Laser Assisted Solution Synthesis of High Performance Graphene Supported Electrocatalysts

Simple, yet versatile, methods to functionalize graphene flakes with metal (oxide) nanoparticles are in demand, particularly for the development of advanced catalysts. Herein, based on light-induced electrochemistry, a laser-assisted, continuous, solution route for the simultaneous reduction and modification of graphene oxide with catalytic nanoparticles is reported. Electrochemical graphene oxide (EGO) is used as starting material and electron-hole pair source due to its low degree of oxidation, which imparts structural integrity and an ability to withstand photodegradation. Simply illuminating a solution stream containing EGO and metal salt (e.g., H2PtCl6 or RuCl3) with a 248 nm wavelength laser produces reduced EGO (rEGO, oxygen content 4.0 at%) flakes, decorated with Pt (~2.0 nm) or RuO2 (~2.8 nm) nanoparticles. The RuO2-rEGO flakes exhibit superior catalytic activity for the oxygen evolution reaction, requiring a small overpotential of 225 mV to reach a current density of 10 mA cm-2. The Pt-rEGO flakes (10.2 wt% of Pt) show enhanced mass activity for the hydrogen evolution reaction, and similar performance for oxygen reduction reaction compared to a commercial 20 wt% Pt/C catalyst. This simple production method is also used to deposit PtPd alloy and MnOx nanoparticles on rEGO, demonstrating its versatility in synthesizing functional nanoparticle-modified graphene materials.


Introduction
The urgent need for sustainable and clean energy to reduce the usage of traditional fossil fuels has promoted enormous interest in the field of energy storage and conversion. [1] The use of hydrogen as an intermediate for energy storage and power generation has been considered as one of the most promising alternatives to the current non-renewable fossil fuels. Within the hydrogen economy, molecular hydrogen links power grids to other energy sectors through a zero-emission electrochemical pathway. [2][3] In detail, the pathway is the generation of hydrogen as well as oxygen via electrochemical water splitting by the hydrogen evolution reaction (HER) and oxygen evolution reaction (OER) at the generator and then the consumption of hydrogen using a fuel cell system, in which the oxygen reduction reaction (ORR) and hydrogen oxidation reactions convert hydrogen directly into electricity. However, the sluggish kinetics of these electrochemical energy conversion reactions limits seriously the wide application of hydrogen energy. [4] The development of high-performance electrocatalysts, which can reach a designated current density under minimum overpotential, is desirable for maximizing the hydrogen production and utilization efficiency. To-date, platinum group metals and their oxides remain the state-of-the-art catalysts for electrochemical energy conversion. [2] To reduce the cost of electrocatalysts, carbon nanomaterials (e.g. carbon black, carbon nanotubes, graphene, etc.) are widely used as supports for the platinum group metal (oxide) nanoparticles.
Due to its high electrical conductivity and large specific surface area, the two-dimensional (2D) single atom thick graphene has been considered as a promising supporting material for developing advanced electrocatalysts. [5] Compared with the hydrophobic pristine graphene flakes, graphene oxide (GO) with oxygen groups and thus aqueous solution processability has become a more versatile starting material for loading/supporting functional nanoparticles, including electrocatalyst nanoparticles. [6] In the last decade, various methods have been developed to deposit nanoparticles onto the surface of graphene/GO/reduced GO (rGO) flakes, including wet chemical deposition, [7][8][9][10] electrochemical deposition, [11][12] and plasmaassisted synthesis, [13] etc. However, several of these techniques require elevated temperatures, harsh chemicals, or high voltage bias. A robust and versatile method which is able to synthesis various types of ultrafine nanoparticles on the surface of graphene in a single-step is still desired.
The use of laser technology to prepare nanomaterials has recently attracted increasing attention due to its simple and fast merits. [14][15][16][17][18][19][20] Methods including laser ablation in liquid (PLAL), [14][15] laser pyrolysis, [16,18,[20][21] and photodeposition [22][23][24] have been demonstrated for the synthesis of either carbon supported or unsupported nanoparticles for the applications in electrochemical energy storage and conversion. Briefly, in laser ablation in liquid, nanoparticles are formed by the rapid cooling of a plasma plume compromised of elements from the solid ablation targets and the surrounding liquid; [14][15] while in laser pyrolysis, the laser induced carbonization/graphitization of a precursor (e.g. polyimide) and/or decomposition of metal salt precursor leads to the formation of nanoparticles modified graphene/nanocarbon materials. [16,18,[20][21] Different from the photothermal mechanism of both the laser ablation in liquid and laser pyrolysis, which requires focused and intense laser beam, photodeposition is based on mild light-induced electrochemistry.
Photodeposition of metal (oxide) nanoparticles on the surfaces of semiconductors (metal oxides or sulfides) has been thoroughly studied in the past few decades as indicated in a review paper published recently. [25] Photodeposition is driven by light-induced electron transfer, it occurs simply via illumination of dispersions of semiconductor particles in aqueous solutions containing metal salt precursors. [25] The photoexcitation of semiconductor creates electron-hole pairs, which reduce/oxidize the adsorbed metal ions into metal/metal oxide; the insoluble metal/metal oxide heterogeneously nucleates and grows on the semiconductor substrate. [25][26] In spite of the numerous interests in photodeposition of nanoparticles on the surfaces of metal oxides and sulfides semiconductors, only a few works have studied GO, [22][23][24] which also behaves as a semiconductor with tunable bandgap values depending on the content and type of oxygen groups. [27][28][29] Additionally, the sizes of the photodeposited nanoparticles on rGO reported in these pioneering works using lasers with wavelengths of 355 and 532 nm are relatively large (> 5 nm) and not uniformly distributed. [22][23][24] To date, there is no report regarding the use of photodeposited nanoparticles on GO/rGO as electrocatalysts. One of the possible reasons for the absence of reports on the use of GO for photodeposition is the simultaneous reduction and degradation of GO when exposed to laser irradiation. The reduction of GO by the photon-excited electrons, [25,[30][31] is accompanied by the undesired oxidative GO degradation by photo-generated holes. [31][32] This generally leads to a partial reduction or even degradation of the GO flakes, [31][32][33] particularly for the heavily oxidized GO with a large amount of oxygen groups (oxygen composition ≥ 30 at.%). [32] In addition, the existence of electron scavenging metal ions (e.g. Pd 2+ ) during the photodeposition of metal nanoparticles could further impede the full reduction of GO. [24] A very recent discovery suggests that the use of mildly oxidized, oxygen functionalized graphene as starting material leads to highly reduced high quality graphene flakes via an ultraviolet (UV) light-induced reduction. [31] Compared with the conventional chemical GO (CGO) produced by Hummers' method, the oxygen functionalized graphene has lower oxygen content and a less disrupted graphene honeycomb lattice structure. [34][35] Electrochemical GO (EGO), produced using a scalable, low-cost and environmentally friendly electrochemical oxidation, has a similar structure with oxygen functionalized graphene with a low content of oxygen groups (~20 at.%), especially the unrestorable C=O and −COO− groups. [36] It has been proved that the reduction of EGO via chemical approaches (e.g. hydrazine) can lead to a higher degree of graphene lattice restoration compared to that can be achieved using CGO. [36] In addition, the photo-degradation of carbon lattice of GO is reported to be dependent on the oxidation level/oxygen content, with a more severe degradation for GO has higher oxygen content. [32] Therefore, the higher structural stability and narrower bandgap of EGO due to the lower oxygen content compared with CGO could potentially allow a rapid and full reduction using lasers with high photon energy and fluence (i.e. laser energy density in mJ cm −2 ) without causing significant degradation of the carbon lattice. Meanwhile, the use of laser beam with higher photon energy and fluence benefits the generation of electron-hole pairs via a one photon process, the abundant electrons/holes would potentially lead to a high nucleation rate for the metal/metal oxide particles and thus small particle size.
Herein, in this work, we report a UV (248 nm) laser-induced continuous solution phase strategy for simultaneous reduction and modification of EGO with uniformly distributed ultrafine catalyst nanoparticles. In a typical experiment (Figure 1 and S1), the aqueous precursor solution of metal salt (e.g. H2PtCl6, RuCl3, etc.) and EGO was circulated continuously from a bulk solution tank to a quartz cell reactor, on which a UV laser (KrF excimer; wavelength: 248 nm; pulse width: 10 ns; repetition rate: 100 Hz; photon energy:

Characterization of reduced EGO
The optical bandgap of EGO and their fragments dispersed in water has been estimated by applying the Tauc plot [37] through the UV-Vis absorption spectrum (Figure 2a and b). The combination of the π-state (sp 2 bonded) carbons, and the σ-state (sp 3 bonded) carbons in GO makes it a semiconductor with a bandgap in a range from 2 to 7 eV. [28] The absorption at ~4 eV caused by n-π* transitions of C=O, and the peak at approximately 5 eV is attributed to ππ* transitions of C=C. [38][39] The bandgap values of EGO has been approximated from the linear extrapolation using the Tauc plot, [40] which gives a direct bandgap range from 3.25 to 3.95 eV (Figure 2a), and an indirect bandgap from 2.04-2.4 eV (Figure 2b). Therefore, the photon energy of 248 nm laser (4.99 eV) is sufficiently high to excite EGO and thus create electron-hole pairs. In contrast, due to the higher degree of oxidation compared with EGO, [36] CGO exhibits larger bandgap values for both direct (3.25 to 4.31 eV) and indirect (2.04 to 3.38 eV) bandgaps ( Figure S2).
Raman spectroscopy was used to characterize the quality of the graphene lattice and defect density of rEGO. The excitation laser (He-Ne; 633 nm) output power for recording the Raman spectra was limited to < 0.5 mW to avoid any laser-induced structure alteration. Figure 2c shows typical Raman spectra of EGO and rEGO. Spectra of all samples show a D band at 1333 cm −1 representing the edge planes and disordered structures, and the characteristic G band at 1595 cm −1 ascribed to the ordered sp 2 bonded carbon. [41] The Raman spectra were further analyzed by fitting the D and G bands with Lorentzian function after baseline subtraction (see details in Figure S3). The intensity ratio of D to G band (ID/IG) is closely related to the density of defect/functionality in the graphene lattice, [42][43][44] the ID/IG ratio of rEGO increases linearly from 1.25 to 2.01 with the increasing of laser fluence (0 to 681 mJ cm −2 , Note: zero fluence represents pristine EGO) (Figure 2d). Using the model proposed by Lucchese et al. [42] and Cançado et al., [43] the defect distance (LD) can be determined from the Raman ID/IG ratio ( Figure S4a and Table S1). With the increasing laser fluence, LD rises from 1.17 to 1.32 nm, suggesting a partial restoration of the graphene lattice. Quantification of defect density (θ), defined as the ratio of C(sp 3 ) to C(sp 2 ), using the calculated LD values, [44] suggests a gradual reduction of θ from 2.13% to 1.68% with the increase of laser fluences ( Figure S4b and Table S1). The lattice restoration after laser reduction is also confirmed by the narrowing of full width at half-maximum (FWHM; Γ) of D and G bands (Figure 2e and f). [45] In addition, the photographs of EGO solutions (Figure 2g) indicate that the color of EGO dispersions turned from brown into black after laser irradiation at various fluences of 227, 340, 454, 568 and 681 mJ cm −2 , respectively. X-ray diffraction (XRD) further confirms the increased reduction of EGO at higher laser fluence. The diffraction peak of EGO at ~10° vanished gradually with the increase of laser fluence ( Figure S5).
The X-ray photoelectron spectroscopy (XPS) survey spectra collected with pristine EGO and rEGO after irradiation at a laser energy density of 568 mJ cm −2 reveal a significant decrease in oxygen content, from 24.0 at.% (EGO) to 4.0 at.% (rEGO) ( Figure S6). Figure 2h [46] (details in Table S2). In contrast, the C 1s of rEGO shows an effective removal of oxygen groups and restoration of sp 2 carbon structure. Notably, the sp 2 carbon content increases dramatically from 2.5 at.% for pristine EGO to 61.9 at.% for rEGO. Both the Raman and XPS results indicate an efficient restoration of the sp 2 bonded graphene lattice from the sp 3 oxygenated sites via the facile laser-induced reduction. In comparison, the reduction of EGO by thermal annealing up to 250 °C removes oxygen functionalities but leaves the defects unrestored. The XRD diffraction peak of EGO at ~ 10° decreases in its intensity and disappears after annealing at temperature > 200 °C, revealing the removal of the majority of the oxygen functionalities ( Figure S7). Previous reports also suggested that thermal annealing at 200 °C effectively reduced EGO film as evidenced by the dramatic recovery of electrical conductivity. [36] However, the decrease of Raman ID/IG ratio ( Figure S8) for thermal annealed EGO suggests the sp 2 bonded graphene lattice structure is not restored. Quantification and comparison of defect distance and density using Raman ID/IG ratio indicate that thermal annealing leads to a slightly increased defect density compared with pristine EGO ( Figure   S9). This is due to the thermal decomposition of oxygen functional groups to CO2/CO products, thereby removing oxygen and carbon atoms simultaneously, leaving permanent defects and vacancies in the graphene lattice. [47] The results from thermally reduced EGO also suggest that the UV laser-induced reduction of EGO in aqueous solution follows a dominating photochemical mechanism with minimum photothermal effect.
Due to the hydrophobic nature of graphene, the rEGO dispersion agglomerated and settled at the bottom of the water within 2 hours ( Figure S10). Additionally, there is a gradual increase in weight loss (26.8 to 79.1 wt.%) and Raman ID/IG ratio (1.46 to 2.01) for rEGO irradiated using a laser beam with increasing fluence from 227 to 568 mJ cm −2 ( Figure S11). This suggests a more efficient removal of oxygen groups and restoration of graphene lattice at higher laser fluence. Further increase of the laser fluence to 681 mJ cm −2 causes a sudden drop in the remaining weight of rEGO to 22.4% compared to the starting EGO ( Figure S11), indicating photodegradation occurs at such high laser fluence. In comparison with EGO, there are much less apparent changes in weight loss, Raman ID/IG ratio, ΓD and ΓG for CGO after laser irradiation at increasing fluence from 227 to 681 mJ cm −2 (Figure S12), but a more severe photodegradation (13.8% remaining weight) at the laser fluence of 681 mJ cm −2 . In comparison to the rEGO, the laser treated CGO under the same condition yields a brownish suspension indicating insufficient reduction ( Figure S13). The less efficient reduction and restoration of conventional CGO using the laser-assisted solution approach are consistent with the recent report. [31] To avoid severe photodegradation, the laser fluence of 568 mJ cm −2 was selected to be optimal for the simultaneous reduction of EGO and deposition of catalyst nanoparticles.

Characterization of reduced EGO modified with metal (oxide) nanoparticles
Metal salts of chloroplatinic acid (H2PtCl6) and ruthenium chloride (RuCl3) were selected as precursors for metal (Pt) and metal oxide (RuO2) nanoparticles, respectively. RuO2 via doping effect and/or bond formation. [48] Therefore, the reduced ID/IG ratio and shift of G band position of RuO2-rEGO compared with rEGO suggests the formation of oxygen bridges. The complete reduction of EGO in the Pt-rEGO was confirmed by XPS C 1s spectrum ( Figure S15a). In contrast, according to literature, [24] the photodeposition of Pd nanoparticles from Pd 2+ inhibited the complete reduction of CGO. As the reduction potential vs standard hydrogen electrode (SHE) for PdCl4 2− /Pd (0.591 V) is not significantly different from that of PtCl6 2− /PtCl4 2− (0.68 V) and PtCl4 2− /Pt (0.755 V). [49] The effective laser-induced reduction of EGO in the presence of competing reactions (reduction of PtCl6 2− to Pt) suggests EGO with low contents of oxygen (~20 at.%) and unrestorable C=O/−COO− groups as a promising platform for the laser-assisted synthesizing of nanoparticle-functionalized graphene materials.
Both the Raman spectrum and XRD pattern (Figure S16a and b) of the as-synthesized RuO2-rEGO indicate an amorphous structure of the freshly deposited hydrous RuO2 (RuO2 with structural water, RuO2•xH2O). As the degree of crystallinity of RuO2 affects its OER [50][51] and pseudo-capacitance [52][53][54] [25,49] the reduction potential vs SHE decreases in the order: PtCl4 2− /Pt (0.755 V) > PtCl6 2− /PtCl4 2− (0.68 V) > Ru 2+ /Ru (0.455 V) > Ru 3+/ Ru 2+ (0.249 V). The deposition of metallic Ru requires a more negative potential than that of Pt, which is probably one of the reasons for the formation of RuO2 as the dominant phase rather than metallic Ru. Note that the pH values of EGO-H2PtCl6 (3.18) and EGO-RuCl3 (3.32) precursor solutions are comparable. At this pH range, according to the Pourbaix diagram, the reduction potential for Ru(OH)3/Ru is ~ 0.5 V vs SHE, lower than that of Pt(OH)2/Pt (~0.8 V vs SHE). [55] In addition, Ru is known to be much less noble than other Pt group metals and thus has a stable oxidation state of +4 in the presence of oxygen. [55] Electrochemical investigation of Ru metal electrodes indicated that the surface oxidation of Ru already begins at the potentials in, or close to, the H region, 0.05 to 0.2 V vs reversible hydrogen electrode (RHE) in 0.5 M H2SO4. [56] In contrast, the surface oxidation of Pt starts at a high potential of 0.8 V vs RHE in 0.5 м H2SO4. [57] Hence, another possible reason for the formation of RuO2 as dominating phase is due to the as-formed metallic Ru is prone to be oxidized by photo-generated holes.
Transmission electron microscopy (TEM) and scanning transmission electron microscopy with high angle annular dark field (STEM-HAADF) were used to characterize the particle size and crystal structure of the RuO2-rEGO and Pt-rEGO composites. The TEM and STEM-HAADF images for both the RuO2-rEGO-250HT (heat-treated in air at 250 °C; Figure 3a and c) and Pt-rEGO (Figure 3b and d) show uniformly distributed ultrafine nanoparticles on the rEGO support. Compared with the TEM images of the as-synthesized RuO2-rEGO without heat treatment (Figure S17), annealing at 250 °C shows minor effects in the morphology of RuO2-rEGO sample. Statistical particle size analysis was conducted for the Pt and RuO2 nanoparticles, with 1000 and 833 particles in random areas being analyzed, respectively, and a log-normal function being used for data fitting. The results show mono-dispersed particle sizes of 2.0 nm (standard deviation: σ = 0.5) for Pt-rEGO and 2.8 nm (σ = 0.6) for RuO2-rEGO-250HT. One of the possible reasons, that the smaller size of nanoparticles (2~3 nm) in this work compared with the sizes of photodeposited nanoparticles (> 5 nm) on rGO in the previous reports, [22][23][24] is due to the use of laser with higher photon energy and fluence. Analog to electrodeposition, higher laser photon energy and fluence correspond to larger current density (overpotential) and thus increased supersaturation, leading to higher nucleation rate and reduced critical cluster size, therefore smaller particle size. [58] In addition, the possible bond formation (e.g. oxygen bridges) between rEGO and nanoparticles would provide anchoring effect and thus inhibit the agglomeration and growth XPS was performed to investigate the oxidation state of the elements in the as-formed nanoparticles, and their coupling with the rEGO support. As shown in Figure 3m, the Ru 3d high-resolution spectrum of the as-synthesized RuO2-rEGO without heat treatment shows a set of doublet peaks located at 280.9 and 284.9 eV, corresponding to the doublet peaks for Ru (IV) 3d5/2 and 3d3/2, respectively. [59] Owing to the strong interference of Ru 3d and C 1s signals, the comparison of survey, Ru 4d and Ru 3p spectra (Figure S20a, b and c, respectively) for RuO2-rEGO, commercial RuO2 (CM RuO2) and metallic Ru was conducted to identify the oxidation state of Ru. In comparison to the Ru 3p3/2 peaks of CM RuO2 at 462.5 eV and metallic Ru at 461.5 eV (Figure S20c), the Ru 3p3/2 of RuO2-rEGO contains only Ru (IV) with Ru 3p3/2 peak located at 462.7 eV. This indicates that the RuO2-rEGO is dominated by Ru (IV) with a negligible amount of Ru (0) (details in Table S3). The comparison of Ru 3d (Figure S20b) also indicates the dominance of Ru (IV) in RuO2-rEGO.
In addition, although the C 1s signal is overlapped with that of Ru 3d3/2, three components can be deconvoluted as shown in Figure 3m (Figure S20d), suggesting a lower electron density at the oxygen sites due to the electron transfer from the oxygen to Ru atoms; [60] (3) the peak identified at 528.6 eV in O 1s spectrum is in agreement with the bridged O connecting Ru and C as-reported in the literature. [61] For the Pt 4f high-resolution spectrum of Pt-rEGO (Figure 3n), the set of doublet peaks located at 71.2 and 74.6 eV are ascribed to the surface Pt atoms, while the doublet peaks at higher binding energies of 72.3 and 75.6 eV correspond to the bulk atoms of Pt. [62] Further, inductively coupled plasmaoptical emission spectrometry (ICP-OES) reveals that the mass loadings of RuO2 and Pt in the as-prepared composites are 41.6 ± 0.9 wt.% and 10.2 ± 1.0 wt.%, respectively.
The laser-induced solution approach leads to deeply reduced rEGO modified with ultrafine catalyst nanoparticles in a single step, as demonstrated above using RuO2 and Pt as model materials. To further prove the versatility of this laser-assisted approach in deposition of various types of functional nanoparticles on the surface of rEGO sheets, PtPd alloy and MnOx nanoparticles have also been successfully deposited using H2PtCl6/Na2PdCl4 and MnCl2 as precursor metal salts, respectively. The corresponding TEM characterizations for PtPd/rEGO and MnOx/rEGO are available in Figure S21 and Figure S22, respectively. Based on the experimental results and literature knowledge, [23,25,31]  where hvb + and ecb − are photo-generated holes and electrons, respectively. The subsequent reduction of EGO by photo-generated electrons occurs following: EGO + e cb − → rEGO + OH − For reductive photodeposition of metal (M) nanoparticles, the reduction of metal ions happens as: While for the oxidative photodeposition of metal oxide nanoparticles, the reaction could occur through: The excess photo-generated holes and electrons are consumed by the sacrificial electron donor (D) and acceptor (A), respectively, as follow: In an actual photodeposition process, the consumption of excess electrons leads to the reduction of protons to form H2 gas, while the consumption of excess holes could oxidize water to from O2. [25] If sacrificial organic agents (e.g. methanol, isopropanol) are added, their reactions with photo-generated holes/electrons form highly reducing radical species, which participate in the reduction of metal ions. [25,31] In the present work, isopropanol and acetone were added as sacrificial electron donor and acceptor, respectively, leading to the formation of highly reducing carbon centered isopropanol radicals, [31] which further benefit a fast and thorough reduction of EGO and metal ions. In addition, these highly reducing radicals could also be the reason that metallic Ru is present in the RuO2-rEGO product.

Electrocatalytic performance
To demonstrate the applications of the as-prepared RuO2-rEGO and Pt-rEGO composites, their performances as electrocatalysts have been measured and evaluated. To obtain the intrinsic catalytic activities of the samples for OER, HER and ORR, the ohmic-drop correction was carried out to minimize the effects of solution resistance (Figure S23).  4b). These results suggest the OER activity is profoundly affected by the crystallinity of RuO2. Note that it is known that metallic Ru nanoparticles are unstable and dissolve completely during the first OER polarization, [63] and this dissolution of Ru is more severe in alkaline electrolytes than in acidic electrolytes. [64] Therefore, the small content of Ru nanoparticles in RuO2-rEGO catalyst would dissolve rapidly in the first OER polarization and have minimum influence on the subsequent evaluation of catalytic activity. This has been confirmed by the first 11 OER CV scans recorded with the RuO2-rEGO catalyst (Figure S25), the current due to Ru oxidation/dissolution appears only in the first anodic scan, and the CVs overlap with each other after the second cycle. As a benchmark, the OER performance of CM RuO2 (Premion, Alfa Aesar) was measured under the same conditions. Notably, CM RuO2 requires an overpotential of 283 mV to reach the current density of 10 mA cm −2 (Figure 4a), with RuO2 loading 2.4 times higher than that of RuO2-rEGO (42.5 wt.% of RuO2). In addition, as shown in Figure 4b, the RuO2-rEGO-250HT shows a smaller Tafel slope (50 mV dec −1 ) than that of CM RuO2 (53 mV dec −1 ). Chronopotentiometric testing at 10 mA cm −2 was used to evaluate the durability of RuO2-rEGO-250HT catalyst, as shown in the inset of Figure   4a, the overpotential of RuO2-rEGO-250HT increased by only 14.1 mV after 3 hours testing.
The comparison of linear sweep voltammograms (LSVs) (Figure S26a) and Tafel plots ( Figure S26b) for the RuO2-rEGO-250HT catalyst before and after chronopotentiometric test shows only slight degradation of catalytic activity.
The electrochemically active surface area (ECSA) of the catalysts has been estimated by measuring the double-layer capacitance as reported in the literature. [65] The CDL was determined from the cyclic voltammetry (CV) scans in the non-Faradaic region between −0.1 and 0.1 V vs Ag/AgCl in 1 м KOH aqueous electrolyte ( Figure S23 c and d). The ECSA was estimated to be 50.2 and 47.4 cm 2 for RuO2-rEGO-250HT and CM RuO2, respectively, indicating that the RuO2-rEGO composite with 42.5 wt.% of RuO2 has a slightly higher accessible surface area to electrolyte in comparison with CM RuO2 (100 wt.% RuO2 loading).
In addition, the specific surface areas derived from nitrogen adsorption/desorption isotherms ( Figure S27) also suggest that the RuO2/rEGO has a higher specific surface area (261.8 m 2 g −1 ) than that of CM RuO2 (49.1 m 2 g −1 ). Electrochemical impedance spectroscopy (EIS) was conducted at 1.5 V vs RHE for the as-prepared catalysts ( Figure S28). Interestingly, the charge transfer resistances derived from the Nyquist plots for RuO2-rEGO composites (6.72 and 10.59 Ω for RuO2-rEGO-250HT and RuO2-rEGO-200HT, respectively) are significantly smaller than that of CM RuO2 (18.05 Ω). The enhanced charge transfer in the RuO-rEGO composites is attributed to the highly conductive rEGO support, which provides fast electron transfer routes. In addition, the smaller charge transfer resistance for RuO2-rEGO composite annealed at 250 °C, compared to that annealed at 200 °C, is probably due to increased electrical conductivity of RuO2 itself at higher annealing temperatures. [66] As revealed in Figure 4c, the RuO2-rEGO-250HT exhibits much higher (one order of magnitude) mass activity and turnover frequency (TOF) than that of the CM RuO2 (A detailed TOF calculation is given in Supporting Information). This further confirms the enhancement of intrinsic catalyst activity for the RuO2-EGO composites. The OER performance of RuO2-rEGO-250HT has been further compared with the state-of-the-art Ru-based catalysts reported in the literature. The overpotential of RuO2-rEGO-250HT at 10 mA cm −2 and the Tafel slope are outperforming the majority of the literature values (Figure 4d and Table S4). The low overpotential, small Tafel slope, high mass activity and TOF, together with the good durability of the RuO2-rEGO composites indicate their great potential as electrocatalysts for OER.
RuO2 is known to be a very promising electrode material for supercapacitors due to its large specific capacitance (experimental values up to 720 F g −1 ) and good rate capability. [66] However, the rarity of Ru in Earth's crust results in the high cost of RuO2, which limits its wider application in the real world. This challenge could be possibly solved by developing composites of RuO2 and carbon, which can reduce the usage of RuO2 while maintaining high specific capacitance. Hence, the electrochemical capacitance of RuO2-rEGO composites (RuO2 loading: 41.6 ± 0.9 wt.%) has been measured and evaluated.
Initially, the effect of annealing temperature on the electrochemical capacitance of the RuO2-rEGO composite was investigated ( Figure S29). The specific capacitance increases with the rise of annealing temperature and maximizes at 200 °C, further increase of the annealing temperature causes deterioration of the specific capacitance. This phenomenon is consistent with previous works and can be explained by the balancing of ionic (proton), and electronic conductivity of hydrous RuO2 during annealing (detailed discussion in Figure S29). [66]  mA cm −2 ). This is caused by retarded diffusion of dissolved O2 through the stacked graphene sheets. [67] Effective prevention of graphene restacking could lead to further enhancement in performance, but it remains a big challenge. Nevertheless, the 2D rEGO flakes act as barriers to prevent the leaching and dissolution of Pt, [67] leading to significantly improved ORR durability for Pt-rEGO catalyst (76% retention after ca. 5.6 hours test) compared with that of CM CM Pt/C (63% retention after ca. 5.6 hours; see Figure S31a for details).
The Pt-rEGO exhibits superior catalytic activity for HER when compared with the CM Pt/C.
The HER activity of Pt-rEGO was measured in N2 saturated 0.5 м H2SO4 electrolyte. Figure   5c shows

Supporting Information
Supporting Information is available from the Wiley Online Library or from the author. b) The photoexcitation of the semiconducting EGO creates electron-hole pairs, which reduce/oxidize the metal ions (M n+ ) into metal/metal oxide; the metal/metal oxide nucleates and grows on the EGO substrate as nanoparticles; simultaneously, the EGO is reduced by photo-generated electrons to rEGO; Pt (left) and RuO2 (right) nanoparticels (NPs) are shown as expamples for the reductive and oxidative photodeposition, respectively; the photo-generated holes for reductive photodeposition are consumed by sacrifical electron donor (D). c) and d) The as-synthesized graphene supported RuO2 and Pt nanoparticles, respectively, are used for electrochemical energy storage and conversion. Figure 2. a) and b) Tauc plots derived from the UV-vis spectra of pristine EGO for determination of bandgaps for direct and indirect transitions, respectively. c) Typical Raman spectra of EGO after 248 nm laser irradiation at various fluences, d) the evolution of ID/IG ratio with the increase of laser fluence. e) and f) FWHM values of D and G band for EGO treated with various laser fluences. g) Photographs of pristine EGO and rEGO after laser irradiation at various fluences. h) XPS high resolution C 1s spectra of pristine EGO and rEGO reduced by laser irradiation at 568 mJ cm −2 . i) Comparison of typical Raman spectra for pure rEGO, RuO2-rEGO and Pt-rEGO composites prepared using aser irradiation at 568 mJ cm −2 , the ID/IG ratio displayed is an average value of five spectra.   Table S1 in Supporting Information. e) CVs recorded at 20 mV s −1 and f) specific capacitances at various current densities for the RuO2-rEGO-200HT and CM RuO2 electrodes in 1 м H2SO4 aqueous electrolyte. Graphene supported electrocatalysts, including RuO2 and Pt, have been synthesized by laser (248 nm) irradiation of semiconducting electrochemical graphene oxide (EGO) and metal salts precursor solutions. The catalysts show superior electrocatalytic activities for oxygen evolution reaction (OER) and hydrogen evolution reaction (HER), which is attributed to the homogeneous distribution of ultrafine nanoparticles (~2 nm) on the deeply reduced EGO support.