Scholarly article on topic 'Observation of Lithium Dendrites at Ambient Temperature and Below'

Observation of Lithium Dendrites at Ambient Temperature and Below Academic research paper on "Nano-technology"

Share paper
Academic journal
ECS Electrochemistry Letters
OECD Field of science

Academic research paper on topic "Observation of Lithium Dendrites at Ambient Temperature and Below"

p1 r j Observation of Lithium Dendrites at Ambient Temperature

V —1 ¿_J I Í Corey T. Love,*'z Olga A. Baturina,* and Karen E. Swider-Lyons*

Chemistry Division, U.S. Naval Research Laboratory, Washington, DC 20375, USA

Lithium-ion batteries are prone to failure at low temperatures and dendrite growth during charging is one suspect. We attempt to understand lithium dendrite growth by observing their number, initiation time and growth rate at ambient and sub-ambient temperatures: — 10°C, 5°C, and 20°C using an in-situ optical microscopy cell (Li0|Li0). We find that while dendrites initiate quickly at — 10°C, the cells at 5°C short-circuit most rapidly due in part to a favorable morphology at this temperature. The experimental approach has broad applicability to other electrochemical energy storage technologies where mass transport limitations are present at low temperatures, particularly Li-air, Li-S, and Zn-air batteries.

© The Author(s) 2014. Published by ECS. This is an open access article distributed under the terms of the Creative Commons Attribution Non-Commercial No Derivatives 4.0 License (CC BY-NC-ND,, which permits non-commercial reuse, distribution, and reproduction in any medium, provided the original work is not changed in any way and is properly cited. For permission for commercial reuse, please email: [DOI: 10.1149/2.0041502eel] All rights reserved.

Manuscript submitted October 29, 2014; revised manuscript received November 24, 2014. Published December 11, 2014.

Lithium dendrite formation at the anode during charging of lithium-ion batteries can initiate internal short circuits, a known failure mechanism. Dendrite-induced short circuits at low temperature have been identified as a causal factor in recent lithium-ion battery failures aboard commercial aircraft.1 Low temperature charging affects the kinetic processes occurring within the cell causing lithium ions to reduce and plate metallic lithium (Li0) on the carbon anode rather than favoring ion intercalation into the carbon electrode, according to Equation 1. At best, this Li0 deposition is manifested as uniform plating leading to a loss of useful capacity of the battery. In the worst case the Li0 electrodeposits grow uncontrollably to form an internal short to the cathode capable of initiating the thermal runaway reaction.2 Factors which favor unstable Li+ reduction to form dendrites rather than ion intercalation at low temperature include: lower solubility of Li+ within the liquid electrolyte, greater ion pairing between Li+ and the anion (PF6—), and increased viscosity of electrolyte causing slowed ion diffusion.3

Li+ + e—= Li0 [1]

Akolkar recently developed a temperature-dependent diffusion-reaction model to predict dendrite growth rate based upon the ratio of the dendrite tip current density to the current density on a flat lithium surface.4 The model is built upon the principle that dendrite growth is more pronounced at low temperature due to increased mass transport resistance of lithium ions through a viscous electrolyte and reduced charge transfer resistance created by a thinner solid electrolyte interface (SEI) layer. The model predicts — 10°C as the critical temperature below which dendrites grow uncontrollably based upon an applied current density of 5 mA • cm-2.

Lithium electrodeposition and dendrite formation are observable in a Li0|Li0 cell using in-situ optical microscopy.5-11 We use a custom open-faced Li0|Li0 cell to characterize lithium electrodeposition during low temperature charging. Harris outlined the advantages of using an "open-faced" electrochemical cell; take advantage of color changes associated with lithiation of graphite, introduce large concentration gradients for making transport measurements, and the ability to convert the system from 3 D to 2 D5 which in our experiments creates a uniform edge for directing and observing dendrite formation. We observe the character and morphology of lithium electrodeposition and then dendrite initiation and growth rates at three temperatures: — 10°C, 5°C and 20°C representing 3 distinct regimes of "uncontrolled", "intermediate" and "suppressed" growth according to the Akolkar model, which are related to large, moderate and small ratios of tip-to-surface current densities, respectively. We also monitor the

* Electrochemical Society Active Member. zE-mail:

number of dendrites formed at each temperature and compare dendrite morphology to changes in the charge curve. The quantifiable features are: (1) the observed initiation time, ti, the time at which the first dendrite is optically apparent from the lithium electrode surface at a distance of greater than 10 microns; (2) the dendrite growth rate, vd; and (3) short-circuit time, tsc, the time required for lithium dendrites to extend through the 2 mm gap in the Li0|Li0 cell. The combination of results sheds practical information on the roles that kinetics and mass transport play in deleterious dendrite formation and provide experimental data to support future modeling efforts.


Symmetric cells of Li0|Li0 positive and negative electrodes were prepared by rolling lithium metal (Aldrich, t = 38 |xm) on 5 mm wide copper strips (All Foils, t = 25.4 |im) etched prior to the experiment in 0.1 M HNO3. In-situ optical microscopy cells were constructed by mounting two "open-faced" electrodes between optically transparent quartz windows, as shown schematically in the inset of Figure 1A and in other works.5,11,12 A PTFE gasket was placed between the two optical windows in order to prevent Li deposition to the front and back surfaces of the Li electrode, leaving only the rectangular cross-section exposed to electrolyte. The cell was sealed with a rubber gasket to contain the liquid electrolyte, 1 MLiPF6 inEC:DMC (1:1 w/w), which was injected into the closed cell via a syringe.

The in-situ cells were introduced into environmental chambers held constant at —10°C, 5°C and 20°C. The cells remained in the isothermal chamber during application of charge current and microscopy image collection. Constant current charging was performed with a Mac-cor Model 4300 battery tester at a current density of approximately, i = 5 mA • cm—2. Optical micrographs were collected using a Navitar Zoom 6000 zoom lens system outfitted with a Luminera 2 megapixel color digital camera with 1.0x lens adapter and 0.25x lens attachment. Digital images were collected at 50 s intervals until the electrodeposits had grown sufficiently to short the cell, dropping the cell voltage to 0 V. The initial dendrite growth rate was calculated from the progression of micrographs collected during charging.

Results and Discussion

Figure 1A shows the voltage vs. time data for representative cells at the three temperatures and Figure 1B illustrates the dendrite morphology and growth pattern at the noted points along the voltage curves (Points 1-15). The open circuit potential for the Li0|Li0 electrochemical cells at each temperature condition is nearly 0 V at the beginning of the test. As current is applied the cell voltage increases rapidly (Points 1 and 6) due to the increased surface overpotential needed to drive the charging process, and then the voltage starts to

Figure 1. (A) Voltage vs. time under an applied constant current for three temperature conditions. Inset shows a schematic of the Li0|Li0 cell used for in-situ optical viewing. (B) Optical micrographs taken from the imaging window illustrating dendritic growth and morphology at select points throughout the charge cycle (identified numerically) at three temperatures: — 10°C (1-5), 5°C (6-10) and 20°C (11-16). White arrows are used to identify individual dendrite nucleation and to illustrate growth over time. White asterisks are used to identify needle-like dendrite morphologies. Red "X" marks illustrate the location where the dendrite makes electrical contact with the cathode forming a hard short-circuit. The distance between the electrodes is 2 mm.

decay, indicative of a decrease in cell resistance (V = iR). Several features in the voltage vs. time plot appear during dendrite initiation and growth periods. For instance, at —10°C after the large initial polarization loss a slight voltage shoulder emerges which coincides with the end of the dendrite initiation period. At 5°C, a small polarization voltage drop is followed by a voltage plateau which endures until just after the last dendrite has initiated, Point 7 in Figure 1A. These effects

are less noticeable at 20°C where 2 small voltage plateaus (Points 11 and 12) occur within the initiation range but do not bookend the initiation times of the first and last dendrite. Also, electrodeposition at 20° C begins as uniform plating which does not produce a steep drop in cell voltage. Voltage plateaus and "peaks" have also been described by Fan to indicate lithium electrodeposition on graphite at low temperatures.13 In Figure 1A these voltage features occur within the initiation times observed for lithium electrodeposition. Similar variations in cell potential have been shown by Brissot et al. which reflect the evolution of dendrite initiation and growth as well as the formation of a short-circuit.14 The time required for dendrites to short-circuit the cell is easily distinguished at Points 5, 10 and 15 (—10°C, 5°C and 20°C, respectively).

Close inspection of Figure 1B shows four temperature dependent morphologies observed during the electrodeposition of lithium as a result of formation temperature: dense mushroom-shaped electrode-posits at — 10°C, jagged edged particulates at 5°C and thin needle-like wires and films at 5°C and 20° C. The dense mushroom-like dendrites at — 10°C grow outwardly in a uniform manner. Specific dendrites are noted in the figure with white arrows. Similar morphology was observed by Lu et al. where low current density produced mushroomlike dendrites while higher current densities formed more needle-like protrusions.15 The formation of a short is noted with a red X. At 5°C and 20°C, the lithium dendrites are micron-scale lithium needles (noted with an asterisk, *) which kink to form needles and wound-ball morphologies. The diameter of the wound wire balls is higher at 5°C than 20°C. The morphology of the wound wire balls at 5°C and 20°C is similar to the "aggregate" and "tangled" morphologies of dendrites formed on the surface of lithium at room temperature, as described by Orsini.16

Figure 2 quantifies the results of dendrite formation in replicate cells at —10°C, 5°C and 20°C. The initiation times, ti, for the dendrite formation are plotted in Figure 2A. As expected, ti increases with increasing temperature, as dendrite initiation time is known to increase proportionally to the diffusion constant of the ionic species.17 The ti is smallest at —10°C, where dendrites form rapidly and no planar plating-type electrodeposition is observed. By comparison at 20°C, dendrites form much later after an initial period of uniform Li-plating. Figure 2B shows the growth rates, vd, of individual dendrites formed over the three charging temperatures. The cumulative number of dendrites at —10°C is the highest (23 dendrites) followed by 5°C (10 dendrites) and 20°C (9 dendrites) for three replicate cells at each temperature condition. Generally a large number of den-drite nucleates is due to high overpotential, the result of high current density18 or as we observe, low temperature. After an initial period of uniform plating at 20°C, a transition to dendritic growth occurs where the growth rates are closely grouped between 0.10.2 |xm • sec—1. At 5°C the growth rates are bimodal showing slow (0.03-0.15 |im • sec—1) and quick (0.25-0.4 |im • sec—1) growth rates. Needle-like dendrites with high aspect ratios grow faster than blunt rounded dendrites since electrodeposition occurs largely unidirectional on a tip rather than expanding in the x and y-directions of a flat surface.

The short-circuit time, tsc, is the time required for lithium den-drites to extend through the 2 mm gap between the positive and negative Li electrodes, resulting in an internal short. Figure 2C shows that the cells at 5°C have the lowest tsc. While dendrites form more quickly at —10° C, the morphology of the dendrites also contribute to tsc, implying that at 5°C the dendrites are the most cohesive. The mushroom-like morphologies at —5°C struggle to span the gap, while the micron-scale Li needles formed at 20° C easily break free due to kinks creating isolated metallic particulates. Cell shorting only occurs when these needles are oriented in a percolation network capable of providing an electrical pathway between the positive and negative electrodes.

The temperature dependency of initiation time along with den-drite morphology effects on the time to short-circuit a cell are given qualitatively as an empirical multi-variable relationship in Equation 2, where l is the distance between electrodes, i is the applied current and

Figure 2. Dendrite attributes at — 10°C, 5°C and 20°C for (A) initiation times, ti, in hours required to form the first 3 Li dendrites over three replicate cells; (B) growth rates, Vd, of individual dendrites formed at each temperature condition (— 10°C: 23 dendrites; 5°C: 10 dendrites; 20°C: 9 dendrites) for three replicate cells; and (C) short-circuit times, tsc, in hours required for dendrites to span the 2 mm gap between the electrodes and short the cell.

T is temperature.

tsc = tif (T ) +

vdf (i, T, morphology)

+ morphology f (T) [2]

Initiation time and growth rate are dependent upon ionic mobility and electric field; however, the time to short-circuit is also dependent upon dendrite morphology (which also affects growth rate) and varies with temperature. Not considered here in Equation 2 is the influence of the SEI layer on dendrite morphology. Growth of the Li needle-like morphology, favorable to cell shorting, is due to local variations in the thickness and corresponding impedance of the SEI layer surrounding the electrodeposit.3,19 SEI layers covering dendrite tips are typically thinner, which translates into a higher tip current density and accelerated growth. As a result, Li electrodeposition occurs at a much higher rate. We observe that the interrelation of fast dendrite formation (low ti), quick growth (high vd) and a jagged and spiky morphology capable of bridging the gap between the electrodes at 5°C results in the shortest time to short-circuit.


The growth and morphology of lithium electrodeposits are compared at —10°C, 5°C and 20°C and the Li0|Li0 cells fail most rapidly at 5°C. Dendrite initiation times were rapid at —10°C and 5°, while longer initiation times were found at 20° C. The morphology of elec-trodeposits change with temperature where low temperature favored mushroom-shaped deposition while Li needles formed wound-balls and particulates at 5°C and 20°C, respectively. Qualitatively, our results support the Akolkar model. We observe fast dendrite formation (i.e. high current at dendrite tip) and no Li plating at —10°C (i.e. low Li plating current), a combination of the two at 5°C and more pronounced Li plating at 20° C. However, additional factors such as dendrite morphology must be considered to more accurately predict the time required for dendrites to form an internal short circuit.

Expansion of this work is underway to characterize the mechanical properties of the various morphologies of lithium dendrites as well as the electro-mechanical coupling of lithium dendrites and polymer battery separators at low temperature to gain insight into short-induced failures in lithium-ion batteries. Electrolyte additives which slow the kinetics of dendrite growth, improve mass transport through the electrolyte and favor rounded or mushroom-shaped dendrite morphologies may improve safety of lithium-ion batteries during low temperature charging. Additionally, the effects of charge

rate, cathode-to-anode mass ratios and changes in internal cell resistance should be explored using the in-situ optical microscopy approach. Low temperature charging should be avoided to prevent fast initiation and growth of lithium dendrites capable of forming internal short circuits. The general observations of this work, the need for improved mass transport and favorable dendrite morphology, can be applied to other lithium systems, Li-air and Li-S, while the experimental approach has broad applicability to other electrochemical energy storage technologies including Zn-air and flow batteries.


The authors thank the Office of Naval Research for financial support of this work. The authors acknowledge the assistance of Dr. Benjamin Gould and Drew Rogers for fabrication of gaskets used to properly seal the in-situ optical microscopy cell.


1. N. Williard, et al., "Lessons Learned from the 787 Dreamliner Issue on Lithium-Ion Battery, Reliability," Energies, 6(9), 4682 (2013).

2. M. C. Smart and B. V. Ratnakumar, "Effects of Electrolyte Composition on Lithium Plating in Lithium-Ion Cells," Journal of the Electrochemical Society, 158(4), A379 (2011).

3. A. Heller, "The G. S. Yuasa-Boeing 787 Li-ion Battery: Test It at a Low Temperature and Keep It Warm in Flight," in Interface., The Electrochemical Society (2013).

4. R. Akolkar, "Modeling dendrite growth during lithium electrodeposition at sub-ambient temperature," Journal of Power Sources, 246, 84 (2014).

5. S. J. Harris, et al., "Direct in situ measurements of Li transport in Li-ion battery negative electrodes," Chemical Physics Letters, 485(4—6), 265 (2010).

6. J. Steiger, D. Kramer, and R. Monig, "Mechanisms of dendritic growth investigated by in situ light microscopy during electrodeposition and dissolution of lithium," Journal of Power Sources, 261, 112 (2014).

7. M.S. Park, et al., "A Highly Reversible Lithium Metal Anode," Sci. Rep., 4 (2014).

8. M. Rosso, et al., "Dendrite short-circuit and fuse effect on Li/polymer/Li cells," Electrochimica Acta, 51(25), 5334 (2006).

9. P. C. Howlett, D. R. MacFarlane, and A. F. Hollenkamp, "A sealed optical cell for the study of lithium-electrode electrolyte interfaces," Journal of Power Sources, 114(2), 277 (2003).

10. K. Nishikawa, et al., "In Situ Observation of Dendrite Growth of Electrodeposited Li Metal," Journal of the Electrochemical Society, 157(11), A1212 (2010).

11. T. Nishida, et al., "Optical observation of Li dendrite growth in ionic liquid," Electrochimica Acta, 100, 333 (2013).

12. C. Brissot, et al., "Concentration measurements in lithium/polymer-electrolyte/ lithium cells during cycling," Journal of Power Sources, 94(2), 212 (2001).

13. J. Fan and S. Tan, "Studies on charging lithium-ion cells at low temperatures," Journal of the Electrochemical Society, 153(6), A1081 (2006).

14. C. Brissot, et al., "Dendritic growth mechanisms in lithium/polymer cells," Journal of Power Sources, 81 925 (1999).

15. Y. Lu, Z. Tu, and L. A. Archer, "Stable lithium electrodeposition in liquid and nanoporous solid electrolytes," Nat Mater, 13(10), 961 (2014).

16. F. Orsini, etal., "In situ Scanning Electron Microscopy (SEM) observation of interfaces within plastic lithium batteries," Journal of Power Sources, 76(1), 19 (1998).

17. A. Rosso, etal., "Onset of dendritic growth in lithium/polymer cells," Journal of Power Sources, 97-98, 804 (2001).

18. K. Nishikawa, et al., "Li dendrite growth and Li+ ionic mass transfer phenomenon," Journal of Electroanalytical Chemistry, 661(1), 84 (2011).

19. R. A. Huggins, Advanced Batteries, Springer (2009), p. 474.