Preprint of article published at Chemistry of Materials https://doi.org/ dx.doi.org/10.1021/cm3006887 1 Solid Electrolyte Interphase in Li-ion Batteries: Evolving Structures Measured In Situ by Neutron Reflectometry Jeanette E. Owejan* a , Jon P. Owejan a , Steven C. DeCaluwe b,c and Joseph A Dura* b , a Electrochemical Energy Research Laboratory, General Motors, Honeoye Falls, NY 14472, USA b NIST Center for Neutron Research, Gaithersburg, MD 20899-6102, USA c Department of Materials Science and Engineering, University of Maryland, College Park, MD 20742, USA Author Contact Information: Steven C. DeCaluwe: NIST Center for Neutron Research 100 Bureau Dr. MS 6102 Bldg. 235 A124 Gaithersburg, MD 20899-6102 Phone: 301-975-8811 Fax: 301-921-9847 email: [email protected]Joseph A. Dura: NIST Center for Neutron Research 100 Bureau Dr. MS 6102 Bldg. 235 A126 Gaithersburg, MD 20899-6102 Phone: 301-975-6251 Fax: 301-921-9847 email: [email protected]Jeanette E. Owejan: Electrochemical Energy Research Laboratory, General Motors 10 Carriage St. Honeoye Falls, NY 14472 Phone: 585-259-2130 Fax: 585-624-6680 email: [email protected]Jon P. Owejan: Electrochemical Energy Research Laboratory, General Motors 10 Carriage St. Honeoye Falls, NY 14472 Phone: 585-953-5558 Fax: 585-624-6680 email: [email protected]
22
Embed
Solid Electrolyte Interphase in Li-ion Batteries: …inside.mines.edu › ~decaluwe › Papers › OwejanEtAl_ChemMat...Solid Electrolyte Interphase in Li-ion Batteries: Evolving Structures
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Solid Electrolyte Interphase in Li-ion Batteries: Evolving Structures Measured In Situ by Neutron Reflectometry Jeanette E. Owejan*a, Jon P. Owejana, Steven C. DeCaluweb,c and Joseph A Dura*b, aElectrochemical Energy Research Laboratory, General Motors, Honeoye Falls, NY 14472, USA b NIST Center for Neutron Research, Gaithersburg, MD 20899-6102, USA c Department of Materials Science and Engineering, University of Maryland, College Park, MD
20742, USA
Author Contact Information:
Steven C. DeCaluwe:
NIST Center for Neutron Research 100 Bureau Dr. MS 6102 Bldg. 235 A124 Gaithersburg, MD 20899-6102 Phone: 301-975-8811 Fax: 301-921-9847 email: [email protected] Joseph A. Dura:
NIST Center for Neutron Research 100 Bureau Dr. MS 6102 Bldg. 235 A126 Gaithersburg, MD 20899-6102 Phone: 301-975-6251 Fax: 301-921-9847 email: [email protected]
Jeanette E. Owejan:
Electrochemical Energy Research Laboratory, General Motors 10 Carriage St. Honeoye Falls, NY 14472 Phone: 585-259-2130 Fax: 585-624-6680 email: [email protected]
Jon P. Owejan:
Electrochemical Energy Research Laboratory, General Motors 10 Carriage St. Honeoye Falls, NY 14472 Phone: 585-953-5558 Fax: 585-624-6680 email: [email protected]
Li-ion batteries are made possible by the solid electrolyte interphase, SEI, a self-forming
passivation layer, generated due to electrolyte instability with respect to the anode chemical
potential. Ideally it offers sufficient electronic resistance to limit electrolyte decomposition to
the amount needed for its formation. However, slow continued SEI growth leads to capacity
fade and increased cell resistance. Despite the SEI’s critical significance, currently structural
characterization is incomplete due to the reactive and delicate nature of the SEI and the
electrolyte system in which it’s formed. Here we present, for the first time, in situ neutron
reflectometry measurements of the SEI layer as function of potential in a working lithium half-
cell. The SEI layer after 10 and 20 CV cycles is 4.0 and 4.5 nm, respectively, growing to 8.9 nm
after a series of potentiostatic holds that approximates a charge / discharge cycle. Specified
datasets show uniform mixing of SEI components.
Keywords:
Li ion Battery, Solid Electrolyte Interphase, In-situ Neutron Reflectometry
Synopsis
These first in situ neutron reflectometry measurements of a solid electrolyte interphase, SEI, layer vs. potential in a working Li battery show its structure, including Li incorporation and thickness, from 4.0 nm to 8.9 nm after a charge / discharge cycle. Uniform mixing of SEI components occurs at some potentials.
chosen as working electrode as it does not intercalate lithium, and thus all electrochemical
charge is attributed to decomposition of the electrolyte to form the SEI layer. Copper is
frequently used as an ‘ideal’ electrode for SEI studies31, 39-45, and it has been established that
films grown on non-intercalating substrates are similar to those grown on carbon materials at low
potentials in Li salts46,47. Aurbach et al. have also shown that the thermodynamics of reduction
processes of the electrolyte solution species are governed by the cation used43. The electrolyte
consisted of 1mol/L LiPF6 in a 1:2 (v/v) ratio of deuterated ethylene carbonate, EC, and diethyl
carbonate, DEC. The deuteration was chosen for two reasons. It increases the SLD of the
electrolyte to provide greater contrast to the SEI, which is expected to have a relatively low SLD
due to Li incorporation. Additionally, isotopic labeling of the EC allows one to attempt to
identify the possible preferential decomposition of cyclic over acyclic carbonates.
Experimental Methods
Half Cell Fabrication
The native oxide was first removed by HF cleaning from a single crystal silicon wafer
(Institute of Electronic Materials Technologyd) that is 7.6cm in diameter by 0.95 cm thick with
N/P/100 type/dopant/orientation, respectively. Immediately a 7.5 nm titanium adhesion layer
and 40 nm copper layer were deposited via electron beam evaporation, at the Semiconductor and
Microsystems Fabrication Laboratory at Rochester Institute of Technology. Layer composition,
uniformity, and roughness of a sacrificial wafer were verified using X-ray reflectometry. The
lithium foil (Alfa Aesar), 5.72 cm diameter, was polished and pressed onto a copper/titanium dCertaincommercialequipment,instruments,ormaterialsareidentifiedinthispapertofosterunderstanding.SuchidentificationdoesnotimplyrecommendationorendorsementbytheNationalInstituteofStandardsandTechnology,nordoesitimplythatthematerialsorequipmentidentifiedarenecessarilythebestavailableforthepurpose.
Figure 1. a) Neutron reflectivity vs. Q is shown for the sample at OCV and after 10 CV cycles during a hold at 250 mV. The solid lines are the best fit, from a simultaneous fit to the two data sets. b) The SLD profiles for the two best fits. The SLD values of Si, Cu and Ti (calculated from known densities) are indicated, and Electrolyte, SEI, and TiSix layers are identified. For both parts, the darker and lighter shaded regions are the 68 % and 95 % confidence intervals, as discussed in the methods section.
simultaneous fit excluding the SEI was performed. This fit (not shown) resulted in not only an
increase in χ2 but also an increase in the Bayesian information criterion, thus indicating that the
model with an SEI is statistically more likely than the model excluding it. The next test point (c-
150mV) repeated the cyclic voltammograms, followed by a potentiostatic hold at 150 mV.
Charge data for all test points may be found in the Supporting Information.
The SEI layer at test point c is remarkably similar to the previous test point, b, having
similar SLD and interface widths, but slightly thicker at 4.5 nm [4.0, 4.7] nm due to the lower
reduction potential and/or additional 10 CV cycles. These results also agree with a 3.6 nm [3.2,
4.2] nm thick SEI deposited by a similar procedure in a second cell with a fully protonated
electrolyte (see Supporting Information). Subsequently, six data sets (d-i) were taken at various
potentials by slowly ramping the potential at a rate of 10 mV/s to the next value and holding
during NR data collection, as summarized in Figure 2, which shows representative CV curves
taken throughout the experiments. Note that complex chemistries occur as demonstrated by the
various peaks in the CV curves, the
origins of which have been
discussed in the literature31,43,44,51.
They serve for us as points at
which to execute the potential
holds and NR measurements. The
decrease in the magnitude of the
current density with increased
cycling in Figure 2 demonstrates
the passivating nature of the
Figure 2. Electrochemical measurements including cyclic voltammogram results for selected scans. Test points b-i denote the location of potentiostatic holds for NR testing.
Figure 3. Neutron reflectivity (expressed as RxQ^4) vs. Q is shown for the sample measured during holds at the potentials indicated. The solid lines are the best individual fit, with several parameters kept constant at values determined from the simultaneous fit of the OCV and 250 mV data (see text for details). The darker and lighter shaded regions are the 68 % and 95 % confidence intervals, as discussed in the methods section.
Figure 4. The SLD as a function of depth for the SEI deposited on Cu, showing the evolution of the thickness, SLD and interface roughness with hold potential, as indicated. The lines for b-i, are best individual fits, with several parameters kept constant constant at values determined from the simultaneous fit of the OCV and 250 mV data (see text for details). The inset shows the full SLD profile, with the fits co-aligned on the Ti layer.
increased from a starting thickness of 41.51 nm [41.39, 41.57] nm in test point b to a thickness of
41.81nm [41.76, 41.88] nm and 42.17 nm [42.10, 42.23] nm for test points e and f, respectively,
due to the addition of a few monolayers of material, presumably Cu2O (which has a high SLD
sufficiently close to that of Cu). This layer is too thin to be accurately modeled as a separate
layer32,33, and is subsequently removed during cathodic potential holds in test points g-i, where
the Cu layer thickness decreases to a final value of 41.43 nm [41.40, 41.49] nm, which is the
same as the initial value to within uncertainty. Sufficient published electrochemical data of
copper in non-aqueous systems identifies Cu+1 oxidation occurring near 2.0 V vs. Li and Cu+2
oxidation occurring at 3.5V vs. Li.51 . These increases in apparent copper thickness originate
from the demand of current-generating chemical reactions to sustain the potential holds. Thus,
fitting results. X-ray Photoelectron Spectroscopy, XPS, data (contained in Supporting
Information) indicates the presence of LiF, LiOH, lithium alkyl carbonates, and non-lithium
containing polyethylene oxide (SLD = 3.86 x10-4 nm-2), PEO. The fractional amount of lithium
was assigned to each molecule according to the peak areas in the deconvoluted XPS data. The
relative amount of PEO was determined from the C1s and O1s peaks in the XPS data. Porosity
of the SEI is handled by adding electrolyte to the composition, which affects both SLD and
overall thickness. The thickness is determined from the number of molecules in the model and
the molecular volume. The amounts of each component were adjusted within the variability
range observed in the post mortem XPS measurements, and guided by known chemical reactions
producing SEI components21, until the model thickness and composition (red and blue dashed
lines, respectively) matched the thickness and SLD measured by NR (red squares and blue
circles respectively) as seen in Figure 5. The results of this model are shown in Figure 6, where
the modeled partial thickness are summed and compared to the NR measurement.
The changes of the SEI
composition when moving to
successively more reducing
potential include the increase in
concentration of LiOH and LiF
molecules, and the decrease of
lithium alkyl carbonates. The
porosities determined in these
models (9-10%) are similar to the
11.0% [10.6%, 15.9%] porosity
Figure 5. Selected fitting parameters as a function of test point. For reference, the hold potential vs. Li is also shown. The red and blue dashed lines are the total thickness and SLD from the composition modeling, as described in the text, and match well with values measured by NR. Cu thickness is plotted relative to the thickness measured in test point a – OCV.
Figure 6. SEI thickness and composition modeled by use of XPS, electrochemical, and NR measured parameters compared to the thickness measured by NR. Each segment in the bar represents the partial thickness (not gradient) that each molecule contributes to the SEI thickness. The hold potential and charge ratio (measured during the holds / modeled) are also shown.
determined by another cell that utilized contrast variation via fluid exchange (see Supporting
Information). The quantity of LiOH and LiF grows during test duration, while the lithium alkyl
carbonate concentration fluctuates during testing. The molecules identified above are consistent
with previous ex situ studies, however the lack of segregation of certain molecules shows that
our in situ evaluation of the SEI structure and composition sheds significant light on how the SEI
initiates and grows. Furthermore, others have modeled that ageing of the SEI over time will
reduce hydroxides and lithium alkyl and dicarbonates to the more stable Li2O, LiF, and Li2CO3
components55. However, this combined electrochemical/reflectivity test was run over a period of
investigations will enable future systematic improvements for commercial device performance,
affordability, and sustainability.
Acknowledgements
The authors gratefully acknowledge funding for this work through NIST ARRA Award Number: 60NANB10D027, part of the American Recovery and Reinvestment Act of 2009, as well as the National Research Council for funding through the NRC Research Associateship Program. Nicholas P. Irish of GM Global R&D is acknowledged for XPS data collection and peak deconvolution and Paul A. Kienzle, NCNR, for useful discussions of NR fitting.
Supporting Information
Additional Figures in the form of charge accumulation, XPS overlay plots, reflectivity plot for the fluid exchange cell, and scattering length density profile as a function of test point for the fluid exchange cell. Tabular data for atomic % composition by XPS for the non fluid exchange cell, description of ex situ sample preparation, cell build and test for the fluid exchange cell. Neutron Reflectivity data collection and analysis details, results of data from fluid exchange cell. This material is available free of charge via the Internet at http://pubs.acs.org.
23. Kanamura, K., Tamura, H., Shiraishi, S., & Takehara, Z. J. Electrochem. Soc. 1995, 142, 340-7.
24. Kanamura, K., Shiraishi, S., & Takehara, Z. J. Electrochem. Soc. 1996, 143, 2187-2197.
25. Lee, C., Mun, B., & Ross Jr., P. N. J. Electrochem. Soc. 2002, 149, A1286-A1292.
26. Andersson, A. M., & Edström, K. J. Electrochem. Soc. 2001, 148, A1100-9.
27. Ein-Eli, Y., McDevitt, S. F., & Laura, R. J. Electrochem. Soc. 1998, 145, L1-3.
28. Aurbach, D., Markovsky, B., Shechter, A., Ein-Eli, Y., & Cohen, H. J. Electrochem. Soc. 1996, 143, 3809-3820.
29. Xu, K., Zhuang, G. V., Allen, J. L., Lee, U., Zhang, S. S., Ross Jr., P. N., et al. J. Phys. Chem. B. 2006, 110, 7708-7719.
30. Xu, K. J. Electrochem. Soc. 2009, 156, A751-5.
31. Zhuang, G. V., Xu, K., Yang, H., Jow, T. R., & Ross Jr., P. N. J. Phys. Chem. B. 2005, 109, 17567-17573.
32. Seah, M. P., et al. Surf. Interface Anal. 2004, 36, 1269-1303. and
Seah, M. P., et al Surf. Interface Anal. 2009, 41, 430-9.
33. Dura, J. A., C A. Richter, C. F. Majkrzak, and N. V. Nguyen, Appl. Phys. Let. 1998, 73, 2131.
34. Fitzsimmons M. R., & Majkrzak, C. F. Modern Techniques for Characterizing Magnetic Materials, Ed. Zhu, Y., Kluwer Acedemic Publishers: Boston, 2005.
35. Dura, J. A., Murthi, V. S., Hartman, M., Satija, S. K., & Majkrzak, C. F. Macromolecules 2009, 42, 4769-4774.
36. Majkrzak, C.F., et al. Neutron Scattering in Biology: Techniques and Applications. Eds. Fitter, J., Gutberlet, T., Katsaras, J., Springer Publishing: New York, 2006.
37. Wacklin, H. P. Curr. Opin. Colloid In. 2010, 15, 445-454.
38. Wang, H., Downing, R. G., Dura, J.A., & Hussey, D.S. In Situ Neutron Techniques for Studying Lithium Ion Batteries, in Polymers for Energy Storage and Delivery: Polyelectrolytes for Batteries and Fuel Cells, Amer. Chem Soc. 2011, in press.
41. Peng, C., et al. J Appl Electrochem 2010, 40,653–662.
42. Kawakita, J., Kobayashi, K., J.Power Sources 2001, 101, 47-52.
43 .Moshkovich, M., Gofer Y., Aurbach, D. J Electrochem Soc, 2001,148, E155-E167.
44. Zhao, M. et al. J Electrochem. Soc., 2000,147, 2874-2879.
45. Kanamura, K., Tamura, H, Shiraishi, S. Takehara, Z. J Electroanal Chem 1995, 394, 49-62.
46. D. Aurbach and Y. Gofer, in Nonaqueous Batteries, D. Aurbach, Editor, Chap. 4, p. 137, Marcel Dekker, Inc., New York: 1999. 47. E. Peled, in Li Batteries, J. P. Gabano, Editor, Chap. 3, p. 43, Academic Press, New York: 1983.
48. Kienzle P. A., Krycka J. A., & Patel, N. Refl1D: Interactive depth profile modeler. http://www.reflectometry.org/danse/software
49. Vrugt J. A., ter Braak C. J. F., Diks C. G. H., Higdon D., Robinson B. A., & Hyman J. M. Int. J. Nonlin. Sci. 2009, 10, 271-288.
50. Dura, J. A., et al. Rev. Sci. Instrum. 2006, 77.