Synthetic dimensions let physicists use spin states, frequencies, momentum modes, and driven quantum dynamics as extra coordinates, creating experimental worlds whose geometry cannot be built in ordinary three-dimensional space.
Every experiment occupies ordinary space. An atom can move along three independent spatial directions, and no laboratory can simply place a fourth perpendicular ruler beside the first three. Yet many of the most interesting models in mathematical physics are not limited to three dimensions. Higher-dimensional quantum Hall systems, unusual topological phases, exotic localization transitions, and gauge theories often become richer in four or more dimensions, even though such spaces appear impossible to realize directly.
Quantum mechanics offers a way around this limitation because a dimension is not defined only by what a ruler measures. In a lattice model, what matters is that states can be arranged into an ordered set and connected by controlled transitions. If an atom possesses several internal states, or a resonator supports many frequencies, those states can be made to obey the same hopping rules as particles occupying neighboring sites in physical space. The state label then acts as a coordinate. This is the central idea of a synthetic dimension, a technique now used across ultracold atoms, Rydberg systems, photonics, superconducting circuits, momentum-space lattices, and driven quantum systems.
Synthetic dimensions are not evidence that hidden macroscopic directions of space have been discovered. An atom moving from one synthetic site to another has usually changed its spin, frequency, momentum, or another internal quantum number rather than physically leaving the three-dimensional laboratory. What makes the method powerful is that the Hamiltonian can be engineered so that these transitions are mathematically indistinguishable from motion through an additional lattice direction. Dimensionality, in that sense, becomes an experimentally adjustable property of the dynamics rather than an immutable property of the apparatus.
The basic mathematics is almost deceptively simple. Consider a particle on a one-dimensional lattice whose sites are labeled by an integer $n$. If it can tunnel only between neighboring sites, a minimal tight-binding Hamiltonian is
$$ H=-J\sum_n(|n+1\rangle\langle n|+|n\rangle\langle n+1|) $$
Here $J$ sets the hopping strength. The equation does not actually care whether $n$ labels positions separated by nanometers, atomic spin levels, photon frequencies, or some other collection of distinguishable quantum states. It cares about the connectivity of the states. If state $n$ couples to states $n-1$ and $n+1$ with the appropriate amplitudes, the resulting dynamics possess the geometry of a line.
Suppose an atom has several internal levels $|m\rangle$. Carefully chosen laser or microwave fields can coherently drive transitions from $|m\rangle$ to $|m+1\rangle$. The atom may remain nearly stationary in real space while its wavefunction spreads through the internal states exactly as a particle wavepacket would spread through lattice sites. The internal quantum number has effectively become a coordinate. This strategy has been demonstrated using hyperfine states, metastable clock states, Rydberg levels, momentum states, and other atomic degrees of freedom.
The analogy becomes more powerful when real and synthetic coordinates are combined. A gas that can physically move along $x$ while its spin states form a synthetic direction $m$ behaves as a particle living on a two-dimensional strip $(x,m)$, even if the physical experiment is essentially one-dimensional. In 2015, landmark experiments used this construction to realize synthetic Hall ribbons with ultracold atoms and directly observe chiral edge behavior associated with quantum Hall physics. Neutral atoms, which do not ordinarily undergo the electronic Lorentz force responsible for the conventional Hall effect, were made to behave as if they lived in a magnetic two-dimensional lattice.
Creating the extra direction is only the beginning. The phase of the hopping can also be engineered. A general transition between neighboring synthetic sites can carry a complex phase,
$$ H_{\mathrm{hop}}=-\sum_n J_ne^{i\phi_n}|n+1\rangle\langle n|+\mathrm{h.c.} $$
In ordinary quantum mechanics, complex hopping phases are deeply connected to magnetic fields. When a charged particle travels around a closed path in a magnetic field, its wavefunction accumulates an Aharonov-Bohm phase related to the magnetic flux enclosed by the loop. A synthetic lattice can reproduce the same effect without requiring the particle to physically circle a real magnetic flux tube.
If the hopping phases around a closed synthetic plaquette add to a nonzero value,
$$ \Phi=\sum_{\mathrm{loop}}\phi_{ij} $$
then $\Phi$ behaves as an artificial gauge flux. The phase can be changed by adjusting laser or microwave phases, allowing the effective magnetic field itself to be programmed. This controllability is one reason synthetic dimensions became closely tied to topological quantum simulation: the topology does not have to be discovered in a naturally occurring crystal first. It can be constructed directly from the couplings.
The phrase “artificial magnetic field” should be interpreted carefully. A neutral atom in these experiments may not experience an ordinary electromagnetic magnetic field corresponding to the synthetic flux. Instead, its quantum amplitudes acquire the same relative phases that a charged particle would acquire in a magnetic field. The resulting interference, band structure, and edge dynamics can therefore reproduce Hall-type physics even though the microscopic mechanism is entirely different.
Rydberg atoms have pushed this idea into a regime where the synthetic particles also interact strongly. In a 2024 Nature Communications experiment, researchers trapped potassium atoms in optical tweezers and used four Rydberg levels as the sites of a synthetic lattice. Four phase-stable microwave fields connected those levels into a diamond-shaped plaquette, and the microwave phases controlled a tunable $U(1)$ gauge flux through that synthetic loop. A single atom could therefore undergo interference around a lattice that existed entirely within its own internal Rydberg-state structure.
The more important step came when several atoms were introduced. Rydberg atoms possess strong dipole interactions, so their motion through the synthetic lattice was no longer independent. The experiment observed substantial changes in population dynamics as the interaction strength increased, including strong suppression of motion in the most interacting regime. This matters because much of the early synthetic-dimension literature focused on effectively noninteracting particles, whereas many of the hardest problems in condensed matter and gauge theory arise precisely when topology and strong interactions coexist.
The combination creates an unusual geometry. Two atoms can be separated at different positions in ordinary space while each also possesses a coordinate inside an internal synthetic lattice. Their interaction depends on their physical separation and internal states simultaneously. The resulting many-body problem does not simply reproduce an ordinary higher-dimensional crystal, because interactions along synthetic coordinates can acquire structures that would be difficult or impossible to engineer with conventional solids. Synthetic dimensions are therefore becoming useful not merely for reproducing familiar higher-dimensional models, but for constructing Hamiltonians with forms that have no simple material counterpart.
The most dramatic application so far is the experimental construction of four-dimensional quantum Hall physics. The ordinary quantum Hall effect is intrinsically associated with two-dimensional motion. Electrons moving through a plane under a strong magnetic field develop topologically quantized Hall responses governed by a first Chern number. Mathematically, however, quantum Hall physics has a higher-dimensional hierarchy, and four dimensions support a distinct topological invariant called the second Chern number.
The second Chern number can be represented schematically as
$$ C_2=\frac{1}{8\pi^2}\int F\wedge F $$
where $F$ is the Berry-curvature two-form over the relevant four-dimensional parameter space. In two dimensions, topology links an applied force to a transverse Hall current. In four dimensions, the response becomes inherently nonlinear and can involve fields or perturbations acting in different planes. The topology is no longer encoded by the first Chern number alone.
Earlier experiments had accessed aspects of four-dimensional Hall topology indirectly through lower-dimensional topological pumps and synthetic parameter spaces. In 2024, a team working with ultracold dysprosium atoms reported a system whose effective dynamics were genuinely four-dimensional in the Hamiltonian: two dimensions were ordinary spatial coordinates, while two additional coordinates were encoded using the atoms' large electronic spin. The experiment measured a quantized nonlinear response associated with the second Chern number, observed unusual three-dimensional hyperedge modes, and excited nonplanar cyclotron motion characteristic of the effective four-dimensional system.
The atoms did not enter a literal fourth direction of the room. Their quantum states evolved according to equations whose independent coordinates formed a four-dimensional space. That distinction allows researchers to investigate predictions that would otherwise remain purely mathematical. The experiment turns higher-dimensional topology from an abstraction into something that can be prepared, perturbed, and measured with laboratory instruments.
Four-dimensional Hall physics also changes the idea of a boundary. A two-dimensional quantum Hall system has one-dimensional edges, and those edges can carry chiral modes that travel in only one direction. A four-dimensional bulk instead has a three-dimensional boundary. The 2024 atomic experiment observed what the authors describe as anisotropic hyperedge modes, with ballistic behavior along one orientation and insulating behavior along others. These are not merely enlarged copies of familiar two-dimensional edge channels; their structure reflects the higher-dimensional topology of the bulk.
This illustrates one of the main scientific uses of synthetic dimensions. Higher-dimensional theories often contain qualitative phenomena that cannot simply be inferred by drawing a three-dimensional analogy. Topological invariants change, boundary dimensionality changes, response laws change, and the possible singularities of gauge fields change. A synthetic platform allows those mathematical differences to generate actual experimental signals instead of remaining extensions written only on paper.
The same approach has already connected quantum simulation to more exotic objects. Previous experiments have measured second Chern numbers associated with a quantum-simulated non-Abelian Yang monopole, while other synthetic systems have investigated tensor monopoles and higher-dimensional topological structures. Synthetic dimensions are valuable here because parameters that would normally be abstract coordinates in a Hamiltonian can be turned into independently controllable experimental knobs.
Synthetic dimensions are also revealing physics that has little to do with topology. In 2025, an ultracold-atom experiment used periodic driving to construct an effective four-dimensional Anderson localization problem. Anderson localization describes how disorder can halt wave transport through interference. In three dimensions, increasing the strength of hopping relative to disorder can drive a transition between localized and diffusive states. The nature of this transition depends strongly on dimensionality, making four dimensions theoretically important but inaccessible to ordinary bulk materials.
The experiment used a kicked atomic gas whose motion was physically one-dimensional, but the kick strength was modulated with additional incommensurate frequencies. Those independent drive phases act mathematically like additional coordinates, increasing the effective dimensionality of the disorder problem. By designing the modulation appropriately, the researchers realized a four-dimensional version of the Anderson transition and measured its critical scaling. They observed scale-invariant momentum distributions at the transition and extracted behavior consistent with the four-dimensional universality class.
This implementation demonstrates an especially subtle form of synthetic space. The extra dimensions were not literal rows of internal atomic levels. They emerged from quasiperiodic time dependence. Independent drive phases can play the role of momenta in extra dimensions, so a system evolving along one physical coordinate can reproduce localization physics belonging to a space with several more coordinates. Dimensionality here is encoded partly in the structure of time.
That flexibility makes synthetic dimensions broader than a single experimental trick. The recurring idea is to identify a collection of distinguishable states or phases, couple them in a controlled pattern, and arrange the resulting Hamiltonian so that its connectivity reproduces the target geometry.
=
Light offers an entirely different way to build synthetic space. An optical resonator supports a ladder of resonant frequencies. Instead of treating those frequencies merely as different colors of light, researchers can regard them as lattice sites. If adjacent resonances are separated by a frequency interval $\Omega$, their values can be labeled approximately as
$$ \omega_m=\omega_0+m\Omega $$
An electro-optic modulator driven at the appropriate frequency can convert photons between neighboring modes. In the synthetic description, frequency conversion becomes hopping along a lattice. The photon has moved in frequency space rather than physical space.
This approach has a practical advantage: a single small resonator can contain a long synthetic lattice because many frequency modes fit inside the same physical structure. Additional modulation tones can connect distant frequencies, create next-nearest-neighbor hopping, or imprint complex phases. Several resonators can then provide additional physical or synthetic coordinates. The dimensionality and connectivity of the simulated lattice can be changed largely through the applied electronic control signals rather than by fabricating a completely new chip.
In 2025, researchers demonstrated a programmable thin-film lithium-niobate device capable of constructing several kinds of synthetic frequency lattices on chip. The platform supported different coupling ranges and lattice structures, continuing a shift from synthetic dimensions as fixed demonstrations toward reconfigurable Hamiltonian engineering. In 2026, a hybrid-frequency architecture extended this idea by combining different classes of synthetic frequency sites and realizing Hall ladders, Creutz ladders, SSH-type lattices, topological flat-band behavior, and long-range coupling within one photonic platform.
Frequency lattices are especially interesting because they blur the distinction between moving a photon and changing the photon itself. In physical space, hopping from one lattice site to another changes position while leaving frequency approximately fixed. In a frequency synthetic dimension, a step along the coordinate changes the photon's energy. The lattice coordinate is therefore also a physical observable. This gives synthetic lattices forms of control that conventional crystals generally lack.
Boundaries provide a simple example. In an ordinary crystal, the edge is created by physically ending the material. In a frequency lattice, the experiment can engineer a boundary by suppressing selected transitions beyond a chosen frequency. Changing the control signal can move or reshape that boundary without cutting or rebuilding the device. Earlier photonic work explicitly demonstrated boundary engineering in synthetic frequency dimensions, and newer on-chip architectures are designed to combine large synthetic spaces with more flexible boundary control.
Long-range hopping is similarly natural. Connecting two distant atomic sites in a crystal can require complex hardware or interactions. Connecting distant synthetic frequency sites may require adding another modulation tone. The graph describing which sites are neighbors can therefore be changed while the physical chip remains the same. Synthetic geometry is closer to software than ordinary geometry.
The word geometry becomes especially interesting in this context. Ordinary materials inherit their local connectivity from where atoms are physically placed. A square lattice has four nearest neighbors, a triangular lattice has six, and a three-dimensional cubic crystal follows another fixed arrangement. Synthetic lattices can loosen this constraint because the notion of neighborhood is determined by which states are coupled.
A state can be connected to one neighbor, five neighbors, or a distant collection of states without requiring those states to be physically adjacent. Researchers have therefore proposed and studied synthetic systems corresponding to cylinders, ladders, tubes, curved spaces, high-dimensional lattices, and other nonstandard geometries. Recent perspectives on the field highlight proposals extending synthetic-dimension quantum simulators toward curved spaces, gauge theories, twistronics, quantum walks, and geometries that would be difficult to reproduce through conventional lattice construction.
This is where synthetic dimensions become conceptually broader than “extra dimensions.” They amount to a method for rewriting locality. In ordinary matter, local interactions are usually strongest between things that are physically nearby. In a synthetic lattice, locality can instead mean adjacency in spin number, frequency, momentum, or another engineered graph. The experimenter decides which states count as neighbors.
That freedom has limits. Not every arbitrary graph can be realized efficiently, and the physical interactions underlying synthetic hopping still obey the constraints of the actual device. Nevertheless, the ability to redesign effective locality is already enough to access models that would require much more elaborate hardware in real space.
The distinction becomes particularly important for many-body physics. Synthetic dimensions are often easier to control than real dimensions, but interactions can behave strangely in them. Two atoms occupying different synthetic spin states may still sit on top of each other in physical space, so they can interact strongly even though they appear far apart in the synthetic coordinate. Conversely, two synthetic neighbors may correspond to internal levels of atoms that are physically separated.
This creates interaction structures with no straightforward analogue in ordinary solids. Long-range interactions in synthetic space can arise naturally, and the same physical interaction can couple many synthetic sites at once. Such features were initially an obstacle to reproducing textbook lattice models exactly, but they are increasingly treated as an opportunity to study unfamiliar many-body Hamiltonians. The 2024 Rydberg experiment illustrates this shift by deliberately combining strong dipolar interactions with a microwave-built synthetic gauge lattice.
Polar molecules offer another promising route because a molecule can possess a large ladder of rotational states. Those states can serve as many sites along a synthetic coordinate while dipole-dipole interactions provide strong coupling between molecules in real space. Theoretical work has predicted unusual phases in which interacting molecules form strings and other extended objects across combined real and synthetic dimensions. The attraction of such systems lies partly in the large number of accessible internal states, which could produce synthetic directions much wider than the few-state ladders used in early experiments.
Synthetic dimensions also offer unusually direct access to edges. In a real material, bulk atoms vastly outnumber boundary atoms, and probing an individual topological edge can require sophisticated microscopy or transport contacts. In a spin synthetic dimension containing several magnetic sublevels, the lowest and highest spin states are automatically the two boundaries. Population in every synthetic site can sometimes be measured separately.
This was one reason the early synthetic Hall experiments were so revealing. Chiral edge motion could be associated with particular internal states rather than inferred only from an electrical current at the edge of a solid. Synthetic dimensions turn abstract lattice coordinates into experimentally addressable labels.
The same logic applies to topology in higher dimensions. A four-dimensional system has boundary structures that cannot be embedded intuitively in three-dimensional space without distortion, yet synthetic-state readout can still distinguish their characteristic dynamics. The mathematics does not require the human visual system to picture four perpendicular axes before the experiment can measure their consequences.
This is one of the quiet strengths of quantum simulation: physical understanding does not always require constructing a literal visual replica of the target system. It requires reproducing the relevant Hamiltonian and identifying observables that distinguish its phases.
There is an important philosophical temptation here that is best avoided. Because synthetic-dimensional experiments can reproduce four-dimensional equations, it can sound as though they are showing that our universe secretly contains accessible higher spatial dimensions. They are not. A synthetic coordinate does not have all the physical properties of an ordinary spatial direction. Interactions, causality, measurement, and locality may behave differently depending on how that coordinate is encoded.
What these experiments demonstrate is arguably more useful. Many observable phenomena do not care about the microscopic origin of a coordinate as much as they care about the structure of the Hamiltonian defined over it. A topological invariant can be measured in a parameter space. A Hall response can arise from engineered gauge phases. A four-dimensional localization transition can be reproduced through quasiperiodic drives. The mathematical relationships that define these phenomena can survive transplantation into a completely different physical platform.
This separation between physical space and dynamical space is familiar elsewhere in physics. Momentum space is not another ordinary room, yet band topology is naturally described there. Phase space combines position and momentum into a larger mathematical space. Hilbert space can have an enormous dimension unrelated to physical dimensionality. Synthetic dimensions differ because researchers deliberately turn selected nonspatial coordinates into dynamical lattice directions and then engineer transport through them.
The field is now moving toward greater programmability. Early synthetic-dimension experiments often demonstrated one carefully designed model. Newer systems increasingly aim to make the coupling graph reconfigurable so that a single platform can represent many Hamiltonians. Recent hybrid-frequency photonic chips are one example: modulation settings determine whether the device represents a Hall ladder, Creutz ladder, SSH chain, or more elaborate long-range lattice.
This suggests a different model for quantum simulation from the familiar goal of building a universal digital quantum computer. Instead of encoding every problem into gates on qubits, one can build a highly controllable physical system whose natural modes already form the relevant lattice. Changing frequencies, phases, and coupling amplitudes then changes the simulated geometry. Such analog simulators are specialized, but specialization can make difficult Hamiltonians far more accessible.
The synthetic-dimension approach is particularly attractive when the target problem involves topology, gauge fields, or high-dimensional geometry, because these are precisely the features that are hard to obtain by physically arranging components. An artificial coordinate sidesteps the need to construct the corresponding material literally.
The next challenge is to combine three capabilities that have often appeared separately: large synthetic dimensions, strong interactions, and high-fidelity local control. Large internal-state lattices make higher-dimensional geometry more convincing. Strong interactions are required for fractionalized and genuinely many-body phases. Local control is needed to prepare and measure complicated states rather than merely observe single-particle band structure. Rydberg arrays, polar molecules, ultracold atoms with large spin, and programmable photonics each provide different parts of this combination.
If these capabilities converge, synthetic systems could explore higher-dimensional interacting quantum Hall liquids rather than only their noninteracting counterparts. The recent four-dimensional atomic Hall work itself identifies strongly correlated 4D liquids as a future direction, but reaching that regime requires synthetic dimensions large enough for localized many-body excitations and interactions to develop cleanly.
Such experiments would be valuable partly because four-dimensional interacting topological matter is not simply a speculative material waiting to be discovered. It belongs to a mathematical landscape that cannot exist as an ordinary bulk crystal in three-dimensional space. Synthetic dimensions may therefore become the only realistic way to turn parts of that landscape into experimental physics.
There is also a practical side to synthetic dimensions, especially in photonics. Frequency lattices naturally connect to frequency conversion, routing, spectral shaping, and nonreciprocal transport. Topological ideas developed initially for fundamental physics can potentially make optical signals more robust against certain imperfections or allow them to move predictably through large networks of frequency channels.
It is too early to treat synthetic dimensions as a general replacement for ordinary photonic circuitry. Loss, fabrication imperfections, finite mode numbers, modulation bandwidth, readout complexity, and unwanted couplings remain real constraints. The more meaningful point is that frequency can be engineered not only as information carried by light but also as a space through which light moves. That change in viewpoint creates design possibilities unavailable when every optical frequency is treated as an independent channel.
A similar conceptual shift is occurring in atomic physics. Spin was traditionally a degree of freedom attached to a particle moving through space. Synthetic dimensions allow spin itself to become a place where dynamics occurs. Periodic driving was traditionally a way of perturbing a system in time; it can now provide coordinates of an effective higher-dimensional disorder problem. The components of the quantum system are familiar, but their roles are being reassigned.
That reassignment is the deeper reason synthetic dimensions are interesting. Physics normally begins with a geometry and asks what matter does inside it. Synthetic-dimension experiments partially reverse the problem. They begin with controllable quantum states and design the geometry those states experience.
A microwave phase can become magnetic flux. A sequence of atomic levels can become a boundary. A set of photon colors can become a lattice. Several incommensurate drive phases can become additional dimensions. A large atomic spin can help realize four-dimensional Hall dynamics. None of these constructions changes the dimensionality of ordinary space, but each changes the dimensionality relevant to the quantum equations being tested.
The distinction between simulation and reality becomes less important than it first appears. A quantum Hall edge state is recognized through its dynamics and topology, not because its lattice sites must be made from one particular substance. If an engineered atomic system has the same Hamiltonian and displays the corresponding quantized response, then the higher-dimensional physics is experimentally present in precisely the sense relevant to the model.
Synthetic dimensions therefore extend the experimental reach of physics without requiring the universe to provide new directions. They turn internal structure into geography and coupling graphs into geometry. The laboratory remains three-dimensional, but the quantum systems inside it no longer have to be.