---
title: Optical Pump-Probe Reflectance (OPPR)
url: https://www.emergentmind.com/topics/optical-pump-probe-reflectance-oppr
type: topic
---

# Optical Pump-Probe Reflectance (OPPR)

Searching arXiv for recent and foundational OPPR-related papers to ground the article.
Optical pump–probe reflectance (OPPR) is an ultrafast optical spectroscopy method in which a pump pulse generates a transient change in a material and a delayed probe pulse measures the corresponding relative change in reflectance, typically written as $\Delta R/R$ or $\Delta R/R_{0}$. Across the literature, OPPR is used to interrogate non-equilibrium electronic dynamics, coherent collective modes, carrier relaxation, spin-light coupling, transient conductivity, and selective-reflection nonlinearities in solids, semiconductors, correlated materials, and atomic vapors [2108.10605], [1608.03503], [1110.6643], [1701.05376], [2410.21496], [2312.06243]. A central methodological development is the model-independent inversion of differential reflectance into transient changes of the complex dielectric function, $\Delta \epsilon(\omega,t)=\Delta\epsilon_{1}(\omega,t)+i\Delta\epsilon_{2}(\omega,t)$, which is intended to remove ambiguities caused by the simultaneous influence of reflection, transmission, and scattering on conventional pump–probe observables [2108.10605].

## 1. Definition and observable structure

In its most common form, OPPR records the pump-induced, time-dependent change in reflectance after photoexcitation by a femtosecond pump pulse, for example
\[
\Delta R/R(\omega,t)=\frac{R_{\mathrm{pumped}}(\omega,t)-R_{\mathrm{unpumped}}(\omega)}{R_{\mathrm{unpumped}}(\omega)}
\]
or equivalently
\[
\Delta R/R_{0}\equiv (R_{\mathrm{on}}-R_{\mathrm{off}})/R_{\mathrm{off}}\,,
\]
depending on the notation adopted in a given experiment [1608.03503], [1110.6643]. In reflection geometry, the measured signal is often interpreted as arising from pump-induced changes in the real and imaginary parts of the dielectric function, $\epsilon_{1}$ and $\epsilon_{2}$, or equivalently in the refractive-index components $n$ and $k$ [2108.10605], [1407.0073].

For a planar interface, the equilibrium optical response is expressed through Fresnel relations. In the semi-infinite normal-incidence limit from air into a medium,
\[
R(\omega)=\frac{(n-1)^2+k^2}{(n+1)^2+k^2}\,,
\]
with
\[
\epsilon(\omega)=\epsilon_{1}(\omega)+i\epsilon_{2}(\omega)=(n(\omega)+ik(\omega))^{2}\,.
\]
For thin films on substrates, transfer-matrix or analytic expressions $R(\epsilon_{1},\epsilon_{2})$ are used instead [2108.10605]. A linearized treatment of small photoinduced perturbations gives
\[
\Delta R(\omega)\simeq \frac{\partial R}{\partial \epsilon_{1}}\Delta\epsilon_{1}
+\frac{\partial R}{\partial \epsilon_{2}}\Delta\epsilon_{2}
\]
and therefore
\[
\frac{\Delta R}{R}=A(\omega)\Delta\epsilon_{1}+B(\omega)\Delta\epsilon_{2}\,,
\]
where $A(\omega)=R^{-1}\partial R/\partial\epsilon_{1}$ and $B(\omega)=R^{-1}\partial R/\partial\epsilon_{2}$ are computed from the static dielectric function [2108.10605].

This formulation establishes the basic interpretive problem of OPPR: $\Delta R/R$ is not itself a material parameter, but a projection of the transient optical state through equilibrium optics and geometry. The motivation for quantitative OPPR is therefore to reconstruct $\Delta\epsilon$, $\Delta n$, and $\Delta k$ rather than to interpret raw reflectance changes phenomenologically [2108.10605].

## 2. Dielectric-function inversion and quantitative OPPR

A model-independent route to transform differential transmission or reflection spectra into changes in the transient dielectric function is given by a variational Kramers–Kronig procedure [2108.10605]. The protocol begins by measuring the equilibrium reflectance $R_{0}(\omega)$, and if available transmission $T_{0}(\omega)$, then retrieving $\epsilon_{1,0}(\omega)$ and $\epsilon_{2,0}(\omega)$ with a variational dielectric function composed of a small set of Lorentz oscillators plus a dense grid of weak KK-constrained oscillators [2108.10605]. This supplies analytic $\epsilon_{1,0}$ and $\epsilon_{2,0}$ and permits analytical differentiation of $R(\epsilon_{1},\epsilon_{2})$ to obtain $A(\omega)$ and $B(\omega)$ across the measured spectral window.

The transient step is then formulated as a constrained inverse problem. One records $\Delta R/R(\omega,t)$ at selected delays, represents $\Delta\epsilon_{2}(\omega)$ on a dense energy grid of KK-paired basis functions, enforces KK consistency so that $\Delta\epsilon_{1}(\omega)$ is the Hilbert transform of $\Delta\epsilon_{2}(\omega)$, and solves a least-squares problem of the form
\[
\chi^{2}=\sum_{\omega}\Bigl[\frac{\Delta R}{R}(\omega,t)-A(\omega)\Delta\epsilon_{1}(\omega,t)-B(\omega)\Delta\epsilon_{2}(\omega,t)\Bigr]^{2}
\]
with respect to the variational amplitudes [2108.10605]. The result is a delay-dependent reconstruction of $\Delta\epsilon_{1}(\omega,t)$ and $\Delta\epsilon_{2}(\omega,t)$ without invoking an assumed bandshape or kinetic model.

The same work states that the dielectric perturbation can then be converted to refractive-index changes using
\[
n+ik=\sqrt{\epsilon_{1}+i\epsilon_{2}}\,.
\]
The detailed expression printed in the supplied technical guide for $\Delta n+i\Delta k$ appears in truncated form; however, the text explicitly states that the inversion yields $\Delta n$ and $\Delta k$ from the recovered dielectric response [2108.10605]. Benchmarking against time-resolved Frequency Domain Interferometry (FDI), which directly measures $\Delta n$ with sub-femtosecond resolution, shows quantitative agreement within experimental uncertainty [2108.10605]. This is the principal basis for treating OPPR as a quantitative dielectric probe rather than only a differential observable.

A worked example is provided for a CsPbBr$_3$ thin film at $t=500$ fs. At $\hbar\omega=2.37$ eV, static ellipsometry gives $n_{0}=1.80$ and $k_{0}=0.02$, corresponding to $\epsilon_{1,0}\approx 3.24$ and $\epsilon_{2,0}=0.072$. The static Fresnel reflectance is $R_{0}\approx 0.0817$, the derivatives are $A(\omega)\approx -0.45$ and $B(\omega)\approx -1.20$, and a measured $\Delta R/R=+0.015$ is inverted to $\Delta\epsilon_{2}\approx +0.010$ and $\Delta\epsilon_{1}\approx -0.005$. The derived refractive-index changes are $\Delta n\approx -0.0014$ and $\Delta k\approx +0.0028$, consistent with FDI-measured $\Delta n\approx -0.0015$ at the same delay [2108.10605].

## 3. Experimental architectures and detection strategies

OPPR instrumentation spans broadband visible reflectivity, spectrally filtered detection, and reflection-geometry THz probing. A representative femtosecond broadband transient reflectivity instrument for solids combines a Ti:Sapphire oscillator at 800 nm with 45 fs pulses, an amplified 3–10 kHz output up to 3 mJ/pulse, pump–probe beam splitting, broadband white-light generation in 3 mm CaF$_2$, and collection into a spectrometer coupled to a single-shot Hamamatsu CMOS line array with 1024 pixels [1608.03503]. In that configuration, the probe spans 1.72–2.92 eV, the effective time resolution is approximately 45 fs, and a single-shot readout synchronized up to 10 kHz resolves coherent oscillations with amplitudes below $10^{-4}$ in $\Delta R/R$ [1608.03503]. Group-velocity-dispersion correction is performed by fitting the wavelength dependence of time zero to a polynomial up to third order and shifting each trace by its extracted $\Delta t_{0}(\omega)$ [1608.03503].

The same apparatus illustrates several standard OPPR design choices: a motorized retroreflector delay line with step size 1 $\mu$m corresponding to 6.7 fs, pump spot dimensions of about $150\,\mu\mathrm{m}\times150\,\mu\mathrm{m}$, probe spot dimensions of about $50\,\mu\mathrm{m}\times50\,\mu\mathrm{m}$, and cryogenic sample control from 8 to 340 K in a closed-cycle He cryostat [1608.03503]. The raw observables are the pumped and unpumped reflectances, from which the differential reflectivity is computed directly [1608.03503].

A distinct implementation is reflection-geometry pump–THz probe spectroscopy. In the system reported for tunable near- to mid-infrared pumping, a Ti:sapphire amplifier at 800 nm and about 35 fs drives a two-output optical parametric amplifier and difference-frequency generation in GaSe to generate pump wavelengths from 3 to 15 $\mu$m and beyond, while a THz probe is generated in ZnTe and reflected from the sample at about $30^\circ$ incidence [1708.01991]. Detection uses electro-optic sampling in a 1 mm ZnTe crystal. The measured quantity is the transient change in reflected THz electric field, $\Delta E(t,\tau)=E'(t,\tau)-E_{0}(t)$, with normalized response $\Delta E/E_{0}\simeq \Delta R(\omega,\tau)/R_{0}(\omega)$ in the small-change approximation [1708.01991]. The stated frequency coverage is about 0.25 to 2.5 THz, time resolution is set by the optical gate at about 35 fs, and a $0.5\%$ transient change in reflected THz electric field is resolved in EuSbTe$_3$ under 15 $\mu$m excitation [1708.01991].

Spectrally resolved detection can substantially alter OPPR sensitivity. In diamond, transient reflectivity from coherent optical phonons measured with optical band-pass filters shows an approximately 35-fold enhancement in coherent oscillation sensitivity relative to spectrally integrated detection [1603.06023]. The theoretical basis is that phonon-induced sidebands at $\Omega+\omega$ and $\Omega-\omega$ enter the reflectivity modulation with opposite sign, producing substantial cancellation upon full spectral integration, whereas filtering isolates one sideband and prevents cancellation [1603.06023]. In the reported experiment, the unfiltered oscillation amplitude is of order $10^{-5}$, while filtered detection with FB900 yields a much larger oscillatory signal at the same 40.0 THz phonon frequency and about 6 ps damping time [1603.06023].

## 4. Dynamical regimes accessed by OPPR

### Carrier and excitonic dynamics

Few-layer MoS$_2$ provides a canonical example of exciton-sensitive OPPR. Using a 780 nm, approximately 100 fs Ti:Sapphire pump and a tunable 645–700 nm probe derived from an OPO and BBO doubling, differential reflection was measured near an excitonic resonance with pump fluence from about 15 to 100 $\mu\mathrm{J/cm}^{2}$ [1110.6643]. The $\Delta R/R$ spectrum closely resembles the derivative of the resonance, and the best-fit perturbation of a Gaussian resonance involves an approximate red shift of $+0.14$ nm, a peak amplitude increase of about $+0.37\%$, and a linewidth narrowing of about $-0.57\%$ [1110.6643]. At fixed wavelength the dynamics are well fitted by a single exponential,
\[
\Delta R/R(t)=(\Delta R/R)_{\mathrm{peak}}e^{-t/\tau}\,,
\]
with decay times increasing monotonically from about 150 ps on the blue side of resonance to about 400 ps on the red side [1110.6643]. The peak signal is proportional to pump fluence, while the decay time is fluence independent between 15 and 100 $\mu\mathrm{J/cm}^{2}$, which the authors interpret as linear carrier dynamics without significant many-body effects at those densities [1110.6643].

### Polaron formation and spin-correlated relaxation

In Eu$_{0.75}$Y$_{0.25}$MnO$_3$, OPPR at 1.55 and 3.10 eV was used to track Jahn–Teller relaxation, polaron trapping, and localized spin excitation after photoexcitation [1407.0073]. The experiment employed 70 fs pulses from a Ti:Sapphire regenerative amplifier at 250 kHz, pump fluence of approximately $50\,\mu\mathrm{J/cm}^{2}$ with linearity confirmed down to $5\,\mu\mathrm{J/cm}^{2}$, and temperatures from 5 to 300 K [1407.0073]. The transient response at low temperature is described by a double-exponential form
\[
\Delta R/R(t)=A_{1}e^{-t/\tau_{1}}+A_{2}e^{-t/\tau_{2}}\,,
\]
where the first term parameterizes the rise of a slow component associated with localized spin excitation and the second describes nanosecond-scale recombination [1407.0073]. Representative values at 5 K are $|A_{1}|\approx 3\times10^{-3}$ and $\tau_{1}\approx 55$ ps at 1.55 eV, and $A_{1}\approx +3\times10^{-2}$ and $\tau_{1}\approx 75$ ps at 3.10 eV, with $\tau_{2}\sim 1$–5 ns [1407.0073]. The temperature dependence shows that $A_{1}(T)$ and $\tau_{1}(T)$ are essentially zero above about 150 K, begin to grow below about 100–120 K, and do not display sharp anomalies at $T_{N}=47$ K or $T_{FE}=30$ K, instead tracking the onset of short-range spin correlations near 100 K [1407.0073].

### Coherent phonons and collective modes

Broadband OPPR on La$_2$CuO$_4$ demonstrates the ability of transient reflectivity to resolve coherent collective modes in strongly correlated quantum systems [1608.03503]. With a 3.10 eV, 45 fs pump at absorbed fluence 4.4 mJ/cm$^{2}$ and probe continuum from 1.72 to 2.92 eV, the transient map shows a resolution-limited rise of about 50 fs with superimposed oscillations [1608.03503]. Temporal traces are fitted with multicomponent oscillators containing five frequencies, 3.68, 4.55, 6.93, 8.23, and 13.0 THz with uncertainty $\pm 0.22$ THz, and Fourier analysis identifies $A_{g}$ phonon modes up to about 13 THz in agreement with spontaneous Raman data [1608.03503]. The extracted amplitude versus probe photon energy is described as a “Raman matrix element” and is used for quantitative electron-phonon coupling analysis [1608.03503].

In diamond, the coherent optical phonon at 40.0 THz with period $25.0\pm0.4$ fs and damping time about 6 ps is detected by transient reflectivity, and the experiment explicitly shows that the phase of the oscillatory signal depends on whether the high-energy or low-energy sideband of the probe spectrum is selected [1603.06023]. This establishes that not only the magnitude but also the sign and phase of OPPR signals can depend strongly on spectral detection window [1603.06023].

## 5. Coherence, polarization selection, and magneto-optical effects

OPPR in reflection geometry can detect perturbed free-induction decay (PFID) at negative pump–probe delay. For a two-level ensemble described by optical Bloch equations, a weak probe first generates a free-induction-decay polarization $P^{(1)}(t)$, and a later pump perturbs that residual polarization, generating a third-order component $P^{(3)}(t)$ that radiates into the probe-reflection direction [1809.00292]. After monochromator filtering, double chopping, and lock-in detection, the measured spectrally resolved signal takes the form
\[
S(\tau,\omega_{m})\propto
\frac{I_{pr}I_{pm}}{\Gamma^{2}+(\omega_{0}-\omega_{m})^{2}}
\,e^{-\tau/T_{2}}
\cos[(\omega_{0}-\omega_{m})\tau]
\]
for negative delay $\tau>0$, with $\Gamma=1/T_{2}$ [1809.00292]. At resonance, $\omega_{m}=\omega_{0}$, the negative-delay signal becomes a pure exponential $e^{-\tau/T_{2}}$, permitting direct extraction of the coherence time. In GaAs/AlGaAs multi-quantum wells at 4 K, degenerate 100 fs pulses tuned to the heavy-hole exciton yielded $T_{2}\approx 1.5$ ps, in good agreement with the value inferred from the continuous-wave photoluminescence-excitation linewidth $\Delta E\approx 0.4$ meV via $T_{2}\approx 2\hbar/\Delta E\approx 1.68$ ps [1809.00292].

Magneto-optical OPPR introduces an additional polarization-selective mechanism through relativistic spin–light coupling. Starting from the Foldy–Wouthuysen-transformed Dirac–Kohn–Sham Hamiltonian, Mondal et al. identify an Angular Magneto-Electric Hamiltonian
\[
\mathcal{H}_{\rm AME}=
\frac{e^{2}\hbar}{4m^{2}c^{2}}\,
\boldsymbol{\sigma}\cdot(\mathbf{E}\times\mathbf{A})
\]
which can be recast as a Zeeman-like coupling to an induced opto-magnetic field $\mathbf{B}_{\mathrm{opt}}$ [1701.05376]. In a pump–probe experiment the total field contains pump–pump, probe–probe, and pump–probe cross terms, and the cross term is transiently active during pulse overlap [1701.05376]. For parallel linear pump and probe polarizations, $\mathbf{E}_{\rm pu}^{0}\times \mathbf{E}_{\rm pr}^{0*}=0$ and $\mathbf{B}_{\mathrm{opt}}=0$. For perpendicular linear polarizations, the theory predicts a nonzero opto-magnetic field along $\pm \hat z$ proportional to $\sin(\omega\tau)$, so that at short delays $\omega\tau\ll 1$ the coherent field grows linearly in $\tau$ [1701.05376]. The paper states that this explains why a sizable coherent Kerr signal appears in the perpendicular configuration but not in the parallel one, and gives an estimated field of approximately $0.8\,\mu$T for the experiment of Bigot et al. [1701.05376].

A common misconception is that negative-delay or polarization-dependent reflectivity signals necessarily require a more complex nonlinear wave-mixing geometry. The cited works show, more narrowly, that reflection-geometry pump–probe experiments can access PFID and polarization-dependent coherent opto-magnetic fields within pump–probe architectures, provided spectral filtering, polarization control, and lock-in selection are handled explicitly [1809.00292], [1701.05376].

## 6. Nonlinear depth profiles, selective reflection, and theoretical generalizations

When the photoinduced response depends nonlinearly on incident pump intensity, the depth profile of photoconductivity can no longer be assumed to retain a fixed shape as fluence changes. In the formalism of “Nonlinear photoconductivity in pump-probe spectroscopy. I. Optical coefficients,” the inhomogeneous medium is written for $z\ge 0$ with
\[
\epsilon_{r}(z)=\bar n^{2}+i\,\sigma(z)/(\omega\epsilon_{0})
\]
and the probe obeys the scalar wave equation
\[
\frac{d^{2}U}{dz^{2}}+k_{0}^{2}[\epsilon_{r}(z)-\beta^{2}]U=0\,,
\]
with reflection amplitude determined by the surface logarithmic derivative $\Phi(0)$ [2410.21496]. The paper gives analytic profiles for linear exponential photoconductivity, saturable profiles, two-photon-absorption (TPA) profiles, and combined saturable+TPA profiles [2410.21496]. Depending on the profile, the wave equation reduces to the Bessel equation, Gauss hypergeometric equation, or Heun equation, yielding exact analytic expressions for the optical coefficients [2410.21496]. In the thin-film limit, a saturable profile produces $\Delta R/R\propto \ln(1+F/F_{\mathrm{sat}})$, while the TPA thin-film form is stated to be indistinguishable from linear behavior in $\Delta R/R(F)$ at low fluence [2410.21496]. This provides a systematic framework for constraining systematic uncertainty associated with nonlinear photoconductivity profile distortion [2410.21496].

A different nonlinear reflectance regime is selective reflection from a YAG window–high-density rubidium vapor interface. There, the reflection coefficient is written as
\[
r(\omega)=\frac{n_{1}-n_{2}(\omega)}{n_{1}+n_{2}(\omega)}
\]
with $n_{1}=1.821$ for YAG at 780 nm and $n_{2}(\omega)=\sqrt{1+\chi(\omega)}$ for the vapor [2312.06243]. For $|\chi|\ll 1$, the resonant change in reflectance is proportional to $\mathrm{Re}[\chi(\omega)]$, making the signal dispersion-like [2312.06243]. The linear susceptibility near the Rb D$_2$ resonance is approximated by a Lorentzian with linewidth dominated by dipole–dipole broadening,
\[
\Gamma(N)=KN,\qquad K/2\pi=(1.10\pm0.17)\times10^{-16}\ \mathrm{GHz\cdot cm}^{3}\,,
\]
so that for $N=(1.2$–$3.6)\times 10^{17}\ \mathrm{cm}^{-3}$ the broadening spans 13.2–39.6 GHz [2312.06243]. Pump intensities from 1.2 to 8.8 kW$\cdot$cm$^{-2}$ reduce both resonance magnitude and width through resonant saturation, with the measured selective-reflection width collapsing from about 16 GHz to about 8 GHz at the lowest density and from about 42 GHz to about 20 GHz at the highest density over the reported intensity range [2312.06243]. For $N=2.5\times10^{17}\ \mathrm{cm}^{-3}$, the resonance amplitude decreases from about $10\%$ at zero pump to about $3\%$ at 8.8 kW$\cdot$cm$^{-2}$ [2312.06243]. The paper suggests all-optical modulation as an application, with estimated modulation depth 50–60% and response times 4–15 ps inferred from $\Gamma$ [2312.06243].

On the theoretical side, a generalized two-time linear-response theory for pumped systems extends OPPR analysis to generic pump–probe delays and probe frequencies [2404.10768]. In that framework, the total Hamiltonian is split into a pumped part $H'(t)=H_{0}+H_{\mathrm{pu}}(t)$ and a weak probe coupling $H''(t)$, leading to a two-time susceptibility $\chi(t,t')$ and a two-time optical conductivity $\sigma(t,t_{\mathrm{pr}})$ [2404.10768]. The out-of-equilibrium dielectric function is then
\[
\epsilon(\omega,t_{\mathrm{pr}})=1+\frac{i}{\omega\epsilon_{0}}\sigma(\omega,t_{\mathrm{pr}})
\]
and for $s$-polarized light at incidence angle $\theta$ the reflectance is approximated in the XUV by the usual Fresnel formula involving $\epsilon(\omega,t_{\mathrm{pr}})$ [2404.10768]. The work emphasizes that inter-band transitions dominate sharp resonance features, intra-band motion contributes smoother background modulations, and inclusion of a local dipole term introduces odd-order pump-frequency oscillations in $\Delta R(\omega,\omega')$, whereas pure Peierls coupling yields only even harmonics [2404.10768]. The associated Dynamical Projective Operatorial Approach permits one-shot pump-only time propagation followed by postprocessing for all pump–probe delays and probe spectra [2404.10768].

## 7. Assumptions, limitations, and methodological practice

Quantitative OPPR depends on a sequence of approximations and calibration choices that are explicit in the cited works. The dielectric inversion of $\Delta R/R$ assumes the perturbative limit, namely that $\Delta\epsilon(\omega,t)$ remains small enough for first-order Taylor expansion to hold [2108.10605]. It also assumes that KK reconstruction is sufficiently accurate within a finite measured spectral window, which requires that evolving features be local and well bounded or, as a practical recommendation, that the spectral range include the full feature and its tails beyond band edges [2108.10605]. Errors in the static dielectric baseline propagate directly into $A(\omega)$ and $B(\omega)$, and neglected Fabry–Pérot fringes, film-thickness uncertainty, or inhomogeneities can distort the inversion if not built into $R(\epsilon)$ [2108.10605]. Strong spectral shifts or high-fluence effects such as bandgap renormalization may require repeated expansion around a shifted baseline or inclusion of higher-order terms [2108.10605].

Instrumental practices likewise determine reliability. The broadband OPPR guide recommends high static-spectrum signal-to-noise ratio, stated as $\ge 10^{4}$ counts, careful chirp calibration, homogeneous excitation densities, and fluence-linearity tests [2108.10605]. The 45 fs broadband apparatus reports a noise floor of $\lesssim 10^{-4}$ RMS after 1000 shots per integration, at least 50 repeats per matrix, and binning of about 5–20 pixels in photon energy [1608.03503]. In THz reflection geometry, the principal error sources are laser intensity and timing jitter, beam-pointing drift, residual water-vapor absorption, detector-crystal alignment, and pump-induced sample heating or damage at high fluence [1708.01991]. In PFID measurements, the signal may vanish under spectral integration, so monochromator selection is not incidental but essential to observing the negative-delay coherence term [1809.00292]. In coherent-phonon reflectivity, spectral integration can suppress the oscillatory signal by destructive interference between sidebands, whereas band-pass filtering can enhance the detected oscillation amplitude by a factor of about 35 [1603.06023].

Several broader implications follow from the literature but should be stated cautiously. The collected results suggest that OPPR is not a single fixed observable but a family of reflectance-based measurements whose interpretation depends on geometry, spectral resolution, polarization configuration, depth profile, and inversion model. They also suggest that reflectance-only data can be elevated from qualitative kinetics to quantitative dielectric reconstruction when equilibrium optical constants, derivative optics, KK constraints, and independent benchmarks such as FDI are available [2108.10605]. A plausible implication is that future OPPR work will continue to converge with transfer-matrix modeling, two-time nonequilibrium response theory, and spectrally selective detection rather than relying only on empirical fitting of raw $\Delta R/R$ traces [2108.10605], [2404.10768], [2410.21496].

Source: https://www.emergentmind.com/topics/optical-pump-probe-reflectance-oppr