On-Surface Azide-Alkyne Cycloaddition Reaction: Does It Click with Ruthenium Catalysts?

Abstract

Owing to its simplicity, selectivity, high yield, and the absence of byproducts, the "click" azide-alkyne reaction is widely used in many areas. The reaction is usually catalyzed by copper(I), which selectively produces the 1,4-disubstituted 1,2,3-triazole regioisomer. Ruthenium-based catalysts were later developed to selectively produce the opposite regioselectivity─the 1,5-disubstituted 1,2,3-triazole isomer. Ruthenium-based catalysis, however, remains only tested for click reactions in solution, and the suitability of ruthenium catalysts for surface-based click reactions remains unknown. Also unknown are the electrical properties of the 1,4- and 1,5-regioisomers, and to measure them, both isomers need to be assembled on the electrode surface. Here, we test whether ruthenium catalysts can be used to catalyze surface azide-alkyne reactions to produce 1,5-disubstituted 1,2,3-triazole, and compare their electrochemical properties, in terms of surface coverages and electron transfer kinetics, to those of the compound formed by copper catalysis, 1,4-disubstituted 1,2,3-triazole isomer. Results show that ruthenium(II) complexes catalyze the click reaction on surfaces yielding the 1,5-disubstituted isomer, but the rate of the reaction is remarkably slower than that of the copper-catalyzed reaction, and this is related to the size of the catalyst involved as an intermediate in the reaction. The electron transfer rate constant (k_et) for the ruthenium-catalyzed reaction is 30% of that measured for the copper-catalyzed 1,4-isomer. The lower conductivity of the 1,5-isomer is confirmed by performing nonequilibrium Green's function computations on relevant model systems. These findings demonstrate the feasibility of ruthenium-based catalysis of surface click reactions and point toward an electrical method for detecting the isomers of click reactions.

Figure 1

Schematic of the SAMs studied. Oxide-free silicon (Si–H) electrodes are reacted with 1,8-nonadiyne via a hydrosilylation reaction to form SAM S-1. A ferrocene moiety is attached to the distal end of the monolayer by (a) CuAAC reaction to yield the redox-active SAM S-2 and (b) RuAAC reaction to yield the redox-active SAM S-3.

Figure 2

Most stable conformations of (a, b) 1,4-isomer and (c, d) 1,5-isomer on silicon surface. (a, c) Sparsely substituted Si surface. (b, d) Conformations of the isomers on a densely substituted Si surface. See the Methods section for computational details.

Figure 3

Static image of a water droplet on (a) SAM S-2, which was formed by a CuAAC reaction; (b) SAM S-3, which was formed by a RuAAC reaction; and (c) SAM S-1 before the click reaction. (d) Water contact angles for SAM S-1, S-2, and S-3. The error bars in (d) are the standard deviation of water contact angles from the mean value of three different surfaces.

Figure 4

AFM topography images of (a) SAM S-2, which was formed by CuAAC reaction, and (b) SAM S-3, which was formed by RuAAC reaction. The insets in (a) and (b) show cross-sectional profile (line) roughness.

Figure 5

(a) XPS survey spectra of the monolayer of S-2 SAM. XPS high-resolution spectra for S-2 SAM for (b) Fe 2p, (c) N 1s, (d) Si 2p, and (e) C 1s. (f) XPS survey spectra of the monolayer of S-3 SAM with a reaction time of 24 h. XPS high-resolution spectra of S-3 SAM for (g) Fe 2p, (h) N 1s, (i) Si 2p, and (j) C 1s for which the reaction time was 24 h.

Figure 6

Electrochemical characterization of SAM S-2 and S-3, which were formed by CuAAC and RuAAC reactions, respectively. CVs for (a) a click reaction without any catalyst; (b) SAM S-2, which was formed by the CuAAC reaction; and (c) SAM S-3, which was formed by the RuAAC reaction at the scan rate of 0.1 V/s. (d) Corresponding surface coverages calculated from the oxidation waves of the CVs in (b) and (c). The error bars in (d) are the standard deviation of the surface coverages obtained from the mean value of three different surfaces.

Figure 7

Nyquist plot from EIS measurement for (a) SAM S-2 which was formed by CuAAC reaction, and (b) SAM S-3, which was formed by RuAAC reaction. (c) Bode plots for CuAAC and RuAAC reactions with a frequency range of 4–60,000 Hz. Scattered dots (black and blue) are experimental data and lines (red and magenta) that are best fit to the experimental data. (d) Evolution of k_et obtained by fitting the impedance data to a Randles circuit (Table S3, Supporting Information). The error bars in (a) are the standard deviation of k_et values obtained from the mean value of three different surfaces.

Figure 8

(a) Transmission spectra at zero-bias and (b) current–voltage characteristics of (c) 1,4- and (d) 1,5-isomers of 1,2,3-triazole model systems computed between two gold nanowires. Transmission spectra suggest that the main active orbitals in the conductance are the highest occupied molecular orbital (HOMO)-type orbitals that have energies lower than the Fermi level. The conductivity of the model systems is computed in a range of −1 to +1 V, and in the full range, the 1,4-isomer (red trace) remains more conductive than the 1,5-isomer (blue trace). The conductivity of the 1,4-isomer in the range of −0.1 to +0.1 V is computed to be 15 μS and that of the 1,5-isomer in the same range is predicted to be about 3.1–4.8 μS. Thus, the 1,5-isomer’s conductivity is ∼30% of that of the 1,4-isomer in the aforementioned voltage range.