Discretization-Controlled Reliability of Gold Nanogap Dimer Plasmon Modes in DDSCAT

Introduction: Gold nanosphere dimers are used as model plasmonic systems because their localized surface plasmon resonance is sensitive to interparticle coupling. Previous discrete dipole approximation studies established that calculated optical responses depend on geometric discretization and converge as dipole spacing is reduced. However, this effect is often treated only as a quantitative error. Here, we show that insufficient resolution can reverse the apparent resonance ordering of neighbouring nanogaps and lead to an incorrect interpretation of coupling strength.

Methods: DDSCAT 7.3.4 was used to calculate extinction spectra of 50 nm-radius gold nanosphere dimers in water for interparticle gaps from 20 to 2 nm. A single gold sphere was benchmarked against Mie theory. Longitudinal and transverse excitations were then examined using SHPAR values of 36, 48, and 64, with SHPAR = 80 calculations for the smallest gaps.

Results: The single-sphere extinction maximum occurred at 576 nm in DDSCAT and 570 nm in Mie theory on the same wavelength grid. Under longitudinal excitation, the dimer resonance shifted from 640 nm at a 20 nm gap to 720 nm at 5 nm. In the 2–3 nm regime, SHPAR = 36 and 48 incorrectly predicted the 3 nm gap to be more red-shifted than the 2 nm gap. At SHPAR = 64, monotonic ordering was recovered, with peaks at 724, 756, and 810 nm for 5, 3, and 2 nm gaps, respectively. SHPAR = 80 yielded 755 and 815 nm for the 3 and 2 nm gaps, confirming the ordering.

Conclusions: The key advance is the demonstration that an under-resolved nanogap can generate a false physical trend rather than merely a small wavelength deviation. A smooth and plausible spectrum is therefore not sufficient evidence of numerical reliability. Reliable DDSCAT nanogap studies should report the physical gap, dipole spacing, gap-to-spacing ratio, and explicit peak-convergence evidence.