7/30/2019 96603724 5th Chapter Thesis 2nd http://slidepdf.com/reader/full/96603724-5th-chapter-thesis-2nd 1/47 HEMT Simulations 1 Department of ECE, NITC CHAPTER 1 INTRODUCTION 1.1 Introduction: High Electron Mobility Transistor (HEMT) is also known as Heterostructure Field Effect Transistor (HFET) or Modulation-doped Field Effect Transistor (MODFET), is a Field Effect Transistor (FET) incorporating a junction between two different materials with different band gaps (i.e. a heterojunction) as the channel instead of a doped region, as is generally the case for Metal Oxide Field Effect Transistor (MOSFET) . It is a device that exploits the high electron mobility in an undoped region to achieve high speed operation. Film deposition technique like epitaxy is used to create undoped region, then narrow undoped region, when applied a MBE electron well which forms the channel for current flow. Electrons from suitable bias, work as a quantum mechanical surrounding doped regions of the device are trapped in the quantum well resulting in a high concentration of electrons in the channel. This channel is below the surface of the device and separated from surface which reduces surface scattering. As the doping in the channel is uninitiated, lack of scattering sites in the channel results in high electron mobility. In addition, the channel itself is normally constructed from a material which possesses high mobility such as InGaAs. A commonly used material combination is GaAs with AlGaAs, in this work combination of GaAs (substrate), InGaAs (channel) with AlGaAs used. This combination is known as pseudomorphic HEMT (PHEMT) [1]. InGaAs channel is epitaxially grown on a GaAs substrate. Second variant is AlInAs/GaInAs HEMTS grown lattice-matched on InP substrates. These two combinations used in pseudomorphic HEMT.
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.
High Electron Mobility Transistor (HEMT) is also known as Heterostructure
Field Effect Transistor (HFET) or Modulation-doped Field Effect Transistor
(MODFET), is a Field Effect Transistor (FET) incorporating a junction between two
different materials with different band gaps (i.e. a heterojunction) as the channel
instead of a doped region, as is generally the case for Metal Oxide Field Effect
Transistor (MOSFET). It is a device that exploits the high electron mobility in an
undoped region to achieve high speed operation. Film deposition technique like
epitaxy is used to create undoped region, then narrow undoped region, when applied
a MBE electron well which forms the channel for current flow. Electrons from
suitable bias, work as a quantum mechanical surrounding doped regions of the device
are trapped in the quantum well resulting in a high concentration of electrons in the
channel. This channel is below the surface of the device and separated from surface
which reduces surface scattering. As the doping in the channel is uninitiated, lack of
scattering sites in the channel results in high electron mobility. In addition, the
channel itself is normally constructed from a material which possesses high mobilitysuch as InGaAs. A commonly used material combination is GaAs with AlGaAs, in
this work combination of GaAs (substrate), InGaAs (channel) with AlGaAs used.
This combination is known as pseudomorphic HEMT (PHEMT) [1]. InGaAs channel
is epitaxially grown on a GaAs substrate. Second variant is AlInAs/GaInAs HEMTS
grown lattice-matched on InP substrates. These two combinations used in
This chapter discusses the classification of Semiconductors and then
compound semiconductor properties are discussed. Hetero junction function isdiscussed along with HEMT operation. Semiconductors can be broadly classified intotwo categories.
Elemental semiconductors C, Si, Ge, Sn are widely used in VLSI application.However, Due to their inferior properties of compound semiconductors are when bothelectronic and mechanical properties are better when compared to elementalsemiconductors.
2.1 Types of Compound Semiconductors:
Compound semi conductors are semiconductor materials obtained by mixing
two or more elements. To get a compound semiconductor from different elements the
following criterions must be satisfied.
1. The average valence should be four.
2. Mole fraction contribution group should be same that of its complimentary
group.
3. There shouldn’t be any chemical any chemical compound formation or
chemical reaction between different elements which are used in compound
semiconductor.
4. Band gap of resultant material should be in the range of semiconductor
band gap.
5. Conductivity modulation by doping.
6. The bonding should be of more covalent than ionic.
Depending up on the number of elements, compound semiconductors are classified
Quaternary compound semiconductors are made by combining from elements
generally from columns III and V are mixed to obtain compound semiconductor.
Widely used compound semiconductor is GaxIn1-xAs1-yPy. As the mole fraction isvaries different properties can be varies. The In/Ga and As/P ratios may be varied
independently. GaInAsP is used for near-infrared laser diodes employed in optical
communication [2].
2.2 General Properties:
GaAs, InxGa1-xAs and AlxGa1-xAs are the semiconductors considered in this
study. So the properties of their semiconductors are studied in detail. Fig. 2.1 shows
energy band gap versus lattice constant for various semiconductors. It can be seen
that lattice constant of Si and GaAs are very close. In HEMT devices, generally GaAs
is used as substrate and InGaAs or GaAs is used as the channel region. In this work
mole fraction discusses In 0.15 and mole fraction of Al is 0.2.
Different semiconductors are formed between different materials, there
junctions will be discussed.
2.3 Hetero junctions:
Hetero junction is a junction formed by two dissimilar semiconductormaterials, across the junction, one semiconductor material will have larger band gap
(denoted by capital letter) and other material will have narrow band gap (denoted by
small letter). Depending upon the conductivity type across the junction, hetero
junction can be classified into iso type and aniso-type. Depending upon the band
alignment, Hetero junctions can be classified into three categories. They are
1. Type I
2. Type II3. Type III
2.3.1 Type I:
Type I heterostructure is the most common. In type I hetero junction, both
conduction band and valence band of narrow band gap material lie completely
between conduction band and valence band of wide band gap material. Schematic
band gap diagram of type I hetero junction is shown in Fig.2.4. An important example
of a type I heterostructure is the GaAs–AlGaAs materials system. In a type I
Heterostructure, the sum of the conduction band and valence band edge
discontinuities is equal to the energy gap difference,∆E = ∆E + ∆E (2.11)
In Type II hetero junction, only one of conduction band or valence band of
narrow band material is in between the conduction band and valence band of the wide
band gap material. Schematic band representation of type II material is shown in
Fig.2.5. The band gap discontinuity in this case is given as the difference between the
conduction band and valence band edge discontinuities. A type II Heterostructure is
formed by Al0:48In0:52As and InP.
Fig.2.5. Type II hetero junction
2.3.3 Type III:
Type III heterostructure, both conduction band and valence band of narrowband material never lie between conduction band and valence band of wide band gap
material. Schematic band representation of type III is shown in Fig.2.7. Example of
type III hetero structure is GaSb and InAs.
As for the type II case, the band gap discontinuity is equal to the difference
between the conduction band and valence band edge discontinuities. Certainly, one of
the most important Heterostructure is that formed between GaAs and AlAs or its
related ternary compounds, AlGaAs. The GaAs–AlGaAs heterostructure has the
additional feature of close lattice matching. Two materials that have nearly identical
lattice constants are when used to form a Heterostructure. If the materials are not
lattice matched, the lattice mismatch can be accommodated through strain or by the
With respect to acceptor level, unoccupied is neutral and occupied is negative.
The neutral level lineup does not occur, charge transfer takes place between the states
on either side of the interface and this would create a dipole field. From the energy
considerations the system favors dipole minimization and therefore equalization of E0 levels. In the hetero junction a band bending take place such that the conduction band
edge in the neutral region of InGaAs is raised by ∆Ec so that net electron flow in
thermal equilibrium is zero. Therefore the potential drop S in InGaAs in thermal
equilibrium is higher than in the homogeneous junction by ∆Ec /q.
cEHetero Homo
S S q
(2.12)
Due to the discontinuity ∆Ec in the conduction band of InGaAs and GaAs, theband bending in the undoped InGaAs is more in the GaAs homojunctions of similar
doping levels. Due to this effect, large concentrations of electrons are present at the
InGaAs surface adjacent to GaAs and they remain there due to notch in the
conduction band. The electrons have actually been supplied from GaAs layer which
has been doped. These electrons are in the In0.2Ga0.8As region where doping
concentration is low.
2.5 Strain at Heterointerfaces:
Bulk crystalline semiconductor is that it exhibits perfect or nearly perfect
translational symmetry. In other words, suitable translations of the basic unit cell of a
crystal restore the crystal back into itself. Atoms within the crystal are regularly
spaced throughout the entire bulk sample. This assumption is true for bulk materials.
However, two important exceptions can arise. The first is that a bulk crystal can
include impurities and dislocations such that the perfect periodicity of the material is
disrupted locally. The crystal can still retain its overall highly ordered structure, yet
contain local regions in which perfect periodicity is disrupted by impurities ordislocations. These impurities and dislocations can significantly affect the properties
of the material. The second situation arises in multilayered structures. Crystal growth
technology has enabled the growth of thin layers of heterogeneous semiconductor
material called Heterostructure. Using exacting crystal growth procedures,
Heterostructure can be grown with atomic layer precision. A very thin layer of
material can be grown on top of or sandwiched between layers grown with a different
type of semiconductor material, even materials in which the lattice constant is
different. When a thin layer of material is grown either on or between layers of a
different semiconductor that has a significantly different lattice constant, the thin,
epitaxial layer will adopt the lattice constant of the neighboring layers provided that
the lattice mismatch is less than about 10% as can be seen from Figure 2.9, When the
thin, epitaxial layer adopts the lattice constant of the surrounding layers, it becomes
strained, i.e., it is either compressed or expanded from its usual bulk crystal shape.
There exists a maximum thickness of the thin layer below which the lattice mismatch
can be accommodated through strain. For layer thickness above the critical thickness,the lattice mismatch cannot be accommodated through strain, dislocations are
produced and the strain relaxes as is seen in Fig.2.10.
The strain within the layer is homogeneous. The strained layer can be in either
compressive or tensile strain. If the lattice constant of the strained layer is less than
that of the surrounding layers the system is in tension. Conversely, if the lattice
constant of the strained layer is greater than that of the surrounding layers, the
strained layer is in compression. Heterojunctions may be comprised of
Pseudomorphic layers in which one of the layers is lattice mismatched and coherent,
or elastically strained. For a given strain, determined by the lattice mismatch, a
maximum thickness exists, called the critical thickness hc, for the strained layer to be
completely coherent with the substrate. The critical thickness has a minimum under
conditions of thermodynamic equilibrium, but can be enhanced by the epitaxial
process. For thicknesses greater than the critical thickness, dislocations are created to
reduce the energy of the system. These dislocations can significantly degrade device
Pseudomorphic growth was first investegated by methews and
Blakeslee.he is experiments with alternating layers of GaAs and GaAsP to derive a
theoritical expression for critical layer thickness (hc) which canot be exceeded if misfit dislocations must be avoided.the thickness hc upto which the mismatch of the
InxGa1-xAs layer can be accomodated by elastic strain is given by eqn. (2.13)
0 .0 7 x =[ √ ]
√ ( )(2.13)
Where x is the indium content, a is GaAs lattice constant (a=0.565nm), and is
poisson’s ratio (0.23 for GaAs). For x=0, the growth is lattice matched, magnitude of hC is infinite (same as AlGaAs/GaAs HEMT). For large x critical thickness is only a
few nanometers. Quantum wells that are too narrow contain only a small number of
charge carriers, and the transport properties of these carriers are degraded due to
excessive interface scatering. Transistor performnance is found x=0.2 and a thickness
12nm. Transsistor performnance is number of transistors in the channel and carrier
mobility.
2.7 Compound Semiconductor Properties:
Properties of compound semiconductors will be discussed below. In this work
I will discuss GaAs, AlxGa1-xAs, InxGa1-xAs material properties. Mole fraction of
2.8 Potential Well and Two Dimensional Electron Gas:
Different compound semiconductors are epitaxially grown on device their
Band gaps and electron affinity are different so that the interface forms a
discontinuity and potential well is created. AlGaAs and InGaAs forms Triangularpotential well. Electrons in the interface of two regions are confined in this well it
forms Two-Dimensional Electron Gas (2DEG). 2DEG is a gas of electrons free to
move in two dimensions, but tightly confined in the third. This tight confinement
leads to quantized energy levels for motion in that direction, which can then be
ignored for most problems. Thus the electrons appear to be a 2D sheet embedded in a
three dimensional (3D) world. Dopants placed in AlGaAs layer. Due to the energy
band structure, electrons from these dopants are confined to the channel layer where
they form 2DEG. The sheet carrier density of 2DEG and the electron mobility
measured at room temperature are 2.1x1012 cm2 and 8600 cm2 /V s, respectively.
Electrons are confined in the channel device operated in inversion region.
2.9 HEMT Structure:
HEMT channel is undoped and physically removed from the ionized donors,
and because the electrons travel in the quantum well parallel to the heterointerface,
the HEMT electron mobilities are more typical of ultra-pure bulk semiconductors. Itshould be noted that p-type modulation doping is also feasible though there is wide
variation, dependent on the application of the device. Devices incorporating more
indium generally show better high-frequency performance [9], while in recent years;
GaN HEMTs have attracted attention due to their high-power performance [10]. To
allow conduction, semiconductors are doped with impurities which donate mobile
electrons (or holes). However, these electrons are slowed down through collisions
with the impurities (dopants) used to generate them in the first place. HEMTs avoidthis through the use of high mobility electrons generated using the heterojunction of a
highly-doped wide-band gap n-type donor-supply layer (AlGaAs in our example) and
a non-doped narrow-band gap channel layer with no dopant impurities (InGaAs in
The electrons generated in the thin n-type AlGaAs layer drop completely into
the InGaAs layer to form a depleted AlGaAs layer, because the heterojunction created
by different band-gap materials forms a quantum well (a steep canyon) in the
conduction band on the InGaAs side where the electrons can move quickly without
colliding with any impurities because the InGaAs layer is undoped, and from which
they cannot escape. The effect of this is to create a very thin layer of highly mobile
conducting electrons with very high concentration, giving the channel very low
resistivity (or to put it another way, "high electron mobility"). This layer is called a
Two-Dimensional Electron Gas (2DEG). As with all the other types of FETs, a
voltage applied to the gate alters the conductivity of this layer. Ordinarily, the two
different materials used for a heterojunction must have the same lattice constant
(spacing between the atoms). In semiconductors, these discontinuities form deep-level traps, and greatly reduce device performance. HEMT formed from different
components of elements are called compound semiconductors. Basic HEMT diagram
is shown in below fig.2.11.
Fig.2.11. HEMT basic block diagram
Different combinations of HEMT are as shown in Table.2.4.
In this work single hetero junction pseudomorphic HEMT technique is used.
2.10 Working of HEMT:
In this topic Band diagram, Gate functionality, operation of HEMT discussed.
2.10.1 Band Diagram:
Two semiconductors with different energy band gaps are grown on each other,
discontinuities of the conduction and valence band edges arise at the heterointerface.
Two semiconductors with different band gap energies are joined together the
difference is divided up into a band gap offset in the valence band ∆EV and a band
gap offset in the conduction band ∆EC. In most cases conduction band offset between
adjacent layers ∆EC is close to 2/3 of the total band gap difference ∆Eg. One of the
most common made assumptions for the AlGaAs/InGaAs material system is 40 %
valence band offset and 60 % conduction band offset. This is only valid for Alcontents below about 45 %. For higher Al contents the band gap of AlGaAs changes
from direct to indirect. Energy band diagram with gate voltage is -0.6 V as shown in
Fig.2.12. AlGaAs layer electron concentration increased from top to bottom. Sudden
discontinuity is due to change of compound semiconductors. The conduction band of
the channel relative to the Fermi level EF is determined by ∆EC, the doping level ND,
the barrier height of the Schottky contact (qFB), the gate to channel separation, and
the applied voltage on the gate VGS. Gate voltage is 0 compare to -0.6 volts
conduction band discontinuity is more and more number of electrons are confined in
the channel. Up to 0.6 um AlGaAs layer there electron concentration increases and
later 0.6 to 0.7 um InGaAs channel where 2DEG formed and after GaAs substrate
region electron concentration decreases. InGaAs layer is to provide better electron
confinement and superior electron transport characteristics. In these devices, the sheet
concentration of 2DEG is increased over comparable AlGaAs /GaAs structures due to
the increase of the conduction band-edge discontinuities at the heterojunctions.
Electron mobility and steady state saturation velocity in InGaAs are intrinsicallyhigher than those in GaAs. There are only two controllable parameters that exert a
strong influence on the electron density: the donor density and the conduction-band
discontinuity. Band diagram of HEMT applied voltages -0.6 and 0 volts as shown in
Fig.2.12 and Fig.2.13.
Fig.2.12 Band Diagram applied gate voltage is -0.5 V
Fig.2.13. Band Diagram applied gate voltage is 0 volts
2.10.2 Gate Functionality:
A Schottky barrier is a rectifying metal semiconductor contact while ohmic
contact has linear relationship between the voltage and current, Schottky contact in
equilibrium. There is no current flow in equilibrium and that gradient of Fermi levelis zero. So Fermi level is flat everywhere in equilibrium. At the metal-semiconductor
in equilibrium the Fermi levels must align. Consider the behavior of metal first Fermi
level lies above the conduction band edge. The energy needed that an electron be
ionized, escape from the metal and enter the vacuum level, is called the work function
energy or qm. Most of the conduction electrons within the metal are at energy
reasonably close to the Fermi level. Therefore Work function is the energy difference
between the vacuum level and Fermi level. In a nondegenerate semiconductor theirFermi level doesn’t lie above the conduction band edge but some is somewhere in the
forbidden gap. The semiconductor work function is also difference between the Fermi
level and vacuum level. Since most of the electrons in semiconductors are not at the
Current flowing through the device is depends on electron concentration in the
channel. Its controlled gate Schottky barrier diode formed on AlGaAs. Drain current
flow due to electron transport through the n-type AlGaAs layer should be avoided byensuring that this layer is fully depleted wider in gate region. This voltage is known
OFF-voltage. It is same as like threshold voltage in MOSFET. Voltage required to
turnoff the depletion. HEMT AlGaAs doped layer should not be use for conducting
otherwise current flows through it. Only channel layer is used for current transport.
Current flow in the device should take place only through electron in the notch. They
are high mobility electron. Electron concentration can be reduced by decreasing the
potential drop on the P-side. It is possible by forward bias the AlGaAs and InGaAs.That voltage is EC /q. OFF-voltage is the difference between voltage required to
depleted to entire region is Vbi-Vpo where Vpo is pinch of voltage, Vbi is the build
in potential and Remove charges in InGaAs(no conduction of electrons). voltage
required to depleted to entire region is Vbi-Vpo where Vpo is pinch of voltage, Vbi is
the build in potential. and Remove charges in InGaAs(no conduction of electrons) is
EC /q. Vpo is pinch of voltage depletes the channel completely.
V( OFF) = V
−V
−(
∆) (2.11)
V = l n
(2.12)
V = ∈∈ (2.13)
where a is GaAs layer thickness and ND is dopant density of AlGaAs. Whenever
reverse gate voltage is applied, the conducting channel between the source and drain
can be choked off completely, under these conditions, the device is said to be in
pinch-off.
Depends on OFF-voltage weather the device is depletion or enhancement decided.
Gate voltage is less than OFF-voltage device is depletion mode and gate voltage is
greater than OFF-voltage device in enhancement mode. –ve voltage on gate will
The HEMT was originally developed for high speed applications. It was only
when the first devices were fabricated that it was discovered they exhibited a very
low noise figure. This is related to the nature of the 2DEG and the fact that there are
less electron collisions. As a result of their noise performance they are widely used in
low noise small signal amplifiers [13], power amplifiers, oscillators [14] and mixers
operating at frequencies up to 60 GHz and more and it are anticipated that ultimately
devices will be widely available for frequencies up to about 100 GHz. In fact HEMT
devices are used in a wide range of RF design applications including cellular
telecommunications, direct broadcast receivers - DBS, radar, radio astronomy, and
any RF design application that requires a combination of low noise and very high
frequency performance. HEMTs are manufactured by many semiconductor devicemanufacturers around the globe. They may be in the form of discrete transistors, but
now a days they are more usually incorporated into integrated circuits. These
Monolithic Microwave Integrated Circuit chips (MMIC) are widely used for RF
design applications, and HEMT based MMICs are widely used to provide the
required level of performance in many areas.
The GaAs-based PHEMTs are used extensively for solid state power amplifier
[15] applications at microwave and millimeter-wave frequencies.
Table.3.1 Specifications of HEMT process simulations
3.3 Device Simulations:
The process simulated structure is then exported in to device simulator
MEDICI. The exported file consists region information, dopant information, contact
information and structure information of the process simulated structure. To
successfully simulate the electrical characteristics, one has to consider and choose
appropriate model. The following subsection discusses various models.
3.3.1 Mobility Models [18]
Carrier mobility in estimating the currents play an important role. There arevarious mobility models which can be used in simulating the currents. The following
mobility models are considered in this study.
1. Constant mobility
2. Concentration dependent mobility
3. Analytical mobility
4. Arora mobility
5. Carrier-carrier mobility
3.3.1.1 Constant Mobility:
In the constant mobility model, mobility values are constant with respect to
doping concentration, temperature etc. Electron mobility for InGaAs layer (µ on) is
2.73x104 cm2 /V.sec and hole mobility for InGaAs layer (µ op) is 480 cm2 /V.sec.
Values in to eqn. (3.7) to eqn. (3.10) the mobility model result isμ = M UN 1. AR0 (3.11)
μ = M UP1 . AR0 (3.12)
Then from eq. (3.11) and eqn. (3.12) it can be observed that arora mobility
model transform in to constant mobility model for all temperature and doping
concentration. Then arora mobility model is not considered for device simulations.
3.3.5 Carrier-carrier mobility
The carrier-carrier consideration mobility model also takes carrier-carrier
scattering effects, in to this is based on work by Dorkel and Leturcq [19]. While
estimating mobility values the carrier-carrier scattering effects are important whenhigh concentrations of electrons and holes are present in a device. The model also
takes into account the effects of lattice scattering and ionized impurity scattering. The
electron hole mobilities are modeled by eqn. (3.13) and eqn. (3.14) respectively.
μ = μ (.
. μμ. − C.LIC) (3.13)
μ = μ ( .. μμ. − C.LIC) (3.14)
Where the superscripts L, I and C stand for lattice scattering, ionized impurity
scattering and carrier-carrier scattering, respectively; n and p stand for electron and
hole respectively.
μ = μ +μ (3.15)
μ = μ + μ (3.16)
Eqn. (3.14) and eqn. (3.15) is Effective mobility calculated using matthiessen’s rule.
[12] Frank Ellinger, “Radio Frequency Integrated Circuits and Technologies”, 2nd
edition, springer , 2008.
[13] Malmkvist M., Lefebvre E. , Borg M. , Desplanque L. , Wallart x., DambrineG., Bollaert S., Grahn, J., “Electrical Characterization and Small-Signal Modelingof InAs/AlSb HEMTs for Low-Noise and High-Frequency Applications,” Microwave Theory and Techniques,56, pp.2685-2691, 2008.
[14] Kudszus, S. Haydl, W.H. Tessmann, A. Bronner, W. Schlechtweg, M.Fraunhofer Inst. for Appl. Solid State Phys., Freiburg, “Push-push oscillators for94 and 140 GHz applications using standard pseudomorphic GaAs HEMTs,” Microwave Symposium Digest , 3, 2001.
[15] “Taurus TSUPREM-Taurus Process Reference Manual” synopsis, Version X-
2005.10, October 2005.
[16] Sorab K Ghandhi,” Vlsi Fabrication Principles: Gallium Arsenide” 2nd edition John Wiley & sons, New York, 1994.
[17] “Taurus Medici Medici User Guide” synopsis, Version D-2010.03, March2010.
[18] M. Dorkel and Ph. Leturcq, “Carrier Mobilities in Silicon Semi-EmpiricallyRelated to Temperature, Doping, and Injection Level,” Solid-State Electronics,24,pp. 821-825, 1981.
[19] Nhan, E. , Sheng Cheng, Jose M.J. , Fortney S.O. , Penn J.E. , “ Recent test
results of a flight X-band solid-state power amplifier utilizing GaAs MESFET,HFET, and PHEMT technologies,” GaAs Reliability Workshop,2002.