Synergy of Nanocrystalline Carbon Nitride with Cu Single Atom Catalyst Leads to Selective Photocatalytic Reduction of CO 2 to Methanol

Carbon nitride (C3N4) possesses both a band gap in the visible range and a low-lying conduction band potential, suitable for water splitting and CO2 reduction reactions (CO2RR). Yet, bulk C3N4...


Figure S2 .
Figure S2.TGA in air shows oxidation onset temperature increases for nc-C 3 N 4 (from 540 °C to 700 °C).

Figure S3 .
Figure S3.Powder XRD comparison of all materials: Cu/nc-C 3 N 4 (dark blue) and Cu/b-C 3 N 4 (light green) show no diffractions related to Cu and only have those pertinent to C 3 N 4 materials.nc-C 3 N 4 (dark green) and b-C 3 N 4 (purple) show diffractions (100) related to the in-plane tri-striazine and (002) related to the interlayer stacking.

Figure S4 .
Figure S4.TGA of nc-C 3 N 4 emulating reaction conditions (500 °C, 2hr) in air (black) and Ar (Red).It is observed (insert) that thermal treatment in air shows higher residual weight.Suggesting that some oxidation has occurred.

Figure S6 .
Figure S6.AC-STEM images from which particle analysis reveals single atoms (white circles) have a prevalence of around 70 % with the remainder of clusters not exceeding 3 atoms.

Figure S7 .
Figure S7.UV-Vis absorption spectra are shown for all materials to have an edge commencing at 450 nm, with a peak at 375 nm (black dotted line).Cu-containing materials show higher absorption intensities.

Figure S8 .
Figure S8.Tauc Plot analysis shows materials have a bang gap between 2.6 and 2.67 eV (dot/dash intercept extrapolation).

Figure S9 .
Figure S9.Linear region of flat band potential of b-C 3 N 4 (black dots).Intercept extrapolation (red line) shows a conduction band potential of -0.87 V vs Ag/AgCl.

Figure S10 .
Figure S10.Change in current density observed from photocurrent response measurements.

Figure S11 .
Figure S11.The stability of Cu/nc-C 3 N 4 for prolonged methanol production was tested by measuring the methanol concentration for 8 h (composed of 4 consecutives 2 h runs).A 77% decrease in activity is observed during the second cycle (70 μmol g cat -1 h -1 ) and further by 50% during the third cycle (34 μmol g cat -1 h -1 ).Methanol is only found in trace amounts upon the fourth cycle.These results suggest that the present Cu/nc-C 3 N 4 catalyst might have changed over time, or the methanol produced during the reaction may be consumed as a hole-scavenger by a long-term reaction.It strongly suggests that product separation is crucial as soon as it is produced during the reaction.

Figure S12 .
Figure S12.The AQY was obtained using Cu/nc-C 3 N 4 (red line) for CH 4 and CH 3 OH and the absorption profile of Cu/nc-C 3 N 4 (black dotted line).

Figure S13 .
Figure S13.Tri-s-triazine vacancy of C 3 N 4 with spacing corresponding to i) 0.26 nm; ii) 0.42 nm and iii) 0.23 nm.Cu(I) with radius 0.074 nm (red circle) is most likely to bind at sites i) assuming theoretical Cu-N bond length values of 0.196 nm. 1

Figure S14 .
Figure S14.Schematic representation of the Cu binding to nc-C 3 N 4 sites achieved through the increased crystallinity, and dangling bonds in these regions.

Figure S15 .
Figure S15.Schematic of custom-built Pyrex continuous flow photoreactor equipped with two mass flow controllers.Reactor volume: 28.5 mL, total pipeline volume 5.3 mL.

Figure S16 .
Figure S16.Control reaction under Ar (blue) conditions show trace amounts of methanol observed by NMR spectroscopy when compared to the catalyst under CO 2 saturation (green).

Figure S17. 1 H
Figure S17.1 H NMR spectroscopy analysis of methanol formed during CO 2 photoreduction across Cu/nc-C 3 N 4 and a control reaction under Ar across Cu/nc-C 3 N 4 which evidences the carbon source for methanol formation is CO 2 .

Table S1 :
N 1s XPS results for C 3 N 4 materials, including % peak area.

Table S3 :
C 1s XPS results for C 3 N 4 materials, including % peak area.