Time-Resolved Cathodoluminescence

• Time-resolved cathodoluminescence is an advanced spectroscopic method that uses pulsed electron beams to excite and temporally resolve luminescence with nanometric spatial precision.
• It employs ultrafast electron pulses, high-numerical-aperture photon collection, and time-correlated single-photon counting to measure sub-nanosecond to picosecond emission dynamics.
• The technique enables detailed mapping of quantum emitters, defect states, and plasmonic responses, offering actionable insights for nanoscale material and device characterization.

Time-resolved cathodoluminescence (TR-CL) is an advanced spectroscopic technique that measures the temporal evolution of luminescence induced by high-energy electron beams in solids, enabling the study of ultrafast optical processes at nanometric spatial resolution. By combining pulsed electron excitation with time-correlated photon detection, TR-CL directly quantifies the excited-state lifetimes and relaxation dynamics of electronic, excitonic, and plasmonic modes, revealing both coherent and incoherent emission pathways with spatial precision far below the optical diffraction limit.

1. Physical Principles and Conceptual Framework

Cathodoluminescence (CL) arises when incident electrons interact with a material, promoting electronic excitations that may relax radiatively and emit photons. The temporal response of this emission, encoded in the CL decay trace $I(t)$, provides insight into carrier recombination, defect physics, local density of optical states (LDOS), energy transfer, and non-radiative processes.

Time-resolved CL leverages the rapid time structure of pulsed or synchronized electron sources, typically generated via femtosecond lasers or rapid beam blankers, to act as a well-defined excitation event ("pump"). The subsequent photon emission ("probe") is registered by ultrafast single-photon detectors, with timing referenced either to the electron pulse or, in continuous beams, to the passage of individual electrons detected downstream. Histogramming photon arrival delays $t$ with respect to electron arrival (or trigger) yields a decay curve $I(t)$, which can be deconvoluted from the instrumental response to extract the intrinsic emitter lifetimes and analyze multipath relaxation channels (Meuret et al., 2021, Yanagimoto et al., 2023, Varkentina et al., 2023, Tizei et al., 10 Nov 2025).

Central features of TR-CL include:

• Sub-nanometer to few-nanometer spatial excitation region, dictated by the electron-beam diameter and scattering.
• Temporal resolution limited by the convolution of electron-pulse duration, detector jitter, and timing electronics, ranging from tens of picoseconds (ps) to sub-femtoseconds in optimal configurations (Solà-Garcia et al., 2021, Meuret et al., 2021).
• Capability to separate coherent processes (e.g., transition radiation, Cherenkov emission) from incoherent bands (e.g., defect, excitonic, or impurity luminescence) via photon correlation analysis (Scheucher et al., 2021).
• Temporal coincidence methods—either electron–photon or photon–photon—to assign causality to each emission event and suppress background.

2. Instrumentation and Experimental Methodologies

Electron Excitation and Pulsed Sources

TR-CL requires electron beams with temporally localized excitation. Key modalities include:

• Femtosecond photoemission sources in ultrafast transmission electron microscopes (UTEMs): fs-laser (e.g., λ=515 nm) illumination of a cold-field emitter tip yields $<1$ ps electron pulses at MHz repetition rates. With probe sizes $<1$ nm and effective mapping resolutions of $\sim$12 nm, such setups achieve lifetime mapping with sub-nanosecond time resolution (Meuret et al., 2021).
• Electrostatic/magnetic beam blankers or RF cavity chopping for ns–ps pulses in SEM/STEM (Tizei et al., 10 Nov 2025, Solà-Garcia et al., 2021).
• Conventional beams in continuous mode can be correlated with photon detection via fast electron counters (scintillator-photomultiplier sequences or direct detectors) for event-by-event TR-CL (Yanagimoto et al., 2023).

Photon Collection and Time-tagged Detection

• High-numerical-aperture parabolic or ellipsoidal mirrors placed close to the sample (sub-mm distances) maximize solid-angle photon collection and couple emission into multimode fibers or directly to detectors (Meuret et al., 2021, Bittorf et al., 29 Jan 2025).
• Detection schemes include hybrid photomultiplier tubes, single-photon avalanche diodes (SPADs), or superconducting nanowire single-photon detectors (SNSPDs), offering timing jitter from $\sim$20 ps (SNSPD) to several hundred ps (PMT/SPAD).
• Time-to-digital converters (TDCs) or time-correlated single-photon counting (TCSPC) correlate electron/photon arrival times with <10 ps–1 ns binning, constructing histograms of emission delay.

A summary table of typical configurations and parameters:

Instrument	Electron Pulse	Detector Jitter	TCSPC Bin	Spatial Resolution	Temporal Resolution
UTEM (cold-FEG)	400 fs	350 ps	400 ps	$\sim$12 nm	0.5–0.9 ns
STEM + Timepix3	$\sim$1 ns	1.6 ns (TP3/PMT)	1.6 ns	$\sim$1 nm	$\sim$2 ns
SEM + fiber/PMT	25 ns–$\mu$s	40–100 ps	10–100 ps	$<$100 nm	$<$0.1 ns

(Parameters represent typical values as reported in (Meuret et al., 2021, Varkentina et al., 2023, Bittorf et al., 29 Jan 2025).)

3. Data Analysis and Lifetime Extraction

The measured TR-CL decay $I_{\rm meas}(t)$ is modeled as the convolution of the sample's true temporal response $I_{\rm sample}(t)$ and the instrument response function (IRF):

$$I_{\rm meas}(t) = \int_{-\infty}^\infty IRF(\tau) I_{\rm sample}(t - \tau) d\tau$$

For single-exponential emitters (neglecting rise time):

$$I_{\rm sample}(t) = A \exp\left[-(t-t_0)/\tau\right] + B$$

The IRF may often be well-approximated as Gaussian, with width determined by both the photon detector and timing electronics: $$\sigma_{\rm tot} = \sqrt{\sigma_{\rm detector}^2 + \sigma_{\rm bin}^2}$$. Full width at half maximum is given by $$FWHM_{\rm IRF} \approx 2 \sqrt{2\ln2}\, \sigma_{\rm tot}$$ (Meuret et al., 2021).

In multi-channel or heterogeneous systems, a sum of exponential decays is fitted:

$$I_{\rm sample}(t) = \sum_{i} A_i \exp\left[-(t-t_0)/\tau_i\right]$$

Least-squares fitting of the convolution yields pixel-wise or spectrally-resolved maps of $\tau_i$, $A_i$, and associated uncertainties.

Alternative approaches leverage electron–photon (EELS–CL) coincidence counting, generating two-dimensional histograms $H(E, \tau)$ binned by electron energy loss $E$ and photon delay $\tau$, from which excitation-specific lifetimes $\tau(E)$ can be extracted by mono- or multi-exponential modeling (Varkentina et al., 2023, Varkentina et al., 2022).

Photon autocorrelation (Hanbury–Brown–Twiss) methods, with fiber-based HBT setups and TCSPC electronics, enable the extraction of luminescence decay times $\tau_d$ by fitting

$$g^{(2)}(\tau) = 1 + g_0 \exp\left(-|\tau|/\tau_d\right)$$

where $g_0$ quantifies photon bunching (Bittorf et al., 29 Jan 2025).

4. Representative Results and Case Studies

Quantum Emitters: NV Centers in Nanodiamonds

UTEM-based TR-CL mapping of nanodiamonds with dense NV$^0$ centers (150 keV, $\sim$5 nm spot, 12 nm pixel binning) revealed luminescence lifetimes $\tau = 21.4 \pm 0.5$ ns and spatial heterogeneity from 15.8 to 23.4 ns over $<50$ nm. Variations are attributed to local environment, surface proximity, impurity effects, and carrier diffusion length ($\sim$50 nm) (Meuret et al., 2021, Varkentina et al., 2023, Tizei et al., 10 Nov 2025).

CLE Spectroscopy of Plasmonic and Defect States

Time-resolved CLE assigns each detected photon a coincident EELS energy, enabling construction of excitation-to-emission maps, and the measurement of quantum efficiency as a function of excitation pathway (Varkentina et al., 2022, Varkentina et al., 2023). For plasmonic nanoparticles, phase-locked processes (SP, TR) yield instantaneous emission (within IRF), while defect-band emission in h-BN displays nanosecond-scale lifetimes, resolved as broadened features in the $\tau$ domain.

Nanothermometry and Environmental Sensitivity

Lifetime-based TR-CL enables robust, intensity-independent nanoscopic thermometry, as demonstrated for NaYF$_4$:Yb$^{3+}$,Er$^{3+}$ nanoparticles. The emission lifetime's temperature dependence ($\sim$0.9 %/°C, sensitivity $\sim$30 mK) was exploited, with spatial resolution set by the electron-beam spot and achieved FWHM for the red emission band $\approx$28.6 nm (Aiello et al., 2018).

5. Differentiation of Emission Mechanisms

Time-resolved photon correlation functions are used to discriminate coherent vs. incoherent CL components. Coherent contributions (e.g., Cherenkov, transition radiation) manifest as narrow, highly bunched peaks at $\tau=0$, with the bunching amplitude $g^{(2)}(0)$ scaling inversely with beam current and saturating for purely coherent emission. Quantitative determination of coherent/incoherent ratios enables isolation of defect-state luminescence in complex spectra, enhancing reliability for nanocharacterization (Scheucher et al., 2021).

Momentum-resolved electron–photon coincidence schemes further select for phase-related (entangled or momentum-conserved) pairs, as indicated by an excitation correlation factor $\xi_{ep}>1$ in momentum-matched geometries (Yanagimoto et al., 2023, Preimesberger et al., 2024).

6. Applications and Outlook

Time-resolved cathodoluminescence enables nanoscale mapping of excited-state lifetimes in quantum emitters, defects, quantum dots, and 2D materials, with demonstrated spatial resolutions down to $\sim$12 nm and temporal resolutions at or below the nanosecond level (Meuret et al., 2021, Bittorf et al., 29 Jan 2025). Correlative measurements unifying structural, chemical, and optical dynamic data are possible in modern (S)TEM/SEM platforms (Tizei et al., 10 Nov 2025).

Emerging directions include:

• Integration of TR-CL with time-resolved EELS for coupled electronic/optical studies.
• Exploitation of multi-exponential fitting for separation of radiative and non-radiative recombination channels (Meuret et al., 2021).
• Ultrafast pump–probe architectures (fs electron and laser pulses) for femtosecond time resolution (Solà-Garcia et al., 2021).
• Proof-of-principle coincidence imaging of energy–momentum conservation and progress toward electron–photon entanglement studies (Preimesberger et al., 2024).
• Quantitative photon statistics mapping (e.g., antibunching in single-photon emitter investigations).

Main technological advances required include faster, lower-jitter electron and photon detectors (e.g., Timepix4, SNSPD), greater collection solid angle, and efficient signal processing to allow picosecond and potentially sub-picosecond TR-CL (Tizei et al., 10 Nov 2025).

7. Limitations and Practical Considerations

• Temporal resolution is fundamentally limited by instrument response: detector jitter, TDC binning, and electron pulse duration. Current state-of-the-art is $\sim$50–100 ps, with path to tens of ps using SNSPDs and direct electron detectors (Meuret et al., 2021, Tizei et al., 10 Nov 2025).
• Coincidence rates for EELS–CL or TEPCoM methods are low ($\sim$10^{-5}$ per electron), requiring long acquisitions for weak signals. Improvements in quantum efficiency and SNR are essential for rapid mapping in low-yield materials (Preimesberger et al., 2024, Varkentina et al., 2023).
• Trade-off exists between spatial resolution (favored by low electron-beam current) and photon statistics/SNR, with higher current risking beam-induced damage, non-linear effects, or statistical broadening (Meuret et al., 2021, Aiello et al., 2018).
• Spectral and energy filtering may introduce detection biases, so multimodal or broadband detection is often preferred.
• For materials with extremely short (ps-scale) intrinsic lifetimes (e.g., transition radiation, plasmon decay), IRF deconvolution is mandatory.

TR-CL is a convergent technique at the intersection of electron microscopy, ultrafast optics, and quantum photonic measurement, offering unprecedented capabilities for resolving dynamic processes at the nanometer–femtosecond nexus (Tizei et al., 10 Nov 2025, Preimesberger et al., 2024, Meuret et al., 2021).