Corona Yank in Edged Cathode Particles for Li-Ion Batteries

Feb 18, 2016 - In a heuristic endeavor based on electrostatics-dynamics considerations, an attempt has been made to explain the significant enhancemen...
2 downloads 7 Views 2MB Size
Article pubs.acs.org/Langmuir

Corona Yank in Edged Cathode Particles for Li-Ion Batteries Jagannathan Rangarajan* Functional Materials Division, CECRI-CSIR, Karaikudi-630006, Tamil Nadu, India ABSTRACT: In a heuristic endeavor based on electrostatics-dynamics considerations, an attempt has been made to explain the significant enhancement (∼20%) observed in initial capacity for the as-synthesized, sharp-edged cathode particles of lithium nickel manganese cobalt oxide, and lithium nickel manganese oxide cathode systems which are considered here as illustrative examples. Also, the impressive capacity retention (∼90%) observed for the milled blend of these, comprising smooth surfaced particles with cells made using respective cathode particles, has been explained. Enhancement in initial capacity observed for the edged particles can be squarely attributed to a corona-gain process. On the contrary, impressive capacity retention possible for the milled blend (of near-spherical particles) can be rationalized in terms of the absence of predominant energy drain encountered while shuttling charges across edges of the as-synthesized cathode particles, a process described as corona yank

1. INTRODUCTION Over the years, the Li-ion battery community, with the goal of achieving high-capacity rate capability coupled to capacity retention, safety issues, and so on, has been endeavoring through different methods such as ingenious choice-combination of cathode materials/systems, tuning material (cathode) properties, improved battery design, and so on.1−3 In particular, for achieving cathode materials with desired material properties, these are synthesized through numerous novel syntheses strategies, namely, reverse micelle process and molten salt synthesis to cite a few, yielding cathode particles of desired morphologies such as desert rose morphology having high surface area,4 platelet morphology,5 rod-like particles6 and, more notably, spherical particles.7 Also, the importance of the shape of cathode particles, in particular, particles having spherical shape coupled to high tap density, has been amply demonstrated by other researchers.8 It is now well established that the composite cathode materials prepared as milled blend or through other chemical synthesis routes have the unique advantage of offering both high capacity and good thermal stability. There have been several efforts underlying the importance of composite cathodes in terms of high capacity and thermal stability. Some significant contributions, as illustrative examples, are Li[(Ni0.8Co0.1Mn0.1)0.8(Ni0.5Mn0.5)0.2]O2 core−shell structure through coprecipitation method;9 advantages of blending LiNi0.8Co0.2O2 into Li1+xMn2−xO4 cathodes;10 composite cathodes based on blends of LiMn2O4 and LiNi0.8Co0.2O2;11 a mixture of Li−Mn spinel and Li−Ni−Co−Mn oxide as a positive electrode material for high-temperature storage;12 the mixed electrode (Li1.1Mn1.9O4:LiNi0.8Co0.15Al0.05O2 in 1:1 weight ratio as a promising positive electrode;13 and intergrown LiNi0.5Mn1.5O4·LiNi1/3Co1/3Mn1/3O2 composite nanorods as high-energy density cathode materials.14 © XXXX American Chemical Society

Notwithstanding impressive electrochemical performance achieved with these cathode materials, a clear perception of the exact mechanism leading to such high capacity still seems to be a daunting task, with little agreement on different mechanisms proposed. In the backdrop of this, we recently proposed a more plausible mechanism based on electrostatic considerations that can give precise clues on realizing high initial capacity,15 and subsequently we adduced some experimental evidence of a fade-control mechanism possible for the milled spherical particles.16 It seems reasonable to hypothesize that cathode particles with sharp edges stemming from polyhedral, faceted particle morphology have the advantage of realizing higher initial capacity over cathode particles having smooth surface. This generalization seems to gain credence with many other similar results, reporting higher capacity for cathode particles having sharp edges in comparison with (near) spherical particles.17,18 In the same breath, we should make note of an important observation that edged particles do exhibit the pronounced propensity to fade rather faster and more drastically over spherical or near-spherical milled particles having no edges. In this context, it is significant to make comparison with the interesting work of Zhu et al.19 in convincingly demonstrating the importance of particle morphology for achieving initial capacity and capacity retention for spinel LiMn2O4 system. In this scenario, we consider that a heuristic approach based on electrostatic considerations would be more purposeful here. Received: January 29, 2016 Revised: February 4, 2016

A

DOI: 10.1021/acs.langmuir.6b00336 Langmuir XXXX, XXX, XXX−XXX

Article

Langmuir

Figure 1. Particle morphologies using field-emission scanning electron microscope images of (a) LNMC as-prepared, (b) LNM as-prepared, and (c) milled blend of LNMC and LNM component in 3:1 ratio. Broken arrows (in red) indicate the presence of sharp edges and facets in as-synthesized cathode particles. The corresponding left side images and sketches illustrate the representative single-particle image and expected charge/electrontransport process.

2. EXPERIMENTAL SECTION

copy (SEM) images, we have that the as-synthesized LNMC and LNM components having sharp edges arising from faceted morphology of the cathode particles with their mean particle size being 200 and 150 nm, respectively (Figure 1a,b). These Voronoi particles, henceforth referred to as Voronoi cells, having defined edges and tips are expected to be several orders more active for charge transport through diffusion, also even through drift over the bulk.21 Although in a given electrochemical system charge/ion-transport, movement, and potential eventually determine charge−discharge capacity, cell voltage and maintenance have crucial dependence on the chemical system of the cathode materials. Further augmentation in cathode performance becomes feasible through effective physical mixture of two chosen cathode systems so that preferred electrochemical features can be harnessed for gainful applications. Bearing this important point in mind, a blend of LNMC and LNM as-synthesized cathode particles was made through comminution using a pulverizer. In this process there is a considerable reduction in particle size (∼75 nm) accompanied by smoothening of the edges of these cathode particles (Figure 1c). Although the synthesis of composite cathode materials through comminution apparently looks simple, consequences following comminution of the cathode particles are manifold: (i) Reduction in particle size accompanied by change in particle morphology can bring about drop in electrical resistivity and

This analysis in large part centers around charge−discharge electrochemical capacity results of lithium nickel manganese cobalt oxide (LNMC) and lithium nickel manganese oxide (LNM) cathode particles having regular polyhedral morphologies with defined edges synthesized through sol−gel thermolysis, as previously described by our group.15 Also, these results are compared with those of the composite cathode materials made as blend using LNMC and LNM components taken in the preoptimized ratio of 3:1 obtained through intense pulverizing using a rotor ball mill (for 1 h @ 400 rpm). Obviously, in the milled particles, the particle size gets reduced considerably through comminution by a factor of 3 to 4 with most of the individual particles lacking clear definition of edges. Also, with the goal of getting a general perception of the role of morphology of cathode particles vis a vis electrochemical capacity and retention, this report takes into consideration similar results on electrochemical capacity correlated with a gamut of morphologies of cathode particles available in open literature.

3. RESULTS AND DISCUSSION 3.1. As-Synthesized Cathode Particles and the Blend through Comminution and Defect Chemistry. The synthesis of LNMC and LNM cathode particles essentially based on a facile sol−gel thermolysis followed by postcalcination at 900 °C results in particles having polyhedral morphology (Figure 1a,b), consistent with Voronoi distribution20 From Figure 1 (left panels) depicting the particle morphology through field-emission scanning electron microsB

DOI: 10.1021/acs.langmuir.6b00336 Langmuir XXXX, XXX, XXX−XXX

Article

Langmuir also totally new electrostatic, dynamic conditions. (ii) Introduction of new defect centers that can either aid (through excitons) or hamper (defects generated acting as energy sink) can totally alter the complexion of defect chemistry of the blend. (iii) There is enhanced mobility for charge transport (of electrons and ions) and better surface reactivity. All in all, the immediate consequence of particle size reduction is shortening of the Li ion diffusion length and also decrease in the antisite defect, leading to increase in the Li ion conductivity.22 Furthermore, it is pertinent to note that Gaberscek et al.23 found that the discharge capacity of LiFePO4-based electrodes shows linear drop with average particle size, having little dependence on carbon coating. Also, they observed that the electrode resistance, Rm, varies as a square of particle size, and it also can be generalized that the capacity decreases linearly with the increase in particle size (iv) More significantly, it is reasonable to expect that in the comminuted cathode blend particles obtained through application of local stress on the particles/crystals there is large scope for the introduction of edge dislocations (represented by ⊥) and also local defects (represented by °) in the form of vacancies. These will become electrochemically active under charge−discharge conditions, eventually aiding the augmentation of electrochemical capacity. These foregoing points can be consolidated and can be collectively represented as an illustrative model diagram, as will be discussed in the following sections. 3.2. Particle Morphology vis a vis Electrostatic and Dynamic Considerations. From Figure 1 (left panels) depicting the particle morphology through field-emission SEM images, we have found that the as-synthesized LNMC and LNM samples have sharp edges arising from faceted morphology of the cathode particles, while in the blend obtained after milling these edges get smothered out, resulting in smooth, curve-headed particles. The blend, predominantly comprising more spherical particles with sharp edges of particles, have been lost as a result of milling. In terms of electrochemical performance, as revealed from charge− discharge cycle studies, an impressive capacity retention/fade control is possible with the smooth-surfaced milled blend (Figure 2, curve c), while the as-prepared LNMC and LNM components having sharp edges showed high initial capacity, followed by pronounced capacity fade (∼20−35%). For explaining the present results let us first start from the celebrated Maxwell’s electrodynamics equation. Accordingly, the electric-field intensity E turns out to be a gradient of the electric potential V, the net work done in transporting charges over the given surface.

E f = ∇V

Figure 2. Charge−discharge cycle profiles (capacity vs cycle number performance) of as-prepared (a) LNMC, (b) LNM cathode components, and (c) milled LNMC+LNM (75:25) blend cathode materials ΔQLi fade/drop in discharge capacity upon cycling is indicated using a negative sign. Also, observed edge-induced gain in initial capacity is indicated using a bold arrow.

as-synthesized particles having defined edges and facets, the charge-transport processes comprise multiple components involving different factors, while for the milled spherical particles the charge-transport process is isotropic and little work needs to be done in moving charges over the spherical surface because the sphere happens to be an equipotential surface with potential gradient turning out to be zero. On the contrary, charge transport over facet(s) of cathode particle is bound to be totally different from that across edges and tips of the particles. Hence as-synthesized particles are bound to exhibit electric field comprising different electric-field intensity components arising from n faces and 2n edges, but this will be reduced to a single isotropic contribution in the milled particles, as can be visualized through Green−Stokes theorem vis a vis Gauss’s divergence theorem.24 The electric-field intensity (Eedge) from the polyhedral particles is much higher than that of spherical particles Espherical for the simple reason that electric-field contribution from edges of the as-synthesized particles is much stronger than that of the spherical surface of the particles. It turns out that for the as-synthesized cuboid particles there are 6 faces and 12 edges, which upon smudging through milling lose all faces and edges, eventually becoming reduced to a sphere, that is nEfaces + 2nEedge > > Espherical

(2)

For the cuboidal particles, the total electrical-field intensity is the linear sum of electric-field intensity contributions from both faces and edges

(1)

In explaining the working of Li-ion batteries, charge-transport considerations become more appropriate while describing intercalation/deintercalation of electrons and Li+ ions during charge−discharge cycles over cathode particulate surfaces, which are limited to electrode−electrolyte interface regime. By viewing the intercalation−deintercalation process microscopically or even confining our attention to single particles, there is bound to be a drastic difference between polyhedral or Voronoi LNMC and LNM particles versus the milled LNMC+ LNM blend having spherical particles in terms of electrostatic potential and allied charge-transport processes. For the former,

Ecuboid =

∑ 6 ∬ Efaces ds + 12 ∮ Eedges dl

(3)

with their respective field intensity values given by their potential gradient term determined by the operator ∇(∂/∂x + ∂/∂y)Vfaces and so on, showing significant, finite values. However, in the case of milled (spherical) particles, the field intensity becomes zero Emilled = C

∭ ∇V = 0,

∵ ∂V / ∂x = ∂V / ∂y = 0

(4)

DOI: 10.1021/acs.langmuir.6b00336 Langmuir XXXX, XXX, XXX−XXX

Article

Langmuir

intercalation process can be limited to the surface of the particle. In a typical intercalation process, the flux density of the intercalation process20 can be given by

This, in turn, suggests little cost in energy, eventually determining capacity, in contrast with that encountering sharp edges. To explain higher initial capacity observed for the edged particles, it is pertinent to invoke the relation between charge density vis a vis curvature of a given conducting (also dielectric) surface explained through electrostatic considerations. It is well established that electrical charges have profound propensity to crowd along edges/tips of a conductor having steep curvatures over a flat surface for the simple reason that charge density bears direct relation with curvature.24 Hence it is quite logical that the unmilled cathode particles having sharp edges show higher initial capacity over milled blend, having near-spherical particles with their edges smothered out as a consequence of milling. For milled spherical particles, the curvature happens to be reciprocal of radius (half the particle size) of the spherical particles. Also, the particle morphology of unmilled edged particles can be analogized to Voronoi cells, and hence the expression for curvature25 can be given by kij = 2ni ·(pi − pj )/|pi − pj |2

J = −Dgb∇gb

(6)

highlighting the importance of the shape of the grain boundary. It is significant to note that in the case of sharp edges, by virtue of the steep, pronounced curvature of the edges, electrical charges have the pronounced tendency to become crowded in the vicinity of steep edges (having high curvature) over smooth surface. This may explain the higher (∼20%) initial capacity realized for the as-synthesized cathode sample. This situation can be analogized to the working of sharp tips/edges of a lightning conductor and ignition spark-plug acting as a spout for the copious supply of electrical charges, thus accounting for higher initial capacity of the as-synthesized cathode particles. Notwithstanding the higher initial capacity realized for the sharp-edged particles upon cycling, especially during discharge condition, the (de)intercalating charges acquire or are expected to acquire a more complex picture so that there would be a steep field intensity gradient for the charge transport between plain surface and edges of the cathode particles. In particular, shuttling between edges and the plain surface of the given cathode particle would entail energy loss, so as to be consistent with the partial derivative continuity equation conserving linear momentum26

(5)

suggesting a many fold increase in curvature and hence eventually paving the way for high charge density, in turn, leading to higher initial capacity, as schematically illustrated in Figure 3. 3.3. Charge Transport During (De)intercalation and Capacity Fade. Sharp edges of the cathode particles constituting grain boundary surface may play a very dominant role over bulk in terms of charge transport, especially during the intercalation process Hence the discussion pertaining to

∂g mech /∂t + ∂g field /∂t + ∇T = 0

(7)

with g and T being linear momentum density and flux tensors, respectively. This would describe the complex scenario of the trajectories of itinerant charges experiencing drastic change under field conditions. In particular, at the edges or sharp points there will be pronounced shooting of field, which eventually predicts much energy loss for the charge transport or while shuttling across edges during deintercalation. This could eventually result in energy loss and hence capacity fade for the sharp-edged particles. On the contrary, in the case of milled blend having particles of smooth spherical shape, such energy drain can be expected to be totally absent, warranting little energy drain. To illustrate the extent of energy loss when charges (both e− and Li+) encounter edges during the (de)intercalation process, we can estimate an energy loss of 0.32 × 10−19 eV for a single event of an electron encountering an edge of a cathode particle having an edge of 200 nm, with a nominal diffusion length (Ldiff) of 10 nm.27 Accordingly, the total energy loss due to edges or sharp tips of cathode particles christened as corona yank can be estimated using the expression ∇E= ∑ 2nEedgeLdiff e nI

(8)

Our hypothesis that there is energy loss when charges encounter edges of cathode particles becomes corroborated with the observation of electrochemical irreversibility or electrochemical hysteresis in cyclic voltammogram (CV) traces corresponding to sharp-edged particles (Figure 4a,b), while there is clear electrochemical reversibility for the milled blend (Figure 4c). For an electrochemical system exhibiting reversible electrochemical reaction, we can expect equal or comparable rate constant values for forward and backward electron-transfer

Figure 3. Schematics illustrating the occurrence of corona-yank in edged LNMC and LNM cathode particles (a) compared with impressive capacity maintenance in smooth surface milled blend (b). D

DOI: 10.1021/acs.langmuir.6b00336 Langmuir XXXX, XXX, XXX−XXX

Article

Langmuir

Figure 4. Cyclic voltammogram traces of cells made with (a) Li vs LNMC (as prepared), (b) Li vs LNM (as-prepared), and (c) Li vs LNMC and LNM (milled blend) in the voltage range from 2.5 to 4.6 V (1st to 5th cycles). Pronounced electrochemical irreversibility present in the first two cases (a,b) is indicated using double-headed arrows.

Notes

reactions. This is indeed the case observed for the CV traces in Figure 4c corresponding to milled blend. On the contrary, in an irreversible electrochemical system, the backward electrontransfer rate is considerably slowed, owing to the presence of barriers (thermodynamic/electrostatic, dynamic in nature), and hence it would be unfavorable to maintain the Nernstian equilibrium. Quite expectedly, charge transport across edges of the cathode particles might entail energy loss or decrease in flux density, with the edge serving as a potential barrier to be surmounted. Furthermore, a strong intersection of CV traces for the sharp-edged particles (Figure 4a,b) might again suggest a decreased electron/charge-transfer rate. In any case, the sharp-edged cathode particles do command higher initial capacity, which upon cycling undergo pronounced fade, which we prefer to call a corona yank.

The authors declare no competing financial interest.



ACKNOWLEDGMENTS Sincere thanks to Dr. Periasamy and Dr. P. Manikandan of CECRI-CSIR, Karaikudi-630006 (TN), India for the data support.



(1) Afyon, S.; Krumeich, F.; Mensing, C.; Borgschulte, A.; Nesper, R. New High Capacity Cathode Materials for Rechargeable Li-ion Batteries: Vanadate-Borate Glasses. Sci. Rep. 2014, 4, 7113. (2) Wang, M.; Yang, M.; Ma, L.; Shen, X. The high capacity and excellent rate capability of Ti-doped Li2MnSiO4 as a cathode material for Li-ion batteries. RSC Adv. 2015, 5, 1612−1618. (3) Lee, H.-W.; Muralidharan, P.; Ruffo, R.; Mari, C. M.; Cui, Y.; Kim, D. K. Ultrathin Spinel LiMn2O4 Nanowires as High Power Cathode Materials for Li-Ion Batteries. Nano Lett. 2010, 10, 3852− 3856. (4) Chen, H.; Grey, C. P. Grey. Molten Salt Synthesis and High Rate Performance of the “Desert-Rose” form of LiCoO2. Adv. Mater. 2008, 20, 2206−2210. (5) Kim, D. H.; Kim, J. Synthesis of LiFePO4 nano-particles in polyol medium and their electrochemical properties. Electrochem. Solid-State Lett. 2006, 9, A439−A442. (6) Fang, H.; Li, L.; Yang, Y.; Yan, G.; Li, G. Low-temperature synthesis of highly crystallized LiMn2O4 from alpha manganese dioxide nanorods. J. Power Sources 2008, 184, 494−497. (7) Liang, H.; Qiu, X. P.; Zhang, S. C.; He, Z. Q.; Zhu, W. T.; Chen, L. Q. High performance lithium cobalt oxides prepared in molten KCl for rechargeable lithium-ion batteries. Electrochem. Commun. 2004, 6, 505−509. (8) Chang, Z.; Chen, Z.; Wu, F.; Tang, H.; Zhu, Z.; Yuan, X. Z.; Wang, H. Synthesis and properties of high tap-density cathode material for lithium ion battery by the eutectic molten-salt method. Solid State Ionics 2008, 179, 2274−2277.

4. CONCLUSIONS Foregoing results and related results in literature lend support to the conclusion that as-synthesized cathode particles for Liion batteries having sharp edges do command high initial capacity attributed to crowding of electrical charges at steep curvatures or sharp edges explained on the basis of electrostatic considerations. Simultaneously, under discharge conditions, these edges have the drawback of losing energy because these act as potential barriers for charge transport, a situation quite different from edge-free milled particles offering hassle-free charge transport, explaining good capacity retention for the latter.



REFERENCES

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]; [email protected]. E

DOI: 10.1021/acs.langmuir.6b00336 Langmuir XXXX, XXX, XXX−XXX

Article

Langmuir (9) Park, B.-C.; Bang, H. J.; Amine, K.; Jung, E.; Sun, Y.-K. Electrochemical stability of core−shell structure electrode for high voltage cycling as positive electrode for lithium ion batteries. J. Power Sources 2007, 174, 658−662. (10) Numata, T.; Amemiya, C.; Kumeuchi, T.; Shirakata, M.; Yonezawa, M. Advantages of blending LiNi 0.8 Co 0.2 O 2 into Li1+xMn2−xO4cathodes. J. Power Sources 2001, 97−98, 358−360. (11) Ma, Z. F.; Yang, X. Q.; Liao, X. Z.; Sun, X.; McBreen, J. Electrochemical evaluation of composite cathodes base on blends of LiMn2O4 and LiNi0.8Co0.2O2. Electrochem. Commun. 2001, 3, 425− 428. (12) Kitao, H.; Fujihara, T.; Takeda, K.; Nakanishi, N.; Nohma, T. High-Temperature Storage Performance of Li-Ion Batteries Using a Mixture of Li-Mn Spinel and Li-Ni-Co-Mn Oxide as a Positive Electrode Material Electrochem. Electrochem. Solid-State Lett. 2005, 8, A87−A90. (13) Myung, S. T.; Cho, M. H.; Hong, H. T.; Kang, T. H.; Kim, C. S. Electrochemical evaluation of mixed oxide electrode for Li-ion secondary batteries: Li1.1Mn1.9O4 and LiNi0.8Co0.15Al0.05O2 J. J. Power Sources 2005, 146, 222−225. (14) Yang, J.; Zhang, X.; Han, X.; Cheng, F.; Tao, Z.; Chen, J. Intergrown LiNi0.5Mn1.5O4·LiNi1/3Co1/3Mn1/3O2 composite nanorods as high-energy density cathode materials for lithium-ion batteries. J. Mater. Chem. A 2013, 1, 13742−13745. (15) Manikandan, P.; Periasamy, P.; Jagannathan, R. Faceted shapedrive cathode particles using mixed hydroxy-carbonate precursor for mesocarbon microbeads versusLiNi1/3Mn1/3Co1/3O2 Li-ion pouch cell. J. Power Sources 2014, 245, 501−509. (16) Manikandan, P.; Periasamy, P.; Jagannathan, R. Grain boundary driven capacity fade/hysteresis abated in composite cathode material for lithium-ion batteries/pouch cell. J. Power Sources 2014, 264, 299− 310. (17) Chang, Z.; Chen, Z.; Wu, F.; Tang, H.; Yuan, X. Z.; Wang, H. Synthesis and characterization of nonspherical LiCoO2 with high tap density by two-step drying method Electrochem. Electrochem. SolidState Lett. 2008, 11, A229−A232. (18) Chang, Z.; Chen, Z.; Wu, F.; Tang, H.; Zhu, Z.; Yuan, X. Z.; Wang, H. Solid State Ionics 2008, 179, 2274−2277. (19) Zhu, H. L.; Chen, Z. Y.; Ji, S.; Linkov, V. Influence of different morphologies on electrochemical performance of spinel LiMn2O4. Solid State Ionics 2008, 179, 1788−1793. (20) Han, S.; park, J.; Lu, W.; Sastry, A. M. Numerical study of grain boundary effect on Li+ effective diffusivity and intercalation induced stresses in Li-ion battery active Materials. J. Power Sources 2013, 240, 155−167. (21) Loeb, L. B. Electrical Coronas, Their Basic Physical Mechanisms; University of California Press, 1965. (22) Xu, B.; Qian, D.; Wang, Z.; Meng, Y. S. Recent progress in cathode materials research for advanced lithium ion batteries. Mater. Sci. Eng., R 2012, 73, 51−65. (23) Gaberscek, M.; Dominko, R.; Jamnik, J. Electrochem. Commun. 2007, 9, 2778−2783. (24) Riley, K. F., Hobson, M. P.; Bence, S. J. Mathematical Methods for Physics and Engineering; Cambridge University Press, 2010. (25) Rusinkiewicz, S. Estimating Curvatures and Their Derivatives on Triangle Meshes; Princeton University Publications, 2004; pp 1−8. (26) Thidé, B. Electromagnetic Field Theory, 2nd ed.; Upsilon Books: Uppsala, Sweden, 2010. https://www.calvin.edu/~pribeiro/courses/ engr315/EMFT_Book.pdf. (27) Bueno, P. R.; Leite, E. R. Nanostructured Li Ion Insertion Electrodes. 1. Discussion on Fast Transport and Short Path for Ion Diffusion. J. Phys. Chem. B 2003, 107, 8868−8877.

F

DOI: 10.1021/acs.langmuir.6b00336 Langmuir XXXX, XXX, XXX−XXX