NANOELECTRONICS

Module 1
“A biological system can be exceedingly
small. Many of the cells are very tiny, but
they are very active; they manufacture
various substances; they walk around; they
wiggle; and they do all kinds of marvelous
things—all on a very small scale. Also, they
store information. Consider the possibility
that we too can make a thing very small that
does what we want—that we can
manufacture an object that maneuvers at
that level.”
(From the talk “There’s Plenty of Room at
the Bottom,” delivered by Richard P.
Feynman at the annual meeting of the
American Physical Society at the California
Institute of Technology, Pasadena, CA, on
December 29, 1959.)
INTRODUCTION TO NANATECHNOLOGY

• Nano – a billionth (1 x 10-9)

• Nano technology deals with various structures
of matter having dimensions of the order of a
billionth of a meter.
• The word nanotechnology may be new but
nanostructures have existed on Earth as long
as life itself.
• Present Microelectronic Technology
To produce smaller, faster, and more powerful logic
devices.
The number of transistors per integrated circuit will
double every 18–24 months due to advancements in
technology is commonly referred to as Moore’s Law,
after Intel founder Gordon Moore, who made the
prediction in a 1965 paper with the
prophetic title “Cramming More Components onto
Integrated Circuits. ”
At the time he thought that his prediction would hold until at
least 1975; however, the exponentially increasing rate of
circuit densification has continued into the present (Graph
4.1).

In 2000, Intel introduced the Pentium 4 containing 42 million

transistors, an amazing engineering achievement.
The increases in packing density of the circuitry are achieved
by shrinking the line widths of the metal interconnects, by
decreasing the size of other features, and by producing
thinner layers in the multilevel device structures.
These changes are only brought about by the development of
new fabrication techniques and materials of construction.
Fundamental Physical Limitations of Present
Technology
This top-down method of producing faster and more
powerful computer circuitry by shrinking features
cannot continue because there are fundamental
physical limitations, related to the material of
construction of the solid-state-based devices, that
cannot be overcome by engineering.
For instance, charge leakage becomes a problem when
the insulating silicon oxide layers are thinned to about
three silicon atoms deep, which will be reached
commercially by 2003–2004.
Moreover, silicon loses its original band structure when
it is restricted to very small sizes.
The lithography techniques used to create the circuitry
on the wafers has also neared its technological limits,
although derivative technologies such as e-beam
lithography, extreme ultraviolet lithography (EUV), and
x-ray lithography are being developed for commercial
applications.
A tool capable of x-ray lithography in the sub-100 nm
range has been patented.
The staggering increase in cost is due to the extremely
sophisticated tools that will be needed to form the
increasingly small features of the devices.
As devices increase in complexity, defect and
contamination control become even more important as
defect tolerance is very low — nearly every device must
work perfectly.
For instance, cationic metallic impurities in the wet
chemicals such as sulfuric acid used in the fabrication
process are measured in the part per billion (ppb) range.
With decreases in line width and feature size, the
presence of a few ppb of metal contamination could lead
to low chip yields.
Therefore, the industry has been driving suppliers to
produce chemicals with part per trillion (ppt)
contamination levels, raising the cost of the chemicals
used.
Depending on the complexity of the device, the
number of individual processing steps used to make
them can be in the thousands.
It can take 30–40 days for a single wafer to make it
through the manufacturing process. Many of these
steps are cleaning steps, requiring some fabs to use
thousands of gallons of ultra-pure water per minute.
The reclaim of waste water is gaining importance in
semiconductor fab operations.
The huge consumption of water and its subsequent
disposal can lead to problems where aquifers are low
and waste emission standards require expensive
treatment technology.
A new technology that addressed only one of the
potential problems we have discussed would be of
interest to the semiconductor industry.
A new technology would be revolutionary if it produced
faster and smaller logic and memory chips, reduced
complexity, saved days to weeks of manufacturing time,
and reduced the consumption of natural resources.
 Applications
• Nanoparticles
Nanoscopic particles for coatings, as catalysts in chemical
processing, in organ-specific drug delivery etc.
• Nanomaterials
Materials with improved strength and weight, perhaps
incorporating carbon nanotubes.
• Nanoelectronics
Extremely small, low-power electronic devices, such as
transistors.
• Nano-optics
Forming optical circuits at the nano-scale, often using
plasmonic structures.
• Nano-magnetics
Using nanoscopic magnetic materials and devices for
magnetic storage technologies, logic devices etc.
 Applications
• Nanofluidics
Devices based on fluid properties at the nanoscale
(leading to what are called chemical and biological
laboratories on a chip).
• Nanobioelectronics
Combinations of nanodevices and biological
structures for, examples , chemical/bioweapons
detection.
 Limitations
• Device Fabrication
Difficult to extend optical lithography into the realm of
low tens of nanometer and other fabrication method
not as yet mature, commercially.
For extremely small devices, random process variations
can lead to device characteristics that are
unpredictable and thus complicating circuit design.
• Device operation
As device dimensions are reduced, voltage levels also
need to be reduced accordingly. This lowers the
threshold voltage of MOSFET devices, and makes it
difficult to completely turn the device off, wasting
power.
Furthermore, due to the law of quantum physics,
conventional MOSFETs will behave differently
(tunneling and ballistic transport will arise).
• Heat Dissipation
As device density increases, the dissipation of the heat
becomes a major problem, reducing circuit reliability and
leading to shorter device lifetimes, or to device failure.
With current technologies, if the rate of increasing device
density were to continue, microprocessors would soon be
producing more heat per cm2 than the surface of the sun.
10 W/cm2 – Hot plate
100 W/cm2 – Current IC power density
7000 W/cm2 – Surface of the Sun
• TRENDS IN NANOELECTRONICS AND OPTOELECTRONICS
• The evolution of microelectronic devices is influenced by
factors such as
– growing demands in memory capacity of integrated circuits
– high transmission data speed
– optical communications etc.
• This requires electronic logic devices (DRAM etc.) with
– faster speed operation
– smaller size
– Powerful operation.
• Figure 1.1 shows the reduction of feature size of metal-
oxide-semiconductor (MOS) transistors, as well as the
number of bits per chip for the period 1970–2000.
Figure 1.1. Evolution of the minimum feature size of a Si DRAM.
• For example, a 256 M-bit DRAM contains about 109
transistors with a feature size L ∼ 100nm. For structures
with these dimensions, transport can still be treated
classically.
• Today, expects feature sizes still one order of magnitude
lower, i.e. L∼10nm; but below this size, transistors based
on new concepts like single electron transistors, resonant
tunneling devices, etc. (based on quantum mechanics).
• The operation of this new kind of devices has to be
described by the concepts of mesoscopic and quantum
physics.
• In the near future, and due to the growing demands
of calculus from industries like communications,
information, military, space, etc. microelectronics will
be replaced by nanoelectronics since the feature size
of electronic devices will be reduced to about 10nm.
• But, the progress in the development of new system
architectures dealing with billions of transistors is
scarce.
• In this sense, new architectures for parallel data
processing are under current development.
• The evolution towards nanoelectronics has been possible
because of advances in the deposition of very thin films
to form heterostructures in which electrons could be
con- fined to a 2D mesoscopic system.
• During the 1980s both molecular beam epitaxy (MBE)
and metal organic chemical vapour deposition (MOCVD)
became available.
• In MBE, a film of the desired material is epitaxially grown
over a substrate under ultrahigh vacuum conditions.
• Very thin films of elemental semiconductors (Si, Ge), III-V
(AlGa, AlGaAs, InP), and II-VI compounds (CdTe, PbS) can
be deposited by MBE.
• Although the deposited films are of very high quality, the
technique is somewhat slow and expensive.
• MOCVD is the preferred industrial tech nique
to produce semiconductor heterostructures.
• This technique allows for simultaneous
deposition on several wafers and is used for
the commercial production of semiconductor
lasers.
• In general, mesoscopic systems require the formation of
nanostructures in the range close to 100nm.
• Therefore at present we are close to the limits of
conventional optical lithography, and other high resolution
nanolithographic techniques (electron beam, ion beam, x-
ray, etc.) have to be industrially implemented.
• As for resists, the most commonly used in nanolithography
is the positive tone resist known as PMMA (poly-
methylmethacrylate). Although the molecular weight of
PMMA is close to 106, its roughness, once spin coated onto
the substrate, is only about 2nm.
• ULTIMATE LIMIT
• It seems reasonable that the rate of scaling
down predicted by Moore’s law [2] will have
to slow down.
• It is expected that the limits of further
miniaturization, from an industrial and
economic point of view, will be reached in
about one decade.
• The technological limits are related to several
factors of which we will mention only two.
• The first has to do with the amount of heat generated
by the power consumed, and which cannot be
eliminated because of thermal conductivity limits of
the materials and the increasing number of overlayers.
• The other factor is related to the so-called “parameter
spread” in fabrication.
For instance many of the electrical parameters of the MOS
transistors are set by doping; however, as the size of the
region to be doped is decreased to about 0.1µm3, the
number of doping atoms becomes so low (about 10) that the
parameter spread cannot be controlled appropriately.
• In addition to the above mentioned technological
limits, there are others of fundamental nature, which
are called physical limits.
– Thermal limit
– Relativistic limit
– Quantum-Mechanical limit
• (i) Thermal limit
The energy necessary to write a bit should be several
times kT which is the average energy of thermal
fluctuations.
In CMOS the lowest values contemplated to write a bit
should not be smaller than about 2eV, i.e. 100kT at
room temperature or ∼3 × 10−19 J.
• (ii) Relativistic limit: Signals cannot propagate faster
than the speed of light. Therefore assuming that the
nucleus of a microprocessor has a size of a few cm, it
takes 10−10 sec. for the signal to propagate, which
corresponds to a frequency of about 10GHz.
• (iii) Uncertainty principle
According to the Heisenberg’s uncertainty principle, the
energy and time needed to write or read a bit should
be related by E·t ≥ h. To be safe, we ask the product E·t
to be 100h. Since for future circuits E could be as low as
10−19 J, we can appreciate that we can approach the
quantum limit as the frequency increases.
• Solid state nanoelectronics consists of heterostructures
of well-known materials (Si, SiO2, III-V compounds),
and several types of transistors: heterojunction, single-
electron, resonant-tunnelling, ballistic, etc.
• However, the amount of computational capacity for
some tasks like speech and visual recognition is so
large that other radically different alternatives are
being sought.
• Some of these alternatives, like superconductivity
electronics and spintronics, use fabrication techniques
not too different from those employed in present
integrated circuit technologies.
• Superconducting electronics, proposed in the 1970s,
and developed to the prototype stage during the
1980s, is based on the switching properties of the
Josephson junction, which consists of two
superconducting layers separated by a very thin oxide
insulating film that can be tunnelled by
superconducting pairs of electrons.
• The advantages of superconducting electronics are
based on the fact that Josephson junctions can operate
at high switching speeds (switching times between 1
and 10ps), the amount of dissipated power is very low
and the resistance of interconnect superconducting
lines is practically null.
• Another technology being pursued is spintronics
which exploits the spin orientation of electrons.
• Electron-spin transistors are built by enclosing a
semiconductor layer (base) sandwiched between
two ferromagnetic layers (emitter and collector).
• Electrons acquiring the magnetization state of the
emitter can only travel through the collector
across the base if their spins are aligned with the
magnetization of the collector.
• These developments canbe considered as
running parallel to the efforts in
magnetoelectronics to develop MRAM
memories, based both in the giant
magnetoresistance effect and in the magnetic
tunnel junctions developed in 1995.
• Electron-spin transistors have a big potential if
successfully integrated with CMOS circuits.
• Other radically new alternatives for future
nanoelectronics have been proposed. Molecular
electronics is based on the different states or
configurations that molecules can take, like “trans” or
“cis”, as well as parallel or antiparallel alignment of the
spin of unpaired electrons.
• The change between states must be fast, consume little
energy, should be addressed by some external signal,
and should be readable by a probe.
• If this technology is able to be put to work, we will
have the ultimate step in miniaturization, since
molecules are much smaller than present feature sizes
in integrated circuits.
• Besides, molecules have the advantage of being
capable of self-organization in 3D supramolecular
entities, although, after the development of scanning
atomic force microscopy, molecules can be, in
principle, individually manipulated.
• Examples of molecules that can be used in molecular
electronics are: azobenzene, hydrazobenzene, etc.
• One advantage of organic molecules, when
compared to inorganic ones, is that it is easier to
isolate them in single molecular systems.
• Although in the future it is expected that
molecular wires or nanotubes can be
developed for contacting molecules, today
only metallic or semiconducting electrodes are
contemplated.
• Even with this limitation, the interfacing of
molecules to the external world for
addressability in the case of large systems
seems, at present, an enormous problem.
• Lastly, we mention biology-inspired electronics,
also called bioelectronics.
• In trying to copy nature, we are not concerned
with the size of the building blocks, since for
instance, a neuron is very large for
nanotechnology standards.
• What nanoelectronics would like to imitate of
biological neurons is the capabilities in parallel
processing as well as their 3D architectures and
the topology of the interconnects.
• This results from the large number of
computations, needed for instance in pattern
recognition of the visual systems of humans and
animals, which has to be performed
simultaneously at different sites.
• In addition to parallel processing, neural
networks try to simulate the integration of
computing and memory functions, which in
CMOS microprocessors are performed separately.
• To conclude this section, the present situation in
optoelectronics will be shortly considered.
• Optoelectronic devices, based mainly in III-V direct gap
semiconductors, have received a great upsurge since
the development of optical fibre communications.
• In addition, there is at present a tendency to replace,
whenever possible, electronic devices by photonic
ones.
• The III-V semiconductors more frequently used are
based on AlGaAs–GaAs and GaInAsP–InP
heterostructures which cover the 0.8–1.6µm
wavelength range.
• GaN blue lasers for short wavelength applications
were developed about ten years ago.
• In the last two decades, quantum well
semiconductor lasers with very low threshold
currents and photodetectors are replacing the
conventional ones, especially in long distance optical
communications.
• One very interesting type of quantum well lasers,
which operate at still lower threshold currents, is
based on strained-layer quantum heterostructures.
• At present, laser diodes are manufactured in chips by
standard integrated circuit technology, coupled to
transistors and optical interconnects, constituting the
so-called optoelectronic integrated circuits (OEIC).
• In all cases, the trend in optoelectronic devices is to
achieve a high level of integration which implies
smaller sizes, but still in the micron range.
• Here again there is a lot of research in efficient
integration architectures.
• Perhaps, the greatest advances in optoelectronics
based on quantum semiconductor heterostructures
can be found in the field of electro-optical signal
modulation.
• In effect, modulators based on the confined quantum
Stark effect (Section 8.4) are several orders of
magnitude more effective than their bulk counterparts.
• As we will see in Section 10.8, this is due to the fact
that excitons in quantum wells have much higher
ionization energy than in the bulk, and therefore can
sustain much higher electric fields.
• 4 important aspects where nanomaterials
differ from macroscale materials.
– Gravitational force becomes negligible and
electromagnetic forces dominate.
– Quantum mechanics is used to describe motion
and energy instead of the rules of classical
mehanics.
– Greater surface-to-volume ratio.
– Random molecular motion becomes more
important.
• Mesoscopic
– Refers to solid state structures and devices that
are large compared to the atomic scale but but
small compared to macroscopic scale.
– This dimension is comparable to the elastic
scattering length of electrons. An electron can
travel a large distance before suffering a scattering
event.
• Properties of Some nano materials.
– Carbon Nano Tube (CNT) is stronger than steel. It
exhibits good conducting and semiconducting
behaviour.
– Nanostructured ceramics are stronger than coarser
grained ceramics.
– Gold is inert in bulk form, but catalyses CO to CO2, if
the diameter is a few nm.
– Gold appears as deep red to black in solution
depending on its size.
– Melt at much lower temperature than gold slabs.
• Properties of Some nano materials.
– Absorption of solar radiation in photovoltaic cell is
much higher in materials composed of nanoparticles
than same in thin films of continuous sheets of
material.
– Titanium dioxide particles of nanosize in sun screen
film can absorb harmful ultraviolet radiation.
– The buckyball contains 60 carbon atoms (70, 84, ….,
500 etc.)and its properties are quite different from
that of carbon bulk.
• Properties of Some nano materials.
– A cluster of few atoms behaves quite differently
from a bulkmaterial. Copper is an opaque
substance in bulk but it becomes transparent
when the size reaches about a few nanometer. A
stable material like Aluminium becomes
combustible one. Platinum is inert but at
nanoscale it becomes a good catalyst.
• Properties of Some nano materials.
– Two important classes of nanomaterials are
fullerenes and inorganic nanoparticles.
– The fullerenes are a class of allotropes of carbon
rolled into tubes or spheres. These include carbon
nano tubes (CNT), a widely used nano material. It
has a large mechanical strength and electrical
property.
– Very low cost solar panels may be fabricated using
CNT.
NANOELECTRONICS

Module 1
Present Microelectronic Technology
• The evolution of microelectronic devices is influenced
by factors such as
– growing demands in memory capacity of integrated
circuits
– high transmission data speed
– optical communications etc.
• This requires electronic logic devices (DRAM etc.) with
– faster speed operation
– smaller size
– Powerful operation.
The number of transistors per integrated circuit will
double every 18–24 months due to advancements in
technology is commonly referred to as Moore’s Law,
after Intel founder Gordon Moore, who made the
prediction in a 1965 paper with the prophetic title
“Cramming More Components onto Integrated
Circuits. ”
At the time he thought that his prediction would hold until at
least 1975; however, the exponentially increasing rate of
circuit densification has continued into the present (Graph
4.1).

In 2000, Intel introduced the Pentium 4 containing 42 million

transistors, an amazing engineering achievement.
The increases in packing density of the circuitry are achieved
by shrinking the line widths of the metal interconnects, by
decreasing the size of other features, and by producing
thinner layers in the multilevel device structures.
These changes are only brought about by the development of
new fabrication techniques and materials of construction.
• Figure 1.1 shows the reduction of feature size of
metal-oxide-semiconductor (MOS) transistors, as
well as the number of bits per chip for the period
1970–2000.
• For example, a 256 M-bit DRAM contains about 109
transistors with a feature size L ∼ 100nm. For
structures with these dimensions, transport can still
be treated classically.
Figure 1.1. Evolution of the minimum feature size of a Si DRAM.
• Today, expects feature sizes still one order of
magnitude lower, i.e. L∼10nm; but below this size,
transistors based on new concepts like single
electron transistors, resonant tunneling devices, etc.
(based on quantum mechanics).
• The operation of this new kind of devices has to be
described by the concepts of mesoscopic and
quantum physics.
• In the near future, and due to the growing demands
of calculus from industries like communications,
information, military, space, etc. microelectronics will
be replaced by nanoelectronics since the feature size
of electronic devices will be reduced to about 10nm.
Fundamental Physical Limitations of Present
Technology
The top-down method of producing faster and more
powerful computer circuitry by shrinking features
cannot continue because there are fundamental
physical limitations, related to the material of
construction of the solid-state-based devices, that
cannot be overcome by engineering.
• It seems reasonable that the rate of scaling down
predicted by Moore’s law will have to slow down.
• It is expected that the limits of further
miniaturization, from an industrial and economic
point of view, will be reached in about one decade
because of technological limits and fundamental
limits.
• Technological Limits
 Limitations
• Device Fabrication
Difficult to extend optical lithography into the realm of
low tens of nanometer and other fabrication method
not as yet mature, commercially.
(e-beam lithography, extreme ultraviolet lithography
(EUV), and x-ray lithography are being developed for
commercial applications).
For extremely small devices, random process variations
can lead to device characteristics that are
unpredictable and thus complicating circuit design.
• For extremely small devices, random process
variations (“parameter spread” in fabrication) can
lead to device characteristics that are unpredictable
and thus complicating circuit design.
For instance many of the electrical parameters of
the MOS transistors are set by doping; however, as
the size of the region to be doped is decreased to
about 0.1µm3, the number of doping atoms
becomes so low (about 10) that the parameter
spread cannot be controlled appropriately.
As devices increase in complexity, defect and
contamination control become even more important as
defect tolerance is very low — nearly every device must
work perfectly.
For instance, cationic metallic impurities in the wet
chemicals such as sulfuric acid used in the fabrication
process are measured in the part per billion (ppb) range.
With decreases in line width and feature size, the
presence of a few ppb of metal contamination could lead
to low chip yields.
Therefore, the industry has been driving suppliers to
produce chemicals with part per trillion (ppt)
contamination levels, raising the cost of the chemicals
used.
As complexity of the device increases processing steps
also increases. Many of these steps are cleaning steps,
requiring thousands of gallons of ultra-pure water per
minute.
The huge consumption of water and its subsequent
disposal can lead to problems where aquifers are low
and waste emission standards require expensive
treatment technology.
• Device operation
As device dimensions are reduced, voltage levels also
need to be reduced accordingly. This lowers the
threshold voltage of MOSFET devices, and makes it
difficult to completely turn the device off, wasting
power.
Furthermore, due to the law of quantum physics,
conventional MOSFETs will behave differently
(tunneling and ballistic transport will arise).
• Charge leakage becomes a problem when the
insulating silicon oxide layers are thinned to about
three silicon atoms deep.
• Moreover, silicon loses its original band structure
when it is restricted to very small sizes.
• Heat Dissipation
As device density increases, the dissipation of the heat
becomes a major problem, reducing circuit reliability and
leading to shorter device lifetimes, or to device failure.
With current technologies, if the rate of increasing device
density were to continue, microprocessors would soon be
producing more heat per cm2 than the surface of the sun.
10 W/cm2 – Hot plate
100 W/cm2 – Current IC power density
7000 W/cm2 – Surface of the Sun
• The amount of heat generated by the power
consumed, and which cannot be eliminated because
of
– thermal conductivity limits of the materials and
– increasing number of overlayers.
• Material Limit
Independent of the particular devices and circuits, if
devices are built with same material.
The parameters depend on material are
Mobility
Carrier saturation velocity
Self-ionizing electric field strength
Thermal conductivity
• Circuit Limit
For example, in order to distinguish between the zero
and one logic levels with nearly zero error, a CMOS
inverter should have 𝑉𝑑𝑑 ≥ 𝑉𝑑𝑑(𝑚𝑖𝑛)
2𝑘𝑇 𝐶𝑓𝑠 𝐶𝑜
𝑉𝑑𝑑(𝑚𝑖𝑛) = 1+ 𝑙𝑛 2 +
𝑞 𝐶𝑜 + 𝐶𝑑 𝐶𝑑
𝛽𝑘𝑇
≥
𝑞
Where
𝑉𝑑𝑑 = 𝑆𝑢𝑝𝑝𝑙𝑦 𝑣𝑜𝑙𝑡𝑎𝑔𝑒
𝑉𝑑𝑑(𝑚𝑖𝑛) = 𝑀𝑖𝑛. 𝑙𝑖𝑚𝑖𝑡 𝑜𝑓 𝑉𝑑𝑑
𝐶𝑓𝑠 = 𝐶𝑕𝑎𝑛𝑛𝑒𝑙 𝑓𝑎𝑠𝑡 𝑠𝑢𝑟𝑓𝑎𝑐𝑒 𝑠𝑡𝑎𝑡𝑒
𝐶𝑜 = 𝐺𝑎𝑡𝑒 𝑜𝑥𝑖𝑑𝑒 𝑐𝑝𝑎𝑐𝑖𝑡𝑎𝑛𝑐𝑒
𝐶𝑑 = 𝐶𝑕𝑎𝑛𝑛𝑒𝑙 𝑑𝑒𝑝𝑙𝑒𝑡𝑖𝑜𝑛 𝑟𝑒𝑔𝑖𝑜𝑛 𝑐𝑎𝑝𝑎𝑐𝑖𝑡𝑎𝑛𝑐𝑒
𝛽 = 2 to 4
Usually, at T=300 ˚K, 𝑉𝑑𝑑(𝑚𝑖𝑛) ≈ 0.1𝑉
• Cost
The staggering increase in cost is due to the extremely
sophisticated tools that will be needed to form the
increasingly small features of the devices.
• Small Dimensional Effects
Short channel and Narrow width Effects
Sub-threshold current
Increase of leakage/off current
Hot Electrons
Velocity saturation
Source drain resistance
Doping fluctuations
• Fundamental Limits
• In addition to the above mentioned technological
limits, there are others of fundamental nature, which
are called physical limits.
– Thermal limit
– Relativistic limit
– Quantum-Mechanical limit
 Thermal limit (Thermodynamic limit)
• (i) Thermal limit
The energy necessary to write a bit should be several
times kT which is the average energy of thermal
fluctuations.
In CMOS the lowest values anticipated to write a bit
should not be smaller than about 2eV, i.e. 100kT at
room temperature or ∼3 × 10−19 J.
• (ii) Relativistic limit (Electromagnetic Limit)
Signals cannot propagate through an interconnect faster
than the speed of light.
If 𝐿 is the length of the interconnect and 𝜏 is the
interconnect transit time, and 𝑐𝑜 is the velocity of light in
free space, the constraint is represented as
𝐿
≤ 𝑐𝑜
𝜏
Therefore assuming that the nucleus of a microprocessor
has a size of a few cm, it takes 10−10 sec. for the signal to
propagate, which corresponds to a frequency of about
10GHz.
• (iii) Uncertainty principle (Quantum theoretical Limit)
According to the Heisenberg’s uncertainty principle, the
energy and time needed to write or read a bit should be
related by Et ≥ h.
Or the power P should be
𝑕
𝑃≥ 2
𝑡
To be safe, we ask the product E·t to be 100h. Since for
future circuits E could be as low as 10−19 J, and can
approach the quantum limit as the frequency increases.
Thermal limit (Thermodynamic limit)
NANOELECTRONICS

Module 1
• Mesoscopic physics deals with structures which have a size
between the macroscopic world and the microscopic or atomic
one.
• These structures are also called mesoscopic systems, or
nanostructures since their size usually ranges from a few
nanometres to about 100nm.
• The electrons in such mesoscopic systems show their wavelike
properties and therefore their behaviour is dependent on the
geometry of the samples.
• In this case, the states of the electrons are wavelike and
somewhat similar to electromagnetic radiation in waveguides.
• For the description of the behaviour of electrons in solids it is
very convenient to define a series of characteristic lengths.
• If the dimensions of the solid is comparable or smaller than
these characteristic lengths, the material might show new
properties, which in general are more interesting than the
corresponding ones in macroscopic materials.
• In fact, the physics needed to explain these new properties is
based on quantum mechanics.
• In other words, a mesoscopic system approaches its
macroscopic limit if its size is several times its characteristic
length.
 CHARACTERISITC LENGTHS IN
MESOSCOPIC SYSTEMS
• Most commonly used characteristic lengths in
mesoscopic systems are
– de Broglie wavelength
– Mean free path
– Diffusion length
– Screening length
– Localization length
• de Broglie wavelength
For an electron having momentum 𝑝, there corresponds
a wavelength given by the de Broglie wavelength,
ℎ
𝜆𝐵 =
𝑝
ℎ
= ∗ … … . (1)
𝑚 𝑣
where 𝑚∗ is the electron effective mass.
Inside a semiconductor electron behaves as if their mass
is 𝑚∗ instead of the mass 𝑚𝑂 of the electron in vacuum.
In certain semiconductors like GaAs or InSb, 𝑚∗ is
much smaller than 𝑚𝑂 .
𝑚∗ = 0.067𝑚𝑂 and (for GaAs)
𝑚∗ = 0.014𝑚𝑂 (for InSb)
Smaller the value of the 𝑚∗ , the easier will be to observe
the size quantum effects in nanostructures of a given size.
So, with present lithographic techniques, it is relatively
easy to construct semiconductor nanostructures with one
or two of their dimensions (length, width and thickness)
of the order of , or smaller than 𝜆𝐵 .
• Mean free path
As the electron moves inside a solid, it is usually scattered by
interactions with crystal imperfections like impurities, defects,
lattice vibrations (phonons), etc.
In most cases, these scattering events or “collisions” are
inelastic, that is, the values of energy and momentum of the
system after the interaction, differ from the corresponding
ones before they interact.
The distance covered by the electron between two inelastic
collisions is usually called the mean free path 𝑙𝑒 of the
electron in the solid. If 𝑣 is the speed of the electron, then
𝑙𝑒 = 𝑣𝜏𝑒 … … (1.2)
where 𝜏𝑒 is known as relaxation time.
• Diffusion length
In a mesoscopic system of typical size 𝐿, and if 𝑙𝑒 ≫ 𝐿, the
particle moves throughout the structure without scattering.
This is called ballistic transport regime in which the
surfaces are the main scattering entities.
In ballistic regime, electron can reach energies much higher
than the ones corresponding to the lattice thermal energy
(Hot electron transistors).
If 𝑙𝑒 ≪ 𝐿, the transport can be explained as a diffusion
process, and the system is characterized by a diffusion co-
efficient 𝐷. In terms of 𝐷, the diffusion length 𝐿𝑒 is defined
by,
𝐿𝑒 = 𝐷𝜏𝑒 … … (1.3)
where 𝜏𝑒 is known as relaxation time.
In semiconductor theory, the concept of diffusion length
is used very often. Diffusion length of electron is the
average distance that an electron diffuses before
recombining with a hole.
If electrons diffuse within a p-type semiconductor, their
concentration diminishes exponentially with distance
with a decay length equal to 𝐿𝑒 .
In the diffusive regime, transport in the mesoscopic
system is usually explained by means of the Boltzmann
equation, as in the bulk.
In the ballistic regime, the Boltzmann transport model is
not valid, and electrons move through the structure
essentially without scattering.
• Screening length
In extrinsic semiconductors, the dopants or impurities
are usually ionized and constitute a main factor
contributing to scattering.
However, in general we cannot consider that the
electrical potential produced by these impurities varies
with distance as 1/r.
Because of the screening of free carriers by charges of
the opposite polarity, the effect of the impurity over the
distance is partially reduced.
It is found that the variation of the potential is modulated
by the term 𝑒 −𝑟 𝜆𝑠 where λs is called the screening
length and is given by
1
𝜖𝑘𝑇 2
𝜆𝑠 = 2 … … (1.4)
𝑒 𝑛
where
𝑒 is the electronic charge
𝜖 the dielectric constant of the semiconductor
𝑛 the mean background carrier concentration
One should be careful about nomenclature because
some authors call λs the Debye length or the Fermi–
Thomas length.
In a typical semiconductor, λs is in the range 10–100nm,
and is an indication of the attenuation of charge
disturbances in a semiconductor.
From Eq. (1.4) it is determined that λs should be much
smaller in metals than for semiconductors.
Figure 1.2 shows both the Coulomb potential and the
screened potential.
Screening potential, 𝜙𝑠𝑝 is
𝑒 −𝑟 𝜆𝑠
𝜙𝑠𝑝 = 𝛾 … … (1.5)
𝑟
where 𝛾 = 1 4𝜋𝜖𝑜
As λs →∞ the screening effect disappears, yielding the
Coulombic potential.
It can be observed from Figure 1.2 that for distances
from the impurity larger than about 2λs, the screening is
almost complete.
In the above discussion, we have considered that the
charge originating the potential was an impurity, but in
general it can be any perturbation in the charge
concentration.
Another interesting observation is that the above
equations imply a dielectric function 𝜖(𝑟) which is
distance dependent.
• Localization length
The localization length can be understood in terms of
transport in disordered materials. There can be localized
states in the disordered material and the movement of
electrons is by “hopping” transport between localized
states, or from a localized state to a bound state.
• In order to describe the hopping transport and other
mesoscopic properties of the localized states, it is
assumed that the electron wave function is described
by
𝜓 = 𝑒 −𝑟 𝜆𝑙𝑜𝑐 … … (1.6)

where 𝜆𝑙𝑜𝑐 is known as the localization length.

• Evidently, the electrical conductivity of a material will
be proportional, among other factors, to the overlap
between the wave functions. If the sample dimensions
are of the order of 𝜆𝑙𝑜𝑐 , we can say that our system is
mesoscopic.
NANOELECTRONICS

Module 1
QUANTUM MECHANICAL COHERENCE
• In a mesoscopic structure of dimensions similar to the
electron de Broglie wavelength 𝜆𝐵 , the behaviour of
the electron should be described quantum
mechanically, i.e. by using Schrödinger equation.
• If the electron interacts inelastically with a defect, or
any impurity, the electrons change their energy and
momentum, as well as the phase of their wave
function.
• The phase coherence length Lφ is defined as the
distance travelled by the electron without the carrier
wave changing its phase.
• Evidently, interference effects in the electron waves
should only be observed if the particles move over
distances of the order of, or smaller than, 𝐿𝜙 .
• Since we are usually interested in mesoscopic systems
in the quasi-ballistic regime, in which electrons are
practically unscattered, 𝐿𝜙 should be a length similar
to the inelastic scattering mean free path 𝑙𝑒 previously
defined.
• Coherent states can evidently show interference
effects. On the contrary, once the coherent states loose
their coherence, by inelastic scattering, the
corresponding waves cannot be superposed and
cannot interfere (in some way, the matter–wave
becomes a particle).
• In mesoscopic physics, the loss of coherence is usually
called dephasing. Evidently, the coherence processes
are characteristic of mesoscopic systems.
• From the above definition of 𝐿𝜙 , electrons can show
interference effects over distances smaller than 𝐿𝜙 .
• If electrons with phase 𝜙1 interfere with electrons of
phase 𝜙2 , we know from the wave theory that the
amplitude of the resultant wave varies as
𝐶𝑜𝑠 𝜙1 − 𝜙2 and the amplitudes can add up to each
other, or they can be subtracted depending on the
phase difference.
NANOELECTRONICS

Module 1
 Classification of Nano Structures
Low Dimensional Structures
• When one of the three spatial dimensions of a solid,
usually a semiconductor material, is of a size
comparable to 𝜆𝐵 , the material or a structure can be
called low dimensional structure.
• In some situations, other characteristic lengths,
different from 𝜆𝐵 , are more convenient to use.
• For most important semiconductors, 𝜆𝐵 usually ranges
between 10 and 100nm and therefore we have to deal
with solids of size in the nanometre range in order to
observe quantum effects such as QHE, Coulomb
blockade, quantized conductance, etc.
 Classification of Nano Structures

• Low-dimensional materials or structures are classified

according to the number of spatial dimensions of
nanometric size. These are
– Quantum well (2D)
– Quantum wires (1D)
– Quantum dot (0D).
QUANTUM WELLS, WIRES, AND DOTS
The series of characteristic lengths 𝜆 correspond to
physical properties of electrons which are size
dependent.
When the dimensions of the solid get reduced to a size
comparable with, or smaller than 𝜆, then the particles
behave wave-like and quantum mechanics should be
used.
Assume that an electron confined within a box of
dimensions 𝐿𝑥 , 𝐿𝑦 , 𝐿𝑧 . If the characteristic length is 𝜆,
we can have the following situations:
(i) 𝜆 ≪ 𝐿𝑥 , 𝐿𝑦 , 𝐿𝑧
In this case the electron behaves as in a regular 3D bulk
semiconductor.
(ii) 𝜆 > 𝐿𝑥 and 𝐿𝑥 ≪ 𝐿𝑦 , 𝐿𝑧
In this situation we have a 2D semiconductor
perpendicular to the x-axis. This mesoscopic system is
also called a quantum well.
(iii) 𝜆 > 𝐿𝑥 , 𝐿𝑦 and 𝐿𝑥 , 𝐿𝑦 ≪, 𝐿𝑧
This corresponds to a 1D semiconductor or quantum
wire, located along the z-axis.
(iv) 𝜆 ≫ 𝐿𝑥 , 𝐿𝑦 , 𝐿𝑧
In this case it is said that we have a 0D semiconductor or
a quantum dot .
 Ly

Lx
Lz

3D 2D 1D 0D
Bulk Quantum Quantum
Quantum
wire dot
well
NANOELECTRONICS

Module 1
Consider an electron in a one dimensional potential well
of width a.
The potential at the sides of the well is ∞ and the inside
the well is zero.
V=∞ V=∞

V=0

0 a

Fig.2: An infinite potential well of width a.

The wave function of the particle may be obtained by
solving the time independent Schrodinger wave equation
of
𝜕 2 𝜓 2𝑚
2
+ 2 𝐸 − 𝑉 𝜓 = 0 … … (1.1)
𝜕𝑥 ℏ
Inside the well V =0, then Eqn.1.1 becomes,
𝜕 2 𝜓 2𝑚
2
+ 2 𝐸𝜓 = 0 … … (1.2)
𝜕𝑥 ℏ
That is,
𝜕2𝜓 2 𝜓 = 0 … … (1.3)
+ 𝑘
𝜕𝑥 2
where
2𝑚𝐸
𝑘2 = 2
ℏ
8𝜋 2 𝑚𝐸
= 2
… … (1.3𝑎)
ℎ
Eqn. 1.1 has solution of the form
𝜓𝑥 = 𝐴 𝑆𝑖𝑛 𝑘𝑥 + 𝐵 𝐶𝑜𝑠 𝑘𝑥 … … (1.4)
Since 𝜓 = 0, at x = 0, 𝐵 = 0. Therefore,
𝜓𝑥 = 𝐴 𝑆𝑖𝑛 𝑘𝑥 … … (1.5)
The electron can not penetrate the infinite potential barrier.
Hence, at 𝑥 = 𝑎, 𝜓𝑎 = 0, so that
𝐴 𝑆𝑖𝑛 𝑘𝑎 = 0 … … (1.6)
Therefore, 𝑘𝑎 = 𝑛𝜋
𝑛 2𝜋 2
𝑘 2 = 2 … … (1.7)
𝑎
where 𝑛 = 1, 2, 3 … . .
From Eqn. 1.3a,
ℎ2 𝑘 2
𝐸 = 2 … … (1.8)
8𝜋 𝑚
Using Eqn.1.7 into 1.8, energy of electron in 𝑛𝑡ℎ state is
ℎ2 𝑛2
𝐸𝑛 = 2
… … (1.9)
8𝑚𝑎
From the basic postulates of quantum mechanics, the wave function 𝜓 is
normalized, and is
+∞
𝜓𝜓 ∗ 𝑑𝑥 = 1
−∞

where 𝜓 ∗ is the complex conjugate of 𝜓.

Here, this would become,
𝑎
𝐴𝐴∗ 𝑆𝑖𝑛2 𝑘𝑥 𝑑𝑥 = 1
0

𝑎
1 − 𝐶𝑜𝑠 2𝑘𝑥
𝐴2 𝑑𝑥 = 1
0 2

𝑎
𝑥 𝑆𝑖𝑛 2𝑘𝑥
𝐴2 − =1
2 4𝑘 0
But, 𝑆𝑖𝑛 2𝑘𝑎 = 0 𝑏𝑒𝑐𝑎𝑢𝑠𝑒 𝑘𝑎 = 𝑛𝜋.
2 2 𝑛𝜋𝑥
Therefore, 𝐴= and 𝜓= 𝑆𝑖𝑛
𝑎 𝑎 𝑎
The normalized wave function in the lowest three energy
states are
2 𝜋𝑥 𝜆
𝜓1 = 𝑆𝑖𝑛 , half cycle variation over 𝑎 (𝑎 = )
𝑎 𝑎 2

2 2𝜋𝑥
𝜓2 = 𝑆𝑖𝑛 , full cycle variation over 𝑎 (𝑎 = 𝜆)
𝑎 𝑎

2 3𝜋𝑥 3𝜆
𝜓3 = 𝑆𝑖𝑛 , three half cycle variation over 𝑎 (𝑎 = )
𝑎 𝑎 2
𝑛𝜆
For nth state, 𝑎 =
2
The energy levels and corresponding wave
function are shown in the Figure.
 Finite Square Quantum Well

• In practice, instead of infinite potential well, we will

encounter with potential well of finite height. Such a
structure can be modeled as follows.
 V(x)
Vo Vo

Region 1 Region 2 Region 3

V=0

-a/2 0 a/2 x
 Finite Square Quantum Well

This is the simplest possible potential which gives bound

state solutions to the Schrödinger equation. This
potential can be expressed as
𝑎
𝑉0 𝑤ℎ𝑒𝑛 𝑥 > (𝑜𝑢𝑡𝑠𝑖𝑑𝑒 𝑡ℎ𝑒 𝑤𝑒𝑙𝑙)
𝑉 𝑥 = 2
𝑎
0 𝑤ℎ𝑒𝑛 𝑥 ≤ (𝑖𝑛𝑠𝑖𝑑𝑒 𝑡ℎ𝑒 𝑤𝑒𝑙𝑙)
2

• There are three regions characterized by potentials V0,

0 and V0 for the regions 1,2 and 3. Schrödinger’s
equation for the 3 regions,
 Finite Square Quantum Well

• Outside the well

𝜕 2 𝜓1 2𝑚
2
+ 2 𝐸 − 𝑉0 𝜓1 = 0 … … (1.1)
𝜕𝑥 ℏ
Or
𝜕 2 𝜓1 2𝑚
2
− 2 𝑉0 − 𝐸 𝜓1 = 0
𝜕𝑥 ℏ
Or
𝜕 2 𝜓1 2 𝜓 =0
−∝ 1
𝜕𝑥 2
where
2
2𝑚
∝ = 2 𝑉0 − 𝐸
ℏ
 Finite Square Quantum Well

• Inside the well

𝜕 2 𝜓2 2𝑚
2
+ 2 𝐸 − 𝑉0 𝜓2 = 0 … … (1.1)
𝜕𝑥 ℏ

Since we take 𝑉0 = 0 inside the well,

𝜕 2 𝜓2 2𝑚
2
+ 2 𝐸𝜓2 = 0
𝜕𝑥 ℏ
Or
𝜕 2 𝜓2 2
2
+ 𝑘 𝜓2 = 0
𝜕𝑥
where
2
2𝑚
𝑘 = 2𝐸
ℏ
 Finite Square Quantum Well

• If the energy of the particle 𝐸 < 𝑉0, then it leads to

bound state solutions with quantized energy levels. In
the well region the solutions are of the form,
𝑎
𝜓2 𝑥 = 𝐴 𝑆𝑖𝑛 𝑘𝑥 + 𝐵 𝐶𝑜𝑠 𝑘𝑥, if 𝑥 <
2
Outside the well region, the solutions must vanish with
increasing 𝑥 and are hence of the form
𝛼𝑥
𝑎
𝐶𝑒 𝑤ℎ𝑒𝑛 𝑥 < −
𝜓1 𝑥 = 2
−𝛼𝑥
𝑎
𝐷𝑒 𝑤ℎ𝑒𝑛 𝑥 >
2
• The values of the constants 𝐴, 𝐵, 𝐶 and 𝐷 can be
determined by requiring that 𝜓(𝑥) and its first
derivative are continuous at ± 𝑎 2.
• The solution leads to various possible energy levels
and wave functions.
• Energy levels and corresponding wave functions are
shown in the Fig. below. Dotted lines represent energy
levels in infinite square well.
 Vo

5.2E1
E2 = 4E1

2.5E1
E1
0.6E1

Fig. First three energy levels in the finite well.

 V(x)

E (eV)

-a/2 0 a/2
x (nm)

Fig. Wave functions of first three states.

• Features
1. The energy of the particle is discrete. That is,
electron can move with certain energies only in the
well if E < V0.
2. The particle is free when E > V0. The particle is not
at all trapped in the well. The energy of the particle
is continuous.
3. The wave function penetrates into classically
forbidden region. There is a small probability to see
the particle outside the well. Hence the particle
shows tunneling.
4. Penetration of electron wave function into the
barriers is small for the lowest state.
5. It can be shown that there is always atleast one
bound state exists for the particle inside the well,
however small V0 be.
6. All these results can be proved through the
complete solution of Schrödinger equation for three
regions.
NANOELECTRONICS

Module 1
 Parabolic well
Parabolic well
• The potential profile due to harmonic oscillator is
1 2
𝑉 𝑧 = 𝑘𝑧 … … (4.10)
2
In simple harmonic motion particle oscillates under
parabolic potential. The plot of these potential energy
with z-co-ordinate is parabolic in shape and hence the
name.
 PARABOLIC AND TRIANGULAR
QUANTUM WELLS
Parabolic well
• In nano electronic systems, parabolic potential occurs
in many situation.
• The case of the parabolic well is well known in solid
state physics since the vibrations of the atoms in a
crystal lattice, whose quantification gives rise to
phonons, are described in a first approximation by
harmonic oscillators.
• In addition, a magnetic field applied to a two-
dimensional electron system gives rise to a parabolic
potential, and the electrons oscillate at the so-called
cyclotron frequency.
 PARABOLIC AND TRIANGULAR
QUANTUM WELLS

Parabolic quantum well profiles can also be produced by

the MBE growth technique.
In this case, alternate layers of GaAs and AlxGa(1−x)As of
varying thickness are deposited, increasing the thickness
of the AlGaAs layer quadratically with distance, while the
thickness corresponding to the GaAs layer is
proportionally reduced.
 Parabolic well
Parabolic well
Eqn.4.10 describes a classical harmonic oscillator.
For example, a mass 𝑚 on a spring sees the force (4.10),
where 𝑘 is the spring constant and 𝑧 is the displacement
from the equilibrium. This gives rise to harmonic motion
𝑧 𝑡 = 𝐴𝐶𝑜𝑠𝜔0 𝑡
where the frequency of oscillation is 𝜔0 = 𝑘/𝑚
• For the harmonic potential, Schroedinger equation for the
problem is
𝑑 2 𝜓(𝑧) 2𝑚 1 2
2
+ 2 𝐸 − 𝑘𝑧 𝜓(𝑧) = 0
𝑑𝑧 ℏ 2
• The solution of this equation is
𝑚𝜔0 1/2 𝑚𝜔0 2
𝜓 𝑧 = 𝐶𝑛 𝐻𝑛 𝑧 𝑒 2ℏ 𝑧
−
ℏ
where 𝐶𝑛 is a constant
1/2 1/4
1 𝑚𝜔0
𝐶𝑛 =
𝜋2𝑛 𝑛! ℏ
and where 𝐻𝑛 are Hermite polynomials, the first several of
which are
𝐻0 (𝑧) = 1
𝐻1 (𝑧) = 2𝑧
𝐻2 𝑧 = 4𝑧 2 − 2
𝐻3 𝑧 = 8𝑧 3 − 12𝑥,
etc.
The first several wave functions for the parabolic potential
resemble those for the square well potential.
• The energy values are equidistant and given by
1
𝐸𝑛 = 𝑛 − ℏ𝜔0 , 𝑛 = 1,2,3 … … … (4.11)
2
where ω0 is natural angular frequency.
The potential and wave functions are shown in the
Fig.4.5
• Fig.4.5: Harmonic oscillator potential well V(z). The first three
levels and corresponding wavefunctions are shown.
• Figure 4.5 shows the case of a parabolic well potential
and the wave functions of three electron bound states.
• The wave functions are mathematically expressed in
terms of the Hermite polynomials.
• Note, as in the case of the square well, the symmetric
or antisymmetric character of the wave functions and
their exponential decay in the forbidden energy zone.
.
Features
• Energy is quantized
• Minimum energy is non-zero.
• Energy increases in steps of ℏ𝜔0
• Wave functions show penetration into classically forbidden
region.
• For very large quantum number, the particle behaves like a
classical harmonic oscillator.
Triangular wells
• The triangular potential well is one of the most
common geometries, since the potential profile across
quantum heterojunctions , such as the well-known
modulation- doped AlGaAs–GaAs heterojunction, is
almost triangular in shape for electrons within GaAs.
• Another very important case, where an almost
triangular-shaped well is formed, occurs at the
semiconductor in a MOS structure.
• Figure 4.6 shows a triangular potential well, in which,
for simplicity, it is assumed that the left barrier is
infinite in energy and it increases linearly for z > 0:
𝑉 𝑧 = 𝑞𝐹𝑧, 𝑓𝑜𝑟 𝑧 > 0 … … (4.12𝑎)
𝑉 𝑧 = ∞, 𝑓𝑜𝑟 𝑧 ≤ 0 … … (4.12𝑏)
where q is a constant equal to the electron charge and F
is a uniform electric field.
 V(z)=qFz

• Fig.4.6: Triangular potential well V(z). The first three levels and
corresponding wave functions are shown.
• As in the other cases, the electron energies and states are
found by solving the Schrödinger equation subject to the
boundary condition ψ(z= 0) = 0.
• In this case, the eigenvalues are given in terms of the Airy
functions. However, for small values of n, it can be
demonstrated by applying the WKB quantum-mechanical
approximation (Wentzel-Kramers-Brillouin: for finding
approximate solutions to linear differential equations), that

2 3 1 3
3 1 𝑞 2 𝐹 2 ℏ2
𝐸𝑛 ≈ 𝜋 𝑛 − , 𝑛 = 1,2 … … … (4.13)
2 4 2𝑚
• Figure 4.6 shows the spacing between the energy
levels, which get a little closer as n increases, in
contrast to the square well where the levels become
further apart as n increases (in the parabolic case they
were equally spaced).
• Also observe that as n increases, the wave functions
add one more half- cycle.
• However at difference with the previous parabolic
case, the wave functions are neither symmetric or
antisymmetric due to the asymmetry of the potential
well.
NANOELECTRONICS

Module 1
 Density Of States (DOS) And Dimensionality

• The number of quantum states available per unit energy

interval per unit volume is called density of states (DOS).
• Density – x/unit volume
• States – Allowable solutions of Schrödinger wave equation.
• In order to calculate the optical properties like photon
absorption and emission in optical devices, transport of
electric charge, specific heat of solid, superconducting
energy gap, tunneling current and other electrical
behaviour, we need density of states.
 Density Of States (DOS) And Dimensionality

• DOS for one, two and three dimensions are denoted by

𝑛1𝐷 , 𝑛2𝐷 and 𝑛3𝐷 respectively.
• If consider the number of quantum states available per unit
energy interval only, then it is denoted by 𝜌1𝐷 , 𝜌2𝐷 and 𝜌3𝐷
respectively.
• In the definition, unit volume refers to length, area and
volume in one, two and three dimension respectively.
 Density Of States (DOS) And Dimensionality
• The consequences of the dimensionality of the system in the
DOS observe here.
• Most physical properties significantly depend on the DOS
function 𝜌.
• The DOS function, at a given value 𝐸 of energy, is defined
such that 𝜌 𝐸 Δ𝐸 is equal to the number of states (i.e.
solutions of Schrödinger equation) in the interval energy Δ𝐸
around 𝐸.
• If the dimensions 𝐿𝑖 𝑖 = 𝑥, 𝑦, 𝑧 are macroscopic and if
proper boundary conditions are chosen, the energy levels can
be treated as a quasi-continuous.
• On the other hand, in the case where any of the dimensions
𝐿𝑖 gets small enough, the DOS function becomes
discontinuous.
DOS function for different low-dimensional solids.
• For the case of a cube of size 𝐿 and volume 𝑉 = 𝐿3 , the
DOS function for 3D is
3
𝑉 2𝑚∗ 2
𝜌3𝐷 𝐸 = 2 𝐸 … … (1.9𝑎)
2𝜋 ℏ2
In terms of ℎ,
4𝜋𝑉
𝜌3𝐷 𝐸 = 3 2𝑚∗ 3/2
𝐸 … … (1.9𝑏)
ℎ
The DOS functions for 2D and 1D semiconductors of area A
and length L, are 𝜌2𝐷 𝐸 and 𝜌1𝐷 𝐸 respectively. The
expressions are
𝐴 𝑚∗
𝜌2𝐷 𝐸 = 2
… … (1.10)
𝜋 ℏ
1
𝐿 2𝑚∗ 2 1
𝜌1𝐷 𝐸 = 2
… … (1.11)
2𝜋 ℏ 𝐸
By considering spin, Eqn. 1.11 becomes,
1
𝐿 2𝑚∗ 2 1
𝜌1𝐷 𝐸 = … … (1.12)
𝜋 ℏ2 𝐸
Some important considerations can already be made:
the DOS function,
in 3D semiconductors is proportional to √E,
in 2D is constant, and
in 1D varies inversely proportional to √E.
This implies in the last case that at the bottom of bands, the
DOS plays a very important role, because there is a singularity
for E = 0.
Fig. Density of states function for a 1D electron system, as a function of energy.
Fig. Density of states function for a 2D electron system, as a function of
energy.
DOS

Fig. Density of states function for a 3D electron system, as a function of

energy.
Derivation
• Consider the motion of a particle in one dimensional
box of width 𝑎. The possible energies are given by
𝑛2 𝜋 2 ℏ2
𝐸= 2
… … (1𝑎)
2𝑚𝑎

= 𝑛2 𝐸1 … … (1𝑏)
ℎ 𝜋 2 ℏ2
where ℏ = and 𝐸1 = is the ground state energy.
2𝜋 2𝑚𝑎2
The positive integer 𝑛 is called quantum number.
There is an energy state for each value of the quantum
number 𝑛.
• Allowed energies are discrete or quantized.
• If the width of the box increases, the separation
between energy states decreases. Hence energy
spectrum grows nearly continuous.
• Considering n and E as continuous variables, from
𝐸 = 𝑛2 𝐸1
we get
𝑑𝐸 = 2𝑛𝑑𝑛𝐸1
• The number of quantum states per unit energy interval,
𝑑𝑛
= 𝜌1𝐷 (𝐸)
𝑑𝐸
1 1
= = Refer
2𝑛𝐸1 2𝐸1 𝐸 𝐸1 Eqn.1b.
1
= … . . (2)
2 𝐸𝐸1
This expression gives the number of quantum states for unit
energy interval for a particle. But one quantum state can be
occupied by two electrons due to two spin state. Therefore,
the number of quantum states is twice the above expression.
That is,
1
𝜌1𝐷 𝐸 = 2 ×
2 𝐸𝐸1
1 𝑎 2𝑚
= = … … (3)
𝜋 2 ℏ2 𝜋ℏ 𝐸
𝐸×
2𝑚𝑎2
The number of quantum states per unit energy interval per
unit volume (in this case unit length) 𝑛1𝐷 is
𝜌1𝐷 𝐸
= 𝑛1𝐷 𝐸
𝑎
1 2𝑚
= … … (4)
𝜋ℏ 𝐸
The mass of electron in a semiconductor is different from the
rest mass. It is called effective mass. It accounts for various forces
experienced by electron (hole) in a semiconductor. Effective
mass of electron is denoted by 𝑚𝑒∗ or 𝑚∗ . Therefore,
1 2𝑚∗
𝑛1𝐷 𝐸 = … … (5)
𝜋ℏ 𝐸
Hence
1
𝑛1𝐷 𝐸 ∝ … … (6)
𝐸
• Thus for a particle confined to one dimensional box, the
density of states vanishes in the limit of large energy (or
large quantum number).
• DOS diverges at 𝐸 = 0.

Fig.
 Particle in a 2D Box
Consider the motion of a particle of mass m in 2D box of side
𝑎 and 𝑏. Potential energy inside the box is zero and 𝑉 = ∞
outside the box.
y

(a,b)
(0,b)

V=0 V=∞
(x,y)
(0,0) (a,0) x
Fig. Particle in a two dimensional box.
We can write time independent Schrodinger wave
equation for two dimension as
𝜕 2 𝜓 𝜕 2 𝜓 2𝑚
2
+ 2 + 2 𝐸 − 𝑉 𝜓 = 0 … … (1.1)
𝜕𝑥 𝜕𝑦 ℏ
Inside the well 𝑉 = 0, then Eqn.1.1 becomes,
𝜕 2 𝜓 𝜕 2 𝜓 2𝑚
2
+ 2 + 2 𝐸𝜓 = 0 … … (1.2)
𝜕𝑥 𝜕𝑦 ℏ
We assume that the wave function 𝜓 can be written as a product
of two functions 𝑋(𝑥) and 𝑌(𝑦) .
𝜓 𝑥,𝑦 = 𝑋(𝑥) 𝑌(𝑦) … … (1.2𝑎)
Therefore,
𝜕2𝜓 𝑑2 𝑋
2
=𝑌 2
𝜕𝑥 𝑑𝑥
and
𝜕2𝜓 𝑑2 𝑌
2
=𝑋 2
𝜕𝑦 𝑑𝑦
Now, Eqn. 1.2 becomes
𝑑2 𝑋 𝑑2 𝑌 2𝑚
𝑌 2 + 𝑋 2 + 2 𝐸𝑋𝑌 = 0
𝑑𝑥 𝑑𝑦 ℏ
Dividing by XY
1 𝑑2 𝑋 1 𝑑2 𝑌 2 = 0 … … (1.3)
+ + 𝑘
𝑋 𝑑𝑥 2 𝑌 𝑑𝑦 2
2𝑚𝐸
where 𝑘2 = 2 … … (1.4)
ℏ
Each term is independent of the remaining terms. This is
possible only if each term is a constant. Let
1 𝑑2 𝑋 2 … … (1.5)
2
= −𝑘 𝑥
𝑋 𝑑𝑥
And
1 𝑑2 𝑌 2
2
= −𝑘 𝑦 … … (1.6)
𝑌 𝑑𝑦
𝑘 2 = 𝑘𝑥2 + 𝑘𝑦2 … … (1.7)
From Eqns. (1.5) and (1.6)
 𝑑2 𝑋 2
2
+ 𝑘 𝑥𝑋 = 0
𝑑𝑥
𝑑2 𝑌 2𝑌 = 0
+ 𝑘 𝑦
𝑑𝑦 2
Solutions are
𝑋 = 𝐶1 𝑆𝑖𝑛𝑘𝑥 𝑥 … … (1.8)
𝑌 = 𝐶2 𝑆𝑖𝑛𝑘𝑦 𝑦 … … (1.9)
Therefore, as per Eqn. (1.2a) and using Eqns. (1.8) and (1.9),
wave function 𝜓 is
𝜓 = 𝐶1 𝐶2 𝑆𝑖𝑛𝑘𝑥 𝑥 𝑆𝑖𝑛𝑘𝑦 𝑦
𝜓 = 𝐶 𝑆𝑖𝑛𝑘𝑥 𝑥 𝑆𝑖𝑛𝑘𝑦 𝑦 … … (1.10)
Applying the boundary condition 𝜓 = 0 at 𝑥 = 0 and 𝑥 = 𝑎,
as well as 𝜓 = 0 at y = 0 and y = 𝑏, to the Eqn.(1.10),
yields
𝑘𝑥 𝑎 = 𝑛1 𝜋
𝑘 𝑦 𝑏 = 𝑛2 𝜋
Or
𝑛1 𝜋
𝑘𝑥 = … … (1.11)
𝑎
𝑛2 𝜋
𝑘𝑦 = … … (1.12)
𝑏
where 𝑛1 , 𝑛2 = 1,2,3, …
From Eqn.1.7,
2𝑚𝐸 𝑛12 𝜋 2 𝑛22 𝜋 2
2
= 2 + 2
ℏ 𝑎 𝑏
Or
ℏ2 𝜋 2 𝑛12 𝑛22
𝐸𝑛1 ,𝑛2 = 2
+ 2 … … (1.13)
2𝑚 𝑎 𝑏
For a square box 𝑎 = 𝑏,
ℏ2 𝜋 2 2 2
𝐸𝑛1 ,𝑛2 = 2
𝑛1 + 𝑛 2 … … (1.14)
2𝑚𝑎
Ground state energy,
ℏ2 𝜋 2 2 2
𝐸1,1 = 1 + 1
2𝑚𝑎2
ℏ2 𝜋 2
= 2
× 2 … … (1.15)
2𝑚𝑎
The first excited state, for 𝑛1 = 1, 𝑛2 = 2
ℏ2 𝜋 2 2 2 … … (1.16)
𝐸1,2 = 1 + 2
2𝑚𝑎2
for 𝑛1 = 2, 𝑛2 = 1
ℏ2 𝜋 2 2 2 … … (1.17)
𝐸2,1 = 2 + 1
2𝑚𝑎2
The above two states have same energy. The number of
quantum states corresponding to the same energy is called
degeneracy. Hence, the above energy level is two fold
degenerate.
Eqn. 1.14, can be written as
𝐸 = 𝐸1 𝑛12 + 𝑛22 … … (1.18)
where
ℏ2 𝜋 2
𝐸1 =
2𝑚𝑎2
For the calculation of DOS in 2D, we can use the following
Fig.
There is an energy state for any pair of quantum numbers (𝑛1
and 𝑛2 ). In other words, every point with an integer
coordinates in the plane with axes 𝑛1 and 𝑛2 represents a
state.
From the Fig., we can write
𝑛2 = 𝑛12 + 𝑛22 … … (1.19)
All points in the washer of radii 𝑛 and 𝑑𝑛 have the same
energy.
 n2
• • • • • • • • • • • • • • • • • •
• • • • • • • • • • • • • • • • • •
• • • • • • • • • • • • • • • • • •
• • • • • • • • • • • • • • • • • •
• • • • • • • • • • • • • • • • • •
• • • • • • • • • • • • • • • • • •
• • • • • • • • • • •
n
• • • •
dn
• • •
• • • • • • • • • • • • • • • • • •
• • • • • • • • • • • • • • • • • • n1
• • • • • • • • • • • • • • • • • •
• • • • • • • • • • • • • • • • • •
• • • • • • • • • • • • • • • • • •
• • • • • • • • • • • • • • • • • •
• • • • • • • • • • • • • • • • • •
• • • • • • • • • • • • • • • • • •
• • • • • • • • • • • • • • • • • •

Fig. For the calculation of DOS in 2D.

Since 𝑛1 and 𝑛2 are positive, we take the area of washer in the
first quadrant.
1
𝐴𝑟𝑒𝑎 = × 2𝜋𝑛𝑑𝑛 … … (1.20)
4
For electrons, the allowed states will be twice the above
expression due to spin. Then,
1
𝐴𝑟𝑒𝑎 = 2 × × 2𝜋𝑛𝑑𝑛 … … (1.21)
4
This area is nothing but the number of states having energy
between 𝐸 and 𝐸 + 𝑑𝐸, that is, 𝜌2𝐷 𝐸 𝑑𝐸.
𝜌2𝐷 𝐸 𝑑𝐸 = 𝜋𝑛𝑑𝑛
𝜋𝑛𝑑𝑛
𝜌2𝐷 𝐸 = … … (1.22)
𝑑𝐸
Using Eqn. (1.18) and (1.19),
𝐸 = 𝑛2 𝐸1
𝑑𝐸
= 2𝑛𝐸1
𝑑𝑛
ℏ2 𝜋 2 𝑛ℏ2 𝜋 2
= 2𝑛 2
=
2𝑚𝑎 𝑚𝑎2
𝑑𝑛 𝑚𝑎2
= 2 2 … … (1.23)
𝑑𝐸 𝑛ℏ 𝜋
Substitute Eqn. (1.23) into (1.22),
𝜋𝑛𝑚𝑎2
𝜌2𝐷 𝐸 =
𝑛ℏ2 𝜋 2
𝑚𝑎2
= 2
… … (1.24)
𝜋ℏ
Now, DOS for 2D is 𝑛2𝐷 (𝐸), which is defined as number of
states per unit energy interval per unit volume, is
𝑚
𝑛2𝐷 𝐸 = 2 … … (1.25𝑎)
𝜋ℏ
If we consider the effective mass of the particle, then
𝑚∗
𝑛2𝐷 𝐸 = 2 … … (1.25𝑏)
𝜋ℏ
So, for 2D, DOS function is independent of energy.
NANOELECTRONICS

Module 1
Two-Dimensional Electronic Systems
• One of the most important recent developments in
semiconductors, both from the point of view of physics and
for the purpose of device developments, has been the
achievement of structures in which the electronic behavior
is essentially two-dimensional (2D).
• This means that, at least for some phases of operation of the
device, the carriers are confined in a potential such that
their motion in one direction is restricted and thus is
quantized, leaving only a two-dimensional momentum or k-
vector which characterizes motion in a plane normal to the
confining potential.
• The major systems where such 2D behavior has been
studied are MOS structures, quantum wells and
superlattices.
• More recently, quantization has been achieved in 1-
dimension (the quantum wires) and “zero”–dimensions (the
quantum dots).
Two-Dimensional Behavior
• Other systems where two-dimensional behavior has been
observed include heterojunctions of III-V compounds such
as GaAs/Ga1−xAlxAs, layer compounds such as GaSe, GaSe2
and related III-VI compounds, graphite and intercalated
graphite, and electrons on the surface of liquid helium.
• The GaAs/Ga1−xAlxAs heterojunctions are important for
device for applications because the lattice constants and the
coefficient of expansion of GaAs and Ga1−xAlxAs are very
similar.
• This lattice matching permits the growth of high mobility
thin films of Ga1−xAlxAs on a GaAs substrate.
 Basic Properties of 2D Semiconductor Nano Structures
• One of the most practical two-dimensional semiconductor
structures consists of a sandwich of gallium arsenide
(GaAs), with a thickness in the nanometre range,
surrounded on each side by a semiconductor such as
aluminium gallium arsenide (Alx Ga1− x As) of higher
bandgap (as shown in the Fig.). Here, x is the mole fraction
of Al.
• The band gap can be empirically approximated as
𝐸𝑔 ≅ 1.426 + 1.247𝑥, eV
for 𝑥 < 0.45.
• For this range of x, Alx Ga1− x As is a direct band gap
material, whereas for 0.45 < 𝑥 < 1.0, it is an indirect
bandgap material.
Fig.4: A quantum well structure.
 Basic Properties of 2D Semiconductor Nano
Structures
• The bandgap of Alx Ga1− x As(x ∼ 0.3) is close to 1.8 eV
while that of GaAs is 1.426 eV.
After forming the structure, the conduction band difference of
Alx Ga1− x As and GaAs is
∆𝐸𝐶 = 𝐸𝐶(𝐴𝑙𝑥 𝐺𝑎1−𝑥 𝐴𝑠) − 𝐸𝐶(𝐺𝑎𝐴𝑠)
= 𝜒(𝐺𝑎𝐴𝑠) − 𝜒 𝐴𝑙𝑥 𝐺𝑎1−𝑥 𝐴𝑠
𝜒 is called electron affinity of the material. 𝜒 = 4.07 for
GaAs, and for 𝐴𝑙𝑥 𝐺𝑎1−𝑥 𝐴𝑠 it is
𝜒 = 4.07 − 1.1𝑥, (𝑓𝑜𝑟 𝑥 < 0.45)
𝜒 = 3.64 − 0.14𝑥, (𝑓𝑜𝑟 𝑥 > 0.45)
For 𝑥 = 0.3,
𝜒 𝐴𝑙𝑥 𝐺𝑎1−𝑥 𝐴𝑠 = 3.74 𝑒𝑉
Therefore, ∆𝐸𝐶 = 4.07 − 3.74
= 0.33 𝑒𝑉
∆𝐸𝑉 = 𝐸𝑔(𝐴𝑙𝑥 𝐺𝑎1−𝑥 𝐴𝑠) − 𝐸𝑔(𝐺𝑎𝐴𝑠) − ∆𝐸𝐶
= 1.8 − 1.426 − 0.33 ≅ 0.04 𝑒𝑉
• As a consequence, the potential energy profile has the shape
of a square well (Figure 4.1(a)), with a barrier height of
0.33eV for electrons and 0.04eV for holes.
Basic Properties of 2D Semiconductor Nano Structures
• In reality, the profile of the potential barrier is
somewhat more complicated, since the potential varies
with atomic distances, which also affects the wave
functions.
• As seen in Figure 4.1(a), carrier motion for both
electrons and holes is not allowed in the direction
perpendicular to the well, usually taken as the z-
direction because of the potential walls.
Basic Properties of 2D Semiconductor Nano Structures
• However in the other two spatial directions (x, y),
parallel to the semiconductor interfaces, the motion is
not restricted, i.e. the electrons behave as free
electrons.
2D Semiconductor DOS
• Let us study the possible energy states in a quantum well. A
simplified model of quantum well is shown in the Fig.6.
• Let 𝑚 be the mass of the particle in the slab.
• The dimension in the z- direction is assumed to be small as
compared to length and breadth, that is,
𝑎≪𝐿
• Quantum sized effect is observed in the z-direction due to
the smallness of size 𝑎.
 Fig.6: Simplified quantum well model.

• Potential energy is defined by

0≤𝑥≤𝐿
𝑉=0 0≤𝑦≤𝐿
0≤𝑧≤𝑎
𝑉 = ∞ 𝑒𝑙𝑠𝑒𝑤ℎ𝑒𝑟𝑒
• The motion of electrons is confined only in the z-direction.
Electron behaves like a free particle in the xy plane.
• The wave function of the electron is the product of two
functions.
𝜓 𝑥, 𝑦, 𝑧 = 𝜓𝐿 (𝑥, 𝑦)𝜓𝑎 (𝑧)
• For z-directional motion (particle in one dimension),
2 𝑛1 𝜋𝑧
𝜓𝑎 (𝑧) = 𝑆𝑖𝑛
𝑎 𝑎
and
2 𝑛2 𝜋𝑥 𝑛3 𝜋𝑦
𝜓𝐿 (𝑥, 𝑦) = 𝑆𝑖𝑛 𝑆𝑖𝑛
𝐿 𝐿 𝐿
• Corresponding energies are
ℎ2 𝑛12
𝐸𝑛1 = 2
(𝑎𝑙𝑜𝑛𝑔 𝑧 𝑑𝑖𝑟𝑒𝑐𝑡𝑖𝑜𝑛: 𝑡ℎ𝑖𝑐𝑘𝑛𝑒𝑠𝑠)
8𝑚𝑎
and for 2D box
ℎ2 𝑛22 𝑛32
𝐸𝑛2 𝑛3 = 2
+ 2
8𝑚 𝐿 𝐿
Therefore the total energy
ℎ2 𝑛12 𝑛22 + 𝑛32
𝐸𝑛1 𝑛2 𝑛3 = 2
+
8𝑚 𝑎 𝐿2
• Let us compare the ground state energies in z and xy
motion. Ground state energy of xy motion
𝐸𝐿 = 𝐸11
ℎ2 12 12 2ℎ2
= 2
+ 2 =
8𝑚 𝐿 𝐿 8𝑚𝐿2
Suppose 𝑎 = 𝐿/20,
𝐸𝑎 = 𝐸1
ℎ2 12 ℎ2
= 2
=
8𝑚𝑎 8𝑚 𝐿/20 2
400ℎ2 2ℎ2
= 2
= 200 ×
8𝑚𝐿 8𝑚𝐿2
= 200𝐸𝐿
• This means that the ground state energy in the z-direction is
200 times that of in the xy motion
• Since the wave function is the product 𝜓𝑎 𝜓𝐿 , the wave
function exists only if the particle has minimum energy
𝐸1 = 200𝐸𝐿 in the z-direction. Otherwise, the wave
function does not exist.
• Let us see the density of states in two dimensional
structure.
i) 0 < 𝐸 < 𝐸1
If the energy E lies between 0 and 𝐸1 , no energy state is
allowed for electron in the slab. The density of state is zero.
ii) 𝐸1 < 𝐸 < 𝐸2
Since the total energy (E) of the particle is greater than 𝐸1 ,
motion is possible in the box.
Electron will be located in the sub-band corresponding to
energy 𝐸1 .
The density of the states is the same as that of 2D case derived
previously, that is,
𝑚𝑒∗
𝑛2𝐷 (𝐸1 ) = 2
𝜋ℏ
which is independent of energy.
iii) 𝐸2 < 𝐸 < 𝐸3
The electron can be located in sub-band corresponding to
energy 𝐸1 (𝑛 = 1) and 𝐸2 (𝑛 = 2). As DOS of 2D is
independent of energy, 𝑛2𝐷 𝐸1 = 𝑛2𝐷 𝐸2 .
Now total DOS in this case is
𝑚𝑒∗ 𝑚𝑒∗
𝑛2𝐷 𝐸1 + 𝑛2𝐷 𝐸2 = 2 + 2
𝜋ℏ 𝜋ℏ
These two sub-bands may overlap for some levels. It is shown
in the Fig.8. Note that E-k diagram is parabolic.
Fig.8: (a) Two sub-bands in the well.
Fig.8: Quantum well (b) energy levels; (c) energy sub-bands
iv) 𝐸3 < 𝐸 < 𝐸4
The electron can occupy 3 sub-bands corresponding to
𝐸1 , 𝐸2 𝑎𝑛𝑑 𝐸3 . Hence, total DOS is
𝑚𝑒∗ 𝑚𝑒∗ 𝑚𝑒∗
2
+ 2+ 2
𝜋ℏ 𝜋ℏ 𝜋ℏ
Fig.9. shows the energy levels in the confined region (thickness)
and corresponding DOS.
The DOS function in the 2D semiconductor structure is similar
to staircase shape. For every jump into possible energy in the
confined direction a sub-band of energy states is realized.
Fig.9: (a) Energy versus wave vector k, showing three sub-bands, (b) Energy
versus two-dimensional DOS. For comparison, the continuous parabolic curve
is the DOS for three-dimensional free electrons.
Fig.10: The same Fig.9b with axis swapping. Density of states function for
a 2D electron system, as a function of energy.
Quantum Wires and Quantum Dots
Quantum wire
• For a quantum wire quantum confinement occurs in two
directions. Particle is free along the length of the wire and
restricted in other two directions.
• It is similar to an optic fibre cable for the transmission of
photons.
• We assume that the electrons confined in two directions,
(x, y), and can freely propagate along the z-direction,
usually perpendicular to the plane defined by x and y. See
Fig.1.
 𝑎𝑥

𝑎𝑦

Fig.1: A quantum wire which confines electrons in 2 dimensions.

For a two dimensional space, Schrodinger’s equation is
ℏ2 𝜕 2 𝜕2
− ∗ 2
+ 2 + 𝑉 𝑥,𝑦 𝜓 𝑥, 𝑦 = 𝐸𝜓 𝑥, 𝑦
2𝑚𝑒 𝜕𝑥 𝜕𝑦
We assume that the potential inside the wire is zero and
infinite for outside the wire.
𝑉 𝑥,𝑦 = 0, 0 < 𝑥 < 𝑎𝑥 𝑎𝑛𝑑 0 < 𝑦 < 𝑎𝑦
= ∞, 𝑒𝑙𝑠𝑒𝑤ℎ𝑒𝑟𝑒
So inside the wire, above Schrodinger’s equation becomes
ℏ2 𝜕 2 𝜕2
− ∗ 2
+ 2 𝜓 𝑥, 𝑦 = 𝐸𝜓 𝑥, 𝑦
2𝑚𝑒 𝜕𝑥 𝜕𝑦
The total energy in a quantum wire is in the form,
𝐸 = 𝐸𝑛𝑥 ,𝑛𝑦 + 𝐸𝑐𝑜𝑛𝑡 (𝑘𝑧 )
Here 𝐸𝑛𝑥 ,𝑛𝑦 represents discrete two dimensional sub-band
energies because 𝑎𝑥 and 𝑎𝑦 are relatively small. But,
𝐸𝑐𝑜𝑛𝑡 (𝑘𝑧 ) represents an approximately continuous energy
profile because the dimension 𝐿 is relatively large.
The energy in the confined region (𝑥, 𝑦 plane) is given by the
result of 2D box (derived earlier session) for rectangular cross
section, as
ℏ2 𝜋 2 𝑛𝑥2 𝑛𝑦2
𝐸𝑛𝑥 , 𝑛𝑦 = ∗ 2 + 2 𝑛𝑥 , 𝑛𝑦 = 1,2,3 … .
2𝑚𝑒 𝑎𝑥 𝑎𝑦
• Energy in the free direction (along the length of the
quantum wire) is given by
ℏ2 𝑘𝑧2
2𝑚𝑒∗
Total energy
ℏ2 𝑘𝑧2
𝐸 = 𝐸𝑛𝑥 , 𝑛𝑦 +
2𝑚𝑒∗
ℏ2 𝜋 2 𝑛𝑥2 𝑛𝑦2 ℏ2 𝑘𝑧2
= ∗ + 2 +
2𝑚𝑒 𝑎𝑥 𝑎𝑦2 2𝑚𝑒∗
DOS for quantum wire
• Basically, a quantum wire can be approximated to a 1D
system.
• We have already derived the DOS function for a 1D
system, in previous sections and is
1 2𝑚𝑒∗
𝑛1𝐷 𝐸 =
𝜋ℏ 𝐸
This is shown in the Fig.2. Note that DOS diverges at
𝐸 = 0.
Fig.2: The DOS for a one dimensional system, as a function of energy
• When we consider the free propagation of wave vector
along the length, z- direction, of the quantum wire, then the
DOS for the quantum wire is
1 2𝑚𝑒∗
𝑛1𝐷 𝐸 =
𝜋ℏ 𝐸 − 𝐸𝑛𝑥 , 𝑛𝑦
𝑛𝑥 , 𝑛𝑦
The plot is shown in the Fig.3.
Features
• Wave functions are given in terms of Bessel functions.
• The energy levels corresponding to transverse direction of
the wire specified by two quantum numbers.
• Fig.3: DOS for a quantum wire (1d).
Quantum Dots
• Quantum dots are zero dimensional structures in which the
carriers are confined in all three dimensions.
• Their energy states are quantized in all directions and the
density of states is represented by series of discrete sharp
peaks.
• A quantum dot may be modeled as an electron in a
spherical well with perfectly reflecting walls.
• Quantum dots are often nanocrystals with all three spatial
dimensions in the nanometre range (Refer Fig.4).
• The nanocrystals can be grown from liquid phase solutions
at well-specified temperatures.
• Conversely, they can also be prepared by lithographic
etching techniques from macroscopic materials.
• A quantum dot may have about 105 atoms.

Fig.4: A quantum dot. It can also be spherical shape.

• de Broglie wavelength is comparable to the size of the dot.
Hence, quantum sized effects are dominant.
• An interesting feature is that we can control the size of the
quantum dot and hence its energy levels can be suitably
tailored.
• Since the energy spectrum of quantum dot is similar to that
of an atom, quantum dot is called an artificial atom.
• Quantum dot can also be modeled as a cube of edge having
nano size.
• In both cases, the ground state energy is the same.
• A quantum dot can be approximately described by an
electron inside an infinitely deep spherical well.
𝑉 𝑟 =0 0<𝑟<𝑎
=∞ 𝑟>𝑎
Possible energy values are
ℏ2 𝑔𝜄,𝑛 2
𝐸𝑛,𝜄 =
2𝑚𝑒∗ 𝑎
where 𝜄 = 0, 1, 2, … 𝑎𝑛𝑑 𝑛 = 1, 2, 3 … . . and
𝑔𝜄,𝑛 represent the zeros of the Bessel functions 𝐽𝜄 𝑥 .
 𝑔0,1 = 𝜋, 𝑔0,2 = 2𝜋, 𝑔0,3 = 3𝜋, …
𝑔1,1 = 4.4934, 𝑔1,2 = 7.7253, 𝑔1,3 = 10.9041, …
𝑔2,1 = 5.7635, 𝑔2,2 = 9.0950, 𝑔2,3 = 12.3229, …

Ground state energy

ℏ2 𝜋 2
𝐸1,0 =
2𝑚𝑒∗ 𝑎
is same as in an infinite potential.
Features
• Energy levels are discrete.
• Energy levels are similar to that of an atom.
• Quantum confinement is observed in three dimensions.
• Density of state is sharply peaked and is given Dirac delta
function.
• The wave functions are expressed in terms of spherical
functions.
• The simplest model of quantum dot is a box of sides 𝑎𝑥 , 𝑎𝑦
and 𝑎𝑧
• We have solved Schrodinger's equation for a particle in a
box. The energy is given by
ℎ2 𝑛𝑥2 𝑛𝑦2 𝑛𝑧2
∗ 2 + 2+ 2
8𝑚𝑒 𝑎𝑥 𝑎𝑦 𝑎𝑧

where 𝑛𝑥 , 𝑛𝑦 , 𝑛𝑧 = 1, 2, 3 … . .
• In a bulk material, 𝑎𝑥 , 𝑎𝑦 and 𝑎𝑧 are large. On the other
hand in a mesoscopic system the particle is confined by all
dimensions. Quantum size effect is observed in all
directions.
• In a quantum well, confinement is observed in a single
direction.
• In spherical dot, with potential energy zero inside and
infinite outside, gives spherical functions as wavefunctions.
DOS
• No free motion is possible in a quantum box, since the
electron is confined in all three spatial dimensions.
• Consequently, there is no k space available which could be
filled up with electrons.
• Each quantum state of a zero dimensional system can
therefore be occupied by only two electrons. The density of
states is therefore described by Dirac delta function
𝑛0𝐷 (𝐸) = 2𝛿 𝐸 − 𝐸0
• For more than one quantum state, the density of states is
given by
𝑛0𝐷 (𝐸) = 2𝛿 𝐸 − 𝐸𝑛
𝑛
• Since in the case of quantum dots the electrons are totally
confined, the energy spectrum is totally discrete and the
DOS function is formed by a set of peaks in theory with no
width and with infinite height (Fig.5).
• Evidently, in practice, the peaks should have a finite width,
as a consequence, for instance, of the interaction of
electrons with lattice phonons and impurities.
Fig.5: Density of states function for a 0D electron system.