Introduction

During the solvation treatment, the solvent plays a critical role in determining the stability, morphology, and reactivity of functional nanomaterials in colloidal solution1,2,3,4,5. In this regard, the solvent governs the fundamental thermodynamic and/or kinetic processes in the synthesis of nanomaterials. This further influences the nucleation and growth dynamics, and thereby dictates the properties of nanomaterials, such as shape, size distribution, and crystallinity6,7,8. More importantly, the persisting solvation shell dynamically interacts with the colloids by accessing the surface/interface landscapes9,10,11,12. In this scenario, the polarity of the solvent and its capacity to create an encompassing electric field would dictate the stability, reactivity, and supramolecular interactions (e.g., hydrogen bonds, dipoles, van der Waals forces, and electrostatic interactions) of colloids. This further facilitates the precise control of material isomerization and related total structures, which is responsible for the performance of functional nanomaterials13,14,15,16,17.

Metal NCs are composed of monolayer-protected few-to-hundred-atom and feature the metal(0)@metal(I)-ligand core-shell structure, which can be well-described by the “divide-and-protect” model18. The metal(0) core is dynamically wrapped by the “staple motifs” of metal(I)-ligand complexes, with uncoordinated dangling bonds on the surface. This structural characteristic renders them highly accessible to surrounding solvents19,20,21,22. Recent reports have well-recognized the significant role of solvent molecules in determining the structural and optical properties of metal NCs: 1) Solvent can act as a reducing agent to promote the growth of the metal core. For example, Sun et al. obtained Ag70 NCs with a five-fold tetrahedral Ag10 core through in-situ partial reduction using dimethylformamide23. 2) Solvent can participate in proton and/or electron transfer paths. For example, proton solvent triggers the non-radiative proton-coupled electron transfer (PCET) reaction of Au10 NCs, resulting in fluorescence quenching24. 3) Solvent can regulate crystallization and assembly. As a typical case, Zheng et al. discovered that a 34-atom gold-silver intermetallic cluster can be assembled to form a 1D polymer through solvent-mediated Ag-Au-Ag bonds between adjacent clusters25. 4) The polarity and coordination of the solvent can induce the structural transformation of the cluster. In a seminal work, Wu et al. reported that two Au28 structural isomers can undergo at least 10 reversible conversion cycles depending on the dielectric constant of the solvent26. In addition, Wang et al. reported the methanol-induced inter-cluster transformation from Ag98 to Ag74 NCs. The transformation is mainly due to the polar and coordination ability of the methanol molecules, which enable the partial replacement of protecting alkynyl ligands to induce the rearrangement of the ligand shell in Ag98 NCs27. Despite sustained progress in leveraging solvents to modulate the total structure, the fundamental principles governing solvation in determining the excited-state electronic dynamics and optical properties of metal NCs remain ambiguous and poorly understood28,29,30,31. It, therefore, is of significance for elucidating the solvent-mediated regulation mechanism and deepening the fundamental understanding toward rational design of the PL performance of metal NCs.

Herein, 6-MPR is adopted as the protecting ligand to synthesize color-tunable Au NCs, considering their well-accessed isomerism ability dictated by the surrounding solvent. We first proved that the conversion from the thione (protonated N) to the thiol (deprotonated N) isomers of 6-MPR molecules can be achieved through gradually increasing the pH from 6.5 to 13.5 in aqueous solution. This ligand isomerization is accompanied by the blue shift of the 6-MPR molecules’ absorption peak from 321 to 310 nm (Fig. 1a and Supplementary Fig. 1)32,33. Therefore, the isomerization capability of 6-MPR molecules enables the adjustment of the proton transfer process, and accordingly we envisage that the color tuning can be achieved in 6-MPR-protected Au NCs by rationally controlling the solvation of the Au NCs. As proof of concept, we have synthesized the 6-MPR-protected Au NCs and found their dual-phosphorescent emission bands located at 500–700 nm (PL I) and 600–1000 nm (PL II), arising from the R1 domain and R2 domain isomer terminals, respectively. We further investigated the impact of solvation with different organic solvent molecules on the optical properties of the Au NCs. The addition of strong coordination aprotic solvents (i.e., dimethyl sulfoxide (DMSO), N,N-dimethylacetamide (DMAc), and N,N-dimethylformamide (DMF)) can act as hydrogen bond acceptors and preferentially bind to the R1 domain in Au NCs to transform them into the R2 domain. As expected, the opposite phenomenon is observed for high-viscosity proton solvents (i.e., glycerol (GL), 1, 2-propylene glycol (PG), and ethylene glycol (EG)), which are in the nature of hydrogen bond donors and will preferentially bind to the R2 domain in Au NCs to transform them into the R1 domain. The solvent-induced surface ligand isomerization between R1 and R2 domains combined with their high viscosity property collectively triggers the formation of a lower-lying electron transfer state, which further enables the regulation of PCET dynamics and the transformation of the two PL peaks. As a result, the emission color tuning is achieved in the Au NCs spanning the entire visible range with a tunable correlated color temperature from 10,738 to 1877 K.

Fig. 1: Structural characterization and optical properties of AuH₂O NCs.
Fig. 1: Structural characterization and optical properties of AuH₂O NCs.The alternative text for this image may have been generated using AI.
Full size image

a Chemical structures of thione and thiol 6-MPR isomers. b Schematic illustration of two 6-MPR isomers protected AuH₂O NCs. c Schematic illustration of the isomerization of the surface terminal for AuH₂O NCs. R represents the ribosyl at the terminal of the 6-MPR ligand. d High-resolution N 1 s XPS and S 2p XPS spectra of AuH₂O NCs synthesized in pH = 6.5 and 13.5 aqueous solutions, respectively. For N 1 s XPS, the yellow, gray, and reddish-brown shaded areas represent the N-H, pyrrolic N, and pyridinic N, respectively. For S 2p XPS, the yellow, gray, and reddish-brown shaded areas represent the Au-S, C = S, and C-S bonds, respectively. e Positive MALDI-TOF mass spectra of AuH₂O NCs synthesized in different pH aqueous solutions (from bottom to top: pH = 6.5, 7, 9, 11, 13, and 13.5). The gray shaded areas represent the location of the MALDI-TOF mass spectrometry signal. f PL spectra of serial NCs synthesized at pH = 6.5, 7, 9, 11, 13, and 13.5, respectively (λex = 365 nm). Illustrations are the evolution of terminal ligands in the surface motifs on the AuH₂O NCs. The gray and reddish-brown shaded areas represent the PL I and PL II bands, respectively. g Comparison of UV-vis absorption spectra of AuH₂O NCs synthesized in different pH aqueous solutions (from bottom to top: pH = 6.5, 7, 9, 11, 13, and 13.5). The pink shaded area represents the surface absorbance of AuH₂O NCs. The gray shaded area represents the core absorbance of AuH₂O NCs. Source data are provided as a Source Data file.

Results

Structural characterization of dual-emission Au NCs

Based on the pH-responsive transformation of 6-MPR ligands between the thione and thiol isomers, we first synthesized a series of aqueous Au NCs protected by 6-MPR ligands within different pH conditions (pH = 6.5, 7, 9, 11, 13, and 13.5, denoted as AuH₂O NCs) to investigate the impact of pH-induced surface isomerization on the optical properties (Fig. 1b,c). We have tried to grow their high-quality single crystals to obtain atomically precise structures. However, all our attempts failed to acquire X-ray-quality ones. This can be rationally assigned to the strong dipolar interactions coupled between the 6-MPR ligands and the intrinsically high flexibility and rotationality of the ribofuranoside ring in 6-MPR ligands, which can hinder the formation of necessary stable inter-ligand interactions for long-range ordered structures34,35. We therefore conducted the structural and compositional dissection of serial AuH₂O NCs by the X-ray and mass spectra techniques. The X-ray photoelectron spectra (XPS) of the series of AuH₂O NCs were tested, and their N 1 s and S 2p orbitals binding energies were fitted. Two AuH₂O NCs synthesized with pH = 6.5 and 13.5 were selected as models to investigate the configuration of 6-MPR ligands in AuH₂O NCs. When the pH is 6.5, the N 1 s peaks located at 401.7, 400.97, and 399.9 eV were assigned to N-H, pyrrolic N, and pyridinic N species, respectively36,37,38. When the pH is increased to 13.5, the pyridinic N binding energy shifts to 399.28 eV (Fig. 1d and Supplementary Fig. 2). This shift is attributed to the deprotonation of the 6-MPR ligands, which alters the microenvironment of the pyridinic N atoms39. More importantly, the proportion of N-H peaks gradually disappears from 6.5 to 13.5 pH, attributed to the conversion of thione isomer to thiol isomer on the surface of the AuH₂O NCs. In addition, the S 2p spectra of AuH₂O NCs synthesized at pH = 6.5 are assigned to the C = S (164.81 and 163.62 eV) and Au-S ( ≈ 163.15 eV) species, respectively. However, S 2p spectra of AuH₂O NCs synthesized at pH = 13.5 are assigned to the C-S (164.14 and 162.64 eV), and Au-S ( ≈ 163.15 eV) species, respectively40,41. Therefore, with the increase of pH from 6.5 to 13.5, the proportion of C = S peak gradually decreases, and finally disappears, accompanied by the gradually increase proportion of C-S peaks (Fig. 1d, Supplementary Fig. 3). These results indicate that the 6-MPR ligands on AuH₂O NCs mainly exist in the form of thione isomer when they were synthesized in aqueous solution under low pH conditions. On the contrary, the surface 6-MPR ligands would convert to the thiol isomer with the increase of pH in the synthesis of AuH₂O NCs (Fig. 1d insert). We further carried out the electrospray ionization mass spectrometry (ESI-MS) analysis of serial AuH₂O NCs in both positive and negative-ion modes to identify their composition information. We found that the mass signals of intact NCs were hard to acquire for all AuH₂O NCs under both measuring modes, which was mainly due to the presence of excess small ions (e.g., Na+, Cl-, and OH-) and the assembly form of clusters seriously impede their ionization in ESI-MS measurements42. Alternatively, we have implemented the matrix-assisted laser desorption/ionization-time-of-flight (MALDI-TOF) mass spectra to identify the composition of as-synthesized NCs. The corresponding MALDI-TOF results manifest a common mass peak at 3275.6 g mol−1 under different pH conditions, which matches well with the chemical composition of Au8(MPR)6 (Fig. 1e). The Au 4f5/2 and 4f7/2 peaks of a series of AuH₂O NCs synthesized at different pH lie between that of Au(0) NPs and the Au(I)-p-MBA complex (Supplementary Fig. 4), indicating a characteristic core-shell structure with Au(0) core and Au(I)-S staple motif shell in AuH₂O NCs. In addition, the ratio of Au(I) to Au(0) was found to remain at ≈3:1 in a series of AuH₂O NCs (Supplementary Fig. 5), which agrees well with the as-informed chemical formula (Au8(MPR)6). The Au L3-edge Fourier-transformed extended X-ray absorption fine structure (FT-EXAFS) of serial AuH₂O NCs with corresponding fits illustrate that both Au-S and Au-N bonds are formed in the AuH₂O NCs synthesized at different pH (Supplementary Fig. 6–8 and Supplementary Table 1). Therefore, the surface 6-MPR ligands can bond on the surface of AuH₂O NCs by chelating Au(I) species with S and N atoms.

Consistent with the two-isomers-dominant landscape of AuH₂O NCs, two emission bands in the wavelength range of 500–700 nm (PLI) and 600–1000 nm (PL II), along with a pH-dependent interconversion feature, were observed in the PL spectra of serial AuH₂O NCs. To the best of our knowledge, the PL regulation of metal NCs within such a wide wavelength range has rarely been achieved (Supplementary Table 2)43,44,45. The AuH₂O NCs synthesized at pH = 6.5 and 13.5 only have PL I and PL II, respectively. With the decrease of pH from 13 to 7, the integrated emission proportion of PL I band gradually increases from 8.9% to 89.3%, while that of PL II band decreases accordingly from 91.1% to 10.7% (Fig. 1f). Combined with the above pH-dependent structural and compositional dissection, PL I and PL II are reasonably linked with the R1 domain and R2 domain, respectively. To further prove this point, the corresponding absorption spectra were measured. As shown in Fig. 1g, the absorption bands in the wavelength range of 230–300 nm and 300–400 nm are contributed by the characteristic absorption transitions of the surface staple motif and the metal core of AuH₂O NCs, respectively. When pH decreased from 13.5 to 6.5, the red shift in the absorption spectra of the AuH₂O NCs can be observed, which is caused by the formation of a lower-lying electron transfer state and the transformation between R1 and R2 domains24,46,47,48.

The PL origin and dynamics of dual-emission Au NCs

To analyze the origin of the two emission bands, the temperature-dependent PL spectra of AuH₂O NCs with pure PL I and II bands, synthesized at pH conditions of 6.5 and 13.5, respectively, were first tested. Both the emission peak intensity of PL I and PL II increases under the excitation of 365 nm with the decrease of temperature, which indicates the strong electron-vibration interaction in the AuH₂O NCs (Supplementary Figs. 9, 10)49. The plots of integrated PL intensity profiles of PL I and PL II against temperature are shown in Fig. 2a, b. To analyze the non-radiative relaxation pathway of thermal activation, we fitted the function of integrated PL intensity versus temperature according to the following Arrhenius expression50:

$$I\left(T\right)=\frac{{I}_{0}}{1+a{{{{\rm{e}}}}}^{{-E}_{{{{\rm{a}}}}}/({{{{\rm{k}}}}}_{{{{\rm{B}}}}}T)}}$$
(1)

where I0 and I(T) stand for the emission intensity at the initial temperature and other temperatures, respectively. a refers to the ratio of nonradiative and radiative probabilities. Ea denotes the activation energy for the nonradiative relaxation pathway. The fitting results give the activation energies that the acoustic phonon modes coupled with the PL I and PL II of 38.5 and 50.4 meV, respectively. Accordingly, the corresponding phonon-coupled vibration frequencies were calculated to be 308 and 403 cm−1, respectively. Based on previous theoretical and experimental studies51, the possible origin of the relatively low-frequency vibrational modes can be reasonably attributed to the Au-Au bonding vibrations (typically smaller than 150 cm−1). The high-frequency vibrational modes can be attributed to Au-S bond vibrations (typically larger than 200 cm−1). Therefore, both the PL I and PL II emission bands of the AuH₂O NCs structurally originate from the Au(I)-S interface motif (Fig. 2e). The lifetimes of PL I and PL II emission of the AuH₂O NCs were measured to be 1.19 and 2.28 μs, respectively (Supplementary Fig. 11). We further employed 1,3-diphenylisobenzofuran (DPBF) to investigate the PL type of PL I and PL II based on the consciousness that DPBF can easily react with photosensitized 1O2 to form 1,2-dibenzoylbenzene (DBB), which leads to bleaching of the 412 nm absorption band52. In our case, the 412 nm absorption intensity of oxygen-saturated DPBF (10 μM) in the aqueous solution of AuH₂O NCs (pH = 6.5 and 13.5) decreased rapidly with the delay of detection time (Fig. 2c, d), suggesting the underlying energy transfer from the triplet state of AuH₂O NCs to 3O2 to generate the singlet state 1O2. Combined with the above relatively high phonon-coupled vibration frequency (308 and 403 cm−1) and μs-level lifetimes of the PL I and PL II, it is confirmed that both emission bands of the AuH₂O NCs belong to the phosphorescent emission nature. Of note, given the sense that the 6-MPR ligand functions not merely as an electron acceptor, but more importantly as a proton acceptor, the pH-triggered terminal isomerization and PCET process is expectable or inevitable during the triplet-state phosphorescence emission process (Fig. 2f). Therefore, we conducted a series of photophysical measurements to demonstrate the existence of the PCET process and its influence on the PL properties of serial AuH₂O NCs under different pH environments. To this end, the time-resolved photoluminescence (TRPL) spectra of the AuH₂O NCs synthesized at pH = 6.5 and 13.5 (which especially give individual PL I and PL II emission bands, respectively) were first measured and analyzed (Fig. 2g). It is worth noting that the emission band of PL I of AuH₂O NCs continuously red-shifted from 509 to 590 nm within a delay time of 150 ns (Supplementary Fig. 12a). On the contrary, the PL II emission band of the AuH₂O NCs remains at 770 nm without any shift in a delay time window of 150 ns (Supplementary Fig. 12b). In addition, we further substantiate our proposed mechanistic interpretation of the bonding mode of sulfur alters conjugation and thus switches the PCET pathway by carrying out the density functional theory (DFT) simulation (Supplementary Figs. 13, 14). We designed several Au8(MPR)6 structures by directly attaching 6-MPR ligands to Au8 kernel of different morphologies, and performed structural optimizations on them53,54,55 (Supplementary Fig. 13). Ultimately, the most stable configuration is selected for further investigation (Supplementary Fig. 13d). It should be clarified that the interaction between the 6-MPR ligands and the gold core surface is highly complex. In addition to the formation of Au-S bonds, Au-N bonds are also present. Furthermore, the isomerization of the 6-MPR ligands involves the transformation of a C-S single bond into a C = S double bond. Next, we performed a frontier orbital analysis of Au8(MPR)6 NCs based on their stable isomerized forms (Supplementary Fig. 14a). As shown in the Supplementary Fig. 14b, the distributions of both the highest occupied molecular orbital (HOMO) and lowest unoccupied molecular orbital (LUMO) of Au8(MPR)6 NCs with the R2 domain are primarily localized on the gold core. In contrast, for the Au8(MPR)6 NCs with the R1 domain, the HOMO remains mainly on the gold core, while the LUMO is predominantly distributed over the ligands. The as-observed significant differences in the frontier orbital analysis suggest different excitation characteristics: the Au8(MPR)6 NCs with R2 domain involve electronic excitation largely within the gold core, whereas the Au8(MPR)6 NCs with R1 domain exhibit electronic excitation between the gold core and the ligands and therefore lead to the PCET process in the AuH₂O NCs with R1 domain56. This result strongly supports our proposed mechanistic interpretation that the bonding mode of sulfur alters conjugation and thus switches the PCET pathway.

Fig. 2: Vibrational modes and PCET mechanism of AuH₂O NCs.
Fig. 2: Vibrational modes and PCET mechanism of AuH₂O NCs.The alternative text for this image may have been generated using AI.
Full size image

a,b Normalized integrated PL intensity of the PL I (AuH₂O NCs synthesized in pH = 6.5 aqueous solution) and PL II (AuH₂O NCs synthesized in pH = 13.5 aqueous solution) as a function of temperature and their corresponding fits by using the Arrhenius equation. The gray and reddish-brown line represents AuH₂O NCs with R1 domain (pH = 6.5) and AuH₂O NCs with R2 domain (pH = 13.5), respectively, in a and b. c,d Time-dependent absorption spectra of the DPBF including AuH₂O NCs (pH = 6.5 and 13.5) under ambient conditions. e Schematic illustration of phonon-coupled vibrations of Au core and Au-S motifs. f Schematic illustration of photo-induced PCET reaction for AuH₂O NCs, R represents the ribosyl at the terminal of the 6-MPR ligand. g TRPL spectra of the AuH₂O NCs synthesized at pH = 6.5 and 13.5 (dominated by PL I and PL II, respectively) upon 375 nm laser excitation. h Cathodic cyclic voltammetry profiles of AuH₂O NCs (synthesized at pH = 7, 9, 11). The inset shows the cathodic potential of AuH₂O NCs versus pH. i Excited-state PCET rate of AuH₂O NCs (synthesized at pH = 6.5, 7, 9, 11) dissolved in H2O and D2O extracted from their corresponding PL lifetimes. j Plots of ln(kH/kD) vs. 1/T for AuH₂O NCs synthesized at pH = 6.5 and the corresponding linear fit. Source data are provided as a Source Data file.

The corresponding cathodic cyclic voltammetry scans of serial AuH₂O NCs reveal a pH-dependent reduction peak in the pH range of 7–11. Fitting the cathodic potential versus pH relationship using the Nernst equation gives a slope of 55 mV per pH unit (Fig. 2h), which is consistent with the theoretical 59 mV pH-1 slope predicted for a 1e-/1H+ PCET process57,58. All the PL I and PL II decay curves recorded by the time-correlated single-photon counting (TCSPC) method were then subjected to fitting to extract the PL lifetimes according to the following double-exponential decay model (Supplementary Figs. 11, 15, and Supplementary Tables 3–5)59:

$${I}_{(t)}=\,{I}_{0}+\,{A}_{1}{{{{\rm{e}}}}}^{(\frac{-t}{{\tau }_{1}})}+\,{A}_{2}{{{{\rm{e}}}}}^{(\frac{-t}{{\tau }_{2}})}$$
(2)

Based on the PL lifetime measurements, the PCET rate for gold NCs is calculated by Eq. (3)60:

$${k}_{{{{\rm{PCET}}}}}=\frac{1}{{\tau }_{{{{\rm{app}}}}.}}-\,\frac{1}{{\tau }_{{{{\rm{ave}}}}.}}$$
(3)

where τapp. and τave. represent the apparent and intrinsic PL lifetimes of AuH₂O NCs, respectively. τapp. = ΣAiτi, where Ai and τi are the amplitude and lifetime components, respectively. The τave. is the average lifetime of AuH₂O NCs. The excited-state PCET rates for PL I of AuH₂O NCs (synthesized in the aqueous solution with pH = 6.5, 7, 9, and 11, respectively) dissolved in H2O and D2O were calculated to be 1.42 × 106, 1.16 × 106, 1.13 × 106, 1.04 × 106 and 0.96 × 106, 0.80 × 106, 0.78 × 106, 0.71 × 106 s-1, respectively (Fig. 2i). It can be informed that the AuH₂O NCs synthesized at a lower pH will give faster PCET rate both in H2O and D2O solvent.

In addition, for the AuH₂O NCs synthesized at the same pH, the H2O solvent would give a faster PCET rate than that in D2O.To further verify the excited-state electron transfer mechanism, the kinetic isotope effect (KIE) values for PL I of AuH₂O NCs were determined by Eq. (4)61:

$${{{\rm{KIE}}}}=\frac{{k}_{{{{\rm{PCET}}}}}^{{{{\rm{H}}}}}}{{k}_{{{{\rm{PCET}}}}}^{{{{\rm{D}}}}}}$$
(4)

where \({k}_{{{{\rm{PCET}}}}}^{{{{\rm{H}}}}}\) and \({k}_{{{{\rm{PCET}}}}}^{{{{\rm{D}}}}}\) are the PCET rate constants in H2O and D2O, respectively. The KIE values were calculated to be 1.48, 1.47, 1.44, and 1.46 in pH = 6.5, 7, 9, and 11 aqueous solutions, respectively. Such large KIE values (>1.0) further confirmed that the PCET process is involved and can be controlled in the emission of the PL I band in AuH₂O NCs with different pH conditions. Then, the temperature-dependent PL lifetimes of Au NCs (pH = 6.5) were measured in both H₂O and D₂O solutions (Supplementary Fig. 16). The KIE exhibits an exponential dependence on the absolute temperature (1/T), and is governed by Eq. (5)62:

$$\frac{{k}_{{{{\rm{H}}}}}}{{k}_{{{{\rm{D}}}}}}=\,\frac{{A}_{{{{\rm{H}}}}}}{{A}_{{{{\rm{D}}}}}}{{{{\rm{e}}}}}^{{\Delta E}_{{{{\rm{a}}}}({{{\rm{D}}}}-{{{\rm{H}}}})}/({{{\rm{R}}}}T)}$$
(5)

where \(\frac{{A}_{{{{\rm{H}}}}}}{{A}_{{{{\rm{D}}}}}}\) is the pre-exponential factor, Ea denotes the activation energy, \({\Delta E}_{{{{\rm{a}}}}({{{\rm{D}}}}-{{{\rm{H}}}})}\) stands for the difference in activation energy between deuterium and hydrogen. R is the universal gas constant. When plotting \({{\mathrm{ln}}}(\frac{{k}_{{{{\rm{H}}}}}}{{k}_{{{{\rm{D}}}}}})\) versus 1/T, the Arrhenius equation is transformed Eq. (6)62:

$${{\mathrm{ln}}}\left(\frac{{k}_{{{{\rm{H}}}}}}{{k}_{{{{\rm{D}}}}}}\right)=\frac{{\Delta E}_{{{{\rm{a}}}}\left(D-H\right)}}{{{{\rm{R}}}}T}+{{\mathrm{ln}}}\left(\frac{{A}_{{{{\rm{H}}}}}}{{A}_{{{{\rm{D}}}}}}\right)$$
(6)

where the slope \(\frac{{\Delta E}_{{{{\rm{a}}}}\left(D-H\right)}}{{{{\rm{R}}}}}\) and the intercept \({{\mathrm{ln}}}(\frac{{A}_{{{{\rm{H}}}}}}{{A}_{{{{\rm{D}}}}}})\) derived from the plot’s tangent are sensitive parameters for hydrogen tunneling, providing direct insight into the degree of tunneling that occurs during the PCET reaction. Therefore, we plotted and fitted the function of \({{\mathrm{ln}}}(\frac{{k}_{{{{\rm{H}}}}}}{{k}_{{{{\rm{D}}}}}})\) versus 1/T according to the Eq. (6) to analyze hydrogen tunneling contributions (Fig. 2j). The fitting results give the Arrhenius pre-exponential factor \(\frac{{A}_{{{{\rm{H}}}}}}{{A}_{{{{\rm{D}}}}}}\) of 0.14 and the difference in activation energy \({\Delta E}_{{{{\rm{a}}}}\left({{{\rm{D}}}}-{{{\rm{H}}}}\right)}\) of 5.91 kJ mol−1, respectively. The relatively larger \({\Delta E}_{{{{\rm{a}}}}\left({{{\rm{D}}}}-{{{\rm{H}}}}\right)}\) (>5.0 kJ mol−1 of vibrational zero-point energies for O-H and O-D bonds) and the smaller Arrhenius pre-exponential factor (AH/AD < 0.5) suggest the presence of a prominent proton tunneling effect in PL I, which strongly supports the PCET assignment62,63. The presence of the PCET process can lower the reaction energy barrier by coupling the electron and proton transfer processes, enabling efficient electron transfer under alkaline conditions64,65. Hence, the PCET process in AuH₂O NCs is essential to achieve the stable, continuous, and reversible broad PL modulation even under alkaline conditions (up to pH = 11), which cannot be achieved merely through simple proton transfer at the N‑sites of the ligand limited in the relatively lower pH condition24.

Solvent-directed emission tuning in Au NCs

The sensitive optical responses of 6-MPR-protected Au NCs to the pH condition in aqueous solution greatly inspire a further investigation of their solvation-triggered PL tuning properties. In this context, we have mapped out the impact of solvation with different organic solvents on the optical properties of the as-obtained AuH₂O NCs. Fourteen organic solvents in four categories, namely weak coordination aprotic solvents (including pyridine (Py), tetrahydrofuran (THF), acetone (ACE), and dichloromethane (DCM)), strong coordination aprotic solvents (including DMSO, DMAc, and DMF), low-viscosity proton solvents (including tert-butyl alcohol (t-BuOH), isopropyl alcohol (IPA), ethanol (EtOH), and n-propanol (n-PrOH)) and high-viscosity proton solvents (including GL, PG, and EG), were selected and added into the aqueous AuH₂O NCs to tailor their PL properties (Fig. 3a). Especially, we found that the AuH₂O NCs (synthesized at pH = 11 as example) are insoluble in weak coordination aprotic solvents and low-viscosity protic solvents, and their emitting color unchanged due to their intrinsic low polarity. The PL I peak shows a clear blue shift when the AuH₂O NCs are added to the strong coordination aprotic solvents (i.e., DMSO, DMAc, DMF) and high-viscosity protic solvents (i.e., GL, PG, EG). With the increase of the volume ratio of strong coordination aprotic solvents to water from 1.5:1 to 10:1, the proportion of PL I emission band gradually decreases. In contrast, that of the PL II emission band gradually increases. Accordingly, the AuH₂O NCs exhibit multicolor emissions from green to red under 365 nm near-ultraviolet light illumination (Fig. 3c). However, adding the high-viscosity proton solvent can decrease the proportion of PL II emission band and increase the proportion of PL I emission band, which is opposite to the results of adding strong coordination aprotic solvents. In this case, the AuH₂O NCs show multicolor emissions from orange to blue with increasing the volume ratio of high-viscosity proton solvents to water from 1.5:1 to 10:1 (Fig. 3d). The emission color spans almost the full-spectrum range, and the corresponding correlated color temperature (CCT) varies between 1877 and 10738 K (Supplementary Fig. 17 and Supplementary Table 6). In addition, the PLQY of the Au NCs shows significant enhancement from 5.86% in aqueous solution to 29.4% and 30.6% in H2O/DMSO and H2O/GL mixed solvents, respectively (Supplementary Figs. 18, 19, and Supplementary Table 7).

Fig. 3: Steady-state optical properties of AuH₂O NCs in different solvents.
Fig. 3: Steady-state optical properties of AuH₂O NCs in different solvents.The alternative text for this image may have been generated using AI.
Full size image

a Structural formulas of the adopted fourteen organic solvents in four categories, the red font is the dielectric constant (ε) of the corresponding solvents. b Schematic illustrates the form of AuH₂O NCs in organic solvents (DMSO and GL as representatives).c Steady-state emission spectra of AuH₂O NCs in H2O/DMSO, H2O/DMAc, and H2O/DMF, respectively (the volume ratio of aprotic solvent increased from top to bottom, i.e., AuDMSO-1 to AuDMSO-6, AuDMAc-1 to AuDMAc-6, and AuDMF-1 to AuDMF-6, λex = 365 nm). d Steady-state emission spectra of AuH₂O NCs in H2O/GL, H2O/PG, and H2O/EG, respectively (the volume ratio of proton solvent increased from top to bottom: i.e., AuGL-1 to AuGL-6; AuPG-1 to AuPG-6; and AuEG-1 to AuEG-6, λex = 365 nm). Insets are photographs taken under 365 nm UV light. R represents the ribosyl at the terminal of the 6-MPR ligand. Source data are provided as a Source Data file.

We further performed dynamic light scattering (DLS) to analyze the particle size of the AuH₂O NCs in aqueous solution and a mixture of water and different organic solvents (Supplementary Fig. 20). It can be found that the size distributions of serial AuH₂O NCs in different solvent mixtures are almost identical, suggesting that the as-observed solvent-dependent multicolor emission is not derived from the solvent-triggered aggregation-induced emission of NCs. For aprotic solvents, such as DMSO, they can act as hydrogen bond acceptors, which preferentially interact with the R1 domain in AuH₂O NCs and form hydrogen bonds with the H-N on the mercaptopurine ring (Fig. 3b, c insert)66,67,68. The increase of aprotic solvents in the aqueous solution of AuH₂O NCs thus can promote the conversion from R1 to R2 domain in AuH₂O NCs and result in the gradual blue shift and silence of PL I band but consistent enhancement of PL II band. For proton solvents, such as GL, they can act as hydrogen bond donors, which preferentially interact with the R2 domain in AuH₂O NCs and form hydrogen bonds with the N site on the mercaptopurine ring (Fig. 3b, d insert)69,70,71. The increase of proton solvents thus can forward the conversion from R2 to R1 domain in AuH₂O NCs, leading to the gradual blue shift and increased intensity of the PL I band, along with the disappearance of the PL II band.

Mechanism insights into the emission tuning in Au NCs

To further explore the mechanism of tailorable multiple-peak emission and PLQY enhancement of AuH₂O NCs in different pH and organic solvent mixture conditions, by employing the typical H2O/DMSO and H2O/GL (i.e., AuDMSO1-6 NCs and AuGL1-6 NCs, named according to the order of increase in the amount of organic solvent) systems as models, the Au L3-edge X-ray absorption near-edge structure (XANES), Au L3-edge FT-EXAFS, Wide-angle X-ray scattering (WAXS), and the femtosecond-transient absorption (fs-TA) spectra were carried out. In detail, the Au L3-edge XANES spectra show that the valence states of AuH₂O NCs in H2O/DMSO and H2O/GL are lower than those of AuH₂O NCs in aqueous solution (Fig. 4a). This indicates that the organic solvent can inhibit the PCET from the metal core to the interface R1 domain72. In addition, the average coordination number (C.N.) of Au NCs extracted from corresponding FT-EXAFS spectra in different solvents has been changed (Fig. 4b, Supplementary Figs. 21, 22, and Supplementary Table 1). Concretely, the fitted C.N. of Au-N, Au-S, and Au-Au bonds in AuH₂O NCs (pH = 11) are 0.22, 1.16, and 0.18, respectively. However, the average C.N. of Au-S bonds in AuDMSO-3 NCs increased to 1.82, and the average C.N. of Au-N/O bonds in AuGL-3 NCs increased to 0.45, which strongly indicates that the S in DMSO or O in GL has been coordinated with Au(I) species in AuH₂O NCs11,73,74. This coordination effect will result in the contraction of the gold core (as manifested by the Au-Au bond length shrinking from 2.94 to 2.89 and 2.83 Å). In addition, organic solvents will also change the structural and electronic configurations of serial Au NCs. As shown in Fig. 4c, a diffraction peak appears in the q range of 1.0–1.5 Å-1, indicating the existence of π-π stacking between the 6-MPR ligands on the surface of AuH₂O NCs75. The distance of π-π stacking was found to get smaller from 0.65 (in aqueous solution) to 0.50 (in H2O/DMSO mixing solvent) and to 0.43 nm (in H2O/GL mixing solvent). The stronger π-π interaction between adjacent 6-MPR ligands can thus suppress the vibration and rotation of the surface landscape and further increase the rigidity of AuH₂O NCs76.

Fig. 4: The ultrafast electron dynamics and local structure of AuH₂O NCs in different solvents.
Fig. 4: The ultrafast electron dynamics and local structure of AuH₂O NCs in different solvents.The alternative text for this image may have been generated using AI.
Full size image

a Au L3-edge XANES spectra of Au foil, AuH₂O, AuDMSO-3, and AuGL-3 NCs. The inset shows the magnified XANES spectra in the energy range of 11.915–11.925 keV. b Standard Au-Au, Au-S, Au-N, and Au-O paths, and the experimental data and corresponding best fits of Au L3-edge FT-EXAFS (shown in k2-weighted R-space) of AuH₂O, AuDMSO-3, and AuGL-3 NCs by using these scattering paths. The reddish-brown line represents the curve fitting of AuH₂O, AuDMSO-3, and AuGL-3 NCs. c Transmission mode WAXS maps and linear spectra of AuH₂O, AuDMSO-3, and AuGL-3 NCs. d The 430 nm ESA kinetic decays and corresponding fits of AuH₂O NCs (pH = 11), (drew in Adobe Illustrator). e The 430 nm ESA kinetic decays and corresponding fits of AuDMSO-1, AuDMSO-6, AuGL-1, and AuGL-6 NCs. Source data are provided as a Source Data file.

We then tested the fs-TA spectra of serial Au NCs synthesized in different pH aqueous solutions (synthesized under pH = 6.5, 7, 9, 11, 13, and 13.5) to investigate their electron relaxation dynamic properties. As shown in Supplementary Fig. 23a–f, under the excitation of 350 nm pulsed laser, a dominant excited-state absorption (ESA) band centering at 450 nm is found for Au NCs synthesized at pH = 6.5 (with pure PL I emission). And accordingly, these 450 nm ESA bands originate from the pure R1 domain. Furthermore, a dominant ESA band centering at 580 nm is found for Au NCs synthesized at pH = 13.5 (with pure PL II emission). Similarly, these 580 nm ESA bands originate from the pure R2 domain. While for the Au NCs synthesized at pH = 7–13, the ESA signals of the R1 and R2 domains overlap each other. Especially, as the pH increases, the proportion of the ESA signal of the R2 domain increases, and that of the R1 domain decreases accordingly. To avoid the overlap of two ESA from different domains, we selected the 430 (pH = 6.5–11) and 610 nm (pH = 13 and 13.5) ESA signals as the decay kinetics of R1 and R2 domains, respectively, and further conducted kinetic fits to them. It was found that the 430 and 610 nm ESA signals can be decomposed into different decay components: 0.33 ps, 4 ps, 28 ps, and > 1 ns for pH =  6.5 AuH₂O NCs; 0.32 ps, 3.8 ps, 36 ps, and >1 ns for pH = 7 AuH₂O NCs; 0.30 ps, 4 ps, 42 ps, and >1 ns for pH = 9 AuH₂O NCs; 0.30 ps, 3.6 ps, 55 ps, and >1 ns for pH = 11 AuH₂O NCs; 0.31 ps, 3.7 ps, and >1 ns for pH = 13 AuH₂O NCs; 3.8 ps, and >1 ns for pH = 13.5 AuH₂O NCs (Supplementary Fig. 23g–m and Supplementary Table 8). Specifically, the decay process of about 0.3 ps in the ESA signals can be assigned to the internal conversion (IC) process of hot electrons from the high-energy singlet state (Sn) to the lowest singlet state (S1) (Sn → S1). The several picosecond time components are attributed to the ultrafast core-shell transition of electrons from the S1 to the lowest triplet state (T1), that is, the intersystem crossing (ISC) process. The few dozen picoseconds decay component is attributed to the PCET process from the metal core to the R1 domain in AuH₂O NCs77. Finally, the time component of >1 ns belongs to the PL process of the AuH₂O NCs. Interestingly, as the pH increases from 6.5 to 13.5, the time components of PCET gradually increase, indicating that a high pH environment would suppress PCET in AuH₂O NCs. The PCET process has even vanished and cannot be detected in the AuH₂O NCs synthesized at pH = 13 and 13.5. Especially, when pH = 13.5, the 350 nm pump energy is close to the band gap of corresponding AuH₂O NCs (3.534 eV), giving no IC process.

We then tested and fitted the 430 nm ESA decay signals of AuH₂O NCs (pH = 11) in different H2O/organic mixed solvents (i.e., H2O/DMSO, H2O/DMAc, H2O/DMF, H2O/GL, H2O/PG, and H2O/EG) to check the difference of excited state electron relaxation process (Fig. 4e, Table 1, Supplementary Fig. 24, and Supplementary Table 9). Within about 8 ns delay time window, the corresponding fits give three or two decay components depending on the type of mixed solvent. Taking DMSO and GL as an example, the time components were fitted as follows: 3.2 ps, 248 ps, and >1 ns for AuDMSO-1; 3.2 ps, and >1 ns for AuDMSO-6; 3.6 ps, 265 ps, and >1 ns for AuGL-1; 3.6 ps, and >1 ns for AuGL-6. More importantly, with the addition of organic solvents to the AuH₂O NCs, the time component of PCET increased from the initial 55 ps to hundreds of picoseconds and finally disappeared. This is mainly due to the coordination effect between the solvent molecules and the NCs and the enhanced structural rigidity of the NCs, which leads to the inhibition of the PCET process.

Table 1 Fitting parameters of fs-TA of AuH₂O NCs in different solvents
Full size table

On the one hand, due to the hydrogen bond acceptor and coordination properties of strong coordination aprotic solvents, the R1 domain related to PCET is gradually reduced. And therefore, the PCET process from the metal core to the R1 domain gradually disappears. On the other hand, although the hydrogen-bond-donating of the high viscosity proton solvent gradually strengthens the R1 domain involved in PCET, the concomitant high viscosity impedes surface/interface vibrational dynamics in AuH₂O NCs78,79. Consequently, this suppression leads to a progressive decline in the PCET process from the metal core to the R1 domain. The interconversion of PL I and PL II bands and the blue shift of PL I in steady-state spectra support this finding. Fs-TA results further indicate that organic solvents induce electronic reorganization in the AuH₂O NCs. Crucially, the introduction of organic solvents drives three key changes: 1) isomerization of surface 6-MPR ligands, 2) regulation of electron dynamics, and 3) suppression of surface/interface vibrational relaxation. This combination of effects is responsible for the enhanced PLQY and emission tuning of Au NCs.

Combined with the above steady-state absorption, emission, and fs-TA spectra, we can reasonably explain the as-observed emission tuning mechanism of AuH₂O NCs in different solvent mixtures (Fig. 5). In aqueous solution, the electrons in the metal core of AuH₂O NCs composed of pure R1 domain can be excited from the S0 to the Sn under high-energy (3.54 eV) laser pumping. Then, the excited electrons in the metal core decay rapidly to the T1 state of the surface state through the IC process and the ISC process, respectively. Then, the excited-state electrons in the T1 state undergo PECT relaxation to the electron transfer state (ETS). Finally, the excited electrons in the ETS relax to the lowest S0 level through radiative relaxation, accompanied by the microsecond phosphorescence emission of 590 nm (PL I). Different from the AuH₂O NCs composed of pure R1 domain, the entire electronic transition of AuH₂O NCs composed of pure R2 domain does not involve the PCET and IC process. The excited electrons in the T1 state relax to the lowest S0 level through radiative relaxation, accompanied by the microsecond phosphorescence emission of 770 nm (PL II). In H2O/organic solvents, the S, N, or O in the solvent molecules can coordinate with the Au(I) species at the interface of the AuH₂O NCs, resulting in the transformation of the R1 and R2 domains. Specifically, aprotic solvents with strong coordination will transform R1 into R2 domain in AuH₂O NCs, which is characterized by a decrease in the proportion of PL I and an increase in the proportion of PL II in the PL spectrum. The decrease of PL I related to PCET determined by the R1 domain further led to the disappearance of the PCET process. On the contrary, proton solvent with high viscosity causes the transformation of R2 into R1 domain in AuH₂O NCs, which can increase the proportion of PL I and decrease the proportion of PL II in the spectrum. However, the high viscosity of the proton solvent inhibits the vibration and rotation of the electron donor and acceptor of the AuH₂O NCs. And therefore, it also shows the disappearance of the PCET process along with a blue shift of the PL I band in the PL spectra. By precisely regulating the surface ligand configuration and excited-state electron dynamics through decorating the surrounding solvent chemical environment of the AuH₂O NCs, the customizable multiple PL peaks and color-tunable emissions can be realized in these gold NCs.

Fig. 5: Proposed photophysical mechanism for solvent-modulated emission tuning in AuH₂O NCs.
Fig. 5: Proposed photophysical mechanism for solvent-modulated emission tuning in AuH₂O NCs.The alternative text for this image may have been generated using AI.
Full size image

a The excited-state electron relaxation of PL I and PL II in AuH₂O NCs in aqueous solution (drew in Adobe Illustrator). b The excited-state electron relaxation of PL I and the evolution of the terminal domain on the AuH₂O NCs in organic solvents (drew in Adobe Illustrator). In the energy level diagram, the dark and light gray shaded areas represent the intermediate energy level of the T1 state and the light-emitting energy level of ETS in AuH₂O NCs with R1 domain, respectively. The red shaded area represents the light-emitting energy level of the T1 state in AuH₂O NCs with R2 domain. In the spectrum, the gray and reddish-brown shaded areas represent the PLI and PLII bands, respectively. S0: ground state, Sn: high-energy singlet state, S1: lowest singlet state, T1: lowest triplet state, ETS: electron transfer state, IC: internal conversion, ISC: the intersystem crossing, PCET: proton-coupled electron transfer. Source data are provided as a Source Data file.

Discussion

In summary, we proposed the solvation-mediated isomerization of surface terminals and electron transfer rate management in 6-MPR-protected AuH₂O NCs to realize the emission tuning spanning from 495 nm sky-blue to 800 nm near-infrared light. The thione and thiol isomers of 6-MPR ligands endow AuH₂O with R1 and R2 domains, which give the PL I band related to the electron transfer emission state and the PL II band from the triplet state, respectively. In aqueous solution, tuning pH can adjust the relative content of R1 and R2 domains in the AuH₂O NCs to regulate the PCET rate related to the PL I. In addition, the solvation of AuH₂O NCs in organic solvents allows the strong coordinated aprotic (i.e., DMSO, DMAc, DMF) and high-viscosity protic solvent molecules (i.e., GL, PG, and EG) to strongly coordinate with Au(I) species in the AuH₂O NCs, inducing the transformation between R1 and R2 domains. This effect leads to the prohibition of PCET from the metal core to the R1 domain. Therefore, the collective effects of solvation-triggered surface terminal isomerization and management of PCET dynamics direct the contribution and peak position of PL I and PL II, and finally result in the customizable emissive color of Au NCs.

Methods

Materials and reagents

Hydrogen tetrachloroaurate (III) trihydrate (HAuCl4·3H2O, ≥49.0% Au basis) was purchased from Sigma Aldrich. 6-MPR (99%), sodium hydroxide (NaOH, 96%), hydrochloric acid (HCl, 36%), and DPBF (≥97%) were purchased from Aladdin Reagent Co. (Shanghai, China). Py (≥99.9%), THF (≥99.9%), ACE (≥99.7%), DCM (≥99.9%), DMSO (≥99.9%), DMAc (≥99.9%), DMF (≥99.9%), t-BuOH (≥ 99.9%), IPA (≥99.9%), EtOH (≥ 99.9%), n-PrOH (≥99.9%), GL (≥99.9%), PG ( ≥99.5%), and EG (≥99.7%) were purchased from Xilong Science Co. (Shanghai, China). Ultrapure water (18.2 MΩ cm) was purchased from Aladdin Reagent Co. (Shanghai, China). All chemicals were used as received without additional purification. All glassware was washed with aqua regia and rinsed with ultrapure water before use.

Preparation of AuH₂O NCs

In a typical synthesis of AuH₂O NCs, 625 μL aqueous solution of HAuCl4 (20 mM) and 1 mL aqueous solution of 6-MPR (25 mM in 450 mM NaOH) were added to 10 mL of water in sequence under 1000 × g stirring at room temperature. After stirring for 5 min, a certain amount of HCl (1 M) was dropped into the reaction solution to tune the pH of the aqueous solution to 6.5, 7, 9, 11, 13, and 13.5, respectively. After incubating for 12 h, the above NCs aqueous solution was dialyzed for 12 h with water at pH 6.5, 7, 9, 11, 13, and 13.5, respectively. Finally, the dialyzed NCs aqueous solution was freeze-dried for later use.

Preparation of a series of organic solvents-introduced AuH₂O NCs

0.2 mL AuH₂O NCs (0.5 mg mL-1) aqueous solution is added into 0.3, 0.4, 0.5, 0.7, 1, and 2 mL organic solvents, respectively (organic solvents are DMSO, DMAc, DMF, GL, PG, and EG, the volume ratio of organic solvents to water are 1.5:1, 2:1, 2.5:1, 3.5:1, 5:1, and 10:1, respectively). The corresponding samples are named as AuDMSO-1, AuDMSO-2, AuDMSO-3, AuDMSO-4, AuDMSO-5, AuDMSO-6 NCs, AuDMAc-1, AuDMAc-2, AuDMAc-3, AuDMAc-4, AuDMAc-5, AuDMAc-6 NCs, AuDMF-1, AuDMF-2, AuDMF-3, AuDMF-4, AuDMF-5, AuDMF-6 NCs, AuGL-1, AuGL-2, AuGL-3, AuGL-4, AuGL-5, AuGL-6 NCs, AuPG-1, AuPG-2, AuPG-3, AuPG-4, AuPG-5, AuPG-6 NCs, AuEG-1, AuEG-2, AuEG-3, AuEG-4, AuEG-5, AuEG-6 NCs, respectively.

Characterization

MALDI-TOF mass spectra were tested on a Bruker Autoflex speed TOF under a positive linear mode. All samples were tested by using trans-2-[3-(4-tert-Butylphenyl)-2-methyl-2-propenyldidene] malononitrile (DCTB) as the matrix. XPS of the AuH₂O NCs synthesized in pH = 6.5–13.5 aqueous solutions were collected on an ESCALAB250 spectrometer (Thermo Fisher). Cyclic voltammetry was performed in a three-electrode cell using a glassy carbon working electrode (3 mm), an Ag/AgCl reference electrode, and a platinum wire counter electrode. The electrolyte was aqueous 1 M NaCl containing the AuH₂O NCs synthesized at pH = 7–11. Hydrodynamic diameters of series Au NCs were measured by DLS with Zetasizer Nano S (Malvern Instruments Ltd.) at room temperature. UV-vis absorption spectra were captured on the Shimadzu UV-1900i spectrometer. PL spectra were recorded on a Edinburgh FLS1000 spectrometer. The PLQY of the AuH₂O, AuDMSO-1, and AuGL-4 NCs was measured and calculated on an Edinburgh FLS1000 spectrometer equipped with an integrating sphere coating with a BaSO4 reflective layer. For the detailed measurements of the PLQY of all samples, an Xe lamp with a fixed emission wavelength at 365 nm was used as the excitation source. Pure H2O, H2O/DMSO, and H2O/GL mixed solvents support was measured first in the integrating sphere and used as blank references for the samples in an aqueous solution. The calculation of absolute PLQY values was conducted on the built-in “Fluoracle” software (version 2.13.2). The values of PLQY were calculated by Eq. (7):

$${{{\rm{PLQY}}}}=\,\frac{\int \lambda P\left(\lambda \right){{{\rm{d}}}}\lambda }{\int \lambda \left\{E\left(\lambda \right)-R(\lambda )\right\}{{{\rm{d}}}}\lambda }$$
(7)

where E(λ)/(hν), R(λ)/(hν), and P(λ)/(hν) are the number of photons in the spectrum of excitation, reflectance, and emission, respectively. Since the PL intensity of water fluctuates greatly in the near-infrared detection region, the PLQY of AuH₂O NCs is combined with absolute (visible region: <800 nm) and relative (near-infrared region: > 800 nm) measurement methods. The absolute PLQY of AuH₂O NCs in the visible region is 4.67%, and because the peak area ratio of AuH₂O NCs in the visible region to the near-infrared spectrum is 3.92, the PLQY of AuH₂O NCs in the near-infrared region is 1.19%. Therefore, the PLQY of AuH₂O NCs is 5.86%. For AuDMSO-1 and AuGL-4 NCs, the PL in the near-infrared region accounts for a small proportion, and the PLQY in the near-infrared region is neglected when calculating the PLQY. The TRPL spectra and PL lifetime measurements were carried out through a TCSPC method on an Edinburgh FLS1000 spectrofluorometer with 375 nm pulsed laser excitation sources. The average PL lifetime (τave.) is calculated according to Eq. (8):

$${\tau }_{{{{\rm{ave}}}}.}=\frac{{A}_{1}{\tau }_{1}^{2}+{A}_{2}{\tau }_{2}^{2}}{{A}_{1}{{\tau }_{1}+{A}_{2}\tau }_{2}}$$
(8)

Temperature-dependent PL spectra of NCs film were performed on FLS1000 with an Oxford Optistat attemperator cooled by liquid nitrogen. fs-TA spectroscopy was measured under a pulsed laser of 350 nm on HELIOS (Ultrafast systems) spectrometers. WAXS and X-ray absorption fine structure (XAFS) spectra of the samples were collected at the BL17B beamline station of Shanghai Synchrotron Radiation Center. WAXS is conducted in transmission mode.

Computational details

The geometric and electronic structures of the Au8(MPR)6 NCs were investigated using DFT with the Perdew-Burke-Ernzerhof (PBE) exchange-correlation functional80, as implemented in the ORCA package (version 6.0.1)81. All-electron 6-31 G(d) basis sets were employed for C, H, N, and S atoms, while the LANL2DZ pseudopotential basis set was used for Au. To analyze charge transfer characteristics, the HOMO and LUMO were visualized using Multiwfn (version 3.8)82,83, and rendered with VMD (version 1.9.3)84.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.