Understanding the Role of Overpotentials inLithium Ion Conversion Reactions: Visualizingthe InterfaceGuennadi Evmenenko,† Robert E. Warburton,‡ Handan Yildirim,‡ Jeffrey P. Greeley,‡

Maria K. Y. Chan,§ D. Bruce Buchholz,† Paul Fenter,§ Michael J. Bedzyk,†

and Timothy T. Fister*,§

†Northwestern University, Evanston, Illinois 60208, United States‡Purdue University, West Lafayette, Indiana 47907, United States§Argonne National Laboratory, Lemont, Illinois 60439, United States

*S Supporting Information

ABSTRACT: Oxide conversion reactions are known to have substantiallyhigher specific capacities than intercalation materials used in Li-ion batteries,but universally suffer from large overpotentials associated with the formationof interfaces between the resulting nanoscale metal and Li2O products. Herewe use the interfacial sensitivity of operando X-ray reflectivity to visualize thestructural evolution of ultrathin NiO electrodes and their interfaces duringconversion. We observe two additional reactions prior to the well-known bulk,three-dimensional conversion occurring at 0.6 V: an accumulation of lithium atthe buried metal/oxide interface (at 2.2 V) followed by interfacial lithiation ofthe buried NiO/Ni interface at the theoretical potential for conversion (at 1.9V). To understand the mechanisms for bulk and interfacial lithiation, wecalculate interfacial energies using density functional theory to build apotential-dependent nucleation model for conversion. These calculations showthat the additional space charge layer of lithium is a crucial component forreducing energy barriers for conversion in NiO.KEYWORDS: battery, reflectivity, reflectometry, nucleation, interfaces

Lithium ion batteries are a key component of manymodern technologies but are still predicated onmaterials capable of intercalating lithium.1,2 While

these materials provide improved reversibility and highervoltages than other energy storage technologies, they areintrinsically limited in specific capacity by lithium site densityin their crystal structures. Reaching higher lithium contentrequires a shift from classic insertion materials to the broaderclass of conversion chemistries where lithium reacts directlywith the host material. By removing the crystallographicconstraint required for intercalation, conversion reactions, suchas those found in a lithium sulfur battery3 or displacementreactions in binary metal oxides4 and fluorides,5 can achievespecific capacities many times higher than materials currentlyused in lithium ion batteries.6 However, these chemistriesuniversally suffer from poor reversibility, both in their poorCoulombic efficiency and significant overpotentials (η = Eeq −Eexp), which have slowed their commercialization.While irreversible capacity is a known byproduct of the large

volume change associated with conversion and can bemitigated by nanoengineering or conductive additives, over-

potentials are intrinsic and less understood.7 In this study, wedetermine the origins of this overpotential of the well-knowndisplacement reaction of a metal oxide into nanoscale metaland lithium species. The energy density of these materials,given by the product of their reaction voltage and specificcapacity, is shown in Figure 1 in comparison to standardcathode materials (LCO = LiCoO2, LMNO = LiMn1.5Ni0.5O2,LFP = LiFePO4). Despite their lower intrinsic voltages, thetheoretical energy density of conversion reactions is oftenhigher than intercalation materials. The significant over-potentials for oxide conversion severely reduce their energydensity, as illustrated in Figure 1.8,9 As a result, oxideconversion materials are largely considered to be anodeswith limited practical viability.The overpotential, which is especially pronounced during

the first discharge, has been attributed to the formation of

Received: March 13, 2019Accepted: May 20, 2019Published: May 22, 2019

Artic

lewww.acsnano.orgCite This: ACS Nano 2019, 13, 7825−7832

© 2019 American Chemical Society 7825 DOI: 10.1021/acsnano.9b02007ACS Nano 2019, 13, 7825−7832

Dow

nloa

ded

via

NO

RT

HW

EST

ER

N U

NIV

on

Nov

embe

r 15

, 201

9 at

19:

12:1

1 (U

TC

).Se

e ht

tps:

//pub

s.ac

s.or

g/sh

arin

ggui

delin

es f

or o

ptio

ns o

n ho

w to

legi

timat

ely

shar

e pu

blis

hed

artic

les.

interfaces between the metal and Li2O products.7 This energybarrier can involve components related to strain imposed bysignificant volume change upon conversion, transport ofelectrons and Li ions through the complex network ofnanoparticles formed during lithiation, and the intrinsicnucleation barrier required for phase separation. Many groupshave improved the overall reversibility and kinetics ofconversion by nanoengineering10−12 or with conductiveadditives,13,14 butwith a few exceptions15the overpotentialappears to be intrinsic. In order to understand the origin of theoverpotentials and possible contributions from interfaces,nanoscale operando characterization techniques are sorelyneeded. The spatial resolution of transmission electronmicroscopy (TEM) has been used to visualize the nucleationand evolution of a conversion reaction.16−18 However, theelectrochemical conditions of in situ TEM are often ill-definedand the complex, three-dimensional nature of conversion canobscure the atomic scale properties of the interface.Here we take an alternative route by studying the lithiation

of ultrathin (<2 nm thick) films of NiO at well-definedpotentials using X-ray reflectivity (XR). X-ray and neutronreflectivity has previously been used to study interfacialreactions in a variety of battery materials, including cathodeand solid electrolyte thin films,19−26 alloying reactions insilicon and silicide composites,27−34 and reactions associatedwith lithium plating and solid electrolyte interphase com-pounds.35−41 Here, we reduce the electrode thickness to thetypical size of a conversion product. This eliminates much ofthe 3D complexity of the reaction and can accentuate the sub-nanometer depth resolution of XR. Such model electrodeslargely minimize the contributions of strain and transport tothe overpotential and so provide direct insights into theintrinsic barriers for lithiation.NiO was chosen as a model system due to the simplicity of

the reaction, NiO + 2Li+ + 2e− ↔ Ni + Li2O, which does notform any intermediates. In spite of this chemical simplicity, thevoltammetry and XR data both reveal two distinct reactions:one near the theoretical potential for lithiation (1.9 V), i.e.,with little or no overpotential, and one at the typicallyobserved discharge plateau for the reaction at 0.6 V. Thehigher potential reaction is initiated at the buried Ni/NiO

interface and is preceded by significant accumulation of Li+ atthis interface. This accumulation is a direct measurement ofcapacitive lithium storage, which was previously speculated tobe a space charge layer giving rise to additional capacity duringthe first discharge that is ubiquitous for oxide conversionreactions.42,43 Using density functional theory (DFT) calcu-lations, we correlate these effects to the dramatically reducedinterfacial energy of Li2O/Ni in the presence of excess lithiumat the interface. Similar findings were found for a Ni/NiO/Nitrilayer with improved reversibility. These results help to clarifythe fundamental origin of overpotentials in metal oxideconversion reactions and demonstrate that, at least in principle,conversion can occur at the theoretical limit.

RESULTS AND DISCUSSIONAmorphous Ni and NiO films were grown on annealed R-plane sapphire substrates by pulsed laser deposition (PLD)and measured in a fully immersed, three-electrode cell (FigureS1) at the Advanced Photon Source, sector 12ID-D. The NiOlayer was masked on the sides to enable electrical contact tothe bottom Ni current collector. Specular X-ray scatteringmeasurements were performed in the Fresnel regime up to amomentum transfer (q) of 1.2 Å−1, providing a vertical spatialresolution of 2.7 Å. Rocking curves measured between eachscan (every 0.1 V) showed no change in width (Δθ = 0.04°),suggesting that the lateral structure over the 1 × 3 mm2

footprint of the beam remained homogeneous throughout thereaction. Similarly prepared NiO/Ni interfaces have beenshown by TEM to have nearly ideal and uniform interfaces,consistent with the results of the XR analysis of the as-deposited samples.44 Samples were lithiated and delithiated bycyclic voltammetry between 3 and 0.3 V vs Li/Li+ at 0.2 mV/sin 1.0 M LiPF6 in 1:1 ethylene carbonate (EC)/dimethylcarbonate (DMC). Given the extremely small amount of NiOpresent (0.28 μg), we carefully shielded the working electrodeconnections to minimize the amount of side reactions with theelectrolyte.Fresnel-normalized reflectivity data taken during the

lithiation of a 17 Å thick NiO film grown on a 25 Å Nicurrent collector are shown in Figure 2a. The logarithmicintensity of the XR is denoted by color to highlight structuralchanges during voltammetry, with data and fits overlaying theplot at select potentials. Data were collected continuouslyduring voltammetry so that each XR scan actually takes placeover a 0.09 V range. Fits were performed at potentials wherethe reflectivity was changing slowly. Changes in the XR datacorrelate well with the current response of the simultaneouslymeasured cyclic voltammetry, shown in Figure 2b, withstructural changes correlating to the two well-definedreduction events centered at 1.7 and 0.6 V. The cyclicvoltammograms (CVs) appear to have some capacitiveresponse as well and are shifted to negative currents due toside-reactions related to solid electrolyte interface processes orelectrolyte decomposition on the exposed current collector.From the raw data alone, the initiation of the lithation reactionat 1.9 V is observed both with a shift of Kiessig fringes in theXR data to lower q, reflecting the overall vertical expansion ofthe heterostructure, and with an abrupt increase in scatteringpower due to heightened interfacial contrast. To quantify thesestructural changes, we show selected fits to the XR data inFigure 2c. The extracted electron density profiles show that thehigher potential phase change is most pronounced at theburied Ni/NiO interface, followed by expansion and lithiation

Figure 1. Comparison of the energy density of intercalation andmetal oxide conversion materials. Note that the energy loss inconversion (denoted by arrows) is largely due to the largeoverpotentials during lithiation, as highlighted for NiO, the subjectof this study. Equilibrium (η = 0) potentials are calculated usingbulk, experimental enthalpies; experimental capacities andobserved lithiation potentials are taken from ref 7.

ACS Nano Article

DOI: 10.1021/acsnano.9b02007ACS Nano 2019, 13, 7825−7832

7826

of the remaining NiO at lower potentials. Even at potentialsexceeding Eeq, we find a small reduction in electron density atthe Ni/NiO interface, consistent with the accumulation of Li+

at this buried interface. This capacitive effect reflects thedouble-layer current measured between 1.9 and 2.5 V andsubtle changes in the XR that are most pronounced at high q.The accumulation of Li+ at this interface, rather than at thesurface, suggests that Li+ conductivity through the nanocrystal-line NiO film is faster than electron transport (consistent withnickel oxide’s large bandgap of 4.3 eV). Below 1.9 V, thisinterlayer density drops quickly to the combined density of theconversion products, indicating an interfacial conversionreaction. As the rest of the NiO film lithiates at lowerpotentials, this interfacial layer eventually reaches the densityof pure Li2O.During delithiation, the NiO surface continued to roughen,

complicating fitting of the bilayer. The first cycle reversibilityof both reactions can be seen in the oxidative sweep in Figure2b, but the overall current is noticeably reduced on the secondcycle. This latter effect could be tied to partial delamination of

the film during delithiation of the planar Li2O layer at thebottom interface of the electrode.Similar measurements on a Ni/NiO/Ni trilayer (14 Å thick

NiO) were performed to better resolve changes in the NiOfilm and to improve the structural reversibility of the planarelectrodes. Such layered heterostructures help confine thereaction products and provide a stronger interfacial contrast forthe overall thickness of the heterostructure.44−46 Previous workin multilayer electrodes suggested that Li+ transport is possiblethrough sufficiently thin metal interlayers.47 With the added Nioverlayer, the two-step lithiation is even more apparent in boththe XR data and the CV (as seen in Figure 3). The CV shows

improved reversibility, both in the delithiation reaction and inthe second lithiation cycle. The density profiles shown inFigure 3c also show the presence of a capacitive Li+ doublelayer at both Ni/NiO interfaces at elevated potentials (2.2 V).This is clearly seen in the raw data (Figure 3a), which show anabrupt increase in thickness and scattering power at 1.9 V. Aswith the simpler bilayer, the interfacial component of thereaction continues until ∼1.5 V. The derived density profiles(Figure 3b) reveal that the bottom interface develops the space

Figure 2. (a) Fresnel normalized reflectivity (R/RF) is shownduring the first lithiation sweep. The XR intensity is denoted bycolor so that potential-dependent changes can be compared withthe CV in (b). Data were measured continuously duringvoltammetry with each XR measurement occurring over 0.09 V.Select curves from the fitting overlay the XR map. (c) Electrondensity profiles and their propagated error bars extracted from fitsto the XR data (visible for the 0.3 V data) are shown for selectpotentials. Reference densities for NiO, Ni, and Li2O and theaverage density of the conversion products (Ni + Li2O) areprovided for reference. The area under the open circuit densityprofile (3.1 V) is colored to emphasize the initial Ni, NiO, andelectrolyte layers. Note the strong changes at the Ni/NiO interfacenear Eeq = 1.9 V, followed by overall expansion, roughening, anddensity reduction in the remaining NiO layer approaching Ebulk =0.6 V.

Figure 3. (a) Normalized reflectivity during lithiation (top) anddelithiation (bottom). As in Figure 2, select curves overlay thedata. CVs taken during the first cycles are shown on the samepotential scale as the XR lithiation data for comparison. (b)Electron density profiles from fits to XR data at select potentialsshow a decrease in density at the interfaces consistent with Liaccumulation, followed by lithiation of the NiO interlayer. Minorchanges also occur at the surface of the top nickel layer due to itsnative oxide. Like Figure 2, the area under the open circuit densityprofile (3.1 V) is colored to highlight the initial trilayer structure.

ACS Nano Article

DOI: 10.1021/acsnano.9b02007ACS Nano 2019, 13, 7825−7832

7827

charge layer first and begins to undergo conversion at 1.9 V.Both buried interfaces react fully, eventually approaching thedensity of pure Li2O (0.57 e−/Å3) with a more Ni-rich (butreacted) interlayer in between. Unlike the first sample, thesurface of the exposed Ni overlayer also undergoes a separateconversion reaction with its native oxide. Much like theunderlying NiO layer, this native oxide develops a slight dip inelectron density consistent with a lithium-rich region thateventually approaches the density of pure Li2O near 0.6 V.The present results from both the bilayer and trilayer

present a consistent picture for the reaction pathway for theNiO conversion reaction (illustrated schematically in Figure4). The first step involves Li+ diffusion through the

electronically insulating NiO to the charged Ni surface,forming a solid-state electric double layer. This layer is notvisible by XR until ∼2.5 V, possibly due to the finite energybarrier to migrate through the ultrathin NiO. At theequilibrium potential for conversion (1.9 V), we see a sharptransition in the XR data that we associate with the suddennucleation of a conversion reaction at the Ni/NiO interface(including growth of the bottom nickel layer and sub-nanometer thick Li2O layer at its interface with the currentcollector). However, the CVs in each case do not show asimilar discontinuity. The more gradual changes in theinterfacial structure (near the 1.7 V peak in the CV) suggestthat NiO redox may occur more gradually as the Li+ spacecharge layer is consumed to form a Li2O-rich layer that alsoblocks interfacial sites for Ni reduction. The formation of thislayer appears to create a barrier for interaction between Li+ ande− species, limiting the extent of reaction to <1 nm from theoriginal interface. At lower potentials (ca. 0.6 V), the remainingNiO reacts to form a rougher film with average density

matching the Ni + Li2O conversion products. Here, the densityof the interfacial layer decreases even further, consistent with apure Li2O layer at the interface, suggesting that the isolated Niproducts from the first reaction aggregate into the Ni electrode.Given the limited thickness of the interfacial region, our

results suggest that the associated lithiation reactions are arelatively small component of the discharge in bulk conversionelectrodes, which could be inferred from previous studies ofNiO.48 However, it is striking that in such nanoscale electrodesthere is still a strong separation between “bulk” and interfacialreactions given the nanoscale dimensions of the film. To betterunderstand the origins of this interfacial reaction, we used DFTto calculate interfacial energies (γ) associated with the NiO,Ni, and Li2O species involved in the reaction. Coupled withthe bulk free energy of reaction (ΔGbulk) for NiO conversion,we calculate the intrinsic nucleation barriers and the size of thenucleation products for conversion. This model incorporatesgeometric considerations from the formation/destruction ofdifferent solid−solid interfaces as a result of nucleation at theNi/NiO interface (ΔAi). Additionally, the model is potential-dependent due to the presence of a (Li+ + e−) pair in reactantsof the bulk reaction stoichiometry:

∑ γΔ = [Δ = − ] + ΔG E n n G E eE A( , ) ( 0 V) 2i

i ibulk

interfaces

(1)

Note that the first term (bulk thermodynamics, which scaleswith the number of NiO sites, n) becomes negative when E isless than the equilibrium potential (Eeq), and the second termaccounts for all new interfaces (with area ΔA) formed uponconversion. While this is a simplification of conversion in amacroscopic electrode, these intrinsic quantities are intendedto resemble the simplified geometry of these ultrathinelectrodes, where energy barriers due to strain and masstransport are largely minimized.DFT calculations were performed for all binary permuta-

tions of solid−solid interface formation between Ni(111),NiO(100), and amorphous Li2O. The resulting interfacialenergies were incorporated into the nucleation model in eq 1(further details are provided in the Supporting Information).Based on the XR data, we incorporated a Li2O layer around anassumed Ni nucleus forming at the Ni/NiO interface into ourtheoretical model, as illustrated in Figure 5a. Using eq 1, wecalculated a nucleation barrier of 7.32 eV at 1.4 V, while thisbarrier is gradually reduced to 0.52 eV near the observed bulkpotential of 0.6 V. Clearly, the barrier associated with thismodel is unrealistically large at 1.4 V, whereas the significantlyreduced 0.52 eV barrier at 0.6 V suggests only a modest barrierto formation of the discharge product at these potentials.Although the nucleation model suggests possible mecha-

nisms for conversion and growth at 0.6 V, wherein much of thecurrent density from conversion is observed (reference CVs inFigures 2b and 3a), it does not address the observed interfacialreactions that occur near Eeq. To more comprehensivelyinterpret the XR results in this region, we additionally modeledthe presence of excess interfacial lithium (shown by theillustration in Figure 5b), as observed through XR at E > Eeq(see Figures 2c and 3b). Incorporation of a single lithiummonolayer between the Ni(111) substrate and the amorphousLi2O reduces the calculated interfacial energy to approximately0.41 J/m2. We note that the exact coverage of excess lithium isnot known, and the value of one monolayer assumed in the

Figure 4. Schematic of the interfacial and bulk conversionprocesses during the first lithiation including (a) lithiumaccumulation, (b) nucleation of conversion products at the buriedNi/NiO interface, (c) formation of Ni and Li2O layers at theburied interface, and (d) bulk three-dimensional conversion. Forreference, an electron density profile (similar to those measured inFigure 2) is shown with each picture. In each density profile, theoriginal film is denoted by the black dashed line, and changes inthe density profile at each step are highlighted by therepresentative color of the reacting species.

ACS Nano Article

DOI: 10.1021/acsnano.9b02007ACS Nano 2019, 13, 7825−7832

7828

calculations represents a likely upper bound to this quantity,implying that the magnitude of the calculated change ininterfacial energy is also an upper bound. Applying the lowerLi2O/Li/Ni interfacial energy to the free energy expression ineq 1 leads to a net interfacial contribution to the total ΔG thatis negative, since the higher energy Ni/NiO interface ispartially destroyed (γNi/NiO = 1.44 J/m2, ΔA < 0), while lowerenergy Li2O/Li/Ni (γLi2O/Li/Ni = 0.41 J/m2) and NiO/Li2O(γNiO/Li2O = 0.35 J/m2) interfaces are formed (ΔA > 0). Thenegative interfacial component of ΔG leads to prediction of abarrierless nucleation reaction, as shown in Figure 5c,d. Thisresult implies that the formation of the Li2O/Li/Ni interface islikely to be facile upon introduction of an electrochemical bias,as is also suggested by the interfacial conversion observed byXR at high potentials. We note that the presence of excesslithium appears to be consistent with the formation of a spacecharge layer that catalyzes the initial conversion reaction at theburied Ni/NiO interface. In addition, we expect thatconversion at interfacial defects (e.g., steps, kinks) on thenickel surface is likely to reduce nucleation barriers andoverpotentials even further, as previously suggested.49 Thesebarriers are consistent with the observed conversion processesin the region E > 0.6 V and underline the importance of excesslithium in reducing interfacial energies between Ni- and Li2O-rich regions of the conversion product. Furthermore, theconsumption of the Li+ to form Li2O in this region is likelywhy this phenomenon is self-limiting until bulk conversion isobserved near 0.6 V.

CONCLUSIONSThis study enhances the understanding and visualization of thenanoscale interfacial reactions in lithium ion conversionreactions through the use of model NiO thin-film electrodes.We identified an interfacially controlled reaction step in theconversion mechanism, in addition to the known bulkconversion processes. We developed a simple model to

understand how the interplay between capacitive and Faradaicprocesses can lead to an interfacial conversion reaction nearthe theoretically predicted potentials, revealing that theobserved behavior is controlled by the presence of an excessof lithium at the buried Ni/NiO interface. These resultssuggest that a viable route toward reducing overpotentials inoxide conversion reactions can be achieved by engineeringmetal/oxide interfaces as a solid-state pseudocapacitor.50

These principles could be extended to the study andoptimization of other conversion battery materials such asfluorides,5,51 which have higher theoretical equilibriumpotentials than oxides in general, but still experience significantoverpotentials and voltage hysteresis. While the inherentformation of interfaces remains a fundamental obstacle foroxide conversion reactions, the use of tailored, pre-existinginterfaces that can systematically deliver Li to the reaction sitecould, ironically, be the solution.

METHODS AND EXPERIMENTALFilm Growth. Nickel/nickel oxide bilayer and trilayer thin films

were grown by PLD on the 10 × 3 mm2 R-face surfaces of 1 mm thicksapphire α-Al2O3 (102) substrates (CrysTec GmbH, Germany) usinga PLD/MBE 2300 (PLD Products) system in the NorthwesternUniversity Pulsed Laser Deposition Facility. The system employed a248 nm KrF excimer laser with a 25 ns pulse operating at 5 Hz. Thelaser was focused to a 1.5 × 3.5 mm2 spot size on the targets, whichwere rotated at 5 rpm to prevent localized heating. The target−substrate separation was fixed at 6 cm. Nickel was deposited from ametallic nickel target at a laser energy of 300 mJ per pulse at thechamber base pressure of ∼5 × 10−7 Torr. There was a 30% energyloss along the optical train to yield an energy density at the target of 4J/(pulse × cm2). Nickel oxide was deposited from a dense hot-pressednickel oxide target at an energy of 200 mJ per pulse in a depositionambient of 5 × 10−4 Torr ultra-high-purity (99.994%) oxygen. Therewas a 30% energy loss along the optical train to yield an energydensity at the target of 2.7 J/(pulse × cm2). The thickness of eachdeposited layer was controlled by adjusting the number of laserpulses: nickel deposited at 0.01 Å/pulse; NiO deposited at 0.1 Å/

Figure 5. (a, b) Illustrations of the models used to calculate the total energy required to undergo conversion. (c, d) Solid lines: ΔG (per NiOsite), including bulk, interfacial, and total contributions, calculated using γNi/NiO = 1.44 J/m2, γNiO/Li2O = 0.35 J/m2, and γLi2O/Ni = 1.86 J/m2

at 1.4 and 0.6 V, respectively. The dashed lines represent the same model with excess lithium incorporated at the Ni/Li2O interface, whichreduces γLi2O/Li/Ni to 0.41 J/m2. In each case, the presence of excess interfacial lithium leads to a negative interfacial contribution to the freeenergy change (second term of the energy expression in eq 1) due to the destruction of the higher energy Ni/NiO interface and formation ofthe lower energy NiO/Li2O and Li2O/Li/Ni interfaces. In the presence of excess lithium, which resembles a space charge layer at the Li2O/Ni interface, the conversion reaction occurring directly at the buried Ni/NiO interface would be expected to be facile upon introduction ofan electrochemical bias.

ACS Nano Article

DOI: 10.1021/acsnano.9b02007ACS Nano 2019, 13, 7825−7832

7829

pulse. The bottom nickel layer of each heterostructure was maskedprior to deposition of subsequent layers to form a 3 × 3 mm2

electrical contact with the spring-loaded electrode of the electro-chemical cell.32

X-ray Reflectivity. The operando X-ray reflectivity experimentswere performed with an X-ray energy of 20.00 keV at the 12ID-DAdvanced Photon Source station at Argonne National Laboratory.The X-ray beam was defocused and collimated to 1.0 × 0.18 mm2,and the scattered X-ray pattern was acquired with a Pilatus 100kdetector. Full reflectivity data scans were measured in 15 min (every0.09 V) and were collected repeatedly during electrochemical cycling.Data were analyzed by subtracting the background (linear fit in thesample-χ direction) in each Pilatus image. Reflectivity was normalizedby the incident intensity and beam spill-off, which was proportional toq. The electrochemical cell32 had separate lithium metal counter andreference electrodes, and the samples were fully immersed in a 1 Msolution of LiClO4 in a 1:1 ratio by volume of EC and DMC. ACHI760E electrochemical workstation was used for electrochemicalcontrol of lithiation. Ex situ XR studies of test samples were carriedout at a Rigaku ATXG diffractometer (NU X-ray Diffraction Facility)with E = 8.04 keV (λ = 1.54 Å) X-rays collimated to a 0.1 × 2.0 mm2

spot. All XR measurements were performed at ambient laboratorytemperature, which ranged between 20 and 25 °C.XR analysis used Motofit52 with a multiple-slab model that

included a sapphire substrate, Ni and NiO layers, and an electrolyte(operando) or air (ex situ experiments). Structural parameters forsapphire and the electrolyte were fixed, whereas the parameters for thebuffer and active layers (electron density, interface roughness, andlayer thickness) were allowed to vary. The electron densities wereinitially estimated based on the chemical composition of themultilayer electrode components. As discussed in the SI, the errorsin the structural parameters and their covariances were used tocalculate the uncertainty for the overall density profile (i.e., as afunction of height). In most cases, the errors in derived electrondensity are actually smaller than the line-width used in the plots.Individual layer parameters and their errors are reported in the SI.First-Principles Calculations. Spin-polarized DFT calculations

were performed using the Vienna ab Initio Software Package(VASP).53−55 The generalized gradient approximation (GGA) ofPerdew−Burke−Ernzerhof (PBE) was used as the exchange−correlation functional.56 Γ-centered k-point grids were used forBrillouin zone sampling, where the product of the number of k-pointsand the corresponding lattice vector was ∼30 Å. The projector-augmented wave method57,58 was used to treat the effective corepotentials, with Li 1s electrons treated explicitly as valence in allcalculations. For interfacial models between Ni and amorphouslithium (a-Li2O), the Kohn−Sham wave functions were expanded in aplane-wave basis set up to kinetic energy of 400 eV. All modelscontaining NiO (interfacial models between NiO/a-Li2O and Ni/NiO) used a kinetic energy cutoff of 520 eV. Calculations containingNiO used the DFT+U method59−62 with an effective U value of 6.2eV applied to the Ni 3d electrons.63 A U value of 4.0 eV was appliedto the 3d states of metallic Ni for Ni/NiO interfacial models toaccurately reproduce experimental formation enthalpies for NiO andLiNiO2 as in previous work.

64,65 Further details are provided in the SI.Interfacial energies were calculated for interfaces between Ni/a-

Li2O, NiO/a-Li2O, and Ni/NiO. For models containing amorphouslithium, an ensemble of structures is constructed through sampling abulk supercell of a-Li2O. The a-Li2O supercell was generated from amelt and quench approach within classical molecular dynamicssimulations performed using the DL_POLY code66 (see SI for furtherdetails). Crystalline Ni/NiO interfaces were made using an in-houselattice matching algorithm to minimize strain between Ni(111) andNiO(100). We sampled several lattice matching ratios about theminimum strain model output from the algorithm to ensure theinterfacial energy is well optimized. Interfacial energies, γ, betweentwo compounds, α and β, were evaluated using the general formula

γ = [ − + ]α α β βAE n g n g

12

( )totalDFT bulk bulk

(2)

where EtotalDFT, ni, and gi

bulk are the total DFT energy of an interfacialmodel, the number of formula units of compound i, and the performula unit bulk free energy of compound i, respectively, for aninterface with interfacial area A. Structures that included an extra layerof lithium were treated using bulk metallic lithium as reference, sinceanalysis of interfacial electron density suggested a metallic nature fornearly all of the excess lithium in this layer. We used an amorphousLi2O reference state, as described further in the SI.

ASSOCIATED CONTENT*S Supporting InformationThe Supporting Information is available free of charge on theACS Publications website at DOI: 10.1021/acsnano.9b02007.

Further details related to the sample cell, data analysisprocedures, and theoretical calculations; results for allparameters extracted from fits to the reflectivity (PDF)

AUTHOR INFORMATIONCorresponding Author*E-mail: fister@anl.gov.ORCIDGuennadi Evmenenko: 0000-0002-6580-2237Handan Yildirim: 0000-0002-3060-7896Jeffrey P. Greeley: 0000-0001-8469-1715Maria K. Y. Chan: 0000-0003-0922-1363Paul Fenter: 0000-0002-6672-9748Michael J. Bedzyk: 0000-0002-1026-4558Timothy T. Fister: 0000-0001-6537-6170Author ContributionsG.E., T.T.F., M.J.B., and P.F. performed the XR measurementsand analyzed the data. D.B.B grew the thin films. R.E.W., H.Y.,M.K.Y.C., and J.P.G. performed the DFT and MD calculations.All authors were involved in the writing of the manuscript.NotesThe authors declare no competing financial interest.

ACKNOWLEDGMENTSThis work was supported by the Center for ElectrochemicalEnergy Science, an Energy Frontier Research Centre funded bythe U.S. Department of Energy, Office of Science, Office ofBasic Energy Sciences, under award number DE-AC02-06CH11. Use of the Advanced Photon Source and the Centerfor Nanoscale Materials, both Office of Science user facilitiesoperated for the DOE, Office of Science by Argonne NationalLaboratory, was supported by the U.S. DOE under ContractNo. DE-AC02-06CH11357. Use of the Pulse Laser Depositionand X-ray Diffraction Facilities at Northwestern University aresupported by the MRSEC under National Science Foundationgrant DMR-1720139.

REFERENCES(1) Li, M.; Lu, J.; Chen, Z. W.; Amine, K. 30 Years of Lithium-IonBatteries. Adv. Mater. 2018, 30, 24.(2) Whittingham, M. S. Lithium Batteries and Cathode Materials.Chem. Rev. 2004, 104, 4271−4301.(3) Bruce, P. G.; Freunberger, S. A.; Hardwick, L. J.; Tarascon, J. M.Li-O-2 and Li-S Batteries with High Energy Storage. Nat. Mater.2012, 11, 19−29.(4) Poizot, P.; Laruelle, S.; Grugeon, S.; Dupont, L.; Tarascon, J. M.Nano-Sized Transition-Metaloxides as Negative-Electrode Materialsfor Lithium-Ion Batteries. Nature 2000, 407, 496−499.

ACS Nano Article

DOI: 10.1021/acsnano.9b02007ACS Nano 2019, 13, 7825−7832

7830

(5) Wang, F.; Robert, R.; Chernova, N. A.; Pereira, N.; Omenya, F.;Badway, F.; Hua, X.; Ruotolo, M.; Zhang, R.; Wu, L.; Volkov, V.; Su,D.; Key, B.; Whittingham, M. S.; Grey, C. P.; Amatucci, G. G.; Zhu,Y.; Graetz, J. Conversion Reaction Mechanisms in Lithium IonBatteries: Study of the Binary Metal Fluoride Electrodes. J. Am. Chem.Soc. 2011, 133, 18828−18836.(6) Wu, F. X.; Yushin, G. Conversion Cathodes for RechargeableLithium and Lithium-Ion Batteries. Energy Environ. Sci. 2017, 10,435−459.(7) Cabana, J.; Monconduit, L.; Larcher, D.; Palacin, M. R. BeyondIntercalation-Based Li-Ion Batteries: The State of the Art andChallenges of Electrode Materials Reacting through ConversionReactions. Adv. Mater. 2010, 22, E170−E192.(8) Nitta, N.; Wu, F. X.; Lee, J. T.; Yushin, G. Li-Ion BatteryMaterials: Present and Future. Mater. Mater. Today 2015, 18, 252−264.(9) Aravindan, V.; Lee, Y. S.; Madhavi, S. Research Progress onNegative Electrodes for Practical Li-Ion Batteries: Beyond Carbona-ceous Anodes. Adv. Energy Mater. 2015, 5, 1402225(10) Cherian, C. T.; Sundaramurthy, J.; Kalaivani, M.; Ragupathy,P.; Kumar, P. S.; Thavasi, V.; Reddy, M. V.; Sow, C. H.; Mhaisalkar, S.G.; Ramakrishna, S.; Chowdari, B. V. R. Electrospun Alpha-Fe2o3Nanorods as a Stable, High Capacity Anode Material for Li-IonBatteries. J. Mater. Chem. 2012, 22, 12198−12204.(11) Huang, Y.; Huang, X. L.; Lian, J. S.; Xu, D.; Wang, L. M.;Zhang, X. B. Self-Assembly of Ultrathin Porous Nio Nanosheets/Graphene Hierarchical Structure for High-Capacity and High-RateLithium Storage. J. Mater. Chem. 2012, 22, 2844−2847.(12) Wu, H. B.; Chen, J. S.; Hng, H. H.; Lou, X. W. NanostructuredMetal Oxide-Based Materials as Advanced Anodes for Lithium-IonBatteries. Nanoscale 2012, 4, 2526−2542.(13) Huang, X. L.; Wang, R. Z.; Xu, D.; Wang, Z. L.; Wang, H. G.;Xu, J. J.; Wu, Z.; Liu, Q. C.; Zhang, Y.; Zhang, X. B. HomogeneousCoo on Graphene for Binder-Free and Ultralong-Life Lithium IonBatteries. Adv. Funct. Mater. 2013, 23, 4345−4353.(14) Wu, Y.; Wei, Y.; Wang, J. P.; Jiang, K. L.; Fan, S. S. ConformalFe3o4 Sheath on Aligned Carbon Nanotube Scaffolds as High-Performance Anodes for Lithium Ion Batteries. Nano Lett. 2013, 13,818−823.(15) Larcher, D.; Masquelier, C.; Bonnin, D.; Chabre, Y.; Masson,V.; Leriche, J. B.; Tarascon, J. M. Effect of Particle Size on LithiumIntercalation into Alpha-Fe2o3. J. Electrochem. Soc. 2003, 150, A133−A139.(16) Wang, F.; Robert, R.; Chernova, N. A.; Pereira, N.; Omenya, F.;Badway, F.; Hua, X.; Ruotolo, M.; Zhang, R. G.; Wu, L. J.; Volkov, V.;Su, D.; Key, B.; Whittingharn, M. S.; Grey, C. P.; Amatucci, G. G.;Zhu, Y. M.; Graetz, J. Conversion Reaction Mechanisms in LithiumIon Batteries: Study of the Binary Metal Fluoride Electrodes. J. Am.Chem. Soc. 2011, 133, 18828−18836.(17) Luo, L. L.; Wu, J. S.; Xu, J. M.; Dravid, V. P. Atomic ResolutionStudy of Reversible Conversion Reaction in Metal Oxide Electrodesfor Lithium-Ion Battery. ACS Nano 2014, 8, 11560−11566.(18) Lin, F.; Nordlund, D.; Weng, T. C.; Zhu, Y.; Ban, C. M.;Richards, R. M.; Xin, H. L. Phase Evolution for Conversion ReactionElectrodes in Lithium-Ion Batteries. Nat. Commun. 2014, 5,DOI: 10.1038/ncomms4358.(19) Dupre, N.; Cuisinier, M.; Zheng, Y.; Fernandez, V.; Hamon, J.;Hirayama, M.; Kanno, R.; Guyomard, D. Evolution of Lifepo4 ThinFilms Interphase with Electrolyte. J. Power Sources 2018, 382, 45−55.(20) Kim, S.; Hirayama, M.; Suzuki, K.; Kanno, R. Hetero-EpitaxialGrowth of Li0.17la0.6tio3 Solid Electrolyte on Limn2o4 Electrode forAll Solid-State Batteries. Solid State Ionics 2014, 262, 578−581.(21) Minato, T.; Kawaura, H.; Hirayama, M.; Taminato, S.; Suzuki,K.; Yamada, N. L.; Sugaya, H.; Yamamoto, K.; Nakanishi, K.; Orikasa,Y.; Tanida, H.; Kanno, R.; Arai, H.; Uchimoto, Y.; Ogumi, Z.Dynamic Behavior at the Interface between Lithium Cobalt Oxideand an Organic Electrolyte Monitored by Neutron ReflectivityMeasurements. J. Phys. Chem. C 2016, 120, 20082−20088.

(22) Hirayama, M.; Sonoyama, N.; Abe, T.; Minoura, M.; Ito, M.;Mori, D.; Yamada, A.; Kanno, R.; Terashima, T.; Takano, M.;Tamura, K.; Mizuki, J. Characterization of Electrode/ElectrolyteInterface for Lithium Batteries Using in Situ Synchrotron X-RayReflectometry - a New Experimental Technique for Licoo2ModelElectrode. J. Power Sources 2007, 168, 493−500.(23) Hirayama, M.; Yonemura, M.; Suzuki, K.; Torikai, N.; Smith,H.; Watkinsand, E.; Majewski, J.; Kanno, R. Surface Characterizationof Lifepo4 Epitaxial Thin Films by X-Ray/Neutron Reflectometry.Electrochemistry 2010, 78, 413−415.(24) Ishikawa, H.; Hirano, T.; Nagasaka, Y.; Kawakami, K.; Ohta,H.; Nanba, T.; Hirano, A.; Kanno, R. Reflectivity Measurements ofSuperionic Conductors for Li Ion Secondary Battery Materials. J. Phys.Chem. Solids 2005, 66, 2065−2067.(25) Kim, K.; Toujigamori, T.; Suzuki, K.; Taminato, S.; Tamura, K.;Mizuki, J.; Hirayama, M.; Kanno, R. Characterization of Nano-SizedEpitaxial Li4ti5o12(110) Film Electrode for Lithium Batteries.Electrochemistry 2012, 80, 800−803.(26) Zheng, Y. M.; Taminato, S.; Suzuki, K.; Hirayama, M.; Kanno,R. Fabrication and Lithium Intercalation Properties of EpitaxialLi2ruo3 Thin Films. Thin Solid Films 2012, 520, 4889−4893.(27) Jerliu, B.; Huger, E.; Dorrer, L.; Seidlhofer, B. K.; Steitz, R.;Horisberger, M.; Schmidt, H. Lithium Insertion into SiliconElectrodes Studied by Cyclic Voltammetry and Operando NeutronReflectometry. Phys. Chem. Chem. Phys. 2018, 20, 23480−23491.(28) Jerliu, B.; Huger, E.; Horisberger, M.; Stahn, J.; Schmidt, H.Irreversible Lithium Storage During Lithiation of Amorphous SiliconThin Film Electrodes Studied by in-Situ Neutron Reflectometry. J.Power Sources 2017, 359, 415−421.(29) Seidlhofer, B. K.; Jerliu, B.; Trapp, M.; Huger, E.; Risse, S.;Cubitt, R.; Schmidt, H.; Steitz, R.; Ballauff, M. Lithiation ofCrystalline Silicon as Analyzed by Operando Neutron Reflectivity.ACS Nano 2016, 10, 7458−7466.(30) Cao, C. T.; Steinrck, H. G.; Shyam, B.; Toney, M. F. TheAtomic Scale Electrochemical Lithiation and Delithiation Process ofSilicon. Adv. Mater. Interfaces 2017, 4, 1700771.(31) Cao, C. T.; Steinruck, H. G.; Shyam, B.; Stone, K. H.; Toney,M. F. In Situ Study of Silicon Electrode Lithiation with X-RayReflectivity. Nano Lett. 2016, 16, 7394−7401.(32) Fister, T. T.; Long, B. R.; Gewirth, A. A.; Shi, B.; Assoufid, L.;Lee, S. S.; Fenter, P. Real-Time Observations of Interfacial Lithiationin a Metal Silicide Thin Film. J. Phys. Chem. C 2012, 116, 22341−22345.(33) Jerliu, B.; Huger, E.; Dorrer, L.; Seidlhofer, B. K.; Steitz, R.;Oberst, V.; Geckle, U.; Bruns, M.; Schmidt, H. Volume ExpansionDuring Lithiation of Amorphous Silicon Thin Film ElectrodesStudied by in-Operando Neutron Reflectometry. J. Phys. Chem. C2014, 118, 9395−9399.(34) Huger, E.; Dorrer, L.; Rahn, J.; Panzner, T.; Stahn, J.;Lilienkamp, G.; Schmidt, H. Lithium Transport through NanosizedAmorphous Silicon Layers. Nano Lett. 2013, 13, 1237−1244.(35) Steinhauer, M.; Stich, M.; Kurniawan, M.; Seidlhofer, B. K.;Trapp, M.; Bund, A.; Wagner, N.; Friedrich, K. A. In Situ Studies ofSolid Electrolyte Interphase (Sei) Formation on Crystalline CarbonSurfaces by Neutron Reflectometry and Atomic Force Microscopy.ACS Appl. Mater. Interfaces 2017, 9, 35794−35801.(36) Steinruck, H. G.; Cao, C. T.; Tsao, Y.; Takacs, C. J.; Konovalov,O.; Vatamanu, J.; Borodin, O.; Toney, M. F. The Nanoscale Structureof the Electrolyte-Metal Oxide Interface. Energy Environ. Sci. 2018, 11,594−602.(37) Avdeev, M. V.; Rulev, A. A.; Bodnarchuk, V. I.; Ushakova, E. E.;Petrenko, V. I.; Gapon, I. V.; Tomchuk, O. V.; Matveev, V. A.;Pleshanov, N. K.; Kataev, E. Y.; Yashina, L. V.; Itkis, D. M.Monitoring of Lithium Plating by Neutron Reflectometry. Appl. Surf.Sci. 2017, 424, 378−382.(38) Veith, G. M.; Doucet, M.; Sacci, R. L.; Vacaliuc, B.; Baldwin, J.K.; Browning, J. F. Determination of the Solid Electrolyte InterphaseStructure Grown on a Silicon Electrode Using a Fluoroethylene

ACS Nano Article

DOI: 10.1021/acsnano.9b02007ACS Nano 2019, 13, 7825−7832

7831

Carbonate Additive. Sci. Rep. 2017, 7, DOI: 10.1038/s41598-017-06555-8.(39) Fears, T. M.; Doucet, M.; Browning, J. F.; Baldwin, J. K. S.;Winiarz, J. G.; Kaiser, H.; Taub, H.; Sacci, R. L.; Veith, G. M.Evaluating the Solid Electrolyte Interphase Formed on SiliconElectrodes: A Comparison of Ex Situ X-Ray Photoelectron Spectros-copy and in Situ Neutron Reflectometry. Phys. Chem. Chem. Phys.2016, 18, 13927−13940.(40) Kawaura, H.; Harada, M.; Kondo, Y.; Kondo, H.; Suganuma,Y.; Takahashi, N.; Sugiyama, J.; Seno, Y.; Yamada, N. L. OperandoMeasurement of Solid Electrolyte Interphase Formation at WorkingElectrode of Li-Ion Battery by Time-Slicing Neutron Reflectometry.ACS Appl. Mater. Interfaces 2016, 8, 9540−9544.(41) Owejan, J. E.; Owejan, J. P.; DeCaluwe, S. C.; Dura, J. A. SolidElectrolyte Interphase in Li-Ion Batteries: Evolving StructuresMeasured in Situ by Neutron Reflectometry. Chem. Mater. 2012,24, 2133−2140.(42) Maier, J. Thermodynamics of Electrochemical Lithium Storage.Angew. Chem., Int. Ed. 2013, 52, 4998−5026.(43) Chen, C. C.; Maier, J. Decoupling Electron and Ion Storageand the Path from Interfacial Storage to Artificial Electrodes. Nat.Energy 2018, 3, 102−108.(44) Evmenenko, G.; Fister, T. T.; Buchholz, D. B.; Li, Q. Q.; Chen,K. S.; Wu, J. S.; Dravid, V. P.; Hersam, M. C.; Fenter, P.; Bedzyk, M. J.Morphological Evolution of Multilayer Ni/Nio Thin Film ElectrodesDuring Lithiation. ACS Appl. Mater. Interfaces 2016, 8, 19979−19986.(45) Fister, T. T.; Esbenshade, J.; Chen, X.; Long, B. R.; Shi, B.;Schleputz, C. M.; Gewirth, A. A.; Bedzyk, M. J.; Fenter, P. LithiumIntercalation Behavior in Multilayer Silicon Electrodes. Adv. EnergyMater. 2014, 4,.(46) Fister, T. T.; Hu, X. Y.; Esbenshade, J.; Chen, X.; Wu, J. S.;Dravid, V.; Bedzyk, M.; Long, B.; Gewirth, A. A.; Shi, B.; Schleputz,C. M.; Fenter, P. Dimensionally Controlled Lithiation of ChromiumOxide. Chem. Mater. 2016, 28, 47−54.(47) Evmenenko, G.; Fister, T. T.; Buchholz, D. B.; Castro, F. C.; Li,Q. Q.; Wu, J. S.; Dravid, V. P.; Fenter, P.; Bedzyk, M. J. Lithiation ofMultilayer Ni/Nio Electrodes: Criticality of Nickel Layer Thicknesseson Conversion Reaction Kinetics. Phys. Chem. Chem. Phys. 2017, 19,20029−20039.(48) Ponrouch, A.; Cabana, J.; Dugas, R.; Slack, J. L.; Palacin, M. R.Electroanalytical Study of the Viability of Conversion Reactions asEnergy Storage Mechanisms. RSC Adv. 2014, 4, 35988−35996.(49) Zhukovskii, Y. F.; Balaya, P.; Kotomin, E. A.; Maier, J. Evidencefor Interfacial-Storage Anomaly in Nanocomposites for LithiumBatteries from First-Principles Simulations. Phys. Rev. Lett. 2006, 96,DOI: 10.1103/PhysRevLett.96.058302.(50) Augustyn, V.; Come, J.; Lowe, M. A.; Kim, J. W.; Taberna, P.L.; Tolbert, S. H.; Abruna, H. D.; Simon, P.; Dunn, B. High-RateElectrochemical Energy Storage through Li+ Intercalation Pseudoca-pacitance. Nat. Mater. 2013, 12, 518−522.(51) Kim, S. W.; Pereira, N.; Chernova, N. A.; Omenya, F.; Gao, P.;Whittingham, M. S.; Amatucci, G. G.; Su, D.; Wang, F. StructureStabilization by Mixed Anions in Oxyfluoride Cathodes for High-Energy Lithium Batteries. ACS Nano 2015, 9, 10076−10084.(52) Nelson, A. Co-Refinement of Multiple-Contrast Neutron/X-Ray Reflectivity Data Using Motofit. J. Appl. Crystallogr. 2006, 39,273−276.(53) Kresse, G.; Furthmuller, J. Efficiency of Ab-Initio Total EnergyCalculations for Metals and Semiconductors Using a Plane-WaveBasis Set. Comput. Mater. Sci. 1996, 6, 15−50.(54) Kresse, G.; Furthmuller, J. Efficient Iterative Schemes for AbInitio Total-Energy Calculations Using a Plane-Wave Basis Set. Phys.Rev. B: Condens. Matter Mater. Phys. 1996, 54, 11169−11186.(55) Kresse, G.; Hafner, J. Ab Initio Molecular Dynamics for LiquidMetals. Phys. Rev. B: Condens. Matter Mater. Phys. 1993, 47, 558−561.(56) Perdew, J. P.; Burke, K.; Ernzerhof, M. Generalized GradientApproximation Made Simple. Phys. Rev. Lett. 1996, 77, 3865−3868.(57) Blochl, P. E. Projector Augmented-Wave Method. Phys. Rev. B:Condens. Matter Mater. Phys. 1994, 50, 17953−17979.

(58) Kresse, G.; Joubert, D. From Ultrasoft Pseudopotentials to theProjector Augmented-Wave Method. Phys. Rev. B: Condens. MatterMater. Phys. 1999, 59, 1758−1775.(59) Anisimov, V. I.; Solovyev, I. V.; Korotin, M. A.; Czyzyk, M. T.;Sawatzky, G. A. Density-Functional Theory and Nio PhotoemissionSpectra. Phys. Rev. B: Condens. Matter Mater. Phys. 1993, 48, 16929−16934.(60) Anisimov, V. I.; Zaanen, J.; Andersen, O. K. Band Theory andMott Insulators: Hubbard U Instead of Stoner I. Phys. Rev. B:Condens. Matter Mater. Phys. 1991, 44, 943−954.(61) Liechtenstein, A. I.; Anisimov, V. I.; Zaanen, J. Density-Functional Theory and Strong Interactions: Orbital Ordering inMott-Hubbard Insulators. Phys. Rev. B: Condens. Matter Mater. Phys.1995, 52, R5467−R5470.(62) Dudarev, S. L.; Botton, G. A.; Savrasov, S. Y.; Humphreys, C. J.;Sutton, A. P. Electron-Energy-Loss Spectra and the Structural Stabilityof Nickel Oxide: An Lsda+U Study. Phys. Rev. B: Condens. MatterMater. Phys. 1998, 57, 1505−1509.(63) Jain, A.; Ong, S. P.; Hautier, G.; Chen, W.; Richards, W. D.;Dacek, S. T.; Cholia, S.; Gunter, D.; Skinner, D.; Ceder, G.; Persson,K. A. Commentary: The Materials Project: A Materials GenomeApproach to Accelerating Materials Innovation. APL Mater. 2013, 1,011002.(64) Jain, A.; Hautier, G.; Ong, S. P.; Moore, C. J.; Fischer, C. C.;Persson, K. A.; Ceder, G. Formation Enthalpies by Mixing Gga andGga + U Calculations. Phys. Rev. B: Condens. Matter Mater. Phys.2011, 84, 045115.(65) Zeng, Z.; Chan, M. K. Y.; Zhao, Z.-J.; Kubal, J.; Fan, D.;Greeley, J. Towards First Principles-Based Prediction of HighlyAccurate Electrochemical Pourbaix Diagrams. J. Phys. Chem. C 2015,119, 18177−18187.(66) Smith, W.; Forester, T. R. Dl_Poly_2.0: A General-PurposeParallel Molecular Dynamics Simulation Package. J. Mol. Graphics1996, 14, 136−141.

ACS Nano Article

DOI: 10.1021/acsnano.9b02007ACS Nano 2019, 13, 7825−7832

7832