Zakhele B. Ndala*abc,
Ndivhuwo P. Shumbulad,
Seiso E. Tsoeua,
Thelma W. Majolaa,
Siziwe S. Gqobaa,
Cebisa E. Linganisoabc,
Zikhona N. Tetanae and
Nosipho Moloto
*a
aMolecular Science Institute, School of Chemistry, University of the Witwatersrand, Private Bag 3, Wits 2050, South Africa. E-mail: zakhele.ndala@wits.ac.za; Nosipho.Moloto@wits.ac.za; Tel: +2711 717 1339
bDST/NRF Centre of Excellence in Strong Materials, University of the Witwatersrand, Private Bag 3, Wits 2050, South Africa
cMicroscopy and Microanalysis Unit, University of the Witwatersrand, Private Bag 3, Wits 2050, Johannesburg, South Africa
dHealth Platform, Advanced Materials Division, Mintek, 200 Malibongwe Drive, Randburg, South Africa
eInstitute for Nanotechnology and Water Sustainability, College of Science, Engineering and Technology, University of South Africa, Florida Campus, Johannesburg 1709, South Africa
First published on 2nd May 2025
The synthesis of heterostructures composed of transition metal dichalcogenides (TMDs) and carbon nanostructures has garnered a lot of attention in recent years. This is due to the synergistic effects that arise from these heterostructures that are advantageous in various applications. This includes but is not limited to the improvement in electron conductivity of TMDs that are grown on carbon nanostructures. This improvement in electron conductivity can increase the catalytic activity of TMDs towards the hydrogen evolution reaction (HER). Therefore, it is crucial to understand the formation of these heterostructures and how the interaction of the component materials can improve their performance as electrocatalysts in the HER. This study highlights how surface chemistry affects heterostructure formation and the catalytic performance of heterostructures in the HER. ReSe2 nanocrystals were grown on 2D carbon nanostructures, specifically reduced graphene oxide (rGO), nitrogen doped reduced graphene oxide (N-rGO), and graphitic carbon nitride (g-C3N4). FTIR, XPS, and TEM analyses showed that functional groups on carbon surfaces play a key role in the formation of the heterostructures. Among the materials tested, rGO had the highest ReSe2 loading due to the availability of oxygen containing functional groups on the surface of rGO. However, the performance of the heterostructures as catalysts in the HER showed that ReSe2-N-rGO had the highest catalytic activity with the lowest onset potential (115 mV), Tafel slope (72 mV dec−1), and overpotential (218 mV). The enhanced performance of the ReSe2-N-rGO catalyst was due to the modulation of rGO by nitrogen doping which improved the electron transfer between ReSe2 and N-rGO, this was further confirmed using computational studies and by ReSe2-N-rGO having the lowest Rct (65 Ω).
ERHE = EAg/AgCl + 00059 pH + E0Ag/AgCl | (1) |
Eb = (EGO/ReSe2 − EGO − EReSe2)/S | (2) |
Eb = (EGO-graphiticN/ReSe2 − EGO-graphiticN − EReSe2)/S | (3) |
The (002) peak is observed for both the rGO and N-rGO at 26.6° and 26.8° respectively (JCPDS card no: 75-2078). This peak is indicates a high degree of graphitization, which alludes to the formation of a well-arranged layer structure.27 The appearance of the (100), (004), and (110) peaks at 43.8°, 54.7°, and 77.8° degrees respectively indicates that the rGO nanosheets are not monolayer.27,28 There is an observable shift to the right in the (002) peak in the N-rGO sample as shown in Fig. 1(b), this has been attributed to the doping of nitrogen in the graphene oxide lattice structure.28 Sherrer's equation can be used to determine crystallite size along the (002) plane which can be used as an estimate for the thickness of the rGO nanosheets, the thickness was determines to be 16.34 nm for rGO and 14.4 nm for N-rGO. The thickness is reduced in N-rGO because of the removal of oxygen containing functional groups during the doping process. The PXRD pattern of g-C3N4 is shown in Fig. 1(c), the characteristic diffraction peaks of g-C3N4 were observed (JCPDS: 87-1526).29 The main peak at 2θ = 27.8° corresponds to the (002) plane, which is ascribed to the distance between the layers of the graphitic material with an interlayer spacing of 0.323 nm. The peak at 2θ = 13.7° corresponds to the (001) diffraction plane, this peak is attributed to the intra-layer d-spacing (0.651 nm). The g-C3N4 synthesized through poly-condensation is typically considered to have low crystallinity, as is the case in this study. This is due to the incomplete deamination of the precursors which results in the formation of a tri-s-triazine-based melon structure which consists of in-plane infinite one-dimensional chains of NH-bridged melem oligomers. The diffractogram for the graphitic carbon nitride synthesized in this study is similar to that of bulk g-C3N4 also signifying the formation of the tri-s-triazine melon structure.29 The powder patterns of the ReSe2 and carbon heterostructures is shown in Fig. 1(d). The diffraction peaks as shown in Fig. 1(d) are observed at 13.7°, 31.9°, 35.2°, 42.6°, 47.1°, and 55.2°. These diffraction peaks correspond to the lattice planes (001), (1−22), (−220), (012), (003) and (1−42) of ReSe2. The broad diffraction peak between 2θ = 30–40° is attributed to the overlap of smaller peaks in that region, this has been observed in nanosized ReSe2 materials.21 The PXRD confirms that the ReSe2 nanostructures have crystallized in the distorted 1T phase triclinic system with the P space group (JCPDS card no: 01-065-4275). The (002) peak of the rGO, N-rGO, and g-C3N4 can still be observed in the powder patterns of the heterostructures which alludes to the incorporation of the ReSe2 nanostructures on the 2D carbon nanostructures. Raman spectroscopy was also used to characterize the rGO and N-rGO. This was done by analyzing the characteristic D, G, and 2D bands of the nanostructures. The Raman spectra of the rGO and N-rGO are shown in Fig. 1(e). The position of the G-band in rGO and N-rGO was observed at ∼1588 cm−1 and 1591 cm−1 respectively, this is in accordance with what has been previously reported in the literature.30 The G-band is attributed to the in-plane stretching vibrations of the sp2 hybridized carbon–carbon bonds. The D band is associated with the defects that exist in the structure of the materials. The D-bands of the rGO and N-rGO were observed at 1316 cm−1 and 1307 cm−1 respectively. The broad and high-intensity D band observed in both the undoped rGO and N doped rGO signifies a high level of lattice distortion and sp3 like defects which show that there are functional groups on the surface of the rGO and N-rGO.27 The rGO was further reduced in the doping process which occurred at 800 °C. The ID/IG ratio of the N-rGO was found to be 2.09 which was higher than that of the rGO at 1.86, this indicates that there is an increase in the number of defects in the structure of the N-rGO which is due to the nitrogen-doping.31 The 2D peak occurs due to two lattice vibrational processes, this peak can be used to assess the quality of graphene because it is very sensitive to disorder. It is normally very prominent in highly crystalline high-quality graphene. The peak experiences a decrease in intensity when there are defects in the structure of the materials.30 The notable decrease in the intensity of the 2D peak of the N-rGO compared to the rGO is indicative of the introduction of more disorder in the structure through the introduction of nitrogen defects.
X-ray photoelectron spectroscopy (XPS) was used to evaluate the bonding configuration and the elemental composition of the nanostructures. The XPS spectra of rGO and N-rGO are shown in Fig. S1(a) and (b).† The C1s and O1s peaks were detected in the survey scan of the rGO nanostructures. The survey scan of the N-rGO nanomaterials also shows the N1s peak which confirms the doping of the rGO with nitrogen.32 The elemental composition of rGO and N-rGO is shown in Table 1.
Sample | Element | Binding energy (eV) | Atomic (%) |
---|---|---|---|
rGO | C | 284.9 | 86.4 |
O | 532.8 | 13.6 | |
N | — | — | |
N-rGO | C | 284.4 | 97.5 |
O | 532.7 | 1.6 | |
N | 399.1 | 0.9 |
The elemental composition shows that the percentage of oxygen is much larger in rGO compared to the N-rGO. This shows that the rGO contains more of oxygen-containing functional groups and that some of these functional groups are removed in the doping process. The elemental composition of the N-rGO only shows a nitrogen doping concentration of 0.9%. The C1s spectra of the rGO and N-rGO are shown in Fig. S2(a) and (b)† and a summary of the atomic composition is shown in Table 2.
Sample | Assignment | Binding energy (eV) | Atomic (%) |
---|---|---|---|
rGO | C–C sp2 | 284.2 | 35.5 |
C–C sp3 | 284.9 | 23.0 | |
C–O | 285.8 | 20.3 | |
C![]() |
533.5 | 7.1 | |
O–C![]() |
288.7 | 6.8 | |
N-rGO | C–C sp2 | 284.4 | 69.7 |
C–C sp3 | 532.7 | 15.6 | |
C–O | — | — | |
C![]() |
— | — | |
O–C![]() |
399.1 | 0.9 |
The peak binding energies were found to be 284.9 eV, 285.8 eV, and 288.7 for C–C, C–O, and O–CO functional groups respectively for rGO.33 The peak binding energies were found to be 284.4 eV and 285.7 for C–C and C–O functional groups respectively for N-rGO.33 Table 2 shows there is a notable decrease in the amount of oxygen-containing functional groups in N-rGO, this is due to the removal of the functional groups during doping. The N1s spectrum of N-rGO is shown in Fig. 2(a), the N1s peak can be deconvoluted into several individual peaks, namely the pyridinic-N (398.1–399.3 eV), pyrrolic-N (399.8–401.2 eV), and graphitic-N (401.1–402.7 eV).33,34 These characteristic peaks are observed in the N1s XPS spectrum and are shown in Fig. 2(a), the nitrogen-doped graphene oxide seems to be dominated by pyridinic-N sites followed by pyrrolic-N sites. The graphitic-N sites are the least prevalent, while the peak at 405 eV is attributed to chemisorbed nitrogen oxide. The types of nitrogen defects introduced are shown graphically in Fig. 2(b). The XPS survey spectra of g-C3N4 is shown in Fig. S1(c).† The high-resolution N1s and C1s spectra are shown in Fig. 2(c) and (d) respectively. The N1s spectrum shows the 3 characteristic peaks of g-C3N4.35 The main peak at 398.0 eV is due to sp2 hybridized nitrogen in the triazine ring (C–C
N) (N1), the peak at 399.8 eV was ascribed to the tertiary nitrogen in the heptazine unit (N2), the peak at 401.4 eV was ascribed to the amino groups carrying hydrogen (N3), the peak at 403.8 eV was ascribed to the NOx species (N4).35 The C1s spectra shows the three characteristic peaks. The peak at 284.4 eV has been shown to be due to the C–C (C1) bonds in graphitic carbon while the peak at 288.4 eV was ascribed to the sp2 bonded carbon (N–C
N) of the triazine and heptazine rings (C2).35 The peak at 285.8 can be attributed to a nitrile species –C
N (C3). The peak at 288.5 eV is due to the C–O bonding which suggest that there exist some oxygen groups on the surface of the graphene sheets (C4), this is confirmed in the XPS survey spectrum as shown in Fig. S1(c).† The XPS survey spectra of pristine ReSe2 is shown in Fig. S1(d).† The XPS spectra of the Re4f and Se3d are shown in Fig. 2(e) and (f) respectively.
The high-resolution XPS spectrum of Re4+ is shown in Fig. 2(e), the core level peaks 4f7/2 and 4f5/2 were measured at 41.2 eV and 43.6 eV respectively.21 The Se high-resolution XPS spectrum of the ReSe2 nanomaterials is shown in Fig. 2(f). The deconvoluted peak at ∼54 eV shows the characteristic core level 3d5/2 and 3d3/2 peaks Se2− at 53.7 eV and 54.7 eV respectively.21 The XPS survey spectra of the heterostructures is shown in Fig. S3(a)–(c),† the high-resolution spectra of the Re4+ and Se2− is also shown in Fig. S4(a)–(f).† The XPS spectra of the analysis of the O1s spectra was used to investigate possible interactions between the ReSe2 and the carbon nanostructures. Table 3 shows a summary of the binding energy of the O1s spectra of the rGO, N-rGO, and the heterostructures, the O1s spectra is shown in Fig. S5(a)–(d).†
Sample | Assignment | Binding energy (eV) |
---|---|---|
rGO | C–O | 532.0 |
C![]() |
533.5 | |
N-rGO | C–O | 531.2 |
C![]() |
532.9 | |
ReSe2-rGO | C–O | 531.7 |
C![]() |
533.2 | |
ReSe2-N-rGO | C–O | 532.2 |
C![]() |
533.5 |
Table 3 shows that the binding energy of the O1s spectrum of ReSe2-rGO shifted by ∼0.3 eV from the pristine rGO. The same phenomenon was observed for the ReSe2-N-rGO, a noticeable shift in the binding of the assigned peaks. This indicates that the chemical environment rGO functional group had been altered due to the growth of the ReSe2 nanostructures on the surface of the rGO and N-rGO. Table S1† shows a summary of the binding energy of Re4f7/2. A shift in the binding energy of 0.3 eV and 0.5 eV in the Re4f7/2 for ReSe2-rGO and ReSe2-N-rGO further indicates an interaction between the oxygen functional groups in rGO and N-rGO. The Re4+ XPS spectra are shown in Fig. S4(b) and (d).† The high-resolution N1s and C1s spectrum of ReSe2-g-C3N4 are shown in Fig. S4(g) and (h).† The characteristic peaks of carbon nitride are observed in the N1s and C1s spectra of the heterostructure. However, there is an approximately 0.5 eV shift in the characteristic N1s peaks of the ReSe2-g-C3N4 compared to the g-C3N4. This shows a change in the chemical environment of the nitrogen groups in the carbon nitride structure. This indicates that there is a chemical interaction between the nitrogen groups in the carbon nitride cavities and the ReSe2 nanostructures. The C1s spectrum of the heterostructures shows a slight shift and repression of the C2 peak which confirms that the ReSe2 nanostructures interact with the nitrogen groups in the heptazine and triazine structure of the carbon nitride. This was further explored using FTIR. The FTIR spectra of the g-C3N4 and ReSe2-g-C3N4 are shown in Fig. S6.† The typical characteristic peaks of g-C3N4 observed at 1318 cm−1, 1393 cm−1, 1534 cm−1, and 1621 cm−1 were assigned to the stretching modes of the aromatic C–N heterocycle and the peak at 805 cm−1 was assigned to the breathing mode of the triazine units. The peak assigned to the breathing mode of the triazine units was observed to have shifted to 797 cm−1 in the heterostructure and the peak was also significantly supressed compared to the pristine g-C3N4. This suggests an interaction between the nitrogen groups in the triazine units and the ReSe2 nanostructures as suggested by the XPS analysis.
The TEM images of the rGO, N-rGO, g-C3N4 the 2D carbon–ReSe2 heterostructures are shown in Fig. 3.
![]() | ||
Fig. 3 TEM images of (a) ReSe2 nanoflowers, (b) rGO, (c) N-rGO, (d) g-C3N4, (e) ReSe2-rGO, (f) ReSe2-N-rGO, and (g) ReSe2-g-C3N4. |
The TEM image of the ReSe2 nanostructures is shown in Fig. 3(a), the image shows that the ReSe2 nanostructures form in a flower-like morphology. The TEM images of rGO and N-rGO are shown in Fig. 3(b) and (c) show that they are both composed of nanosheets that are wrinkled and folded in some regions. The nanosheets are not single sheets, as was already confirmed by the PXRD analysis. There are no significant and notable changes in the morphology of the rGO before and after nitrogen doping as shown in Fig. 3(b) and (c). Fig. 3(d) shows the TEM image of g-C3N4, the image shows that the g-C3N4 nanosheets that show some degree of folding in some regions, similar to the rGO and N-rGO morphology. The images of ReSe2-rGO and ReSe2-N-rGO are shown in Fig. 3(e) and (f), respectively. The images show that few-layered ReSe2 nanostructures have grown on the surface of the rGO and N-rGO. For the ReSe2-rGO, the ReSe2 nanostructures seem to have completely covered the area of the rGO nanosheets. The same is true for the ReSe2-N-rGO, with some of the areas showing more defined nanoflower structures. The HRTEM image of the ReSe2-g-C3N4 in Fig. 3(g) shows that ultrathin ReSe2 nanosheets were able to form on the surface of the g-C3N4 nanosheets.
An initial study was conducted to determine the interaction of the Re monomer with the 2D carbon nanostructures, this study was conducted using rGO as the 2D carbon nanostructure of interest. The reaction for the synthesis of ReSe2 was conducted without selenization (introduction of selenium precursor), this was done to specifically determine the interaction of the Re monomer with the rGO nanosheets. The Re monomer – 2D carbon material was then selenized to observe the growth of the ReSe2 nanostructures on the surface of the rGO. A schematic representation of the process before and after selenizations is shown in Fig. 4.
Previous reports have shown that the ReO42− anion is formed when NH2ReO4 is heated at high temperature.36 The ReO42− is then able to interact with the functional groups on the surface of the rGO, this is in line with what is reported in literature.36 The TEM image in Fig. 5(a), shows that after the reaction with NH4ReO4 the sheet like structure of the rGO does not change. The powder pattern of the rGO-ReO42− shows no peaks from anything other than rGO as shown in Fig. S7(a),† this suggests that the NH4ReO4 did not convert to any crystalline oxide species such as ReO3. However, EDS of the rGO-ReO42− (Fig. S8†) shows that Re is indeed present which suggests the presence of the ReO42− anion. The UV-vis spectrum of the rGO-ReO42− (Fig. S7(b)†) shows that characteristic peak for ReO42− at 291 nm, which confirms the presence of ReO42−.37 Elemental mapping of rGO-ReO42− is shown in Fig. 5(b)–(d). The uniform distribution of the O and Re atoms on the surface of the rGO confirms the interaction of ReO42− with the oxygen functional groups on the surface of the rGO.
![]() | ||
Fig. 5 (a) Sheetlike structure of rGO after introduction of Re monomer. (b) Grey image of rGO-ReO42−. Elemental mapping of (c) oxygen and (d) rhenium. |
After selenization (when the selenium precursor in introduced), the formation of the ReSe2 nanostructures can clearly be observed on the surface of the rGO nanostructures. This is shown in Fig. 6(a). An interface can be formed between the ReSe2 nanostructures and the rGO can be clearly observed in Fig. 6(b). A similar phenomenon is observed for the N-rGO 2D carbon nanostructures, after selenization the ReSe2 would grow on the surface of the N-rGO as shown in Fig. 6(c). When g-C3N4 was used the type of interaction would involve the ReO42− would interact with the nitrogen groups in the triazine ring as alluded to by the XPS and FTIR analysis. After selenization the ReSe2 nanostructures were able to grow on the on the g-C3N4 nanostructures as shown in Fig. 6(d). The growth of the ReSe2 nanostructures seems to distort the structure of g-C3N4. The well-defined g-C3N4 nanosheets seem to form a distorted flocculate-like morphology with ReSe2 nanosheets that have grown on the g-C3N4.
Scanning electron microscopy (SEM) was also performed on the heterostructures. Fig. 7(a), (b), and (c) shows the ReSe2-rGO, ReSe2-N-rGO, and ReSe2-g-C3N4 heterostructures respectively. The images clearly show that the ReSe2 nanocrystals have grown on the surface of the carbon nanostructures.
The TEM image in Fig. 3(a) already showed that the ReSe2 nanostructures had a pseudo-spherical flower-like morphology. The SEM images show that the flower-like ReSe2 nanostructures have grown on the different 2D carbon nanostructures. The SEM image of the ReSe2-rGO (Fig. 7(a)) shows that the rGO has been completely covered by the ReSe2 nanosheets with agglomeration of the spherical flower-like ReSe2 being observed. This was not the case for the N-rGO (Fig. 7(b)) and g-C3N4 (Fig. 7(c)), there are clearly visible regions where the uncovered N-rGO and g-C3N4 can be observed, this is due to the readily available oxygen containing functional groups on the rGO that ensure maximal formation and growth of the ReSe2 nanostructures on its surface. Energy-dispersive X-ray spectroscopy (EDX) was used to investigate the elemental composition of the heterostructures and the distribution of the ReSe2 nanocrystals on the surface of the 2D carbon nanostructures. The EDX spectra of the heterostructures is shown in Fig. S9 (a), (b), and (c)† for the ReSe2-rGO, ReSe2-N-rGO, and ReSe2-g-C3N4 heterostructures respectively. The spectra show the presence of Re, Se, C, N, Ag, and Pd. Ag and Pd was used to coat the samples to prevent charging. The elemental mapping of the heterostructures is shown in Fig. S10(a)–(c).† The mapping shows that the Se atom distribution overlaps well with the Re mapping, this confirms the formation of ReSe2 nanocrystals on the surface of the 2D carbon nanostructures. The maps also show that the carbon nanostructures are covered by the ReSe2 nanocrystals with some areas of exposed carbon being observed. The maps also show the presence of oxygen from the surface functional groups.
Thermogravimetric analysis (TGA) was used to investigate the decomposition of the rGO, N-rGO, g-C3N4 and the heterostructures under an airflow of 20 ml min−1. The decomposition profiles are shown in Fig. 8.
![]() | ||
Fig. 8 Decomposition profiles of (a) rGO and N-rGO, (b) rGO and ReSe2-rGO, (c) N-rGO and ReSe2-N-rGO and (d), g-C3N4 and ReSe2-g-C3N4. |
The TGA profiles of the rGO and N-rGO are shown in Fig. 8(a). The initial weight losses on both profiles at ∼150 °C are attributed to the evaporation of water molecules adsorbed on the rGO and N-rGO.38 The second largest mass loss was measured at 6% for rGO and 3% for N-rGO which was attributed to the oxygen-containing functional groups.39 There is twice the loss of these functional groups in rGO which suggests that the doping process was able to remove functional groups on the surface of the N-rGO beforehand. The largest weight loss between 400–700 °C which accounted for 94% in rGO and 97% in N-rGO was attributed to the decomposition of the carbon skeleton of graphene.36 The decomposition temperature for this weight loss on N-rGO was slightly higher than that of rGO which was expected as doped carbon structures have been known to be more stable than undoped rGO structures.40 The difference in decomposition temperatures can be clearly seen in the derivative decomposition profile in Fig. S11.† The decomposition profile of g-C3N4 is shown in Fig. 8(d), the initial mass loss at ∼150 °C was due to the evaporation of water molecules on adsorbed on the g-C3N4. The large mass loss of 96% observed at 520 °C was attributed to the decomposition of the g-C3N4 to carbon and nitrogen containing gases. The TGA curves ReSe2-rGO and ReSe2-N-rGO are shown in Fig. 8(b) and (c), these are shown with the TGA curves of rGO and N-rGO for comparison. The heterostructures show two mass losses, one between 400–500 °C and another between 500 to 600 °C. The first mass loss was measured at 60% for rGO and 63% for N-rGO. The mass loss was attributed to the decomposition of the graphene backbone. The second mass loss was attributed to the decomposition of the oleic acid that is bound to the ReSe2 nanostructures which accounted for 37% and 36% of the ReSe2-rGO and ReSe2-N-rGO respectively. The increase in the temperature of the decomposition compared to literature was attributed to the bound oleic acid (oleic acid interacting with the ReSe2 nanostructures) as opposed to unbound oleic acid.39 The decomposition profile of ReSe2-g-C3N4 is shown in Fig. 8(d). The large mass loss accounting to 77% between 400 °C and 600 °C was attributed to the decomposition of g-C3N4. The decomposition of the bound oleic acid occurs between 500 °C and 600 °C and accounts for 20% of the mass loss. The final mass loss between 700 °C and 800 °C accounting for ∼3% was attributed to the rest of the g-C3N4 decomposition. The residues that remained after decomposition accounted for 2.7%, 1.7%, and 0.5% for ReSe2-rGO, ReSe2-N-rGO, and ReSe2-g-C3N4 respectively. This suggests that the rGO had a larger catalyst loading of the ReSe2 compared to the N-rGO and g-C3N4, this is due to the higher availability of the oxygen containing functional groups on the rGO. The ReSe2-g-C3N4 heterostructure had the lowest mass loading of ReSe2, this suggest that the nitrogen defects in the structure of the material may not be as effective at anchoring the ReSe2 compared to the oxygen containing functional groups.
The adsorption–desorption curves of all the nanostructures show a type-IV isotherm with a distinct hysteresis loop (IUPAC classification) as shown in Fig. S12† and 9, this suggests that the nanostructures assume a mesoporous structure.40
![]() | ||
Fig. 9 Adsorption–desorption isotherms of (a) ReSe2, (b) ReSe2-rGO, (c) ReSe2-N-rGO, and (d) ReSe2-g-C3N4 . |
The BET surface area of the rGO and N-rGO was calculated to be 231.3 m2 g−1 and 236.3 m2 g−1 as shown in Fig. S12(a) and (b).† The surface area of the doped and undoped area is virtually identical which suggests that there was no significant folding or wrinkling of the nanosheets during the doping process which would have reduced the surface area. The surface area of the g-C3N4 was measured to be 7.9 m2 g−1, the surface area is quite low as expected for g-C3N4 nanosheets. The BET surface area of the pristine ReSe2 nanoflowers was measured to be 16.6 m2 g−1 as shown in Fig. 9(a). The surface area of the ReSe2-rGO, ReSe2-N-rGO, and ReSe2-g-C3N4 was measured to be 36.2 m2 g−1, 42.2 m2 g−1, and 26.6 m2 g−1 respectively as shown in Fig. 9(b), (c), and (d). The boosted surface area of the pristine ReSe2 nanoflowers is due to the confined growth of vertically aligned nanosheets on the 2D carbon nanostructures.41
Several parameters are used to evaluate the catalytic activity of nanomaterials towards the HER. The polarization curves can be used to determine two useful parameters, which are the onset potential and the overpotential at 10 mA cm−2 (η10). The polarisation curves for the pristine ReSe2 and the heterostructures is shown in Fig. 10(a). The onset potential is the lowest potential at which the HER can be produced. A good HER catalyst is characterized by a low onset potential. The onset potential of the pristine ReSe2 nanostructures was determined to be 159 mV. The incorporation of the ReSe2 nanostructures on the carbon nanostructures significantly reduces the onset potential. The onset potential of the ReSe2-rGO, ReSe2-N-rGO, and ReSe2-g-C3N4 was measured to be 128 mV, 115 mV, and 135 mV respectively. Forming a heterostructure with the 2D carbon nanostructures seems to improve the catalytic activity. The overpotential at 10 mA cm−2 (η10) is a measure of the overpotential at a specified current density, 10 mA cm−2 is chosen because it represents the operating current density of a commercial cost-competitive photoelectrochemical cell (PEC). This parameter is useful when comparing the catalytic activity of nanomaterials, an ideal HER catalyst has the lowest possible η10. The η10 of the pristine ReSe2 nanostructures was recorded at 331 mV. The η10 of the ReSe2 heterostructures were recorded at 259 mV, 218 mV, and 358 mV for the ReSe2-rGO, ReSe2-N-rGO, and ReSe2-g-C3N4 respectively. Incorporating the ReSe2 nanostructures on rGO improves the overpotential which is attributed to the improved electron conductivity of the heterostructures. The doping of the rGO with nitrogen further improves the overpotential as expected. This is not the case when g-C3N4 was used, which was attributed to the unsuitable interaction between the ReSe2 and the g-C3N4. The Tafel plot analysis is crucial in evaluating the catalytic activity of nanomaterials towards the HER. The Tafel equation (η = b log J + a, where η is the overpotential, J is the current density and b is the Tafel slope) fitted to the Tafel plots to determine the Tafel slope.42 The Tafel slope can be used to gain insights into the mechanistic details of the HER on the specified catalyst. The Tafel slope can be used to determine the rate-limiting step of the reaction and the reaction pathway of the HER on the surface of the nanostructures. The HER in acidic media can be described in three steps:
(Volmer step) H3O+ + e− + M ⇌ M–H + H2O, (120 mV dec−1) | (4) |
(Heyrovsky step) M–H + H3O+ + e− ⇌ H2 + H2O + M, (40 mV dec−1) | (5) |
(Tafel step) 2M–H ⇌ H2 + 2M, (30 mV dec−1) | (6) |
The Volmer reaction is the primary discharge step in which protons are adsorbed to active sites on the surface of the catalysts and combine with electrons to form adsorbed hydrogen atoms. This step is followed by either an electrochemical desorption step (Heyrovsky reaction) or a recombination step (Tafel reaction).21 The Tafel plots of the pristine ReSe2 and the heterostructure are shown in Fig. 10(b). The Tafel slopes of the ReSe2 nanostructures and the heterostructures were determined to be 97 mV dec−1, 84 mV dec−1, 72 mV dec−1, and 142 mV dec−1 for the pristine ReSe2 nanostructures, ReSe2-rGO, ReSe2-N-rGO, ReSe2-g-C3N4 respectively. This shows that the HER on the surface of the ReSe2-rGO and ReSe2-N-rGO heterostructures proceeds under the Volmer–Heyrovsky pathway. This pathway involves the adsorption of protons on the surface of the catalyst which then combines with electrons to form adsorbed hydrogen atoms (Volmer step), this is followed by the electrochemical desorption of the hydrogen from the surface of the electrode (Heyrovsky step). The electrochemical desorption is the rate-limiting step in this case. However, the Volmer step is the rate determining step for the HER on the ReSe2-g-C3N4 heterostructure. The exchange current density is a useful parameter to probe reaction kinetics, it represents the rate of electron transfer at zero overpotential and is represented in the form of current density. The exchange current density can be determined from the Tafel plot by using the Tafel equation (log Jo = a), where a is the intercept of the Tafel plot. The higher the exchange current density, the better the catalytic performance. The exchange current density was determined to be 0.61 mA cm−2, 0.69 mA cm−2, 0.71 mA cm−2, and 0.38 mA cm−2 for ReSe2, ReSe2-rGO, ReSe2-N-rGO, and ReSe2-g-C3N4. The exchange current density increases when the ReSe2 nanostructures are grown on rGO and N-rGO. However, the exchange current density decreases when ReSe2 is grown on g-C3N4. This is in keeping with what has already been observed with the analysis of the overpotential and the Tafel slope.
The electrochemically active surface area (ECSA) was determined using the Cdl (double layer capacitance). The Cdl has been shown to be proportional to the ECSA, the Cdl was determined using cyclic voltammetry (CV) as shown in Fig. S13.† 37,43 The estimation is done by plotting ΔJ (Ja − Jc) vs. the scan rate at 0.41 V vs. RHE which is shown in Fig. 10(c), in this linear plot, the slope is the Cdl. The Cdl of the nanostructures was determined to be 1.85 mF cm−2, 1.98 mF cm−2, 4.45 mF cm−2, 1.98 mF cm−2 for ReSe2, ReSe2-rGO, ReSe2-N-rGO, and ReSe2-g-C3N4 respectively. The rGO and N-rGO based heterostructures have a higher ECSA which indicates that the exposure of the active sites was enhanced when the rGO and N-rGO were introduced. The ReSe2-g-C3N4 heterostructure had the lowest ECSA, which was expected considering the heterostructure had the lowest loading of the ReSe2 on the surface of g-C3N4. The ReSe2 with the nitrogen-doped rGO had a lower ECSA than that of the undoped rGO. This is expected considering the TGA analysis showed that there was much less loading of the ReSe2 catalyst on the nitrogen-doped rGO. However, by every other metric, the ReSe2 catalyst with the nitrogen-doped rGO has better electrocatalytic activity than the ReSe2 with the undoped rGO. This suggests that the improved catalytic activity of the ReSe2 with the nitrogen-doped rGO is not just due to the improved exposure of the active sites. Since the determination of the ECSA is made using the Cdl, the Cdl contribution of the 2D carbon nanostructures must also be considered, the Cdl contribution of the carbon nanostructures are sure to over inflate the Cdl values. There is another approach that can be used to evaluate the ReSe2 catalyst available on the surface of the 2D carbon nanostructures. The irreversible surface oxidation of the ReSe2 nanostructures can be used to investigate ECSA. The anodic peaks shown in Fig. 10(d) represent the surface oxidation process, the area under the peaks represents the amount of charge needed to fully oxidize the surface of the nanostructures.44 There is significantly more charge required for the oxidation of the ReSe2-rGO, which suggests that there is far more available surface area to oxidize than for the ReSe2-N-rGO and ReSe2-g-C3N4. This indicates that the higher catalyst loading on the rGO leads to more available active sites compared to the ReSe2-N-rGO and ReSe2-g-C3N4, this is corroborated by the ECSA and TGA. The electrochemical impedance spectroscopy (EIS) was used to study the electron transfer properties of the heterostructures during the HER. The EIS results are represented as Nyquist plots as shown in Fig. 10(e). The equivalence circuit fitted to the data is shown as an insert in Fig. 10(e), the circuit components are shown as Rs (solution resistance), Q2 (constant phase element), and the Rct which is the charge transfer resistance. The charge transfer resistance represents the ease or difficulty of charge transfer at the interface between the electrode and the electrolyte. The lower the Rct the better the rate of electron transfer and thus the better the performance of the catalyst. The Rct of the ReSe2 nanostructures was measured to be 251 Ω. The electron transfer processes improve when the ReSe2 is incorporated into the doped and undoped rGO. The Rct of the ReSe2-rGO, ReSe2-N-rGO, and ReSe2-g-C3N4 was determined to be 117 Ω, 65 Ω and 275 Ω respectively. The electron transfer of the HER processes is more effective in the ReSe2-N-rGO as compared to the ReSe2-rGO. This is due to the improved electron conductivity of the rGO after nitrogen doping. This explains the superior performance of ReSe2-N-rGO as compared to ReSe2-rGO even though the ReSe2-rGO has been shown to have a higher ECSA. This is further confirmed through the computational analysis of the ReSe2-rGO and ReSe2-N-rGO. The heterostructures were constructed by attaching the GO sheet and N-rGO to the optimized structure of the ReSe2 structure. The build structures are shown in Fig. 11. Fig. 11(C) and (F) show the side view of the optimized heterostructure, while Fig. 11(F) shows the top view.
The charge transfer of the N-rGO and ReSe2 heterostructures were examined using the charge density difference, with an isosurface value of 0.009 e Å−3 (Fig. 12).
We contend that there is clear accumulation/depletion of electrons within both heterostructures. The gain of electrons (shown in yellow) and the loss of electrons (shown in blue) from the ReSe2-rGO and ReSe2-N-rGO heterostructures—demonstrates the obvious anisotropic redistribution of electronic charges that occurs upon the formation heterointerfaces (as shown in Fig. 12).45 ReSe2-rGO heterostructure (Fig. 12(A)), reveals that the more localised 4p state of the Se-atoms and the delocalised Re-atoms in the 6s are the sources of the charge redistribution. The ReSe2 monolayer shows a very balanced electron accumulation/depletion as electrons are equally distributed within the system. In the upper layer (GO monolayer), the influence of the O-atom shows a high localised charge redistribution of electrons and makes C-atoms delocalised, this is due to the high electronegativity effect of O-atom in the GO/ReSe2 heterostructure system. We observe a high-intensity accumulation of electron charge density within the sphere of O-atom. The oxygen 2p state acts as the charge redistribution localization and the carbon 2p state acts as the charge delocalisation. In the formed π-bonds, oxygen turns to achieve high electron efficiencies as it shows a substantial yellow area.46
Fig. 12(B) shows the GO-graphiticN/ReSe2 heterostructure, where we can observe the yellow and blue areas that demonstrate the accumulation/depletion. In the ReSe2 monolayer, the Se-atoms illustrate the depletion of electrons in all sites of the monolayer, this influences Re-atoms that show the high accumulation spheres.47 As expected in this system, we can observe a high electron charge transportation around the Re-atoms, suggesting that Re-atoms gain electrons through the electron charge reduction of Se-atoms. In the upper layer (GO-graphitic layer), as the N-atom is introduced into the system, we can observe no change in the reduction of high-intensity charge distribution within the O-atom. O-atom shows a very high accumulation of electrons compared to N-atom, this is because oxygen has higher electronegativity than nitrogen. Fig. 12(C) shows the top-view of the GO-graphiticN/ReSe2 heterostructure, although N-atom is gaining electrons as well, oxygen has more accumulation sphere than nitrogen.47 The accumulation of charge on the O-atom on the surface functional groups observed in the charge difference maps facilitates the transfer of charge (electron transfer) from the ReSe2 to the rGO. The electron transfer is further enhanced by the incorporation on N-doping, this is shown in the accumulation of charge on the N-atom in N-doped rGO. This clearly shows the mechanism by which the N-doping can improve the electron transfer from ReSe2 and N-rGO.48,49 Coupled with the observed improvement in the Rct of ReSe2-N-rGO, it is evident that the N-doping of the rGO can significantly affect the electrocatalytic activity of the heterostructures through the modulation of the electron transfer characteristics of the 2D nanostructures. The ReSe2-g-C3N4 heterostructure had the highest Rct, this is due to the weak interaction of the ReSe2 with the nitrogen groups on the g-C3N4. The stability of the electrocatalysts was evaluated using chronoamperometry (CA), where the current response of the catalyst after an applied potential is evaluated over time as shown in Fig. 10(f). The ReSe2 and ReSe2-g-C3N4 catalysts experience a current drop in the first two hours and then the current begins to stabilize. However, the current response for ReSe2-rGO and ReSe2-N-rGO seems to continue dropping until 4 hours. A comparison of the catalytic activity of the as synthesized ReSe2 heterostructures and their counterparts found in literature that are synthesized with other methods is shown in Table 4. The electrochemical characterisation of the nanostructures compared were conducted in 0.5 M H2SO4.
HER catalyst | Synthesis method | η10 (mV) | Tafel slope (mV dec−1) | Onset potential (mV) | Ref. |
---|---|---|---|---|---|
ReSe2@rGO | Hydrothermal | 145 | 40.7 | — | 36 |
ReSe2 on carbon cloth | CVD | 265 | 69 | — | 20 |
ReSe2 on porous carbon cloth | CVD | 200 | 108 | — | 50 |
Treated ReSe2 on porous carbon cloth | CVD | 140 | 64 | — | 50 |
ReSe2-rGO | Colloidal | 259 | 84 | 128 | This work |
ReSe2-N-rGO | Colloidal | 218 | 72 | 115 | This work |
ReSe2-g-C3N4 | Colloidal | 385 | 142 | 135 | This work |
Table 4 The catalytic activity of ReSe2 grown on carbon cloth in other studies is comparable to the catalytic activity of the ReSe2-N-rGO heterostructure. There is currently no work that studies the use of nitrogen doped reduced graphene oxide. The ReSe2@rGO synthesized by Yan et al. had higher catalytic activity than the heterostructures synthesized in this work. This was attributed to the catalyst loading; the loading was recorded as 9.5 wt%.36 This was much higher than the catalyst loading recorded in this work, the catalyst loading was recorded at 2.7% and 1.7% for ReSe2-rGO and ReSe2-N-rGO respectively. This indicates that there is still a lot of work to be done in optimizing the reaction conditions in order to ensure higher catalyst loading. The work by Li et al. also showed higher catalytic activity and this was due to heat treatment after the ReSe2 nanostructures on the carbon cloth.50
Footnote |
† Electronic supplementary information (ESI) available. See DOI: https://doi.org/10.1039/d5ra01096a |
This journal is © The Royal Society of Chemistry 2025 |