xARPES: Expanded Photoemission Spectroscopy

• xARPES is an expanded ARPES method that augments traditional measurements with spin, micro/nano-resolution, femtosecond time resolution, and enhanced depth sensitivity.
• It integrates advanced detection modalities and in-situ synthesis workflows to enable detailed multi-dimensional probing of electronic structures in quantum materials.
• xARPES also leverages computational tools to extract self-energies and Eliashberg functions, offering a quantitative approach to many-body renormalization analysis.

Searching arXiv for recent and foundational papers on xARPES and advanced ARPES modalities. xARPES denotes an expanded angle-resolved photoemission spectroscopy repertoire in which the standard measurement of occupied electronic structure is augmented by additional observables or operating dimensions, including spin, micrometer or nanometer lateral resolution, femtosecond time resolution, soft- and hard-X-ray depth sensitivity, and in-situ synthesis/characterization workflows; in recent literature, the name also designates a Python code for extracting self-energies and Eliashberg functions from ARPES data (Zhang et al., 2022, Waas et al., 19 Aug 2025). Across these uses, the unifying objective is to extend the measurable electronic structure from $E(\mathbf{k})$ alone to a higher-dimensional description involving $A(\mathbf{k},\omega)$, matrix-element control, spatial heterogeneity, nonequilibrium dynamics, buried interfaces, and many-body renormalization. The term should be distinguished from X-ARAPUCA, a liquid-argon photon detector for DUNE that is unrelated to photoemission spectroscopy (Palomares et al., 2022).

1. Scope and nomenclature

In the modern condensed-matter literature, ARPES is presented as a technique whose core observables have been extended “to spin (SpinARPES), micrometer or nanometer lateral dimensions (MicroARPES/NanoARPES), and femtosecond timescales (TrARPES)” (Zhang et al., 2022). Soft- and hard-X-ray implementations add enhanced probing depth for buried layers and improved $k_z$ definition, while microprobe and PEEM-based variants localize the measurement in real space (Strocov et al., 2019, Menteş et al., 2012). In parallel, integrated growth-and-measurement platforms connect oxide molecular beam epitaxy, ARPES, and scanning tunneling microscopy under ultrahigh vacuum, extending xARPES toward engineered heterostructures and contamination-free workflows (Kim et al., 2018, Yang et al., 2021).

A plausible implication is that xARPES is best understood not as a single hardware class but as a family of ARPES extensions that add new conjugate variables, new length or time scales, or new analysis formalisms. That breadth is explicit in the coexistence of experimental modalities such as tr-SARPES, SX-ARPES, autonomous microARPES, and angle-resolved XPEEM, together with the computational \textsc{xARPES} code for self-energy inversion (Kawaguchi et al., 2023, Agustsson et al., 2024, Waas et al., 19 Aug 2025).

Modality	Added capability	Representative details
SpinARPES / tr-SARPES	Spin polarization vector and ultrafast spin dynamics	VLEED detectors; 1-MHz, 10.7-eV probe; 25 meV and 360 fs (Kawaguchi et al., 2023)
MicroARPES / NanoARPES	Spatially resolved band mapping	Focused beam to micron or nanometer scales (Zhang et al., 2022)
TrARPES	Pump-probe access to nonequilibrium states	$\Delta E\,\Delta t \gtrsim \hbar/2$ tradeoff (Zhang et al., 2022)
SX/HX-ARPES	Buried interfaces, improved $k_z$, chemical specificity	$h\nu \sim 0.2\!-\!2$ keV or $>2$ keV (Zhang et al., 2022)
$\mu$-ARPES / df-XPEEM	Local $k$-space imaging in inhomogeneous materials	Few tens of nm lateral resolution in df-XPEEM (Menteş et al., 2012)
In-situ ARPES	Growth-to-measurement without vacuum break	OMBE–ARPES–STM integration (Kim et al., 2018, Yang et al., 2021)

2. Measurement framework and observables

The common basis of all xARPES variants remains the photoelectric effect and the reconstruction of initial-state electronic structure from measured electron energy and emission angle. In the standard formulation,

$E_k = h\nu - \phi - E_B$

and

$A(\mathbf{k},\omega)$0

with more explicit geometric relations for $A(\mathbf{k},\omega)$1 and $A(\mathbf{k},\omega)$2 available for specific analyzer geometries (Zhang et al., 2022). The measured photocurrent is written as

$A(\mathbf{k},\omega)$3

so xARPES always couples the spectral function to dipole matrix elements and experimental resolution (Zhang et al., 2022, Kordyuk, 2014).

The spectral function formalism is central because it makes xARPES a direct probe of many-body renormalization rather than only of bare band dispersion. A standard expression is

$A(\mathbf{k},\omega)$4

where $A(\mathbf{k},\omega)$5 encodes quasiparticle renormalization and lifetime effects (Zhang et al., 2022). For fermiology-oriented analyses, this framework underlies MDC and EDC fitting, the extraction of kinks and linewidths, and the identification of ordering-induced Fermi-surface reconstruction (Kordyuk, 2014).

What differentiates xARPES modalities is how they enlarge the observable manifold around this same kernel. SpinARPES resolves the spin polarization vector as a function of $A(\mathbf{k},\omega)$6 and $A(\mathbf{k},\omega)$7; MicroARPES and NanoARPES add real-space coordinates; TrARPES adds pump-probe delay $A(\mathbf{k},\omega)$8; SX/HX-ARPES modify probing depth and $A(\mathbf{k},\omega)$9 sensitivity through higher kinetic energies; resonant photoemission adds elemental and chemical selectivity at absorption edges (Zhang et al., 2022, Strocov et al., 2019). In superconducting and nonequilibrium contexts, the observable intensity may be modeled as a sum of surface and bulk spectral functions convolved with instrumental resolution and supplemented by empirical background terms, making explicit that gap visibility depends on both intrinsic self-energy and extrinsic background conditions (Kuibarov et al., 14 May 2025).

3. Instrumentation and end-station architectures

High-performance xARPES depends on beamline brightness, polarization control, energy resolution, focusing optics, sample manipulation, and analyzer throughput. At Diamond Light Source, a synchrotron radiation beamline in the photon energy range of 18–240 eV was constructed with a 5-meter long Apple-II type undulator capable of linear horizontal, linear vertical, circular left, and circular right polarization at all photon energies, a collimated plane grating monochromator, a re-focussing system forming a 50 $k_z$0 50 $k_z$1 beam spot, and a six-degrees-of-freedom cryogenic manipulator (Hoesch et al., 2016). The facility reports energy resolution below 2 meV at 60 eV with the smallest slit aperture, combined ARPES resolution of $k_z$2 meV demonstrated as $k_z$3 meV in an Au Fermi-edge measurement at 20 eV and 6 K, 10–15 meV for fast $k_z$4-space mapping, and photon flux up to $k_z$5 ph/s for mapping (Hoesch et al., 2016). The same platform emphasizes polarization-dependent ARPES, dichroism, precise sample rotations reproducible to better than $k_z$6, and temperature control from 7 K to 350 K (Hoesch et al., 2016).

The BL03U endstation at SSRF exemplifies a complementary architecture optimized for small spot size, super-high energy resolution, and in-situ sample handling. Its synchrotron beamline covers 7–165 eV in the fundamental, with access up to 798 eV via higher-order monochromator gratings, a spot size of 7.5 $k_z$7m (vertical) $k_z$8 67 $k_z$9m (horizontal), monochromator resolution of 0.47 meV at 21.6 eV, total ARPES resolution of 2.67 meV at 21.2 eV, and a six-axis cryogenic manipulator reaching 7 K (Yang et al., 2021). Integration with oxide molecular beam epitaxy and low-temperature STM/STS allows direct transfer of as-grown films under UHV, with ozone delivery, RHEED monitoring, and alkali-doping capability (Yang et al., 2021).

Brookhaven’s OASIS system extends this integrated logic by interconnecting oxide MBE, ARPES, and spectroscopic-imaging STM through UHV transfer lines and a “Grand Central Station” chamber, using a universal sample holder compatible with high-temperature ozone-rich growth and cryogenic spectroscopy (Kim et al., 2018). This architecture was designed specifically to enlarge the range of complex oxides and artificial heterostructures accessible to ARPES and STM without breaking vacuum (Kim et al., 2018). A plausible implication is that advanced xARPES increasingly depends on facility-scale co-design of growth, transfer, and spectroscopy rather than on analyzer performance alone.

4. Principal modalities of xARPES

SpinARPES augments ARPES with spin polarimetry. The standard detector classes are Mott, very-low-energy electron diffraction, and exchange-scattering polarimeters; the signal undergoes substantial intensity loss, with a figure of merit $\Delta E\,\Delta t \gtrsim \hbar/2$0, although “recent two-order-of-magnitude efficiency improvements” have extended applicability (Zhang et al., 2022). A concrete realization is the tr-SARPES setup employing a 10.7-eV probe laser at 1 MHz together with VLEED spin detectors. That system combines a Yb:fiber laser, harmonic generation in Xe gas, and a DA30L hemispherical analyzer; it achieves photon flux of $\Delta E\,\Delta t \gtrsim \hbar/2$1 photons/s, ARPES-mode resolution of 22 meV, SARPES-mode resolution of 25 meV, 360 fs pump-probe cross-correlation, and momentum coverage up to about $\Delta E\,\Delta t \gtrsim \hbar/2$2, while using the high repetition rate to reduce space-charge effects (Kawaguchi et al., 2023).

MicroARPES and NanoARPES focus the beam to micron or nanometer scales, typically using Fresnel zone plates or related focusing optics and piezo-driven stages for nanometric positioning (Zhang et al., 2022). Autonomous microARPES at the SGM4 micro-focus beamline of ASTRID2 adds Bayesian experimental control: a focused spot better than $\Delta E\,\Delta t \gtrsim \hbar/2$3m, a 6-axis piezoelectric stick-slip stage with $\Delta E\,\Delta t \gtrsim \hbar/2$4 nm repeatability, and Gaussian process regression over $\Delta E\,\Delta t \gtrsim \hbar/2$5 to select measurement points that maximize mean intensity, sharpness, curvature, or combined task functions relative to movement cost (Agustsson et al., 2024). This protocol searches both real and $\Delta E\,\Delta t \gtrsim \hbar/2$6-space and is explicitly designed to reduce the time burden of high-dimensional mapping (Agustsson et al., 2024).

SX-ARPES and HX-ARPES shift the photon energy into the soft- and hard-X-ray regimes to access buried layers, impurities, and three-dimensional bulk band structure (Zhang et al., 2022). Around 1 keV, the inelastic mean free path increases by a factor of 3–5 relative to VUV-ARPES, the probing depth is approximately $\Delta E\,\Delta t \gtrsim \hbar/2$7, and the larger escape depth improves the definition of $\Delta E\,\Delta t \gtrsim \hbar/2$8 through $\Delta E\,\Delta t \gtrsim \hbar/2$9 (Strocov et al., 2019). In one representative experiment, SX-ARPES at $k_z$0 eV exposed coherent GaAs:Be bulk band dispersion through an amorphous As passivation layer of about 1 nm without surface cleaning, while any interface-state contribution remained below the detection limit (Kobayashi et al., 2012). In oxide interfaces, resonant SX-ARPES at the Ti $k_z$1-edge resolves $k_z$2, $k_z$3, and $k_z$4 subbands of the buried LaAlO$k_z$5/SrTiO$k_z$6 interface 2DES and identifies peak-dip-hump spectral structure associated with Holstein-type large polarons (Strocov et al., 2016).

Angle-resolved XPEEM and microprobe diffraction provide a distinct branch of xARPES for laterally inhomogeneous materials. Energy-filtered PEEM can switch between real-space imaging and localized $k_z$7-ARPES, while darkfield XPEEM selects off-normal photoelectrons to image real-space variations of specific $k_z$8-space states away from $k_z$9, with few tens of nm lateral resolution (Menteş et al., 2012). This makes local electronic-structure imaging possible in systems where conventional ARPES would average over domains, interfaces, or adsorption-site inequivalence (Menteş et al., 2012).

5. Analysis formalisms, denoising, and inversion

Theoretical interpretation of xARPES has evolved in parallel with instrumentation. The one-step model of photoemission treats excitation, transport, and escape as a single coherent multiple-scattering process, with the photocurrent written using Pendry’s formula,

$h\nu \sim 0.2\!-\!2$0

Within this framework, fully relativistic KKR captures Rashba splitting, CPA treats substitutional disorder, DMFT supplies local correlation self-energies, CPA+DMFT treats correlated alloys, and explicit electron-phonon self-energies and thermal vibrations extend the model toward quantitative HAXPES and temperature-dependent spectra (Minár et al., 2010).

The \textsc{xARPES} code systematizes many-body extraction from measured ARPES line shapes. It models the spectral function as

$h\nu \sim 0.2\!-\!2$1

decomposes the self-energy into phonon, electron-electron, and impurity contributions, and uses a maximum-entropy method with Bayesian inference to optimize the bare dispersion and interaction parameters for curved as well as linear bands (Waas et al., 19 Aug 2025). A key methodological point is explicit: Lorentzian MDC fitting is exact only for linear bands with $h\nu \sim 0.2\!-\!2$2-independent self-energy, whereas direct fitting of parabolic or polynomial dispersions avoids systematic errors for curved bands (Waas et al., 19 Aug 2025). The code was benchmarked on synthetic data and applied to TiO$h\nu \sim 0.2\!-\!2$3-terminated SrTiO$h\nu \sim 0.2\!-\!2$4 and Li-doped graphene, where it achieved “unprecedented agreement” between Eliashberg functions extracted from symmetry-related dispersions (Waas et al., 19 Aug 2025).

Data-quality enhancement is now an additional analytical axis of xARPES. In soft-X-ray ARPES at SPring-8 BL25SU, a deep prior-based denoising method using a 4-layer U-shaped CNN with skip connections was integrated into a micro-focused SX-ARPES system to eliminate grid and spike noise in voltage Fixed-mode images within about 30 seconds (Yamagami et al., 28 Nov 2025). For CeRu$h\nu \sim 0.2\!-\!2$5Si$h\nu \sim 0.2\!-\!2$6, statistically reliable ARPES data were accumulated in about 40 seconds and denoised within another 30 seconds; compared with Swept Mode requiring more than 45 minutes, this was reported as “over a 60x speedup” (Yamagami et al., 28 Nov 2025). In polycrystalline Au at $h\nu \sim 0.2\!-\!2$7 eV, the same workflow achieved a total energy resolution of 51.6 meV (Yamagami et al., 28 Nov 2025). A plausible implication is that analysis pipelines are becoming part of the instrument, not merely post-processing.

6. Scientific applications, interpretive limits, and outlook

One prominent xARPES application is the detection of weak superconducting gaps under constrained experimental conditions. In synchrotron ARPES on $h\nu \sim 0.2\!-\!2$8-PtBi$h\nu \sim 0.2\!-\!2$9, horizontally polarized photons of 13–20 eV and an energy resolution of 2–4 meV were used at 1.5 K; photon energies of 17–20 eV probed near the $>2$0 point where surface Fermi arcs are separated from bulk states, enabling detection of a surface superconducting gap with $>2$1 meV even under more relaxed conditions than previous laser ARPES (Kuibarov et al., 14 May 2025). The same work states minimum requirements for observing the gap: energy resolution $>2$2 meV, low background and high surface selectivity, optimized geometry, and favorable signal-to-background ratio; it also emphasizes that the leading-edge shift does not directly equal the true gap size (Kuibarov et al., 14 May 2025). This is an instructive methodological limit that generalizes beyond the specific material.

Buried heterostructures and oxide interfaces form another major domain. SX-ARPES revealed coherent three-dimensional GaAs:Be bulk bands through a $>2$3 nm amorphous As cap with no surface treatment (Kobayashi et al., 2012). In buried semiconductor and oxide systems more generally, SX-ARPES was shown to image quantum-well states through barriers of order 30 Å, resolve bosonic coupling in oxide 2DESs, and use resonant enhancement to isolate dilute impurity bands (Strocov et al., 2019). At the LaAlO$>2$4/SrTiO$>2$5 interface, resonant SX-ARPES at the Ti $>2$6-edge exposed orbital-selective subbands, peak-dip-hump structure, temperature-dependent quasiparticle collapse, oxygen-vacancy-derived in-gap states, and spectroscopic signatures of electronic phase separation (Strocov et al., 2016).

Thin-film and engineered-heterostructure studies illustrate the in-situ branch of xARPES. The OASIS platform at Brookhaven enabled ARPES on a 10-unit-cell La$>2$7Sr$>2$8CuO$>2$9 film grown by oxide MBE, showing a sharply defined closed Fermi-surface contour around the Brillouin-zone center and well-defined quasiparticle dispersions consistent with a strongly overdoped metallic LSCO film (Kim et al., 2018). Similar integrated logic appears at SSRF, where OMBE, ARPES, and STM/STS are co-located to study iridates, magnetic topological insulators, and in-situ alkali-doped superconductors (Yang et al., 2021).

The principal controversies and misconceptions are mostly interpretive rather than definitional. xARPES does not remove matrix-element effects; rather, tunable polarization and photon energy make them experimentally useful while also making intensity non-universal (Zhang et al., 2022, Kordyuk, 2014). Greater probing depth in SX/HX-ARPES is not a monotonic improvement, because it comes with lower photoexcitation cross-sections and reduced energy and $\mu$0-resolution unless offset by very high photon flux and efficient detection (Strocov et al., 2019, Zhang et al., 2022). Likewise, higher-dimensional xARPES does not make acquisition intrinsically easier: spin detection remains inefficient, NanoARPES is flux-limited and alignment-sensitive, TrARPES is constrained by the $\mu$1 tradeoff, and SX/HX-ARPES is photon-hungry (Zhang et al., 2022). The contemporary response to these constraints is visible in autonomous acquisition, denoising, Bayesian inversion, integrated endstations, and proposals for spin-resolved SX-ARPES and in-operando field-effect measurements (Agustsson et al., 2024, Yamagami et al., 28 Nov 2025, Strocov et al., 2016).

A plausible implication is that xARPES is evolving toward a genuinely multidimensional spectroscopy in which source engineering, endstation integration, adaptive control, and inverse modeling are co-equal with the photoemission measurement itself. In that sense, the subject is less a single technique than a continuously expanding framework for experimentally resolving the momentum-, spin-, space-, time-, and depth-dependent electronic structure of quantum materials (Zhang et al., 2022).