Spectrally Resolved Time-Gated Heralded Spectroscopy

• Spectrally resolved time-gated heralded spectroscopy is a measurement paradigm that uses a herald photon to define a temporal reference, a narrow gate to condition detections, and spectral mapping to assign wavelength coordinates.
• It enables joint spectral intensity reconstruction, fluorescence lifetime measurements, and multiexciton cascade isolation across visible, near-infrared, and telecom bands.
• The method leverages precise time gating and frequency-to-time mapping to suppress accidental coincidences while optimizing spectral resolution and overall signal-to-noise performance.

Spectrally resolved time-gated heralded spectroscopy denotes a family of measurement schemes in which detection of a herald photon or event defines a reference time, a narrow temporal gate selects correlated detections within that reference, and spectral information is recovered from the accepted events by dispersive, imaging, or interferometric means. In the literature represented here, this framework is used for joint spectral intensity reconstruction, conditional spectral analysis, fluorescence lifetime measurement, and multiexciton cascade isolation, across visible, near-infrared, and telecom bands and under both pulsed and continuous-wave excitation (Iso et al., 2 Oct 2025, Davis et al., 2016, Alvarez-Mendoza et al., 5 May 2025).

1. Conceptual basis

The defining elements are heralding, time gating, and spectral resolution. Heralding means that one detector channel identifies the presence or timing of a correlated partner photon or of a specific step in a radiative cascade. Time gating means that only events within a selected temporal window relative to that herald are accepted. Spectral resolution means that the accepted events are further assigned a wavelength or frequency coordinate, either directly or through calibrated mapping. In practice, the gate may be implemented in hardware or by offline selection from time-tagged data, but in both cases its role is to reject accidentals and to condition the spectrum on a physically meaningful temporal reference (Davis et al., 2016, Iso et al., 2 Oct 2025).

This logic appears in several distinct experimental regimes. In biphoton source characterization, the target observable is commonly the joint spectral intensity, or a conditional spectrum derived from it. In entangled-photon spectroscopy, the herald may define the excitation time of a sample at the single-photon level, so that fluorescence arrival histograms become time-resolved spectroscopic observables even under continuous-wave pumping (Alvarez-Mendoza et al., 5 May 2025). In quantum-dot cascade studies, the first detected photon in a multiexciton decay chain serves as a herald for later photons from the same excitation cycle, enabling conditional spectra for biexciton and triexciton emission from ensembles (Scharf et al., 7 Aug 2025).

A broader theoretical formulation treats time- and frequency-gated coincidence counting as a multidimensional spectroscopy in detection space. In that description, an $N$-photon coincidence signal becomes a $2N$-dimensional function of detection times and frequencies, with the gates explicitly entering the measured signal through detector spectrograms rather than being treated as incidental post-processing (Dorfman et al., 2016).

2. Formal structure and observables

For biphoton spectroscopy, the starting point is the joint spectral amplitude

$$f(\omega_s,\omega_i)=\alpha(\omega_s+\omega_i)\,\phi(\omega_s,\omega_i),$$

where $\alpha(\omega_s+\omega_i)$ is the pump envelope function and $\phi(\omega_s,\omega_i)$ is the phase-matching function. The joint spectral intensity is

$$\mathrm{JSI}(\omega_s,\omega_i)=|f(\omega_s,\omega_i)|^2.$$

A heralded or conditional spectrum is then obtained by fixing one partner frequency, for example

$$S_s(\omega_s\,|\,\omega_i)\propto \mathrm{JSI}(\omega_s,\omega_i).$$

With time gating, the coincidence probability becomes

$$P(\omega_s,\omega_i,\Delta t)=|f(\omega_s,\omega_i)|^2\,W(\Delta t),$$

where $W(\Delta t)$ selects a narrow temporal window around the relevant correlation peak (Iso et al., 2 Oct 2025).

Frequency-to-time mapping is the most common route to spectral resolution in single-photon implementations. Near a center wavelength $\lambda_0$, the calibration is written as

$2N$0

with dispersion $2N$1. If the effective timing resolution is $2N$2, the smallest resolvable wavelength interval is

$2N$3

This relation underlies chirped-fiber-Bragg-grating spectrometers, fiber-dispersion spectrometers, and other time-of-flight architectures (Davis et al., 2016).

The statistics of the gate are equally important. In coincidence-based measurements, narrowing the window suppresses accidentals approximately linearly with gate width,

$2N$4

where $2N$5 and $2N$6 are herald and signal singles rates and $2N$7 is the gate width. That scaling is used explicitly in both heralded spectroscopy and high-throughput time-tagging systems (Davis et al., 2016, Wahl et al., 2020). For position-sensitive visible spectrometers based on delay-line anodes, the spectral coordinate itself is reconstructed from timing differences,

$2N$8

after which the horizontal coordinate is mapped to wavelength by grating dispersion (Iso et al., 2 Oct 2025).

Schmidt analysis connects measured spectra to modal purity. If

$2N$9

then the Schmidt number and purity are

$$f(\omega_s,\omega_i)=\alpha(\omega_s+\omega_i)\,\phi(\omega_s,\omega_i),$$0

In practice, several implementations estimate $$f(\omega_s,\omega_i)=\alpha(\omega_s+\omega_i)\,\phi(\omega_s,\omega_i),$$1 and $$f(\omega_s,\omega_i)=\alpha(\omega_s+\omega_i)\,\phi(\omega_s,\omega_i),$$2 from measured JSIs, sometimes augmented by time-domain information when spectral phase is not directly measured (Iso et al., 2 Oct 2025, Allgaier et al., 2017).

3. Instrumental realizations

One major class uses dispersive time-of-flight spectrometers. In a pulsed single-photon spectrometer based on chirped fiber Bragg gratings, frequency-to-time mapping over 825–835 nm was achieved with nearly uniform efficiency, 55 pm resolution at 830 nm, and approximately 24 GHz frequency resolution using a fast detector, while a two-arm JSI configuration with slower detectors gave approximately 0.2 nm resolution (Davis et al., 2016). A related architecture uses long dispersive fibers at telecom wavelengths to convert arrival time into wavelength, often in tandem with SNSPDs or other fast detectors.

A second class uses spatially dispersive spectrographs with single-photon imagers. A hybrid biphoton spectrometer spanning visible and near-infrared bands combines a grating spectrometer and delay-line-anode single-photon imager for signal photons near 515 nm with a 1.5 km dispersion-compensating fiber and SNSPD for idler photons near 1550 nm. All channels are time-tagged against a laser synchronization signal with 25 ps binning, and time-resolved JSI frames are formed by slicing the herald–sync histogram into 150 ps windows (Iso et al., 2 Oct 2025). This architecture explicitly bridges visible and telecom bands without scanning.

A third class uses detector arrays that directly register both position and time. A data-driven linear SPAD-array spectrometer reported approximately 0.04 nm spectral resolution and approximately 40 ps temporal resolution, with direct time- and frequency-stamping of each photon on a 512-pixel line sensor (Jirsa et al., 2023). An earlier visible-range multiphoton correlator using a monochromator and SPAD array provided 300–900 nm coverage, about 2 nm/pixel spectral resolution, and 110 ps time resolution, enabling time-resolved multiphoton spectral correlation measurements (Johnsen et al., 2014).

Spectral resolution need not be dispersive. In entangled-photon optical spectroscopy, a common-path birefringent interferometer of the TWINS type creates two delayed replicas of a fluorescence wave packet, and a Fourier transform of the interferogram reconstructs the spectrum. Here the herald provides the time reference, while the spectral dimension is obtained interferometrically with a single-pixel SPAD rather than a detector array (Alvarez-Mendoza et al., 5 May 2025).

The time-tagging backbone can itself be a decisive part of the apparatus. FPGA-based TCSPC electronics supporting up to 16 photon timing channels plus synchronization input, 80 ps timing resolution, 650 ps dead-time per channel, and sustained aggregate throughput greater than 80 Mcps have been used for spectrally resolved detection across multiple channels. The same platform supports T2 and T3 modes, channel-wise offsets, and White Rabbit synchronization with a measured contribution of approximately 39 ps rms over 5 km (Wahl et al., 2020).

4. Representative experiments

A central demonstration of the field is the hybrid visible/telecom biphoton spectrometer. In that work, highly non-degenerate photon pairs were generated by type-I SPDC in LBO, pumped by second-harmonic generation from a mode-locked Ti:Sapphire laser at 773.2 nm. The phase-matching configuration yielded daughter photons centered at 515 nm and 1550 nm, with measured widths of 1.60 nm and 12.9 nm, respectively. The instrument response function of the herald path was approximately 263 ps, the time-resolved JSI was sliced into five 150 ps windows, and the sum of time-sliced counts per pixel equaled the static JSI count per pixel. The measured Schmidt number was $$f(\omega_s,\omega_i)=\alpha(\omega_s+\omega_i)\,\phi(\omega_s,\omega_i),$$3 with purity $$f(\omega_s,\omega_i)=\alpha(\omega_s+\omega_i)\,\phi(\omega_s,\omega_i),$$4, compared with simulated values $$f(\omega_s,\omega_i)=\alpha(\omega_s+\omega_i)\,\phi(\omega_s,\omega_i),$$5 and $$f(\omega_s,\omega_i)=\alpha(\omega_s+\omega_i)\,\phi(\omega_s,\omega_i),$$6 (Iso et al., 2 Oct 2025).

The chirped-fiber-Bragg-grating spectrometer provided an earlier non-scanning implementation of the same logic. It measured spectral interference of heralded single photons and the JSI of a type-II SPDC source near 830 nm. The heralded interferogram showed approximately 0.2 nm fringes corresponding to a delay of about 11 ps and a best-fit visibility of approximately 24%. In the JSI measurement, the normalized coincidence histogram acquired over 600 s contained about 28,000 joint counts (Davis et al., 2016).

The method has also been extended from source characterization to spectroscopy on matter. In entangled-photon time- and frequency-resolved optical spectroscopy, a cw 413 nm pump drove SPDC in ppKTP to generate photons near 800 nm and 860 nm. Idler detection provided $$f(\omega_s,\omega_i)=\alpha(\omega_s+\omega_i)\,\phi(\omega_s,\omega_i),$$7 for excitation, while fluorescence detection yielded TCSPC histograms and TWINS interferograms. Reported instrument-response widths were about 260 ps with a fast SPAD and about 600 ps with a higher-efficiency SPAD, with effective temporal resolution of about 100–200 ps after deconvolution. The method resolved mono-exponential decays of 800CW dye at $$f(\omega_s,\omega_i)=\alpha(\omega_s+\omega_i)\,\phi(\omega_s,\omega_i),$$8 ns and IR143 at $$f(\omega_s,\omega_i)=\alpha(\omega_s+\omega_i)\,\phi(\omega_s,\omega_i),$$9 ns, distinguished spectral evolution in a dye mixture, and recorded wavelength-integrated LH2 traces in 50 s, 10 s, 2 s, and even 0.6 s with acceptable signal-to-noise ratio (Alvarez-Mendoza et al., 5 May 2025).

In solid-state spectroscopy, heralded emission detection under twin-photon excitation was demonstrated for near-infrared quantum-dot solids. The coincidence histogram was modeled as the convolution of the true decay with an instrument response function plus flat accidental background, and the measured IRF was about 72 ps FWHM. Herald singles were about $\alpha(\omega_s+\omega_i)$0 cps at 4.7 K, emission singles were about 89 cps with an SNSPD emission arm, and coincidence counts reached up to approximately 230 ccps when an APD was used for emission detection (Tsao et al., 15 Sep 2025).

Ensemble SR-TGHS has additionally been used to isolate multiexciton cascades in CdSe/CdS core/shell quantum dots. By selecting early photons in the first nanosecond after a laser pulse and conditioning on same-pulse pairs or triplets, biexciton and triexciton spectra were separated from exciton emission. The reported biexciton binding energies ranged from $\alpha(\omega_s+\omega_i)$1 meV for 2.2/1.5 nm dots to $\alpha(\omega_s+\omega_i)$2 meV for 1.8/5.2 nm dots, and triexciton pathways at about 2.01 eV and 2.10 eV were assigned lifetimes of $\alpha(\omega_s+\omega_i)$3 ps and $\alpha(\omega_s+\omega_i)$4 ps in first-photon decays (Scharf et al., 7 Aug 2025).

5. Performance metrics and limiting factors

Temporal resolution is generally set by detector jitter, gate width, and any dispersive mapping used to infer wavelength. In the visible/telecom hybrid JSI spectrometer, the gate width was 150 ps, the herald IRF was about 263 ps, and the idler-arm timing contribution broadened the coincidence response to 310 ps in start/stop correlation terms (Iso et al., 2 Oct 2025). In the CFBG spectrometer, 52 ps detector jitter and about 950 ps/nm dispersion yielded 55 pm resolution, whereas slower detectors broadened the JSI resolution to about 0.2 nm (Davis et al., 2016). Direct SPAD-array instruments have reported approximately 40 ps single-photon timing and 0.042–0.054 nm Gaussian spectral widths (Jirsa et al., 2023). Waveguided SFG temporal sampling in LiNbO$\alpha(\omega_s+\omega_i)$5 has pushed the time-domain side to about 200 fs after deconvolution, while using fiber spectrometers for the complementary spectral data (Allgaier et al., 2017).

Signal-to-noise ratio depends strongly on accidental suppression. In the hybrid biphoton spectrometer, the single-detector count rate was $\alpha(\omega_s+\omega_i)$6 Hz, the coincidence rate was $\alpha(\omega_s+\omega_i)$7 Hz, and the coincidence-to-accidental ratio was 0.36, with the low CAR attributed to thermal photocathode background in the delay-line detector and the broad idler arrival-time distribution (Iso et al., 2 Oct 2025). The CFBG spectrometer incurred about 10 dB total loss through the circulator-plus-grating path, comprising about 3.4 dB from the circulator round trip, about 3 dB from the grating, and the remainder from splice mismatch (Davis et al., 2016). In the integrated telecom heralded source based on a 1 GHz gated InGaAs/InP SPAD, detector operation at 15.5% PDE gave a dark-count probability of $\alpha(\omega_s+\omega_i)$8 per gate and afterpulse probability below 1.0% with 5 $\alpha(\omega_s+\omega_i)$9s dead time; the measured unheralded $\phi(\omega_s,\omega_i)$0 corresponded to $\phi(\omega_s,\omega_i)$1 when the heralding arm used the gated SPAD and to $\phi(\omega_s,\omega_i)$2 when two SNSPDs were used (Pereira et al., 19 Sep 2025).

Throughput and scaling impose additional constraints. Multi-channel TCSPC electronics can sustain aggregate rates above 80 Mcps and histogramming rates of 78 Mcps per channel, which is relevant when spectroscopy is distributed across multiple spectral bins or detector pixels (Wahl et al., 2020). By contrast, entangled three-photon spectroscopy with cascaded PDC was reported with detected triplet rates of approximately 7 counts per hour, illustrating the severe count-rate penalty that accompanies higher-order coincidence spectroscopy even when the theoretical information content approaches coherent 2D spectroscopy (Fujihashi et al., 2021).

A recurrent trade-off is that narrowing the gate suppresses accidentals while increasing demands on detector timing and synchronization. Another is that increasing dispersion improves spectral resolution but reduces usable bandwidth or introduces additional loss. A plausible implication is that the most effective architectures are those in which the temporal gate is matched to the detector IRF and the source bandwidth, rather than being made arbitrarily narrow.

6. Relation to adjacent methods and current trajectory

SR-TGHS developed partly in contrast to scanning spectrometers and stimulated emission tomography. Scanning grating or filter methods are effective but slow and low contrast, while SET is fast and high resolution but generally not time-resolved and requires chromatic laser scanning (Iso et al., 2 Oct 2025). Fiber spectrographs, SPAD arrays, and data-driven cameras have therefore been adopted to acquire spectra without scanning, although spectral range, detector quantum efficiency, and dead time remain practical constraints (Iso et al., 2 Oct 2025, Jirsa et al., 2023).

The method also intersects with quantum interference spectroscopy. In spectrally resolved Hong–Ou–Mandel interference between independent sources, dispersive fiber spectrometers recovered correlated spectral intensity maps as a function of delay, and time-window narrowing from 5 ns to 1 ns improved heralded HOM visibility from $\phi(\omega_s,\omega_i)$3 to $\phi(\omega_s,\omega_i)$4 by acting as a spectral post-filter (Jin et al., 2015). This is not sample spectroscopy in the usual sense, but it exemplifies the same operational triad: heralding, time gating, and spectral conditioning.

On the theoretical side, time-and-frequency gated photon coincidence counting has been formulated as a detection-based multidimensional spectroscopy in which detector gates, rather than excitation pulse sequences alone, define the measured information content (Dorfman et al., 2016). Entangled three-photon spectroscopy makes this connection explicit: frequency-dispersed two-photon coincidence counting with monochromatic pumping can reproduce the information content of coherent 2D optical spectra along a constrained slice, while the phase-matching function acts as a spectral filter and the entanglement time defines the intrinsic temporal gate (Fujihashi et al., 2021).

Current trajectories in the experimental literature are concrete rather than speculative. The visible/telecom hybrid spectrometer proposes use of the unused $\phi(\omega_s,\omega_i)$5 coordinate of the delay-line detector for streak-tube integration or fiber-bundle encoding, potentially adding sub-picosecond temporal imaging or spatial information (Iso et al., 2 Oct 2025). High-speed TDCs with few-picosecond resolution, SNSPD arrays, and advanced imaging electronics are identified as routes to sharper spectral projections and higher signal-to-noise ratios (Iso et al., 2 Oct 2025). Broader-band or cascaded chirped fiber Bragg gratings, lower-jitter detectors, and multi-channel parallel detection have likewise been proposed to extend frequency-to-time mapping beyond its current bandwidth–loss compromise (Davis et al., 2016). In distributed experiments, White Rabbit synchronization already provides tens-of-picoseconds alignment across kilometer scales, which suggests a practical path toward remotely synchronized heralded spectroscopy (Wahl et al., 2020).

Taken together, these developments define SR-TGHS not as a single instrument design but as a measurement paradigm: correlated detection establishes the clock, temporal gating enforces causal and statistical selectivity, and spectral resolution is supplied by whichever optical or computational transducer best matches the photon flux, wavelength range, and target observable.