Model and range

Use atomic units and two separated systems A,BA,B in an ordinary spatial three-plane, with nondegenerate ground states, positive excitation gaps Δa,Δb\Delta_a,\Delta_b, and zero permanent dipoles. Assume separation RR large compared with their sizes, negligible exchange, and propagation delay short compared with the relevant response times. The leading instantaneous interaction is

V=R−3dAiTijdBj,Tij=δij−3R^iR^j. V=R^{-3}d_A^i T_{ij}d_B^j,\qquad T_{ij}=\delta_{ij}-3\hat R_i\hat R_j.

This is a conditional ground-state response calculation. A classical field application must supply its fluctuation state and normalization; zero-point excitation weight is not implicit in the geometry.

Spectral sum

First order vanishes. The second-order ground energy is

ΔE(2)=−C6R6,C6=∑a,b∣⟨0A∣dAi∣a⟩Tij⟨0B∣dBj∣b⟩∣2Δa+Δb≥0. \Delta E^{(2)}=-\frac{C_6}{R^6},\qquad C_6=\sum_{a,b}\frac{ |\langle0_A|d_A^i|a\rangle T_{ij} \langle0_B|d_B^j|b\rangle|^2}{\Delta_a+\Delta_b}\ge0.

All denominators are positive, fixing the attractive sign. Sums include complete excited multiplets and, where necessary, continuum integrals; convergence is an assumption on the response. For isotropic systems each spectral tensor is proportional to δij\delta_{ij} and Tr⁡T2=6\operatorname{Tr}T^2=6. Writing Da=∑i∣⟨0∣di∣a⟩∣2D_a=\sum_i|\langle0|d^i|a\rangle|^2 with degenerate states grouped gives

C6=23∑a,bDaDbΔa+Δb. C_6=\frac23\sum_{a,b}\frac{D_aD_b}{\Delta_a+\Delta_b}.

The sign and power have a direct second-order oscillator derivation in the van der Waals teaching article.

Imaginary-frequency form

Define isotropic polarizability at imaginary frequency iuiu by

α(iu)=23∑aΔaDaΔa2+u2. \alpha(iu)=\frac23\sum_a\frac{\Delta_aD_a}{\Delta_a^2+u^2}.

Partial fractions, including the continuous equal-gap limit, give

∫0∞ab du(a2+u2)(b2+u2)=π2(a+b),a,b>0. \int_0^\infty\frac{ab\,du}{(a^2+u^2)(b^2+u^2)} =\frac{\pi}{2(a+b)},\qquad a,b>0.

Substitution into the spectral sum therefore proves

C6=3π∫0∞αA(iu)αB(iu) du. \boxed{C_6=\frac3\pi\int_0^\infty\alpha_A(iu)\alpha_B(iu)\,du.}

For a single-resonance response αX(iu)=αX(0)ΩX2/(ΩX2+u2)\alpha_X(iu)=\alpha_X(0)\Omega_X^2/(\Omega_X^2+u^2),

C6=32αA(0)αB(0)ΩAΩBΩA+ΩB. C_6=\frac32\alpha_A(0)\alpha_B(0) \frac{\Omega_A\Omega_B}{\Omega_A+\Omega_B}.

For equal resonances this is 3αA(0)αB(0)Ω/43\alpha_A(0)\alpha_B(0)\Omega/4. Replacing the resonance by an ionization scale is an additional approximation. These equations give a target for a V2 response calculation once its actual modes, dipole vertex, and state have been determined.

resonance_profile_checks.py checks finite examples by quadrature, matrix evolution, and independent algebra.