Quantum Lab Concepts

In 1922, Otto Stern and Walther Gerlach sent a beam of silver atoms through a strongly inhomogeneous magnetic field (produced by a specially shaped magnet with one pole ground to a sharp edge) and observed where the atoms landed on a detector screen downstream. Classically, if each atom’s intrinsic magnetic moment could point in any random direction, the field gradient’s resulting deflecting force should have produced a continuous vertical smear of landing positions, spread across a range proportional to each possible magnetic-moment orientation. Instead, the atoms landed in exactly two, sharply-separated discrete spots – direct, unambiguous experimental evidence that the atoms’ magnetic moment (and the underlying angular momentum it originates from) is quantized, restricted to only two possible values along the field’s measurement axis, rather than being free to point continuously in any direction – a landmark result later understood as the first direct experimental demonstration of electron spin, predating the theoretical concept of spin itself by several years.

Passing a beam through a first Stern-Gerlach magnet oriented along z and selecting only the atoms deflected ”up” prepares a beam of atoms with a definite, well-defined Sz = +¯h/2 spin state. If this already-filtered, Sz-definite beam is then passed through a second Stern-Gerlach magnet oriented along a different axis, x, the atoms split again into two new spots (corresponding to Sx = ±¯h/2) with roughly equal probability – demonstrating directly and unambiguously that Sz and Sx are incompatible, non-commuting observables (exactly as introduced in the Advanced Quantum Mechanics deck): even though the atoms entering the second magnet had a perfectly definite, previously-measured Sz value, that same beam has no definite value of Sx at all, and measuring Sx genuinely, physically randomizes the atoms’ outcomes along that new axis – making this simple, sequential apparatus one of the cleanest, most directly intuitive laboratory demonstrations of the general quantum-measurement principle that measuring one observable can fundamentally disturb another, incompatible one.

The double-slit experiment sends particles (originally photons; later demonstrated with electrons, and eventually with much larger, massive molecules) toward a screen with two narrow, closely-spaced slits, and observes the resulting pattern accumulated on a detector screen behind the slits. Sending particles through one at a time – waiting long enough between each individual particle that no two could possibly be present in the apparatus simultaneously – and still observing an interference pattern gradually build up, spot by spot, over many individual, independent detection events directly rules out any naive classical explanation involving particles somehow physically interacting with or influencing each other in transit; instead, it demonstrates that each individual particle’s own probability of landing at a given position is itself already governed by wave-like interference, as if each single particle’s own possible paths

Interference in the double-slit experiment fundamentally requires the two possible paths (through slit A, or through slit B) to remain genuinely indistinguishable at the point of detection, so that their corresponding quantum amplitudes can coherently add together (or destructively cancel) at the final screen. Installing any apparatus capable of reliably recording which specific slit a given particle passed through – even one engineered to disturb the particle’s own subsequent trajectory as gently and minimally as physically possible – necessarily requires that apparatus to become correlated (in the sense of quantum entanglement) with the particle’s which-path information, and this correlation alone is sufficient to eliminate the interference pattern entirely, regardless of whether anyone actually ever looks at or reads out the detector’s recorded result. This is a direct, foundational illustration of decoherence (introduced generally in the Advanced Quantum Mechanics deck) in its simplest, most pedagogically clear possible setting – the mere existence of accessible which-path information, not any observer’s conscious awareness of it, is what destroys the interference.

In the photoelectric effect, light shining on a metal surface ejects electrons, and careful measurement reveals two experimental facts fundamentally incompatible with classical electromagnetic wave theory: increasing the incident light’s intensity (classically, its wave amplitude) increases the number of ejected electrons (the resulting photocur- rent) but leaves their maximum kinetic energy completely unchanged, while increasing the light’s frequency directly and proportionally increases each ejected electron’s maximum kinetic energy, with a sharp threshold frequency be- low which no electrons are ejected at all, regardless of how intense the light is made. Classical wave theory predicts that a sufficiently intense wave, given enough time to deliver adequate cumulative energy, should always eventually eject electrons regardless of frequency, and that a more intense wave should straightforwardly deliver more energy per ejected electron – neither is observed. Einstein’s 1905 explanation (that light itself consists of discrete quanta, or photons, each carrying energy E = hν, with the photoelectric process removed of exactly one photon per one ejected electron) correctly and quantitatively accounted for both otherwise-inexplicable observations, providing foundational early evidence for the quantization of light itself.

A Mach-Zehnder interferometer splits an incoming beam at a first beamsplitter into two separate spatial paths (”arms”), lets each path accumulate its own independent phase (potentially adjustable via a phase shifter placed in one arm), and then recombines both paths at a second beamsplitter before final detection. When operated with genuinely single photons (one at a time, exactly as in the double-slit discussion above), each individual photon’s detection probability at the two final output ports depends on the relative phase difference between the two arms – demonstrating that each single photon’s quantum state genuinely, coherently traverses both arms simultaneously as a superposition, with the specific final detection probabilities determined by interference between these two coher- ent path amplitudes – making the Mach-Zehnder interferometer the standard, most widely used minimal laboratory testbed for exploring and demonstrating single-photon interference, complementary state preparation, and (with a which-path detector inserted into one arm) the interference-destroying effect of which-path information discussed above.

In a delayed-choice version of the Mach-Zehnder setup, the experimenter’s decision whether to insert the final, re- combining beamsplitter (enabling interference, and thus ”wave-like” behavior) or leave it out (allowing a simple, direct which-path measurement, revealing ”particle-like” behavior) is made after the photon has already entered the interferometer and passed the first beamsplitter – genuinely too late, in principle, for the photon to have some- how ”already decided in advance,” at the moment it passed the first beamsplitter, which specific type of experiment (wave-like or particle-like) it was ultimately going to be subjected to. The experimentally confirmed result – that the photon’s observed behavior always correctly, retroactively matches whichever experimental configuration was ultimately chosen, regardless of how late that choice is made – strongly disfavors naive ”hidden variable” pictures in which a photon carries some definite, predetermined internal property (a fixed, particle-like or wave-like ”nature”) from the very start, instead reinforcing the standard quantum-mechanical view that a system’s properties are not definite and fixed until the specific, complete experimental context is actually, fully determined.

Building on the which-path-destroys-interference principle (Section 1), a quantum eraser experiment first marks a particle’s which-path information (e.g. by entangling its path with an auxiliary quantum system’s state, destroying the primary interference pattern exactly as expected), but then performs an additional, subsequent measurement on that auxiliary marking system, specifically chosen to erase (or ”randomize away”) the previously-recorded which-path information rather than simply reading it out. Sorting the primary particle’s detection events according to the auxiliary system’s erasure-measurement outcome reveals that interference can be recovered, appearing in specific correlated

A Michelson interferometer splits light at a beamsplitter into two perpendicular arms, each terminated by a mirror that reflects the light back to recombine at the same beamsplitter – structurally similar to, but historically predat- ing, the Mach-Zehnder configuration, and famously used in the 1887 Michelson-Morley experiment’s unsuccessful search for the hypothesized ”luminiferous ether” (the null result of that classical experiment being one of the pivotal pieces of evidence eventually motivating special relativity). While originally a purely classical-optics experiment, the Michelson configuration’s basic two-arm, path-length-sensitive interferometric structure directly underlies nu- merous modern quantum-optical and precision-metrology applications discussed throughout this deck series – most notably, LIGO’s gravitational-wave detector (introduced in the Quantum Sensing deck) is, at its architectural core, an enormous, exquisitely precision-engineered Michelson interferometer, making the basic Michelson configuration a valuable, foundational conceptual stepping stone connecting classical 19th-century optics directly to today’s most sensitive quantum-enhanced measurement instruments.

In their 1935 paper, Einstein, Podolsky, and Rosen considered a pair of particles prepared in a specific correlated (what would today be called entangled) state, such that measuring one particle’s property (position, or in later, simplified spin versions of the argument, its spin along a given axis) allows one to immediately, with certainty, predict the corresponding measurement outcome on the distant, second particle – without ever having directly disturbed that second particle at all. Reasoning from the principle that a distant, causally-disconnected measurement should not be able to instantaneously create or determine a genuine physical property at a spatially separated location (a principle they termed ”local realism”), EPR argued that this distant particle must therefore already have possessed that definite, predictable property all along, even before any measurement occurred – and since standard quantum mechanics’ wavefunction description does not assign such a definite, pre-existing value to the unmeasured particle, EPR concluded that quantum mechanics, as then formulated, must be an incomplete description of physical reality, presumably to be supplemented by some additional, currently-missing ”hidden variables.”

Bell’s theorem proved, remarkably, that any theory satisfying EPR’s own ”local realism” assumptions (that measure- ment outcomes are determined by pre-existing, local hidden properties, unaffected by instantaneously distant events) must satisfy a specific, quantitative mathematical constraint – a Bell inequality – bounding the strength of correlations achievable between measurements on the two separated particles for various different combinations of measurement settings. Crucially, Bell also showed that quantum mechanics itself predicts that appropriately-prepared entangled particles can violate this inequality, exceeding the maximum correlation strength any local hidden variable theory could ever produce. This was a genuinely transformative result specifically because it converted what had seemed, for three decades, like a purely philosophical disagreement about the interpretation and completeness of quantum mechanics into a concrete, sharply falsifiable, and directly experimentally testable numerical prediction – if real ex- periments could be performed and their measured correlations exceeded Bell’s bound, local hidden variable theories of the EPR type would be definitively ruled out by direct experimental evidence, not mere philosophical argument.

Aspect’s landmark experiments (building on earlier Bell tests by Freedman and Clauser) used pairs of entangled pho- tons generated via atomic cascade emission, measuring their polarization correlations at two well-separated detector stations and confirming a clear, statistically significant violation of a Bell inequality, closely matching quantum me- chanics’ predictions and inconsistent with any local hidden variable theory. Aspect’s most celebrated refinement specifically addressed the locality loophole – the concern that if each measurement station’s chosen setting were somehow fixed and predetermined well before the photons left their shared source, some conceivable (if physically implausible) local signal could, in principle, have communicated each station’s eventual setting to the photon source (or to the other station) in time to influence the outcome, without requiring any genuinely faster-than-light influence.

Beyond the locality loophole Aspect’s experiments addressed, a separate, distinct detection loophole (also introduced in the Quantum Key Distribution deck’s discussion of device-independent QKD) arises whenever real detectors have imperfect efficiency, missing some fraction of the entangled pairs – since only the subset of pairs that were actu- ally successfully detected on both sides contribute to the measured correlation statistics, it remains mathematically conceivable that this specific, detected subset is not a fair, representative sample of the full underlying population, potentially allowing a cleverly constructed local hidden variable model to mimic an apparent Bell violation purely through this selective, efficiency-related sampling bias. The landmark 2015 experiments (using, variously, entangled electron spins in diamond color centers, or entangled photons combined with high-efficiency detection) achieved sufficiently high overall detection efficiency, combined with Aspect-style fast setting-randomization, to close both loopholes simultaneously within a single, unified experiment for the first time – providing what is now widely re- garded as the first genuinely definitive, loophole-free experimental confirmation that our universe’s actual correlations cannot be explained by any local hidden variable theory whatsoever.

Doppler cooling illuminates atoms from multiple directions with laser light tuned slightly below (red-detuned from) the atoms’ own resonant absorption frequency. An atom moving toward a given laser beam sees that beam’s frequency Doppler-shifted upward, closer to (and thus more strongly absorbed at) the atom’s true resonance – while an atom moving away from that same beam sees it shifted further away from resonance, absorbing correspondingly less. Since each absorbed photon imparts a small momentum kick directly opposing the atom’s motion toward the laser source, and atoms moving in any direction preferentially absorb more strongly from whichever beam directly opposes that specific motion, the net effect (with beams arranged along all relevant axes, in a configuration commonly called ”optical molasses”) is a velocity-dependent damping force that decelerates atoms regardless of which direction they happen to be moving in – directly analogous to a viscous drag force, but generated entirely through the atoms’ own resonant light absorption and the Doppler effect, providing the foundational cooling mechanism underlying essentially all subsequent laser-cooling and trapping techniques.

A MOT combines the same red-detuned, multi-directional laser-cooling beams used in Doppler cooling with a pair of magnetic coils arranged in an ”anti-Helmholtz” configuration, producing a magnetic field that is exactly zero at the trap’s center and increases linearly with distance away from that center in every direction. This spatially-varying field Zeeman-shifts the atoms’ relevant internal energy levels by an amount that depends on the atom’s specific position – engineered so that an atom displaced away from the trap center experiences a Zeeman shift that brings it closer to resonance specifically with whichever laser beam is pointed back toward the center, and further from resonance with the beam pointing away – adding a genuine, position-dependent restoring force pulling atoms back toward the trap’s center, on top of the velocity-dependent damping the optical molasses configuration alone already provides. This combination of simultaneous velocity damping (cooling) and spatial confinement (trapping) makes the MOT the standard, near-universal starting point for essentially all modern cold-atom experiments, routinely cooling and confining a cloud of atoms down to temperatures in the microkelvin range.

Standard Doppler-cooling theory predicts a fundamental minimum achievable temperature (the ”Doppler limit,” set by the intrinsic random recoil heating from spontaneously emitted photons balancing against the cooling force) – but careful early experiments measuring actual optical-molasses temperatures found atoms routinely, reproducibly cooled to temperatures significantly below this theoretical limit, an unexpected result that Doppler theory alone could not explain. The resolution came from recognizing that real optical-molasses laser configurations, using multiple in- tersecting beams with different polarizations, create a spatially-varying polarization gradient (rather than a spatially uniform light field) – and atoms moving through this varying polarization landscape experience an additional, more

An optical tweezer exploits the AC Stark shift (light shift): an atom placed in an intense, off-resonant laser field has its internal energy levels shifted by an amount proportional to the local light intensity, and, for laser light detuned appropriately (typically red-detuned, below the relevant atomic resonance), this shift lowers the atom’s ground-state energy specifically in regions of higher light intensity – meaning the atom is drawn toward, and can become trapped at, the point of maximum intensity created by tightly focusing a laser beam through a high-numerical-aperture lens. This creates a genuine, three-dimensional, conservative potential-energy well strong enough to capture and hold a single neutral atom (or, with suitable adaptation, single molecules or nanoparticles) at the laser’s focal point – and, critically, arrays of many independently-positioned optical tweezers (created using programmable holographic beam- shaping techniques) have become one of the leading modern platforms for assembling large, precisely-controlled, individually-addressable arrays of neutral-atom qubits for quantum simulation and quantum computing.

Earnshaw’s theorem proves that no arrangement of purely static electric fields can create a stable, three-dimensional confining potential minimum for a charged particle in free space (a static field configuration that confines along one axis necessarily de-confines along another). A Paul trap circumvents this fundamental limitation by using a rapidly oscillating (radio-frequency) electric field instead of a static one – while the instantaneous field at any given moment does indeed only confine along some axes while de-confining along others (exactly as Earnshaw’s theorem demands), rapidly alternating which axes are confining versus de-confining, faster than the trapped ion can significantly respond and escape during the ”wrong” half of each cycle, produces a net, time-averaged effective confining potential (a ”pseudopotential”) that is genuinely stable in all three dimensions – this elegant trick is the foundational technol- ogy underlying essentially all modern trapped-ion quantum computing and precision-spectroscopy experiments, and earned Wolfgang Paul a share of the 1989 Nobel Prize in Physics.

Haroche’s group at ENS Paris developed an extraordinarily high-quality-factor superconducting microwave cavity (constructed from superconducting niobium, cooled to near absolute zero) capable of storing individual microwave photons for remarkably long times, and sent a controlled stream of highly-excited ”Rydberg” atoms (atoms with one electron excited to an extremely large principal quantum number, making them exceptionally strongly coupled to microwave fields) through the cavity one at a time. By carefully measuring each Rydberg atom’s own inter- nal state after it passed through the cavity (a technique exploiting the same dispersive-shift principle discussed for superconducting-qubit readout in the Superconducting Quantum Circuits deck, here applied to directly count photons non-destructively), Haroche’s team could repeatedly probe the cavity’s photon number without absorbing or destroy- ing the photons being measured – directly observing, for the first time, individual photons persisting in the cavity and occasionally undergoing sudden ”quantum jumps” between definite photon-number states, providing some of the most direct and striking experimental confirmations of core quantum measurement postulates ever achieved.

Wineland’s group at NIST developed techniques for trapping individual ions in Paul traps (discussed above), laser- cooling them to their motional ground state, and using precisely controlled laser pulses to manipulate and read out individual ions’ internal quantum states (and their coupled motional states) with extraordinary precision – demon- strating, among many other landmark results, some of the earliest quantum logic gates between individual trapped ions, and pioneering techniques for using one ion’s motional state as a controllable, ”quantum bus” to mediate gen- uine multi-ion entangling interactions. This represented a conceptual mirror image of Haroche’s approach: where Haroche used controlled atoms to non-destructively measure and manipulate trapped photons in a cavity, Wineland

A Penning trap achieves stable confinement using a fundamentally different strategy than a Paul trap’s oscillating- field trick: it combines a static (DC) electric quadrupole field, which alone provides stable confinement along the trap’s central axis but is inherently de-confining in the perpendicular radial plane (exactly the Earnshaw’s-theorem limitation discussed above), with a strong, uniform static magnetic field aligned along that same central axis – the magnetic field causes any ion attempting to escape radially to instead execute a stable, bound cyclotron orbit (via the Lorentz force) rather than actually escaping, converting the otherwise radially-de-confining electric field into an overall stable, fully three-dimensional trap. Penning traps are particularly prized for extraordinarily precise mass spectrometry and fundamental-physics tests (measuring the electron’s magnetic moment/g-factor to remarkable pre- cision, for instance, directly testing the Dirac-equation predictions discussed in the Advanced Quantum Mechanics deck) specifically because they avoid the RF micromotion (a small, residual oscillatory motion driven by the oscil- lating trapping field itself) that Paul traps inherently introduce, at the cost of requiring a large, strong, and technically demanding superconducting magnet as part of the apparatus.

Quantum state tomography is the experimental procedure for fully reconstructing an unknown quantum state’s com- plete density matrix ρ (introduced in the Quantum Information deck) from a set of measurements performed on many identically-prepared copies of that state. A single fixed measurement basis (say, the computational Z basis alone) only reveals the diagonal elements of ρ in that specific basis – providing no information whatsoever about the crucial off-diagonal coherence terms that distinguish a genuine quantum superposition from a simple classical statistical mixture of the same diagonal populations. Reconstructing the complete, unambiguous density matrix therefore re- quires repeating the measurement across several different, mutually incompatible bases (e.g. the X, Y , and Z Pauli bases for a single qubit) on freshly-prepared, identical copies of the state, and then combining the resulting statistics from all these different measurement settings – via a suitable reconstruction algorithm – into the single, complete density matrix consistent with every one of the individually-measured basis outcomes.

Given a finite, necessarily noisy set of experimentally measured tomography statistics, a naive linear inversion ap- proach (directly, algebraically solving for the density matrix consistent with the observed data) can easily produce a mathematically ”reconstructed” matrix that is unphysical – failing to be a genuinely valid density matrix (for in- stance, having a negative eigenvalue, which is impossible for any real, physical quantum state) purely due to ordinary statistical measurement noise and finite sample size. Maximum-likelihood estimation instead searches, among only the space of genuinely physically valid density matrices (correctly Hermitian, positive semi-definite, and trace-one), for the specific one that makes the actually-observed experimental measurement data most probable/likely – guaran- teeing, by construction, that the final reconstructed state is always a valid, physically sensible density matrix, which is precisely why MLE (or related, more modern Bayesian reconstruction techniques) has become the standard, gener- ally preferred approach for quantum state tomography in real experimental settings, rather than simple, unconstrained linear inversion.

The Wigner function represents a quantum state as a function defined over the same classical phase space (position and momentum, or equivalently the two field quadratures for a light mode) that a classical probability distribution would occupy – and for states with a good classical analogue (like an ordinary coherent state of light), the Wigner function closely resembles an ordinary, non-negative classical probability distribution. Crucially, however, for genuinely non- classical states (Fock states, Schrödinger-cat-like superposition states, and other states with no meaningful classical counterpart), the Wigner function can take on negative values in certain regions of phase space – a mathematical impossibility for any genuine classical probability distribution, but perfectly well-defined and physically meaningful for this specific quantum quasi-probability construction. Because of this, experimentally reconstructing a state’s Wigner function (typically via homodyne-detection-based tomography, Section 8) and directly observing negative

While quantum state tomography reconstructs an unknown, but fixed, quantum state, quantum process tomography instead aims to fully characterize an unknown quantum process – a gate, channel, or general CPTP map (intro- duced in the Quantum Information deck) – by systematically preparing a complete, informationally sufficient set of known input states, sending each one through the unknown process, and then performing full state tomography on each resulting output state. Combining the complete set of known-input/measured-output pairs allows the un- known process itself (represented, for instance, via its Kraus operators, or equivalently its full process matrix) to be fully, uniquely reconstructed – an essential, standard experimental characterization technique for validating that an actually-implemented quantum gate or channel genuinely performs its intended operation with high fidelity, though its resource cost (both in the number of distinct input states needed and the number of tomographic measurements per output state) scales considerably worse than simple state tomography alone, motivating more resource-efficient, partial-characterization alternatives (like randomized benchmarking, introduced in the Quantum Gates deck) for rou- tine, large-scale hardware characterization where full process tomography quickly becomes impractical.

BEC is a phase transition, predicted theoretically by Bose and Einstein in the 1920s, in which a dilute gas of identical bosonic atoms, cooled below a specific critical temperature, undergoes a dramatic, qualitative change in behavior: rather than the atoms’ individual quantum states remaining thermally spread across a broad range of available energy levels (as an ordinary classical or even a typical quantum gas would), a macroscopically large fraction of all the atoms in the entire sample spontaneously occupies the exact same lowest-energy single-particle quantum state simul- taneously. This dramatic collective behavior occurs specifically because, unlike fermions (which are forbidden by the Pauli exclusion principle – discussed in the Advanced Quantum Mechanics deck – from ever occupying the same quantum state), bosons face no such restriction and, once their individual thermal de Broglie wavelengths become comparable to the typical inter-particle spacing (occurring at sufficiently low temperature and/or high density), the atoms’ individual quantum wavefunctions begin to overlap substantially, and the system’s lowest-energy configura- tion becomes overwhelmingly favored to have essentially all the atoms condense into that single, shared quantum ground state – producing, in effect, a single, macroscopic quantum object directly visible and manipulable at a scale ordinarily associated only with classical, everyday objects.

Laser cooling techniques (Doppler and sub-Doppler cooling, Section 4) are fundamentally limited by the recoil an atom receives from each spontaneously emitted photon during the cooling process itself, setting a practical floor (the ”recoil limit”) on achievable temperatures that, while impressively low (microkelvin range), remains far too warm to actually achieve Bose-Einstein condensation for typical atomic densities. Evaporative cooling provides the essential final cooling stage: atoms are held in a (typically magnetic or optical) trap, and the trap’s confining potential is deliberately, gradually lowered over time, selectively allowing only the hottest, most energetic atoms in the trapped sample’s thermal distribution to escape entirely (carrying away a disproportionate share of the sample’s total energy) – while the remaining, cooler atoms collide and re-thermalize among themselves to a progressively lower overall temperature, directly analogous to how a cup of hot coffee cools as its most energetic molecules preferentially evaporate from its surface. This technique, applied after an initial laser-cooling and trapping stage, was precisely the missing final ingredient that allowed Eric Cornell, Carl Wieman, and Wolfgang Ketterle to achieve the first laboratory Bose-Einstein condensates in 1995, earning the 2001 Nobel Prize in Physics.

An optical lattice is created by interfering two or more counter-propagating laser beams, producing a spatially periodic standing-wave pattern of light intensity – via the same AC-Stark-shift trapping mechanism underlying optical tweez- ers (Section 4), this periodic intensity pattern creates a correspondingly periodic potential-energy landscape that traps ultracold atoms at each of the pattern’s regularly-spaced intensity maxima or minima, directly mimicking the periodic potential that electrons experience moving through a real crystalline solid’s regular atomic lattice. Because essentially

A Feshbach resonance occurs when the energy of two colliding, free atoms happens to become nearly degenerate with the energy of a distinct, weakly-bound molecular state accessible via a different (closed) collision channel – and because different atomic and molecular spin states generally experience different Zeeman shifts under an applied magnetic field, this near-degeneracy condition (and thus the resulting scattering behavior) can be precisely tuned simply by adjusting the strength of an externally applied magnetic field. Near such a resonance, the effective atom- atom interaction strength (characterized by the scattering length) can be continuously tuned across an enormous range – including being tuned through zero (turning off atomic interactions almost entirely) and even flipped in sign (switching between effectively repulsive and effectively attractive interactions) – purely by adjusting an external magnetic field, with no need to change the atoms’ actual physical density or any other experimental parameter. This exceptional degree of interaction control is precisely what makes ultracold-atom platforms (Section 7’s optical-lattice quantum simulators, in particular) such uniquely powerful and flexible tools for exploring how interaction strength governs quantum many-body physics.

Homodyne detection interferes the weak optical signal field to be measured with a much stronger reference beam (the ”local oscillator,” derived from, and phase-locked to, the exact same laser source generating the signal) on a 50:50 beamsplitter, then measures the difference in photocurrent between the two resulting output ports using a balanced pair of photodetectors. The resulting difference-photocurrent signal is directly, linearly proportional to a specific field quadrature of the original signal field – with which particular quadrature actually measured determined by the local oscillator’s own relative phase, which the experimenter can freely and precisely control – and because the local oscillator’s much greater intensity effectively amplifies the otherwise very weak signal-field information above the electronic detector noise floor, homodyne detection allows the precise measurement of even extremely weak quantum optical fields (including, critically, squeezed-light quadratures, Section 1 of the Quantum Nanophotonics deck) that would be far too weak to measure directly and reliably with an ordinary single, unassisted photodetector alone.

Extending simple homodyne detection (which measures just one specific field quadrature, set by a single fixed local- oscillator phase) into a full tomographic reconstruction technique, optical homodyne tomography systematically scans the local oscillator’s phase across a full range of settings (from 0 to 2π), recording the resulting quadrature- measurement probability distribution at each distinct phase setting. Combining this complete set of phase-resolved quadrature distributions via a mathematical reconstruction procedure (closely analogous to the inverse Radon trans- form used in classical medical X-ray computed-tomography imaging, from which the technique borrows both its con- ceptual approach and its ”tomography” name) allows the full, complete Wigner function (Section 6) of the original optical field to be reconstructed – making optical homodyne tomography the standard, most widely used experimental technique specifically for directly measuring and visualizing genuinely non-classical light states’ full Wigner-function representation, including any characteristic negative-valued regions signaling non-classicality.

An ordinary projective measurement, per the standard measurement postulate (introduced in the Advanced Quan- tum Mechanics deck), generally collapses a quantum state – while a repeated, immediate re-measurement of the exact same observable does reproduce the same outcome (since the state has already collapsed to that observable’s corresponding eigenstate), the measurement process itself typically disturbs other, complementary properties of the system, and for many observables of genuine practical interest (particularly photon number, measured via ordinary photon absorption in a standard detector), the measurement process itself inherently destroys the very quantity being measured (an absorbed photon is, definitionally, gone, precluding any further measurement of it at all). A QND measurement is specifically engineered so that the same observable can be measured repeatedly, with each subse- quent measurement reliably reproducing the same result and leaving that specific observable’s value undisturbed for future measurement – Haroche’s cavity-photon-counting experiments (Section 5), which measured photon number

Building on the general entanglement-swapping concept introduced in the Quantum Cryptography deck’s quantum- repeater discussion, a laboratory entanglement-swapping experiment typically starts with two independent entangled photon pairs (each generated via SPDC, as introduced in the Quantum Nanophotonics deck) – photon pair (A,B) from one source, and photon pair (C,D) from a separate, independent source. Bringing photons B and C together at a beamsplitter and performing a joint Bell-state measurement on them (a specific joint measurement that projects the B-C pair onto one of the four Bell states, exploiting the same Hong-Ou-Mandel-type two-photon interference in- troduced in the Quantum Nanophotonics deck) causes photons A and D – which were generated at entirely separate sources and have never directly interacted with each other at all – to become genuinely entangled with one another, conditioned specifically on that particular Bell-measurement outcome having actually occurred (hence ”heralded”). This laboratory technique is precisely the core, essential building block underlying both quantum teleportation ex- periments and the quantum-repeater architectures discussed throughout the Quantum Cryptography and Quantum Information decks, and has been experimentally demonstrated over progressively longer distances as a key milestone on the path toward a practical, extended-range quantum internet. End of deck.