Introduction

Chiral interactions between (supra)molecular excitations and localized surface plasmons in metal nanostructures generate a plethora of emerging photophysical/photochemical phenomena. For instance, chiral molecules coulombically induce chiral displacement currents on the surface of metal nanoparticles, resulting in chirality transfer from molecules to nanoparticles1,2,3. Conversely, localized chiral plasmonic fields alter the electromagnetic fields within achiral molecules, leading to induction of optical chirality in the molecules4,5. These discoveries have profound significance for chiral nano/quantum photonics6,7,8, asymmetric synthesis9,10,11, biological diagnostics12,13,14,15, etc16,17. Nevertheless, the experimental and theoretical studies in the past primarily focused on the mechanisms of electromagnetic couplings between molecules and metals, which mainly involve local field enhancement and/or superchiral field18. The potential effects of electronic excitations and transfer at the plasmon-molecule interface have often been ignored.

When the separation between molecules and metal nanostructures is within the sub-nanometer range or when they are directly chemically bonded, interfacial charge transfer (CT) may occur. This can be attributed to the overlap and hybridization of electronic states between molecules and metals, and/or the generation and transport of photo-induced hot carriers19,20. This process is widely demonstrated to alter the optical and optoelectronic properties of various resulting hybrid systems21,22,23, for example, generating enhanced Raman signals or rapid formation of triplet excited states of molecules, and others24,25,26,27. However, chiral molecule-metal complexes undergoing interfacial CT has been rarely explored. Consequently, the impact of potential mixing of electronic states of chiral molecules with the metals as well as the carriers’ transport at their interfaces in the chiroptical responses is still confusing. Moreover, since the near-field electromagnetic coupling could play a role, methods to distinguish the respective influence of the local field effect and the interfacial CT process is demanding, but have not been established yet.

There already exist theoretical reports employing full quantum models based on quantum electrodynamics (QED) or time-dependent density functional theory (TD-DFT) to describe the electron density distribution and spectral behavior of molecule-metal hybrid complexes28,29. These studies have also compared the obtained conclusions with results based on classical electromagnetic theory. Despite these advances, current full quantum approaches are still limited to hybrid structures close the size of atomic clusters. For larger hybrid systems, it is challenging to consider the molecular-level details such as electron orbitals and hybridizations at the molecule-metal interface. Furthermore, intermolecular orbital interactions add another complexity in the collectively-coupled hybrid systems. Therefore, due to the lack of unifying multiscale framework describing plasmons, CT, and molecular excitons, elucidating the CT-involved collective optical responses of chiral molecule-metal complexes is a significant theoretical challenge. Unfortunately, experimental construction of relevant model structures has been rare due to the stringent requirements in systematic molecular designs, such as tunable control over the molecule-surface bindings, optical transition energy, and even rotational strength. These challenges have greatly hampered to clarify the entangled physical and chemical mechanisms governing the chiral interactions at the interface between molecules and metals.

In this study, we atomically tune the chemical adsorption of chiral molecules on the surface of plasmonic metal nanoparticles. These two components are designed to be optically resonant in the visible region. Interfacial CT between them is unveiled through ultrafast transient absorption and Raman scattering measurement combined with quantum chemical calculations. Ground-state electronic hybridization between molecules and metals as well as photo-induced carriers’ transport are revealed. The interfacial CT greatly impacts the delocalized excitations of chiral molecules and hence excitonic couplings between molecules. This delocalization generates drastically altered circular dichroism (CD) responses, such as strong spectral shifting, broadening, and signal intensity variations. Furthermore, the CD signals associated with the intensity ratio between excitations at different energy levels is seemingly dependent on the metal nanoparticles’ sizes, indicating the contribution of local electromagnetic field enhancement in the CD generation. However, the consistent sign and broadening of the CD spectra, irrespective of nanoparticle sizes and molecular orientations, illustrate the leading role of interfacial CT in altering the chiroptical responses.

Results

Optical activity of chiral complexes beyond electromagnetic coupling

Figure 1a illustrates the formation of a chiral molecule-metal complex. The chiral molecule (Supplementary Figs. 1–11) is a derivative of BINOL-BODIPY dye (BDP)30, denoted as BDP1, which contains a pyridine group capable of chemically bonding with the surface atoms of gold nanoparticles (AuNPs)31,32. The CD of BODIPY core originates from BINOL-to-BODIPY perturbation accompanied by intramolecular CT33 and occurs at approximately 504 nm (in organic solvent), which is proximal to the wavelength of localized surface plasmon resonances of AuNPs (520 nm) (Supplementary Figs. 12–17). Therefore, BDP1 and AuNPs can form an optically resonant chiral plasmonic complex (denoted as Complex1). To examine the adsorption of BDP1 on the surfaces of AuNPs, a surface ligand exchange was conducted on Complex1, and the alterations in the surface-enhanced Raman scattering (SERS) signals were monitored (Fig. 1b). Complex1 exhibits characteristic SERS signals in the range of 800 ~ 1620 cm-1. Following replacement by 4-mercaptobenzoic acid (4-MBA), the signals of Complex1 disappeared, and the characteristic signals of 4-MBA appeared at 1074 cm−1 and 1587 cm−1, suggesting pyridine-mediated chemisorption of BDP1 on the surfaces of AuNPs. The dispersion stability of chiral complexes in aqueous solution was also investigated. Time-traced measurements of dynamic light scattering (DLS) reveal that, under moderate BDP1/AuNP molar ratios, there is no notable increase in the hydrodynamic diameter, indicating negligible aggregation of the complexes (Fig. 1c, Supplementary Figs. 18–20). This dispersive stability can be further corroborated by very small or even hardly observable peak shift in the absorption spectrum (Supplementary Fig. 21).

Fig. 1: Preparation, characterization and CD responses of Complex1.
Fig. 1: Preparation, characterization and CD responses of Complex1.The alternative text for this image may have been generated using AI.
Full size image

a Schematic illustration of the synthesis of Complex1. b SERS spectra measured under 532 nm laser illumination. The dry samples are (i) AuNPs (citrate-capped), (ii) BDP1, (iii) Complex1, (iv) AuNP@4-MBA, and (v) Complex1@4-MBA (BDP1 on AuNPs’ surface was replaced by 4-MBA). c Time-dependent changes in the DLS of 16 nm AuNPs and Complex1. The concentration of BDP1 is 1.4 μM. d CD spectra of BDP1, AuNPs, Complex1, and AuNP+BDP1. Note that Complex1 is the chemically-bonded AuNPs and BDP1 complexes while AuNP+BDP1 is a superposition of individual CD of isolated AuNPs and BDP1. e–g CD spectra of Complex1 with varied concentrations of BDP1. The BDP1 concentrations are changed from 0 μM to 0.28 μM, 0.56 μM, 1.12 μM, 1.4 μM, and 2.8 μM. The diameter of AuNPs is (e) 3 ± 0.5 nm, (f) 7 ± 0.9 nm, and (g) 16 ± 2 nm, respectively. The samples in (d–g) were incubated overnight.

The CD spectrum of Complex1 exhibits significant differences compared to BDP1 (Fig. 1d). The main peak position associated with the BODIPY core is red-shifted from 525 nm to ~550 nm and is considerably broadened with full width at half maximum (FWHM) from around 25 nm to 110 nm. New CD transitions appear in the long-wavelength region between 550 and 650 nm. Besides, the intensities of main CD peaks are substantially decreased compared to the pure molecular CD (Supplementary Fig. 12). In contrast, the pure chiral molecules retain the same CD lineshape as in organic solvent, although the molecules are prone to aggregation in aqueous solution (Supplementary Figs. 12–16). These results suggest the importance of the presence of AuNPs and the possible effect of interfacial CT between BDP1 and AuNPs under strong chemisorption conditions34 in modifying the molecular CD.

The AuNPs’ size effect was studied in order to further analyze the molecular CD alterations (Fig. 1e–g, Supplementary Fig. 22). With increasing BDP1 concentration, the CD spectra exhibited consistent change trends. At unsaturated surface concentrations of BDP1, all CD spectra exhibit similar red-shifting and broadening, regardless of the change in AuNPs’ sizes. These consistent spectral changes may be intrinsically related to the same chemical bonding interaction between BDP1 and AuNPs of different sizes. Notably, as AuNPs’ size increased, the intensity of the short-wavelength (around 525 nm) CD is seemingly increased compared to that in the long-wavelength region (around 550–600 nm). As the size of AuNPs increases, the intensity of the local electromagnetic field (hot spot) becomes stronger35,36. This enhanced near-field electromagnetic coupling between molecules and metals could contribute to the slight enhancement of CD near the surface plasmon resonance wavelengths (around 525 nm). Similar enhancement of optical absorption of chiral molecules by metal plasmons and hence the CD signals has been well interpreted theoretically37, although possible interfacial CT was not taken into account. Besides, many studies pointed out that due to near-field electromagnetic coupling (dipolar or multipolar) between chiral molecules and plasmonic nanoparticles, an induced CD response can be generated at the surface plasmon resonance wavelengths38. Thus, the induced plasmonic CD could be another factor contributing to the weak enhancement of CD signals near 525 nm for larger-sized complexes (Supplementary Figs. 23–24).

Furthermore, the chiral molecular structures were finely tailored to modulate the coupling interactions between molecules and metals. Specifically, by changing the position of pyridine group in the chiral molecules, their average adsorption orientation on the surface of AuNPs was adjusted (Fig. 2a, b). Besides, the strength of molecule-surface binding was enhanced by replacing the anchoring group from pyridine to thiol (Fig. 2c). It should be mentioned that BDP1-3 share the same chiral source group, i.e., BINOL, thus exhibiting similar CD spectral shapes and signal intensities. Crucially, these adjustments in molecular structure were found to strongly influence the chiroptical responses of the molecule-metal complexes, as detailed below.

Fig. 2: CD and Raman scattering spectra of plasmonic complexes with different chiral molecules.
Fig. 2: CD and Raman scattering spectra of plasmonic complexes with different chiral molecules.The alternative text for this image may have been generated using AI.
Full size image

a–c Schematic structures of Complex1-3 that has chiral molecules BDP1-3 on the AuNP’s surfaces, respectively. The surface charges and dipole plasmons of AuNPs are depicted. d–f CD spectra of Complex1-3 with different concentrations of chiral molecules. Complex1-2 and their AuNP controls both in aqueous solution, Complex3 and its AuNP control both in DMF. The AuNPs and the chiral molecules are incubated overnight for sample preparation. g–i Time-dependent changes in the CD spectra after mixing the chiral molecules and the AuNPs. The molecular concentrations are 1.4 μM in (g, h), and 12.8 μM in (i). The concentrations of AuNPs (3 nm) are ~600 nM in (g, h), and ~270 nM in (i). j–l SERS spectra of Complex1-2 in aqueous solution and Complex3 in DMF, under 532 or 633 nm laser excitation. The diameter of AuNPs is 3 nm. The samples in (d–f) and (j–l) were incubated overnight.

As compared to Complex1, Complex2 exhibits strong CD signals of unadsorbed molecules as well as weak signals in the long-wavelength region. This might be because BDP2 has high steric hindrance, reducing its adsorption on the AuNP surfaces (Fig. 2d, e). The orientation of chiral molecules relative to the surface of plasmonic nanoparticles may determine the sign (positive or negative) of emergent plasmonic CD induced by the molecules39. For instance, chiral molecules perpendicular or parallel to the nanoparticle surface could induce reverse plasmonic CD signals38,40. Here, the BINOL-to-BODIPY direction of BDP2 on AuNPs’ surfaces is rotated by 90 degrees as compared to that in BDP1 because of the different pyridine group positions. However, the CD spectra of the two complexes still maintain high similarity in terms of the sign and broadening of CD signals at the long-wavelength region, indicating that the molecule-induced plasmonic CD is weak, provided it exists. (Fig. 2d, e, and Supplementary Figs. 23–26). This finding is also substantiated by the close similarity of largely broadened CD lineshape at varied AuNPs’ sizes, albeit with slight CD enhancement near the surface plasmon resonance wavelengths with increasing AuNP’s sizes (Fig. 1e–g). In brief, the effects of molecular orientation and nanoparticle size together illustrate that the influence of near-field electromagnetic coupling between molecules and metals is limited in this situation, which is distinctly different from that of potential interfacial CT between them.

The effect of interfacial CT on chiral molecular excitations and interactions

Complex1-2 exhibited broadened exciton-coupled CD at high molecular concentrations (Fig. 2d, e), different from those in various supramolecular assemblies due to conventional dipolar interactions41 (Supplementary Fig. 15). This broadening may be ascribed to delocalization of BDP1 and BDP2 excitations arising from CT mixing effects42,43 (Supplementary Figs. 27–29). In contrast, the CD line of Complex3 is only slightly broadened near 500 nm accompanied by signal intensity decrease compared to pure molecules (Supplementary Fig. 17). CD signals were hardly observed in the long-wavelength region. These results illustrate less delocalization of BDP3 excitations and indicate wavelength-dependent interfacial responses (Fig. 2f). Our previous studies suggested the presence of CT state between closely-spaced BDP derivatives39. Given the distance dependence of intermolecular CT (3–4 Å), evident CT is expected at sufficiently high molecular surface densities. We conducted in situ time-dependent CD measurements at a very low molecular concentration (BDP/AuNP molar ratio about 2) to suppress the possible intermolecular CT (Fig. 2g, h). Complex1 showed progressive CD spectral broadening and splitting within 30 minutes, confirming strong delocalized nature of BDP1 excitations and couplings. Complex2 displayed minimal exciton-coupled CD with broadening, potentially due to the steric hindrance of BDP2 which causes a lower rate of adsorption. However, the pure chiral molecules (BDP1-2) in aqueous solution did not display such broadening of the CD spectrum as observed here. Thus, it can be deduced that the strong delocalization of BDP1-2 excitations primarily originates from the interfacial CT between the molecules and the nanoparticles. On the contrary, Complex3 consistently exhibited narrow CD lineshape (Fig. 2i), further demonstrating more localized nature of BDP3 excitations. This difference emphasizes the crucial role of the interfacial chemical bindings between the chiral molecules and metal surfaces.

To examine the interfacial CT within the chiral complexes, we measured their fingerprint Raman scattering (Fig. 2j–l). Interfacial CT between molecules and metals, that is a chemical enhancement mechanism, can strongly modify the SERS spectral shape associated with varied extents of enhancement among different vibrational modes of molecules26. As exhibited in Supplementary Figs. 30–32, the Raman spectral shapes of dried samples (Complex1-2) have undergone great changes at different laser excitation wavelengths. For example, as compared to 633 nm, the 532 nm excitation significantly enhances Raman scattering from pyridine or pyrrole ring stretching and deformation in plane (1520 cm−1), C-C single bond asymmetric stretching (1423 cm−1), C-H deformation in plane (1167 cm−1) and BODIPY core (964 cm−1)44,45. The monodispersed solution of Complex1-3 was also measured in order to exclude the potential influence of local electric field enhancement in dried nanoparticle aggregates. The small-sized, isolated AuNPs (3 nm) support considerably weak local electric field, thus contributing very minor physical enhancement of Raman scattering. We found that the Raman signal intensities of Complex1-3 under 532 nm laser condition are significantly increased, while 633 nm irradiation generates hardly observable signals (Fig. 2j–l). These results indicate the contribution of interfacial CT transitions between the chiral molecules and the AuNPs under excitation at 532 nm wavelength. Note that, resonant electronic transitions of molecules can occur at 532 nm wavelength, although no characteristic signals were detected from pure molecules (BDP1-3). Thus, we can speculate that interfacial CT transitions are synergistically operating with resonant excitations of molecules in the SERS of the complexes46,47.

Ultrafast transient absorption (TA) spectroscopy was employed to further verify the interfacial CT in the chiral complexes. Figure 3a, b and Supplementary Figs. 33–40 show the TA spectra of Complex1-3 (bump wavelength, 504 nm), which all appear similar to that of AuNPs. The TA spectrum exhibits two positive peaks and one negative peak between them. The negative peak is mainly attributed to the surface plasmon resonances,48 while the two positive peaks are associated with the intra/interband transitions. Upon photoexcitation, hot carriers are generated during the plasmon decay process49,50, which are subsequently consumed through electron-electron scattering (e-e, ~10 to 100 fs), electron-phonon scattering (e-ph, ~1 ps), and eventual thermalization51,52. Figure 3c depicts the carrier dynamics curves probed at a wavelength (525 nm) near the surface plasmon resonance wavelengths. It was found that the carrier relaxation rate in Complex1-2 is faster than that of AuNPs. The dynamics curves were fitted with a two-exponential decay after the rising edge, showing that a time component of carriers (possibly related to e-ph scattering) is approximately 2.01 ps in AuNPs but is reduced to around 1.65 ps in the two complexes. This faster carrier relaxation denotes an additional decay channel, through transferring from the AuNPs to the chiral molecules53,54,55.

Fig. 3: Ultrafast TA measurements of chiral complexes.
Fig. 3: Ultrafast TA measurements of chiral complexes.The alternative text for this image may have been generated using AI.
Full size image

a, b Full wavelength TA spectra of AuNPs and Complex1 in aqueous solution, respectively. c, d TA dynamics curves of AuNPs and Complex1-2 in aqueous solution monitored at 525 nm and 575 nm with the corresponding fittings. e, f TA dynamics curves of AuNPs and Complex3 both in DMF monitored at 525 nm and 575 nm with the corresponding fittings. A pump beam of 504 nm was used. The average diameter of AuNPs is 7 nm.

Moreover, we examined the carriers’ dynamics probed at a long wavelength (575 nm) to avoid strong background interference of surface plasmons (Fig. 3d). It is evident that Complex1-2 exhibit prolonged carriers’ relaxation times (800 fs and 900 fs, respectively) compared to AuNPs (500 fs). As demonstrated in Fig. 3c and Fig. 4, the interfacial CT is bidirectional, exchanging between the chiral molecules and the AuNPs. Thus, the prolonged carrier’s lifetimes in Complex1-2 observed at 575 nm suggest slower or longer lasting carrier’s exchange between molecules and metals. This result is consistent with the above conclusion, that is, the interfacial CT is primarily responsible for the delocalization of BDP1-2 excitations. As compared to Complex1-2, the carrier relaxation of Complex3 probed at 525 nm is closer to that of AuNPs. The lack of evident lifetime shortening indicates that interfacial CT from AuNP to BDP3 does not dominate the dynamics, though it may still occur (Fig. 2l). Also, the carrier lifetime of Complex3 is little changed compared to that of AuNPs at 575 nm probing wavelengths (Fig. 3e, f). The minimal interfacial CT in the long-wavelength region might be correlated with the less delocalization of BDP3 molecules and hence no apparent broadening of CD spectra in this region (Fig. 2f, i).

Fig. 4: DFT calculation results of Complex1.
Fig. 4: DFT calculation results of Complex1.The alternative text for this image may have been generated using AI.
Full size image

The electronic density of (a) HOMO-3/HOMO-4 orbital, (b) HOMO orbital, (c) LUMO orbital and (d) LUMO + 1 orbital of isolated and adsorbed BDP1. e Differential charge density of adsorbed BDP1 on Au (111). f Projected density of states (PDOS) of isolated and adsorbed BDP1 on Au (111) surface, respectively. g Differential charge density of dimeric BDP1 (BDP1-d) in close separation on Au (111). h PDOS of free and adsorbed BDP1-d on Au (111) surface, respectively. The blue arrow shows the net charge transfer from BDP1 to Au substrate and the blue and yellow isosurfaces indicate electron depletion and electron accumulation with isovalues of 10−3 e bohr−3 in (e, g). The red dotted line shows the difference in energy levels of HOMO-1 → LUMO in (f, h).

Quantum chemistry calculation on the chiral molecular excitations

We performed density functional theory (DFT) calculations of the molecular excitations at the molecule-metal interface. As a representative example, the calculation results of Complex1 were presented. To understand the contribution of transitions between molecular orbitals to the CD of Complex1, the electronic CD spectrum of the isolated BDP1 molecule was firstly calculated. Although a slight blue shift in peak frequencies was observed, the theoretically predicted CD spectrum aligned with experimental observations. The discrepancy is likely attributable to factors such as molecular concentration, environment, and computational methodology. The most intense peak in the CD spectrum of BDP1 (Supplementary Fig. 41) primarily originates from the HOMO-1 → LUMO transition. The HOMO-LUMO transition exhibits relatively low intensity, possibly due to minimal spatial overlap between these orbitals (Supplementary Figs. 42–44 and Table 1). When modeling the interaction between the BDP1 molecule and AuNPs, emphasis was placed on the adsorption of the molecule’s N-terminus at the top site of the Au (111) surface. The Au-N bond length was determined to be 2.19 Å, with an adsorption energy of −0.99 eV, indicating the formation of a chemical bond between the BDP1 molecule and the Au (111) surface.

Electronic structure analysis demonstrates significant interactions between the HOMO, LUMO, and the nearly degenerate HOMO-3/HOMO-4 orbitals (which appear as one peak in PDOS due to broadening) of BDP1 and Au (Fig. 4a–d, Supplementary Figs. 45–50). In contrast, the nearly degenerate HOMO-1/HOMO-2 orbitals show minimal overlap with Au states. Electrons on HOMO and HOMO-3/HOMO-4 orbitals of BDP1 transfer to Au, while the interaction between the unoccupied LUMO/LUMO + 1/LUMO + 2 orbitals and Au enhances electron distribution on the N-containing six-membered ring, indicative of back-donation from Au to BDP1. Notably, the adsorbed BDP1 on the Au surface induces an orbital exchange between the molecule’s LUMO + 2 and LUMO + 3 (Supplementary Fig. 54), facilitating the back-donation process. The differential charge density indicates that electron transfer between BDP1 and Au is bidirectional, involving electron exchange. Bader charge analysis reveals a net electron transfer of 0.21e from BDP1 to Au. Small PDOS peaks emerge between the HOMO and LUMO (Supplementary Fig. 53), primarily localized on the N-containing six-membered ring, potentially contributing to low-energy electronic transitions. Moreover, strong hybridization between BDP1’s LUMO + 1 orbital and Au shifts the LUMO + 1 orbital towards the Fermi level (~0.58 eV), bringing the HOMO → LUMO + 1 electron transition into the low-energy region (1.72 eV). The energy gap of the HOMO-1 → LUMO electron transition is also decreased (Fig. 4e, f). These mechanisms could be partially responsible for the observed red-shift of CD wavelengths as a result of the interfacial CT interactions.

Further, the adsorption of BDP1 with different surface coverages was investigated. As the surface coverage increased from 7% to 12% and 30%, accompanied by enhanced intermolecular interactions, the adsorption energy of BDP1 on the Au surface was observed to strengthen (from −0.99 eV to −1.12 eV and −1.34 eV). This comparison suggests preferential adsorption of molecules on the Au surface, aligning with the experimental observations. Furthermore, with increasing surface coverage, a decrease in the energy gap between HOMO-1 and LUMO orbitals was observed (Supplementary Fig. 57). Concurrently, local electron density plots reveal a further enhancement of electron exchange (Supplementary Figs. 54–56). At a surface coverage of 7% of individual molecules, the case of two BDP1 adsorbing simultaneously (denoted as BDP1-d) was examined. In this configuration, the shortest distance between two BDP1 molecules was 3.50 Å, and the co-adsorption energy was calculated to be −2.27 eV. The results showed that this configuration also exhibited an increase of electron exchange between molecular and metal states (Fig. 4g, h and Supplementary Fig. 58).

Similarly, Complex2 presents strong orbital hybridizations between molecules and metals, with net electron transfer of 0.24 e from BDP2 to Au (Supplementary Figs. 41–50). But the value was reduced to 0.15 e from BDP3 to Au possibly due to stronger back transfer of electrons as a result of stable interfacial chemical bindings (Au-S). This comparison suggests that the interfacial CT variation determined by the distinct molecule-surface bindings is pivotal for the observation of localized/delocalized chiral molecular excitations.

Discussion

The experimental and theoretical results point to two types of interfacial CT in the chiral complexes: electronic hybridization between molecular and metal states, and photo-induced carriers’ transport from metals to molecules. In the former, direct, resonant photoexcitation occurs between different hybridized states34,56. In the latter, hot electrons are scattered/diffused into the unoccupied orbitals of the molecules57,58. These two processes (or at least one of them) interfere with the molecular excitations, and, consequently, give rise to distinctly different collective CD of the molecular layer on nanoparticles’ surfaces (Fig. 5a, b). Particularly, the theoretical results and our previous studies39 indicate that at high molecular surface densities, intermolecular and interfacial CT processes may exert mutual influence, collectively contributing to the modifications of molecular excitations and couplings (Fig. 5c). Note that, because of the superposition of different chirality origins (individual vs. collective), further investigation is required to reach a conclusion on their separate contributions to the CD response, for instance, through controlling single chiral molecule on the nanoparticle surface.

Fig. 5: Schematic of possible chirality origins in the background of interfacial CT.
Fig. 5: Schematic of possible chirality origins in the background of interfacial CT.The alternative text for this image may have been generated using AI.
Full size image

a Individual molecular excitations. b, c CT mixed collective intermolecular couplings. At low molecular surface densities (b), long-range and delocalized excitonic couplings can occur, as indicated by the gray arrow. “E” represents energy transfer. Besides, at high molecular surface densities (c), short-range intermolecular CT can occur simultaneously, as denoted by the blue straight arrow. Only the molecules chemically-adsorbed to the metal surface are displayed for clarity.

The possible contribution of the Stark effect in the CD alterations is worth mentioning. Since the Stark effect is positively correlated with the electric field strength, it can be examined by comparing the AuNPs’ size that determines the local electric field enhancement. However, the red-shift and broadening of the CD spectra of chiral complexes show little correlation with the size of AuNPs (Fig. 1e–g). This suggests that the Stark effect is not dominant, although it may play a role. DFT calculations do not show new electronic transitions under the applied electric field, further corroborating the minor Stark effect (Supplementary Figs. 59–61 and Table 2). Nevertheless, we found that the surface plasmonic field has a non-negligible influence on the CD spectral shape, as it can modulate the interfacial CT transitions59. Also, due to the optical resonance between molecular excitons and localized surface plasmons, the photoexcitation of molecules is enhanced by the local near-field. Ultimately, this electromagnetic coupling effect can alter the respective intensity of CD at different energy levels, as suggested in Fig. 1e–g.

In summary, rich photophysical phenomena have been discovered in the chiral interactions at the plasmon-molecule interface. Through a combined experimental and theoretical study, the respective role of interfacial CT and near-field electromagnetic coupling between chiral molecules and plasmonic nanoparticles is revealed. It has been shown that the interfacial CT between molecules and metals has a significant impact on the delocalized molecular excitations and intermolecular couplings, leading to distinct chiroptical responses. These results are a departure from the concept of plasmon-induced chirality transfer or conferral solely considering the electromagnetic coupling between molecules and metals. Thus, this study facilitates discovery on the chiral light-matter interactions at the nanometer to sub-nanometer scale and in the collective coupling regime, and provide evidences for promoting multiscale theoretical framework incorporating plasmons, CT, and molecular excitons. Besides, the findings will inspire the study of the relationship between chirality-induced spin selectivity (CISS) in the interfacial CT and chiral optical/optoelectronic responses. It provides insights for manipulating charge/spin transport to control the quantum states of chiral molecules, and to develop potential applications in catalysis, quantum information, and more.

Methods

Synthesis and characterization of BDP1

The synthesis of BDP1 is as follows: To a solution of 1,3,5,7-tetramethyl-8-(pyridin-4-yl)-F-BODIPY (0.18 g, 0.5536 mmol) and (S)-1,1’-bi-2-naphthol ((S)-BINOL) (0.7925 g, 2.768 mmol) in dry CH2Cl2 (20 mL) was added dropwise Et2AlCl (1.00 M in Hexane, 1.1072 mL, 1.1072 mmol) under nitrogen atmosphere and the solution was stirred in the dark for 2 h. After the solvent was removed, the crude residue purified by silica-gel column chromatography (CH2Cl2/hexane, 1/1 to 4/1, v/v) afforded BDP1 (36.2 mg, 0.0634 mmol) as red solid; 1H NMR (400 MHz, Chloroform-d): δ 8.84–8.78 (m, 2H), 7.84 (d, J = 8.0 Hz, 2H), 7.78 (d, J = 8.7 Hz, 2H), 7.44 –7.38 (m, 2H), 7.32 (ddd, J = 8.0, 6.6, 1.2 Hz, 2H), 7.25–7.11 (m, 6H), 5.83 (s, 2H), 1.66 (s, 6H), 1.40 (s, 6H). 13C NMR (101 MHz, CDCl3): δ 157.55, 154.45, 150.59, 144.25, 142.18, 133.78, 131.09, 129.95, 129.32, 127.91, 127.18, 125.28, 123.78, 123.68, 123.40, 123.04, 121.49, 31.53, 30.15, 16.03, 14.97. MS (ESI): calculated: 571.4 found: m/z 572.4 (M + H)+.

Synthesis and characterization of BDP2

The synthesis of BDP2 is as follows: To a solution of 1,3,5,7-tetramethyl-6-(pyridin-4-yl)-8-phenyl-F-BODIPY (0.3 g, 0.748 mmol) and (S)-1,1’-bi-2-naphthol ((S)-BINOL) (0.4718 g, 1.6456 mmol) in dry CH2Cl2 (20 mL) was added dropwise Et2AlCl (1.00 M in Hexane, 1.5 mL, 1.5 mmol) under nitrogen atmosphere and the solution was stirred in the dark for 2 h. After the solvent was removed, the crude residue purified by silica-gel column chromatography (CH2Cl2/Hexane, 1/1 to 4/1, v/v) afforded BDP2 (51.4 mg, 0.0795 mmol) as red solid; 1H NMR (400 MHz, Chloroform-d) δ 8.49 (d, J = 5.0 Hz, 2H), 7.91–7.72 (m, 4H), 7.60–7.27 (m, 8H), 7.25–7.08 (m, 4H), 6.87 (d, J = 5.0 Hz, 2H), 5.85 (s, 1H), 1.70 (d, J = 3.6 Hz, 6H), 1.40 (s, 3H), 1.31 (s, 3H). 13C NMR (101 MHz, CDCl3): δ 157.75, 153.36, 153.21, 151.79, 148.47, 143.23, 141.59, 141.01, 137.22, 134.46, 132.76, 132.72, 132.18, 130.53, 129.73, 128.97, 128.82, 128.30, 128.23, 128.18, 128.15, 128.05, 127.21, 127.17, 126.87, 126.78, 126.13, 126.06, 124.30, 124.26, 124.17, 122.67, 122.59, 122.40, 122.30, 120.62, 120.20, 52.39, 30.55, 30.40, 29.16, 28.67, 21.62, 15.01, 13.79, 13.57, 13.10, 11.84. MS (ESI): calculated: 647.5 found: m/z 648.5 (M+H)+.

Synthesis and characterization of dimeric BDP3

The synthesis of dimeric BDP3 is as follows: To a solution of 1,2-bis[4-(1,3,5,7-tetramethyl-F-BODIPY-8-yl)phenyl]disulfide (1.49 g, 2.097 mmol) and (S)-1,1’-bi-2-naphthol ((S)-BINOL) (3.6 g, 12.582 mmol) in dry CH2Cl2 (30 mL) was added dropwise Et2AlCl (1.00 M in Hexane, 8.39 mL, 8.39 mmol) under nitrogen atmosphere and the solution was stirred in the dark for 2 h. After the solvent was removed, the crude residue purified by silica-gel column chromatography (Ethyl Acetate/Hexane, 1/2, v/v) afforded BDP3 (1.65 g, 1.37 mmol) as red solid; 1H NMR (400 MHz, Chloroform-d): δ 7.83 (d, J = 6.8 Hz, 4H), 7.77 (d, J = 8.7 Hz, 4H), 7.66 (d, J = 8.2 Hz, 4H), 7.38–7.28 (m, 8H), 7.20 (td, J = 16.6, 8.5 Hz, 12H), 5.80 (s, 4H), 1.64 (s, 12H), 1.37 (s, 12H). 13C NMR (101 MHz, CDCl3): δ 156.95, 154.54, 142.32, 137.71, 135.09, 133.78, 132.03, 129.92, 129.26, 128.86, 127.88, 127.18, 125.23, 123.74, 123.34, 122.68, 121.49, 77.35, 77.03, 76.71, 15.98, 14.84.

Circular dichroism and absorption measurements

The CD spectra were recorded using a Chirascan V100 CD spectrometer, scanning from 300 to 700 nm with a 2 nm bandwidth, a 1 cm path length, and a scan speed of 120 nm·min⁻¹. Absorption spectra were simultaneously recorded on the same instrument. Baseline correction for CD and absorption spectra was performed using background measurements of deionized water.

Raman scattering measurements

Raman scattering measurements were performed using a Renishaw inVia Reflex Raman microscope equipped with a 50× objective lens and a 532/633 nm laser. Spectra were acquired in air using a 5 s integration time.

Femtosecond and nanosecond transient absorption spectroscopy

A titanium sapphire femtosecond laser (Astrella, Coherent Inc.) was employed as the excitation source, generating 100 fs pulses centered at 800 nm with a repetition rate of 1 kHz. The 800 nm output was split into two beams using a beamsplitter. One beam was directed into an optical parametric amplifier (OPerA Solo, Coherent Inc.) to generate pump light at the desired wavelength for the experiment. The other beam was passed through a delay line and focused onto a nonlinear sapphire crystal within a transient absorption spectrometer (Helios, Ultrafast Systems) to generate a supercontinuum probe beam spanning 450 to 750 nm. This probe beam was then focused onto the sample using an off-axis parabolic mirror. The pump beam was modulated to 500 Hz by a chopper and polarized to 54.7° relative to the probe beam using a half-wave plate before being focused onto the sample. Liquid samples were placed in a 2 mm quartz cuvette. A Glan-Taylor prism, positioned after the sample, filtered out scattered pump light due to the perpendicular polarization of the pump and probe beams, enabling only the probe beam to pass through. The probe beam, carrying information about the excited and ground states of the sample, was detected by a fiber-optic probe after passing through the sample. By adjusting the delay between the probe and pump beams using a delay line, the spectral information of the sample at different time delays after excitation could be obtained. The instrumental response of the system was approximately 120 fs. The sample solutions were prepared in deionized water with a steady-state absorbance of approximately 0.2–0.3 OD at the excitation wavelength (504 nm) in a 2 mm quartz cuvette.

Density functional theory calculation

All the calculations were performed using the Vienna Ab initio Simulation Package (VASP)60,61. The projected augmented wave62 method and the Perdew-Burke-Ernzerhof (PBE) functional63 within the generalized gradient approximation (GGA) were employed to describe nuclei-electron and electron-electron interactions. Weak interactions were described using the DFT-D3 method64. The plane-wave cutoff energy was set as 400 eV. Convergence criteria for energy and force in self-consistent iterations were set to 10−5 eV and 0.02 eV/Å, respectively.

The Au nanoparticle was simulated as a three-layer (8 × 8) Au (111) surface with bottom two layers fixed at their bulk positions and the distance between adsorbed BDP1 molecules was larger than 15 Å with surface coverage of about 7%. To avoid interlayer interactions, a 30 Å vacuum layer was implemented in the z-direction. Only the gamma point was used for the sampling of the Brillouin zone for the large system65. To examine the surface coverage effects, we reduced the Au (111) unit cell to (6 × 6) and (3 × 5) surfaces using a 2 × 2 × 1 and 3 × 3 × 1 k-point mesh65, resulting in minimum distances between BDP1 molecules of 8.72 Å and 2.51 Å with molecular surface coverage of 12%, and 30%, respectively.

The adsorption energy (Ea) was calculated using the following formula:

$${E}_{a}={E}_{{sub}}{{{\rm{\hbox{-}}}}}{E}_{{slab}}{{{\rm{\hbox{-}}}}}{E}_{{mole}}$$

where Esub represents the total energy of the adsorbed system, Eslab denotes the energy of the Au (111), and Emole is the energy of the isolated molecules.

The molecular surface coverage is the maximum length of the molecule multiplied by the widest point divided by the surface area of the Au (111) surface.

Molecular orbital analyses and CD spectrum calculations for isolated BDP1 molecules were performed using the Gaussian 09 program66. All calculations within this program utilized the B3LYP functional67,68. The 6-311 G** basis set69,70,71,72 was applied to all atoms (non-metallic) for structure optimization and orbital analysis, followed by CD spectrum calculations using the 6-311 + + G** basis set73,74,75. The CD spectrum was analyzed using Multiwfn 3.8 software76.