Electronics Guide

Articles in This Category

Quantum Mechanical Foundations

Quantum technologies exploit a small set of properties that separate quantum systems from classical ones. Superposition allows a quantum bit, or qubit, to occupy a weighted combination of its two basis states rather than one or the other. Entanglement creates correlations between particles that cannot be reproduced by any classical description assigning independent properties to each particle, and those correlations persist regardless of separation. Quantum interference allows probability amplitudes to add constructively or destructively, which is the mechanism every useful quantum algorithm relies on to concentrate probability on correct answers and cancel incorrect ones.

Two further principles constrain what any quantum machine can do. Measurement collapses a superposition to a single classical outcome, so an n-qubit register described by 2ⁿ amplitudes yields only n classical bits when read out. The no-cloning theorem forbids making an independent copy of an unknown quantum state, which rules out the simplest form of redundancy used in classical error correction while simultaneously making eavesdropping detectable in quantum key distribution. The same physics is therefore both the constraint and the resource, depending on the application.

These correlations are not merely a theoretical convenience. Experiments closing successive loopholes in Bell inequality tests established that no local hidden-variable theory can reproduce quantum predictions, work recognized by the 2022 Nobel Prize in Physics awarded to Alain Aspect, John Clauser, and Anton Zeilinger. The engineering programs described in this section rest on that experimental foundation.

Qubits, Gates, and Coherence

A working quantum computer must satisfy a demanding set of requirements, articulated by David DiVincenzo around 2000 and still used as a checklist: a scalable register of well-characterized qubits, the ability to initialize them to a known state, coherence times far longer than gate times, a universal set of gates, and reliable qubit-specific measurement. Every hardware platform trades these requirements against one another, and no platform yet satisfies all of them comfortably at scale.

The two figures that dominate engineering discussion are gate fidelity and the ratio of coherence time to gate time. Superconducting transmon qubits operate in dilution refrigerators near ten to twenty millikelvin, are driven by microwave pulses in roughly the four-to-eight-gigahertz band, and execute two-qubit gates in tens to a few hundred nanoseconds against coherence times measured in hundreds of microseconds. Trapped-ion qubits hold coherence for seconds or longer but take tens to hundreds of microseconds per two-qubit gate, so the ratio of useful operations to coherence is broadly comparable. Leading platforms now report two-qubit gate fidelities in the range of 99.5 to 99.9 percent, which sounds excellent until one observes that a circuit of a few thousand gates then has little chance of finishing without an error.

That arithmetic is why raw qubit count is a poor summary of capability. A processor with a thousand noisy qubits and a two-qubit error rate near one percent cannot run circuits deep enough to be interesting, while a smaller register with an order of magnitude lower error rate can. Composite metrics such as quantum volume, circuit layer operations per second, and application-level benchmarks attempt to capture this, though no single number has achieved consensus across platforms, and vendor-published figures should be read with the measurement conditions attached.

Where Quantum Computers Help

Quantum computers are not universally faster than classical computers. They offer meaningful speedups only for problem classes whose structure quantum interference can exploit. Simulating quantum systems is the most natural application, and the one Feynman proposed in 1982: a quantum processor represents molecular and material systems whose exact classical simulation cost grows exponentially with size, which is why chemistry, catalysis, and materials science are the leading candidates for early scientific value.

Certain number-theoretic problems are also vulnerable. Shor's algorithm, published in 1994, factors integers and computes discrete logarithms in polynomial time, which breaks the RSA and elliptic-curve schemes underpinning most public-key infrastructure in use today. No machine remotely capable of running it at cryptographic scale exists, but the threat is prospective rather than hypothetical, because encrypted traffic captured today can be stored and decrypted later. That "harvest now, decrypt later" exposure, not any current hardware, is what drives migration schedules. Grover's algorithm, published in 1996, provides a quadratic speedup for unstructured search, which is real but far less dramatic; doubling the key length of a symmetric cipher restores the lost margin.

The practical response has been post-quantum cryptography rather than quantum hardware. NIST published its first post-quantum standards on August 13, 2024: FIPS 203 (ML-KEM, derived from CRYSTALS-Kyber) for key encapsulation, FIPS 204 (ML-DSA, derived from CRYSTALS-Dilithium), and FIPS 205 (SLH-DSA, derived from SPHINCS+) for signatures. In March 2025 NIST selected HQC as a backup key-encapsulation mechanism built on different mathematical assumptions. NIST's published transition guidance plans to deprecate 112-bit-classical-strength algorithms such as RSA-2048 around 2030 and to disallow them after 2035, and the NSA's CNSA 2.0 suite sets a comparable deadline for national security systems. Electronics engineers encounter this as a hardware problem: larger keys and signatures, new accelerator requirements, and long-lived embedded devices that must be able to accept a cryptographic update years after shipping.

Optimization is the most oversold category. Quantum approaches to combinatorial optimization exist, including variational methods and annealing, but rigorous evidence of practical advantage over well-tuned classical heuristics remains scarce, and several early claims did not survive classical re-analysis. Treat optimization claims skeptically until the classical baseline is specified.

Error Correction and the Path to Fault Tolerance

Today's machines operate in what is commonly called the noisy intermediate-scale quantum era, in which qubit counts are large enough for interesting experiments but error rates cap circuit depth. The route beyond it is quantum error correction, which encodes one logical qubit across many physical qubits and continuously measures error syndromes without measuring, and thereby destroying, the encoded information itself. The threshold theorem guarantees that if physical error rates fall below a code-dependent threshold, roughly one percent for the widely studied surface code, then adding more physical qubits reduces the logical error rate arbitrarily.

Crossing that threshold experimentally was the field's central open question for two decades, and it has now been answered in the affirmative. In December 2024 Google reported in Nature that its 105-qubit Willow processor operated below threshold: increasing the surface-code distance from three to five to seven roughly halved the logical error rate at each step, the scaling behavior that fault tolerance requires. Neutral-atom groups reached comparable milestones by a different route, including a Harvard, MIT, and QuEra demonstration of 48 logical qubits on a 280-atom processor in 2023, and a Microsoft and Atom Computing demonstration entangling 24 logical qubits in 2024.

The remaining problem is overhead. Surface-code encoding of a single high-quality logical qubit can require on the order of a thousand physical qubits, which places a cryptographically relevant machine well beyond current fabrication. Reducing that ratio is now a primary research target, and quantum low-density parity-check codes are the leading candidate because they encode more logical qubits per physical qubit at the cost of requiring couplings between distant qubits rather than only nearest neighbors. IBM's published roadmap follows exactly this line: the experimental Loon chip tests the long-range coupling and multilayer routing such codes need, while the Starling system, targeted for 2029, is specified at roughly 200 logical qubits built from on the order of 10,000 physical qubits. Vendor roadmaps are statements of intent rather than results, and should be read as such.

Hardware Platforms

Several physical platforms compete to implement practical quantum processors, and none has yet won. Superconducting circuits lead in gate speed and industrial maturity. IBM's 1,121-qubit Condor processor demonstrated large-scale integration in 2023, after which the company shifted emphasis from qubit count to quality, with the Heron family using tunable couplers to suppress crosstalk and the 120-qubit Nighthawk design adopting a square lattice with 218 tunable couplers to raise the circuit depth a processor can sustain. Google's Willow follows the same philosophy. The field's shift from advertising qubit counts to reporting error rates is the single clearest sign of its maturation.

Trapped-ion systems, built by companies including Quantinuum and IonQ, offer the highest reported gate fidelities and all-to-all connectivity within a trap, since any pair of ions can interact through shared motional modes. Their weaknesses are slower gates and the engineering difficulty of scaling beyond a single trap, which motivates architectures that shuttle ions between zones or link traps photonically. Neutral-atom arrays scale most easily in raw atom number, because optical tweezers can hold and rearrange large registers: Atom Computing has operated 1,180 atoms in a 1,225-site array, and Pasqal has demonstrated registers of roughly a thousand atoms. Reconfigurable atom movement gives these machines flexible connectivity that suits high-rate error-correcting codes.

Photonic processors, pursued by PsiQuantum and Xanadu among others, manipulate qubits encoded in light and avoid cryogenic operation for the optical path itself, though practical designs still require superconducting nanowire single-photon detectors that run cold. They integrate naturally with optical fiber, which makes them attractive for networked architectures. Silicon spin qubits aim to reuse established CMOS manufacturing and offer an extremely small qubit footprint, but lag in demonstrated system size. Topological qubits promise error resistance built into the physics rather than added by software; Microsoft's 2025 report of a Majorana-based device attracted wide attention and also substantial scientific scrutiny, and the approach remains unproven. Quantum annealers, notably D-Wave's, form a separate branch: they are not universal gate-model machines and target optimization and sampling problems specifically.

Quantum Sensing and Metrology

Sensing is the most commercially mature quantum technology, and it advances quietly compared with computing. The underlying idea is that a quantum system with well-defined energy levels makes an exceptionally stable reference, and that superposition and entanglement can push measurement precision toward and past the standard quantum limit. Optical lattice and single-ion clocks now reach fractional frequency uncertainties near one part in 10¹⁸, well beyond the caesium microwave standard that has defined the SI second since 1967. International metrology bodies have set out a roadmap toward redefining the second on an optical transition, with adoption anticipated around 2030 once agreed technical conditions on clock comparisons and time-scale contributions are met.

Beyond timekeeping, nitrogen-vacancy centers in diamond provide magnetometers with nanoscale spatial resolution that operate at room temperature; SQUIDs and optically pumped magnetometers detect the femtotesla fields produced by the brain and heart; and atom-interferometric gravimeters measure gravitational acceleration precisely enough to map subsurface density, with applications in geophysics, civil engineering, and navigation where satellite positioning is denied. Squeezed light is already deployed operationally: gravitational-wave observatories inject squeezed vacuum states to reduce quantum noise and improve detector sensitivity. These devices need no error correction and no logical qubits, which is why they reached fielded products first.

Quantum Communication and Networking

Quantum communication uses measurement disturbance and the no-cloning theorem to make interception detectable. In the BB84 protocol, proposed by Charles Bennett and Gilles Brassard in 1984, two parties exchange single photons prepared in randomly chosen bases; any eavesdropper necessarily perturbs the transmitted states and raises the observed error rate, allowing the legitimate parties to discard a compromised key. Practical systems add decoy states to close the loophole created by imperfect single-photon sources.

Loss is the fundamental constraint. Photons cannot be amplified without destroying their quantum state, so key rates fall exponentially with fiber length, and direct fiber links are practical over hundreds of kilometers rather than thousands. Two responses dominate. Satellites bypass fiber loss by transmitting through the low-loss upper atmosphere and vacuum: China's Micius satellite distributed entangled photon pairs over more than 1,200 kilometers and supported intercontinental key exchange, and Europe's EuroQCI initiative, including the SES-led EAGLE-1 mission developed with ESA, pursues a comparable capability. Quantum repeaters, which use entanglement swapping and quantum memories to extend range without measuring the key, remain a laboratory technology.

An honest assessment must note that quantum key distribution is not the mainstream answer to the quantum threat. Several national security agencies, including those of the United States and the United Kingdom, recommend post-quantum cryptography rather than quantum key distribution for general-purpose protection, citing the cost of dedicated fiber and hardware, the limited scope of what the physics actually guarantees, and the difficulty of authenticating the classical channel the protocol still requires. Quantum networking retains a strong long-term rationale for linking quantum processors and distributing entanglement, which is a task no classical network can perform.

Engineering and Economic Challenges

Scaling a quantum processor is largely a classical electronics problem. Superconducting machines have historically required at least one coaxial control line per qubit, running from room-temperature instruments through several refrigerator stages, where heat load and physical space both become limiting. The cooling budget is unforgiving: a dilution refrigerator supplies on the order of a milliwatt at one hundred millikelvin and only tens of microwatts near its base temperature, so every added wire and amplifier competes for a vanishingly small thermal allowance. Cryogenic CMOS control chips, multiplexed readout, and photonic interconnects are all under development to break that dependence. Trapped-ion and neutral-atom systems face an analogous problem in optics rather than wiring: hundreds of stabilized laser beams must be routed, switched, and held to tight frequency tolerances, which drives interest in integrated photonics.

Supporting constraints are equally practical. Helium-3, essential to dilution refrigeration, comes from a constrained supply chain. Syndrome decoding for error correction must complete faster than errors accumulate, which imposes hard real-time requirements on classical decoders, typically implemented in FPGAs or dedicated ASICs beside the cryostat. Capital and operating costs keep most access on cloud platforms rather than on premises, and hybrid workflows that pair a quantum processor with classical high-performance computing are the normal mode of use, with each resource handling the part of a problem it does best.

Outlook

Quantum technologies advance on several fronts on different timelines. Sensing and metrology already deliver fielded instruments and are close to redefining a base SI unit. Quantum communication has working point-to-point and satellite links but an unsettled role relative to post-quantum cryptography. Quantum computing has passed a genuine scientific threshold with the demonstration of below-threshold error correction, and vendors now target verified computational advantage on useful problems and fault-tolerant machines near the end of the decade. Those targets remain forecasts.

The realistic near-term picture is not one machine that outperforms classical computers everywhere, but a set of specialized capabilities in simulation, sensing, and secure communication that mature separately. For electronics practitioners, the immediate obligations are concrete rather than speculative: plan cryptographic agility into long-lived products now, and follow error rates and logical qubit counts rather than headline qubit numbers. The subcategories in this section examine the hardware, software, error correction, sensing, and communication pillars in depth, along with the engineering trade-offs that will determine how quickly these systems move from demonstration to dependable practice.

Related Topics