Accommodating High Transformation Strains in Battery Electrodes via

Feb 21, 2017 - NaxFePO4 (0 < x < 1, NFP) olivine, of interest as a cathode for sodium-ion batteries, is a model for ... Energy Storage Materials 2018 ...
0 downloads 0 Views 3MB Size
Download PDF
Letter pubs.acs.org/NanoLett

Accommodating High Transformation Strains in Battery Electrodes via the Formation of Nanoscale Intermediate Phases: Operando Investigation of Olivine NaFePO4 Kai Xiang,† Wenting Xing,† Dorthe B. Ravnsbæk,†,‡ Liang Hong,∥ Ming Tang,∥ Zheng Li,† Kamila M. Wiaderek,⊥ Olaf J. Borkiewicz,⊥ Karena W. Chapman,⊥ Peter J. Chupas,⊥ and Yet-Ming Chiang*,† †

Department of Material Science and Engineering, Massachusetts Institute of Technology, 77 Massachusetts Avenue, Cambridge, Massachusetts 02139, United States ‡ Department of Physics, Chemistry and Pharmaci, University of Southern Denmark, Campusvej 55, 5320 Odense M, Denmark ∥ Department of Material Science and Nanoengineering, Rice University, 6100 Main Street, Houston, Texas 77005, United States ⊥ X-ray Science Division, Advanced Photon Source, Argonne National Laboratory, 9700 South Cass Avenue, Argonne, Illinois 60439, United States S Supporting Information *

ABSTRACT: Virtually all intercalation compounds exhibit significant changes in unit cell volume as the working ion concentration varies. NaxFePO4 (0 < x < 1, NFP) olivine, of interest as a cathode for sodium-ion batteries, is a model for topotactic, high-strain systems as it exhibits one of the largest discontinuous volume changes (∼17% by volume) during its first-order transition between two otherwise isostructural phases. Using synchrotron radiation powder X-ray diffraction (PXD) and pair distribution function (PDF) analysis, we discover a new strain-accommodation mechanism wherein a third, amorphous phase forms to buffer the large lattice mismatch between primary phases. The amorphous phase has short-range order over ∼1nm domains that is characterized by a and b parameters matching one crystalline end-member phase and a c parameter matching the other, but is not detectable by powder diffraction alone. We suggest that this strain-accommodation mechanism may generally apply to systems with large transformation strains. KEYWORDS: Phase transformations, operando, batteries, olivines, sodium iron phosphate, pair-distribution function

I

bers,5−8 the reality has proven to be considerably more complex. Various TEM studies confirmed the existence of phase boundaries in micron-scale particles.9−13 In situ TEM studies have shown the phase boundary to be a solid solution zone12 or semicoherent.13 With reduction of particle size into the nanoscale regime, the misfit strain is reduced, corresponding to an increase in the mutual solid solubility between LFP and FP,14−17 even reaching complete solid solution for certain materials.18 A theoretical study also suggested the possibility of bypassing the phase transition via formation of a metastable solid solution above a critical overpotential.19 Utilizing operando characterization techniques, recent studies show that the transformation mechanism is current-rate-dependent; at low rates, the transformation follows a classical first-order phase-transition pathway, whereas at higher rates, metastable phases (which may subsequently relax) are obtained.12,20−22

ntercalation compounds such as those used for ion storage in advanced batteries commonly undergo phase transformations as the working ion concentration varies. Under scenarios in which phase transformation takes place inside a single particle or crystallite, a volume mismatch is generally created. The phase transition induced strain, if sufficiently large, can significantly impact the reversibility of ion insertion and extraction inside the electrode material. Electrochemical− mechanical coupling also commonly induces fracture or “electrochemical shock”;1,2 there are also beneficial uses, such as the harnessing of intercalation strain in electrochemical actuators.3,4 For fundamental study, the alkali metal phosphates AMPO4 (A, alkali metal; M, first-row transition metal) of olivine structure are ideal models due to the ability to systematically tune the transformation strain and, therefore, the phase-transformation pathway. Within this family of compounds, LiFePO4 (LFP) has been most widely studied. Although earlier work suggested that the phase separation process in LFP follows a classical first-order phase transition pathway between fully lithiated and fully delithiated endmem© XXXX American Chemical Society

Received: November 29, 2016 Revised: February 12, 2017 Published: February 21, 2017 A

DOI: 10.1021/acs.nanolett.6b04971 Nano Lett. XXXX, XXX, XXX−XXX

Letter

Nano Letters

electrochemical lithiation of FePO4 (FP) at particle size of micrometer scale.13 Thus, at higher transformation strains, plasticity seems assured. Also included in Figure 1 is silicon, representing materials of extraordinarily high transformation strain (up to 300 vol %), and in which lithiation was shown (originally by Limthongkul et al.26) to cause amorphization. Subsequent cycling within the amorphous regime is possible and even desirable.27 Plastic deformation accommodated by amorphization is not an isotropic response when the starting material is crystalline, however, and the anisotropy and uniformity of electrochemical expansion can be controlled to mitigate mechanical failure.28−30 In this context, accommodation of the 17% volume transformation strain of olivine NaFePO4 (NFP) requires at least plasticity, if not amorphization. A pair of research groups have previously utilized X-ray diffraction to study structure changes during electrochemical cycling of olivine NFP. CasasCabanas et al. studied NFP with primary particle size of ∼800 nm and observed two and three crystalline phases during charge and discharge, respectively. Gaubicher et al. found a notable variation in the Na solubility of crystalline phases even within two-phase coexistence.31,32 However, the microscopic behavior was not elucidated in these studies. In the present study, we use operando synchrotron radiation powder X-ray diffraction (SR-PXD), ex situ pair distribution function (PDF) analysis, and transmission electron microscopy (TEM) to characterize the phase transition behavior of nanocrystalline NFP. Because NFP cannot presently be synthesized directly in the olivine phase (maricite being the stable phase), we prepared olivine NFP starting from an LFP powder of ∼50 nm particle size (extensively characterized in previous work,23,24) and chemically delithiated the powder to FP (with less than 1% residual Li, as characterized by ICP analysis). Electrochemical sodiation of the FP phase in half-cells using sodium metal negative electrodes then produced olivine NFP. A comparison of the phase evolution path for NFP with that for LMFP and LFP is shown in Figure 2. A bird’s eye view shows the evolution in the (200) reflections for all three olivines during charge and discharge, collected using the AMPIX cell,33,34 specifically designed for operando SR-PXD. It is clearly seen that the evolution of composition, as indicated by the 2θ position, and phase quantity, as indicated by relative intensity of the (200) reflections, follows a different path for each olivine. LiFePO4 shows two distinct peaks for LFP and FP, the positions of which remain fixed during charge and discharge while the intensities vary, indicating classical two-phase coexistence with very limited solid solution of either phase. LMFP (y = 0.4) exhibits three distinct crystalline phases during charge, LMFP, MFP, and an intermediate solid-solution phase LxMFP. The peak positions of LMFP and LxMFP both change with state-of-charge, showing solid-solution composition change while in two-phase coexistence. The behavior is asymmetric; during discharge, a continuous transformation (solid solution) is observed between LxMFP and LMFP limiting compositions. This behavior has been discussed in detail elsewhere.23,24 The behavior of NFP is different from both LFP and LMFP. While there are two crystalline phases, according to a published phase diagram,35 the sodiated phase exists over a wide range of solid solution, from the fully sodiated composition, here denoted NFP, to a solid solution limit NxFP for the Na-rich phase, which has a limiting Na concentration of x ≈ 0.66. The bird’s eye view of the NFP transition in Figure 2 is qualitatively consistent with the

In contrast to LFP, the mixed olivine LiFeyMn1−yPO4 (LMFP) has transformation strains that are highly tunable with composition, ranging from 1 to 3% for y < 0.5 to ∼10% for y = 0 (pure LMP).23−25 Operando X-ray diffraction studies of LMFP have shown, even under low rate cycling, a continuous change in the compositions of coexisting lithiated and delithiated phases, demonstrating a nonequilibrium, extended solid solution behavior that reduces the volume strain during the phase transition. Indeed, the minimum in transformation strain occurring at Mn concentration y = 0.2−0.4 has been shown to coincide with the maximum in rate capability, e.g., for nano-LMFP with 50C rate capability.23,24 Its high power as well as higher cell voltage has generated commercial interest in LMFP as a second-generation olivine cathode after LFP. At higher Mn concentrations, however (y > 0.5), LMFP exhibits sluggish first-order phase transition behavior similar to that of coarse-grained LFP.23,24 (Hereafter in this manuscript, “LMFP” will refer to the high rate, low-Mn compositions (y < 0.5) unless otherwise specified.) Intercalation compounds with still higher transformation strains exist and for which the strain accommodation mechanism is unclear. Here, we focus on the compound with the largest known transformation strain (∼17 vol %) in the olivine series, NaFePO4 (NFP). Figure 1 compares the volume

Figure 1. Systems studied, corresponding transformation strains, and phase-transformation behaviors observed. Phospho-olivines can be tuned over a wide range of transformation strain, corresponding to which are several possible strain-accommodation mechanisms. The focus of this work, NaFePO4, is unique among intercalation compounds in having ∼17% unit cell volume change during the first-order phase transition between nominally isostructural sodiated and desodiated endmembers. Silicon is included as an example of a system with very large volume change between structurally dissimilar phases. (The transformation strain is the percentage difference between the unit cell volume of oxidized phase and that of reduced phase during phase transition.).

transformation strains of LMFP (y < 0.5), LFP, and NFP, with their known or suspected materials responses during electrochemical cycling. In each instance, the response referred to what occurs in micro- to nanoscopic particle systems; at a larger crystal-size scale, even modest transformation strains are typically accommodated by brittle fracture. For LMFP, the volume transformation strain ranges from 1% to 3%, leading to coherent phase-transition behavior.23,24 Upon reaching transformations strains around 5%, as for LFP, plasticity begins to play a role. Dislocations have been directly observed during B

DOI: 10.1021/acs.nanolett.6b04971 Nano Lett. XXXX, XXX, XXX−XXX

Letter

Nano Letters

Figure 2. Bird’s eye view of the evolution in the (200) reflections for LiMnyFe1−yPO4 (y = 0.4, LMFP), LiFePO4 (LFP), and NaFePO4 (NFP) during charge and discharge. Shown is the first electrochemical cycle for both LMFP and LFP, starting with a fully lithiated positive electrode. For NFP, a starting LFP olivine was chemically delithiated to FP and then electrochemically sodiated. The results shown correspond to the first desodiation (charge) and the second electrochemical sodiation (discharge). LFP (center), an exemplar of first-order transition, shows distinct peaks with fixed positions for both the LFP phase and the FP phase. LMFP (left) has three phases, and the peak positions for LMFP and LxMFP change with overall state of charge of the sample. NFP (right) follows a binary phase transition although the position for the sodium-rich phase, NxFP, changes with state-of-charge due to its broad solid solution range.

Figure 3. Voltage profile of sodium iron phosphate cell, refined unit cell volume of observed crystalline phases, and calculated mole percentage of crystalline phases against initial amount of crystalline FP phase. (A) Voltage of NaxFePO4 vs Na+/Na cells vs scan number during the operando SR-PXD experiment while electrochemically cycling at constant current rate of C/20. (B) Unit cell volume data for sodium-poor phase (FP) and sodium-rich phase (NFP or NxFP) obtained via Rietveld refinement of the diffraction data. See the Supporting Information for a full bird’s eye view of diffraction data. (C) Mole percentage of crystalline FP, crystalline NFP, and the sum of crystalline phases, normalized to the initial FP content, which is assumed to be 100% crystalline. Arrows indicate composition values xNa = 0.3 and xNa = 0.6, at which ex situ PXD and PDF analyses were carried out.

reported phase diagram, showing continuous variation of the NFP peak position over a wide SOC range, consistent with solid solution behavior, whereas the FP peak position remains fixed. However, closer scrutiny of the data revealed that the combined intensity of (200) reflections for these two phases increases significantly with state of charge and decreases to a comparable extent with state of discharge. This interesting behavior suggests that an additional phase or phases not detectable by PXD appeared and disappeared during the discharge−charge cycles, prompting us to investigate further. To quantify the phase behavior, Rietveld refinement was carried out for the operando SR-PXD data using a previously described protocol.23,24 Figure 3 displays (A) the voltage of the AMPIX cell during electrochemical cycling, (B) the unit cell volume of the crystalline phases, and (C) the molar percentage of each of the crystalline phases, as well as their sum, normalized to the starting FP material as a 100% crystalline material (referring to experimental section for details). The horizontal axis is the Na content; in total 212 individual PXD scans were made to collect the data in Figure 3, with the sodium content varying by ∼1.25% over the duration of each scan. A full bird’s eye view of the diffraction intensities during this experiment is shown in Figure S1. The unit cell volumes from literature36,37 for NFP and N0.66FP are shown as horizontal lines in Figure 3B. During the first discharge (first Na insertion, represented by the left column in Figure 3), crystalline NFP is formed with a similar unit cell volume to the literature value for NaFePO4. The difference in unit cell volume between this phase, and the

FP phase is ∼17 vol %, as expected. Concurrent with the first discharge, however, there is 20% decline in the total crystalline phase percentage compared to the starting FP. During the second charge (center column in Figure 3, the first charge being the original chemical delithiation of the starting LFP to FP), a two-step voltage profile emerges (Figure 3A, center), similar to previously reported voltage curves.36,37 The initial part of the voltage curve corresponds to the desodiation of NFP to its solid solution limit, NxFP of x = 0.66. The unit cell volume drops accordingly (Figure 3B). Next, a two-crystalline-phase field (the shaded region) dominated by the coexistence of NxFP and FP is observed. In this regime, the phase-transformation strain (difference in unit cell volume) between the crystalline phases is reduced from ∼17% for NFP−FP to ∼13% for NxFP−FP. Several features show that the sample is not at equilibrium in this regime. First, the NxFP unit cell volume continues to decreases as the sample is charged, while the FP unit cell volume increases slightly instead of each remaining constant, as would be expected for two-phase equilibrium. Second, the total crystalline percentage (Figure 3C) continues to drop in this regime, reaching a minimum value toward the middle of this two-crystalline-phase field. As the cell reaches its fully charged C

DOI: 10.1021/acs.nanolett.6b04971 Nano Lett. XXXX, XXX, XXX−XXX

Letter

Nano Letters

Figure 4. Pair-distribution function (PDF) analysis of xNa = 0.3 and xNa = 0.6 samples after first sodiation of a starting FP powder. (A) A two-phase structure model (red lines) optimized for the simultaneously collected powder diffraction data analyzed using Rietveld methods did not yield a satisfactory fit to the corresponding PDF data (blue lines). Note the large residual (black lines) at a small pair distance, r. (B) A three-phase structure model provided a good fit to the data, with the lattice dimensions of the FP and NFP phases and their relative abundance matching the diffraction data obtained simultaneously for the same samples. The third phase has short-range periodicity with a and b lattice parameters (a = 9.84 Å; b = 5.78 Å) that closely match the prototypical FP phase, while the c lattice dimension (c = 4.93 Å) that closely matches that in the NFP phase.

The loss of crystalline phase fraction, as determined from the refinement of the operando PXD data, begins during the first discharge and continues during the second charge. Formation of a disordered phase is suggested, but the nature of this phase is not clear from the SR-PXD results alone. Prior studies on electrochemically cycled LFP have also noted the loss of crystallinity in operando characterization.38−40 Thus, we used pair distribution function (PDF) analysis to characterize the short-range order and to distinguish possibly amorphous or nanocrystalline phases.41,42 To our knowledge, this is the first instance in which the electrochemically transformed product in olivine cathodes has been characterized by PDF. Ex situ PXD and PDF analyses were simultaneously conducted on two samples discharged to compositions xNa = 0.3 and xNa = 0.6, respectively, between which the operando results in Figure 3A show a declining crystalline fraction. The weight and crystalline

state, and the crystalline phase is predominantly a single phase (FP), the crystalline percentage rises again. Thus, there is a correlation between the minimum in crystalline percentage, and the coexistence of two phases between which there is a large misfit. The NxFP unit cell volume (Figure 3B) actually drops below the value for N0.66FP in the shaded region; applying Vegard’s law to calibrate the unit cell volumes, the sodium content this phase drops to as low as xNa = 0.6. Simultaneously, the unit cell volume of the FP phase increases slightly to a value corresponding to a partially sodiated composition with xNa = 0.08. Therefore, it appears that the solid solution limits of NxFP and FP are slightly extended when the two phases coexist. Between the second charge and second discharge, the phase behavior is largely symmetrical, suggesting that a reversible state has been reached. D

DOI: 10.1021/acs.nanolett.6b04971 Nano Lett. XXXX, XXX, XXX−XXX

Letter

Nano Letters

Figure 5. TEM images of NaxFePO4 at different stages of sodiation. (A) Starting FP powder prepared by chemically delithiating nano LiFePO4 of 47 nm mean BET particle size to xNa = 0. (B) After sodiation to xNa = 0.6, crystallites of the original size remain, as indicated by arrows, but are more sparsely distributed. Similar results are seen for sample sodiated to xNa = 0.3. (C) The larger crystallites are identified to be either highly sodiated NFP or desodiated FP; this example for xNa = 0.3 is NFP. (D and E) The material in between the larger nanocrystallites contains material exhibiting no distinct lattice fringes as well as 5−7 nm diameter crystallites. The former corresponds to the amorphous phase characterized in the PDF analysis. The latter crystallites are also produced as a result of the transformation but, in the PDF analysis, they are included in the refinement of crystalline FP or NFP.

scattering signal. The average lattice parameters refined for the amorphous phase closely match the prototypical FP phase in the a and b directions (a = 9.84 Å, b = 5.78 Å) and the NFP phase in the c direction (c = 4.93 Å). The parameters obtained from PDF refinement for the amorphous phase are given in Table S2, and those for the bulk FP and NFP phases are given in Table S3. The “unit cell” volume for the amorphous phase (280.4 Å3) is between that for NxFP and FP (see Figure 3B). The correspondence of the lattice dimensions and intermediate lattice volume suggests that the amorphous phase may be able to mediate the lattice strain between both of the more-ordered crystalline components, Na-rich NFP and Na-deficient FP, and thereby lower the net strain energy. The abundance of the amorphous third phase from the PDF analysis matches the “missing” diffraction intensity noted in the diffraction analysis. The dimensions of structurally ordered domains was quantified using a spherical envelope model. The FP phase showed progressive reduction in structural coherence with ordered domains up to 5.9 and 3.7 nm for xNa = 0.3 and xNa = 0.6 samples, respectively. Conversely, the NFP phase showed a progressive increase in structural coherence with ordered domains increasing from 12.5 nm. This result suggests the presence of FP and NFP domains that are much smaller than the original olivine crystallites. We then used transmission electron microscopy (TEM) to directly observe the phase distributions corresponding to the PXD and PDF results, and to determine whether the disordered third phase identified in the PDF results could be imaged. Figure 5A shows the starting chemically delithiated FP powder, in which crystallites of ∼50 nm diameter are seen, consistent with the equivalent spherical particle size of 47 nm obtained from the BET specific surface area (35.6 m2/g) of the starting LFP material. After sodiation, crystallites of the same

percentages of FP and NxFP in these samples, and the corresponding lattice parameters, are given in Table S1. As shown in Figure 4A, the calculated two-phase structure model optimized for the powder diffraction data using Rietveld methods did not yield a satisfactory fit to the corresponding PDF data for either the xNa = 0.3 or xNa = 0.6 samples. While the model describes long-range distances well, significant misfits in a local structure are clearly visible (below ca. 12 Å). The presence of these low r features of the residual suggests the existence of an additional phase with only short-range structural coherence. Because diffraction can only probe long-range order in crystalline materials, diffraction would be blind to this component. Attempting to model the whole PDF without this third component biases the PDF fits to short distances and results in a poorer overall fit and discrepancies in the unit cell parameters of the FP phase in comparison to Rietveld refinement analysis. These findings strongly support the presence of a third phase with only short-range structural coherence (that is, an amorphous phase). To test the three-phase model, the structural parameters for a two-phase NFP−FP model were optimized on the basis of the long-range correlations in the PDF (above 20 Å). Keeping the parameters for the initial two-phase fit unchanged, an additional phase, derived from already present FP (Pnma) but with smaller ordered structural domains was added to the model, and the PDF data were calculated over the full r-range. The ordered domain size for this amorphous phase refined to 1.1 nm. This three-phase model (Figure 4B) yielded a good fit to the data. The xNa = 0.3 sample was calculated to contain 31% FP, 40% NFP, and 28% of the amorphous phase, while the xNa = 0.6 sample contained 11% FP, 50% NFP, and 39% of the amorphous phase. Thus, the amorphous phase is not a minor constituent but has substantial contributions to the total E

DOI: 10.1021/acs.nanolett.6b04971 Nano Lett. XXXX, XXX, XXX−XXX

Letter

Nano Letters ORCID

size remain but are now more sparsely distributed, as shown in Figure 5B for the xNa = 0.6 sample. These crystallites were identified, using FFT analysis of lattice fringe images, an example of which is shown in Figure 5C, to be either NFP or FP. The amorphous phase produced during sodiation, not detectable by PXD but characterized by the PDF analysis, appears in substantial volume fraction among the larger nanocrystallites. However, the TEM reveals that the amorphous phase also has embedded within it nanocrystallites of approximately spherical morphology, as seen in Figure 5D,E, that are smaller in size than the starting crystallites by about a factor of 10. Measurements of over 50 particles each in the xNa = 0.3 and 0.6 samples yielded average particles sizes of 7 and 5 nm, respectively, for these smaller crystallites embedded within the amorphous phase. (The respective particle size distributions, and other TEM images of these samples, are given in the Supporting Information.) Thus, based on experimental evidence (loss of crystallinephase fraction upon sodiation and desodiation (Figure 3), PDF analysis of electrochemically cycled samples (Figure 4), and direct observations by TEM (Figure 5)), we identify a unique strain-accommodation mechanism in NaxFePO4 that has not been previously reported for any intercalation compound. An amorphous phase is formed in substantial volume fractions upon electrochemical cycling; this phase does have short-range periodicity close to that of the crystalline phases, suggesting the ability to orient topotaxially against FP or NFP, forming coherent or semicoherent interfaces. By ex situ TEM, we observe smaller nanocrystallites of ∼5 nm size embedded within the amorphous phase. Under these internal constraints, the crystalline sodiated phase has decreased its Na content to about Na0.6FePO4, while the crystalline desodiated phase has increased its Na content to about Na0.08FePO4, presumably driven by coherence strain energy. Although we do not have an independent measurement of the Na content of the amorphous phase, based on its periodicity, it is reasonable to assume that it has Na content intermediate between the crystalline phases. We suggest that the mediating amorphous phase is able to mitigate the otherwise high misfit strain energy of the NFP−FP system at the expense of an increase in structural, interfacial, and chemical energy. The formation of metastable disordered and nanocrystalline phases to alleviate misfit strain energy may be a general response in ion-storage electrodes exhibiting hightransformation strains.

■

Kai Xiang: 0000-0001-5933-4644 Yet-Ming Chiang: 0000-0002-0833-7674 Author Contributions

K.X. and W.X. contributed equally to this work. All authors have given approval to the final version of the manuscript. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS This work was supported by DOE project no. DE-SC0002626. M.T. acknowledges support from the DOE BES Physical Behavior of Materials Program under grant no. DE-SC0014435. Use of the Advanced Photon Source, an Office of Science User Facility operated for the U.S. Department of Energy (DOE) Office of Science by Argonne National Laboratory was supported by the U.S. DOE under contract no. DE-AC0206CH11357. We thank Paul Gionet of A123 Systems LLC, Waltham, Massachusetts for providing the starting lithium iron phosphate powder used in this work and an anonymous reviewer for useful comments.

■

(1) Woodford, W. H.; Chiang, Y.-M.; Carter, W. C. J. Electrochem. Soc. 2010, 157, A1052−A1059. (2) Woodford, W. H.; Carter, W. C.; Chiang, Y.-M. Energy Environ. Sci. 2012, 5, 8014−8024. (3) Massey, C.; McKnight, G.; Barvosa-Carter, W.; Liu, P. Proc. SPIE 2005, 322−330. (4) Koyama, Y.; Chin, T. E.; Rhyner, U.; Holman, R. K.; Hall, S. R.; Chiang, Y.-M. Adv. Funct. Mater. 2006, 16, 492−498. (5) Padhi, A. K.; Nanjundaswamy, K. S.; Goodenough, J. B. D. J. Electrochem. Soc. 1997, 144, 1188−1194. (6) Yamada, A.; Kudo, Y.; Liu, K.-Y. J. Electrochem. Soc. 2001, 148, A1153−A1158. (7) Dodd, J. L.; Yazami, R.; Fultz, B. Electrochem. Solid-State Lett. 2006, 9, A151−A155. (8) Laffont, L.; Delacourt, C.; Gibot, P.; Wu, M. Y.; Kooyman, P.; Masquelier, C.; Tarascon, J. M. Chem. Mater. 2006, 18, 5520−5529. (9) Chen, G.; Song, X.; Richardson, T. J. Electrochem. Solid-State Lett. 2006, 9, A295−A298. (10) Gu, L.; Zhu, C.; Li, H.; Yu, Y.; Li, C.; Tsukimoto, S.; Maier, J.; Ikuhara, Y. J. Am. Chem. Soc. 2011, 133, 4661−4663. (11) Suo, L.; Han, W.; Lu, X.; Gu, L.; Hu, Y.; Li, H.; Chen, D.; Chen, L.; Tsukimoto, S.; Ikuhara, Y. Phys. Chem. Chem. Phys. 2012, 14, 5363−5367. (12) Niu, J.; Kushima, A.; Qian, X.; Qi, L.; Xiang, K.; Chiang, Y.-M.; Li, J. Nano Lett. 2014, 14, 4005−4010. (13) Zhu, Y.; Wang, J.; Liu, Y.; Liu, X.; Kushima, A.; Liu, Y.; Xu, Y.; Mao, S. X.; Li, J.; Wang, C.; Huang, J. Adv. Mater. 2013, 25, 5461− 5466. (14) Yamada, A.; Koizumi, H.; Nishimura, S.; Sonoyama, N.; Kanno, R.; Yonemura, M.; Nakamura, T.; Kobayashi, Y. Nat. Mater. 2006, 5, 357−360. (15) Meethong, N.; Huang, H. S.; Carter, W. C.; Chiang, Y.-M. Electrochem. Solid-State Lett. 2007, 10, A134−A138. (16) Meethong, N.; Huang, H. S.; Speakman, S. A.; Carter, W. C.; Chiang, Y.-M. Adv. Funct. Mater. 2007, 17, 1115−1123. (17) Omenya, F.; Chernova, N. A.; Upreti, S.; Zavalij, P. Y.; Nam, K.; Yang, X.; Whittingham, M. S. Chem. Mater. 2011, 23, 4733−4740. (18) Gibot, P.; Casas-Cabanas, M.; Laffont, L.; Levasseur, S.; Carlach, P.; Hamelet, S.; Tarascon, J.-M.; Masquelier, C. Nat. Mater. 2008, 7, 741−747. (19) Malik, R.; Zhou, F.; Ceder, G. Kinetics of non-equilibrium lithium incorporation in LiFePO4. Nat. Mater. 2011, 10, 587−590.

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.nanolett.6b04971. Additional details of experimental methods. Tables showing characteristics of FP and NFP and a refined structure of the non-crystalline and crystalline components. Figures showing a voltage profile of sodium iron phosphate cell and PXD data, TEM images, and data from Vegard’s law calculations. (PDF)

■

REFERENCES

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Phone: (617)253-6471. Fax: (617) 253-6201. F

DOI: 10.1021/acs.nanolett.6b04971 Nano Lett. XXXX, XXX, XXX−XXX

Letter

Nano Letters (20) Sharma, N.; Guo, X.; Du, G.; Guo, Z.; Wang, J.; Wang, Z.; Peterson, V. K. J. Am. Chem. Soc. 2012, 134, 7867−7873. (21) Liu, H.; Strobridge, F. C.; Borkiewicz, O. J.; Wiaderek, K. M.; Chapman, K. W.; Chupas, P. J.; Grey, C. P. Science 2014, 344, 1252817. (22) Lim, J.; Li, Y.; Alsem, D. H.; So, H.; Lee, S. C.; Bai, P.; Cogswell, D. A.; Liu, X.; Jin, N.; Yu, Y.; Salmon, N. J.; Shapiro, D. A.; Bazant, M. Z.; Tyliszczak, T.; Chueh, W. C. Science 2016, 353, 566−571. (23) Ravnsbæk, D. B.; Xiang, K.; Xing, W.; Borkiewicz, O. J.; Wiaderek, K. M.; Gionet, P.; Chapman, K. W.; Chupas, P. J.; Chiang, Y.-M. Nano Lett. 2014, 14, 1484−1491. (24) Ravnsbæk, D. B.; Xiang, K.; Xing, W.; Borkiewicz, O. J.; Wiaderek, K. M.; Gionet, P.; Chapman, K. W.; Chupas, P. J.; Tang, M.; Chiang, Y.-M. Nano Lett. 2016, 16, 2375−2380. (25) Li, G.; Azuma, H.; Tohda, M. Electrochem. Solid-State Lett. 2002, 5, A135−A137. (26) Limthongkul, P.; Jang, Y.; Dudney, N. J.; Chiang, Y.-M. Acta Mater. 2003, 51, 1103−1113. (27) Obrovac, M. N.; Krause, L. J. J. Electrochem. Soc. 2007, 154, A103−A108. (28) McDowell, M. T.; Lee, S.; Nix, W. D.; Cui, Y. Adv. Mater. 2013, 25, 4966−4985. (29) Yang, H.; Huang, S.; Huang, X.; Fan, F.; Liang, W.; Liu, X. H.; Chen, L. Q.; Huang, J. Y.; Li, J.; Zhu, T.; Zhang, S. Nano Lett. 2012, 12, 1953−1958. (30) An, Y.; Wood, B. C.; Ye, J.; Chiang, Y.-M.; Wang, Y. M.; Tang, M.; Jiang, H. Phys. Chem. Chem. Phys. 2015, 17, 17718−17728. (31) Gaubicher, J.; Boucher, F.; Moreau, P.; Cuisinier, M.; Soudan, P.; Elkaïm, E.; Guyomard, D. Electrochem. Commun. 2014, 38, 104− 106. (32) Galceran, M.; Saurel, D.; Acebedo, B.; Roddatis, V. V.; Martin, E.; Rojo, T.; Casas-Cabanas, M. Phys. Chem. Chem. Phys. 2014, 16, 8837−8842. (33) Borkiewicz, O. J.; Shyam, B.; Wiaderek, K. M.; Kurtz, C.; Chupas, P. J.; Chapman, K. W. J. Appl. Crystallogr. 2012, 45, 1261− 1269. (34) Borkiewicz, O. J.; Wiaderek, K. M.; Chupas, P. J.; Chapman, K. W. J. Phys. Chem. Lett. 2015, 6, 2081−2085. (35) Lu, J.; Chung, S. C.; Nishimura, S.; Yamada, A. Chem. Mater. 2013, 25, 4557−4565. (36) Moreau, P.; Guyomard, D.; Gaubicher, J.; Boucher, F. Chem. Mater. 2010, 22, 4126−4128. (37) Casas-Cabanas, M.; Roddatis, V. V.; Saurel, D.; Kubiak, P.; Carretero-González, J.; Palomares, V.; Serras, P.; Rojo, T. J. Mater. Chem. 2012, 22, 17421−17423. (38) Kao, Y.-H.; Tang, M.; Meethong, N.; Bai, J.; Carter, W. C.; Chiang, Y.-M. Chem. Mater. 2010, 22, 5845−5855. (39) Tang, M.; Carter, W. C.; Belak, J. F.; Chiang, Y.-M. Electrochim. Acta 2010, 56, 969−976. (40) Tang, M.; Carter, W. C.; Chiang, Y.-M. Annu. Rev. Mater. Res. 2010, 40, 501−529. (41) Chupas, P. J.; Chapman, K. W.; Lee, P. L. J. Appl. Crystallogr. 2007, 40, 463−470. (42) Bates, S.; Zografi, G.; Engers, D.; Morris, K.; Crowley, K.; Newman, A. Pharm. Res. 2006, 23, 2333−2349.

G

DOI: 10.1021/acs.nanolett.6b04971 Nano Lett. XXXX, XXX, XXX−XXX