Srinivasa N.a,
Jack P. Hughesbc,
Prashanth S. Adarakatticd,
Manjunatha C.e,
Samuel J. Rowley-Nealebc,
Ashoka S.*a and
Craig E. Banks*bc
aDepartment of Chemistry, School of Engineering, Dayananda Sagar University, Bengaluru, India. E-mail: ashok022@gmail.com
bFaculty of Science and Engineering, Manchester Metropolitan University, Chester Street, M1 5GD, UK. E-mail: c.banks@mmu.ac.uk; Tel: +44 (0)1612471196
cManchester Fuel Cell Innovation Centre, Manchester Metropolitan University, Chester Street, Manchester M1 5GD, UK
dDepartment of Chemistry, SVM Arts, Science and Commerce College, Ilkal 587125, India
eDepartment of Chemistry, RV College of Engineering, Bengaluru 590059, India
First published on 21st April 2021
We present the facile synthesis of Ni/NiO nanocomposites, via a solution combustion methodology, where the composition of metallic Ni within NiO is controlled by varying the annealing time, from 4 minutes up to 8 hours. The various Ni/NiO nanocomposites are studied via electrically wiring them upon screen-printed graphite macroelectrodes by physical deposition. Subsequently their electrochemical activity, towards the oxygen evolution reaction (OER), is assessed within (ultra-pure) alkaline media (1.0 M KOH). An optimal annealing time of 2 hours is found, which gives rise to an electrochemical oxidation potential (recorded at 10 mA cm−2) of 231 mV (vs. Ag/AgCl 1.46 vs. RHE). These values show the Ni/NiO nanocomposites to be significantly more electrocatalytic than a bare/unmodified SPE (460 mV vs. Ag/AgCl). A remarkable percentage increase (134%) in achievable current density is realised by the former over that of the latter. Tafel analysis and turn over frequency is reported with a likely underlying mechanism for the Ni/NiO nanocomposites towards the OER proposed. In the former case, Tafel analysis is overviewed for general multi-step overall electrochemical reaction processes, which can be used to assist other researchers in determining mechanistic information, such as electron transfer and rate determining steps, when exploring the OER. The optimal Ni/NiO nanocomposite exhibits promising stability at the potential of +231 mV, retaining near 100% of its achievable current density after 28 hours. Due to the facile and rapid fabrication methodology of the Ni/NiO nanocomposites, such an approach is ideally suited towards the mass production of highly active and stable electrocatalysts for application within the anodic catalyst layers of commercial alkaline electrolysers.
Electrolytic water splitting within alkaline conditions is potentially the cheapest method of hydrogen production, due to its low energy consumption, low cost and the long lifetime of the electrolyser cell components.2 This is in comparison to electrolysis in acidic conditions or proton exchange membrane (PEM) electrolysis, where there is typically a requirement for costly cell components such as the catalyst layers and bipolar plates; this is exacerbated by the operating conditions within a PEM cell leading to component corrosion.3,4 Electrolytic water splitting within an alkaline electrolyser, where ideally, water is turned into hydrogen and oxygen, requires two major reactions, namely the cathodic hydrogen evolution reaction (HER) and the anodic oxygen evolution reaction (OER). A theoretical thermodynamic cell voltage of +1.23 V (vs. RHE) is required for the overall reaction to occur, where any additional potential over this, is termed as the overpotential and represents the thermodynamic inefficiencies within the electrolyser cell.5 Many studies have compared the alkaline OER activity of non-precious metal (NPM) based anodic electrocatalysts to benchmark precious metal catalysts, where superior catalysis has been observed in NPM catalysts such as perovskites,6 spinels7 and metal oxides.8 However, these OER catalysts typically require complex synthesis protocols, which limits their scalability within commercial electrolysis, thus the research into mass producible NPM based catalysts is imperative to the advancement of electrolyser technologies.
In the search for NPM OER electrocatalysts, researchers have focused on a range of abundant and cheap oxides, such as nickel, iron, manganese and cobalt.9–15 The OER is generally catalysed by a metal oxide rather than a pure metal, with the mechanism different for oxides with different surface morphologies. It has been reported that the OER activity of metal oxides follows the trend of NiOx > CoOx > FeOx > MnOx,16,17 with Ni based oxides reported to exhibit the most promising OER catalysis owing to their high intrinsic activity;18,19 In exemplifying the case of a metal surface versus a metal oxide, Babar et al.17 explored a thermally oxidized porous NiO supported on nickel foam (NF) in 1.0 M KOH towards the OER. The NiO/NF annealed at 400 °C required an overpotential of 310 mV to reach +10 mA cm−2, compared to the bare NF, which exhibited an overpotential of 400 mV. The beneficial OER performance exhibited by NiO/NF is attributed to the formation of porous NiO within the NF substrate, which provides a large electrochemical surface area with a larger number of exposed active sites, therefore an enhanced signal output. On the other hand, various studies have reported that NiO possesses a large bandgap and therefore poor electrical conductivity, resulting in limited kinetics. This can be mitigated by alloying NiO with metallic Ni, where the bandgap associated with NiO is reduced by the addition of Ni, thus increasing electrochemical charge transfer rates.20,21
The synthesis of Ni/NiO composites have previously been reported in literature; however, these studies have failed to control the quantity and homogeneous distribution of Ni within the NiO matrix.21–25 Zhou et al.26 reported a Ni/NiO composite embedded in graphitic carbon (Ni–NiO/C) that displayed excellent activity and stability as a water splitting catalyst in alkaline conditions. However, the Ni–NiO/C composite was synthesised using complex experimental conditions, involving the hydrothermal growth of nickel organic frameworks on nickel foam at 120 °C for 36 h, followed by calcination at 600 °C for 2 h within an inert atmosphere. The development of scalable, rapid and facile methodologies to synthesise Ni/NiO is essential if OER catalysts are to be utilised within commercial AEM electrolysers. Consequently, we report a facile synthesis methodology encompassing a solution combustion method, which allows for the content of Ni to be controlled that meets the requirements specified above for the large-scale production of Ni/NiO nanocomposites.
XRD was utilised to assess the crystallinity of the synthesised Ni/NiO nanocomposites. The XRD patterns of the Ni/NiO nanocomposites annealed at 500 °C for 4 min, 30 min, 2 hours and 8 hours are displayed within Fig. 1. The XRD standards for Ni (JCPDS 65-2865), exhibits diffraction peaks at 2θ values of 45.1°, 52.0° and 76.4° corresponding to the single crystal faces (111), (200) and (220), while the NiO standard (JCPDS 65-5745) relates to 2θ values of 37.2°, 43.2°, 62.8°, 75.6° and 79.4° corresponding to the single crystal faces (111), (200), (220), (311) and (222). Analysis of the XRD peak areas, using the 43.2° (200), NiO and 45.1° (111), Ni single crystal faces as a function of annealing time reveals the following ratios: 0.071:0.307; 0.374:0.499; 0.372:0.011; 0.368:0 for 4 min, 30 min, 2 hours and 8 hours respectively. In comparison of the various XRD patterns presented in Fig. 1, it is clear that an annealing time of 4 min results in a sample that contains a small amount of NiO, which is predominantly comprised of Ni. As the annealing time is increased the Ni/NiO material transforms to one where the NiO predominates over the Ni content, which is readily achieved within 2 hours. Note that the (111) diffraction peak decreases as the annealing time in increased, which is absent in the longest annealing time of 8 hours, which has resulted in a material comprised purely of NiO. The surface morphology of an electrocatalyst significantly affects the electrochemical performance towards water splitting, which was visually assessed using SEM. ESI Fig. S1† shows SEM images of the Ni/NiO nanocomposites where it is clear that as the annealing time increases, agglomeration of the fine Ni/NiO particles of size 30–40 nm, also increases. Such agglomeration will reduce the accessible surface area for electrocatalytic sites to be assessable, but it is the nanocomposites composition that is more dominant; electrochemical performance and ToF as evaluated later will give more insights.
XPS was also utilised, with Fig. 2(A) showing the XPS survey spectrum confirming the presence of Ni, O, and C, where the carbon content is a result of the adsorbed carbon dioxide or small amount of carbon remaining from the fabrication process (fuel), as detailed in the experimental section. Also shown is the key Ni 2s and Ni 2p regions. Fig. 2(B) shows the high-resolution XPS spectra of the different Ni/NiO nanocomposites, which exhibits a characteristic XPS spectra for NiO displaying a multiplet-split Ni 2p3/2 at 853.6 and 855.4 eV, and a Ni 2p3/2 satellite at 860.6 eV with a multiplet-split. Additionally a Ni 2p3/2 at 871.1 eV and Ni 2p1/2 satellite at 879.9 eV is also observable, which is agreement with previous studies.31 It is worth noting that the Ni(0) peak is not readily identified in the XPS spectra, whilst it is identified in the XRD patterns. This may be due to the formation of NiO upon the Ni particles or where the Ni is embedded within/throughout the NiO, also worth noting is the different depth penetration profiles of the two techniques and also that the XPS binding energy of Ni metal, NiO are very close.
The voltammetric responses in terms of overpotential, for the Ni/NiO nanocomposites, towards the OER were benchmarked at +10 mA cm−2, as is common within the literature. The potentials required to reach +10 mA cm−2 were found to be 420 (1.65 vs. RHE), 279 (1.50 vs. RHE), 254 (1.48 vs. RHE), 231 (1.46 vs. RHE) and 288 mV (1.51 vs. RHE) for the bare/unmodified SPE, 4m-Ni/NiO SPE, 30m-Ni/NiO SPE, 2h-Ni/NiO SPE and 8h-Ni/NiO SPE, respectively. It is evident that as the annealing time is increased, the voltammetric response reduces in overpotential, with the 2h-Ni/NiO SPE giving rise to the optimal response towards the OER. Of interest the 30m-Ni/NiO SPE shows an oxidation peak of quite higher intensity and shifted towards higher potentials compared to the most performant sample, the 2h-Ni/NiO SPE. This is likely due to composition still being dominate towards NiO. Note that these results are favourable, if not better than other nickel based composites, as overviewed in Table 1. A notable percentage increase in achievable current density of 134% is observed in the 2h-Ni/NiO SPE in comparison to a bare/unmodified SPE. The 30m-Ni/NiO SPE and 2h-Ni/NiO SPE display high achievable current densities of between 160–175 mA cm−2, which is due to a high content of conductive metallic Ni atoms. It is interesting to note that increasing the annealing time up to 8 h results in a voltammetric response displaying less beneficial electrocatalytic activity towards the OER, which might be on first sight counter-intuitive. This decrease in activity can be explained by inspection of the physicochemical analysis presented above, particularly the XRD patterns (see Fig. 1), and the electrochemical data reported in Fig. 3, it is evident that a pure NiO phase is a poor material towards the OER. This is consistent with literature reports and the introduction of metallic Ni atoms reduce the bandgap associated with NiO, thus increase the charge transfer rate between electrode and electrolyte.35
Catalyst | Supporting electrode | Electrolyte | Stability | Deposition technique | Catalyst loading | OER overpotential at 10 mA cm−2 (mV vs. RHE) | Tafel value (mV dec−1) | Ref. |
---|---|---|---|---|---|---|---|---|
a Key: SPEs – screen-printed electrodes; RHE – reversible hydrogen electrode; AAO – anodic aluminium oxide (AAO) membrane; GCE – glassy carbon electrode; GC – graphitic carbon; NF – nickel Foam; rGO – reduced Graphene oxide; AC – acid cleaned; CC – carbon cloth; LDH – layer double hydroxide; CP – carbon fiber paper; CF – copper foil; - = not reported. | ||||||||
NiCo2O4/CoO | SPEs | 1.0 M KOH | 10 h at +700 mV (vs. Ag/AgCl) | Drop-cast | 0.53 mg cm−2 | 323 | 118 | 58 |
Ni90Fe10 | AAO | 1.0 M KOH | 24 h at +400 mA cm−2 (vs. RHE) | Electrodeposition | — | 236 | 45 | 59 |
Ni | AAO | 1.0 M KOH | — | Electrodeposition | — | 405 | 117 | 59 |
NiO | NF | 1.0 M KOH | 12 h at +153 mV (vs. RHE) | Drop casting | — | 356 | 77 | 60 |
NiO nanowalls | Quartz/Ti/Au | 1.0 M KOH | — | Sputtering deposition | — | 345 | 48 | 56 |
Ni/NiO | CP | 1.0 M KOH | 13.8 h at +158 mV (vs. RHE) | Drop casting | — | 353 | 97 | 61 |
Ni3S2/NiS | GCE | 1.0 M KOH | 15 h at +10 mA cm−2 (vs. RHE) | Drop cast | 0.20 mg cm−2 | 298 | 58 | 62 |
NiO@NiMoO4 | NF | 1.0 M KOH | 12 h at +10 mA cm−2 (vs. RHE) | Chemical growth | 1.40 mg cm−2 | 280 | 32 | 63 |
NiO nanosheets | NF | 1.0 M KOH | 12 h at +10 mA cm−2 (vs. RHE) | Chemical growth | 0.30 mg cm−2 | 340 | 97 | 63 |
NF | N/A | 1.0 M KOH | 12 h at +10 mA cm−2 (vs. RHE) | N/A | N/A | 340 | 109 | 63 |
rGO/Ni2P | GCE | 1.0 M KOH | 30h at + 10 mA cm−2 (vs. RHE) | Drop cast | 0.10 mg cm−2 | 283 | 44 | 64 |
NiOx–Fe | NF | 1.0 M KOH | 18 h at +10 mA cm−2 (vs. RHE) | Chemical growth | 0.014 mg cm−2 | 266 | 36 | 11 |
P–NiFe2O4 | CC | 1.0 M KOH | 50 h at +10 mA cm−2 (vs. RHE) | Chemical vapour deposition | — | 231 | 49 | 65 |
Ni32Fe | GC | 1.0 M KOH | 50 h at +10 mA cm−2 (vs. RHE) | Drop casting | 0.12 mg cm−2 | 291 | 58 | 66 |
NiCeOx | GCE | 1.0 M KOH | 200 h at +10 mA cm−2 (vs. RHE) | Chemical growth | — | 295 | 66 | 67 |
NiO–NiFe–LDH | NF | 1.0 M KOH | 16.5 h at +10 mA cm−2 (vs. RHE) | Chemical growth | — | 265 | 72 | 68 |
Ni–Ni(OH)2 | CF | 1.0 M KOH | 24h at + 10 mA cm−2 (vs. RHE | Electrodeposition | — | 290 | 97 | 69 |
Ni/NiOx | NF | 1.0 M KOH | 5.5 h at +500 mV (vs. SCE) | Chemical growth | — | 390 | 70 | 70 |
Ni/NiO | NF | 1.0 M KOH | 30 h at +10 mA cm−2 (vs. RHE) | Drop casting | 0.51 mg cm−2 | 295 | 74 | 26 |
4m-Ni/NiO | SPEs | 1.0 M KOH | — | Drop casting | 0.004 mg cm−2 | 279 | 118 | This work |
30m-Ni/NiO | SPEs | 1.0 M KOH | — | Drop casting | 0.004 mg cm−2 | 254 | 123 | This work |
2h-Ni/NiO | SPEs | 1.0 M KOH | 26 h at +700 mV (vs. RHE) | Drop casting | 0.004 mg cm−2 | 231 | 108 | This work |
8h-Ni/NiO | SPEs | 1.0 M KOH | — | Drop casting | 0.004 mg cm−2 | 288 | 119 | This work |
assuming n = 1:
(1.1) |
(1.2) |
Under extreme potential, e.g. E ≫ E0f or E ≪ E0f, the eqn (1.2) can be simplified where one term or another is neglected. In the case of an electrochemical reduction, i.e..
A + e− → B, eqn (1.2) becomes:39
(1.3) |
while for in the case of an electrochemical oxidation, i.e.
B – e− → A, eqn (1.2) becomes:
(1.4) |
The corresponding Tafel44–46 equations are then, from eqn (1.3) and (1.4):
(1.5) |
(1.6) |
Experimentally, and historically, the analysis of Tafel slopes are plotted as (E − E0f) vs. log10I due to the way they were originally measured, i.e. galvanostatic rather than potentiostatic, thus we re-state (1.5) and (1.6) as the following Tafel equations:
(1.7) |
(1.8) |
Note that all the above is for a simple one-step, one-electron transfer process, which is the most commonly reported in the academic literature. However, multi-step electrochemical processes, such as that encountered in the OER, as explored in this paper, require a different approach. The Butler–Volmer equation can be modified for a multi-step overall electrochemical reaction process, which comprises electron transfer steps in addition to the rate determining step:36
(1.9) |
Now, let us consider some general electrochemical processes that may be encountered, such as in the OER, in determining a reaction mechanism where Tafel analysis is routinely utilised. If we consider a multi-step reaction, A + ne− → Q, which has the following pre-steps:
In this approach, we define npre to be the number of electrons transferred before the rate-determining step (rds). npost is the number of electrons transferred after the rate-determining step. nRDS is defined as the number of electrons transferred in the rate-determining step and ν the stoichiometric number is the number of times the rate-determining step occurs. Note that the total number of electrons, n, in the overall electrode reaction is given by: n = npre + npost + nRDS. Thus, the overall transfer coefficients can be expressed as: where αcathode is the transfer coefficient of the overall forward/anodic reaction, npre is the number of electrons take up by the electrode before the rds and nRDS is the number of electrons involved in the rate-determining step. Conversely: where αanode is the transfer coefficient of the overall backward/cathodic reaction, leading to:
(1.10) |
Noting that where the rds occurs ν times in the electrode reaction. A full derivation is presented by Bockris and Reddy,36 and later Fletcher47 but summarised here for the convenience for material scientists developing new OER electrode materials. We next consider some scenarios, to elaborate on the above information. The above is written for an electrochemical reduction/cathodic process, i.e. but is applicable for an oxidation/anodic process. Thus if we consider the case of the OER, an anodic process, i.e. A – ne− → B, we should consider the following:
(1.11) |
If we now consider some scenarios, for the electrochemical process: A – ne− → B, which is considered to be the rate-determining step, it follows from above that we use: where, v = 1; npost = 0; nRDS = 1; β = 0.5 and assuming T = 298 K, F is 96485.3; R = 8.31, a Tafel slope of 118 mV decade−1 should be experimentally observed. This is known to be the classical result for a single-step one-electron transfer processes. In practice, values are not always precisely 118 mV due to experimental errors. Table 2 overviews a range of multi-step electrochemical mechanisms and the predicted Tafel plots using the notation of Testa and Reinmuth48 (where E is an electrochemical step, C is a chemical step etc.); however, the above approach should allow most, if not all, possible mechanisms encountered in such processes such as the OER. Note that while Tafel analysis and that of Table 2 provides more evidence of the overall electrochemical mechanism, further insights will be needed from physiochemical analysis and/or in situ measurements to provide an unambiguous determination.
Returning to the voltammetric responses presented within Fig. 3, Tafel analysis was performed (Fig. 3(B)) with the bare SPE, 4m-Ni/NiO SPE, 30m-Ni/NiO SPE, 2h-Ni/NiO SPE and 8h-Ni/NiO SPE exhibiting Tafel values (slopes) of 191, 118, 123, 108 and 119 mV dec−1, respectively. Using a recently updated Pourbaix diagram for Nickel,49 and noting that the XPS analysis indicated that the surface morphology is comprised of NiO, which will be the major component exposed at the interfacial surface, while the Ni is embedded through the NiO samples, the underlying electrochemical mechanism, as observed in Fig. 3, is tentatively proposed as follows:
NiO + OH → NiOOH + e−, rds | (1.12) |
NiOOH + OH → NiOO− + H2O | (1.13) |
NiOO− → Ni + O2 + e− | (1.14) |
Ni + 2OH → Ni(OH)2 | (1.15) |
Ni(OH)2 + OH → NiOOH + H2O + e−, rds | (1.16) |
NiOOH + OH → NiOO− + H2O | (1.17) |
NiOO− → Ni + O2 + e− | (1.18) |
Further characterisation of the 2h-Ni/NiO nanocomposite was performed, with Fig. 3(B) showing a transition electron microscope (TEM) image indicating the Ni/NiO particles are in the range of 25–45 nm. The particles exhibit a large surface area with irregular morphologies where the observed OER kinetics, enhanced by the formation of the Ni–NiO hetero-structures are due to the synergistic effect between NiO and metallic Ni, where the surface layer of NiO exhibits a high number of active sites. These active sites allow oxygen to readily form, and the surrounding metallic Ni atoms serve as short diffusion pathways and channels for rapid electron transport.52,53
The electrochemically active surface area (ECSA) was determined by the specific capacitance method, as advocated within the academic literature via the following equation: ECSA = Cd/Cs where a literature value of 0.04 mF cm−2 is used for Cs, while the value of Cd is determined experimentally. The latter is determined for the various Ni/NiO nanocomposites with cyclic voltammograms recorded in 0.1 M KOH within a non-Faradaic region (i.e. 0.04 to 0.15 V (vs. Ag/AgCl)) over the range of scan rates from 10 mV s−1 to 50 mV s−1; see ESI Fig. S2.† The capacitance of an electrode surface can be determined via cyclic voltammetry within a potential window (see ESI Fig. S2†) where no Faradaic reactions occur. ESI Fig. S2† shows typical results where a plot of current vs. scan rate, ν is constructed allowing the capacitance, C, of the electrode surface to be deduced from: I = Cν. Consequently, ECSA values of 2.5, 19.3, 21.3, 125.0, 8.1 were determined for bare SPE, 4m-Ni/NiO SPE, 30m-Ni/NiO SPE, 2h-Ni/NiO SPE and 8h-Ni/NiO SPE respectively. While these values likely do not reflect the true electrochemical area, they provide some form of benchmark. The effective resistance, or impedance, of the electrical circuit was measured for the various Ni/NiO SPEs. Electrochemical impedance spectra (EIS) were recorded at an amplitude of 10 mV (vs. RHE) and frequency range 0.1–100000 Hz in 1.0 M KOH solution (shown in ESI Fig. S3†). The charge transfer resistance (Ω) values of 581.1, 337.7, 325.5, 218.4 and 472.9 Ω are exhibited by the bare SPE, 4m-Ni/NiO SPE, 30m-Ni/NiO SPE, 2h-Ni/NiO SPE and 8h-Ni/NiO SPE, respectively. It is clear that the 2h-Ni/NiO SPE exhibits the lowest charge transfer resistance, which coincides with the water splitting catalysis displayed by the nanocomposites, in regards to the 2h-Ni/NiO SPE exhibiting the fastest charge transfer rate and lowest OER overpotential.
Last, the intrinsic catalytic activity of the Ni/NiO nanocomposites towards the OER was estimated by determining the Turnover Frequency (ToF), which is the number of molecules (e.g. O2) produce per second per site:
ToFmin = jAgeo/nFm | (1.19) |
One issues is, where to measure the ToF from, some literature advocates the use of using an overpotential of 100 mV or 350 mV, from which the current density is measured.54–56 Typically the latter value should be utilised as this will correspond to the correct portion of the cyclic voltammetric profile where the OER is in full operation. Using the current density at 350 mV and determining the number of moles of the Ni/NiO catalyst (molecular weight: 133.385 g mol−1) upon the electrode surface, ToFmin was deducted to be: 0.00138, 0.00180, 0.0204 and 0.0012 s−1 for the 4m-Ni/NiO SPE, 30m-Ni/NiO SPE, 2h-Ni/NiO SPE and 8h-Ni/NiO SPE, respectively. The 2h-Ni/NiO SPE displays the fastest rate of O2 production and thus exhibits the highest faradaic efficiency. These values are of similar magnitude to that reported for similar nickel systems, but fabricated via different approach but without varying the Ni/NiO ratio21,57 but an order of magnitude slower than others.54–56
Footnote |
† Electronic supplementary information (ESI) available. See DOI: 10.1039/d0ra10597j |
This journal is © The Royal Society of Chemistry 2021 |