Silver Vanadium Diphosphate Ag₂VP₂O₈: Electrochemistry and Characterization of Reduced Material providing Mechanistic Insights

Abstract

Silver vanadium phosphorous oxides (Ag_wV_xP_yO_z) are notable battery cathode materials due to their high energy density and demonstrated ability to form in-situ Ag metal nanostructured electrically conductive networks within the cathode. While analogous silver vanadium diphosphate materials have been prepared, electrochemical evaluations of these diphosphate based materials have been limited. We report here the first electrochemical study of a silver vanadium diphosphate, Ag₂VP₂O₈, where the structural differences associated with phosphorous oxides versus diphosphates profoundly affect the associated electrochemistry. Reminiscent of Ag₂VO₂PO₄ reduction, in-situ formation of silver metal nanoparticles was observed with reduction of Ag₂VP₂O₈. However, counter to Ag₂VO₂PO₄ reduction, Ag₂VP₂O₈ demonstrates a significant decrease in conductivity upon continued electrochemical reduction. Structural analysis contrasting the crystallography of the parent Ag₂VP₂O₈ with that of the proposed Li₂VP₂O₈ reduction product is employed to gain insight into the observed electrochemical reduction behavior, where the structural rigidity associated with the diphosphate anion may be associated with the observed particle fracturing upon deep electrochemical reduction. Further, the diphosphate anion structure may be associated with the high thermal stability of the partially reduced Ag₂VP₂O₈ materials, which bodes well for enhanced safety of batteries incorporating this material.

1. INTRODUCTION

Advances in energy storage materials have taken on expanded significance due to the needed ability to store electricity for many applications including aerospace, transportation, portable electronics, and biomedical devices.¹ Along with requirements for energy density and power, safety of the power sources under use and abuse conditions is a key consideration for implementation. While several components of a battery contribute to the overall safety of the system, cathode thermal stability has been noted as an important factor.² Recently, the use of phosphate based materials has led to increased cathode thermal stability, the most notable example being lithium iron phosphate, LiFePO_4, which has demonstrated increased thermal stability over oxide based battery materials.^{3, 4} While the voltage and energy density of phosphate based materials may be lower than oxide materials when used for lithium based systems, the chemical and electrochemical stability of the phosphate based materials are viewed as a significant asset. In order to achieve practical deployment of phosphate based materials, their inherently low electrical conductivity must be addressed.

Bimetallic cathode materials have been employed in batteries in part because they can provide multiple electron reduction per formula unit. A notable example is the silver vanadium oxide (Ag₂V₄O₁₁) system used in biomedical batteries that power implantable cardiac defibrillators (ICDs).^5-7 Similar to other vanadium oxides, Ag₂V₄O₁₁ has high voltage,⁵ and provides the opportunity for multiple electron reduction per Ag₂V₄O₁₁ formula unit. Additionally, an important factor in the success of silver vanadium oxide (Ag₂V₄O₁₁) for the ICD application is its ability to form in-situ silver metal nanoparticles upon reduction of Ag⁺ to Ag⁰, contributing to the high electrical conductivity of partially discharged silver vanadium oxide, necessary for the high pulse power needed for the defibrillation function.^7-12

We have embarked on the rational study of a new family of cathode materials for lithium batteries, namely silver vanadium phosphorous oxides (Ag_wV_xP_yO_z), with the broad goal of combining the thermal stability of lithium iron phosphate with the enhanced electrical conductivity of partially reduced silver vanadium oxide. We recently introduced the first member of the silver vanadium phosphorous oxide family for electrochemical study, Ag₂VO₂PO₄. When used as a cathode in a lithium based battery, Ag₂VO₂PO₄ displayed high discharge capacity and high current pulse capability, both promising attributes toward future use in high power biomedical applications.¹³ The chemical changes in the Ag₂VO₂PO₄ cathode as a function of discharge were examined.¹⁴ Most notably, as reduction was initiated, in-situ formation of silver nanoparticles in the cathode matrix was observed with an accompanying 15,000 fold increase in cathode conductivity. The impact of discharge on the resistance and impedance behavior of the cell was also studied.¹⁵ We also reported a novel ambient pressure synthesis method for Ag₂VO₂PO₄ preparation with accompanying changes in morphology, particle size, and surface area over the previously reported preparation method providing enhanced behavior under pulse discharge conditions.¹⁶ More recently, we examined a silver vanadium phosphorous oxide Ag_0.48VOPO₄·1.9H₂O with a silver to vanadium ratio of <1.0. ¹⁷ Lithium based electrochemical cells displayed a multi-plateau voltage profile and analysis of electrochemically reduced cathodes indicated significant formation of silver metal by reduction of Ag⁺ to Ag^o after ~ 0.37 electron equivalents were added. Thus, in this case, the in-situ formation of silver metal on electrochemical reduction was delayed compared to Ag₂VO₂PO₄ where formation of silver metal is the predominant process upon initial reduction.

While interest in phosphate based materials continues, there have been few reports of metal ion vanadyl diphosphate, M_xVP₂O₈, materials examined for lithium based batteries. Preparation of M = Na have been reported.^{18, 19} More recently, ion exchange of the Na based compound was utilized to prepare Li₂VP₂O₈ which discharged effectively and demonstrated reversibility over a limited range of Li ion insertion.²⁰ Some reports have focused on vanadium phosphate materials with a higher P/V ratio including Li₉V₃(P₂O₇)₃(PO₄)₂ where the material was highlighted as promising as a battery cathode.^{2, 21, 22} Diphosphate materials are chemically related to phosphate materials, but due to the P-O-P bond found in the diphosphate anion, we anticipate that the structural rigidity of diphosphate relative to phosphate might affect properties such as discharge profiles and rechargeability, especially in terms of usable depth of discharge and current capability. Thus, we report here the first electrochemical study of a diphosphate composition SVPO type material, Ag₂VP₂O₈. Notably, this material has a silver to vanadium (Ag/V) ratio of 2/1 similar to Ag₂VO₂PO₄, yet a vanadium to phosphorous (V/P) ratio of 1/2. The electrochemical reduction of Ag₂VP₂O₈ and the characterization of electrochemically reduced Ag₂VP₂O₈ material by x-ray diffraction, scanning electron microscopy, four point probe conductivity and differential scanning calorimetry are described herein. Rietveld analysis of the parent Ag₂VP₂O₈ material is presented. In addition, structural analysis contrasting the crystallography of the parent Ag₂VP₂O₈ with that of the proposed Li₂VP₂O₈ reduction product is employed to gain insight into the observed electrochemical reduction behavior. This work continues to demonstrate the importance of chemical composition and structure to tune electrochemical properties of cathode materials, including bimetallic phosphates and diphosphates.

2. MATERIAL AND METHODS

Ag₂VP₂O₈ was prepared following a previously reported synthesis method.²³ X-ray diffraction (XRD) data were collected at room temperature with a Rigaku Ultima-IV system (Cu KD radiation) with Bragg-Brentano configuration and a monochromator. The X-ray diffraction patterns were indexed to Ag₂VP₂O₈ (PDF 01-088-0436).²⁴ XRD data of a representative sample with a fixed-step width of 0.01q at 8 s/step for Rietveld analysis. Structure refinement was performed using the software package GSAS and the EXPGUI interface.^{25, 26} The profile was fit using Thompson-Cox-Hastings pseudo-Voigt profile function and the background was fit using a shifted Chebyshev polynomial with 6 terms.^{27, 28} Peak Fit software from Seasolve Software, Inc. was used to measure the full width at half-maximum of the silver peak located at 38.1° 2-theta. Crystallite sizes were determined using the Scherrer equation.^{29, 30}

Inductively coupled plasma optical emission spectroscopy (ICP-OES) was performed using a Thermo Scientific iCAP 6000 series ICP spectrometer. Differential scanning calorimetry (DSC) was conducted using a TA Instruments Model Q20. Scanning electron microscope (SEM) images were recorded using a Hitachi SU-70 field emitting scanning electron microscope. Four point conductivity measurements were obtained using a standard linear four-point probe arrangement. Magnetic susceptibility measurements were made using a Guoy balance, calibrated using powdered HgCo(SCN)₄ samples prior to use. Magnetic susceptibility data are based on replicate samples for each oxidation state.

Cyclic voltammetry was conducted at room temperature at a scan rate of 1 × 10^-5 V/sec using lithium reference and auxiliary electrodes. Lithium anodes were used in lab scale experimental cells with 1 M LiPF₆ in ethylene carbonate and dimethylcarbonate (30/70 by volume) as the electrolyte. Discharge tests were performed at 37°C. Reduction of the material under galvanostatic control was conducted at a current density of 0.05 mA/cm². A 0.19 mA/cm² current was applied for Galvanostatic Intermittent Titration Technique type testing. For some experiments, cathodes were recovered for further ex-situ characterization after the desired level of reduction was achieved.

3. RESULTS

3.1. Structure analysis

Structural analysis of the silver vanadium diphosphate material was undertaken. We report here the Rietveld refinement of Ag₂VP₂O₈ (Tables 1-3). Good agreement is observed between the calculated and observed XRD profiles (Figure 1). Atomic positions and U_iso values from the Rietveld refinement of the previously reported Ag₂VP₂O₈ were used as starting parameters.²⁴ Silver vanadium diphosphate has the P21/c space group with the structure having several notable features. The structure is formed by layers which stack along the y-direction which can be considered as VPO₈ chains comprised of VO₆ octahedra and PO₄ tetrahedra where each VO₆ octahedron shares one corner with P(2)O₄ (Figure 2). Each VO₆ octahedron and each P₂O₇ group has one free apex directed toward the interlayer, providing flexibility for the structure. Each layer consists of interconnected chains of [V₂P₂O₁₄]_∞ which run parallel to the x-axis. The chains are formed by two unique, polyhedral units (chain links). One unit is built of alternating VO₆-PO₄-VO₆-PO₄ polyhedral units which yield a rectangular cavity with a 3.005 Å shorter diagonal. The second unit [VP₂O₁₁] may be described as a P₂O₇ unit which shares terminal corners with a VO₆ octahedron and yields a triangular cavity with an area of 2.84 Ų. At the same time, the structure has a large 6 sided window providing tunnels that run along the b axis.

Table 1.

Crystallographic data and Rietveld refinement results of Ag₂VP₂O₈.

Crystal system	monoclinic
Space Group	P 21/c
Z	4
a (Å)	7.73734(6)
b (Å)	13.6060(11)
c (Å)	6.29592(5)
β (°)	99.0077(8)
V (Å3)	654.624(9)
Rwp	0.1219
Rp	0.0924
χ 2	1.978
S	1.41
2 θrange (°)	12-90
2 θstep width (°)	0.01
Number of reflections	522

The R factors and S are defined in Ref.⁴⁴

Table 3.

Selected Bond Distances (Å) of Ag₂VP₂O₈

V octahedron		P1 tetrahedron		P2 tetrahedron
V-O3	1.590(11)	P1-O1	1.468(12)	P2-O5	1.521(12)
V-O4	1.934(13)	P1-O2	1.499(11)	P2-O6	1.545(12)
V-O2	1.989(12)	P1-O8	1.503(12)	P2-O7	1.549(13)
V-O6	2.002(13)	P1-O7	1.678(13)	P2-O4	1.550(13)
V-O5	2.014(12)
V-O1	2.205(12)

Agl environment		Ag2 environment
Ag1-O8	2.414(12)	Ag2-O1	2.391(13)
Ag1-O8	2.418(12)	Ag2-O8	2.421(12)
Ag1-O3	2.464(11)	Ag2-O6	2.428(12)
Ag1-O3	2.509(12)	Ag2-O2	2.628(11)
Ag1-O5	2.657(12)

Figure 1.

Rietveld refinement results for Ag₂VP₂O₈

Figure 2.

Structure of Ag₂VP₂O₈

The polyhedral subunits comprising the silver vanadium diphosphate Ag₂VP₂O₈ exhibit unusual, but not unprecedented geometry (Figure 3A). The two, crystallographically unique tetrahedral units, P1-and P2-centered, can be considered a P₂O₇ unit. The P1-centered polyhedron is characterized by an unusually long P-O bond (1.678 (13) Å) to the bridging O (P₂O₇) accompanied by an unusually short PO bond (1.468(12) Å) to the “free” oxygen. The P2-centered polyhedron shares all corners with neighboring polyhedral and contains more isotropic bonds, with the bond to the bridging O (P₂O₇) slightly longer (1.550(13) Å) than the opposite P-O bond (1.521(12) Å). The seemingly unusual bond lengths and angles for this subunit (P₂O₇) are consistent with the local charge distribution surrounding the phosphorus environment. The P2-centered tetrahedron is electron deficient with P⁵⁺ surrounded by four corner-shared O^2- which yields a local charge of +3. The P1-centered tetrahedron is surrounded by three corner-shared O^2- and one free O^2- which yields a balanced charge (0) locally. Therefore, the long P1-O bond to the bridging oxygen is not unexpected.

Figure 3.

Structure of Ag₂VP₂O₈, A) emphasizing coordination environments of V and P polyhedra, B) emphasizing channels for Ag⁺ ions

The distorted VO₆ octahedron shares 5 corners with 5 P⁵⁺-tetrahedral units with the remaining oxygen “free” or not shared with neighboring polyhedra. The vanadium environment has been described as a regular O₆ octahedron with the vanadium atom displaced towards the free oxygen atom.³¹ The off-center displacement of the V⁴⁺ atom yields long (2.205(12) Å) and short (1.590(11) Å) apical V⁴⁺-O^2- bond distances. These distances are characteristic of vanadyl(IV) phosphates.^{24, 31-38} The unusual bond lengths may be rationalized by considering the local charge distribution surrounding the local vanadium environment. The V⁴⁺-centered octahedron is surrounded by five corner-shared O^2- and one free O^2- which yields an electron-rich region with a local charge of -3. The V⁴⁺ atom is displaced toward the center of the electron-deficient, triangular region formed by three, electron-poor 2P-tetrahedral units.

The monovalent silver ions, Ag⁺, that are part of the structure are considered to be located in two different crystallographically unique environments where the (A1) silver ions are seven coordinate and located between the layers while the remainder of the (A2) silver ions are located in the tunnels. When viewed down the x-axis, both Ag⁺ ions are able to conduct along both the x- and the z-direction. The Ag⁺ ionic conduction along the x- and z-directions may be described as ionic movement between the interconnected chains of [V₂P₂O₁₄]_∞ . When viewed down the y-axis, the Ag⁺ ions, Ag1 and Ag2, are more clearly distinguished as occupying the cavities created by the interconnection of [VPO₈]_∞ chains. The Ag1 atoms occupy a smaller (4-sided), yet more collimated cavity than the Ag2 atoms. The Ag2 atoms occupy a larger (6-sided), yet less collimated cavity (Figure 3B). Here the chemical uniqueness of these two atoms is evident and may result in differences in the conduction pathways and rates of conduction for each type of Ag⁺ ion.

3.2. Material Characterization

The silver to vanadium ratio of 2:1 and the vanadium to phosphorous ratio of 1:2 were confirmed with the use of inductively coupled plasma-optical emission spectroscopy (ICP-OES). Differential scanning calorimetry (DSC) was used to assess the thermal stability of the material and showed a featureless profile up to 600°C. Scanning electron micrographs (SEM) indicated micron sized particles. The surface area of the samples as determined by the BET method was typically 1.3 m²/g with a range of 1.2 to 1.5 m²/g.

3.3. Electrochemical Evaluation

Initial electrochemical performance was assessed utilizing a three electrode cell configuration to conduct cyclic voltammetry over a voltage range of 2.85 to 1.30 V (Figure 4). The first scan shows a major reduction peak at 2.07 V with a small shoulder at 1.97 V. The return scan has no oxidative peak indicating lack of reversibility over this voltage range. The second scan shows no major reduction or oxidation peaks with subsequent scans overlaying the second scan.

Figure 4.

Cyclic voltammetry of Ag₂VP₂O₈

The material was electrochemically reduced under galvanostatic control. After three electron equivalents of reduction, the voltage of the Ag₂VP₂O₈ material was above 1.5 V (Figure 5). The discharge curve shows an initial high voltage region until 0.2 electrons per formula unit followed by a long plateau at 2.3 V until a total of 1.8 electrons per formula unit were incorporated. From 1.9 electrons until 2.6 electrons per formula unit a slightly lower voltage plateau is observed. In total, 176 mAh/g of capacity were delivered.

Figure 5.

Galvanostatic discharge of Ag₂VP₂O₈

Galvanostatic Intermittent Titration Technique type testing was conducted. The data illustrate significant changes in the level of polarization with state of reduction (Figure 6). During the initial stages of the reduction, the polarization is larger and decreases continuously as the discharge progresses. During reduction between x = 1 electrons and x = 2 electrons per formula unit, a continuous decrease in polarization is notable, but at a lower slope than below x = 1 electrons. A further change to a smaller slope occurs with reduction past x = 2 electrons.

Figure 6.

Galvanostatic intermittent titration type test on Ag₂VP₂O₈

3.4. Characterization of reduced material

Partially electrochemically reduced material samples were interrogated using ex-situ XRD. Samples of Ag₂VP₂O₈ were reduced to 0.2, 0.5, 2.0, 1.5, and 2.0 electron equivalents and recovered for further analysis. X-ray powder diffraction patterns show that the presence of silver metal, Ag^o, can be observed in samples reduced by 0.2 electrons (Figure 7). Silver metal becomes more clearly visible in samples reduced by ≥0.5 electrons per formula unit. With reduction of 2.0 electrons per formula unit, the XRD pattern of the parent material pattern is no longer clearly visible with only silver metal observable. Note that the peak at 26.4° 2-theta is due to graphite added to the composite pellets.

Figure 7.

X-ray diffraction (XRD) of electrochemically reduced Ag₂VP₂O₈, where x = A) 0.2, B) 0.5, C) 1.0, D) 2.0, and E) 3.0

The integrated area of the silver peak at 38.1° 2-theta was determined for each sample and plotted versus the state of reduction (Figure 8). The area of the silver peak increases throughout the three electron reduction process. The crystallite size of the 38.1° 2-theta silver metal peak was also determined (Figure 8). Examination of the silver metal crystallite size as a function of reduction indicated a continuous decrease in the crystallite size of silver metal with electrochemical reduction.

Figure 8.

Integral area and crystallite size of Ag⁰ peak at 38.1° 2 theta versus electron equivalents for electrochemically reduced Ag₂VP₂O₈

Composite electrodes containing as-prepared and electrochemically reduced Ag₂VP₂O₈ were examined by scanning electron microscopy (SEM). As noted above, the parent material consists of micron sized particles (Figure 9A). After electrochemical reduction, small nanometer sized nodules appear dispersed on the surface of the Ag₂VP₂O₈ particles (Figure 9Bi). The same sample was imaged using backscatter imaging where the nodules now appear as bright spots dispersed among the particles (Figure 9Bii). These bright spots were assessed using energy dispersive spectroscopy (EDS) and identified as silver metal. SEM also shows evidence of particle fracture during the reduction process, as seen in the difference between the parent and reduced materials (Figure 9). This fracture was consistently observed throughout the electrochemically reduced sample, where the sample imaged prior to electrochemical reduction showed no evidence of fracture (Figure 10).

Figure 9.

Scanning electron microscope (SEM) images of composite electrodes containing Ag₂VP₂O₈ acquired at 20 kX magnification, A) prior to electrochemical reduction and B) after electrochemical reduction where x = 3.0. Images acquired using the secondary electron detector (i) and the backscatter electron detector (ii).

Figure 10.

Scanning electron microscope (SEM) images of composite electrodes containing Ag₂VP₂O₈ acquired at 1 kX magnification, A) prior to electrochemical reduction and B) after electrochemical reduction where x = 3.0. Images acquired using the secondary electron detector (i) and the backscatter electron detector (ii).

Conductivity measurements were made of composite electrodes containing as-prepared and electrochemically reduced Ag₂VP₂O₈ (Figure 11). The data showed an initial conductivity rise up to 0.2 electron equivalents. Further electrochemical reduction resulted in a decrease in conductivity, where the material reduced to x ≥ 1.0 showed a lower conductivity than the parent sample.

Figure 11.

Conductivity versus electron equivalents for parent and electrochemically reduced Ag₂VP₂O₈

Samples of electrochemically reduced Ag₂VP₂O₈ were recovered for ex situ magnetic susceptibility measurements. To our knowledge, the magnetic susceptibility data for the partially reduced Ag₂VP₂O₈ has not been previously investigated. The Ag₂VP₂O₈ material displayed a regular increase in magnetic susceptibility with discharge (Figure 12).

Figure 12.

Magnetic susceptibility versus electron equivalents for parent and electrochemically reduced Ag₂VP₂O₈

Thermal stability was examined by performing differential scanning calorimetry on as-synthesized and electrochemically reduced Ag₂VP₂O₈ samples (Figure 13). The parent material shows a featureless profile up to 580°C. The reduced materials show exothermic behavior above 500°C. Even the partially reduced samples show thermal stability up to 500°C establishing the promise for thermal stability of batteries using Ag₂VP₂O₈ as a cathode material.

Figure 13.

Differential scanning calorimetry for parent and electrochemically reduced Ag₂VP₂O₈

4. DISCUSSION

We anticipated that the silver vanadium diphosphate material, Ag₂VP₂O₈, would demonstrate conversion of Ag⁺ to Ag⁰ metal upon electrochemical reduction, reminiscent of our observations during the electrochemical reduction processes of silver vanadium phosphorous oxides, Ag₂VO₂PO₄ and Ag_0.48VOPO₄.^{13, 17, 39} Consistent with our hypothesis, the XRD data indicate the formation of silver metal with the onset of Ag₂VP₂O₈ reduction, where clear Ag⁰ formation is observed after 0.2 electron equivalents, with increased silver formation throughout the discharge (Figures 7, 8). Observation of nanosized silver metal nodules on the surface of the Ag₂VP₂O₈ particles by SEM is consistent with the XRD observation (Figure 9). Notably, the crystallite size of the silver metal decreases with continued reduction. Examination of the SEM images of partially reduced Ag₂VP₂O₈ provide rationalization for the observation in the XRD data of smaller crystallite size for the silver metal that is formed with discharge. Images of material samples reduced by 3.0 electron equivalents reveal significant material fracture where cracks along the long dimension of a particle are apparent (Figures 9, 10). The significant cracking of the material exposes new surface sites of the material providing additional locations for the formation of silver metal as it is displaced from the parent material. Thus, when the Ag₂VP₂O₈ material is initially reduced, fewer surface sites for the formation of silver metal islands are present. As the material fractures during the discharge process, additional surface sites appear providing the opportunity for formation of additional islands of silver metal. The increase in the number of silver formation locations would reduce the size of the deposit at each location resulting in determination of smaller average crystallite size for the overall population. Thus, these data support a reduction-displacement reaction where Ag⁺ is reduced to Ag^o and exits the structure, allowing detection of silver metal by XRD (Figures 7, 8) and observation of silver nodules on the surface of the Ag₂VP₂O₈ particles by SEM (Figures 9, 10).

The magnetic properties and electronic structure of Ag₂VP₂O₈ have been previously investigated.⁴⁰ The measured magnetic susceptibilities of V³⁺ and V⁴⁺ oxides and phosphates were determined to be vary from those anticipated from simple spin only calculations. Despite the layered crustal structure suggesting one- or two-dimensional magnetic behavior, Ag₂VP₂O₈ has been described as a spin dimer system. Due to the distortion of the octahedrally coordinated V⁴⁺ in this material, the single d electron occupies the dxy orbital oriented perpendicular to the short bond. The result is a one-dimensional chain system, which has been described as a frustrated alternating chain.⁴¹

Ex situ magnetic susceptibility data is expected to change as a function of discharge and thus, provide insight into the progression of discharge. Prior to discharge, the vanadium in Ag₂VP₂O₈ has an oxidation state of 4+, with one d electron per magnetic atom. Upon reduction from V⁴⁺ → V³⁺ the vanadium d electron configuration changes from d¹ → d², while reduction from Ag⁺ → Ag⁰ has no significant change, retaining a d¹⁰ electron configuration. Therefore, the changes in the vanadium oxidation state should be reflected by the magnetic susceptibility measurements in the presence of the silver reduction process.

We can rationalize the increase in magnetic susceptibility as a function of electrochemical reduction (Figure 12) which supports the reduction of vanadium in conjunction with the reduction of silver as observed from the XRD data. If the Ag⁺ were entirely reduced fully prior to the V⁴⁺ reduction, then we would anticipate no significant magnetic susceptibility changes on initial reduction from 0 to 2 electron equivalents (Ag₂VP₂O₈ → Li₂VP₂O₈ + 2Ag⁰). Notably, the data show a ~3 fold increase in magnetic susceptibility in this range consistent with the previously reported reduction of Ag₂VO₂PO₄,^{39, 42, 43} where some reduction of V⁵⁺ occurs in parallel with the Ag⁺ reduction.

Thus, based on the XRD data as well as the magnetic susceptibility data, the reactions taking place during the reduction of Ag₂VP₂O₈ can be represented by the following scheme where in the first reaction the reduction of Ag⁺ is noted and in the second reaction reduction of V⁴⁺ is noted. While the reactions are shown as sequential, in fact, they are occurring concurrently.

$${\mathrm{Ag}}_2{\mathrm{VP}}_2{\mathrm{O}}_8+\mathrm{yLi}\to {\mathrm{Li}}_{\mathrm{y}}{\mathrm{Ag}}_{2\text{-}\mathrm{y}}{\mathrm{VP}}_2{\mathrm{O}}_8+{\mathrm{yAg}}^0$$ (1)

$$+\mathrm{zLi}\to {\mathrm{Li}}_{\mathrm{z}}{\mathrm{VP}}_2{\mathrm{O}}_8+2{\mathrm{Ag}}^0$$ (2)

Comparison of the isostructural materials Ag₂VP₂O₈ and Li₂VP₂O₈ provides insight into the behavior of the Ag₂VP₂O₈ material as a function of electrochemical reduction. As the reduction progresses, we would expect that as the Ag⁺ ions are displaced from the structure, the sites where the Ag⁺ ions were located would be occupied by Li⁺ ions. The structure of Li₂VP₂O₈ has been previously reported²⁰ and is shown here (Figure 14). While Li₂VP₂O₈ is isostructural with the parent Ag₂VP₂O₈ material, there are significant differences between the two materials. From analysis of the local polyhedra, the Li-analogue exhibits larger cavities within each V-O-P-O chain, but smaller cavities between the chains. The Li₂VP₂O₈ compound shows more distortion in its VO₆ polyhedra, with a V-O1 bond length of 2.32 Å, a >5% increase relative to the V-01 bond length of Ag₂VP₂O₈. In addition, the complete replacement of Ag⁺ by Li⁺ would result in >8% (>1.1 Å) decrease in the interlayer spacing (b-lattice parameter). Thus, the total replacement of Ag⁺ by Li⁺ would result in significant structural modification. This structural analysis is consistent with the observed loss of signal in the XRD patterns of the reduced material samples, where the increased structural distortion upon conversion from Ag₂VP₂O₈ to Li₂VP₂O₈ may result in a loss of crystallinity for the material. Further, the SEM images show evidence of macroscopic decrepitation of the particles upon electrochemical reduction, where the particle fracture may result from the crystallographic stress induced by the electrochemical reduction and the replacement of Ag⁺ by Li⁺. Similar decrepitation of the particle structure has been previously observed upon conversion from Na₂VP₂O₈ to Li₂VP₂O₈, where a 0.9 Å change in the interlayer spacing (b-lattice parameter) was noted.²⁰

Figure 14.

Structure of Li₂VP₂O₈, A) emphasizing coordination environments of V and P polyhedra, B) emphasizing channels for Li⁺ ions

The proposed mechanism can also be used to rationalize the conductivity data recorded for composite electrodes of the reduced material (Figure 11). On initial reduction, the initial rise in conductivity can be rationalized by the formation of silver metal and the increase in electronic conductivity due to the formation of highly conductive silver particles within the electrode structure. However, as the reduction progresses the parent particles fracture. The SEM images show that crevices form as a result of particle fracture. These fracture points result in gaps leading to loss of electrical contact within the electrode matrix, resulting in lower conductivity at higher levels of electrochemical reduction, and lower conductivity for the electrodes containing highly reduced material relevant to those containing the parent Ag₂V₂PO₈ material. This magnitude of decreased conductivity at high stages of reduction is not observed for Ag₂VO₂PO₄,³⁹ which also does not demonstrate particle fracture upon electrochemical reduction by SEM. Thus, the observed particle fracture and significant decrease in conductivity may be associated with the structural rigidity of the diphosphate anion.

Finally, the thermal stability of the Ag₂VP₂O₈ material is of note (Figure 13). The parent material shows a featureless profile as measured by DSC up to 580°C. Even the partially reduced samples show thermal stability up to 500°C establishing the promise for thermal stability of batteries using Ag₂VP₂O₈ as a cathode material.

5. CONCLUSIONS

In summary, the first electrochemical evaluation of a silver vanadium diphosphate, Ag₂VP₂O₈ has been reported herein. Ex-situ characterization of the reduced material via x-ray diffraction, magnetic susceptibility, scanning electron microscopy, and conductivity measurement has provided mechanistic insights. Reminiscent of Ag₂VO₂PO₄ reduction, in-situ formation of silver metal nanoparticles was observed upon reduction of Ag₂VP₂O₈. However, counter to Ag₂VO₂PO₄, the Ag₂VP₂O₈ material demonstrates a significant decrease in conductivity and a concomitant fracture of the material particles upon continued electrochemical reduction. A structural analysis contrasting the crystallography of the parent Ag₂VP₂O₈ with that of the proposed Li₂VP₂O₈ reduction product reveals significant distortion and interlayer collapse upon replacement of Ag⁺ with Li⁺. Thus, the presence of the rigid disphosphate anion under partial discharge allows for the retention of the anionic superstructure during discharge, while under more complete discharge, the anionic superstructure displays a pronounced structural collapse resulting in a disruption of the conducting silver matrix. These initial observations demonstrate the structural flexibility demanded under deep discharge applications. Finally, a notable feature of Ag₂VP₂O₈ including partially reduced samples is high thermal stability which bodes well for safety of batteries incorporating this material.

Highlight.

• First electrochemical study of a silver vanadium diphosphate, Ag₂VP₂O₈

• in-situ formation of Ag⁰ nanoparticles was observed upon electrochemical reduction

• structural analysis used to provide insight of the electrochemical behavior

Table 2.

Atomic positions of Ag₂VP₂O₈.

Atom	Site	x	y	z	aUiso(Å2)	bOcc.
V	4e	0.2253(6)	0.4226(3)	0.2754(7)	0.025	1.00
P1	4e	0.3641(9)	0.6273(5)	0.5467(11)	0.025	1.00
P2	4e	0.1761(9)	0.4784(5)	0.7489(11)	0.025	1.00
Ag1	4e	0.0642(2)	0.70236(11)	0.1494(3)	0.025	1.00
Ag2	4e	0.5130(3)	0.35356(11)	0.9135(3)	0.025	1.00
O1	4e	0.3637(16)	0.5623(9)	0.361(2)	0.025	1.00
O2	4e	0.4607(16)	0.3602(8)	0.3158(18)	0.025	1.00
O3	4e	0.1278(14)	0.3232(8)	0.1970(18)	0.025	1.00
O4	4e	0.2312(15)	0.4060(9)	0.5815(22)	0.025	1.00
O5	4e	0.0179(17)	0.5140(8)	0.2368(18)	0.025	1.00
O6	4e	0.2695(17)	0.4558(8)	-0.021(2)	0.025	1.00
O7	4e	0.2218(17)	0.5863(10)	0.702(2)	0.025	1.00
O8	4e	0.2826(16)	0.7231(9)	0.465(2)	0.025	1.00

^aU_iso is the isotropic atomic displacement parameter when the Debye-Waller factor is represented as exp(-8π²Usin² θ/λ²).

^bOccupancy