skip to content

Teaching Light
Where to Go

The mathematics underlying 13 publications at the intersection of materials chemistry and nanophotonics. 171 machine-checked theorems. 2 open goals remaining.


171
theorems
2
sorry
13
papers
30
files

On the mathematics underlying perovskite nanocrystal photonics — orientation, cooperative emission, and spectrum splitting.

scroll

I put your papers into a theorem prover and made a computer check every step of the math — from how a polyhedral reflector sorts sunlight across seven subcells, to how the color, polarization, and direction of light can be tuned by the structure of a nanocrystal, to how a thousand emitters synchronize their dipoles into a single coherent pulse. Along the way, the formalization revealed a few things.

Orientation

What happens when you flatten a nanocrystal into a platelet?

Every paper on dipole orientation compares to Θ = 2/3. The standard derivation integrates sin²θ over the sphere. But it follows from symmetry alone:

Proof by symmetry
On the unit sphere, ⟨x²⟩ = ⟨y²⟩ = ⟨z²⟩ by rotation invariance.
Their sum is 1. Therefore each is 1/3.
Θ = ⟨x² + y²⟩ = 2/3. ∎

Three lines. No integration. Generalizes to ddimensions: Θ = (d − 1)/d.

The doubly beneficial parameter

Nanoplatelets outperform nanocubes. Two independent mechanisms, both increasing with aspect ratio:

Intrinsic — D4h symmetry breaking: Θh= 1 − 1/(3·AR²)
Alignment — van der Waals face-down fraction: f = 1/(1 + exp(−ΔE·AR/kT))
Insight

AR is the onlyparameter that improves both intrinsic orientation and alignment simultaneously. There is no tradeoff — aspect ratio is a free lunch.

Eisler/Conjectures/OrientationOptimization.lean · 26 theorems

Quantum Optics

One emitter radiates. Two can interfere. A thousand can synchronize.

Dicke superradiance gives a peak emission rate that scales as N², not N. The standard claim assumes the symmetric Dicke state. The formalization proves something stronger:

Upper bound

Γ(J, M) ≤ γ0· (N/2)(N/2 + 1)  on the physical Dicke ladder

It's a proved upper bound, not an assumption. And the buildup time τD = ln(N)/(N·γ0) is strictly decreasing for N ≥ 3. Bigger domains give both more intensity and faster onset. Named and proved:

theorem
theorem delay_decreasing_per_emitter : ∀ (N₁ N₂ : ℕ) (γ₀ : ℝ), 3 ≤ N₁ → 0 < γ₀ → N₁ < N₂ → superradiantDelay N₂ γ₀ < superradiantDelay N₁ γ₀
#print axioms delay_decreasing_per_emitter
› propext, Quot.sound, Classical.choice
Eisler/QuantumOptics/DickeSuperradiance.lean:82

Key step uses Mathlib's MonotonOn.strictMono for the ln(N)/N decay above e.

Eisler/QuantumOptics/DickeSuperradiance.lean · 7 theorems · view proof →

Novel Theorem

Nobody had composed these two effects.

Transition dipole orientation and Dicke superradiance had been studied independently for decades. The formalization proves: for N ≥ 3 emitters with Θ > 2/3, directed superradiant power always exceeds isotropic independent power.

Θ · γ0· N(N+2)/4  > (2/3) · N · γ0

At the measured Jurow 2019 value Θ = 0.89 — i.e. 〈sin²θ〉 = 0.89, meaning 89% of the dipole power sits in the in-plane component. The far-field forward/backward ratio additionally depends on substrate index and outcoupling geometry (Brütting 2013) and is not claimed here.

Insight

Two independently demonstrated effects — orientation (Θ) and cooperative emission (N²) — compose multiplicatively. The directed superradiant dominance theorem proves they never cancel.

Eisler/Conjectures/DirectionalSuperfluorescence.lean · 6 theorems

A useful heuristic

A Three-Factor Figure of Merit

A first-order product score that captures the direction of improvement across solar cells, LEDs, and directed quantum sources when the three factors are held approximately independent. This is not a derived thermodynamic bound — it is a useful design-space gradient:

FoM = PLQY × Θ × (1 / Δλnorm)

current score
0.846
0.95 · 0.89 · 1
heuristic max
1.000
1.0 · 1.0 · 1
gap
0.154
Θ: 0.105 • QY: 0.045

Dimensionless score, not an efficiency. Linewidth held at reference.

Proved

Under the product FoM, improving Θ from 0.89 → 1.0 yields ΔFoM = 0.95·0.11 = 0.1045; improving PLQY from 0.95 → 1.0 yields 0.05·0.89 = 0.0445. Ratio 2.35×— orientation is the bigger lever under this heuristic. Even with worst-case 5% PLQY loss from the shape change, the net FoM still increases. Orientation is the bigger lever.

Eisler/Conjectures/OptimalDesign.lean · 15 theorems

Thermodynamics

Shockley-Queisser is a limit on single-junction cells.

The SQ bound is a rigorous thermodynamic result for a single junction illuminated by a blackbody spectrum under detailed balance — it is not a bound on solar conversion in general. Relax the single-junction assumption and the bound lifts. The polyhedral specular reflector sorts photons by energy onto seven subcells; the paper reports:

Experimental (Eisler 2019, IEEE JPV): 7 subcells, 88% optical splitting efficiency, >50% total module efficiency measured.
Formalized structural monotonicity: under the record model, adjoining a zero-power band to an N-band module cannot decrease efficiency. The sharper thermodynamic multijunction bound (optimal N+1 ≥ optimal N; DeVos 1980, Martí & Araújo 1996) is standard and is not formalized here.

The blackbody flux above any bandgap is proved strictly decreasing in Eg. The proof: the Wien function f(x) = (x² + 2x + 2)·e−x loses to exponential decay because 1 + h + h²/2 ≤ eh. Foundation for Stokes shift concentration.

Every photon that enters the module must first cross a dielectric interface. The loss is set by Fresnel—drag the sliders and watch Rs, Rp, and the unpolarized average evolve. Brewster's angle is the zero of the p-curve.

Fresnel reflectance

unpolarized R = (Rs + Rp) / 2

Rs19.6%
Rp11.7%
R15.7%
θB66.5°
θc—

Eisler/Optics/SnellFresnel.lean

Eisler/Thermodynamics/ · 29 theorems

Formalization

30 Files, 7 Modules, 171 Theorems

Thermodynamics

32 thm

5 files

  • ●Shockley-Queisser limit
  • ●Detailed balance
  • ●Spectrum splitting
  • ●Luminescent concentrator
  • ●Physical constants

Optics

11 thm

4 files

  • ●Etendue conservation
  • ●Fresnel equations
  • ●Brewster’s angle
  • ●Ray tracing

Quantum Optics

36 thm

5 files

  • ●Dicke superradiance
  • ●Superfluorescence
  • ●Einstein coefficients
  • ●Transition dipoles

Materials

8 thm

3 files

  • ●Perovskite structure
  • ●Exciton physics
  • ●Surface passivation
  • ●Quantum confinement

Mechanics

6 thm

2 files

  • ●Schmid factor
  • ●Pseudoelastic deformation

Experimental

3 thm

2 files

  • ●Dipole alignment
  • ●Solar efficiency

Conjectures

75 thm

8 files

  • ●Design Triangle
  • ●Directed superradiance
  • ●Orientation optimization
  • ●Stokes shift concentration
Publications

Papers Formalized

Solar Photovoltaics
Multijunction solar cell efficiencies: Effect of spectral window, optical environment and radiative coupling
Eisler, Abrams, Sheldon, Zhang, Atwater
Energy Environ. Sci. 7, 3600 (2014)
Design of photovoltaics for modules with 50% efficiency
Warmann, Flowers, Lloyd, Eisler, Escarra, Atwater
Energy Sci. Eng. 5, 69 (2017)
The polyhedral specular reflector: ultrahigh (>50%) solar module efficiencies
Eisler et al.
IEEE J. Photovolt. 9(1), 174–183 (2018)
Photonic luminescent solar concentrator design
Eisler, Parsons, Nett, Love, Schwartzberg, Alivisatos
Frontiers in Photonics 3, 932913 (2022)
Transition Dipole Orientation
Manipulating the transition dipole moment of CsPbBr₃ perovskite nanocrystals
Jurow, Morgenstern, Eisler et al.
Nano Lett. 19, 2489 (2019)
Tunable angular light emission via solution-processed substrate treatment
Parsons, Russ, Eisler
ACS Nanoscience Au (2025)
Oriented dipoles in ordered ensembles of confined perovskite nanocrystals
Parsons, Grishchenko, Eisler
J. Phys. Chem. C 130, 1996 (2026)
Quantifying TDM alignment from realistic angular emission data
Russ, Lin, Elenteny, Eisler
ACS Photonics (2025)
Quantum Cooperative Emission
Superfluorescence from solution-processed, tunable materials
Russ, Eisler
Nanophotonics 13, 1927 (2024)
Materials & Nanocrystal Physics
Suppression of surface recombination in CuInSe₂ via TOP:S passivation
Luo, Eisler et al.
Acta Mater. 106, 171 (2016)
Pseudoelasticity at large strains in Au nanocrystals
Gu, Hanson, Eisler, Koc, Alivisatos
Phys. Rev. Lett. 121, 056102 (2018)
Anisotropic 2D excitons unveiled in organic-inorganic quantum wells
Maserati et al.
Mater. Horiz. 8, 197–208 (2021)
Perovskite nanowire composites with programmable polarization anisotropy
Zhou, Bekenstein, Eisler et al.
Science Adv. 5, eaav8141 (2019)
Verification

Check It Yourself

# Clone and verify
git clone https://github.com/awkronos/eisler
cd eisler
lake exe cache get
lake build Eisler

# Inspect any theorem
#print axioms delay_decreasing_per_emitter

Lean 4

Kernel-checked. propext, Quot.sound, Classical.choice. Nothing else.

Mathlib v4.30

200K+ theorems of community-maintained formalized mathematics.

Zero Trust

2 sorry. 0 axioms. 3 physical hypotheses, documented.

The code is the proof.
The proof is the paper.

For Carissa — who taught light where to go.

— Tim

Verified by Awkronos, Inc. · Formally verified mathematical infrastructure

SourceScholarLabVerify