Size-dependent activity of carbon dots for photocatalytic H2 generation in combination with a molecular Ni cocatalyst

Carbon dots (CDs) are low-cost light-absorbers in photocatalytic multicomponent systems, but their wide size distribution has hampered rational design and the identification of the factors that lead to their best performance. To address this challenge, we report herein the use of gel filtration size exclusion chromatography to separate amorphous, graphitic, and graphitic N-doped CDs depending on their lateral size to study the effect of their size on photocatalytic H2 evolution with a DuBois-type Ni cocatalyst. Transmission electron microscopy and dynamic light scattering confirm the size-dependent separation of the CDs, whereas UV-vis and fluorescence spectroscopy of the more monodisperse fractions show a distinct response which computational modelling attributes to a complex interplay between CD size and optical properties. A size-dependent effect on the photocatalytic H2 evolution performance of the CDs in combination with a molecular Ni cocatalyst is demonstrated with a maximum activity at approximately 2–3 nm CD diameter. Overall, size separation leads to a two-fold increase in the specific photocatalytic activity for H2 evolution using the monodisperse CDs compared to the as synthesized polydisperse samples, highlighting the size-dependent effect on photocatalytic performance.

Electronic Supplementary Material (ESI) for Nanoscale.This journal is © The Royal Society of Chemistry 2023 Table S1.Summary of separated fractions and sizes of the GF-SEC separated CDs.

GF-SEC fractions
TEM particle size (nm) Average size for CDs as synthesized are: i) 6.4 ± 2.1 nm for a-CD, ii) 3.2 ± 1.1 nm for g-CD and 3.0 ± 1.1 nm for g-N-CD.A HiLoad® 26/600 Superdex® 200 pg prepacked column (Cytiva) was used to purify the CDs in 20 mM Borate buffer (pH 8), and the column was eluted at a flow rate of 2 mL min −1 .
Table S2.DLS and zeta potential measurements of the eluted size separated CDs.), tend to aggregate in the DLS cuvette while measuring since there is no stirring, so the measures should be taken as an idea of the general trend observed for the particle size decrease, not the real particle diameter.
Table S3.Singlet vertical electronic excitation energy, oscillator strengths (f), CI coefficients and primary character (|CI| > 0.2) based on the frontier molecular orbitals calculated for the 4×4-1L system.Only transitions with finite oscillator strengths are shown.Note: The powder XRD pattern for g-N-CD and g-CD show a broad but well-defined peak consistent with a nanocrystalline graphitic structure (graphitic core). For g-CD there is a well-defined peak (200) centered at 26.8° 2θ (lattice spacing of 3.29 Å) confirming the nanocrystalline graphitic core in g-CD.a-CD do not display this feature, in agreement with their predominant amorphous carbon core. [3]
Figure S1.(a) IR and b) UV-vis characterization of the synthesized a-CD, g-CD and g-N-CD in H2O at 298 K.

Figure S2 .
Figure S2.Powder X-Ray diffraction (XRD) patters of g-N-CD (black line), g-CD (blue line) and a-CD (red line) (a), and emission spectra of the synthesized (b) a-CD, (c) g-CD and (d) g-N-CD in H2O at 298 K at different excitation wavelengths (see legends).Note: The powder XRD pattern for g-N-CD and g-CD show a broad but well-defined peak consistent with a nanocrystalline graphitic structure (graphitic core).For g-N-CD the peak is centered at ca. 27.0° 2θ, which corresponds to the (002) reflection in graphitic structures and a lattice spacing of 3.25 Å (d002 = 3.35 Å in bulk graphite).[1][2]For g-CD there is a well-defined peak (200) centered at 26.8° 2θ (lattice spacing of 3.29 Å) confirming the nanocrystalline graphitic core in g-CD.a-CD do not display this feature, in agreement with their predominant amorphous carbon core.[3]

Figure S3 .
Figure S3.TEM images and the distribution of particle sizes of the synthesized (a) a-CD, (b) g-CD and (c) g-N-CD bulk materials before purification.

Figure S4 .Figure S5 .
Figure S4.GF-SEC traces of the separation of a-CD, g-CD and g-N-CD in (a) tris•HCl buffer pH 7.3 (100 mM) NaCl (100 mM) and (b) borate buffer pH 8 (20 mM) using a Superdex 200 pg Increase 10/300 GL SEC column.Broad GF traces with tris•HCl buffer pH 7.3 (100 mM) NaCl (100 mM) suggest interaction between the column matrix and the CDs.Note: According to SEC principles, a-CDs should elute earlier than g-CD and g-N-CD due to their bigger size.The elution times can therefore be explained by the stronger interactions between the a-CDs and the dextran gel matrix compared to the other CDs, even when using borate buffer as eluent.Importantly, these CD-gel interactions do not change for the same type of CDs with different sizes and we can still employ GF-SEC for size-separation.This is supported by our TEM analyses and confirms separation between 2 and 7.6 nm, within the calculated pore size of the resin.

Figure S6 .
Figure S6.TEM images and corresponding FFT pattern of the bulk g-CD (a) and the different size-separated fractions (b-f).

Figure S7 .S10Figure S8 .
Figure S7.Scale-up GF-SEC traces of the separation of a-CD in borate buffer pH 8 (20 mM) using a Superdex 200 pg Hi Load 26/600 GL (a), TEM images (b), UV-vis spectra (c), and emission spectra (lex = 405 nm) (d) of the separated fractions (0.14 mg•mL -1 concentration).Color codes correspond to the color band associated to each fraction.Red lines in the TEM images show an individual particle.

Figure S9 .
Figure S9.Scale-up GF-SEC traces of the separation of g-N-CD in borate buffer pH 8 (20 mM) using a Superdex 200 pg Hi Load 26/600 GL (a), TEM images (b), UV-vis spectra (c), and emission spectra (lex = 405 nm) (d) of the separated fractions (0.14 mg•mL -1 concentration).Color codes correspond to the color band associated to each fraction.

Figure S10 .
Figure S10.TEM images and FFT pattern of the bulk g-N-CD (a) and the different size-separated fractions (b-h).

Figure S11 .
Figure S11.Absorption peak energy as function of the CDs size for (a), a-CD (b), g-CD and (c) g-N-CD.

Figure S12 .
Figure S12.Excitation (lem = 460 nm) and emission spectra (lex = 360 nm) with a magnified region of the emission spectra of the separated fractions (a) a-CD, (b) g-CD and (c) g-N-CD.

Figure S13 .
Figure S13.Photoluminescent (PL) peak energy shift as function of the CDs size at 360 and 405 nm excitation for (a), a-CD (b), g-CD and (c) g-N-CD.

Figure S14 .
Figure S14.FTIR spectra of the bulk CDs (top spectra of a, b and c) and of some of the size separated fractions (from bigger to smaller) for (a), a-CD (b), g-CD and (c) g-N-CD.

Figure S20 .
Figure S20.Simulated absorption spectra of graphene systems with varying sizes and two atomic layers.

Figure S21 .
Figure S21.Simulated absorption spectra of the systems containing 2×2 (a) and 3×3 (b) graphene sheets with different number of layers.

Figure S22 .
Figure S22.Side view representation of the isosurfaces (isovalue = 0.02 a.u.) of the LUMO of the 2×2 system with different number of layers.The calculated HOMO-LUMO gap (∆EHOMO-LUMO) is also provided.

Figure S23 .
Figure S23.Side view representation of the isosurfaces (isovalue = 0.02 a.u.) of the LUMO of the 3×3 system with different number of layers.The calculated HOMO-LUMO gap (∆EHOMO-LUMO) is also provided.

Figure S24 .
Figure S24.Side view representation of the isosurfaces (isovalue = 0.02 a.u.) of the LUMO of the 4×4 system with different number of layers.The calculated HOMO-LUMO gap (∆EHOMO-LUMO) is also provided.

Figure S25 .
Figure S25.(a)H2 evolution studies with the optimum size for each type of CD and control studies without NiP, using NiP without CDs, and using NiCl2 as control.(b) Experiments with different concentration of EDTA for the selected best and worst performing CDs.Catalytic conditions: a-CD, g-CD and g-N-CD (0.5 mg), NiP (50 nmol), NiCl2 (30 nmol) and EDTA (0.1 M, pH 6, unless otherwise indicated) in water irradiation at 405 ± 10 nm for 24 h, at 25 ºC under N2 atmosphere.

Table S4 .
Singlet vertical electronic excitation energy, oscillator strengths (f), CI coefficients and primary character (|CI| > 0.2) based on the frontier molecular orbitals calculated for the 4×4-4L system.Only transitions with finite oscillator strengths are shown.