Y DIODES ON GaN (GALLIUM NITRIDE) SEMICONDUCTORS SCHOTTK' by VEENA VIJAYAN, B.Tech. A THESIS IN ELECTRICAL ENGINEERING Submitted to the Graduate Faculty of Texas Tech University in Partial Fulfillment of the Requirements for the Degree of MASTER OF SCIENCE IN ELECTRICAL ENGINEERING Approved Chairperson of the Committee Accepted "bean of tl^ Graduate Scho^ August, 2000
111
Embed
Y DIODES ON GaN (GALLIUM NITRIDE) SEMICONDUCTORS …
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Y DIODES ON GaN (GALLIUM NITRIDE) SEMICONDUCTORS SCHOTTK'
by
VEENA VIJAYAN, B.Tech.
A THESIS
IN
ELECTRICAL ENGINEERING
Submitted to the Graduate Faculty of Texas Tech University in
Partial Fulfillment of the Requirements for
the Degree of
MASTER OF SCIENCE
IN
ELECTRICAL ENGINEERING
Approved
Chairperson of the Committee
Accepted
"bean of t l ^ Graduate Scho^
August, 2000
ACKNOWLEDGEMENTS
I would like to express my sincere gratitude to everyone who assisted me during
this research work. As a firm believer in God, I would first thank him for giving me
support through all these years.
I would first like to thank Dr. Henryk Temkin, advisor and Jack Maddox
Distinguished Engineering Chair, for giving me an opportunity to work with highly
skilled scientists in a world-class laboratory. I am thankful for having benefited from his
confidence and was able to contribute to a major research project. The project done under
his guidance has enriched me with some priceless experiences, which I will cherish
forever.
I would like to thank Dr. Jon Bredeson and Dr. Tim Dallas for serving as my
committee members and spending their valuable time in reviewing this thesis.
I would like to thank my colleagues Sergey Nikisin, Caixia Jin, Vladimir
Kuryatkov and Marvin Cummings for valuable hours of discussions.
I would like to thank my parents and my siblings Vini and Vijesh for their support
and patience since I left home. Finally, I would like to express my gratitude to Shahin, for
After exposure, the photoresist is developed to produce the image, which is used
as a mask for lift-off. The main purpose of the development process is to effectively
produce the specified pattern on the wafer with minimum distortions or swelling. The
development time is normally kept short.
The positive resist developers are alkaline solutions diluted with water. SHIPLEY
MICROPOSIT DEVELOPER CONCENTRATE positive photo resist developer was
used for the development of the pattern for the diodes in the lab. The development
solution contained developer and water in 1:1 ratio. As a result, the wafer required only a
water rinse before it was blown dry. During the develop cycle, in the exposed areas, the
carboxylic acid reacts with the developer to form amines and metallic salts, which then
rapidly dissolve in the developer solution. In the unexposed areas, no such groups are
formed and these regions are unaffected by the developer. The development time used in
the process was 20 sec.
After development, the wafers are rinsed in DI water. Proper rinsing in DI water
is required as the developing action continues until the developer is completely removed
from the resist surface. The substrate is rinsed for more than 30 sec and then is blown
dry.
4.6 Choice of the Metals Used for Making the Ohmic and Schottky Contacts
According to the Schottky-Mott model, the barrier height equals the difference
between the metal workfimction and the electron affmity of the semiconductor x • The
51
electron affmity of GaN - 4.1 eV. The workfimction of Au and Ni is 5.1 and 5.15,
respectively, and thus was choosen to form Schottky diodes. Schottky contacts were
fabricated by evaporating Au and Ni respectively. The thickness of the Schottky contacts
were 1000 A and were circular in shape with a diameter of 200//w .
A composite metal layer of Ti/Al/Ti/Au(300 A /700 A /300 A /500 A) formed the
ohmic contact configuration. Electron beam evaporation was used to deposit the layer.
4.7 Deposition Technology
4.7.1 Physical Vapor Deposition by Evaporation
The ohmic and Schottky contacts necessary for the fabrication of a Schottky diode
are deposited using evaporation technique. In this process the source material is heated
thereby causing evaporation. The evaporation is done in high-vacuum conditions
typically in the range of 10~ to 10 torr. The vaporized molecules from the source
reach the substrate surface without collisions with any other gas molecules in the
deposition chamber. As a result the deposited film is as pure as the source material. The
main advantage of the evaporation process is that the substrate surface is not damaged
due to the low energy of the impinging atoms and moreover the films can be deposited at
higher rates. During the evaporation process as the source is heated, the pressure rises
somewhat as contaminants are desorbed from its surface.
4.8 The Evaporation Svstem Used for the Experiment
The evaporator used in the lab for the fabrication of the Schottky diode is shown
in Figure 4.11. The main parts of the evaporator can be divided into three main sections:
52
• Vacuum system,
• Chamber which includes the evaporating source, substrate holder, the crucible liners,
• Evaporation thickness controller.
This section explains the evaporation system followed by a brief summary of the
system configuration.
4.8.1 Vacuum system
The vacuum system consists of the rotary roughing pump and the cryogenic pump
as shown in Figure 4.11. The rotary roughing pump is a mechanical pump that transports,
compresses, and expels the gas. The gas expelled from a roughing pump is known to be
carcinogenic. Therefore a ventilation system is devised from 3-1/2 inch PVC tubing that
is directed to a roof vent. The rotary roughing pumps serves two purposes, first they are
used to pump vacuum chambers from atmosphere to the medium vacuum range in the
order of 10 torr. And second they serve as forepumps for the cryo pump.
A cryogenic pump, Cryo-Torr 8 Cryopump, is used to improve the vacuum. It
operates on the principle that gases can be condensed and held at extremely low vapor
pressures, achieving high speeds and throughputs. Cryogenic pumps are closed cycle
refrigerator pumps that removes gases from the vacuum chamber by capturing them on a
cold surface. The process of gas capture by high-vacuum cryopump involves both
cryocondensation and cryosorption. Cryocondensation refers to the condensation on a
surface whose temperature is cold enough that the vapor pressure of the condensed
substance is so low that the vapor is effectively removed from the system. Cryosorption
involves the adsorption without condensation of a gas on a cold surface.
53
Nj Puige
NjPuige
Vent Valve
Absorption Trap
Ron^ineValve
--•@TCGauge#2
Gate Valve
Ion Gauge
<5G)
TC Gauge i
Roughing Piunp
^ 7 \ / ) CiyoPump
Figure 4.11 The evaporation system used.
4.8.1.1 Vacuum procedure.Operating procedure requires that the cryopump
pressure is below 50 torr before turning on the compressor unit. The pressure in the
chamber is brought down to the operating level of the cyro pump by first turning the
roughing pump on. The procedure for the operation of the vacuum chamber is as follows.
• After loading the samples into the chamber the vacuum jar is completely lowered.
54
• With the Vent Valve and Collar Valve closed and the roughing pump on, the
roughing valve is opened. The pump is allowed to evacuate the chamber until the
pressure is about 10"^ torr.
• When adequate pressure is achieved, the roughing valve is closed and the gate valve
is opened. The cryo pump is turned on.
• The roughing pump is turned off and the vent valve is opened. The foreline pressure
is allowed to reach atmospheric pressure and then the vent valve is closed. This is
done to prevent the back-flow of oil vapors from the roughing pump to the cahmber.
• The pressure is allowed to reach the baseline pressure for deposition(~10 ~ torr)
Monitoring of the vacuum pressure is an important factor in the deposition of the
thin films. Since it is impossible to find a vacuum gauge that can give quantitative
measurements over the entire vacuum region of interest (10^^ - 10"^ torr), two vacuum
gauges are used, an ion gauge and a vacuum gauge. One is placed above the gate valve
and one below. This arrangement allows the monitoring of the pressure inside the cryo
pump and the chamber. The ionization gauge measures pressure in terms of the number
of density of molecules and has a measuring range from 10 to 10 torr. The Varian
low vacuum guage is a heat loss manometer, inferring the pressure of the gas by
measuring the thermal loss of a heated wire, and has a measuring range from atmospheric
pressure to 10~^torr. This gauge is used to indicate the actual pressure in the chamber
turning the roughing stage.
55
4.8.2 Vacuum Chamber
The vacuum chamber consists of the crucible liners, substrate holder, shutter, the
crystal detector and the evaporation source as shown in Figure 4.13.
4.8.2. ICrucible Liners.The crucible liners used for this work was a 2.2 cc 3kW e-
Gun liner with the following dimensions and is shown in Figure 4.12.
Table 4.1 Dimension of the crucible liners.
Dimensions
(A) Top Diameter
(B) Bottom Diameter
(C) Height
(D) Wall Angle
Measurements
.75"
.56"
.37"
15"
' ^
T
Figure 4.12 Crucible Liner
k«-
56
The Glassy Coated Graphite crucible liners were used for the evaporation of Au,
Al, and Ti source material. The glassy coated carbon graphite liners are made from high
grade carbon with a coated graphite process that is baked above 1,400^C to eliminate any
porosity in the carbon.
4.8.2.2 Evaporation source. There are three different types of crucible heating
systems: resistive, inductive, and electron beam systems. For the evaporation of the
source material in the lab, electron beam evaporation was used. In this technique, a
stream of electrons is accelerated to high kinetic energy. The beam is directed at the
material to be evaporated, and the kinetic energy is transformed to thermal energy upon
impact. The electron stream can melt and evaporate any material, provided the beam can
supply sufficient energy. The electron beam heating is an efficient and practical way of
achieving temperatures in excess of 3500^ C for uniform thin film deposition process.
The Thermionics e-Gun^^ source, which is bakeable to 250° C, and its control units, have
been chosen to be part of our evaporation system. Its simplicity and non-contaminating
operation makes it suitable for research activities. The features of the e-Gun is as
follows:
• The e-Gun^^ can be used to evaporate the metals required for the ohmic and
Schottky contacts.
• The source control unit has three safety interlocks that work in conjection with the
e-Gun evaporation source.
1. Water-flow interlock, to insure proper water flow to the unit.
2. Pressure interlock, to make sure the confrol unit won't operate above 1x10" torr.
57
3. Cabinet interlock, to make sure the power is safely disconnected when the rack
panel door is opened.
TM • The maximum output power of the e-Gun source is 3000 Watts
TM • The e-Gun source has 5 crucible positions.
• Each crucible will hold up to 2.5 c c of source material.
TM The e-Gun source control unit has two important accessories, which will allow
precise application of any standard evaporant.
• A deposition rate controller, which monitors the rate of deposition in angstroms per
minute.
• A thickness control monitor, which indicates the film thickness that is deposited.
'< > <: ^
^^^jr,a_jr*t,jr^^jn^j^^j^
m
Substrate Holder
fc. Shutter
_^ Crystal
^ Filament
Figure 4.13 Evaporation Chamber
58
4.8.3 Evaporation thickness control monitor
The block diagram of the thickness control monitor is shown in Figure 4.14.
Inficon Deposition Thickness Controller.
Vacuum guage Controller Ion Guage Controller
Thermionics e-Gun Source Control Unit.
Figure 4.14 Control Cabinet
The Inficon evaporation thickness controUer is placed at the highest position. The
Inficon IC/5 thin film deposition controller is equipped with a high frequency crystal
(6.0 MHz) and has a 0.00577 A rate resolution. It is a precise controller with a high
resolution capability, can define up to 50 processes, store recipes up to 250 layers for a
single process, and can use up to 24 materials. It also has a co-operation fimction, which
means two sources may operate concurrently.
59
4.8.4 Summary of system configuration
The e-gun beam evaporation system is configured as follows:
• Pumping system: An Alcatel M2008A, mechanical pump is employed to rough the
chamber. It takes about 10 minutes to drop the pressure inside the chamber to 1x10"
torr. A Helix CRYO-TORR 8 high vacuum pump is used to reach the high vacuum of
Q
10' torr. It takes about 5 hours to drop the pressure to high vacuum after switching
from the Alcatel mechanical pump to Helix CRYO-TORR pump.
• Chamber: It consists of a metal jar with a opening through which the crucible can be
viewed, power lines, vacuum gauges, cooling water circulation, displacer of the
crucibles, and crystal wires.
• Evaporation source. Thermionic e-Gun source is a 3 KW five position e-beam
source with 2.5 c c crucibles. The different metals are loaded in the crucibles and are
evaporated sequencially for the ohmic contacts.
• Deposition controller: The Inficon IC/5 Thin Film Deposition Controller is able to
provide small power rise steps for the Thermionic e-Gun^^ source control unit. The
rate of deposition was set to 5 A / sec.
• Substrate holder. The substrate holder was a square metallic holder with four circular
opening to hold four, two inch samples as shown in the Figure 4.13.
• Cooling system: A cooling system which provides constant water cooling when the
evaporation is in progress.
• Instrumentation: Two vacuum gauge controllers and a Thermionic e-Gun source
control unit constitutes the whole system.
60
4.9 Lift-Off Technique
After the evaporation of the contact, the wafer is subjected to the lift-off
technique. The wafers are immersed in a solution capable of dissolving the photoresist.
Semiconductor grade acetone is used for this purpose. The wafer is placed in acetone for
ten minutes and then is subjected to ultrasonic agitation. The metal that was deposited
directly on the semiconductor remains, while the metal deposited on the resist lifts off of
the wafer as the resist dissolves. The wafers are then cleaned in semiconductor grade
methanol, rinsed in de ionized water and are dried in N2.
For the two-mask process, after the deposition of the ohmic contacts, the wafers
were annealed for 2 minutes at 700°C in nitrogen ambient. After the annealing and
subsequent lift-off process, the wafers are again subjected to the photoresist procesing
steps to pattern the Schottky contacts, and thus are subjected to a total of two lift-off
processes.
4.10 Other Work Involving Different Device Geomentries
Several other device and contact geomentries were used by different research
groups in making Schottky diodes. These included lateral, mesa, and Schottky metal field
plate devices. The lateral Schottky rectifiers were similar to the one used in our lab for
this work. In some devices, large area Au metallization was used as a low resistivity
contact.
61
The mesa devices used by other research groups had the geometry shown in
Figure 4.15. In these devices, the mesa edge termination was done by chemically assisted
ion beam etching, ^ ^ using Xe ions accelerated with lOOOV and 25 seem of CI2. Mesa
devices offers lower resistivity since the ohmic contacts were deposited closer to the high
conductivity interface layer. In some cases the mesa structure was formed by CVAr ICP
etching, followed by 750°C annealing to remove the etch damage, and are then treated
in (NH4)2 solution at 25°C for 20 min. ^ ^
Schottky Contact
Ohmic Contact
Figure 4.15 Schematic of a mesa diode.
The metal field plate device is shown in Figure 4.16. In these structures, Si02 was
sputtered using Si02 targets and 10 seem of O2 flow, and was then patterned.
SiO,
Figure 4.16 Schematic of metal field plate device.
62
In some cases, a conventional photolithographic lift-off technique was used to define
Schottky diodes using the configuration as shown in Figure 4.17. ^^ The diodes had a
donut configuration with a diameter of 125 /jm . The configuration gave a relatively low
series resistance due to the small spacing between the contacts {-40 jam ) . The outer dark
colored ring formed the ohmic contact, while the circular region at the center formed the
Schottky contact.
OO GO
"• Schottky contact
-•Ohmic contact.
Figure 4.17 Schematic of a Schottky diode.
63
CHAPTER V
INVESTIGATION OF RESULTS AND DISCUSSION
In this thesis work, I report on the investigation of Schottky characteristics of Au
and Ni on n-GaN. Three different contact geometries were tested. The characteristics
were determined using I-V and C-V measurements. The measurements were done at
room temperature. The details of the fabrication of the diodes are dealt with in Chapter
III.
The metals Ni and Au was chosen for this study because both have a work
fiinction of 5.15eV and 5.1eV, respectively and could form good Schottky diodes with n-
GaN. Ni was chosen because it is more chemically active and provides better adhesions
to the samples.
5.1 Experiment
The GaN sample used in this study was n-GaN grown by gas-source molecular
beam epitaxy. The details of the grov^h procedure and fabrication process were discussed
in Chapter HI. Three different contact geomentries were taken into consideration. In each
of the different process, the electrical characteristics of Ni and Au Schottky diodes on n-
GaN were investigated. In the one mask process large area metallization formed the
ohmic contacts. In the two mask process, Ti(300A )/Al(700 A)/Ti(300 A)/Au(500 A)
multilayers formed the ohmic contact. Each diode fabricated was 200//w in diameter.
The distance between the adjacent ohmic and Schottky contacts is about 50 jum
64
The Capacitance-Voltage measurements were done with a HP 4275A Multi
frequency LCR meter. The data was captured with a Lab VIEW program. The I-V
measurements were done with a Keithley 2400 source meter and recorded using a
Lab VIEW program.
5.1.1 Capacitance -Voltage Measurements
The capacitance measurements of Ni and Au Schottky diodes were measured at
different frequencies raging from 10 KHz to 1 MHz between -6 V and 0 V.
The C-V relationship for a Schottky barrier is given by^ '
1 2 (.. „ kT-^ F., - V,
where s^ is permittivity of free space, Ks the dielectric constant of GaN, Vbi is the flat
band voltage, A is the area of the diodes, and VR is the applied voltage.
The capacitance and Cm is related according to
C = ^ " {\ + RGf+{27(fCRf
where f is the measurement frequency, R is the series resistance, and G is the junction
conductance. The value of the conductance was found to be very small and thus was
neglected.
Thus the capacitance was related by
c = ^
65
The value of Cm approaches C, when RG « 1 and (2;rfilC)^ « 1 . The series resistance at
lower frequencies was considered negligible. The value of series resistance was then
calculated for each frequency range. The barrier height and carrier concentration was
calculated from these measurements.
5.1.2 Current-Voltage Characteristics
The forwrad-biased I-V characteristics are described by the following equation:
-i'^'-'''ynkr)-^ I = Is
and
.**,
where Is is the saturation current, q is the electron charge, V is the applied voltage, Rs is
the series resistance, n is the ideality factor, A the effective Richardson constant, and
6g is the measured barrier height. The barrier height was determined using the theoretical
value of the Richardson constant.
The Richardson constant is given by
.** 47rem„k^
** .
A = h'
The electron effective mass for GaN is 0.22mo.The theoretical value of A' ' is 26 Acm'
It has been reported that A** value varies over a wide range of values for Schottky diodes
on GaN. ^ " ' ' Since a large variation in A** does not have a significant influence on the
66
deduced Barrier height (for example a variation of 100% in A * will resuh only in 1.8%
reduction in barrier height) and therefore the theoretical value was taken for calculations
5.2 Results and Discussions of Capacitance-Voltage Measurements using One-Mask process
The Capacitance-Voltage measurements for Au/n-GaN and Ni/n-GaN Schottky
diode for different frequencies are shown in Figure 5.1 and Figure 5.3, respectively. The
plot of 1/C vs V for different frequencies for Au and Ni Schottky diodes using the one
mask process are given in Figure 5.2 and Figure 5.4, respectively. The carrier
concentration and barrier height are calculated for Au and Ni Schottky diodes and are
summarized in Table 5.1 and Table 5.2.
67
Voltage (V)
Figure 5.1 Capacitance-Voltage measurements of Au/n-GaN using the one-mask process for different frequencies.
68
Voltage(V)
Figure 5.2 Capacitance-Vohage, i.e., (1/C^ vs. V) of Au/n-GaN Schottky diodes measured at different frequencies.
69
Voltage (V)
Figure 5.3 Capacitance-Voltage measurements of Ni/n-GaN using the one-mask process for different frequencies.
70
Figure 5.4 Capacitance-Voltage, i.e., (1/C^ vs. V) of Au/n-GaN Schottky diodes measured at different frequencies.
71
The measured capacitance for lOKHz, 20 KHz, 40 KHz, and lOOKHz were found
to be nearly equal. The slopes were equal for all these frequencies. The capacitance for
400KHz and 1 MHz was different from that of the other frequencies. Table 5.1 and
Table 5.2 show the variation of the parameters with frequency measurements for Au/n-
GaN and Ni/n-GaN Schottky diodes.
The variation in capacitance with frequency could be due to the presence of deep
level defects or a large resistance of the GaN sample. At lower frequencies the measured
capacitance is due to the depletion capacitance, i.e., it is frequency is independent of the
bias. As the frequencies is increased, the diode capacitance depends on the depletion
capacitance, the bulk resistance, and dispersion capacitance. If the change in capacitance
was due to the presence of defects, all donor levels above the Fermi level will be ionized
under the bias, giving a higher doping concentration near the interface. At low
frequencies, the deep levels can follow the ac signal. As a result, there will be larger
contribution to the capacitance, resulting in an increase in carrier concentration which
will decrease the slope of the 1/C^ plot. However, from the experimental results of the
Au/n-GaN and Ni/n-GaN Schottky diodes, the slopes remained nearly same for most
frequencies used and the carrier concentration for Au/n-Gan Schottky diodes at
frequencies from 10 KHz to 100 KHz was found to be in the range of 5.72 x lO^^cm' -
5.75 X lO^^cm" . For Ni/n-GaN diodes, the carrier concentration was in the range of
5.05 X lO^ cm" -5.20 x lO^^cm" . It was concluded that the presence of deep levels was
not the primary reason for the observed dependence of the capacitance on frequency.
72
Table 5.1 Summary of C-V measurements on Au/n-GaN using the one-mask process
Measurement Frequency
IMHz
400KHZ
200KHZ
lOOKHz
40KHz
20KHz
lOKHz
Barrier Height
1.25eV
1.14eV
1.22eV
1.22eV
1.22eV
1.20eV
1.20eV
Carrier Concentration
5.50xl0^^cm'^
5.22 X lO^^cm'
5.66xl0^^cm'^
5.74xl0'^cm''
5.75 X lO^ cm'
5.72xl0'^cm-'
5.72 X lO' cm'
Table 5.2 Summary of C-V measurements on Ni/n-GaN using the one-mask process
Measurement Frequency
IMHz
400KHZ
200KHZ
lOOKHz
40KHz
20KHz
lOKHz
Barrier Height
1.12eV
1.08eV
1.18eV
1.24eV
1.22eV
1.18eV
1.17eV
Carrier Concentration
4.26 X lO'^cm'
4.36 X lO' cm'
4.78 X lO^^cm'
5.20xlO'^cm'^
5.15xlO'^cm"'
5.04 X lO^^cm'
5.03 X lO^^cm'
73
The conductance measured was found to be very small and was neglected. The
resistance measured was in the range of 220-260 Q for the Au/n-GaN and in the 400 Q.
range for Ni/n-GaN Schottky diodes. The condition {^TtfRsCY <1 was satisfied. The
barrier height and carrier concentration was calculated.
5.3 Results and Discussions of Current-Voltage Measurements using One-Mask process
The current-voltage characteristics of Au/n-GaN and Ni/n-GaN are shown in
Figure 5.5 and Figure 5.6. The reverse bias current was found to be in the nA range at -
10 V. the effective barrier height and ideality factor were determined.
In the measurements, it was observed that the current-voltage charactersitics
deviated from the linear behaviour in the forward bias region due to the series resistance.
As a result, accurate determination of IQ and n cannot be obtained from the I-V plot.
Series resistance has been found to dominate current conduction process in large-band-
gap semiconductors materials. At large currents apart from the voltage drop at the
junction there will a voltage drop across the series resistance. The value of series
resistance was found using the method of Cheung and Cheung. Once the series resistance
is calculated the voltage is corrected and the correct voltage across the junction is given
by Vd= V-IRs.
74
1E-3 ^
1E-4 .
^ 1E-5 <
g> 1E-6
o 1E-7 ^
1E.8 r
1E-9 =;•
-
—
-"
^ " -m
---
2 2 •
-"
; • ™
" -•
:-•
" " 1
i
i
t
U
1
• t
« «
• •
t
1
« •
•
1
«
#
T
tf
1
tf
1
tf
1
* «
- I
• m
1
«
1
« «
f
1
«
1
« •
" • •<
*
a
I •
^ --
^ H -
--
I •• a
. -•
_ B
-_ .
;
-
I
0.0 0.4 0.8 1.2 1.6
Voltage (V)
2.0
Figure 5.5 Forward I-V characteristics of Au/n-GaN diodes.
75
04 as 1.2
NADHage(V)
1.6 20
Figure 5.6 Forward I-V characteristics of Ni/n-GaN diodes using the one-mask process.
76
1E€
1&7r
1B8r
1&9r
1E-10
-20 -16 -12
Volatge(V)
Figure 5.7 Reverse I-V characteristics from Au/n-GaN Schottky diodes using the one-mask process.
77
1E5.
1E6
^
1E-7
1E6
1B9
1E-10
T—1
1 [III
1 1 T—1
1 [1
11 1
1 1
• 1
1 1
Iniii 1
• ' • • • • •
•
• •
• • • -
• 3
[ % -
•
* • _
• •
_ J
: ] • a
: . 1 . 1 . 1 . 1 . 1 1 :
^ -25 -20 -15 -10
Voltage (\0
Figure 5.8 Reverse I-V characteristics from Ni/n-GaN Schottky diodes using the one-mask process
78
A plot of the natural log of current versus forward bias voltage for small forward
currents are plotted, the family of ln(I) versus Vd plot is investigated. The curves are
linear over the entire vohage range. Accurate values of the ideality factor n. Is and barrier
height are calculated. The results are summarized in Table 5.3.
Table 5.3 Summary of the characteristics of the Schottky barriers from I-V measurements using one-mask
Metal
Au/n-GaN
Ni/n-GaN
Barrier height
0.9eV
0.901eV
Ideality factor
1.01
1.2
In the measurements, the value of the series resistance has been found to be 350-
550 Q. The large series resistance is attributed to the large spacing between the contacts.
The values of n calculated are in general agreement with those reported for Schottky
diodes on n-GaN samples. "^ The ideality factor of neariy unity in one diode indicated
the current flow by thermionics in the small forward bias region. In the measurements,
the ideality factor was found to range between 1 OleV - 2.4 eV for some diodes. Higher
values of ideality factor indicate the nonideal behaviour of the diodes. It suggests that
apart from thermionic, other ttansport mechanism may be present in the diode.
The reverse breakdown voltage for Au and Ni Schottky diodes were measured.
The range was between -lOV and -30V for Au/n-GaN and in the range of-lOV and -40
V for Ni/n-GaN diodes. The reverse bias current was found to be ~90nA for -20V for
79
Ni/nGaN and ~60nA for Au/n-GaN. There was a variation in the breakdown voltage for
different diodes. This variation was associated with the non-uniformities in the electron
concentration of the sample. However, once the reverse bias vohage exceeded the range,
subsequent measurements showed degradation of the reverse characteristics of the
diodes. The reverse characteristics were not investigated in detail because the diodes
were not passivated by a mesa etch or similar measure.
The barrier height is calculated using the theoretical value of the Richardson
constant. The electron affinity 2^ of the semiconductor is defined as the energy difference
between an electron at the bottom of the conduction band and an electron at the vacuum
level. It is related to the work fimction of the metal and the barrier height by
^s-^m~ ^Bn • Due to the charges present at the junction, this relation is only true for
the ideal case. Taking the value of the barrier height from the measurements, the electron
affinity is calculated to be 4.2eV for Au/n-GaN and 4.249 eV for Ni/n-GaN
The barrier height obtained from CV is larger than that from I-V . The value of
the barrier height from the capacitance-voltage measurements are found to be 0.2 eV
more than that measured from the I-V characteristics. This behavior is attributed to the
fact that the transport mechanism in these diodes is not purely due to thermionic
emmisions alone, other transport mechanism are also present in the diodes. The barrier
height from the I-V measurements are voltage or electric field sensitive while that
measured from C-V are not. The difference between the barrier height can also be due to
the presence of a native oxide layer at the metal semiconductor interface. Almost all
Schottky diodes have a thin oxide layer or insulatmg layer. These layers contribute to the
80
larger value of Vo and thus, give a higher value for the barrier height. An other possible
reason for the increased barrier height can also be due to the deep level. If the deep levels
emit carriers to the conduction band, it can affect the capacitance measurements. Since
the capacitance-vohage measurements signified that the deep levels are not so prominent
in the measurements, this reason was not taken. According to Werner and Guttler, "^
spatial inhomogeneities at the metal-semiconductor interface of the Schottky contacts
also can contribute to the difference in the barrier height determined from the I-V and C-
V measurements.
5.4 Results and Discussions of Capacitance-Voltage Measurements using Two-Mask process
During the measurements it was observed that the sample was not uniform with
regard to the concentration of electrons. It ranged from 3x10 cm' to 10 cm" .
The areas where Au contacts were made were regions which showed more variation in
the distribution of electron concentration than the areas where the Ni contacts were made.
This section describes the various capacitance voltage measurements done on the Au/n-
GaN and N-/n-GaN schottky diodes fabricated using two masks. Figure 5.9 and Figure
5.10 shows the capacitance-vohage curves for Au/n-GaN Schottky contacts where a
carrier concentration of 10^ cm' and 10^^ cm' were observed. The barrier height and
carrier concentrations for these ranges are given in Table 5.4 and Table 5.5, respectively.
The capacitance-Vohage measurements for Ni/n-GaN is shown in Figure 5.11 and
the barrier height and carrier concentrations are summarized in Table 5.6.
81
Figure 5.9 Capacitance-Vohage measurements of Au/n-GraN using the two-mask process for different frequencies. The carrier concentration was approximately 10'*cm-l
82
Figure 5.10 Capacitance-Vohage measurements of Au/n-GaN using the two mask process for different frequencies. The carrier concentration was approximately 10 cm' -3
83
Figure 5.11 Capacitance-Voltage measurements of Ni/n-GaN using the two-mask process for different frequencies. The carrier concentration was approximately 10 ^ cm' .
84
Table 5.4 Summary of C-V measurements on Au/n-GaN using two-mask process
Measurement Frequency
IMHz
400KHZ
200KHz
lOOKHz
40KHz
20KHz
lOKHz
Barrier Height
2.18eV
2.05eV
2.04eV
2.02eV
1.96eV
2.01eV
2.02eV
Carrier Concentration
3.71 x lO' cm'
3.64xl0'^cm'^
3.66xl0^^cm'^
3.62xl0*^cm'^
3.60xl0^^cm'^
3.63 x lO^ cm'
3.41 x lO^ cm"
Table 5.5 Summary of C-V measurements on Au/n-GaN using two-mask process
Measurement Frequency
IMHz
400KHZ
200KHz
lOOKHz
Barrier Height
1.28eV
1.08eV
1.19eV
1.20eV
Carrier Concentration
5.628 x lO'^cm''
4.36xl0'^cm'^
5.75 X lO' cm'
5.85 x lO^^cm'
85
Table 5.6 Summary of C-V measurements on Ni/n-GaN using the two-mask process.
Measurement Frequency
IMHz
400KHZ
200KHZ
lOOKHz
40KHz
20KHz
lOKHz
Barrier Height
2.45eV
2.40eV
2.38eV
2.33eV
2.32eV
2.3 leV
2.30eV
Carrier Concentration
2.62xl0^^cm'^
2.57xl0'''cm'^
2.61 X lO^ cm'
2.53 X lO' 'cm'
2.54 X lO' 'cm'
2.32 X lO' 'cm'
2.51 X lO' 'cm'
From Figure 5.9, it was observed that for the Au/n-GaN Schottky diodes, the
measured capacitances in the kilohertz range was approximately the same. The carrier
concentration was found to be in the vicinity of 3x 10 ^ cm'l The barrier height was
found to be large, -2.0 eV, whereas, the Au/n-GaN Schottky diodes, which had a carrier
concentartion near 10^ cm'^ the barrier height was 1.2eV. The Ni/nGaN also showed the
same characteristics as that of the Au/n-GaN shown in Figure 5.9. The barrier height was
found to be large.
The presence of interfacial layers was assumed as a reason. The interfacial
reactions may have occurred during annealing of the ohmic contacts before the deposition
of the Schottky contacts. An other possible reason for the formation of the interfacial
layer may be during the evaporation of the Schottky contacts. During the procedure of
86
evacuation of the evaporation chamber, the oil vapor from the oil-pumped roughing
systems might have condensed on the semiconductor before the metal was deposited,
resuhing in the formation of an interfacial layer. The interfacial layer has an effect on the
capacitance. The layer modifies the dependence of charge in the depletion region on the
bias vohage. The capacitance of the depletion region is in series with that of the
interfacial layer. Thus, the capacitance is a complicated parameter.
The interfacial layer may have a different work fimction, which is responsible for
the increase in barrier heights. The effect of an interfacial layer on capacitance has been
analyized by Cowley, the analysis assumes that the interfacial layer is thin. The
interfacial sates that may exist at the interfacial insulator/semiconductor interface are
filled or emptied by the tunneling of electrons from the metal contact. Under such
conditions, the interface states need not be taken into account in calculating the
capacitance. The slope of such states remained straight. From the measurements of the
Au/n-GaN with a concentration of 10 cm', it was observed that the 1/C versus V plot
remained linear.
When an interface layer is present, the flat vohage is modified as
Vft-Oto-^ + ^1 IrT \/f I^T\/2
'2 kT y2 0
+—*-V ^) "^
where VA, is the flatband vohage in the presence of a insulating layer, ^ is the separation
between the Fermi level and the conduction band minimum, 0^^ is the zero-bias barrier
height, Vdo the band bending at zero bias, and 0^ = la^q y^ . The interfacial layer is
characterized by a parameter a - 5s l^s^ + qSDg), where S is the thickness of the layer,
87
£s and £•, are the permittivity of the semiconductor and the interfacial layer, Ds is the
interface state density. The barrier height increased due to the presence of the interfacial
layer.
In the observation, h was found that some of the diodes the interfacial layer was
so thick that the graph of C' vs V was not linear. It was concluded that the thickness of
the interfacial layer was also not uniform in the sample.
5.5 Results and Discussions of Current-Voltage Measurements using Two-Mask process
The forward current-vohage characteristics of Au/n-GaN and Ni/n-GaN are
shown in Figure 5.12 and Figure 5.13, respectively. Figure 5.14 and Figure 5.15 shows
the reverse current-voltage characteristics of the Au/n-GaN and Ni/n-GaN Schottky
diodes. The effective barrier height and ideality factor were determined.
In the measurements, it was observed that the current-voltage characteristics
deviated from the linear behaviour in the forward bias region due to the series resistance,
similar to the condition in the one mask process. As a result, accurate determination of lo
and n cannot be obtained from the I-V plot. The series resistance was calculated and the
vohage was corrected and the corrected vohage across the junction is given by Vd= V-
IRs.
A plot of the natural log of current versus forward bias vohage for small forward
currents have been plotted. The family of ln(I) vs Vd plot was investigated. The curves
are linear over the entire voltage range. Accurate value of ideality factor n. Is and barrier
height were calculated. The results are summarized in Table 5.7
88
1E4
1E«.
1E€.
1E-7
1E6.
^E9r.
1
'•
"•
•'
•
-
0
0
\ 0
0
0
0
* 1
1 > 1 • 1 \ . .V ' T • • •
• * *
/ 0
0
0
0
0
0
0
0
1 . 1 . 1 . 1 .
-i 1
. 1
1 ll
lll
l_.J
1
1
1 1
llll
l 1
1 1 ll
lll
-
-1
1
1 1
1 ll
ll
[ill 1 1
1
--
-;
04 08 12
Voltage (V)
1.6 20
Figure 5.12 Forward I-V characteristics of Au/n-GaN diodes using the two-mask process.
89
1E4
1B6.
1B6
1B-7.
1E«
l E Q .
-" •
^ z . -
lllll
-
"
r
-
_
-
~ -
-
-
~
4
-" 0
0
r»
?
' 1
0
0
0
$
$
0
0
0
0
0
0
0
0
0
*
1
* »
1
^
1
» *
1
«
1
«
— r
•
,
# «
1 . •
1
• •
T -
•
1
« • .1
1
• " • - ' - T —
,
_
•
. z --•
;
--
_ ;
-
-
^ ; "
-
; . --
-:
-
04 08 ^2
\A)ltage(\0
1.6 20
Figure 5.13 Forward I-V characteristics of Ni/n-GaN Schottky diode using the two- mask process.
90
I C Q
1E«
1E-7
1E8
1E8
1E-10
1 1
1 iiiii|
c . . -
_ • • •
m
' m m
m
• •
• * • -
• " m "
- ' I 1
• •
: i
1 . 1 . 1 . 1 . •
-10 -8
Volt£ge(V)
-2
Figure 5.14 Reverse I-V characteristics of Au/n-GaN Schottky diodes using the two-mask process.
91
1E6
1&7
1E8
1EQ
E-10
_•
!
-
— I —
« «
_J
1 > 1 1 1 1 1 r 1 1 1 ,
«
* %
* «
\
i , 1 . 1 . 1 . 1 1 1 .
1 ' -
-
; --
-
-
-
-
-
-
•
K 1 1 -
- « - ^ -30 -25 -20 -15 -10
Voltage (V)
Figure 5.15 Reverse I-V characteristics of Ni/n-GaN Schottky diodes using the two-mask process.
92
Table 5.7 Summary of the characteristics of the Schottky barriers from I-V measurements using the two-mask process.
Metal
Au/n-GaN
Ni/n-GaN
Barrier height
0.85eV
0.86eV
Ideality factor
1.3
1.02
In the measurements, the value of the series resistance has been found to be in the
kilo ohm range. The large series resistance is attributed to the large spacing between the
contacts. The values of n calculated are in general agreement with those reported for
Schottky diodes on n-GaN samples. "^ In the measurements, the ideality factor was found
to range between 1.OleV - 2.47 eV for some diodes. Higher values of ideality factor
indicate the nonideal behaviour of the diodes. It suggests that in addition to the
thermionic effect, other transport mechanism may be present in the diode.
The reverse breakdovm vohage for Au and Ni Schottky diodes were observed.
The range was between -lOV and -30V for Au/n-GaN and in the range of-lOV and -50
V for Ni/n-GaN diodes. The reverse bias current was found to be ~50nA for -20V for
Ni/nGaN and ~80nA at -5 V for Au/n-GaN. There was a variation in the breakdown
voltage for different diodes. This variation was associated with the non-uniformities in
the electron concentration of the sample. However, once the reverse bias voltage
exceeded the range, subsequent measurements showed degradation of reverse
93
characteristics of the diodes. The reverse characteristics were not investigated in detail
because the diodes were not passivated by a mesa etch, or similar measure.
The barrier height is calculated using the theoretical value of Richardson constant.
The barrier height obtained from CV is larger than that from I-V.
5.6 Conclusion
In summary, two different processing techniques for the fabrication of Schottky
diodes have been compared. The one-mask process have given better results in terms of
barrier height. A barrier height of 1.2 eV for Au/n-GaN and 1.18 eV for Ni/n-GaN is
obtained using this method and is better than the results published so far for Au and Ni.
The results obtained from the two-mask process was not very good when compared with
that of one mask, due to the possible formation of an interfacial layer between the contact
and the semiconductor. As a result, the barrier height increased. It was also observed that
the interfacial layer was not uniform giving rise to some good diodes with a barrier height
of 1.2eV for Au/n-GaN Schottky diodes.
94
CHAPTER VI
CONCLUSIONS
Gallium Nitride (GaN) is one of the few instances in the history of
semiconductors where the device applications have even come before the properties of
the semiconductor were understood completely. The technology of GaN is progressing
quhe rapidly which suggests that the electronic potential of III-V based nitrides can be
realized; GaN based electronics will be a reality.
The application of GaN includes various fields like displays, data storage, UV
detectors, high power microwave applications, new sensor technologies. A recent report
of an implanted GaN FET suggests that implanted IC technology is possible. As a better
understanding of the properties of this material evolve, more rapid development in the
device applications will be seen. The interest in GaN has exploded in the past few years,
leading some to predict that GaN will eventually become a very important semiconductor
next to Si and GaAs.
The research done in this thesis promises good Schottky and ohmic contacts to n-
type GaN semiconductor. The multilayer (Ti/Al/Ti/Au) formed good ohmic contact to n-
type GaN. The Au and Ni proved to be good Schottky contacts.
GaN device development has recently emphasized GaN based FETs. Good
Schottky contacts are a prerequishe. The processing technique for the fabrication of the
Schottky diodes and the metal semiconductor contacts used in this work are now used for
17. I. Akasaki and H. Amano, in GaN, edited by J.I.Pankove and T.D.Moustakas. Academic, New York. (1998), Vol. 1, pp 459-72.
18. H.Amano, N.Sawaki, I.Akasaki, and Y.Toyoda, Appl. Phys. Lett. 48,353 (1986); H.Amano, I.Akasaki, T.Kozawa, K.Hiramatsu, N.Sawaki, K.Ikeda, and Y.Ishii, J.Lumin. 40&41, 121 (1988); H.Amano, M.kito,K.Hiramatsu, and I.Akasaki, Jpn. J. appl. Phys., Part 2 28, L2112 (1989
19. S.Nakamura and G.Fasol, The Blue Laser Diodes. Springer, Heidelberg. (1997).
20. J.Pearton, J.C.Zolper, RJ.Shul, and F.Ren, J. Appl. Phys.86, 1 (1999).
21. M. A.Khan, J.N.Kuznia, A.R. Bhattarai and D.T.Olson, "Metal semiconductor field effect transistor based on single crystal GaN," Appl. Phys. Lett. 62, 1786 (1993).
22. C.H.Qiu, J.I.Pankove, C.Rossington: In Nitride Based X-Ray Detectors, ed. By J.I. Pankove, T.D.Moustakas. Semiconductors and Semimetals, Vol. 50B (Academic, San Siedo 1998) pp. 1-24.
23. Robert F.Pierret, Semiconductor Device Fundamentals . Addison-Wesley PubHcation. (1996).
65. M.W.Cole, F.Ren, and S.J.Pearton, J.Electrochem. Soc. 144, L275(1997); Appl. Phys. Lett. 71,3004(1997).
66. D.Qiao, L.S.Yu and S.S.Lau, Appl. Phys. Lett. 87, 801(2000).
67. X.A.Cao, H.Cho, and S.J.Pearton, Appl. Phys. Lett. 75, 232 (1999).
68. S.K.Cheung and N.W.Cheung, Appl. Phys. Lett. 49, 85 (1986).
69. James H.Edgar, Samuel Strite, Isamu Akasaki, Hirushi Amano and Christian Wetzel, "Properties, Processing and Applications of Gallium Nitride and related Semiconductors." INSPEC, London, United Kingdom. (1999).
70. Q.Chen, J.W.Wang, A.Osinsky, S.Gangopadhyay, B.Lim, M.Z.Anwar, and M.A.Khan, Appl. Phys. Lett. 70, 2277 (1997).
71. Dieter K.Schroder, ""Semiconductor Material and Device Characterization." John Wiley «fe Sons, Inc. (1998)
72. S.M.Sze, ""Physics of Semiconductor Devices. " John Wiely & Sons, New York. (1981).
73. J.H.Wemer and H.H.Guttler, J.Appl.Phys. 69, 1522 (1991).
100
PERMISSION TO COPY
In presenting this thesis in partial fulfillment of the requirements for a
master's degree at Texas Tech University or Texas Tech University Health Sciences
Center, I agree that the Library and my major department shall make it freely
available for research purposes. Permission to copy this thesis for scholarly
purposes may be granted by the Director of the Library or my major professor.
It is understood that any copying or publication of this thesis for financial gain
shall not be allowed without my further written permission and that any user