Introduction

A sustainable production of H2 as a chemical vector by a reliable and cost-effective way is one of the major challenges for the hydrogen economy. The electrolysis of water is a clean way to produce H2 and O2, without production of impurity by-product gasses. However, these reactions are not spontaneous at standard temperature and pressure, thus requiring the use of a catalyst and energy input to afford that purpose. From the rational point of view, electro-splitting of water is the only feasible option when electricity is obtained from non-conventional renewable energy sources, by utilizing as small as possible amounts of the electrical power [1]. The development of cheap catalysts, based on abundant materials, able to compete with noble metal based and expensive catalysts, such as RuO2 or IrO2, has been the focus for this application during the last years [2].

Current efforts are concentrated on the development of noble metal-free metal oxide and oxy-hydroxide-based catalysts [2,3,4,5,6]. Among them, perovskite-structure [7], spinel-structure [8, 9], and hydroxide-based [10, 11] materials are studied most intensively. Nickel-based materials exhibit the highest activity [12]. However, cobalt [13] and iron [6] containing materials can also be successfully optimized. Manganese oxides are gaining more and more interest as oxygen evolution electrocatalysts because of their proven high activity of biological photosynthesis [14]. Recently, several papers reported investigations of simple manganese oxides in acidic, neutral, or basic conditions [15,16,17,18]. Moreover, alkali nanostructured cryptomelane and birnessite materials were investigated with respect to possible enhancement of oxygen evolution activity by controlled introduction of potassium [19, 20], transition metals [21], and noble metals [22]. Although the reported activity is usually lower than that of nickel-based materials, the nature-provided examples of Mn-based compounds inspire new research directions for the optimization of Mn-based electrocatalysts.

Cryptomelane is a nanoscale alkali-doped manganese octahedral molecular sieve (OMS). The high activity in oxidation reactions of Mn-based nanostructured materials is well known in heterogeneous catalysis [23,24,25,26,27,28]. Their high reactivity can be assigned to a combination of several features, such as beneficial porous structure, particular nanorod-type morphology, redox properties due to the mixed valence framework, and high oxygen mobility [29,30,31,32]. Cryptomelane is a hollandite type manganese oxide, with a tunneled structure (OMS-2) nanostructured by potassium (K-OMS-2). The tunnels are built up of four double-wide (2 × 2) slabs of octahedral MnO6 units, connected by edges, forming tunnel dimensions of 4.6 × 4.6 Å [33]. Moreover, the introduction of heteroatoms in the structure of tunneled or layered oxides can bring about desired modification of their properties without changing their structure [34,35,36].

The possible activity enhancement of pristine cryptomelane was the motivation for the current study. The influence of the introduction of Fe, Co, and Ni structural promoters into the cryptomelane structure was investigated with respect to the oxygen evolution reaction.

Experimental

Material Preparation

Undoped cryptomelane (KMn8O16, K-OMS-2) was prepared by dissolving 11 g of manganese acetate in 40 ml of water, to which 5 ml of glacial acetic acid was added to adjust the pH. The solution was heated to boiling under reflux and maintained for 45 min. Potassium permanganate (6.5 g), dissolved in 150 ml of water, was then added. The mixture was stirred and heated for 24 h. The obtained solid phase was filtered and washed with distilled water until neutral pH, dried at 80 °C, and calcined at 450 °C for 2 h.

For the preparation of Co and Fe-substituted K-OMS-2 material (KMn8 − xCoxO16 and KMn8 − xFeXO16), 0.01 mol of potassium nitrate was dissolved in 50 ml of distilled water along with a combined 0.0046 mol of manganese nitrate and the appropriately substituted metal nitrate in a 10:1 ratio. A cross-linking reagent, glucose, was dissolved in distilled water and added into the solution of salts. The molar ratio of salts and cross-linking reagent was adjusted to be 1:2. The mixture was stirred and heated to form a homogeneous clear sol. Then, the sol was evaporated and heated at 180 °C for 2 h in an electric oven. The resulting powder was calcined at 800 °C for 2 h in static air.

The nickel substituted K-OMS-2 material (KMn8 − xNixO16) was prepared by adding a 50-ml water solution containing nickel (0.663 g) and manganese (5.726 g) nitrates to 40 ml of potassium permanganate (2.768 g) mixed with 10 ml of acetic acid. Both solutions were previously heated and mixed with a temperature of ~ 60 °C. The mixtures were kept heated and mixed under a cooling condenser for 6 h. The formed precipitate was then filtered, washed with distilled water, and transferred to a crucible to be calcined at 400 °C for 2 h.

Physicochemical Characterization

The XRD patterns were recorded by a Rigaku MiniFlex powder diffractometer with Cu Kα radiation at 10 mA and 10 kV, 2θ step scans of 0.02°, and a counting time of 1 s per step. The micro-Raman spectra were recorded in ambient conditions using a Renishaw InVia spectrometer equipped with a Leica DMLM confocal microscope and a CCD detector, with an excitation wavelength of 785 nm. The laser power at the sample position was 1.5 mW (0.5% of total power) with a magnification of × 20. The Raman scattered light was collected in the spectral range of 200–850 cm−1. At least six scans, 10 s each, were accumulated to ensure a sufficient signal to noise ratio.

Specific surface areas (SSA) were obtained by the three-point nitrogen adsorption method at liquid nitrogen temperatures using a Quantachrome Autosorb-1 automated gas sorption instrument. Samples were degassed at 120 °C under a flow of helium for 2 h before the analysis.

The relative ratio of metal ions was determined by using an energy-dispersive X-ray fluorescence (XRF) spectrometer (Thermo Scientific, ARL QUANT’X with the Rh anode, the X-rays of 4–50 kV). A 3.5-mm Si(Li) drifted crystal with a Peltier cooling (∼ 185 K) was used as a detector. Quantitative analysis of the results was done semi-automatically with the use of UniQuant software, calibrated with a series of metallic standards.

X-ray photoelectron spectroscopy (XPS) was performed to determine the surface elemental composition of the catalysts. The analysis was carried out using a Thermo Fisher Scientific K-Alpha+ X-ray photoelectron spectrometer with monochromated Al Kα radiation operating at a power of 72 W (6 mA × 12 kV). A combination of low-energy electrons and argon ions were used to compensate charge. High-resolution scans were performed at a pass energy of 40 eV, while the survey spectra were acquired at a pass energy of 150 eV. Survey scans were recorded from 0 to 1350 eV. The high-resolution O 1s spectra were collected from 525 to 545 eV, C 1s from 279 to 305 eV, Mn 3s from 73 to 100 eV, Mn 2p from 632 to 660 eV, Fe 2p from 700 to 740 eV, and Co 2p from 771 to 812 eV. All the spectra were calibrated against a value of the C 1s binding energy of 284.8 eV. The procedure for the determination of an average Mn oxidation state is based on [37] and described in the Supplementary Information, p. 1.

SEM observations were carried out by TESCAN Vega 3 LUM electron microscope equipped with a LaB6 cathode.

H2-TPR measurements were obtained using a Chembet-300 apparatus with a TCD detector. About 0.025 g of sample was placed in an U-shaped quartz flow reactor (diameter ca. 5 mm) and kept in a helium stream for 1 h at 100 °C to remove adsorbed water. Next, the samples were cooled to ambient temperature and the TPR analysis was performed in the range of 25–600 °C, with a temperature ramp of 10 °C min−1 and a 30-cm3 min−1 flow rate of reducing mixture (5% H2 in Ar).

Electrocatalytic Activity

Reactivity in the oxygen evolution reaction (OER) of the investigated catalysts was evaluated using rotating disk electrode (RDE). All measurements were conducted using a Biologic S-150 instrument and a standard three-electrode electrochemical glass cell (Als-Japan). A Hg/HgO electrode and a gold wire were used as the reference and counter electrodes, respectively. After measurements, all potentials were converted to the reversible hydrogen electrode (RHE) scale. OER polarization curves were obtained at room temperature using steady-state staircase voltammetry testing in a nitrogen-saturated 0.1 M KOH. The electrode was rotated at 2000 rpm, and a step size of 10 mV and time period of 10 s step−1 was used. All tests were performed at room temperature. The counter was a spiral Gold wire with a length of 23 cm and diameter of 0.5 mm, with a total surface area of 3.6 cm2 (Biologic Science Instruments, A-0.1263). A typical ink formulation was 20 mg of the catalyst powder, 210 μl of Nafion®, 600 μl of water, and 2400 μl of ethanol. The mixture was sonicated for 10 min, and 4 μl of the suspension was deposited on an electrode with an area of 0.125 cm2.

Results and Discussion

The structural characterization performed with the use of X-ray diffractometer indicates that the bulk of the samples is composed of cryptomelane type materials (identified using the PDF database: 59159-ICSD), as shown in Fig. 1. All diffractograms feature broad peaks, which indicate small nanocrystal domains, for the investigated samples.

Fig. 1
figure 1

XRD patterns of the investigated cryptomelane-based catalysts

Full size image

Raman spectra, which are more sensitive to the local coordination of metal cations and presence of heterophases, show that some structural changes occur for the Fe- and Co-doped samples, Fig. 2. Upon doping with Co and Fe, a partial change of the structure into the birnessite type takes place, as indicated by an intensity increase of the high wavenumber peak and disappearance of the lowest wavenumber peak [38]. Unfortunately, the quantification of the cryptomelane and birnessite phases present in the samples is not feasible basing on the available XRD and Raman spectroscopy data. The nitrogen adsorption measurements indicate that for the Fe and Co-doped samples, SSA is lower by one and two orders of magnitude than for the undoped and Ni-doped samples: K-OMS-2 73 m2 g−1, Fe/K-OMS-2 10 m2 g−1, Co/K-OMS-2 ~ 1 m2 g−1, and Ni/K-OMS-2 120 m2 g−1. The changes of the crystals’ morphology from cryptomelane nanorods [32] to flower-like birnessite material [39] may be the reason for the observed decrease. This interpretation is supported by the SEM pictures shown in Fig. 3, where the Co/K-OMS-2 and Fe/K-OMS-2 are visibly more compact, without the sharp features of the undoped cryptomelane.

Fig. 2
figure 2

Raman spectra of the investigated cryptomelane-based catalysts

Full size image
Fig. 3
figure 3

SEM pictures of the investigated cryptomelane-based catalysts

Full size image

The amounts of transition metals (TM) introduced into the K-OMS-2 structure vary considerably, being the highest for Fe and the lowest for Ni. A general trend is that the more doping TM in the structure, the more potassium presence is also observed (Table 1).

Table 1 XRF analysis results of cryptomelane-based catalysts
Full size table

Elemental analysis of the samples’ surface was performed with the use of XPS. The elemental analysis with and without adventitious carbon is summarized in Tables S1 and S2 in the Supplementary Information. The high-resolution spectra for the regions of interest are collected in Figs. S1–S4 in the Supplementary Information. Table 2 summarizes the manganese normalized relative composition. In general, the relative cationic composition of the surface varies from sample to sample. The most striking observation is that the Ni-doped sample exhibits relatively low amount of oxygen. This indicates a reduced surface, with nickel in the form Ni2+. Fe- and Co-doped samples, where the doping cations easily adopt multiple oxidation states, show higher surface oxygen content. The average manganese oxidation state is correlated with potassium content (Table 2; Fig. S5), which is reasonable from the crystal electronegativity point of view—the higher the relative content of low oxidation state K, the higher Mn oxidation state to compensate the oxygen negative charges. Relatively, low oxygen content in the Ni/K-OMS-2 sample results in the reduction of the sample surface. A provisory charge balance analysis, comparing the sum of negatively and positively charged ions (Table S3), indicates that the positive charges exceed the sum of negative O2− only for the Ni/K-OMS-2 sample. For the undoped cryptomelane, the surface is about neutral, while for the Co- and Fe-doped samples, the surface is oxidized.

Table 2 XPS surface analysis results of cryptomelane-based catalysts
Full size table

The comparison of bulk composition of the cryptomelane materials determined with the XRF with the XPS results shows that in the case of the undoped and Fe- and Co-doped materials, the surface is enriched with potassium. Furthermore, it can be inferred that the surface content of Co in Co-doped cryptomelane is much lower than the bulk one.

Temperature-programmed reducibility tests with hydrogen, presented in Fig. 4, indicate that for the doped samples, the Co-doped one features reducibility similar to that of cryptomelane, which can be rationalized by the low Co content. On the other hand, the Fe- and Ni-doped samples show a decreased reducibility—the reduction peaks are shifted towards higher temperatures. The effect of iron doping can be rationalized by the observations that the presence of manganese retards the reduction of iron in mixed iron-manganese oxides [40]. The reduced state of the surface of Ni-doped sample can be responsible for the decreased reducibility of Ni/K-OMS-2.

Fig. 4
figure 4

H2-TPR profiles of the investigated cryptomelane-based catalysts

Full size image

Figure 5a shows the OER IR-corrected polarization curves of the synthesized catalysts under pseudo-steady state conditions in N2-saturated 0.1 M KOH. Not surprisingly, undoped cryptomelane shows a negligible OER reactivity. However, doping the K-OMS-2 with Fe and Ni causes even further deterioration of the electrocatalytic activity. On the other hand, the addition of Co leads to a striking enhancement in oxygen evolution. The Tafel analysis of the polarization curves reveals that, indeed, Co/K-OMS-2 is characterized by the lowest slope value (Fig. 5b). The synthesized materials exhibit different Tafel slopes, KMn8O16 (134 mV decade−1), Fe/KMn8O16 (152 mV decade−1), Co/KMn8O16 (93 mV decade−1), and Ni/KMn8O16 (116 mV decade−1). Tafel slopes of about 120 mV decade−1 indicate the predominance of M-OH or M-O surface groups before the rate-determining step. When the surface adsorbed species produced in the early stage of the OER remains predominant, the Tafel slope decreases. The lower thermodynamic limits in the Butler-Volmer formalism predict Tafel slopes of 24, 40 ,or 60 mV decade−1, which occur in the case of Ni, Co, and Ru surface oxides [1]. Specifically, for the case of Co/KMn8O16 (93 mV decade−1), the sample shows similar Tafel plot to Ir/cryptomelane-type sample (74 mV decade−1) [22]. This comparison may suggest that using Ir or Co to form surface active species will be limited due to the chemistry of cryptomelane coordination, limiting activities with respect to pure oxides. However, the advantage of using doped cryptomelane materials is their stability at relevant reaction temperatures. The specific activity (normalized by the specific surface area) and mass activity plots are presented in the Fig. S7A, B in the Supplementary Information. Since a similar mass of the catalyst was used for all the measurements, the mass activity plot is almost identical to the Fig. 5a. The specific activity plot in Fig. S7B amplifies the superior reactivity of the Co/cryptomelane sample due to its very low SSA.

Fig. 5
figure 5

a OER polarization curves and b Tafel plots of cryptomelane-based samples obtained at room temperature using steady-state staircase voltammetry, under nitrogen-saturated 0.1 M KOH

Full size image

The interesting characteristic of the Co/K-OMS-2 to bear in mind is that there is very little Co present in the material, especially on the surface (Table 2) and that the SSA of this material is very low. There are several possible interpretations of the reactivity enhancement upon cobalt doping. The formation of the active cobalt spinel, Co3O4, whose very good reactivity in OER is widely reported [9], can be excluded on the basis of the XRD and Raman spectroscopic results. The formation of birnessite type phase with intercalated Co was recently reported to improve oxygen evolution catalysis [21], and this could occur to some limited extent in the investigated sample. Furthermore, the formation of layered cobaltane phase, K0.25CoO2, can be responsible for the enhanced reactivity. Figure S6 in the Supplementary Information shows a comparison of such material with the Co/K-OMS-2 sample. Lastly, the increase of the OER activity can be due to the intrinsic activity of the Co-doped cryptomelane structure—not induced by the formation of other phases. For this material, the XPS results suggest the highest average oxidation state of manganese, in line with the proposed active centers for OER catalysts based on Mn, Ni, Co, and Ir [41]. H2-TPR results show the highest lattice oxygen mobility (reducibility) of the Co/K-OMS-2, which can also contribute to the reactivity enhancement. It is noteworthy that the Co/K-OMS-2 catalyst also shows the highest capacitance, in line with the observation that good supercapacitors are also good electrocatalysts (see cyclic voltammograms in Fig. S8). Interestingly, the stability of the Co-doped cryptomelane is very high and show very little or no degradation after 1000 polarization cycles, Fig. S6. Also, in the Fig. S6, a comparison with reference catalysts is presented (Ni/NiO, Co3O4, NiCo2O4) [42].

Conclusions

Fe-, Co-, and Ni-doped K-OMS-2 materials were successfully synthesized and characterized. The electrocatalytic reactivity in oxygen evolution is substantially enhanced for the cobalt-doped material. On the other hand, iron and nickel have no, or even, negative effect on the K-OMS-2 electrocatalytic activity. The possible negative effect can be related to the increased stability of the lattice oxygen, evidenced by the H2-TPR. The structure of the material bulk is preserved in all cases, as evidenced by the X-ray diffractograms, but the formation of additional birnessite phases can be evidenced for the iron and cobalt dopants. It is shown that the reactivity enhancement of the Co/K-OMS-2 can be related not only to the formation of cobalt-doped heterophases (cobaltane, birnessite) but also to the induced modifications of the properties of pristine cryptomelane. The characteristic features of Co-doped K-OMS-2 are the low-temperature reducibility, with an increase of the average manganese oxidation state and an increased capacitance. The considerably better electrocatalytic activity of Co/K-OMS-2 is achieved despite disadvantageous very low Co content and negligible specific surface area.