Spectroscopic Limited Maximum Efficiency (SLME)

• SLME is a photovoltaic figure of merit that estimates the maximum efficiency of a single-junction solar absorber by incorporating realistic, energy-dependent absorption spectra and finite film thickness.
• It refines the Shockley–Queisser limit by replacing the ideal step-function absorptivity with material-specific optical responses derived from first-principles calculations.
• SLME serves as a practical design tool, linking microscopic optical transitions to device-level efficiency and guiding materials screening and thickness engineering.

Searching arXiv for recent and foundational papers on Spectroscopic Limited Maximum Efficiency (SLME). Spectroscopic Limited Maximum Efficiency (SLME) is a photovoltaic figure of merit that estimates the maximum power-conversion efficiency of a single-junction solar absorber under detailed balance while replacing the idealized absorptivity of the Shockley–Queisser (SQ) limit with the actual optical response of a specific material. In the literature summarized here, SLME is consistently defined as a spectroscopically constrained, thickness-dependent efficiency limit derived from the absorption coefficient $\alpha(E)$, the AM1.5G solar spectrum, and a radiative-limit current–voltage relation, with extensions in some works to include non-radiative recombination through phenomenological radiative fractions or explicit Shockley–Read–Hall terms (Choudhary et al., 2019). Across recent first-principles studies, SLME has become a standard screening metric linking microscopic optical spectra to device-level photovoltaic potential for bulk semiconductors, perovskites, chalcogenides, antiperovskites, and low-dimensional absorbers (Adhikari et al., 2 Apr 2026).

1. Conceptual definition and relation to the Shockley–Queisser limit

SLME is a detailed-balance efficiency limit like the SQ limit, but it uses the actual, energy-dependent absorption coefficient $\alpha(E)$, a finite absorber thickness $L$, and the real nature of the band gap through the absorption onset rather than assuming an ideal step-function absorber (Dhariwal et al., 20 Jun 2026). In this sense, it is a materials-specific upper bound on photovoltaic efficiency rather than one determined solely by the band gap (Behera et al., 2023).

The distinction from SQ is central in essentially all treatments. SQ assumes a step-like absorptivity $A_{\text{SQ}}(E)=\Theta(E-E_g)$, infinite thickness or perfect absorption above $E_g$, and purely radiative recombination, so only the band gap matters (Dhariwal et al., 20 Jun 2026). SLME instead uses an absorptivity derived from first-principles $\alpha(E)$, accounts for incomplete absorption in thin films, includes the spectral shape, magnitude, and onset of absorption, and distinguishes materials with the same $E_g$ but different absorption strength (Dhariwal et al., 20 Jun 2026). This difference is not merely formal. Several studies explicitly show that band-gap-only reasoning can fail: materials with near-optimal gaps may exhibit reduced SLME because of weak or anisotropic absorption, while materials with somewhat non-optimal gaps can still achieve high SLME if their optical transitions are strong (Irfan et al., 18 Jun 2025).

The literature also treats SLME as an “improved version” or “more realistic assessment” than SQ because it incorporates real absorptance and finite thickness (Behera et al., 2023). Some works emphasize that SLME is typically below the SQ limit for a given gap because real absorptivity is not ideal and thickness is finite (Dhariwal et al., 20 Jun 2026). However, a notable exception is the analysis of CuAu-like chalcogenides, which argues that SLME can exceed the SQ limit within the same detailed-balance framework because the SQ step-function absorptivity can overestimate the radiative recombination current $J_0$ relative to a finite-thickness realistic absorber (Bercx et al., 2016). This establishes a methodological nuance rather than a contradiction: SLME and SQ differ not only in photogeneration but also in the radiative dark current through their different absorptivity models.

Within the broader screening literature, SLME is used precisely because it captures the difference between fundamental and direct-allowed gaps, weak versus strong optical onset, and finite-thickness optics, all of which are invisible to gap-only SQ screening (Choudhary et al., 2019). A plausible implication is that SLME is best understood not as a replacement for SQ’s ideal thermodynamic bound in the abstract, but as a more discriminating upper bound for a concrete absorber realized as a finite film.

2. Formalism and core equations

In the standard Yu–Zunger formalism used throughout this literature, the absorptivity of a film of thickness $L$ is written as

$A(E,L)=1-e^{-2\alpha(E)L},$

with the factor of 2 corresponding to the double-pass approximation used in standard SLME implementations (Dhariwal et al., 20 Jun 2026). Some treatments also present the single-pass variant

$\alpha(E)$0

noting that the convention depends on the optical model adopted (Behera et al., 2023). The dominant formulation in the cited computational studies is the double-pass form (Hansraj et al., 2021).

The short-circuit current density is obtained by integrating the absorbed solar photon flux: $\alpha(E)$1 where $\alpha(E)$2 is the elementary charge and $\alpha(E)$3 is the AM1.5G solar photon flux (Dhariwal et al., 20 Jun 2026). The radiative dark current is computed analogously using the black-body photon flux: $\alpha(E)$4 with

$\alpha(E)$5

in one common convention (Dhariwal et al., 20 Jun 2026). The current–voltage relation in the radiative limit is then

$\alpha(E)$6

The output power density is

$\alpha(E)$7

and the efficiency is

$\alpha(E)$8

or, equivalently, the energy-weighted integral over the solar photon flux depending on notation (Dhariwal et al., 20 Jun 2026).

A recurrent refinement relative to the strict radiative-limit form is the introduction of a radiative fraction $\alpha(E)$9 or an equivalent factor accounting for the difference between the fundamental gap $L$0 and the direct-allowed gap $L$1. In the JARVIS-based high-throughput framework, this is written as

$L$2

so materials with large $L$3 are strongly penalized (Choudhary et al., 2019). Related formulations appear in older chalcogenide studies through a radiative-fraction model parameterized by $L$4 (Sarmadian et al., 2016). This suggests that the practical definition of SLME in the literature spans a family of closely related detailed-balance models rather than a single invariant formula, though all preserve the same central idea: realistic absorptivity replaces the SQ step function.

A more elaborate extension appears in the study of Mg$L$5Si and Ca$L$6Si, where the radiative SLME is supplemented by an explicit non-radiative Shockley–Read–Hall current,

$L$7

to define a trap-limited conversion efficiency beyond SLME (Solet et al., 2024). That work keeps SLME as the radiative baseline and then adds defect-mediated recombination to assess the drop from the spectroscopic limit to a more realistic defect-limited efficiency.

3. First-principles inputs and computational pipeline

Across the cited studies, SLME is not computed from empirical optical data alone but from first-principles electronic structure and optical spectra. The typical workflow begins with a band-structure calculation using DFT or many-body perturbation theory, followed by computation of the complex dielectric function

$L$8

from which the refractive index, extinction coefficient, and absorption coefficient are derived (Dhariwal et al., 20 Jun 2026).

Several methodological tiers appear in the literature. A widely used semilocal strategy combines PBE structural optimization with the Tran–Blaha modified Becke–Johnson potential for improved gaps, as in Ca$L$9PX$A_{\text{SQ}}(E)=\Theta(E-E_g)$0 and in the JARVIS screening work (Dhariwal et al., 20 Jun 2026). Hybrid-functional pipelines are also common, for example HSE06 for ACuX chalcogenides and AsNCa$A_{\text{SQ}}(E)=\Theta(E-E_g)$1 optical interpolation via Wannier-based methods (Behera et al., 2023). At the highest level, GW and Bethe–Salpeter equation calculations are used to include quasiparticle and excitonic effects explicitly, as in GeSe$A_{\text{SQ}}(E)=\Theta(E-E_g)$2, LiZnAs and ScAgC, Si$A_{\text{SQ}}(E)=\Theta(E-E_g)$3O, AZrS$A_{\text{SQ}}(E)=\Theta(E-E_g)$4, and Ba$A_{\text{SQ}}(E)=\Theta(E-E_g)$5MA$A_{\text{SQ}}(E)=\Theta(E-E_g)$6 antiperovskites (Kang et al., 10 Feb 2026).

The optical input is crucial because SLME is highly sensitive to the shape of $A_{\text{SQ}}(E)=\Theta(E-E_g)$7 near the band edge. In the ACuX study, direct allowed transitions in the hexagonal Se/Te phases produce $A_{\text{SQ}}(E)=\Theta(E-E_g)$8 in the visible, which is directly tied to higher SLME (Behera et al., 2023). In LiZnAs and ScAgC, BSE yields high absorption coefficients of $A_{\text{SQ}}(E)=\Theta(E-E_g)$9 and correspondingly high SLME at sub-micron thickness (Solet et al., 1 Apr 2025). In Si$E_g$0O, excitonic effects shift the optical gap to about 1.2 eV and help produce an SLME of about 27% at $E_g$1 (Kim et al., 2020). These examples show that many-body optical spectra can materially change SLME relative to independent-particle calculations.

Thickness is treated as an explicit device parameter rather than a material constant. Many papers present $E_g$2 curves that rise with $E_g$3 and then saturate once absorption becomes effectively complete (Dhariwal et al., 20 Jun 2026). Typical saturation thicknesses vary substantially. Ca$E_g$4PI$E_g$5 reaches approximately 30% SLME at $E_g$6 (Dhariwal et al., 20 Jun 2026), NaCuTe reaches about 18% at 8–10 $E_g$7m (Behera et al., 2023), and type-II GeSe$E_g$8 reaches about 25.6% already at $E_g$9 (Kang et al., 10 Feb 2026). This variability underscores one of SLME’s main uses: it translates optical strength into a thickness requirement.

4. Physical determinants of SLME

The most important material determinant of SLME in the cited literature is not the band gap alone but the combination of band-gap magnitude, directness, and spectral absorption strength. Direct allowed transitions at the fundamental gap or very near it are repeatedly associated with high SLME because they produce strong near-edge absorption and therefore high absorptivity even in thin films (Behera et al., 2023). Conversely, indirect gaps or optically forbidden direct transitions delay the onset of strong $\alpha(E)$0, depressing $\alpha(E)$1 and often increasing the effective dark current penalty in formulations that distinguish $\alpha(E)$2 from $\alpha(E)$3 (Choudhary et al., 2019).

This distinction is made especially clearly in several case studies. In ACuX, all compounds are direct-gap in the band-structure sense, but orthorhombic NaCuS, KCuS, and KCuSe have zero transition dipole matrix element at $\alpha(E)$4, indicating optically forbidden VBM→CBM transitions; their SLME is therefore not emphasized despite direct band gaps (Behera et al., 2023). In AsNCa$\alpha(E)$5, the yellow phase has a band gap of 1.33 eV, essentially at the SQ optimum, yet its SLME is only 27.25% because optical anisotropy and uneven absorption reduce photon harvesting efficiency, especially at lower energies (Irfan et al., 18 Jun 2025). In Mg$\alpha(E)$6Si, the large direct optical gap of about 2.45 eV relative to the much smaller indirect fundamental gap yields an SLME of only about 1.3%, far below its SQ limit of 19.37% (Solet et al., 2024).

Excitonic effects can either sharpen the absorption onset or shift it to lower energy and thereby enhance SLME. In type-II GeSe$\alpha(E)$7, a pronounced excitonic absorption peak near 2 eV enhances visible-light absorption and contributes to an SLME of $\alpha(E)$8 at $\alpha(E)$9 (Kang et al., 10 Feb 2026). In LiZnAs and ScAgC, BSE-based absorption is used directly in SLME, and bright excitons near the gap are part of the reason both compounds reach radiative-limit SLMEs of about 32% and 31%, respectively, at $E_g$0 (Solet et al., 1 Apr 2025). In Si$E_g$1O, the optical gap remains near 1.2 eV from monolayer to bulk because excitonic red shifts compensate the change in quasiparticle gap, and this near-optimal optical onset helps produce a very high SLME (Kim et al., 2020).

Several papers also connect dielectric screening, exciton binding, and SLME indirectly. In Ba$E_g$2MA$E_g$3, BSE reveals moderate exciton binding energies of 0.254–0.352 eV and intermediate-radius excitons, while SLME reaches roughly 19–32% depending on composition (Adhikari et al., 2 Apr 2026). The interpretation is that excitons enhance absorption near the band edge while remaining sufficiently extended that efficient separation remains plausible. This suggests a broader design rule: moderate excitonic enhancement can be beneficial for SLME provided it is not accompanied by unrealistically poor carrier separation.

5. Representative materials and numerical ranges

The numerical SLME values reported in the cited literature span a very broad range, from effectively negligible values for poor absorbers to values close to the SQ limit for near-ideal direct absorbers. The range itself is informative because it demonstrates how strongly SLME discriminates among candidate photovoltaic materials.

Material system	Reported SLME	Thickness context
Ca$E_g$4PF$E_g$5	0.60%	Saturated value (Dhariwal et al., 20 Jun 2026)
Mg$E_g$6Si	1.3%	BSE absorption, 300 K (Solet et al., 2024)
NaCuTe	about 18%	8 $E_g$7m film (Behera et al., 2023)
Ca$E_g$8PI$E_g$9	29.60%	Saturated near $J_0$0 (Dhariwal et al., 20 Jun 2026)
Ba$J_0$1SbNbO$J_0$2	26.8%	$J_0$3, 293.15 K (Hansraj et al., 2021)
type-II GeSe$J_0$4	$J_0$5	$J_0$6 (Kang et al., 10 Feb 2026)
LiZnAs	32%	$J_0$7 (Solet et al., 1 Apr 2025)
ScAgC	31%	$J_0$8 (Solet et al., 1 Apr 2025)
Ba$J_0$9BiI$L$0	32.41%	Saturation thickness (Adhikari et al., 2 Apr 2026)

In halide-substituted Ca$L$1PX$L$2, SLME increases monotonically from F to I as the band gap narrows from 3.788 eV to 2.000 eV and visible absorption strengthens, with Ca$L$3PF$L$4 at 0.60% and Ca$L$5PI$L$6 at 29.60% (Dhariwal et al., 20 Jun 2026). In the ACuX family, NaCuTe reaches about 18% at 8–10 $L$7m, while the other favorable hexagonal Se/Te phases lie between about 10% and 13% (Behera et al., 2023). In AsNCa$L$8, stable phases span 27.25–31.23%, with only modest variation across polymorphs (Irfan et al., 18 Jun 2025). In Ba$L$9MA$A(E,L)=1-e^{-2\alpha(E)L},$0, SLME ranges from 19.10% for Ba$A(E,L)=1-e^{-2\alpha(E)L},$1PCl$A(E,L)=1-e^{-2\alpha(E)L},$2 to 32.41% for Ba$A(E,L)=1-e^{-2\alpha(E)L},$3BiI$A(E,L)=1-e^{-2\alpha(E)L},$4 (Adhikari et al., 2 Apr 2026).

A recurring pattern is that materials with direct gaps in roughly the 1.0–1.6 eV range and strong visible absorption can reach SLME values above 25%, often close to the SQ optimum. LiZnAs at 1.5 eV and ScAgC at 1.0 eV are explicit examples (Solet et al., 1 Apr 2025). Ca$A(E,L)=1-e^{-2\alpha(E)L},$5Si, with a direct gap near 0.96 eV and strong BSE absorption, reaches an SLME of 31.2%, essentially its SQ limit of 31.25% (Solet et al., 2024). By contrast, materials with indirect gaps, large direct-allowed gaps, or ultraviolet-only absorption show dramatically lower SLME even when the fundamental gap alone might look promising (Solet et al., 2024).

6. Thickness dependence and device-level interpretation

Thickness dependence is one of SLME’s defining features. For small $A(E,L)=1-e^{-2\alpha(E)L},$6, absorptivity remains well below unity for much of the above-gap solar spectrum, so $A(E,L)=1-e^{-2\alpha(E)L},$7 and the efficiency are low (Dhariwal et al., 20 Jun 2026). As $A(E,L)=1-e^{-2\alpha(E)L},$8 increases, $A(E,L)=1-e^{-2\alpha(E)L},$9 approaches unity over larger energy ranges, and $\alpha(E)$00 rises. Beyond a material-specific saturation thickness, further increases yield little gain because optical absorption is already nearly complete (Behera et al., 2023).

The cited studies use this behavior to infer practical absorber thicknesses. In Ca$\alpha(E)$01PI$\alpha(E)$02, a minimum film thickness of $\alpha(E)$03 is needed to reach the maximum SLME of about 30% (Dhariwal et al., 20 Jun 2026). In Ba$\alpha(E)$04SbNbO$\alpha(E)$05, SLME saturates beyond $\alpha(E)$06 (Hansraj et al., 2021). In type-II GeSe$\alpha(E)$07, the high SLME at $\alpha(E)$08 suggests a truly thin-film absorber (Kang et al., 10 Feb 2026). In NaCuTe and related ACuX compounds, saturation requires 8–10 $\alpha(E)$09m, which is still relevant to thin-film processing but substantially thicker than the best direct absorbers (Behera et al., 2023).

One important implication drawn in the literature is that SLME can guide thickness engineering, not merely composition selection. Strongly absorbing materials saturate at micrometer or sub-micrometer thicknesses; weak absorbers may require impractically thick films to realize even their modest SLME (Dhariwal et al., 20 Jun 2026). This makes the $\alpha(E)$10 curve a compact descriptor of both intrinsic optical quality and technological practicality.

In device-level studies, SLME can align surprisingly well with more explicit simulations. For MgSnN$\alpha(E)$11, a 2 $\alpha(E)$12m film yields a room-temperature SLME of 13.17%, while a SCAPS simulation for a single-junction device yields 12.80%, close to the SLME prediction (Mahraj et al., 19 Mar 2026). This does not eliminate the idealizations in SLME, but it suggests that when optical absorption is the dominant intrinsic limitation and transport losses are modest, SLME can be an informative proxy for realistic upper-bound performance.

7. Limitations, controversies, and methodological caveats

The literature is explicit that SLME is an upper bound under idealized assumptions. The most common caveat is the radiative-limit assumption: non-radiative recombination through defects, surfaces, interfaces, trap-assisted pathways, and Auger processes is neglected unless added phenomenologically or explicitly in extended models (Dhariwal et al., 20 Jun 2026). Real device efficiencies are therefore expected to fall below SLME.

A second limitation is the use of perfect bulk optical spectra with no disorder, band tailing, grain boundaries, contact resistance, or parasitic absorption in transport layers (Dhariwal et al., 20 Jun 2026). The AsNCa$\alpha(E)$13 and Ba$\alpha(E)$14SbXO$\alpha(E)$15 studies, for example, treat absorber properties in isolation from full device architectures (Irfan et al., 18 Jun 2025). Likewise, many calculations neglect temperature-induced band-gap renormalization and sometimes spin–orbit coupling, even when heavy elements are present (Dhariwal et al., 20 Jun 2026).

A more specific controversy concerns indirect-gap materials. The CuAu-like chalcogenide and silicon analysis argues that the standard SLME recombination model can be unfair to indirect absorbers because the exponential radiative-fraction factor based on $\alpha(E)$16 can drive the reverse saturation current to unrealistically large values, drastically depressing the calculated efficiency (Bercx et al., 2016). Silicon is the extreme example: using the standard $\alpha(E)$17 model gives effectively zero efficiency, whereas assigning a more realistic $\alpha(E)$18 yields sensible results (Bercx et al., 2016). This suggests that SLME is most robust as a screening metric for direct or nearly direct absorbers and more problematic for strongly indirect materials unless the recombination model is refined.

Another methodological caveat is that SLME values can depend noticeably on the level of theory used to generate $\alpha(E)$19. In Mg$\alpha(E)$20Si and Ca$\alpha(E)$21Si, BSE absorption is essential to capture optical spectroscopy accurately (Solet et al., 2024). In Si$\alpha(E)$22O and GeSe$\alpha(E)$23, excitonic effects reshape the absorption onset and thus the SLME materially (Kim et al., 2020). This suggests that high-SLME predictions based on semilocal optics should be interpreted more cautiously than those based on GW+BSE, especially in low-dimensional or excitonic systems.

Finally, the high-throughput literature shows that SLME can be integrated into screening pipelines and even machine-learning models, but such workflows inherit any approximations in the underlying dielectric functions and band gaps (Ginter et al., 9 Oct 2025). They remain powerful for ranking and triage, but not substitutes for full device-level validation.

8. Role in screening, machine learning, and materials discovery

SLME has become a standard screening metric in large-scale materials discovery. In the JARVIS study, SLME was computed for 5097 non-metallic materials, of which 1997 had SLME higher than 10%; after effective-mass and convex-hull filters, 934 candidates remained (Choudhary et al., 2019). The same work used SLME as the label for a binary machine-learning classifier, with $\alpha(E)$24 defining the positive class, and then applied the model to hundreds of thousands of materials (Choudhary et al., 2019). This established SLME as both a physically meaningful metric and a scalable target for high-throughput discovery.

More recent work has shifted from predicting SLME directly to predicting the dielectric function or learning SLME-correlated latent representations from spectra. An ALIGNN model trained on $\alpha(E)$25 dielectric functions from JARVIS reproduced spectral features well enough that SLME computed from predicted spectra had a mean absolute error of 1.95 percentage points relative to DFT-based SLME, with $\alpha(E)$26 (Ginter et al., 9 Oct 2025). Applied to the Alexandria database, this workflow found that 16.3% of the screened insulating materials had $\alpha(E)$27, and that 22.9% of stoichiometrically defined perovskites exceeded this threshold (Ginter et al., 9 Oct 2025). Vanadium emerged as a strong indicator of high-SLME compounds, particularly in perovskites (Ginter et al., 9 Oct 2025).

In a different unsupervised direction, disentangling autoencoders trained on 17,283 optical absorption spectra learned a latent variable strongly correlated with SLME, with Pearson correlation $\alpha(E)$28, despite never seeing SLME labels during training (Cha et al., 25 Jul 2025). Latent traversal revealed that this dimension encoded the transition from concave, direct-gap-like absorption onset to convex, indirect-gap-like onset, which is precisely the spectral feature to which SLME is highly sensitive (Cha et al., 25 Jul 2025). This result suggests that SLME-relevant physics can emerge naturally from unsupervised learning on spectra, reinforcing the interpretation of SLME as fundamentally a functional of the absorption onset and line shape.

These developments indicate that SLME now occupies two complementary roles. In first-principles materials science, it is a spectroscopically grounded efficiency limit for specific compounds. In data-driven materials discovery, it is an effective bridge between spectral representations and target photovoltaic performance.

9. Design principles implied by the SLME literature

Across the cited work, a consistent set of design principles emerges. High SLME is associated with direct or nearly direct gaps, strong dipole-allowed transitions at the fundamental edge, large absorption coefficients in the visible, and film thicknesses that allow $\alpha(E)$29 to saturate without becoming impractically large (Adhikari et al., 2 Apr 2026). Band gaps in roughly the 1.1–1.6 eV range are repeatedly favored for single-junction applications, though higher-gap materials can still achieve high SLME when used as top cells or when optical strength is exceptional (Solet et al., 1 Apr 2025).

Halide or chalcogen substitution is often used to tune SLME systematically. In Ca$\alpha(E)$30PX$\alpha(E)$31, moving from F to I narrows the gap and increases visible absorption, producing a monotonic SLME increase (Dhariwal et al., 20 Jun 2026). In Ba$\alpha(E)$32MA$\alpha(E)$33, moving toward heavier pnictogens and halides drives the gap toward the optimal range and raises SLME from about 19% to above 32% (Adhikari et al., 2 Apr 2026). In mixed perovskites such as RbPb$\alpha(E)$34Ge$\alpha(E)$35I$\alpha(E)$36, intermediate alloying levels can maximize SLME by balancing gap reduction, stability, and absorption (Nyayban et al., 2022).

A plausible implication is that SLME is most valuable when used not as an isolated score but as a design map linking composition, structure, optical transitions, and thickness requirements. This is how it is used in the best of the cited studies: not merely to rank materials, but to explain why some compositions or phases are promising and how those properties might be tuned further.

10. Encyclopedia summary

SLME is a spectroscopically informed photovoltaic efficiency limit derived from detailed balance but constrained by a material’s real absorption spectrum and finite thickness. It replaces the ideal absorptivity of the Shockley–Queisser model with a first-principles absorptivity built from $\alpha(E)$37, thereby incorporating band-edge transition strength, direct versus indirect character, and thickness-dependent incomplete absorption (Dhariwal et al., 20 Jun 2026). In its standard form, it remains a radiative-limit metric; in extended forms, it can be supplemented by empirical radiative fractions or explicit SRH recombination to estimate more realistic defect-limited efficiencies (Solet et al., 2024).

The recent arXiv literature shows that SLME has matured into a standard tool for screening photovoltaic absorbers, from halide-substituted perovskites and oxide double perovskites to chalcogenides, antiperovskites, and low-dimensional van der Waals materials (Adhikari et al., 2 Apr 2026). It consistently reveals distinctions that band-gap-only SQ analysis misses: direct allowed absorbers with strong visible $\alpha(E)$38 can approach or match the SQ limit, while indirect or optically weak absorbers can fall far below it even if their fundamental gaps seem favorable (Behera et al., 2023). At the same time, the literature identifies clear caveats, especially for strongly indirect materials and for calculations lacking explicit treatment of defects, interfaces, and non-radiative losses (Bercx et al., 2016).

In contemporary materials discovery, SLME functions both as a rigorous first-principles screening metric and as a target or latent property in machine-learning workflows, enabling exploration of vast compositional spaces while preserving a direct connection to device-relevant optical physics (Ginter et al., 9 Oct 2025). Its enduring importance lies in that dual role: it is both a compact scalar efficiency estimate and a physically transparent framework for translating spectroscopic detail into photovoltaic design.