Highly defective graphene quantum dots-doped 1T/2H-MoS2 as an efficient composite catalyst for the hydrogen evolution reaction

We present a new composite catalyst system of highly defective graphene quantum dots (HDGQDs)-doped 1T/2H-MoS2 for efficient hydrogen evolution reactions (HER). The high electrocatalytic activity, represented by an overpotential of 136.9 mV and a Tafel slope of 57.1 mV/decade, is due to improved conductivity, a larger number of active sites in 1T-MoS2 compared to that in 2H-MoS2, and additional defects introduced by HDGQDs. High-resolution transmission electron microscopy (HRTEM), Raman spectroscopy, x-ray diffraction (XRD) and x-ray photoelectron spectroscopy (XPS) were used to characterize both the 1T/2H-MoS2 and GQDs components while Fourier-transform infrared spectroscopy (FTIR) was employed to identify the functional groups on the edge and defect sites in the HDGQDs. The morphology of the composite catalyst was also examined by field emission scanning electron microscopy (FESEM). All experimental data demonstrated that each component contributes unique advantages that synergistically lead to the significantly improved electrocatalytic activity for HER in the composite catalyst system.


Methods
The 1T/2H-MoS 2 /HDGQDs composite catalyst system was fabricated by sequentially dropping the HDGQDs (0.2 ml, 6 mg/ml) or GQDs (0.0545 ml, 22 mg/ml) 14 and 1T/2H-MoS 2 solutions (0.2 ml) onto the carbon paper substrate (CeTech, GPP035), covering a circular area of radius 0.3 cm, and then letting them air dry.The dosages for 1T/2H-MoS 2 and HDGQDs were tested to achieve optimal loading, with further details provided in the supplementary information.As shown in Supplementary Figure S1, S2 and Table S1, we determined that 0.2 ml of MoS 2 is the optimal loading on the Carbon paper as the primary catalytic agent.When the MoS 2 loading exceeds 0.2 ml, the efficiency decreases due to increased resistance, which hinders the hydrogen evolution reaction.As depicted in Supplementary Figure S3, we found that 0.2 ml of HDGQDs is the optimal loading for enhancing the MoS 2 catalyst.Further addition of HDGQDs does not increase efficiency, as the active sites of MoS 2 are already saturated.
To prepare the HDGQDs solution, citric acid powders (2 g) in a glass flask were oil-bath heated to 200℃ for 30 min.The orange liquid of pyrolyzed citric acid was mixed with deionized water (20 ml) to form a HDGQDs solution, which was then purified by filtering using a 0.22 μm microporous membrane and dialyzed using a cellulose ester membrane bag (flat width 54 mm, diameter 34 mm, MWCO 3.5 KD) for two days, with several water changes during the dialysis process.For comparison, we also fabricated standard GQDs by increasing the heating time from 30 to 90 min in the above process.
The 1T/2H-MoS 2 solution was prepared by using an improved lithium ion intercalation-based exfoliation process.Molybdenum (IV) sulfide 99% powders of 90 nm-diameter (0.1 g) and n-Butyl lithium (5 ml) were mixed in an autoclave reactor and heated to 110℃ in an oven for 12 h.In contrast to the previous method, where 0.5 g bulk MoS 2 powder was soaked in 4 ml of 1.6 M n-butyllithium/hexane for 48 hours 15 , this approach can significantly enhance the efficiency of lithium ion insertion.After cooling down to room temperature, the mixture was diluted in n-hexane and centrifuged to remove excessive n-Butyl lithium not intercalated with MoS 2 , leaving the Li x MoS 2 as precipitates.Hydrochloric acid solution (150 ml, pH = 3) was then added to exfoliate MoS 2 few layer sheets from the Li x MoS 2 precipitates.Finally, the MoS 2 solution was purified by centrifugation, filtration (pore size 0.22 μm, CA membrane filter) and dialysis (Peristaltic pump rapid dialysis device, MWCO 12-14 kD hollow fiber) to obtain a neutral 1T/2H-MoS 2 solution.Compared to the previous method, where hydrochloric acid solution was dropwise added to an alkaline MoS 2 colloidal solution containing trace amounts of lithium hydroxide until the pH reached around 7 15 , this milder dialysis approach avoids residual chloride ions and exothermic acid-base reactions.This prevents phase change and oxidation of MoS 2 .On the other hand, the milder dialysis approach can effectively remove LiOH and concentrate the sample according to experimental requirements.For the Raman and XPS measurements, the solution was dropped on a carbon paper substrate and air dried to form the 1T/2H-MoS 2 sample.A more thermally stable 2H-MoS 2 sample was also prepared by heating the 1T/2H-MoS 2 sample at 300℃ for 1 h in vacuum 16 .
To characterize the GQDs and 1T/2H-MoS 2 components of the composite catalysts, a high-resolution transmission electron microscope (HR-TEM) (JEOL JEM-2010) with acceleration voltage 200 kV was used to probe the microstructures.Raman spectra were obtained using a Raman spectrometer with a CCD camera (HORIBA, iHR-550) and a 514 nm excitation laser.The Raman band for Si at ~ 520.7 cm −1 was used as a reference to calibrate the spectrometer.The x-ray diffraction (XRD) patterns in the 2θ range of 5-40° were recorded on an in-house x-ray diffractometer (Bruker, D8 discover plus) using the Cu-Kα radiation (λ = 0.15418 nm).Fourier-transform infrared spectroscopy (FTIR) was also obtained, by using an FT-IR spectrophotometer (Bruker, Vertex80v), to identify the functional groups on the edge and defect sites in the HDGQDs sample.The chemical environments surrounding C and O in GQDs and Mo and S in MoS 2 were investigated by x-ray photoelectron spectroscopy (ULVAC-PHI, PHI Quantera II, with binding energies corrected using the C 1 s peak of 284.5 eV).Finally, the morphologies of the composite catalyst systems were examined by field emission scanning electron microscopy (FE-SEM) (JEOL 7900F).
Electrochemical analyses including linear sweep voltammetry (LSV), Tafel plots, electrochemical impedance spectroscopy (EIS), and stability test were carried out to evaluate the electrocatalytic activity for the new composite catalyst system of 1T/2H-MoS 2 /HDGQDs on carbon paper substrate.An electrolytic cell with the electrocatalyst, an Ag/AgCl electrode, and a glassy carbon electrode as the working, reference, and counter electrodes, respectively, submerged in a 0.5 M H 2 SO 4 electrolyte solution was connected to a CHI electrochemical workstation (CHI Instruments 760D) for all the electrochemical measurements.The LSV curves were scanned with a scan rate of 5 mV/s.The LSV curve of a blank carbon paper substrate was also measured to perform background subtraction and baseline correction for all samples 17 .A 1.5 Ω small series resistance is used to treat LSV curves for iR correction 18 .The electrochemical potentials in these measurements were converted to the reversible hydrogen electrode (RHE) scale 19 .Using the Tafel equation, Tafel plots were also obtained from the LSV data to calculate the Tafel slopes 20 .The EIS data were measured at an overpotential of − 0.3 V vs. RHE in the frequency range of 0.001 ~ 100 kHz with an amplitude of 5 mV 10,12 .Chronopotentiometry experiments were carried out in a 0.5 M H 2 SO 4 solution with a constant current density of 10 mA/cm 2 .

Results and discussion
As demonstrated in Fig. 1a, the TEM micrograph shows that the particle size of the HDGQDs is around 3-5 nm.The graphite (1120) planes with a d-spacing of 0.22nm 21 can also be clearly seen in the high-resolution TEM (HRTEM) micrograph (Fig. 1b).However, no clear pattern was observed in the fast Fourier transform of the micrograph, probably due to the material's highly defective structure.In contrast, as shown in Fig. 1c, the TEM micrograph of the standard GQDs sample exhibits an appreciably larger particle size of 20 nm and its fast Fourier transform shows clear pattern of six-fold symmetry 22 (Fig. 1d), indicating that the increased heating time of 90 min can effectively eliminate defects and enlarge graphene particles at the same time.
The Raman spectroscopy measurements can be used to compare the number of defects in the HDGQDs with that in the standard GQDs.As demonstrated in Fig. 2a, two peaks are present in the spectra of both the HDGQDs and standard GQD samples.The peaks at 1355 cm −1 and 1595 cm −1 are attributed to the D band and G band which represent the defect-related out-of-plane vibration and the in-plane sp 2 mode of graphene, respectively.Therefore, the ratio of peak intensities of the D and G bands can be used to evaluate the defective structure of the sample 23 .The D/G ratio of 2.56 for the HDGQDs sample is appreciably higher than that of 1.70 for the standard GQDs sample.The Raman spectroscopy results demonstrate that the number of defects in GQDs samples fabricated from pyrolyzing citric acid can be controlled by adjusting the heating time.
X-ray photoelectron spectroscopy (XPS) was used to identify the chemical environments of carbon and oxygen atoms in the HDGQDs sample.As shown in Fig. 2c, three peaks centered at 288.5 eV, 286.3 eV and 284.6 eV assigned to O = C-O, C-O/C-O-C, and sp 2 C = C bonding, respectively, were observed in the C1s spectrum 11,13,21,[27][28][29] .In the O1s spectrum shown in Fig. 2d, the peaks at 535.6 eV, 532.3 and 530.9 eV are attributed to the C-O-H, C-O, and C = O bonding, respectively 27 .The chemical bonding of C and O revealed by XPS measurements is consistent with the presence of the functional groups observed from FTIR data in the HDGQDs samples (Fig. 2b).
As shown in Fig. 3a, the particle size of 1T/2H-MoS 2 revealed by the TEM micrograph is around 20-40 nm, substantially larger than the HDGQDs size of 3-5 nm.The Raman data plotted in Fig. 3b shows that both the 1T/2H-MoS 2 and 2H-MoS 2 spectra have two major peaks: a peak at 379.4 or 381.5 cm −1 (E 1 2g ), representing the vibration of two sulfur atoms with respect to molybdenum, and a peak at 401.0 or 4.03,8 cm −1 (A 1g ), representing the relative vibration of S atoms in the out of plane direction 8 .However, the 1T/2H-MoS 2 spectrum exhibits four additional peaks at 146.8 cm −1 (J 1 ), 237.6 cm −1 (J 2 ), 283.7 cm −1 (E 1g ), and 334.7 cm −1 (J 3 ) 6 .The E 1g mode is www.nature.com/scientificreports/attributed to the octahedral coordination of Mo in 1T-MoS 2 30 while the J 1 , J 2 , and J 3 modes are due to the superlattice structure of 1T-MoS 2 31 .Therefore, the Raman results indicate that our 1T/2H-MoS 2 sample indeed contains both the 1T-MoS 2 and 2H-MoS 2 phases.
Figure 3c shows the XRD patterns of the 1T/2H-MoS 2 and 2H-MoS 2 samples.A diffraction peak centered at around 15.0° corresponding to the (002) planes of the 2H-MoS 2 structure appeared in the data of the 2H-MoS 2 samples.On the other hand, three peaks at around 7.8°, 11.6° and 15.1° representing the 1T (001), 1T (002) and 2H (002) planes, respectively, were observed in the data of the 1T/2H-MoS 2 sample [32][33][34] .According to the Scherrer equation, D = κλ/βcos(θ), where κ = 0.9 and λ = 0.154056 (nm), the grain size for the 2H-MoS 2 sample with 2H (002) peak position 2θ = 15.0° and peak width β = 1.44° was estimated to be 5.6 nm.For the 1T/2H-MoS 2 sample, the 1T-MoS 2 grain size calculated from the 1T(001) peak at 2θ = 7.8° with β = 1.87° is 4.3 nm while the 2H-MoS 2 grain size calculated from the 2H (002) peak at 2θ = 15.1° with β = 1.84° is 4.4 nm.We note that the grain size obtained from the XRD patterns for the MoS 2 samples are a lot smaller than the particle size observed in the TEM micrograph.This indicates that the smaller MoS 2 grains that give rise to the broader XRD peaks have combined to form larger aggregates seen in the TEM micrograph.From the above XRD results, we can confirm that our 2H-MoS 2 sample has pure 2H phase and our 1T/2H-MoS 2 sample has both 1T phase and 2H phase.
The LSV curves and Tafel plots for all samples are shown in Fig. 6a,b, respectively.The overpotential and Tafel slope that showcase the electrocatalytic activity for each sample are plotted in Fig. 6c 40 .As shown in Fig. 6c, the  We can also see that the electrocatalytic activities for both 1T/2H-MoS 2 and 2H-MoS 2 were significantly improved when the standard GQDs were incorporated into the samples, while the catalytic activity of the pure HDGQDs and GQDs electrode can be considered negligible (Supplementary Figure S4).Further activity improvement can be achieved by replacing the standard GQDs with the HDGQDs.The best electrocatalytic activity for the MoS 2 /GQDs systems was found in the 1T/2H-MoS 2 / HDGQDs sample, of which the overpotential and Tafel slope of 136.9 mV and 57.1 mV/decade were significantly improved towards the Pt/C values of 64.1 mV and 43.6 mV/decade, respectively.
Figure 6d shows the Nyquist diagrams obtained from the electrochemical impedance spectroscopy (EIS) for all samples, as well as the carbon paper substrate.The corresponding equivalent circuit which consists of solution resistance, R s , charge-transfer resistance, R ct , and constant phase element (CPE) is also shown in the inset of Fig. 6d.3][44][45] The charge transfer resistance (R ct ), that originates from the electronic and ionic resistances at the electrode-electrolyte interface, reflects the kinetics of the catalyzed hydrogen evolution reaction 10 .We can see that, as expected, the metallic-1T-MoS 2 -rich 1T/2H-MoS 2 sample has a lower R ct compared to that of the semiconducting 2H-MoS 2 sample.When standard GQDs are added to the MoS 2 samples, both the 2H-MoS 2 and 1T/2H-MoS 2 samples show reduced charge-transfer resistance, indicating GQDs have effectively improved the electrode-to-electrolyte charge transfer of the composite catalytic systems.The conductivity can be further improved by replacing the standard GQDs with HDGQDs.We observed the lowest R ct of 537.4 Ω in the 1T/2H-MoS 2 /HDGQDs sample, which was dramatically lower than the 2H-MoS 2 value of 1256.3Ω.
It has been reported that GQDs can generate abundant defect sites on the basal plane and edge plane of MoS 2 . 10These defect sites facilitate electron transfer from GQDs to MoS 2 . 10It has also been reported that enhanced GQDs-to-MoS 2 charge transfer can improve MoS 2 's efficiency in adsorbing protons. 12As demonstrated in our XPS analysis above, we did observe charge transfer from HDGODs to MoS 2 in our sample.In this work, we use highly defective GQDs to increase the number of defect sites introduced by GQDs, so as to enhance the GQDs-to-MoS 2 charge transfer, further promoting MoS 2 's efficiency in adsorbing protons and therefore greatly reducing R ct in our catalyzed HER.
According to the above EIS analysis, by selecting 1T/2H-MoS 2 and HDGQDs to construct our composite catalyst system, we were able to greatly reduce the charge transfer resistance in the catalyzed hydrogen evolution reaction.Furthermore, while 2H-MoS 2 only has active sites on the edge plane, 1T-MoS 2 has active sites on the edge plane and basal plane. 3,4,46,47When doped with the HDGQDs, 1T/2H-MoS 2 is likely to have more defects near the active sites.][7][8][9][10]46,47 .
Finally, a continuous long-term operation of 1000 cycles and a 24-h chronopotentiometry 29,48-50 were conducted for the 1T/2H-MoS 2 /HDGQDs sample to test its stability.As shown in Fig. 7a, no appreciable difference was found between the LSV curve obtained after 1000 cycles and the initial curve.The chronopotentiometry curve in Fig. 7b shows that the overpotential with a constant current density of 10 mA/cm 2 was maintained at a stable level for 24 h.As shown in the inset in Fig. 7b, the LSV curves before and after the 24-h chronopotentiometry measurement showed no appreciable difference either.Therefore, our 1T/2H-MoS 2 /HDGQDs composite catalyst appears to be very stable for HER applications.

Conclusion
We have designed and successfully fabricated an efficient composite catalyst system, consisting of a metallic-1T-MoS 2 -rich 1T/2H-MoS 2 and a HDGQDs components of particle sizes around 20-40 and 3-5 nm, respectively, for the hydrogen evolution reaction.The 1T/2H-MoS 2 component was prepared by using an improved process involving a heated mixing of MoS 2 powder with n-Butyl lithium in an autoclave reactor and a dialysis procedure to significantly expedite the production of 1T-MoS 2 .By using a shortened heating time in pyrolyzing citric acid, we were able to introduce a large number of defects in the GODs product to form HDGQDs, as shown by Raman spectroscopy.While FTIR and XPS data revealed the presence of many functional groups in the HDGQDs that may promote catalytic activity, the XPS data demonstrated that the 1T phase of MoS
and 161.06 eV, while those for the 2H-MoS 2 phase are at 163.19 and 162.06 eV, respectively.From the fitted peak

Figure 7 .
Figure 7. Stability test: (a) 1000 LSV cycles and (b) The chronopotentiometry curve recorded at a constant cathodic current density of -10 (mA cm -2 ).Inset shows the linear sweep voltammetry polarization curves before and after the 24-h chronopotentiometry.