Comparative study of infrared photodetectors based on quantum wells (QWIPs) and quantum dots (QDIPs)

(1)

IDE0626,Technical Report,IDE2006

COMPARATIVE STUDY OF INFRARED PHOTODETECTORS BASED ON QUANTUM

WELLS (QWIPS) AND QUANTUM DOTS (QDIPS)

Master´s Thesis in Electrical Engineering Conny Hansson, Krishna Kishore Rachavula

School of Information Science, Computer and Electrical Engineering Halmstad University

(2)

(3)

Comparative study of infrared photodetectors based on quantum wells (QWIPs) and quantum dots (QDIPs)

Master´s Thesis in Electrical Engineering

School of Information Science, Computer and Electrical Engineering Halmstad University

Box 823, S-301 18 Halmstad, Sweden

1 2006

c 2006

(4)

Description of cover page picture: IR image of a face taken with a commercial QWIP camera.

(5)

Abbreviations

QD Quantum Dot QW Quantum Well

QDIP Quantum Dot Infrared Photodetector QWIP Quantum Well Infrared Photodetector MBE Molecular Beam Epitaxy

MOCVD Metal Organic Chemical Vapor Deposition MOVPD Metal Organic Vapor Phase Deposition CBE Chemical Beam Epitaxy

MOMBE Metal Organic Molecular Beam Epitaxy DWELL Dot-in-a-Well

FTIR Fourier Transform Infrared IR Infrared

W Wavenumber

SNR signal-to-noise ratio OPD Optical Path Difference ZPD Zero Path Difference

(6)

(7)

CONTENTS

1 Abstract

This master’s thesis deals with studies of lateral and vertical carrier transport Dot-in- a-Well (DWELL) Quantum Dot Infrared Photodetectors (QDIPs). During the project, devices have been developed and tested using a Fourier Transform Infrared (FTIR) spectrometer with the purpose to find the processes governing the flow of photocurrent in the different kinds of detectors, the dark current magnitude in the vertical Quantum Dot Infrared Photodetector (QDIP) and the Quantum Well Infrared Photodetector (QWIP) and the light polarization dependences for the vertical QDIP and the QWIP.

The lateral carrier transport DWELL QDIP was found to have poor conduction in the well mainly due to re-trapping of electrons in this region. The main process governing the flow of photocurrent for this type of device at 77K is photo-excitation from the Quantum Dot (QD)s to the excited state in the Quantum Well (QW) and further thermal excitation. If the electrons are mainly transported in the matrix or the well at 77K is presently not clear.

For the vertical carrier transport DWELL QDIP at 77K, the wavelength response could be tuned by altering the applied voltage. At higher voltages, the dominant process was found to be photo-excitation from the QDs to the excited state in the QW followed by thermal assisted tunneling into the GaAs-matrix. At lower voltages, photo-excitation from the QDs directly into the the GaAs-matrix was the predominant process. The dark current level in the vertical QDIPs was found to be 1.5 to 5 orders of magnitude smaller than for the QWIP measured at 77K. Furthermore, the QDIP was found to be close to polarization independent. As expected the QWIP had a reduced sensitivity to normal incident light. The existence of this signal was attributed to interface scattering of light inside the device.

(10)

(11)

CHAPTER 2. ACKNOWLEDGMENTS

2 Acknowledgments

This project has been made possible by the combined effort of a number of people. First we would like to thank our supervisor Prof. H˚akan Pettersson for giving us the opportunity to work in such an interesting area of physics. H˚akan has been a continuous source of help and support. His enthusiasm and love for physics is apparent and has pushed us to make, what we believe to be, some of our best work up to date. We would also like to thank Dr. Jan Andersson and Linda H¨oglund from Acreo AB for their participation in this project. Linda has been a great source of help and interesting discussions. Her seamless never-ending optimism and her impressive attitude to research have influenced us greatly throughout this project. We also gratefully acknowledge the support from Acreo AB for fabrication of all the devices investigated in this project.

Finally, we express our gratitude to the division of solid state physics at Lund University for the possibility to carry out the optical measurements in this project. It has been a true privilege to be able to work at such a state of the art research facility.

Conny Hansson Krishna Kishore Rachavula 2006-01-19

(12)

(13)

CHAPTER 3. INTRODUCTION

3 Introduction

An electromagnetic spectrum is the distribution of all electromagnetic waves arranged according to their wavelength or corresponding energy. It includes all types of radiation from γ- rays to radio waves which are used in different applications. The visible light is just a part of this spectrum which extends from 0.39µm to 0.77µm [1],as shown in the expanded scale of Fig 3.1.

Figure 3.1: Chart of Electromagnetic Spectrum

The spectrum shown in Fig 3.1 is represented in different wavelength bands. It can also be represented in the form of energy by using the following relation,

E(eV ) = h.ν = h.c

λ = 1.23984

λ(µm) (EQ 3.1)

The radiation with higher photon energy or shorter wavelength, like γ-rays, X-rays and ultraviolet radiation is usually harmful for a human eye, whereas the lower photon energy radiation or longer wavelength radiation, like infrared, microwaves and radio waves is normally less harmful.

The Infrared (IR) region is located in-between the visible and microwave spectral regions. This thesis focuses on the design and operation of photodetectors for the IR region.

(14)

Parts of the IR spectrum is less suitable for applications due to the absorption of radiation by water or CO₂ which is present in the atmosphere. Several wavelength bands are currently in use in e.g., telecommunication applications due to better transmission.

1. The near IR region (0.7-1.5µm) suffers from high atmospheric absorption.

2. The medium wavelength IR region (1.5-6µm) offers nearly 100% transmission between 3 to 5 µm.

3. The long wavelength IR (6-14µm) and far IR regions (14-40µm) offer excellent vision of all terrestrial objects. The LWIR region offers nearly 100% transmission between 8 to 12 µm.

(15)

CHAPTER 4. BACKGROUND AND RELATED WORK

4 Background and Related Work

First, we would like to present a general background on different types of detectors in use and introduce the importance of infrared detectors that are incorporated in various commercial applications. The basic working principle of a detector is to absorb photons of certain energy (or a particular wavelength) and produce an electrical response that can be amplified and converted into a human-readable form. The detectors for infrared or long wavelength radiation show some promising features in night vision, space surveillance, space exploration, remote monitoring of environment etc. Such types of detectors are known as infrared detectors, and were originally developed during the Second World War by the German military from a compound called lead sulphide (PbS) [2].

Due to atmospheric absorption of infrared radiation, there has been some limita- tions in the usefulness of the infrared detectors to either 3 to 5 µm band or 8 to 12µm bands. There has been extensive research on how to optimize the detector in these two bands. Over the past three decades, highly effective commercial detectors have evolved in the areas of spectrometry, protein analysis, fire detection systems, preventive maintenance, process control and astrophysical studies. Extensive efforts have been directed towards the environmental applications, such as pollution detection and medical applications for blood testing.

Two different types of infrared detectors can be identified, photon- and thermal detectors [2,3, 4].

1. Photon Detectors: In a photon detector, the infrared radiation is absorbed in two different optical excitation processes as illustrated in Fig 4.1. The spectral response of a photon detector depends both on the wavelength and on the power of the radiation striking it. Except for the detectors working in the near infrared range, cryogenic cooling is normally required to prevent the thermal generation of charge carriers. These cooling requirements have made the IR systems expensive and bulky in nature [5].

The class of photon detectors is divided into several types depending on the kind of interaction within the material. Of them, intrinsic, extrinsic and quantum well detectors are discussed in this report.

a). Intrinsic Detectors: In these detectors, photons with higher energy than the bandgap are converted to electron-hole pairs. Different mechanisms for extracting an output signal based on the generated electron-hole pairs lead to further classification into photoconductive and photovoltaic detectors.

• Photoconductive detector: In its simplest form, this detector is a piece of a semiconductor with two contacts. The electron-hole pairs create a change in the electrical conductivity of the material. The change in conductivity can e.g., be measured as a change in current when a constant voltage is applied over the detector.

• Photovoltaic detector: Detectors with a built-in electric field are usually referred to as photovoltaic or p-n junction photodiodes. When photons with higher energy than the bandgap energy impinge on the surface of the device, electron-hole pairs are created in the depletion region of the p-n junction. Due to the strong built-in electric

(16)

Figure 4.1: Optical excitation processes in a semiconductor: 1).intrinsic absorption and 2).extrinsic absorption.

field, the electron-hole pairs are separated in the space charge region generating a photocurrent. In contrast to a photoconductive detector, no external voltage source is required.

b). Extrinsic Detectors: When doping a semiconductor, an extra allowed state is introduced in the bandgap (refer Fig 4.1 for extrinsic absorption). This extra state can be located close to the conduction band. The electrons can be excited by infrared radiation from this extra state into the conduction band, where they contribute to an increased conductivity. As for the previously described photoconductive detector, an external voltage is needed.

Since the energy needed for exciting the electrons to the conduction band is small, the electrons can readily be thermally excited at elevated temperatures. If all the electrons are excited into the conduction band due to thermal excitation, none can be optically excited and no change occurs in the current due to illumination. Therefore, the device has to be cooled down to very low temperatures using e.g., liquid helium. This is the main drawback of extrinsic type detectors compared to intrinsic.

The theory and working principle of quantum well detectors is discussed in later chapters.Lucent Technologies Bell Laboratories pioneered quantum well array technology in the mid-80s. Acreo AB got involved in QWIP research and development in 1986. At this time a small group lead by Dr. Jan Andersson started studies on QWIPs and means to couple radiation into the QW-structure. One result of this research was a patented 2- dimensional grating coupler that increases the absorption quantum efficiency to typically 50%. In 1992 FLIR Systems, FMV, FOA and NUTEK started sponsoring a research and development project at ACREO AB aimed at developing QWIP detectors and start up production. In 1997 SAAB Technologies joined the list of sponsors. In 2000 Acreo AB started weekly delivery of 320*240 pixel detectors to FLIR Systems. Acreo AB is one of three companies who supply QWIP detectors for IR-cameras in volume. Acreo ABs partner FLIR Systems is with 26% market share by far the leader on the infrared image

(17)

CHAPTER 4. BACKGROUND AND RELATED WORK

equipment market. For further reading on the research and development Acreo AB has done concerning QWIPs see [6]

Figure 4.2: Spectral response for a photon- and a thermal detector.

2. Thermal Detectors: In a thermal detector, the detector element is suspended on two lags, which are connected to a heat sink that provides continuous constant temperature to the device [5]. When the infrared radiation strikes the surface of the device, an increase in the temperature occurs which determines the change in the output signal. The output signal is generated by temperature-dependent mechanisms such as thermoelectric effect, resistance and pyroelectric effect. The spectral response of a thermal detector is wavelength independent whereas a photon detector is dependent on the wavelength as shown in Fig 4.2. Thermal detectors have shown their capabilities in some of the commercial applications shown below.

• Bolometer: A bolometer consists of a thin resistive element constructed to have a very small thermal capacity to sustain a large variation in the resistance when infrared radiation strikes it. The radiant power produces heat within the material, which changes the resistance of the device. The change in resistance is similar to that of a photoconductor; however, the detection mechanisms are different.

(18)

Figure 4.3: Two dissimilar metals connected in series for a thermopile.

• Thermopile: A thermoelectric device (thermocouple or thermopile) is made up of two dissimilar conductors (metal or semiconductor) connected in series as shown in Fig 4.3. Two distinct junctions are formed, one is the reference junction which is kept at a constant reference temperature while the other junction varies in temperature when the infrared radiation strikes the absorber. The temperature difference between the two junctions generates an output voltage, ∆V. The thermoelectric effect is also called the Seebeck-effect.

(19)

CHAPTER 5. THEORY

5 Theory

5.1 Basics of semiconductor heterostructures

A semiconductor heterostructure is obtained by growing one semiconductor layer on top of another semiconductor. Several epitaxial growth techniques were introduced for growing heterostructures with layers as thin as a monolayer. Molecular Beam Epitaxy (MBE) was the technique used initially, and later on Metal Organic Chemical Vapor Deposition (MOCVD), Metal Organic Vapor Phase Deposition (MOVPD), Chemical Beam Epitaxy (CBE), Metal Organic Molecular Beam Epitaxy (MOMBE) techniques were developed. To achieve the epitaxial growth of a semiconductor heterostructure, certain selective materials must be chosen which have compatible crystal structure and lattice spacings forming a perfect continuation of the bottom material in terms of lattice matching [7]. Three different types of heterostructures can be formed [7] as illustrated in Figure 5.1:

Figure 5.1: Different types of semiconductor heterostructures

1. Type I: In this type, the total gap of the smaller bandgap material resides in-between the conduction and valence bands of the larger bandgap material.

2. Type II: In this type, one of the band offsets is larger than the difference between the two semiconductor bandgaps, but it is smaller than the largest bandgap.

3. Type III: In this type, one of the band offsets is larger than the bandgap of the higher bandgap material.

(20)

To obtain coherent (defect free) heterostructures of the above mentioned types, five distinct families of composite material systems have been developed.

The GaAs/Al_xGa_1−xAs composite material system is one of most important systems which don’t produce any drastic changes in the lattice constant by varying the value of x. Other important heterostructure systems are Ga_xIn_1−xAs_yP_1−y and Al_xIn_1−xAs on InP, In_xAs_1−xSb and Al_xGa_1−xSb on GaSb.

5.2 Semiconductor quantum wells

A quantum well is formed when a thin layer of a semiconductor material with lower bandgap is sandwiched in-between the two larger bandgap materials to obtain a potential step as shown in Figure 5.2. The formation of the well confines the motion of electrons in one direction (say z) while the electrons are free to move in the other two directions (x and y). When the width of the well is sufficiently small, the motion of electrons in the well is quantized in the growth direction and the energy levels in this direction become discrete [7, 8]. One of the basic quantum well devices in which we are interested is the QWIP.

Figure 5.2: Illustration of QW in physical- and band-structure.

5.2.1 Electronic structure of a quantum well

The electronic structure of nanostructures is of great interest since it gives the allowed energy levels which the electrons can occupy. The energy levels can be found by solving the Schr¨odinger equation (EQ 5.1) :

HΨ_n = E_nΨ_n (EQ 5.1)

(21)

CHAPTER 5. THEORY

where H is the Hamiltonian,representing the sum of kinetic and potential energy, ψ_nis the electron wave function and E_n is the corresponding energy. Assuming a one dimensional potential in the growth direction (z-direction) with V = 0 in the well and V = ∞ outside the well results in a free electron Hamiltonian inside the well:

H = p²

2m∗ (EQ 5.2)

Here p is the electron momentum and m* the effective electron mass. In quantum theory the momentum can be represented as:

p = −i~ d

dz (EQ 5.3)

When inserting (EQ 5.3) into (EQ 5.2) we find that the Schr¨odinger equation becomes:

HΨ_n= − ~² 2m∗

d²Ψ_n

dz² = E_nΨ_n (EQ 5.4)

Figure 5.3: Allowed energy states and wave functions in the growth direction of a QW

Since the barriers are infinite the boundary conditions for the wave function are:

ψn(0) = ψn(L) = 0 (EQ 5.5)

The following normalized solutions are obtained (illustrated in Figure 5.3):

ψn= r2

Lsin

nπ L z

(EQ 5.6) where n is an integer known as a quantum number and L is the width of the well. As evident from (EQ 5.6) only multiples of half wavelengths can exist inside the QW. By using the equation (EQ 5.6), and its second derivative, in (EQ 5.1):

− ~² 2m∗

d²Ψ_n

dz² = E_nΨ_n⇒ ~² 2m∗

nπ L

2

sinnπ L z

= E_nsinnπ L z

(EQ 5.7)

(22)

the energy levels E_n are given by:

En= ~² 2m∗

nπ L

2

= ~²

2m∗k_z² (EQ 5.8)

where k_z is the wavevector. It should be noted that the solutions given by (EQ 5.6) are only approximate. Since the potential outside the well in reality is finite and equal to the conduction band offset the wave function will penetrate into the barrier, resulting in shifts in the position of the energy levels. The analysis does however give us an explanation of which parameters are of interest. In the other two dimensions (y- and z-direction) of the well the electron levels are not quantized since there is no confinement. Solving the Schr¨odinger equation for free electrons gives a quasi-continuous energy spectrum:

E_n= ~²

2m∗ k_x²+ k²_y

(EQ 5.9) For the quantized growth direction the boundary conditions were given by the need for vanishing electron wave functions at the barriers. In the case of the two non-quantized directions, so called periodic boundary conditions are invoked. The reason for applying boundary conditions to free electrons is to be able to count the number of states and to determine the important quantity known as density of states D (E). Periodic boundary conditions can be visualized as electron waves traveling in a circle. In order for the wave to be able to travel round the circle a multiple of the electron wavelength has to be equal to the circumference L of the circle. Here L is taken as the in-plane size of the well. This implies that the distance between adjacent allowed k values will be ^2π_L. Taking this into account the number of allowed electron states can be expressed as:

n = 2 L 2π

2

πk²_xy = 2

A 4π²

πk_xy² (EQ 5.10)

Here, the factor _2π^L2

gives the number of states per unit volume in k-space accounting for both allowed dimensions, and πk_yz² is the area of the circle in k-space created by the allowed k-values for the two dimensions. It should be noted that the multiplication factor 2 in (EQ 5.10) is present to account for the degeneracy of the system, allowing each orbital to accept two electrons. The allowed energy levels in two dimensions can be written as:

E_n = ~²

2m∗k_xy² (EQ 5.11)

By combining (EQ 5.10) and (EQ 5.11) we can calculate the density of states for the two

”free” dimensions:

D (E) = dn

dE = mA

~²π = Const. (EQ 5.12)

It should be noted that the density of states for two dimensions without confinement is constant. Since the quantum well exists in three dimensions the energy structure of the QW is a combination of the 1D- quantized energy structure and the 2D- non-quantized energy structure. The resulting energy structure can be seen in Figure 5.4

(23)

CHAPTER 5. THEORY

Figure 5.4: a) Energy levels in the growth direction of a QW, b) Density of states in a QW, c) Energy to k-value relationship in a QW

5.4(a) shows the allowed energy states of the QW in the growth direction. In 5.4(b) is seen that even though the density of states in two dimensions is constant, the density of states for a quantum well is a step function with steps occurring at the energy of each quantized level. Figure 5.4(c) shows that the electrons can have a higher energy than the discrete quantum energy level shown in Figure 5.4(a). This is a result of the freedom of the electrons to move in the plane of the well.

5.3 Semiconductor quantum dots

A QD can be described as a small semiconductor box with dimensions less than 100 nm, which is incorporated into a semiconductor matrix of a different material. The great interest in QDs is partly due to the similarities between the QD and the atom. In fact, the QD is sometimes referred to as an artificial atom. The wavelike nature of electrons allows them to exist only in discrete states in an atom. In the atom, the electrons are attracted by the positive nucleus. Similarly, electrons in a QD are confined in all three dimensions. In this case, however, the confinement results from the potential pocket due to the small bandgap of the QD compared to the bandgap of the surrounding matrix. The electrons in the QDs are trapped in discrete states similarly to free atoms. The electrons trapped in a QD stem from doping of the semiconductor matrix. The main difference between a real atom and a QD is the size. The QD consists of a very large number of atoms. One driving force behind the development and extensive studies of the QDs has been the possibility of tuning quantum and many body effects by changing the size of the dots and the number of electrons in the dots. Changing the size of the dot varies the electronic structure inside the dot, while changing the number of electrons in the dot enables us to change the atomic like structure. For example, having one electron trapped in the dot gives us a ”hydrogen-like” system, two trapped electrons gives us a ”helium-like”

system and so on.

(24)

5.3.1 Electronic structure of a quantum dot

When analysing the electronic structure of a quantum dot one has to keep in mind that the electrons are confined in all three dimensions. The arguments leading to (EQ 5.8) in Section 5.2.1 are valid for all three dimensions. This results in the discrete energy spectrum and sharp density of states profile seen in Figure 5.5:

Figure 5.5: (a) Allowed electronic states in a QD (b)Density of state profile for a quantum dot

5.3.2 The Stranski-Krastanow growth method

One interesting way of forming QDs is a self-organising method called the Stranski- Krastanow growth method which has the great advantage of fabricating QDs with a high density and rather uniform shape and size. The method uses lattice mismatching to spontaneously create the QDs. Lattice mismatch means that the crystal structure of the existing semiconductor and that of the semiconductor being grown differ in lattice constant as can be seen in Figure 5.6.

Figure 5.6: Process of Stranski-Krastanow growth method

The method starts with the growth of a single atomic layer called the wetting layer. The growth is then continued on this wetting layer, inducing strain into the material

(25)

CHAPTER 5. THEORY

being grown as it deforms to fit the lattice of the substrate (material being grown upon).

This growth continues until a critical thickness is reached. As the growth continues, a metastable growth is in action and the wetting layer thickness becomes supercritical.

When growth continues more strain is induced into the wetting layer and when the energy due to the strain reaches the activation energy, dots are spontaneously formed. As the dots form, the strain in the material is reduced and the thickness of the wetting layer in-between the dots goes down to approximately the level of the critical thickness. The strain energy profile during the growth can be seen in Figure 5.7:

Figure 5.7: The strain energy profile during growth of dots.

(26)

5.3.2.1 Density and size of quantum dots

The control of the size and density of the dots is a topic that is not completely understood at present. It is known that increasing the material deposited on the substrate increases the density of dots. If the amount of material deposited on the substrate is kept constant there are three factors that strongly influence the density and size of the dots. These are:

-Temperature -Deposition rate

-The surface diffusion constant.

Once the wetting layer has passed the critical thickness and formation of dots has been made possible, it is the surface diffusion length that determines the distance between the dots. The surface diffusion length usually increases with increased temperature and decreased deposition rate, which also means that the distance between the dots increases in accordance with these parameter changes. In the case when only the deposited material is involved in the formation of dots (no diffusion of substrate material into the dots during formation), then the size of the dots and the density of the dots are inversely related to each other. For a fixed amount of deposited material, an increase in dot density leads to each dot having less material available, effectively limiting the maximum size of the dot.

5.4 The interaction of light with matter

The interaction of light with matter is governed by three basic mechanisms, as illustrated in the two level system in Figure 5.8.

Figure 5.8: Illustration of the three basic light-matter interaction mechanisms.

The first mechanism (labeled a) is absorption. The electron in the ground state absorbs a photon with the energy equal to the energy difference between the two allowed energy states, E1 and E2. This results in the electron being excited from state E1 to state E2.

For atoms, only photons with exactly the same energy as the energy difference between two states can be absorbed.

In the case of semiconductors, the energy structure consists of energy bands rather than discrete energy levels as can be seen in Figure 5.9. The energy bands are separated by a bandgap. A photon of slightly higher energy than the bandgap can then be absorbed which creates an electron-hole pair. Both the electron and the hole have higher energy than the corresponding bandedge. Since this state is not a stable state, the electron and hole swiftly relaxes to the bottom of the bands, dissipating the excess energy in the form

(27)

CHAPTER 5. THEORY

Figure 5.9: Illustration of electron-hole pair generation in a semiconductor due to absorption.

of phonons. The second mechanism (labeled b) is stimulated emission. In this case the electron is already positioned in the higher energy level. If a photon with energy equal to that of the energy gap interacts with the electron, the result will be that the electron recombines into the low energy level, while releasing one photon with the same phase and energy as the incoming photon. The energy of each photon will be equal to the energy difference E2-E1. The third mechanism (labeled c) is spontaneous emission. An electron in a higher energy level is not in a stable state. It will, therefore, after a short time recombine into the lower energy level giving off a photon with the energy equal to the energy difference, and in so doing, preserving the conservation of energy and momentum.

Spontaneous emission and stimulated emission processes take place in semiconductors invoking electrons in the conduction band and holes in the valence band.

(28)

(29)

CHAPTER 6. THE PROJECT

6 The project

6.1 Scope of project

This master’s thesis deals with studies of lateral and vertical carrier transport DWELL QDIPs and QWIPs. During the project, devices have been developed and tested using a FTIR spectrometer with the purpose to find the processes governing the flow of photo current in the three kinds of detectors, the dark current level, and the light polarization dependence for the vertical QDIP and the QWIP. The measurements were done at the department of solid state physics at Lund University and the devices were fabricated by ACREO AB.

6.2 Quantum Well Infrared Photodetectors (QWIPs)

In 1969, Esaki and Tsu [9], discovered novel electronic properties in devices where two semiconductor materials with different bandgaps were joined together forming a semiconductor heterostructure. This discovery led to the foundation of quantum well based devices. The potential landscape of the heterostructures varies from one layer to another depending on the bandgap of the different materials. By tremendous achievements in developing advanced growth techniques such as MBE, CBE and MOMBE, heterostructures can now be grown with a crystalline perfection at the interfaces. Hence, such structures have made it easy to form the basis for the development of quantum well devices [9].

A QWIP is a device designed to detect infrared or long wavelength radiation.

The basic layout of a QWIP can be seen in Figure 6.1.

Figure 6.1: Basic layout of QWIP. In a real device the number of AlGaAs/GaAs layers would be around 50.

The layout is a sandwich structure grown on a substrate. The potential profile of the QWIP is shown in Figure 6.2.

(30)

Figure 6.2: Electronic structure of unbiased QWIP

The low bandgap material acts as a trap for electrons. The electrons which are attracted to the low energy area are trapped inside the QW by the potential barriers created by the surrounding material. An electron stays in the well until it is excited by incoming photons, or thermally excited at elevated temperatures. When the energy of the electron is raised enough to be able to be excited over the energy barrier, the electron leaves the well and contributes to the current through the device. It should be noted that the QWIP is a vertical device, i.e. the current flow perpendicular to the plane of the quantum wells.

A lateral QWIP would only result in a short circuit, since the trapped electrons in the wells can move freely in the plane of the wells.

By varying the depth and width of the well, it is possible to change the position, and thus also the number of discrete energy states in the well. Absorption of long wavelength radiation in a quantum well is due to either a)bound to bound state transitions in a wide well or b)bound to continuum transitions in a narrow well. Figure 6.3 presents an overview of these two excitation mechanisms. In a wide well, photons with energy greater than or equal to the energy difference between the two states are absorbed. Thus when a photon strikes the quantum well, an electron in the ground state is excited to the first excited state thereafter an applied external voltage extracts these electrons to contribute to the flow of photocurrent. In a QWIP, the flow of photocurrent can be attributed to either tunneling or thermal excitation of carriers into the continuum as shown in Fig 6.3(a).

In QWIPs, one or more regions of the detector have to be doped. This has to be done in order to have free electrons (or holes) that can be trapped in the wells.

GaAs/Al_xGa_1−xAs is one of the most renowned material systems used for the detection of light in QWIPs at wavelengths longer than 6µm.

One of the parameters from which QWIPs can be judged is the dark current which is the current that flows in a biased detector in the absence of photons striking the well. To improve the performance of a QWIP, this dark current must be low.

(31)

Figure 6.3: (a) bound to bound state transition in a wide well under bias. (b) bound to continuum transition in a narrow well under bias.

There are several drawbacks of QWIPs:

From quantum mechanics, a set of selection rules is obtained suggesting that intersubband absorption is allowed when the incident light is polarized parallel to the growth direction. This creates a lot of difficulties in fabrication of a detector, since there is no absorption of light impinging perpendicular to the quantum well plane. In addition, large dark currents, and extremely fast recapture of photoexcited electrons are the main drawbacks of QWIPs. To overcome these problems, QDIPs have been developed and research is in progress to optimize these parameters [9, 7,8, 10].

(32)

6.3 Dark current in infrared photodetectors

Dark current is one of the most important parameters which limit the performance of an IR photodetector. It is the current which flows along a biased detector in the absence of IR radiation. In QWIPs, the dark current is the major limiting factor of the detector’s performance at higher temperatures. Hence, the detectors have to be cooled in the range of 50-77K to reduce the dark current. This raises the cost of a commercial photodetector.

To overcome such problems, QDIPs were introduced to reduce the effects of dark current.

In the case of a QW, electrons are quantized in the growth direction (say z) and free to move in the other two directions (x and y). The electrons moving in the x-y plane have a higher energy than the ground state (E_o) of the QW. Considering the case of a QD, the electrons are confined in all three dimensions and the ground state energy is lower compared to a QW. From this state, the electrons need more energy to escape out of the dot which reduces the dark current.

The three different dark current mechanisms illustrated in Fig 6.4 are predominant in an intersubband photodetector:

1. Ground state tunneling 2. Thermally assisted tunneling 3. Thermionic emission

Figure 6.4: Three dark current mechanisms in an intersubband photodetector:(1)ground state tunneling,(2)thermally assisted tunneling and(3)thermionic emission

Ground state tunneling occurs when the electron in the ground state move through the barrier into the neighboring well. To overcome this tunneling effect, thick barriers are incorporated in between two wells. The barrier width should be in the range of 250-500˚A units [11] to avoid tunneling. The thickness of the barriers must be smaller than the critical thickness for the actual material system; otherwise the sample quality will be poor.

Thermally assisted tunneling is a two-step process : at first, the electrons are thermally excited to a higher state thereafter they might tunnel out. Thermally assisted tunneling is considered as the major source of dark current [12]. To overcome this effect, the bias voltage (refer Fig 6.5(b)) and the tunneling barriers must be optimized.

(33)

When the electrons get thermally excited out of the well into the continuum, thermionic emission takes place. To calculate the actual dark current density flowing through a biased detector, one needs to find out the current density produced by the thermionic emission (j_emission) as well as the reduction in the photocurrent caused by the trapping of electrons (j_trapping) into the neighboring well. Under steady-state conditions, both current densities are assumed to be equal. The total dark current density can be calculated from the following relation:

J_Dark = j_emission

p (EQ 6.1)

Where p is the trapping probability for an electron traversing a potential well. The trapping probability can be expressed as:

p = τtransit

τ_transit+ τ_{lif e} (EQ 6.2)

The complete theoretical expression of the dark current for QDIPs is presented by V.Ryzhii et al., in [13].

Figure 6.5: Conduction band edge profile of an intersubband photodetector under zero bias(a) and finite bias(b)

(34)

6.4 Quantum Dot Infrared Photodetectors (QDIPs)

The basic design of the QDIP used in this project is the Dot-in-a-WELL (DWELL) structure, of which one type is schematically depicted in Figure 6.6.

Figure 6.6: Basic design of lateral DWELL-QDIP. In an actual device, multiple dot-in-well layers are inserted.

In contrast to a QWIP, the electrons are trapped into the Stranski-Krastanow grown dots incorporated into the wells. In an ideal device, no free charge carriers (electrons) would remain in the wells. The band structure of the device is shown in Figure 6.7.

Figure 6.7: Bandstructure of DWELL-QDIP

By introducing QDs we expect a much higher sensitivity to normal incident radiation compared to QWIPs, and also a lower dark current. The DWELL QDIP will be a lateral or vertical carrier transport device with the dots situated inside the QWs. The vertical carrier transport device will have the same surface contact geometry as the QWIP.

(35)

6.4.1 The lateral carrier transport DWELL QDIPs

The physical structure of the lateral carrier transport DWELL QDIP can be seen in Figure 6.6. In this device the electrons are excited from the dot into the ground state or excited state of the well where the transport is expected to take place due to a lateral (in plane) applied electric field as can be seen in Figure 6.8.

Figure 6.8: Electron transition in a lateral carrier transport DWELL QDIP

By varying the width or composition of the QW, the lower states are more effected than the higher states. This enables the longest wavelength for which a photocurrent occurs to be effectively tuned. One of the main drawbacks with the lateral design is that fabricating devices and incorporating them into a focal plane-array is very hard.

6.4.2 The vertical carrier transport DWELL QDIPs

The physical structure of the vertical carrier transport DWELL QDIP can be seen in Figure 6.9

Figure 6.9: The physical structure of the vertical carrier transport DWELL QDIP

(36)

In this device, the electrons are primarily excited from the ground state of the QD into the excited state of the QW whereafter they get transported away from the QW due to a vertically applied bias just as in a QWIP device. The process can be seen in Figure 6.10.

Figure 6.10: Electron transition for a vertical carrier transport DWELL QDIP

Since the higher energy states in the QW does not change so much with the width or composition of the QW, the onset energy in vertical carrier transport DWELL QDIPs can not be tuned as effectively as for the lateral carrier transport DWELL QDIPs. But the vertical carrier transport devices are on the other hand the norm of design for implemen- tation in focal plane-arrays. The advantage of vertical carrier transport DWELL QDIPs in comparison to QWIPs is a lower ground state energy due to the dot, which decreases the dark current. The dot also introduces polarization independence of the device.

(37)

CHAPTER 7. FOURIER TRANSFORM INFRARED SPECTROSCOPY

7 Fourier Transform Infrared Spectroscopy

7.1 Introduction to FTIR spectroscopy

A detailed understanding of the interaction of infrared radiation with matter is very important for realization of IR photodetectors. The FTIR spectrometer is the most reliable instrument for demanding spectroscopy in the infrared region. It has two major advantages as compared to a conventional grating spectrometer. All wavelengths are present at the same time and a larger infrared radiation flux is present in the spectrometer for a given resolution. These two facts enable fast measurements with outstanding signal-to-noise ratio. This spectrometer is ideal for absorption measurements as well as photoconductivity measurements in which the sample itself acts as a detector.

The reciprocal of the wavelength is known as the wavenumber (W) measured in cm⁻¹. Concisely, it measures the number of waves present in a centimeter in terms of the number of crests and troughs.

W = 1

λ (EQ 7.1)

The wavenumber of a light wave is directly related to the photon energy as:

E = hcW (EQ 7.2)

where,

E = photon energy

c = speed of light in vacuum (2.9979 ∗ 10⁸m/s) h = Plancks constant (6.62607 ∗ 10⁻³⁴J.s)

From this relation, a larger wavenumber corresponds to a higher photon energy, and a shorter wavelength.

The pattern of absorption of infrared radiation by matter as a function of wavelength or wavenumber is called an infrared absorption spectrum (Fig 7.1).

An important parameter for a measured spectrum is the signal-to-noise ratio (SNR). SNR is a measure of the peak height in absorbance to the noise level present at the baseline point in the spectrum as shown in Fig 7.1. Noise is usually observed in the form of fluctuating signals on either side of the main peak. A FTIR spectrometer has an excellent SNR, which facilitates detection of weak interesting spectral features above the noise level.

(38)

Figure 7.1: Infrared absorption spectrum illustrating the absorption peak with baseline noise.The SNR of this spectrum is 0.5/0.025 = 20.

7.2 Description and working principle of the FTIR spectrometer

The physical layout of the FTIR spectrometer is shown in Fig 7.2. The spectrometer consists of various key components discussed in the following sub-sections.

The beam of light from a broad-band source is focused onto an adjustable circular aperture. Then the light is focused on the collimating mirror2 which reflects a parallel beam to the entrance of the interferometer. After having passed the interferometer, the modulated light is focused onto the sample compartment where in our measurements the sample itself acts as a detector for the photoconductivity measurements [14].

(39)

Figure 7.2: Schematic layout of a Fourier transform infrared spectrometer

(40)

The key components of an FTIR spectrometer are discussed in the following:

7.2.1 Broad band source

An infrared source consists of a coil of nichrome wire or a small ceramic piece through which electric current is passed. Due to resistance of the element, it gets heated up giving off the infrared radiation. In our experiments, a globar is used as the broad-band infrared source. The globar source is heated above 1400K to give enough radiation to improve the SNR.

7.2.2 Beamsplitter

A beamsplitter splits the beam of radiation into two parts , one being transmitted and the other reflected. The beam splitter used in the experiments consists of a thin Ge coating sandwiched in-between two pieces of KBr. This type of beamsplitter covers the entire mid-infrared region i.e., from 400 to 4000 cm⁻¹.

7.2.3 The Michelson interferometer setup

The Michelson interferometer basically consists of a beamsplitter, a fixed mirror and a movable mirror as shown in Fig 7.2. The Michelson interferometer is the most used type in commercially available FTIR instruments.

An interferometer works on a principle of splitting an incoming light beam into two parts. The relative difference in the distance propagated by the two partial beams is referred to as the Optical Path Difference (OPD) denoted by δ.

The beam of light which is transmitted through the beamsplitter strikes the fixed mirror whereas the other beam which is reflected by the beamsplitter is allowed to strike the moving mirror. After getting reflected from their respective mirrors, the two light beams are again transmitted or reflected by the beamsplitter. The output light from the interferometer is focused by the collimating mirror3. If the distance between the movable and fixed mirrors from the beamsplitter is the same, then the two light beams travel the same distance to attain a Zero Path Difference (ZPD) condition. When these beams interfere at the exit, the resulting beam has higher amplitude than the individual beams. This phenomenon is called constructive interference. All wavelengths of light constructively interfere at ZPD. Constructive interference also takes place when the OPD is equal to an integer of λ [15].

When the light beams are completely out of phase at the exit i.e.,

δ = (n + 1/2)λ (EQ 7.3)

where destructive interference occurs resulting in a vanishing intensity for n = 0,1,2,3,. . . . Depending on the wavelength of the light, constructive and destructive interference occur at a given OPD. If the movable mirror is scanned with a constant velocity, the intensity at a given wavelength will be uniquely coded in the total intensity distribution as a function of the OPD. The variation of light intensity with OPD is observed as a complex waveform on a plot called the interferogram shown in Fig 7.3. The interferogram

(41)

is subsequently Fourier transformed to give the infrared spectrum. For generating one interferogram, the movable mirror is allowed to make one scan between two end points.

In our measurements, typically 1000 scans were superimposed to achieve a good SNR.

Figure 7.3: An interferogram showing the centerburst at ZPD.

The interferogram shown in the Fig 7.3 is obtained for a broad-band infrared source in which the centerburst shown corresponds to a maximum intensity at ZPD [15]

which is the signal obtained for all the wavelengths constructively interfering at ZPD. As we move away from ZPD, the signal intensity becomes weaker showing the destructive interference.

In our measurements, the modulated infrared radiation from the interferometer is focused onto the sample which itself acts as a detector generating a modulated photocurrent that is subsequently amplified and converted to a voltage. After Fourier trans- formation of this ”voltage” interferogram, the photocurrent versus wavenumber is deduced.

(42)

(43)

CHAPTER 8. RESULTS AND DISCUSSION

8 Results and discussion

The absorption of light in an n-type photodetector releases mobile electrons to the conduction band which increases the conductivity in the device. The photoconductivity measurements in this project were performed on different DWELL QDIP and QWIP devices at 10K and 77K using a Bomem DA8 Fourier transform infrared spectrometer at Lund University, Sweden. Fig 8.1 presents a list of components investigated in the project.

Figure 8.1: List of components used in the measurements

(44)

8.1 Photoconductivity studies on lateral transport DWELL QDIPs

Fig 8.2 shows the spectral distribution of the photocurrent obtained at different temperatures in a lateral DWELL QDIP structure. The weak peak at 5µm, corresponds to electron transitions from the ground state of the dot into the matrix, whereas the peak at 9µm corresponds to the transition from the dot to the excited state in the well. The electrons excited into the well are expected to contribute to a photocurrent as it is a lateral device. However, comparing the spectra at 77K to the one at 10K it is evident that thermal excitation of the electrons from the excited state plays a dominant role. The 9µm peak at 10K is small in comparison to the one obtained at 77K, which is most likely due to re-trapping of electrons in the well into the dots reducing the conduction inside the well. The electrons therefore need to have a higher energy than the excited state energy of the well in order to form a large photocurrent signal. It is still an open question if the conduction actually takes place in highly excited states in the well, or in the matrix outside of the well. If the electrons escape out of the well completely or not is thus not yet understood.

Figure 8.2: Temperature dependence of the PC normalized w.r.t. the photon flux for a lateral DWELL QDIP structure

(45)

8.2 Photoconductivity studies on vertical transport DWELL QDIPs

Figure 8.3 shows the spectral distribution of the PC obtained at 77K and different bias in a vertical transport DWELL QDIP (QD83).

Figure 8.3: Bias dependence of the PC at 77K, normalized with respect to the photonflux, for QD83.

At 4 volts, a strong signal was detected with the main peak in the 8.5µm-region corresponding to an excitation from the QD ground state to the excited state in the QW. The peak centered at about 5.5µm corresponds to an excitation from the QD to the GaAs- matrix. At 3 volts the signal has been significantly reduced and the relationship between the two peaks has changed towards equal magnitude. The fact that the signal becomes increasingly dependent on excitation from the QD to the excited state in the QW at larger bias points to the fact that a tunneling process from the excited well state to the matrix occurs.

(46)

Figure 8.4: Temperature dependence of the photocurrent signal at 4V for QD83.

In Figure 8.4 the temperature dependence of the signal at 4 volts is shown. As can be seen, the peak centered at 8.5µm is reduced when lowering the temperature to 10K. This indicates that the electrons excited into the excited state of the QW needs further thermal excitation in order for the tunneling process to become efficient. This two-step excitation process is referred to as thermally assisted tunneling (see Section 6.2).

During the project, two devices with the exact same structure, QD83 and QD82, were fabricated. The only difference between the two devices was an extra annealing step for the QD82 device. The same processes are therefore predominant in the components wherefore they both can be used in a comparative analysis of the photocurrent response in vertical structures.

(47)

Figure 8.5 shows the temperature dependence for the QD82 device and evidently the same thermally assisted tunneling process is observed as in QD83.

The voltage dependence of the photocurrent signal at 77K in QD82 is illustrated in Figure 8.6.

Figure 8.6: Bias dependence of the photocurrent at 77K for QD82.

At 4 volts the tunneling process out from the excited state of the well is still not efficient.

A bias of 5V is needed to see the strong enhancement of the tunneling efficiency seen at 4V for sample QD83.

(48)

Figure 8.7 shows the temperature dependence of the PC at 4V. Evidently, the temperature has little effect in the region where the applied voltage has not yet enabled tunneling from the excited state in the well. This result shows that the contribution to the 8.5µm peak from thermal excitation of the electrons directly from the excited state into the GaAs- matrix is very small. This also indicates that the recapture of photoexcited electrons in the well is fast. The interpretation of the combined result from the data is that generation of photocurrent in a vertical transport DWELL QDIP becomes more and more dominated by thermally assisted tunneling from the excited state of the QW with increased voltage.

This process is illustrated in Figure 8.8.

Figure 8.8: Dominant process leading to photocurrent in vertical transport DWELL QDIPs.

(49)

8.3 Comparison of vertical carrier transport DWELL QDIPs and lateral carrier transport DWELL QDIPs

One of the most notable differences between the two DWELL QDIP structures studied in this project is their tuning possibilities. While the lateral carrier transport DWELL QDIP, QD56, has no tuning possibility the vertical carrier transport DWELL QDIP, QD83 and QD82, do. This is illustrated in Figure 8.9.

Figure 8.9: Voltage dependence of vertical QD83 and lateral QD56 transport DWELL QDIPs

For the lateral DWELL QDIP, excitation from the QD into highly excited states in the well, or into the GaAs matrix is the main process leading to a photocurrent, regardless of applied voltage. A shift in the threshold wavelength is therefore not possible.

For the vertical DWELL QDIPs, the processes contributing to the photocurrent varies with voltage. At lower voltages the main process is excitation from the QD into the GaAs-matrix. At higher voltages the main process is excitation from the QD to the excited state in the QW where further thermal excitation leads to tunneling into the matrix. Both of these effects are visible in Figure 8.9. By varying the voltage applied over the vertical carrier transport DWELL QDIP, it is therefore possible to shift the main wavelength detection peak of the device between 8.5µm and 5.5µm. Further studies and development of this device might lead to a dual- or multi-wavelength region detector that is voltage controlled.

(50)

8.4 Comparison of photoconductivity in QWIPs and DWELL QDIPs

Figures 8.10 and 8.11 shows the spectral distribution of the photocurrent at different temperatures obtained in a QWIP1282. The IR radiation impinges on a 45 degree edge- polished surface of the sample. Edge-polished surfaces are commonly used for QWIPs due to their inherently lower sensitivity to normal incidence radiation.

Figure 8.10: Normalized photocurrent at 10K in a QWIP

Figure 8.11: Normalized photocurrent at 77K in a QWIP

(51)

Figure 8.12: Temperature dependence of the normalized PC at different temperatures in a vertical transport DWELL QDIP(QD83)

The corresponding photocurrent signals obtained for QD83 at different temperatures is shown in Fig 8.12. At 4V, the strength of the signal at 77K is large compared to 10K.

Fig 8.13 shows dark current measurements taken for QWIP and DWELL QDIP samples at different temperatures. It is evident that the dark current for QD83 is much smaller as compared to QWIP1282.

Figure 8.13: Comparison of dark current measurements for QWIPs and vertical transport DWELL QDIPs (Mesasize QWIP:200µm ∗ 200µm, Mesasize DWELL QWIP:360µm ∗ 360µm).

The main problem using QWIPs at higher temperatures is the high dark current levels they exhibit. At low photon flux, such as the amount of heat radiating from a human being, only weak photocurrent signals are generated. The signal-to-noise ratio for a QWIP detecting a human being, would therefore be poor, even when the detector is cooled to 77K. The increased signal in combination with the low dark current levels at 77K for the DWELL QDIP shown in Figure 8.13 is noticeable since it gives an indication that DWELL QDIPs can be used at higher temperatures such as 77K with a good signal-to- noise ratio. This attributes to a lower operating cost of the detector as the cooling power

(52)

for associated cryocoolers can be reduced. The reason for the lower dark current levels in DWELL QDIPs is partially due to a lower ground state energy of the dots which reduces thermally induced excitation processes.

(53)

8.5 Comparison of polarization dependence in QWIPs and DWELL QDIPs Fig 8.14 shows the spectral distribution of the photocurrent obtained in an edge polished sample for incident radiation at 45 degrees relative to normal incidence, and at normal incidence at 10K. Theory predicts that intersubband excitations are forbidden for normal incidence, and hence that no PC is generated. In our case, a weak PC signal is observed at normal incidence. This signal is due to effective scattering of the radiation by different layers inside the device. This enables some of the electrons to be absorbed by the incident radiation. This scattering is dependent on the exact sample geometry and therefore it varies between different components. In fact, in commercial QWIPs, artificial gratings are embedded in the sample to optimize the scattering.

Figure 8.14: Polarization dependence in QWIPs

(54)

Fig 8.15 shows the photocurrent signals obtained for edge incidence and normal incidence radiation of DWELL QDIPs. There is a 20% difference in amplitude between the peaks in the two measurements. It has to be noted that the normal incidence measurements were done from the backside of the device. It is therefore expected that absorption takes place before the radiation strikes the active (dot-in-well) layers of the device. This leads to a loss of signal for the normal incidence measurements compared to the edge polished surface measurements. Hence, we conclude that the DWELL QDIPs are more polarization independent than the QWIPs. This important property of DWELL QDIPs eliminates the additional fabrication step of adding grating couplers.

Figure 8.15: Polarization dependence in DWELL QDIPs

(55)

CHAPTER 9. CONCLUSIONS

9 Conclusions

The lateral carrier transport DWELL QDIP was found to have poor conduction in the QW, mainly due to re-trapping of electrons into dots in this region. The main process governing the flow of photocurrent in this type of device at 77K is photo-excitation from the QDs to the excited state in the QW and further thermal excitation. If the electrons are mainly transported in the matrix or in the well at 77K is not known.

For the vertical carrier transport DWELL QDIP at 77K, the wavelength response could be tuned by altering the applied voltage. At higher voltages, the dominant process was attributed to photo-excitation from the QDs to the excited state in the QW where thermal assisted tunneling into the GaAs-matrix took place. At lower voltages, photo- excitation from the QDs directly into the GaAs-matrix was the predominant process.

The dark current levels in the vertical QDIPs were found to be 1.5 to 5 orders of magnitude smaller than the corresponding levels for the QWIPs measured at 77K. The reason for the decrease in dark current for the QDIPs is still being investigated.

The QDIPs were found to be close to polarization independent. The QWIPs had a reduced signal at normal incidence radiation. The existence of this signal was attributed to interface scattering inside the device.

The results presented in this thesis points to important advantages gained by replacing QWIPs with QDIPs. Further research is necessary to quantify important parameters not covered in this work e.g. responsivity and speed. Still, we believe that QDIPs will be introduced in the next-coming generations of high performance IR photodetectors.

(56)

(57)

BIBLIOGRAPHY

Bibliography

[1] S. M. Sze, Semiconductor Devices: Physics and Technology, 2nd ed. John Wiley and Sons,Inc., Newyork, 2002.

[2] Infrared-Detectors-Tutorial, “Ir-detectors,” 2005-05-17. [Online]. Available: http:

//www.electro-optical.com/bb rad/detector.htm

[3] Acreo-AB, “Infrared detector arrays for thermal imaging,” 2005-05-17. [Online].

Available: http://www.acreo.se/templates/Page 217.aspx

[4] G. Rieke, Detection of Light: From the Ultraviolet to the Submillimeter, 1st ed.

Cambridge University Press, 2002.

[5] A. Rogalski, Infrared Detectors. Taylor and Francis, 2000.

[6] J. Andersson et al, “R and d and production of qwip fpa at acreo, sweden * not only a nightvision,” SPIE’s International Symposium on Optical Science and Technology, 2002.

[7] E. Rosencher, Optoelectronics, 1st ed. Cambridge University Press, 2002.

[8] M. Razeghi, Fundamentals of Solid State Engineering, 1st ed. Kluwer Academic Publishers, 2002.

[9] B. R. Nag, Physics of Quantum Well Devices, 1st ed. Kluwer Academic Publishers, 2001.

[10] S. D. Gunapala et al, “640x512 pixel four-band, broad-band, and narrow-band quantum well infrared photodetector focal plane arrays.” Proceedings of SPIE 4820, pp.

306–317, 2003.

[11] M. Razeghi, Chapter 1-H.C.Liu of Long Wavelength Infrared Detectors. Taylor and Francis, 1997.

[12] B. F. Levine et al, “High sensitivity low dark current 10 µm gaas quantum well infrared photodetectors,” Applied Physics Letters 56(9), pp. 851–853, 1990.

[13] V. Ryzhii, “The theory of quantum dot infrared phototransistors,” Semicon. Sci.

Technol. 11, pp. 759–765, 1996.

[14] L. Landin, Optical studies of InAs Quantum Dots in III-V Semiconductors. Ph.D Thesis, Lund University, Sweden, 2000.

[15] B. C. Smith, Fundamentals of Fourier Transform Infrared Spectroscopy. CRC Press, 1995.

Comparative study of infrared photodetectors based on quantum wells (QWIPs) and quantum dots (QDIPs)

IDE0626,Technical Report,IDE2006

COMPARATIVE STUDY OF INFRARED PHOTODETECTORS BASED ON QUANTUM

WELLS (QWIPS) AND QUANTUM DOTS (QDIPS)

Master´s Thesis in Electrical Engineering Conny Hansson, Krishna Kishore Rachavula

Comparative study of infrared photodetectors based on quantum wells (QWIPs) and quantum dots (QDIPs)

Master´s Thesis in Electrical Engineering

Abbreviations

Contents

1 Abstract

2 Acknowledgments

3 Introduction

4 Background and Related Work

5 Theory

6 The project

7 Fourier Transform Infrared Spectroscopy

8 Results and discussion

9 Conclusions

Bibliography