How the donor closes the gap: para-substituent effects in a minimal push-pull dye

1. What this note explains

Para-fluorine is a weak net donor in push-pull chromophores. It withdraws electron density through σ bonds (negative inductive effect, −I) and donates electron density through π conjugation (positive resonance effect, +R). In most push-pull scaffolds the +R effect dominates slightly.

This note explains how that net donor effect appears in the frontier orbitals of para-X-benzylidenemalononitrile, Ar–CH=C(CN)₂, with X = H, F, NH₂, and NMe₂. It reports how donor strength moves the HOMO and the LUMO separately on the same molecular framework. In these chromophores, donor strength raises the HOMO and acceptor strength lowers the LUMO; the relative size of those two moves is measured directly.

2. The scaffold

The series is para-X-benzylidenemalononitrile, Ar–CH=C(CN)₂. Each molecule has a benzene ring with a dicyanovinyl acceptor at one para position and a variable substituent X at the other para position. The acceptor is held constant, so changes in the frontier-orbital energies are due to X.

X = H, F, NH₂, and NMe₂. The substituents are ordered by the Hammett σ_p⁺ constant: H (0.00), F (−0.07), NH₂ (−1.30), NMe₂ (−1.70).¹ The small negative value for F is the empirical signature of a weak net donor effect. The calculation reproduces that signature and shows how it is expressed in the HOMO and LUMO energies.

3. What was computed

Every molecule was optimized at B3LYP/def2-SVP, then the 12 lowest singlets were computed with full-response TD-DFT at def2-TZVP with both CAM-B3LYP and B3LYP, in Psi4 1.9.1.^2–5 The same protocol was used in the recent DCDHF-Me2 manifold note; here it is read for the Kohn-Sham orbital energies rather than for the manifold structure.

Two quantities are reported for each molecule and functional: the self-consistent HOMO and LUMO energies (from which the Kohn-Sham gap is ε(LUMO) − ε(HOMO)), and the lowest TD-DFT vertical excitation energy S₁. The HOMO–LUMO gap is a ground-state orbital spacing; S₁ is an excited-state energy. They trend together here, but they are not the same quantity.

4. The frontier-orbital picture

CAM-B3LYP / def2-TZVP

B3LYP / def2-TZVP

HOMO LUMO HOMO−1 · LUMO+1 connector value: ε(LUMO) − ε(HOMO), eV

Figure 1. Frontier-orbital energies of the four BMN molecules under CAM-B3LYP and B3LYP (def2-TZVP). The panels share a vertical scale. Solid bars: HOMO and LUMO; dashed bars: HOMO−1 and LUMO+1. The connector value is the Kohn-Sham HOMO–LUMO gap. Exact values are in Table 1.

The absolute gap is strongly functional-dependent. For BMN-H the CAM-B3LYP gap is 6.57 eV while the B3LYP gap is 4.12 eV — a spread of 2.46 eV for the same molecule and geometry. Range-separated hybrids such as CAM-B3LYP lower the donor-localized HOMO and raise the acceptor-localized LUMO relative to B3LYP, widening the gap; this is the characteristic behavior for a charge-transfer system. The trend with donor strength is nearly identical under the two functionals: both show the gap closing as X becomes more electron-donating, and both order the substituents the same way.

5. The numbers

Table 1 gives the frontier energies, the Kohn-Sham gap, and the lowest TD-DFT excitation for every molecule and functional. The HOMO and LUMO columns contain the same values that appear in Figure 1. The S₁ column is the lowest vertical excitation energy; it is always smaller than the Kohn-Sham gap because the excited-state calculation includes electron-hole interaction and relaxation.

X	σ_p⁺	functional	ε(HOMO) (eV)	ε(LUMO) (eV)	gap (eV)	S₁ (eV)	S₁ character
H	0.00	CAM-B3LYP	-8.69	-2.12	6.57	4.23	HOMO→LUMO
F	−0.07	CAM-B3LYP	-8.68	-2.17	6.51	4.19	HOMO→LUMO
NH₂	−1.30	CAM-B3LYP	-7.57	-1.62	5.95	3.78	HOMO→LUMO
NMe₂	−1.70	CAM-B3LYP	-7.32	-1.58	5.74	3.62	HOMO→LUMO
H	0.00	B3LYP	-7.38	-3.26	4.12	4.01	HOMO→LUMO
F	−0.07	B3LYP	-7.36	-3.30	4.06	3.97	HOMO→LUMO
NH₂	−1.30	B3LYP	-6.33	-2.72	3.61	3.58	HOMO→LUMO
NMe₂	−1.70	B3LYP	-6.10	-2.66	3.44	3.38	HOMO→LUMO

Table 1. Kohn-Sham frontier-orbital energies, Kohn-Sham gaps, and lowest TD-DFT vertical excitations for the BMN series. The S₁ character column is the dominant amplitude; every S₁ in the series is HOMO→LUMO, with the smallest weight 96% under CAM-B3LYP and 76% under B3LYP.

Two derived quantities summarize the trend. From H to NMe₂ the HOMO rises by 1.37 eV under CAM-B3LYP and 1.28 eV under B3LYP, while the LUMO rises by only 0.54 eV and 0.60 eV respectively. The HOMO moves roughly 2.5 times as far as the LUMO under CAM-B3LYP and 2.1 times as far under B3LYP. Most of the gap closure comes from raising the HOMO, not from lowering the LUMO.

A linear fit of the gap to σ_p⁺ is excellent under both functionals: slope 0.48 eV per unit σ_p⁺ with R² = 0.999 under CAM-B3LYP, and 0.38 eV per unit σ_p⁺ with R² = 0.998 under B3LYP. The near-perfect linearity is partly a consequence of using only four points, two of which are far apart; it is not a claim about a universal law, only a statement that at this level of theory the gap responds smoothly to the donor-strength parameter.

6. Fluorine is a weak net donor

Replacing H by F closes the Kohn-Sham gap by 0.07 eV under CAM-B3LYP and 0.06 eV under B3LYP, and the S₁ excitation red-shifts by a comparable amount. The direction is the same as for NH₂ and NMe₂; only the magnitude is much smaller. This matches the small negative σ_p⁺ value of fluorine.

The orbital decomposition shows how the competing effects resolve. Fluorine leaves the HOMO almost unchanged relative to H. The LUMO is lowered slightly. The HOMO is localized on the donor end, so fluorine’s −I withdrawal and +R donation largely cancel there. The LUMO is localized on the dicyanovinyl acceptor; the small net stabilization means inductive withdrawal dominates over resonance donation for that orbital. Both the tiny HOMO rise and the slightly larger LUMO lowering close the gap.

The 0.07 eV (CAM-B3LYP) and 0.06 eV (B3LYP) shifts are comparable to the typical TD-DFT error for charge-transfer states,⁶ so the magnitude is near the noise floor; the sign is reproduced across both functionals. The result is specific to para-X-benzylidenemalononitrile and the stated level of theory.

7. The HOMO moves more than the LUMO

From H to NMe₂ the HOMO rises by 1.37 eV under CAM-B3LYP and 1.28 eV under B3LYP, while the LUMO rises by only 0.54 eV and 0.60 eV respectively. The HOMO is localized on the donor end of the molecule and is therefore directly perturbed by the substituent. The LUMO is localized on the dicyanovinyl acceptor, which is identical across the series, so it feels the substituent only through the conjugated bridge.

The NH₂ and NMe₂ substituents are close on the donor-strength axis (σ_p⁺ = −1.30 and −1.70) but the gap still closes by approximately 0.2 eV between them. The calculation does not decompose the methyl-group effect into inductive and geometric contributions, but the direction of the shift is consistent with NMe₂ being a stronger donor than NH₂.

One caveat: the Kohn-Sham gap is not the excitation energy. For BMN-H the CAM-B3LYP Kohn-Sham gap is 6.57 eV while S₁ is 4.23 eV, a difference of 2.35 eV. The gap is the single-particle orbital spacing; S₁ is the response of the full system to an electron promotion. Both trend the same way here, but they are not the same quantity.

8. Reproducibility

All calculations use Psi4 1.9.1⁵ in the recorded conda environment (environment-psi4_19.yml), on Linux, gas phase, no solvent model. The protocol is identical to the recent DCDHF-Me2 manifold note: geometry optimization at B3LYP/def2-SVP (C1, gau_tight), then full-response TD-DFT for the 12 lowest singlets at def2-TZVP with CAM-B3LYP and B3LYP.^2–4

Starting structures were built analytically because no force-field toolkit is available in the environment; the acceptor was twisted 30° and each amine nitrogen was displaced 0.20 Å out of plane to avoid landing on a symmetry-imposed stationary point. The canonical run command is research/bmn-frontier-orbitals/run_all.sh. It is resumable: each stage skips files that already exist, so a fresh clone can rebuild every committed artifact from the starting geometries and the environment.

The frontier-orbital energies were extracted from the Psi4 output logs by orbital_gaps.py (occupied) and extract_frontier.py (virtual), with the latter cross-checking occupied energies against the former to 1 meV. The logs themselves are not committed, but their sha256 hashes are recorded in results/frontier_orbitals.json. The inline figure was rendered by make-figure.mjs and every cited number resolves from metrics.json via generate-metrics.mjs; both generators support --check to verify that the committed artifacts match the generator’s output.

9. Where the model stops

These are gas-phase vertical excitations of isolated molecules with no solvent model, and charge-transfer states are strongly solvatochromic. Nothing here has vibrational structure or computed band widths. The conclusion about fluorine is specific to para-X-benzylidenemalononitrile and may not transfer to other scaffolds. TD-DFT carries a typical 0.2–0.3 eV error for charge-transfer states,⁶ so the 0.07 eV (CAM-B3LYP) and 0.06 eV (B3LYP) F-vs-H shifts are near the method’s noise floor; the sign is reproduced across both functionals, but the numbers should be treated as indicative. The near-linear fit to σ_p⁺ uses only four points, two of which are close together; it describes this series, not a general physical law. Measured solution spectra of these compounds would test the computed trend in a way this calculation cannot.

References

Hansch, C.; Leo, A.; Taft, R. W. A Survey of Hammett Substituent Constants and Resonance and Field Parameters. Chemical Reviews 1991, 91 (2), 165–195. https://doi.org/10.1021/cr00002a004.

Becke, A. D. Density-Functional Thermochemistry. III. The Role of Exact Exchange. The Journal of Chemical Physics 1993, 98 (7), 5648–5652. https://doi.org/10.1063/1.464913.

Weigend, F.; Ahlrichs, R. Balanced Basis Sets of Split Valence, Triple Zeta Valence and Quadruple Zeta Valence Quality for h to Rn: Design and Assessment of Accuracy. Physical Chemistry Chemical Physics 2005, 7 (18), 3297–3305. https://doi.org/10.1039/b508541a.

Yanai, T.; Tew, D. P.; Handy, N. C. A New Hybrid Exchange–Correlation Functional Using the Coulomb-Attenuating Method (CAM-B3LYP). Chemical Physics Letters 2004, 393 (1–3), 51–57. https://doi.org/10.1016/j.cplett.2004.06.011.

Smith, D. G. A.; Burns, L. A.; Simmonett, A. C.; Parrish, R. M.; Schieber, M. C.; Galvelis, R.; Kraus, P.; Kruse, H.; Di Remigio, R.; Alenaizan, A.; James, A. M.; Lehtola, S.; Misiewicz, J. P.; Scheurer, M.; Shaw, R. A.; Schriber, J. B.; Xie, Y.; Glick, Z. L.; Sirianni, D. A.; O’Brien, J. S.; Waldrop, J. M.; Kumar, A.; Hohenstein, E. G.; Pritchard, B. P.; Brooks, B. R.; Schaefer, H. F.; Sokolov, A. Yu.; Patkowski, K.; DePrince, A. E.; Bozkaya, U.; King, R. A.; Evangelista, F. A.; Turney, J. M.; Crawford, T. D.; Sherrill, C. D. PSI4 1.4: Open-Source Software for High-Throughput Quantum Chemistry. The Journal of Chemical Physics 2020, 152 (18), 184108. https://doi.org/10.1063/5.0006002.

Santoro, F.; Jacquemin, D. Going Beyond the Vertical Approximation with Time-Dependent Density Functional Theory. Wiley Interdisciplinary Reviews: Computational Molecular Science 2016, 6 (5), 460–486. https://doi.org/10.1002/wcms.1260.