|Home | About | Journals | Submit | Contact Us | Français|
As a cathode material, silver vanadium phosphorous oxide (Ag2VO2PO4) displays several notable electrochemical properties: large capacity, high current capability, and an effective delivery of high current pulses. These cell performance characteristics can be attributed to the presence of silver nanoparticles formed in-situ during the electrochemical reduction of Ag2VO2PO4. Specifically, changes in the composition and structure of Ag2VO2PO4 with reduction, especially the formation of silver nanoparticles, are detailed to rationalize a 15,000 fold increase in conductivity with initial discharge, which can be related to the power characteristics associated with Ag2VO2PO4 cathodes in primary lithium batteries.
Polyanion compounds in the general class of MPO4 are receiving significant attention in journal publications as materials for energy storage applications, most generally with compositions such as LixMPO4 where M equals Fe, Mn or VO.1–9 The materials demonstrate exceptional stability and can provide high voltage and capacity when used in lithium based batteries, but their characteristically low electrical conductivities10–14 are their biggest challenge for implementation as battery materials. In order to enhance LixMPO4 electrical conductivity, several strategies have emerged in the scientific literature, including coating of LixMPO4 particles with carbon, 15, 16 co-synthesizing the LixMPO4 materials with carbon to achieve intimate contact of the particles with the conductive material,13, 17 adding silver and copper powder to the LixMPO4 matrix to achieve improved conductivity,18, 19 or the solid-solution doping of LiFePO4.20 Unfortunately, the strategies involving the addition of an external conducting material require additional processing steps and can significantly reduce energy density due to the presence of the extraneous inert conducting material.
Another family of materials, copper vanadium oxides, have also been the subject of a number of recent published studies as cathode materials for primary and rechargeable systems.21–29 Notably, the electrochemical reduction of copper vanadium oxide occurs with the formation of very small metallic copper dendrites on the electrode surface.30, 31 Since the copper vanadium oxide powders retained much of their structural integrity with the formation of copper dendrites, it was proposed that copper vanadium oxides are well suited for rechargeable systems, due to the facility associated with copper insertion upon electrode oxidation (recharging).
Although not specifically stated in the copper vanadium oxide literature, we hypothesized that the formation of electrically conducting metal particles upon metal oxide reduction might result in an enhancement of electrical conductivity of the metal oxide, which should address conductivity problems for poorly conducting materials. Thus, we recently identified silver vanadium phosphorous oxide (Ag2VO2PO4)32 as a material of interest for next generation batteries, based on the desire to obtain the chemical stability observed in other phosphate cathode materials, achieve multiple electron transfer inherent in bimetallic materials, and provide the opportunity for the in-situ generation of a conductive silver matrix. In our initial communication, we reported that Ag2VO2PO4 displays multiple electron transfer per formula unit with the associated high capacity as well as high current capability on discharge.33
With this report, we describe our approach to conductivity improvement via in-situ silver metal formation, and thus rationalize the unexpected high current capabilities observed with Ag2VO2PO4. Our strategy differs from prior methods of conductivity enhancement of metal oxides, as the Ag2VO2PO4 conductivity is increased in-situ with the initial reduction of the electroactive Ag2VO2PO4 material. The electrochemical discharge process of Li/Ag2VO2PO4 electrochemical cells is fully characterized, where structural and chemical changes in the Ag2VO2PO4 cathode as a function of discharge are detailed, including formation of metallic silver nanoparticles and nanowires during discharge. To our knowledge, this is the first study of this strategy applied to the MPO4 class of materials, and while limited here to Ag2VO2PO4, the concept should be applicable to families of metal oxides and phosphorous oxides which form conducting metal particles upon reduction.
The preparation of silver vanadium phosphorous oxide (Ag2VO2PO4) via a topotactic exchange reaction of an organically template vanadyl phosphate has been previously described.34 A previously reported hydrothermal synthesis method was used to prepare Ag2VO2PO4 for this study.35 Ag2VO2PO4 was produced on a 0.5 g scale, by heating Ag2O, V2O5, and H3PO4 in aqueous solution at 230°C in a Teflon-lined autoclave for 96 hours. The yellow powdered sample was collected using vacuum filtration, then dried under vacuum.
X-ray diffraction (XRD) was measured using a Rigaku Ultima IV x-ray powder diffractometer, with Cu Kα radiation and Bragg-Brentano focusing geometry. Data was collected from 5 – 90° 2θ in continuous measurement mode, with a scan speed of 0.600 degrees/minute and a sampling width of 0.0200. Crystallite sizes were determined using the Scherrer equation after correcting for instrumental broadening using a LaB6 standard.
Scanning electron microscopy (SEM) data was collected on a Hitachi SU-70 field emitting scanning electron microscope equipped with an Oxford Inca energy dispersive x-ray spectroscopy (EDS) system. Secondary electron images were acquired at 5 kV. Backscatter electron images were observed and EDS data was collected at 20 kV.
Coin cells were fabricated within an Argon filled glove box. Some cathodes were prepared containing pure, as synthesized Ag2VO2PO4 material. For the pulsing study, Ag2VO2PO4 was mixed with graphite and poly(tetrafluoroethane). Polypropylene separator material and 1 M LiAsF6 in 50/50 (v/v) propylene carbonate/dimethoxyethane electrolyte were employed in all coin cells.
For cells containing pure Ag2VO2PO4 cathodes, AC impedance (ACI) measurements were performed using a CH Instruments Model 604C Electrochemical Analyzer. Fresh coin cells were measured using the frequency range of 0.1 mHz to 100 kHz at 37°C. Each coin cell was then discharged using a C/200 rate at 37°C on a Maccor Series 4000 battery tester. Discharge was stopped at various depths of discharge (DODs) and cells were allowed to recover at open circuit potential until their rate of voltage change minimized. The open circuit voltage data shown below (Figure 1) represent final voltages measured at the end of voltage recovery. AC impedance was then measured at each DOD to a low frequency of 10 mHz. An amplitude of 5 mV was used for all ACI measurements. Analysis of the ACI data involved the instant fit approach where the segments of the initial semicircles of the data were fit to a basic equivalent circuit (Figure 8).
A three-electrode setup was used for slow scan rate voltammetry (SSRV). The working electrode was a platinum disk coated with a Ag2VO2PO4:graphite mixture in the presence of polyacrylic acid. Lithium metal was used for the counter and pseudo-reference electrodes. SSRV was performed inside an argon filled inert atmosphere glovebox, using electrolyte of 1 M LiAsF6 in 50/50 (v/v) propylene carbonate/dimethoxyethane. A CH Instruments CHI604B potentiostat was used. Initial and final potentials of 3.5 V, and a lower potential of 1.5 V were used, with a scan rate of 2.00 × 10−5 V/s.
For cells containing Ag2VO2PO4 composite cathodes, each coin cell was discharged at 37°C using a pulse discharge test on a Maccor Series 4000 battery tester. For the pulse discharge test, cells were discharged for five seconds at alternating high current densities of 10, 20, and 30 mA/cm2, with low current discharge between pulses. The composite cathode data shown below (Figures 9 and and10)10) represent average values from three coin cells.
The objective of this study was to investigate the changes in Ag2VO2PO4 as it is reduced and lithiated during the discharge process in a primary lithium anode battery. As discussed below, metallic silver is a product of the Ag2VO2PO4 reduction, so the lithiation process could be conceptualized as shown below (Scheme 1). For simplicity, the general notation LixAg2-xVO2PO4 is used when describing multiple levels of discharge. The notation LiyAg2-yVO2PO4 + yAg0 is used specifically for y < 2, and the notation LizVO2PO4 + 2Ag0 is used specifically for z > 2 to describe the reduced and lithiated material.
Samples of LixAg2-xVO2PO4 where x ranged from 0 to 3.6 were electrochemically generated. Specifically, the Ag2VO2PO4 material was electrochemically reduced in a two electrode configuration using a lithium counter electrode and then isolated at various states of lithiation. The material samples were characterized to gain structural and compositional information about the discharge process as a function of lithiation.
The lithiation of Ag2VO2PO4 was conducted at a current density of 0.32 mA/cm2. As the cathodes reached the desired level of composition, discharge was halted and they were allowed to recover for several days at room temperature under zero current (Figure 1). The average initial open circuit voltage for the cells when first assembled was 3.44 V with a range of 3.41 to 3.46 V for the population of cells. The 24 hour recovered open circuit voltage steadily decreased with discharge where at a 90% discharge level the open circuit voltage reached 1.54 V. The electrochemically lithiated LixAg2-xVO2PO4 samples were used for further analysis.
The LixAg2-xVO2PO4 samples recovered from the discharged cells were used for analysis by Scanning Electron Microscopy (SEM). Note the as-synthesized Ag2VO2PO4 material has a uniform morphology, consisting of micron sized bladed particles (Figure 2). The individual particles are highly acicular, many with an aspect ratio > 20. The images also show striations along the long axis of the particles indicating that many of the rods consist of assemblies of thinner planes of materials.
SEM of the lithiated LixAg2-xVO2PO4 material was collected at different levels of lithiation. A backscatter electron detector was used to provide atomic number contrast. Even at early states of lithiation, small bright nanoparticles were visible on the surface of the micron sized bladed particles. A representative micrograph for Li0.8Ag1.2VO2PO4 + 0.8Ag0 is shown (Figure 3). Nanoparticles appeared on the surface of the micron sized bladed particles, with a regular, layered arrangement where they were arranged along the long dimension of the particles.
The bright nanoparticles persisted at high depths of discharge for LixAg2-xVO2PO4, and a smaller number of larger white particles were also observed. A representative micrograph for Li1.6Ag0.4VO2PO4 + 1.6Ag0 is shown (Figure 4). Localized energy-dispersive x-ray spectroscopy (EDS) was collected at three locations of interest (Table 1). For locations A and C, the Ag:V:P ratio was approximately 2:1:1, consistent with an Ag2VO2PO4 stoichiometry both on the flat surface and on the surface of the smaller gray micron sized particles. For the bright nanosized particle in location B, a higher Ag content was observed, consistent with our interpretation that the bright white regions are Ag0 nanoparticles and nanowires formed on discharge.
X-ray powder diffraction (XRD) data was collected for the Ag2VO2PO4 material as synthesized and for the lithiated material. A comparison of as synthesized material (Ag2VO2PO4), a partially lithiated material (Li1.2Ag0.8VO2PO4 + 1.2Ag0) and a highly lithiated material (Li3.2VO2PO4 + 2Ag0) is shown (Figure 5). The as-synthesized material showed excellent correlation to a literature pattern for Ag2VO2PO4, with no extra peaks observed in our data.36 As discharge progressed, the major Ag2VO2PO4 peaks decreased in intensity or disappeared, and new peaks emerged. The newly formed peaks showed excellent correspondence to a literature file for silver metal (Figure 5). This observation is consistent with the formation of silver metal upon discharge of Ag2V4O11.37 To monitor changes as a function of discharge, the changes in the major Ag2VO2PO4 peak (400) and the major Ag peak (111) intensity were plotted as a function of x in LixAg2-xVO2PO4 (Figure 6). The Ag2VO2PO4 peak intensity decreased linearly (open triangles in Figure 6) while the Ago peak intensity increased linearly (open circles in Figure 6) up to x = 2.4. If the reduction mechanism of Ag2VO2PO4 were to proceed stepwise with complete reduction of Ag+ → Ag0, followed by reduction of V5+, then the Ag0 peak intensity would be expected to reach its maximum at x = 2.0. Therefore, this XRD data suggests that some reduction of Ag+ and V5+ occurs in parallel.
Crystallite size changes for Ag0 were monitored. For crystallite size determination, the Ag0 (220) and (311) peaks were selected. The Ag0 crystallite size was small, and showed no systematic changes in either direction, with average sizes of 8.0 ± 0.8 nm in the (220) direction and 6.8 ± 1.1 nm in the (311) direction for x = 2.0 to 3.2 in LixAg2-x VO2PO4, where the ± values represent two standard deviations.
Since the material analysis of the reduced Ag2VO2PO4 confirmed the generation of silver nanoparticles, it was of interest to probe whether the nanoparticles had any functional impact on the resistance of the cathode material. Two electrode cells were constructed with pure Ag2VO2PO4 versus a lithium metal electrode. AC impedance was measured for the Li/Ag2VO2PO4 cells as first assembled, and when they had been only lightly discharged to compositions of x = 0.08 and 0.2 in LixAg2-xVO2PO4 (Figure 7). The Zview software package was used to determine the change in resistance of the cells with discharge. A simple equivalent circuit using a resistor, Rs, in series with two resistor/capacitor parallel combinations R1, C1 and R2, C2, was used (Figure 8). R s, representing electrolyte and other ohmic resistance elements, stayed relatively constant at 6.3, 5.1 and 6.3 ohms for the three compositions measured. The resistance of the first semicircle, R1, also stayed fairly constant at 18, 22 and 19 ohms at x = 0, 0.08 and 0.2 respectively. However, there was a significant change in the resistance, R2, of the second parallel circuit element as the resistor value was 810,000 at x = 0 and decreased to 54.9 and 20.5 ohms at x = 0.08 and 0.2. This illustrates an increase in conductivity of >15,000 times as the cell was discharged from x = 0 to 0.08 or 2% of the cell theoretical capacity. The small change in R1, yet profound change in R2, is fully consistent with the significant increase in the LixAg2-xVO2PO4 cathode conductivity early in the discharge of the cell. The change in R2 from x = 0.08 or 2% discharge to x = 0.2 or 5% discharge was much smaller as the resistance decreased by about a factor of 2 times. The significant decrease in resistance with initial discharge of the cells is consistent with the generation of a silver metal matrix within the cathode material as the Ag2VO2PO4 cathodes are discharged with the reduction of Ag+ to Ag0 and the accompanying displacement of the silver from the vanadium phosphate structure.
The confirmation of silver metal as a discharge product and the significant increase in conductivity with only modest levels of lithiation provided incentive to further probe the electrochemical performance of Ag2VO2PO4 as configured as a composite cathode versus a lithium metal anode. Two electrode coin cells were discharged using a pulse discharge test designed to test the ability of the Ag2VO2PO4 material to deliver high pulse currents in a primary lithium battery. The cells delivered 215 mAh/g above 1.3 V pulse loaded voltage at 30 mA/cm2 (Figure 9). We assigned the theoretical capacity as 272 mAh/g where it is assumed that reduction of Ag+ to Ag0 followed by reduction of the vanadium centers from V5+ to V4+ to V3+ takes place. If reduction of Ag+ to Ag0 is followed by reduction of the vanadium from only V5+ to V4+, the theoretical capacity would be 204 mAh/g. Thus, based on a theoretical capacity of 272 mAh/g, ~80% of the Ag2VO2PO4 material was accessed at the highest current density tested, yielding a material composition of Li3.2VO2PO4 + 2Ag0. The loaded voltage during the background segments of the discharge was fairly flat during the majority of discharge with some slope near the end of discharge.
The Li/Ag2VO2PO4 battery resistance throughout cell life was calculated and assessed (Figure 10). Resistance was calculated using the difference between the pre-pulse voltage and the minimum voltage of the pulse, divided by the pulse current. Detailed examination of the RDC as a function of depth of discharge reveals a significant change as the cell discharge initiates. The calculated RDC during the first applied pulse was 119 ohms. The second pulse applied at a current density of 20 mA/cm 2 or 3% DOD showed an RDC of 35 ohms and the third pulse under 30 mA/cm2 or 7% DOD showed a value of 24 ohms. On return to a pulse current of 10 mA/cm2 at 10% DOD, the RDC had decreased to 31 ohms, a factor of 3 reduction from its original value. This illustrates that the Ag2VO2PO4 is initially more resistive, but rapidly becomes more conductive as discharge progresses consistent with the AC impedance results and the identification of silver metal as a discharge product. The decrease in resistance between the first and second pulse is most notable, indicating that only a small level of discharge is needed to significantly improve overall conductivity.
A three-electrode cell configuration was used to study the electrochemistry of the Ag2VO2PO4 material when configured in a very thin layer on the surface of an electrode. Slow scan rate voltammetry data was collected for Ag2VO2PO4 versus metallic lithium counter and pseudo-reference electrodes (Figure 11). A major reduction peak was observed at 2.6 V, with a minor reduction peak at 2.9 V. These potentials are consistent with our previous observations of inflection points in an analysis of differential capacity versus voltage for the discharge of the Li/Ag2VO2PO4 cell system.33 Though not quantitative, the presence of one very large wave at 2.6 V in the voltammogram may indicate that under the three-electrode cell conditions, the reductions of silver and vanadium occur largely concurrently as was indicated by the XRD analysis. If the silver and vanadium reductions were significantly separated, we would expect reduction waves in a ratio of 2:1 for Ag+ to Ag0 reduction when compared with the reduction of V5+ to V4+ or V4+ to V3+.
The reduction of silver vanadium phosphorous oxide (Ag2VO2PO4) demonstrates the fundamental phenomenon that structural incorporation of a metal ion that is reduced to metallic particles or nanowires on initiation of reduction, leads to a significant increase in conductivity, in this case a factor of 15,000 times. This conductivity increase is of the material matrix itself independent of conductive additives. Additionally, silver vanadium phosphorous oxide can access four electrons per formula unit, resulting from the reduction of Ag+ → Ag0 and V5+ → V4+ → V3+. Conductivity increases achievable in bimetallic systems due to the formation of metallic nanowires or particles on reduction may be applicable not only to battery applications, but also for other systems benefitting from inherent electronic conductivity.
This project was supported by the New York State Foundation for Science, Technology, and Innovation, the University at Buffalo (SUNY), and the National Institutes of Health under Grant 1R01HL093044-01A1 from the National Heart, Lung, and Blood Institute.
Formation of silver nanoparticles from the electrochemical reduction of Ag2VO2PO4.