 7.5. (b) Mode hopping observed at a constant current by simply immersing the Bragg grating in fluid (n = 1.735). The wavelength shift is reproducible when the procedure is repeated after evaporation of the liquid.


99

cations based on absorption of the mid-infrared laser radiation (IR spectroscopy on a chip) by the analyte. The DFB lasers used in this study allow liquid specimens to be directly placed within the laser cavity, where they form an integral part of the laser itself. This intra-cavity sensing allows the injected liquid analyte to directly influence the laser characteristics, which when combined with electrical injection should also allow in the future to perform detectorless IR spectroscopy by monitoring the current-voltage characteristics of the laser. When combined with microfluidic sample delivery, the compact size and electrical read-out should enable ir spectroscopy to be used in completely new ways and in new arenas outside of the laboratory.

Chapter 8 Conclusions In this thesis I have presented the development of distributed feedback quantum cascade laser arrays, and shown their applications to chemical sensing. Distributed feedback quantum cascade laser arrays were developed that operated as broadband sources of mid-infrared coherent radiation. We investigated the design, fabrication and characteristics of DFB-QCL arrays, particularly focussing on achieving good single-mode selection despite the e↵ect of feedback from end mirror reflections. Single-mode selection is defined as the ability to enforce lasing at a specific desired frequency among the many modes potentially supported by a DFB grating. We noted that the random position of the end facet mirrors relative to the grating affects both the single-mode selection and the slope efficiency (and thus output power) of the lasers. We tried various methods to minimize the impact of the end facet mirrors, including anti-reflection coatings and highly over-coupled DFBs (L

1).

We demonstrated a device with 32 single-mode lasers on a single chip, emitting in a range over 85 cm

1

near 9 µm wavelength, operated pulsed at room temperature.

100

Chapter 8: Conclusions

101

This array was over-coupled with L ⇠ 11. All the lasers operated at the design frequencies, with undesired side-modes suppressed by >20 dB. The output power of the array varied from 20 to 200 mW peak intensity. This array was made using a QCL active region based on a bound-to-continuum design to achieve broad gain. The DFB-QCL array can be continuously tunable, since the separation in nominal emission frequencies is small enough that we can use temperature tuning to span the frequency gaps between adjacent lasers in the array. QCLs are excellent laser sources for infrared spectroscopy, with many potential applications in chemical sensing, including medical diagnostics such as breath analysis, pollution monitoring, and remote detection of toxic chemicals and explosives. To demonstrate the capability of our DFB-QCL array, we used it to perform absorption spectroscopy on fluids. We connected the array to a custom-built microelectronic controller, consisting of pulse generators to power the lasers, direct current bias circuitry to heat individual lasers in the array for temperature tuning, and a serial port interface for computer control of laser firing. We performed absorption spectroscopy by firing the array of lasers one-by-one through an analyte, and looking at the transmitted signal intensity, as compared to a reference case without the analyte. The absorption at any frequency can then be calculated by comparing the two measurements. We obtained absorption spectra for several fluids, with results of comparable quality to conventional, commercial Fourier transform infrared spectrometers. Temperature tuning of the lasers was used to obtain continuous spectra. Spectral beam-combining is a technique that allows angularly divergent beams from a laser array to be overlapped and co-propagated. Achieving overlapped beams


102

at extended distances can be important for a number of applications envisioned for DFB-QCL arrays, particularly remote sensing. With spectral beam combining, we reduced the total angular divergence of our DFB-QCL array to less than 2 milliradians, which is 40 times better than without beam combining. Using the beam-combined array, we performed absorption spectroscopy at a distance of 6 m from the laser chip; this is just a taste of the possibilities for remote sensing. We investigated DFB-QCL arrays using a heterogeneous cascade structure to further increase the coverage and tuning range. A heterogeneous cascade based on two bound-to-continuum designs centered at 8.4 and 9.6 µm was used to fabricate an ultra-broadband DFB-QCL array. This array emitted in a range over 220 cm

1

near

9 µm wavelength, operated pulsed at room temperature. The output power of the array varied between 100 and 1100 mW peak intensity. In both the extent of frequency coverage and the level of output power, this is a significant improvement from our first arrays. A completely di↵erent method of using QCLs for chemical sensing is to place analytes directly in intimate contact with the laser mode in the cavity of the device. This kind of “intra-cavity” sensing can be extremely sensitive to small changes in the composition of analytes. To achieve this objective, we encapsulated a DFB-QCL in a microfluidic chamber. We injected various liquids into this chamber while operating the DFB-QCL. We demonstrated tuning of the emission frequency of the DFB-QCL, by using liquids of di↵erent refractive indices to change the e↵ective index of the laser mode. For the future, we envision that intra-cavity sensing could involve detecting the absorption of analytes through their e↵ect on the losses of the laser cavity. The


103

change in the cavity losses could be detected from variations in the power output or threshold current of the QCL device. The results in this thesis point towards DFB-QCL arrays as a versatile, broadband, tunable mid-infrared laser source. With all the benefits of QCL-based sensors, including their high resolution, brightness, lower cost, size and portability, we hope that they will find many applications in chemical sensing and spectroscopy. In particular, we imagine them as an alternative to FTIR spectrometers. For the future, an important challenge is to increase the reliability of single-mode selection of DFB-QCLs, particularly since this issue is compounded for DFB-QCL arrays. One method for selecting a single mode of emissions in a DFB-QCL is to introduce gain- or loss-coupling in the laser. Gain and loss-coupling are also useful for maintaining good single-mode selection when reflections from end facet mirrors are present. We are investigating a loss-coupled DFB design that includes free-carrier absorption in parts of the grating to selectively absorbed radiation for undesired modes of the laser. Other future work will include the development of continuouswave DFB-QCL arrays and more extensive exploration of the applications to chemical sensing and remote detection.

Bibliography SJ Augst, AK Goyal, RL Aggarwal, TY Fan, and A Sanchez. Wavelength beam combining of ytterbium fiber lasers. Optics Letters, 28:331–333, 2003. Y. Bai, S. R. Darvish, S. Slivken, W. Zhang, A. Evans, J. Nguyen, and M. Razeghi. Room temperature continuous wave operation of quantum cascade lasers with wattlevel optical power. Applied Physics Letters, 92:101105, 2008. M. Beck, D. Hofstetter, T. Aellen, J. Faist, U. Oesterle, M. Ilegems, E. Gini, and H. Melchior. Continuous-wave operation of a mid-infrared semiconductor laser at room-temperature. Science, 295:301, 2002. Mikhail A. Belkin, Federico Capasso, Feng Xie, Alexey Belyanin, Milan Fischer, Andreas Wittmann, and Jerome Faist. Room temperature terahertz quantum cascade laser source based on intracavity di↵erence-frequency generation. Applied Physics Letters, 92:201101, 2008a. Mikhail A. Belkin, Jonathan A. Fan, Sahand Hormoz, Federico Capasso, S. P. Khanna, M. Lachab, A. G. Davies, and E. H. Linfield. Terahertz quantum cascade lasers with copper metal-metal waveguides operating up to 178 k. Optics Express, 16:3242, 2008b. J Buus. Mode selectivity in DFB lasers with cleaved facets. Electronics Letters, 21: 179, 1985. F. Capasso, C. Gmachl, D.L. Sivco, and A.Y. Cho. Quantum cascade lasers. Physics Today, 55:34, 2002. B Chann, AK Goyal, TY Fan, A Sanchez-Rubio, BL Volodin, and VS Ban. Efficient, high-brightness wavelength-beam-combined commercial o↵-the-shelf diode stacks achieved by use of a wavelength-chirped volume Bragg grating. Optics Letters, 31: 1253–1255, 2006. C. Charlton, F. de Melas, A. Inberg, N. Croitoru, and B. Mizaiko↵. IEEE Proc. Optoelectron., 150:306, 2003.

104

Bibliography

105

J.Z. Chen, Z. Liu, Y.S. Rumala, D.L. Sivco, and C. Gmachl. Electronics Letters, 42: 534, 2005a. J.Z. Chen, Z. Liu, Y.S. Rumala, D.L. Sivco, and C. Gmachl. Optics Express, 13:5953, 2005b. V Daneu, A Sanchez, TY Fan, HK Choi, GW Turner, and CC Cook. Spectral beam combining of a broad-stripe diode laser array in an external cavity. Optics Letters, 25:405–407, 2000. L. Diehl, D. Bour, S. Corzine, J. Zhu, G. Hofler, M. Loncar, M. Troccoli, and F. Capasso. High-power quantum cascade lasers grown by low-pressure metalorganic vapor-phase epitaxy operating in continuous wave above 400 k. Applied Physics Letters, 88:201115, 2006a. L. Diehl, Benjamin G. Lee, P. Behroozi, M. Loncar, Mikhail A. Belkin, Federico Capasso, T. Aellen, D. Hofstetter, M. Beck, and J. Faist. Microfluidic tuning of distributed feedback quantum cascade lasers. Optics Express, 14:11660, 2006b. J. Faist, F. Capasso, D.L. Sivco, C. Sirtori, A.L. Hutchinson, and A.Y. Cho. Quantum cascade laser. Science, 264:553, 1994. J. Faist, F. Capasso, C. Sirtori, D. L. Sivco, A. L. Hutchinson, and A. Y. Cho. Vertical transition quantum cascade laser with bragg conned excited state. Applied Physics Letters, 66:538, 1995. J. Faist, C. Gmachl, F. Capasso, C. Sirtori, D. L. Sivco, J. N. Baillargeon, and A. Y. Cho. Distributed feedback quantum cascade lasers. Applied Physics Letters, 70: 2670, 1997. J. Faist, M. Beck, T. Aellen, and E. Gini. Quantum-cascade lasers based on a boundto-continuum transition. Applied Physics Letters, 78:147, 2001. Z. Feit, M. McDonald, R. J. Woods, V. Archambault, and P. Mak. Low threshold pbeusete/pbte separate connement buried heterostructure diode lasers. Applied Physics Letters, 68:738, 1996. C. Gmachl, J. Faist, J.N. Bailargeon, F. Capasso, C. Sirtori, D.L. Sivco, S.N.G. Chu, and A.Y. Cho. Complex-coupled quantum cascade distributed-feedback laser. IEEE Photon. Technol. Lett., 9:1090, 1997. C. Gmachl, D. L. Sivco, R. Colombelli, F. Capasso, and A. Y. Cho. Ultra-broadband semiconductor laser. Nature, 415:883, 2002a.

Bibliography

106

C. Gmachl, A. Straub, R. Colombelli, F. Capasso, D.L. Sivco, A.M. Sergent, and A.Y. Cho. Single-mode, tunable distributed feedback and multiple wavelength quantum cascade lasers. IEEE J. Quantum Electron., 38:569, 2002b. S Golka, C Pflugl, W Schrenk, and G Strasser. Quantum cascade lasers with lateral double-sided distributed feedback grating. Applied Physics Letters, 86, 2005. D. Hofstetter, T. Aellen, M. Beck, and J. Faist. IEEE Photon. Technol. Lett., 12: 1610, 2000. D Hofstetter, J Faist, M Beck, and U Oesterle. Surface-emitting 10.1 µm quantumcascade distributed feedback lasers. Applied Physics Letters, 75:3769, 1999. Robin K. Huang, Bien Chann, Leo J. Missaggia, Joseph P. Donnelly, Christopher T. Harris, George W. Turner, Anish K. Goyal, Tso Yee Fan, and Antonio SanchezRubio. High-brightness wavelength beam combined semiconductor laser diode arrays. IEEE Photonics Technology Letters, 19:209–211, 2007. R. F. Kazarinov and R. A. Suris. Possibility of the amplication of electromagnetic waves in a semiconductor superlattice. Sov. Phys. Semicond., 5:707, 1971. K. Kennedy, A. B. Krysa, J. S. Roberts, K. M. Groom, R. A. Hogg, D. G. Revin, L. R. Wilson, and J. W. Cockburn. High performance InP-based quantum cascade distributed feedback lasers with deeply etched lateral gratings. Applied Physics Letters, 89, 2006. C.B. Kim and C. B. Su. Meas. Sci. Technol., 15:1683, 2004. H. Kogelnik and C. V. Shank. Coupled-wave theory of distributed feedback lasers. J. Appl. Phys., 43:2327, 1972. R. Kohler, A. Tredicucci, F. Beltram, H. E. Beere, E. H. Lineld, A. G. Davies, D. A. Richie, R. C. Lotti, and F. Rossi. Terahertz semiconductor-heterostructure laser. Nature, 417:156, 2002. A. Kosterev and F. Tittel. Chemical sensors based on quantum cascade lasers. IEEE J. Quantum Electron., 38:582, 2002. A. Kosterev, G. Wysocki, Y. Bakhirkin, S. So, R. Lewicki, M. Fraser, F. Tittel, and R. F. Curl. Application of quantum cascade lasers to trace gas analysis. Applied Physics B, 90:165, 2008. Benjamin G. Lee, Mikhail A. Belkin, Ross Audet, Jim MacArthur, Laurent Diehl, Christian Pflugl, Douglas Oakley, David Chapman, Antonio Napoleone, David Bour, Scott Corzine, Gloria Hofler, Jerome Faist, and Federico Capasso. Widely tunable single-mode quantum cascade laser source for mid-infrared spectroscopy. Applied Physics Letters, 91:231101, 2007.

Bibliography

107

Benjamin G Lee, Mikhail Belkin, Christian Pflugl, Laurent Diehl, Haifei A Zhang, Ross M Audet, Jim MacArthur, David Bour, Scott Corzine, Gloria Hofler, and Federico Capasso. Distributed feedback quantum cascade laser arrays. IEEE J. Quantum Electron., accepted, 2008. B. Lendl, J. Frank, R. Schindler, A. Mller, M. Beck, and J. Faist. Anal. Chem., 72: 1645, 2000. M. Loncar, Benjamin G. Lee, L. Diehl, Mikhail A. Belkin, Federico Capasso, M. Giovannini, J. Faist, and E. Gini. Design and fabrication of photonic crystal quantum cascade lasers for optofluidics. Optics Express, 15:4499, 2007. G. Luo, C. Peng, H. Q. Le, S.S. Pei, H. Lee, W.Y. Hwang, B. Ishaug, and J. Zheng. Broadly wavelength-tunable external cavity, mid-infrared quantum cascade lasers. IEEE J. Quantum Electron., 38:486, 2002. G. P. Luo, C. Peng, H. Q. Le, S. S. Pei, W.Y. Hwang, B. Ishaug, J. Um, J. N. Baillargeon, and C.H. Lin. Grating-tuned external-cavity quantum-cascade semiconductor lasers. Applied Physics Letters, 78:2834, 2001. A. Lyakh, C. Pflugl, L. Diehl, Q. J. Wang, Federico Capasso, X. J. Wang, J. Y. Fan, T. Tanbun-Ek, R. Maulini, A. Tsekoun, R. Go, and C. Kumar N. Patel. 1.6 W high wall plug efficiency, continuous-wave room temperature quantum cascade laser emitting at 4.6 µm. Applied Physics Letters, 92:111110, 2008. R. Maulini, M. Beck, J. Faist, and E. Gini. Broadband tuning of external cavity bound-to-continuum quantum-cascade lasers. Applied Physics Letters, 84:1659, 2004. R. Maulini, D. A. Yarekha, J.M. Bulliard, M. Giovannini, J. Faist, and E. Gini. Continuous-wave operation of a broadly tunable thermoelectrically cooled external cavity quantum-cascade laser. Optics Letters, 30:2584, 2005. R. Maulini, A. Mohan, M. Giovannini, J. Faist, and E. Gini. External cavity quantumcascade lasers tunable from 8.2 to 10.4 µm using a gain element with a heterogeneous cascade. Applied Physics Letters, 88:201113, 2006. D.D. Nelson, J.H. Shorter, J.B. McManus, and M.S. Zahniser. Sub-part-per-billion detection of nitric oxide in air using a thermoelectrically cooled mid-infrared quantum cascade laser spectrometer. Appl. Phys. B, 75:343, 2002. A. K. Y. Ngai, S. T. Persijn, G. von Basum, and F. J. M. Harren. Automatically tunable continuous-wave optical parametric oscillator for high-resolution spectroscopy and sensitive trace-gas detection. Appl. Phys. B, 85:173, 2006.

Bibliography

108

K. Ohtani and H. Ohno. Inas/alsb quantum cascade lasers operating at 10 µm. Applied Physics Letters, 82:1003, 2003. C. K. N. Patel. Continuous-wave laser action on vibrational-rotational transitions of co2 . Phys. Rev., 136:A1187, 1964. C Pflugl, M Austerer, W Schrenk, S Golka, G Strasser, RP Green, LR Wilson, JW Cockburn, AB Krysa, and JS Roberts. Single-mode surface-emitting quantumcascade lasers. Applied Physics Letters, 86, 2005. M. Pushkarsky, A. Tsekoun, I. G. Dunayevskiy, R. Go, and C. K. N. Patel. Sub-partsper-billion level detection of no2 using room-temperature quantum cascade lasers. Proc. Natl. Acad. Sci. USA, 103:10846, 2006. D. G. Revin, L. R. Wilson, E. A. Zibik, R. P. Green, J. W. Cockburn, M. J. Steer, R. J. Airey, and M. Hopkinson. Ingaas/alassb quantum cascade lasers. Applied Physics Letters, 85:3992, 2004. J. S. Roberts, R. P. Green, L. R. Wilson, E. A. Zibik, D. G. Revin, J. W. Cockburn, and R. J. Airey. Quantum cascade lasers grown by metalorganic vapor phase epitaxy. Applied Physics Letters, 82:4221, 2003. S. Sanders, R. J. Lang, L. E. Myers, M. M. Fejer, and R. L. Byer. Broadly tunable midir radiation source based on di↵erence frequency mixing of high power wavelengthtunable laser diodes in bulk periodically poled linbo3 . Electron. Lett., 32:218, 1996. G. Scamarcio, F. Capasso, C. Sirtori, J. Faist, A. L. Hutchinson, D. L. Sivco, and A. Y. Cho. High-power infrared (8-micrometer wavelength) superlattice lasers. Science, 276:773, 1997. C. Sirtori, J. Faist, F. Capasso, D. Sivco, A. Hutchinson, and A. Cho. Applied Physics Letters, 66:3242, 1995. C. Sirtori, P. Kruck, S. Barbieri, P. Collot, J. Nagle, M. Beck, J. Faist, and U. Oesterle. Gaas/algaas quantum cascade lasers. Applied Physics Letters, 73:3486, 1998. W. Streifer, R. Burnham, and D. Scifres. E↵ect of external reflectors on longitudinal modes of distributed feedback lasers. IEEE J. Quantum Electron., 11:154, 1975. G. Totschnig, F. Winter, V. Pustogov, J. Faist, and A. Muller. Mid-infrared externalcavity quantum-cascade laser. Optics Letters, 27:1788, 2002. M. M. J. W. van Herpen, S. E. Bisson, A. K. Y. Ngai, and F. J. M. Harren. Combined wide pump tuning and high power of a continuous-wave, singly resonant optical parametric oscillator. Appl. Phys. B, 78:281, 2004.

Bibliography

109

M. E. Webber, M. Pushkarsky, and C. K. N. Patel. Optical detection of chemical warfare agents and toxic industrial chemicals: Simulation. J. Appl. Phys., 97: 113101, 2005. G.M. Whistesides, E. Ostuni, S. Takayama, X. Jiang, and D.E. Ingberg. Annu. Rev. Biomed. Eng., 73:335, 2001. R. M. Williams, J. F. Kelly, J. S. Hartman, S. W. Sharpe, M. S. Taubman, J. L. Hall, F. Capasso, C. Gmachl, D. L. Sivco, J. N. Baillargeon, and A. Y. Cho. Kilohertz linewidth from frequency stabilized mid-infrared quantum cascade lasers. Optics Letters, 24:1844, 1999. A. Wittmann, M. Giovannini, J. Faist, L. Hvozdara, S. Blaser, D. Hofstetter, and E. Gini. Room temperature, continuous wave operation of distributed feedback quantum cascade lasers with widely spaced operation frequencies. Applied Physics Letters, 89:141116, 2006. G. Wysocki, R. F. Curl, F. K. Tittel, R. Maulini, J. M. Bulliard, and J. Faist. Widely tunable mode-hop free external cavity quantum cascade laser for high resolution spectroscopic applications. Applied Physics B, 81:769, 2005. G. Wysocki, R. Lewicki, R. F. Curl, F. K. Tittel, L. Diehl, F. Capasso, M. Troccoli, G. Hofler, D. Bour, S. Corzine, R. Maulini, M. Giovannini, and J. Faist. Widely tunable mode-hop free external cavity quantum cascade lasers for high resolution spectroscopy and chemical sensing. Applied Physics B, 92:305, 2008. J. S. Yu, S. Slivken, A. Evans, L. Doris, and M. Razeghi. High-power continuous-wave operation of a 6 µm quantum-cascade laser at room temperature. Applied Physics Letters, 83:2503, 2003.

Appendix A List of Publications 1. Benjamin G Lee, Haifei A Zhang, Christian Pflugl, Laurent Diehl, Mikhail A Belkin, Milan Fischer, Andreas Wittmann, Jerome Faist, Federico Capasso. Broadband distributed feedback quantum cascade laser array operating from 8.1 to 9.8 microns. Manuscript in preparation (2008). 2. Benjamin G Lee, Jan Kansky, Anish Goyal, Christian Pflugl, Laurent Diehl, Mikhail Belkin, Antonio Sanchez, Federico Capasso. Beam-combining of quantum cascade laser arrays. Manuscript in preparation (2008). 3. Benjamin G Lee, Mikhail Belkin, Christian Pflugl, Laurent Diehl, Haifei A Zhang, Ross M Audet, Jim MacArthur, David Bour, Scott Corzine, Gloria Hofler, Federico Capasso. Distributed feedback quantum cascade laser arrays. Accepted in IEEE Journal of Quantum Electronics (2008). 4. Benjamin G Lee, Mikhail A Belkin, Ross Audet, Jim MacArthur, Laurent Diehl, Christian Pflugl, Douglas Oakley, David Chapman, Antonio Napoleone, David Bour, Scott Corzine, Gloria Hofler, Jerome Faist, Federico Capasso. Widely tunable single-mode quantum cascade laser source for mid-infrared spectroscopy. Applied Physics Letters 91, 231101 (2007). 5. Mikhail A Belkin, Marko Loncar, Benjamin G Lee, Christian Pflugl, Ross Audet, Laurent Diehl, Federico Capasso, David Bour, Scott Corzine, Gloria Hofler. Intra-cavity absorption spectroscopy with narrow-ridge microfluidic quantum cascade lasers. Optics Express 15, 11262 (2007). 6. Ross Audet, Mikhail A Belkin, Jonathan A Fan, Benjamin G Lee, Kai Lin, Federico Capasso, E. E. Narimanov, D. Bour, S. Corzine, J. Zhu, and G. Hofler. Single-mode laser action in quantum cascade lasers with spiral-shaped chaotic resonators. Applied Physics Letters 91, 131106 (2007).

110

Appendix A: List of Publications

111

7. M. Loncar, Benjamin G Lee, L. Diehl, Mikhail A Belkin, Federico Capasso, M. Giovannini and J. Faist, and E. Gini. Design and fabrication of photonic crystal quantum cascade lasers for optofluidics. Optics Express 15, 4499 (2007). 8. L. Diehl, Benjamin G Lee, P. Behroozi, M. Loncar, Mikhail A Belkin, Federico Capasso, T. Aellen, D. Hofstetter, M. Beck, J. Faist. Microfluidic tuning of distributed feedback quantum cascade lasers. Optics Express 14, 11660 (2006). 9. L. Diehl, D. Bour, S. Corzine, J. Zhu, G. Hofler, Benjamin G Lee, Christine Y Wang, M. Troccoli, and Federico Capasso. Pulsed- and continuous-mode operation at high temperature of strained quantum-cascade lasers grown by metalorganic vapor phase epitaxy. Applied Physics Letters 88, 41102 (2006). 10. M. Troccoli, S. Corzine, D. Bour, J. Zhu, O. Assayag, L. Diehl, Benjamin G Lee, G. Hofler, and Federico Capasso. Room temperature continuous-wave operation of quantum-cascade lasers grown by metal organic vapour phase epitaxy. Electronics Letters 41, 1059 (2005).

Appendix B List of Patents 1. Benjamin G Lee, Christian Pflugl, Laurent Diehl, Mikhail Belkin, Federico Capasso. Methods and Apparatus for Single-Mode Selection in Quantum Cascade Lasers. Patent submitted (2008). 2. Christian Pflugl, Benjamin G Lee, Laurent Diehl, Mikhail A Belkin, Federico Capasso, Thomas J Tague. Spectrometers Utilizing Mid-Infrared UltraBroadband and High Brightness Light Sources. Patent submitted (2008). 3. Mikhail Belkin, Benjamin G Lee, Ross M Audet, Jim MacArthur, Laurent Diehl, Christian Pflugl, Federico Capasso. Broadly Tunable Single-Mode Quantum Cascade Laser Sources and Sensors. Patent application #11/611819 (2006).

112

Appendix B: List of Patents

113