Magnesium Metal Anode Interfaces and Performance in Chloride-Free Electrolytes

Thursday, October 15, 2015: 10:00
101-A (Phoenix Convention Center)
N. T. Hahn (Sandia National Laboratories, Joint Center for Energy Storage Research) and K. R. Zavadil (Joint Center for Energy Storage Research, Sandia National Laboratories)
The development of rechargeable Mg batteries, driven by the desire to overcome the limiting specific energy density and high cost of Li ion batteries, is currently frustrated by the lack of stable, functional electrolytes compatible with high voltage cathodes.  A functional electrolyte must deliver the Mg cation to the anode surface at nearly 100% Coulombic efficiency, requiring cation desolvation and accommodation without formation of a cation blocking film.  Successful electrolytes have relied on Lewis acid – base reactions to form Mg cation-solvent complexes that include the traditional Grignard based and more recently reported inorganic MgCl2 based complex electrolytes containing either BR4- or AlRxCl4-x- species.1  However, the stability of organometallic and chloroaluminate electrolytes against traditional current collectors and proposed oxide intercalation cathodes are of significant concern. Novel Mg battery electrolytes based on the weakly coordinating N(SO2CF3)2 (TFSI) anion have recently received interest due to their wide electrochemical window and general chemical compatibility while maintaining the ability to deposit and strip Mg, a feature previously unknown for “conventional” Mg salt solutions.2,3,4  However, Mg deposition in these systems is typically characterized by low coulombic efficiency (< 50%) and Mg surface passivation in the absence of chloride additives. 

In this presentation we describe Mg deposition and stripping from a chloride-free MgTFSI2-glyme electrolyte at nearly 90% coulombic efficiency, demonstrating that such “conventional” electrolytes have the ability to support reversible Mg deposition at application-relevant rates (> 1 mA/cm2).  This finding proves that the use of weakly coordinating salts in glymes need not prevent Mg2+ transport across the anode/electrolyte interface. The MgTFSI2 system further teaches us that non-reducing, non-chloride electrolytes are highly sensitive to impurities, which regulate the interfacial processes governing deposition irreversibility. We will discuss the influence of specific impurities, identified through spectroscopic techniques including GC-MS, on the activity of MgTFSI2-glyme electrolytes and demonstrate the importance of impurity removal for attaining good performance (Figure 1a).  Having addressed the problem of impurities we will further discuss the origins of the technologically insufficient coulombic efficiency (<99.9%) and the passivation of Mg in MgTFSI2-glyme electrolytes.  Characterization of cycled electrolytes by ESI-MS and cycled electrode interfaces by XPS implicates TFSI decomposition during Mg deposition and stripping as the primary source of these problems.  This finding corroborates previous computational work performed by our collaborators, which predicts an electrochemical-chemical TFSI decomposition mechanism involving S-C bond cleavage and liberation of a CF3 group.5  Finally, we observe that films formed on Mg surfaces in MgTFSI2-glyme electrolytes do not prevent the subsequent deposition of Mg.  This surprising result argues that surface films formed in these electrolytes are not necessarily cation blocking layers in the classic sense, but instead exhibit rectifying behavior, facilitating cation transport during deposition and impeding cation transport during dissolution (Figure 1b).  Characterization of these films via visualization and surface-sensitive spectroscopic techniques including ToF-SIMS will be presented, and the extent to which these films regulate Mg2+ transport to and from the anode surface will be discussed.

This work is supported as part of the Joint Center for Energy Storage Research, an Energy Innovation Hub funded by the U.S. Department of Energy, Office of Science. Sandia is a multi-program laboratory managed and operated by Sandia Corporation, a wholly owned subsidiary of Lockheed Martin Company, for the U.S. DOE’s NNSA under contract DE-AC04-94AL85000.

1.  T. Liu, et al. J. Mater. Chem. A 2014, 2 3430

2.  U.S. Patent #US 2013/0252112 A1

3.  S.Y. Ha, et al. ACS Appl. Mater. Interfaces 2014, 6, 4063

4.  Y. Orikasa, et al. Sci. Rep. 2014, 4, 5622

5.  N. Rajput, et al. J. Amer. Chem. Soc. 2015, 137, 3411