Nanometric water channels in water-in-salt lithium-ion battery electrolyte

Subscriber access provided by The University of Texas at El Paso (UTEP)

Article

Nanometric water channels in water-in-salt lithium-ion battery electrolyte JoonHyung Lim, Kwanghee Park, Hochan Lee, Jungyu Kim, Kyungwon Kwak, and Minhaeng Cho J. Am. Chem. Soc., Just Accepted Manuscript • DOI: 10.1021/jacs.8b07696 • Publication Date (Web): 25 Oct 2018 Downloaded from http://pubs.acs.org on October 26, 2018

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 9 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of the American Chemical Society

Nanometric water channels in water-in-salt lithium-ion battery electrolyte †,‡ *,†,‡ Joonhyung , Lim Kwanghee†,‡ Park , Hochan†,‡ Lee , Jungyu†,‡ Kim , Kyungwon Kwak , and *,†,‡ Minhaeng Cho †Center

for Molecular Spectroscopy and Dynamics, Institute for Basic Science (IBS), of Chemistry, Korea University, Seoul 02842, Korea KEYWORDS Lithium-ion battery, water hydrogen-bond, electrolyte, water network, vibrational spectroscopy, aqueous electrolyte ‡Department

ABSTRACT: Lithium-ion batteries (LIB) have been deployed in a wide range of energy-st revolutionize technological development. Recently, a lithium-ion battery that uses has been developed. However, the role of water in facilitating fast ion transport in such not been fully understood yet. Here, femtosecond IR spectroscopy and molecular dynamics like water coexists with interfacial water on ion aggregates. We found that dissolved i networks that are spontaneously intertwined with nanometric water hydrogen-bonding net through bulk-like water channels acting like conducting wires for lithium ion transpo indicate that water structure-breaking chaotropic anion salts with high propensity of f be excellent candidates for water-based LIB electrolytes. We anticipate that the pres developing aqueous LIB electrolytes.

gradient NMR, small angle neutron scatterin spectroscopy, and molecular dynamics (MD) simu that the extraordinary stability of wate Lithium-ion batteries (LIB) are one of therevealed most important energy lithium bis(trifluoromethanesulfonyl)imide ( storage systems and have penetrated deeply into our modern life from the formation of LiF-based SEI through the dec because of their desirable properties for portable electronic devices, 14. LiTFSI salt such as light weight, long cycle life, andconcentrated high power and energy 1,2 3 density . Commercial LIB’s use organic for electrolytes their Yet another important characteristic of wate electrochemical stability and formationperfomance of solid electrolyte is the unexpectedly high lithium ion 4 via interphase (SEI) the decomposition of the electrolyte on anode solutions given their high vis such electrolyte surface, where the SEI acts as a protective the layer and effectively viscosities of salt solutions increase wit prevents any side reactions due to its ion-selective permeability. Thus, a general approach when attempting to i performance LIB electrolytes is determining However, organic solvent-based rechargeable LIB’s of have ofand lithium salt to maximize io drawbacks such as high processing costs, concentration safety issues, Although recent report suggested that the fa environmental problems, mainly due to the highly avolatile, + solvation in transport is due to small the disproportionation of Li flammable, and toxic nature of organic solvents. In particular, 13, the role of water in such highly con aqueous electrolytes amounts of water under high voltages in those organic-based LIB’s salt solutions is not completely understood. H produce hydrogen and oxygen, which then pose a serious safety 5,6 femtosecond IR spectroscopy and MD simulation s risk . Consequently, vigorous investigations to find alternative 7 or polymers 8,9have been undertaken. water in water-in-salt LIB electrolytes we acces electrolytes based on solids 10,11 information on the lithium ion transport mechanis Particularly, water-based , electrolytes which are non-volatile, electrolytes. at LiTFSI salt concentrations non-flammable, and relatively human-friendly, could be aEven useful 21 m, of we LIB’s. found the existence of bulk-like and in alternative for developing the next-generation However, molecules. ions form intricate th their narrow electrochemical stability window (1.23V) The and dissolved the network structures via ion-ion electrostatic i difficulty of achieving efficient SEI formation have long been the spontaneously with the nanometric major obstacles against using aqueous electrolytes in intertwine LIB’s. bonding networks and feature nanometer radius wat Recently, Xu and coworkers found that a highly concentrated hydrated lithium ions move through bulk aqueous electrolyte can withstand a high Thus, voltage (~3.0V) and channels, which act like possess stabilities comparable to the organic electrolytes used in conducting wires for io the aqueous LIB electrolytes. commercial 12LIB’s . Subsequent13 studies using pulsed-field

Introduction

1 ACS Paragon Plus Environment

Journal of the American Chemical Society 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 9

Figure 1 | FTIR and IR pump-probe (IR-PP) of water in water-in-salt LIB electrolytes. (a) IR abs HOD in aqueous LiTFSI solutions (lower panel). Guassian fitting analysis results for aq concentrations (upper panel). (b) Time-resolved IR PP (normalized) spectra of HOD in 5 m, 10 m, The OD stretch mode frequencies (c) and the relative populations (d) of the bulk-like (blue respect to LiTFSI concentration. In the panels (c) and (d), closed symbols represent the data p open symbols from the IR-PP measurements, after the non-Condon effect was taken into accoun intensity to the relative concentration of HOD (see Supporting note 1). (e) The vibrational (red) water OD stretch modes are plotted with respect to LiTFSI concentration.

Results and Discussion

-1 can be assigned to two different HOD s 2630 cm -. Our observation that - the 2630 HOD···OH 2 and HOD···TFSI IR spectroscopy of water. Since all the previous experimental 1 band is not affected by the nature of cations (Fi 12-14 methods used, such as NMR, Raman and IR , are spectroscopy supports this band assignment, consistent with unable to monitor water dynamics in real time, no experimental which suggested that water can form strong H-bo + measurement of the hydrationions states and of their Li sulfonyl oxygen atoms of TFSI-, i.e., O-D···O=S correlation with water dynamics in water-in-salt electrolytes has To estimate the relative populations of these tw been available. IR spectroscopy of the OD stretch mode of HOD Gaussian functions molecules is an incisive approach to the study of the water H-were used to fit the FTIR spect 15-20 m LiTFSI solutions (upper panel of Figure 1a). H bonding structure in the electrolyte . Particularly, solutions oflocal each peak do not directly corr because the OD stretch frequency is highlyintegrated sensitiveareas to its 21,22 relative populations of the two water states, be environment , the spectral distribution provides indisputable strength of OD stretch information on the electrostatic environment surrounding the IR strongly depends on its loc (or frequency). Thus, the frequency-dependent tr probe, the OD group23. of HOD need to be considered to obtain the relative popul For FTIR measurements, aqueous LiTFSI (see the inset of Figure different OD species. Following the correcti 1a for the chemical structure ) solutions of TFSIwith 5 vol% HOD 26, coworkers developed by Tokmakoff and we used the were prepared. The background-subtracted FT-IR spectra of experimentally measured Raman spectra for all solutions of varying LiTFSI concentrations are shown in Figure 1a solutions considered here to convert the rela (lower panel). As LiTFSI concentration increases from 1 m to 21 amplitudes of FTIR and IR PP spectra into the relati -1 m, the intensity of IR the band 2510decreases cm and that of the (for details, see SI). It is interesting to note -1 2630 cmband increases. This notable blue-shift of the OD peak concentration (the molar ratio of water to LiTFSI frequency clearly indicates that the average H-bond number per 2.8) there is still bulk-like water featuring H-b water molecule decreases dramatically as the water H-bonding with 3 to 4 water molecules. Previously, the exis 24. As by network becomes significantly disrupted dissolved ions water in highly concentrated salt solutions was p has been observed in aqueous solutions at high salt the similarity of reduced Raman spectrum of neat 21,25 -1and concentrations , the two bands at approximately 2510 cm


Page 3 of 9 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


solution, which was obtained by carrying out optical Kerr effect symbols). Here, the IR-PP eigenspectra were no 27. measurements considering the frequency-dependent transition (non-Condon effect) In Figure 1c and 1d (closed symbols), the peak frequencies of the of the OD stretch mode, which to obtain the relative populations of HOD···OH -and HOD···OH 2 and HOD···TFSI HOD···TFSI 2 bands and their relative populations (Figure 1d, open symbols). These frequencies are plotted with respect to LiTFSI concentration. Upon increasing are in good agreement with the resul salt concentration, 2the band HOD···OH undergoes a large populations bluecarrying out two-Gaussian fitting analyses of -1, but the shift of approximately 50 cm position of the HOD···TFSI which further support our band assignments. Con band remains almost unchanged (Figure 1c) across the FTIR and that IR-PPthose analysis results, it becomes cle concentration range from 1 m to 21 m. This indicates structures and ion hydration states are not ho water molecules interacting remain with tightly TFSI H-bonded to distributed the highly concentrated LiTFSI sol anions even at relatively low concentrations. Thein relative ions at change such high concentrations have firm tende populations of the two water species gradually with 30 aggregates, e.g., ion clusters , the very or networks existence of increasing LiTFSI concentration. - is strong spectrally distinguishable HOD···TFSI evidence of Femtosecond IR pump-probe study. Although the FTIR study interfacial water on the surface of the large-sca of water in aqueous LiTFSI electrolytes shows the existence of two + and aggregates formed by electrostatic interactions separable domains with either bulk-like water or anion-bound TFSI. water, their vibrational dynamics extracted from time- and Water rotational dynamics. Controlling the relative polari frequency-resolved IR pump-probe (IR-PP) signals provides vital direction of the probe information on the local environment around the two types of water beam with respect to that o canOD measure the parallel and perpendicular IR-PP si molecules. The vibrational lifetime of each stretch mode is anisotropic signal containing informatio highly sensitive to its H-bond acceptor molecules and decay strengths. dynamics of the OD group of the HOD molecules in t Thus, water molecules in heterogeneous environments have (Figure S5). Here, anisotropic IR-PP signals in the distinctively different vibrational lifetimes, such that multi-1 were higher than 2604 cm principally considered becaus component analyses of IR-PP data provide unambiguous research present focus was information on the heterogeneous environments inon the rotational relaxation -. As can be seen in Figure 1a, t water is molecules, solutions even when the OD stretch IR spectrum broad and HOD···TFSI spectral amplitudes of the IR-PP signal from bu 28. multiple featureless and lacking any hint of components -1 are molecules at and above probe frequency of 260 The time-resolved IR-PP spectra of the OD negligible. stretch mode For in high concentration (>5 m) LiTFSI s concentrated aqueous LiTFSI solutions are two shown in Figure 1b. exponential components are needed to fit the Due to the limited spectral bandwidth of ourPP laser pulse, thetheir rotational relaxation data (Figureonly 2a) and positive 0-1 transition that corresponds toare ground state bleaching plotted in Figure 2b. The fast components of and stimulated emission contributions to the injected beam signals areprobe found to is be independent of LiTFSI c clearly visible. Figure 1b shows a complicated decay pattern whereas the slow component becomes slightly sl featuring a shift of the positive peak frequency over time.increases. There are concentration Given that the anisotr components that rapidly decay within a fewhigher picoseconds, but a -1 than 2604 arecm exclusively associated with i more slowly decaying positive peak persists even at 25 ps. The -, the water molecules, i.e., HOD···TFSI rotational relaxation latter is well-known to be a heating effect a caused by the dissipation time constant of about 1 ps in aqueous salt sol of vibrational energy associated with excited OD modes (see originate from thethe inertial and restricted rotatio -1 long-time trace of the positive in the peak IRat PP signal 2500OD cm group in HOD···TFSI - caused by its strong H-bond with of the 5 m LiTFSI solution). After removing such heating anion. The slow rotational relaxation at high LiT contributions from the raw data (for details, see SI), isotropic with the large-scale 31 (> 5 m) can the be associated ro IR-PP signals were used to obtain the vibrational lifetimes of theof -to from one H-bonding site a another given TFSI site of the sam -. Here, it OD stretch modes of HOD···OH 2 and HOD···TFSI - (for or a neighboring TFSI details, see SI). As will be should be noted that the background-corrected isotropic IR-PP below, the rotational relaxation time constant spectra of 1 m LiTFSI solution cannot be decomposed into two that of the spectral diffusion time constant of - is decaying components because the fraction of HOD···TFSI -, suggesting a strong correla stretch mode of HOD···TFSI negligible. between OD frequency fluctuation dynamics and O For solutions with LiTFSI concentrations from 5 m to 21 m, relaxation of the water molecules at the interface or isotropic IR-PP signals at each probe frequency could be easily the bulk-like water and ion aggregate domains. fitted with the two exponentially decaying components (Figure Interestingly, theS3). two rotational relaxation t The probe frequency-averaged lifetimes of the vibrationally excited interfacial water molecules do not strongly dep - are found to OD stretch modes of HOD···OH 2 and HOD···TFSI concentration. This can be understood by noting be 1.9 ps and 9 ps, respectively, regardless of LiTFSI strongly interacting - undergo with water molecules an TFSI concentrations (Figure 1e). This is important evidence supporting 32, approximately 1 ps wobble-in-a-cone-type rotati the notion that the H-bond acceptor of the fast (1.9 component which is ps) then followed by a large amplitude rotat OD group is a second water molecule and that the acceptor of the with relaxation time of approximately 17 ps. This -. Since slow (9 ps) component is TFSI the OD stretch lifetime in another observation that the relative anisotropy 29 neat water is about , our 1.7 1.9 psps component originates from fast (~1 ps) and slow (17 ps) components do not dep those HOD molecules that are H-bonded to surrounding water concentration either. molecules. Two-dimensional IR spectroscopy of water. The time- and The two-component analyses of the isotropic IR-PP signalsIR-PP signals provide the spec frequency-resolved additionally provide probe frequency-dependent amplitude spectra contributing to the time-d of the IR oscillators - (Figure S4). (or eigenspectra) of HOD···OH HOD···TFSI 2 and absorption and their dynamical evolutions in rea The peak frequencies of the two IR-PP eigenspectra corresponding new dimension, excitation pump frequency, allo to HOD···OH are plotted in Figure 1ctwo-dimensional (open 2 and HOD···TFSI IR (2D-IR) measurements (in pu



frequencies) and provides extra information on the dynamical correlation of an IR oscillator selected (tagged) by a particular pump frequency component with that by the same or different probe frequency component at a 33-40 waiting time later.


Page 4 of 9

Page 5 of 9 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure 2 | Rotational and spectral diffusion dynamics of water in water-in-salt LIB electrolyt (= 2(t) r(t)/0.4), from anisotropic IR-PP signals (red squares) and MD simulations (black solid line) solution are plotted with respect to pump-probe delay time. The frequency-frequency correlat line slope of the time-resolved OD stretch 2D-IR spectra of 21 m LiTFSI electrolyte solution is the experimentally measured rotational correlation function. (b) The rotational correlation bi-exponential function and the fast (closed black squares) and slow (closed red circles) d LiTFSI concentration. Here, the bi-exponential decay constants of the nodal line slope (or F representative 2D-IR spectra after waiting times of 0.6 ps, 2.4 ps, 6.5 ps, and 10 ps, for 21 m Li figure. The red line in each 2D-IR spectrum corresponds to the nodal line of which slope is app and horizontal dotted lines on each 2D-IR spectrum correspond to the peak freqeuncies 2 and of the OD -. Although no notable cross peaks indicating chemical exchange processes between b HOD···TFSI water channels and interfacial water molecules interacting with ion networks are found in th change in time due to the difference in lifetimes of the two water states.

is principally determined by the signal from in The excitation frequency-emission frequency correlation molecules. Interestingly, even after 7 ps the 2 decreases over time due to the stochastic nature of solute-solvent interaction at thermal equilibrium states.exhibit a diagonally elongated shape (Figure S heterogeneous environment around each individu Four representative 2D-IR spectra of 21 m LiTFSI solution after water molecule. four different waiting times are shown in Figure 2c. The upper We carefully examined positive peak centered on the fundamental transition frequency ofthe entire set of 2D-IR spe not bleaching find any increase of cross peaks over time. Th the OD stretch mode results from ground-state and suggests that the the chemical exchange proces stimulated emission contributions to the 2D-IR signals, whereas - take longer than our experimen HOD···OH bottom negative peak originates from the excited state absorption 2 and HOD···TFSI time span 16 ps, of which upper limit is dete of the probe beam due to vibrational transition from theof first to the lifetime of the overtone state of the OD mode. The negative peak position isslow lowercomponent. This means that bulk-like) water along the emission frequency y-axis in Figure 2c due to themolecules remain as interfaci -(HO) for water molecules interacting with atTFSI least 16 ps. Ji overtone anharmonicity. 2 41 measured such water chemical exchange time et al. At short waiting times (e.g., ps), the T 2D-IR spectrum is w = 0.6 -, where X represents other salt a HOD···X 2 andtoward elongated along at diagonal line and has a HOD···OH long tail low and found that the exchange time is about 6 ps. Ho frequency side. This tail corresponds to the 2D-IR signal from bulkthat we were not able to observe any water chemical like water molecules, of which the average 2, HOD···OH peaks that2Dthe interaction between wa vibrational lifetime is short (1.9 ps). Thus, the indicates low-frequency -is sufficiently strong enough to make the H-b TFSI -1disappears rapidly so IR spectral feature = = 2550 at  cm that 

m

the 2D-IR spectrum at longer waiting times (> a few picoseconds)



Page 6 of 9

Figure 3 | MD-simulated structures of water H-bonding and ion networks. Snapshot structures of solutions are shown. The structures of the ion network and water H-bonding network are shown in in a and b. A slab of MD simulation box of 21 m LiTFSI electrolyte solution is shown in c, where and ion aggregate, respectively. The water channel diameter is estimated to be 1.1 nm on average grey balls in d represent the same lithium cation in the bulk-like water region at four sequent ion is highly mobile through the bulk-like nanometric water channel. However, over this tim network does not move but just fluctuates at the same position. Furthermore, a small fraction ion networks and they show no notable mobility over the 3 ps in time frame.

45, the us three-dimensional structures of wate longer than 16 ps or that those interfacialby water molecules are highly concentrated salt solutions are intricate relatively confined to the region close to the ion aggregates detailed spontaneously formed in the high concentration (21 morphology m) LiTFSI of ion aggregates, which is intrinsic ion characteristics, such as charge dens solution because each TFSI- anion has multiple H-bond acceptors strongmolecules Coulomb interaction sites, and hydrophobi that are always available for interfacial water to find 30,46 47 30,48-50 - orthe outTFSI femtosecond ,IR-PP NMR , MD , and graphanother H-bonding partners within in a same -. TFSI theoretical studies of a variety of highly concen neighboring we showed chaotropic anions at high concent One important piece of spectral information that canthat be extracted ion networks intertwined with water networks from the 2D-IR lineshape analysis is the time-dependent nodal line anions of low aqueous solubility te 42,43 slope (NLS) , the slope of the line separatingkosmotropic the upper positive crystalline clusters which exclude water mol peak and the lower negative peak. The NLS with respect to ion waiting 45. MD simulation results often strongly dep time is directly related to the OD stretch insides frequency-frequency of following the force-field parameters used. Thu correlation function and reflect the rate ofaccuracy memory loss quantitative validity of the parameters, we fi the initial excitation (pump) frequency after a given finite waiting properties, e.g., ion self-diffusio period. The fitted nodal line (red) is shown transport in Figure 2c, and its conductivities details, see SI), and rotati slope is plotted with respect to waiting time in Figure 2a (for (blue 51 and anisotropic signal of water (HOD) directly molecule circles). The experimental NLS data can be fitted toIR-PP a bicompared them with the experimental results. In exponential function with fast (1.1 ps) and slow (13.4 ps) MD-simulated anisotropy (black solid line) of components. Surprisingly, these two decaying components have -, is plotted and found to be in excellen HOD···TFSI the same time scales as the two anisotropy i.e., decaying components with the experimental (Figure 2b). This is not a coincidence because the two rotational result from the IR-PP da -1. higher than 2604 relaxation processes, inertial wobble-likefrequencies rotational fluctuation and cm In addition, theHquantitative agreements betwe slow rotational diffusion due to jumps between the water ion diffusion coefficients and conductivities f bonding sites-,of can TFSI correspond to the two sources of OD 44 LiTFSI solutions from experimental data indicate stretch frequency fluctuation or .modulations Thus, the in time 2D-IR spectroscopic results also support fields the two-state water are quantitatively acceptable. Figure 3a a snapshotrelaxation structures of hypothesis and the interpretations of rotational andLiTFSI solution (Figure S7 (red), and involving water network (blue) at two different co frequency fluctuation dynamics based on the model -) water molecules closely associated interfacial (HOD···TFSI with m and 21 m, respectively. As LiTFSI concentratio ion aggregates or networks. intricate entanglements between water H-bonding electrostatic interaction networks becomes man Molecular dynamics simulations. FTIR and femtosecond sample-spanning ion aggregate formation can be v nonlinear IR studies all suggest that bulk-like H-bond-networking 45. A slab of the snapshot structur percolation water co-exist with interfacial water thatthe interacts with phenomena domainm LiTFSI solution To ingain Figure 3c exhibits water chan separated ion aggregates in water-in-salt LIB electrolytes. S8 and S9 for more water a deeper in-sight into the morphological aspects of both water H-channel structures and type networks serve as a porous framework providing ope bonding networks and ion aggregates, MD simulations of LiTFSI through which water can flow, which was assumed pre solutions were performed (for details, see SI). As has been shown


Page 7 of 9 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


the thorough analysis of macroscopic transport data within(4) wide these effects. Anions should be able to form i 52,53 range of salt concentration . The average diameter of water lithium ions at high concentrations so chaotro channels is estimated to be about 1.2 nm, andmultiple this is believed to be electrostatic interaction sites are the conducting wire for hydrated lithium ion transport when an aqueous LIB electrolytes. We therefore anticip externally applied electric field is present in a real LIB. Figure 3d experimental and simulation results providing gu shows a mobile lithium (grey) ion at four sequential 1 ps steps for developing, less toxic and volatile, environ through a bulk-like water channel, though one lithium ionwater-based in the ion LIB’s in the future. nonflammable network does not move (Figure 3e). To support the fast lithium ion transport through nanometric water channels, we calculated Experimental Section the diffusion coefficient of lithium ions in bulk-like water and IR Pump-Probe and domain Two-dimensional IR spectroscopy. that in ion network domain separately and found that the rate of Details of the IR pump-probe setup and 2D-IR measu lithium ion diffusion in those water channels is almost 8 times described elsewhere. Briefly, an 800 nm laser puls faster than that in ion networks (see Supporting note 3 and Table by a laser system consisting of a Ti:sapphire osc S3). Spectra physics) and a regenerative amplifier(S 54 studied Recently, Yamada et al.a mixed lithium salt electrolyte physics). With the BBO crystal based OPA system, solution, an eutectic mixture, where LiTFSIpulse is mixed with lithium is divided into two near IR pulses (~1.3 and -1 centered bis(pentafluoroethanesulfonyl)imid (LiBETI) 7:3 molar thenwith obtained the ratio. desired 2550 cmmid-IR pulse by Interestingly, even though the solubilitynonlinear limit of LiTFSI in water isthrough DFG phenomena the AgGaS The 2 crystal. about 22 m, that of BETI LiTFSI is as high as 28 m in water. 0.7 0.3 generated MIR pulse has a ~70 fs pulse duration and Then, the average number of water molecules either selective experiment, the MI Insolvating the polarization LiTFSI or LiBETI is just about 2 so that this hydrate-melt does not into a probe pulse (~ 100 nJ) and a pump pulse (~ 90 have sufficient water molecules that canZnSe form beam bulk-like water splitter(9:1), and the repetition rate channels for lithium ion transport. Consequently, at to such was adjusted 500 high Hz using an electronically syn concentration (~28 BETI m LiTFSI solution), it was suggested 0.7 0.3 chopper. Two wire grid polarizers are placed before that the lithium ion transport mechanism can be dominated by abeam. After the interaction b sample in the probe hopping-type process. Their Raman scattering spectrum shows a pulses and the sample, the probe pulse whose po single O-H band as the mixed lithium salt concentration increases parallel (or perpendicular) with respect to the pu up to 28 m, which was believed to be strong evidence supporting alternatively selected by the motorized polariz their conclusion. On the other hand, in the probe 21 m LiTFSI pulseaqueous was spread spectrally by monochromato solution considered here, where the salt-to-solvent molar array ratio MCT is detector. In the case by 64 element about 1:2.7, there exist sufficient amountmeasurement, of water molecules that three parallel polarized k ) pulses of (k 1, 2kand 3 can participate in and form bulk-like watersimilar channels for vehicleintensity were shed on the sample in boxtype ion transport, which is consistent with our IR the PP data showing create echo signal. The intensity of that the co-existence of two different water states. Nonetheless, it would enhanced by local oscillator pulses and the hete be really interesting to carry out both IR PP and 2D-IR by studies of detector. detected MCT-array the mixed salt xBETI (LiTFSI solutions with varying 1-x) concentrations to further examine the possibility a gradual ASSOCIATEDof CONTENT transition of ion transport mechanism from the vehicle-type process Supporting Information. to the hopping-type process with increasing salt concentration. The Supporting Information is available free of Hypothesis, summary, and a few concluding Internet http://pubs.acs.org. at Supporting informat methods for sample preparation, experimental se remarks. resolved infrared spectroscopies (IR-PP and 2D-I Although Xu and coworkers showed that highly concentrated simulations. It also provides the details of an aqueous LiTFSI solution can be an excellent choice of water-based jump rotation of interfacial water molecules as LIB electrolyte, the interplay of ion transport dynamics and the 3D local heating effect. (PDF) morphological characteristics of water networks and ion aggregates is unclear. A few critical questions that have not been addressed AUTHOR INFORMATION are: (1) what is the functional role of water in aqueous LIB electrolyte?, (2) why anion does the work TFSI particularly well?, Author Corresponding (3) why is a high concentration of LiTFSI needed, or is there an [email protected] (MC) or optimum salt concentration for aqueous LIB*E-mail: electrolyte? and (4) [email protected] what types of counter anions are suitable for aqueous LIB Notes electrolyte? Based on the results of this work, we hypothesize that The authors declare no competing financial intere nanometric water channels intertwined with ion networks promote lithium ion conduction, in which counter anions should be capable ACKNOWLEDGMENT of forming 3D extended ion networks, rather than clusters, through This wasto supported by IBS-R023-D1.. interacting with lithium ions. This helps us towork answer the above questions. (1) Water has a dual role; bulk-like water in ion transport REFERENCES channels within porous ion network grids plays the active role as a medium for lithium ion transport, whereas interfacial wetting water J.-M., Nature 2008, 451, 652-657 (1) Armand, M.; Tarascon, -anions on ion networks behaves as a lubricant. (2) Chaotropic TFSI J.-M.; Armand, M., Nature 2001, 414, 359-36 (2) Tarascon, (3)provide Xu, K.,aChem. Rev. 2004, 104, 4303-4417. at high concentration form ion networks, which porous (4) Verma, P.; Maire, P.; Novak, P., Electrochim. Acta 201 framework for nanometric water channels. (3) High concentrations 6332-6341. of LiTFSI salt is needed for stable SEI formation and reducing (5) is Wang, Q.; Ping, P.; Zhao, X.; Chu, G.; Sun, J water electrolysis, but low salt concentration preferred for high Power Sources 2012, 210-224. ion conductivity. Thus, the optimum concentration has208, to balance



Page 8 of 9

(6) Lu, L.; Han, X.; Li, J.; Hua, J.; Ouyang, (29) M., J.Rezus, Power Sources Y. L. A.; Bakker, H. J., J. Chem. Phys. 2006 2013, 226, 272-288. 144512. (7) Wang, Y.; Richards, W. D.; Ong, S. P.; Miara, L. J.; Kim, C.;H.; Choi, J.-H.; Cho, M., J. Chem (30) Kim, S.;J. Kim, Mo, Y.; Ceder, G., Nat. Mater. 2015, 14, 1026-1031. 141, 124510. (8) Bruce, P. G.; Scrosati, B.; Tarascon, J.-M., Angew. Chem. Int. (31) Laage, D.; Hynes, J. T., Science 2006, 311, 832-835. Ed. 2008, 47, 2930-2946. (32) Moilanen, D. E.; Fenn, E. E.; Wong, D.; Fayer, (9) Meyer, W. H., Adv. Mater. 1998, 10, 439-448. Chem. B 2009, 113, 8560-8568. (10) Kim, H.; Hong, J.; Park, K. Y.; Kim, H.; Kim, S.Hamm, W.; Kang, K., M. H.; Hochstrasser, R. M., J. Ph (33) P.; Lim, Chem. Rev. 2014, 114, 11788-11827. 1998, 102, 6123-6138. (11) Wang, X.; Hou, Y.; Zhu, Y.; Wu, Y.; Holze,(34) R., Sci. Rep. 2013, Cho, M., Chem. Rev. 2008, 108, 1331-1418. 3, 1401. (35) Hamm, P.; Zanni, M., Concepts and Methods of 2D Infrared (12) Suo, L. M.; Borodin, O.; Gao, T.; Olguin, M.; Ho, Cambridge J.; Fan, X. Spectroscopy; University Press: New York, 2011. L.; Luo, C.; Wang, C. S.; Xu, K., Science 2015, 350, 938-943. (36) Mukamel, S., Annu. Rev. Phys. Chem. 2000, 51, 691-729. (13) Borodin, O.; Suo, L.; Gobet, M.; Ren, X.; (37) Wang, F.; Faraone, Ganim, Z.; Chung, H. S.; Smith, A. W.; Deflor A.; Peng, J.; Olguin, M.; Schroeder, M.; Ding, M. S.; Gobrogge, A., E.; Acc. Chem. Res. 2008, 41, 432-441. K. C.; Tokmakoff, von Wald Cresce, A.; Munoz, S.; Dura, J. A.; Greenbaum, S.; Wang, (38) Tian, P.; Keusters, D.; Suzaki, Y.; Warren, W. S C.; Xu, K., ACS Nano 2017, 11, 10462-10471. 300, 1553-1555. (14) Vatamanu, J.; Borodin, O., J. Phys. Chem. Lett. 2017, 8, 4362(39) Zheng, J.; Kwak, K.; Asbury, J.; Chen, X.; Pil 4367. M. D., Science 2005, 309, 1338-1343. (15) Bakker, H. J.; Skinner, J. L., Chem. Rev. 2010, 110, 1498-1517. (40) Lee, K.-K.; Park, K.; Lee, H.; Noh, Y.; Kosso K.; M.; Cho, M., Nat. H. Commun. (16) Tielrooij, K. J.; Garcia-Araez, N.; Bonn, Bakker, J.,2017, 8, 14658. Science 2010, 328, 1006-1009. (41) Ji, M. B.; Odelius, M.; Gaffney, K. J., Science 20 (17) Fournier, J. A.; Carpenter, W.; De Marco,1005. L.; Tokmakoff, A., J. Am. Chem. Soc. 2016, 138, 9634-9645. (42) Kwac, K.; Cho, M., J. Phys. Chem. A 2003, 107, 5903-5912 (18) Fecko, C. J.; Eaves, J. D.; Loparo, J. J.; Tokmakoff, A.; (43) Cho, M., Two-Dimensional Optical Spectroscopy; CRC press: Geissler, P. L., Science 2003, 301, 1698-1702. Boca Raton, 2009. (19) Khatib, R.; Backus, E. H.; Bonn, M.; Perez-Haro, M. J.; (44) Laage, D.; Hynes, J. T., Chem. Phys. Lett. 2006, 433, 80 Gaigeot, M. P.; Sulpizi, M., Sci. Rep. 2016, 6, 24287. (45) Choi, J.-H.; Lee, H.; Choi, H. R.; Cho, M., A (20) Sovago, M.; Campen, R. K.; Wurpel, G.Chem. W.;2018,69, Muller, M.; 125-149. Bakker, H. J.; Bonn, M., Phys. Rev. Lett. 2008, 100, 173901. (46) Kim, H.; Park, S.; Cho, M., Phys. Chem. Chem. Phys. 2012 (21) Omta, A. W.; Kropman, M. F.; Woutersen, S.; Bakker, H. J., 6233-6240. (47) Dillon, S. R.; Dougherty, R. C., J.Phys. Chem. A 2 Science 2003, 301, 347-349. (22) Smith, J. D.; Saykally, R. J.; Geissler,10217-10220. P. L., J. Am. Chem. Soc. (48) Choi, J.-H.; Cho, M., J. Chem. Phys. 2014, 141, 154502. 2007, 129, 13847-13856. (23) Fayer, M. D., Acc. Chem. Res. 2012, 45, 3-14. (49) Choi, J.-H.; Cho, M., J. Chem. Phys. 2015, 143, 104110 (24) Auer, B.; Kumar, R.; Schmidt, J. R.; Skinner, J.Choi, L., Proc. Natl. Cho, M., J. Chem. Phys. 2016, 144, 204126. (50) J.-H.; Acad. Sci. U. S. A. 2007, 104, 14215-14220. (51) Lin, Y. S.; Pieniazek, P. A.; Yang, M.; Skinne (25) Moilanen, D. E.; Wong, D.; Rosenfeld, Phys. D. E.; E. E.; 2010,Fenn, 132, 174505. Fayer, M. D., Proc. Natl. Acad. Sci. U. S. A. 2009, 106, 375-380. (52) Angell, C. A.; Bressel, R. D., J. Phys. Chem. 1972, 76 (26) Loparo, J. J.; Roberts, S. T.; Nicodemus,3253. R. A.; Tokmakoff, A., (53) Angell, C. A., Chem. Rev. 2002, 102, 2627-2650. Chem. Phys. 2007, 341, 218-229. (27) Turton, D. A.; Corsaro, C.; Martin, D. F.; Mallamace, F.; (54) Yamada, Y.; Usui, K.; Sodeyama, K.; Ko, S.; Wynne, K., Phys. Chem. Chem. Phys. 2012, 14, 8067-8073. Yamada, A., Nat. Energy 2016, 1, 16129. (28) Piletic, I. R.; Moilanen, D. E.; Levinger, N. E.; Fayer, M. D., J. Am. Chem. Soc. 2006, 128, 10366-10367.


Page 9 of 9 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


ACS Paragon Plus Environment

9

Nanometric water channels in water-in-salt lithium-ion battery electrolyte

Recommend Documents