When lanthanum is partially substituted for calcium sites in the perovskite‑type oxide of CaMnO₃, high electronic conductivity is observed at room temperature due to the valence change of manganese and the easy change of oxygen content in the lattice [1] [2] . Therefore, this oxide is a candidate for air electrode of SOFCs [3] and/or active material in alkaline [4] -[11] and sea water batteries [12] . If this oxide was used for practical alkaline batteries, the discharge capacity density could be made larger. This is because, as the oxides show high electronic conductivity, conductive agents such as graphite are unnecessary in active materials. In practice, the sintered ceramic, which did not contain any conductive material, could be discharged; 1 g sintered ceramic of Ca_0.9La_0.1MnO_3−δ indicated a potential plateau region at −380 mV against Hg/HgO electrode for 17 h on 10 mA cathodicdischarge in KOH solution [7] . Although the ceramic did not work as a cathode material of rechargeable battery, the discharge capacity seemed to be dependent on the electrolyte solution. Considering an incombustible aqueous lithium battery, this phenomenon is very interesting to us.

M. Manickam et al. have already carried out the investigation of batteries using LiOHaq. and MnO₂ [13] - [15] . They used a few kinds of MnO₂ called chemical manganese dioxide (CMD) including chemically prepared battery grade manganese (BGM) and/or electrolytic manganese dioxide (EMD) such as γ-MnO₂. Especially, they reported a phenomenon not observed in KOH aq. solution, a reversible lithium insertion/desertion for γ-MnO₂ to form Li_xMnO₂. MnO₂s including γ-MnO₂ are all structured by linked octahedron, which are completely different from the perovskite-type one, and furthermore, not a good electronic conductor. Here, we synthesized high conductivity perovskite-type Ca_0.9A_0.1MnO_3−δ (A = Y, Sm, Gd and so on) in porous state by a sintering method, and investigated the electrochemical properties, especially the rechargeability of the porous ceramic sample as a cathode material in LiOHaq. solution.

The samples were prepared from reagent grade powders of CaCO₃ and Mn₂O₃ and some rare earth oxides A₂O₃ (A = La, Nd, Sm, Gd and Y) with 99.9% purity. These materials were weighed in the defined molar ratios, mixed, temporarily pressed into pellets, and pre-fired in an alumina boat at 1200˚C in air for 10 h. To obtain porous samples for electrochemical measurements, the pre-fired samples were crushed, mixed with NH₄HCO₃, pressed (600 kg∙cm⁻³) into disks (ca. 12 mm diameter by 1 mm thickness) and sintered again at 1300˚C in air for 10 h. The crystal structure of the samples was checked by X-ray diffraction (XRD) and the sample morphology by scanning electron microscopy (SEM).Electrical conductivity of the sintered porous samples was measured by the 2-terminal ac method with 10 kHz signal in the temperature range from room temperature to 1000˚C in air. The oxygen content, i.e. 3 − δ in Ca_0.9A_0.1MnO_3−δ at room temperature was calculated from the TGA result measured between room temperature and 1000 ˚C using the starting powder mixture.

To investigate the cathodic properties of the sample in LiOHaq. solution, the sintered sample disc (ca. 0.50 g) including no graphite was fixed on a platinum plate with a Pt lead in a polyethylene cell. This was immersed in 5% LiOHaq. solution and the current of 5 mA·g⁻¹ was passed through the test electrode (cathode material) and a counter electrode (platinum plate). In each case, the cathode potential was measured in the window range from −0.80 to +0.60 V against an Hg/HgO electrode at 30˚C. From the discharge curves, the discharge capacities of the cathode materials were calculated.

The oxides shown by Ca_1−xA_xMnO_3−δ are generally known to have fairly wide solid solution ranges. In the present experiment, the sample compositions were fixed to be x = 0.10 for various A elements. As a result, sintered (porous) samples were confirmed to be composed of an orthorhombic perovskite‑type phase in all cases. Of course, the crystal lattices somewhat changed depending on the ionic radius of A element; Ca_0.9Y_0.1MnO_3−δ showed the most packed unit lattice among them and Ca_0.9La_0.1MnO_3−δ the most open parallel structure, which coincided with the size of substituting elements [16] . However, the valence of manganese remaining could also influence the lattice constant. In fact, although the substituent composition is the same in Ca_0.9La_0.1MnO_3−δ and Ca_0.9Y_0.1MnO_3−δ, the values of 3 − δ were 2.88 to 2.91 and 2.83 to 2.85, respectively, which meant that Ca_0.9La_0.1MnO_3−δ had a little more tetravalent manganese than Ca_0.9Y_0.1MnO_3−δ. The samples thus obtained had ca. 60% porosity and showed high electrical conductivity; for example, the electrical conductivity of the sample Ca_0.9­Y_0.1MnO_3−δ was 1.4 × 10² S∙cm⁻¹ at room temperature in air as shown in Figure 1.

The electrodes, constructed with porous ceramic oxides, showed stable EMFs’ over 200 mV (vs­. Hg/HgO) in 5% LiOH solution at 30˚C. Figure 2 represents the typical discharge curves of the ceramic cathodes Ca_0.9A_0.1MnO_3−δ on 5 mA discharge. In these cases, after the initial sharp potential drops, the plateau regions were observed at 350 to 400 mV followed by sharp potential drops. The discharge times corresponding to the plateau regions are not so much different depending on the “A” elements employed; the discharge time for the sample of A = La is the longest and is 2 to 3 hours longer than the previous result in 15% KOH solution. Assuming the end potential of discharge as −800 mV, the specific discharge capacity was calculated to be 200 mAh∙g⁻¹, which was the completely the same as that we previously reported in the comparative experiments

[7] . This meant that this type of experiment was certainly reproducible. Furthermore, the aqueous solution with lithium ion was considered to be favorable to discharge. Considering this lithium insertion, Ca_0.9Y_0.1MnO_3−δ would be inferior to the other oxides because of the most packed structure. However, it showed good performance next to Ca_0.9La_0.1MnO_3−δ, which may be due to the delicate structural distortion by doping of a different trivalent ion to the calcium site.

When the sample composition was represented as CaMnO₃, the electrode reaction in an alkaline solution would be generally shown in the followings,

If LiOH solution was used, the next reaction could also be expected,

In any case, the discharge capacity due to the change of Mn⁴⁺ to Mn³⁺ should be 187.5 mAh·g⁻¹ at maximum. If some manganese were substituted by a rare earth element, tri-valent manganese could be formed to reduce the content of tetra-valent manganese, which would cut down the total discharge capacity. As we observed the actual discharge capacity of 200 mAh·g⁻¹ exceeding the maximum, we thought that the prolonged discharge time was principally due to the change of Mn⁴⁺ down to Mn²⁺, where the lithium insertion might be favorable in addition to hydrogen insertion. Then, we tried to check the lithium insertion into the bulk sample by ICP using the fully rinsed sample after discharge. Unfortunately, we could not strictly determine whether the obtained signal of lithium was based on the sample bulk or in the LiOH adhered tightly in the sample pore (surface). From the discharge result of the sample showing the discharge capacity less than 2 mAh·g⁻¹ done in 1 M LiClO₄/PC solution by 10 mA·g⁻¹ as an additional experiment, the effect of lithium insertion itself would be undesirable. Of course, the XRD diffraction analysis was carried out, the result of which is mentioned in the following paragraph.

Next, we compared the performances within the samples of Ca_0.9La_0.1MnO_3−δ and Ca_0.9Y_0.1MnO_3−δ Figure 3(a), and Figure 3(b) comparatively show the 5 mA discharge & charge curves for the sintered porous samples of Ca_0.9La_0.1MnO_3−δ and Ca_0.9Y_0.1MnO_3−δ with 0.5 g weight in 5% LiOH solution. The former sample showed the 1st discharge plateau with the capacity of ca. 200 mAh·g⁻¹, but the potential easily rose to evolve oxygen gas in the charging process.In the latter case, on the contrary, a somewhat inferior discharge performance was obtained, but the 2nd charge and the subsequent discharge were possible despite the fairly low capacity. This was a new finding on the CaMnO_3−δ based material in alkaline solution. In order to check the reaction, XRD was carried out for the sample of Ca_0.9Y_0.1MnO_3−δ after discharging and charging. As shown in Figure 4, some new peaks were mixed with those of Ca_0.9Y_0.1MnO_3−δ itself after discharge, which could be a reason of a long plateau region during discharging. After charging, such peaks disappeared to regenerate the bulk of Ca_0.9Y_0.1MnO_3−δ. Although the XRD peak near 2θ = 32˚ could be assigned to that of Li_xMnO₂, the exact conclusion could not be done because of its weakness.

Furthermore, these phenomena were recognized by SEM pictures; the sample situation was fairly different before and after discharging as seen in Figure 5, but it seemed to recover to the original situation after charging. As mentioned above, the XRD signals of the newly appeared peaks were so weak that we could not detect the exact phase of them. However, if a new phase with lower conductivity might inhibit continuous charge/discharge cycling, it might be improved by changing cutoff potentials. Figure 6 shows the experimental results done in a shallow cutoff potential range to compare with the previous one; the 1st discharge performance completely the same as before, but the potential for the quasi-plateau region considerably decreased during charging and the corresponding discharge time somewhat increased. As a result, the 3rd discharge was possible, although the total discharge capacity was extremely low. This phenomenon meant that, as expected, the main problem for charging was the property of the compound formed at the lower potential on discharge, probably the electrical conductivity.

In order to check whether the ceramic oxide Ca_0.9Y_0.1MnO_3−δ could be employed as an actual battery material or not, a cell was set up using this oxide as a cathode material, a zinc metal as an anode material and 5% LiOHaq. solution as an electrolyte solution. Figure 7 shows a 5 mA discharge curve at 30 C. The open circuit voltage was 1.45 V. At the beginning of the discharge, the voltage sharply decreased, but a fairly good plateau region was observed at about 1.0 V; assuming the end potential of discharge 0.8 V, the discharge capacity was calculated to be 167 mAh∙g⁻¹. In this case, on the other hand, the charge process did not carry out well, although the end potential was carefully chosen as 0.8 V which was only 0.20 V lower than that of the plateau region considering the result of Figure 6. In any case, the present result denoted that the oxide may be applicable to a kind of a primary battery system.

The sintered ceramic Ca_0.9A_0.1MnO_3−δ (A = La, Nd, Sm, Gd and Y) showed properties as a cathode material with no conductive powder such as graphite in 5% LiOHaq. solution. The discharge capacity changed depending on

the A elements and that Ca_0.9Y_0.1MnO_3−δ was 190 mAh∙g⁻¹, which was greater than 165 mAh∙g⁻¹ of Ca_0.9La_0.1MnO_3−δ in the previous experiment using 15% KOH aq. solution. This result shows the effect of lithium ions to promote the discharge reaction. On the other hand, the discharge capacity indicated that manganese was reduced partly from Mn⁴⁺ to Mn²⁺ in the oxide. By changing the cutoff potential shallow to depress formation of the compound including Mn²⁺, the 2nd or more discharge/charge was found to be available. Finally the cell of using zinc like (−)Zn/LiOHaq./Ca_0.9Y_0.1MnO_3−δ(+) was found to work as a primary battery with ca. 1.0 V plateau voltage.