Tailoring the Ag⁺ Content within the Tunnels and on the Exposed Surfaces of α-MnO₂ Nanowires: Impact on Impedance and Electrochemistry

Silver hollandite was prepared via ambient pressure reflux with reagent mixtures containing different Ag/Mn ratios using a method previously reported.^13,15 The abbreviation Ag_xMn₈O_16-y, where x = Ag content, is used here for simplicity. Ag₂O coated Ag_xMn₈O_16-y was synthesized via a coprecipitation reaction as follows. The above synthesized Ag_xMn₈O_16-y powder was added to a N₂ degassed AgNO₃ solution at 0°C with vigorous stirring to obtain a suspension. A solution of NaOH was added to the Ag_xMn₈O_16-y suspension. Three different molar ratios, 2:1, 1:1 and 1:2 of Ag_xMn₈O_16-y to Ag₂O were prepared with 1:1, 1:2 and 1:4 molar ratios of Ag_xMn₈O_16-y to AgNO₃. The abbreviation Ag_1.13Mn₈O_16-y·aAg₂O is used for simplicity. After chemical precipitation, the as-prepared samples were washed with DI water and dried in vacuo. Pure Ag₂O was also synthesized using the same method without adding silver hollandite.

Powder X-ray diffraction (XRD) was employed to characterize the crystal structure and estimate crystallite sizes. The XRD patterns were measured using a Rigaku Smart Lab X-ray diffractometer with Cu Kα radiation and Bragg−Brentano focusing geometry. An MDI JADE version 8.5.3 software with ICDD and NIST databases was used for search-match analysis. A TA Instruments SDT Q600 was used to collect simultaneous thermogravimetric analysis/differential scanning calorimetry (TGA/DSC) for determination of oxygen content of silver hollandite samples according to the method previously described.²⁷ A Thermo Scientific Nicolet iS10 spectrophotometer with attenuated total reflectance accessory (ATR) was applied in the wave number range from 4000 cm⁻¹ to 500 cm⁻¹ for FT-IR. A Thermo Scientific ICAP 6300 radial inductively coupled plasma optical emission spectrometer (ICP-OES) was used for the quantitative Ag and Mn determination. Transmission electron microscopy (TEM) data was collected by a transmission electron microscope (JEOL JEM 1400) equipped with a Gatan CCD camera.

The ratios of Ag_xMn₈O_16-y to Ag₂O were determined using the following method. First, the Ag/Mn ratio of the parent silver hollandite was determined. Then, the total Ag (from both hollandite and silver oxide) and Mn amounts were determined for the coated materials. The value of a in Ag_xMn₈O_16-y·aAg₂O was then calculated according to the silver amount in hollandite and in silver oxide.

Electrodes were prepared by pressing pellets of the as-synthesized materials and the densities of the pellet electrodes are shown in Table I. Coin type cells were fabricated within an argon glove box with lithium metal anodes and 1 M LiPF₆ in 70/30 (V/V) dimethyl carbonate/ethylene carbonate electrolyte. Intermittent discharge with a current density of 1 mA/g was employed. The discharge process was stopped every 1% of theoretical capacity three times and every 5% of theoretical capacity for the remainder of the test. AC impedance measurements were measured after ∼12 h at OCV using a BioLogic VSP impedance analyzer with a 10 mV amplitude and a frequency range of 100 kHz to 0.1 Hz. The Nyquist plots were normalized assuming a zero intercept at the high frequency x-axis intercept for comparison of R_CT. Equivalent circuit analysis was performed using Z-view software.¹⁶²⁸

The XRD patterns of representative as-prepared hollandites are presented in Figure 1, red traces, for low (a) and high (b) silver content. All peaks can be indexed to the reference pattern (JCPDS card No. 97-006-015). The XRD patterns for samples coated with Ag₂O show intense PXRD reflections at 34°, 37° and 56° 2θ, in agreement with cubic Ag₂O, Figure 1, green traces. Samples prepared with lower Ag₂O content gave less intense XRD peaks consistent with the ability to tune the coating amount. For samples with low levels of Ag₂O, XRD could not readily discern the peaks corresponding to the silver oxide phase.²⁹ The presence of lower amounts of the Ag₂O coating was confirmed by alternate techniques, as described below.

Zoom In Zoom Out Reset image size

Empirical formulas for all the materials were determined by ICP-OES in conjunction with TGA. Typical as-prepared hollandites resulted in of Ag_xMn₈O_16-y, x = 1.13 and x = 1.54, for the low silver and high silver materials, respectively. The oxygen contents of y = 15.15 (x = 1.13, low silver) and 15.48 (x = 1.54, high silver) were determined yielding formulas of Ag_1.13Mn₈O_15.15 and Ag_1.54Mn₈O_15.48.²⁷ The fraction of Ag₂O in the final precipitate was roughly proportional to the Ag_xMn₈O_16-y: Ag ratio used during synthesis, Table II.

FTIR spectra were recorded for all the samples, Figure 2. Pure Ag₂O demonstrates two main peaks at 1380 and 881 cm⁻¹, consistent with previously published data at 1379 and 901 cm⁻¹.³⁰ As the Ag₂O/Ag_1.13Mn₈O_15.15 ratio increases, the FTIR spectra continue to demonstrate the silver hollandite peaks, while the peaks detected at ∼1380 cm⁻¹ increase. The XRD and FTIR data indicate that the silver hollandite structure remains intact throughout the coating procedure³¹ with no apparent change of Ag_1.13Mn₈O_15.15 lattice constants despite the presence of Ag₂O.

Zoom In Zoom Out Reset image size

The presence of a silver rich composition to the nanoparticle coating was confirmed by TEM, Figure 3. As-prepared Ag_1.13Mn₈O_15.15 and Ag_1.54Mn₈O_15.48 (Figures 3a, 3b) are ∼15 nm diameter nanorods with smooth surfaces. With increasing Ag₂O content, nanoparticles up to size 5∼10 nm are visible on the surface of the nanorods (Figures 3g, 3h). HAADF-STEM and STEM-EDS images confirmed these particles contain mostly silver, Figure S1, suggesting these are the deposited Ag₂O. It can be seen in all images that the deposited nanoparticles are distributed uniformly.

Zoom In Zoom Out Reset image size

Lithium anode electrochemical cells were prepared using Ag₂O, Ag_xMn₈O_16-y and the Ag_xMn₈O_16-y·aAg₂O composites as cathodes. Electrochemical impedance spectroscopy was used to characterize the cells before and after reduction. The Nyquist plots of all cells showed a similar characteristic semicircle, Figures 4–8. The data were fitted to an equivalent circuit consisting of a resistor (Rs), a parallel combination of a resistor (Rct1) and a constant phase element (CPE), the other parallel combination of a resistor (Rct2) and CPE and a Warburg element (open, W₀), Table III.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

A pure Ag₂O cathode was characterized as a control. The voltage profile is shown in Figure 4a, where the initial loaded voltage of 2.96 V recovers to 3.1–3.2 V within the first 0.004 molar electron equivalents of reduction. This phenomenon of voltage increase upon reduction has been observed previously in the discharge of Ag₂VO₂PO₄ and was attributed to Ag⁰ nanoparticle formation,³² with concomitant decrease in resistance. For the Li/Ag₂O cell, ∼1000-fold reduction in impedance was observed upon reduction by 0.02 molar electron equivalents (501 Ω) compared to that before discharge (339560 Ω), Figure 4b. Discharge continues with two voltage plateaus, at ∼3.1 V, and 2.2 V until ∼0.52 electron equivalents (∼65 mAh/g) followed by a steeper decrease. Notably, cell resistance continued decreasing with depth of discharge to a final value of 176 Ω at a voltage of 1.5 V. Above 2.2 V, Ag₂O contribution to capacity (20 mAh/g, 0.17 electron equivalents) is small. For the electrode sample with maximum Ag₂O content (Ag_1.54Mn₈O_15.48·1.43Ag₂O), a maximum contribution to capacity of 6 mAh/g above 2.2 V can be ascribed to Ag₂O based on mass.

The discharge profiles of pure pellets comprised of only Ag_1.13Mn₈O_15.15 or Ag_1.54Mn₈O_15.48 materials are shown in Figures 5a and 5b, respectively. Both hollandite materials show abrupt decreases in their voltages upon initial reduction. Above 2.2 V, Ag_1.13Mn₈O_15.15 discharges 0.24 electron equivalents and delivers 8 mAh/g capacity. In contrast, Ag_1.54Mn₈O_15.48 delivers 1.04 electron equivalents and 31 mAh/g. This is the first direct observation of hollandite with high Ag content delivering higher capacity during electrochemical reduction compared to low Ag content hollandite.²⁷ This observation may be attributed to limited electronic conductivity, as shown in the AC impedance of Ag_1.13Mn₈O_15.15 and Ag_1.54Mn₈O_15.48 before and after discharge in Figures 5c and 5d. The initial impedance is ∼3000 Ω and ∼1000 Ω for Ag_1.13Mn₈O_15.15 and Ag_1.54Mn₈O_15.48, respectively_. Upon discharge to 0.08 electron equivalents the impedance for the Ag_1.13Mn₈O_15.15 cell increased to ∼5000 Ω while the impedance for Ag_1.54Mn₈O_15.48 decreased to ∼500 Ω. Further reduction led to increased impedance in both cases with a more significant increase for Ag_1.13Mn₈O_15.15 at a lower molar electron equivalent than Ag_1.54Mn₈O_15.48 (see Table III).

Zoom In Zoom Out Reset image size

Lithium cells using Ag_1.13Mn₈O_15.15·0.25Ag₂O or Ag_1.54Mn₈O_15.48·0.25Ag₂O as cathodes showed sloping voltage profiles with discharge, Figures 6a and 6b. Above 2.2 V, Ag_1.54Mn₈O_15.48·0.25Ag₂O delivered the higher capacity, 46 mAh/g, compared to 30 mAh/g for Ag_1.13Mn₈O_15.15·0.25Ag₂O, higher than that of Ag_1.54Mn₈O_15.48 and Ag_1.13Mn₈O_15.15 (31 mAh/g and 8 mAh/g, respectively). The Nyquist plots for the cells are shown in Figures 6c and 6d. The initial impedance is ∼500 Ω for both materials (Table III), which represents a substantial decrease relative to the uncoated silver hollandite. Notably, the impedance increases with discharge. For Ag_1.13Mn₈O_15.15·0.63Ag₂O and Ag_1.54Mn₈O_15.48·0.63Ag₂O, (Figures 7a and 7b) the initial loaded voltage of 3.19 V and 3.23 V recovers to 3.23 V and 3.32 V within the first 0.01 electron equivalents of inserted Li⁺. With further discharge, the voltage profiles of both materials slope, and the Ag_1.13Mn₈O_15.15 structure now outperforms Ag_1.54Mn₈O_15.48, delivering ca. 90 vs. 65 mAh/g above 2.2 V. This correlates with an observed reduced impedance during initial discharge for Ag_1.13Mn₈O_15.15 (Figure 7c). We note both Ag_1.13Mn₈O_15.15·0.63Ag₂O and Ag_1.54Mn₈O_15.48·0.63Ag₂O showed an increase in impedance with continued discharge (Table III), further suggesting Mn reduction rather than continued formation of Ag⁰.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

The discharge profiles of Ag_1.54Mn₈O_15.48·1.43Ag₂O and Ag_1.13Mn₈O_15.15·1.43Ag₂O are shown in Figures 8a and 8b. Both curves showed sloping profiles, and above 2.2 V Ag_1.13Mn₈O_15.15 yielded higher capacity than Ag_1.54Mn₈O_15.48. Notably, above 1.5 V the capacity of the Ag_1.54Mn₈O_15.48·1.43Ag₂O cell reaches 230 mAh/g, which is the highest value of all the cells discussed above. The high capacity can be attributed to significantly lower impedance at high levels of discharge (Table III), concurrent with the appearance of Ag metal by XRD (Figure 9).

Zoom In Zoom Out Reset image size

A plot of energy density above 2.2 V vs. the Ag₂O content in the pure Ag_xMn₈O_16-y materials and the Ag_xMn₈O_16-y·aAg₂O composites compares the electrochemistry of lower silver (x = 1.13) and higher (x = 1.54) intra-tunnel silver containing materials, Figure 10. For materials with no (a = 0) or low (a = 0.25) Ag₂O content, the higher intra-tunnel silver material (x = 1.54) exhibits higher energy density, which can be ascribed to 2–10x lower impedance values. At higher Ag₂O values (a = 0.63 or 1.43) the lower intra-tunnel silver material (x = 1.13) exhibits higher energy density as well as lower impedance. Thus, despite the lower amount of silver, Li⁺ conduction becomes a more significant contributor to deliverable energy density once electron conduction is available. This trend generally holds for delivered capacity for the series of materials as well, Figure 10 inset.

Zoom In Zoom Out Reset image size