LETTERS PUBLISHED ONLINE: 12 FEBRUARY 2012 | DOI: 10.1038/NMAT3246 7 Li MRI of Li batteries reveals location of microstructural lithium S. Chandrashekar 1,2 , Nicole M. Trease 2 , Hee Jung Chang 2 , Lin-Shu Du 2 , Clare P. Grey 2,3 * and Alexej Jerschow 1 * There is an ever-increasing need for advanced batteries for portable electronics, to power electric vehicles and to facilitate the distribution and storage of energy derived from renewable energy sources 1,2 . The increasing demands on batteries and other electrochemical devices have spurred research into the development of new electrode materials that could lead to better performance and lower cost (increased capacity, stability and cycle life, and safety) 1–3 . These developments have, in turn, given rise to a vigorous search for the development of robust and reliable diagnostic tools to monitor and analyse battery performance, where possible, in situ 4–9 . Yet, a proven, convenient and non-invasive technology, with an ability to image in three dimensions the chemical changes that occur inside a full battery as it cycles, has yet to emerge. Here we demonstrate techniques based on magnetic resonance imaging, which enable a completely non-invasive visualization and characterization of the changes that occur on battery electrodes and in the electrolyte. The current application focuses on lithium-metal batteries and the observation of electrode microstructure build-up as a result of charging. The methods developed here will be highly valuable in the quest for enhanced battery performance and in the evaluation of other electrochemical devices. The development and use of advanced electrochemical systems, such as batteries, are vital to increase both the efficiency of fossil- fuel consumption and the deployment of intermittent renewable energy sources such as solar and wind energy 1,2 . The immensely popular lithium-ion batteries (LIBs), with their high gravimetric and volumetric capacities, have helped to enable the portable electronics revolution, and are now being developed for use in hybrid and all-electric vehicles. New electrode and electrolyte materials are constantly being proposed, which could lead to substantial improvements in the electrochemical properties and capacities of the devices 1–3 . Non-invasive diagnostic tools are of prime importance in the assessment and development of such new materials 4–10 . Magnetic resonance imaging (MRI), also a non- invasive diagnostic tool, is well established in the medical field 11 and is becoming increasingly popular in the materials science field 12–16 . Examples of the latter include MRI in polymer sciences, catalysis 13 , process analysis 14 and more recently in fuel cells 17,18 and corrosion processes 19 . Studies of MRI of batteries have been very sparse 20,21 , and have been limited to either imaging the electrolyte only 20 , or visualizing 7 Li signals from specifically constructed cells that allow for spatial resolution through radiofrequency (RF) field gradients 21 . In this latter work, a two-dimensional chemical-shift imaging (CSI) protocol was employed, with the spatial image limited to one 1 Chemistry Department, 100 Washington Square East, New York, New York 10003, USA, 2 Department of Chemistry, Stony Brook University, Stony Brook, New York 11794, USA, 3 Department of Chemistry, University of Cambridge, Lensfield Road, Cambridge CB2 1EW, UK. *e-mail:cpg27@cam.ac.uk; alexej.jerschow@nyu.edu. dimension, and being nonlinear. Such cells, however, only remotely resemble practical batteries, and the spatial resolution can only be achieved along specific dimensions. In this report, we demonstrate two- and three-dimensional 7 Li MRI and (magnetic resonance) CSI of batteries, with full control over resolution and imaging axes. It is further shown that challenges due to the orientation-dependent RF penetration of the electrodes can be overcome. We have chosen to investigate Li electrodes in a symmetric cell, in their pristine and subsequently charged states. The images reveal the location of the Li microstructures on or in the vicinity of the negative electrode. In principle, the most attractive anode material for a LIB is Li metal, as it is the lightest and most electropositive material, giving the greatest capacity and voltage. Li metal has a specific energy density that is more than ten times greater than that of lithiated graphite, the most commonly used commercial anode material, on the basis of the mass of Li versus fully lithiated carbon, LiC 6 . However, the commercial use of Li metal batteries is beset by serious safety concerns, largely due to the formation of microstructure build-up on the electrodes. Such microstructures may appear in the form of irregular porous microstructures on top of the electrode (‘moss’; refs 6,22), or in dendritic shapes (‘dendrites’; refs 6,10,22). Li metal disassociated from the electrodes and protruding into the electrolyte/separator region can result in short-circuiting, battery overheating and risk of fire 23–25 . Li deposits can also form on the graphitic anodes used at present, representing a significant safety concern that prevents rapid charging of LIBs. The rate of dendrite formation increases with cycle rate 6–8 , although the electrochemical conditions under which Li microstructures (dendrite, moss and so on) form, and the effects of different electrolytes and additives, are still not well characterized. A deeper understanding of the electrochemical processes leading to these degradation mechanisms would accelerate the development of commercial Li-metal batteries for future technologies. Various techniques have been used to study Li microstructure formation. Scanning electron microscopy (SEM) can distinguish dendrites from moss, but is not quantitative 6,10 , whereas 7 Li in situ NMR (refs 5,7,8,21) provides a quantitative method, but lacks spatial information. MRI has the potential to provide both, as demonstrated below, thus representing a unique tool to probe microstructure formation. NMR spectroscopy and imaging of bulk metals is especially challenging, because RF irradiation only penetrates an extremely (microscopically) thin layer of material of the order of the skin depth δ (refs 7,26), which under the conditions considered here is 9.3 μm. Despite these difficulties, it has been recently demonstrated that quantitative information can be gleaned from the NMR NATURE MATERIALS | VOL 11 | APRIL 2012 | www.nature.com/naturematerials 311 © 2012 Macmillan Publishers Limited. All rights reserved