Sulfur-Doped Carbon Nanohorn Bifunctional Electrocatalyst for Water Splitting

Sulfur-doped carbon nanohorns (S-CNHs) were prepared by an easy one-pot solvothermal process and were employed as efficient electrocatalysts towards water splitting. Initially, oxidation of CNHs followed by thermal treatment with the Lawesson’s reagent resulted in the formation of S-CNHs with the sulfur content determined as high as 3%. The S-CNHs were thoroughly characterized by spectroscopic, thermal and electron microscopy imaging means and then electrocatalytically screened. Specifically, S-CNHs showed excellent activity and durability for both O2 and H2 evolution reactions, by showing low overpotential at 1.63 and −0.2 V vs. RHE for oxygen and hydrogen evolution reaction, respectively. Additionally, S-CNHs showed significantly lower Tafel slope value and lower current resistance compared to oxidized and pristine CNHs for both electrocatalytic reactions. The outstanding electrocatalytic properties and high conductivity, along with the high S-doping level, render S-CNHs a promising bifunctional electrocatalyst for water splitting.


Introduction
Electrochemical water splitting is a cost-effective clean method for breaking down water to hydrogen (H 2 ) and oxygen (O 2 ) and is absolutely important in hydrogen economy. In a water electrolyzer, water electrolysis is separated into two half-reactions: hydrogen evolution reaction (HER) happens at the cathode, i.e., 2H + + 2e − → H 2 (under acidic conditions), while oxygen evolution reaction (OER) occurs at the anode, i.e., 2H 2 O → O 2 + 4H + + 4e − . Notably, OER is more kinetically sluggish compared to HER, as it is a four-electron transfer reaction, while HER requires only two electrons. To date, both of these half-reactions have relied on noble-metal-based electrocatalysts to reach their best performance, namely Pt for HER, and IrO 2 and RuO 2 for OER [1][2][3][4]. Despite their high activity, noble-metal-based electrocatalysts are expensive, and this limits their employment in real applications. In this context, progress has been made towards the exploration of non-precious-based metal electrocatalysts, while recent trends want transition metal-based and carbonaceous electrocatalysts to reach low overpotentials for HER and OER [5][6][7][8][9][10][11][12][13]. Moreover, due to the different mechanisms that govern HER and OER, it is quite often a good electrocatalyst that can perform at lower overpotential for HER to show poor activity for OER and vice versa. Therefore, it still remains quite a challenge to effectively employ highly active, stable, abundant and low-cost bifunctional electrocatalysts for water electrolysis that compete the performance of the precious metals ones [1,14].

General
Chemicals, reagents, and solvents were purchased from Sigma-Aldrich and used without further purification. Infrared (IR) spectra were obtained on a Fourier Transform IR spectrometer (Equinox 55 from Bruker Optics, Billerica, Massachusetts, US) equipped with a single reflection diamond attenuated total reflection (ATR) accessory (DuraSamp1IR II by SensIR Technologies, Danbury, CT, USA). Raman measurements were recorded with a Renishaw confocal spectrometer (Wotton-under-Edge, UK) at 514 nm. The data were obtained and analyzed with Renishaw Wire and Origin software (OriginPro 2018, Northampton, MA, USA). Thermogravimetric analysis was performed using a TGA Q500 V20.2 Build 27 instrument by TA (New Castle, DE, USA) in a nitrogen (purity > 99.999%) inert atmosphere. High-resolution TEM was performed employing an imaging-side aberration-corrected FEI Titan-Cube microscope working at 80 kV, equipped with a Cs corrector (CETCOR from CEOS GmbH, Heidelberg, Germany). XPS were recorded on a Kratos Axis Supra spectrometer, (Manchester, UK) equipped with a monochromated Al Kα X-ray source using an analyzer pass energy of 160 eV for survey spectra and 20 eV for the core level spectra. Spectra were recorded by setting the instrument to the hybrid lens mode and the slot mode providing approximately a 700 × 300 µm 2 analysis area using charge neutralization. Regions were calibrated using the reference value BE(C 1s sp2) = 284.5 eV.
All XPS spectra were analyzed using CASA XPS software (2.3.23, Teignmouth, Devon, UK). The XPS peaks were fitted to GL(70) Voigt lineshape (a combination of 70% Gaussian and 30% Lorentzian character), after performing a Shirley background subtraction. For the light-assisted oxidation reaction of CNHs the light source used was a 500 W halogen lamp, which was positioned 20 cm away from the reactor. Autoclave synthesis was undertaken in a Parr autoclave reactor with a 300 mL Teflon-lined stainless-steel vessel. All electrochemical measurements were carried out using an Metrohm Autolab PGSTAT128N potentiostat/galvanostat and were carried out at room temperature in a standard three-compartment electrochemical cell by using a platinum wire as a counter-electrode, an RDE with glassy carbon disk (geometric surface area: 0.0196 cm 2 ) as a working electrode, and Hg/HgSO 4 (0.5 M K 2 SO 4 ) and Hg/HgO (0.1 M KOH) (AMEL S.r.l., Milan, Italy) as reference electrodes for HER and OER, respectively. LSV measurements for HER were carried out at room temperature in N 2 -saturated 0.5 M H 2 SO 4 . LSV measurements for OER were carried out at room temperature in O 2 -saturated 1.0 M KOH. The catalyst ink was prepared by dispersing 4.0 mg of the catalytic powder in a 1 mL mixture of water, isopropanol, and 5% Nafion (v/v/v = 4:1:0.02) and sonicated for 30 min prior to use. Before casting the electrocatalytic ink on the electrode's surface, the working electrode was polished with 6-, 3-and 1-mm diamond pastes, rinsed with deionized water, and sonicated in double-distilled water. Afterwards, 8.5 µL aliquots of the electrocatalyst were casted on the electrode surface and were left to dry at room temperature. Electrochemical Impedance Spectroscopy (EIS) measurements (Metrohm Autolab, Utrecht, The Netherlands)) were conducted from 10 5 to 10 −1 Hz with an AC amplitude of 0.01 V, at a potential where significant HER and OER current was recorded, at −2 mA/cm 2 for HER and at 3 mA/cm 2 for OER. EIS data were fitted to Randles circuit. In EIS fit and simulation Chi square value is <0.001. The data were obtained and analyzed with Nova 2.1.4 software (Metrohm Autolab, Utrecht, The Netherlands).

Results and Discussion
The preparation of S-doped CNHs was accomplished by treatment of oxidized CNHs (ox-CNHs) with Lawesson's reagent, which was used as sulfur source, in the autoclave reactor, as shown in Figure 1. The doping of S atoms occurs mostly at the conical tips of ox-CNHs, where carbonyl groups are abundantly present. The doping mechanism can be briefly described as follows: initially, thioketone groups are formed, and afterwards, S atoms replace C atoms in the graphitic network [20,21]. ATR-IR and Raman spectra give proof of the successful realization of S-doped CNHs. Specifically, while the IR spectrum of ox-CNHs is governed by the characteristic carbonyl-related band at 1705 cm −1 due to stretching vibrations of the carboxylic acid moieties, the corresponding IR spectrum due to S-CNHs shows the decrease of the carbonyl band and the evolution of a new broad band at 1080 cm −1 related to C-S vibrations ( Figure 2a). On the other hand, the Raman spectrum of pristine CNHs exhibits the characteristic D-band at 1341 cm −1 , associated with defects, and the G-band at 1591 cm −1 , related to the sp 2 graphitic network. In the Raman spectrum of ox-CNHs, an increase of the intensity ratio I D/G compared to pristine CNHs from 1.3 to 1.7, due to the change of hybridization from sp 2 to sp 3 upon oxidation, is evident. Concerning S-CNHs, there is a decrease of the I D/G to 1.0 ( Figure 2b) due to S-doping and partial restoration of the graphitic network [31]. Additionally, there is a shift of the G-band to lower frequencies by 8 cm −1 , which is indicative of n-doping in S-CNHs [20]. Nanomaterials 2020, 10, x FOR PEER REVIEW 3 of 12 an analyzer pass energy of 160 eV for survey spectra and 20 eV for the core level spectra. Spectra were recorded by setting the instrument to the hybrid lens mode and the slot mode providing approximately a 700 × 300 μm 2 analysis area using charge neutralization. Regions were calibrated using the reference value BE(C 1s sp2) = 284.5 eV. All XPS spectra were analyzed using CASA XPS software (2.3.23, Teignmouth, Devon, UK). The XPS peaks were fitted to GL(70) Voigt lineshape (a combination of 70% Gaussian and 30% Lorentzian character), after performing a Shirley background subtraction. For the light-assisted oxidation reaction of CNHs the light source used was a 500 W halogen lamp, which was positioned 20 cm away from the reactor. Autoclave synthesis was undertaken in a Parr autoclave reactor with a 300 mL Teflon-lined stainless-steel vessel. All electrochemical measurements were carried out using an Metrohm Autolab PGSTAT128N potentiostat/galvanostat and were carried out at room temperature in a standard three-compartment electrochemical cell by using a platinum wire as a counter-electrode, an RDE with glassy carbon disk (geometric surface area: 0.0196 cm 2 ) as a working electrode, and Hg/HgSO4 (0.5 M K2SO4) and Hg/HgO (0.1 M KOH) (AMEL S.r.l., Milan, Italy) as reference electrodes for HER and OER, respectively. LSV measurements for HER were carried out at room temperature in N2-saturated 0.5 M H2SO4. LSV measurements for OER were carried out at room temperature in O2-saturated 1.0 M KOH. The catalyst ink was prepared by dispersing 4.0 mg of the catalytic powder in a 1 mL mixture of water, isopropanol, and 5% Nafion (v/v/v = 4:1:0.02) and sonicated for 30 min prior to use. Before casting the electrocatalytic ink on the electrode's surface, the working electrode was polished with 6-, 3-and 1-mm diamond pastes, rinsed with deionized water, and sonicated in double-distilled water. Afterwards, 8.5 μL aliquots of the electrocatalyst were casted on the electrode surface and were left to dry at room temperature. Electrochemical Impedance Spectroscopy (EIS) measurements (Metrohm Autolab, Utrecht, The Netherlands)) were conducted from 10 5 to 10 −1 Hz with an AC amplitude of 0.01 V, at a potential where significant HER and OER current was recorded, at −2 mA/cm 2 for HER and at 3 mA/cm 2 for OER. EIS data were fitted to Randles circuit. In EIS fit and simulation Chi square value is <0.001. The data were obtained and analyzed with Nova 2.1.4 software (Metrohm Autolab, Utrecht, The Netherlands).

Results and Discussion
The preparation of S-doped CNHs was accomplished by treatment of oxidized CNHs (ox-CNHs) with Lawesson's reagent, which was used as sulfur source, in the autoclave reactor, as shown in Figure 1. The doping of S atoms occurs mostly at the conical tips of ox-CNHs, where carbonyl groups are abundantly present. The doping mechanism can be briefly described as follows: initially, thioketone groups are formed, and afterwards, S atoms replace C atoms in the graphitic network [20,21]. ATR-IR and Raman spectra give proof of the successful realization of S-doped CNHs. Specifically, while the IR spectrum of ox-CNHs is governed by the characteristic carbonyl-related band at 1705 cm −1 due to stretching vibrations of the carboxylic acid moieties, the corresponding IR spectrum due to S-CNHs shows the decrease of the carbonyl band and the evolution of a new broad band at 1080 cm −1 related to C-S vibrations ( Figure 2a). On the other hand, the Raman spectrum of pristine CNHs exhibits the characteristic D-band at 1341 cm −1 , associated with defects, and the G-band at 1591 cm −1 , related to the sp 2 graphitic network. In the Raman spectrum of ox-CNHs, an increase of the intensity ratio ID/G compared to pristine CNHs from 1.3 to 1.7, due to the change of hybridization from sp 2 to sp 3 upon oxidation, is evident. Concerning S-CNHs, there is a decrease of the ID/G to 1.0 ( Figure 2b) due to S-doping and partial restoration of the graphitic network [31]. Additionally, there is a shift of the G-band to lower frequencies by 8 cm −1 , which is indicative of n-doping in S-CNHs [20]. under nitrogen atmosphere. Pristine CNHs present an almost stable thermal profile up to 800 °C, while ox-CNHs show a 9% mass loss up to 500 °C due to the thermal decomposition of the carboxylic groups. Conversely, this is not shown in the thermograph of S-CNHs, where the thermal mass loss starts after 330 °C. The mass loss at higher temperatures is attributed to the thermal decomposition of the graphitic framework at sp 3 defects close to where sulfur doping takes place. HR-TEM imaging and XPS were employed to determine the morphology and the chemical composition and elemental content of S-CNHs. In more detail, HR-TEM micrographs (Figure 3a,b) show the characteristic morphology of those nanostructures and their aggregates' organization. In the XPS survey spectrum of S-CNHs, characteristic peaks due to S2p, C1s, O1s located at binding energies of ca. 164, 285 and 533, respectively, are evident ( Figure 3c). Moreover, the S content was determined to be 3%, higher than that reported for S-doped graphene [31] and previously prepared S-doped CNHs [26]. To probe the chemical state of sulfur in S-CNHs high-resolution S2p peaks were analyzed ( Figure 3d). The S2p peaks can be resolved into four peaks at binding energies of 162.2 (8.4%), 164.1 (70.6%), 165.3 (11.3%) and 168.4 (9.8%) eV, respectively. While the second and third deconvoluted peaks are related to sulfur bound as heteroatoms to aromatics, such as thiophene [31,32], the peaks at 162.2 and 168.4 eV arise from thiol and sulfone components [31,32]. The high percentage of the peaks corresponding to C-S related bonds confirms that doping of sulfur atoms happens mostly at the conical tips of CNHs where five-membered rings are located [16].  Figure 2c shows the thermographs of pristine CNHs, ox-CNHs and S-CNHs that were recorded under nitrogen atmosphere. Pristine CNHs present an almost stable thermal profile up to 800 • C, while ox-CNHs show a 9% mass loss up to 500 • C due to the thermal decomposition of the carboxylic groups. Conversely, this is not shown in the thermograph of S-CNHs, where the thermal mass loss starts after 330 • C. The mass loss at higher temperatures is attributed to the thermal decomposition of the graphitic framework at sp 3 defects close to where sulfur doping takes place.
HR-TEM imaging and XPS were employed to determine the morphology and the chemical composition and elemental content of S-CNHs. In more detail, HR-TEM micrographs (Figure 3a,b) show the characteristic morphology of those nanostructures and their aggregates' organization. In the XPS survey spectrum of S-CNHs, characteristic peaks due to S 2p , C 1s , O 1s located at binding energies of ca. 164, 285 and 533, respectively, are evident ( Figure 3c). Moreover, the S content was determined to be 3%, higher than that reported for S-doped graphene [31] and previously prepared S-doped CNHs [26]. To probe the chemical state of sulfur in S-CNHs high-resolution S 2p peaks were analyzed (Figure 3d). The S 2p peaks can be resolved into four peaks at binding energies of 162.2 (8.4%), 164.1 (70.6%), 165.3 (11.3%) and 168.4 (9.8%) eV, respectively. While the second and third deconvoluted peaks are related to sulfur bound as heteroatoms to aromatics, such as thiophene [31,32], the peaks at 162.2 and 168.4 eV arise from thiol and sulfone components [31,32]. The high percentage of the peaks corresponding to C-S related bonds confirms that doping of sulfur atoms happens mostly at the conical tips of CNHs where five-membered rings are located [16].
Next, evaluating the electrocatalytic properties of S-CNHs against OER, linear sweep voltammetry (LSV) measurements were performed in O 2 saturated 1.0 M aqueous KOH. A very important factor for evaluating the performance of OER electrocatalysts is the onset potential. However, since it is hard to determine the exact value, the potential at 10 mA/cm 2 is commonly used as a benchmark. In OER, overpotential is the potential difference between the potential registered at a specific current density and the equilibrium potential, i.e., 1.23 V [1,2]. Figure 4a screens the polarization curves of S-CNHs, ox-CNHs and pristine CNHs. It is clear from the LSV curves and the recorded values of overpotentials registered at 10 mA/cm 2 current density ( Table 1) that S-CNHs possess the best electrocatalytic activity with the lowest potential at 1.63 V vs. RHE. Such a low overpotential is considered excellent for the OER, and is close to the overpotentials registered for noble metals [1,2,12]. On the contrary, the corresponding potential for both ox-CNHs and pristine CNHs was registered at 1.82 vs. RHE, 220 mV higher than that of S-CNHs. At the same time, S-CNHs reached high current density values compared to the other two reference materials, which is a prerequisite for an effective OER electrocatalyst. The superior electrocatalytic activity of S-CNHs is attributed to the presence of sulfur dopants, which induce a non-uniform spin density distribution due to the mismatch of the outermost orbitals of sulfur and carbon atoms, while further altering the charge density at nearby sites.  Next, evaluating the electrocatalytic properties of S-CNHs against OER, linear sweep voltammetry (LSV) measurements were performed in O2 saturated 1.0 M aqueous KOH. A very important factor for evaluating the performance of OER electrocatalysts is the onset potential. However, since it is hard to determine the exact value, the potential at 10 mA/cm 2 is commonly used as a benchmark. In OER, overpotential is the potential difference between the potential registered at a specific current density and the equilibrium potential, i.e., 1.23 V [1,2]. Figure 4a screens the polarization curves of S-CNHs, ox-CNHs and pristine CNHs. It is clear from the LSV curves and the recorded values of overpotentials registered at 10 mA/cm 2 current density ( Table 1) that S-CNHs possess the best electrocatalytic activity with the lowest potential at 1.63 V vs. RHE. Such a low overpotential is considered excellent for the OER, and is close to the overpotentials registered for noble metals [1,2,12]. On the contrary, the corresponding potential for both ox-CNHs and pristine CNHs was registered at 1.82 vs. RHE, 220 mV higher than that of S-CNHs. At the same time, S-CNHs reached high current density values compared to the other two reference materials, which is a prerequisite for an effective OER electrocatalyst. The superior electrocatalytic activity of S-CNHs is attributed to the presence of sulfur dopants, which induce a non-uniform spin density distribution due to the mismatch of the outermost orbitals of sulfur and carbon atoms, while further altering the charge density at nearby sites.   Next, evaluating the electrocatalytic properties of S-CNHs against OER, linear sweep voltammetry (LSV) measurements were performed in O2 saturated 1.0 M aqueous KOH. A very important factor for evaluating the performance of OER electrocatalysts is the onset potential. However, since it is hard to determine the exact value, the potential at 10 mA/cm 2 is commonly used as a benchmark. In OER, overpotential is the potential difference between the potential registered at a specific current density and the equilibrium potential, i.e., 1.23 V [1,2]. Figure 4a screens the polarization curves of S-CNHs, ox-CNHs and pristine CNHs. It is clear from the LSV curves and the recorded values of overpotentials registered at 10 mA/cm 2 current density ( Table 1) that S-CNHs possess the best electrocatalytic activity with the lowest potential at 1.63 V vs. RHE. Such a low overpotential is considered excellent for the OER, and is close to the overpotentials registered for noble metals [1,2,12]. On the contrary, the corresponding potential for both ox-CNHs and pristine CNHs was registered at 1.82 vs. RHE, 220 mV higher than that of S-CNHs. At the same time, S-CNHs reached high current density values compared to the other two reference materials, which is a prerequisite for an effective OER electrocatalyst. The superior electrocatalytic activity of S-CNHs is attributed to the presence of sulfur dopants, which induce a non-uniform spin density distribution due to the mismatch of the outermost orbitals of sulfur and carbon atoms, while further altering the charge density at nearby sites.   To define how swiftly the current increases with the applied overpotential, Tafel slopes were extracted from the LSV curves. Specifically, S-CNHs showed a low Tafel slope value of 80 mV/dec, as shown in Figure 4b. Low Tafel slope values are required for ideal electrocatalysts as they result in an improved OER rate under a constant increase of overpotential, which can be beneficial in real applications. The two reference materials, ox-CNHs and pristine CNHs, accordingly showed much higher Tafel slope values of 202 and 333 mV/dec, which correspond to slower OER kinetics. Electrochemical impedance spectroscopy (EIS) assays were performed to evaluate the OER kinetics. S-CNHs showed the smallest frequency semicircle in the Nyquist plot, which corresponds to a small charge transfer resistance (R ct ) value of 10.8 Ω (Figure 4c). On the contrary, the very high R ct values of the ox-CNHs and pristine CNHs, 560.5 and 450.5 Ω, respectively, reveal high charge transfer resistance. The higher R ct value of ox-CNHs compared to pristine CNHs is justified by the disruption of the graphitic network due to the introduction of carboxylic species. Conversely, the low charge transfer resistance that was reported for S-CNHs arises from the increased conductivity due to the improvement of the graphitic degree resulting from the sulfur doping.
Stability is a crucial factor for the evaluation of the overall performance of an electrocatalyst. Figure 4a shows the LSV polarization curves for OER of S-CNHs as compared to that for ox-CNHs and pristine CNHs, after continuous cycling for 2000 cycles. Notably, negligible loss of the anodic current is noted for S-CNHs. As a striking difference, the two reference materials exhibit extremely huge losses of the anodic current that are unable to reach current density values at 10 mA/cm 2 . Electrochemical OER data for S-CNHs, ox-CNHs and pristine CNHs before and after 2000 cycles are summarized in Table 1.
To get a better understanding of the charge transport flow in S-CNHs in OER, the electrochemically active surface area (ECSA) was estimated. In more detail, ECSA values were calculated based on the following equation [20,[33][34][35][36]: ECSA = C dl /C s , where C dl stands for the electrochemical double-layer capacitance, while C s is the specific capacitance of a flat surface with 1 cm 2 of real surface area with a value assumed to be 40 µF/cm 2 for the flat electrode. For the estimation of the ECSA, cyclic voltamographs were recorded in a non-Faradaic region at scan rates of 50, 100, 200, 300, 400 and 500 mV/sec for all tested materials before and after 2000 cycles ( Figure 5). ECSA values were obtained from C dl by plotting the ∆j = (j a − j c ) at a given potential versus the scan rate as stated in the equation C dl = d(∆j)/2dV b . Specifically, S-CNHs showed higher ECSA value of~67.5 cm 2 . Consistent with the LSV polarization curves for OER, ox-CNHs and pristine CNHs present significantly lower ECSA, ca.~60 and~19.4 cm 2 , respectively. These values are in accordance with the electrocatalytic results towards OER, since the efficient accessibility of the active sites is related to higher ECSA values. Moreover, after cycling continuously for 2000 cycles, a slight decrease of the ECSA was noted for S-CNHs,~61.25 cm 2 , while ox-CNHs and pristine CNHs exhibited huge drops at~13.75 and~4.7 cm 2 , respectively.
Next, the electrochemical properties of S-CNHs against HER were evaluated by performing LSV measurements in N 2 -saturated 0.5 M aqueous H 2 SO 4 . The polarization curves of S-CNHs, along with ox-CNHs and pristine CNHs as well as commercially available 20 wt% Pt on carbon black (Pt/C) as reference materials, can be seen in Figure 6a. The lowest onset potential is registered at −0.2 V vs. RHE for S-CNHs, which is more positive by 50 and 120 mV in comparison with the values noted for ox-CNHs (i.e., at −0.25 V) and pristine CNHs (i.e., at −0.32 V), respectively. In addition, potential values for HER are often compared at the functional current density for electrochemical water splitting of −10 mA/cm 2 , where sufficient hydrogen production is achieved, while faulty interpretations from intrinsic electroactivity are prevented. It is clear that S-CNHs show higher electrocatalytic activity at −0.31 V vs. RHE, shifted by 120 and 440 mV to more positive potentials compared to the value recorded for ox-CNHs (i.e., at −0.43 V) and pristine CNHs (i.e., at −0.75 V), respectively. Furthermore, to define the rate-limiting step of HER, Tafel slopes were extracted from the LSV curves (Figure 6b). Additional analysis reveals that S-CNHs present a Tafel slope of 82 mV/dec, which is slightly higher than the one of Pt/C (i.e., 35 mV/dec). The latter demonstrates that the electrochemical desorption of adsorbed hydrogen atoms onto the electrode (Heyrovsky desorption) is the rate-limiting step for HER. On the other hand, HER of ox-CNHs and pristine CNHs is rate-limited by the initial adsorption of a proton onto electrode's surface (Volmer adsorption), which is screened by the much higher Tafel slope values of 240 and 320 mV/dec, respectively. The easier kinetics of HER, revealed from the much lower Tafel slope for S-CNHs, signifies more reactive/accessible protons and water molecules that reach the electrochemical active sites [33]. Next, the electrochemical properties of S-CNHs against HER were evaluated by performing LSV measurements in N2-saturated 0.5 M aqueous H2SO4. The polarization curves of S-CNHs, along with ox-CNHs and pristine CNHs as well as commercially available 20 wt% Pt on carbon black (Pt/C) as reference materials, can be seen in Figure 6a. The lowest onset potential is registered at −0.2 V vs. RHE for S-CNHs, which is more positive by 50 and 120 mV in comparison with the values noted for ox-CNHs (i.e., at −0.25 V) and pristine CNHs (i.e., at −0.32 V), respectively. In addition, potential values for HER are often compared at the functional current density for electrochemical water splitting of −10 mA/cm 2 , where sufficient hydrogen production is achieved, while faulty interpretations from intrinsic electroactivity are prevented. It is clear that S-CNHs show higher electrocatalytic activity at −0.31 V vs. RHE, shifted by 120 and 440 mV to more positive potentials compared to the value recorded for ox-CNHs (i.e., at −0.43 V) and pristine CNHs (i.e., at −0.75 V), respectively. Furthermore, to define the rate-limiting step of HER, Tafel slopes were extracted from the LSV curves (Figure 6b). Additional analysis reveals that S-CNHs present a Tafel slope of 82 mV/dec,  Next, the electrochemical properties of S-CNHs against HER were evaluated by performing LSV measurements in N2-saturated 0.5 M aqueous H2SO4. The polarization curves of S-CNHs, along with ox-CNHs and pristine CNHs as well as commercially available 20 wt% Pt on carbon black (Pt/C) as reference materials, can be seen in Figure 6a. The lowest onset potential is registered at −0.2 V vs. RHE for S-CNHs, which is more positive by 50 and 120 mV in comparison with the values noted for ox-CNHs (i.e., at −0.25 V) and pristine CNHs (i.e., at −0.32 V), respectively. In addition, potential values for HER are often compared at the functional current density for electrochemical water splitting of −10 mA/cm 2 , where sufficient hydrogen production is achieved, while faulty interpretations from intrinsic electroactivity are prevented. It is clear that S-CNHs show higher electrocatalytic activity at −0.31 V vs. RHE, shifted by 120 and 440 mV to more positive potentials compared to the value recorded for ox-CNHs (i.e., at −0.43 V) and pristine CNHs (i.e., at −0.75 V), respectively. Furthermore, to define the rate-limiting step of HER, Tafel slopes were extracted from the LSV curves (Figure 6b). Additional analysis reveals that S-CNHs present a Tafel slope of 82 mV/dec, Figure 6. (a) LSVs, before (solid lines) and after 10,000 cycles (dashed lines) for HER, (b) Tafel slopes, and (c) Nyquist plots of S-CNHs (red), ox-CNHs (black), pristine CNHs (grey) and Pt/C (blue). The LSV polarization curves obtained at 1600 rpm rotation speed and 5 mV/s scan rate.
Next, EIS assays were performed to evaluate the HER kinetics. EIS measurements further confirm the superior HER kinetics of S-CNHs (Figure 6c), which show the smallest R ct value of 30.4 Ω, comparable to that of Pt/C (i.e., 6.1 Ω). On the other hand, ox-CNHs and pristine CNHs exhibited much higher R ct values, 197 and 123.4 Ω, respectively. This tendency is in line with the results of Tafel slope values, which reveal improved kinetics and higher electrocatalyst activity of S-CNHs.
Additionally, the long-term stability of S-CNHs was assessed on the basis of durability studies. Figure 6a shows the LSV polarization curves for HER of S-CNHs in comparison to that for ox-CNHs, pristine CNHs and Pt/C, after ongoing cycling for 10,000 cycles, where negligible loss of the cathodic current is noted. Table 2 summarizes the HER-derived data for all tested materials, before and after 10,000 cycles. Finally, the ECSA values were also calculated. Cyclic voltamographs for HER of S-CNHs, as compared to that for ox-CNHs, pristine CNHs and Pt/C, in a non-Faradaic region were measured at scan rates of 50, 100, 200, 300, 400 and 500 mV/sec (Figure 7). Particularly, S-CNHs exhibit the higher ECSA value,~205 cm 2 , while the equivalent values that are noted for ox-CNHs and pristine CNHs are significantly lower, i.e.,~130 and~31.25 cm 2 , respectively. Similarly, the estimated ECSA values for the same materials after 10,000 cycles are remarkably lower than that recorded for S-CNHs.  Cyclic voltamographs for HER of S-CNHs (a) before, and (b) after 10,000 cycles, ox-CNHs (c) before, and (d) after 10,000 cycles, pristine CNHs (e) before, and (f) after 10,000 cycles, and Pt/C (g) before, and (h) after 10,000 cycles, measured in a nitrogen saturated aqueous 0.5 M H2SO4 electrolyte, at a rotation speed of 1600 rpm and scan rates from 50 to 500 mV/s. Inset: Scan rate dependence of the current densities for the corresponding materials. and (h) after 10,000 cycles, measured in a nitrogen saturated aqueous 0.5 M H 2 SO 4 electrolyte, at a rotation speed of 1600 rpm and scan rates from 50 to 500 mV/s. Inset: Scan rate dependence of the current densities for the corresponding materials.

Conclusions
In summary, we used a facile doping strategy to introduce sulfur atoms into the graphitic network of CNHs, and we tested S-CNHs against OER and HER. The S-CNHs proved to be a highly efficient bifunctional electrocatalyst towards water splitting due to the high S doping levels. Doping with sulfur effectively modified the electronic structure of CNHs, which along with high conductivity helped to achieve low overpotential values for both OER and HER. Specifically, S-CNHs showed low overpotential for oxygen and hydrogen evolution reaction, at 1.63 and −0.2 V vs. RHE, respectively, and excellent durability and stability. Additionally, S-CNHs exhibited significantly lower Tafel slope and lower current resistance compared to oxidized and pristine CNHs for both electrocatalytic reactions. The easy preparation method, high activity and excellent durability of S-CNHs make them a promising alternative for replacing noble-metal-based electrocatalysts for water splitting.