1. Introduction
Environmental problems, such as air pollution and global warming caused by the overuse of fossil fuel, have long been concerning [
1,
2,
3]. To tackle these environmental problems, attention has been shifted from traditional fuels to efficient and clean energies. H
2 is considered as an ideal clean energy and electrolysis of water to produce hydrogen is a very feasible strategy [
4]. The electrochemical overall water splitting process is constituted by the hydrogen evolution reaction (HER) on the cathode side and the oxygen evolution reaction (OER) on the anode side and requires catalysts of outstanding performance promoting the reaction process [
5]. Noble metals with promising catalytic properties, such as Pt, RuO
2, and IrO
2, have recently been considered as ideal catalytic materials for HER and OER; both of them correspondingly are used in the anode and cathode reaction as a feasible strategy for overall water splitting [
6,
7,
8]. However, the utilization of these materials in the electrochemical overall water splitting field is always restricted due to the high cost and rarity [
8,
9]. With the development of functional materials in nanoscale, significant advances have been made in the study of electrocatalysts, but most electrocatalysts exhibit highly efficient catalytic activity only for HER or OER reactions [
4,
5,
10]. Hence, it is significant to design a low-cost bifunctional catalyst with excellent performance in both HER and OER process to achieve high efficiency overall water splitting.
Most bifunctional catalysts were designed by integrating excellent active components of HER and OER [
4]. The as-designed catalysts could efficiently urge active sites to favorably bind towards both oxygen-containing and hydrogen intermediates, reducing the kinetic energy barriers [
5]. Molybdenum disulfide (MoS
2), a representative HER catalyst, has widely attracted attention from researchers [
11,
12,
13] since its free energy of adsorbed H
+ is close to Pt [
14,
15]. However, MoS
2 is not suitable as a catalyst for OER process because of providing intrinsically poor OER activity [
5,
16]. It is a feasible strategy to introduce OER active components into MoS
2-based materials to prepare bifunctional catalysts [
11,
17]. Additionally, nickel sulfides [
18,
19,
20] exhibited excellent catalytic performance in the OER process due to their unique physicochemical properties. Bolin Li et al. [
21] fabricated in situ N-doped Ni
3S
2 nanowires on Ni-foam by high-temperature pyrolysis, which showed excellent performance in OER process, requiring only 271 mV overpotential to achieve a current density of 10 mA/cm
2. Highly durable OER performance was demonstrated by 40 h chronopotentiometry and 10,000 cyclic voltammetry (CV) cycle tests. The rational construction of MoS
2/Ni
3S
2 heterostructures has the potential to achieve highly efficient bifunctional electrocatalysts [
22,
23]. Furthermore, constructing rational heterostructures could be beneficial to fulfill the synergy in heterostructures, exposing more active sites, promoting catalyst performance, and the electronic transfer. In addition, OER activity of catalysts also could be effectively enhanced by properly doping iron [
24,
25]. However, their applications also are hampered due to intrinsically low electrical conductivity [
4,
26]. To prepare a highly conductive catalyst, for a powder catalyst, this is usually achieved by building heterostructures with superior conductive materials, including carbon nanotubes and graphene [
27]. C elements derived from metal-organic frameworks (MOFs) also could improve conductivity of electrocatalysts, provide the efficient pathways for charge transport, reduce resistance to charge transfer, and downgrade overpotential of reactions [
28,
29,
30].
Recently, methods using ferrocene and ferrocenecarboxylic acid to modify electrocatalysts have attracted considerable attention, such as Ziqian Xue et al. [
31], who prepared missing-linker metal-organic (MOF) frameworks through modifying Co-BDC with ferrocenecarboxylic acid (Fc), and thus achieved excellent OER performance with ultra-low overpotential and current density of 241 mV at 100 mA cm
−2, respectively. Pitchai Thangasamy et al. [
32] synthesized ferrocene-incorporated cobalt sulfide nanoarchitecture, investigated OER performance, and finally gained rather good results: a Tafel slope of 54.2 mV dec
−1 and an overpotential as low as 304 mV to drive a current density of 10 mA/cm
2. The authors attribute the enhancing performance of OER to the strong synergistic effect of Fe-based compounds derived from ferrocene. Few studies, however, have been reported for lowering overpotential of HER process by using ferrocenecarboxylic acid. Given all this, considering the different roles of Fe and C elements in catalysts, we decided to prepare an efficient bifunctional catalyst by using the strategy of doped Fc.
Herein, we successfully prepared the flower-like Fe/C-doped-MoS2/Ni3S2-450 (Fe, C-MoS2/Ni3S2-450) sphere structures by doping ferrocenecarboxylic acid and demonstrated the significance of ferrocenecarboxylic acid in the MoS2/Ni3S2 heterostructures on the electrochemical HER and OER. High-resolution transmission electron microscopy (TEM) and X-ray photoelectron spectroscopy (XPS) analysis were carried out to further characterize the structure and surface states of flower-like sphere structures. In our study, the Fe, C-MoS2/Ni3S2-450 represents a low overpotential of 188 mV and 270 mV to drive a current density of 10 mA cm−2 for HER and OER, respectively. In this work, we pioneered the use of ferrocenecarboxylic acid-derived iron-carbon doped metal sulfides to enhance the electrical transport performance and increase electronic site for catalytic reaction. This work provides valuable insights for the rational design of energy storage and conversion materials by incorporation of transition metal and carbon elements into metal sulfide structures utilizing metallocene.
2. Results and Discussion
The synthetic procedure of Fe, C-MoS
2/Ni
3S
2-450 is as follows. First, the α-Ni (OH)
2 nanosheets were prepared on the basis of previous reports [
26]. As shown in
Figure S1(see Supporting Information), it can be seen that the formation of well-crystallized α-Ni (OH)
2 was corroborated successfully by X-ray diffraction (XRD) pattern (see
Figure S1) [
33]. The as-prepared Ni(OH)
2 product represents nanosheets with smooth surfaces, as shown in the scanning electron microscopy (SEM) image (see
Figure S2). The obtained α-Ni (OH)
2 was then mixed with sodium molybdate dihydrate, Ferrocenecarboxylic acid (Fc), and thiourea in a suitable proportion through ultrasonic treatment and magnetic stirring. The solution was then transferred into a 100 mL Teflon-lined autoclave reactor and heated to 200 ℃ for 24 h. After reaction, the as-cooled products were washed several times and collected. To obtain final Fe, C-MoS
2/Ni
3S
2-450, products were annealed at 450 ℃ in Ar/H
2 (90/10%) environment for 2 h, while the control group was prepared by adjusting the annealing temperatures (400, 450, 500, and 550 ℃).
The existence of NiS phase can be confirmed by the XRD pattern shown in
Figure 1, in which the diffraction peaks at 2Theta = 18.4°, 32.2°, 35.6°, 40.4°, 48.8°, 50.1°, 52.6°, and 56.25° corresponding to NiS can be clearly observed. By comparison with Joint Committee on Powder Diffraction Standards ((JCPDS) card (No. 12-0041 NiS) [
34], it can be seen that the diffraction angles and different crystal faces of NiS phase are corresponding respectively. However, peaks corresponding to different crystal faces of MoS
2 phase were not detected, suggesting the existence of other Mo compounds. Moreover, it could be likely that the MoS
2 phase was formed before annealing under 450 ℃ for 2 h in Ar/H
2 (90%/10%) environmental atmosphere. As the results shown in
Figure 1, three legible peaks are found at the position of 2Theta = 14.3°, 33.5°, 58.3°, which accordingly correspond to (002), (101), (110) crystal planes of MoS
2 phases (JCPDS no. 37-1492-MoS
2) [
35]. In addition, other obvious diffraction peaks appear in 2Theta = 21.8°, 31.1°, 37.8°, 44.2°, 49.8°, 50.1°, 55.2°, corresponding to (101), (110), (003), (202), (113), (211), (122) crystal planes of Ni
3S
2 phases (JCPDS no. 44-1418-Ni
3S
2) [
21], respectively, indicating the transformation from NiS phase to Ni
3S
2 phase. This might result from the breaking of partial Ni-S bonds, allowing S- and Mo- sites to combine and form MoS
2 due to the annealing process at 450 ℃ for 2 h in Ar/H
2 (90%/10%) environmental atmosphere. According to previous work [
5], MoS
2 could exhibit excellent catalytic performance in HER process and edge sites of MoS
2 and sulfur anion vacancies could provide sites for increased HER activity. It is shown that annealing temperature is a key factor affecting the hydrogen evolution performance of the catalyst. In order to investigate the elemental surface distribution and valence state of catalyst, Raman measurement was carried out to measure the surface states of flower-like sphere structures. As shown in
Figure 1b, the peaks at 298, 370, and 403 cm
−1 are corresponding to E
1g, E
12g, and A
1g modes of 2H-MoS
2, while the peak at 315 cm
−1 is from the modes of Ni
3S
2 [
4]. The peak located at 345 cm
−1 belongs to the modes of MoS
2/Ni
3S
2 [
22,
23]. In addition, three sharp peaks can be observed between 800–1000 cm
−1, indicating the molecular structure of Mo
3S
13 distributing on edge sites of MoS
2 and the existence of rich under-coordinated Mo-S edge sites in the Fe, C-MoS
2/Ni
3S
2-450.
As shown in the SEM figures (
Figure 2a), the reaction products of Ni(OH)
2, Na
2MoO
4·2H2O, Fc, and thiourea display three-dimensional flower-like sphere structures. A large number of nanosheets on the sphere surface can be observed, which may result in the increase of effective active sites and promote performances of the catalyst. Moreover, once the feeding rate of Ni(OH)
2 was increased, the product evolved from flower-like spherical structures to irregular forms or spheroidal particles distributed on the previous flower-like spherical surface. To maintain the preferable nanosheet structure, an optimal Ni(OH)
2/Na
2MoO
4 (the amount of Na
2MoO
4 remained at 2 mmol) feeding ratio of 0.2g/Na
2MoO
4 was adopted for the fabrication. As shown in the SEM images (
Figure 2b,c), samples at different magnifications clearly reveal the existence of that original three-dimensional flower-like spherical structure, with ultrathin nanosheets distributed evenly on the surface. This result can be further confirmed by the TEM image shown in
Figure 2d–g, in which an irregular sphericity structure of Fe, C-MoS
2/Ni
3S
2-450, with its interior clearly shaded, can be observed, suggesting the whole structure is not in a hollow state. The TEM images in
Figure 2e,f further confirm the formation of flower-like Fe, C-MoS
2/Ni
3S
2-450 spheres. In
Figure 2g, the lattice fringe of 0.62 nm belongs to the (002) lattice plane of MoS
2 [
22], while the lattice distances of 0.28 nm are indexed to the (110) plane of Ni
3S
2, as the (110) plane of Ni
3S
2 and the neighboring (002) plane of MoS
2 constitute the interface of MoS
2/Ni
3S
2 heterostructures [
22]. In conclusion, the flower-like spherical structure of Fe, C-MoS
2/Ni
3S
2-450 synthesized in this work has good uniformity in both structure and size.
The results of XPS measurement are shown in
Figure 3, where in the Ni 2p is composed of the alike characteristic spin-orbit states of Ni 2p
1/2 and Ni 2p
3/2, which are accordingly corresponding to the strong peaks at 857.0 eV and 874.0 eV in the upper half of
Figure 3a, with its two weak satellite peaks located at 862.1 eV and 879.5 eV, respectively. Comparing with the lower half of
Figure 3a, there were significant positive shifts for peaks of Ni 2p and obvious changes of chemical valence states. After annealing in reducing atmosphere, it is possible that the Fe doping brings identifiable chemical state variation of Ni compared to Fe, C-Mo
yS
x/NiS. The result above is similar to the previous report [
22], and can endow catalysts with more active sites for OER process [
25]. For the Mo 3d spectrum in
Figure 3b, the two peaks at about 229.6 and 232.8 eV are corresponding to Mo
4+ 3d
5/2 and Mo
4+ 3d
3/2. In addition, the peak located at 236.2 eV indicates the presence of a high oxidation state of Mo
6+, which might be due to the existing Mo
6+ that has not been fully reduced before annealing. The chemical valence of Mo has significant change: the peaks belonging to Mo
6+ decrease noticeably in Fe, C-MoS
2/Ni
3S
2-450 as compared with Fe, C-Mo
yS
x/NiS, illustrating that MoS
2 is generated. The peak at about 226.7 eV originates from S 2s, indicating the existence of the chemical bindings of Mo-S. The S 2p spectrum in
Figure 3c deconvolutes into two peaks, which accordingly correspond to S 2p
1/2 (163.5 eV) and S 2p
3/2 (162.4 eV), suggesting the existence Mo-S and Ni-S. The nearby peak located at 168.9 eV belongs to the satellite peak of S 2p.
As shown in
Figure 3c, the peak of S 2p
1/2 is belonging to the sulfur with low coordination and has a relationship with sulfur defects, which indicates the possibility of abundant defects in Fe, C-MoS
2/Ni
3S
2-450. It may be significant for HER performances. In
Figure 3d, the Fe 2p peaks of Fe, C-MoS
2/Ni
3S
2-450 can be deconvoluted, four peaks at 708.3 eV (Fe
2+ 2p3/2) and 720.5 eV (Fe
2+2p1/2), 713.5 eV (Fe
3+2p3/2) and 726.0eV (Fe
3+2p1/2), suggesting the Fe
2+ compounds and Fe
3+ compounds. However, comparing with the Fe 2p spectrum in the lower half of
Figure 3d, we clearly find that the binding energy of Fe 2p got shifted further to the left and the amount of Fe
2+ species was decreased, suggesting partial original Fe
2+ transformation into Fe
3+. To investigate the effect of temperature on chemical shifts of Ni, Mo, S, and Fe, XPS measurements were performed on Fe, C-Mo
yS
x/NiS (
Figure 3a–d) for comparison. It can be seen that large shifts of binding energy of Ni and Mo, S, Fe, and the heterostructure relationship of MoS
2 and Ni
3S
2 could be strengthened after annealing [
4,
22]. In addition, annealing changes the chemical environment of Fe, C-MoS
2/Ni
3S
2-450 and treatment chemical state of Fe and Ni elements and may affect catalytic activity.
The electrocatalytic HER and OER performance of the Fe, C-MoS
2/Ni
3S
2-450 flower-like spherical structures were measured in a three-electrode configuration and alkaline electrolyte (1M KOH, PH = 14). Meanwhile, the performance of the commercial Pt/C, RuO
2, and MoS
2, NiS
2, MoS
2/Ni
3S
2 were also investigated.
Figure 4a illustrates the polarization curves of the samples obtained at a scan rate of 5 mV/s (corrected with IR compensation already, unless otherwise noted). It can be seen that Fe, C-MoS
2/Ni
3S
2-450 exhibits superior HER activity compared to those of MoS
2, NiS
2, and MoS
2/Ni
3S
2 as the potential increasing. To achieve a current density of 10 mA/cm
2, a low overpotential of 188 mV should be affordable for Fe, C-MoS
2/Ni
3S
2-450. However, upon comparison among MoS
2/Ni
3S
2 (315mV), NiS
2 (334 mV), and MoS
2 (363 mV), the advantages of Fe and C for effectively promoting the HER activity of MoS
2/Ni
3S
2 could be found. In addition, samples obtained in different annealing temperatures in atmospheres of Ar/H
2 (90%/10%) were also prepared for comparison. As shown in
Figure 4f, the HER performance of the Fe, C-MoS
2/Ni
3S
2-450 were effectively improved when temperature was increased, and reached optimum effect at the temperature of 450 ℃. However, the catalytic reaction process of Fe, C-MoS
2/Ni
3S
2-450 became worse when the annealing temperature exceeded 450 ℃. According to XRD analysis (
Figure 1a), it is possible that Ni
3S
2 and MoS
2 phase were formed, and thus the electrocatalytic performance of catalysts were modified. It is a practicable method to enhance catalytic properties in some materials’ systems through doping catalytic materials with transition-metal atoms. Meanwhile, according to previous reports [
25], the catalytic performance of electrocatalysts could be increased, through adjusting their coordination valence states and chemical environment by doping a small amount of Fe. Therefore, the Fe, C-MoS
2/Ni
3S
2-450 in this study exhibited much higher HER performance than that of MoS
2/Ni
3S
2, which might result from the introduction of Fe and C elements.
To further investigate the HER activity, the Tafel plots from polarization curves were fitted. As shown in
Figure 4b, the Tafel slope of the Fe, C-MoS
2/Ni
3S
2-450 (95 mV/decade) is lower than that of pure MoS
2/Ni
3S
2 (168 mV/decade), MoS
2 (140 mV/decade), and NiS
2 (211 mV/decade), indicating that Fe, C-MoS
2/Ni
3S
2-450 has faster reaction kinetics. According to previous reports [
4,
22], three principal steps are involved in HER, with different Tafel slopes corresponding to associated reaction steps. The first step is Volmer reaction (a Tafel slope of about 120 mV/decade), the second step is Heyrovsky reaction (a Tafel slope of about 40 mV/decade) or Tafel reaction (a Tafel slope of about 30 mV/decade). According to the Tafel slope of Fe, C-MoS
2/Ni
3S
2-450 (−95 mV/decade), its HER could follow the Volmer−Heyrovsky mechanism (H
2O + H
ads + e
- → H
2 +OH
−), where H
ads represents the H atom on an active site. Meanwhile, the smaller Tafel slopes stand for faster electrocatalytic kinetics of HER and OER.
In addition, electrochemical impedance spectroscopy (EIS) analysis at −0.25 V (V vs. reversible hydrogen electrode (RHE)) was carried out for the in-depth study of the activity of electrocatalytic kinetics of HER. As shown in
Figure 4c, it can be seen that charge-transfer resistance in MoS
2, NiS
2, MoS
2/Ni
3S
2, and Fe, C-MoS
2/Ni
3S
2-450 was illustrated by Nyquist. Among all the samples, the lowest charge transfer resistance (R
ct) occurred in Fe, C-MoS
2/Ni
3S
2-450, suggesting faster electrocatalytic kinetics of HER activities and expediting of the Faradaic process on electrode. Hence, the result further reveals that the charge-transfer process of MoS
2/Ni
3S
2 can be accelerated by doping C element.
Electrochemical stability of the catalysts is taken as standard for judging the HER performance of catalyst. The long-term stability of Fe, C-MoS
2/Ni
3S
2-450 was obtained using chronoamperometry and CV test. As shown in
Figure 4d (a constant current density of 10 mA/cm
2 and time of 12 h), potential of the Fe, C-MoS
2/Ni
3S
2-450 was enhanced to only 25 mV over 12 h at 10 mA/cm
2. The Fe, C-MoS
2/Ni
3S
2-450 exhibited superior stability of HER with only 47 mV degradation in potential after 3000 CV cycles (see
Figure 4e).
It is known that catalytic performance is positively correlated with electrochemically active surface areas (ECSA), and ECSA also shows a positive correlation with the double-lay capacitance (C
dl). It is therefore C
dl was obtained from cyclic voltammetry (CV) curves to further evaluate the HER catalytic performance. Based on CV curves (
Figure S3), the C
dl values of the NiS
2 (0.15 mF/cm
2), MoS
2 (0.24 mF/cm
2), MoS
2/Ni
3S
2 (0.85 mF/cm
2), and Fe, C-MoS
2/Ni
3S
2-450 (1.04 mF/cm
2) were obtained. Through comparison, it can be seen that Fe, C-MoS
2/Ni
3S
2-450 shows the highest C
dl value, meaning higher ECSA of as-obtained Fe, C-MoS
2/Ni
3S
2-450 because of the Fe and C incorporation. The result reveals that the Fe and C incorporation effectively provide more electroactive catalytic sites that lead to enhanced HER performance.
Since electrochemical water splitting is composed of HER and OER processes, it is very necessary to further investigate the OER performance of Fe, C-MoS
2/Ni
3S
2-450. In the polarization curves of OER (see
Figure 5a), the lowest overpotential of 273 mV at a current density of 10 mA/cm
2 is observed in Fe, C-MoS
2/Ni
3S
2-450 (350 mV for RuO
2, 400 mV for NiS
2, 460 mV for MoS
2, and 355 mV for MoS
2/Ni
3S
2), suggesting the superior OER performance of flower-like MoS
2/Ni
3S
2 spheres were improved by doping Fe and C. According to previous reports [
25], Fe element has a unique catalytic ability in OER, and suitable doping of Fe helps the low valence state of Ni
0 in metal oxide/(oxy)hydroxide materials to shift to more active Ni2
+ with a high valence state in OER activity and optimize the absorption energies of O and OH binding, causing the OER performance to improve.
Likewise, the Tafel plots from polarization curves were fitted to further investigate OER activities. As shown in
Figure 5b, a Tafel slope of only 66 mV/decade was observed in Fe, C-MoS
2/Ni
3S
2-450, which is superior to the RuO
2 (82 mA/decade), the NiS
2 (104 mA/decade), the MoS
2 (194 mA/decade), and MoS
2/Ni
3S
2 (95 mA/decade). This result reveals the enhanced reaction kinetic of Fe, C-MoS
2/Ni
3S
2-450.
In addition, EIS (at 1.48 V (vs. RHE),
Figure 5c) and chronoamperometry (a constant current density of 10 mA/cm
2 and time of 12 h,
Figure 5d) tests were carried out to investigate the OER performance of the Fe, C-MoS
2/Ni
3S
2-450. As with the HER activity, the lowest R
ct in Fe, C-MoS
2/Ni
3S
2-450 indicates the faster Faradaic process on the Fe, C-MoS
2/Ni
3S
2-450 heterostructures, resulting in excellent OER performance Fe, C-MoS
2/Ni
3S
2-450. As shown in
Figure 4d, the Fe, C-MoS
2/Ni
3S
2-450 maintains a high stability at 10 mA/cm
2, but the left-hand of the chronoamperometry curve has an upward trend. The increased contact resistance between the catalyst and the electrolyte was due to the bubbles from the OER reaction that could not overflow instantly. However, the bubbles busted and disappeared as they became overgrown. After
it test (The stability test was carried out by the
i-t curve) at 10 mA/cm
2, the overpotential of the Fe, C-MoS
2/Ni
3S
2-450 reached 60 mV in
Figure 5e. To find the best annealing temperature, different samples were prepared to test OER activity by changing the annealing temperature. The result shown in
Figure 5f reveals that the lowest overpotential of OER at 10 mA/cm
2 occurred in samples annealed at 450 ℃, and thus indicate the best OER performance.
In conclusion, the improved HER and OER performance of the Fe, C-MoS2/Ni3S2-450 originated from the doping C element, which resulted in better electrical transport performance with lower overpotential and rich electronic sites for catalytic reaction after suitable annealing. In addition, the doping of Fe helped to enhance the interfacial effects between MoS2 and Ni3S2, as the Fe caused the low valence state of coordinated metals to shift to a more active high valence, optimized the heterostructure, and enhanced the reaction kinetics of OER.