Scholarly article on topic 'Roles of Processing, Structural Defects and Ionic Conductivity in the Electrochemical Performance of Na3MnCO3PO4 Cathode Material'

Roles of Processing, Structural Defects and Ionic Conductivity in the Electrochemical Performance of Na3MnCO3PO4 Cathode Material Academic research paper on "Nano-technology"

0
0
Share paper
OECD Field of science
Keywords
{""}

Academic research paper on topic "Roles of Processing, Structural Defects and Ionic Conductivity in the Electrochemical Performance of Na3MnCO3PO4 Cathode Material"

Roles of Processing, Structural Defects and Ionic Conductivity in the Electrochemical Performance of Na3MnCO3PO4 Cathode Material

Chuanlong Wang,a'b Monica Sawicki,a'b James A. Kaduk,c and Leon L. Shawa'b'*'z

aDepartment of Mechanical, Materials and Aerospace Engineering, Chicago, Illinois 60616, USA b Wanger Institute for Sustainable Energy Research, Chicago, Illinois 60616, USA c Department of Chemistry, Illinois Institute of Technology, Chicago, Illinois 60616, USA

Na3MnCO3PO4 with a potential to deliver two-electron transfer reactions per formula via Mn2+/Mn3+ and Mn3+/Mn4+ redox reactions and a high theoretical capacity (191 mAh/g) can play an important role in Na-ion batteries. This study investigates the dependence of the electrochemical performance of Na3MnCO3PO4-based sodium-ion batteries on processing, structural defects and ionic conductivity. Na3MnCO3PO4 has been synthesized via hydrothermal process under various conditions with and without subsequent high-energy ball milling. Particle sizes, structural defects and ionic conductivity have been studied as a function of processing conditions. It is found that Na3MnCO3PO4 nanoparticles (20 nm in diameter) can be produced from hydrothermal synthesis, but the reaction time is critical in obtaining nanoparticles. Nanoparticles exhibit a higher ionic conductivity than agglomerated particles. Further, structural defects also have a strong influence on ionic conductivity which, in turn, affects the charge/discharge capacities of the Na3MnCO3PO4-based sodium-ion batteries. These results provide guidelines for rational design and synthesis of high capacity Na3MnCO3PO4 for Na-ion batteries in the near future.

© The Author(s) 2015. Published by ECS. This is an open access article distributed under the terms of the Creative Commons Attribution 4.0 License (CC BY, http://creativecommons.org/licenses/by/4.0/), which permits unrestricted reuse of the work in any medium, provided the original work is properly cited. [DOI: 10.1149/2.0801508jes] All rights reserved.

Manuscript submitted February 17, 2015; revised manuscript received April 20, 2015. Published May 28, 2015.

High power and high energy density rechargeable batteries with long cycle life and low cost are in urgent demand for electric vehicles and large-scale energy storage devices.1-4 In light of these emerging demands, interest in Na-ion batteries (NIBs) has increased recently5-27 because of the abundance and low cost of Na metal. Although NIBs have the cost advantage over Li-ion batteries (LIBs), it is generally agreed that the gravimetric and volumetric energy densities of NIBs cannot compete with that of LIBs because of two intrinsic shortcomings. First, Li has a higher ionization potential than Na.15 Second, Li+ ions are smaller and lighter than Na+ ions.15 One way to mitigate these intrinsic deficiencies is to search for high capacity anodes and cathodes.7-27 In this context, Na3MnCO3PO4 has been predicted via ab initio calculations28,29 to be capable of delivering two electron transfer redox reactions per formula and thus has a high theoretical capacity (i.e., 191 mAh/g of Na3MnCO3PO4). The earlier experiments, however, were only able to obtain a low specific capacity of 125 mAh/g, i.e., ~65% of the theoretical.13 Recently, our group has shown that high specific capacities, reaching as high as 92% of the theoretical, can be achieved when aided by the addition of 53 wt% (60 vol%) carbon black (CB) which provides a continuous CB network interacting with almost all Na3MnCO3PO4 particles in the cathode and allow them to participate in electrochemical reactions.30

The aforementioned theoretical prediction and experimental investigations13,28-30 reveal clearly that Na3MnCO3PO4 has a potential to be a high capacity cathode material for NIBs. Therefore, in this study, we have investigated factors, other than the electronic conductivity, that can affect the electrochemical properties of Na3MnCO3PO4 as the cathode material for NIBs. Specifically, the effects of materials synthesis conditions and the structural defects within the Na3MnCO3PO4 crystal on the ionic conductivity of Na3MnCO3PO4 have been studied. Furthermore, how the structural defects and the ionic conductivity influence the electrochemical properties of Na3MnCO3PO4 has been investigated. These studies reveal that the characteristics of the Na3MnCO3PO4 powder are significantly affected by materials synthesis and processing conditions. Structural defects can reduce the ionic conductivity of Na3MnCO3PO4 by two orders of magnitude. Furthermore, cells made of Na3MnCO3PO4 with higher ionic conductivities display higher discharge specific capacities

* Electrochemical Society Active Member. zE-mail: lshaw2@iit.edu

than cells made of Na3MnCO3PO4 with lower ionic conductivities. The detailed results of these findings are described below.

Materials and Methods

Na3MnCO3PO4 powder was synthesized via hydrothermal method using Na2CO3 (Alfa Aesar, 98%) as both sodium and carbonate sources, Mn(NO3)2 • 4H2O (Alfa Aesar, 99.98%) as manganese source, and (NH4)2HPO4 (Alfa Aesar, 98%) as phosphate source. 0.02 mol Mn(NO3)2 • 4H2O was first dissolved in 50 ml water to form a clear solution A. At the same time 0.02 mol (NH4)2HPO4 and 20 g Na2CO3 were dissolved in 100 ml water to form a clear solution B. Then solution A was quickly added into solution B under fast magnetic agitation for about half an hour. The mixed slurry was then transferred to an autoclave and heated at 120°C for 15 hours, 25 hours and 70 hours for comparison. After cooling to room temperature, the product was washed 8 times with distilled water and then 3 times with methanol, followed by drying in a vacuum oven at 40° C overnight.

In order to study the ionic conductivity, some of the dried samples were high-energy ball milled without any additive for 10 minutes, 60 minutes, or 180 minutes. High-energy ball milling was performed using a SPEX 8000M Mill with 0.64 cm diameter steel balls. The ball-to-powder charge ratio was 30:1. All the samples with and without high-energy ball milling (Table I) were subjected to ionic conductivity measurement using electrochemical impedance spectroscopy (EIS). A special die with a plastic shell and two copper rods, as shown in Figure 1, was custom made to allow the ionic conductivity measurement. The diameter and length of each copper rod were 0.95 cm and 2.54 cm, respectively. All the Na3MnCO3PO4 samples were pressed under 130 MPa into dense pellets with a cross-section area of 0.712 cm2 and a thickness of 0.07 cm. During the ionic conductivity measurement, the entire die was clamped with a specially designed vise to ensure a good contact of the pellet with two copper rods. The impedance measurements were conducted in the frequency range of 500,000 Hz to 10 Hz with an amplitude of 10 mV using a frequency response analyzer (Parstat 4000, Princeton Applied Research).

All the powder samples processed under different conditions (Table I) were characterized using scanning electron microscopy (SEM, HITACHI S-4700) to determine the particle size and morphology. Further, all the samples were analyzed using X-ray diffraction (XRD, Bruker D2) to identify crystalline phase(s). For those samples without high-energy ball milling (A1, B1 and C1 in Table I),

Table I. Summary of sample processing conditions and the lattice parameters measured from XRD.

Sample Hydrothermal synthesis Ball milling Beta angle Unit cell

ID time (hour) time (minute) A (Ä) B (Ä) C (Ä) (degree) volume (Ä3)

A1 15 0 8.9831 6.7387 5.1659 90.125 312.71

B1 25 0 8.9821 6.7389 5.1676 90.111 312.79

C1 70 0 8.9853 6.7393 5.1665 90.132 312.85

D1 25 10 8.9641 6.7677 5.1781 90.243 314.09

E1 25 60 8.9431 6.7833 5.1609 89.676 313.08

F1 25 180 8.9479 6.7940 5.1660 89.499 314.05

Cu rod

Na3MnCO3PO4 pellet

Plastic shell

Figure 1. Schematic of the cut-away of the custom-made cylindrical die set for measurement of the ionic conductivity of various Na3MnCO3PO4 samples.

hand grinding with the aid of a mortar and pestle was applied for about 5 minutes to make loose powder for SEM and XRD analyses. Rietveld refinement using GSAS software31 was used to determine the lattice parameters, crystallite size, lattice microstrain, and phase identification from the XRD pattern.

To study the dependence of electrochemical performance of batteries on the ionic conductivity and powder processing conditions, various batteries were fabricated and tested, as listed in Table II. To fabricate these batteries, Na3MnCO3PO4 powders were mixed with 53 wt% carbon black (CB, Cabot Corporation) and high-energy ball milled for 10, 60 or 180 minutes to ensure good mixing of the active material with CB and to create different processing conditions (Table II). The mixing of 53 wt% CB with the Na3MnCO3PO4 powder was based on the previous study30 showing that high CB concentration (such as 53 wt%) is needed to obtain high specific capacity from Na3MnCO3PO4. The 53 wt% CB in the cathode is equivalent to 60 vol% CB in the cathode if one assumes the theoretical densities of CB and Na3MnCO3PO4 to be 2.26 g/cm3 and 2.958 g/cm3, respectively, and excludes the volume of voids between powder particles in the con-

version. High-energy ball milling conditions can be found in Table II and other parameters not listed in Table II are the same as those used in the preparation of various powder samples for the ionic conductivity measurement discussed above. The ball milled Na3MnCO3PO4 + CB powders were then mixed with 10 wt% of poly(vinylidene) fluoride (PVDF, Alfa Aesar) in 1 ml of 1-methyl-2-pyrrolidinone (NMP, Alfa Aesar) solvent to make the cathode slurry. After a uniform slurry of Na3MnCO3PO4 + CB + PVDF was prepared, the slurry was pasted onto a clean Al foil which was then transferred to a vacuum oven in an argon-filled glove box (H2O < 10 ppm and O2 < 1 ppm) and heated to dry at 110°C for about 10 hours. The dried cathode electrode was then punched into circular disks of 1.6 cm2 using a disk punching machine and the Na3MnCO3PO4 loading in each circular cathode was ~1.2 mg.

The anodes were prepared with a copper foam (thickness 2 mm) infiltrated with pure sodium (99.9%, Sigma-Aldrich) through melting in the argon-filled glove box. The electrolyte was composed of 1M sodium hexafluorophosphate (NaPF6, 99%, Alfa Aesar) with 2 vol% fluoroethylene carbonate (99%, Sigma-Aldrich) as the additive in propylene carbonate (99.7%, Sigma-Aldrich). Coin cells (CR 2032) were assembled with anodes, separators and cathodes in the argon-filled glove box and tested subsequently with a Neware BTS battery testing instrument. The voltage range for the galvanostatic charge/discharge test was from 2.0 V to 4.5 V. The current for charge was C/30 (where 1C = 191 mA/g), while the current for discharge was either C/30 or C/10 (see Table II).

Since the CB concentrations in the Na3MnCO3PO4 cathodes of this study were very high (53 wt%), it is necessary to find out the contribution of CB in storing charges (such as via the electrical double layer effect) in the cathode. By taking into account of the CB contribution, one can evaluate the specific capacity of Na3MnCO3PO4 more accurately. Towards this objective, cells with the cathodes made of pure CB (i.e., 100% CB) were also prepared and tested for its storage capacity. In preparing these cells, CB was high-energy ball milled for 10 minutes, and then mixed with 10 wt% PVDF and dissolved in 1 ml NMP to form the slurry. This procedure was the same as that used to make cathodes A2, B2 and C2 (Table II) except the cathode being 90 wt% CB + 10 wt% PVDF. The anodes of these CB cells were the same as those used to make the Na3MnCO3PO4 cells. The built CB cells were subjected to charge and discharge testing within a 2.0-4.5 V potential range at a current density of 8.85 mA/g CB. The CB loading per cathode was ~1.13 mg in these CB cells. Note that the current density (8.85 mA/g CB) used for the pure CB cells was close

Table II. The cathode composition, fabrication and testing conditions of various batteries.

Na3MnCO3PO4 used

Battery ID to make the cathode Cathode processing conditions Battery testing conditions

A2 A1 Ball milling 10 min of A1 with 53 wt% CB First charge at C/30 and first discharge at C/30

B2 B1 Ball milling 10 min of B1 with 53 wt% CB First charge at C/30 and first discharge at C/30

B3 B1 Ball milling 10 min of B1 with 53 wt% CB First charge at C/30 and first discharge at C/10

C2 C1 Ball milling 10 min of C1 with 53 wt% CB First charge at C/30 and first discharge at C/30

C3 C1 Ball milling 10 min of C1 with 53 wt% CB First charge at C/30 and first discharge at C/10

E2 B1 Ball milling 60 min of B1 with 53 wt% CB First charge at C/30 and first discharge at C/30

E3 B1 Ball milling 60 min of B1 with 53 wt% CB First charge at C/30 and first discharge at C/10

F2 B1 Ball milling 180 min of B1 with 53 wt% CB First charge at C/30 and first discharge at C/30

Table IV. The residuals and scale factors of the Rietveld refinement for each sample.

Figure 2. XRD patterns of the as-synthesized samples (A1, B1 and C1 in Table I), compared with the XRD pattern of the Na3MnCO3PO4 standard (PDF: 04-012-5266).

to that (6.73 mA/g CB) experienced by CB in the Na3MnCO3PO4 cells with 53 wt% CB at the C/30 rate if one assumes that all the current is carried by CB in the cathode.

Results and Discussion

Figure 2 shows the XRD patterns of the hydrothermally synthesized Na3MnCO3PO4 powders with different synthesis times (i.e., A1, B1 and C1 samples in Table I). Note that the XRD patterns of all the as-synthesized Na3MnCO3PO4 powders match that of the Na3MnCO3PO4 standard (PDF: 04-012-5266) very well. All of the as-synthesized Na3MnCO3PO4 powders, regardless of the synthesis time, contain 99.8% or more crystalline Na3MnCO3PO4. The lattice parameters of these powder samples determined via GSAS analysis are summarized in Table I, while the crystallite sizes and lattice microstrains are listed in Table III. Table IV summarizes the weighted-profile R-factor, Rwp, and x2 of the Rietveld refinement for each sample. Here X2 is defined to be (Rwp/Rexp)2 and Rexp is the expected R-factor. It can be concluded from Table IV that the calculated XRD pattern fits the experimental one very well because the values of Rwp of all samples are less than 0.1 and the values of x2 are less than 4.0.32 From Table I, we note that the lattice parameters of Na3MnCO3PO4 for the three hydrothermal synthesis conditions are nearly identical, and their unit cell volume is 312.68 A3, matching the value reported in Ref. 29 (312.78 A3) very well. The lattice microstrains within these as-synthesized powders are very small (Table III) and the difference among them is within the error range, suggesting that all three syn-

Table III. Summary of the crystallite sizes and lattice microstrains of various Na3MnCO3PO4 samples measured from XRD.

Sample ID

Crystallite size (nm)

(202) strain tensor coefficient*

A1 B1 C1 D1 E1 F1

126.6 187.1 160.6 28.8 31.6 35.1

194.50

222.70

*The lattice microstrain is anisotropic because we have used a tensor model to evaluate it. However, the (202) coefficient is chosen to be shown here because the other coefficients exhibit similar trends.

Sample ID A1

Rwp 0.0718 0.0761 0.0737 0.0805 0.0981 0.0981 X2 1.287 1.399 1.318 1.397 2.112 2.160

Scale factor 38.96 40.62 40.62 41.14 41.84 48.59

thesis conditions lead to crystals with similar and low defect concentrations. However, the XRD analysis reveals that the crystallite size of Sample A1 (15-h synthesis) is slightly smaller than those of Samples B1 and C1 (25-h and 70-h synthesis, respectively). This is not a surprise since crystallites are likely to grow during holding at the hydrothermal synthesis temperature (120°C). However, the crystallite sizes of Samples B1 and C1 should be considered the same because the difference between them is within the error range.

Figure 3 compares the XRD pattern of the as-synthesized Na3MnCO3PO4 powder (B1 in Table I) with those of Sample B1 after high-energy ball milling for different milling times (D1, E1 and F1 in Table I). It is noted that ball milling leads to a decrease in the peak intensity and broadening in the peak widths, suggesting crystallite size refinement and introduction of lattice microstrain.33-35 Moreover, the (100) peak of Na3MnCO3PO4 becomes much smaller after 10-min ball milling and disappears completely after 60- or 180-min ball milling while all other peaks are still present, suggesting the formation of significant structural defect(s) in the Na3MnCO3PO4 crystal. Detailed GSAS analysis reveals that ball milling has resulted in lattice distortion in all three cases (10-min, 60-min and 180-min ball milling), leading to a smaller a parameter, larger b and c parameters, and a slightly larger unit cell volume, changing from ~312.79 to ~314.05 A3 (Table I). Accompanied with the changes in the lattice parameters and unit cell volume, the crystallite size has been reduced from ~150 nm to ~30 nm for all the ball milled samples. Further, the lattice microstrain shows a clear trend, that is, the lattice microstrain increases dramatically with ball milling and the longer the ball milling time, the larger the lattice microstrain (Table III). These results unambiguously reveal that ball milling has introduced significant defects to the Na3MnCO3PO4 crystal.

In order to understand why ball milling results in the disappearance of the (100) peak, additional XRD pattern modeling for Na atom misplacement has been conducted. As shown in Figure 4, Na3MnCO3PO4 has an intricate structure with corner sharing of tetrahedral PO4 groups and MO6 octhedra forming double layers which accommodate Na atoms at two different interstitial sites.13,36 Na(1) sites coordinate with 7 oxygen atoms, and Na(2) sites coordinate with 6 oxygen atoms. The

Figure 3. XRD patterns of Sample B1 and its samples after high-energy ball milling for different times (i.e., D1, E1 and F1 in Table I).

Figure 4. The crystal structure of Na3MnCO3PO4 viewed along different directions: (a) [100], (b) [010], and (c) [001] direction. Color codes: MnO6 octahedra, green; PO4 tetrahedra, pink; CO3 triangular planar, dark green; Na(1) site, purple; and Na(2) site, blue (shown as K in the color code).

triangular planar CO3 groups share an oxygen edge with the MO6 octahedra.13,36 Figure 5 shows the result of XRD pattern modeling, revealing that although all other peaks stay the same, the (100) peak will disappear if Na at site 2 is interchanged with Mn, or the (100) peak intensity will decrease significantly if Na at site 1 is interchanged with Mn. Although the existing XRD data from E1 and F1 samples is not extensive enough to sort out the percentages of Mn interchange with Na(1) and Na(2) unambiguously, a general trend is identified, i.e., as the ball milling time increases, the occupancies of both 4-fold Na(1) and 2-fold Na(2) sites by Mn increase, but the Na(2) site increases more. Therefore, we conclude that the disappearance of the (100) peak after long-time ball milling (60 and 180 min) is due to the formation of the structural defect, i.e., the interchange of Mn with Na induced by high-energy ball milling. Furthermore, there is a tendency for Mn to occupy Na(2) preferentially, but there is some occupancy at the Na(1) site as well.

It is well known that long-time high-energy ball milling can result in amorphization of crystalline phases.35,37-39 Therefore, the possibility of amorphous phase formation has also been considered. Toward this end, we have used the scale factors in the Rietveld refinement as a rough measure of the relative concentrations of the crystalline fraction in various samples. Such estimation is reasonable because all the patterns are collected under exactly the same conditions. As shown in Table IV, the scale factors for all the samples are nearly the same, about 40 - 41, except Sample F1 which has a slightly larger scale factor (~48). Since a larger scale factor in the Rietveld refinement means more crystalline phase,40 we conclude that there is no evidence of amorphization. Further, since Sample F1 with 180-min ball milling cannot have more crystalline phase than other samples without ball milling or with shorter ball milling times, we should treat the scale factors of all the samples in Table IV being similar, i.e., all samples contain 100% crystalline Na3MnCO3PO4.

SEM images of various powders are shown in Figure 6. There are several interesting observations that can be derived from SEM images. First, Sample A1 hydrothermally synthesized for 15 h (Figure 6a) has extremely fine particles with sizes ranging from 10 to 20 nm. Second, high-energy ball milling of Sample A1 with 16 wt% CB (i.e., 20 vol% CB) for 10 min does not change Na3MnCO3PO4 particle sizes (Figure 6b). This is consistent with our expectation because high-energy ball milling is effective in reducing particle sizes from micrometers to sub-micrometers, but not effective in reducing particle sizes to nanometers.34,35,41-44 Third, longer hydrothermal synthesis times (25 and 70 h) lead to larger Na3MnCO3PO4 particles with sizes of ~200 nm and elongated morphology (Figure 6c and 6e). A closer examination of these particles reveals that these large particles are in

No interchange

1/ Mn interchange

Mn interchange j / with Na 2

with Na 1 A /

Two-Theta (deg)

Figure 5. XRD pattern modeling of the (100) peak of Na3MnCO3PO4 with the interchange of Mn with Na at different sites. Note that Mn interchange with Na(2) will result in the disappearance of the (100) peak.

ANL-EMC 10.0kV 11.4mm x100k SE(U1

Figure 6. SEM images of various Na3MnCO3PO4 samples: (a) A1, (b) A1 + 16 wt% CB, ball milled for 10 min, (c) B1, (d) B1, (e) C1, (f) C1, (g) D1, and (h) E1. Note that different magnifications are used and the sample IDs can be found in Table I.

fact agglomerates of smaller particles, as evidenced by small particles attaching to or growing from large agglomerates (Figures 6d and 6f). Interestingly, these agglomerates contain little porosity, i.e., small particles are glued together by subsequently grown Na3MnCO3PO4 from the solution leaving little porosity between them. As a result, sodia-tion and desodiation of these agglomerates during charge/discharge cycles would require diffusion of Na ions over a distance of ~200 nm in order to reach the center of the agglomerates. Finally, high-energy ball milling of these agglomerates for 10 min does not change the agglomerate size noticeably (Figure 6g), but 60-min high-energy ball milling does reduce the agglomerate size to ~150 nm (Figure 6h). These phenomena are, again, in good accordance with our prior high-energy ball milling investigations,34,35,41-44 showing that high-energy ball milling is only effective in reducing agglomerate sizes to sub-micrometers for many materials. Based on these SEM examination, we have summarized the average particle sizes of various Na samples in Table V. It should be emphasized that the average particle size here refers to the size requiring Na ions to diffuse through to get to the center of particles during charge/discharge cycles.

Figure 7 displays the Nyquist plot of Sample E1 in Table I (i.e., 25-hour hydrothermal synthesis and then 60-minute ball milling). The Nyquist plot shows a typical semicircle at the higher frequency region. The intercept of the arc at the higher frequency region on real axis represents the bulk resistance of Na3MnCO3PO4, and the horizontal length of the semicircle, i.e., the distance between the

Figure 7. The impedance spectrum of Sample E1. The arrow indicates the value of the total resistance determined by the intercept of the straight line and the semicircle.

two x-axis intercepts, stands for the electrode polarization.45,46 A straight line spike at the lower frequency region represents the diffusion effect in Na3MnCO3PO4, a characteristic feature expected for pure ionic conductors.47 The intercept of the straight line with the semicircle, as shown in Figure 7, is therefore employed to determine the total resistance, R, of Na3MnCO3PO4. The total conductivity, ct, of Na3MnCO3PO4 is computed from the total resistance, R, using Eq. 146

where the thickness of sample (L) is 0.07 cm and the cross-section area of sample (A) is 0.712 cm2. Note that the total electrical conductivity of any material is generally accepted as the sum of electronic conductivity and ionic conductivity.45,46 However, the electronic conductivity of Na3MnCO3PO4 is very low since 53 wt% CB (60 vol% CB) has to be added in the Na3MnCO3PO4 cathode to obtain a sufficient electronic conductivity, a finding of the previous study.30 Thus, the total conductivity measured here is taken to be the ionic conductivity of Na3MnCO3PO4. Furthermore, it is found that all Na3MnCO3PO4 samples with different synthesis and processing conditions have exhibited the similar Nyquist plot as that shown in Figure 7. Therefore, the measurement and computation procedure described for Sample E1 has been applied to all Na3MnCO3PO4 samples and their measured ionic conductivities are summarized in Table V.

It is interesting to note from Table V that Samples B1, C1 and D1 have similar ionic conductivities (~1.5 x 10-6 s/cm). This is consistent with the fact that B1 and C1 have very similar crystallite sizes (Table II), lattice microstrains (Table II), lattice parameters (Table I) and particle sizes (Figure 6). Therefore, they should have similar ionic conductivities. Sample D1 with high-energy ball milling for 10 min also has similar particle sizes as B1 and C1 (Figure 6), but its crystallite size is smaller and lattice microstrain is higher than those of B1 and C1 (Table II). The similar ionic conductivity of D1 as B1 and C1 suggests that the defects introduced in 10-min high-energy ball milling are not significant enough to change the ionic conductivity. However, if high-energy ball milling time is increased to 60 or 180 min, the ionic conductivity is reduced by about two orders of magnitude to ~0.5 x 10-8 S/cm(E1 andF1 in Table V). As discussed above, long milling times have introduced significant structural defects to the Na3MnCO3PO4 crystal, as evidenced by large lattice microstrains (Table II). Furthermore, long milling time has resulted in the interchange of Mn with Na and disappearance of the (100) peak (Figure 5). Clearly, this mechanically forced misplacement of Na ions has disrupted Na diffusion inside the Na3MnCO3PO4 crystal, leading to very low ionic conductivity. Finally, it is noted that Sample A1 has the

Table V. Summary of measured ionic conductivities and sodium-ion diffusion coefficients.

Sample ID Resistance №) Ionic conductivity (S.cm-1) Sodium-ion diffusion coefficient (cm2.S-1) Average particle size (nm)

A1 17,550 5.60x10-6 7.11 x 10-12 20

B1 70,000 1.40x10-6 1.78x10-12 -200

C1 64,000 1.54x10-6 1.96x10-12 -200

D1 63,000 1.56x10-6 1.98x10-12 -200

E1 200,000 0.492 x10-8 0.625 x10-14 -150

F1 132,000 0.745 x10-8 0.946x 10-14 -150

Time for Na-ion diffusing through an average particle size (s)

0.56 224.72 204.08 202.02 3.6 x 104 2.4 x 10'

highest ionic conductivity among all the samples measured (Table V). In comparison with B1 and C1, A1 has much smaller particle sizes. Other than this, A1, B1 and C1 have similar crystallite sizes, lattice microstrains and lattice parameters (Tables I and III). Thus, we hypothesize that the high ionic conductivity of A1 is likely due to faster diffusion of Na ions on the particle surface and the large surface area provided by nanoparticles of A1.

Based on the Nernst-Einstein equation,48 Eq. 2, the sodium-ion diffusion coefficient (Dff) can be estimated from the ionic conductivity, if we assume that (i) the current is carried by Na ions and (ii) Na3MnCO3PO4 crystals are solids where Na ions are not interacting each other, not even site blocking effects.

Da = -Z-2 a [2]

n( Ze)

where kB, n, Z and e are Boltzmann constant (1.38 x 10-23 J/K ), the number of carriers in unite volume, the valence of charge carrier, and elementary charge (1.60 x 10-19 C), respectively. From the a lattice parameter of 8.98 A for Na3MnCO3PO4 at room temperature, the number of carriers can be determined to be n = 0.0231 mol cm-3. The calculated Na-ion diffusion coefficients for all the samples are included in Table V for comparison.

Furthermore, the time for sodium-ion to diffuse over an average particle can be estimated with the aid of Eq. 349

where D is the average particle size determined from SEM image analysis and listed in Table V. The estimated diffusion time for each Na3MnCO3PO4 sample is listed in Table V. The different ionic conductivities, Na-ion diffusion rates, and times for Na ions to diffuse

Figure 8. The charge/discharge curves of a cell with the cathode made of 100% CB. The specific capacity increases slightly from the first cycle to the 10th cycle. However, no change is observed beyond the 10th cycle, as demonstrated by the complete overlap of the 10th and 20th discharge curves. See the text for detailed discussion.

through an average particle are expected to have impact on electrochemical performance of various Na3MnCO3PO4-based half cells, as discussed below.

The 1st, 5th, 10th and 20th charge/discharge curves of the pure CB cells are displayed in Figure 8. The specific capacity is 35.8 mAh/g CB at the first discharge. We attribute this charge storage to surface absorption of Na ions via the electrical double layer (EDL) mechanism. The evidence of supporting this argument is that dQ/dV of the discharge curve is nearly constant (where Q is the quantity of charge and V is the voltage), indicating that this is an electrochemical capacitor behavior.50 Note that the specific capacity increases slightly from the first discharge to the 10th discharge, most likely due to the improved wetting of the CB particles by the electrolyte. However, beyond the 10th discharge the specific capacity exhibits no changes at all, again displaying the electrochemical capacitor behavior since it is well known that electrochemical capacitors can charge/discharge for several-thousand times with little or no capacity fading.50,51 Note that the first charge exhibits a very large irreversible specific capacity, which is attributed to the formation of the solid electrolyte interphase (SEI) layer on the Na anode. However, the slight non-linear curvature of the subsequent charge curves at the high voltage portion (>4.0 V) is likely due to the gradual decomposition of the electrolyte.

To obtain the accurate estimation of the specific capacity of Na3MnCO3PO4, we have included the EDL effect of CB in all calculations. This is done by subtracting the total capacity of the entire cathode by the capacity derived from the given amount of CB in the cathode (i.e., 53 wt% CB in all of the Na3MnCO3PO4 cells). The resulting capacity after the subtraction is then divided by the weight of Na3MnCO3PO4 in the cathode to obtain the specific capacity of Na3MnCO3PO4.

Figure 9 compares the first charge-discharge curves of various Na3MnCO3PO4 cells tested at the C/30 rate. In this figure the specific capacity is referring to that of Na3MnCO3PO4 because the

Figure 9. Comparisons in the first charge-discharge curves of various Na3MnCO3PO4 cells (after the modification with subtraction of the EDL effect of 53 wt% CB). The first operation for these cells is charging, and all cells are tested at the C/30 rate.

Figure 10. A comparison in the specific capacities of various Na3MnCO3PO4 cells tested at the C/30 rate (after the modification with subtraction of the EDL effect of 53 wt% CB).

CB contribution to the capacity has been subtracted from the capacity of the entire cathode, as mentioned previously. It is noted that the first charge of all samples exhibits a very large irreversible specific capacity, similar to what have been observed in the pure CB cells. Again, we have attributed this large irreversible specific capacity at the first charge to the formation of the SEI layer on the Na anode. To facilitate the comparison, the discharge specific capacities of various Na3MnCO3PO4 cells measured from Figure 9 have been re-plotted in a bar chart as shown in Figure 10.

In analyzing the data of Figure 10, it should be remembered that the different discharge capacities of different cells are not due to the effect of electronic conductivity because all the cathodes in these cells contain 53 wt% CB (60 vol% CB) which is sufficient to provide a

continuous CB network to interact with nearly all of Na3MnCO3PO4 particles in the cathode, as demonstrated in a previous study.30 Therefore, the different discharge capacities of Figure 10 can be mainly attributed to the effects of ionic conductivity and particle size among others. From Table V, one can divide various Na3MnCO3PO4 powders into three groups: Sample A1 as the first group with the highest ionic conductivity (5.6 x 10-6 S/cm) and the shortest time for Na ions to diffuse through an average particle (0.56 seconds), Samples E1 and F1 as the second group with the lowest ionic conductivity (~0.5 x 10-8 S/cm) and the longest time for Na ions to diffuse through an average particle (~7 to 10 hours), and Samples B1, C1 and D1 as the third group with the intermediate properties between the first and second groups. Based on these fundamental properties, one would expect the

Figure 11. A comparison in the specific capacities of Na3MnCO3PO4 cells tested at the C/30 and C/10 rates (see Table II for the battery ID).

cell made of the A1 material to have the highest discharge specific capacity and the cells made of the E1 and F1 materials to have the lowest discharge specific capacities. As shown in Figure 10, this is indeed the case because the A2 cell has the highest discharge specific capacity, whereas E2 and F2 cells have the lowest discharge specific capacities. B2 and C2 cells exhibit intermediate specific capacities because their Na3MnCO3PO4 materials have the intermediate ionic conductivities and times for Na ions to diffuse through an average particle (Table V). These results clearly indicate that that ionic conductivity and the time for Na-ion diffusion through an average particle are important factors in controlling the specific capacity of Na3MnCO3PO4-based cathodes. Further, the higher the ionic conductivity, the larger the specific capacity.

Generally speaking, the change in the ionic conductivity is also expected to affect the rate performance. Thus, several samples have been tested at a higher charge/discharge rate (C/10) and their specific capacities are compared with those obtained tested at the C/30 rate. As shown in Figure 11, the trend of a higher ionic conductivity leading to a larger specific capacity obtained at the C/30 rate still holds at the C/10 rate. Further, as expected, the specific capacity of all cathodes decreases as the rate increases from C/30 to C/10. However, it is noted that Sample B2 has a relatively larger reduction in the specific capacity (12% decrease) than Samples C2 (3% decrease) and E2 (2.9% decrease) when the rate increase from C/30 to C/10. If the specific capacity were only controlled by ionic conductivity, one would have expected that Sample E2 (rather than B2) be more sensitive to the charge/discharge rate (i.e., exhibiting a larger decrease in the specific capacity) since it has a lower ionic conductivity than the other two samples. The discrepancy between the expectation and the result suggests that ionic conductivity is not the only factor affecting the specific capacity. In fact, if one only considers ionic conductivity and the time for Na ions to diffuse through an average particle, then one would also expect B2 and C2 cells to have similar specific capacities because their Na3MnCO3PO4 materials have similar ionic conductivities and times for Na ions to diffuse through an average particle (Table V). Since B2 and C2 cells display significant difference in the discharge specific capacity, we conclude that other factors than electronic conductivity, ionic conductivity, and the time for Na ions to diffuse through an average particle are playing a role in affecting the specific capacity. These factors remain to be investigated in the near future.

Conclusions

The present study unambiguously reveals that hydrothermal synthesis conditions will affect the characteristics of the Na3MnCO3PO4 powder product. Short holding time (15 h) at the hydrothermal synthesis temperature results in nanoparticles (20 nm in diameter), whereas long holding times (25 h an 70 h) lead to significantly agglomerated particles (~200 nm). In spite of their different particle sizes, all of these particles have similar crystallite sizes, lattice microstrains and lattice parameters. However, nanoparticles exhibit a higher ionic conductivity than agglomerated particles, likely due to faster diffusion of Na ions on the particle surface and the large surface area provided by nanoparticles. Structural defects of the Na3MnCO3PO4 crystal also have significant influence on ionic conductivity. The interchange of Mn atoms with Na atoms induced by high-energy ball milling reduces ionic conductivity by two orders of magnitude. When the electronic conductivity of the Na3MnCO3PO4-based cathode is high (achieved by high CB concentrations), the half-cells made of Na3MnCO3PO4 with higher ionic conductivities display higher discharge specific capacities than the half-cells made of Na3MnCO3PO4 with lower ionic conductivities. Thus, ionic conductivity, like electronic conductivity, is a critical parameter in controlling the specific capacity of Na3MnCO3PO4. Ionic conductivity is, in turn, affected by powder synthesis and processing conditions that could alter particle sizes and structural defects.

Acknowledgments

The authors are grateful to the assistance in electrochemical testing offered by Jack Shamie.

References

1. J.-M. Tarascon and M. Armand, "Issues and challenges facing rechargeable lithium batteries," Nature, 414, 359 (2001).

2. V. Palomares, P. Serras, I. Villaluenga, K. B. Hueso, J. Carretero-Gonzales, and T. Rojo, "Na-ion batteries, recent advances and present challenges to become low cost energy storage systems," Energy Environ. Sci., 5, 5884 (2012).

3. J. Liu, "Addressing the grand challenges in energy storage," Adv. Funct. Mater., 23, 924 (2013).

4. L. Chen and L. Shaw, "Recent advances in lithium-sulfur batteries," J. Power Sources, 267, 770 (2014).

5. M. M. Doeff, A. Ferry, Y. Ma, L. Ding, and L. C. DeJonghe, "Effect of electrolyte composition on the performance of sodium/polymer cells," J. Electrochem. Soc., 144(2), L20 (1997).

6. S. Komaba, T. Ishikawa, N. Yabuuchi, W. Murata, A. Ito, and Y. Ohsawa, "Fluorinated ethylene carbonate as electrolyte additive for rechargeable Na batteries," ACS Applied Materials & Interfaces, 3, 4165 (2011).

7. M. M. Doeff, Y. Ma, S. J. Visco, and L. C. DeJonghe, "Electrochemical Insertion of Sodium into Carbon," J. Electrochem. Soc., 140, L169 (1993).

8. H. Xiong, M. D. Slater, M. Balasubramanian, C. S. Johnson, and T. Rajh, "Amorphous TiO2 Nanotube Anode for Rechargeable Sodium Ion Batteries," J. Phys. Chem. Lett., 2, 2560 (2011).

9. D. Kim, E. Lee, M. Slater, W. Lu, S. Rood, and C. S. Johnson, "Layered Na[Ni1/3Fe1/3Mn1/3]O2 cathodes for Na-ion battery application," Electrochem. Commun., 18, 66 (2012).

10. M. D. Slater, D. Kim, E. Lee, and C. S. Johnson, "Sodium-ion batteries," Adv. Funct. Mater., 23, 947 (2013).

11. L. Wang, Y. Lu, J. Liu, M. Xu, J. Cheng, D. Zhang, and J. B. Goodenough, "A superior low-cost cathode for a Na-ion battery," Angew. Chem. Int. Ed. Engl., 52, 1964 (2013).

12. S. P. Ong, V. L. Chevier, G. F. Hautier, A. Jain, C. Moore, S. Kim, X. Ma, and G. Ceder, "Voltage, stability and diffusion barrier differences between sodium-ion and lithium-ion intercalation materials," Energy Environ. Sci., 4, 3680 (2011).

13. H. Chen, Q. Hao, O. Zivkovic, G. Hautier, L.-S. Du, Y. Tang, Y.-Y. Hu, X. Ma, C. P. Grey, and G. Ceder, "Sidorenkite (Na3MnPO4CO3): a new intercalation cathode material for Na-ion batteries," Chem. Mater., 25, 2777 (2013).

14. C. D. Wessells, S. V. Peddada, R. A. Huggins, and Y. Cui, "Nickel hexacyanoferrate nanoparticle electrodes for aqueous sodium and potassium ion batteries," Nano Lett, 11,5421 (2011).

15. S. Komaba, W. Murata, T. Ishikawa, N. Yabuuchi, T. Ozeki, T. Nakayama, A. Ogata, K. Gotoh, and K. Fujiwara, "Electrochemical Na insertion and solid electrolyte interphase for hard-carbon electrodes and application to Na-ion batteries," Adv. Funct. Mater., 21,3859 (2011).

16. XinXiaand J. R. Dahn,"NaCrO2 is a Fundamentally Safe Positive Electrode Material for Sodium-Ion Batteries with Liquid Electrolytes," Electrochemical and Solid-State Letters, 15, A1 (2012).

17. H. Yoshida, N. Yabuuchi, and S. Komaba, "NaFe0.5Co0.5O2 as high energy and power positive electrode for Na-ion batteries," Electrochem. Commun., 34, 60 (2013).

18. J. Zhao, L. Zhao, K. Chihara, S. Okada, J.-i. Yamaki, S. Matsumoto, S. Kuze, and K. Nakane, "Electrochemical and thermal properties of hard carbon-type anodes for Na-ion batteries," J. Power Sources, 244, 752 (2013).

19. P. Senguttuvan, G. Rousse, V. Seznec, J.-M. Tarascon, andM.R. Palacln, "Na2Ti3O7: Lowest Voltage Ever Reported Oxide Insertion Electrode for Sodium Ion Batteries," Chem. Mater., 23,4109 (2011).

20. S. Komaba, C. Takei, T. Nakayama, A. Ogata, and N. Yabuuchi, "Electrochemical intercalation activity of layered NaCrO2 vs. LiCrO2," Electrochem. Commun., 12, 355 (2010).

21. Y. Cao, L. Xiao, M. L. Sushko, W. Wang, B. Schwenzer, J. Xiao, Z. Nie, L. V. Saraf, Z. Yang, and J. Liu, "Sodium ion insertion in hollow carbon nanowires for battery applications," Nano lett., 12, 3783 (2012).

22. D. H. Lee, J. Xu, and Y. S. Meng, "Advanced cathode materials for Na-ion batteries with high capacity and excellent structural stability," Phys. Chem. Chem. Phys., 15, 3304 (2013).

23. A. Langrock, Y. Xu, Y. Liu, S. Ehrman, A. Manivannan, and C. Wang, "Carbon coated hollow Na2FePO4F spheres for Na-ion battery cathodes," J. Power Sources, 233, 62 (2013).

24. S. Kim, D. Seo, X. Ma, G. Ceder, and K. Kang, "Electrode materials for rechargeable sodium-ion batteries: potential alternatives to current lithium-ion batteries," Adv. Energy Mater., 2, 710 (2012).

25. M. Nose, H. Nakayama, K. Nobuhara, H. Yamaguchi, S. Nakanishi, and H. Iba, "Na4Co3(PO4)2P2O7: A novel storage material for sodium-ion batteries," J. Power Sources, 234,175 (2013).

26. Y.-U. Park, D.-H. Seo, H.-S. Kwon, B. Kim, J. Kim, H. Kim, I. Kim, H.-I. Yoo, and K. Kang, "A new high-energy cathode for a Na-ion battery with ultrahigh stability," J. Am. Chem. Soc., 135, 13870 (2013).

27. H. Yoshida, N. Yabuuchi, and S. Komaba, "NaFe0.5Co0.5O2 as high energy and power positive electrode for Na-ion batteries," Electrochem. Commun., 34, 60 (2013).

28. G. Hautier, A. Jain, H. Chen, C. Moore, S. P. Ong, and G. Ceder, "Novel mixed polyanions lithium-ion battery cathode materials predicted by high-throughput ab initio computations," J. Mater. Chem., 21, 17147 (2011).

29. H. Chen, G. Hautier, A. Jain, C. Moore, B. Kang, R. Doe, L. Wu, Y. Zhu, Y. Tang, and G. Ceder, "Carbonophosphates: a new family of cathode materials for Li-ion batteries identified computationally," Chem. Mater, 24, 2009 (2012).

30. C. Wang, M. Sawicki, S. Emani, C. Liu, and L. Shaw, "Na3MnCO3PO4 - a high capacity, multi-electron transfer redox cathode material for sodium ion batteries," Electrochimica Acta, 161, 322 (2015).

31. A. C. Larson and R. B. Von Dreele, "General structure analysis system (GSAS)", Los Alamos National Laboratory Report LAUR 86-784 (2004).

32. L. B. McCusker, R. B. Von Dreele, D. E. Cox, D. Louer, and P. Scardi, "Rietveld refinement guidelines," J. Appl. Cryst., 32, 36 (1999).

33. A. Ortiz, W. Osborn, T. Markmaitree, and L. Shaw, "Crystallite sizes of LiH before and after ball milling and thermal exposure," J. Alloys Compd., 454, 297 (2008).

34. R.-M. Ren, Z.-G. Yang, and L. Shaw, "Synthesis of nanostructured silicon carbide through integrated mechanical and thermal activation process," J. Am. Ceram. Soc., 85,819 (2002).

35. R.-M. Ren, Z.-G. Yang, and L. Shaw, "Polymorphic transformation and powder characteristics of TiO2 during high energy milling," J. Mater. Sci., 35, 6015 (2000).

36. H. Chen, G. Hautier, and G. Ceder, "Synthesis, computed stability, and crystal structure of a new family of inorganic compounds: carbonophosphates," J. Am. Chem. Soc., 134, 19619(2012).

37. A. W. Weeber and H. Bakker, "Amorphization by Ball Milling: A Review," Physica B, 153,93 (1988).

38. U. Mizutani and C. H. Lee, "Effect of Mechanical Alloying Beyond the Completion of Glass Formation for Ni-Zr Alloy Powders," J. Mater. Sci., 25, 399 (1990).

39. C. C. Koch, "Amorphization Reactions during Mechanical Alloying/Milling of Metallic Powders," React. Solids, 8, 283 (1989).

40. H. M. Rietveld, "A profile refinement method for nuclear and magnetic structures," J. Appl. Cryst., 2, 65 (1969).

41. K. Crosby and L. Shaw, "Dehydriding and rehydriding properties of high-energy ball milled LiBH4 + MgH mixtures," Int. J. Hydrogen Energy, 35, 7519 (2010).

42. K. Crosby, L. Shaw, C. Estournes, G. Chevallier, A. W. Fliflet, and M. A. Imam, "Enhancement in Ti-6Al-4V Sintering via Nanostructured Powder and Spark Plasma Sintering," PowderMetall., 57(2), 147 (2014).

43. L. Shaw, R. Ren, T. Markmaitree, and W. Osborn, "Effects of Mechanical Activation on Dehydrogeneation of the Lithium Amide and Hydride System," J. Alloys Compd., 448, 263 (2008).

44. L. Shaw, M. Zawrah, J. Villegas, H. Luo, and D. Miracle, "Effects of Process Control Agents on Mechanical Alloying of Nanostructured Aluminum Alloys,"Metall. Mater. Trans., 34A(1), 159 (2003).

45. S. Wang, T. Kobayashi, M. Dokiya, and T. Hashimoto, "Electrical and ionic conductivity of Gd-doped ceria," J. Electrochem. Soc., 147, 3606 (2000).

46. S. W. Anwane, "Solid electrolytes: principles and applications," in Advanced Energy Materials, edited by A. Tiwari and S. Valyukh, Scrivener Publishing LLC, 2014, pp. 259.

47. G. Sahu, Z. Lin, J. Li, Z. Liu, N. Dudney, and C. Liang, "Air-stable, high-conduction solid electrolytes of arsenic-substituted Li4SnS4," Energy Environ. Sci., 7, 1053 (2014).

48. Y.-J. Bai, C. Gong, Y.-X. Qi, N. Lun, and J. Feng, "Excellent long-term cycling stability of La-doped Li-^TisOn anode material at high current rates," J. Mater. Chem., 22,19054 (2012).

49. Y.-J. Bai, C. Gong, N. Lun, and Y.-X. Qi, "Yttrium-modified Li4Ti5O12 as an effective anode material for lithium ion batteries with outstanding long-term cyclability and rate capabilities," J. Mater. Chem. A, 1, 89 (2013).

50. B. E. Conway, Electrochemical supercapacitors: scientific fundamentals and technological applications, Kluwer Academic, New York, NY, 1999.

51. C. Liu, Z. Yu, D. Neff, A. Zhamu, and B. Z. Jang, "Graphene-based supercapacitor with an ultrahigh energy density," Nano Lett., 10, 4863 (2010).