Tr-ARPES: Ultrafast Quantum Dynamics

• Time-resolved ARPES is an ultrafast spectroscopic technique that provides momentum- and energy-resolved snapshots of transient electronic structures in quantum materials.
• It combines femtosecond pump-probe setups with ARPES to directly observe phenomena like hot-carrier thermalization, phase transitions, and quasiparticle dynamics.
• Advances in light sources, analyzer designs, and data analysis (including machine learning) enhance its ability to study unconventional superconductors, topological systems, and more.

Time-resolved angle-resolved photoemission spectroscopy (Tr-ARPES) is an ultrafast experimental technique that enables direct, momentum-resolved access to the transient electronic structure of quantum materials following optical excitation. By combining femtosecond pump–probe capability with the momentum and energy resolution of ARPES, Tr-ARPES can track the evolution of spectral functions, occupation distributions, and many-body self-energies on femtosecond timescales, thereby providing decisive insight into nonequilibrium phenomena such as hot-carrier thermalization, collective mode dynamics, photoinduced phase transitions, and transient formation of exotic quasiparticles.

1. Theoretical Foundations and Signal Formalism

At the core of Tr-ARPES is the measurement of the instantaneous photocurrent, which, in the sudden approximation and perturbative regime, is expressed via the two-time lesser Green’s function $G^<_{k}(t_1,t_2)$ weighted by the probe envelope:

$$I(k,\omega,T)=\int dt_1 dt_2\, s(t_1-T) s(t_2-T)\, e^{i\omega(t_1-t_2)}\, G^<_{k}(t_1,t_2)$$

where $T$ is the probe center time (pump–probe delay), $s(t)$ the probe envelope (typically Gaussian), and $$G^<_{k}(t_1,t_2) = i\langle c^\dagger_{k}(t_2) c_{k}(t_1)\rangle$$ encodes the time-dependent quantum correlations (Freericks et al., 2021, Boschini et al., 2023). In equilibrium, $G^<$ depends only on $t_1-t_2$, and the signal reduces to conventional ARPES. Out of equilibrium, the full two-time structure encodes both population dynamics and changing self-energies.

Time–frequency uncertainty is intrinsic: the product of probe duration $\Delta t$ and frequency bandwidth $\Delta \omega$ satisfies $\Delta t\,\Delta\omega \gtrsim 1/2$ (Gaussian pulses), so one must balance temporal and energy resolution (Freericks et al., 2021, Boschini et al., 2023). The intensity formula generalizes to multiband systems by including band-dependent and time-dependent dipole matrix elements, which must be tracked under basis transformations for physical nonnegativity and gauge invariance (Freericks et al., 2016).

2. Light Sources, Photon Energy Ranges, and Instrumentation

Tr-ARPES experiments rely on a diverse set of ultrafast light sources to generate the needed pump and probe pulses, with fine control over photon energy, bandwidth, pulse duration, and repetition rate (Na et al., 2023, Mills et al., 2019, Hellbrück et al., 2024, Pan et al., 2023). Probe energies span from deep ultraviolet (UV, 5–7 eV) to vacuum and extreme ultraviolet (VUV, XUV, 7.0–40 eV and beyond); generation involves frequency upconversion in nonlinear crystals (BBO, KBBF), ultraviolet-driven high-harmonic generation (HHG), or cavity-enhanced HHG (Mills et al., 2019, Pan et al., 2023, Lee et al., 2019).

Key performance parameters:

Light Source	Photon Energy (eV)	Energy Res. (meV)	Time Res. (fs)	Rep. Rate	Notable Features
Nonlinear Crystal	5.3–7.2, 6.0, 7.2	8.5–48	72–320	1 Hz–100 MHz	Switchable resolution (Pan et al., 2023)
HHG (fsEC)	8–40	22–32	190	60 MHz	High flux, MHz @ table-top (Mills et al., 2019)
Xe-Gas HHG	10.7	22–25	360	1 MHz	Spin-ARPES with broad $$I(k,\omega,T)=\int dt_1 dt_2\, s(t_1-T) s(t_2-T)\, e^{i\omega(t_1-t_2)}\, G^<_{k}(t_1,t_2)$$0-range (Kawaguchi et al., 2023)
Hollow-core fiber	7.2–10.8	17–48	300–430	0.5–2 MHz	Ribbon-tunable VUV (Hellbrück et al., 2024)
Gas-jet HHG/XUV	17–45	30–150	45–70	1–6 kHz	Full 3D BZ; sub-0.01 1/Å $$I(k,\omega,T)=\int dt_1 dt_2\, s(t_1-T) s(t_2-T)\, e^{i\omega(t_1-t_2)}\, G^<_{k}(t_1,t_2)$$1-res. (Majchrzak et al., 2023)

Momentum and energy resolutions depend on analyzer choice (hemispherical, time-of-flight, or novel analyzer designs such as FeSuMa (Majchrzak et al., 2023)) and, in VUV/XUV, the photon energy (in-plane $$I(k,\omega,T)=\int dt_1 dt_2\, s(t_1-T) s(t_2-T)\, e^{i\omega(t_1-t_2)}\, G^<_{k}(t_1,t_2)$$2). Trade-offs in time-bandwidth product, repetition rate, and sample throughput are a function of source design and conversion efficiency (Mills et al., 2019, Hellbrück et al., 2024, Pan et al., 2023).

3. Experimental Protocols and Data Acquisition

The dominant measurement protocol is pump–probe, wherein a laser pulse excites the sample at $$I(k,\omega,T)=\int dt_1 dt_2\, s(t_1-T) s(t_2-T)\, e^{i\omega(t_1-t_2)}\, G^<_{k}(t_1,t_2)$$3 (pump), and after a variable time delay $$I(k,\omega,T)=\int dt_1 dt_2\, s(t_1-T) s(t_2-T)\, e^{i\omega(t_1-t_2)}\, G^<_{k}(t_1,t_2)$$4, a synchronized probe pulse initiates photoemission. The evolution $$I(k,\omega,T)=\int dt_1 dt_2\, s(t_1-T) s(t_2-T)\, e^{i\omega(t_1-t_2)}\, G^<_{k}(t_1,t_2)$$5 is then recorded for a discrete set of delays, constructing a time-resolved, momentum-resolved "movie" of quasiparticle and collective mode dynamics (Boschini et al., 2023, Zonno et al., 2021). Static and time-resolved data are acquired via energy- and angle-dispersive analyzers (e.g., hemispherical, TOF, momentum microscope, or FeSuMa), often over the entire Brillouin zone when using VUV/XUV probes with sufficiently high $$I(k,\omega,T)=\int dt_1 dt_2\, s(t_1-T) s(t_2-T)\, e^{i\omega(t_1-t_2)}\, G^<_{k}(t_1,t_2)$$6 (Hellbrück et al., 2024, Lee et al., 2019, Majchrzak et al., 2023).

Key procedural aspects:

• Pump–probe cross-correlation establishes effective time resolution: $$I(k,\omega,T)=\int dt_1 dt_2\, s(t_1-T) s(t_2-T)\, e^{i\omega(t_1-t_2)}\, G^<_{k}(t_1,t_2)$$7.
• Polarization, fluence, and wavelength selection allows excitation of specific bands, collective modes, or symmetry channels (Boschini et al., 2023, Schwarz et al., 2020).
• High-repetition rate and low per-pulse photon number minimize space-charge effects; MHz-class sources enable measurements with sub-10 meV energy broadening even at high flux (Mills et al., 2019, Kawaguchi et al., 2023, Hellbrück et al., 2024).
• Analyzer geometry, angular acceptance, and sample rotation/deflection configure the accessible $$I(k,\omega,T)=\int dt_1 dt_2\, s(t_1-T) s(t_2-T)\, e^{i\omega(t_1-t_2)}\, G^<_{k}(t_1,t_2)$$8-space.

4. Physical Interpretation of Spectra and Extracted Quantities

The Tr-ARPES signal is interpreted in terms of the time-dependent one-particle spectral function $$I(k,\omega,T)=\int dt_1 dt_2\, s(t_1-T) s(t_2-T)\, e^{i\omega(t_1-t_2)}\, G^<_{k}(t_1,t_2)$$9, the nonequilibrium occupation $T$0, and, implicitly, the self-energy $T$1 (Freericks et al., 2021, Boschini et al., 2023, Caruso et al., 2019). Out-of-equilibrium, these quantities lose strict time-translation invariance and can encode highly non-thermal, time-dependent phenomena:

• Quasiparticle dynamics: Shifts and broadening in $T$2 can be tracked to resolve transient changes in dispersion, bandgap, or lifetimes (Zonno et al., 2021, Caruso et al., 2019).
• Population relaxation: Changes in $T$3 (e.g., Fermi-edge broadening) are used to extract an effective $T$4, yielding insight into hot-carrier cooling and electron-phonon coupling (Ulstrup et al., 2014, Boschini et al., 2023).
• Collective modes: Oscillations in $T$5 or related observables can identify Higgs (amplitude) oscillations (Schwarz et al., 2020, Xu et al., 2018), order-parameter dynamics, or phononic coherences.
• Multi-particle and exotic phenomena: Floquet-Bloch states, exciton or trion satellites, and nonequilibrium bosonic correlations can give rise to replicas and sidebands, as well as k-dependent renormalizations, observed in frequency and momentum resolved fashion (Boschini et al., 2023, Chan et al., 2023, Wu et al., 24 Nov 2025, Christiansen et al., 2019).

Extraction of underlying parameters such as $T$6 (superconducting gap), $T$7, or modal contributions is typically achieved via fitting established theoretical models to $T$8, using population or spectral characteristics as constraints (Xu et al., 2018, Ulstrup et al., 2014, Zonno et al., 2021).

5. Applications to Quantum Matter and Representative Findings

Tr-ARPES has revealed ultrafast phenomena across a breadth of material classes:

• Unconventional superconductors (cuprates): Direct tracking of nodal and near-nodal gap dynamics, nonthermal pair breaking, and quasiparticle relaxation, exposing robust Fermi-liquid behavior and bosonic bottlenecks (Zonno et al., 2021).
• Charge density waves and excitonic insulators: Observation of ultrafast melting and recovery of CDW gaps, time-dependent filling and flattening of excitonic gaps, and phonon-coupled band modulations (Lee et al., 2019, Pan et al., 2023, Boschini et al., 2023).
• Topological and spin-orbit materials: Full Brillouin zone mapping enables observation of Dirac cones, surface states, and light-induced Floquet band engineering (Boschini et al., 2023, Kawaguchi et al., 2023, Majchrzak et al., 2023).
• 2D semiconductors and heterostructures: Direct imaging of ultrafast exciton and trion formation, phonon-mediated intervalley scattering, and evolutions from coherent polarization to thermalized populations (Christiansen et al., 2019, Wu et al., 24 Nov 2025, Chan et al., 2023).
• Higgs spectroscopy: Momentum-resolved spectral oscillations associated with collective amplitude modes resolve condensate dynamics in superconductors (Schwarz et al., 2020, Xu et al., 2018).

6. Advanced Methodologies, Limitations, and Emerging Directions

Modern Tr-ARPES systems exhibit tunable control over photon energy, temporal and energy resolution, and $T$9-space reach. Advances include:

• Multi-mode analyzer designs: FeSuMa and momentum microscopes combine rapid full-$s(t)$0 coverage and high temporal resolution for 3D Brillouin zone mapping (Majchrzak et al., 2023).
• Polarization and spin resolution: Use of VLEED detectors and multi-axis spin filters enables direct time-, spin-, and $s(t)$1-resolved studies on sub-ps timescales (Kawaguchi et al., 2023).
• Switchable energy/time resolution configurations: Systems employing nonlinear optics and on-the-fly reconfiguration to trade $s(t)$2 for $s(t)$3 expand experimental flexibility (Pan et al., 2023, Hellbrück et al., 2024).
• Machine learning and automated data analysis: High-dimensional data sets in $s(t)$4 space are increasingly handled via advanced denoising, clustering, and feature identification algorithms (Na et al., 2023).

Notable limitations remain: the fundamental time-bandwidth constraint, sample heating at high repetition rate, trade-off between spatial and momentum resolution (for $s(t)$5-Tr-ARPES), and challenges in nontrivial matrix element effects or gauge-invariant interpretation, especially in complex multiband or driven systems (Freericks et al., 2016, Na et al., 2023).

Prospective future directions include attosecond-resolved Tr-ARPES for subcycle dynamics, full polarization control for Berry curvature and orbital texture studies, machine learning-facilitated discovery in large data volumes, and integration of Tr-ARPES with other ultrafast probes (e.g., X-ray or electron diffraction) for comprehensive mapping of electronic, lattice, and magnetic degrees of freedom (Boschini et al., 2023, Na et al., 2023).

---

References:

(Xu et al., 2018, Mills et al., 2019, Pan et al., 2023, Kawaguchi et al., 2023, Bao et al., 2021, Hellbrück et al., 2024, Freericks et al., 2021, Majchrzak et al., 2023, Na et al., 2023, Zonno et al., 2021, Wu et al., 24 Nov 2025, Caruso et al., 2019, Schwarz et al., 2020, Ulstrup et al., 2014, Christiansen et al., 2019, Chan et al., 2023, Freericks et al., 2016, Lee et al., 2019, Boschini et al., 2023, Puppin et al., 2018)