All reagents were analytical grade and used as received without further purification. Distilled water was used in preparation and measurements. KMnO₄ (99.5%) and HNO₃ (65%-68%) were supplied by Guangzhou Haizhu Zone Chemical Reagent Company.

The samples were obtained in a typical synthesis using KMnO₄ and HNO₃ with a molar ratio of about 3.4:1(a1), 0.85:1(b1) and 0.2:1(c1), respectively. The weight of KMnO₄ for synthesizing sample a1, b1 and c1 is 0.79, 0.79 and 0.2 g, respectively, while the concentration of HNO₃is 0.05, 0.2 and 0.2 mol/L (all for the volume of 30 mL). KMnO₄ was dissolved in HNO₃ under stirring for 30 min condition. The solution was transferred to a Teflon-lined stainless-steel autoclave, sealed and maintained at 180℃ for 24 h. The brown-black precipitates were filtered and washed with distilled water for several times to remove the unreacted materials and excess materials. The washing was done until the pH value of the washed water was 7. The resulting products were finally dried at 80℃ for 24 h in air.

Air electrodes for performance evaluation were produced in two layers: the catalyst layer on the electrolyte side and a waterproof diffusion layer on the gas side. The active layer was composed of a catalyst (typically 80wt%), Vulcan XC-72 (10wt%, Hexing Chemical Industry Co., Ltd) and PTFE (10wt%, Dupont). The as-prepared catalysts (a1, b1 and c1) were used as the active material in air electrode. A slurry of the above mixture was made using absolute alcohol as a solvent. Then it was sandwiched in two foamed nickel foils, subsequently compressed for 3 min at 5×10³ kg/cm² under 120℃ to obtain the air electrode.

The phase structure of the as-prepared MnO₂ samples was recorded by X-ray diffraction (Rigaku D/max 2200 vpc). And the morphologies of samples were examined with scanning electron microscope (SEM, JEOL JSM-6490LV). Thermal properties were measured on a thermogravimetric analyzer (NETZSCH 409PC, Germany) with a heating rate of 10 K/min from 50℃ to 280℃. In the test process, nitrogen was used as the purge gas at a flow rate of 50 mL/min. The Brunauer-Emmett-Teller (BET) specific surface area was determined with a specific surface area instrument (ASAP2020M, Japan). The IR spectra were recorded on a Fourier transformation infrared (FTIR) spectrometer (IR Prestige-20, Shimadzu, Japan).

Linear sweep voltammetry were performed in a three-electrode configuration on an electrochemical station (CHI 650, Wuhan, China). 6 mol/L KOH solution was used as the electrolyte. The prepared air electrodes were used as the working electrode with the working area of 2 cm × 2 cm. A homemade Hg/HgO electrode was used as reference electrode. In the linear sweep voltammetry, a platinum sheet was used as the counter electrode and the scan rate was 1 mV/s. In the battery performance determination, a zinc sheet was used as the anode. To prevent moisture penetrating from the outside and flooding-out of the electrolyte from the inside of the cell, a PTFE sheet was attached to the cathode side that was open to the air.

The discharge tests of the simulated cells were performed on CT2001A (WUHAN, CHINA) and were discharged to an end voltage of 0.9 V at different constant currents of 1, 5 and 10 mA/cm².

All measurements were carried out at 25℃ (with an accuracy of 0.05℃) using a water thermostat. To make the results reproducible and reliable, a fresh electrolyte solution was used in every measurement.

Figure 1 shows the XRD patterns of the as-prepared products. From Fig. 1, it is obvious that the sample a1, b1 and c1 belong to δ-MnO₂, α-MnO₂ and β-MnO₂, respectively. The ratio of K⁺ion to H⁺ ion exerts a decisive influence on the crystal structure of the products. When H⁺ ion concentration is lower than that of K⁺ ion (n_k+/n_H+=3.4), the observed product is δ-MnO₂ as described above, which is layered crystal structure. When H⁺ ion concentration is about equal to that of K⁺ ion ( n_k+/ n_H+=0.85), the product is converted to α-MnO₂. Under the condition of lower concentration of K⁺ ion and higher concentration of H⁺ ion ( n_k+/ n_H+=0.24), the obtained product is β-MnO₂. The result indicates that K⁺ion can serve as a template for the formation of the [2×2] tunnel structure of α-MnO₂ while H⁺ ion cannot be the template for it. The formation of β-MnO₂ is no need for template ion, and the tendency to form β-MnO₂ at low K⁺ concentration is rationalized. Main reasons are as follow: the polymorphs of α- and δ-type are different from β-type that the α- and δ-type are constructed from double chains of [MnO₆] octahedral forming [2×2] tunnels, while the β-type is composed of single chain of octahedral forming [1×1] tunnels, and β-MnO₂ is the thermodynamically stable phase and other MnO₂ phases tend to be converted to β-MnO₂ under appropriate condition^{[ 29]}.

Fig. 1

	Figure Option ViewDownloadNew Window
	Fig. 1 XRD patterns for three types of the as-prepared samplesa1:δ-MnO2; b1:α-MnO2; c1:β-MnO2

Therefore, K⁺ served as template for the formation of double chains of [MnO₆] octahedral forming [2×2] tunnels. It can be noted that sample c1 shows better crystalline than samples a1 and b1. This phenomenon can be ascribed to the different H⁺ concentration. The higher H⁺ concentration reduces the formation rate of manganese dioxide and thus favors the formation of better crystalline of the product (c1).

FT-IR spectra of the as-prepared MnO₂ samples are shown in Fig. 2. The bands at around 3420, 2366, 1622 and 1136 cm^-1 correspond to the O-H vibrating mode of absorbed water. The band located at about 580 cm^-1 on curve a1 in Fig. 2(B) is attributed to the metal-oxygen (Mn-O) bending vibration in δ-MnO₂. The bands located at 590 and 510 cm^-1 belong to the metal-oxygen (Mn-O) bending vibration of [MnO₆] octahedral in α-MnO₂ as shown in Fig.2(B) (b1)^{[ 30]}, while the bands located at 714, 523 and 469 cm^-1 correspond to the metal-oxygen (Mn-O) bending vibration of β-MnO₂ as shown in Fig. 2(B)(c1)^{[ 31]}. The result of FT-IR spectra is consistent with the XRD result.

Fig. 2

	Figure Option ViewDownloadNew Window
	Fig. 2 FT-IR spectra of the as-prepared samples

The results of TG analysis of the as-prepared samples are shown in Fig. 3. From the TG profiles, it can be seen that the samples show different thermal behavior. Two major weight losses occur between 50℃ and 280℃. The first weight loss (1.65wt%, 1.73wt% and 1.62wt% for a1, b1 and c1, respectively) below 120℃ can be attributed to desorption of physisorbed water whilst the release of chemisorbed water may be responsible for the significant weight loss (6.43wt%, 6.08wt% and 4.95wt% for a1, b1 and c1, respectively) occurred between 120 and 280℃.

Fig. 3

	Figure Option ViewDownloadNew Window
	Fig. 3 TG analysis for the as-prepared samples

The morphology of the as-prepared samples was ob- served by SEM, as shown in Fig. 4. From Fig. 4(a), it can be seen that the product particles are uniformly distributed and the products are mainly composed of spherical-like particles with 250 nm of average grain size. For the active material, the smaller particles have larger specific surface area and shorter diffusion distance, which provides better rate performance. Fig. 4(b) shows the existence of the opening hydrangea like morphologies and the nanostructued particles in the as-prepared b1 sample (about 500 nm). Figure 4(c) shows that there exists a great deal of large particles of MnO₂, which consisted of aggregation of MnO₂ hystrichosphere particles (about 250 nm). Thus, the morphology can be affected by the concentrations of the reactants. However, the actual mechanism is not clear for the moment and needs to be further investigated.

Fig. 4

	Figure Option ViewDownloadNew Window
	Fig. 4 SEM images of the as-prepared samples

The BET surface areas of the samples, as shown in Table 1, are 68.2, 67.8 and 62.1 m²/g for a1 (δ-MnO₂), b1 (α-MnO₂) and c1 (β-MnO₂), respectively. Compared to c1, the specific surface areas of a1 and b1 are increased by 9.8% and 9.2%, respectively.

Table 1

Table 1

Table 1 Specific surface area of as-prepared samples As-prepared sample Specific surface area/(m2·g-1) a1(δ-MnO2) 68.2 b1(α-MnO2) 67.8 c1(β-MnO2) 62.1

Table 1 Specific surface area of as-prepared samples

Steady state polarization curves for the as-prepared samples are presented in Fig. 5. The current density was obtained based on the area of the work electrode because the electrodes contained the as-prepared catalysts with the same weight. As seen in Fig. 5, the initial oxygen reduction potentials for the electrodes consisting of the as-prepared catalysts are -0.04 V (c1), -0.05 V (a1) and -0.05 V (b1). Subsequently, with scanning potential becoming more negative, the reduction current increases slowly. The oxygen reduction current of the as-prepared samples (a1, b1 and c1) at -0.35 V is 56.28, 56.01 and 40.88 mA/cm², respectively. Compared with b1 (α-MnO₂) and c1 (β-MnO₂), the catalytic oxygen reduction current of a1 (δ-MnO₂) is increased by 0.48% and 37.67%, respectively. The results suggest that the catalytic activity of the as-prepared catalysts is related to not only the specific surface area but also the crystal structure. The activity towards oxygen reduction of the as-prepared samples is in the order: δ-MnO₂ ≈α-MnO₂ >β-MnO₂, which is in agreement with the other reports^{[ 32]}.

Fig. 5

	Figure Option ViewDownloadNew Window
	Fig. 5 Cathodic polarization curves for oxygen reduction on gas diffusion electrodes in 6 mol/L KOH solutiona1: (δ-MnO2); b1: (α-MnO2); c1: (β-MnO2). Scan rate: 1 mV/s

Figure 6 shows the discharge curves of the as-prepared samples a1 (δ-MnO₂), b1 (α-MnO₂) and c1(β-MnO₂) as cathodes in 6 mol/L KOH. From Fig. 6, it is clear that all the three electrodes exhibit smooth discharge curves. At a current rate of 1 mA/cm², a1, b1 and c1 have an open-circuit voltage of 1.38, 1.35 and 1.35 V, and a discharge plateau at 1.22, 1.22 and 1.20 V, respectively. The discharge capacity of a1, b1 and c1 is 231, 229 and 220 mA·h, respectively, as shown in the curves a, d and h. In addition, at a high rate of 5 mA/cm², the open-circuit voltage of a1, b1 and c1 is 1.20, 1.20 and 1.18 V, respectively. And the discharge capacity of the catalysts is 205 mA·h (a1), 203 mA·h (b1) and 186 mA·h (c1). When the current rates was increased to 10 mA/cm², the discharge capacity of a1, b1 and c1 were decreased to 175, 174 and 160 mA·h, respectively. With increasing current rate, the reduction of the discharge capacity of a1 is equal to that of b1, but smaller than that of the c1. This phenomenon can be ascribed to the better crystalline, which does not favor the conduction of proton and electron during oxygen reduction reaction.

Fig. 6

	Figure Option ViewDownloadNew Window
	Fig. 6 Discharge curves of air/Zn batteries using as-prepared δ-MnO2(a, b and c), α-MnO2 (d, e and f) and β-MnO2 (h, I and j) at 1, 5 and 10 mA/cm2, respectively

The effect of the concentration of K⁺ ion and H⁺ ion on the microcrystals type and the surface morphology of the products synthesized by a hydrothermal method, and the relationship between the microcrystals type and electrochemical performance were investigated in the paper. XRD and SEM images show that the as-prepared samples are δ-MnO₂(a1), α-MnO₂(b1) and β-MnO₂(c1), obtained by a hydrothermal method with different concentration ratio of C_k⁺/ C_H⁺=3.4, C_k⁺/ C_H⁺=0.85 and C_k⁺/ C_H⁺=0.2, respectively, indicating that δ-MnO₂ nanorods can be synthesized by controlling the concentration ratio of K⁺/H⁺ in the solution. The all three as-prepared MnO₂ samples contain physisorbed water and chemisorbed water which doesn’t affect the catalystic activity of the samples. The sequence of the as-prepared samples is δ-MnO₂≈α-MnO₂> β-MnO₂ in the catalystic oxygen reduction activity, which is in agreement with the sequence of the surface area of the samples. It is obvious that the concentration of K⁺ ion and H⁺ ion has a strong effect on the structure, morphology and electrochemical properties of the different type MnO₂ products.