What happens when the same light-responsive molecule is placed in different solvents? In these anthracene-based chalcones, the solvent does more than shift an absorption peak—it can reverse which molecule has the stronger computed nonlinear optical response.

Anthracene-based chalcones are pi-conjugated donor–acceptor chromophores with potential second-order nonlinear optical applications, but solvent-field polarization and solvent-induced geometry relaxation can affect their response in different ways. The comparison focuses on ANT–IND and ANT–DMNAF, which differ in donor environment, to identify how the medium controls absorption and static first hyperpolarizability.

Solvation red-shifts the principal ultraviolet band by about fourteen nanometers and substantially increases the computed nonlinear optical response. In methanol, optimized ANT–DMNAF reaches beta parallel equal to ten thousand two hundred sixty-nine atomic units, exceeding ANT–IND under the same conditions and indicating a solvent-dependent crossover in the computed nonlinear optical ranking.

Large second-order nonlinear optical responses attract interest because of potential applications in photonic devices, electro-optic modulators, frequency doublers, and optical data storage systems. Donor–pi–acceptor chromophores link an electron-donating group and an electron-withdrawing group through a conjugated bridge, making them a widely explored architecture for nonlinear optical activity.

In these systems, nonlinear optical efficiency is governed by the extent of intramolecular charge transfer. A critical aspect of molecular nonlinear optical behavior is the influence of the solvent environment. In solution, that surrounding medium becomes the setting in which the molecular response is evaluated.

The medium can modify ground-state charge distribution, molecular geometry, excitation energies, and electric-response properties of push–pull chromophores. Previous theoretical and experimental studies show that solvent-induced polarization and structural relaxation can substantially alter static and frequency-dependent nonlinear optical responses, including first hyperpolarizabilities.

Continuum calculations connect solvent-driven changes in charge separation with bond-length alternation and large variations in nonlinear optical response, while explicit and hybrid treatments can add local solute–solvent effects beyond a bulk dielectric description. That motivates comparing fixed-geometry and solvent-relaxed calculations, while avoiding claims about specific first-shell interactions that the continuum model does not explicitly represent.

The study investigates two structurally related anthracene-based chalcone derivatives: ANT–IND, bearing an indole donor, and ANT–DMNAF, bearing a 4-(dimethylamino)naphthyl donor. Because the two molecules share the same anthracene–enone framework but have distinct donor environments, they provide a direct comparison of donor effects.

Using CAM-B3LYP with the six-media SMD setup, the calculations characterize geometry, frontier orbitals, ultraviolet–visible absorption and excited-state character, static first hyperpolarizabilities, and polarizabilities. GasGeom and SolvGeom are used to assess sensitivity to solvent-field polarization and geometry relaxation, not as experimentally separable contributions.

ANT–IND is a twenty-seven-heavy-atom system with an indolyl electron-donor moiety, while ANT–DMNAF is a thirty-one-heavy-atom system containing a 4-(dimethylamino)naphthyl donor fragment. In both molecules, the alpha,beta-unsaturated carbonyl bridge provides the main pi-conjugated linkage, and the anthracenyl fragment extends the conjugated framework.

Initial geometries were preliminarily optimized with B3LYP, the def2-SVP basis set, and the Grimme D3 correction with Becke–Johnson damping. Those D3BJ structures served only as starting geometries; the reported electronic, excited-state, and nonlinear optical properties came from subsequent fully unconstrained CAM-B3LYP optimizations with gas-phase or SMD conditions.

No systematic conformational search or torsional scan was performed, so the analysis is restricted to the selected E-configured conformer used consistently throughout the study. The blue boxes show the principal torsional coordinates used for each optimized structure: Structure I follows phi one equals twenty-five–twenty-six–twenty-seven–twenty-eight and phi two equals twenty-seven–twenty-eight–thirty–thirty-two, while Structure II uses twenty-four–twenty-five–twenty-six–twenty-seven and twenty-seven–twenty-nine–thirty-one–thirty-two.

The dashed regions visually separate the donor, enone pi bridge, and anthracene or carbonyl acceptor portions. Full geometry optimizations and harmonic frequency calculations used CAM-B3LYP with the six–thirty-one plus G basis set, a range-separated method suited to charge-transfer character and long-range exchange effects relevant to nonlinear optical properties.

The calculations used the Tight self-consistent-field convergence criterion with the XQC option and an ultrafine integration grid. Geometry optimizations were performed in the gas phase and in each of five solvents using the SMD continuum solvation model. The optimized structures were confirmed as stationary points because all cases had no imaginary vibrational frequencies, with the lowest real frequency above eight inverse centimeters.

The solvent environment was modeled with the Solvation Model based on Density, or SMD, which combines a self-consistent electrostatic reaction field with parametrized non-electrostatic contributions. Five solvents were selected to span a broad range of dielectric constants, beginning with n-hexane at a dielectric constant of one point eighty-eight.

Vertical excitation energies, oscillator strengths, and transition compositions were computed with time-dependent density functional theory at the CAM-B3LYP level, requesting thirty excited states. The calculations covered both GasGeom and SolvGeom protocols in all six media for both structures.

CAM-B3LYP was selected because range-separated hybrids provide a more appropriate description of long-range charge-transfer excitations than many global hybrids in donor–pi–acceptor molecules. For the most intense state in each medium, dominant one-electron configurations were extracted from the Gaussian time-dependent density functional theory expansion coefficients, and the ultraviolet–visible spectra were simulated by Gaussian broadening with a half-width at half-maximum of zero point two electronvolts.

Static first hyperpolarizabilities and isotropic polarizabilities were computed analytically using the Polar keyword in Gaussian sixteen. The dipole-oriented first hyperpolarizability beta parallel is the response quantity used in the nonlinear optical analysis. Both structures show a bathochromic shift upon solvation: the principal absorption band moves from approximately two hundred thirty-one nanometers in the gas phase to roughly two hundred forty-three to two hundred forty-six nanometers in most solvents.

Relative to the gas phase, the excitation-energy decrease spans approximately zero point two three to zero point three two electronvolts for Structure I and zero point two five to zero point three five electronvolts for Structure II. Within the solvent set, wavelength variation is modest: toluene gives the longest-wavelength maximum, while methanol gives a shorter-wavelength maximum than several other solvents.

Figure three compares TD-DFT simulated UV–Vis spectra for Structure I and Structure II across the gas phase and five solvents. The dashed gas-phase curves are calculated at gas-phase optimized geometries, while solvent curves use solvent-relaxed geometries, with each spectrum normalized to its maximum intensity.

The solvent-dependent shifts and changes in the longer-wavelength bands matter because they show how the surrounding medium and geometry relaxation modify the molecules’ electronic transitions. Tables six and seven report dipole moment, isotropic polarizability, and dipole-oriented first hyperpolarizability for both structures across the gas phase and five solvents.

GasGeom and SolvGeom are retained as two computational protocols for evaluating sensitivity to geometry relaxation. SolvGeom results are used for solvent-relaxed solution-phase comparisons, whereas GasGeom values are treated as fixed-geometry diagnostic calculations rather than experimentally measurable solvent responses.

Reading down the hyperpolarizability column, the reported value rises from twenty-two point eighty-five in gas to seventy-six point seventy-seven for methanol with Solv geometry, while the corresponding change column reaches plus two hundred thirty-six percent. Within each solvent pair, the Solv row is also slightly larger than the Gas row, and every row identifies beta sub z as the dominant component.

The GasGeom calculations are treated as fixed-geometry diagnostics rather than experimentally measurable solvent responses. In the gas phase, Structure I gives beta parallel equal to two thousand six hundred forty-five atomic units, or twenty-two point eighty-five times ten to the minus thirty esu.

For solvent-relaxed geometries, the increase relative to the gas-phase reference is forty-nine percent in n-hexane, sixty-six percent in toluene, one hundred forty-three percent in DCM, one hundred sixty-five percent in DMSO, and two hundred thirty-six percent in methanol.

These SolvGeom percentages are the values used to discuss the overall solvent-dependent response, while GasGeom values remain fixed-geometry diagnostics. Structure II shows a stronger solvent response after full geometry relaxation.

Its SolvGeom values are three thousand seven hundred eighty-two atomic units in n-hexane, four thousand two hundred eighty-two in toluene, six thousand four hundred twenty-two in DCM, seven thousand nine hundred thirty-six in DMSO, and ten thousand two hundred sixty-nine in methanol, relative to a two thousand three hundred ninety-three atomic-unit gas-phase reference.

The methanol value is the largest in the dataset, and the dipole moment increases from four point nine four four debye in the gas phase to eight point one eight two debye at the methanol-optimized geometry. GasGeom values are retained as fixed-geometry diagnostic results and are not used as the total solution-phase enhancement.

Table seven reports the static nonlinear optical properties for Structure II across gas phase and five solvents, using both fixed gas geometries and solvent-relaxed geometries. It lists dipole moment, isotropic polarizability, and dipole-oriented first hyperpolarizability, with beta z identified as the dominant component in every row.

The solvent-relaxed entries provide the solution-phase comparisons, while gas-geometry results serve as fixed-geometry diagnostics. The comparison between the two structures reveals a crossover in nonlinear optical performance as the medium changes. Structure I is slightly higher in the gas phase and low-polarity media, whereas Structure II becomes larger in DCM, DMSO, and methanol after solvent relaxation.

The GasGeom-to-SolvGeom change is more pronounced for Structure II: in methanol it is two thousand seven hundred sixteen atomic units, about thirty-six percent relative to GasGeom, compared with seven hundred eighty-six atomic units, about nine point seven percent, for Structure I.

This demonstrates greater geometry sensitivity for Structure II within the present protocol, but it is not interpreted as an experimentally isolated geometric contribution. Structure II does not have the larger gas-phase beta parallel, yet it develops the larger solvent-relaxed beta parallel in DCM, DMSO, and methanol.

Therefore, gas-phase donor strength alone does not determine the solution-phase ordering. Because no exhaustive conformational scan was performed, the difference is not assigned uniquely to donor flexibility or to a specific torsional coordinate.

Instead, it is reported as the combined electronic and geometry response of the two molecular frameworks within the SMD protocol. Future work will extend the analysis to explicit methanol-cluster models to test possible hydrogen-bonding effects, frequency-dependent hyperpolarizabilities, and natural-transition-orbital or hole–electron analysis of excited-state charge transfer.

Those extensions matter because SMD provides a continuum description and was not specifically parameterized for nonlinear optical response observables. The central result is that solvent-relaxed geometry changes matter: ANT–DMNAF becomes the stronger computed nonlinear optical material in DCM, DMSO, and methanol, while the study also cautions that this SMD-based crossover is a computational result, not an isolated experimental decomposition.