Please wait, loading content (0% complete)...
This is only a preview of the March 2016 issue of Silicon Chip. You can view 36 of the 96 pages in the full issue and the advertisments. For full access, purchase the issue for $8.20 or subscribe for access to the latest issues. Items relevant to "Ultrasonic Garage Parking Assistant":
Items relevant to "1Wire Digital Temperature Sensor For The Raspberry Pi":
Items relevant to "Delta Throttle Timer For Cars":
Items relevant to "Solar MPPT Charger & Lighting Controller, Pt.2":
Items relevant to "BatteryPack Cell Balancer For Optimum Charging":
Purchase a printed copy of this issue for $10.00. 
QUANTUM
COMPUTERS
“No, you’re not going to
be able to understand it...
You see, my physics students
don’t understand it either.
That is because I don’t
understand it. Nobody does...
The theory of quantum
electrodynamics describes
Nature as absurd from the
point of view of common
sense. And it agrees fully
with experiment. So I hope
you can accept Nature
as She is – absurd.”
(Nobel laureat Richard Feynman, 1985)
Quantum computers are based on a different type of architecture to
conventional computers and can solve problems using the strange
properties of quantum mechanics such as superposition and
entanglement.
By Dr David Maddison
I
n recent times computers have become enormously
powerful and can be used to solve extremely complicated problems such as in fluid dynamics. But the architecture of conventional computers is unsuited for certain classes of problems.
Solving those problems would take impractically long
periods of time or be altogether impossible.
Examples of problems that quantum computers could
or should be good at solving include simulation of quantum mechanical systems. For example, it may be possible
to accurately simulate interactions that occur in a particle
accelerator such as the large Hadron Collider. Chemical re14 Silicon Chip
actions could also be simulated including extremely complex ones like photosynthesis.
New drugs could also be developed more quickly because
large numbers of configurations of drug molecules could
be evaluated quickly and the ones most likely to work selected for further testing.
They could be used for breaking of certain types of encryption codes (with Shor’s algorithm, for example) or
searching very large unsorted databases (with Grover’s algorithm). A classic example given is searching a phone book
for a certain phone number when the name is not known.
If a book had one million numbers it could be searched in
siliconchip.com.au
one thousand steps with a quantum computer instead of
on average one half million steps as with a conventional
computer and a naive search method!
Weather forecasting and climate models are other possible uses. There are huge numbers of ways that different
parts of a weather or climate system can interact. A quantum computer could analyse all possibilities at once and
find the optimal answer.
Travel routes and schedules could be quickly optimised.
For example, for the classic “travelling salesman problem”
the order of which cities to visit in order to minimise the
distance travelled and not to visit any city twice could be
quickly evaluated. A commercial quantum computer, the
DWave Vesuvius solved such a problem in less than half
a second compared with a conventional computer which
took 30 minutes.
Other possible uses include machine learning, pattern
recognition, image classification and analysis, optimisation problems, quantum communication for guaranteed
privacy and quantum teleportation where information is
transferred from one quantum system to another with no
wires, radio or light transmission.
Note that quantum computers are unlikely to replace the
computers we use now but will coexist with them and be
used only for the types of problem they are best at solving.
Quantum mechanics
Before discussing how a quantum computer works it is
first necessary to discuss some basic principles of quantum mechanics. Quantum mechanics is a branch of physics
that describes the behaviour of the very small objects such
as atoms, subatomic particles and photons and is distinct
from traditional classical mechanics that describes the behaviour of larger scale objects.
In classical mechanics, objects exist in a specific and
definite place and time, something we are all used to. But
in quantum mechanics, particles exist in a “cloud of probability” so that the location of a particle is described by a
probability distribution.
In addition, in quantum mechanics, particles are restricted to certain particular values of properties such as
how much energy they have or a property known as spin
The probability distribution
of an electron in a particular
orbital of an atom. The darker
the “cloud”, the higher the
probability of finding the
electron. It does not have an
orbit analogous to a
planet orbiting the sun
as in the traditional
simplistic view that
many people are
familiar with
although electrons
will have the
highest probability of being
at certain energy
levels. For a further
explanation of energy levels
in atoms see SILICON CHIP,
November 2015, page 17.
(Image credit: RJ Hall)
siliconchip.com.au
How safe is encryption against attack by
a Quantum Computer?
It has been claimed that quantum computers will be able to break
certain types of cryptography by their ability to factor large numbers which are the basis of certain types of encryption schemes.
Schemes claimed to be at risk include those based on symmetric key algorithms (block cyphers) and asymmetric public key
algorithms (RSA, DSA, ECC).
Acknowledging the risk, the US National Security Agency has
already announced it will be moving toward using encryption
schemes which are resistant to attack by quantum computers.
It should be noted that for the foreseeable future, there are
no conceivable realistic quantum computers that are able to
come close to factoring the numbers required to break the above
schemes (when the key length is long enough) so they should be
safe for a long time.
For example, using Grover’s algorithm to factor a large number
would enable the calculation to be done in the square root of the
time taken by a classical computer (say 10 days instead of 100
days) but the security of the encryption could be maintained if the
key length were doubled which is relatively easy to do.
where a particle must be either “up” or “down”, much like
the north or south on a compass.
The values of the properties of particles are regulated like
the clicks on a dial and are said to be quantised. Another
main property of quantum mechanics is that elementary
particles sometimes act like waves and at other times like
particles, socalled “waveparticle” duality.
There is also the uncertainty principle which states that
for a given particle we cannot measure with precision its
properties of both position and its momentum. The more
accurately one value is known, the less accurately the other
is known. In fact, accurately knowing these two properties
together is a meaningless concept in nature.
Any attempt at measurement of one property will alter
the other property of the particle so it is impossible to ever
know both values. Incidentally, this uncertainty also applies to macroscopic objects but is so small as to be of no
consequence, eg, the uncertainty of position of a thrown
cricket ball would be around 1030mm.
Superposition is the condition whereby a particle can be
in a combination of two or more quantum states simultaneously. For example, rather than having a spin of “up” or
“down”, an electron can be (3/5) up or (4/5) down. In this
case, the RMS sum of the coefficients must remain as one.
If up and down corresponded to the binary numbers zero
and one we would say (simplistically) that it partially had
the values of zero and one at the same time. Any attempt to
read or measure the value of the particle, however, causes
its quantum state to collapse or decohere into one of the
values it possesses such as a “one” and superposition is
lost. This phenomenon is known as quantum decoherence.
While superposition is a characteristic of oneparticle
systems, a property that pairs or groups of particles can
have is entanglement. In this case the quantum state of
the pair or group is described collectively as it is shared
and it is not possible to describe the state of an individual
particle independently.
Consider a pair of entangled particles which are known
to have a total spin of zero. These entangled particles are
March 2016 15
SOME SIGNIFICANT DEVELOPMENTS IN THE HISTORY OF QUANTUM COMPUTING
There are far too many developments relevant to quantum computers to list them all here, so only a selection is given.
1975 RP Poplavskii showed the impossibility of simulating quantum systems on classic computers due to superposition.
1976 Roman Stanisław Ingarden published work on quantum information theory.
1980 Yuri Manin proposed the idea of a quantum computer in his work “The computable and the noncomputable” (in Russian).
1981 Richard Feynman said in a talk that it seemed impossible to simulate quantum systems on classical computers and proposed a basic theoretical model
for a quantum computer. For those interested they can read a transcript of this talk at www.cs.berkeley.edu/~christos/classics/Feynman.pdf
1982 Paul Benioff proposed a comprehensive theoretical model for a quantum computer.
1985 David Deutsch described a theoretical model of universal quantum computer that can be used to model other quantum computers and the algorithms
they use.
1991 Artur Ekert invents secure communication based on quantum entanglement.
1993 Dan Simon invents a problem that would be exponentially faster for a quantum computer to solve than a classical one.
1994 Peter Shor, incorporating Dan Simon’s ideas from above, discovers a method to factor large integers quickly. The factoring of large integers is the basis
of many modern cryptography systems such as credit card transactions and this algorithm could theoretically break many such systems. This lead to a
tremendous interest in quantum computation.
1995 Peter Shor and Andrew Steane propose a method for quantum error correction. Also Christopher Monroe and David Wineland experimentally produce
the first quantum logic gate based on a trapped atom.
1996 Lov Grover invents a quantum algorithm to search databases that is much faster than would be achievable on classic computers. David P. DiVincenzo
published a list of the physical requirements for a quantum computer.
1998 First demonstration of a quantum algorithm run on a two qubit quantum computer. First three qubit quantum computer invented. Grover’s algorithm
(1996) run on quantum computer.
2000 First five qubit and then seven qubit quantum computer and also partial execution of Shor’s algorithm (1995).
2001 Full execution of Shor’s algorithm (1995) to factor the number 15.
2002 Quantum Computation Roadmap developed to facilitate the future development of quantum computation. The document is constantly updated.
See http://qist.lanl.gov/qcomp_map.shtml
2003 The US Defense Advanced Research Projects Agency (DARPA) implements a quantum network using optical fibres to transmit information securely
using entangled photons. Any attempt to improperly intercept the data will result in a loss of entanglement of the photons and an inability to read the
data. Also the University of Queensland demonstrate quantum NOT gates.
2005 First quantum byte created, known as a qubyte.
2006 First 12 qubit quantum computer.
2007 Commercial company DWave Systems announce working 28 qubit quantum computer.
2008 Qubits based on graphene quantum dots. DWave Systems announce working 128 qubit quantum computer chip.
2009 Qubits with lifetimes of hundreds of milliseconds. Google and DWave Systems collaborate in the area of using quantum computation for image searches.
2010 Single electron qubit demonstrated.
2011 DWave produces first commercial quantum computer. Error correction in quantum processor developed. Decoherence minimised using high magnetic
fields. Record low error rates are achieved for a quantum computer. An error rate of one in 10,000 logic operations was considered a benchmark target
but a rate of one in 50,000 was achieved. A group at the University of New South Wales and the University of Tokyo achieve a breakthough in quantum
teleportation, successfully transmitting quantum information without error or affecting the superpositions of qubits.
2012 DWave produced quantum computer with 84 qubits. Single atom transistor developed. 1QB Information Technologies founded, the world’s first company
to write quantum computer software. See www.1qbit.com/ Decoherence was kept suppressed for 2 seconds. A group at the University of New South
Wales develop the first qubit based on a single atom of silicon which would enable quantum computers to be built in silicon like conventional computers
with similar fabrication technology.
2013 Three billion qubits were held in a state of superposition for 39 minutes, exceeding the previous record of 2 seconds (2012).
2014 Leaked documents show that the US National Security Agency is interested in quantum computing for cryptography purposes. Quantum teleportation
demonstrated over 3 metres. This is necessary for a quantumbased Internet to make it secure and fast. The largest number ever factored on a quantum
computer was achieved, 56,153 exceeding the previous record of 143. University of New South Wales researchers embedded qubits in silicon to protect
them and give them longer decoherence times.
2015 DWave Systems announce a 1,000 qubit system. University of New South Wales researchers build the world’s first quantum logic gate in silicon.
16 Silicon Chip
siliconchip.com.au
The Titan supercomputer at the Oak Ridge National Laboratory, Tennessee, USA is the most powerful classical computer
in the Western world and the most powerful supercomputer that is freely accessible. It uses 18,688 AMD Opteron 6274
16core CPUs and the same number of Nvidia Tesla K20X GPUs or graphics processing units. It has a benchmark of 17.59
petaFLOPs (where peta is 1015 or 1,000,000,000,000,000 and a FLOP is one floating point operation per second). The
computer runs the Cray Linux Environment and it consumes 8.2MW. There is a more powerful Tianhe2 supercomputer
in China; however it uses USmade CPUs, is not freely accessible and has been criticised for its difficulty of use. Quantum
computers will not replace computers such as these but will supplement them.
in a state of superposition. If a measurement is made (thus
destroying superposition) on one particle and it is found
to have an up spin, for example, the other particle will
automatically acquire a down spin as the total spin of the
pair must be zero (an up spin plus a down spin).
The particle that is not measured changes its quantum
state as if to “know” a measurement has been made on
its partner. This happens no matter by what distance the
particles are separated and would happen even if the particles were at opposite ends of the universe.
Furthermore, the change is instantaneous, not propagated at the speed of light as might be expected. The information travels at an infinite speed, although it cannot be
used for fasterthanlight communication. Einstein called
this phenomenon “spooky action at a distance” and felt
it meant that the description of reality by quantum mechanics was incomplete.
Bits and Qubits
The basic unit of information in a conventional computer is the bit which can have a value of either zero or
one. It is typically physically implemented by the use of
a transistor which is in either an “off” or an “on” state
representing either zero or one or a capacitor which is
either charged or discharged. For 2015 the commercial
CPU with the largest number of transistors, 5.5 billion,
was Intel’s 18core Xeon HaswellEP.
A qubit is the quantum equivalent of a bit which when
read (measured) will result in an answer equivalent to 0
Bloch sphere diagram
representing a qubit.
x, y and z represent
the axes of the
sphere, the north
and south poles
represent the
basis states and
the represents
the superposition
of 0> and 1>.
and represent
angles. Image credit:
Glosser.ca [CC BYSA 3.0]
siliconchip.com.au
or 1. Due to the principle of quantum superposition as explained above, the qubit can have a combination of these
values at the same time whereas a conventional bit must
be either zero or one but not both at any given time.
A qubit can be physically represented by the states of
various quantum particles such as the spin of electrons
(which are either up or down) or other quantumdominated systems (see below).
A qubit is regarded as the superposition of two basis states
which are denoted mathematically as 0> and 1> (spoken
as ket 0 or ket 1) and are equivalent to 0 or 1 in classical
computing. While an ordinary bit in classical computing
can be represented in a diagram by either a simple 0 or 1
a qubit is a bit more complicated and is represented by a
Bloch Sphere as shown.
On the Bloch spere, the “north” and “south” poles represent the basis states of 0> and 1> which physically might
Simulating a Quantum Computer
without yet having one!
There are a lot of problems to solve with quantum computers
but algorithms and computer code still need to be developed to
solve these problems.
Microsoft have developed a software simulation tool called
LIQUi> or LanguageIntegrated Quantum Operations (the symbols at the end a notation used in quantum computing) that
transforms a higher level computer language such as F# that is
coded to represent a quantum operation into one specific to low
level operations in quantum computers.
It allows researchers to write and develop quantum code on
conventional computers in the absence of access to full scale
quantum computers that Microsoft judges to be 1020 years
away, notwithstanding the developments described here. If you
are interested in looking at this it can be downloaded free from
https://github.com/msrquarc/liquid
That version allows for the simulation of up to 23 qubits.
Among specific algorithms that can be simulated and which are
included as examples are: simple quantum teleportation, Shor’s
factoring algorithm, quantum chemistry, computing the ground
state energy of a molecule, quantum error correction, quantum
associative memory and quantum linear algebra.
March 2016 17
represent spin up or spin down states. The superposition
of these states – the qubit  is represented by some point
anywhere on the sphere.
When the state of a qubit is measured there is a loss of
superposition and thus the system can no longer be in
two states simultaneously due to quantum decoherence.
The result is 0> or 1>, equivalent to 0 or 1 in classical
computing.
When multiple qubits exists in a system they can possess
the property of entanglement, mentioned above. This means
that, for example, a pair of entangled qubits will maintain
a relationship with each other so if one is measured (thus
causing quantum decoherence) and found to have a spin
up state, the other will automatically have a down spin.
Entanglement is one method by which multiple qubits
can be made to “work together” and thus solve more complex problems.
Information representation in bits and qubits
Consider the information that can be represented in a
2bit system. Two bits can be represented as either 00, 10,
01 or 11. Two bits can therefore represent only one of four
different values and to use all four values in some given
computation the computer would have to execute at least
four cycles so that each value could be loaded and then
used in a calculation.
On the other hand, a 2qubit quantum computer can
contain and utilise for a calculation all those four values
(above) simultaneously so only one computer cycle is necessary to operate on all four items of data.
In other words, two bits contain information about only
one value and two qubits contain information about four
values. In fact, quantum computers scale the information
that can be contained in the qubits exponentially according to 2n where n is the number of qubits.
A 4qubit computer could, for example, simultaneously
hold sixteen values (24), ie, 0000, 0001, 0010, 0011, 0100,
0101, 0110, 0111, 1000, 1001, 1010, 1011, 1100, 1101, 1110
and 1111. In contrast, a conventional 4bit computer could
store only one of those sixteen values and would have to
repeat an operation 16 times to do the same computation
SQUID as used for the
qubit in the DWave
quantum computer. The
horizontal arrows
represent the possible
directions of current
and the vertical arrows
represent the two
possible spin states, up
or down corresponding
to zero or one.
as the quantum computer could do just once.
A multiprocessor classical computer increases its power
directly in proportion to the number of processors it possesses. The ability for data representation to scale exponentially in a quantum computer compared to a classical
computer and subsequent processing of that data is a key
to its theoretical power, providing that can be implemented in a practical manner. In other words a quantum computer is not simply the same as a classical parallel processing computer.
What can be used as a qubit?
Almost any system that displays quantum mechanical
phenomena can be used as the basis of a qubit as long it is
capable of possessing two different quantum mechanical
states, such as spin up or spin down.
Any real quantum computer might have a combination
of different two state systems just as a classical computer
uses the state of a transistor in a CPU, capacitors in RAM,
the pit or absence thereof in optical media such as a DVD
or the state of a magnetic domain on a hard disk.
Systems proposed include but are not limited to:
electrons (spin up or down), light (amplitude or phase
“squeezed”), Josephson junction and SQUIDs (direction
of current), photon (vertically or horizontally polarised),
atomic nucleus (spin up or down), optical lattice (spin up or
down), quantum dot (spin up or down), graphene quantum
dot (spin up or down), trapped ion (state of ions), nuclear
magnetic resonance of liquid molecules (nuclear spin state)
and diamonds (nuclear spin of atomic vacancies). Note:
What would program code for a Quantum Computer look like?
Anyone who has learned to program has probably started
with a simple program such as the classic one that prints “Hello
world”. What would a very simple program on a quantum computer look like?
No one yet knows how quantum computers and their programming languages will evolve but it might look like the following.
Consider a quantum computer language with just four instructions N (create qubit), E (entangle qubit), M (measure qubit) and
X (execute operation). This program creates an ancilla, a special
bit used for quantum error correction, entangles it with the input
qubit, measures the input qubit and conditionally performs an operation on the ancilla.
After the operation, qubit 2 contains the state of qubit 1 after
a Hadamard transformation has been performed. A Hadarmard
transformation is a one qubit rotation whereby two qubit states
are mapped onto two superposition states with the same computational state as the original qubits (more generally it is a class of
Fourier transforms).
Note that this is very low level programming, equivalent to as18 Silicon Chip
sembly language in a conventional computer and coding would not
normally be done at such a low level – much higher level programming languages would be used.
N 2
# create a new quantum bit and identify it as ‘2’
E 1 2 # entangle qubits ‘1’ and ‘2’, qubit 1 already exists
and is considered input
M 1 0 # measure qubit ‘1’ with an angle of zero
(angle can be anything in [0,2pi]
# qubit ‘1’ is destroyed and the result is either
True or False
# operations beyond this point can be dependent
on the signal of ‘1’
X 2 1 # if the signal of qubit ‘1’ is True, execute the
PauliX operation on qubit ‘2’
Reference: http://cstheory.stackexchange.com/questions/9381/
whatwouldaverysimplequantumprogramlooklike
siliconchip.com.au
This graph shows “Rose’s Law” demonstrating the steady increase in the number of qubits in the DWave quantum
computer which is analogous to Moore’s Law with the number of transistors in a classical computer,
SQUID is a superconducting quantum interference device.
Note that a qubit does not have to be physically small,
although that is desirable so many qubits can be placed
on one chip.
Basic elements of a quantum computer
A practical quantum computer must have certain basic
requirements (DiVincenzo’s criteria) some of which also
differ from a conventional computer as explained below.
1) It must be scalable to enable a reasonable number of qubits just as a conventional computer must have a reasonable number of bits for efficient operation.
2) The qubits must be able to be set to a common initial
state such as all zeros, just as are the bits in a conventional computer.
3) The state of the computer must be controllable using
universal gates such as quantum logic gates. They are
analogous to the logic gates in conventional digital computer circuits (but unlike in a conventional computer
they are reversible).
4) To enable logic operations to be performed by the logic
gates the decoherence times of the qubits must be long
enough for the gate operation to complete. Decoherence
can be suppressed by error correction techniques and
fault tolerant computation. The logic state of a conventional digital circuit will remain indefinitely but qubits
are inherently unstable and will eventually revert to an
alternative state. A stability time of somewhere between
nanoseconds and seconds is required.
5) There has to be a means to read the quantum state of the
processor. In quantum mechanics, the very act of taking
a reading or measurement will alter the state of the system. Conventional digital circuits can be read without
altering the state of the system.
siliconchip.com.au
Quantum decoherence
As mentioned above, quantum decoherence can happen
due to making a measurement or reading but it can also
happen for unwanted reasons and this represents one of
the greatest challenges of quantum computing.
A quantum system can decohere due to thermal vibrations in the atomic lattice (if a crystalbased system is used)
or other subatomic or macro scale phenomena. One partial solution is to cool the quantum processor to extremely
low temperatures in order to reduce thermal vibrations.
INTO MODEL RAILWAYS
IN A BIG WAY?
With lots of points, multiple tracks, reversing loops, multiple
locos/trains, – in other words, your model trains are more a
passion than just a hobby? Then you might be interested in
these specialised model train projects from March 2013
Automatic Points Controller
(Supplied with two infrared sensor boards)
(PCB 09103131/2)........................$13.50
Frog Relay Board (09103133)............$4.50
Capacitor Discharge for TwinCoil Points
Motors (PCB 09203131)..................$9.00
See article previews at www.siliconchip.com.au
ORDER NOW AT
www.siliconchip.com.au/shop
March 2016 19
(Above): closeup of the DWave 1000 qubit quantum
processor.
(Right): DWave processor package mounted on dilution
refrigerator to keep it at a temperature close to absolute
zero.
Temperatures as low as 20mK or
twenty thousandths of a degree above
absolute zero are required. This corresponds to 273.15° Celsius and is much
colder than anywhere in the universe,
which doesn’t get much colder than
about 3° above absolute zero.
Cooling won’t necessarily remove
all instances of decoherence and it is
necessary to use quantum error correction to detect and reduce errors
however this comes at the cost of the
requirement for many more qubits in
the system.
Conventional computers, it should
be noted, also use extensive error correction to ensure they operate correctly
and in very early digital computers it
was necessary to run a program several times to ensure the same result
was obtained each time and if it was,
confidence could be had in the result!
Operation of a Quantum
Computer
To operate a quantum computer the
qubits are first set to an initial state representing the problem and then those
qubits are manipulated using quantum logic gates which are operated in
a sequence according to a quantum
algorithm.
Quantum logic gates are like logic
gates in classical computers (although
their operation is reversible). A quantum algorithm consists of the stepbystep instructions for solving the problem but is specifically designed to utilise features of the quantum computer
such as superposition and entangle20 Silicon Chip
ment. Algorithms from classical computers can also be implemented on a
quantum computer.
Two widely known quantum algorithms are Shor’s algorithm for factoring and Grover’s algorithm for searching unstructured databases.
Once a quantum computer has finished running an algorithm, a measurement of the qubits is made which
collapses the qubits into their basis
states, representing a zero or one to
yield the result.
Some quantum algorithms give the
correct answer only with a certain
probability and may give a different
result each time the algorithm is run!
This is the case with some algorithms
run on the DWave computer discussed next. When these algorithms
are run multiple times the most common result is likely to be the correct
one.
The commercial DWave
Quantum Computer
The only company making quantum
computers on a commercial basis is DWave Systems (www.dwavesys.com),
a Canadian company founded in 1999.
DWave’s computers run a very specialised type of process called quantum annealing which is used for solving problems involving optimisation
where a huge number of options are
reduced to the best choice.
One way to think of these problems
is to think of a metaphor involving a
vast landscape with many hills and
valleys. The object is to find the low
est valley (the best choice) and the way
to do it is either to
1) survey the whole landscape by
walking up and down the hills looking for the lowest valley as a conventional computer would do or
2) use the quantum computer to effectively tunnel through the hills to
quickly find the lowest point.
The basis of the qubit in the DWave
computers is a SQUID or Superconducting QUantum Interference Device.
The device is made of a ring of superconducting niobium and a junction.
Current within the ring can flow in
one direction or the other, resulting in
magnetic spin states which are either
up or down although before measurement the device is in a superposition
of both states, effectively meaning
that the current flows in both directions at once.
The DWave computer quantum
processor must be kept at a temperature close to absolute zero to minimise
quantum decoherence and also to ensure that the SQUID devices can operate in their superconducting state. The
large size of the computer is primarily
due to the cooling equipment.
In the quantum annealing process,
the algorithm used to run calculations
tries to predict what states the qubits
will be in when the temperature of the
SQUIDs is increased, thus finding the
solution or set of solutions for the lowest point in the valley in the landscape
metaphor described above.
As mentioned previously, this computer does not necessarily give the
siliconchip.com.au
same answer to a problem if run a second time however the more answers it
repeatedly gives which are the same,
the greater the confidence one has in
the result.
DWave sees this as an advantage as
it assists in determining the confidence
the computer has in the result of complex computerbased decisions in machine learning applications.
The DWave computer is in use by
Google, NASA, Lockheed Martin and
others.
Google hopes to use the computer for image and news classification,
spoken word recognition, machine
learning and understanding natural
language and is doing research into
other possible uses.
The DWave computer has been criticised because it is not a “universal
quantum computer” meaning that it
cannot run any type of calculation but
is limited to just “combinatorial optimisation problems” and it thus cannot run Shor’s algorithm, for example.
Another criticism relates to whether
it truly is a quantum computer andwhether it uses entangled states.
The reality is that no one fully understands how it works in all aspects,
not even the designers, although it is
now generally agreed that it is indeed
a real quantum computer.
Other issues relate to questions of
how to benchmark the speed of such
a computer and compare it to classical computers.
Making single atom qubits, atomic wires
Australia is a world leader in aspects
of quantum computing.
The Centre of Quantum Computation and Communication Technology
(www.cqc2t.org/) is a collaboration
between The University of NSW, The
University of Melbourne, Australian
National University, Griffith University, The University of Queensland
and The University of Sydney.
It is undertaking work involving a
diverse area of quantum communication and quantum computing. One
(1) To make an image of an atomic structure
the probe of a scanning tunnelling microscope
(STM) is moved along the surface of a silicon
crystal and an image of the surface is obtained
by measuring a current flowing between the
crystal and the tip which varies according to
the position on the crystal surface. An STM
can also be used to manipulate single atoms
on the crystal surface. It is important to map
the crystal surface so the exact location of the
qubit is known.
22 Silicon Chip
(4) Phosphine gas, consisting of phosphorus
(red) and hydrogen, is introduced and the
molecule of gas settles in the place where
the two hydrogen atoms were removed.
(5) The phosphorus atom of the gas molecule
now lies on the surface of the silicon crystal.
Conclusions
The dream of quantum computing
has been around for a while and now
there is one type of specialised quantum computer in commercial production with major research in other areas
of quantum computing, with Australia
being a key player.
Quantum computers will not replace classical computers but will
supplement them by solving specialised types of problems for which they
are suited.
It is also important to distinguish
hyperbole from reality.
Most likely quantum computers will
be introduced slowly, at first solving a
limited number of problems and then,
perhaps, the market will expand as
they solve problems with widespread
demand, such as understanding and
interpreting spoken language, recognising objects or even artificial intelligence.
SC
particular project is the Precision Qubit Program.
This program involves making qubits using single atoms and aims to
“position, control and read out the
electron spin on a single (phosphorus)
atom in silicon which acts as a quantum bit or qubit”. Single electron transistors and microwave strip lines are
used to both read and manipulate the
electron spin on a single phosphorus
atom embedded in a crystal of silicon.
The ability to create a single atom
(2) A layer of hydrogen atoms (light colour)
is laid down on the silicon surface to create
the desired types of surface chemical bonds.
A pulse of current is then applied to the STM
probe which removes one hydrogen atom.
(3) A second pulse of current is then applied to the STM tip to remove a second hydrogen atom.
(6) The hydrogen atoms are removed.
(7) More silicon atoms are added to the surface, embedding the phosphorus atom deep
in the atomic structure where it is not affected
by undesired interference from the crystal surface. (Diagrams captured from https://youtu.
be/0dXNmbiGPS4)
siliconchip.com.au
and the single atom transistor
qubit and support structures such as
nanowires to access the qubit is a remarkable achievement and only possible due to the recent development
of techniques to reproducibly manipulate single atoms and also to know
exactly where those single atoms are
located within the crystal lattice.
The illustrations in the numbered
images in the box show how a single atom of phosphorus is embedded
into a specific location within a silicon crystal.
Actual STM image of a phosphorus
atom (centre) located on the surface
of a silicon crystal at step 6. The
scale bar represents one nanometre,
one millionth of a millimetre.
The ability to accurately place a single
atom at a precise location plus the ability to
make an atomic scale wire allow the fabrication of a single atom transistor. Such a transistor can be used as a qubit or as a component of a classical computer.
While making such a device is a fantastic
start, practical computers need large numbers of devices on the one chip. Also, according to Moore’s Law for classical computing
which says that the number of transistors
on a chip doubles every 12 to 18 months,
the size will need to reach the atomic scale
by 2020 if that rate of advancement is to be
maintained. Obviously beyond the point of
a single atom transistor, no further size reduction is possible.
STM image of single atom transistor.
The single phosphorous atom is at the
centre and the atomic scale wires are
shown in pink.
siliconchip.com.au
Dr Matthew House with
Honours student
Kirsti Date studying
deterministic placement
of single donors in
silicon at the Atomic
Fabrication Facility
at the University
of New South
Wales.
An atomic scale
wire just one atom
tall and four atoms
wide. This is the type
of wire that may be
used to connect to
single atom qubits.
It was made by
using an STM to
create a channel in
the silicon and then
exposing the area
to phosphine gas
to make a line of
phosphorus
atoms and then
depositing silicon
atoms on top of the
phosphorus atoms
(similar to with the
numbered images). The phosphorus
atoms, which were placed at a spacing
of less than one nanometer, doped the
region around their vicinity causing
it to become conductive and act as a
wire.
A similar conductivity and current
carrying capability as copper was
achieved. This particular work also
proved that Ohm’s law operates at
the atomic scale which was not an
expected result as quantum effects
were though to dominate at this size
scale. On the other hand, a concern
that has been raised
from the knowledge
that Ohm’s law still
works at this scale
is that nonquantum affects
may dominate
making a qubit
difficult to
implement.
Another important outcome of
this work relates
to conventional
silicon chip fabrication.
Companies
such as Intel have
become increasingly worried that
the feature size on microprocessors
is becoming so small that quantum
effects will soon start to dominate
and no further miniaturisation can
occur.
Already transistor gate sizes are at
22nm which is about 100 times the
spacing of silicon atoms. This work
suggests that miniaturisation can
continue for some time and down to
much smaller feature sizes.
Image Courtesy of the Centre
for Quantum Computation &
Communication Technology.
March 2016 23
