Quantum simulation
I. M. Georgescu
CEMS, RIKEN, Saitama, 351-0198, Japan

S. Ashhab*
CEMS, RIKEN, Saitama, 351-0198, Japan
and Qatar Environment and Energy Research Institute, Doha, Qatar

Franco Nori†
CEMS, RIKEN, Saitama, 351-0198, Japan, Department of Physics, University of Michigan,
Ann Arbor, Michigan 48109-1040, USA, and Department of Physics, Korea University,
Seoul 136-713, Korea

(published 10 March 2014)
Simulating quantum mechanics is known to be a difficult computational problem, especially when
dealing with large systems. However, this difficulty may be overcome by using some controllable
quantum system to study another less controllable or accessible quantum system, i.e., quantum
simulation. Quantum simulation promises to have applications in the study of many problems in, e.g.,
condensed-matter physics, high-energy physics, atomic physics, quantum chemistry, and cosmology.
Quantum simulation could be implemented using quantum computers, but also with simpler, analog
devices that would require less control, and therefore, would be easier to construct. A number of
quantum systems such as neutral atoms, ions, polar molecules, electrons in semiconductors,
superconducting circuits, nuclear spins, and photons have been proposed as quantum simulators.
This review outlines the main theoretical and experimental aspects of quantum simulation and
emphasizes some of the challenges and promises of this fast-growing field.
DOI: 10.1103/RevModPhys.86.153

PACS numbers: 03.65.−w, 03.67.Ac

5. Spin glasses
6. Superconductivity
7. Metamaterials
8. Topological order
B. High-energy physics
C. Cosmology
D. Atomic physics
E. Quantum chemistry
F. Open quantum systems
G. Quantum chaos
H. Nuclear physics
I. Interferometry
J. Other applications
VIII. Concluding Remarks

I. Introduction
II. The Problem
III. Definitions
IV. Digital and Analog Quantum Simulation
A. Digital quantum simulation
B. Analog quantum simulation
C. Quantum-information-inspired algorithms for the
classical simulation of quantum systems
V. Resource Estimation and Fault Tolerance
A. Resource estimation
B. Decoherence and errors
VI. Physical Realizations
A. Atoms and ions
B. Nuclear and electronic spins
C. Superconducting circuits
D. Photons
E. Other systems
F. Current state of the art
VII. Applications
A. Condensed-matter physics
1. Hubbard model
2. Spin models
3. Quantum phase transitions
4. Disordered and frustrated systems



Thirty years ago, in the early 1980s, it had become clear that
simulating quantum mechanics was a very challenging problem
(Manin, 1980; Feynman, 1982). One obvious difficulty is the large
large physical system. This state is described by a number of
parameters that grows exponentially with the system size (which is
generally defined as the number of particles or degrees of freedom
in the system). In particular, one needs to keep track of the
probability amplitudes for all the possible classical configurations




© 2014 American Physical Society


I. M. Georgescu, S. Ashhab, and Franco Nori: Quantum simulation


Strengths and weaknesses of some of the proposed and demonstrated quantum simulators.

Quantum simulator
Neutral atoms
Trapped ions
Cavity arrays
Electronic spins (quantum dots)
Superconducting circuits
Photons (linear optics)
Nuclear spins (NMR)

Individual control and readout*
Individual control and readout
Individual control and readout,* tunability
Individual control and readout,* tunability
Well-established, readily available technology*

Individual control and readout
Scaling (some recent progress)
Scaling, no individual control


An asterisk means that the feature has been experimentally demonstrated. By scaling we mean controlling an array of at least a few
tens of qubits. By individual control we refer to the ability of controlling and measuring each individual qubit. The weaknesses refer to
the actual experimental implementations.

of the system. Furthermore, simulating the temporal evolution of
the system requires a number of operations that also increases
exponentially with the size of the system. This exponential
explosion is unavoidable, unless approximation methods (e.g.,
Monte Carlo methods) are used. However, depending on the
specifics of the problem under study, good approximations are not
always available or they also face some limitations. Therefore, the
simulation of quantum systemsin general remains a hardtask even
for today’s supercomputers.
A proposed solution to this problem came in the new
type of computer envisaged by Richard Feynman (Feynman,
1982)—the quantum computer. In fact, as has become clear
over the past three decades, a quantum computer promises to
do much more than simulating quantum mechanics, and today
quantum computation and quantum information theory are
very active research fields [see, e.g., Nielsen and Chuang
(2000), Schleich and Walther (2008), and Stolze and Suter
(2008))]. Feynman realized at the time that a quantum
machine would itself experience an exponential explosion,
but with good consequences. The machine would have the
capacity to contain an exponentially large amount of information without using an exponentially large amount of
physical resources, thus making it a natural tool to perform
quantum simulation. Despite the great importance of his
insight, however, he was not very specific about how his
proposed quantum mechanical computer was supposed to
function or how the simulation itself would be realized, as can
be seen from the following quote:
“Let the computer itself be built of quantum mechanical elements which obey quantum mechanical laws.”
(Feynman, 1982).
More than a decade later, it was shown (Lloyd, 1996) that a
quantum computer (i.e., an ensemble of well-defined qubits
that can be initialized, measured, and on which universal
quantum gates can be performed) can indeed act as a universal
quantum simulator. Here the word universal refers to the fact
that, except for changes in the programs that it runs, the same
machine is capable of tackling vastly different problems.
However, a quantum computer (as defined above) is not
necessarily required for implementing quantum simulation.
Simpler quantum devices that mimic the evolution of other
quantum systems in an analog manner could be used for this
task (note that these are not universal simulators, but rather
problem-specific machines). It is therefore expected that
Rev. Mod. Phys., Vol. 86, No. 1, January–March 2014

practical quantum simulation will become a reality well
before full-fledged quantum computers.
In recent years, the interest in quantum simulation has been
growing rapidly, and the reason for this is twofold. First, there
are a large number of potential applications of quantum
simulation in physics, chemistry, and even biology. Second,
the technologies required for the coherent control of quantum
systems have matured enough to allow for the physical
implementation of practical quantum simulation in the very
near future. In fact, some proof-of-principle experiments on
quantum simulation have already been realized [see, e.g.,
Greiner et al. (2002), Leibfried et al. (2002), Friedenauer
et al. (2008), Neeley et al. (2009), Gerritsma et al. (2010),
Kim et al. (2010), and Lanyon et al. (2010)].
Quantum simulation will provide a valuable tool that
researchers from numerous fields will want to add to their
toolbox of research methods. For instance, in condensedmatter physics, quantum simulation would allow the study
of many difficult problems, such as quantum phase transitions, quantum magnetism, or high-T c superconductivity.
Other potential application areas include high-energy
physics, quantum chemistry, cosmology, and nuclear
With the latest advances in the coherent manipulation of
quantum systems (Ladd et al., 2010; Buluta, Ashhab, and Nori,
2011), such as atoms in optical lattices, trapped ions, nuclear
spins, superconducting circuits, or spins in semiconductors,
practical applications of quantum simulation can be expected in
the coming years. Several research groups are now actively
aiming at the experimental realization of quantum simulators
with tens of qubits, which would be the first practical applications in which quantum computers outperform their classical
There is by now a sizable literature on quantum simulation,
especially papers published in the past decade. However,
besides the brief overview of quantum simulators by Buluta
and Nori (2009) and the specialized reviews focused on cold
atoms (Jaksch and Zoller, 2005; Lewenstein et al., 2007;
Bloch, Dalibard, and Nascimbene, 2012), ions (Blatt and
Roos, 2012; Schneider, Porras, and Schaetz, 2012), photons
(Aspuru-Guzik and Walther, 2012), superconducting circuits
(Houck, Türeci, and Koch, 2012), and quantum chemistry
(Kassal et al., 2011; Lu et al., 2012), a global review of the
field is missing. Moreover, a comprehensive, pedagogic
introduction to the subject would benefit researchers just
entering the field, as well as those already working on

I. M. Georgescu, S. Ashhab, and Franco Nori: Quantum simulation


TABLE II. Potential applications of quantum simulators and the physical systems in which they could be implemented, along with relevant
references. We note that this is not an exhaustive list.

Proposed implementation

Condensed-matter physics:
Hubbard models

Spin models

Quantum phase transitions

Spin glasses
Disordered systems

Frustrated systems
High-T c superconductivity
BCS pairing
BCS-BEC crossover
Time-symmetry breaking
Topological order


Atoms (Jaksch et al., 1998; Greiner et al., 2002)*
Ions (Deng, Porras, and Cirac, 2008)
Polar molecules (Ortner et al., 2009)
Quantum dots (Byrnes et al., 2008)
Cavities (Greentree et al., 2006; Hartmann, Brandao, and Plenio, 2006;
Angelakis, Santos, and Bose, 2007)
Atoms (Jané et al., 2003; Garcia-Ripoll, Martin-Delgado, and
Cirac, 2004, Simon et al., 2011; Struck et al., 2011)*
Ions (Jané et al., 2003; Porras and Cirac, 2004b; Deng, Porras, and Cirac, 2005;
Bermudez, Porras, and Martin-Delgado, 2009; Edwards et al., 2010; Lanyon
et al., 2011*; Kim et al., 2011; Britton et al., 2012)*
Cavities (Cho, Angelakis, and Bose, 2008a; Chen et al., 2010)
Nuclear spins on diamond surface (Cai et al., 2013)
Superconducting circuits (Tsokomos, Ashhab, and Nori, 2010)
Electrons on helium (Mostame and Schützhold, 2008)
Atoms (Greiner et al., 2002)*
Polar molecules (Capogrosso-Sansone et al., 2010; Pollet et al., 2010)
Ions (Retzker et al., 2008; Friedenauer et al., 2008)*;
Ivanov et al., 2009; Giorgi, Paganelli, and Galve, 2010
NMR (Peng, Du, and Suter, 2005; Roumpos, Master,
and Yamamoto, 2007; Zhang et al., 2008)
Superconducting circuits (van Oudenaarden and Mooij, 1996)*
DQS (Lidar and Biham, 1997)
Superconducting circuits (Tsomokos, Ashhab, and Nori, 2008)
Atoms (Schulte et al., 2005; Fallani et al., 2007*;
Billy et al., 2008; Roati et al., 2008)*
Ions (Bermudez, Martin-Delgado, and Porras, 2010)
Superconducting circuits (Garcia-Ripoll, Solano,
and Martin-Delgado, 2008)
NMR (Álvarez and Suter, 2010; Banerjee et al., 2013)*
Ions (Porras and Cirac, 2006b; Kim et al., 2010)*
Photons (Ma et al., 2011)*
DQS (Yamaguchi and Yamamoto, 2002)
Quantum dots (Manousakis, 2002)
NMR (Yang et al., 2006)*
Atoms (Regal, Greiner, and Jin, 2004; Zwierlein et al., 2005)*
Superconducting circuits (Rakhmanov et al., 2008)
Superconducting circuits (Koch et al., 2010)
Atoms (Aguado et al., 2008)
Polar molecules (Micheli, Brennen, and Zoller, 2006)
Linear optics (Lu et al., 2009)*
Superconducting circuits (You et al., 2010)

Experimental realizations.

quantum simulation and looking for a quick reference guide.
Since quantum simulation is a subject of interest to a broad
audience, this review attempts to provide a self-contained
description of the current status of theoretical and experimental research on the subject. However, given the breadth
of the topics touched by quantum simulation, not all technical
details can be provided here, and the interested reader is
directed to the relevant references instead.
The remainder of this review is organized as follows.
Sections II–V discuss in some detail the basic theory.
Readers interested only in the physical implementations of
quantum simulation can concentrate on Sec. VI, while those
interested in the applications of quantum simulation can
concentrate on Sec. VII. Tables I–III provide quick reference
guides for the content of Secs. VI and VII. In Sec. VIII
Rev. Mod. Phys., Vol. 86, No. 1, January–March 2014

we discuss the challenges and prospects of quantum

“The rule of simulation that I would like to have is
that the number of computer elements required to
simulate a large physical system is only to be
proportional to the space-time volume of the physical system. I don’t want to have an explosion”
(Feynman, 1982).
We consider a rather general quantum simulation problem,
namely, that of finding the state of a quantum system

1999) Interferometry: Ions (Leibfried et al. As mentioned earlier. 2008. 2004) Ions (Schützhold and Mostame. 2013*.... Library of . 2000)* Ions (Piilo and Maniscalco. 2010) Universe expansion Atomic physics: Cavity QED Cooling Superconducting circuits (You and Nori. Phys. about ∼3. 2009. Hou. 2010)* DQS (Kassal et al.. the Rev. the solution of the Schrödinger equation: iℏ d jϕi ¼ Hjφi dt (1) is given by jφðtÞi ¼ expf−iℏHtgjφð0Þi. Casanova et al..... 2002*.. 2007. the amount of memory required for representing quantum systems grows exponentially with the system size. 2007) * Experimental realizations. We take as an example the standard “threshold” N ¼ 40 frequently cited in the literature (Lloyd. 2011. because for realistic physical problems the Hamiltonian has a very regular structure and just encoding it in the computer memory does not suffer form the exponential-explosion problem. 2005) Atoms (Giovanazzi.. 2008)*. 86. Application Proposed implementation High-energy physics: Lattice gauge theories DQS (Byrnes and Yamamoto. Olson. You and Nori. 2007. 2012) Dirac particles Nucleons Cosmology: Unruh effect Hawking radiation Ions (Alsing. 2010) Superconducting circuits (Nation et al. and Franco Nori: Quantum simulation TABLE III. Dong et al.. 1998a) Superconducting circuits (Quan et al. 2008) Quantum dots (Smirnov et al. No. and Liu. 2013. the U. 2011) Open systems: NMR (Tseng et al. Georgescu. Hu. and Milburn. 2010) Atoms (Goldman et al. 2010) Photons (Semiao and Paternostro. 2003. 2008). we will not worry about the Hamiltonian with its 240 × 240 ≈ 1024 matrix elements.. and Lee. de Raedt et al. Cirac... Ashhab. Mod. 2004)* Superconducting circuits (Grajcar et al. S.. 1. Spring et al. Cirac and Zoller. 2013)* Superconducting circuits (Zhou.. 2005) Ions (Lamata et al. this is not an exhaustive list.. 2011.. 2003. Menicucci. As in Table II. 2010.. 2007) Chemical reactions Quantum chaos: NMR (Weinstein et al..2 × 1013 bits. representing the state of N spin-1=2 particles requires 2N numbers (namely. Feng. Continuation of Table II.S. Gerritsma et al. Wallraff et al. Broome et al. 2010)* NMR (Du et al. 2012. 2006. 2006) Atoms (Büchler et al.. (For the moment... Yang. 2005) Ions (Horstmann et al. January–March 2014 complex probability amplitudes for the different spin configurations). 2009) BEC (Fischer and Schützhold. Calculating the time evolution of this system requires exponentiating a 2N × 2N matrix.. 2011)* Superconducting circuits (Li et al. 2005. 2010) Other applications: Schrödinger equation Quantum thermodynamics DQS (Boghosian and Taylor. and Pachos. Lau and James. Maraner. 2013)* Chemistry: Thermal rate calculations Molecular energies DQS (Lidar and Wang. 2005) Linear optics (Lanyon et al. Rusin and Zawadzki. and so does the number of operations required to simulate the time evolution. 2010*. This implies storing 240 ≈ 1012 numbers for jφi.156 I. For instance. Friedenauer et al. Vol. In order to compute jϕðtÞi numerically. but focused on applications other than condensed-matter physics. 2009.. 1996. and this is without including the particles’ motional degrees of freedom. In order to put this in perspective. 2002)* Linear optics (Howell and Yeaze. 2012) Photons (Aaronson and Arkhipov... 1999) DQS (Aspuru-Guzik et al. described by the wave function jφi at some time t and computing the value of some physical quantity of interest. Barreiro et al.. M.. 2006. and Milburn.) Assuming single precision. 2013. Crespi et al. Tillmann et al. that is 4 TB (terabytes) are required to represent the spin state of 40 particles in a computer memory. Dowling. it is first necessary to discretize the problem such that it can be encoded in the computer memory. Liao et al. Focusing for simplicity on time-independent Hamiltonians (denoted H).

meanfield theories. Going back to the example given in Sec. The system evolves from the initial state jφð0Þi to jφðtÞi via the unitary transformation U ¼ expf−iℏHsys tg. Mod. If the time evolution of the simulator reproduces the time evolution of the simulated system. and Franco Nori: Quantum simulation Congress has almost 160 TB of data. 1 (color online). This sounds very promising.. The importance of measurement must be stressed because the success of quantum simulation ultimately depends on the ability to extract useful information from the simulator. No. S. the unitary evolution U 0 can be engineered. 2005). • analog quantum simulation. which cancels the advantage of the Monte Carlo methods. and between jφðtÞi and jψðtÞi] exists. the numerical evaluation of the integrals encounters the problem of sampling with nonpositive-semidefinite weight functions. Zagoskin (1998). which is the so-called sign problem (Troyer and Wiese. For example. and hence the required simulation time. quantum simulation. III. These methods allow the evaluation of phase space integrals for many-body problems in a time that scales polynomially with the number of particles. This very general definition allows us to include three types of simulation: • digital quantum simulation. 5-quantum-byte) register. and • quantum-information-inspired algorithms for the classical simulation of quantum systems. The dashed arrows indicate the correspondence between the quantum states of the simulator and the simulated system. the desired unitary evolution U 0 ¼ expf−iℏHsim tg with H sim being the controllable Hamiltonian of the simulator can be engineered. between jφð0Þi and jψð0Þi. the quantum simulator is. The result of this measurement provides information about the simulated system..e. the quantum state of N ¼ 40 spin-1=2 particles. Other methods of solving quantum many-body problems such as density functional theory. namely. January–March 2014 157 FIG. IV. Such stochastic methods generally work well when the functions being integrated do not change sign (and ideally vary slowly with respect to the relevant variables). the desired final state can be obtained without the need for numerically exponentiating a 2N × 2N matrix. The quantum simulator is a controllable system: the initial state jψð0Þi can be prepared. Namely. coupled clusters. 2011). [see Thouless (1972). Schematic representation of a quantum system and a corresponding quantum simulator. they are capable of storing large amounts of information in a relatively small amount of physical space. We denote the state of the simulated system by jφi. and Fetter and Walecka (2003)] have similar validity criteria that restrict their applicability to wellbehaved systems.8 × 1025 bits (or 5 × 1012 TB) would be required. which would require a 4 TB classical memory register. Monte Carlo algorithms (Suzuki. with the number of particles. and the final state jψðtÞi can be measured. have been developed as a way of tackling the difficult problem of simulating large quantum systems. The colored arrows denote the controllable operations. Classical stochastic methods. The quantum simulator evolves from the state jψð0Þi to jψðtÞi via U 0 ¼ expf−iℏHsim tg. many-body perturbation theories or Green’s function-based methods. This is roughly 104 times more than the amount of information stored by humankind in 2007. e. which was estimated to be 2. we understand a controllable quantum system used to simulate or emulate other quantum systems [see. The solid black arrows describe the time evolution of the system and the simulator. can be loosely defined as simulating a quantum system by quantum mechanical means. By quantum simulator. As discussed later.I. in particular.g. These are discussed in some detail in the following sections. 86. . DEFINITIONS The alternative simulation method initially proposed by Feynman. The simulator is designed such that there is a mapping between the simulator and the simulated system. even for quantum simulators... the mappings jφð0Þi↔jψð0Þi. and the measurement are realized using only polynomial resources. Double the number of spins. Ashhab. where Hsys is the Hamiltonian of the system. While the simulated system may not be controllable (or not experimentally accessible in some cases). especially fermionic and frustrated systems. 1. the storage capacity of N qubits is exponentially larger than that of N classical bits. If a mapping between the system and the simulator [i. The basic Rev. idea of quantum simulation is represented schematically in Fig. Georgescu.e. 1993). i. these are not easy tasks. This results in an exponential growth of the statistical error. 1. the initial state jψð0Þi can be prepared. but the quantum simulation problem is not really solved unless the initial-state preparation. Buluta and Nori (2009)]. and U↔U 0 . The quantum state jφð0Þi evolves to jφðtÞi via the unitary transformation U ¼ expf−iℏHsys tg. the implementation of the time evolution. DIGITAL AND ANALOG QUANTUM SIMULATION The advantage of quantum simulators over classical devices is that.4 × 1021 bits (Hilbert and Lopez. jφðtÞi↔jψðtÞi. etc. III. and ∼3. Vol. can be represented by a 40-qubit (i. M.. Phys. For some quantum systems. then the system can be simulated. such that sampling the function at a relatively small number of points gives a good approximation to the integral of the function. being quantum systems themselves.e. and the final state jψðtÞi can be measured.

and Nori (2009) (see Fig. and all Hamiltonians that can be efficiently mapped to such systems. The complicated many-qubit unitary transformation U is implemented through the application of a sequence of single. January–March 2014 In general. U ¼ expf−iℏHtg must be applied to the initial state. Mod. Phys. M. A very simple example is the simulation of spin-1=2 particles. 1Þi into j0i⊗ðn−2Þ . although not all mathematically allowed Hamiltonians can be simulated efficiently. Each particle is represented by a qubit: the spin-up state j↑i is encoded as the qubit state j1i.158 I. Ashhab. Wiebe. This reverse problem allows an intuitive. Quantum circuit for the recursive procedure used to find an efficient gate sequence for preparing a given target state jψ T ðn. S.e. Ortiz et al. and Kendon (2010)]. 2002) and Somma et al. No.e. 2005.and two-qubit gates) is generally an approximation of the desired unitary evolution. 2011).. Some of the representative studies on DQS are Lloyd (1996). where one considers the problem of transforming the target state to the initial state j0i⊗n .—The first step of the simulation is to initialize the quantum register to the state jψð0Þi. and the spin-down state j↓i as j0i. was discussed by Ortiz et al. H. j000    0i) with polynomial resources was given by Abrams and Lloyd (1997). Kassal. Such a circuit-based quantum simulation recreating the evolution jψð0Þi → jψðtÞi is usually referred to as digital quantum simulation (DQS). therefore. Nevertheless. DQS is universal (Lloyd.. Cirac. Wang. Applications of DQS also include obtaining certain properties of a given quantum system [e. that obtained from the decomposition into single. there are Hamiltonians that cannot be efficiently simulated in this way. Ashhab. and Nori (2011) proposed a state-preparation algorithm that incorporates quantum simulation: the time evolution of the quantum system is simulated including the interaction with ancilla.. Although DQS algorithms rely on applying a time-ordered sequence of gates. The procedure uses reverse engineering.e. and Sanders (2012). Terhal and DiVincenzo (2000). it can be inverted in order to prepare the target state ~ jψ T ðn. and Franco Nori: Quantum simulation A. The wave function jφi has to be encoded using the computational basis. qubits that can inject or absorb any specified amount of energy from the system. for particular cases of interest efficient state preparation is possible. Initial-state preparation. 1997)]. These steps are discussed in detail in Sec. Zalka (1998a. Initial-state preparation. Lidar and Biham (1997). 1996). A quantum algorithm for preparing a pure state of a molecular system with a given number m of electrons occupying a given number n of spin orbitals that exhibits polynomial scaling in m (regardless of n) was proposed by Wang. and the final measurement. Moreover. (2001). 1Þ transforms jψ T ðn − 2. However. according to Meyer (2002) it should also be possible to use quantum computers to simulate classical physics more efficiently [see also Sinha and Russer (2010) and Yung et al. FIG.g. (2001. Wang et al. a method for generating a state that encodes the antisymmetrized many-particle state of fermions (including all the possible permutations). i. and Latorre (2009). (2) j¼1 where jvaci is the vacuum state and b†j and bj are the fermionic creation and annihilation operators. For example. Ashhab. 1Þi of one electron occupying n possible orbitals. Rev. and Nori.. phase estimation for computing eigenvalues of operators. (2002). Abrams and Lloyd (1997). 1. systematic solution. the three-spin state jφi ¼ j↑↑↓i is represented in the simulator by jψi ¼ j110i.A [see also (Brown.e. The unitary operations H ~ 0 can be calculated easily from the given target state: each and H one of them transforms the known state of the corresponding qubit to j0i. 1999. and Aspuru-Guzik (2009).. 2.g.. 2). (2010)]. (2002. while Kassal et al. and Raeisi. i. are included in this class. this approximation can be made arbitrarily accurate (by refining the decomposition). it is possible to efficiently simulate any finite-dimensional local Hamiltonian. Wiesner (1996). it must be stressed that the implemented unitary operation (i. For example. auxiliary. 2000). thus preparing any desired energy eigenstate. The unitary operation Qðn − 2. or computing partition functions (Lidar and Biham. particularly the Hamiltonian (Abrams and Lloyd. Note that finding an efficient decomposition in terms of universal gates can in itself be a difficult problem (Daskin and Kais.. This is particularly important since all local spin systems. Digital quantum simulation We consider the well-known circuit model for quantum computation (Nielsen and Chuang. A quantum algorithm for the efficient preparation of physically realistic quantum states on a lattice (arbitrary pure or mixed many-particle states with an arbitrary number of particles) was proposed by Ward. as a superposition of binary bit strings. In many cases the preparation of the initial state is difficult and an efficient algorithm may not be available.. unitary evolution U. Since any unitary operation can be written in terms of universal quantum gates. but this comes at the cost of an ever-increasing number of gates.. Georgescu. it must be noted that not any unitary operation can be efficiently simulated (that is with polynomial resources) and. Marzuoli and Rasetti (2002). those that appear in most physical theories can be simulated efficiently. . In principle. Furthermore. Fortunately. DQS consists of three steps: initial-state preparation jψð0Þi. Verstraete. IV. 1Þi from the initial state j0i⊗n . starting from an unsymmetrized state (e. The preparation of N-particle fermionic states of the following form: jψð0Þi ¼ N Y b†j jvaci. 2009. 2003). Munro. 86. (2008) showed that the most commonly used chemical wave functions can be efficiently prepared.and two-qubit gates (we will come back and discuss the decomposition of U into these simple gates shortly). Adapted from Wang. Once the solution of this inverse problem is found. 1998b). 2010). In other words. Somma et al. it follows that in principle “anything” can be simulated. thus implementing a unitary evolution of the simulator. Vol. In order to obtain jψðtÞi ¼ expf−iℏHtgjψð0Þi. Ashhab. i. Aspuru-Guzik et al. DQS is not restricted to recreating the temporal evolution of the simulated system.

Unfortunately. M. the efficient simulation of a general many-body interaction Hamiltonian using two-body interactions is by no means easy (Bennett et al. We now take a look at another example: the algorithm given by Aspuru-Guzik et al. note that ancilla qubits are required. Recent results have reemphasized the shortcomings of using this first-order Trotter formula (Brown. Examples of Hamiltonians of this form include the Hubbard and Ising Hamiltonians. Clark. e. the corresponding Pauli matrix acting on spin (qubit) i. the decomposition of U cannot be obtained efficiently using classical methods.. Whitfield. where it was shown that by using randomness it is possible to efficiently simulate local bounded Hamiltonians with arbitrary time dependence.—We now discuss in more detail how to obtain U. Hl0  ≠ 0 in general. with α ¼ x. Wang and Zanardi (2002).g. (2001). with the accuracy increasing with increasing number of iterations.e. Hastings (2006). e. Wocjan. The circuit contains six CNOT gates and utilizes a fourth. Each subsequent iteration incorporates the previous estimate and increases the accuracy by one bit. Hl0  ¼ 0 for all l and l0 . reduces the uncertainty by a factor of 2. The first iteration gives a rough estimate for the energy. We now consider an example of constructing rather complex operations from simple quantum gates. Ortiz et al. No. and Beth (2002). Bremne.. An important step in this regard is breaking up the evolution time into a large number of small time steps of duration Δt each: U ¼ ðexpf−iℏHΔtgÞt=Δt : (5) There are approximations available for decomposing expf−iℏHΔtg into local gates. and Chuang. Dür. 1. 4. (2011). FIG. which adds to the resource requirements (see Sec. the decomposition of U into a sequence of local gates is straightforward. 2000. Georgescu. Nielsen et al. Wocjan. Bremner. For example.. Bremne. QFT þ is the inverse quantum Fourier transform. . e. Here H denotes the Hadamard gate. 2002b). and Briegel (2008)]. and several methods have been developed [see. This procedure provides an arbitrarily accurate estimate of the energy. (2008). This estimate is then used as a reference point for the next iteration. If ½Hl . Ashhab. 3. The circuit implements the recursive phase estimation algorithm. Recently quantum algorithms for simulating time-dependent Hamiltonian evolutions on a quantum computer have also been investigated (Wiebe et al. (2002). (8) where σ zi is the Pauli matrix acting on spin (qubit) i. the first-order Trotter formula [see. Throughout the paper we denote by σ αi . 2011). Moreover. S. The quantum circuit in Fig. Dür. V. ancilla qubit (bottom line) in order to achieve the desired effective Hamiltonian. and Nielsen (2005). This question has been thoroughly studied. when Δt → 0. C. 2002).. (2005) for the calculation of molecular energies using a recursive phase-estimation algorithm. (2005). Take the Hamiltonian H ¼ σ z1 ⊗ σ z2 ⊗    ⊗ σ zN . Note that an ancilla qubit is used. We assume that the Hamiltonian can be written as a sum of many terms that describe local interactions: M X Hl : (3) H¼ l¼1 U¼ Y expf−iℏHl tg: Quantum circuit for simulating the three-body Hamiltonian H ¼ σ z1 ⊗ σ z2 ⊗ σ z3 . The quantum circuit is shown in Fig. January–March 2014 is composed of six two-qubit (CNOT) gates and one singlequbit gate. Berry et al.g. and Somma et al. and Franco Nori: Quantum simulation 159 Unitary evolution. It Rev. z. and Aspuru-Guzik. Bravyi et al. (2007). Dodd et al. which FIG.. y. 2011). and Brown et al. and Beth (2002a. 2009. From Nielsen and Chuang.. 3 realizes the unitary transformation U ¼ expf−iℏHtg for N ¼ 3 (Nielsen and Chuang. Rötteler. (2011)].B). Clark et al. Vol. i. The topic was further discussed by Poulin et al.. and Briegel (2008). Phys. showing that higherorder decompositions can be more efficient [see. 2006. 2002. Similar quantum circuits can be written for any product of Pauli matrices H ¼⊗Nl¼1 σ αl : (9) Although the example above might look simple. 2000).I. Bacon. (2002)] gives UðΔtÞ ¼ e−iℏ P l H l Δt ¼ Y e−iℏHl Δt þ OððΔtÞ2 Þ: (6) l As a result. Adapted from Aspuru-Guzik et al. Janzing. 2005. Janzing. and V k ¼ ½expð−i2πφk−1 ÞVˆ k−1 2 . R. 4. but it still remains a difficult problem. Nielsen and Chuang (2000). 86. then (4) l In this case.g. Quantum circuit for the calculation of molecular energies in Aspuru-Guzik et al. The first iteration gives the phase φ (which represents the molecular energy) to four bits of accuracy. Biamonte. in most cases of practical interest ½Hl . UðΔtÞ ≈ Y expf−iℏH l Δtg: (7) l The drawback of this approach is that high accuracy comes at the cost of very small Δt and therefore a very large number of quantum gates.. Mod.. when taken as a whole. As a result.

Georgescu. e. 5. up to a certain tolerance level. but this is not always the case. A detailed discussion is given by Ortiz et al. They showed that if both local and nearestneighbor interactions are controllable. 5. The Hamiltonian of the system to be simulated Hsys is directly mapped onto the Hamiltonian of the simulator H sim. The measurement circuit is shown in Fig. Sometimes the mapping is indeed straightforward. Phys.. quantum state tomography (QST) (D’Ariano. sometimes involving additional externally applied fields or ancillary systems to mediate various interactions. The Rev. for spins this is still an open problem. 2007). S. The black dot represents a j1icontrolled gate and the white dot a j0i-controlled gate. Measurement. where U and V are unitary operators. Manousakis. the dynamics or ground state. An important advantage of AQS is that it could be useful even in the presence of errors. 2004b. hence providing the answer to the question of interest. Kraus. then the simulation of interactions in quadratic fermionic and bosonic systems is possible. and Cirac (2007) discussed the class of Hamiltonians that can be simulated when one is restricted to applying translationally invariant Hamiltonians. Mosca. Analog quantum simulation “. Even without having the full quantitative details.. 2009. in which one quantum system mimics (emulates) another (Wei and Xue. which is initially prepared in the ffiffiffi state jþi ¼ ðj0i þ j1iÞ= 2. 2011). Finding the mapping in an AQS might. and the desired spectrum is obtained by analyzing the time dependence of h2σ aþ i. 1999). and (b) for ˆ Both the measurement of the spectrum of a Hermitian operator Q. look simpler than obtaining the most efficient gate decomposition for a given Hamiltonian in DQS.e. algorithms use one ancillapqubit. In general. The measurement circuit is shown in Fig. Wolf. Vol. January–March 2014 Another approach to simulating quantum systems by quantum mechanical means is analog quantum simulation (AQS). (2002). Adapted from Somma et al. pffiffiffi in the state jþi ¼ ðj0i þ j1iÞ= 2 is needed. One pffiffiffi ancilla qubit that is initially in the state jþi ¼ ðj0i þ j1iÞ= 2 is needed. If the quantum simulator suffers from uncertainties in the control parameters. the direct estimation of certain physical quantities such as correlation functions or spectra of operators is more desirable than taking the long route through QST. the literature has generally focused on the discrete evolution of a quantum system. Then jφð0Þi can be mapped to jψð0Þi via an operator f (jψð0Þi ¼ fjφð0Þi). 1982).—After obtaining jψðtÞi ¼ Ujψð0Þi via the unitary evolution. 2004. 2002. Mod. 5. M. and Sacchi.. a qualitative answer can be quite valuable in this context. The first one refers to measurements of quantities that can be written in the form hU † Vi.there is to be an exact simulation. and Gisin. for characterizing a quantum state.. hU † Vi. In AQS one is usually emulating an effective many-body model of the simulated system. continuous evolution has also been discussed (McKague. For example. Paris. We now look at two examples of mappings between quantum systems and the corresponding simulators. QST requires resources that grow exponentially with the size of the system. For example. B. Savel’ev.. No. (2001) and Somma et al. The procedure is repeated until the desired precision is obtained. the phase transition under study could still be observed. one ancilla qubit that is initially Hermitian operator Q. 1997. Fischer and Schützhold. yields a better estimate. 2003) can be used. 86.g. it is usually assumed that there is no restriction in applying one. Porras and Cirac.. but recently. 2007. i. that the computer will do exactly the same as nature” (Feynman. A controllable “toy model” of the system is used to reproduce the property of interest. The choice of the mapping depends on what needs to be simulated and on the capabilities of the simulator. Zagoskin. and Franco Nori: Quantum simulation (a) (b) second example pertains to measuring the spectrum of a ˆ Again. An interesting question is what Hamiltonians can be simulated under certain control constraints. is given by the expectation value h2σ aþ i of the ancilla at the end of the simulation (here 2σ aþ ¼ σ ax þ iσ ay ). The desired quantity.. 1.. FIG. In order to avoid this problem. which can be controlled at least to some extent: Hsys ↔Hsim : (10) This can be done if there is a mapping between the system and the simulator (Somaroo et al.. Ashhab. However.and twoqubit gates and that all qubits of the simulator can be individually addressed and measured.160 I. Biamonte et al. For Hamiltonians Hsim ¼ fH sys f −1 . one is sometimes interested in knowing whether a certain set of physical conditions leads to a given quantum phase transition. The first is the Hamiltonian describing a gas of interacting bosonic atoms in a periodic potential . Note that the simulator may only partly reproduce the dynamics of the system. we need to perform the final measurement in order to extract the desired information. We consider two examples. and quite often clever mappings have to be devised. Quantum circuits (a) for the measurement of the quantity hU † Vi for two unitary operators U and V. 2002. making it inefficient for large quantum systems. and Nori. Smirnov et al. However. Furthermore. at first glance. and jψðtÞi can be mapped back to jφðtÞi via f −1 . So far.

Mod. Porras.. and Ω intensity of the bichromatic light field. S. in the case of an array of Josephson junctions as in van Oudenaarden and Mooij (1996) the system is described by the quantum phase model. The amount of physical resources (i. With the identifications ˆ and mc2 ≡ ℏΩ. This new class of algorithms makes it possible to simulate spin systems for longer times and to study physical phenomena which would have been inaccessible with previous methods. number of qubits. No. Phys. and Cirac. 2 i. Murg. 2011). evolve it. The first steps in this direction were taken by Verstraete and Cirac (2004) and Verstraete. overcoming in this way the sign problem (Temme et al. Moreover. fully characterizing a quantum system requires an exponentially large number of parameters. pˆ is the momentum operator. January–March 2014 161 measuring some physical quantity of the simulator would yield information about its analog in the simulated system. 2009. some techniques from quantum information theory have been used in recently developed algorithms.. number of operations... For example.) needed for the quantum simulation in the case of an N-body problem strongly depends on the type of problem and the particularities of the simulator. AQS has the additional advantage that physical quantities can be measured directly. Quantum-information-inspired algorithms for the classical simulation of quantum systems In an interesting recent development. ¼ HD φ ¼ ðcpσ ∂t (13) where c is the speed of light. and references therein). these methods can be combined with Monte Carlo techniques. 86. 2008. Using matrix product states and projected entangled-pair states one can simulate more efficiently infinite-size quantum lattice systems in one and two dimensions.g. 2011). Gerritsma et al.. It would be useful if many-particle states could be represented in such a way that some physical quantities could be classically calculated in a more efficient way. In this section. V. classical numerical algorithms for the simulation of quantum many-body systems came out of research on quantum information theory. see. HI ¼ 2ηΔΩσ (14) where η is the Lamb-Dicke parameter.. The coefficient J quantifies the hopping strength between lattice sites. and Cirac (2004) and thereafter a significant effort has been made to explore this idea.. it is expected that the preparation of the initial state can occur naturally in processes mimicking the natural relaxation of the simulated system to an equilibrium state. Porras. H I has the same form as HD . 2010. and Vidal (2008)]. Ashhab. The analog simulation of the Bose-Hubbard model using atoms in optical lattices is therefore straightforward. How many particles or qubits are needed to realize useful quantum simulations? The answer to this question depends on . and εi denotes the energy offset of the ith lattice site due to an external confining potential. M. RESOURCE ESTIMATION AND FAULT TOLERANCE A. and σ x and σ z are the Pauli matrices.e. [For detailed studies on the subject. and Cirac (2004). The Dirac equation in (1þ1) dimensions for a spin-1=2 particle with rest mass m is iℏ ∂φ ˆ x þ mc2 σ z Þφ. Δ is the spatial size of ¯ is controlled via the the ground-state wave function. With c ≡ 2ηωΔ this analogy. This is because it is not always easy to prepare the initial state. both the initial-state preparation and measurement process in AQS will need to be studied in more detail as AQS becomes a widely used research tool. e. and U quantifies the interaction strength between atoms occupying the same lattice site. Because the system and simulator are presumed to be very similar.I. The initial-state preparation and measurement in AQS have not been thoroughly discussed in the literature. nˆ i ¼ aˆ †i aˆ i is the atomic number operator counting the number of atoms on the ith lattice site. etc. Its quantum version allows for direct sampling from the eigenstates of the Hamiltonian. we review some results on the estimation of the required resources for some particular cases. in other situations one must rewrite H sim in order to obtain the mapping Hsys ↔Hsim . Nevertheless. Vol. aˆ i aˆ j þ 2 i. effects such as Zitterbewegung and the Klein paradox can be studied in a nonrelativistic quantum system (Gerritsma et al.j i i (11) where aˆ †i and aˆ i correspond to the bosonic creation and annihilation operators of atoms on the ith lattice site. The second example of a mapping between a quantum system and its simulator is the trapped-ion simulation of the Dirac equation (Lamata et al. 2007. In order to achieve this goal. C. Verstraete. Cirac and Verstraete. 2010). Moreover. and μ is the chemical potential. Verstraete and Cirac (2004). and Franco Nori: Quantum simulation H sim ¼ −J X † X 1 X εi nˆ i þ U nˆ i ðnˆ i − 1Þ. II and IV. which can be connected to the Bose-Hubbard model via a mapping where the field operators aˆ i are reformulated in terms of the amplitude and phase of the superconducting order parameter at different points in the circuit. without the need for computational manipulation of measurement results as in DQS. 1. directly Rev. For more details we direct the interested reader to the two reviews (Verstraete. number of steps. Resource estimation Using a quantum simulator instead of a classical computer does not necessarily provide an efficient solution to the problem of simulating quantum systems. Georgescu. (10) will become clearer for each particular case. As discussed in Secs. Another widely used stochastic method is the Metropolis algorithm. In the following sections the Hamiltonians of several proposed quantum simulators and those of the systems to be simulated are discussed in more detail and the relation in Eq. However. The Hamiltonian of a single trapped ion interacting with a bichromatic light field can be written as ¯ x pˆ þ ℏΩσ z . This Hamiltonian has a similar form to the Bose-Hubbard Hamiltonian H BH ¼ −J X X 1 X bˆ †i bˆ j þ U nˆ i ðnˆ i − 1Þ − μ nˆ i . and measure it with polynomial resources. In this sense.j i i (12) where J and U are the same as above.

some simple chemical reactions (Smirnov et al. where the error in the ground state . and Chuang (2006) that several current algorithms for quantum simulation exhibit poor scaling as a function of desired accuracy. (2008).. Dowling. No. as functions of the system size and the numerical precision. They found that the computation time grows exponentially with desired precision. it follows that in order to outperform current classical computers at least 100 qubits and over 200 000 quantum gates per step would be required. 2013) compared the resource requirements for two alternatives that could be used in fault-tolerant DQS: topologically protected surface codes and circuit models with quantum error correction. there are some interesting applications that could be realized with fewer qubits... 2009). or anyons (Lu et al. Lamata et al. and Franco Nori: Quantum simulation FIG. Another example of resource estimation for a DQS implementation is given by Lanyon et al. One should also note here that. Georgescu. it is sometimes said that in order to outperform classical computers quantum simulators require somewhere between 40 and 100 qubits (Buluta and Nori. even if they seem efficient based on the scaling with system size. Compared with the studies on scaling with the system size. R. and the total number of qubits and the total number of steps were estimated as functions of the desired precision.. and Milburn. Nation et al. In order to obtain polynomial scaling. 1999. 2010). and Stancil. Vol.. (Note that these few-qubit simulations can in principle be readily performed on a present-day classical computer. 1. A discrete-variable representation of the wave function in an n qubit basis is used. were investigated by C. The Coulomb potential can be evaluated in OðN 2 m2 Þ steps.. By analyzing the Ising model as a representative example.. 7.162 I. so chemical dynamics could be simulated on a quantum computer in OðN 2 m2 Þ steps. Martin-Delgado. Kim et al. 6. 2009. 2012). the type of simulation one wants to implement. Indeed. 2006b. at first sight. This gives a total of nð3N − 6Þ þ 4m qubits. 2007. and this number can increase rapidly with decreasing step size. 2010. if one considers a case where increasing the accuracy of the answer (i. four of which give a reasonably high accuracy for the Coulomb potential. The estimation of the requirements for simulating N particles interacting through a pairwise potential Rev. For example. the Unruh effect (Alsing. and using parameters that are relevant to presentday experiments. 2007. As a general rule of thumb. 2012). The vertical dashed lines represent the current limit of numericallyexact quantum simulations on classical computers on a grid. Clark et al. 2002). because the gates can be implemented more quickly for small time steps. Weinstein et al. You et al. less attention has been paid in the literature to the scaling of required resources with desired accuracy. Ashhab. Clark et al. 86.e.. Clark. Dirac particles (Bermudez. The quantum circuit was decomposed into fault-tolerant operations. 2011.. S. Resource requirements for the quantum simulation of the dynamics of N particles interacting through a pairwise potential. one could perform frustrated-spin simulations or molecular-energy calculations at the limits of present-day supercomputers.) With a few tens of qubits. quantum chaos (Howell and Yeaze. The results are reproduced in Fig. The results are reproduced in Fig. There have been rather extensive studies on the resource estimation for DQS. Alternatively. 2007). Mod. they concluded that surface codes are superior for quantum simulation. the desired number of bits in the final answer) leads to an exponential increase in the number of quantum gates. M. 2009) could be used. there is typically an overhead that is proportional to the number of gates that need to be implemented. R. it is not obvious that the quantum simulation can be called efficient. a number m of ancilla qubits is required to represent the desired range of potential values with a certain precision. and Solano. new quantum simulation algorithms are needed. However. Furthermore. maintaining a relatively small error level. The precision requirement in a given quantum simulation is therefore an important question for purposes of resource estimation. From Kassal et al. Gerritsma et al. The chemical symbols are a guide to show what type of problem can be simulated with a given computation size.. including the simulation of frustrated spin systems (Porras and Cirac. Recently. it was pointed out by Brown. from the data in Fig. A related recent study (You.. 2008. Geller. 2010). (2009). (2010). with ten or fewer qubits one could perform proof-of-principle simulations. Phys. For instance. the resource requirements (total number of physical qubits and computation time) for computing the ground state energy of the one-dimensional quantum transverse Ising model with N spin-1=2 particles. 6. systems where the phase estimation algorithm can be implemented without the Trotter formula (C... which is exponentially faster than known classical algorithms. 2005. Ma et al. January–March 2014 has been performed by Kassal et al. However. 6. However. one might think that making the step size smaller for the Trotter decomposition in DQS does not affect the total run time of the algorithm.

through calculations and a nuclear magnetic resonance (NMR) experiment. one could characterize how decoherence affects a simulation. 2007). and Aspuru-Guzik (2011). 2013).. 2006).I. Note that the simulation of open quantum systems does not necessarily require the inclusion of the decoherence of the simulator (Schneider and Milburn. that in the quantum simulation of open systems it is possible to exploit the natural decoherence of the simulator by varying the choice of mapping between the simulated system and the simulator. Rev. This idea has in fact been demonstrated recently in experiment (Li et al. FIG. Second. A precision of 10−4 Eh . Phys.. M. S. one can take advantage of the natural symmetries in order to modify the effective decoherence of the simulator. then it is rather straightforward to artificially supplement noise in the simulator so that the combined noise in the simulator faithfully mimics that present in the simulated system. there are certain limitations and the inclusion of the simulator’s decoherence in the simulation must be carefully considered. The interaction between the system and the environment could be qualitatively different from that between the simulator and its environment (Brown. For example. An extended discussion of the resource estimation for such molecular energy simulations is provided by Whitfield. when simulating spin Hamiltonians with degenerate ground states using trapped ions. This statement does not necessarily imply that few particles are sufficient in order to obtain results in AQS. by an appropriate choice of the mapping between the system and simulator. it is sometimes stated that AQS has less stringent resource requirements in order to produce useful results that are intractable for classical computational methods. There has not been much work in the literature on resource estimation for AQS. spontaneous emission of the ions drives the system to states outside the Hilbert space used in the system-simulator mapping (Brown. one needs to pay attention to the way the system and simulator are described.. Decoherence may also provide a useful tool for extracting information about a critical system (spectral structure or critical point of its quantum phase transition) as suggested by Cucchietti. Georgescu. one needs to understand how decoherence will affect the simulation and. Ashhab. and Franco Nori: Quantum simulation 163 B. a few imperfections in an ensemble of particles performing AQS might not affect the overall behavior of the ensemble. such that the AQS might still produce useful results even in the presence of these imperfections. The ground state energies were obtained via direct diagonalization on a classical computer. it has been suggested that the decoherence of the simulator might be useful (Lloyd. (2002) trapped hundreds of thousands of atoms using three laser beams. and Paz (2007). Tseng et al. it is generally believed that the effects of decoherence are less dramatic. First. where only limited precision (or just a qualitative answer) might be required.. whenever possible. It was also suggested (Tseng et al. 2000) that it should be possible to simplify decoherence effects in a simulation within certain subspaces. This idea was investigated in an NMR setting (Zhang et al. A simple argument could go as follows: if the noise level that is naturally present in the simulator is lower than the noise level in the simulated system. Greiner et al. No.. The horizontal lines indicate the bounds for 10−4 Eh precision. 86. Then. 2001. energy as a function of the time step duration Δt is shown. 7 (color online). For example. January–March 2014 Although quantum simulators are affected by the interactions with the environment in the same way as quantum computers. 2010. Unfortunately. Piilo and Maniscalco. Moreover. Decoherence and errors Trotter error analysis and gate count for a simulation of the hydrogen molecule using a DQS algorithm. (a) The calculated ground state energy of the hydrogen molecule as a function of the time step duration Δt. However. 2007). From Lanyon et al. Fernandez-Vidal. It is therefore necessary to pay more attention to the role of errors in AQS than has been done so far in the literature. find clever mappings in order to take advantage of the uncontrollable properties of the simulator.and two-qubit operations and the estimate does not take into consideration error correction for the qubits. 2009) with the simulation of the Ising Hamiltonian. (b) Total number of gates for a single construction of the approximate unitary as a function of Δt. This shows that one should be cautious when trying to include decoherence in the simulation. . The gate count includes both one. where Eh ≈ 27:21 eV. but rather that large numbers of particles could be collectively manipulated in AQS using a small number of controls. Biamonte. As a result. (2000) demonstrated. Vol. This is most clearly seen with AQS. In principle. Mod. More sophisticated methods of dealing with noise are also possible. 1996) as it could serve as a rough way of modeling the decoherence of the simulated system. 1. is achieved for about 522 gates.

2008). Examples of such analog quantum simulators include atoms. The desired evolution of the system would be induced by the simulator’s control fields. 2007). 8) would make an ideal quantum simulator because it can be thought of as the simplified. complexity. In this context.g. Brown. in trapped ions (Porras and Cirac. the reliability. Buluta et al. VII) this is not a necessary requirement. Byrnes et al. 2006. Clark et al. 2009) or in twodimensional crystals (Porras and Cirac. an array of qubits plus their controls (see Fig. 2002. photons.. 2008) on the effect of noise (timing errors in pairwise interactions and noisy pairwise interactions described by master equations of Lindblad form) in two-body interactions and local control operations used for the simulation of many-body interaction Hamiltonians. but this is possible only for small systems. J. with atoms in optical lattices (Greiner and Fölling. 2004b. M. So far there are few studies investigating how the simulator’s imperfections affect the quantum simulation [e.. January–March 2014 Polar molecules Neutral atoms in optical lattices Arrays of quantum dots Quantum simulator Arrays of trapped ions Others Superconducting circuits Arrays of cavities Nuclear spins FIG. Different systems that could implement a specialized quantum simulator for the study of problems in condensed-matter physics. No. 2012). and Plenio. 2002).” with a magnification factor of a few orders of magnitude. for example. Recently. 2007. but focus on a more widely studied design of a quantum simulator: quantum simulators for various models in condensed-matter physics. Santos. This sensitivity is due to the natural ordering introduced on the qubits by the coding of the simulated system.. and a rotating Fermi gas could be used to understand nuclear physics phenomena (Georgescu et al. Disorder. Vol. 2010. Wesenberg. 2008)]. Mod. Brandao. Angelakis. Georgescu. 2008. Such an array could be realized. The quantum simulation results can also be validated against analytical and numerical predictions. They can be thought of as toy models of the magnified lattice structure of a “solid. as well as superconducting circuits.e. the specific limitations of each system should be considered in more detail. 2008). This can . Brandao. Note that in the case of the quantum simulation of experimentally challenging problems (see Sec. noise. and Leibfried.. 2011 and references therein). However. 2007. Stolze and Suter. Any physical system that can be used as a quantum computer would also be a universal machine for DQS. Cross validation over a number of different physical systems could be used. However. a quantum system that is not a potential quantum computer could still implement AQS. and in this way the particular imperfections of each implementation could be ruled out as possible sources of error. 2011). Ladd et al. Ashhab. These systems could be designed such that they form one. Clark. 2006. More recently. This issue is illustrated in the case of a disordered quantum spin chain where strong disorder introduces large errors. PHYSICAL REALIZATIONS The physical implementation of a quantum simulator requires a controllable quantum mechanical system. and efficiency of analog quantum simulations have been considered in more detail than in past studies (Hauke et al. the propagation of sound waves in a two-component Bose-Einstein condensate (BEC) was proposed for the study of cosmic inflation (Fischer and Schützhold. Schmied. Reliability refers to the need to ensure that the results of the simulation faithfully reflect the simulated system. 2008.. Schleich and Walther. 2006b). and Chuang (2006) conducted a detailed study of the algorithm for finding the low-lying spectrum of a pairing Hamiltonian in an NMR implementation. there have been studies (Dür. R. Further problems may arise for each physical implementation. the simulator is more efficient than a classical computer). Buluta. nuclear and electronic spins. For many problems in this class. Bremne...or two-dimensional arrays of qubits that can be manipulated in different manners. For instance. 2012)..164 I. and Bose. and so on. 1. Rev. Each qubit resides in its own potential energy well and is used to encode a spin-1=2 particle. Complexity and/or efficiency refers to the requirement that the quantum simulator is able to solve problems that cannot be solved on a classical computer in polynomial time (i. atoms in arrays of cavities (Greentree et al. Other studies have shown that for simulating the Schrödinger equation the minimization of amplitude errors would be required (Strini. 2006b. 8 (color online). 2008. Hartmann. magnified lattice structure of a “solid” that can be manipulated in a number of different ways in order to test various models. 2009. We will not discuss the physical realization of such highly specialized quantum simulators. Phys. ions either in microtrap arrays (Chiaverini and Lybarger. and Plenio.. S.. and Nori. 86. and Franco Nori: Quantum simulation The ideal situation therefore remains that uncontrollable errors should be minimized as much as possible. this approach is limited as implementations in different systems are not always available. and other imperfections might affect the reliability of the quantum simulation (Hauke et al. Such simulations were found to be sensitive to systematic errors in the applied Hamiltonian and fault-tolerant implementations to be inefficient with respect to precision in the current Trotter approximation methods. The array is configurable in the sense that its dimensionality and geometry can be changed. Possible routes and experimental progress toward building a quantum computer have been thoroughly discussed in the last decade (Chen et al. Montangero (2004) found the two-qubit entanglement in a simulation of a dynamically localized system to be exponentially sensitive both to small changes of the Hamiltonian and to the locations of the chosen qubits. VI. and Briegel... electrons in arrays of quantum dots (Manousakis. Hartmann. Ashhab. 2004). ions. 2007.

. Rev. triangular lattice (Struck et al. Currently. there are both bosonic and fermionic elements that can be used for quantum simulation with atoms in optical lattices. or by controlling the on-site interactions via Feschbach resonances (Lewenstein et al. or phase of the applied lasers. neutral atoms have long decay times of the order of seconds. 2006. Simon et al. and Zoller. Bloch. see Ladd et al. Furthermore.8 μm). Hinteraction is the interaction part. In the case of Rydberg atoms in optical lattices or magnetic traps the lattice spacing is ∼μm or higher.. 9(a)].. which mainly modifies the tunneling strength and to a lesser extent modifies the on-site interaction strength. (2011) achieved the simulation of an antiferromagnetically coupled spin chain in an external magnetic field. 2002). Atoms and ions Neutral atoms in optical lattices are very well suited for mimicking solid-state systems. Zwierlein et al. no such simulations have been performed to date. 2004. Other reviews (Bloch. addressing individual atoms in optical lattices is difficult because the separation between neighboring lattice sites is comparable to the best achievable focusing widths of laser beams (both typically being 0. kagome lattice (Liu et al. In particular. and coupling between internal quantum states. Dalibard. Using laser-assisted tunneling and lattice tilting. Atoms in optical lattices are flexible systems with several controllable parameters: tunneling strength.. and Spielman... Alternative systems that can be used for quantum simulation include Rydberg atoms (Weimer et al. an overall rotation of the trapping potential can be used to simulate a magnetic field (for the orbital degree of freedom).and two-qubit gates (DQS). Li. thanks to their weak interaction with the environment. etc. The control can be applied individually or to the entire array. 2007. we look at different physical systems and describe how the array and controls can be realized experimentally. One possible method for implementing conditional quantum operations on atoms in optical lattices (Jané et al. 2010) and polar molecules (Lewenstein. 9. dipole-dipole and van der Waals interactions offer a means for implementing effective spin-spin interactions. (2010) and Buluta. long-range and multiparticle interactions.. A rather general type of Hubbard Hamiltonian that can be realized in these systems is H ¼ H hop þ H interaction þ H pot þ Hinternal . At a basic level. 86. Furthermore. Dalibard. the optical potential can be tuned in situ rather easily via the intensity. This regime was previously inaccessible and served as one more example demonstrating the power of atoms as quantum simulators. Since the first experiment on the simulation of the quantum phase transition from a superfluid to a Mott insulator using a cold atomic gas in an optical lattice (Greiner et al. In Secs. Interestingly. 10) microwave excitations.. meaning that there is a single energy scale in the problem. Ashhab. 2009) (see Fig. 2011. The optical potentials can be adjusted to allow the change of the geometry and dimensionality of the lattice [e. 2007) (see Sec. More intricate techniques that rely on additional lasers have been devised in the past few years for the simulation of various types of gauge fields. interactions between more distant atoms can be achieved. there has been increasing interest in the study of condensedmatter physics with atoms in optical lattices. 2010). 2010. 2012) describe recent experimental progress. The advantage of using polar molecules is that . but recent progress suggests that there may be methods for overcoming this difficulty (Nelson. In this simulation the occupation of lattice sites was mapped onto the spin states of a quantum magnet. With a sufficiently large relative displacement of the two lattices. frequency.]. Atoms could also be used for DQS. The interaction between neighboring atoms is realized by displacing one of the lattices with respect to the other [Fig. and Nascimbene... Micheli. allowing the observation of spin-orbit coupling in a BEC (Lin. Ashhab. Sherson et al. Phys. 1.. and spinrotation couplings provide a universal toolbox for effective spin models. VII. and Franco Nori: Quantum simulation either directly realize the desired Hamiltonian (AQS) or reconstruct it out of elementary one. January–March 2014 165 Greiner. A theoretical review (Lewenstein et al. Hpot combines all the effects of the nonuniform potentials felt by the atoms. Mod.. A. 2011). 2008. By tuning interatomic interactions using Feshbach resonances. 2011). and Zwerger. dipole-dipole interactions. optical lattices provide the highly desirable properties of being easily tunable and essentially defect free. 2010.1 on the simulation of the Hubbard model). 2005). The continuous tunability of the interaction strength also allowed access to the so-called unitarity regime (O’Hara et al. It should also be possible to utilize the intrinsic spins of atoms in optical lattices for this purpose. M. Indeed. Würtz et al. However. Jiménez-Garcia. (15) where H hop describes the tunneling of atoms from one lattice site to another. and Hinternal describes coherent on-site transitions between the internal levels of the atoms. one for each of the two internal atomic states (which represent the qubit states). where the interaction strength is comparable to the Fermi energy. and Weiss. Another recently emerging direction for quantum simulation using atomic gases is the simulation of artificial gauge fields (Dalibard et al.A–VI.. nonuniform potentials. 2009. Brennen.A. Bakr et al. 2002). S. For a recent review of the state-ofthe-art capabilities of the physical systems that we consider here. 2006). Gibbons et al. Moreover. allowing single-site addressability. Vol. 2011.. providing various examples of quantum systems that could be simulated. 2011).. With polar molecules (Pupillo et al. and Nori (2011). Weitenberg et al.. it is possible to investigate the crossover from a Bardeen-Cooper-Schrieffer (BCS) state of weakly attractive fermions to a BEC of tightly bound fermion pairs (Regal. nearestneighbor. 2009. 2011)...5–0. No.. VI. on-site. Fuhrmanek et al. Two optical lattice potentials are applied to the atomic ensembles. 2003) is schematically illustrated in Fig.I. and Jin. a pair of neighboring lattice sites sharing a single particle is mapped onto a (static) spin-1=2 particle in the quantum magnet. A quantum simulation of the Mott insulator-superfluid phase transition in the Hubbard model can be realized by adjusting the depth of the optical lattice.F.g. 2007) discusses in detail atoms in optical lattices as potential quantum simulators. Georgescu.

166 I. 2011). 2010) and of relativistic single-particle motion (Gerritsma et al. R. Ashhab. Vol. 2009)]. 2004. 1. a variety of effective Hamiltonians for AQS or quantum gates for DQS can be realized.. Shaded circles denote the state j1i and white ones denote the state j0i. 2009) or arrays of microtraps (Chiaverini and Lybarger. Rev. 2006b. interact rather strongly via Coulomb repulsion. The quantum states of trapped ions are typically manipulated by either resonantly driving transitions between different internal states of the ions or resonantly driving sideband transitions involving the internal states and the vibrational states of the ions in the external trapping potential. 2003. and Franco Nori: Quantum simulation Polar molecules trapped in a plane by an optical potential created by two counterpropagating laser beams with wave vectors indicated by the middle arrows. It is possible. S. Georgescu. extended Hubbard models (Ortner et al. as well as those corresponding to blue-sideband driving and to resonant driving of the ionic internal states. Ions can be trapped by electric (or magnetic) fields. Manipulation of atoms and ions as proposed by Jané et al. Schmied. Shaded circles denote the state j1i and white ones denote the state j0i. ions. Furthermore. This facilitates the implementation of two-qubit gates and the control of the qubit positions and motion. M. The dipoles are aligned perpendicular to the plane by a dc electric field.. and φ is the laser phase. 2008. Adapted from Jané et al. 2007. Blatt and Wineland. and the supersolid phase in a triangular lattice (Pollet et al. Lanyon et al. FIG. No. 2010) could also be simulated with these systems. which interact weakly with each other.. to obtain different arrangements of many ions in anharmonic one-dimensional traps [long ion strings (Lin et al. however. 2008. In contrast with neutral atoms. Johanning. their large electric dipole moments produce strong dipoledipole interactions that can be manipulated relatively easily via external dc and ac microwave fields. 2010).. 2011. J. The high-fidelity one-... 1998).. indicated by the left arrow. and sequences of high-fidelity quantum gates have been demonstrated in experiment (Hanneke et al. 2009.. (a) Neutral atoms in a double optical potential: the interaction between neighboring atoms is realized by displacing one of the lattices with respect to the other. two-dimensional traps [planar crystals (Porras and Cirac. In fact. Clark et al. 2009. FIG. . σ þ and σ − are the two-level atom transition operators. being charged. The right arrow represents the ac microwave field. Lanyon et al.. Ion qubits also have long coherence times of the order of seconds... Phys. 2011) have also been recently demonstrated. Buluta et al. laser cooled and manipulated with high precision for realizing quantum simulation (Schätz et al. a† and a are the creation and annihilation operators of the vibrational mode. Using this Hamiltonian. (2003). 2009). Varon. and even three-qubit (Toffoli) gates implemented with trapped ions have resulted in the most advanced implementations of DQS to date (Barreiro et al. 9 (color online). 2008. η is the Lamb-Dicke parameter (which is assumed to be small here). the Hamiltonian describing the coupling between the internal and vibrational modes due to the laser driving at the red-sideband frequency can be written in the following form: H ¼ iℏηΩ½expðiφÞσ þ a − expð−iφÞσ − a† .. 2009.. (b) Ions in independent trapping potentials: the interaction between two neighboring ions is achieved by conditionally displacing the corresponding ions with a state-dependent force. 86. Analog quantum simulations of frustrated spin systems (Kim et al.. This control can be used to study strongly correlated systems.. Mod.. 2011). quantum phase transitions (Capogrosso-Sansone et al. and Wunderlich. From Pupillo et al.. 10 (color online). one of the earliest theoretical studies on the physical implementation of quantum simulation dealt with trapped ions (Wineland et al. For example. Both the internal energy levels and the vibrational modes of the trapped ions can be exploited for encoding quantum information. Ions have generally been trapped using linear harmonic traps in quantum-simulation experiments to date. 2009). Wesenberg. January–March 2014 (16) where Ω is the Rabi frequency of the transition between the internal states. Biercuk et al. two-. and Leibfried..

while others considered the possibility of using atom-cavity systems for simulating the high-spin Heisenberg model (Cho.I. (2008). 2012). and these atoms can be driven by external lasers. 2009. Hartmann. This system provides an alternative way of simulating the Bose-Hubbard model and quantum phase transitions as well as spin models (Kay and Angelakis. Hartmann. the cavity array is described by the following Hamiltonian: H ¼ ωc X R a†R aR  X 1 þ 2ωc α ða†R aR0 þ H:c:Þ. Wesenberg. the general form of the NMR Hamiltonian is H ¼ −ℏγB X X I zi þ J ij I zi I zj . Ashhab.. 2008).g. and Bose. One now introduces an atom or ensemble of atoms in each cavity... 2009). 2007. ωc is the frequency of the relevant cavity mode. Arrangement of electrodes for the realization of a bilayer honeycomb lattice. 2012). January–March 2014 in the cavities. With optimized electrode structures. which can be utilized for simulating. Atoms in cavity arrays (see Fig. the manipulation of ions in a one-dimensional array of microtraps is depicted schematically. (2011) and Timoney et al. 11 (color online). The atoms trapped in a cavity together with the photons in the same cavity form hybrid excitations called polaritons. 2011. Nuclear and electronic spins Nuclear spins manipulated by means of NMR have been among the first experimental systems to implement small quantum algorithms and quantum simulation (Peng et al. in arrays of microtraps).. and Leibfried. 12) (Greentree et al.R0 Rev. 2012)] or threedimensional traps. Brandao. It would also be possible to combine trapped ions with optical lattices as suggested by Schmied et al. Georgescu. The microtrapping regions are shown as ellipsoids.. and Bose. 2009. Arrays of cavities could also be used to study the quantum analog of Fabry-Pérot interferometers as suggested by Zhou. 2006. and Plenio. the Mott-insulator-superfluid phase transition. but in most of the experiments done to date it has been realized with ions in the same potential. Brandao. Arrays of cavities in an arbitrary geometry may be realized with photonic band gap cavities and toroidal or spherical microcavities coupled via tapered optical fibers (Greentree et al. Peng and Suter. However. 9(b). The nonlinearity introduced by the atoms results in effective polariton-polariton interactions. 86. 2009. and Plenio. Nuclear spin qubits have long coherence times (> 1 s). In the presence of a strong magnetic field pointing along the z axis. Dong et al. M. i i>j (18) . Wesenberg. and Bose. Photons can hop between the different cavities because of the overlap between the light modes in adjacent cavities. 2009.. One such possibility is to use a state-dependent displacement. In the absence of atoms. while the locations of unwanted spurious microtraps are indicated with small spheres. From Schmied. Angelakis. Brandao. 2008b). 2008). and α is the intercavity coupling coefficient (Hartmann. for example. The polariton-polariton interaction strength can. Mod. Brandao. various microtrap arrays where ions are arranged in different lattice configurations (see Fig. Li et al. In Fig. 2010. these might be quite challenging to realize experimentally. 1. A one-dimensional array of cavities with atoms trapped FIG. Angelakis. and Franco Nori: Quantum simulation 167 FIG. where a†R creates a photon in the cavity at site R. No. (17) þ 2 R. Santos.. The atoms can be driven using externally applied lasers. Phys. 2008a) and the fractional quantum Hall effect (Cho. Zhang et al. 2006. which can be implemented by applying optical dipole forces (Blatt and Wineland. This method is particularly useful for ions trapped in different harmonic potentials (e. In fact a twodimensional ion array was recently used to implement a quantum simulation of a spin system with hundreds of ions (Britton et al. 2010. 2006. There are also alternative ways to generate two-qubit interactions between ions. Angelakis. 2006). Vol. for example.. S. The atoms and cavities form hybrid excitations (polaritons) that can hop between the different cavities and that effectively interact with each other. A recent proposal suggested measurement and feedback control as tools for realizing quantum simulation in atomcavity systems (Vollbrecht and Cirac. 12. Schmied et al. Two-qubit interactions are usually realized with optical forces. B. 2006). and high-fidelity quantum gates and the coherent control of up to 12 qubits have been demonstrated. 2007) could also be used as quantum simulators [see also Kimble (2008)]. 2011). and Leibfried. (2008). and Plenio. From Hartmann.. Lau and James. (2011). but a method for laserless simulation (avoiding the problem of scattering) with ions in arrays of microtraps has been proposed by Chiaverini and Lybarger (2008) and recently demonstrated by Ospelkaus et al. 11) can be constructed (Schmied. be tuned via the atom-cavity detuning. and Plenio.

The field of NMR benefits from very welldeveloped control techniques.. Superconducting circuits Superconducting circuits (You and Nori. C. An array of quantum dots realized by a metallic gate with an array of two different size holes placed on top of an Alx Ga1−x As=GaAs heterostructure.. (19) 1≤j<k≤n where the first term is the energy due to an applied magnetic field Bj .. 2002. Alternatively. This was one of the early theoretical proposals for AQS. various lattice geometries could be created. various one-. No. Mod. However. 2008) can also be used as quantum simulators (see Fig. B is the magnetic field. quantum dots allow flexible control over the confinement potential and they can also be excited optically. Byrnes et al. 2011. individual addressing and measurement would still be impractical. 1. If the region is roughly the same size as the wavelength of the charge carrier.. the predictions of the two-dimensional Mott-Hubbard model were tested for electrons in an artificial honeycomb lattice (Singha et al.and possibly multiqubit gates can be implemented. Quantum information can be encoded in different ways: in the number of superconducting electrons on a small island. 2008). Furthermore. The Hamiltonian for an array of quantum dots is given by H¼ n X μB gj ðtÞBj ðtÞ · Sj þ j¼1 X J jk ðtÞSj · Sk . The gaps represent tunnel junctions. In a recent experiment. 14). a recent proposal suggested that nuclear spins attached to a diamond surface and addressed through nitrogen-vacancy (NV) centers could offer an attractive route toward a large-scale quantum simulator for strongly correlated systems (Cai et al... nuclear spins provide a very good test bed for various small simulation problems and allow the implementation of both DQS and AQS. and the second term is the exchange interaction energy. 2008). Here Sj is the spin of the electric charge quanta of the jth dot (Chen et al. 2013). A negative gate voltage is applied to this gate in order to create a potential for the twodimensional electron gas that is otherwise free to move at the Alx Ga1−x As=GaAs interface. S. Nevertheless.. Quantum dots can be defined at fabrication or by applying bias voltages using electrodes placed above a two-dimensional electron gas. or in oscillatory states of the circuit. which is a result of virtual tunneling between the quantum dots.. By adjusting the mesh-gate design and voltage. but this latter approach has the disadvantage of heating and complicated engineering. Electron spins in quantum dots may provide an advantage over atoms in optical lattices due to the very low temperatures (relative to the Fermi temperature) that can be reached and the natural long-range Coulomb interaction. Using rf pulses.” These features make electron spins in quantum dots particularly attractive for quantum simulation. 2007). From Manousakis. 14 (color online). 2008) (see Fig. The interactions between the qubits can be engineered by adjusting the gate voltages together with a careful choice of the mesh hole sizes and doping (Manousakis. Clarke and Wilhelm. Vol. January–March 2014 Schematic diagram of an array of superconducting flux qubits (Ashhab et al. and Jij are the spin-spin coupling coefficients (Chen et al. one of the main reasons being the spectral crowding that occurs as the number of energy levels increases exponentially with increasing number of spins. it is not very flexible and its main disadvantage is the lack of scalability. Moreover. and Franco Nori: Quantum simulation where γ is the gyromagnetic ratio. while the φ-shaped circuits are used to read out the states of the qubits. The T-shaped circuits are used to couple the flux qubits. Georgescu. two. FIG. Ashhab. Quantum dots are semiconductor systems in which the excitations are confined in a very small region in one or two dimensions. Furthermore. Rev. 86. 2005. They can be designed to have certain characteristics and assembled in large arrays. 13). 2002). 2002. 2011). 2008). allowing the addressing of the individual transitions based on their frequencies. Although in solid-state NMR the scalability drawback may be overcome to some extent. The different transitions between pairs of energy levels generally have distinct resonance frequencies. The circuit can be manipulated by applied voltages and currents (including both dc and microwave-frequency ac signals) and measured with high FIG. 2007). State-of-the-art quantum-dot qubits now have long decay times of > 1 s (Amasha et al. M. 13. (2007) proposed that interfering acoustic waves could be used to form an analog of optical lattices in a two-dimensional electron gas. Another system that could be used for quantum simulation is electron spins in semiconductor quantum dots (Hanson and Awschalom. quantum dots with large tunnel coupling can act as “artificial molecules. in the direction of a current around a loop. The manipulation and readout can be done both electrically and optically. Arrays of quantum dots could be realized using twodimensional mesh gates (Manousakis. . the energy levels are quantized and the quantum dot becomes very similar to a real atom (and can therefore be referred to as an “artificial atom”).168 I. I is the angular momentum operator. Byrnes et al. Phys.

2009). Indeed. 2013).e. 2012). Tillmann et al. superconducting circuits can be coupled to “cavities. Furthermore highfidelity one-. 15). DQS could be implemented in a superconducting circuit in the near future. From Koch et al. 2010). S. 2009).. 2010.. and Koch. and corresponding experiments with up to four photons were performed (Broome et al.g.... networks for simulating Anderson and Kondo models (Garcia-Ripoll. Although unscalable. Although the difficulty in implementing two-qubit gates is a serious drawback for photonic systems in the context of quantum computation. (20) ði. 2005. this approach may still enable a universal quantum simulation speedup relative to a present-day classical computer. Ashhab. 2013. It has also been proposed that artificial gauge fields could be simulated with such circuits (Koch et al. 2011) and relativistic field theories (Angelakis et al. Solano. with limited flexibility and scalability. With this level of control. two-. 2008). Although macroscopic in size.. A related study proposed an approach to universal quantum computation and simulation using the single-excitation subspace of an array of coupled superconducting qubits (Geller et al. and fractals (Tsokomos. Harris. Angelakis and Noh. This setup is very useful for the study of electric-circuit analogs of cavity quantum electrodynamics (circuit QED) (You and Nori.. and Nori. Schoelkopf and Girvin.. there have been some notable achievements for quantum simulation using photons. 2012). Crespi et al. performed with artificial atoms and cavities (see Fig.. The Hamiltonian for N charge (flux) qubits biased at their symmetry points (which is optimal for quantum coherence) coupled capacitively (inductively) is H¼− N X Δi i¼1 2 σ zi − X J ij σ xi σ xj . although quantum coherence was not tested on these circuits in the same way that coherence is commonly tested in other experiments using small numbers of qubits. highly connected networks (Tsomokos.. Ashhab. for the most part. it was shown that the propagation of photons in a network of beam splitters is in general a computationally difficult task for classical computers even for a few tens of photons (Aaronson and Arkhipov. January–March 2014 169 A possible design for implementing the Jaynes-Cummings lattice using superconducting qubits. 2008). The fact that superconducting circuits can be produced in large numbers and “wired” together on a chip offers a rather straightforward way of realizing various lattice geometries. interaction strengths. Furthermore. electron-density oscillations). and so on. a recent experiment demonstrated AQS of a spin larger than 1=2 using this approach (Neeley et al.. circuits containing 512 qubits have been fabricated (Harris et al. Furthermore. The colored strips are superconducting resonators and the dots are superconducting qubits. decoherence rates below 10 kHz). Examples include the Kitaev model on a honeycomb lattice (You et al. Rev. 2010. D. No. As for scalability. it could realize the proposal of atom-cavity arrays (or Jaynes-Cummings lattices). Spring et al. It should be noted. E. Georgescu. M. 2010). Ashhab. 2008).. An effective particleparticle interaction is created by the qubit-cavity interactions. and the coupling energy between two qubits can be turned on and off at will.. and to simulate frustrated spin systems (Ma et al. that superconducting circuits have more than two energy levels. these circuits can display quantum behavior and can be seen as artificial atoms. and Martin-Delgado. and one-qubit gates can be easily realized with linear optical components. 2010). 2013.. Phys. and Franco Nori: Quantum simulation accuracy using integrated on-chip instruments. 15 (color online). to calculate the energy spectrum of the hydrogen molecule to 20 bits of precision (Lanyon et al. 2011). It has also been proposed that photons propagating or trapped inside materials doped with atoms that have suitable energy-level structures could be used for the simulation of Luttinger liquids (Angelakis et al. hardly being affected by noise or decoherence. Vol. Photons have been used to calculate the possible fractional statistics of anyons using a six-photon graph state (Lu et al. e. Recently.jÞ where Δi is the level splitting and J ij is the strength of the coupling between qubits i and j. Mod. However... however. in the polarization of the photon. more than 200 superconducting resonators were recently fabricated on a single chip (Houck. 2011. The advantage over real atoms is that superconducting circuits can be designed and constructed to tailor their characteristic frequencies. State-of-the-art superconducting qubits have coherence times exceeding 100 μs (i. Individual control and measurement have been demonstrated (Mariantoni et al. 2010). They naturally possess the ability to encode qubit states.I. 2011). 2013. 86. Photons Photons can carry quantum information over long distances. and Nori.” which are actually electrical resonators (and the “photons” are. 1. If qubits are integrated into such a circuit. FIG. 2013). which is quite high considering that other energy scales in the circuit are typically in the range 10 MHz– 10 GHz. The frequencies can also be tuned in situ by adjusting an external parameter (typically an external magnetic field). it remains to be seen how far photon-based quantum simulation can go. Other systems One of the systems that are being considered as candidate platforms for implementing quantum computation is NV ... 2012). and three-qubit quantum gates have been demonstrated. and these additional levels could also be utilized.. Türeci. 2011). 2013.

Ashhab. In arrays of quantum dots. and in Fig. 2008).or two-dimensional array of qubits which can be manipulated in different manners. NV centers in diamond have not received much attention as potential quantum simulators. Therefore. Mod. (b) onedimensional or (c) two-dimensional arrays of cavities (Greentree et al. the spacing between dots is about 0.. Georgescu.1 to 10 μm) are to be compared with the far smaller average interatomic distances in solids. Ashhab. or (f) two-dimensional Coulomb crystals (Porras and Cirac. January–March 2014 . 86.or two-dimensional arrays of qubits and controls can be used to simulate various models in condensedmatter physics. A rather unconventional example of a quantum simulator is a two-component BEC in which the propagation of sound waves could simulate some aspects of cosmic inflation (Fischer and Schützhold. the distance between qubits is typically a few microns. M. 2006b). Currently only with neutral atoms in an optical lattice is it possible to perform quantum simulations with more than a few particles. These interqubit distances (from 0. From Buluta and Nori. Vol. recent progress in this FIG. and Franco Nori: Quantum simulation centers in diamond (Ladd et al. 16 (color online). The larger distances between qubits make quantum simulators more controllable and easier to measure.. Unlike other systems studied in the context of quantum computation. 2008). The systems shown above realize a one. 2010. Cavity arrays based on photonic band gap cavities would also have submicron intersite separations. F. they can be thought of as models of the magnified lattice structure of a solid. 2005).1 μm.. Yao. Another system that could be used for the quantum simulation of condensed-matter physics is electrons trapped on the surface of liquid helium (Mostame and Schützhold. 16. Although individual control and readout is not yet available. which lists the strengths and weaknesses of each potential quantum simulator. Brandao. 2013). (d) ions in linear chains.. and these systems can at present be considered the most advanced platform for AQS.. and Nori. which might be due to difficulties in coupling the NV centers to each other and future scalability. S. Hartmann. The average distance between the atoms in optical lattices is less than 1 μm. 2009. In the case of electrons on helium the distance between neighboring sites would be about 1 μm. and Plenio. there have been recent studies attempting to alleviate these difficulties (Weimer. 2006. however. they are about 10–15 μm and about the same for two-dimensional Coulomb crystals. (e) two-dimensional arrays of planar traps (Chiaverini and Lybarger. 2008). Current state of the art The above discussion of the physical systems that could implement quantum simulation is summarized in Table I. One. Phys. 2008). Rev. Examples of such analog quantum simulators include (a) atoms in optical lattices (Jaksch and Zoller. 2006). with a magnification factor of 2 to 4 orders of magnitude. but the control would be very difficult.170 I. Nevertheless. which are ≤ 1 nm. 2011). 2002. Although such a quantum simulator would be limited to a narrow class of problems. Byrnes et al. 2004). Buluta. In superconducting circuits. A related system that was proposed recently to implement DQS is a chain of molecular nanomagnets controlled by external magnetic fields (Santini et al. In principle it could be scaled up. No. and Lukin. 1. This setup could be used to simulate the Ising model. it provides an alternative possibility for the simulation of such systems. 2011). (h) in arrays of superconducting circuits. or (i) trapped on the surface of liquid helium (Mostame and Schützhold. As for the inter-ion distances in ion trap arrays. (g) electrons in quantum dot arrays created by a mesh gate (Manousakis.

2008. and Nori.. and superconducting circuits (Mei et al. and Cirac. Liao et al.. the DQS approach allows the simulation of complex spin-spin interactions and could potentially be combined with AQS techniques. evolution. Gerritsma et al. Vol.↓ . 2013). It is also worth noting that in certain limits the Hubbard model reduces to spin models.σ Þ X ~ nði. and measurement steps are described in detail.. with one experiment requiring sequences of more than 100 quantum gates. Moreover.. being quantum systems themselves. 2010. A.jÞ. They explicitly gave the mapping between the operators of the Hubbard model and the DQS. Recent experiments on the subject include the realization of a TonksGirardeau gas in one dimension (Paredes et al.σ aðiþ1...jþ1Þ. Wilson et al. The simulation of the Hubbard model with attractive interactions using atoms in optical lattices was discussed by Ho. Hartmann. (2002). With the current experimental state-of-the-art techniques. Gerritsma et al. others can be treated classically (e. In these experiments.jÞ. 2007. 2010). These include simulations using polar molecules (Herrera and Krems. Dirac particles (Lamata et al.. James-Cummings Hamiltonian and interferometry) but are discussed here as benchmarks for the progress of quantum simulation. No. and Franco Nori: Quantum simulation direction has been made and alternative approaches using Rydberg atoms or polar molecules are now being pursued.σ aði. 2006a. Condensed-matter physics 2. Moreover. The simulation of the Hubbard model has also been considered in the context of DQS. Gong et al. 2008). þU (21) ði. VII. quantum simulations with > 10 qubits and hundreds of high-fidelity gates seem possible in the coming years. Deng.I. with a quantum simulator one could address problems that are either intractable on classical computers or experimentally challenging.↑ nði. the fidelity of the DQS implementation was somewhat lower than for the AQS implementation in the same system (Friedenauer et al.jÞ. Zhou.σ Þ þ ty ða†ði. Somma et al...g.σ ½tx ða†ði. Since no error correction was used.σ aði.. 2011)] and superconducting circuits provide a way of studying intriguing quantum phenomena such as the dynamical Casimir effect (Johansson et al. trapped ions (Porras and Cirac. 2009. Georgescu. 2013).. 2011. an AQS of the Bose-Hubbard model using atoms in optical lattices was proposed by Jaksch et al. 2010. (2007.σ þ a†ði. Among the best-known challenges in this field are understanding high-T c superconductivity and disordered and frustrated systems. January–March 2014 HXYZ ¼ N X ½J x σ xi σ xiþ1 þ Jy σ yi σ yiþ1 þ J z σ zi σ ziþ1 : (22) i¼1 As mentioned in Sec. The simulation of the Fermi-Hubbard model in quantum-dot arrays was proposed by Byrnes et al.jÞ. 1.g... and Giamarchi (2009). Herrera et al.g. 2011b). . 2011.σ þ a†ðiþ1. simulations of various spin models with up to six spins have been demonstrated. 2011a... and Plenio.jÞ. Phys. A possible implementation of DQS for the Holstein model was analyzed by Mezzacapo et al.. 2004a. While some problems are classically intractable (e. M. 2012).jÞ. the model is difficult to treat using classical computers. In this section we discuss how different problems can be studied using the quantum simulators described in Sec. 2008. 2009).. Hubbard models. along with the Hubbard Hamiltonian spin Hamiltonians can be simulated on a DQS. the effects of decoherence and the transition from quantum to classical). (2002) showed how to obtain the energy spectrum of the FermiHubbard Hamiltonian: HH ¼ − X ði. VI. 2008) or atoms in arrays of coupled cavities (Brandao. 2007.jÞ. 86. spin frustration and disorder. 2008). As discussed in Sec. ions (Stojanović et al..A. recent experiments with trapped ions have demonstrated exotic quantum simulations beyond condensedmatter physics [e.. 171 in more than one dimension.g. the most advanced DQS has been realized with trapped ions (Lanyon et al. and possibly even biology (Ghosh. 2011). quantum simulators are able to provide more insight into quantum phenomena than classical simulators (e.A. and the last term (U) repulsion potential energy. IV. S. Hubbard model The Hubbard model is the simplest model of interacting particles on a lattice. Note that the mapping between the fermionic operators and the Pauli matrices employs the Jordan-Wigner transformation. 2010) or the Klein paradox (Casanova et al. However. In this section we discuss the quantum simulation of these problems. APPLICATIONS Quantum simulators will find numerous applications in diverse areas of physics and chemistry (Lanyon et al. Zhou. Ashhab. With error correction and improved control of laser intensity fluctuations. 2013). Porras. the initialization.jÞ..σ aði. Mod. (1998) and implemented by Greiner et al. and quantum chemistry calculations). Smirnov... Meanwhile. Spin systems are typically described by Hamiltonians of the form 1. 2013). In general.jþ1Þ. Cazalilla. Quantum simulation leads to new results that cannot be otherwise simulated and also allows the testing of various theoretical models.jÞ. 2004) and the investigation of atoms trapped in a graphenelike lattice (Uehlinger et al.jÞ where tx and ty are the hopping matrix elements allowing fermions to move on the lattice.. Proposals for the AQS of the related Holstein model have also been put forward. 2010). Lähteenmäki et al. lattice gauge theories. Spin models The difficulty in solving quantum many-body problems is reflected in the open questions in condensed-matter physics. (2012). for large numbers of particles Rev. The first terms describe ~ is the on-site the kinetic energy. There have also been proposals to simulate this model with polar molecules (Ortner et al. VI. Dong et al..

(2009) demonstrated the emulation of the dynamics of single spins with principal quantum numbers s ¼ 1=2. respectively. Ashhab. and Cirac. Phys. and Bose (2008a).172 I. spatial anisotropy. S. Mod. 3. 2011). 2s þ 1 quantum states of the circuit are used. an AQS for the Ising [Eq. Martin-Delgado. 18 (color online). 1. Porras. This direction was pursued further by Deng. Ashhab.. Rev. Edwards et al. with the goalkeeper serving as the anchor. 17 (color online). Spin systems with s > 1=2 can also be naturally realized using superconducting circuits. 86. and Britton et al. The effective interactions between the spins S1 and S2 of the molecules in their rovibrational ground states are generated with a microwave field EðtÞ inducing dipole-dipole interactions between the molecules with dipole moments D1 and D2 . and Franco Nori: Quantum simulation FIG. (2008). (2012). Earlier Milburn (1999) showed that the conditional displacement of the vibrational mode of trapped ions can be used to simulate nonlinear collective and interacting spin systems. The ability to access arbitrary dimensions is made possible by the flexible connectivity. which derives from the flexibility in designing tunable couplers between superconducting qubits and resonators. 2011). Porras and Cirac (2005a. Kim et al. The antiferromagnetic Heisenberg model with longrange interactions can be realized with solid-state NMR (Roumpos. Bermudez. (2003). while a fifth state was used in the experiment as a reference state for phase measurement. and Nori (2010) proposed the simulation of spin and Hubbard models in higher or fractal dimensions with superconducting qubits and resonators. For simulating a spin s. Recently. 2005b). we now discuss quantum phase transitions. (2012) and put into practice in the experiments of Friedenauer et al. No. 2009). Neeley et al. and Martin-Delgado (2009). (22) with J z ¼ 0]. spin chains and ladders can be investigated with atoms in optical lattices (Garcia-Ripoll.. Furthermore. Micheli. An analogy with a classical system (a soccer game). (22) with Jx ¼ J y ≠ J z ] and isotropic XXX [Eq. up to s ¼ 3=2. These interactions can be switched and tuned by lasers and by the choice of trapping conditions. (2009). From Nori. 17). and Zoller (2006) discussed the possibility of engineering Hamiltonians of spin lattice models with polar molecules stored in optical lattices (see Fig. 2007). can also be used to give a simple explanation of the redistribution of the population among the different energy levels in terms of players passing a ball among each other. The spins are represented by single electrons of heteronuclear molecules. January–March 2014 FIG. 2004). M. shown on the left side. 2006. dipole-dipole interactions. 2004b). XY [Eq. 1. and 3=2 (see Fig. Quantum phase transitions occur when one varies a . 18). Illustration of the experiment of Neeley et al. (2010). and spin-rotation couplings enables the realization of effective two-spin interactions with designable range. or XYZ [Eq. The DQS implementation of various spin Hamiltonians with atoms in optical lattices or trapped ions was discussed by Jané et al. For instance. (22) with Jx ¼ Jy ¼ J z ] Heisenberg spin Hamiltonians in an arbitrary array of coupled cavities was proposed by Cho. Quantum phase transitions Having introduced the Hubbard model and spin models. Islam et al. Itinerant ferromagnetism can be studied in a two-component Fermi gas (Jo et al. The AQS of spin models is also possible. From Micheli. 2009. A recent trapped-ion experiment demonstrated the DQS of different spin Hamiltonians using sequences of elementary quantum gates (Lanyon et al. Brennen. in which the states of a superconducting circuit were used to emulate a particle with spin s. and coupling strength. and Cirac (2005). and Duan (2011). and Korenblit et al. Georgescu. and Zoller.. because these circuits generally have more than two quantum states that can be employed in the simulation. especially in these models. in (d) four of the circuit’s quantum states are mapped onto the four quantum states of a spin-3=2 particle. and Yamamoto. Angelakis. (2010. and a scheme to realize the anisotropic XXZ [Eq. Using a combination of microwave excitation. Master. (2011). By applying carefully tuned microwave-frequency drive signals. (2008). (22)] interactions can be realized using the collective vibrational modes (Porras and Cirac. Monroe. Schematic diagram of a proposal for the simulation of anisotropic spin models using polar molecules trapped in an optical lattice: a two-dimensional square lattice with nearest-neighbor orientation-dependent Ising interactions. Brennen. Vol. with trapped ions. Tsokomos. For example. Lin. (23)]. the circuit can simulate the interaction between the spin and a magnetic field with arbitrary strength and direction. Porras et al. Porras.

but are difficult to investigate both by classical simulation and via experimental methods. (2008). 1. This state is sometimes called strongly correlated. or arrays of Josephson junctions (van Oudenaarden and Mooij. (2008) using trapped calcium ions. (12).. Ashhab. Quantum phase transitions are an interesting and fundamental subject. the energy is minimized when the particles are U distributed evenly among the lattice sites and the ground Q state is approximately given by i bˆ †i j0i. the transition from a paramagnet to an antiferromagnet was emulated using two trapped calcium ions by Friedenauer et al.e. The phase transition is perhaps easiest to understand in the case of a unit filling factor. Georgescu. The first term. Then an abrupt change in the ground state of the system is observed (Greiner et al. i. i. A recent experiment demonstrated the quantum simulation of antiferromagnetic . is simulated by coupling the internal levels (representing the two spin-1=2 states) with a resonant rf field. physical parameter at absolute zero temperature. Deep in the superfluid phase. the system undergoes a transition from paramagnetic (j →ij →i    j →i) to ferromagnetic (j↓ij↓i    j↓i þ j↑ij↑i    j↑i) or antiferromagnetic order (j↓ij↑i    j↓ij↑i þ j↑ij↓i    j↑ij↓i). 2006). (2002) using rubidium atoms trapped in an optical lattice.I. Phys. and Plenio. 2002) in an ensemble of atoms trapped in an optical lattice.. 19 (color online). when there are as many particles as lattice sites. The same technique for realizing the spin-spin interaction. Mod. and Franco Nori: Quantum simulation 173 FIG. The interactions between individual spins and an external magnetic field were simulated by coupling the internal levels (representing the spin-1=2 states) with a resonant rf field. implemented by a walking wave formed by two perpendicular laser beams. From Buluta and Nori. when ~ delocalization minimizes the kinetic energy and the J ≫ U. (a) The quantum phase transition from a superfluid to a Mott insulator phase realized by Greiner et al. 2002). was first observed in 2002 (Greiner et al. 19(b). (2010) and Islam et al. January–March 2014 diagram of this quantum phase transition is illustrated in Fig. while the spin-spin interactions were simulated using a statedependent optical dipole force implemented by a walking wave. S. The physical model is described by the Bose-Hubbard Hamiltonian given in Eq. Examples of analog quantum simulators of quantum phase transitions using (a) ultracold neutral atoms and (b) trapped ions. The Hamiltonian for a chain (with two or more spins) is given by H I ¼ −Bx X X σ xi þ J ij σ zi σ zj . the quantum phase transition from a superfluid to a Mott insulator phase. and it corresponds to the Mott phase. i (23) i<j where Bx is the magnetic field strength and J ij are the spinspin coupling coefficients. 19(a). system is said to be in the weakly interacting P regime [the ground state being approximately given by ð i bˆ †i ÞN j0i. Vol. the state-dependent force. such that the quantum phase transition was observed. 86. (1989). 1996). The transition describes an abrupt change in the ground state of the many-body system governed by its quantum fluctuations. When the strength of the spin-spin interaction is increased adiabatically (increasing J while keeping Bx constant). The spin-spin interaction in the second term is simulated by a statedependent optical dipole force. By adjusting the lattice potential depth. M. all the particles are in the single-particle ground state]. (2002) by adjusting the lattice potential depth. 2009. This is illustrated in Fig. Analog quantum simulators can help in understanding this purely quantum phenomenon.e. There are other alternative ways of simulating this quantum phase transition with arrays of cavities (Hartmann.. denoting the interactions of individual spins with an external magnetic field. predicted by Fisher et al. The system is described by the quantum Ising model. Further experimental investigations in this direction have been performed by Kim et al. namely. and the first steps in this direction have already been explored. the ratio between the tunneling energy J and the on-site interaction ~ can be controlled and brought to the point where the energy U transition between the Mott insulator phase and the superfluid phase is induced. can be applied for many ions in a string or ion array. For example.. (b) Magnetic quantum phase transition simulated by Friedenauer et al.. A schematic Rev. The ratio between the tunneling energy and the on-site interaction energy was controlled in Greiner et al. No. when all thermal fluctuations have ceased. (2011). When ~ ≫ J. In quantum magnets. Brandao.

2010). Becker et al. (2005). 2008. (24) can be diagonalized exactly. Solano. the wave functions of particles can be localized in disordered media in spite of the presence of hopping terms in the Hamiltonian.. 2013). For example. Since the theoretical treatment of these problems is particularly challenging. Furthermore. A rich phase diagram with different types of phase transitions was observed... GarciaRipoll. January–March 2014 where the fermionic operators c†kl (ckl ) and b† (b) create (destroy) a spinless fermion in the conduction mode kl and in the impurity.. A review focused on the simulation of disordered quantum systems with quantum gases has been given by Bouyer (2010). optical lattices have been used to experimentally realize a disordered system that may lead to the observation of a Bose glass phase (Fallani et al. Recently. Disordered and frustrated systems Disordered systems appear in many difficult problems in condensed-matter physics such as transport. M. Ashhab. Liu et al. 2010. This phenomenon can occur at the single-particle level by coherent backscattering from random impurities. Bermudez et al. A chain of trapped ions could also be used to explore the physics of disordered quantum systems (Bermudez. The coupling parameters can be tuned to ferromagnetic (solid lines) or antiferromagnetic values (dashed lines). antiferromagnetic. As for AQS. (2007) and references therein. 2005. 2006b. the motional degrees of freedom of atoms trapped in an optical lattice were used to simulate ferromagnetic. Kim et al. (2010). Note that in principle the Hamiltonian in Eq. The fermions can hop between nearest-neighbor sites on the ring or between pffiffiffia site on the ring and the impurity (with matrix element V= n for the latter). Frustrated antiferromagnets are materials whose quantum Monte Carlo simulation suffers from a severe sign problem. (2011)]. NMR (Peng et al.. simulated using the motional degrees of freedom of atoms trapped in an optical lattice. (2010). Ivanov and Schmidt-Kaler. 5. spin glasses. exhibited an analog of universal conductance fluctuations. and Oliver. Vol. 2011).. From Struck et al. Figure 20 shows examples of spin configurations in a triangular lattice and their experimental signatures. Rev.174 I. S. Martin-Delgado.. which are typically studied in the context of particle propagation in twodimensional disordered media. and Struck et al. and Martin-Delgado (2008) proposed a mapping between superconducting circuits and the Hamiltonians describing magnetic impurities in conduction bands (Anderson and Kondo models). The Fano-Anderson Hamiltonian can be studied using DQS. Nevertheless. photons (Ma et al. respectively.. 2008). 1. FIG. A recent experiment on a driven superconducting qubit (Gustavsson.. Geometric frustration refers to the situation in which the geometric properties of the crystal lattice forbid the simultaneous minimization of all the interaction energies acting in a given region (see Fig. 2011). 4. Roati et al. 86. 2010). Spin glasses can be . No. Georgescu. Billy et al. conductivity. and Porras.. The Fano-Anderson Hamiltonian can be written as follows: HFA ¼ n−1 X εkl c†kl ckl þ ϵb† b þ Vðc†k0 b þ ck0 b† Þ. 2011). Thermal states of frustrated spin systems were also simulated recently using NMR (Zhang et al. In a recent experiment. Bylander. so the spin orientation cannot be uniform in space even at low temperatures. 2012). The fermionic Fano-Anderson model in one dimension consists of an n-site ring with an impurity in the center. theoretical and experimental studies have investigated possible routes toward Anderson-like localization of atomic BECs in disordered potentials (Schulte et al. averaged over several experimental runs. several proposals for quantum simulation have been put forward. Mod. and Franco Nori: Quantum simulation spin chains using neutral atoms in an optical lattice (Simon et al. the decoherence-induced localization of the size of spin clusters was investigated in an NMR quantum simulator (Álvarez and Suter. 2009). 2011. and atoms in optical lattices [see Lewenstein et al. and the spin orientation can become random and almost frozen in time. Well-known examples include the antiferromagnetic Ising model on a triangular lattice or the antiferromagnetic Heisenberg model on a kagome lattice.. 20 (color online). 2011). and these determine the resulting spin configuration. and some models of high-T c superconductivity. There have been studies on the AQS of frustrated spin models with trapped ions (Porras and Cirac.. This idea was pursued theoretically and experimentally (with liquid-state NMR) by Negrevergne et al. Phys. along with the appropriate mapping. and frustrated classical spin configurations (Struck et al.. (24) l¼0 Spin configurations in a triangular lattice and their experimental signatures. 21). 2007). They can exhibit characteristic phenomena that are not present in perfectly ordered systems. the experiment serves as a benchmark for the experimental progress of quantum simulators. Spin glasses Spin glasses typically occur when the interactions between spins are ferromagnetic for some bonds and antiferromagnetic for others. 2011.

several groups have observed signatures of the crossover between a BCS and a BEC superfluid as the strength of attractive interactions between fermionic particles is varied [see. 2008. (2005). 2000) can be simulated in an analog manner by an array of electrostatically defined quantum dots. An analog approach to the simulation of spin glasses using magnetic impurities embedded in inert matrices. 2006). (c) The expected form of the Ising interactions J 1 and J 2 . 2010). e. …. the small squares defined by four neighboring vertices) in a square lattice and the products involve the qubits . In another direction related to the quantum simulation of superconductivity. Mod. as the parameter μ~ is defined relative to their frequencies. where v runs over the vertices and p over the plaquettes (i. 2007). and Franco Nori: Quantum simulation 175 understand high-T c superconductivity. From Kim et al. 2010. Greiner. The Hamiltonian HK ¼ − X X Av − Bp v (26) p Q is the sum commuting stabilizers Av ¼ i∈v σ xi and Q of mutually z Bp ¼ i∈p σ i .. One example of a quantum metamaterial is an infinite chain of identical qubits inside a resonator. J 1 is the nearest-neighbor coupling and J 2 is the next-nearest-neighbor coupling. Lidar and Biham (1997) and Lidar (2004) developed an algorithm for the construction of the Gibbs distribution of configurations in the Ising model. efficiently simulated using DQS. No. Superconductivity The high-temperature superconductivity of compounds containing copper-oxide planes is still a puzzle that might be solved using large-scale simulations. The t-J model describes strongly correlated fermions on a lattice and it is used in various attempts to Rev. −li. Regal. 2009). and to models with a magnetic field. the analog simulation of the t-J model was proposed by Yamaguchi and Yamamoto (2002). and Jin (2004). 1. N. to diluted Ising models. 7. Metamaterials The behavior of tunable metamaterials (regular structures obtained from the periodic arrangement of mesoscopic building blocks) in the quantum regime can be seen as a quantum simulation of materials composed of regular atomic structures (Bliokh et al. the number operator nFm ¼ c†m cm. Spin glasses could also be studied using Fermi-Bose mixtures in inhomogeneous and random optical lattices as suggested by Sanpera et al... Ashhab. Ashhab. jmj ¼ 1. S. Rakhmanov et al. Raussendorf. and Lewenstein et al. which are not available to standard materials.. controlled through the laser detuning μ.I.. Such a system offers new ways of controlling the propagation of electromagnetic fields.e. Here t represents the size of the kinetic energy of a hole disrupting an antiferromagnet with spin-spin interaction energy J. 2. as suggested by Manousakis (2002) in one of the early proposals for AQS. Georgescu. M. Moreover. Zwierlein et al.g. using superconducting qubits in circuit QED was given by Tsomokos. and references therein]. The algorithm can be applied to any dimension. This simulation was realized with (b) three ytterbium ions in a onedimensional trap. 2008. Moreover. they have been proposed as media for implementing topological quantum computation (Kitaev. −mjVjl.. 21. such as SherringtonKirkpatrick spin glasses. (2013). January–March 2014 2 ðnFm þ nF−m Þ þ N X † † Vþ ml cm c−m c−l cl .. 8. and Duan. and Lidar (2002) and Wang and Yang (2006) and a two-qubit version of the algorithm was experimentally realized using NMR (Yang et al. 6. (2007). The three oscillation modes shown at the top of (c) are used as markers. is depicted. was proposed by Lemeshko et al. (25) m. A polynomial-time algorithm to model this BCS Hamiltonian was proposed by Wu. 86. Vol. The latter approach has been experimentally realized with linear optics (Lu et al. Anyons have been used to describe two-dimensional systems such as sheets of graphene or the quantum Hall effect. to the class of spin-glass Ising models with a finite portion of frustrated plaquettes. such as solid helium. Phys. denotes all relevant quantum numbers.l¼1 where the fermionic c†m (cm ) are creation (annihilation) operators. and Nori (2008). (2004) and Ahufinger et al. 2003) involving fourbody interactions can be studied using AQS with cold atoms in optical lattices or alternatively using DQS (Han. and the matrix element V þ ml ¼ hm. Byrd. ~ used to generate an optical spin-dependent force. A proposal for the analog simulation of the Lipkin-Meshkov-Glick model and complex quantum systems. An example of the simulation of the simplest case of spin frustration: (a) three antiferromagnetically coupled spins. 2003). (2005). The study of BCS pairing in superconductors could be done using DQS.. The general BCS pairing Hamiltonian has the following form: HBCS ¼ N X ϵm m¼1 FIG. The CuO2 plane in a high-T c superconductor (Orenstein and Millis. Topological order Anyons are two-dimensional particles whose quantum statistics are neither bosonic nor fermionic. The fractional statistics of anyons in the Kitaev model (Kitaev.

. and Franco Nori: Quantum simulation FIG. Tagliacozzo. Micheli. Stannigel et al. The ring exchange Hamiltonian. A recent paper extended this work to the study of Zitterbewegung in a magnetic field (Rusin and Zawadzki. From You et al. a ring of hard-core bosons collectively rotate like a ring around a central boson. 2013. 2010). In optical lattices the creation and manipulation of Abelian and non-Abelian anyons in topological lattice models can be realized using ancilla particles (Aguado et al. M. They gave the mapping between the operators in the lattice gauge Hamiltonian for the U(1). Maraner.. January–March 2014 Dirac equation. This simulation offers the possibility of studying effects such as Zitterbewegung and the Klein paradox. which was used to simulate the generation of photons in the dynamical casimir effect. There have also been related proposals using trapped ions (Casanova et al. and Lewenstein. Ashhab. 2010.. (ii) qubits 1 and 3 are coupled via an LC oscillator. 2009. A method for simulating the Dirac equation in 3 þ 1 dimensions for a free spin-1=2 particle using a single trapped ion was proposed by Lamata et al. 22 (color online). 2010. 2013) lattice gauge theories using neutral atoms. and Z bonds. Stanescu et al. 2010. Note that it is also possible to classically simulate the Zitterbewegung of a free Dirac electron in an optical superlattice (Dreisow et al.. 2011. Aguado. Kitaev. 2013). (2009). Orland. and superconducting circuits (Marcos et al. B.. Novoselov. The implementation of these proposals can also contribute to the development of topological quantum computation. Wang et al.. High-energy physics Another area for the application of quantum simulation that is already showing promising developments is high-energy physics. Zamora. 2005. Yang. The simulation of gauge theories can be a very computationally intensive quantum many-body problem. and Kitagawa (2012). and there have been several recent proposals for the quantum simulation of such theories. A review on the subject was given by Wiese (2013). Pedernales et al. Celi.. Zitterbewegung has never been observed in its original form with free relativistic particles. and Reznik. 2011). where. 2013). and SU(3) (which is particularly important for quantum chromodynamics) lattice gauge theories and spin operators. Zohar. Witthaut. 2011. it is possible to construct a Kitaev honeycomb lattice with superconducting circuits (You et al. 1. Brennen. 2010. Hauke et al. and (iii) qubits 1 and 4 are capacitively coupled via a mutual capacitance Cm . and Geim (2006). Ring-exchange models describe elementary excitations in a solid. The simulation of lattice gauge theories on a DQS was investigated by Byrnes and Yamamoto (2006)... The Klein paradox refers to the situation in relativistic quantum mechanics where a potential barrier of the order of the electron mass becomes nearly transparent for the electron. Cirac.176 I. 2012). 2012. Inset: These three types of interqubit couplings are denoted as X. (2013) considered the possible simulation of the Schwinger model using a chain of trapped ions. 2012. a mapping between the dynamics of the 2 þ 1 Dirac oscillator and the Jaynes-Cummings model (in particular. and Solano (2007). Another proposal for the DQS of field theories was given by Jordan. No. surrounding the vertices or plaquettes (note that in this setup the qubits are placed at the centers of lines connecting neighboring vertices). Vol. and Reznik. Tagliacozzo.. but it has been simulated with a trapped ion (Gerritsma et al. and they showed that the algorithm is efficient.. Lee.. 2011). and Zoller (2006). Furthermore. in connection with trapped-ion experiments) was proposed by Bermudez... 2013. More recently. (2007). The study of relativistic quantum systems such as gauge fields or Dirac fermions with quantum simulators was first suggested by Boghosian and Taylor (1998b). 2010) (see Fig. 86. 2010) [see also K. Hou. 2010). For example. Zohar.. Georgescu. Cirac. Brennen. cavity QED systems (Barrett et al.. Banerjee et al. Casanova et al. several theoretical studies considered Abelian (Kapit and Mueller.. A proposal for a simulation using graphene was put forward by Katsnelson. SU(2). Dirac particles could also be investigated with neutral-atom quantum simulators (Goldman et al. 2008). Topological models have also been investigated by Freedman. and Preskill (2012). Mod. A recent experiment (Lähteenmäki et al. Atoms in optical lattices offer the possibility of realizing the AQS of ring-exchange models (Büchler et al. 2013) used a superconducting circuit to simulate a massless Klein-Gordon field with tunable speed of propagation. 2013. Cirac. Y. Braun.. Such mappings would allow the study of relativistic quantum mechanics described by the Dirac equation in a nonrelativistic quantum system. (2010)]. 23(a)] and described by the . consisting of four superconducting charge qubits (labeled 1 to 4): (i) Qubits 1 and 2 are inductively coupled via a mutual inductance M. X H RE ¼ K ðb†1 b2 b†3 b4 þ b†4 b3 b†2 b1 − n1 n2 − n3 n4 Þ. plaquettes (27) can be realized using atoms with two internal states: one state trapped in a square lattice [see Fig. S.. Schematic diagram of the basic building block of a Kitaev lattice. 2006). Zitterbewegung refers to the rapid oscillatory motion of a free particle obeying the Rev. Xue (2011). Phys. and Lewenstein. A quantum simulation of this phenomenon was recently implemented with trapped ions (Gerritsma et al. Tewari et al. Martin-Delgado. 2009. and Cirac (2009). 22). Celi. 2013) and non-Abelian (Banerjee et al. and Liu. and Wang (2002). and Pachos. 2013.

Georgescu. The effects of those fields on the circuits are the analogs of the Stark and Zeeman effects in atoms.. and singlephoton generators (You et al. 2006) [see Fig. 1. 2006. water surface waves (Weinfurtner et al. 2009. artificial atoms can be lithographically designed to have specific characteristics. Furthermore. and a system of liquid helium could be used for the quantum simulation of problems in cosmology. Noninteracting relativistic fermionic theories or topological insulators could also be investigated using these systems (Mazza et al. and Nori... Semiao and Paternostro (2012) proposed a quantum-circuit simulation of the state of a nucleon. and Franco Nori: Quantum simulation FIG. This tunability is an important advantage over natural atoms. acoustic waves in a two-component BEC can be used to investigate scalar fields within the curved space-time structure of an expanding universe (Fischer and Schützhold. Furthermore.. produce Schrödinger-cat states.. Harrabi et al. (2007). (b) Simulation of compact U(1) lattice gauge theories. Superconducting circuits have also been used to engineer selection rules and thus create combinations of selection rules that are not possible with natural atoms (Liu et al. S..e.. 2011. Examples include the possible observation of an Unruh-like effect (i.or three-dimensional lattices (for example. Cosmology Quantum simulation can also find applications in analog gravity or cosmology models. Adapted from Tewari et al. January–March 2014 177 early Universe. 2009). for example. Dowling. 2005) and the simulation of the Schwinger effect (i. 23. such as a large dipole moment or particular transition frequencies. and simulate the Einstein-Podolsky-Rosen experiment (You and Nori. These circuits can provide analogs of sideband cooling and Sisyphus cooling (Grajcar et al.. Dipolar bosons reside on the sites of the kagome lattice (black dots). 2010) and superconducting circuits (Nation et al.. and Milburn (2010). The study of the analog of cosmological particle creation with trapped ions was proposed by Schützhold et al. 2008) (see Fig. 86. the analog of quantum field effects in cosmological space-times was investigated by Menicucci.. 2010). and more recently. Ashhab.. Atomic physics There is a deep analogy between natural atoms and the artificial atoms formed. as discussed by Schützhold and Mostame (2005). ions (Horstmann et al. D. and Milburn. 2005). Superconducting circuits can also be used for new types of masers. the artificial atoms in the circuits are driven by currents. . With analog models it is possible to test predicted phenomena that have not yet been observed in experiment. voltages. 23(b)]. In contrast with naturally occurring atoms. 2008) [see also (Nation et al. 24). a twodimensional kagome lattice) the simulation of a U(1) lattice gauge theory would be possible (Tewari et al. 2011). by electrons in small superconducting circuits (You and Nori. 2005). the production of electron-positron pairs from the vacuum under the action of a strong electric field) with atoms in an optical lattice (Szpak and Schützhold. In trapped atomic gases a microscopic Hamiltonian can be implemented and its phase diagram can be studied experimentally via controlling the strength of the interaction terms (Büchler et al. 2004). Differences between quantum circuits and natural atoms include how strongly each system couples to its environment (the coupling is weak for atoms and strong for circuits) and the energy scales of the two systems. the observation by an accelerating observer of a thermal flux of particles in vacuum) using the phonon excitation of trapped ions (Alsing. Mod. Nori. This simulation could be implemented using a photonic network. Deppe et al. Ashhab. and microwave photons.. 2005. This idea might be experimentally challenging but it opens up a new possible way to study cosmology. or even with ultrashort pulses of light in microstructured optical fibers (Philbin et al. 2010). There are also numerous similarities between the behavior of superfluid helium and cosmological phenomena. M. (2012)].. 2012). and a second one trapped at the centers of the plaquettes. Phys. lasers. These similarities could be exploited. Both have discrete energy levels and exhibit coherent quantum oscillations between those levels.. study Landau-ZenerStückelberg interferometry (Shevchenko. 2012). The simulation can be performed by varying the interparticle coupling and/ or expanding the condensate in a temporal ramp. (a) Proposed setup for the simulation of ringexchange models . But whereas natural atoms are driven using visible or microwave photons that excite electrons from one state to another. excitonpolariton superfluids in semiconductors (Gerace and Carusotto. we also note the possible simulation of the O(3) nonlinear sigma model using an array of superconducting and insulating spheres with electrons trapped in the insulating spheres. By implementing a Hubbard model with an additional... 2011). simple Bose-Hubbard model. 2005.. The bosons (black dots) reside on a square lattice with the molecules (gray dots) at the centers of the plaquettes. Moreover.e. 2009). and in doing so. such as processes in the Rev. Vol.I.. The resulting electric and magnetic fields control the tunneling of electrons across Josephson junctions. the hexagonal dual lattice is the lattice formed by the centers of the corner-sharing triangles (gray dots). Finally. No. 2008. 2007). Adapted from Büchler et al. study exotic phases with strong correlations. namely. Recently it was demonstrated that Hawking radiation can be measured in a classical analog system.. 2005. C. they can be used to test Bell inequalities (Ansmann et al. strong nearest-neighbor interaction on certain two. 2012). For example. de Groot et al. as discussed in detail by Volovik (2009). analogs of Hawking radiation can be investigated with atoms (Giovanazzi. 2008. with this system it is possible to simulate a certain class of strong coupling Hamiltonians. Olson. 2005).

. as first suggested by Lloyd (1996) and investigated experimentally by Tseng et al. First. 2010). 2008). Depending on the tunnel coupling strengths. 2003.. and Franco Nori: Quantum simulation Sisyphus cooling and amplification in a superconducting circuit (Grajcar et al. 2008. Furthermore. 2007). 86. In the analogy with Sisyphus in Greek mythology the current in the resonator plays the role of Sisyphus and the qubit plays the role of the rock. (2005) showed how to compute molecular energies using DQS and the required number of qubits was estimated (see also Fig. The algorithm runs in polynomial time while any Rev. electron distribution. 2005. The molecular energies of the hydrogen molecule have been calculated using a small photonic DQS (Lanyon et al. January–March 2014 exact classical simulation is bound to exhibit exponential scaling. it is also possible to simulate open quantum systems with closed quantum systems. and σ þ and σ − are the atomic raising and lowering operators.. one could exploit the natural decoherence of the simulator.. V. FIG. 24 (color online). Georgescu. the operators a† and a are the bosonic creation and annihilation operators. 2004. Other studies considered the Dicke model (Schneider and Milburn. Chen..178 I. The simulation of chemical dynamics is also achievable in polynomial time (Kassal et al. The number of qubits scales linearly with the number of basis functions. The algorithm involves the initialization of the DQS to an equal superposition of position eigenstates. Sisyphus amplification. 2007). which describes the interaction of a single quantized mode of the electromagnetic field with a two-level atom: HJC ¼ iℏg0 ða† σ − − aσ þ Þ. M. and the rock keeps going back to the bottom of the hill. Various reaction regimes and different reaction products can be obtained by varying the speed of voltage changes applied to the gates forming the quantum dots. 2001. 2009). Simulating open quantum systems with quantum simulators can be done in two ways.B). Wallraff et al. Maniscalco et al. More recently.. the open-system dynamics of a harmonic oscillator coupled to an artificially engineered reservoir (Lütkenhaus. (2013). A simulation of decoherence caused by classical noise that was artificially added to the control signal of a superconducting qubit was reported by Li et al. a sequence of measurements yielding the energy spectrum and eigenstates. For example. and Zoller. Vol. F. Sisyphus is constantly pushing the rock uphill (as can be seen in steps a and c). 2008. 2008). 1998. chemical reaction dynamics have been investigated using NMR DQS (Lu et al. 2011). 2011). Mod. Moreover. This promises the modeling of various chemical reactions (Smirnov et al. Quantum chemistry Quantum simulators can also have an important contribution in quantum chemistry (Kassal et al. The current in a nearby superconducting resonator drives the qubit. (2000) (see also Sec. Schoelkopf and Girvin. No. 2004). The two energy levels of a superconducting qubit are shown as a function of the current applied to the circuit. Ashhab. 1. The Jaynes-Cummings Hamiltonian can be realized with superconducting circuits (You and Nori.. Using AQS it would be possible to simulate chemical reactions. Phys.. Second. 2011). or open quantum systems in thermal equilibrium . Zueco et al. Open quantum systems Simulating the dynamics of open quantum systems is even more costly than that of closed quantum systems because solving the Lindblad equation requires quadratically more resources than solving the Schrödinger equation for the same physical system. and the number of gates grows polynomially with the number of qubits. The cycle a → b → c → d on the right side mimics Sisyphus cooling in atomic physics: the energy of emitted photons is higher than the energy of photons absorbed from the drive signal. a recent proposal suggests that chemical reactions can also be simulated with ultracold atoms in a waveguide (Torrontegui et al. so simulations with a few tens or a hundred qubits would already outperform classical computers.. Another example is the multiconfiguration selfconsistent field wave-function-based quantum algorithm proposed by Wang et al. the dots can form both ioniclike and covalent-like bonds. Chen. Adapted from Nori. a quantum computer could be used to simulate the static and dynamical chemical properties of molecules. S. an efficient algorithm for calculating the thermal rate constant was given by Lidar and Wang (1999). and Liang.. causing a constant removal of energy from the resonator. and shell structure. The cycle e → f → g → h on the left side represents the opposite scenario. Aspuru-Guzik et al. where energy is constantly added to the resonator. (2008) for obtaining the energy spectrum of a molecular system. E. They showed how a quantum dot with one electron can be considered an artificial hydrogen atom and a quantum dot containing four electrons can be viewed as an artificial oxygen atom. (28) where g0 is the atom-field coupling strength. and a microwave drive signal is applied to the qubit. One of the most important models in atomic physics and quantum optics is the Jaynes-Cummings Hamiltonian. 4). Cirac.. Piilo and Maniscalco (2006) theoretically showed that a driven harmonic oscillator can act as a quantum simulator for the non-Markovian damped harmonic oscillator. 2011. 2008. Nori. followed by a unitary evolution that makes use of the quantum Fourier transform and finally. Smirnov et al. (2007) proposed using the redistribution of electrons between semiconductor quantum dots to simulate the redistribution of electrons in a chemical reaction.

This was among the first experimental realizations of quantum simulation. Feng. 2002) and with linear optics (Howell and Yeaze.. Ashhab. You et al. Vol. Along the way. Georgeot. January–March 2014 179 Boson sampling (Aaronson and Arkhipov. H. The reviews by Zinner and Jensen (2008. graphene (Gibertini et al. and quantum thermodynamics (Maruyama. Another direction of interest is quantum heat engines. the photons interact with either one or two qubits acting as tunable mirrors. 1. Even though N is not as large as in condensed-matter physics (in this case N. Terraneo.. General methods for simulating the Markovian dynamics of open quantum systems have been investigated by Terhal and DiVincenzo (2000). 2010. and Nori. it is to be expected that more disciplines will want to add quantum simulation to their toolbox of research methods and many new applications will be uncovered. Superconducting qubits can be used. Georgeot. operating as tunable mirrors. 2013) detail similarities between atomic nuclei and cold atomic gases. 86. These latter phenomena can be realized using quasi-one-dimensional open systems where photons are injected from one side and move toward the opposite side of the device. a prototypical example in quantum chaos. and Franco Nori: Quantum simulation (Terraneo. 2013. Several recent experiments have demonstrated boson sampling using photons (Broome et al. As mentioned at the beginning of this section. external trapping potential. controlled by changing the applied electric and/or magnetic fields on the qubits. and Shepelyansky.. has an efficient realization in terms of quantum gates (Schack. the calculation of the nuclear force is difficult and therefore nuclear physics simulations require significant computing power. A proposal for the implementation of a quantum heat engine assisted by a Maxwell demon with superconducting circuits was given by Quan et al. 2002). 2010). S. and Nori (2011). As suggested by Georgescu et al. Rev. 2013. G. More recently. 1998a). with quantum simulation one can address problems that are either classically intractable (such as the numerous examples from condensed-matter physics discussed above) or experimentally difficult or inaccessible (such as the examples from high-energy physics and cosmology). No. J. can change the reflection and transmission coefficients of photons confined in waveguides. Given the ease of rotating atomic clouds. More recently. to perform Landau-Zener-Stückelberg interferometry (Shevchenko.. 2008. and Lau and James (2012) proposed an implementation using trapped ions. and Lee (2012).. 2009). Another recent experiment simulated non-Markovian dynamics using a linear-optics setup (Chiuri et al. 2004). A recent experiment with trapped ions demonstrated the possibility of engineering open-system dynamics using the dissipative preparation of entangled states (Barreiro et al. and Vedral. the study of Mach-Zehnder interferometry with an array of trapped ions was explored by Hu. one of which is the superfluid model of the atomic nucleus. Mod. the atomic mass number.. and Shepelyansky. and Wang. Quantum chaos An application of DQS with a few qubits is the study of the dynamics of simple quantum maps (Schack. Ashhab. 2004). 1996). for example. It has been shown that in some cases the quantum approach to the kicked Harper model can provide a polynomial gain as compared to classical algorithms. Dong et al. Rodriguez-Lara. 1998. Ashhab. Phys. 2011) is another interferometry problem that is closely related to quantum simulation. are now discussed in the context of quantum simulation. (2011) discussed an open-system quantum simulator with Rydberg states of neutral atoms in an optical lattice. Interferometry Nonlinear interferometers were first investigated in trappedion experiments (Leibfried et al. Liu. Marchesoni.. Nuclear physics In nuclear physics one must solve an N-body quantum problem. and Angelakis.. and Nori. VIII. where it was shown that six to ten qubits would be sufficient for a simulation of the single-particle Schrödinger equation. 2011. As practical quantum simulators become available. 2011). Levi and Georgeot. with concrete algorithms given in 1998 (Boghosian and Taylor. I. Other applications Some recent topics in physics research. There are many potential applications for quantum simulators in various fields of physics and chemistry. Ian. Several phenomenological models have been developed. Weimer et al. The possibility of simulating the Schrödinger equation on a DQS was first discussed in 1996 (Lloyd. and Nori. Tillmann et al. (2006).. For instance. Nori. 2009) or neutrino oscillations (Noh. 2012). These qubits. 1999). 2005. 2009). (2001).. Georgescu. Another example is the kicked Harper model (Levi and Georgeot. M. (2007). 1998). (2011). and rotation frequency. Liao et al. where a quantum lattice-gas model for the many-particle Schrödinger equation in d dimensions was proposed..I. It should exhibit observable interesting behavior even with only eight qubits. The quantum baker’s map has been experimentally realized in NMR (Weinstein et al. Spring et al. 2012). Hänggi and Marchesoni.. this approach offers the possibility of studying the response of nuclei to increasing angular momentum with unprecedented control over the relevant parameters such as the interaction strength. particle number. 2011). which could be utilized for designing quantum simulators of nuclei using atomic ensembles. 2013. CONCLUDING REMARKS Recent theoretical and experimental results on quantum simulation lead us to believe that practical quantum simulators . 2004. 2003). such as Majorana fermions (Casanova et al. 2003. The quantum versions of the Carnot and other heat engines were discussed by Quan et al.. the quantum baker’s map. this model could be simulated using an analog controllable system: a superfluid gas of fermionic atoms. the single-particle Schrödinger equation was considered by Benenti and Strini (2008). Bacon et al. is smaller than 300). Georgeot.. 2013). quantum Brownian motion (Hänggi. 2010) and Fano and Fabry-Pérot interferometry (Zhou. Crespi et al.

Milburn. M.. Niskanen. Chang. Hanson. M.. Phys. This would be a milestone for physics. Sanpera.. Lett. A. 110. Phys. Y. Lett. Rev. Lett. E. Angelakis. G. J.. M. A. Amasha. G. 333. A. Arkhipov. and A.. Rev. I. M. 014510. and Franco Nori: Quantum simulation will be built in the near future. Osterloh. Rev. Rev. 2006). Radu. 2008. Noh. W. I. 106. Álvarez. S. Zoller. de Groot. arXiv:1306. It is also of both theoretical and practical importance to investigate when and to what extent one can make use of the simulator’s decoherence. 1999. and M. However. 230403. 063616. 8. Ashhab. Love.. L. 1704. 2008. The demonstration of quantum simulations using more than a few tens of qubits would mark the point where quantum computers (whether digital or analog) surpass their classical counterparts. M. P. 2009. With the exception of atoms in optical lattices. 2001). Mooij.. and M.. Huo. Phys. E. 100502. 74. 052330. Lett. Phys. Greiner. L. J. The study of entanglement in many-body systems and its relation with quantum phase transitions (Amico et al. Phys. Rev. Gillen. I. A. J. individual control and readout are difficult to realize for atoms in optical lattices. and F. S. 104. ACKNOWLEDGMENTS We thank N. 2003). and G. W. J. Lett. Nat. Ahufinger. Kantian. 101. This work was partially supported by the ARO. F. S.2179.. A. and C. L. Browne.. The ability to simulate various Hamiltonians could then be useful for realizing practical adiabatic quantum computation (Ashhab. Rev. and S. Johansson. Even in the absence of a theory to be tested or realistically motivated conditions under which the behavior of a physical system is to be predicted. Fölling. S.. M. MacLean. Angelakis. Gossard. the RIKEN iTHES Project. J. P. D. Alsing. Phys. L. S.. Childs. 125303. and Briegel. Rev. D. 2011. and V. 2012) could lead to new possibilities for studying quantum simulation and quantum systems in general. and A. in addition to allowing access to Rev. 504. 86. A. 517. S. and F. M. S. C.-J. G.. From the theoretical point of view. Progress on the experimental implementation of quantum simulation would also be relevant for measurement-based quantum computation. Vol. Bakr. M. Zumbuhl. Lloyd. J. 110.. and M. computer science. Dowling. Georgescu. D. D. adiabatic quantum computation can be viewed as a special case of quantum simulation where the ground state of the simulated Hamiltonian encodes the solution to a computational problem.. A 64. E. Rev. For example. P. M. 547. there are still issues that must be addressed. and Adams. Cirac. Amico. the ability to incorporate nonphysical operations into quantum simulations of physical systems (Mezzacapo et al. P. Phys. 2005. Nature (London) 462. P. Greiner. Phys. P. 5162. Dalmonte. 2008. Rev.. Johansson. Science 309. M. Nation for useful comments on the manuscript. 2005. Phys. Pachos.. D. Quantum simulation will profoundly impact physics research. quantum simulators cannot yet handle large arrays of qubits. 285. Angelakis. G. January–March 2014 new physical regimes that are currently beyond experimental reach. Angelakis. Abrams. 12-02-92100. Kwek. and Nori. Lloyd. F. Phys. A 76. 83.. From the experimental point of view. and the theory of quantum computation (Vollbrecht and Cirac. A. A. Proceedings of the 43rd Annual ACM Symposium on Theory of Computing (ACM. Kyoseva. Bacon. Lett.. 79. Kempe. Lett. A. Nakamura. Rev. and L. Aguado. Bögli. J. Tai. 2008). Lett. Suter. 1. S. and P. Aspuru-Guzik. 2001. K. C. U. 2005. 220401.. Finally. Phys. Chuang. 100. and S. in all proposed quantum simulators improved controllability and scalability are required. Harmans. . A. P. E. Lewenstein. Ansmann.180 I. 2010. 2003). Leung. 260501. 2010. a Grant-in-Aid for Scientific Research (S). Fazio. Rev. L. We note here that for some problems where bulk properties are of interest. Vedral. theoretical and experimental progress in quantum simulation will also have a positive impact on the development of other fields such as adiabatic quantum computation (Farhi et al. Rico. Stebler. For instance. It will provide a new tool for testing physical theories or predicting the behavior of physical systems in various possible conditions.. Rev. as some recent experiments with trapped ions and superconducting qubits have demonstrated. Science 329. C.. 80. Ashhab. 2008. and D.. at least for certain applications. Phys.. O. 062302. measurement-based quantum computation (Raussendorf. Fölling. 2586. Nature (London) 461. Korepin. A 74. 031805. W. Phys. 2006. Huo. Aspuru-Guzik. 2013. Nori. Rev. and X. C. Lett. D. Verstraete. 153601. topological quantum computation (Kitaev. and J. et al. R. Abrams. Nori. 046803. 2013. Ma. Pollet. M. Dutoi. Quantum simulators would not only provide insights into new physical phenomena. R. Picot. A. Wunderlich. Walther. Wiese. M. Bose. D. ions in planar Coulomb crystals (Taylor and Calarco. Phys. 2007. the JSPS-RFBR Contract No. especially the estimation of the experimental requirements for each quantum simulator. 2008. the MURI Center for Dynamic Magneto-Optics. New York). Bakr. D. 94. J. Peng. even with a small-scale quantum simulator exciting physical phenomena could be explored. Simon. 2011. Gillen. M. Mod. further studies of decoherence and control would be very useful. Banerjee. 1997. Moreover. 2009. I. Rev. quantum simulators will also offer scientists a new realm of exploration. 2013. M. V. It is also important to note that. Santos. Brennen. S. while for other systems that is not a problem. and science in general. and S. Sanchez-Palencia. 2009) and atoms in optical lattices (Kay. Phys. REFERENCES Aaronson. D.. K. and the JSPS via its FIRST program. Head-Gordon. G. G. Peng. C. 2006) have been proposed for implementing measurement-based quantum computation. D. No. Ashhab. Phys. Harrabi. Mod. Zhou. On the other hand. but also help solve difficult quantum many-body problems. T. individual control and measurement may not be required. K. It might very well happen that unpredicted discoveries are made through the curiosity-driven experimentation with future quantum simulators. Phys. Rev. A. D. Kwek. D. MEXT Kakenhi on Quantum Cybernetics. 2012.. and V. A 72. Lambert and P. B 77.. and M. 2008) should also be mentioned as an exciting direction closely related to quantum simulation. J. and P. Kastner.

Clark. V. 041801. Rev. G. 104401.. Phys. and C. B. Taylor. 123016. Deng. I. Angelakis. D. Commun. C. A. and S. Rev. A. 277. K. 2010. Lett. Nielsen. Lett.. J. and F. Marx. Nat. Lett. Cirac. Rev. Huber. Chen. Phys. Brandao. H. and B.. Benenti. Phys.. Y. and W. 2009. Hermele. and N. White. Phys. Blatt.. 2006. C. 2010. H. Phys.. F.cfm?id=2012100]. and J.. S. Math.. Rev. S. Cirac. Bacon. Porras. Bliokh. M. Photonics 7. M. I. and K. J. Rev.... 83.. and B.. Rev. Phys. Ashhab. Lett. Brown. Casanova. 2003. Physica (Amsterdam) 120D. Bliokh. Mazzola. A. F. Ralph. 2010. Rev. and F. J. Kendon. G. Retzker. Lippert. and P. Vidal. M. T. R. Munro. and W. 033403. D. Hartmann. R. K. E. P. Roos. J. Nat. . N. L. Cleve. Rev. 2010. 489. J. 190403. and E. J. A. R. X. Liang. R. Rev. Phys. Osborne. Rev. 12. Jelezko. Maiorino. R. Mod. Casanova. Munro. Angelakis. Phys. Clark. D. Zubairy. Shiga. Brown. Z. Brod. and W. and S. M.. 885. J. S. Vol. W. Rev. 2011. and G. J. A 82. H. Lett. J.. 022324. R. Kitaoka. W. 121. January–March 2014 181 Buluta. H. 1. 022327. 2007. Barreiro.. Dalibard. G. 2008. J. 105.. Sanders. Tanaka. I. R. Rev. 012305. and J. Phys. Dalibard. DiVincenzo. J. Prog. P. M. and G. S. A. P. U. and F. Zwerger. Lett. Rico. K. A 72. Trefzger. C. Lett. Rev. A. 80. Schouten. 74. B. 12. 2008. R. K. G. M. P. Linden. Savel’ev. Itano. 8. Rev. Bergholm.. P. P. Brown. J. Sciarrino. S. J... 109. Shou. Blatt. Phys. 2008. and R. 504004. J. S. M. 2009. and Y. I. Rev. T. Porras. 73. Phys. Phys. Bermudez. Deng. Ashhab. A. A 78. and F. Rev. S. Fisher. Aspuru-Guzik. and M. B. 4. Plenio. K. A. 2011. 2010. A 66. Brennen. M. Rohwedder. 2007. X. M. N. D. B 78. B. 101. 125301. Bravyi. 2012. C. Mooij.. A. Hammerer. Plenio. Sawyer. Rev. 2003.. and Franco Nori: Quantum simulation Banerjee. 657. and J. 2013.. Utsunomiya. Science 326. Phys. L.. London). Mod. Bose. D. C. Phys. Buluta. and A. Schindler. 040402. 062401. H. 246809. C. and J. Rep.. K. 2007. A. C. Cirac. J. 2010. Lybarger. and P. F. Rev. Loss. Gerritsma. L.. P. 090501. D. 052312.. Chwalla. Bouyer. Lupascu. Barrett. M. 2007. Rev. 1031. R. Martin-Delgado. Nat. Rev. Fitzsimons. A. J.acm . G. 2011. 794. Imaging Electron Phys. E. 205.. E 57. 2009. R. Dove. Uys. Roos. and E. and D. and and T. Bloch.. Kendon. and J. and G. Gerritsma. Gross. P. AIP Adv. 11. P. J. Commun. Bollinger. R. Cirac. Z. Roos. K. A 77. M. Fernandez-Vidal. J. M. Ramponi. P. Rev. 1. E. A 79. 2011. J. G. T. L. 070503. and S. Solano. Phys.. 022303. Lett. S. D. R. and P. de Groot. Nigg. H. G. P. A 77.. M. Leung. 2007. M. J. J. Nori.. Lugan. D. Cirac. Phys. R. Phys. E. Uys. Leon. J. 2005. Biercuk. Mod. Martin-Delgado. Egusquiza. J. R. Rev. Capogrosso-Sansone. 032337. Phys. and S. Phys. Zoller. D. 2012. Britton. B. T. Phys. Broome. S. Chuang. Bermudez. 2013. S. 2012. Phys. Aaronson. Porras. K. Zuo. Solano. 920 [http://dl. Paz. Zoller. Josse. Öhberg. Byrnes. A 76. and D. J. Nature (London) 470. Y. Müller. Rev. J. Mataloni. M. Adv. Clement. 2002. 1201. Entropy 12.. J. M. Braun. 9. Solano. Fedrizzi. and P. Vitelli. Semba. 062320. 270. Chem.-L. 1998a. Rev. 2008a. Church. Solano. Clarke. A 75. K. Yamamoto. Doscher.. 2011. Phys. N. M. Bollinger. Chiaverini. Casanova. Monz. Kais. M. Phys.. A. D. de Raedt. T. Sci. Bennett. 686. Pachos. A 79. J. S. F.. Clark. K. Biercuk. Hasegawa. J. New J. Kim. J.. C. Science 339. B. 013617. Phys. Ito. Boghosian. 022328. 13. Scully. Nature (London) 453. D’Ariano... Rev. Biamonte. 2008b. Gerbier. C. 104.. Phys. Quantum computing devices: principles. P. V. P. 968. N. and J. Rahimi-Keshari. V. N. Phys. and R. Phys. Yamamoto. S.. Phys. G. E. L. 30. J. M. 2013. Retzker. 1998b. Zoller. Phys. Leifer. J. S... 101. J. R. Brown. B.. Zoller. K. Osellame. Zhou. De. C. 2011. and M. T. Blatt. A. V. Chen. Sabin. Schmidt-Kaler. 54. C. 2005. F.. Terhal. 144112. M. Spagnolo. Phys. F. Brown. 2009. Keith. J. and S. 2009. Mataloni. 2008. 97. Phys. Watanabe. Soltan-Panahi. Mueller. Whitfield.. Freericks. Phys. Dalmonte. 176. H. 267. Georgescu. Rev. 062338. and P. Rev. Chuang. 2008. Bermudez. A 82. Aguado. Englert. Rev.I. Am. Menzel. Nascimbene. Rep. No. J. J. A.. Phys. Freilikher. K. L. Roos. Porras. M. Guo. Bongs. Michielsen. 020101.. Phys. 2. A 76. J. Garcia-Ripoll. D. Appl. New J.. C. and K. and I. F. B. Juzeliūnas. 128. and M. J. 013114. and J. M. J. T. 2008. T. Byrnes. C. G.. Ashhab. Chen. Lisenfeld. A. L. P. K. 99. J. Y. Zoller. M. M. S. 060303. Crespi. Chiuri. Arnold. 6. 053009. Northup.. Strini. J. Lett. Cho. Hambrecht. 2268. Quantum Inf. P. K. 110. Phys. I. Bernard. Phys. X 1. 2008. Cai. Phys. Phys. Science 301. M. Cirac. Billy. J. Trieu. 86. B.. Galvao. Cucchietti. A 76. Wiese. X. A. 8. Hennrich. Comput. W. 2010. S. Lett. D. 9. Garcia-Ripoll. arXiv: quant-ph/0702003. Kakuyanagi. and M. G. Daskin. Harrison. M.. S. C. Harmans. 763. Verstraete. Mariantoni. Phys. Wang. A. E. Phys. 063407. and A. A. A. C. Phys. S. D. 2011. M. Recher. Martin-Delgado. Zhou. A 81. S. Georgescu. X. 2006. A 42. and M. Brown..-G. Meno. Mod. 2008. S. 108. Nori. Phys. Y. Phys. 2012. Vandevender. J. Almeida. E. Rev. Taylor. Henkel. 2008. 055803. Nat. Z. 207209. S. 190502. Wilhelm.. 2007. 168. A 71. Nature (London) 453. Sacchi. L. Boghosian. 2008. 021018. 062314. and E. H. A 77. I. B. Bouyer. Nature (London) 453.. S. Z. 2007. I. D. and J. M. T. M. Cho. J.. 2007. Phys. Nat. P. M. Berry. and E. Becker. F. Metodi. 545. 107. A. 108. and M. I. D. P.. Büchler. 175302.-L. 2010. 2013. Lamata. P. Sanchez-Palencia. 134. and J. J. G. 022126. Rev. and W.. 2012. 105. Rev. Kusudo. Rev. G.. Kronjager. L. S. C. W. Kim. Popescu. 2010. R.. Phys. and V. Gasster. W. Rev. A. M.. 1008. and W. Prog. Lett. and Y. A.. J. 095007. Greganti. M. Cirac. 176. 2009. 2012. Bremner. Pupillo. Ahokas.. 95. Phys. F. Buluta. Deppe. Sengstock. F. S. Mezzacapo. Maraner. T. 486. 2008. I. M. New J. Solano. Lin. N. Paternostro. I. Nat. and M. 2010.. Chen. Phys. Phys. 2011. Phys. M. D. Rep.-J. T.. M. E. and A. Phys. M. C. 2009. 1523. Nori. de Raedt. Dalibard. Lewenstein. Bermudez. T. 016405. 76. 2008. G. New J. and F. J. and F. Aspect. Nature (London) 484. J. 359. Paris. Rev.. and Y. 075320. 80. F. Phys. Bose. and G.. designs and analysis (Chapman and Hall. B. Stebler. Byrnes. 2007. C. C. K. P. 065025. Plenio.. 2005. Rev. Wineland. Yamamoto. D. 891. Comput. 050504. D. J. Bloch. I. Phys. A 73.

Kassal. Phys. Phys. Rev. Ivanov.. 2005. 2004. Lett. J. Pfeiffer. Nori. Türeci. M. 125008. Johansson. 2011. Peng. 2008... Solano. Cirac. D. M. 18681. Pinczuk. 110.. Christensen. A. Niskanen. 063615.. X 3. S. and J. Liu. A 83.. 387. 2013. Rev. G.. 2010.. G. Phys. B 79. 292. Rev. Islam. Nielsen. A. Wu. J. Plenio. H. I. Lett. Schmidt-Kaler. M. M.. Glückert. Lundgren. and F. Harris. Lin...-X. Comput. 2007. G.. B. Walecka. W. Y. Preda. I. Grajcar. Mod. Phys. J. and P. Lett. D. A. F. Phys. 377. 020507. Rev. 2010. T. Phys. I. P. and C. Fischer. and L. 2003. J. Paganelli. Phys. 2009...-B.. Shih. and R. D.. Yang.. S.. S. and D. Sci. 2009. Smirnov. Greiner. Home. 2010. 223002. A 78. J. Nat. Schützhold. J. Rev. Lett. A. S. A. Phys. H. A 80. Ivanov. and J. and M. B 86. Dür. 2012. Thew. Phys. (Berlin) 14. F. and G. and C. A. S. Briegel. C. Nori. Phys. A. Cazalilla. D. Zoller. 2011. A 70. 2006.. 143902. Kassal.. Jo.. 2013. I. D. A. 2012 (private communication). Gerritsma. I. Korenblit. Deutsch. B. Hempel. 2008. 104. Georgescu. Kirchmair. 2008. F. R. J. Fisher. Leibfried. Zähringer. and Z. L. S. Nat.) 315. M. Wilson. 035301. 2006. Islam. 250405. J.. and M. Mandel. Quantum theory of manyparticle systems (Dover Publications. Edwards. Gaspard. 2001. and A. L. A. Hänggi.. New York). Rev. Bourgain. Rev. Carusotto. J. Chem. Cole. Lee. Phys. 2010. E. Bruder.. Rev. Rev. Lett. Phys. Houck. S. and I. J. Lee. 2011. Smirnov. 244103. A.-J. 81. 043621. W. Sortais. Y. Liu. R. Phys. L. B. R. Phys. R. Rev. J. Nature (London) 415. Phys. F. R. F. and D. S. E. D. Nori. and W. J. 052509. 39. G. 2. Rev. A. Lapan. D. Jaksch.. 467. Rev. E. Galve. 75.. S. Int. Aspuru-Guzik. 2008. 2012. No. Georgescu. A.-X. Marcos. 472. Kapit. H.. 016603. T. Awschalom. Phys. A. M. B 77. 2004. R. P. G. Nature (London) 463.. and C. Ho. J. X. F. P. C. (N. Rev. . 2. Europhys. Nolte. D.. Annu. Polini. 98. J. 60. and J. Lett. J. Fisher. 2012.. J. Guarrera.. Lopez. Giorgi. 546. and M. 033620. and F. Nat. and M. and M.. January–March 2014 Gustavsson. Phys. 085115. Science 325. Reznik. Phys.. Nat. Geim... M. A. 2011a. Choi.5260. B 82. 736. Hou. Science 292. P. 2011. Perdomo-Ortiz. S. A. Lanyon. 241406. Izmalkov. Singha. Pritchett. R. W. Ghosh. 62. Phys. D. Nat. C. L. Shnirman. Rev. Rev. S. M. T. Hamley. Farhi. Harrabi. N. Porras. E. 2011. M. 1989.-J. Science 336. 106. Horstmann. P. Freedman. Keil. H. and D. Dalmonte.. Hastings. Georgeot. 2011. I. Wineland. Vignale. Phys. Heinrich. Rev. S. A.. and A. V. Solano. M. 227. D. and M. H. 2002. 041018. 21. A. Vol. 849. Phys. and W. K. Phys. 2011 (unpublished). T. 98... M. W. B 73. Giovanazzi. et al. B.. G. L. Phys.. 2. Martin-Delgado. Goldstone. Phys. P. J. Singer. 061302. Lett. Phys. M. 2011b. arXiv:1210.Y. A. Herrera. Marchesoni. O. Tahan. 1521. M. Varon. C. P. Du. Proc. E 84.. 133003. Phys. P. Fetter. 84.. Lett. Stancil. and R. P. Schätz. M. Mohseni.-G. 2006.. Blatt. Rev. Nakamura. Zoller. 612. Nori.. 032301. Galiautdinov. Wunderlich. Zoller. Rev. U. Kubasiak. J. Gibbons. S. Phys. Love. and A. Phys. Nori.. 2012. 4.. 86. and L. Martin-Delgado. Mueller. Phys. Madison. Phys. B 40. Phys. Phys. J. Hanson. Howell. K. Science 332. 6. Phys. Bloch. H. 13. Zahringer. 2003. Hauke.. C. Phys.. P. A. Blatt. Vitanov. 3. M. Lu. Yoshihara.. Ashhab. Geller. Feng. Phys. M. R. Nori. Rev. Jordan. N. Rev. Rev. Gardiner. P. Bremner. A 79. Nakatsukasa. R. Phys. Chapman. R. Tagliacozzo. Rep. Fölling. Ashhab. Garcı́ a-Ripoll. and X. 2009. Jaksch. Bermudez. J. D. Rev. 154009. B. 93. D. E. J. Thywissen. 2011. K. Bylander. Phys. S. 185. Friedenauer. J. J. M. Rev. R. Fallani. 134. 757. Nature (London) 453.. M. Kim. Novoselov.. Kim.. Commun. Kay. J. U. V. 4. J. Gerace. A 85. Dür. Martin-Delgado.A. 2010. 2006. Rev. B. Browaeys. M. 063823. and Franco Nori: Quantum simulation Dodd. J. 2012. Pellegrini. M. and T. A. J. Hänggi.-H. 2011.-Y. Natl. 051401. Rev. Hänsch. S. A 61. Jost. N. Ketterle. Schmitz. 2010. 131. Nature (London) 453. Il’ichev. A. 043604.-H. K. M. G. J. 2008. Lett. Prog. 1. New J. West. Angelakis. A. A. AspuruGuzik. 2010. Whitfield. Vidal. E.182 I. Freericks. Mering. 060301. G.. Wang. Ghosh. R. 2005. Hilbert. 3108. J. R. 033625. M.S. 2008. Phys. Phys. L. Phys. 250403. and J. Tünnermann. 856. A. W.. Pritchard. 1043. Weichman. M.. Bremne. Rev. 13. Phys. Fort. Monroe. 2004. and P. A. F. 105. V. I. C. Katsnelson. F. and D. Chem. Martinis. J. 2009. D. Phys. W. Y. Greentree. and M. P. Nat.. W. Y.. 2009. A. 104. Solano. 144505. Wang. Duan. 81. Phys. Hartmann. 52. and F.. Casanova. Feynman. P.-M. Phys. Ann. L. Rev. E. Fleischhauer. 2013. 012304. 082401. 1982. 105. 061138. M. A 82. Phys. Raussendorf.. Marchesoni. 68. C.. Gibertini. M. F. Kitaev. 133002. 587. Grinstein. G. 060412. Hu. Roos. 2002. M. and F. E. and S. Lewenstein. Y. Yung. 2009. B 79. S. A 81. C. M. P. 2002. B. 2011.. 024511. N. and A. Fuhrmanek. Phys. 2009. 130404. Phys. Phys. M.. 024522. Nat. J. M. 2. and J. K. P. E.. Phys. 110. Han. Longhi. 150404. and F.. S. A 81. 106. Hollenberg. Tsai. van der Ploeg. Phys. K. Johansson. 1999. Nori. 2009. Cirac. Roos. Lett. Gutmann. A 69. S. Schon. Phys. 1130. Ann. Nat. 2008. 2005. 060503. 2009. Mod. Greiner. Lett. P. C. P. M. Jordan. and F. and K. and F. R. Phys. Sornborger. Rev. P. P.-M. Ivanov. 2012. V. Rev. 106. Lett. J. Krems. and D. B 42. Rev.. Rev. and H. M. Lett. Xu. Goldman. Rev. E. Harris. and P. A. Dreisow. L. and A. T. Acad. M. M. Rev. Theor. Fedorov. I. 2009. A. Lett. Rev.. 2010. T. 1998. W. Hanneke. Rev. Phys. Herrera. Phys. J. and T. and C. A 79. Rev.. Inguscio.. Lewenstein. Y. C. Johanning. Phys. C. Preskill. Yeaze. Cucchietti. 51. 2010. Chem. V. Chang. Brandao. J. Krems. Lee. et al. J. A 84. J. Gerritsma. and F. S. J. A. and D. D. Berciu. G. H. Esslinger. 620. Jané. and F. R. Lett. Phys. and P. 8. Rev. P. 030502. P. and R. 040301. Rev.-M. Duan. and K. Cirac. 20001. Ian.. Lye. A.. A. H. Commun. Smirnov.. 94. Cirac. Giamarchi. Phys.. and I. and A. Oliver. Math. A. R. R. Fagnocchi. 052118. B 82. C. Rev. 052325. Amini. Kirchmair. M. Hauke. and C.-R. D. 2007. E. and E. and J. Szameit. and J. F. Garcia-Ripoll. 2009. Quantum Inf. R. S. O. A 65. J. Zoller. Rev. T. 035102. A. and J.. J. S. 15. Koch. A. A. Meyer. 2008. Rev. E. Ghosh. Garcia-Ripoll. and F. S. 103. F.

Hill. and C. Phys. Rev. P. J. Rev. 2012. and D.Y. and W. A. 2010. Nat. Gong. Zhou. L. D. Science 298. Schätz. Zhu. 230402. Laflamme.. A. F. D.. and O. Chen. Walther. and D. E. S. Y. Manin. W. L. E 59. 2007. Science 325. J. Meyer. X. Rev. N. Rev. Rev. Lemeshko. O. New J. Ortiz. K. McKague. 2013.S. Natl. Sanpera. New J. Thomas. Chan. 243. M.-S. 2010. 2006. Yung. 1996. Marzuoli. and R. 548. 1998. 4. Sen(De). Gong. and P. X. Lidar. K. D. A.-L. J. and I. 247901. Walther. Sun. Zoller. J. Adams. X. 224502. Phys. Rev. M. J. A 64. N. and H. Du. K. S. N. Wang. Adv. B. K. Chen. Phys. and P. 2. Lidar. Phys. Front. G.. Vol. 105003.. Dakic. Science 319. 277. V. Phys. 2012. Jelezko. 2002. Nation. Phys. Phys. Mostame. 101. E 70. Hemmer. Rev.. Suter. Widera. Dodd. J. 2012. 2002. A. Li. K.. Lamata. G. E 56. and J.. 126. 2002. M. arXiv:1205. Du. Phys. S. Aspuru-Guzik. Marcos. Islam. (N. Li. Korenblit. 589. Zhou. Li. Lau.. P. P. Ma. 3. C. Phys. Phil. Phys. and R. Knill. S. 1420. Lett. . Maruyama. Maniscalco. Y. 86. Phys. 022303. Naylor. 107. 2010. Rodriguez-Lara. Liu. 042304. Petruccione. 95. 2010. 309. Cirac. Laflamme. Proc. 2012. 2013. 110504. Colombe. 1023. 468. Cirac. 4. 11. 399. 722. Phys. and V. J. Chem. Peng. New J. Z. Xu. 2001. 341. R. C. January–March 2014 183 Maniscalco. B. Lett. 14. D. Wolf. I. 7. 253005. 6. and S. F. K.. J.. 98. B 88. T. Lin. Rev. Nori. Duan. Phys. L. Kraus. Peng. M. 095019. and F. Lidar. K.. Solano. V. 2010.. D. Mezzacapo. Robertson. C. J. T. 1367. 2013. 1107. L. 2004. A 74. M. Chen. Gisin. Leibfried. Du. Liu. Houck. Somma. J. R. et al.. M. Phys. Rev. Ballester. Kropatsche. Rev.. 087001. and N. Xu.. Amini. Phys. A. 2013. and U. A. 2004. 1. M. M. Gubernatis. D. M. S. T. D. D. B. X. C. Lu. X. Trans. V. 302. 2005. 2. Xu. Nation. Rimberg. R. L. 2011. Lewenstein. Phys. G. Lett. A. Nat. Phys. Manousakis.. Laflamme. Chen. 2013. Rev. and P. Lett. Weiss. J. Rev. Nori. Sci. Kitaev. S..-H. Goldman. Sun. and S. K. Phys. Ann. 83. V. G. Acad. 4234.. 2009.. M. Leonhardt. Levi.. et al. 1. Europhys.. Ortner. and L. Phys.. Edwards.. Candia. and D. 15. E. and I. Commun. Low Temp. P. 45.. Kimble. Lin. 015007.. 1. 2009. S. 11. 2007. 2008. Y. A. 81. L. Science 273. Lloyd. and D. Phys. M. A 71. Hakonen. 2009. James. and S.-W. and B. and J. 032303. 2429. Pan. Brennen... Sun. A. B 88. Duan.-M. S. S. 2012. Lett. A.. Micheli. L.. Zoller. Childs. 2010. Lin. Phys.. F. Pachos. S. A 360.. 89. 2010... 1073. N. A. 9411. 1999. Granade. Y. O. L. 2005.. 102. 014426. 2. Y. X. Lett. F. Cirac. Lanyon. J. Monroe. Solano. F. C.. L. M. Science 334. 56. Rev. Phys. Dakic. 395. Lu. Ladd.. J. V. Koch. 2004. Whitfield. S.. and E. Soc. Orenstein. and P. Phys. 220501... and P. 2002. Zoller. A... Kuklewicz.-D. Phys. Yao. Hur. Science 325. Tian. Y. K. S. Mod. Commun. Lewenstein. Nature (London) 465. and F. Mandel. G. Mei. T. M. New J. H. Rasetti. J. Milburn. Quantum Inf. Nielsen. R. and C. Sci. et al.. 14. Kitagawa. and J. et al. D. 2009. Z. B. Ahufinger. 3661. Duan. Konig. and F. Ashhab. J. 020505.. Nakamura. A 57.. Neeley. Buks. Rev. A. Z. China 5. Lett. J. Suter. Phys. J. M. Z. Laflamme... 2011. Lamata. G. 2010.. A. G. Phys. and C. 2007.-Y. 043811. D. Jiménez-Garcia. 2008. E. J.. Pedernales. 2006. A 82.I. Suter.. D. Cambridge.. Peng. M. Moscow) (in Russian). Science 334. I. Leibfried. Phys. A. Phys. M. Phys. Nature (London) 429. Piilo. P. 2011. A. Zeilinger. L. S. Stojanovic. M. 12. C.. and R. I.. Gong. Blencowe. Nelson. 2012. T. 2002. Spielman. Phys. S. H. Dawson. and P. S. and E. Lewenstein. and M. Micheli. and R. Zhu. 103.. Milburn. Wei. F. N. New J. 106. Johansson. Lu. Phys. Y. Murg. A. 2005. D. and J. K. Piilo. 2004. V. Rabl. Nature (London) 471.. S. Shlyapnikov. M... B. You. E. Rev. Phys. 14. Rev. Zhang. J. Georgescu. Chuang. C. 590. 2006. 055045. Islam. Vedral. Hänsch. W. Lett. 111. P. Peng. R. E. 57. Mod. and U. F. G. G. Mosca. Phys.. Phys. E. M. and J.. Nature (London) 464. and J. 61. 60004. et al.. P.. P. G. Liao. Lütkenhaus. H. and G.. 181. Phys. Brown. Gorshkov. Phys.-B. R. F. B. I. Georgeot. E. A 82. Nat. arXiv:quant-ph/9908037. 062329. No. New J. S. Lin. Phys.. Rev. 2011. C. H. G. Liu. Kim.. Solano. Phys. Lu. 110. Rev. 200501. Nori. Rev. et al. Monroe. Girvin. Casanova.. 14. and D. Naylor. Mod. S. Mariantoni.A. Rev. Paraoanu.. F. A. Bloch. Monroe. Phys. J. W. 2002.. Zeilinger. Fölling. J. 1997. and I.. X. A 81. Z. 2011.. 167. 2009. F. 087004. Sen. A 85.. 2000. M. Ortiz. B. England). 2011. 2011. B. 020501. E. A 75. Phys. Angelakis. Phys. 2002. B. Science 288. and D. 1999. A. Phys.-Q. 2006. T. 146. Phys. Bennett. Jiang. J. A 66. 2003. 2009. 2000. Nori. Phys.2801. A 73. 056218. Weimer. 1980. 2012. 13. and A. et al. Rev. A 71. Nielsen. A 70. S. 2007. D. Gubernatis. 1. 1501. 2. Phys. 109. J. C. Process. 2009. 095024. New J. and M. D. Chem. D. Millis. Rico. 103. Pupillo. and E. Olson. R. Phys. M. Blencowe. 055008.. Nat.. O’Brien. A. 84. 030502. X. D. D. Rev. Xu. Rev. Phys. Zoller. New J. Knill.-M. Y. Knill. Rep. K.-D. and L. Rev. Q. Montangero. 88. Messina. Lett. Rev. 022310. E. Biham. A 306. Rev. M. C.. L. 012307. Rev. X. et al.) 303. R.. Comput.. Paredes. Hassel. G. Negrevergne. M. X. and Franco Nori: Quantum simulation Kay. León. Lähteenmäki. J. Q. Phys. J. Phys.-M. and A. Wineland. and E. Gehm.. 689. J. Chang. Schützhold. and J. Korenblit. Menicucci. 2004. Nori. 79. G. R. 556. Rizzi. 2008. Rev. Gühne... Bermudez. J.-B. Intravaia. M. Li. J. 140501. Ma. B. Nat. D. Lett. Monroe. Gao. Rev. H.-X. Lanyon. Peng. Rev. 2011. M. 033028. Lett. 032344. Nori.. 053605. Momose. P. Ortiz. 1.-K. Lett. MartinDelgado. A. Rev. Rev. Peng. 86. 022317.-D. C. Xu. 2008. Damski. Gopalakrishnan. and D. 2011. P.. Freericks. S. F. Kim.. Computable and Uncomputable (Sovetskoye Radio Press.. Rev. Phys. A. 2009. Ospelkaus. 2009. 2179. Duan. J. 032311. R. J. Nat. Phys. Nature (London) 476. Mazza. M. Phys. Du. M... O’Hara. R. Liu.. U. Nature (London) 453. Siddiqi. N. 2013. 052101. J. Phys. 2012. Rev. S. Quantum computation and quantum information (Cambridge University Press.. U. A. M. Philbin. Warring. Chem. Noh. 102. Nat. X. 022319. S. H. Du. A 69. 106. Lett. Phys. and P. Lett. and L. Phys. Bremner. 2012.

2006. Cirac. Weitenberg. Troiani. Soc. Knill. Phys. Rev. E. R. 92. 2011. Phys. Bruder. Cirac. E 76. I. Porras. Barrett. L. Nori. S. L.. T. Vollbrecht. and M. Fort. and T. Eng. Wesenberg. L. Galitski. Rev. 2011.. 2013. Porras. Phys. 2008. Szameit.Inguscio. 2009. 87. T. and W. J.. M. 050101. Rev. P. and D. D. Vol. Leibfried. 022312.. 230502. D. G. 2006b. 95. C. Phys. Wesenberg. Bakr. D. 2007. Segal.. New J. S. K. I. Troyer. Schmied. New J. L. Boca Raton. 94. Verstraete. and J. 798. G. Zagoskin. 250501. C. V. N. M. B. A. P.. Phys. P. Photonics 7. 195302. J. C. and J. B. E. A 65. Somma. Lett. T. B. Rev. K. 1. M. 2012. Georgescu. S. I. Schätz. 50. H. Modugno. 2011. 13. and D. Lewenstein. Phys. A 68. T. C. Quantum Inf. 250501. Lett. S. G. Phys. Schneider. Schmitz. 2008. D. Phys. A 78. and P. 75.. 101. Cold Molecules: Theory. Zoller. Poulin. 1999.. M. Friedenauer. R.. Phys. 260504. 99. 2008. Thouless. D.. D’Errico. Cirac. S. Orland. Lett. et al. Phys.. 2010. Prog. T. Porras. Rev. A 78. R. 2006. Roscilde. R.. J. A. and Y. M. and J.. F. Targat. Phys. Science 332.. Yamamoto. 2009. 1998. Phys. Semiao. D. 185. J. S. Rev. DiVincenzo. Leibfried. Schneider. Sherson. Nature (London) 467. J. E. Kuhr. SPIE Int. Science 333. and D.. M. 92. Shi. Quantum Inf. Eur. and M. A. H.) 330. Cory. New York). Laflamme. and H. and K. 1972. and J. J. Arlt. Rev. Quantum Monte Carlo Methods in Condensed Matter Physics (World Scientific. M. and S. T. M. 170201. 2004. Plenio. Phys. 93. Mourokh. 96. M. A. Phys. 201301. J. S. V. Kahra. 2010. L. Preiss. and M. Rev. A.. Nat. and D. Phys. (N. Murg. N. Rev. 5833. Walther. R. G. V. Laflamme. Porras. 2008.. Schmitz. Poulin. Rev. 106. R. Stannigel. 2010. Singha. Dakić. P. 2013. Lewenstein. and F. S. Tewari. and J. A. 2012. Rev. Rev. 2007. Mod. Sanpera. Dalmonte. 180402. A. Schmied. Vaishnav. L. Endres. 010101. P. E. C. I. Demarco. Schützhold. 2010. 125302. Gubernatis. I.. D. 1. and F. 11. W. C. Rep. Lett. 9. 2002. 2008. Regal. Porras.. Thompson. Somma. L. R. Europhys. 895. 2010. Celi. A. Ma. and B. 160. R. 36. Greiner. D. Zaccanti.. 96. A. Rev... 996.. Briegel. 2011.. Ruschhaupt. 170411. 040401. A. 14. 1993. P. R. 170501. Singapore). Hauke. Tai. J.. Phys.. Terraneo. Nature (London) 451. Lett. F. Quan. Stolze. Fan. D. Found. N. 2005. and M.. V. J. Zakrzewski. J. 664. Ashhab... Phys. 2317. Schmied. B 79. A. M. Savel’ev. Schätz. Itano. Cirac.. Raussendorf. Büchler. J. H. Y. Capasso.. Soltan-Panahi. No. A. J. Eckardt. and F. Rakhmanov. Kruse.. L. 040403. Rev. J. Porras.. and J. Yampol’skii. M. Porras. Phys. Spreeuw. 2013. H. Opt. Friedenauer. Timoney.. Simon. Lett. A. Tseng. M. 042323. W. A. Nori. Taylor. 1634. Ölschläger. Plenio. J. L. I.. 2003. Cirac. D. 022301. and D. P. 109... 10. Kantian. A. von Delft. A. Lewenstein. 127. Bloch. Phys. Weinheim). Rep. Muga. and F. Roumpos. P. and M. Petersen. 104. 248. J. J. C.. Rev. 2005b. Struck. 2004. Pupillo. Rev. 2009. Cirac. G. Phys. Zakrzewski. 82. R. Opt. Fattori. A. Soc. New J. P. Lett. Proc. B.. Sanchez-Palencia. G. and G. 233002. Rusin. 2013.. D. Jin. Schützhold. 88. Scarola. R. and M. Wunderlich. Havel. and D. 102. 2004a.. M. Rakhmanov. A. New J. B.0528. C. Troyer. T. and C. 144507. Shevchenko. B. Ann. 075101. and P. and R. B 81. A. N. 2003. Schätz. Varon. 86. 2011. and S. Lett. J.. and M. . Verstraete. Rev. D. Spring. Suzuki. Zoller. Phys. Marquardt. Mod. 540. Retzker. Terhal. Sun. J. J. C.. Lett. Szpak. Phys. Ortiz. Fortschr. FL). 2008. Science 339. Schützhold. S. D. Schaetz. Leibfried. 80. Nori. 171. Phys. Weinheim). F. Langer. Modugno. Stojanović. Tillmann. and R. and F. S. and M. Girvin. Phys. and S. and S. Porras. Commun. Porras. I... 2010. Lett. Cheneau. M. R. 2008. 094415. J. R. Torrontegui. Georgeot. T. 307. Liu. Phys. The Quantum Mechanics of Many-Body Systems (Academic Press. X. J. 4. Phys. R. and R. Hafezi. Phys. I. Picon. Nori. 82. and D. and H. Baumgart. JETP Lett. M. 93.. Phys. 97. Laser Phys. SPIE Int.. Raeisi. Sarma. Lewenstein. T. 125031. Suter. Eng. Paternostro. Somma. and S. Marcos. Applications (CRC Press. D. 2002.. 045017. Lett. Y. 97. Ashhab. V. A 79.. M. L. and T. Zamora. 103017. M. X. and U. J. and T. L. Wineland. Eds. M. and G. Lett. H. 2005. Rev. Ortiz. Lett. Sun. 2012. 2007. C. Strini. Quan. 2006a.. 2005a. Tagliacozzo. 207901. Sacha. 2001. Rev. and J. Diehl. A.. Smirnov. Phys. H. 979. Osborne. Chiaverini.. Y. Wiebe.. 492.. 2010. New J. Lett. Gury-Odelin. D. Somaroo. Gubernatis. 2004. 263602. Lett. S. Ed. Sinha. F. Browne.. Knill. S. Johanning. J. S. Appl. Micheli. W. 053639. Phys. S.. Rev. Nolte. A. Russer. C. Tagliacozzo. Nature (London) 453. Retzker. J. Schack.. Liu. and S. Phys. J. 024401. Rev. D. M. Nori. Walther. Whitlock. and F. and J. Milburn. Quantum Computing (Wiley-VCH. G. Stanescu. and M.. 2011. Clark. Elements of Quantum Information (Wiley-VCH. Schmied. M. and F. F. C. P. A 57. Rev. K. H. Cirac. Roati. Wang. et al. Lett. January–March 2014 Schulte. Condensed Matter Physics with Cold Polar Molecules. 2007. 062331. M. 67008.. 385. M. Phys. Petersen. Rev. T. R. 54.. R.. K. 2012. R. F.. 031105. D. 2011.-J. L. Greiner. 127. B. M. Calarco. M. Phys. Nature (London) 476. Phys. Master. G. Rev. P. Qarry. Nature (London) 472. A. D. 68. Zawadzki. 2011. and P. Phys. A 84.184 I. A. M. J. 115011. Temme. Fallani. 15. 12. Savel’ev. Lett. Britton.Y. and Franco Nori: Quantum simulation Pollet. 2000. 14. 5381.. R. Carretta. W. 1176. J. Sarma. P. H. C. Amoretti. T. Phys. 2008.. Heilmann. B 75. 2013. Santini. Uhlmann. 103029. C. W. Experiment. and J. arXiv:1308. Rev. 2007. I. Leibfried. 2615. Cirac. B 77. and F. 5105. Process. G. Phys. C. Phys. Schoelkopf. Rev. Nori. D. Windpassinger. Sengstock. and D. D 22. Wiese. Nature (London) 471. Phys. 67. Phys. 2012.. and P. Proc. Lett. Senthil. Cirac.. Rev. Nat. arXiv:quant-ph/0112042. 2011. 2011. Drenkelforth. Schleich. 465. 200401. Rev. D. Phys.. Celi.. and J. A 61. E. F. T. Szpak. 2008. Opt.. Process.. Jost. M. Phys. E. and S. B 44. Phys. D. 2005.. S. H. 2004b. Rev. Schützhold. Sanders. 107. Rev. Ertmer. M. Shepelyansky. D 82. Rev. and J. Phys. G.. D.. S. Mostame. Büchler. 035001. Lewenstein. L. D. 2003.

Liu. Beth. 2013. Lett. Schauss. Phys. F. P. H. W. X. 885. Wei. G. 4947. Cirac. Wineland. Zhang. X. Cirac. Messer. Phys. Lett. 2000. Y. Chem. Rev. 100501. P. 2007. Rev.. J. C. C. Wang. 343. L.-Y. C. H. 2009. and J. 185307. A 82. S. M.. Phys. 2002. Tsokomos. Phys. 20007. and F. You. T. I. H. and J. T.1602.. G 40... Yamamoto. Phys.. A 44. Huang. C. Sanders.. Phys. You. N. and J. Duty. Xue. Dong. 10. Pourkabirian. A. and B. Rev.. J. Phys. R. 110501. 2010. Lukin. Lett. You. Nori. G. Chem. Rev.. M. Lett. X. 2008. D. 42. Rev. M. Y. C. and T. Phys. D. Nori. 1. 877. Aspuru-Guzik. and J. Feng. Verstraete. Bloch. Zwierlein. Phys. Phys. Wang. U. Havel. Zoller.. Lett. 10. and J. H. Aspuru-Guzik. I. Superlattices Microstruct. Wu. 042305. M. Fortschr. Nori. Phys.. 98. D. D.. J. Y. D. H. M. 11. 110.. T76. M. R. and R. and D. 885. and J. I. arXiv:quant-ph/9702050. Comput. Rev. No. H. and J. Delsing.. E. Ser. 319. D. and A. New York). arXiv:quant-ph/9603028. 104516. and T. Wallraff. 067601. 2013. 2010. Ashhab. Laflamme. Wiesner. Rev. Du. Phys. Ashhab.-X. 227205. Latorre. 2005. Nori. and J. I. Phys. D. Baugh. Ashhab. W. Girvin. Zagoskin. M.. and F. Cheneau. 2006. Zhang. Wang. 2011. 32. Zhang. Fukuhara. R. 162. Nori.. M. New J. Ott. Stancil. Rev. 1998. S. J. Wang. 2009. 3. Kais. Phys. Whitfield. Yang. Chem. . Gong. Nori. I. Uehlinger. A 82. S. G. 2010. and J. J. Phys. Phys. T. Itano. Weinstein. Lett. Phys. Zohar. Ketterle.. and X. You. and W. 147.. E. M. Nori. A 87. Phys. S. Porras. A. Jensen. 2008.. 052311. 064501. Rev. A 79. Phys.. Majer. 20. S. 2006. Verstraete. Unruh. Greif. Reznik. 2002. T. 2009. H. 5388. 2013. Cucchietti. 133. Rev. M. Rev. Phys. Phys. Q. New York). 2013. A 65. Rev. Phys. L. 2009. Weitenberg. 2002b. Murg. 1997. J. Liu. D. Nature (London) 435. King. 125304. Y. M. Janzing. 064509. A 454. F. 106. Nature (London) 431. 010501. C. 2004. Rev. 113020. Cirac. and S. Phys. 120503. Z. 2003. Yung.. Laflamme. Rev. and A. Nat. 57. Quantum Theory of Many-Body Systems: Techniques and Applications (Springer. 130. J.. C. 2009. Ashhab. Rev. Yao.. I. E. 042303. 2008. Galve. Nat. 109. Emerson. 2007. 2010. Lett. 143. J. J. I. Cirac. 76. F. Phys. A. 2009. 014505. R. Ashhab. 1996. S. Beth. Cirac. J. R. Hubbard. S. A 65. 304. Sun. H. Phys. 2008.. A 352. 032341. Nori. S. Wang. N. A 79. J. and A. Savel’ev. C. W. Adv. 2002. I. Nori. A 80.. Lett. 012305. Langen. Kuhr. 012016. 46. F. J. T. Nagaj.-H. J. 589. Lett. Koglbauer. Z. J. and M. Lett. Proc. T. and X. T. Peng. Phys. Endres. and H. 033602. Lloyd. Rev. Lesanovsky. Yamaguchi. 032309. 2008. Whitfield. I. J. Phys. Hofstetter. Wang. No. G. Yung. Kohler. 2011. S. Laflamme. C. D.. S. B 81. W. Rev. J. Phys. Cory. Process. and A. 057904. Weimer.. I.. Janzing.-X. 2013. 060302. Johansson. J. 2009. R. Rev.. 110. Tsomokos. 2011. J. 117. Wocjan.. P. 2004.. N. Y. P. S. A. Zalka. 445308. Zanardi. 1998b. Rev. Cory. D. M. Phys. Gamble. P. Phys. 2011. 93. G. Wang. 2010. 053101. 313. 89. Hu. F. Frunzio. 1996. Nature (London) 474. and T. Schuster. 042335. Berry. arXiv:cond-mat/0407066v1. D. Müller. Esslinger. A 82.. Phys. H. 032327. Xu. and H. Nori. B 68. W. H. 157902. A. F. Zohar. 382. D. I. Conf. Vol. D. and F. Rajendran. Phys..-A. Rötteler. and Franco Nori: Quantum simulation Tseng. Hänggi.. Phys. and P. and M. Shi.. 2002. Verstraete. 93. Chem. C. Ryan. D. and R. Volovik.. Hoffmann. Q. Lawrence. and F. F. N.-A.. Phys. I. J. Wu. 052324.. 2008. Beth. C. D. M. J. Aspuru-Guzik. Sun. van Oudenaarden. Y. Kumar. and F. 2008. F. 6. Yang. 2002. Ditty. W. Phys. Wang.. and B. Y. E. Liu.. 062317. and B. H. Vollbrecht.. 2010. A 82. G. A. Rev. Q. B. D. and I. Blais. Zhou. Lett. Rev. M. Z. Wood. J. Q. Rev. S. Lett. Geller. Soc. 89. Rev. Zalka. A. 376. Zinner. M. Phys. M. and C. S. N.-X. Janzing. Zhang. 062303. Myatt. Phys. R. 032316. S. and F. Sherson. Rev. 2013. Phys. Nori.-J. Ward. and D. Wocjan. P. J. Gericke. A. 2009. Abo-Shaeer. 2012. B 75. Lett. H.. M.. S. and F. Rev. 2. 2010. K. Xue. and P. E. X.. 2004. and D. C. P. F. R. Quantum Inf. I. 2008. arXiv:1305. Sun. 2008. Schirotzek. 1998. 103. and K.-S. Würtz. 10. 10. V. Europhys. Aspuru-Guzik. Wiebe. 042309. Nature (London) 471. arXiv:1108. and P. Mooij.. You. Rev. 021302.. 2011. Verstraete. Lett. and F.. J. Weimer.. K. A. Cirac.. Biamonte. and Y. J. Phys. L. P. Q. Rev. Phys. Lett. Sharf. Process. H. H. Quantum Inf. Mod. Commun. Kassal. G. 2011. 880. Rötteler. W. Rev. A. Laforest. Zhou. Jotzu. H. Phys. You. Quantum Inf.. Chandrashekar. 1047.. A 78.. Rev. Büchler. Wocjan. Comput. 111. Reznik. Rev. M. and F. W. Phys. Byrd. and P. 063827. Liu. 100. P.. Büchler. Rev. 422. Cirac. M. Q. and F. X.. A 79.. Rev. Wilson. Lett. Nori.. P. H. V. and D. 2002a. Vidal. A 82. Aspuru-Guzik. M. A... 2009. 100. N. I. January–March 2014 185 Witthaut. Jensen. Nori. Nori. U. L. You. 101.. C. Monroe. Wang. and A. K. and G. and T. K. Vollbrecht.. Lett. and F. 2011. 2011. Phys. Hoyer. Zueco. Somaroo. 2.. Tedford. Rev. 86. 2005. Bissbort. Penrice. 080404.. C. Zinner. J. S. Phys. Georgescu.I. Weinfurtner. J. D. and F. M. Schoelkopf. L. The Universe in a Helium Droplet (Oxford University Press. A 83. X. F. Rev. 194105. J. 101. 2011. Phys. S. Rev. Wang. Phys.. C. Johansson. 125302. Yang.. 2012. Scr. Zagoskin. A. 100501. A 79. Suter. A. T. M.. A. Cirac. Wiese. Liu. 2011. Lett. Phys. 1998a. Quantum Inf.. M. and F.. Leibfried. Lidar. Müller.-F... Today 58. and F. A 62. Phys. G. Wunderlich. F. 111. Schunck. A 79. C. Rev. D. E. Weimer. T. Phys. Phys. P. Nori.3712. Nature (London) 479. Knill. Lesanovsky.