Condensed Matter: Experimental Techniques

Condensed Matter Physics » Experimental Techniques

Experimental Techniques

Why probes, and what they each “see”

A theory of a solid is only as good as our ability to test it. Every experimental technique in condensed matter is, at bottom, a way of coupling an external field — photons, electrons, neutrons, a magnetic field, a temperature gradient — to a material and reading off the response. What makes a probe powerful is which correlation function it measures and in what variables it is resolved (energy, momentum, position, or temperature).

The single most useful organizing idea is the linear-response / fluctuation–dissipation picture. A weak external perturbation that couples to some operator $\hat{B}$ produces a measurable change in the conjugate observable $\hat{A}$, governed by the retarded response function

Scattering and spectroscopic experiments measure the dynamical structure factor, the imaginary part of $\chi$ weighted by a Bose factor,

so that a scattering experiment with momentum transfer $\mathbf{q}$ and energy transfer $\hbar\omega$ directly maps out the spectrum of the material’s excitations. The table below summarizes the division of labor; the rest of the page treats each probe in detail.

ARPES — mapping band structure and self-energy

Angle-resolved photoemission spectroscopy is the most direct experimental window onto the electronic band structure $E(\mathbf{k})$. A monochromatic photon of energy $h\nu$ ejects an electron from the solid; by measuring the photoelectron’s kinetic energy and emission angle, one reconstructs the energy and crystal momentum the electron had inside the material.

Kinematics

Energy conservation fixes the binding energy relative to the Fermi level,

where $W$ is the work function. Crucially, the component of momentum parallel to the surface is conserved across the sample boundary (the surface breaks translational symmetry only along the normal):

Sweeping the emission angle $\theta$ thus sweeps $k_\parallel$, and a modern hemispherical analyzer records intensity as a 2D image over $(E_{kin},\theta)$ — i.e. a slice of the band structure $E(k_\parallel)$ in a single shot. The perpendicular component $k_\perp$ is not conserved and must be inferred (e.g. by varying $h\nu$, often using a nearly-free-electron final state), which is why ARPES is at its sharpest for quasi-2D materials (cuprates, graphene, transition-metal dichalcogenides) where $k_\perp$ dispersion is weak.

What is actually measured: the spectral function

In the sudden approximation the photocurrent is proportional to the single-particle spectral function times the Fermi function and a matrix element:

The spectral function is the imaginary part of the retarded Green’s function,

For a non-interacting band $A$ is a delta function pinned to the bare dispersion $\epsilon^0_\mathbf{k}$. Interactions enter entirely through the self-energy $\Sigma = \Sigma’ + i\Sigma’’$, and ARPES is essentially a machine for measuring it:

• The real part $\Sigma’$ shifts and renormalizes the dispersion. The renormalized band crosses $E_F$ where $\omega - \epsilon^0_\mathbf{k} - \Sigma’=0$; the slope ratio gives the mass enhancement $m^*/m = 1 + \lambda$, with $\lambda = -\partial\Sigma’/\partial\omega$. A “kink” in the dispersion at a phonon or magnon energy is a textbook fingerprint of electron–boson coupling.
• The imaginary part $\Sigma’’$ sets the linewidth: the peak width in a momentum distribution curve (MDC, fixed $\omega$) is $\Delta k = 2\Sigma’’/(\hbar v_F)$, giving the quasiparticle scattering rate directly. A Fermi liquid shows $\Sigma’’ \propto \omega^2 + (\pi k_B T)^2$; cuprate “strange metals” famously show $\Sigma’’ \propto |\omega|$ (marginal Fermi liquid).

Practical analysis: MDCs vs EDCs

Two orthogonal cuts through the $I(\mathbf{k},\omega)$ image are used:

• EDC (energy distribution curve, fixed $\mathbf{k}$): natural for reading off gaps — superconducting gaps, the cuprate pseudogap, charge-density-wave gaps — as a suppression of weight at $E_F$ and a coherence peak at $\pm\Delta$.
• MDC (momentum distribution curve, fixed $\omega$): a Lorentzian whose center traces the dispersion and whose width gives $\Sigma’’$. Because the bare dispersion is locally linear, MDCs decouple $\Sigma’$ (peak position) from $\Sigma’’$ (peak width) cleanly, which is why self-energy extraction is usually done from MDCs.

Fermi surfaces and modern variants

Integrating $I(\mathbf{k},\omega)$ over a narrow window at $E_F$ and plotting versus $(k_x,k_y)$ produces a direct image of the Fermi surface — bright contours wherever a band crosses $E_F$. Spin-resolved ARPES adds a Mott or VLEED spin detector to map the spin texture, the decisive evidence for spin–momentum locking on topological-insulator surfaces. Time-resolved ARPES (pump–probe) populates and watches unoccupied states relax, accessing the band structure above $E_F$ and ultrafast dynamics.

STM / STS — real-space local density of states

Where ARPES resolves momentum, scanning tunneling microscopy resolves position — down to single atoms. A sharp metal tip is brought within a nanometer of a conducting surface; the exponentially small overlap of tip and sample wavefunctions lets electrons quantum-tunnel across the vacuum gap, producing a current that is exponentially sensitive to the tip–sample separation $d$:

A typical $\kappa \approx 1\ \text{\AA}^{-1}$ means the current changes by an order of magnitude per Angstrom — the origin of STM’s sub-atomic vertical resolution. Holding $I$ constant with a feedback loop while raster-scanning gives a topographic image of the surface (more precisely a contour of constant integrated LDOS).

Tunneling spectroscopy and the LDOS

Within the Tersoff–Hamann picture and a flat tip DOS, the tunneling current integrates the local density of states of the sample over the bias window:

Differentiating with respect to bias gives the centerpiece of scanning tunneling spectroscopy (STS) — the differential conductance is, to good approximation, the LDOS at energy $eV$ measured at the atomic position $\mathbf{r}$:

By taking a full $dI/dV$ spectrum at every pixel one builds a spectroscopic map: the spatial variation of the density of states at a chosen energy. This is how superconducting gap inhomogeneity in cuprates, vortex cores, and impurity-bound states are imaged atom by atom.

Quasiparticle interference (QPI)

Defects scatter Bloch electrons, and the interference of incoming and outgoing waves prints standing-wave ripples in the LDOS at wavevector $\mathbf{q} = \mathbf{k}_f - \mathbf{k}_i$. Fourier-transforming a $dI/dV$ map turns these ripples into bright spots whose positions encode the joint density of states of the band structure:

QPI thereby recovers momentum-space information (constant-energy contours, gap anisotropy, even the sign structure of an order parameter) from a real-space measurement — a beautiful complement to ARPES, and applicable to buried or non-cleavable Fermi surfaces ARPES cannot reach.

Neutron scattering — structure and magnetic order

Neutrons are the workhorse for structure because they carry no charge (so they penetrate deep into bulk samples and scatter from nuclei, not the electron cloud) and because they carry a magnetic moment (so they scatter directly from electronic spins). Their de Broglie wavelength at thermal energies is $\sim 1\text{–}2\ \text{\AA}$ — comparable to interatomic spacings — and their energy at those wavelengths is $\sim$ meV — comparable to phonon and magnon energies. A single instrument therefore resolves both where atoms and spins sit and how they move.

Elastic scattering: crystal and magnetic structure

In a diffraction (elastic) experiment the scattered intensity is concentrated at reciprocal-lattice vectors $\mathbf{G}$, weighted by the structure factor:

where $b_j$ is the nuclear scattering length of atom $j$ and $e^{-W_j}$ is the Debye–Waller factor. Because $b_j$ varies erratically (not monotonically) with atomic number, neutrons locate light atoms next to heavy ones (hydrogen, oxygen, lithium) where X-rays struggle, and distinguish neighboring elements and isotopes.

The magnetic moment of the neutron adds magnetic Bragg peaks. When spins order with a periodicity different from the lattice, new peaks appear at the magnetic propagation vector $\mathbf{k}$, often at half-integer positions for an antiferromagnet. Their intensity is governed by the magnetic structure factor with the all-important polarization factor:

The transverse projector $\delta_{\alpha\beta}-\hat{Q}\alpha\hat{Q}\beta$ means neutrons only see the spin component perpendicular to $\mathbf{Q}$ — by measuring at several $\mathbf{Q}$ one reconstructs the full spin direction, making neutron diffraction the definitive probe of magnetic order (the very tool with which antiferromagnetism was first confirmed). Polarized neutrons further separate nuclear from magnetic and longitudinal from transverse channels.

Inelastic scattering: phonons and magnons

When the neutron exchanges energy with the sample it measures the dynamical structure factor directly:

with $S(\mathbf{Q},\omega)$ as defined in the overview. Scanning $(\mathbf{Q},\omega)$ on a triple-axis or time-of-flight spectrometer maps dispersion relations: acoustic and optical phonon branches from nuclear (coherent) scattering, and magnon / spin-wave branches from the magnetic cross-section. Spin liquids and quantum-critical systems instead show a broad continuum of $S(\mathbf{Q},\omega)$ — the smoking gun of fractionalized (e.g. spinon) excitations — which is one of the most distinctive results only neutrons can deliver.

Inelastic & Raman scattering — excitations near zero momentum

Light-scattering probes complement neutrons by accessing excitations with extreme energy resolution but at essentially zero momentum transfer (the photon wavevector is tiny on the scale of the Brillouin zone). They are table-top, fast, and exquisitely sensitive to symmetry.

Raman scattering

In Raman scattering a visible photon is inelastically scattered, shifting in frequency by the energy of an excitation it creates (Stokes) or absorbs (anti-Stokes):

The Stokes/anti-Stokes intensity ratio is set by thermal occupation, $I_{aS}/I_{S} = e^{-\hbar\Omega/k_B T}$, providing an internal thermometer. The measured cross-section is governed by the Raman response $\chi’‘_{\gamma}(\omega)$, where the symmetry of the light-polarization geometry selects a particular irreducible representation:

Because each excitation transforms as a definite irreducible representation of the crystal point group, choosing incident/scattered polarizations ($A_{1g}$, $B_{1g}$, $B_{2g}$, …) filters phonons, magnons, and electronic continua by symmetry. This is how Raman fingerprints phonon modes (and through them lattice symmetry, strain, layer number in 2D materials), two-magnon scattering in antiferromagnets, the electronic continuum and pair-breaking $2\Delta$ peak in superconductors, and amplitude (Higgs) modes of order parameters.

Infrared and optical conductivity

Infrared spectroscopy measures absorption/reflection, from which a Kramers–Kronig analysis yields the complex optical conductivity $\sigma(\omega)=\sigma_1+i\sigma_2$. The low-frequency Drude peak,

gives the scattering rate $1/\tau$ and plasma frequency (hence carrier density), while gaps appear as a clean suppression of $\sigma_1$ below a threshold. The optical sum rule $\int_0^\infty \sigma_1(\omega)\,d\omega = \pi n e^2/2m$ ties spectral-weight transfer to correlation physics. IR thus reads off charge gaps, the Drude weight, phonon and interband features, and the redistribution of spectral weight at phase transitions.

Quantum oscillations — the Fermi surface

In a strong magnetic field the electronic states condense into Landau levels, and as the field is swept these levels pass through the Fermi energy one by one. Every time a level crosses $E_F$ the density of states at $E_F$ spikes, and essentially every physical property — magnetization (de Haas–van Alphen), resistivity (Shubnikov–de Haas), magnetostriction, sound velocity — oscillates. These quantum oscillations are the cleanest, most quantitative map of the Fermi surface available.

Onsager relation: oscillation frequency = Fermi-surface area

The oscillations are periodic in $1/B$, and the Onsager relation ties their frequency $F$ directly to an extremal cross-sectional area $A_{ext}$ of the Fermi surface perpendicular to the field:

Rotating the sample maps how $F(\theta)$ varies, tracing out the full 3D shape of the Fermi surface. Each distinct extremal orbit contributes its own frequency, so a Fourier transform of the signal in $1/B$ reveals the Fermi-surface “fingerprint” of the metal.

Lifshitz–Kosevich: mass and scattering

The amplitude of each oscillation is described by the Lifshitz–Kosevich formula,

with three damping factors that turn the amplitude into a measurement of microscopic quantities:

• Thermal factor $R_T = X/\sinh X$ with $X = 2\pi^2 k_B T\, m^/\hbar e B$. Fitting the temperature dependence of the amplitude yields the cyclotron effective mass $m^$ — a direct measure of mass renormalization, hence correlation strength.
• Dingle factor $R_D = e^{-2\pi^2 k_B T_D\, m^*/\hbar e B}$. The field dependence gives the Dingle temperature $T_D$ and thus the quantum scattering rate / mean free path.
• Spin factor $R_S = \cos(\pi g m^*/2 m_e)$, encoding Zeeman splitting and the $g$-factor.

The phase $\phi$ carries a Berry-phase offset: a $\pi$ Berry phase (shifting $\phi$ by $1/2$) is a hallmark of Dirac/Weyl fermions, so quantum oscillations also test topological band structure.

Transport measurements

Transport is the most accessible and historically the first window onto a solid’s electronic state — it asks how charge and heat flow in response to electric fields, magnetic fields, and temperature gradients.

Resistivity and the four-probe method

DC resistivity is measured with a four-probe geometry: current is driven through the outer two contacts and voltage read across the inner two, so that contact and lead resistances drop out of the measurement. The temperature dependence is diagnostic of the ground state and its excitations:

• Metal: $\rho(T) = \rho_0 + AT^2$ (Fermi-liquid electron–electron scattering) or $\rho \propto T^5$ (phonon-limited, Bloch–Grüneisen) at low $T$; the residual $\rho_0$ measures disorder.
• Semiconductor / insulator: activated, $\rho \propto e^{E_g/2k_B T}$, giving the gap; variable-range hopping $\rho \propto e^{(T_0/T)^{1/4}}$ signals localization.
• Superconductor: $\rho \to 0$ below $T_c$.
• Strange metal: linear $\rho \propto T$ over a wide range — a defining anomaly of cuprates and other quantum-critical systems.

Hall effect: carrier type and density

A magnetic field perpendicular to the current deflects carriers, building a transverse Hall voltage. In the simplest single-band picture the Hall coefficient gives the sign and density of carriers directly:

A positive $R_H$ signals hole-like, negative electron-like conduction. Combined with the conductivity it yields the mobility $\mu = |R_H|\,\sigma$. The Hall response is also the gateway to topological transport — the quantized Hall plateaus $\sigma_{xy} = \nu e^2/h$, and the anomalous Hall effect proportional to Berry curvature in magnetic conductors.

Magnetotransport

Beyond the Hall effect, the field dependence of the longitudinal resistance (magnetoresistance) reveals multiband effects, Fermi-surface topology (via Shubnikov–de Haas oscillations, above), weak localization/antilocalization (a quantum-coherence and spin–orbit diagnostic), and chiral anomalies in Weyl semimetals (negative longitudinal magnetoresistance).

Thermodynamic measurements

Thermodynamic probes count degrees of freedom and entropy. They do not resolve momentum or position, but they are unmatched at identifying phase transitions and at counting the states available at the Fermi level.

Specific heat

At low temperature the specific heat of a metal separates cleanly into electronic and lattice parts:

Plotting $C/T$ versus $T^2$ gives a straight line whose intercept $\gamma$ is the Sommerfeld coefficient — proportional to the density of states at the Fermi level, $\gamma = \tfrac{\pi^2}{3} k_B^2\, g(E_F)$. A giant $\gamma$ is the defining signature of heavy-fermion materials (effective masses of hundreds of $m_e$). The slope $\beta$ gives the Debye temperature and thus the phonon spectrum. A phase transition shows up as a sharp anomaly: a mean-field jump $\Delta C$ at a superconducting $T_c$ (whose size, $\Delta C/\gamma T_c \approx 1.43$ in BCS, tests the pairing), a $\lambda$-shaped peak at a continuous magnetic transition, or a latent-heat spike at a first-order one. The entropy $S(T)=\int_0^T (C/T’)\,dT’$ released across a transition counts the participating degrees of freedom (e.g. $R\ln 2$ per spin-$\tfrac12$ moment).

Magnetization and susceptibility

The magnetization $M(H,T)$ and susceptibility $\chi = \partial M/\partial H$ classify magnetic ground states. A Curie–Weiss law,

extracts the local-moment size from the Curie constant $C$ and the dominant exchange (sign and scale) from the Weiss temperature $\theta_{CW}$. A temperature-independent Pauli susceptibility signals itinerant moments; sharp features locate ferromagnetic ($\theta_{CW}>0$) and antiferromagnetic ($\theta_{CW}<0$, Néel kink) transitions; hysteresis loops quantify coercivity and ordered moment. The frustration ratio $|\theta_{CW}|/T_N \gg 1$ flags candidate spin liquids.

Thermal expansion and magnetocalorics

Thermal expansion $\alpha = \tfrac1L\,\partial L/\partial T$ couples the entropy to volume (Grüneisen analysis) and is acutely sensitive to pressure-tuned quantum critical points, while the magnetocaloric effect ($\partial T/\partial H$ at fixed entropy) sharpens the detection of field-induced transitions where specific heat alone is ambiguous.

Choosing and combining probes

No single technique tells the whole story; the art of condensed-matter experiment is triangulation. A Fermi surface inferred from ARPES is confirmed in absolute terms by quantum oscillations and tested for momentum-averaged consistency against the Hall coefficient and $\gamma$ from specific heat. A superconducting gap seen as a coherence peak in ARPES EDCs is mapped in real space by STS, its symmetry pinned down by Raman and its $\Delta C/T_c$ jump by calorimetry. Magnetic order proposed from a Curie–Weiss susceptibility is proven — direction and all — only by neutron diffraction, with its dynamics filled in by inelastic neutron and two-magnon Raman scattering. Reading the same physics through complementary correlation functions is what turns a measurement into an understanding.

See Also

• Graduate-Level Formalism & Experiment — the Green’s-function and self-energy machinery these probes measure.
• Superconductivity, Quantum Hall & Topological Phases — the phases these techniques were built to characterize.
• Condensed Matter Physics (Hub) — crystal structure, band theory, and magnetism.
• Quantum Field Theory — linear response and correlation functions in field-theoretic language.