Quantum hardware overview


Modern classical computers are based on transistors made from semiconductors to implement classical logic operations. Quantum computers take radically different approaches to exploit the properties of quantum systems to implement qubits and quantum operations.

In this unit, you'll learn about the challenges of building a quantum computer. You'll also learn the different types of physical qubit systems and how quantum devices are designed.

Harness quantum systems to perform computations

To build quantum computers, we need full access to the controllable quantum systems that will represent the qubits that we'll use to store and process the quantum information. A qubit can be represented by any physical system with two states or levels that can be prepared in a quantum superposition. In general, a functional quantum computer should fulfill the following criteria:

  • Scalability: Quantum computers are based on the manipulation of quantum information by using qubits. We need a system that allows scalability to a large number of qubits.
  • Ability to initialize the qubits in a specific state: Having many qubits isn't useful if we can't trust their initial state. A fundamental characteristic of a quantum computer is its ability to initialize the qubits reliably to a specific state (usually, the state $|0\rangle$).
  • Resilient qubits: Quantum states are delicate. The smallest interaction with the environment can "contaminate" the state of the qubits and ruin our computations. Qubit registers need to be able to remain in a consistent quantum state long enough for us to perform the computations. The amount of time that a qubit can remain in a specific quantum state is often called coherence time.
  • Ability to perform a universal set of operations: Like in classical computing, any quantum algorithm can be decomposed as the application of a sequence of operations from a universal set of basic operations. This set isn't unique, but a quantum computer should be able to reliably perform all the operations of the chosen universal set.
  • Reliable measurements: To obtain the results of quantum computations, we need to measure each qubit with high precision. Note that the measurement affects the state of the register. It's often performed only at the end of the computation.

These five criteria are often known as the DiVincenzo criteria for quantum computation.

Building devices that meet these five criteria is one of the most demanding engineering challenges ever faced by humankind. However, recent years have seen astounding advances in condensed matter physics and quantum optics. These advances are allowing companies and universities to build the first working versions of quantum computers. Microsoft is partnering with some of the best-in-class quantum computer manufacturers around the world to give you access to the latest quantum computing solutions through Azure Quantum.

Here's an overview of the different technologies used to create quantum computers, accessible via Azure Quantum.

Trapped ion quantum computers

Trapped ion quantum computers use ions (electrically charged atoms) suspended in an electromagnetic field in a vacuum as the basic building block.

Diagram of trapped-ion-quantum-computer.

In particular, these computers use chains of trapped ions and each ion represents a qubit. They use lasers to induce controlled vibrations on the chain to accomplish operations in the qubit states. The qubit states are stored in the internal states of each ion. The qubit states are often stored in two hyperfine levels of the ion.


The hyperfine levels of an atom arise from the different possible configurations of the spin orientations of the electron and the nucleus. Generally, the state with the lower energy is associated with the state $|0\rangle$ and the state with higher energy is associated with the qubit state $|1\rangle$.

These computers use lasers and electromagnetic pulses to cause transitions in the internal states of the ions and to modify the vibrational motion of the ion chain. By carefully combining transformations of the vibrational motion of the ion chain and individual ion transitions, it's possible to get qubits entangled and apply a universal set of quantum operations.

Azure Quantum provides access to trapped ion quantum devices through our partners, IonQ and Honeywell. In this module, you'll submit quantum operations to this kind of devices.

Superconducting quantum computers

Superconducting quantum computers are based on superconducting electronic circuits. Classical computers use circuits of transistors to represent bits and perform classical computations. Superconducting quantum computers use transmons to represent qubits and perform quantum operations.

If certain materials are cooled below a critical temperature, their electric resistance drops to zero. These materials are called superconductors. By linking two superconductors with a thin insulating barrier, you can build transmons. A transmon is an electronic device that possesses quantum states in superposition, with two of the states used as |0> and |1>. Transmons can be considered analogous to transistors. In transistors, you use junctions of semiconductors to represent controllable bits. In transmons, you use junctions of superconductors to represent controllable qubits.

Diagram comparing a transmon with a transistor. The transmon can be prepared in a quantum superposition while the transistor only admits discrete classical levels.

There are other variants of superconducting circuits that don't use transmons but similar components. However, transmons are the key component in most cloud-available superconducting quantum computers.

Since superconducting properties appear at very low temperatures and higher temperatures imply more noise, these quantum computers work at very low temperatures. This is why, with the aid of a cryostat, the superconducting chip is cooled to near absolute zero.

Simplified diagram of the cryostat of a superconducting quantum computer.

Other types of quantum computers

Trapped ion and superconducting quantum computers aren't the only options for quantum computing, although they're implemented the most in industry so far. Here are several other approaches that are currently subjected to active research:

  • Quantum dot computers: Quantum dots are small clusters of semiconductor atoms that behave like macroscopic atoms and can be prepared into superposed states.
  • Photonic quantum computers: Polarized photons used to represent qubits and photon waveguides to implement quantum operations.
  • Topological quantum computers: A new type of particles called anyons, first proposed theoretically. Anyons are now in active research and development to create qubits resilient to noise.

There are many other proposals for quantum computers, since the field is still under development. With Azure Quantum, you have access to an ever-growing collection of quantum computers to run your own quantum algorithms.

To learn about Microsoft's latest efforts to advance quantum computing technologies, see Quantum Computing.

Azure Quantum for cloud quantum computing

Azure Quantum is a cloud computing hub where you can connect to different industry-leading providers of quantum hardware to use their most advanced quantum computers.

Diagram of the Azure Quantum service.

All you need is an Azure subscription to start submitting quantum computing jobs to Azure Quantum. These jobs will be queued by the provider to run on the target device you select. Through Azure Quantum, you can manage your jobs, monitor their status, and get the results after they're completed.

In the next unit, you'll use Azure Quantum to submit your first job to a quantum computer.