Greg Bishop

Candidate

Electrical and Computer Engineering

Department

This thesis is approved, and it is acceptable in quality

and form for publication on microfilm:

Approved by the Thesis Committee:

, Chairperson

Accepted:
Dean, Graduate School

Date
InAs/Ga(In)Sb Superlattice Based Infrared
Detectors using nBn Design

by

Greg Bishop

B. S., Electrical and Computer Engineering,

University of New Mexico, 2006

THESIS

Submitted in Partial Fulfillment of the

Requirements for the Degree of

Master of Science
Electrical Engineering

The University of New Mexico

Albuquerque, New Mexico
May, 2008
© 2008, Greg Bishop

iii
Acknowledgements
I would like to acknowledge my advisor Prof. Sanjay Krishna for encouraging me to
attend graduate school and for help and guidance throughout the course of this work. I
would also like to thank my thesis committee, Prof. Luke Lester and Prof. Ralph Dawson
for their help and questions during the defense and with the manuscript. I am eternally
grateful to Dr. Elena Plis for proofreading my manuscript and the incredible boost that it
provided to the overall structure, readability and flow of my thesis. I would ask like to
acknowledge the group members who contributed to this work either in growth or
processing of samples both of which I did not do myself, Dr. J.B. Rodriguez, Dr. Elena
Plis, Dr. Yagya Sharma, Ha Sul Kim and Arezou Khoshakhalagh.

I would like to thank my family and friends for support and understanding. I would
especially like to thank my wife, Kelli for sacrificing so much so that I could finish this
degree.

iv
InAs/Ga(In)Sb Superlattice Based Infrared
Detectors using nBn Design

by

Greg Bishop

ABSTRACT OF THESIS

Submitted in Partial Fulfillment of the

Requirements for the Degree of

Master of Science
Electrical Engineering

The University of New Mexico

Albuquerque, New Mexico

May, 2008
 InAs/Ga(In)Sb Superlattice Based Infrared Detectors

using nBn Design

by

Greg Bishop

B.S., Electrical and Computer Engineering, University of New Mexico, 2006

M.S., Electrical Engineering, University of New Mexico, 2008

Abstract

Infrared detectors in the mid-wave infrared (MWIR) (3-5µm) and long-wave infrared

(LWIR) (8-12µm) have many applications in military, industrial and medical fields. The

state-of-the-art photodetectors based on Hg1-xCdxTe (MCT) have large signal over noise

ratio and their bandgap can be tuned to span 1-32 µm wavelength range. However, large

tunneling and Auger dark currents due to low electron effective mass in MCT detectors

require operation at cryogenic temperatures (77-100 K). The cooling requirement limits

the lifetime, adds weight and expense and increases the power consumption of the

infrared system. There is a concerted effort to develop photonic detectors operating at

higher temperatures. InAs/GaSb strained layer superlattice (SLS) photodectors are now

considered as a promising technology for both MWIR and LWIR wavelength ranges.

The bandgap of the SLS can be adjusted by controlling the thickness of the constituent

InAs and GaSb layers during the growth process. InAs/GaSb SLS can also span the 3-30

vi
µm wavelength range. Moreover, the large splitting of the energy levels of the different

valence subbands in SLS contribute to the suppression of Auger recombination.

The recently proposed nBn heterostructure design has demonstrated a 100 K increase in

background-limited infrared photodetection (BLIP) for InAs-based device, by decreasing

Shockley-Read-Hall generation currents and by suppressing surface currents using

specific processing.

This work is focused on combining the nBn detector design with InAs/GaSb superlattice

material utilizing an AlGaSb barrier layer to improve detector performance and to

increase the operating temperature of Focal Plane Arrays. This thesis covers three topics:

Optical and electrical characteristics of single pixel devices in both the MWIR and LWIR,

the reduction of surface leakage currents by using a shallow isolation etch for definition

of the top contact mesa and comparison between nBn detector and PIN LWIR detectors.

vii
List of Figures xi

List of Tables xv

1 Introduction 1
1.1 Electromagnetic Spectrum .......................................................................... 1

1.2 Infrared radiation ........................................................................................ 4

1.3 Infrared detection ........................................................................................ 5

1.3.1 Thermal Detectors........................................................................... 5

1.3.2 Photon detectors.............................................................................. 7

1.3.3 Photoconductive detectors .............................................................. 8

1.3.4 Photovoltaic detectors................................................................... 10

1.4 Figures of merit......................................................................................... 11

1.4.1 Responsivity.................................................................................. 12

1.4.2 Noise equivalent power (NEP) ..................................................... 14

1.4.3 Specific Detectivity (D*) .............................................................. 14

1.5 Theoretical limit of detection.................................................................... 15

1.6 Focal Plane Arrays.................................................................................... 18

1.6.1 Thermal Detector FPAs ................................................................ 18

1.6.2 Photon Detector FPAs................................................................... 19

1.7 InAs/Ga(In)Sb Type II Strain Layer Superlattice..................................... 21

1.7.1 Type II strain layer superlattice devices ....................................... 24

viii
 1.8 nBn detector .............................................................................................. 25

1.9 Conclusion ................................................................................................ 27

2 Materials and Methods 29

2.1 Molecular Beam Epitaxy (MBE) .............................................................. 29

2.2 Superlattice device processing.................................................................. 34

2.3 Characteristics of Detector Performance .................................................. 36

2.3.1 Responsivity and D* ..................................................................... 42

2.4 Conclusion ................................................................................................ 45

3 InAs/GaSb SLS based MWIR detector using nBn design 46

3.1 Dark current reduction in nBn SL detector............................................... 46

3.2 nBn Detector Growth................................................................................ 47

3.3 Contact metallization Study...................................................................... 48

3.4 nBn Detector Device Processing .............................................................. 50

3.5 nBn Detector Device Characterization ..................................................... 52

3.6 Superlattice nBn 320 x 256 Focal Plane Array......................................... 57

3.7 Conclusion ................................................................................................ 59

ix
4 InAs/GaInSb SLS based detector using nBn design with 8 µm
cutoff 61
4.1 Surface leakage current in InAs/GaSb LWIR........................................... 61

4.2 Growth and Processing ............................................................................. 62

4.3 Device characterization and comparison .................................................. 63

4.4 Conclusion ................................................................................................ 69

5 Conclusions and Future Work 70

5.1 Conclusions............................................................................................... 70

5.2 Future Work .............................................................................................. 72

5.2.1 Lateral Collection.......................................................................... 72

5.2.2 nBn theoretical modeling.............................................................. 73

5.2.3 Focal Plane Array Fabrication ...................................................... 74

5.2.4 Multispectral nBn devices............................................................. 75

5.2.5 Multiple barrier devices ................................................................ 75

Appendix A 77

Appendix B 80

References 90

x
List of Figures

Figure 1.1 The Electromagnetic Spectrum ....................................................................... 1

Figure 1.2 Planck’s Law: spectral exitance vs. wavelength for BB temperature ranging

from 300 K to 1000 K. Dashed line represents Wien’s Law. .................... 3

Figure 1.3 The Transmittance of the atmosphere ............................................................. 4

Figure 1.4 Schematic of a bolometer ................................................................................ 6

Figure 1.5 Quantum efficiency for ideal and real photon detectors.................................. 8

Figure 1.6 Photogeneration in a) intrinsic, b) extrinsic n-type and c) extrinsic p-type

photoconductors.......................................................................................... 9

Figure 1.7 Schematic photoconductor with externally applied bias ................................. 9

Figure 1.8 Schematic of p-n junction optical absorption ................................................ 10

Figure 1.9 IV curve for ideal photodiode with and without incident radiation .............. 11

Figure 1.10 Peak current responsivity as a function of wavelength for ideal photon

detector with different quantum efficiencies ............................................ 13

Figure 1.11 θ dependent aperture limit ........................................................................... 16

Figure 1.12 BLIP D* for ideal PV, PC and Thermal detector ........................................ 17

Figure 1.13 Comparison of commercially available infrared detectors at given operating

temperature ............................................................................................... 19

Figure 1.14 Superlattice schematic band structure and absorption process..................... 23

Figure 1.15 Band diagram of InAs-based nBn structure, biased under operating and

flatband (inset) conditions34 ...................................................................... 25

Figure 2.1 Schematic of a Molecular Beam Epitaxy growth chamber ............................ 31

xi
Figure 2.2 Schematic of possible interface ...................................................................... 32

Figure 2.3 Shutter sequence for mixed interface growth ................................................. 33

Figure 2.4 Cross-sectional TEM image of type-II superlattice undertaken at JPL.......... 34

Figure 2.5 Mask designs used in fabrication of single pixel detectors ............................ 36

Figure 2.6 Conventionally defined PIN photodiode ........................................................ 36

Figure 2.7 DTGS detector correction curves for different FTIR mirror velocities......... 38

Figure 2.8 Background spectrum measured by DTGS detector ...................................... 39

Figure 2.9 Correction Curve as a function of wavenumber............................................. 40

Figure 2.10 Background spectrum after division by correction curve............................. 40

Figure 2.11 Superlattice detector sample spectral response without correction .............. 41

Figure 2.12 Final normalized and corrected SL detector spectral response (NR(λ))....... 41

Figure 2.13 Responsivity measurement setup................................................................. 43

Figure 2.14 Blackbody source and detector configuration for testing............................. 44

Figure 3.1 Heterostructure schematic of InAs/GaSb SLs MWIR detector after growth 48

Figure 3.2 Contact resistance versus the doping concentration in the contact layer at

different annealing temperatures for Ti/Pt/Au contacts on n-type

InAs/GaSb superlattice Inset: The doping concentration as a function of

the Si cell temperature .............................................................................. 49

Figure 3.3 Schematic of structure A (shallow etch)......................................................... 51

Figure 3.4 Schematic of structure B (deep etch).............................................................. 51

Figure 3.5 The dark current density vs. 1000/T in structures A (shallow etched) and B

(deep etched) (Vb = 0.1V). Right Axis: Ratio of dark current densities for

both structures........................................................................................... 53

xii
Figure 3.6 Current Density vs. Voltage for structures A (shallow etched) and B (deep

eteched) at 77 K ........................................................................................ 54

Figure 3.7 Current density vs. Perimeter/Area at V = 0.5V performed by Santa Barbara

Focal Plane................................................................................................ 55

Figure 3.8 The responsivity (solid black line, left axis) and quantum efficiency (dotted

red line, right axis) for structure A (shallow etch).................................... 56

Figure 3.9 Peak Responsivity and D* as a function of applied bias for structure A

(shallow etch)............................................................................................ 57

Figure 3.10 Scanning Electron Microscope (SEM) image showing part of a fully

processed Focal Plane Array with a shallow etched top contacts and deep

etched bottom contacts with indium bumps.............................................. 58

Figure 3.11 The Noise Equivalent Temperature Difference (NETD) distribution in the

FPA for an integration time of 16.3 ms using f/4 optics........................... 58

Figure 3.12 Thermal image taken with at 77 K with and integration time of 16.3 ms and

using a two-point non-uniformity corretction........................................... 59

Figure 4.1 Schematic structure of nBn (Sample A) and PIN (Sample B) 8 µm λcutoff

detectors .................................................................................................... 63

Figure 4.2 Room temperature background JV plotted at different temperatures for sample

A (nBn) ..................................................................................................... 64

Figure 4.3 Room temperature background JV plotted at different temperatures for sample

B (PIN)...................................................................................................... 65

Figure 4.4 Comparison of sample A (nBn) and B (PIN) current density as a function of

applied bias at T = 77 K and 250 K .......................................................... 65

xiii
Figure 4.5 The current density for sample A (nBn, black) and B (PIN, red) versus 1/kT66

Figure 4.6 Normalized spectral response of samples A (nBn, black) and B (PIN, red) at

77 K........................................................................................................... 67

Figure 4.7 Normalized spectral response of samples A (nBn, black) and B (PIN, red) at

150 K......................................................................................................... 67

Figure 4.8 Responsivity (left axis, black line) and Quantum Efficiency (right axis, red

line) are given as a function of wavelength for sample A (nBn) .............. 68

Figure 5.1 MWIR image captured from the first SL nBn FPA with a non-uniformity

correction applied at T = 77 K with a NETD = 26mK ............................. 74

Figure 5.2 Spectral response of nBn detectors with a bias Vb = -0.5V across two top

contacts at T = 100 K ................................................................................ 76

xiv
List of Tables

Table 1.1 Superlattice device designs and parameters...................................................... 24

xv
Chapter 1. Introduction

1 Introduction

1.1 Electromagnetic Spectrum

The photon is the elementary particle of light. Everything with temperature above

absolute zero (0K) gives off photons. The electromagnetic (EM) spectrum classifies the

entire range of possible photon energy. The different regions of the EM spectrum have

been given different names according to their frequency or corresponding wavelength, as

shown in figure 1.1.

Figure 1.1 The Electromagnetic Spectrum1

The emitting properties of an object depends on its temperature, material, and surface

condition. Blackbody (BB) is a convenient construct for the study of emmitance of

radiation of an object at given temperature since the emitted radiation depends only on

the temperature of the blackbody. BB is a hypothetical entity that absorbs all EM

1
Chapter 1. Introduction

radiation that is incident on it and transmits or reflects none and emits energy with perfect

efficiency.

In 1900, Max Planck created an equation for the emission of radiation of a blackbody

that matched the measured values. This equation, known as Planck’s law2,3, is

2πhc 2
M e (λ , T ) = hc (1.1)
λ (e
5 λkT
− 1)

where k is Boltzmann’s constant, h is Planck’s constant and c is the speed of light where

the units are W/(cm2·µm) defines a probability distribution of blackbody radiation at a

given temperature and is shown in Figure 1.2. The idea behind Plank’s law was the

presupposition of a relationship between photon energy E and photon frequency υ and

frequency of EM radiation in vacuum. Planck assumed that E is proportional to the

oscillation frequency υ giving

c
E = hυ = h (1.2)
λ

Two more important laws that give more information about blackbody radiation are

Wien’s law and the Stefan-Boltzmann law. Wien’s law gives the peak of the distribution

of EM energy emitted by BB at a given temperature and is given by

2898
λ max = (1.3)
T

where T is in Kelvin and λmax is in microns. The Stefan-Boltzmann law estimates the

total energy radiated by a blackbody at a given temperature and is given by

M = σT 4 (1.4)

where σ is the Stefan-Boltzmann constant given by4

2
Chapter 1. Introduction

2π 5 k 4 W
σ= 2 3
= 5.7604 x10 −8 (1.5)
15c h cm 2 K 4

From Wien’s law, which is represented by the dashed line in Figure 1.2, one can see that

with increasing BB temperature the peak of Planck’s function moves to the shorter

wavelengths.

1 T=300K
T=400K
T=500K
0.1 T=700K
T=900K
Me (W/(cm µm))

T=1000K
0.01 Wien's Law
2

1E-3

1E-4

1E-5
1 10 100
Wavelength (µm)

Figure 1.2 Planck’s Law: spectral exitance vs. wavelength for BB temperature ranging from

300 K to 1000 K. Dashed line represents Wien’s Law.

All of the laws related to the idealized blackbody are applicable to the real objects by

introducing a parameter called emissivity. It is defined as the ratio between the exitance

of the actual object and the exitance of the BB. The emissivity of an ideal blackbody is

one while laboratory blackbodies can be within a few percent of unity.

The peak wavelength of emission spectrum of the sun (6000 K) corresponds to 0.5 µm

which is in the middle of the visible spectrum. For objects at or around ambient

temperature (300 K) including humans, the peak wavelength of emission spectrum is

3
Chapter 1. Introduction

close to 10 µm. This wavelength falls in the infrared (IR) part of the EM spectrum and

can be detected as heat.

1.2 Infrared radiation

Infrared or thermal radiation has been defined as the part of the EM spectrum with

wavelength longer than visible light and shorter than radio waves. By convention, the

range of the infrared spectrum is from ~0.7 µm to 1000 µm. Different molecules in the

atmosphere absorb different wavelengths of infrared radiation as shown in Figure 1.3.

Water vapor absorbs infrared radiation from 5.5 µm to 7.5 µm and for various

wavelengths from 0.7 µm to 3 µm. CO2 absorbs in distinct narrow bands at 2.7 µm, 4.25

µm and 15 µm. Ozone absorbs infrared radiation at 9.5 µm. The relatively transparent

bands (“windows”) are 3-5µm and 8-13µm.

Figure 1.3 The Transmittance of the atmosphere5

4
Chapter 1. Introduction

Infrared radiation has been classified according to atmospheric windows into several

bands. Near Infrared Radiation (NIR) spans from the end of the visible radiation, 0.7 µm

to 1 µm; Short Wavelength Infrared Radiation (SWIR) covers the EM spectrum from 1

µm to 3 µm. The Mid Wavelength Infrared Radiation (MWIR) window is from 3 to 5 µm.

The Long Wavelength Infrared Radiation (LWIR) is from 8 to 12 µm and the Very Long

Wavelength Infrared Radiation (VLWIR) covers wavelengths 12 to 30 µm.

The LWIR window corresponds to the peak emission of a blackbody with a temperature

of 300K.

1.3 Infrared detection

By constructing devices that can convert infrared radiation to a measurable signal, in

most cases electrical, one can effectively detect the infrared radiation. Infrared detectors

can be split into two different categories, according to the methods of conversion of EM

radiation, thermal detectors and photon detectors.

1.3.1 Thermal Detectors

Thermal detectors are sensitive to external variation in temperature. This is a two step

process; first the absorbed radiation on the detector changes the temperature of the

detector, then the temperature change induces a measurable parameter change. Because

of the nature of detection (response to the intensity of absorbed power), the response of

thermal detectors is not dependent on the wavelength of incident light. The drawback of

thermal detectors is that the incident radiation must change the temperature of the

detector, so the response of thermal detectors is inherently slow. Three common

5
Chapter 1. Introduction

examples of thermal detectors are bolometers, thermocouples, and pyroelectric.

Operation principles of these detectors are discussed further.

The operation of bolometers is based on the bolometric effect, which is change of the

resistance of the responsive element with temperature change produced by the absorbed

IR radiation. Bolometers are usually made of thin metal or semiconductor film with an

absorbing film deposited on top of it and is suspended over a cavity in a substrate for

thermal isolation. The electrical resistance of the film is monitored. When the

temperature increases due to absorbed radiation, the film resistance changes, it increases

if the film is metallic and decreases if the film is semiconductor.

Figure 1.4 Schematic of a bolometer6

Thermocouples rely on the thermovoltaic effect which is that the separation between two

different metals will change with change in the amount of incident radiation. This

change in separation changes the voltage across the junction. Thermocouples connected

6
Chapter 1. Introduction

in series or parallel is called a thermopile and it used to increase the voltage output to a

measurable level.

Pyroelectric detectors can operate at room temperature and are fairly inexpensive making

them very common. Pyroelectric detectors operate when a change occurs in the

temperature due to change in incident radiation which causes an electrical polarization of

the material. This polarization change occurs on the surface of the element which creates

a current. Pryoelectric devices have a broad spectral response in the infrared spectrum.

The change of the polarization can vary rapidly in pryoelectric devices, but the response

of the detector is still limited by the thermal properties of the material.

1.3.2 Photon detectors

Photon detectors operate by the direct interaction of radiation with the atomic lattice of

the material. This interaction of light and matter produces electrical parameter changes

that are detected by associated circuitry. Electrical parameters that can change in a

photon detector are resistance, inductance, voltage and current. For photon detection the

concept of creating an electron-hole pair from an absorbed photon must be understood.

Quantum efficiency (QE), defined as the number of carriers measured at the output of the

detector per number of incident photons per unit time, is used to understand this concept.

External quantum efficiency is ideally unity for photodetectors, however in reality it is

always less than that. The quantum efficiency relates the incident photons on the active

area of the detector to the number of electrons generated, but excludes electrons

generated by internal gain. If the energy of incident radiation is high enough, photons

excite bound electrons to jump to the conduction band. The point where the photon no

7
Chapter 1. Introduction

longer has sufficient energy to excite an electron is called the cutoff wavelength, denoted

(λc) and as shown in figure 1.5. From equation 1.2 one can calculate the cutoff

wavelength.

hc
λc = (1.6)
Eg

Where Eg is the bandgap energy of the semiconductor material. The two main types of

photon detectors are photoconductive (PC) and photovoltaic (PV) devices.

Ideal Photodetector
Real Photodetector

0
λcutoff
λ

Figure 1.5 Quantum efficiency for ideal and real photon detectors

1.3.3 Photoconductive detectors

Photoconductors respond to radiation by changing the conductivity of the material. The

resistance of the device is then inversely proportional to the amount of incident radiation.

Intrinsic photoconductors absorb photons with energy above the bandgap energy of the

semiconductor material as shown in Fig 1.6 (a). The bandgap of extrinsic materials

depends on concentration of intentional impurities and they have absorption energies

8
Chapter 1. Introduction

much lower and λc much longer than intrinsic photoconductors as shown in Fig 1.6 (b &

c). To prevent thermal ionization, extrinsic semiconductors need to be cooled. This

means that intrinsic PCs can operate at higher temperature than extrinsic PCs, but that the

λc of extrinsic PCs can be much longer.

Figure 1.6 Photogeneration in a) intrinsic, b) extrinsic n-type and c) extrinsic p-type

photoconductors

To measure the change in conductivity, the PC must have an externally applied bias. The

external biasing circuit creates an electric field to collect the photogenerated carriers by

opposite contacts as shown in figure 1.7.

Contact Contact

+ -
Ec

hν Ev

Figure 1.7 Schematic photoconductor with externally applied bias

9
Chapter 1. Introduction

1.3.4 Photovoltaic detectors

Optical detection in a photodiode occurs when a photon with energy greater than the

bandgap excites an electron hole pair and the pair is separated by the built in electric field.

Only carriers in the depletion region are separated by the built in field, so electron hole

pairs outside a diffusion length of the depletion region recombine and do not contribute to

the photocurrent. A PIN junction photodiode is depicted in Fig 1.8

N I P

Depletion Region
-
Ec
+
Vbi

Ef
Ev
+
+

Figure 1.8 Schematic of p-n junction optical absorption

The diode current equation is given as

 bkT
qv

i = i0  e − 1 − i g
 (1.7)
 

where i0 is the reverse saturation current, q is the charge on the electron, v is the voltage

across the diode, b is the nonideality factor, k is Boltzmann’s constant, T is the

temperature and ig is the photogenerated current, given by

10
Chapter 1. Introduction

i g = ηφ q q (1.8)

where η is the quantum efficiency and φq is the photon flux measured in photons/sec.

The addition of the photogenerated current shifts the IV curve of the photodiode so that

there is an open circuit voltage (Voc) and a short circuit current (Isc) as shown in Fig. 1.9.

Figure 1.9 IV curve for ideal photodiode with and without incident radiation

The main advantage of a photodiode over a photoconductor is the dominant noise

mechanism in a photodiode is shot noise which a factor of 2 smaller than the dominant

noise mechanism in a photoconductor. Due to this, the SNR of photodiodes is usually

higher than the SNR for photoconductors.

1.4 Figures of merit

There have been many different ways to describe a detector’s performance through the

course of their history. In an attempt to allow for comparison between different types of

11
Chapter 1. Introduction

detectors, figures of merit have been developed. The figures of merit allow for a

comparison that is independent of detector technology. The general figures of merit

include responsivity, the noise equivalent power, and specific detectivity. While the

figures of merit characterize both thermal detectors and photon detectors, the units are

mostly energy based because thermal detectors were developed first.

1.4.1 Responsivity

The responsivity relates the output signal of a detector to the radiant input. The radiant

input is normally expressed in watts while the output signal is expressed in volts, called

the voltage responsivity, or in amps, called the current responsivity. The spectral

responsivity is defined as the responsivity in terms of wavelength while the blackbody

responsivity gives the peak responsivity value. For photon detectors, the responsivity is

related to the quantum efficiency. Using equation 1.8 as the expression for

photogenerated current and φe as the radiant power (flux) input given in watts, the

equation for the current responsivity is then

ηφ q q
ℜi = (1.9)
φe

given in units of amps per watt. The radiant power is related to the photon flux by

hc
φe = φq (1.10)
λ

so the equation for current spectral responsivity becomes

ηφq qλ λ
ℜ i (λ ) = = qη (1.11)
φq hc hc

12
Chapter 1. Introduction

Equation 1.11 gives the relation between quantum efficiency and current responsivity as

a function of wavelength for photon detectors. The theoretical limit for current

responsivity for different QE is given in Fig. 1.10.

The responsivity of a detector gives the expected signal for a given level of radiation.

The responsivity does not provide information about the sensitivity of the detector,

because no information is given about the noise in the detector. So the responsivity of a

detector could be large but the performance of the detector could still be poor if the noise

of the detector was also large.

10 QE=1
QE=0.6
QE=0.3
8
Responsivity (A/W)

0
0 2 4 6 8 10 12
Wavelength (µm)

Figure 1.10 Peak current responsivity as a function of wavelength for ideal photon detector

with different quantum efficiencies

13
Chapter 1. Introduction

1.4.2 Noise equivalent power (NEP)

The responsivity is the output signal divided by the radiant flux on the detector. The

signal to noise ratio (SNR) of the detector is then

ℜ iφ e
SNR = (1.12)
in

where Ri is the current responsivity, φe is the radiant power, and in is the noise expressed

as current. The noise equivalent power is, as the name suggests, the radiant input power

that makes the noise equivalent to the signal or a SNR equal to 1. Setting the left hand

side of equation 1.12 equal to 1 and solving for φe gives

in φe
NEP = = (1.13)
ℜ i i sig / i n

the units of which are watts. The NEP can also be found using the voltage noise and the

voltage responsivity. The lower the value of NEP, the more sensitive the device. The

NEP cannot be readily compared between different detectors because NEP depends on

detector area and other measurement parameters. For direct comparison of any type or

size detector under different operation conditions Jones 7 suggested a figure of merit

called normalized detectivity.

1.4.3 Specific Detectivity (D*)

The specific detectivity (D*) is the SNR of a detector normalized to an area of 1 cm2,

with 1 W of radiant power incident on the detector with a 1 Hz noise bandwidth. D* is

Ad ∆f
D* = (1.14)
NEP

14
Chapter 1. Introduction

where Ad is the area of the detector and ∆f is the bandwidth of the measurement. D* can

be related to the responsivity by

ℜ i Ad
D* = (1.15)
in / ∆f

The units of D* are cm Hz1/2/watt which is called Jones after the creator of the figure of

merit. D* allows for comparison between detectors regardless of size or measurement

parameters.

1.5 Theoretical limit of detection

Using detector figures of merit defined in the previous section, expressions for the

theoretical limit of detection can be derived. The best performance of a photon detector

occurs when photon noise, noise caused by the absorption of a photon, is the dominant

source of noise. When the photon noise from the signal is dominant it is called signal-

dependant noise. This condition is rare for infrared detectors because the background is

the usual dominant photon noise source. When the background is the dominant noise

source the detector is said to be background limited infrared photodetector or BLIP. To

find the BLIP limited D* for both types of photon detectors, the understanding of the

dominant noise sources for PV and PC detectors is required. Shot noise, the noise from

dc current flowing across a potential barrier is the dominant noise source in PV detectors.

The equation for shot noise is

i shot = 2qi∆f (1.16)

if i is the result of incident photons then the detector is BLIP. One can show that the NEP

for BLIP PV detector is

15
Chapter 1. Introduction

2 E q Ad ∆f hc
NEP = (1.17)
η λ

where Eq is the incident photon flux. Eq depends on the detector geometry as shown in

figure 1.11. Eq is then the photon flux exitance times sin2 θ known as the numerical

aperture.

Aperture

Detector
θ

Figure 1.11 θ dependent aperture limit

From equations 1.14 and 1.17, D* is then

λ η
*
D BLIP = (1.18)
hc 2 E q

For PC detectors generation-recombination noise is the dominant noise source and is

given by

i gr = 2qG E q Adη∆f (1.19)

where G is the photoconductive gain. Following the same steps as for PV detectors

λ η
*
D BLIP = (1.20)
2hc Eq

gives D* BLIP for PC. BLIP D* for PC is lower than BLIP D* for PV by the square root

of two.

16
Chapter 1. Introduction

BLIP D* for an ideal thermal detector is given by

1
*
DBLIP = (1.21)
4 ⋅ kσT 5

where σ is Stefan-Boltzman constant and at 300K is equal to 1.8x1010 Jones.

From equations 1.18, 1.20 and 1.21 and using the limiting factor of the field of view

(FOV) for the detector, the peak D* BLIP is shown as a function of wavelength in Fig.

1.12.

1E12

TBackground=300K
D* (Jones)

1E11 θ=90 FOV

D* PV
D* PC
D* Thermal

1E10
1 10 100 1000
Wavelength (µm)

Figure 1.12 BLIP D* for ideal PV, PC and Thermal detector

While photon noise limited detector is desired, in low photon flux levels situations the

best SNR is achieved when the detector is Johnson noise limited. Johnson noise limited

(JOLI) D* for a PV detector is given by

λη q R0 Ad
*
DJOLI = (1.22)
2hc kTd

where Td is the detector temperature and R0Ad is the dynamic resistance at zero bias.

17
Chapter 1. Introduction

1.6 Focal Plane Arrays

A Focal Plane Array (FPA) is the main component used in advanced infrared imaging

systems. An IR FPA is an array of single pixel devices used to capture an infrared signal.

FPAs increase the efficiency of collection compared to a single pixel by decreasing

acquisition time. Both thermal and photon detectors can be made into FPAs and each has

their own advantages and disadvantages.

1.6.1 Thermal Detector FPAs

As mentioned in section 1.3.1 operation of thermal detectors depend on measurable

parameter change produced by absorbed radiation. The main advantage of thermal

detectors is that it can operate at ambient temperature. Another advantage is that the

response is independent of wavelength, which is important for LWIR application. The

main disadvantages for thermal detectors are the low detectivity (capped at 1.8x1010

Jones at 300K) and the slow response time which eliminates thermal detectors FPAs from

being used for high performance applications. Applications for thermal detector FPAs

include space based astronomy, thermal sensors, industrial monitoring including low cost

handheld imagers for industrial applications.

Microbolometer arrays are made using silicon micromachining. Usually semiconducting

vanadium oxide is used as the detecting material. This is placed atop a thin membrane

typically silicon nitride and connected to the substrate by silicon nitride legs coated with

conductive film. Microbolometers with 640x480 pixels are commercially available.8

18
Chapter 1. Introduction

1.6.2 Photon Detector FPAs

Materials for commercial photon detectors FPAs are HgCdTe, InSb, GaAs/AlGaAs and

InAsSb. Each material system has its own inherent strengths and weaknesses.

Figure 1.13 Comparison of commercially available infrared detectors at given operating

temperature9

HgCdTe (MCT)

Infrared detectors HgCdTe (MCT) were first introduced in 1959 by Lawson et al. 10

Hg1-xCdxTe is made up of elements from column II and VI in the periodic table. It is a

direct gap material whose bandgap can be varied from negative values (-0.1415 and -

0.2608 eV at 300K and 77K) for HgTe to relatively large values for CdTe (1.4895 and

1.6088 eV at 300K and 77K) to give a cutoff wavelength a range from 1µm to 30µm.11

The lattice constant difference between HgTe (6.414Å) and CdTe (6.4809 Å) is only

0.3% so the composition of the material can be chosen in any range between the two and

still be relatively lattice matched. Also, the optical absorption coefficient is large

19
Chapter 1. Introduction

allowing for high quantum efficiency11. HgCdTe is the ideal material system for infrared

detection but certain technological issues keep it from being completely dominant. The

key disadvantages of HgCdTe are the uniformity, producibility, and cost. The growth of

high quality HgCdTe is difficult due to large defect densities and large non-uniformity as

the mole fraction of Hg is increased. The major noise mechanism are large tunneling

currents due to low electron effective mass. Commercially available MCT FPAs have

640 x 480 pixels in LWIR with 27 x 27 µm pixels and 640 x 512 pixels with a pixel size

of 24 x 24 µm in MWIR9. 1280 x 1024 pixel FPAs grown on germanium substrates with

square 15 x 15 µm pixels have been reported12. HgCdTe is the material system against

which other variable bandgap materials are judged.

InSb

Indium Antimonide (InSb) detectors are widely used for MWIR applications because of

the bulk properties of the material. The bandgap of InSb is 0.180 eV at room temperature

and 0.228 eV at 80K corresponding to λcutoff of ~7µm and 5.5µm respectively. Thermally

generated carriers are the dominant noise source at high temperature requiring InSb

detectors to be cryogenically cooled. Large uniform substrates are available allowing for

fabrication of large area FPAs (1024 x 1024). InSb detectors are used in forward looking

infrared heat seeking missiles and thermal imaging. The drawbacks of the material

system include the requirement of cooling, the material specific cutoff wavelength limit

and the tendency of the FPA characteristics to drift between uses.

Quantum Well Infrared Photodetector - QWIP

Quantum Well Infrared Photodetectors are a relatively new type of IR photodetector that

employs bandgap engineering and quantum size effects. Bandgap engineering is made

20
Chapter 1. Introduction

possible by the precise control of constituent material layer thickness and doping

concentration by using molecular beam epitaxy (MBE) and metal-organic chemical-

vapor deposition (MOCVD). Using bandgap engineering LWIR photodetectors can be

made using III-V materials whereas the bulk III-V material with the longest cutoff

wavelength is InSb with a λcutoff of 5.5µm. The most researched QWIP is made using

GaAs/AlGaAs. GaAs/AlGaAs QWIPs are grown on GaAs substrates due to small lattice

mismatch between AlGaAs and GaAs thus taking advantage of mature GaAs growth

technology. GaAs layers sandwiched between AlGaAs layers form quantum wells (QW).

The energy levels of the QW are determined by the height and width of the well.

Electrons in the ground state are photoexcited to the higher energy state. An applied bias

lowers the potential barrier allowing the electron to tunnel out of the well.

MWIR FPAs with 1028 x 1028 pixels with NETD of 17mK at 95K have been reported13.

Because of selection rules in quantum wells normal incidence light cannot be absorbed.

For this reason diffraction gratings must be used to couple light into the QWIP for

detection. The quantum efficiency for QWIPs (~20%) detectors is then much lower than

for HgCdTe and InSb detectors. Thermally generated and quantum mechanical tunneling

of carriers increase the dark current in QWIPs thus lowering the SNR, and limiting the

peak operating temperature.

1.7 InAs/Ga(In)Sb Type II Strain Layer Superlattice

Smith and Mailhiot introduced the application of InAs/GaInSb strain layer superlattice as

a possible alternative to HgCdTe in infrared imaging in 198714 using the semiconductor

superlattice introduced by Esaki et al 15 . The superlattice structure is made up of

21
Chapter 1. Introduction

alternating layers of thin semiconductor material. The electronic structure of the

superlattice depends on the layer thickness and periodicity of the constituent superlattice

material 16 . Technological advancements in epitaxial growth of semiconductors allow

precise control of layer thickness. The valence band of the higher bandgap energy GaSb

is above the conduction band of the lower bandgap energy InAs. This is what is known

as a type-II broken band offset and because of this band lineup the bandgap of the

superlattice can be smaller than the bandgap of either of the constituent materials. The

schematic band structure is shown in Figure 1.14.

The precise control of layer thickness and material composition enables engineering of

the electronic band structure. There have been many different approaches to theoretically

model superlattice semiconductor band structure. In the 1980s Smith and Mailhiot17,18

analyzed semiconductor superlattice band structure using k·p theory. Dente and Tilton

used the empirical psuedopotential method (EPM)19 for superlattice materials including

GaAs/AlGaAs, InAs/GaSb and AlSb/InAs/InGaSb/InAs and updated and applied it to

InAs/GaSb superlattices 20 later. Wei and Razeghi applied an empirical tight-binding

method with consideration of interface engineering to model type-II InAs/GaSb

superlattices.21

Using InAs/Ga(In)Sb material system the cutoff wavelength can be tailored over a wide

range from 3µm to 30µm. Different combinations of superlattice periodicity or

constituent material composition can be used to achieve the same superlattice bandgap.

InAs/Ga(In)Sb is seen as a possible material system for third generation FPAs. 22

InAs/Ga(In)Sb SLSs have high responsivity because of their strong absorption of normal

incidence light and because of this, higher QE than QWIPs. Also, InAs/Ga(In)Sb operate

22
Chapter 1. Introduction

at elevated temperatures and use established III-V technology. The advantages that

InAs/Ga(In)Sb may hold over HgCdTe include lower tunneling current since the band-

edge effective masses in InAs/Ga(In)Sb are not directly dependent on the band gap

energy and are larger than HgCdTe at the same bandgap. Another advantage is

commercially available GaSb substrates with good electro-optical homogeneity without

large cluster defects. Also, in this material system the probability of finding electrons

and holes are localized to InAs and Ga(In)Sb layers respectively. This suppresses Auger

recombination, which leads to longer carrier lifetime and higher QE 23 . Further

advancements are required in InAs/Ga(In)Sb SLSs especially for material growth,

processing, substrate preparation and device passivation.24

Figure 1.14 Superlattice schematic band structure and absorption process25

Device passivation is especially difficult as the device cutoff wavelength increases. The

spatial separation of electrons and holes decrease the optical absorption, but this can be

overcome by the introduction of InSb to the GaSb layers which increases the cutoff

wavelength without increasing the layer thickness.

23
Chapter 1. Introduction

1.7.1 Type II strain layer superlattice devices

Table 1.1 provides an overview of different superlattice single pixel detector performance

results.

Superlattice Type Device Type λc η D* Ref.

9.5 ML InAs / 12 ML GaSb n-on-p 5.4 µm @ 77K 30% 1013 26

8 ML InAs / 11 ML GaSb n-on-p 5 µm @ 300K 25% 109 27

9 ML InAs / 4 ML GaSb n-on-p 7 µm @ 77K 12-32% 1012 28

29
14 ML InAs / 7 ML GaSb n-on-p 11 µm 36%
13 ML InAs / 7 ML GaSb n-on-p 12 µm 54% 2x1011 30
31
W-Structure p-on-n 10.5 µm @ 78K 20-34%
14 ML InAs / 9 ML GaSb / 14 ML
InAs / 7 ML Al0.4Ga0.49In0.11Sb
32
M-Structure n-on-p 10.5 µm @ 77K 27%
13 ML InAs / 7 ML GaSb 6 µm
18 ML InAs / 3 ML GaSb / 5 ML
AlSb / 3 ML GaSb
33
12 ML InAs / 7 ML GaSb p-on-n 11 µm 20-35%

8 ML InAs / 8 ML GaSb nBn 5.2 µm @ 300K 18% 109 34

25
9 ML InAs / 10 ML GaSb n-on-p 5.4 µm @ 73 K 30%

25
8.6 ML InAs / 5 ML Ga0.75In0.25Sb n-on-p ~10 µm @ 77K

35
Dual Band 11.4 µm and 17
µm
36
Dual Band nBn 4.5 µm and
~8µm
Table 1.1 Superlattice device designs and parameters

24
Chapter 1. Introduction

1.8 nBn detector

The nBn detector introduced by S. Maimon and G. W. Wicks is a new class of infrared

photodetectors designed to improve detector performance by reducing the dark current

and increasing the operating temperature. 37 The nBn heterostructure design has

demonstrated a 100 K increase in background-limited infrared photodetection (BLIP) for

an InAs-based device, by decreasing Shockley-Read-Hall (SRH) generation currents and

by suppressing surface currents using specific processing.

The reported nBn detector consists of an n-type narrow bandgap absorbing layer (InAs), a

wide bandgap barrier layer (AlAsxSb1-x ) and an n-type narrow bandgap (InAs) top

contact layer.37 The wide bandgap layer was chosen so that there was zero valence band

offset where x~0.15. The barrier effectively blocks the flow of the majority carriers

between the electrodes but allows the photogenerated carriers to be collected. The

schematic band diagram of the nBn detector is given in Fig. 1.15.

Figure 1.15 Band diagram of InAs-based nBn structure, biased under operating and flatband

(inset) conditions34

25
Chapter 1. Introduction

There are three major dark current mechanisms in photodiodes based on narrow bandgap

semiconductors. The first is SRH generation current in the depletion region of a

photodiode. The second is radiative or nonradiative diffusion current in the extrinsic area.

And third is the surface current that is controlled by passivation of the device. The nBn

detector significantly reduces two of the three dark current sources. SRH generation is

very efficient in the depletion region of a pn junction because of the activation of trap

states and is considered the main source of dark current in MWIR photodiodes below

200K.38 The nBn detector depends on the heterostructure barrier to block the flow of

majority carriers through the device instead of the built-in barrier of the pn junction. The

nBn is designed to operate in either flatband or with very little depletion. The nBn

detector eliminates the pn junction so there is no depletion region and the device relies

primarily on diffusion of the carriers. The surface current in the nBn detector is

eliminated by the way the device is defined and processed. The nBn detector is etched

only to the barrier not to the bottom contact layer like in a conventional mesa etch. This

suppresses the surface current by eliminating the etched sidewalls. Also, the barrier then

acts as a passivation layer for the device and no further passivation is needed. The final

source of dark current in a pn junction is thermal generation in the neutral region of the

device. This current source is still present in the nBn detector and will not be suppressed.

This is a major advantage for the nBn detector over other detectors that are limited by

SRH when operating in BLIP conditions. The expressions for the current density for

SRH generation (JSRH) and diffusion (Jdiff) are

ni
J SRH ≈ q Wdep (1.23)
τ SRH

26
Chapter 1. Introduction

where ni is the intrinsic carrier concentration, Wdep is the depletion width, τSRH is the SRH

lifetime of the carriers which depends on the quality of the material, and

1 ni2 1
J diff ≈ qpn L=q L (1.24)
τ diff N d τ diff

where τdiff is the lifetime, L is the smaller of the width of the neutral region of the device

or the diffusion length of the minority carriers, pn is the hole concentration in the n-type

semiconductor and is equal to ni2/Nd. The activation energy of the SRH current is then

proportional to Eg/2 because of the dependence of the current on ni (ni~exp(-Eg/2kT)),

while the diffusion current is dependent on ni2 so that the activation energy is

proportional to the bandgap Eg (ni2~exp(-Eg/kT)). The two lifetimes are usually the same

order of magnitude and Wdep and L are comparable so as long as ni > ni2/Nd the pn

junction will be SRH limited. The nBn eliminates the SRH current and so its activation

energy is Eg whereas a pn junction has an activation of Eg/2. The factor of two difference

means that the nBn can operate with similar performance at about twice the temperature.

The nBn detector is a class of detector like the pn junction so it can be extended to other

material systems. Recently, the nBn design has been extended to InAs/GaSb superlattice

material utilizing an AlGaSb barrier layer, and an MWIR detector (cutoff wavelength

(λcutoff) at 5.2 µm) has been reported.32 Utilizing band gap tunability of InAs/GaSb SLS

material, the operation wavelength of the nBn detector based on such SLS can be

extended to the LWIR and VLWIR regime.

1.9 Conclusion

In this chapter the fundamentals of infrared detection were presented. IR detector figures

of merit were discussed in detail. Competitive third generation FPAs material systems
27
Chapter 1. Introduction

were covered with their disparate strengths and weaknesses, providing motivation for

continued research using other material systems. Superlattice detectors and their merits

as a possible alternative for third generation FPAs were outlined. This chapter provides

the foundation upon which all of the following chapters rest. The remaining chapters

cover the specifics of SLS IR detectors measurement and the research toward increasing

the operating temperature by using nBn superlattice detectors.

28
Chapter 2. Materials and Methods

2 Materials and Methods

This chapter covers the experimental methods required for testing of infrared

photodetectors. The first section covers Molecular Beam Epitaxy (MBE) technique

specifically for growth of InAs/GaSb SL. MBE growth of superlattice material is the first

step in the realization of IR photodetectors. After growth the next step is device

processing. A basic outline of processing is given while a detailed example is given is

Appendix A. The focus of this section will be the processing of single pixel devices.

Measurement techniques will be given for classifying the performance of single pixel

devices to find the spectral response and responsivity other and figures of merit

introduced in the previous chapter (λcutoff, Responsivity, QE, D*).

2.1 Molecular Beam Epitaxy (MBE)

The Molecular Beam Epitaxy (MBE) technique was invented by Alfred Cho of Bell

Telephone Laboratories in the late 1960’s. MBE is an Ultra High Vacuum (UHV)

system for growing single crystal materials on prepared substrates using elemental

sources. The mean free path of particles in UHV is much greater than the distance from

the source to the substrate so that there is no particle interaction with anything before

reaching the substrate. Due to the low growth rate (less than an atomic monolayer per

second) and the ability to shutter the beam of particles within a fraction of a second,

MBE improved the control of semiconductor growth so that epitaxial layers could be

deposited with good control of thickness, composition and doping.

29
Chapter 2. Materials and Methods

The material presented in this work is grown in a VG-80 MBE solid source two sided

reactor. This MBE system consists of a two identical growth chambers, a preparation

chamber and a load lock.

The load lock is used to bring in epi-ready substrates from the outside for growth or to

remove a grown sample for processing while maintaining the vacuum integrity of the

chambers. The load lock is a canister with seals on the topside towards the atmosphere

and the bottom facing the prep chamber. The canister contains slots for the substrate

holders to be placed. To introduce a sample the load lock must be dumped to atmosphere,

the sample inserted into the cassette and then pumped down to vacuum (1x10-8 Torr)

using a combination of a mechanical roughing pump and a turbomolecular pump.

The preparation chamber is used for substrate transfer, substrate storage and substrate

heating for initial outgassing. The preparation chamber is equipped with a trolley system

used to transfer samples under vacuum conditions to the growth chambers. The

preparation chamber is equipped with an ion pump to maintain vacuum conditions and an

ion gauge to monitor the vacuum level.

Each of the growth chambers contain all the necessary components needed for crystal

growth. To achieve UHV level (pressure ~10-10 Torr) the growth chamber is equipped

with an ion pump, a He closed loop cryopump and a Titanium sublimation pump. The

stainless steel shroud around the growth chamber is filled with Liquid Nitrogen (LN2).

This Cryo-shroud captures residual gases and is considered the most important pump

during growth. The sources of material for growth are conventional effusion and

dissociation effusion (cracker) cells each equipped with an individual shutter. The

manipulator assembly contains the substrate holder that heats and rotates the substrate,

30
Chapter 2. Materials and Methods

Figure 2.1 Schematic of a Molecular Beam Epitaxy growth chamber

thermocouple to monitor substrate temperature, and an ion gauge to monitor the beam

equivalent pressure. The growth chamber also includes an optical pyrometer for accurate

substrate temperature readings, a mass spectrometer to monitor residual gas levels (CO2,

H2O, N2 and CO), a Reflection High Energy Electron Diffraction (RHEED) gun and

fluorescent screen for in situ monitoring of the growth process. Schematic picture of

growth chamber is shown in Fig. 2.1.

Effusion cells are used to evaporate materials during growth. There are two different

types of effusion cells used, a conventional effusion cell and a dissociation (cracker)

effusion cell. The conventional effusion cell consists of a pyrolitic boron nitride crucible

that is surrounded by a heating element. Temperature of the cell is controlled by a

thermocouple. The crucibles are filled with group III elements (Al, Ga and In) as well as

dopant materials (Si, Be, GaTe). The flux rate is controlled by the temperature of the cell.

Dual filament cells offer separate temperature control of the tip and the base of the cell
31
Chapter 2. Materials and Methods

and are used to improve growth uniformity by reducing spitting. Dissociation effusion

(cracker) cells are used to reduce the group V (Arsenic and Antimony) sources tetramer

molecules (As4 and Sb4) to dimeric molecules (As2 and Sb2). Valved crackers allow

control of the Beam Equivalent Pressure without changing the temperature, but by

opening and closing a valve. Shutters in front of the sources are controlled by a computer

program are used to block the sources when they are not in use.

During the growth of InAs/GaSb SL the average lattice constant of the SL is different

from than the GaSb substrate. The slight lattice mismatch is reduced through the control

of heterojunction interfaces during SL growth. MBE technique allows careful control of

interfaces to produce the desired interface composition. The hetero-interface can be

approximated either “InSb-like” or “GaAs-like” interfaces (Ga/Sb/In/As or Sb/Ga/As/In

respectively). InSb causes compressive strain and GaAs caused tensile strain. Mixed

interfaces formed during InAs/GaSb SL growth are illustrated in Fig 2.2.

Growth Direction
Sb As
Ga In
Sb As
In Ga
As Sb
“InSb-like” interface “GaAs-like” interface

Figure 2.2 Schematic of possible interface

The alternating nature of the type of strain introduced either compressive strain for InSb

interface or tensile strain for GaAs interface alters the structural, optical and electrical

properties of the SLS. 39 To verify the high quality detector material growth, lattice

32
Chapter 2. Materials and Methods

mismatch between the SLS and the GaSb substrate has to be equal to zero. Proper choice

of shutter sequence during SLS growth (as illustrated in Fig. 2.3) enables compensation

of lattice mismatch between SL and substrate.

Figure 2.3 Shutter sequence for mixed interface growth

A typical cross-sectional Tunneling Electron Microscope (TEM) image of InAs/GaSb SL

grown on MBE is given in Fig. 2.4.

33
Chapter 2. Materials and Methods

Figure 2.4 Cross-sectional TEM image of type-II superlattice undertaken at JPL40

2.2 Superlattice device processing

After growth the detector sample is processed into a device. Single pixel devices are

used to approximate the performance of a FPA processed out of the same material. The

steps required for conventional single pixel processing are described below. Basic

processing sequence can be divided into three steps, top contact metallization, mesa

isolation etch and bottom contact metallization. See Appendix A for complete processing

details.

Optical photolithography is used for the metal pattern definition. The typical steps in a

photolithographic process are the following: wafer cleaning, dehydration bake, adhesion

34
Chapter 2. Materials and Methods

promoter application, photoresist application, softbake, exposure and develop cycle. The

wafer is cleaned using acetone, isopropanol (IPA) and de-ionized (DI) water. Then

dehydration bake is performed. After pretreatment, the sample is primed with

hexamethyldisilazane (HMDS) on a spinner to promote adhesion and then baked. The

next step negative photoresist is spun on and then baked. Using Mask 1 in Fig. 2.5(a) the

sample is then exposed on the mask aligner, then baked, then flood exposed again

without the mask for resist reversal. After the final exposure, the sample was developed

leaving the photoresist in the desired places for metal deposition. After photolithography,

the metal contact is deposited. We used Ti (500 Å)/Pt (500 Å)/Au (3000 Å) to achieve

ohmic contact.

The next step is the mesa isolation etch. The photolithographic steps for the mesa

isolation etch are similar to the steps used for top contact metallization. First the sample

is pretreated and primed. Then positive resist was spun on and baked. Using Mask 2 in

Fig. 2.5(b) the sample was exposed on the mask aligner. Finally the sample was placed

in developer diluted in four parts water and developed.

The mesa isolation etch used wet chemical etching and etched to the bottom contact layer.

The chemical composition is

H 3 PO4 : H 2O2 : H 2O = 1 : 2 : 20 (2.1)

with an etch rate of 0.08 µm/min. The surface quality after etch shows no considerable

degradation. After the mesa isolation etch, photoresist is applied as it was for the top

contact metallization except Mask 3 in Fig. 2.5(c) is used. The bottom contact metal is Ti

(500 Å)/Pt (500 Å)/Au (3000 Å) is then applied. Finally the sample is soaked in acetone

for lift-off.

35
Chapter 2. Materials and Methods

Figure 2.5 Mask designs used in fabrication of single pixel detectors

The final mesa size of a single-pixel device is 410 x 410 µm. The aperture diameters vary

in size and are 300, 200, 150, 100, 50 and 25 µm. The schematic of the fully processed

single-pixel photodiode with contact metallization is shown in Fig. 2.6.

Metallization

P-Contact

Absorber

N-Contact
Substrate

Figure 2.6 Conventionally defined PIN photodiode

2.3 Characteristics of Detector Performance

The processed devices were cleaved or individually diced and wire-bonded to a leadless

chip carrier (LCC). To evaluate the electrical and optical performance of photodetectors

the device to be tested was placed in a close-cycle Helium cryostat. Temperature

36
Chapter 2. Materials and Methods

dependent IV characteristics of the devices were measured using a HP4145

Semiconductor Parameter Analyzer. Measurements were performed in 10K-300K

temperature range. Dynamic impedance-area product (RdA) is given by

−1
 dI 
Rd A = A  (2.2)
 dV 

where A is the mesa area of the detector and the derivate is calculated from the measured

IV data. For BLIP conditions, the background of the measurement determines the

photocurrent generated. The orientation of the cold shield inside the cryostat can be

either open to the room temperature background (room temperature background IV) or

can be rotated so that the detector sees a cold radiation shield (dark current IV). As long

as the detector is not background limited, there is little difference between the RT IV and

the dark current.

The spectral response of the sample is measured using a Fourier Transform Infrared

(FTIR) spectrometer. The FTIR contains a glow bar source, a Michelson interferometer

with a movable mirror, an internally mounted Deuterated triglycine sulfate (DTGS)

detector and mirrors to control the direction of the IR beam. The IR radiation from the

glow bar source is put through the Michelson interferometer with the movable mirror to

produce an interference pattern (which is the Fourier transform) of the spectrum of the IR

source. For the spectral response measurement, the IR beam is directed outside of the

FTIR off of a parabolic mirror and onto the detector mounted in the close-cycle He

cryostat. The signal from the detector is then amplified by a preamplifer and fedback to

the FTIR. Using computer software the inverse fourier transform is performed and the

spectral response curves are then graphed and saved.

37
Chapter 2. Materials and Methods

The DTGS detector measures the background source power spectrum and has a spectral

response that is essentially flat. However, the intrinsic slow response of the detector

combined with the high frequency modulation at the fringes of the interferometer signal

cause the output of the detector to decrease as the modulation of light increases even

above just a few hundred Hertz.41 This creates a wavelength dependent gain in the DTGS

detector. To create a gain curve, the output signal of the DGTS detector was measured

with a 980 nm diode laser directly illuminating the DTGS detector. The laser was

modulated using a sine wave source from 400 Hz to 10 kHz. Using the equation

f = nv (2.3)

where f is the frequency (Hz), n is the wavenumber (cm-1) and v is the velocity (cm/s) of

the mirror, the DTGS detector correction curves were constructed as a function of

wavenumber as shown is Fig 2.7. The DTGS correction instruction manual is given in

Appendix B.

1.0 Velocity
.1581 cm/s
.3165 cm/s
.4747 cm/s
0.8 .6329 cm/s
.9494 cm/s
1.2659 cm/s
Normalized Signal

0.6

0.4

0.2

0.0

7000 6000 5000 4000 3000 2000 1000

-1
Wavenumber (cm )

Figure 2.7 DTGS detector correction curves for different FTIR mirror velocities

38
Chapter 2. Materials and Methods

The background spectrum (Fig. 2.8) is then divided by the gain curve (Fig. 2.9) to give a

corrected background (Fig. 2.10). The spectral response of the SL detector (Fig. 2.11) is

then divided by the corrected background and normalized to give the relative spectral

response of the SL detector (NR(λ)) (Fig. 2.12). This spectral response is then used to in

combination with the other radiometric measurements like the black body responsivity to

determine the figures of merit for the detector.

Background

25
Spectral Response (A.U.)

20

15

10

7000 6000 5000 4000 3000 2000 1000

-1
Wavenumbers (cm )

Figure 2.8 Background spectrum measured by DTGS detector

39
Chapter 2. Materials and Methods

DTGS Frequency Response

1.2

1.0

Spectral Response (A.U.)

0.8

0.6

0.4

0.2

0.0

7000 6000 5000 4000 3000 2000 1000

-1
Wavenumbers (cm )

Figure 2.9 Correction Curve as a function of wavenumber

180

160

140
Spectral Response (A.U.)

120

100

80

60

40

20

-20
7000 6000 5000 4000 3000 2000 1000
-1
Wavenumbers (cm )

Figure 2.10 Background spectrum after division by correction curve

40
Chapter 2. Materials and Methods

SLS nBn Response at 77 K

15

Spectral Response (A.U.)

10

0
7000 6000 5000 4000 3000 2000 1000
-1
Wavenumbers (cm )

Figure 2.11 Superlattice detector sample spectral response without correction

1.0

0.8
Spectral Response (A.U.)

0.6

0.4

0.2

0.0
7000 6000 5000 4000 3000 2000 1000
-1
Wavenumbers (cm )

Figure 2.12 Final normalized and corrected SL detector spectral response (NR(λ))

41
Chapter 2. Materials and Methods

2.3.1 Responsivity and D*

Responsivity for detectors is defined as the amount of current produced by the incident

radiant power flux as given in Eq. 2.3.

signal
ℜ= (2.3)
φe

The signal of the detector is a measured quantity but the incident photon flux is estimated

using radiometry. One way to find the spectral responsivity of the SL detector is to take

the normalized spectral response (NR(λ)) as measured in Section 2.3 and multiply it by

the peak responsivity (Rmax).

ℜ(λ ) = Rmax NR (λ ) (2.4)

The experimental setup for responsivity measurements is shown in Fig. 2.13. This is

used to measure the detector signal. The infrared radiation from the calibrated blackbody

source at a temperature set by the temperature controller is radiated through a chopper

onto the SL sample that is mounted in a pour filled dewar and cooled with LN2. The

chopper is used to periodically modulate our source so that the signal of the detector can

be treated experimentally as the sum of a number of separate sinusoidal waves. The

result is what is known as a Fourier series. The first term of the series is the fundamental

frequency and is the same as the frequency of the chopper3. The electrical signal

produced by the detector is then amplified by the current amplifier. The signal of the

fundamental frequency is then measured using a FFT network analyzer. The radiant

power (φe) in watts from the blackbody on the detector can be estimated by

λ cut
Adet
φe = ∫ M λ (λ ,T ) ⋅ A
0
e, bb ⋅
r2
dλ (2.5)

42
Chapter 2. Materials and Methods

FFT
Blackbody Network
Source Current
Amplifier Analyzer
Chopper

Dewar

Chopper
Blackbody Controller
Temperature
Controller

PC

Figure 2.13 Responsivity measurement setup

where Me is the radiant exitance, Abb is the area of the blackbody aperture, Adet is the area

of the detector and r is the distance from the blackbody to the detector as shown in Fig.

2.14. The responsivity can then be calculated using the definition of detector

responsivity given in Eq. 2.3, the spectral responsivity given in Eq. 2.4 and the estimate

for the radiant power from the blackbody incident on the detector given in Eq. 2.5.

λ cut
Adet
I ( Amps ) = ∫ M λ (λ ,T ) ⋅ A
e, bb ⋅ ⋅ NR(λ ) ⋅ ℜmax dλ (2.6)
0
r2

The solving for the constant Rmax in Eq. 2.6 gives

I
ℜmax = λ cut
Adet (2.7)
Abb ⋅
r2 ∫ M λ (λ ,T ) ⋅ NR(λ )dλ
0
e,

with units of Amps/Watt.

From equations given in Chapter 1 the detector figures of merit can then be found. From

Eq. 1.11 the quantum efficiency as a function of wavelength can be found.

43
Chapter 2. Materials and Methods

ℜmax NR(λ ) hc
η (λ ) = (2.8)
λ q

r
Blackbody
Detector

Adet
Abb

Figure 2.14 Blackbody source and detector configuration for testing

D*(λ) can be found using

ℜ(λ ) Ad
D* = (2.9)
in / ∆f

where in is the current noise measured by the spectrum analyzer in the when the

blackbody is no longer incident on the detector and ∆f is given by

1
∆f = (2.10)
2τ AVG

where τAVG is the integration time of the measurement.

An alternative approach for measuring responsivity uses a reference detector to estimate

the radiometric quantities using the equation

A ⋅ F ∫ M BB (λ , T ) ⋅ ℜ(λ )dλ = Signal [ Amps ] (2.11)

where A is the area of the detector (cm), F is a geometrical factor that takes into account

the measurement setup geometry, and MBB(λ,T) (W/(cm2·µm)) is the spectral excitance

44
Chapter 2. Materials and Methods

given in equation 1.1. Similar to the above approach using equation 2.11 and rearranging

the terms from the integral in equation 2.4 and solving for the constant Rmax

Signal  A
Rmax =   (2.12)
A ⋅ F ∫ M BB (λ , T ) ⋅ NR (λ )dλ  W 

2.4 Conclusion

In this chapter we discussed the growth, processing and characterization of an InAs/GaSb

SL infrared single-pixel detector. Detailed description of characterization techniques

used to access the optical and electrical properties of SL detectors was provided. In the

next chapter a specific example of an InAs/GaSb SL device will be reported.

45
Chapter 3. InAs/GaSb SLS based MWIR detector using nBn design

3 InAs/GaSb SLS based MWIR detector using

nBn design

As stated in Chapter 1, the nBn detector is a new class of infrared photodetector. The

concept of the nBn design can be extended to a variety of other material systems. In this

chapter we will discuss adapting of the nBn design to InAs/GaSb superlattice.

3.1 Dark current reduction in nBn SL detector

With the exception of our research group, all InAs/GaSb detectors are based on the

photodiode design (see table 1.1). At cryogenic operating temperature, the major noise

sources for photodiodes are Shockley-Read-Hall (SRH) recombination and surface states.

In order to eliminate surface leakage currents different passivation methods have been

investigated, such as the deposition of relatively thick layer of dielectric material 42 ,

polyamide passivation 43 , overgrowth of wide bandgap material 44 , deposition of

passivating sulphur coating electrochemically45 and using ammonium sulfide chemical

solutions46. All of these methods add complexity to the fabrication and the long term

effectiveness is still questionable.

The nBn detector design has been shown to eliminate SRH recombination in bulk

semiconductor detectors37. Special processing of the nBn detector allows the device to be

etched only to the barrier layer and removes the lateral mesa sidewalls.

46
Chapter 3. InAs/GaSb SLS based MWIR detector using nBn design

3.2 nBn Detector Growth

The structures presented in this chapter were grown on n-type Te-doped GaSb (001)

substrates by solid source molecular beam epitaxy in a VG-80 reactor equipped with a

arsenic valved cracker source, an antimony cracker source and Ga and In SUMO® cells.

A detailed description of the growth procedure has been reported elsewhere39. The

device consists of a 0.35 µm thick n-type superlattice bottom contact layer composed of 8

ML InAs doped with Si (n=4 x 1018 cm-3) by 8 ML GaSb followed by the non-

intentionally doped (n.i.d.) 1.4 µm thick 8 ML InAs/8 ML GaSb absorber layer. The

barrier layer consists of Al0.2Ga0.8Sb with layer thickness of 100 nm. The function of the

heterostructure AlGaSb barrier is to block the flow of majority carriers through the

device. The barrier composition was chosen so that the majority carrier current (electron

current) is blocked while the minority carrier current (hole current) can flow unimpeded

because the valence band offset with the surrounding SL layers is minimized. The

energy offset in the conduction band is designed and grown to be a thick enough and

large enough barrier to prevent electron tunneling from the top contact layer to the

absorber layer. The barrier layer is desired to have a zero valence band offset. Due to the

complicated nature of SL effective band lineup and the uncertainty involved with SL

modeling the exact band offsets are difficult to determine. The best valence band offset

and barrier material composition are still being researched. The structure was capped by

100 nm thick top contact layer with the same superlattice composition, thickness and

doping concentration as the bottom contact layer. The proceeding structure was grown

above an AlGaSb etch stop layer that is used for substrate removal after FPA

hybridization.

47
Chapter 3. InAs/GaSb SLS based MWIR detector using nBn design

SL (n) 35 nm

Al0.2GaSb 100 nm

Barrier

SL nid 1.4 µm

SL (n) 35 nm

Al0.2GaSb 100 nm

GaSb:Te 2”

Substrate

Figure 3.1 Heterostructure schematic of InAs/GaSb SLs MWIR detector after growth

3.3 Contact metallization Study

The nBn structure is designed to operate with n-type layers in flatband or with little

depletion. Since SRH generation-recombination current is associated with regions

depleted of carriers the heavily doped contact layers required to make ohmic contacts

create n-n+ homojunctions. This space-charge region could contribute to the dark current.

In order to reduce SRH current, the doping concentration in the contact layer needs to be

decreased while still maintaining a good ohmic contact.

To optimize the doping level of the contact layers, several ~ 0.5 µm thick structures

consisting of 8 ML InAs doped with Si/8 ML GaSb SL with different doping levels were

grown. Doping concentrations as a function of dopant cell (Si) temperature is given in

the inset of Fig. 3.2. Transmission line method (TLM) patterns were defined by

conventional UV lithography. Metal pads with dimensions of 50 µm x 100 µm spaced

from 10 µm to 70 µm apart by increments of 10 µm were used. Commonly used Ti (500

Å)/Pt (500 Å)/Au (3000 Å) metallization was used for the contacts. Samples were
48
Chapter 3. InAs/GaSb SLS based MWIR detector using nBn design

annealed using the following conditions: (I) no annealing (II) 260 °C for 1 minute (III)

350 °C for 1 minute and (IV) 385 °C for 1 minute. Results are summarized in Fig. 3.2.

Ti/Pt/Au metallization on undoped SLs material (n=3x1016 cm-3) did result in ohmic

behavior. Contact resistance was found to be ~55 Ω regardless of annealing temperature.

1000
1E19

3
Si concentration, /cm
1E18

100
1E17
Rc (Ω)

0 950 1000 1050 1100 1150

o
Si cell temperature, C

10
no annealing
ο
260 C, 1 min
ο
350 C, 1 min
ο
385 C, 1 min
1
1E16 1E17 1E18 1E19
-3
Dopant Concentration (cm )

Figure 3.2 Contact resistance versus the doping concentration in the contact layer at different

annealing temperatures for Ti/Pt/Au contacts on n-type InAs/GaSb superlattice Inset: The doping

concentration as a function of the Si cell temperature

With increasing level of doping concentration in the SL layer, ohmic behavior in the

current-voltage characteristics was observed and contact resistance was improved. The

structure with the highest doping level (n = 4 x 1018 cm-3) demonstrated contact resistance

equal to 3.7 Ω (no annealing) and 2.4 Ω (annealing at 350 °C for 1 minute). Annealing at

380 °C did not considerably improve the contact resistance but the surface metal layer

was degraded. The results indicate that the doping concentration needs to exceed 1 x 1018

49
Chapter 3. InAs/GaSb SLS based MWIR detector using nBn design

cm-3 in order to achieve ohmic contact to n-type InAs/GaSb SLs. In order to obtain

ohmic contacts to the InAs/GaSb SL contact layer at lower doping concentration a

different metallization needs to be used. GeAu-based metallization on n-type GaSb has

been shown to provide good ohmic contact by combining the solid phase reactions of Ge-

based contacts with the low ohmic resistance of the Au-based contacts47. For the contact

resistance study a ~0.5 µm thick 8 ML InAs doped with Si/8 ML GaSb (n = 5 x 1017 cm-
3
) SL sample was grown. Ge/Au/Pt/Au (287 Å/568 Å/504 Å/2000 Å) contact was

deposited then the sample was annealed at 380 °C for 1 minute. TLM measurements

revealed contact resistance equal to 1 Ω which indicates that there is good ohmic contact.

While this level of contact resistance is desirable the surface quality after annealing is not

ideal. Good metal-semiconductor contacts are difficult to achieve with lower

semiconductor doping. While there were some promising results from the metal contact

study no conclusive result was reached. Therefore the devices presented in this chapter

will use highly doped (n=4 x 1018 cm-3) contact layers with Ti (500 Å)/Pt (500 Å)/Au

(3000 Å) metallization. Since the post-annealing contact resistance improved by a factor

of 1.5, the heat treatment of the contacts was omitted.

3.4 nBn Detector Device Processing

The devices were processed into two distinct structures (referred to as structure A and

structure B) using 410 µm x 410 µm square mesas with circular apertures ranging from

25 µm to 300 µm. Processing was initiated by standard optical photolithography for top

contact metal deposition. Ti (500 Å)/Pt (500 Å)/Au (3000 Å) was used for both top and

bottom contact metallization. Structure A was etched using H3PO4:H2O2:H2O solution

50
Chapter 3. InAs/GaSb SLS based MWIR detector using nBn design

with a ratio of (1:2:20) to the middle of the barrier layer (etch depth ~ 0.15 nm). Then

both structures A and B were inductively coupled plasma dry etched to the middle of the

bottom contact layer. The bottom contacts for structure A were significantly far away

from the top contact so that the etched sidewall surface current would not contribute to

the device current. Both samples were then patterned and bottom contact metal was

deposited. Schematics of structures A and B are shown in Fig. 3.3 and Fig. 3.4

respectively.

Figure 3.3 Schematic of structure A (shallow etch)

SL n 100 nm
Al0.2GaSb 100 nm

SL nid Absorber

SL n 360 nm

GaSb:Te Substrate

Figure 3.4 Schematic of structure B (deep etch)

51
Chapter 3. InAs/GaSb SLS based MWIR detector using nBn design

For structure A the active area of the device is defined by the diffusion length of the

minority carriers (holes) rather than by a conventional mesa. For structure B dangling

bonds are present on the etched mesa sidewalls and surface leakage current is expected to

be high.

3.5 nBn Detector Device Characterization

After processing, the devices were cleaved then mounted and wire bonded to a leadless

chip carrier (LCC) for characterization. Current voltage characteristics were measured in

the range 77K - 293K range with the detector facing a cold shield using a HP4145

Semiconductor Parameter Analyzer. In this work the forward bias of an nBn detector is

defined as a positive voltage applied to the bottom contact of the detector. The dark

current densities measured at V = 0.1V for structures A and B as a function of

temperature are shown in Fig. 3.5 and the dark current densities as a function of applied

voltage is given in Fig. 3.6.

The ratio of dark current densities measured at the same value of applied bias is also

shown. At high temperature the thermally generated carriers dominate the dark current

and the influence of the surface current is not seen because of the large size of the mesa.

The bulk component of dark current is strongly temperature dependent so as the

temperature decreases the surface current becomes the dominant component of dark

current at low temperature. The levels of dark current in structures A and B are

comparable at 250 K but the dark current in structure A is reduced by two orders of

magnitude at 77 K. Dark current densities were equal to 2.3 x 10-6 A/cm2 and 3.1 x 10-4

52
Chapter 3. InAs/GaSb SLS based MWIR detector using nBn design

A/cm2 for structures A and B respectively at Vb = 0.1V and T = 77K. At 250K, the dark

current density was equal to ~ 80 mA/cm2.

Temperature (K)
300250 200 150 100
1000
Dark Current Density (A/cm )
2 V = 0.1V
0.1
Structure A
Structure B

Ratio (JDeep)/(JShallow)
0.01 Ratio
100

1E-3

1E-4 10

1E-5

1E-6 1
4 6 8 10 12 14
-1
1000/T (K )

Figure 3.5 The dark current density vs. 1000/T in structures A (shallow etched) and B (deep

etched) (Vb = 0.1V). Right Axis: Ratio of dark current densities for both structures

53
Chapter 3. InAs/GaSb SLS based MWIR detector using nBn design

100

10

0.1

J (A/cm )
2
0.01

1E-3

1E-4
Structure A
1E-5 Structure B
1E-6 T=77K
1E-7

1E-8
-3 -2 -1 0 1 2 3

Bias (V)

Figure 3.6 Current Density vs. Voltage for structures A (shallow etched) and B (deep eteched)

at 77 K

To confirm that the surface current is reduced measurements were undertaken at

industrial collaborators Santa Barbara Focal Plane (SBFP). The samples were hybridized

to a silicon fanout using compression indium bump bonds. To separate the surface and

bulk contributions to the current, dark current measurements were performed using

Variable Area Diode Array (VADA) with device size varying from 270 µm to 30 µm.

The surface leakage current is due to the mesa sidewalls so as the ratio of perimeter to

area (P/A) increases the contribution of the surface leakage current increases. To show

that the majority of the current is flowing through the bulk of the semiconductor, there

should be little change in the current density as the P/A increases. The current density is

chosen since its square root is proportional to the limiting Shot Noise contribution. The

current density vs. P/A as shown in Fig. 3.7 is not always completely flat as might be

54
Chapter 3. InAs/GaSb SLS based MWIR detector using nBn design

expected with complete passivation. This could be due to lateral collection because the

devices do not have physically defined areas the devices are collecting carriers from the

surrounding area. The lateral collection would present itself in a similar way as surface

current, as the size of the device decreases and the lateral collection stays the same the

ratio of the device current to the lateral collection would increase. Also, the processed

devices still have exposed sidewalls which can contribute to the dark current. For

complete analysis of the elimination of the surface leakage currents, a comparison of

deep and shallow etched VADA at different temperatures must be performed.

Device Side Length (cm)

0.04 0.02 0.003
10
)
2
Current Density at V=0.5 V(A/cm

0.1

0.01

1E-3

1E-4
77 K
100 K
1E-5
120 K
220 K
1E-6
200 400 600 800 1000 1200 1400
-1
Perimeter/Area (cm )

Figure 3.7 Current density vs. Perimeter/Area at V = 0.5V performed by Santa Barbara Focal

Plane

Temperature dependent spectral response of structure A was measured using a Nicolet

6700 Fourier Transform Infrared (FTIR) Spectrometer from 77K – 250K. The spectral

response was seen all the way up to 250K (Vb=0.5V). The cut-off wavelength shifted

55
Chapter 3. InAs/GaSb SLS based MWIR detector using nBn design

from ~ 4.2 µm at 77K to ~4.8 µm at 250K. The responsivity and quantum efficiency

(QE) were measured for structure A as shown in Fig. 3.7.

1.0 40
Responsivity

Quantum Efficiency (%)

QE 35
0.8
Responsivity (A/W)
Vb = 0.5V
30
T=77 K
0.6 25

20
0.4 15

10
0.2
5

0.0 0
2 3 4 5 6
Wavelength (µm)

Figure 3.8 The responsivity (solid black line, left axis) and quantum efficiency (dotted red line,

right axis) for structure A (shallow etch)

For the responsivity measurement, the device was placed in a pour-fill dewar and cooled

with liquid nitrogen. The peak responsivity was calculated using a calibrated Micron

blackbody source at 500K. The signal was collected using a Stanford Research Systems

Model SR570 Low-Noise Current Preamplifier and Model SR770 FFT Network

Analyzer. The peak responsivity is then calculated using the normalized spectral

response and the peak black body response. For structure A the peak responsivity was

found to be 0.835 A/W at 3.5 µm with Vb = 0.5V. The responsivity and QE at 4 µm was

0.74 A/W and 23% respectively. No Antireflection (AR) coating was applied on the

device.

The spectral specific detectivity D*(λ) was estimated using

56
Chapter 3. InAs/GaSb SLS based MWIR detector using nBn design

ℜ(λ )
D* =
4kT
2qJ (V ) +
(3.1)
Rd Ad

where R(λ) is the spectral responsivity, and 2qJ(V) estimates the shot noise and 4kT/RdAd

estimates the Johnson noise at a given voltage. Peak D* was estimated to be 2.8 x 1011

cm Hz1/2/W at 4 µm at Vb = 0.5 V.

3.0
1.2 R
D*
2.5
1.0

D* x10 cm Hz /W
Responsivity (A/W)

1/2
0.8 2.0

0.6
1.5

11
0.4
1.0
0.2

0.0 0.5

0 1 2

Voltage (V)

Figure 3.9 Peak Responsivity and D* as a function of applied bias for structure A (shallow etch)

3.6 Superlattice nBn 320 x 256 Focal Plane Array

In collaboration with QmagiQ LLC, we produced a 320 x 256 InAs/GaSb superlattice

nBn focal plane array. The SEM image in Fig. 3.10 shows the shallow etched top

contacts and deep etched bottom contacts performed at CHTM. The noise equivalent

temperature difference (NETD) is given in Fig. 3.11. The mean NETD was 23.8 mK at

77 K for an integration time of 16.3 ms using f/4 optics. Fig. 3.12 is a thermal image

57
Chapter 3. InAs/GaSb SLS based MWIR detector using nBn design

taken with the 320 x 256 InAs/GaSB superlattice nBn camera with detector temperature

at 77 K and an integration time of 16.3 ms, a two-point non-uniformity correction was

utilized.

Shallow etch
30µ m

Deep etch

Figure 3.10 Scanning Electron Microscope (SEM) image showing part of a fully processed

Focal Plane Array with a shallow etched top contacts and deep etched bottom contacts with

indium bumps

Figure 3.11 The Noise Equivalent Temperature Difference (NETD) distribution in the FPA for

an integration time of 16.3 ms using f/4 optics

58
Chapter 3. InAs/GaSb SLS based MWIR detector using nBn design

Figure 3.12 Thermal image taken with at 77 K with and integration time of 16.3 ms and using a

two-point non-uniformity corretction

3.7 Conclusion

In this chapter, we discussed high performance nBn detector based on InAs/GaSb SL

with a cut-off wavelength of ~ 4.8 µm at 250K. Due to the way the device was defined

and processed, low temperature dark current was reduced by two orders of magnitude due

to elimination of surface currents in comparison with conventional photodiode processing.

Dark current densities were equal to 2.3 x 10-6 A/cm2 and 3.1 x 10-4 A/cm2 for detectors

with shallow etch defined mesa (structure A) and conventionally defined mesa (structure

B), respectively. At 250K the dark current density was ~ 80 mA/cm2 for both devices.

VADA measurements revealed that the surface current does not increase linearly with the

increase of the P/A ratio. Comparison of conventional mesa etched and shallow isolation

etched VADA current density at operational temperature would clearly show whether the

surface currents were eliminated. Removal of surface leakage currents is important for

59
Chapter 3. InAs/GaSb SLS based MWIR detector using nBn design

fabrication of high performance focal plane arrays (FPAs) based on InAs/GaSb SL

operating in MWIR.

60
Chapter 4. InAs/GaInSb SLS based detector using nBn design with 8 µm cutoff

4 InAs/GaInSb SLS based detector using nBn

design with 8 µm cutoff

In this chapter we discuss the implementation of InAs/GaInSb strain-layer superlattice

detectors with cutoff wavelength of 8 µm. The motivation for why surface leakage

current is the limiting factor in development of large area LWIR photodetector FPAs is

discussed. Then growth, processing and characterization of an 8 µm cutoff InAs/GaInSb

nBn detector much is detailed much like as in chapter 3. IV characteristics of the nBn

detector are compared with a PIN photodiode with the same absorber region structure.

The nBn exhibits significantly lower current density than the PIN diode at equivalent

applied electric field. The special processing of the nBn detector and the removal of the

sidewall surface is effective regardless of the increase in the cutoff wavelength of the

detector.

4.1 Surface leakage current in InAs/GaSb LWIR

The surface leakage current of low bandgap devices increases when the bandgap of the

device is lowered. Most of the surface leakage is due to tunneling of electrons. In most

higher bandgap (> 1 eV) III-V devices, dielectric materials like silicon dioxide or silicon

nitride are used for passivation. The interaction between the insulator-semiconductor

interface affects the Fermi level at the boundary of the device by electrostatic interaction

and can cause band bending of ~ 10 meV. For narrow bandgap semiconductors like the

120 meV bandgap corresponding to 10 µm cutoff band bending caused by interface states

results in a strong increase of surface tunneling currents44.

61
Chapter 4. InAs/GaInSb SLS based detector using nBn design with 8 µm cutoff

4.2 Growth and Processing

The structures presented in this chapter were grown on n-type Te-doped GaSb (001)

substrates by solid source molecular beam epitaxy in a VG-80 reactor equipped with a

arsenic valved cracker source, an antimony cracker source and Ga and In SUMO® cells.

Two separate devices were grown an nBn detector and a PIN detector in subsequent

growth runs. The schematic structures are shown in Fig. 4.1. The nBn device (sample A)

consists of a 0.35 µm thick n-type superlattice bottom contact layer composed of 8 ML

InAs doped with Si (n=4 x 1018 cm-3) by 8 ML GaSb followed by the non-intentionally

doped (n.i.d.) 2.35 µm thick 9 ML InAs/5 ML Ga0.75In0.25Sb absorber layer. The

absorber superlattice period structure was taken from Ref. 44 and the expected cutoff

wavelength was 10 µm at 77 K. The barrier layer consists of Al0.2Ga0.8Sb with layer

thickness of 100 nm. A barrier that gives zero valence band offset for this superlattice

structure is under experimental and theoretical examination in our group. The structure

was capped by 100 nm thick top contact layer with the same superlattice composition and

doping concentration as the bottom contact layer.

The PIN device (sample B) consists of 15 periods of 8 ML InAs:Si/8 ML GaSb SL

bottom contact layer then 15 periods 9 ML InAs:Si/5 ML GaInSb for the n-type layer.

The absorber layer shares the same size and growth as the absorber for the nBn and is

2.35 µm thick n.i.d. 9 ML InAs/5 ML GaInSb. The p-type side is 15 periods

InAs/GaInSb:Be and is capped by a 100 µm GaSb:Be top contact layer.

The nBn device was processed using the standard optical photolithography and the

devices were defined with a shallow isolation etch as described in chapter 3. The PIN

device was processed into a conventional mesa structure as defined in chapter 2. Both

62
Chapter 4. InAs/GaInSb SLS based detector using nBn design with 8 µm cutoff

samples A and B were not passivated so that the effect of the surface current would be

evident.

Sample A Sample B

SL (n) 8x8 35 nm GaSb:Be 100 µm

Al0.2GaSb 100 nm SL (p) InAs/GaInSb:Be

X 15
Barrier
SL nid 2.35 µm SL nid 2.35 µm
InAs/Ga0.75In0.25Sb InAs/Ga0.75In0.25Sb
Absorber Absorber

SL (n) 8x8 35 nm SL (n) InAs:Si/GaInSb

X 15
Al0.2GaSb 100 nm SL (n) 8x8 X15

GaSb:Te 2” GaSb:Te 2”

Substrate Substrate

Figure 4.1 Schematic structure of nBn (Sample A) and PIN (Sample B) 8 µm λcutoff detectors

4.3 Device characterization and comparison

After processing the samples were individually wire bonded to an LCC and mounted in a

cryostat as described in Chapter 3. Temperature dependent IV characteristics with room

temperature background illumination were performed for samples A and B and are shown

in Fig 4.2 and Fig. 4.3 and are compared on the same graphs in Fig. 4.4. The current

density at 0.2V is 6.24 x 10-5 A/cm2 for the nBn and for the PIN at -0.2V is 1.03 A/cm2

both at T = 77 K as shown in Fig. 4.5. While both devices were not optimized, the nBn

device exhibited lower current density while the PIN is dominated by a large noise

current. The significantly lower current density for the nBn detector compared to the PIN

diode can be attributed to both the efficiency of the device structure and the special

processing that removes the effect of the surface currents. Enough though the PIN device

63
Chapter 4. InAs/GaInSb SLS based detector using nBn design with 8 µm cutoff

is by no means comparable to published state of the art detectors at this cutoff

wavelength, the fact that both samples were grown in consecutive growth runs and was

processed in the same clean room at the same time show the advantages provided by the

nBn device.

100

10

0.1
J (A/cm )
2

0.01

1E-3
50 K
1E-4
77 K
1E-5 100 K
150 K
1E-6 200 K
250 K
1E-7 293 K
1E-8
-2 0 2
Bias (V)

Figure 4.2 Room temperature background JV plotted at different temperatures for sample A

(nBn)

64
Chapter 4. InAs/GaInSb SLS based detector using nBn design with 8 µm cutoff

100

10

0.1

J (A/cm )
2
0.01

1E-3
50 K
1E-4
77 K
1E-5
100 K
150 K
1E-6 200 K
250 K
1E-7 293 K
1E-8
-2 0 2
Bias (V)

Figure 4.3 Room temperature background JV plotted at different temperatures for sample B

(PIN)

100

10

1
J (A/cm )

0.1
2

0.01

1E-3

1E-4
77 K nBn
1E-5 250 K nBn
77 K PIN
1E-6 250 K PIN

1E-7
-3 -2 -1 0 1 2 3

Bias (V)

Figure 4.4 Comparison of sample A (nBn) and B (PIN) current density as a function of applied

bias at T = 77 K and 250 K

65
Chapter 4. InAs/GaInSb SLS based detector using nBn design with 8 µm cutoff

Temperature (K)
250200 150 100 50
100
nBn at 0.2 V
10 PIN at -0.2 V

Current Density
0.1

0.01

1E-3

1E-4

1E-5
50 100 150 200 250

1/kT (eV)

Figure 4.5 The current density for sample A (nBn, black) and B (PIN, red) versus 1/kT

The spectral response of both devices is given in Fig. 4.6 and Fig. 4.7. The bias for the

nBn device is Vb = 0.2 V while the PIN device spectral response is given at zero bias (Vb

= 0V) V. The SNR at 77 K is slightly better for the nBn device. At 150 K the SNR of

the PIN device is significantly degraded compared to the nBn SNR response.

66
Chapter 4. InAs/GaInSb SLS based detector using nBn design with 8 µm cutoff

1.0 77 K nBn

Spectral Response (a.u.)

77 K PIN

0.8

0.6

0.4

0.2

0.0
3 4 5 6 7 8 9 10

Wavelength (µm)

Figure 4.6 Normalized spectral response of samples A (nBn, black) and B (PIN, red) at 77 K

1.0
150 K nBn
Spectral Response (a.u.)

150 K PIN

0.8

0.6

0.4

0.2

0.0
3 4 5 6 7 8 9 10

Wavelength (µm)

Figure 4.7 Normalized spectral response of samples A (nBn, black) and B (PIN, red) at

150 K

The responsivity and quantum efficiency were measured for sample A using the same

setup as in Chapter 3. The external quantum efficiency was calculated to be 12% and D*

67
Chapter 4. InAs/GaInSb SLS based detector using nBn design with 8 µm cutoff

was calculated to be 2.5 x 1010 cm Hz1/2/W at 5.5 µm. The peak responsivity was 0.51

A/W also at 5.5 µm. The responsivity and quantum efficiency is shown in Fig. 4.6. One

possible explanation for the relatively low quantum efficiency could be the poor

alignment of the valence band of the barrier with the 9 ML InAs/5 ML Ga0.75In0.25Sb SL.

0.7 14
Responsivity QE

0.6 12

Quantum Efficiency (%)

Responsivity (A/W)

0.5 10

0.4 8

0.3 6

0.2 4

0.1 2

0.0 0
2 3 4 5 6 7 8 9 10 11

Wavelength (µm)

Figure 4.8 Responsivity (left axis, black line) and Quantum Efficiency (right axis, red line) are

given as a function of wavelength for sample A (nBn)

Atmospheric absorbance above 5 µm to 8 µm limits the response of the detector. Further

study is currently underway on different superlattice period composition that will

increase the cutoff wavelength. A similar study of PIN and nBn devices using the new

superlattice absorber region is currently under investigation in our group. The barrier

material and composition is also currently being studied on 8x8 SL devices. Ref. 32

suggests that the valence energy level in standard type II superlattice stays almost the

same regardless of superlattice design. While our own theoretical analysis is underway,

68
Chapter 4. InAs/GaInSb SLS based detector using nBn design with 8 µm cutoff

this could mean that if an ideal barrier could be found, it could be used in any standard

type II superlattice structure with little optimization.

4.4 Conclusion

In this chapter we showed an nBn detector with 8 µm cutoff. The nBn device was

compared to a PIN device with the same λcutoff. The lower current density of the nBn

device shows the advantages of the device structure. While these results are promising,

the low quantum efficiency of the device must be improved if this material is to be made

into an FPA.

69
Chapter 5. Conclusions and Future Work

5 Conclusions and Future Work

5.1 Conclusions

This main focus of this paper is to introduce InAs/GaSb SL nBn detectors as a way to

increase the performance of SL detectors and to utilize the following advantages of the

material system. InAs/GaSb SL has high responsivity because of strong absorption of

normal incidence photons. Larger electron effective masses that are not directly

dependent on the bandgap energy lower the tunneling current compared with bulk

materials with the same bandgap energy. Different combinations of superlattice

periodicity or constituent material composition can be utilized to tune the superlattice

bandgap which can be used to provide multicolor capability. Also, the light-hole and

heavy-hole band separation can be adjusted to reduce Auger recombination.

In chapter 1 of this thesis we introduced the general concept of infrared detection and

distinguished between two different classes of IR detectors, thermal and photon detectors.

Photon detectors were further split up into either photoconductive or photovoltaic

detectors. Traditional detector figures of merit responsivity, noise equivalent power and

specific detectivity and their theoretical limits were discussed. The strengths and

weaknesses of existing FPA technology were briefly discussed. InAs/GaSb based

detectors were covered as a possible alternative to existing FPA technology. Finally nBn

detectors were introduced.

In chapter 2 the growth, processing and optical and electrical characterization of

InAs/GaSb IR detectors was explained. Material growth using solid source molecular

70
Chapter 5. Conclusions and Future Work

beam epitaxy was briefly explained. Processing of a single-pixel device was also briefly

covered. The optical and electrical characterization of the devices section included the

explanation of temperature dependent IV measurements, spectral response measurements

using a FTIR spectrometer and the responsivity measurement setup.

In chapter 3 the growth, processing and characterization of an InAs/GaSb SL nBn

detector with a cutoff wavelength of ~ 4.25µm at 77 K was given. Two structures were

fabricated out of the same growth material, one with a shallow top contact isolation etch

(structure A) and one with a conventional mesa structure (structure B). The shallow

etched device dark current was reduced by two orders of magnitude at 77 K compared

with the deep etched device. The current density was equal to 2.3 x 10-6 A/cm2 for and

3.1 x 10-4 A/cm2 for structures A and B respectively at Vb = 0.1V and T = 77K. At 250K,

the dark current density was equal to ~ 80 mA/cm2. The responsivity and QE of structure

A were measured at 4 µm to be 0.74 A/W and 23% respectively.

In chapter 4 we compared the electrical properties of an nBn detector and a PIN detector

with the same absorber thickness and structure (λcutoff = 8 µm) from subsequent growth

runs. The current density at 0.2V is 6.24 x 10-5 A/cm2 for the nBn and for the PIN at -

0.2V is 1.03 A/cm2 both at T = 77 K. While the growth for neither device was optimized,

the large difference between the devices shows that the nBn detector structure provides

significant advantages in the electrical properties (lower dark current) which we expect to

carry over into the optical properties (better signal to noise ratio) of the device. While the

responsivity of the nBn detector was low, 0.51 A/W at 5.5 µm, the expectation is that this

will be increased in future devices.

71
Chapter 5. Conclusions and Future Work

In conclusion, we showed that the nBn detector can be extended to InAs/Ga(In)Sb

superlattice material. Shallow etching of the devices lowered the dark current at lower

temperatures by removing the surface current component. nBn detectors with λcutoff of

4.25 and 8 µm at 77 K were demonstrated.

5.2 Future Work

nBn InAs/GaSb superlattice detectors show promise for increasing the performance of SL

FPAs. Questions behind the physics of this material system still remain unsolved. The

integration of the nBn detector design raises even more questions on the physical

mechanisms behind detection. There are still many topics for future research with these

devices.

5.2.1 Lateral Collection

The issue of lateral collection due to in-plane carriers in the shallow etched devices needs

to be thoroughly investigated. Preliminary investigations by our industrial collaborators

Santa Barbara Focal Plane show that there is that lateral collection of 20 to 60 µm in

measured devices. The lateral collection is a significant issue when incorporating this

material into a FPA where one square pixel has a side length of ~25 µm and the pitch is

~30 µm. Optical and electrical crosstalk would then limit the performance of the FPA. A

trade off exists then between defining devices with a shallow isolation etch to remove the

surface leakage currents and the amount of crosstalk that will can be tolerated in the FPA.

Studies of superlattice in-plane mobility have been preformed by L. Faraone group in

48
Western Australia using quantitative mobility spectrum analysis (QMSA) on

72
Chapter 5. Conclusions and Future Work

conductive GaSb substrates. Further work is being performed on the mobility in the

growth direction. Better understanding of the in-plane and growth mobility could allow

for optimization of the superlattice material in an attempt to lower the lateral collection in

the devices. Additionally, research into the minority carrier diffusion length and minority

carrier lifetime in superlattice material could be incorporated in improving nBn detector

performance.

5.2.2 nBn theoretical modeling

The theory behind the nBn detector is by no means fully understood. Theoretical

modeling of InAs with InAsSb barrier nBn detectors could provide insight that could be

helpful in understanding InAs/GaSb SL nBn detectors. For superlattice simulation

current techniques used in our research group for modeling of InAs/GaSb superlattice

provide useful information about the expected cutoff wavelength, separation of the

conduction and valence bands and expected wavefunctions for the electrons and holes.

Information that is not currently available is the position of the conduction and valence

bands with respect to a reference level. The alignment of the barrier to the first valence

band in InAs/GaSb is critical to the operation of the nBn detector, so the position of the

conduction and valence levels to a reference level could provide insight as to what

barriers could be used. While experimental barrier composition studies are underway,

further study incorporating theoretical SL data into nBn device modeling will be crucial

in understanding the physics behind the nBn detector.

73
Chapter 5. Conclusions and Future Work

5.2.3 Focal Plane Array Fabrication

The ultimate goal of this research is to make the nBn detector into a Focal Plane Array

(FPA) that can be hybridized to a read-out integrated circuit (ROIC). The first

InAs/GaSb SL nBn detector based FPA has been with made with collaboration from

QmagiQ LLC.

Figure 5.1 MWIR image captured from the first SL nBn FPA with a non-uniformity correction

applied at T = 77 K with a NETD = 26mK

Future FPA will be hybridized at The University of New Mexico at The Center for High

Technology Materials. This will allow both MWIR and LWIR SL nBn detectors to be

hybridized onsite. This technology can also be used to hybridize fanout circuits to

ROICs so that small area single pixel devices that are too small to consistently wire-bond

74
Chapter 5. Conclusions and Future Work

(< 100 µm). The fanout could then be used for temperature dependent measurement of

VADA IV.

5.2.4 Multispectral nBn devices

Multicolor nBn detectors have been demonstrated using nBn SL detectors36, but are still

just single pixel devices. Unlike other InAs/GaSb SL multispectral detectors that require

two In bumps per pixel and custom designed ROIC 49 , the multispectral nBn cutoff

wavelength would be bias dependent. The processing steps for conventional FPAs would

not need to be modified and off the shelf ROICs could be used.

5.2.5 Multiple barrier devices

Recent research has shown that 10 µm cutoff can be achieved while measuring the

spectral response with a bias applied across two top contacts. While the mechanism of

detection is not fully understood, the only material with 9 µm cutoff is in the absorber

layer meaning that the material is truly absorbing. More research needs to be done both

experimentally and theoretically to understand the physics behind this.

75
Chapter 5. Conclusions and Future Work

15

D5 S2 V = -0.5V

Spectral Response (a.u.)

S2 D5 T = 100 K

10

0
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

Wavelength (µm)

Figure 5.2 Spectral response of nBn detectors with a bias Vb = -0.5V across two top contacts at

T = 100 K

76
 Appendix A

Appendix A

Processing steps for nBn_L8_11_12_13_14_SLS

1. Mesa Shallow 1.1. Wafer clean Acetone/Methanol/DI rinse (5 minute each)

etching Nitrogen dry

6 minute dehydration bake @ 150°C

1.2. Photolithography Mask: mesa etch—positive PR

HMDS, 5k rpm, 30 sec

150°C / 30 sec

AZp4330, 5k rpm, 30 sec,

90°C / 90 sec

Mask Exposure time = 11 sec, CI2 setting

Develop AZ 400k (1:4), time ~70 sec(????)

1.3 Dry Etching ICP, Gas : BCl3,

Recipe : Zia-InP

ICP power : 500W, RF power : 80W

Etch time (40~45SEC.)

(Etch rate : ~4.5nm/Sec and after 5 min each rate

might differ with time)

1.4 Wet Etching H202:H3PO4:H2O=2:1:20 For 30 sec to 1min

1.5 PR removal PR removal using Acetone or RIE cleaning

1.6. Quality control Inspect etch pattern with alfa-step.

77
 Appendix A

2.Mesa Deep etching 2.1. Photolithography Acetone/Methanol/DI rinse (5 minute each)

Nitrogen dry

6 minute dehydration bake @ 150°C

2.2. Photolithography Mask: passivation—positive PR

HMDS, 5k rpm, 30 sec

150°C / 30 sec

AZp4330, 4k rpm, 30 sec,

90°C / 90 sec

Mask Exposure time = 11 sec, CI2 setting

Develop AZ 400k (1:4), time ~70 sec(????)

2.3 Etching ICP, Gas : BCl3,

Recipe : Zia-InP

ICP power : 500W, RF power : 80W

Etch time (??)

(Etch rate : ~2600 Å/min and after 5 min each

rate might differ with time)

2.4 PR removal PR removal using Acetone or RIE cleaning

2.5. Quality control Inspect etch pattern with alfa-step.

3. Top and bottom 3.1. Wafer clean Acetone/Methanol/DI rinse (5 minute each)

metal deposition Nitrogen dry

6 minute dehydration bake @ 150°C

78
Appendix A

3.2. Photolithography Mask: cmetal—negative PR

HMDS, 5k rpm, 30 sec

150°C / 30 sec

AZ5214E-IR, 5k rpm, 30 sec~45Sec

90°C / 90 sec

Mask Exposure time = 3.5 sec, CI2 setting

112°C / 60 sec (image reverse bake)

Flood exposure time = 60 sec, CI2 setting

Develop AZ400K 1:5, DI water, time = ~45 sec

3.3. Metal deposition Ti : 500 Å @ 2Å /sec

Pt : 500Å @ 1 Å /sec

Au: 3,000 Å @ 2.5 Å /sec

3.4. Life -off Acetone, Over 10 minute. Acetone spary.

3.5 Quality control Inspect patterns using microscope

79
Appendix B

Appendix B

Background Correction Instructions

Part 1

Explanation

The DTGS is a pryoelectric detector. The DTGS detector measures the power output

from the IR source inside the FTIR. The response of the DTGS is not independent of

frequency. As the frequency increases the magnitude of the response decreases. To

correct for this, the response of the DTGS was measured using a 980 nm diode laser in

the frequency range 400 Hz to 10 kHz. Using f = nν where f is the frequency, n is the

wavenumber and ν is the velocity, correction curves were plotted for the wavenumber

range 7400 – 400 cm-1. Higher velocities match the wavenumber range of the DTGS

better.

1.0 Velocity
.1581
.3165
.4747
0.8 .6329
.9494
1.2659
Normalized Signal

0.6

0.4

0.2

0.0

7000 6000 5000 4000 3000 2000 1000

-1
Wavenumber (cm )

Figure App. B. 1

80
Appendix B

1. Mount and cool device either as explained in Cryogenic Spectral Response

Measurements using FTIR in lab 175A or in a pour-fill dewar in the MICA lab.

Omnic.lnk
2. Open OMNIC application.

3. Open the Experiment Setup under the Collect tab or press Ctrl-E.

4. In the Experiment Setup window under the Collect tab choose the final format as

SingleBeam.

Figure App. B. 2

5. Under the Bench tab in the Experiment Setup window verify that the settings for

the sample compartment is Main, the Detector is DTGS TEC, the Beamsplitter is

XT-KBr, the source is IR, and the proper accessory is selected. Set the gain to one

(1.0) and select a velocity from the Velocity drop down menu as shown in Figure 2.

The Max and Min (top left circled in red) values should be in the range –10 to 10

otherwise the measurement will saturate. Also, the lower the velocity the smaller the

Aperture should be.

81
Appendix B

Max

Min

Figure App. B. 3

6. Under the advance tab in the Experiment Setup window set the Phase Correction

drop down menu to Power Spectrum. Power Spectrum is used to prevent negative

values during spectral response measurements.

Figure App. B. 4

7. Open the sample compartment on the FTIR to let in the Atmosphere then Collect the

background spectrum. The sample compartment has a nitrogen purge that removes

the atmospheric interference from H20 and CO2. Opening the sample compartment

reintroduces the atmospheric interference to the DTGS detector. Since the sample to

82
Appendix B

be measured is outside the FTIR this helps approximate the atmospheric interference

from the IR source to the sample detector.

8. Save the collected background by clicking File then save or press F12.

9. Open the correction curve for the Velocity of the background spectrum by

choosing .csv text as the type of file and matching the file name to the velocity

chosen in step 5. In Figure 2 the velocity was set at .6329. To match that open

the .csv file .6329. C:\My Documents\OMNIC\Spectra\Velocity Correction Curves

Figure App. B. 5

10. Set the parameters Y units to Single beam and the resolution to match the resolution

of the background measurement and press OK.

Figure App. B. 6

83
Appendix B

11. In the OMNIC window, first select the background curve then select the correction

curve. The order of the selection is important 1) background 2) correction. To do

this first click on the uncorrected background spectrum, the line should be red. Next

hold down Ctrl and click the correction curve. Both lines should be red.

1st

2nd

Figure App. B. 7

12. With both lines highlighted red click on the Process tab and select Spectral Math…

13. In the operation drop down menu select A / B. A is the background file, B is the

correction curve.

84
Appendix B Calculate

Background

Correction Curve

Figure App. B. 8

14. Press the Calculate button next to the Operation menu.

15. Add the result to the desired window by selecting the window number or Add to a

new window in the drop down menu.

Figure App. B. 9

16. Save the new corrected background.

85
Appendix B

17. In the Experiment Setup window under the Collect tab in the Background Handling

part select Use specified background file: and click Browse to find and select the

corrected background that was saved in Step 16. Then change the Final Format to %

Transmittance so that each collected sample will be automatically divided by the

corrected background.

Final Format Background

Handling
Use Specified
background file:

Figure App. B. 10

86
Appendix B

Part 2 Example

1. Collect Background.

Figure App. B. 11

2. Open Velocity Correction Curve. C:\My Documents\OMNIC\Spectra\Velocity

Correction Curves. The spectra file is the default save/open file when OMNIC is

started. Change Y units to Single Beam and Resolution to match the background

spectrum.

Figure App. B. 12

87
Appendix B

3. Use spectral math to divide Background by Correction Curve.

Figure App. B. 13

4. Save corrected background.

Figure App. B. 14

5. Use Specified background and change Final format to % Transmittance.

88
Appendix B

Figure App. B. 15

6. Collect and save samples

Figure App. B. 16

89
References

References

1
http://en.wikipedia.org/wiki/Image:EM_spectrum.svg
2
Planck, Max. On the Law of Distribution of Energy in the Normal Spectrum.

Annalen der Physik vol. 4, p. 553, 1901.

3
Vincent, John David. Fundamentals of Infrared Detector Operation and Testing. John

Wiley & Sons, New York, 1990.

4
Dereniak, E. L., G. D. Boreman. Infrared Detectors and Systems. John Wiley and Sons,

New York, 1996.

5
http://upload.wikimedia.org/wikipedia/en/5/5b/Atmospheric_transmittance_infrared.gif
6
Rogalski, A. Quantum well photodetectors in infrared detector technology. J. Appl.

Phys., Vol. 93, No. 8, 15 April 2003.

7
Jones, R. Clark. Performance of Detectors for Visible and Infrared Radiation,”

Advances in Electronics, Vol. 5, Academic Press, New York (1953).

8
http://www.infratec.de/en/infratec.html
9
Rogalski, A. Infrared detectors: status and trends, Progess in Quantum Electronics 27 66

(2003)
10
Lawson, W. D., S. Nielson, E. H. Putley, and A. S. Young, Preparation and properties

of HgTe-CdTe, J. Phys. Chem. Solids 9, 325-329 (1959).

11
Rogalski, A. Infrared Detectors. Gordon and Breach Science Publishers 2000

90
References

12
Tribolet, P., S. Blondel, P. Costa, A. Combette, G. Destefanis, P. Ballet, J.-P. Zanatta,

O. Gravrand, C. Langeron, J.-P. Chamonal and A. Million. MWIR focal plane arrays

made with HgCdTe grown by MBE on germanium substrates. SPIE, 2006

13
Gunapala, S.D., S.V. Bandara, J.K. Liu, C.J. Hill, S.B. Rafol, J.M. Mumolo, J.T. Trinh,

M.Z. Tidrow, P.D. LeVan. 1024 x 1024 pixel mid-wavelength and long-wavelength

infrared QWIP focal plane arrays for imaging applications. Semicond. Sci. Technol. 20

No. 5 (2005)
14
Smith, D.L. and C. Mailhiot. Proposal for strained type II superlattice detectors, J.

Appl. Phys. 62, (6), 2545-2548 (1987)

15
Sai-Halasz, G. A., R. Tsu, L. Esaki. A new semiconductor superlattice, Appl. Phys.

Lett. 30, (12), 651-653 (1977)

16
Esaki, L. Proceedings of the 17th International Conference on the Physics of

Semiconductors, edited by J.D. Chadi, and W.A. Harrison (Springer, New York), p. 473

1985
17
Smith, D.L. and C. Mailhiot. k·p theory of semiconductor superlattice electronic

structure. I. Formal results, Phys. Rev. B 33, No. 12 8345-8359 1986

18
Mailhiot, C., and D.L. Smith. k·p theory of semiconductor superlattice electronic

structure. II. Application to Ga1-xInxAs-Al1-yInyAs [100] superlattices Phys. Rev. B 33,

No. 12 8360-8372 1986

19
Dente, G.C., and M.L. Tilton. Pseudopotential methods for superlattices: Applications

to mid-infrared semiconductor lasers, J. Appl. Phys. 86, No. 3, 1420-1429 1999

91
References

20
Dente, G.C., and M.L. Tilton. Comparing pseudopotential predictions for InAs/GaSb

superlattices, Phys. Rev. B 66, 165307 (2002)

21
Wei, Y. and M. Razeghi. Modeling of type-II InAs/GaSb superlattices using an

empirical tight-binding method and interface engineering, Phys. Rev. B 69, 085316

(2004)
22
Rogalski, A. Competitive technologies of third generation infrared photon detectors,

Opto-Electron. Rev., 14, no. 1, 87-101, 2006

23
Youngdale, E. R., J. R. Meyer, C. A. Hoffman, F. J. Bartoli. Auger lifetime

enhancement in InAs-Ga1-xInxSb superlattices, Appl. Phys. Lett. 64, 3160, 1994

24
Bürkle, L., and F. Fuchs. InAs/(GaIn)Sb superlattices: a promising material system for

infrared detection, Handbook of Infrared Detection and Technoligies, Elsevier, Oxford,

2002
25
Rehm, R., M. Walther, J. Schmitz, J. Fleibner, F. Fuchs, W. Cabanski and J. Ziegler.

InAs/(GaIn)Sb short period superlattices for focal plane arrays, Proc. of SPIE Vol. 5783,

123-130, 2006
26
Rehm, R., M. Walther, J. Schmitz, J. Fleibner, S. Kopta, F. Fuchs, W. Cabanski and J.

Ziegler. Growth of InAs/GaSb short-period superlattices for high-resolution mid-

wavelength infrared focal plane array detectors, J. Crys. Grow. 278, 1560161, 2005
27
Yau, H., Y. Wei, M.Z. Tidrow, V. Yazdanpanah, M. Razeghi and V. Nathan.

Uncooled operation of type-II InAs/GaSb superlattice photodiodes in the mid-wavelength

infrared range, Appl. Phys. Lett. 86 233106-233109, 2005

92
References

28
Yau, H., Y. Wei, A. Hood, M.Z. Tidrow, V. Yazdanpanah, M. Razeghi and V. Nathan.

High-performance type-II InAs/GaSb superlattice photodiodes with cutoff wavelength

around 7 µm. Appl. Phys. Lett., 86, 091109-091112, 2005

29
Bergman, J., G.J. Sullivan, A. Ikhlassi, C. Grein, R.E. DeWames, J.R. Waldrop, H.

Yang, M. Flatte, K. Mahalingam, M. Zhong, M. Weimer. Molecular beam epitaxy

growth of high quantum effiency InAs/GaSb superlattice detectors, J. Vac. Sci. Technol.

B, 23, no. 3 1144-1148 (2005)

30
Nguyen, B.-M., D. Hoffman, Y. Wei, P.-Y. Delaunay, A. Hood and M. Razeghi. Very

high quantum efficiency in type-II InAs/GaSb superlattice photodiode with cutoff of 12

µm, Appl. Phys. Lett. 90, 231108 (2007)

31
Aifer, E.H., J.G. Tischler, J.H. Warner, I. Vurgaftman, W. W. Bewley, J.R. Meyer, J.C.

Kim, L.J. Whitman, C.L. Canedy, and E.M. Jackson. W-structured type-II superlattice

long-wave infrared photodiodes with high quantum efficiency, Appl. Phys. Lett. 89,

053519 (2006)
32
B.-M. Nguyen, D. Hoffman, P.-Y. Delaunay and M. Razeghi. Dark Current

suppression in type II InAs/GaSb superlattice long wavelength infrared photodiodes with

M-structure barrier, Appl. Phys. Lett. 91, 163511 (2007)

33
B.-M. Nguyen, D. Hoffman, P.-Y. Delaunay, M. Razeghi and V. Nathan. Polarity

inversion of type-II InAs/GaSb superlattice photodiodes, Appl. Phys. Lett. 91, 103503

(2007)

93
References

34
Rodriguez, J.B., E. Plis, G. Bishop, Y.D. Sharma, H. Kim, L.R. Dawson, and S.

Krishna. nBn structure based on InAs/GaSb type-II strained layer superlattices, Appl.

Phys. Lett. 91, 043514 (2007)

35
E. H. Aifer, J. G. Tischler, J. H. Warner, I. Vurgaftman, J. R. Meyer, C. L. Canedy, and

E. M. Jackson. Dual band LWIR/VLWIR type-II superlattice photodiodes, SPIE, v5783,

112–122 (2005)
36
Khoshakhlagh, A., J. B. Rodriguez, E. Plis, G. D. Bishop, Y. D. Sharma, H. S. Kim, L.

R. Dawson and S. Krishna. Bias dependent dual band response from InAs/Ga(In)Sb type

II strain layer superlattice detectors, Appl. Phys. Lett. 91, 263504, (2007)
37
Maimon, S. and G.W. Wicks. nBn detector, an infrared detector with reduced dark

current and higher operating temperature, Appl. Phys. Lett. 89, 151109 (2006)
38
Rogalski, A. Infrared Phys. 28, 139 (1988)
39
Plis, E., S. Annamalai, K.T. Posani, S. Krishna, R.A. Rupani, and S. Ghosh. Midwave

infrared type-II InAs/GaSb superlattice detectors with mixed interfaces, J. Appl. Phys.

100, 014510 (2006)

40
Hill, C. J., J. V. Li, J. M. Mumolo, S. D. Gunapala. MBE grown type-II MWIR and

LWIR superlattice photodiodes, Infrared Phys. and Tech. 50 v-vii, 189 (2007)
41
Little, J.W., S.P. Svensson, W.A. Beck, A.C. Goldberg, S.W. Kennerly, T. Hongsmatip,

M. Winn and P. Uppal. Thin active region, type II superlattice photodiode arrays:

Single-pixel and focal plane array characterization, J. Appl. Phys. 101, 044514 (2007)
42
Gin. A., Y. Wei, J. Bae, A. Hood, J. Nah and M. Razeghi. Passivation of Type II

InAs/GaSb superlattice photodiodes, Thin Solid Films, (447-448):489-492, (2004)

94
References

43
Hood, A., P.-Y. Delaunay, D. Hoffman, B.-M. Nguyen, Y. Wei, M. Razeghi and V.

Nathan. Near bulk-limited R0A of long-wavelength infrared type-II InAs/GaSb

superlattice photodiodes with polyimide surface passivation, Appl. Phys. Lett., 90,

233513 (2007)
44
Rehm, R. M. Walther, F. Fuchs, J. Schmitz and J. Fleissner. Passivation of

InAs/(GaIn)Sb short-period superlattice photodiodes with 10 µm cutoff wavelength by

epitaxial overgrowth with AlxGa1-xAsySb1-y, Appl. Phys. Lett. 86, 173501 (2005)
45
Plis. E., J.-B. Rodriguez, S. J. Lee and S. Krishna. Electrochemical sulphur passivation

of InAs/GaSb strain layer superlattice detectors, Electron. Lett. 42, 21 (2006)

46
Gin, A., Y. Wei, A. Hood, A. Bajowala, V. Yazdanpanah, M. Razeghi and M. Tidrow.

Ammonium sulfide passivation of Type-II InAs/GaSb superlattice photodiodes, Appl.

Phys. Lett. 84, No. 12 2037 (2004)

47
Vogt, A., A. Simon, H. L. Hartnagel, J. Schikora, V. Buschmann, M. Rodewald, H.

Fuess, S. Facsko, C. Koerdt and H. Kurz. Ohmic contact formation mechanism of the

PdGeAu system on n-type GaSb grown by molecular beam epitaxy, J. Appl. Phys. 83, No.

12, 7715-7719 (1998)

48
Chandrasekhar Rao, T. V., J. Antoszewski, L. Faraone, J. B. Rodriguez, E. Plis and S.

Krishna. Characterization of carriers in GaSb/InAs superlattice grown on conductive

GaSb substrate, Appl. Phys. Lett. 92, 012121 (2008)

49
Rehm, R., M. Walther, J. Fleibner, J. Schmitz, J. Ziegler, W. Cabanski and R. Breiter.

Bispectral Thermal Imaging with Quantum Well Infrared Photodetectors and InAs/GaSb

Type-II Superlattices, Proc. of SPIE, Vol. 6206, (2006)

95