Pith. sign in

REVIEW 3 major objections 5 minor 1 cited by

Sum-of-Squares Bounds on Surface-Enhanced Raman Scattering

T0 review · 3 major / 5 minor · reviewed 2026-08-07 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read SERS bounds come within 3x of best designs

desk verdict A genuinely novel SOS-based method for quartic SERS bounds, but the reported 'bounds' are not yet proven because the Green's-function tail is dropped without a rigorous error estimate. read the letter →

arxiv 2502.09821 v1 pith:SHXMPR43 submitted 2025-02-13 physics.optics

Pengning Chao , Ian M. Hammond , Steven G. Johnson This is my paper · ORCID
classification physics.optics
keywords surface-enhancedRamanscatteringsum-of-squaresprogrammingmetasurfacedesignfundamentallimitsinphotonicstopologyoptimizationquarticfigureofmeritsurfaceplasmonpolaritonsguidedmodes
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper seeks to establish that the maximum spatially averaged surface-enhanced Raman scattering (SERS) achievable by any periodic metasurface has a computable finite upper bound, even though the figure of merit is the quartic integral $\int \lVert \mathbf{E} \rVert^4 \, d\mathbf{r}$ and therefore outside the quadratic bounds previously available in photonics. It derives the bound by relaxing Maxwell's equations to global power-conservation constraints on the induced polarization and enforcing nonnegativity of a generalized Lagrangian through sum-of-squares (SOS) programming, the first use of SOS techniques in optics. Against topology-optimized inverse designs in two dimensions, the bounds are typically within a factor of ten and often within a factor of three for $E_z$-polarized fields. If the bounds are correct, SERS metasurface designers gain a target: whenever an inverse design approaches the bound, the structure is near the global optimum, and the bounds also expose which physical mechanisms are fundamentally worth exploiting.

What carries the argument

The key object is the generalized Lagrangian dual of the polarization-field optimization problem, with the multiplier promoted from a scalar to a quadratic polynomial so that the product with the quadratic power-conservation constraint has the same quartic degree as the objective. Nonnegativity of the resulting polynomial is imposed by the stronger sum-of-squares condition, which converts the dual into a semidefinite program. To keep the program small, the vacuum Green's function $\mathbf{G}_{RD}$ is compressed by a singular-value expansion and the polarization is represented in the $q$ dominant right singular vectors; with a molecule-metasurface separation $d>0$ the singular values decay exponentially, and the paper reports that $q<20$ gives converged bounds.

What would settle it

For a reported geometry, run a much denser inverse-design search (many random initializations and a finer discretization) and evaluate the true $\int \lVert \mathbf{E} \rVert^4 \, d\mathbf{r}$ figure of merit; a single structure whose value exceeds the corresponding SOS bound by more than numerical tolerance would show the bound is not an upper bound. Alternatively, recompute the bound with a rigorous bound on the discarded tail and check whether the stated factor-of-ten tightness survives.

Watch

Extended reading notes

Core claim

The central discovery is that a quartic, spatially averaged SERS figure of merit can be bounded from above over all allowed material distributions by solving a semidefinite program, rather than by factoring the figure of merit into quadratic pieces as earlier nonlinear bounds did. Working in the induced-polarization representation, the paper keeps only the dominant singular-value modes of the vacuum Green's function connecting the design region to the Raman-molecule region, which makes the SOS program tractable; numerical convergence is reached for a cutoff below about twenty modes. The computed bounds agree closely with inverse-designed dielectric and metallic structures, and they reveal two concrete physical conclusions. First, a suspended lossy metasurface can achieve theoretically diverging SERS enhancement as its period approaches the vacuum wavelength from below, because arbitrarily thin waveguides support progressively more delocalized high-$Q$ guided modes whose quality factor grows fast enough to overcome the reduced field overlap. Second, for metallic structures the $E_z$ polarization is fundamentally limited because it cannot excite surface plasmon polaritons, while $H_z$-polarized fields retain the large enhancements.

Load-bearing premise

The bounds rely on the assumption that the parts of the field left out of the truncated calculation are too small to matter; the paper checks this numerically but does not prove it, so the results are converged estimates rather than guaranteed upper limits.

Editorial extensions

If this is right

  • Inverse designs that come within the reported bounds are near global optima, so the bounds give topology optimization a practical stopping criterion for SERS metasurfaces.
  • Suspended thin-waveguide grating designs with period approaching the vacuum wavelength from below are identified as the optimal route to large-area SERS, with enhancement that theoretically diverges even for lossy materials.
  • For metallic metasurfaces, $E_z$-polarized pumping is a low-ceiling strategy; experimenters should use $H_z$ polarization or accept the reflection-limited performance.
  • Adding a uniform loss factor of $1+i/600$ caps the modal quality factor, removes the divergence, and produces a sharp finite transition at $L=\lambda$, matching realistic fabrication and bandwidth limits.
  • The SOS approach extends to any nonlinear photonics figure of merit that is a polynomial or rational function of the fields, so the same machinery can bound second-harmonic generation, bistability thresholds, or lasing thresholds.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • Testable extension: apply the same generalized-Lagrangian SOS relaxation to a Kerr-nonlinearity figure of merit, such as a bistability threshold, and compare the bound against inverse-designed cavities; agreement would transfer the technique's apparent tightness to a new nonlinear regime.
  • The near-$\chi$-independence of the $E_z$ bound away from the divergence suggests that a simple multilayer reflector may already saturate the bound at small periods, leaving little headroom for complex topology optimization of that polarization.
  • Because the tail of the Green's function expansion is dropped without a rigorous error bound, the printed numbers are converged estimates; constructing a certified upper bound by bounding the tail in $L^4$ would settle whether the factor-of-ten tightness holds rigorously.
  • The substrated logarithmic divergence, whose physical origin the paper leaves open, could be probed by varying substrate loss and watching the prefactor of $\log(\lambda-L)$; tracking that prefactor against substrate absorption would test whether the divergence is a loss-softened remnant of the suspended-structure power-law divergence.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

3 major / 5 minor

Summary. This paper proposes a sum-of-squares (SOS) programming framework for computing upper bounds on the spatially averaged SERS figure of merit ∫_{Ω_R} ||E||^4 dr for periodic 2D metasurfaces. The authors relax the structural Maxwell design problem to an optimization over induced polarizations subject to global power-conservation constraints, promote the Lagrange multiplier to a quadratic polynomial to handle the quartic objective, and map the resulting nonnegativity condition to a semidefinite program. To make the SDP tractable, they expand the vacuum Green's function in singular vectors and retain only a dominant subspace. The computed bounds are compared with topology-optimized designs for suspended and substrate-supported structures, leading to claims of tightness, a divergence of enhancement as the period approaches the pump wavelength, a suppression of Ez-polarized SERS for metals, and singularity-strength scaling with molecule–surface separation.

Significance. If the bounds are rigorous, the paper contributes the first application of SOS programming to optics and a general method for bounding non-quadratic, nonlinear photonic FOMs. The physical conclusions—particularly the delocalized high-Q guided-mode route to divergence and the Ez-polarization ceiling for metals—are valuable and consistent with existing intuition. The derivation of the generalized Lagrangian dual and the SOS-to-SDP mapping is mathematically sound, and the comparison against independent inverse-design benchmarks is a strong feature. The main obstacle is that the practical computations do not currently establish the claimed upper-bound status, because the truncation of the Green's function expansion is not certified.

major comments (3)
  1. [Sec. 3.2, Eq. (6); SI Eq. (10)] The triangle inequality gives f(P) ≤ (||E_i + A_q||_4 + ||B_q||_4)^4, but the SOS program (7) bounds only ||E_i + A_q||_4^4. The statement that the tail 'can in practice be neglected once bounds on the first term have converged' is not a proof: convergence of the SOS value for the first term does not control ||B_q||_4, and the fourth power of the sum is not the sum of fourth powers. Feasible polarizations may have components in the discarded singular vectors, and the projected constraints do not force those components to vanish. Without an explicit bound on ||B_q||_4 (e.g., from the power-conservation bound on ||P|| and the exponential decay of s_j), the values reported in Figs. 3 and 4 and the tightness factors in Sec. 4.1 are not certified upper bounds.
  2. [Secs. 4.1–4.3] The physical claims of a diverging enhancement as L→λ and of a fundamental Ez-polarization limitation for metals are invoked as consequences of the computed upper bounds. Because the bound status is not yet established, these claims are currently supported only as numerically converged estimates. The authors should either supply a certified tail bound (or solve the SOS program over the full tail) or clearly re-label the results as approximate bounds in the abstract and throughout the paper.
  3. [SI Table 1] The convergence study in the SI shows that the SOS bound on the first term in Eq. (10) saturates for q≥5, but it provides no information on the discarded tail ||B_q||_4. This does not address the objection raised above; a separate tail estimate or a rigorous bound on the tail operator restricted to the feasible set is needed.
minor comments (5)
  1. [Eq. (4)] The displayed generalized Lagrangian in Eq. (4) is garbled in the arXiv rendering; the expression should be typeset cleanly.
  2. [References] Reference [59] for SumOfSquares.py is incomplete; it should list the authors, version, and a stable URL or DOI.
  3. [Data Availability] The data availability statement says data are not publicly available; given the strongly numerical character of the paper, releasing the code and datasets would substantially improve reproducibility.
  4. [Fig. 3(d) inset] The log-log slopes in the inset of Fig. 3(d) are reported without describing the fitting range or any uncertainty; please specify how the exponents were extracted.
  5. [Sec. 3.2] The sentence 'the bounds generally saturate with q < 20' should be quantified per panel: please state the q used for each curve and the saturation criterion applied.

Circularity Check

0 steps flagged · score 2.0 of 10

No significant circularity: the SOS bound values come from a Maxwell-based SDP relaxation, not from the inverse-design benchmarks; the low-rank tail truncation is an unproven-certificate issue, not a circular reduction.

full rationale

The derivation chain is self-contained against the external topology-optimization benchmark. Starting from Maxwell's equations, the paper relaxes the structural optimization problem to an optimization over polarization P with the single quadratic power-conservation constraint C(P)=0 (Eq. 3), then forms a generalized Lagrangian dual (Eq. 4), replaces non-negativity by the sufficient SOS condition (Eq. 5), and projects onto the dominant singular-vector subspace of G_RD (Eqs. 6-7). Each step is a relaxation or dual bound, so no feasible structure is excluded; the bound value is the optimum of an SDP solved with MOSEK/YALMIP, not a number fitted to the inverse-designed structures. The inverse designs in Figs. 3-4 are independent topology-optimization results and are not used as inputs to the SOS program. The only step that could be mistaken for circularity is the treatment of the Green's-function tail in Eq. (6): the paper drops the second triangle-inequality term after observing numerical convergence of the first term, and the SI states that the SOS bound "converges to a fixed value which is the SERS bound." That is a gap in certifiedness (the reported values are numerically converged estimates, not proven upper bounds), but it is not a definitional or fitted-input circularity: the discarded tail is not used to define the bound as equal to a target, and no parameter is calibrated against the FOM being predicted. The paper cites prior work by the same authors for the general bounding framework (refs. 6, 35) and for related physical motifs (refs. 9, 44), but those citations provide independent, parameter-free formulations and do not by themselves force the numerical results; the SOS derivation and computations stand on the equations presented here. Accordingly, no load-bearing circular step is exhibited, and the score reflects only the presence of minor, non-load-bearing self-citation.

Assumptions & free parameters 3 free parameters · 7 assumptions · 0 invented entities

The central claim rests on standard Maxwell constraints and polynomial-optimization theory, plus a computational truncation assumption. The only hand-chosen numbers are the SVD cutoff q, the constraint interpolation gamma, and the artificial loss cap; none are fitted to the inverse-design results. No invented physical entities appear.

free parameters (3)
  • mode cutoff q = 5 to 26 (converged by q=5 in SI Table 1; typical q<20)
    Truncation order of the Green's function SVD kept in the SOS program. Bounds converge with q, but the tail is neglected without a certified error bound.
  • constraint interpolation parameter gamma = optimized via Brent's method, no single value reported
    Interpolates resistive and reactive power constraints in Eq. (7). Chosen to tighten the bound, not fitted to inverse-design results.
  • uniform loss factor 1 + i/600 = 1 + i/600
    Hand-selected artificial loss applied to all permittivities in Sec. 4.4 to cap modal Q at about 300 and regularize the divergence. Not derived from the physics and not fitted to inverse designs.
assumptions (7)
  • domain assumption The global power-conservation constraint C(P)=0 (Eq. 3b) is necessary for every physical polarization field P in the design region.
    This is the entire basis for the relaxation in Eq. (3). If it were not necessary, the relaxed maximum would not bound the physical design problem.
  • domain assumption The FOM ∫_{Ω_R}||E||^4 dr correctly represents spatially averaged SERS for isotropic molecules, reciprocal materials, and negligible pump/Raman frequency shift.
    Taken from Refs. [1,2]. All bounds and inverse designs target this FOM, so the physical relevance rests on this prior result.
  • ad hoc to paper For q large enough, the tail of G_RD's SVD contributes negligibly to the FOM.
    Eq. (6) bounds the full FOM by the dominant-subspace term plus a tail, and the tail is dropped with convergence heuristics rather than a rigorous error bound. This makes the reported values numerically converged estimates rather than certified bounds.
  • standard math Equality constraints C_I=0 and C_γ=0 project to inequality constraints C_I(x)≥0 and C_γ(x)≥0 in the dominant subspace.
    Justified in Supplementary Section 1 via Schur complements. The projection of an ellipsoid surface onto a subspace is a filled ellipsoid.
  • domain assumption A positive separation d>0 between Ω_R and Ω_D makes G_RD approximately low rank and prevents field-singularity divergences.
    This separation is introduced in Eqs. (1) and used throughout. It is physically motivated by nonlocal response and fabrication constraints, which are not modeled.
  • standard math Replacing polynomial nonnegativity with a sum-of-squares condition preserves a valid upper bound via weak duality.
    Standard result in polynomial optimization (Refs. [3-5]). It is what converts the generalized Lagrangian into a solvable SDP.
  • domain assumption Materials are described by local, frequency-independent permittivities; nonlocal response is not modeled.
    The paper states nonlocal polarizability is an open problem. Bounds are computed for local χ values such as Ag at 540 nm, so results apply at separations where nonlocal effects are negligible.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Sum-of-Squares Bounds on Surface-Enhanced Raman Scattering." pith.science (2026). https://pith.science/paper/SHXMPR43

@misc{pith2026250209821,
  author       = {Pith},
  title        = {Pith review of: Sum-of-Squares Bounds on Surface-Enhanced Raman Scattering},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/SHXMPR43}},
  note         = {Machine review of arXiv:2502.09821}
}
abstract

Surface-enhanced Raman scattering (SERS) is a critical tool for chemical sensing and spectroscopy, and a key question is how to optimally design nanostructures for maximizing SERS. We present fundamental limits on spatially-averaged SERS via periodic metasurfaces, derived using sum-of-squares (SOS) programming. This work represents the first use of SOS techniques to optics, overcoming difficulties that prior bounding techniques have with regards to non-linear photonic processes with higher order figures of merit. Our bounds on the $\int \lVert \mathbf{E} \rVert^4 \text{d} \mathbf{r}$ SERS enhancement factor for 2D examples demonstrate remarkable tightness when compared with inverse-designed dielectric and metallic structures for both electrical field out-of-plane ($E_z$) and in-plane ($H_z$) polarizations. We show that delocalized high-Q guided modes can achieve significant, theoretically diverging SERS enhancement even in the presence of material loss. For metallic structures, we demonstrate a fundamental performance limitation for $E_z$ polarized drive fields due to surface plasmon excitation restrictions. By varying the separation between Raman-active molecules and the metasurface design region, we also find material-dependent bounds on the maximum strength of field singularities. Our results offer insights into optimal metasurface design strategies for enhancing light-matter interactions, and our methodology may be adapted to the study of other nonlinear photonics design problems.

Figures

Figures reproduced from arXiv: 2502.09821 by the authors.

Figure 1
Figure 1. Schematic of the spatially-averaged SERS setting investigated in this work: [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. Plot of the singular values of 𝐺𝑅𝐷 with 𝐻𝐷 = 𝐻𝑅 = 0.5𝜆 and 𝑑 = 0.2𝜆. Given finite separation between Ω𝑅 and Ω𝐷, 𝐺𝑅𝐷 is effectively low-rank with the singular values decaying exponentially. To reduce the computational cost, we restrict SOS analysis to the subspace spanned by the dominant singular vectors, marked schematically in red. Inset: equality constraints (blue ellipsoid surface) for 𝑃 are projected down into t… view at source ↗
Figure 3
Figure 3. Bounds and inverse designs for 2D metasurfaces periodic in the [PITH_FULL_IMAGE:figures/full_fig_p007_3.png] view at source ↗
Figures from the paper (3 more)
Figure 4
Figure 4. Figure 4: Bounds (lines) and inverse designs (shapes) for substrated metasurfaces [PITH_FULL_IMAGE:figures/full_fig_p009_4.png]
Figure 1
Figure 1. Figure 1: Schematic of the ellipsoidal surface projection: the ellipsoidal surface [PITH_FULL_IMAGE:figures/full_fig_p015_1.png]
Figure 2
Figure 2. Figure 2: SERS bounds given (params here) as L → λ from below, for both substrated (a) and suspended (b) design regions. (a) has linear scales for the power enhancement FOM and log scales for the separation λ − L, demonstrating a logarithmic divergence FOM ∼ log(λ − L) as L → λ …

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. Inferring Structure via Duality for Photonic Inverse Design

    math.OC 2025-04 conditional novelty 7.0 of 10

    The paper derives four minimax-based lemmas about scattering QCQPs and introduces a scrape, contract, and expand protocol that transforms a design problem into a strongly dual form whose solution seeds a near-optimal ...

Reference graph

Works this paper leans on

79 extracted references · 77 canonical work pages · cited by 1 Pith paper

  1. [1]

    Rigorous justification of the |E|4 enhancement factor in Surface Enhanced Raman Spectroscopy,

    E. C. Le Ru and P. G. Etchegoin, “Rigorous justification of the |E|4 enhancement factor in Surface Enhanced Raman Spectroscopy,” Chem. Phys. Lett.423, 63–66 (2006)

    work page 2006

  2. [2]

    Designing structures that maximize spatially averaged surface-enhanced Raman spectra,

    W. Yao, F. Verdugo, H. O. Everitt,et al., “Designing structures that maximize spatially averaged surface-enhanced Raman spectra,” Opt. Express31, 4964–4977 (2023)

    work page 2023

  3. [3]

    Structuredsemidefiniteprogramsandsemialgebraicgeometrymethodsinrobustnessandoptimization,

    P.A.Parrilo,“Structuredsemidefiniteprogramsandsemialgebraicgeometrymethodsinrobustnessandoptimization,” Ph.D. thesis, California Institute of Technology (2000)

    work page 2000

  4. [4]

    Global Optimization with Polynomials and the Problem of Moments,

    J. B. Lasserre, “Global Optimization with Polynomials and the Problem of Moments,” SIAM J. on Optim.11, 796–817 (2001)

    work page 2001

  5. [5]

    Blekherman, P

    G. Blekherman, P. A. Parrilo, and R. R. Thomas, eds.,Semidefinite Optimization and Convex Algebraic Geometry, MOS-SIAM Series on Optimization (Society for Industrial and Applied Mathematics : Mathematical Programming Society, Philadelphia, 2013)

    work page 2013

  6. [6]

    Physical limits in electromagnetism,

    P. Chao, B. Strekha, R. Kuate Defo,et al., “Physical limits in electromagnetism,” Nat. Rev. Phys.4, 543–559 (2022)

    work page 2022

  7. [7]

    Heuristic methods and performance bounds for photonic design,

    G. Angeris, J. Vučković, and S. Boyd, “Heuristic methods and performance bounds for photonic design,” Opt. Express 29, 2827–2854 (2021)

    work page 2021

  8. [8]

    Many photonic design problems are sparse QCQPs,

    S. Gertler, Z. Kuang, C. Christie,et al., “Many photonic design problems are sparse QCQPs,” Sci. Adv. (2025)

    work page 2025

Show all 79 references
  1. [9]

    Suppressing electromagnetic local density of states via slow light in lossy quasi-one-dimensional gratings,

    B. Strekha, P. Chao, R. K. Defo,et al., “Suppressing electromagnetic local density of states via slow light in lossy quasi-one-dimensional gratings,” Phys. Rev. A109, L041501 (2024)

    2024

  2. [10]

    S. A. Maier,Plasmonics: Fundamentals and Applications(Springer US, New York, NY, 2007)

    2007

  3. [11]

    L. B. Felsen and N. Marcuvitz,Radiation and Scattering of Waves, IEEE Press Series on Electromagnetic Waves (Inst. of Electrical and Electronics Engineers, New York, 2003)

    2003

  4. [12]

    Confluent tip singularity of the electromagnetic field at the apex of a material cone,

    M. Idemen, “Confluent tip singularity of the electromagnetic field at the apex of a material cone,” Wave Motion38, 251–277 (2003)

    2003

  5. [13]

    On the electromagnetic field singularities near the vertex of a dielectric wedge,

    B. V. Budaev and D. B. Bogy, “On the electromagnetic field singularities near the vertex of a dielectric wedge,” Rádió Sci.42 (2007)

    2007

  6. [14]

    Spectral response of plasmon resonant nanoparticles with a non-regular shape,

    J. P. Kottmann, O. J. F. Martin, D. R. Smith, and S. Schultz, “Spectral response of plasmon resonant nanoparticles with a non-regular shape,” Opt. Express6, 213–219 (2000)

    2000

  7. [15]

    Surface-enhanced Raman spectroscopy,

    R. L. Garrell, “Surface-enhanced Raman spectroscopy,” Anal. Chem.61, 401A–411A (1989)

    1989

  8. [16]

    Kneipp, M

    K. Kneipp, M. Moskovits, and H. Kneipp, eds.,Surface-Enhanced Raman Scattering, vol. 103 ofTopics in Applied Physics(Springer Berlin Heidelberg, 2006)

    2006

  9. [17]

    Raman Techniques: Fundamentals and Frontiers,

    R. R. Jones, D. C. Hooper, L. Zhang,et al., “Raman Techniques: Fundamentals and Frontiers,” Nanoscale Res. Lett. 14, 231 (2019)

    2019

  10. [18]

    Surface-enhanced spectroscopy,

    M. Moskovits, “Surface-enhanced spectroscopy,” Rev. Mod. Phys.57, 783–826 (1985)

    1985

  11. [19]

    Present and Future of Surface-Enhanced Raman Scattering,

    J. Langer, D. Jimenez de Aberasturi, J. Aizpurua,et al., “Present and Future of Surface-Enhanced Raman Scattering,” ACS Nano14, 28–117 (2020)

    2020

  12. [20]

    Inverse design in nanophotonics,

    S. Molesky, Z. Lin, A. Y. Piggott,et al., “Inverse design in nanophotonics,” Nat. Photonics12, 659–670 (2018)

    2018

  13. [21]

    Inverse design of nanoparticles for enhanced Raman scattering,

    R. E. Christiansen, J. Michon, M. Benzaouia,et al., “Inverse design of nanoparticles for enhanced Raman scattering,” Opt. Express28, 4444–4462 (2020)

    2020

  14. [22]

    Topologyoptimizationofsurface-enhancedRamanscatteringsubstrates,

    Y.Pan,R.E.Christiansen,J.Michon, etal.,“Topologyoptimizationofsurface-enhancedRamanscatteringsubstrates,” Appl. Phys. Lett.119, 061601 (2021)

    2021

  15. [23]

    Field behavior near a dielectric wedge,

    J. Andersen and V. Solodukhov, “Field behavior near a dielectric wedge,” IEEE Trans. on Antennas Propag.26, 598–602 (1978)

    1978

  16. [24]

    Nonlocal Optical Response of Metal Nanostructures with Arbitrary Shape,

    J. M. McMahon, S. K. Gray, and G. C. Schatz, “Nonlocal Optical Response of Metal Nanostructures with Arbitrary Shape,” Phys. Rev. Lett.103, 097403 (2009)

    2009

  17. [25]

    Probing the Ultimate Limits of Plasmonic Enhancement,

    C. Ciracì, R. T. Hill, J. J. Mock,et al., “Probing the Ultimate Limits of Plasmonic Enhancement,” Science337, 1072–1074 (2012)

    2012

  18. [26]

    Optical Nonlocality in Polar Dielectrics,

    C. R. Gubbin and S. De Liberato, “Optical Nonlocality in Polar Dielectrics,” Phys. Rev. X10, 021027 (2020)

    2020

  19. [27]

    Minimum length scale in topology optimization by geometric constraints,

    M. Zhou, B. S. Lazarov, F. Wang, and O. Sigmund, “Minimum length scale in topology optimization by geometric constraints,” Comput. Methods Appl. Mech. Eng.293, 266–282 (2015)

    2015

  20. [28]

    Analytical level set fabrication constraints for inverse design,

    D. Vercruysse, N. V. Sapra, L. Su,et al., “Analytical level set fabrication constraints for inverse design,” Sci. Reports 9, 8999 (2019)

    2019

  21. [29]

    Photonic topology optimization with semiconductor- foundry design-rule constraints,

    A. M. Hammond, A. Oskooi, S. G. Johnson, and S. E. Ralph, “Photonic topology optimization with semiconductor- foundry design-rule constraints,” Opt. Express29, 23916 (2021)

    2021

  22. [30]

    Inverse Design of Photonic Devices with Strict Foundry Fabrication Constraints,

    M. F. Schubert, A. K. C. Cheung, I. A. D. Williamson,et al., “Inverse Design of Photonic Devices with Strict Foundry Fabrication Constraints,” ACS Photonics9, 2327–2336 (2022)

    2022

  23. [31]

    Hackbusch,Hierarchical Matrices: Algorithms and Analysis, vol

    W. Hackbusch,Hierarchical Matrices: Algorithms and Analysis, vol. 49 ofSpringer Series in Computational Mathematics(Springer, Berlin, Heidelberg, 2015)

    2015

  24. [32]

    Fluctuating volume-current formulation of electromagnetic fluctuations in inhomogeneous media: Incandescence and luminescence in arbitrary geometries,

    A. G. Polimeridis, M. T. H. Reid, W. Jin,et al., “Fluctuating volume-current formulation of electromagnetic fluctuations in inhomogeneous media: Incandescence and luminescence in arbitrary geometries,” Phys. Rev. B92, 134202 (2015)

    2015

  25. [33]

    Maximum electromagnetic local density of states via material structuring,

    P. Chao, R. K. Defo, S. Molesky, and A. Rodriguez, “Maximum electromagnetic local density of states via material structuring,” Nanophotonics (2022)

    2022

  26. [34]

    Upper bounds on absorption and scattering,

    M. Gustafsson, K. Schab, L. Jelinek, and M. Capek, “Upper bounds on absorption and scattering,” New J. Phys.22, 073013 (2020)

    2020

  27. [35]

    GlobalT-operator bounds on electromagnetic scattering: Upper bounds on far-field cross sections,

    S. Molesky, P. Chao, W. Jin, and A. W. Rodriguez, “GlobalT-operator bounds on electromagnetic scattering: Upper bounds on far-field cross sections,” Phys. Rev. Res.2, 033172 (2020)

    2020

  28. [36]

    Computational bounds to light–matter interactions via local conservation laws,

    Z. Kuang and O. D. Miller, “Computational bounds to light–matter interactions via local conservation laws,” Phys. Rev. Lett.125, 263607 (2020)

    2020

  29. [37]

    Fundamental Limits to Near-Field Optical Response over Any Bandwidth,

    H. Shim, L. Fan, S. G. Johnson, and O. D. Miller, “Fundamental Limits to Near-Field Optical Response over Any Bandwidth,” Phys. Rev. X9, 011043 (2019)

    2019

  30. [38]

    T-operator limits on optical communication: Metaoptics, computation, and input-output transformations,

    S. Molesky, P. Chao, J. Mohajan,et al., “T-operator limits on optical communication: Metaoptics, computation, and input-output transformations,” Phys. Rev. Res.4, 013020 (2022)

    2022

  31. [39]

    Fundamental limits to multi-functional and tunable nanophotonic response,

    H. Shim, Z. Kuang, Z. Lin, and O. D. Miller, “Fundamental limits to multi-functional and tunable nanophotonic response,” (2021)

    2021

  32. [40]

    Fundamental limits to attractive and repulsive Casimir-Polder forces,

    P. S. Venkataram, S. Molesky, P. Chao, and A. W. Rodriguez, “Fundamental limits to attractive and repulsive Casimir-Polder forces,” Phys. Rev. A101, 052115 (2020)

    2020

  33. [41]

    Trace expressions and associated limits for equilibrium Casimir torque,

    B. Strekha, M. Krüger, and A. W. Rodriguez, “Trace expressions and associated limits for equilibrium Casimir torque,” Phys. Rev. A109, 012813 (2024)

    2024

  34. [42]

    Fundamental limits on radiative𝜒(2) second harmonic generation,

    J. Mohajan, P. Chao, W. Jin,et al., “Fundamental limits on radiative𝜒(2) second harmonic generation,” Opt. Express 31, 44212–44223 (2023)

    2023

  35. [43]

    Limits to surface-enhanced Raman scattering near arbitrary-shape scatterers,

    J. Michon, M. Benzaouia, W. Yao,et al., “Limits to surface-enhanced Raman scattering near arbitrary-shape scatterers,” Opt. Express27, 35189–35202 (2019)

    2019

  36. [44]

    Physical limits on Raman scattering: The critical role of pump and signal co-design,

    A. Amaolo, P. Chao, T. J. Maldonado,et al., “Physical limits on Raman scattering: The critical role of pump and signal co-design,” Phys. Rev. A110, L061501 (2024)

    2024

  37. [45]

    Cavity-enhancedsecond-harmonicgenerationvianonlinear-overlapoptimization,

    Z.Lin,X.Liang,M.Lončar, etal.,“Cavity-enhancedsecond-harmonicgenerationvianonlinear-overlapoptimization,” Optica 3, 233–238 (2016)

    2016

  38. [46]

    Inverse-designed photonic fibers and metasurfaces for nonlinear frequency conversion [Invited],

    C. Sitawarin, W. Jin, Z. Lin, and A. W. Rodriguez, “Inverse-designed photonic fibers and metasurfaces for nonlinear frequency conversion [Invited],” Photonics Res.6, B82–B89 (2018)

    2018

  39. [47]

    Adjoint Method and Inverse Design for Nonlinear Nanophotonic Devices,

    T. W. Hughes, M. Minkov, I. A. D. Williamson, and S. Fan, “Adjoint Method and Inverse Design for Nonlinear Nanophotonic Devices,” ACS Photonics5, 4781–4787 (2018)

    2018

  40. [48]

    Inverse-designed silicon carbide quantum and nonlinear photonics,

    J. Yang, M. A. Guidry, D. M. Lukin,et al., “Inverse-designed silicon carbide quantum and nonlinear photonics,” Light. Sci. & Appl.12, 201 (2023)

    2023

  41. [49]

    Inverse Design of Nonlinear Polaritonic Metasurfaces for Second Harmonic Generation,

    S. A. Mann, H. Goh, and A. Alù, “Inverse Design of Nonlinear Polaritonic Metasurfaces for Second Harmonic Generation,” ACS Photonics10, 993–1000 (2023)

    2023

  42. [50]

    Inverse Design of an All-Dielectric Nonlinear Polaritonic Metasurface,

    S. Stich, J. Mohajan, D. de Ceglia,et al., “Inverse Design of an All-Dielectric Nonlinear Polaritonic Metasurface,” (2024)

    2024

  43. [51]

    Designing structures that maximize spatially averaged surface-enhanced Raman spectra: Erratum,

    I. M. Hammond, P. Chao, W. Yao,et al., “Designing structures that maximize spatially averaged surface-enhanced Raman spectra: Erratum,” Opt. Express32, 44754–44755 (2024)

    2024

  44. [52]

    General Heuristics for Nonconvex Quadratically Constrained Quadratic Programming,

    J. Park and S. Boyd, “General Heuristics for Nonconvex Quadratically Constrained Quadratic Programming,” (2017)

    2017

  45. [53]

    S. P. Boyd and L. Vandenberghe,Convex Optimization(Cambridge University Press, Cambridge, UK ; New York, 2004)

    2004

  46. [54]

    NP-hardness of deciding convexity of quartic polynomials and related problems,

    A. A. Ahmadi, A. Olshevsky, P. A. Parrilo, and J. N. Tsitsiklis, “NP-hardness of deciding convexity of quartic polynomials and related problems,” Math. Program.137, 453–476 (2013)

    2013

  47. [55]

    Minimizing Polynomial Functions,

    P. A. Parrilo and B. Sturmfels, “Minimizing Polynomial Functions,” (2001)

    2001

  48. [56]

    DSOS and SDSOS Optimization: More Tractable Alternatives to Sum of Squares and Semidefinite Optimization,

    A. A. Ahmadi and A. Majumdar, “DSOS and SDSOS Optimization: More Tractable Alternatives to Sum of Squares and Semidefinite Optimization,” SIAM J. on Appl. Algebr. Geom.3, 193–230 (2019)

    2019

  49. [57]

    ApS,MOSEK optimizer API for python manual

    M. ApS,MOSEK optimizer API for python manual. Release 10.2.13(2025)

    2025

  50. [58]

    Pre- and post-processing sum-of-squares programs in practice,

    J. Löfberg, “Pre- and post-processing sum-of-squares programs in practice,” IEEE Trans. on Autom. Control.54, 1007–1011 (2009)

    2009

  51. [59]

    SumOfSquares.py,

    C. Yuan, “SumOfSquares.py,”

  52. [60]

    SciPy 1.0: Fundamental Algorithms for Scientific Computing in Python,

    P. Virtanen, R. Gommers, T. E. Oliphant,et al., “SciPy 1.0: Fundamental Algorithms for Scientific Computing in Python,” Nat. Methods17, 261–272 (2020)

    2020

  53. [61]

    R. P. Brent,Algorithms for Minimization without Derivatives(Dover Publications, Mineola, N.Y, 2002)

    2002

  54. [62]

    J. D. Joannopoulos, J. G. Steven, J. N. Winn, and R. D. Meade,Photonic Crystals: Molding the Flow of Light (Princeton University Press, Princeton, 2008), 2nd ed

    2008

  55. [63]

    A non-linear material interpolation for design of metallic nano-particles using topology optimization,

    R. E. Christiansen, J. Vester-Petersen, S. P. Madsen, and O. Sigmund, “A non-linear material interpolation for design of metallic nano-particles using topology optimization,” Comput. Methods Appl. Mech. Eng.343, 23–39 (2019)

    2019

  56. [64]

    The structural basis for giant enhancement enabling single-molecule Raman scattering,

    Z. Wang, S. Pan, T. D. Krauss,et al., “The structural basis for giant enhancement enabling single-molecule Raman scattering,” Proc. National Acad. Sci.100, 8638–8643 (2003)

    2003

  57. [65]

    Self-Similar Chain of Metal Nanospheres as an Efficient Nanolens,

    K. Li, M. I. Stockman, and D. J. Bergman, “Self-Similar Chain of Metal Nanospheres as an Efficient Nanolens,” Phys. Rev. Lett.91, 227402 (2003)

    2003

  58. [66]

    Formulation for scalable optimization of microcavities via the frequency-averaged local density of states,

    X. Liang and S. G. Johnson, “Formulation for scalable optimization of microcavities via the frequency-averaged local density of states,” Opt. Express21, 30812–30841 (2013)

    2013

  59. [67]

    Hierarchicalmean-fieldT-operatorboundsonelectromagneticscattering: Upper bounds on near-field radiative Purcell enhancement,

    S.Molesky,P.Chao,andA.W.Rodriguez,“Hierarchicalmean-fieldT-operatorboundsonelectromagneticscattering: Upper bounds on near-field radiative Purcell enhancement,” Phys. Rev. Res.2, 043398 (2020)

    2020

  60. [68]

    Bounds on Efficiency Metrics in Photonics,

    G. Angeris, T. Diamandis, J. Vučković, and S. P. Boyd, “Bounds on Efficiency Metrics in Photonics,” ACS Photonics 10, 2521–2529 (2023)

    2023

  61. [69]

    Optical bistability on a silicon chip,

    V. R. Almeida and M. Lipson, “Optical bistability on a silicon chip,” Opt. Lett.29, 2387–2389 (2004)

    2004

  62. [70]

    Optical bistability and multistability in one-dimensional periodic metal-dielectric photonic crystal,

    F. Y. Wang, G. X. Li, H. L. Tam,et al., “Optical bistability and multistability in one-dimensional periodic metal-dielectric photonic crystal,” Appl. Phys. Lett.92, 211109 (2008)

    2008

  63. [71]

    Low threshold optical bistability at terahertz frequencies with graphene surface plasmons,

    X. Dai, L. Jiang, and Y. Xiang, “Low threshold optical bistability at terahertz frequencies with graphene surface plasmons,” Sci. Reports5, 12271 (2015)

    2015

  64. [72]

    Plasmonic Metasurfaces for Nonlinear Optics and Quantitative SERS,

    S. Gwo, C.-Y. Wang, H.-Y. Chen,et al., “Plasmonic Metasurfaces for Nonlinear Optics and Quantitative SERS,” ACS Photonics 3, 1371–1384 (2016)

    2016

  65. [73]

    Nonlinear metasurfaces: A paradigm shift in nonlinear optics,

    A. Krasnok, M. Tymchenko, and A. Alù, “Nonlinear metasurfaces: A paradigm shift in nonlinear optics,” Mater. Today 21, 8–21 (2018)

    2018

  66. [74]

    Planar nonlinear metasurface optics and their applications,

    T. Huang, X. Zhao, S. Zeng,et al., “Planar nonlinear metasurface optics and their applications,” Reports on Prog. Phys. 83, 126101 (2020)

    2020

  67. [75]

    Exceptional Precision of a Nonlinear Optical Sensor at a Square-Root Singularity,

    K. J. H. Peters and S. R. K. Rodriguez, “Exceptional Precision of a Nonlinear Optical Sensor at a Square-Root Singularity,” Phys. Rev. Lett.129, 013901 (2022)

    2022

  68. [76]

    Bayesian optimization of Fisher Information in nonlinear multiresonant quantum photonics gyroscopes,

    M. Sun, V. Kovanis, M. Lončar, and Z. Lin, “Bayesian optimization of Fisher Information in nonlinear multiresonant quantum photonics gyroscopes,” Nanophotonics13, 2401–2416 (2024)

    2024

  69. [77]

    Ultralow-threshold electrically pumped quantum-dot photonic-crystal nanocavity laser,

    B. Ellis, M. A. Mayer, G. Shambat,et al., “Ultralow-threshold electrically pumped quantum-dot photonic-crystal nanocavity laser,” Nat. Photonics5, 297–300 (2011)

    2011

  70. [78]

    Monolayer semiconductor nanocavity lasers with ultralow thresholds,

    S. Wu, S. Buckley, J. R. Schaibley,et al., “Monolayer semiconductor nanocavity lasers with ultralow thresholds,” Nature 520, 69–72 (2015)

    2015

  71. [79]

    Inverse Design of Whispering-Gallery Nanolasers with Tailored Beam Shape and Polarization,

    I. Diez, A. Krysa, and I. J. Luxmoore, “Inverse Design of Whispering-Gallery Nanolasers with Tailored Beam Shape and Polarization,” ACS Photonics10, 968–976 (2023). Sum-of-Squares Bounds on Surface-Enhanced Raman Scattering: Supplementary Info Pengning Chao, Ian M. Hammond, an...

    2023 arXiv

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.