Optically excited metallic nanostructures may function as robust photocatalysts triggering plasmon-mediated interfacial molecule-transforming processes that are mechanistically distinct from the catalytic reactions under thermal conditions. [1][2][3][4][5][6][7][8][9][10][11][12][13][14][15][16] The plasmonic hot electrons, which are nonthermally distributed above the Fermi level of the metals, can be injected into unpopulated molecular orbitals of surface adsorbates to induce a series of intriguing bond-breaking and bond-forming photochemical reactions. [3][4][5][6][7][8][9] The photoexcitation efficiency, the lifetime of plasmonic electron oscillations, the energy distribution of hot electrons, and the electronic band energies of the coupled metal-adsorbate systems are all critical factors profoundly influencing the efficiency of hot electron injection into the molecular adsorbates. The photoexcitation, metal-to-adsorbate transfer, and thermal relaxation of hot electrons are all ultrafast photophysical processes on the femtosecond to nanosecond time-scales. 5,6,17 However, hot electron-driven photocatalysis, in many cases, involves drastically slower photochemical transformations spanning the time-scales from milliseconds to minutes, 5,6,10 implying that the rate-limiting kinetic bottle-necks are most likely associated with photo-induced chemical transformations of molecular adsorbates rather than photoexcitation and injection of hot electrons. In this work, we show that the plasmonic reactivity of molecular adsorbates is intimately tied to the chemical nature of metal-adsorbate interactions through a detailed case study of the reductive coupling of nitrophenyl derivatives chemisorbed to Ag nanoparticle surfaces.
In the ever-expanding family of plasmon-driven photoreactions, the reductive coupling of paranitrothiophenol (pNTP) has become a model reaction ideal for detailed mechanistic studies of plasmon-mediated surface chemistry. [18][19][20] Plasmon-enhanced Raman spectroscopies, including both surface-enhanced Raman scattering (SERS) and tip-enhanced Raman scattering (TERS), have been utilized as surface-sensitive spectroscopic tools to precisely characterize the detailed structural evolution of molecular adsorbates during plasmon-driven photocatalytic reactions. [20][21][22][23][24][25][26] As revealed by SERS-and TERS-based spectroscopic studies, chemisorbed pNTP molecules undergo plasmon-driven reductive coupling reactions to produce p,p′-dimercaptoazobenzene (DMAB) on the surfaces of a diverse range of metallic nanostructures, [18][19][20][27][28][29][30][31][32][33][34][35][36][37][38][39][40][41][42] providing a paradigm-shifting strategy for synthesizing aromatic azo compounds. However, the kinetic features and reaction pathways vary substantially from case to case, depending sensitively on the intrinsic properties of the metallic nanocatalysts (materials compositions, plasmon resonance frequencies, local-field enhancements, and surface structures), the photoexcitation conditions (excitation wavelength, excitation power density, and light illumination geometry), and the local environment in which the reactions occur (local temperature, pH, and presence of charge carrier acceptors). Several critical aspects of the detailed reaction mechanisms, such as the rate-limiting elementary steps, 29,33,38,40 the exact roles of plasmonic photothermal heating, 28,37,41 and the relationship between reaction rates and local-field enhancements, 27,38 are still under vibrant debate and remain open to further scrutiny. The seemingly divergent kinetic results and perplexing mechanistic characteristics reported in the literature strongly suggest that the plasmonic reactivity of pNTP adsorbates is not only related to the plasmon-derived photophysical effects but may also be modulated by the chemical interactions between the molecular adsorbates and the plasmonic photocatalysts. The metal-adsorbate interactions rigorously dictate the surface-orientation, structural ordering, and conformational flexibility of the chemisorbed molecules, all of which have been recently found to be key factors affecting the reactivity of thiolated nitrophenyl adsorbates toward the plasmon-driven reductive coupling reactions. 30,42,43 While the thiolated molecules represent so far the most intensively studied chemisorbates on metal surfaces, 44 organic molecules may also use other non-thiolated surface-anchoring groups, such as diazonium, [45][46][47][48][49][50] terminal alkyne, [51][52][53][54][55][56] N-heterocyclic carbene, [57][58][59][60][61][62] and isocyano groups, [63][64][65][66][67][68][69] to form covalent interactions with metallic substrates. Previous studies of plasmon-driven nitro-coupling reactions, however, have been exclusively focusing on thiolated nitrophenyl adsorbates. [18][19][20][21][22][23][24][25][27][28][29][30][31][32][33][34][35][36][37][38][39][40][41][42][43] How the non-thiolated nitrophenyl derivatives behave differently from their thiolated counterparts during the plasmon-driven coupling reactions remains unexplored yet. Here we systematically compare the plasmonic reactivity and transforming kinetics of pNTP, para-nitrophenylacetylene (pNPA), and para-nitrophenylisocyanide (pNPI), which chemisorb to Ag nanoparticle surfaces using the thiol, ethynyl, and isocyano groups, respectively. The chemical structures of the three nitrophenyl derivatives and the azo molecules produced through the plasmon-driven coupling reactions are shown in Scheme 1. Thiolated molecules are covalently bound to Ag surfaces through σ-donation of the lone-pair electrons of S to Ag, whereas the Ag-ethynyl interactions involve the contributions of both σ-donation and π-back donation. The σ-component is associated with the overlap between the conduction band of Ag and the π orbital of the ethynyl group, whereas the π-component arises from the electron back donation from the Ag valence band to the ethynyl π* orbital. [53][54][55] A particularly interesting discovery made in this work is that chemisorption of pNPI involves unique pH-dependent bimodal interactions between the isocyano group and the Ag surfaces, exhibiting the characteristics of the σ-donation mode in an acidic environment but becoming dominated by the π-back donation mode in an alkaline environment. Employing SERS as a molecular fingerprinting tool enables us not only to fully capture the essential spectral features reflecting the bonding nature of the Ag-adsorbate interactions, but also to precisely monitor the plasmon-mediated molecule-transforming processes in real time. The results of SERS-based kinetic measurements conducted under deliberately controlled reaction conditions provide important mechanistic insights into how the metal-adsorbate interactions modulate the plasmonic reactivity of the molecular adsorbates on nanostructured metal surfaces. In this work, the as-assembled Ag nanoparticle arrays played a dual role as both the SERS substrates and the plasmonic photocatalysts. SERS spectra were collected using a BaySpec Nomadic TM confocal Raman microscope built on an Olympus BX51 reflected optical system. Two continuous wave (cw) lasers with emission wavelengths of 532 and 785 nm were used as the excitation sources for both SERS and photocatalysis. The excitation lasers were focused onto a spot size of 2 μm in diameter in the focal plane using a 50× dark field objective (NA = 0.5, WD = 10.6 mm, Olympus LMPLFLN-BD), and the same objective was also used to collect the Raman signals in a back scattering configuration. The laser power focused onto the samples was adjusted either using neutral density filters or by changing the power output of the lasers. The integration time for spectral acquisition was typically in the range of 0.3-1 s. A homebuilt reaction chamber 71 was assembled on top of the ITO-supported Ag nanoparticle arrays, and was filled with an aqueous solution of either 5 mM H2SO4 (pH = 2) or 10 mM KOH (pH = 12) to control the pH of the reaction medium. The reaction progress was tracked in real time by collecting the temporally evolving SERS spectra under continuous laser illumination during the plasmon-driven coupling reactions.
Density functional theory (DFT) calculations were carried out using QChem 5.4 software package, with the functional of B3LYP applied in all simulations. The 6-31+G** basis set was used for all the non-metal atoms in the molecules, while the def2-ecp basis set was used for the Ag atoms and their effective core potential. More details of DFT calculations were described in the Supporting Information.
Ag nanoparticles served as both the adsorbents for molecular chemisorption and the building blocks for the assembly of nanoparticle arrays. The as-synthesized Ag nanoparticles were quasispherical in shape with particle diameters narrowly distributed around 44 nm (Figures S1A and S1B
in the Supporting Information). Colloidal Ag nanoparticles suspended in water displayed a welldefined light extinction peak centered at the wavelength of 415 nm (Figures S1C in the Supporting Information), which was the spectral signature of the dipole plasmon resonance of the nanoparticles.
The CTAC ligands on the surfaces of the as-synthesized Ag nanoparticles could be fully displaced by pNTP, pNPA, and pNPI through ligand exchange. Aqueous colloidal inks containing nitrophenyl derivative-coated Ag nanoparticles underwent a slow solvent evaporation process on hydrophobic substrates, such as ITO-coated glass slides, to form close-packed arrays with sub-10 nm interparticle gaps (Figure 1A). The strong plasmon coupling among the neighboring nanoparticles resulted in a broad-band plasmonic feature spanning the entire visible and much of the near-infrared in the extinction spectrum (Figure 1B). Optical excitations of the broad-band plasmons by a visible (532 nm, 2.33 eV) or a near-infrared (785 nm, 1. Careful analysis of the SERS spectral features enabled us to extract detailed information concerning the molecular structures of the adsorbates and the chemical nature of the Ag-adsorbate interactions. In Figure 1C, we compare the SERS spectra collected from pNTP adsorbates on the Ag nanoparticle arrays to the normal Raman spectrum of neat pNTP at an excitation wavelength (λex) of 785 nm. To eliminate the plasmon-driven coupling reactions, the SERS spectra were collected at a low excitation power (Pex) of 0.29 mW within a spectral integration time of 5 s. Several characteristic vibrational modes of pNTP, including the benzene ring mode (νCC ring, 1576 cm -1 ), the nitro stretching mode (νNO2, 1334 cm -1 ), the C-S stretching mode (νCS, 1076 cm -1 ), the nitro scissoring mode (βNO2, 856 cm -1 ), and the C-H bending mode (βCH, 1105 cm -1 ), were clearly resolved in the normal Raman spectrum. All these vibrational modes were also well-reproduced in the SERS spectra of chemisorbed pNTP. The structures of pNTP chemisorbed on Ag appeared pH-independent, giving rise to essentially the same SERS spectral features in acidic (pH = 2) and alkaline (pH = 12) environments.
The vCS mode in the SERS spectra remained unshifted in comparison to that in the normal Raman spectrum, indicating that the formation of Ag-S covalent bond introduced negligible modification of the length and order of the C-S bond in pNTP. As shown in Figure 1D, the SERS features of chemisorbed pNPA also matched the normal Raman features of neat pNPA very well, except that the C≡C stretching (vC≡C) mode in the SERS spectra was downshifted by ~113 cm -1 (from ~2105 cm -1 in normal Raman spectrum to ~1992 cm -1 in SERS spectra) and became significantly broadened in comparison to that in the normal Raman spectrum. The spectral downshift and lineshape broadening of the vC≡C mode were essentially caused by the π-back donation of electrons form the Ag valence band to the antibonding π* orbital of the terminal alkyne. The experimental results were further corroborated by DFT calculations, which not only confirmed the Raman peak assignments but also provided insights into the origin of the spectral shift of the vC≡C mode. We calculated the Raman spectra of pNTP-Ag4 and pNPA-Ag4 compounds, which were composed of an Ag4 atomic cluster covalently linked to pNTP and pNPA through Ag-S and Ag-C bonds, respectively. The chemical structures of pNTP-Ag4 and pNPA-Ag4 optimized by DFT are shown in Figure 1E and 1F, respectively. The calculated Raman shifts of all the major vibration modes were in very good agreement with the experimental results (Figures 1C and1D). According to the DFT results, the νC≡C mode of pNPA was downshifted by 106 cm -1 upon chemisorption to Ag due to the π-back donation effect associated with the Ag-ethynyl interactions.
pNTP and pNPA adsorbates exhibited strikingly different levels of reactivity toward the plasmondriven coupling reactions, as evidenced by the results of time-resolved SERS measurements. Figure 2 shows the time-resolved SERS spectra collected from pNTP and pNPA on the Ag nanoparticle arrays during the coupling reactions at pHs of 2 and 12 under continuous illuminations by a 532 nm laser at a Pex of 1.74 mW and by a 785 nm laser at a Pex of 2.64 mW. In the time-resolved SERS spectra, the intensities of the spectroscopic signals were shown as the detected counts normalized against the spectral integration time and the excitation laser power, carrying the unit of counts per second per milliwatts (cps mW -1 ). During the coupling reactions, the intensities of the νNO2 and βNO2 modes progressively decreased, while several SERS peaks signifying the azo products, such as the C-N stretching mode (νCN, 1146 cm -1 ) and the vibrational modes of the azo bond (νNN, 1438 and 1393 cm -1 ), emerged and became gradually more intense. Therefore, the reaction progress could be tracked in real time based on the temporal evolution of the SERS spectral features. We used DFT to calculate the Raman spectra of the azo products, p,p′-dimercaptoazobenzene (DMAB) for pNTP coupling and p,p′-diynylazobenzene (DYAB) for pNPA coupling reactions, respectively (Figure S3 in the Supporting Information). Comparison of the experimental spectra to the DFT results clearly showed that the DMAB and DYAB molecules produced through the plasmon-driven coupling reactions adopted the thermodynamically favored trans-conformation rather than the meta-stable cisconformation. Several interesting observations are particularly noteworthy. First, both pNTP and pNPA were observed to be substantially more reactive at λex of 532 nm than at λex of 785 nm. Our DFT calculations showed that the lowest unoccupied molecular orbital (LUMO) of pNTP chemisorbed to Ag were located 1.9 eV above the Fermi level of Ag (Figure S4 in the Supporting Information), in agreement with previously calculated results. 72 At the λex of 532 nm, a fraction of the photoexcited hot electrons were energetic enough to get injected into the LUMO of chemisorbed pNTP to initiate the reductive coupling reactions. At the λex of 785 nm, however, the hot electrons became energetically insufficient for the Ag-to-pNTP transfer, and the reaction mechanism switched to an alternative pathway mediated by the molecular oxygen dissolved in water. The hot electrons were injected into molecular oxygen
to produce anionic O2 -species capable of inducing the nitro-coupling reactions. 38 In this case, the molecular oxygen could be considered as a charge carrier-relaying co-catalyst triggering the multistep reductive coupling reactions. As shown by our previous work, 38 pNTP chemisorbed on Ag nanoparticle surfaces remained unreactive without any observable level of DMAB formation when illuminated by a 785 nm excitation laser in an anaerobic reaction environment. The chemisorption of pNPA to Ag resulted in energy upshifts of the frontier orbitals of the adsorbates, and the calculated LUMO energy of pNPA-Ag4 was 2.8 eV above the Ag Fermi level (Figure S5 in the Supporting Information). Therefore, even at the λex of 532 nm, the hot electrons were energetically insufficient to get injected into the LUMO of chemisorbed pNPA. The pNPA coupling reactions could occur only through the O2-mediated pathway at both excitation wavelengths. Our time-resolved SERS results revealed that the electron transfer from Ag to nitrophenyl adsorbates under visible light illumination was a kinetically more efficient reaction pathway in comparison to the O2-mediated pathway under near-infrared excitations. Second, pNPA was observed to be substantially less reactive than pNTP under identical reaction conditions, suggesting that the π-back donation bonding mode reduced the reactivity of the nitrophenyl derivative adsorbates. The formation of the azo products required specific orientational arrangements of nitrophenyl adsorbates on the Ag nanoparticle surfaces. 30,42,43 Therefore, the orientations and structural flexibility, both of which were directly related to the bonding nature of the metal-adsorbate interactions, became critical factors modulating the reactivity of molecular adsorbates. We recently found that the π-back donation effect in Ag-ethynyl interactions made para-ethynylaniline adsorbates more tilted toward the Ag surface with significantly reduced structural flexibility, which led to declined reactivity toward oxidative coupling reactions. 55 The SERS results shown in Figure 2 revealed that the same rule could also be applied to the reductive coupling of chemisorbed nitrophenyl derivatives. Third, at λex of 532 nm, pNTP appeared more reactive in an acidic environment than in an alkaline environment (Figure 2A) because this reductive coupling reaction involved protons. 39 However, elevated reactivity of chemisorbed pNPA was observed at λex of 532 nm when switching from an acidic to an alkaline environment (Figure 2B), indicating that the effects of metal-adsorbate interactions should also be taken into consideration when interpreting the pH-dependence of the molecular reactivity. After taking a closer look at the detailed spectral lineshapes of the vC≡C mode in the wavenumber range of 1930-2060 cm -1 (Figure S6 in the Supporting Information), we found that the asymmetrically broaden spectral features could be deconvoluted into at least three peaks centered at 1960, 1990, and 2025 cm -1 , corresponding to three subpopulations labeled as I, II, and III, respectively. Higher degree of π-back donation in the Agethynyl interactions led to larger spectral downshift of the vC≡C mode in the SERS spectra. The degree of π-back donation was highly likely to be related to the local surface curvature of Ag nanoparticles, which varied significantly from site to site. The chemisorbed pNPA with the lowest degree of π-back donation (subpopulation III) appeared to be the most reactive subpopulation, whereas the other two subpopulations remained drastically less reactive, further verifying the negative correlation between π-back donation and molecular reactivity. Within the time-duration of our SERS measurements, only a fraction of the chemisorbed pNPA was converted into DYAB. As the coupling reaction proceeded, the fraction of subpopulation III gradually decreased, while the spectral feature of the subpopulation I became increasingly more pronounced, indicating that the coupling of pNPA in the subpopulation III resulted in augmented π-back donation effect in the Ag-ethynyl interactions.
Through time-resolved SERS measurements, we systematically studied the kinetics of pNTP and pNPA coupling reactions in acidic (pH = 2) and alkaline (pH = 12) aqueous environments under continuous illumination by 532 nm and 758 nm lasers at various Pexs. The reaction progress was tracked in real time based on the temporal evolution of Q, which was defined as the relative intensity of the νNN mode (at 1438 cm -1 for DMAB and 1448 cm -1 for DYAB, respectively) with respect to the νCC ring mode in the SERS spectra:
(Equation 1).
In our case, the as-defined Q served as a descriptor of the apparent fractions of DMAB and DYAB during the pNTP and pNPA coupling reactions, respectively. Under each specific reaction condition, the SERS-based kinetic measurements were repeated at 5 different spots on the Ag nanoparticle arrays. Detailed kinetic results, including the temporally evolving Q trajectories collected from individual spots on the samples, the ensemble-averaged Q trajectories (<Q>) under each reaction condition, and the least squares curve-fitting results, are shown in Figures S7-S14 in the Supporting Information. In all cases, the kinetics of the bimolecular coupling reactions could be well-described by the following second-order rate law: (Equation 2), in which t is the reaction time, kobs is the apparent second-order rate constant, and Qmax is the maximal Q achievable when t approaches infinity. The values of Qmax and kobs under various reaction conditions were obtained by fitting the temporally evolving <Q> trajectories with Equation 2.
Figure 3 shows the Pex-dependent Qmax and kobs of the pNTP and pNPA coupling reactions under various reaction conditions. At the λex of 532 nm, Qmax increased monotonically as Pex increased in both acidic and alkaline aqueous reaction media (upper panels in Figure 3A and3B). Previous studies revealed that chemisorbed pNTP molecules could be activated for the reductive coupling reactions only when the local field intensities at the adsorbate-occupying sites exceeded a certain threshold value. 38 Increasing Pex led to larger Qmax values as higher fractions of the pNTP adsorbates started to experience the local-field intensities above the threshold value. Such Pex thresholds have also been previously observed on other plasmon-driven photoreactions, such as the oxidative coupling of aniline derivatives 55,73,74 and the decarboxylation of mercaptobenzobate. 75 The characteristic threshold Pex values for different reactions varied significantly and were intimately tied to the chemical nature of the reactions, the excitation wavelengths, the local-field enhancements, and the local reaction environments. In contrast to the Pex-dependence of Qmax, kobs appeared almost independent of Pex in both acidic and alkaline environments when Pex was varied in the range of 1-5 mW (lower panels in Figure 3A and3B). Although the coupling reactions were initiated by injection of hot electrons into the chemisorbed pNTP at the λex of 532 nm, the apparent second-order kinetic features and Pex-independent kobs values strongly suggested that the rate-limiting step was associated with the bimolecular coupling of photo-activated adsorbates on the Ag surfaces under thermal conditions. Under continuous illumination by the 532 nm laser, pNTP was considerably more reactive than pNPA toward the coupling reactions, as evidenced by the fact that both the Qmax and kobs of the pNTP coupling reactions exhibited higher values than those of the pNPA coupling reactions under each specific reaction condition. When switching the λex from 532 nm to 785 nm, a similar Pexdependence of Qmax (Qmax increased with Pex) was observed on both pNTP and pNPA coupling reactions, except that the Qmax values at λex of 785 nm were significantly lower than those at λex of 532 nm (upper panels in Figure 3C and3D). The Pex-dependence of kobs at λex of 785 nm (lower panels in Figure 3C and 3D) became substantially different from that at λex of 532 nm. We fitted the Pexdependence of kobs with the following power function:
(Equation 3), in which n is an exponent and α is a fractional coefficient. As shown in the lower panel of Figure 3C, the n value obtained from curve fitting for the pNTP coupling reactions at λex of 785 nm and pH of 2 was very close to 1. Although kobs appeared to be linearly proportional to Pex for the pNTP coupling reactions, pNPA showed limited reactivity featured by much smaller Qmax and kobs values within the same Pex range. The origin of the linear power dependence of kobs observed on the pNTP coupling reactions could be well-interpreted in the context of the O2-mediated reaction mechanism. At λex of 785 nm, the plasmonic hot electrons were injected into the π* orbital of surface-adsorbed O2 instead of the LUMO of surface-adsorbed pNTP. As shown in detail by our previous work, 74 the overall reaction kinetics were essentially modulated by the photoactivated O2 -radicals at its steady state concentrations, which exhibited a linear dependence on Pex. The linear Pex-dependence of kobs also indicated that under this reaction condition, the pNTP coupling reaction was driven by plasmonic hot electrons without any significant influence by photothermal heating. 9,15,76 Because of the small excitation volume of the confocal illumination and high thermal conductivity of water, the heat produced through photothermal transduction was rapidly dissipated to the aqueous reaction environment, resulting in rather insignificant elevation of the local temperatures on the photocatalyst surfaces. 75,77 In alkaline environments, 785 nm excitations gave rise to sublinear Pex-dependence of kobs for both pNTP and pNPA coupling reactions (lower panel of Figure 3D). Such sublinearity in power dependence is a common feature of semiconductor-driven photocatalytic reactions when the bulk electron-hole pair recombination becomes prevalent but has been rarely observed in plasmondriven photocatalysis on metallic nanostructures. 9 The plasmon-driven reductive coupling of nitrophenyl derivatives under near-infrared excitations involves not only the photoactivated O2 -but 𝑜𝑏𝑠 𝑒𝑥 𝑛 also protons. We hypothesized that the limited availability of protons in an alkaline reaction environment might become a kinetic bottleneck in the high Pex regime. The detailed mechanisms underpinning the sublinear power-dependence of kobs, however, still need to be further investigated.
Although higher kobs values were observed on pNPA than on pNTP at same Pexs (lower panel of Figure 3D), the Qmax values of the pNPA coupling reactions were substantially lower than those of the pNTP coupling reactions (upper panel of Figure 3D), suggesting that only a small fraction of surfaceadsorbed pNPA, probably the above-mentioned subpopulation III, remained reactive in an alkaline environment at λex of 785 nm.
The covalent interactions between the isocyano group of pNPI and the Ag nanoparticle surfaces were chemically more versatile than the Ag-thiol and Ag-ethynyl interactions. We found that pNPI molecules were chemisorbed to Ag surfaces through unique pH-dependent bimodal interactions, which were σ-bonding in nature in an acidic environment but became dominated by π-back donation in an alkaline environment. As shown in Figure 4A, the Raman shifts of various major vibrational modes, including the νCC ring, νNO2, βNO2, and βCH modes, in the SERS spectra of chemisorbed pNPI at pH of 2 matched those in the normal Raman spectrum of neat pNPI, except for the peak position of the stretching mode of N≡C bond (νNC). The νNC mode was upshifted from 2130 cm -1 in the normal Raman spectrum to 2186 cm -1 in the SERS spectrum collected at pH of 2 due to σ-donation of electrons from Ag to the isocyano group. In contrast, at pH of 12, the νNC mode in the SERS spectrum was downshifted to 1974 cm -1 and became significantly weakened and broadened due to π-back donation of electrons from the valence band of Ag to the antibonding π* orbital of the isocyano group.
Through DFT calculations, we identified two optimized structures of the pNPI-Ag4 complex, denoted as pNPI-Ag4-i and pNPI-Ag4-ii, respectively. pNPI-Ag4-i adopted an atop conformation with the C atom in the isocyano group forming a covalent σ-bond with an Ag atom (Figure 4B). The σ-donation of electrons from Ag to the isocyano group led to shortening of the N≡C bond. The upshift of the νNC mode in SERS with respect to that in the normal Raman spectra signified the σ-donation mode of the Ag-isocyano interactions, as exemplified not only by the chemisorbed pNPI studied in this work but also by other previously investigated isocyanyl derivatives, such as 2,6-dimethylphenyl isocyanide and 1,4-phenylene diisocyanide. [67][68][69] In contrast, pNPI-Ag4-ii adopted a bridge conformation with the C atom interacting with two Ag atoms simultaneously (Figure 4C). The LUMO energies of pNPI-Ag4-i and pNPI-Ag4-ii were calculated to be 1. (300 ms). At the λex of 532 nm, the Qmax values associated with the pNPI coupling reactions at the pH of 2 remained around 0.37, almost independent of Pex (Figure 5C). Such Pex-independent Qmax indicated that the Pex-threshold for photoactivation of chemisorbed pNPI was significantly lower than those of pNTP and pNPA, and all the reactive subpopulations of pNPI adsorbates were rapidly converted into DIAB in the Pex range we investigated. At the λex of 785 nm and the pH of 2, the Qmax values also appeared Pex-dependent (Figure 5D), whereas the kinetics of the pNPI coupling reactions became clearly resolvable by our SERS measurements. A nearly linear relationship between kobs and Pex was observed when Pex was varied in the range of 1-5 mW (Figure 5E). At the pH of 12, however, the coupling reactions became kinetically sluggish with very limited yields of DIAB achieved within the time-durations of our SERS measurements (Figure 5C and5D). The SERS results on the pNPI coupling reactions clearly showed that the π-back donation in Ag-isocyano interactions led to significantly decreased plasmonic reactivity of chemisorbed pNPI toward the reductive coupling reactions, fully in line with our observations on the pNTP and pNPA coupling reactions. The key findings reported in this work clearly indicate that the plasmonic reactivity and transforming kinetics of molecular adsorbates are not only related to the plasmonic properties of the nanostructured metallic adsorbents, but also rigorously dictated by the bonding nature of the metal-adsorbate interactions.