Effect of Precursor-Solution pH on the Phase Composition and Microstructure of Precipitated NiO, and By-Product Retention in Hydrothermally Derived NiO–MnO₂ Powders

2.1 Materials

Nickel(II) chloride hexahydrate (NiCl₂·6H₂O), manganese(II) sulfate monohydrate (MnSO₄·H₂O), potassium peroxodisulfate (K₂S₂O₈), sulfuric acid (H₂SO₄, 95–97%), sodium hydroxide (NaOH), distilled water, and deionized water were used as received.

2.2 Synthesis of NiO Nanoparticles (Precipitation)

NiO nanoparticles were prepared by precipitation. A 10 M NaOH solution was prepared by dissolving 4.000 g of NaOH in 10 mL of distilled water. Separately, 11.883 g of NiCl₂·6H₂O (0.2 M) was dissolved in 250 mL of distilled water on a hot plate at 50 °C under stirring. The NaOH solution was then added dropwise while the pH was monitored until the target value was reached; four precursor pH values were prepared—pH 8, 9, 10, and 11. The solution was held on the hot plate for 40 min, after which the green precipitate was separated by filtration through a poly(vinylidene fluoride) (PVDF) membrane filter and washed on the filter with deionized water until the filtrate reached neutral pH, and dried in an oven at 60 °C for 14 h. The dried solid was ground into a powder and calcined at 500 °C for 2 h.

The pH was brought to the target value by dropwise addition and was not controlled thereafter: no buffer or pH-stat was used, and because the precipitation of Ni(OH)₂ consumes OH⁻, the pH will have drifted downward over the 40 min hold. The washing end-point was the neutrality of the filtrate; the number of washes was therefore not fixed but set by that criterion. This criterion tracks the removal of excess hydroxide, not of chloride: NaCl is a neutral salt and does not change the pH of the filtrate, so a neutral filtrate does not indicate that the NaCl has been removed. No chloride test (e.g. with AgNO₃) or conductivity measurement of the filtrate was made. One batch was prepared at each precursor pH and the syntheses were not independently repeated; one diffraction pattern was collected per sample. The uncertainties reported throughout are accordingly those of the profile fits and of the regressions—measurement and fitting precision—and do not include batch-to-batch reproducibility.

2.3 Hydrothermal MnO₂-Growth Step

To prepare the MnO₂ precursor solution, 1.352 g of MnSO₄·H₂O (0.2 M) and 2.163 g of K₂S₂O₈ (0.2 M) were weighed, 2 mL of H₂SO₄ was added, and the mixture was dissolved in 40 mL of distilled water and stirred at 600 rpm at room temperature until homogeneous. Then 1 g of the calcined NiO powder was added and stirred for 1 h at room temperature. The mixture was transferred to an autoclave and held at 120 °C for 8 h. The resulting precipitate was dried at 80 °C for 8 h, ground into a powder, and calcined at 500 °C for 2 h. This step was intended to yield a NiO/MnO₂ composite; because the phase analysis of Section 3.1 does not establish MnO₂, its product is referred to below as the hydrothermal product. The overall two-stage synthesis route is summarized in Fig. 1.

2.4 Characterization

Phase composition and crystallite size were determined using a PANalytical X’Pert³ Powder diffractometer (Cu Kα, λ = 0.15406 nm). The pH 8 powders were measured over 2θ = 10–100° with a 0.05° step and a counting time of 31 s per step, whereas the pH 9–11 powders were measured over 5–80° with a 0.01° step and 1 s per step. The two protocols differ in sampling rather than in quality: the pH 8 scan has the better counting statistics per point (31 s against 1 s) but samples each reflection about five times more coarsely, so that a reflection of FWHM ~0.4° is spanned by some 8 points rather than some 40. The derived quantities are affected very unevenly, and we therefore compare them on the following basis. Quantities obtained from the peak width (D, ε) depend on how well the profile is sampled and are compared only across the identically measured pH 9–11 samples; the pH 8 values are reported for completeness but are excluded from the trend discussion. Quantities obtained from the peak position (the Nelson–Riley lattice parameter) are insensitive to the step size at this resolution, and quantities obtained from intensity ratios measured within a single pattern (the reference-intensity-ratio estimate and the texture coefficients) are likewise insensitive, because both phases are recorded in the same scan and the acquisition factors cancel in the ratio; the pH 8 sample is retained for both. Surface morphology was examined with a Thermo Scientific Quattro S field-emission scanning electron microscope at 15 kV, and elemental composition was obtained from energy-dispersive X-ray spectroscopy (EDS) map-sum spectra acquired in the same instrument. For the pH-8 samples, EDS elemental maps and point spectra were additionally acquired (at 20 kV) to resolve the spatial distribution of the by-product elements. Electron microscopy was performed only on the pH 8 and pH 9 samples; no FE-SEM or EDS data were acquired for the pH 10 and pH 11 powders or their hydrothermal products, and the two sessions differ in accelerating voltage, so the pH 8 and pH 9 micrographs are compared qualitatively only. Particle/cluster sizes were estimated from the 100 000× micrographs (2.70 nm per pixel, from the instrument metadata) by watershed segmentation (scikit-image; contrast-limited equalization, Gaussian smoothing, Otsu thresholding, and a distance-transform watershed), the size of each feature being reported as its equivalent circle diameter; because the material is agglomerated, these are semi-quantitative feature sizes rather than primary-particle diameters.

Each NiO reflection was fitted with a pseudo-Voigt profile on a locally linear background by non-linear least squares (SciPy) to extract the peak position 2θ and the full width at half maximum β (in radians). The cubic lattice parameter was obtained from the interplanar spacing of each reflection, a=dhklh2+k2+l2, and refined to a₀ by a Nelson–Riley extrapolation [74] of a against F(θ)=12(cos⁡2θ/sinθ+cos⁡2θ/θ) to θ=90°. The crystallite size D was first estimated from the NiO (200) reflection (~43.3°) with the Scherrer equation [75],

The relative amount of NaCl was estimated from the integrated (200) intensities of NiO and NaCl by the reference-intensity-ratio method [77], using reference intensities R = m|F|² LP/V² computed from the rock-salt structures with Cromer–Mann atomic form factors and the Lorentz–polarization factor (a two-phase NiO+NaCl assumption, with no microabsorption correction, so the result is semi-quantitative). The NiO texture was assessed from the coefficient [78] TC(hkl)=[I(hkl)/R(hkl)]/⟨I/R⟩, which equals unity for a randomly oriented powder.

3.1 XRD Analysis

Figure 2a shows the XRD patterns of the NiO powders prepared at precursor pH 8–11. The diffraction peaks at 2θ ≈ 37.2, 43.3, 62.9, 75.4, and 79.4°, indexed to the (111), (200), (220), (311), and (222) planes, confirm the formation of cubic NiO (PDF 00-073-0450). Additional reflections at 2θ ≈ 27, 31.7, 45.4, and 56.4° correspond to a cubic halite (NaCl) phase (PDF 00-005-0628). This by-product forms during precipitation according to NiCl₂ + 2NaOH → Ni(OH)₂ ↓ + 2NaCl; the Ni(OH)₂ then dehydrates to NiO during calcination. This NaCl is a major crystalline component of the powders rather than a trace impurity: for the pH 9–11 samples the NaCl (200) reflection at ~31.6° is the most intense peak in the whole pattern, and its intensity relative to NiO varies non-monotonically with pH—largest at pH 9 and comparable at pH 10–11 (a semi-quantitative weight fraction for all four samples, including pH 8, is given below, together with the reason why this intensity-ratio quantity, unlike the width-derived ones, can be compared across the two acquisition protocols). The persistence of this strong NaCl signal shows that the washing step did not remove the chloride/sodium by-product.

Figure 2b shows the XRD patterns of the hydrothermal products. The NiO (111) reflection at 2θ ≈ 37° is retained. The patterns are otherwise dominated by a single by-product phase. Eleven of the resolved reflections index on one primitive cubic cell, refined to a = 10.1175 Å with a root-mean-square deviation of 0.068° in 2θ and ten of the eleven matched to within 0.06° (Table 1). The strong reflection near 28° is the (310) reflection of this cell, which is the strongest reflection of the langbeinite structure type. Both the cell size and the K₂S₂O₈/H₂SO₄/MnSO₄ reagent chemistry, followed by calcination at 500 °C, point to a langbeinite-type potassium–manganese sulfate of the K₂Mn₂(SO₄)₃ type: the refined cell agrees to within 0.03% with the reference value for synthetic cubic K₂Mn₂(SO₄)₃ (a = 10.1143 Å, space group P2₁3 [79]), and it is clearly distinct from the Ni-langbeinite K₂Ni₂(SO₄)₃ (a = 9.8436 Å [79]). Because substitution of the smaller Ni²⁺ would contract the cell, the agreement also suggests that little nickel is incorporated in this phase. One weak reflection at 24.11° (~5% relative intensity) remains unindexed, and a definitive assignment would still require whole-pattern (Rietveld) refinement or a complementary technique, so this is presented as an indexing result rather than a completed phase identification.

The indexing bears directly on the assignment of MnO₂. The evidence for tetragonal MnO₂ consisted of two reflections, measured at 45.78° and 56.12°. The cubic cell above generates reflections at 45.69° and 55.98°, against 46.10° and 56.66° for tetragonal MnO₂ (PDF 01-072-0071), so the by-product cell fits these two positions at least as well. We do not read that as positive evidence for the sulfate at those positions, because a cubic cell of this size generates a dense set of reflections above 45° and coincidental agreement is likely; the conclusion that follows is the weaker and safer one, that the presence of MnO₂ is not established by these two reflections alone. The third expected MnO₂ reflection, near 37°, overlaps NiO (111) and cannot be resolved. Because the sulfate is itself Mn-bearing, the detection of Mn by EDS (Section 3.3) does not settle the question either. We therefore describe the hydrothermal product as a multiphase, NiO-bearing powder in which MnO₂ is the intended but unconfirmed Mn-bearing phase.

The weak NiO signal is consistent with dilution by these by-product phases and with the low overall crystallinity of the hydrothermal-product patterns.

Beyond the single-peak Scherrer estimate, the microstructure of the NiO powders was analysed by refining the cubic lattice parameter and by separating the size and strain contributions to the broadening with the Williamson–Hall and size–strain-plot methods; the results are collected in Table 2.

The Nelson–Riley refined lattice parameter is a₀ = 4.178 Å (individual values 4.1771–4.1781 Å; Table 2) for all four precursor pH values (Fig. 3c), equal to the bulk NiO value (4.1769 Å) within uncertainty and without a systematic dependence on pH. The macrostrain (unit-cell dilation, Δ a/a ≈ 0) is therefore negligible—distinct from the non-zero microstrain (~1×10⁻³) obtained below from line broadening—so the precursor pH does not distort the NiO lattice.

The three size estimates agree in magnitude: the single-peak Scherrer value (NiO (200)) is 20–23 nm, the Williamson–Hall value 25–33 nm, and the size–strain-plot value 23–26 nm (Fig. 3a). The offset between the single-peak and multi-peak sizes reflects the microstrain (~1×10⁻³; Fig. 3b) that the single-peak Scherrer analysis otherwise attributes to size. Across the identically measured pH 9–11 series the pH-to-pH variations in both D and a₀ are comparable to their standard errors (Table 2, Fig. 3); the apparent decrease of the four-peak Williamson–Hall size at pH 11 (Fig. 4d) is, moreover, an artefact of the weak high-angle (222) reflection and disappears (~30 nm, constant within error) once that reflection is excluded. The NiO crystallite size therefore shows no systematic dependence on precursor pH and stays of the order of 20–30 nm, consistent with the pH-insensitive lattice parameter and of the same order as NiO nanoparticle sizes reported for related precipitation syntheses [13,18]. This robustness contrasts with reports in which the precursor pH markedly altered the phase, size, or morphology of other precipitated oxides such as TiO₂, Fe₃O₄, and ZnO [54,55,56], and indicates that, over the pH 8–11 range studied, the NiO microstructure is comparatively insensitive to this parameter.

Why the precursor pH should have so little effect over this window is worth setting out, since the contrast with those oxides is otherwise puzzling. Four factors act in the same direction. First, the window lies entirely above the completion of precipitation: for 0.2 M Ni²⁺ and K_sp(Ni(OH)₂) ≈ 5.5×10⁻¹⁶, precipitation begins near pH 6.7 and is more than 99.7% complete by pH 8 [80]. The four endpoints therefore differ in residual Ni²⁺ by orders of magnitude, but all correspond to essentially complete precipitation. Second, nucleation is governed by local rather than bulk supersaturation: 10 M NaOH was added dropwise into 0.2 M NiCl₂, so that the supersaturation at the point of contact is extreme and much the same in all four syntheses, and the bulk endpoint pH is reached only once the nucleation burst is over. Third, Ni²⁺ is not amphoteric in this window. The systems in which the precursor pH strongly alters the product—TiO₂, Fe₃O₄, ZnO—involve either a phase selection or a soluble hydroxo-complex equilibrium, Zn(OH)₂ in particular redissolving as hydroxo-zincate complexes at high pH [81]; Ni(OH)₂ has no competing phase and no soluble complex between pH 8 and 11, so there is no chemical lever for the pH to act on. Fourth, the calcination at 500 °C for 2 h largely resets the microstructure, so that the final crystallite size reflects the solid-state step rather than the as-precipitated hydroxide. Taken together, these indicate that pH 8–11 is a plateau, and that a pH effect in this system would have to be sought below the completion of precipitation or, for the by-product content, in the washing step rather than in the precipitation itself.

Two further checks reinforce this picture. First, a reference-intensity-ratio estimate from the (200) reflections puts the NaCl content of the powders at 32–41 wt% (Fig. 5a), confirming quantitatively that the by-product is a major fraction of the crystalline material. The individual values are 32, 41, 34 and 34 wt% at pH 8, 9, 10 and 11 respectively: the content is largest at pH 9 and does not vary monotonically with pH. Because this estimate is a ratio of integrated intensities of two phases recorded in the same scan, the acquisition conditions cancel, and the pH 8 sample is therefore included here on the same footing as the others, unlike the width-derived quantities of Table 2.

The estimate is semi-quantitative and its limitations should be stated. It assumes a two-phase NiO+NaCl mixture, so that any amorphous or unindexed material is ignored and both phases are over-weighted in proportion; the reference intensities are computed from the rock-salt structures rather than measured against a standard; no microabsorption correction is applied, although the linear attenuation coefficients of the two phases differ appreciably at Cu Kα; and the Cu Kα₂ component is not stripped. Microabsorption should be minor for the NiO, whose features lie well below the scale at which the Brindley correction becomes significant, but it cannot be excluded for the NaCl, whose grain size after drying is not known. Fluorescence does not contribute, since the Ni K edge (8.33 keV) lies above Cu Kα (8.04 keV). We therefore attach an uncertainty of order 5 wt% absolute to these values, and use them only to establish that NaCl is a major crystalline component, not to assign a precise weight fraction or a pH trend.

An independent cross-check supports the magnitude. The 0.05 mol of NiCl₂·6H₂O charged yields at most 0.1 mol of NaCl (5.84 g) alongside 0.05 mol of NiO (3.73 g), that is 61 wt% NaCl if none were removed. The measured 32–41 wt% therefore corresponds to the washing step having removed 56–70% of the NaCl formed, a physically sensible figure obtained without reference to the diffraction analysis.

Second, the NiO texture coefficients (Fig. 5b) lie close to unity (0.9–1.2 for the strong (111), (200), and (220) reflections), showing that the NiO is essentially randomly oriented; this validates the random-orientation assumption underlying the intensity-based size and phase analyses. The lower value for the weak (222) reflection reflects its poor counting statistics rather than genuine texture.

For the hydrothermal products a reliable crystallite size could not be obtained. The hydrothermal-product patterns are weak (maximum intensity 1600–3400 counts, against ~6500 counts for the NiO powders), and in the NiO (200) region the signal-to-noise ratio is only 1.4–1.7. Moreover, the strongest feature in this region lies at 43.8–44.1°—at the by-product phases, consistent with the (422) reflection of the cubic cell of Table 1, calculated at 43.80°—rather than at the NiO (200) position (43.3°). Figure 6 compares the hydrothermal-product and NiO-powder patterns over 40–48° and shows that the sharp NiO (200) reflection of the powder has no counterpart in the hydrothermal product. A single-peak fit in this window therefore characterizes a secondary phase, not NiO, and no NiO crystallite size is reported for the hydrothermal products. The much lower peak intensities and higher relative background of the hydrothermal-product patterns further indicate that the hydrothermal products are considerably less crystalline than the NiO powders, which would be consistent with a poorly crystalline Mn-bearing phase, although that phase is not identified here. The dominance of the NaCl and sulfate by-products in the diffraction patterns underlines that improved washing and purification are the priority for future work.

3.2 FE-SEM Analysis

Figure 7 compares the morphology of the NiO powder and the hydrothermal product for a representative sample (pH 9 precursor). The NiO powder (Fig. 7a) consists of agglomerated, approximately spherical nanoclusters formed by the coalescence of smaller grains [13], interspersed with occasional elongated rod- and fibre-like features. Because the EDS spectrum of this powder contains only Ni and O (Section 3.3), these elongated features are not a manganese phase; they are most plausibly Ni(OH)₂-derived NiO filaments retained through calcination, or thin fibres from the conductive sample mount. The clusters span tens to several hundred nanometres (quantified below), whereas the crystallite size from XRD is only 20–30 nm, so each cluster is a polycrystalline aggregate of many NiO domains rather than a single crystal. Such agglomeration is characteristic of oxides obtained by hydroxide precipitation and calcination, in which capillary and surface forces during drying and solid-state coarsening during calcination draw the primary particles together.

After the hydrothermal step (Fig. 7b) the fine morphology changes: the product is dominated by sheet- and flake-like features carrying nodular particles, which are absent from the bare NiO powder. We describe these features without assigning them to a phase. Neither technique used here resolves the interfacial structure—FE-SEM because no cross-sections were prepared and only surface topography was imaged, and EDS because the map-sum spectra have no spatial resolution at the relevant scale—and, because the langbeinite-type sulfate of Section 3.1 is present in quantity and its habit under these conditions is not independently known, the morphology cannot distinguish it from MnO₂. Under the acidic hydrothermal conditions used here the peroxodisulfate would be expected to oxidize Mn²⁺ to MnO₂ (MnSO₄ + K₂S₂O₈ + 2H₂O → MnO₂ + K₂SO₄ + 2H₂SO₄), with the pre-formed NiO clusters acting as heterogeneous substrates for that growth and the morphology depending on the hydrothermal temperature and time [35]. The present data do not confirm that this is what occurred: as set out in Section 3.1, the dominant crystalline product of the hydrothermal step is a potassium–manganese sulfate, the reflections attributable to MnO₂ are not distinguishable from those of that phase, and EDS cannot say which phase hosts the manganese it detects. What can be said is that the product is an intimate multiphase mixture rather than the ordered, free-standing core/shell arrays reported for template- or foam-supported NiO/MnO₂ electrodes [11,12,25,34], and that elongated rod/fibre features are already present in the bare NiO, so that the morphology by itself is not diagnostic of composite formation. Because FE-SEM and EDS were acquired only for the pH 8 and pH 9 samples, they document the hydrothermal step at fixed precursor pH and are not, on their own, evidence for a dependence of morphology on that variable.

Micrographs of the pH-8 samples (Fig. 8) are included for completeness, although they do not settle the question either. The pH-8 NiO powder (Fig. 8a) shows the same granular nanoparticles together with the elongated rod/fibre features already noted at pH 9, whereas the pH-8 hydrothermal product (Fig. 8b) is dominated by dense elongated needles rather than by the sheet- and flake-like features seen at pH 9. We do not read this difference as an effect of the precursor pH. The two sets were acquired in separate sessions at different accelerating voltages (20 kV for pH 8, 15 kV for pH 9); each is a single field of a strongly agglomerated and demonstrably heterogeneous powder (Section 3.3); and rod/fibre features are present in the bare NiO at both pH values, so they cannot be attributed to the hydrothermal step. XRD, which samples the bulk, finds the same dominant by-product phase at all four precursor pH values (Section 3.1) and gives no indication of a pH-driven change in phase content. These micrographs are therefore reported as documentation of the material rather than as evidence on the pH dependence of the morphology, which the present data cannot establish either way.

A watershed segmentation of the 100 000× micrographs (Fig. 9) provides a semi-quantitative measure of the apparent particle/cluster size. The equivalent circle diameter has a median of 65 nm for the NiO powder (n=478) and 83 nm for the hydrothermal product (n=335), with broad, right-skewed distributions extending beyond 300 nm. Both medians greatly exceed the NiO crystallite size from XRD (20–30 nm), quantitatively confirming that the imaged particles are polycrystalline aggregates; the somewhat larger hydrothermal-product values are consistent with the sheet- and flake-like features formed in the hydrothermal step, whatever their phase. Because the particles are agglomerated, these values are approximate and represent feature or cluster sizes rather than primary-particle diameters.

3.3 EDS Analysis

The elemental composition obtained from the EDS map-sum spectra (pH 9 samples) is given in Table 3. The measurement is standardless and semi-quantitative, and light-element (oxygen) quantification is unreliable, so the metal ratios are more informative than the absolute oxygen contents. For the NiO powder, only Ni and O were detected, with a Ni-rich metal:oxygen ratio (Ni 55, O 45 at.%) consistent with NiO within EDS accuracy and with cubic NiO being the only Ni-bearing phase seen by XRD. No sodium or chlorine was detected, although NaCl is a major crystalline component of the bulk powder; the analysed field was therefore NiO-rich and not representative of the bulk, which is itself direct evidence of spatial heterogeneity.

For the hydrothermal product, Ni, Mn, and O were all detected, confirming that manganese is present in the analysed region, although, as noted in Section 3.1, EDS cannot establish which phase hosts it. The manganese content exceeds the nickel content (Mn:Ni ~1.6:1 by atom). This is not a consequence of manganese having been charged in excess. The 1.352 g of MnSO₄·H₂O corresponds to 8.00 mmol of Mn, whereas 1 g of the calcined powder, taken as pure NiO, corresponds to 13.39 mmol of Ni—a charged ratio of 0.60:1, that is, manganese in deficit. Correcting for the residual NaCl measured in that same powder (41 wt% at pH 9) lowers the nickel content to 7.90 mmol and raises the charged ratio to 1.01:1, or approximately equimolar. The measured 1.6:1 therefore exceeds the charged ratio, and we attribute the difference to the measurement rather than to the synthesis: the spectrum is a single map-sum over one field that is demonstrably unrepresentative, as the absence of K and S from it shows; EDS samples only the outermost micrometre, so that a Mn-bearing phase formed on the outside of the NiO clusters is over-weighted relative to the nickel it covers; and the quantification is standardless. We accordingly treat the Mn:Ni ratio as qualitative evidence that manganese is present, not as a measure of the amounts charged. No potassium or sulfur was detected in this map-sum spectrum, even though the hydrothermal-product diffraction pattern is dominated by potassium–manganese–sulfate by-products; as for the NiO powder, this reflects a single, non-representative analysis field rather than the absence of those phases. The oxygen content is over-estimated here (O 59 at.%) whereas it was under-estimated for the NiO powder, confirming that the EDS oxygen values cannot be read as stoichiometry in either sample; the metal ratio (Mn:Ni ~1.6:1) is the more robust output, and the analysis is best regarded as qualitative confirmation that manganese is present in the product.

The spatial character of the by-products was examined directly on the pH-8 samples. For the pH-8 hydrothermal product, elemental maps (Fig. 10) show Mn and O distributed across the whole field with Ni interspersed and—in contrast to the pH-9 map-sum spectrum—clear K and S signals, confirming that the potassium–manganese–sulfate by-products lie within the analysed region. Point spectra on the same sample bear out the heterogeneity: some regions contain about 11 at.% K and 13 at.% S (a sulfate-rich domain) whereas others show only Mn, O, and Ni; likewise, point spectra on the pH-8 NiO powder detect Na and Cl (residual NaCl) in some regions but not in others. The by-product phases seen by XRD are therefore real but spatially localized, which explains why a single map-sum field can miss them and confirms that the EDS analysis is representative only of the region sampled.

3.4 Growth Mechanism

The structural and compositional results are drawn together in the growth scheme of Fig. 11. In the precipitation stage, Ni²⁺ and OH⁻ form Ni(OH)₂ nuclei that grow, aggregate, and convert to cubic NiO on calcination, while the Na⁺ and Cl⁻ counter-ions crystallize into the NaCl by-product that survives the washing step (32–41 wt%; Section 3.1). Across pH 8–11 this sequence yields a NiO microstructure that is indistinguishable within the measurement uncertainty of a single batch per condition (D ≈ 20–30 nm, a₀ ≈ 4.178 Å), and a by-product content that varies non-monotonically, so that within this window the precursor pH acts on neither the oxide nor, systematically, on the amount of by-product. In the hydrothermal stage, Mn²⁺ is expected to be oxidized by peroxodisulfate to MnO₂ (Mn²⁺ + S₂O₈²⁻ + 2H₂O → MnO₂ + 2SO₄²⁻ + 4H⁺), the classic persulfate route to MnO₂ nanostructures [82], alongside the pre-formed NiO. The observed product, however, is a multiphase, NiO-bearing powder whose crystalline fraction is dominated by a langbeinite-type potassium–manganese sulfate; whether MnO₂ is also present is not resolved by the present data (Section 3.1). Because neither the phases nor their spatial relationship are fully established, the scheme deliberately depicts coexisting phases rather than a core–shell or coated architecture.