ArticlePDF Available

Mapping Mechanisms and Growth Regimes of Magnesium Electrodeposition at High Current Densities

Authors:

Abstract and Figures

The utilization of metallic anodes holds promise for unlocking high gravimetric and volumetric energy densities and is pivotal to the adoption of ‘beyond Li’ battery chemistries. Much of the promise of magnesium batteries stems from claims regarding their invulnerability to dendrite growth. Whilst considerable effort has been invested in the design of novel electrolytes and cathodes, detailed studies of Mg plating are scarce. Using galvanostatic electrodeposition of metallic Mg from Grignard reagents in symmetric Mg-Mg cells, we establish a phase map characterized by disparate morphologies spanning the range from fractal aggregates of 2D nanoplatelets to highly anisotropic dendrites with singular growth fronts and nanowires entangled in the form of mats. The effects of electrolyte concentration, applied current density, and coordinating ligands have been explored. The study demonstrates a complex range of electrodeposited morphologies including canonical dendrites with shear moduli conducive to penetration through typical polymeric separators. We further demonstrate a strategy for mitigating Mg dendrite formation based on the addition of molecular Lewis bases that promote nanowire growth through selective surface coordination.
Content may be subject to copyright.
This journal is ©The Royal Society of Chemistry 2020 Mater. Horiz., 2020, 7,843--854 | 843
Cite this: Mater. Horiz., 2020,
7,843
Mapping mechanisms and growth regimes of
magnesium electrodeposition at high current
densities
Rachel Davidson,
ab
Ankit Verma,
c
David Santos,
ab
Feng Hao,
c
Cole D. Fincher,
d
Dexin Zhao,
b
Vahid Attari,
bd
Parker Schofield,
ab
Jonathan Van Buskirk,
ab
Antonio Fraticelli-Cartagena,
a
Theodore E. G. Alivio,
ab
Raymundo Arroyave,
bd
Kelvin Xie,
b
Matt Pharr,
d
Partha P. Mukherjee *
c
and Sarbajit Banerjee *
ab
The utilization of metallic anodes holds promise for unlocking high
gravimetric and volumetric energy densities and is pivotal to the
adoption of ‘beyond Li’ battery chemistries. Much of the promise of
magnesium batteries stems from claims regarding their lower pre-
dilection for dendrite growth. Whilst considerable effort has been
invested in the design of novel electrolytes and cathodes, detailed
studies of Mg plating are scarce. Using galvanostatic electrodeposition
of metallic Mg from Grignard reagents in symmetric Mg–Mg cells, we
establish a phase map characterized by disparate morphologies span-
ning the range from fractal aggregates of 2D nanoplatelets to highly
anisotropic dendrites with singular growth fronts and nanowires
entangled in the form of mats. The effects of electrolyte concen-
tration, applied current density, and coordinating ligands have been
explored. The study demonstrates a complex range of electrodepos-
ited morphologies including canonical dendrites with shear moduli
conducive to penetration through typical polymeric separators. We
further demonstrate a strategy for mitigating Mg dendrite formation
based on the addition of molecular Lewis bases that promote nano-
wire growth through selective surface coordination.
Introduction
Lithium-ion batteries are currently the dominant electrochemical
energy storage technology with accessible gravimetric and
volumetric energy densities approaching 250 W h kg
1
and
600 W h L
1
, respectively.
1,2
Current Li-ion batteries pair transi-
tion metal oxide cathodes with graphite anodes;
3
supplanting the
latter with metallic lithium would yield theoretical capacities as
a
Department of Chemistry, Texas A&M University, College Station, TX 77843, USA. E-mail: banerjee@chem.tam u.edu
b
Department of Materials Science & Engineering, Texas A&M University, College Station, TX 77843, USA
c
School of Mechanical Engineering, Purdue University, West Lafayette, IN 47907, USA. E-mail: pmukherjee@purdue.edu
d
Department of Mechanical Engineering, Texas A&M University, College Station, TX 77843, USA
Electronic supplementary information (ESI) available: Videos: S1. Time lapse video of growth from 0.921 mA cm
2
in 0.25 M MeMgCl solution for 24 h reactions
shown at 4000speed. S2. Time lapse video of growth from 0.921 mA cm
2
in 0.5 M MeMgCl solution for 24 h reactions shown at 4000speed. S3. Time lapse video of
growth from 0.921 mA cm
2
in 1.0 M MeMgCl solution for 24 h reactions shown at 4000speed. S4. Time lapse video of growth from 0.921 mA cm
2
in 1.5 M MeMgCl
solution for 24 h reactions shown at 4000speed. S5. Time lapse video of growth from 0.921 mA cm
2
in 2.0 M MeMgCl solution for 24 h reactions shown at 4000
speed. S6. Time lapse video of growth from 0.921 mA cm
2
in 0.5 M MeMgCl solution with addition of 0.0626 M dodecanethiol for 24 h reactions shown at 4000
speed. S7. Time lapse video of growth from 0.921 mA cm
2
in 0.5 M MeMgCl solution with addition of 0.125 M dodecanethiol for 24 h reactions shown at 4000speed.
S8. Time lapse video of growth from 0.921 mA cm
2
in 0.5 M MeMgCl solution with addition of 0.188 M dodecanethiol for 24 h reactions shown at 4000x speed. S9. The
alighted tilt series of soft X-ray microscopy images of a fractal grown at 0.921 mA cm
2
in 0.5 M MeMgCl for 24 h obtained at the Mg K-edge in transmission (left) and
optical density (right). S10. 3D reconstruction series of soft X-ray microscopy images of a fractal grown at 0.921 mA cm
2
in 0.5 M MeMgCl for 24 h obtained at the Mg
K-edge in transmission (left) and optical density (right). S11. 3D digital tomography of growth from 0.921 mA cm
2
in 0.5 M MeMgCl solution for 24 h reactions. S12. 3D
digital tomography of growth from 0.921 mA cm
2
in 2.0 M MeMgCl solution for 24 h reactions. See DOI: 10.1039/c9mh01367a
Received 28th August 2019,
Accepted 25th November 2019
DOI: 10.1039/c9mh01367a
rsc.li/materials-horizons
New concepts
Here, we explore electrodeposition of magnesium under varying electric
fields, electrolyte concentrations, and added ligands. Distinctive growth
mechanisms are differentiated including fractal and dendritic growth
regimes, which are rationalized based on the dynamical interplay between
electrochemical reaction and self-diffusion rates. Limitations of current
batteries represent perhaps the largest roadblock to the continued
advancement of renewable energy technologies. Supplanting the graphite
used in Li-ion batteries with metallic anodes holds promise for significantly
enhanced capacity and energy density but requires mitigating the proclivity
of lithium to deposit as dendrites. The ‘beyond Li’ paradigm of energy
storage has attracted consideration attention with much of its promise
derived from the utilization of metallic anodes that are saferin comparison
to lithium. The manuscript presents characterization of electrodeposition
products across multiple length scales.We note unprecedentedsingle crystal
growth of Mg dendrites, which has not heretofore been reported and has no
parallels in the lithium dendrite literature. Mg dendrites are found to be
substantially harder than their lithium counterparts, which further
underscores the need for stiffer separators. The addition of dodecanethiol
alters growth dynamics leading to consistent isolation of nanowires and
mitigation of dendritic growth.
Materials
Horizons
COMMUNICATION
Published on 25 November 2019. Downloaded by Texas A & M University on 7/24/2020 4:50:20 PM.
View Article Online
View Journal
| View Issue
844 |Mater. Horiz., 2020, 7, 843--854 This journal is ©The Royal Society of Chemistry 2020
high as 3860 mA h g
1
.
4
However, Li metal has a high propensity
for dendrite formation; the plating of lithium as anisotropic
fractal structures that can bridge across liquid and solid electro-
lytes, thereby short-circuiting the cell, represents a major safety
hazard. Consequently, the paucity of scalable methods to achieve
reproducible electroplating of metallic lithium has emerged as a
substantial roadblock to accessing improved storage capacities.
5,6
Dendrite formation has been the scourge even when utilizing
graphite anodes wherein under specific temperature, voltage, and
electrolyte decomposition conditions, dendritic growth regimes
become more favourable as compared to insertion reactions.
Indeed, numerous high-profile incidents have underscored the
importance of understanding the accumulative impact of low-
probability, stochastic processes in electrochemical energy storage
systems wherein fundamental processes operate across multiple
decades of time and length scales. Developing experimental
conditions that replicate such local far-from-equilibrium beha-
viour has thus emerged as an urgent imperative. Considerable
effort has been invested in the development of ‘‘beyond Li’’
intercalation systems that derive a considerable portion of their
promise from the potential to utilize their respective metallic
anodes. Sodium, magnesium, calcium, and zinc are considered to
deposit with much lower propensities for dendrite formation as
compared to lithium owing to their more facile self-diffusion,
which thereby results in the plating of relatively homogeneous
deposits.
7–9
Magnesium batteries are considered a promising alternative
given the divalent charge of Mg, which has been proposed as a
means of achieving higher energy densities since most cathode
materials are limited in terms of their available redox sites and
not accessible redox states. In addition, magnesium holds
promise for enabling use of metal anodes as a result of
its supposed ‘‘non-dendrite’’ forming nature.
10–14
Grobhas
attributed the low propensity for dendrite formation to small
self-diffusion barriers and vanishingly small Ehrlich–Schwo
¨bel
barriers for 3D diffusion. Much research has targeted the
development of novel cathode materials that can readily diffuse
highly polarizing divalent Mg-ions as well as in the develop-
ment of electrolytes stable across extended potential windows
that allow for effective desolvation of magnesium at electrode
interfaces.
12,15–20
Ideas regarding the permeability or lack
thereof of divalent Mg-ions through solid electrolyte interfaces
(SEI), which may form through degradation of electrolytes
during cycling, have inspired the design of several stable
classes of electrolytes.
12,21–23
Several experimental observations of homogeneous plating
as compared to agglomerate formation support the idea of
a reduced predilection of magnesium towards formation of
dendritic structures.
10,13,14,24–26
Dual-salt electrolytes contain-
ing both Li and Mg components have been considered as a
means of utilizing the faster kinetics of Li at the cathode whilst
avoiding Li dendrite formation through preferential plating of
Mg at the anode.
27,28
The faster surface diffusion of Mg-ions
along the Mg(0001) plane predicted from first-principles calcu-
lations has been put forth as the intrinsic basis for reduced
propensity for dendritic growth and is further corroborated by
the prediction of low diffusion barriers for diffusion across
steps and terraces.
7
Self-diffusion coefficients, Ehrlich–Schwo
¨bel
barriers, and anisotropy resulting from the intrinsic crystal struc-
ture have emerged as some putative descriptors for comparing the
dendrite-forming nature of different anode materials.
26,29–32
While
reports of reduced propensity for dendrite growth in magnesium
are well founded, it is worth noting that electrodeposition processes
often occur far from equilibrium wherein otherwise reliable
descriptors can be thwarted by other vectors.
33
Inhomogeneities
in magnesium deposition are not unprecedented
34–36
and capacity
fading analogous to the problems discussed with lithium has been
observed.
37,38
Recently Bitenc and co-workers showed highly
uneven deposition in MgCl
2
–AlCl
3
–DME electrolyte systems.
36
Grob
and co-workers have pointed out that surface self-diffusion in itself
cannot explain the deposition characteristics; the applied current
density is an equally important measure, which determines the
incoming reactant flux.
39–41
Yet, comprehensive investigations of
non-equilibrium phase spaces and Mg electrometallurgy are scarce
even though reports of fractal Mg microstructures within alloys are
abundant in the metallurgy literature.
42,43
Fractal and dendritic magnesium deposits have indeed been
observed upon the electrodeposition of Grignard reagents
12
in
Mg–Mg symmetric cells monitored in situ with videomicroscopy
under galvanostatic conditions. In this article, overpotentials
required for electrocrystallization of Mg at varying concentra-
tions and current densities are explored, and distinctive growth
morphologies are delineated including unambiguous fractal
and dendritic growth regimes. Deposition is seen to be under-
pinned by diffusion-limited aggregation (DLA) mechanisms
across much of the examined reaction space.
6,44–51
The Mg
deposits have been extensively explored across different length
scales utilizing a combination of electron and X-ray micro-
scopy. The experimental observations are explained with refer-
ence to an analytical framework contrasting the Mg
2+
diffusive
transport and reaction rates wherein exacerbated electrodepo-
sition instabilities are anticipated beyond the ‘‘Sand’s time’’
limit at elevated current densities.
52
Furthermore, phase-field
modeling studies have been used to unravel the mechanistic
underpinnings of the observed electrodeposited morphologies.
Results and discussion
Formation and characterization of fractal Mg structures:
developing a phenomenological map of deposition regimes
Electrodeposition of metallic Mg from MeMgCl and EtMgCl in
tetrahydrofuran (THF) has previously been shown to yield
continuous thin film and nanowire array morphologies; the
latter has been proposed to result from a modified faces, steps,
and kinks mechanism governed primarily by the deposition
rate.
53
While these electrolytes have limited stability windows,
they have been extensively used for Mg electrodeposition and
serve as effective model systems as compared to multicomponent
electrolytes. The utilization of a symmetric cell geometry to
examine electrocrystallization of Mg as will be discussed here
mitigates the influence of convoluting factors such as insertion
Communication Materials Horizons
Published on 25 November 2019. Downloaded by Texas A & M University on 7/24/2020 4:50:20 PM.
View Article Online
This journal is ©The Royal Society of Chemistry 2020 Mater. Horiz., 2020, 7,843--854 | 845
reactions, electrolyte decomposition at the cathode, and dissolu-
tion of the cathode as a result of parasitic reactions. The use of Mg
ribbon electrodes further allows for direct observation of intrinsic
phenomena without potential confounding factors such as electro-
catalytic processes at transition metal electrodes. Nevertheless,
similar results are obtained for Pt, stainless steel, A36 steel, and
galvanized steel. Application of a voltage in a parallel-plate
geometry yields a variety of morphologies of Mg spanning the
range from aggregated polycrystalline quasi-spherical deposits to
dendrites spanning millimeters in length, aggregated platelets,
and nanowires, depending on the current density, concentration,
and presence of coordinating ligands (vide infra). Videos S1–S8
(ESI) illustrate time-lapse images of Mg deposition as a function
of varying concentration of MeMgCl (Videos S1–S5, ESI)and
concentration of added dodecanethiol (Videos S6–S8, ESI).
Fig. 1A shows a phenomenological map illustrating the
different observed growth regimes for electroplating of Mg,
indicating considerable complexity as well as clear dendritic
growth windows in the multidimensional space. The plot charts
out correlations between processing conditions and mesoscale
texture and microstructure evolving from the interplay between
thermodynamics and kinetics of Mg electrodeposition. Intrigu-
ingly, this richness of electrodeposited Mg morphologies does
not appear to have been previously reported in the literature
even for these common electrolytes. Generally, upon increase in
concentration of the electrolyte, an increase in the grain size of
the deposit is observed resulting in a transition from highly
fractal growths formed from aggregation of hexagonal platelets
to aggregates of quasi-spherical deposits and finally converging
towards stabilization of highly crystalline dendritic deposits
with singular dominant growth fronts. Such morphologies
represent anisotropic growth regimes, which could detrimen-
tally impact battery performance; mapping such mechanisms is
imperative in order to systematically tune the nature of electro-
deposited films and to enable identification of consistent,
controllable, and stable plating windows. Fig. 1 depicts, as will
be discussed below, that the inclusion of dodecanethiol yields
nanowire morphologies in the form of mats, which may offer a
route to the design of cyclable high-surface-area metal anodes. In
the sections below, we will discuss this phase space across multi-
ple length scales while delineating observations from monitoring
the evolution of mesoscale morphologies, resulting microstruc-
ture, and crystal structure for each distinctive regime.
Mesoscale and higher length scale plating morphologies
have been monitored using videomicroscopy (Videos S1–S8,
ESI). Fig. 1B depicts a typical fractal deposit formed from the
electrodeposition of Mg from a 0.5 M solution of MeMgCl
in THF at a constant current density of 0.921 mA cm
2
. The
deposits span several millimeters in length, are highly
branched, and grow from the edges of the Mg ribbon. Fig. 1C
shows a SEM image of the same deposits depicted in Fig. 1B.
SEM images of the fractal deposits indicate aggregates of
hexagonal platelets characteristic of the intrinsic habit of hcp
Mg. Crystallographic information has further been derived
from high-resolution TEM and XRD in order to understand
the electrocrystallization process.
Powder XRD patterns of all deposits exhibits sharp reflec-
tions that can be readily indexed to PDF 35-0821, corres-
ponding to metallic magnesium as is shown in Fig. 1D for
the fractal and dendritic deposits. XPS spectra have further
been acquired for fractal deposits to examine the elemental
composition of their surfaces. Samples were exposed briefly to
ambient environments during loading of the substrates within
the instrument. Fig. S1A (ESI) shows a survey scan, whereas
high-resolution scans for Mg 2p, O 1s, C 1s, and Cl 2p are
shown in Fig. 1SB–E (ESI), respectively. The Mg 2p high-
resolution XPS spectrum exhibits the presence of zero-valent
Mg at 49.5 eV. Some samples additionally show a smaller
second peak at 52.6 eV, which can be ascribed to surficial
Mg–Cl known to exist as a key passivating species in the
electrodeposition of Grignard reagents,
54
as well as a feature
centered at 55.9 eV arising from the Fe 3p spectrum of
impurities resulting from the steel electrode clips. As the clips
were not submerged in solution during the reaction, the
influence of Fe on the characteristics of deposits was considered
to be negligible and is an artifact of washing the electrodes
following the reaction (the Fe signal is not observed in samples
where just the electrodes are washed). The oxygen 1s XPS spec-
trum shows a prominent peak centered at 531.4 eV, which can be
assigned to Mg(OH)
2
. A weak shoulder at 529.9 eV is additionally
observed likely arising from MgO and at 533.5 eV ascribed to the
presence of surface-bound ether species given the strong com-
plexation of THF and ethers to magnesium.
55,56
High resolution
Fig. 1 Fractal growth of electrodeposited Mg. (A) Phenomenological map
depicting several differentiated growth regimes as a function of reaction
variables. 2D diffusion-limited-aggregation-type growth, regions with
spherical diffusion-limited aggregation growth, dendritic growth, and
nanowire growth are distinguishable across this parameter space. Char-
acterization of Mg deposits obtained at a constant current density of
0.921 mA cm
2
from a 0.5 M solution of MeMgCl in THF. (B) Digital
photograph of a magnesium fractal deposit; (C) SEM image showing a
high-magnification view of the fractal surface; clear hexagonal habits can
be discerned. (D) Powder XRD patterns acquired for detached Mg deposits
grown from 0.5 and 1.5 M MeMgCl in THF.
Materials Horizons Communication
Published on 25 November 2019. Downloaded by Texas A & M University on 7/24/2020 4:50:20 PM.
View Article Online
846 |Mater. Horiz., 2020, 7, 843--854 This journal is ©The Royal Society of Chemistry 20 20
scansoftheC1sregionshowadventitiouscarbonaswellas
smaller peaks at 288.2 eV and 289.4 eV, which can be assigned to
carboxylates and carbonates, respectively.
57
Fig. S2A–C (ESI) indicate projections of 3D tomography
maps constructed using soft-X-ray microscopy at the Mg
K-edge. Videos S9 and S10 (ESI) show the resulting aligned
tilt series and the 3D reconstruction, respectively, in terms of
the transmission intensity (left) and optical density (right).
The fractal aggregate structures are observed to be solid with
faceted surfaces.
In situ observations of dendrite growth under varying
deposition conditions
Studies of fractal growth in metallic copper and zinc deposits
have shown that various experimental parameters affecting the
reactivity or diffusion of the electrolyte allow for tuning of
the crystallinity as well as the compactness of the plated
deposits.
58–61
Bazant noted that considerations such as the
anisotropy of crystal structures or the high activity of light
metals add complexity but do not fundamentally alter the
influence of these parameters.
48
Magnesium electrodeposition
from Grignard’s agents in THF solution has been first mon-
itored as a function of the applied current density for an overall
duration of 8 h from 0.5 M THF solutions of MeMgCl. Digital
photographs indicating the formation of fractal structures at 2,
4, 6, and 8 h time points are depicted in Fig. 2. Increasing the
current density increases the extent of deposition and yields
more heavily branched deposits. This observation as well as the
lack of extended crystalline order within the deposits suggests
the operation of a diffusion-limited aggregation (DLA) mecha-
nism, as has been observed for dendritic lithium growth.
48,62
Higher certainty of reduction of metal ions at a given site
(oftentimes quantified using a ‘‘sticking coefficient’’
63,64
)
resulting from the increased driving force for deposition at
higher current densities results in more extensive fractal
growth. The flux and reaction rates under these conditions
overcome the relatively fast self-diffusion predicted for Mg.
7
Table 1 shows the resulting weights of the fractal product and
overpotentials required to maintain the constant current con-
ditions. Generally, there is an increase in the overpotential with
increasing current density; the resulting mass of fractal depos-
its is furthermore increased. The analytically predicted total Mg
deposition is also tabulated as anticipated from Faraday’s law;
detailed analysis is presented in the latter half of this article.
The conditions correspond to relatively high current densities,
but it is worth noting that proposed fast charging applications
will indeed necessitate high current fluxes. Corresponding
voltage over time plots are shown in Fig. S3 (ESI).
The growth regimes have been additionally monitored as a
function of electrolyte concentration. Time lapse digital photo-
graphs acquired at 6, 8, 12, and 16 h intervals are shown in
Fig. 3 for different electrolyte concentrations in THF. Videos
exhibiting the progression of dendrite growth as a function of
time are shown in Videos S1–S5 (ESI) and the characteristics of
the deposited products are noted in Table 1. Corresponding
voltage versus time plots are shown in Fig. S4 (ESI). Increasing
Fig. 2 In situ videomicroscopy observations of fractal growth as a func-
tion of applied current density. Digital photographs have been acquired at
2, 4, 6, and 8 h time points for deposition from 0.5 M THF solutions of
MeMgCl solutions under different applied current densities (0.307, 0.921,
and 1.54 mA cm
2
).
Table 1 Mass of dendritic product and overpotential as a function of applied current density and concentration of MeMgCl. Depositions with varying
current density were performed for 8 h and in 0.5 M MeMgCl solutions. Reactions with varying electrolyte concentration were performed for 24 h at an
applied current density of 0.921 mA cm
2
Variation of applied current density
Current density (mA cm
2
)Predicted total deposition
mass of Mg (mg) Measured mass of dendritic
Mg (mg) Vh Volts (average) E(V mm
1
)
0.307 3.63 6.8 0.2 100.7 12.6 0.220
0.921 10.9 6.2 1.3 158.7 19.8 0.347
1.54 18.1 14.2 5.0 222.7 27.8 0.487
Variation of electrolyte concentration
MeMgCl concentration (M) Predicted total deposition
mass of Mg (mg) Measured mass of dendritic
Mg (mg) Vh Volts (average) E(V mm
1
)
0.25 32.64 21.6 9.0 568.0 23.7 0.414
0.50 32.64 27.8 14.3 466.3 19.4 0.340
1.0 32.64 9.1 1.6 37.7 1.6 0.027
1.5 32.64 13.9 3.9 10.3 0.4 0.008
2.0 32.64 12.1 6.8 7.0 0.3 0.005
Communication Materials Horizons
Published on 25 November 2019. Downloaded by Texas A & M University on 7/24/2020 4:50:20 PM.
View Article Online
This journal is ©The Royal Society of Chemistry 2020 Mater. Horiz., 2020, 7,843--854 | 847
MeMgCl concentration results in the formation of thicker, less
branched deposits, which is thought to be reflective of modifica-
tion in the growth mechanism. In addition, the microstructure of
the deposits is modified upon going from 0.25 to 0.5 M with the
0.25 M reactions yielding fractals constituted from much smaller
grains as can be seen more clearly in Fig. 4. The overpotential
generally decreases with increasing concentration for all samples
asaresultofthehighersolutionconductivity. Typically, electrolyte
ionic conductivity exhibits a non-monotonic trend with concen-
tration, increasing until an optimal concentration is reached,
beyond which it is diminished.
65
For MeMgCl in THF, a
steady decrease in overpotential is observed even up to con-
centrations of 2 M.
The morphologies observed upon non-equilibrium, fractal
growth are governed by a balance between local surface dynamics,
long-range diffusion, nucleation probabilities, and anisotropic
growth rates along different crystallographic directions.
66
Fig. 4
shows SEM images acquired at different magnifications for
deposits obtained from 0.25, 0.5, and 1.5 M solutions of MeMgCl
in THF (at a constant current density of 0.921 mA cm
2
), which
allow for different types of microstructures constituting the fractal
morphologies to be differentiated. Fig. 5 shows more extensive
crystallographic and nanomechanical characterization of the
deposits.
Three distinctive growth regimes can be distinguished with
considerable differences in the mode of aggregation and direc-
tionality of growth. The deposits are constituted from hexagonal
platelets as fundamental building blocks, preserving the
symmetry of the underlying crystal lattice. Energy minimized
Wulff reconstructed surfaces are discernible (Fig. 4C, F, and I),
which suggest that the low self-diffusion barriers in this system
indeed allow for thermodynamic shapes to be stabilized. How-
ever, the mesoscale orientation and attachment of the shapes
are highly variable as a function of the concentration and
current density. At low concentrations of 0.25 M MeMgCl and
high overpotentials, nucleation of new particles dominates over
growth of incipient nuclei resulting in fractals comprising
aggregates of numerous thin hexagonal platelets on the order
of around 3–6 mm in diameter. An increase in concentration of
MeMgCl results in a decrease in overpotential and greater
availability of ions at reactive sites. Consequently, the growth
rates are accelerated and the individual crystallites are sub-
stantially larger with a more spherical appearance (with end-to-
end dimensions of 30–60 mm, albeit still with some clearly
defined hexagonal facets) resulting in a considerably altered
Fig. 3 In situ videomicroscopy observations of fractal growth as a func-
tion of electrolyte concentration. Digital images acquired at 6, 8, 12, and
16 h time intervals for 0.25, 0.5, 1, 1.5, and 2.0 M concentrations of MeMgCl
in THF at a constant current density of 0.921 mA cm
2
.
Fig. 4 Fractal to dendrite transformation. SEM images acquired at varying
magnifications for deposits obtained at a constant current density of
0.921 mA cm
2
for (A–C) 0.25 M; (D–F) 0.5 M; and (G–I) 1.5 M solutions
of MeMgCl in THF. The top two rows exhibit fractal growth, whereas the
bottom row corresponds to a dendritic growth regime.
Fig. 5 Microstructural characterization of Mg dendrites. (A) SEM image of
a Mg dendrite electrodeposited under 0.921 mA cm
2
applied constant
current in a 1.5 M MeMgCl for 24 h; (B) higher magnification SEM image of
(A) illustrating regions from which EBSD and TEM specimens have been
extracted using FIB; (C) EBSD IPF map and 3D crystallographic schematic
of the Mg dendrite; (D) bright-field TEM image of the Mg dendrite and
corresponding SAED pattern. Representative nanoindentation (E) load-
depth curves, (F) elastic modulus versus depth, and (G) hardness versus
depth for Mg electrodeposits grown from 0.5 M and 2 M MeMgCl solutions
under 0.921 mA cm
2
applied constant current for 24 h.
Materials Horizons Communication
Published on 25 November 2019. Downloaded by Texas A & M University on 7/24/2020 4:50:20 PM.
View Article Online
848 |Mater. Horiz., 2020, 7, 843--854 This journal is ©The Royal Society of Chemistry 2020
fractal morphology as seen in Fig. 4D–F.
67
As described below,
growth under these conditions corresponds to a diffusion-
limited regime; as a result, the observed morphologies are
characteristic of diffusion-limited aggregation. At a still higher
concentration of 1.0 M MeMgCl, Fig. 3 suggests a notable
alteration of the deposition mechanism. SEM images of depos-
its obtained from 1.5 M THF solutions of MeMgCl (Fig. 4G–I)
indicate that increasing concentration brings about a transition
from fractal growth to stabilization of dendrites. The deposits
exhibit a singular dominant growth tip, albeit with somewhat
irregular branches (Fig. 4G–I). Video S4 and Fig. S5 (ESI)
depict lower magnification views of the growth tip (delineated
by red arrows in Fig. S5, ESI). It is worth noting that such
growth is distinctly different from the root-growing, needle-like
growth observed in lithium.
48,52
Dendritic growth with the
observed dominance of a finite number of growth fronts
requires the influence of anisotropy, which may be derived in
this case from the intrinsic asymmetry of the hcp crystal
structure or, extrinsically, as a result of preferential passivation
owing to electrolyte decomposition.
66,68,69
With diminishing
diffusion limitations, the effects of anisotropy are clearly dis-
cernible at both the micron- and mesoscale levels.
Thin platelet growth is furthermore observed upon the
addition of oleylamine (0.121 M) to the 0.5 M THF solution of
MeMgCl at a current density of 0.921 mA cm
2
, as shown in
Fig. 6A–C. Oleylamine, a Lewis basic ligand that weakly coordi-
nates to Mg-ions, is thought to buffer the monomer super-
saturation and allows for nucleation-dominated growth.
70,71
Surface passivation necessitates diffusion of monomer ions
through the capping layer and likely also alters self-diffusion rates.
XPS spectra for deposits formed through addition of oleylamine
are shown in Fig. S6 (ESI) and are very similar to that of spectra
observed for dendrites formed without the addition of oleylamine
with the addition of a characteristic N 1s signal and a shoulder
centered around 283.5 eV for the C 1s.
Characterization of Mg dendrites
The microstructure and the growth direction of the Mg
dendrites electrodeposited from 1.5 M MeMgCl solutions in
THF under 0.921 mA cm
2
constant current densities have
been examined by electron backscatter diffraction (EBSD) and
transmission electron microscopy (TEM) (Fig. 5). The dendrites
obtained under these conditions span hundreds of microns in
width and millimeters in length. Each dendrite comprises a
number of Mg crystals with well-defined crystal facets (Fig. 5A).
The EBSD and TEM samples have been prepared from an
individual branch of a Mg dendrite as shown in Fig. 5B,
obtained from the region in Fig. 5A highlighted with the white
rectangle; the lengths of the lift-out specimens are parallel to
the growth direction of the dendrite. The EBSD map, based on
the growth direction of the inverse pole figure (IPF) map and
IPF triangular reference, displays a uniform hue, indicating
that the examined part of the Mg dendrite is single crystalline.
The EBSD map reveals a growth direction of h11%
20i(Fig. 5C).
The single crystalline nature and growth direction of the Mg
dendrites have been further corroborated by TEM observations
in Fig. 5D. The corresponding selected area electron diffraction
(SAED) pattern (Fig. 5D, inset) confirms the h11%
20igrowth
direction. This growth preference can be rationalized consider-
ing that the most dense packing of atoms in hexagonal close-
packed Mg is along h11%
20i.
As seen in Fig. 5E–G, indentation measurements have been
used to derive elastic and plastic properties for bulk Mg as well
as Mg dendrites electrodeposited from 0.5 and 2 M concentra-
tions of MeMgCl in THF. Indentation of bulk Mg in Fig. 5F
yields an elastic modulus of 39.4 0.9 GPa, similar to pre-
viously reported values of ca. 40–45 GPa in the literature.
72,73
In
contrast, the 0.5 and 2 M electrodeposited Mg deposits exhibit
elastic moduli of 23.8 1.6 and 22.5 1.8 GPa, respectively. In
other words, the electrodeposited Mg structures possess an
elastic modulus nearly 60% that of bulk Mg. Optical observa-
tion of the indents (Fig. S7, ESI) does not reveal excessive
pile-up. Furthermore, consistent and flat E
2
/Hvalues at sub-
stantial depths as well as the frame stiffnesses’
74
favorable
comparison with that of the calibration material (fused silica)
provides further verification of the validity of these results.
Possible origins of the reduced elastic moduli observed for the
dendrites include the presence of porosity, impurities in the
electrodeposited Mg, and/or the influence of the grain size and
orientation of the electrodeposited Mg.
Analysis of plastic properties suggests that the electrodeposi-
tion parameters furthermore influence the resulting mechanical
properties of the Mg deposits. As seen in Fig. 5G, the indentation
of bulk Mg yields a hardness of 665 33 MPa. Assuming a Tabor
factor of 2.8, the yield strength of the bulk Mg can be estimated to
be B235 MPa.
73,75,76
At an indentation depth of 1500 nm, the Mg
Fig. 6 Ligand modification of Mg Morphologies. SEM images of electro-
deposited Mg obtained through addition of (A–C) oleylamine (0.121 M) or
(D–I) varying concentrations of dodecanethiol. Spherical clusters of
shorter wires have been observed upon addition of (D) 0.0626 M, (E)
0.125 M, and (F) 0.188 M dodecanethiol. These form extended structures
as can be observed in (G), which shows a representative example from a
reaction containing 0.0626 M dodecanethiol. In addition to clusters,
extended 1D wires are observed upon addition of higher concentrations
of dodecanethiol as observed upon the addition of (H) 0.125 M and
(I) 0.188 M dodecanethiol.
Communication Materials Horizons
Published on 25 November 2019. Downloaded by Texas A & M University on 7/24/2020 4:50:20 PM.
View Article Online
This journal is ©The Royal Society of Chemistry 2020 Mater. Horiz., 2020, 7,843--854 | 849
electrodeposited from 0.5 and 2 M MeMgCl in THF displayed
hardness values of 525 38 MPa and 415 18 MPa (corres-
ponding to yield strengths of B190 and 150 MPa), respectively.
Theoriginsofthedierencesinplasticpropertiesfrombulk
Mg remain unclear but again may be related to impurities (e.g.,
precipitates) or specifics of the microstructure (e.g., grain sizes,
dislocation densities) that form during electrodeposition under
different conditions.
A popular model for predicting conditions to prevent
dendrite formation is that of Newman and Monroe, who
considered electrode stability of electrode (lithium)/separator
(or solid electrolyte) interfaces using linear elasticity theory.
According to their model, dendrites can be suppressed by a
separator or solid electrolyte that has a shear modulus approxi-
mately twice that of the electrode itself.
77
Taking the elastic
modulus for the dendritic Mg as 25 GPa and the Poisson’s ratio
as 0.35,
78
the shear modulus of a dendrite can be calculated as
m=E/[2(1 + n)] = 10.0 GPa. As a result, the Newman and
Monroe
77
model predicts that a separator or solid state electro-
lyte with a shear modulus of more than B20 GPa will be
necessary to prevent the formation of Mg dendrites within a
battery. Since polymer separators typically have moduli on the
order of 1 GPa and a Poisson’s ratio of 0.46,
79
their shear
modulus of B340 MPa is much too small to prevent the
propagation of Mg dendrites. However, stiff ceramic solid-
state electrolytes with large shear moduli (425 GPa) may
suppress dendrites and thereby warrant further investigation.
Notably, both of these electrodeposited Mg morphologies
possess significantly larger elastic moduli and hardness values
as compared to Li (modulus of B9 GPa and bulk indentation
hardness of 4.5 MPa).
80,81
As a result, mechanically suppressing
dendritic growth may prove substantially more challenging
than in the case of Li.
Ligand modification of electrodeposition morphologies
The addition of dodecanethiol yields a pronounced change in
appearance, a gray powder is obtained at low concentrations of
dodecanethiol, whereas an entangled fibrous mat is recovered
at high concentrations. Fig. 6D–I show a pronounced modifica-
tion of the morphology upon the addition of dodecanethiol at
different concentrations. Powder XRD patterns for deposits
grown with addition of dodecanethiol can be indexed to
metallic Mg (PDF 35-0821, Fig. S8, ESI). XPS spectra of the
nanowires formed through the addition of 0.125 M dodeca-
nethiol are shown in Fig. S9 (ESI) and show similar features to
that of the dendrites formed without addition of the alkyl thiol,
with the addition of a S 2p band and a shoulder at around
283.5 eV for the C 1s spectrum. An initial reaction between
MeMgCl and dodecanethiol produces a thiolate species and
MgCl
+
; as such the dynamics of deposition is substantially
altered. Selective adsorption of the thiolate molecules on
specific growth facets and the ability of the Lewis basic ligands
to buffer the monomer supersaturation substantially reduces the
effective monomer flux.
70,82,83
Under these conditions, the self-
diffusion characteristics are comparable to the flux rate; conse-
quently, arrays of faceted nanowires with lateral dimensions of
250–800 nm are observed. Nanowires appear in two primary
forms; spherical clusters of shorter wires around 10–20 mmin
length are observed upon addition of 0.0626 M, 0.125 M, and
0.188 M dodecanethiol as shown inFig.6DF,respectively.As
shown in Fig. 6G, such nanowires furthermore form mesoscale
patterns through aggregation of the spheres. Still higher concen-
trations of dodecanethiol result in the stabilization of long Mg
nanowires on the order of many tens to hundreds of micrometers
in length (Fig. 6H and I); the nanowires form entangled mats
without the higher order aggregation observed at lower dodeca-
nethiol concentrations. This method of achieving the controlled
deposition of nanowire arrays furthermore provides a route to
nanotextured metallic anode films directly integrated onto the
current collector. The results demonstrate the ability to prepare a
disparate range of highly textured Mg anode films from electro-
platingofGrignardsreagents.Cyclingofnanowirearraysis
expected to yield improved reaction kinetics and a reduced local
overpotential owing to the greater availability of deposition sites,
thereby reducing the predilection for dendrite formation. The
utilization of such anodes in conjunction with dual salt electro-
lytes portends intriguing battery architectures designed to
mitigate dendrite formation.
27,28
Plating phase maps and mechanistic underpinnings
The morphology of electrodeposited Mg is governed by the
interplay of electrochemistry, ion transport, nucleation, and
crystal growth. Specifically, the balance between ion transport
in the electrolyte, Mg surface diffusion on the plating electrode,
and the electrochemical reaction rate dictate the observed
morphologies. At applied current rates, i
app
(A m
2
), exceeding
the limiting current density, i
lim
, for the electrochemical system
under observation, diffusional transport in the electrolyte can
become the limiting mechanism, resulting in the depletion of
Mg
2+
ions from the proximity of the plating electrode. As such,
transformation from smooth to dendritic structures is corre-
lated with this scarcity of Mg
2+
occurring at Sand’s time, t
Sand
,
given by
tSand ¼pDzc
0FðÞ
2
4iappta

2)D¼4iappta

2tsand
pzc0FðÞ
2(1)
Here, zis the cationic charge number, c
0
is the bulk salt
concentration in the electrolyte (mol m
3
), Fis Faraday’s
constant (C mol
1
), Dis the binary diffusion coefficient
(m
2
s
1
)andt
a
is the anionic transference number. Determination
of Sand’s time can help in accurate quantification of electrolyte
diffusivity, which is generally a monotonically decreasing
function of concentration owing to concentrated solution
effects and hence cannot be taken as constant. Further electro-
deposition beyond Sand’s time results in preferential growth of
dendritic structures. For our experiments, the Sand’s time
parameter values can be directly correlated to the amount
of dendritic magnesium, m, tabulated in Table 1 as per
Faraday’s law:
Ittotal tsand
ðÞ¼
zFmden
M(2)
Materials Horizons Communication
Published on 25 November 2019. Downloaded by Texas A & M University on 7/24/2020 4:50:20 PM.
View Article Online
850 |Mater. Horiz., 2020, 7, 843--854 This journal is ©The Royal Society of Chemistry 20 20
where Iis the applied current (A), t
total
is the total temporal
duration of the experiment (s), m
den
is the amount of dendritic
magnesium and Mis the molar mass of magnesium. Table 1
reports the mass of electroplated dendritic Mg deposits for
constant current electroplating at 0.921 mA cm
2
over a 24 h
total time period for varying electrolyte concentrations. Conse-
quently, equivalent Sand’s time can be computed for each of
the experimental conditions reported in Table 1. This further
enables the estimation of the electrolyte diffusion coefficient,
which is required in order to compute the limiting current
density.
Limiting current density is estimated from the computed
diffusivity as per:
ilim ¼2zc0FD
taL(3)
Here, Lis the inter-electrode distance (5.715 cm in the system
under consideration). The computed diffusivities and limiting
current densities are reported in Table 2, and the corres-
ponding variation with electrolyte concentration is also shown
explicitly in Fig. 7B and C. As pointed out earlier, the diffusivity
shows a decreasing trend with concentration. However, the
limiting current density has a non-monotonic trend owing to
the competing effects of increasing salt concentration and
decreasing diffusivity. Notably, the regimes evaluated here are
consistently above this limiting current density, which enables
mapping of non-equilibrium deposition regimes.
It is notable that while the calculations here pertain to
global conditions, diffusion limitations can further play an
important role in mediating localized heterogeneous deposi-
tion. Electrode interfacial inhomogeneities arising from inade-
quate electrolyte wetting, a heterogeneous solid electrolyte
interphase (SEI), and rough electrode surfaces can create
localized reaction zones governed by local diffusion considera-
tions. While poor electrolyte wetting is generally a result of
electrolyte–electrode mismatch in terms of interfacial wettabil-
ity or low concentration electrolyte operation, spatial variability
of the chemical constituents in a multicomponent SEI can
result in a non-uniform Mg-ion flux. Surface perturbations
can furthermore serve as preferential deposition sites as a
result of the warping of the electric field adjacent to surface
protrusions, evidenced by the preferred formation of Mg
dendrites near the edges in Fig. 2 and 3. Given that this is an
open system, a similar effect is observed with disk electrodes
(Fig. S10, ESI) where fields are localized and concentration
gradients are amplified at the edges. The subsequent steep
increase in local reaction rates can far surpass Mg self-diffusion
on the electrode surface.
35
In particular, electrolyte diffusion
limitations at high currents beget dendritic Mg morphologies
with the specific surface diffusion rates dictating fractal-like or
needle-like growth regimes as mapped in Fig. 7. The addition of
ligand molecules buffers the electrolyte concentration and
alters the effective diffusivity, whilst promoting preferential
growth morphologies as a result of selective binding to specific
facets. Consequently, the dynamic interplay between the elec-
trochemical Damkohler number (Da) contrasting the reaction
and self-diffusion rates
84
and the electrochemical Biot number
(Bi) contrasting the reaction and electrolyte transport rates
governs the morphologies of electrodeposited Mg stabilized at
high current densities.
85
Further insight into the growth of dendritic structures has
been derived from phase-field modeling calculations. The
quaternary phase diagram in Fig. S11A (ESI) illustrates the
equilibrium relationship between the different components of
the system under consideration. A plane is defined to illustrate
zero charge conditions and the respective tie lines depict the
equilibria varying between Mg(M)–THF at negative electrode
potentials and Mg(M)–MeMgCl at positive electrode potentials.
MgCl
2
species known to form passivation layers on surfaces of
Mg electrodes are further considered.
86
The dynamical model is
initiated by seeding a nucleation event at the electrolyte–
electrode surface situated at the bottom center of the domain.
Fig. S11B (ESI) shows a dendrite evolved from an initial seed.
Fig. 7C–E shows progression of dendrite growth as a function
of time. Fig. S11C (ESI) indicates the extracted information
from the overview microstructure along the blue arrow. The
three extracted curves correspond to the phase-field order
Table 2 Calculated values for mean diffusivity and limiting current
densities for reactions with varying concentrations of MeMgCl in THF
based on Sand’s time calculations
Concentration (M) Mean diffusivity
(m
2
s
1
)Mean limiting current
density (mA cm
2
)
0.25 1.43 10
9
0.22
0.5 2.41 10
10
0.05
1.0 1.93 10
10
0.13
1.5 4.95 10
11
0.06
2.0 3.14 10
10
0.07
Fig. 7 (A) Variation of diffusion coefficient with bulk electrolyte concen-
tration. Electrolyte diffusivity decreases with concentration. (B) Variation of
limiting current density with electrolyte concentration. Limiting current
density shows a non-monotonic trend because of the competing effects
of electrolyte concentration and electrolyte diffusivity. (C–E) Evolution of
dendritic growth from an initial seed located in the bottom center of the
domain based on phase field modeling for a dendrite grown in 1 M
MeMgCl with three time points representing t*=5,t*=10,andt* = 15.
Communication Materials Horizons
Published on 25 November 2019. Downloaded by Texas A & M University on 7/24/2020 4:50:20 PM.
View Article Online
This journal is ©The Royal Society of Chemistry 2020 Mater. Horiz., 2020, 7,843--854 | 851
parameter (z), Mg
2+
concentration and electrostatic potential (c).
The local variations of Mg
2+
concentration and electrostatic
potential at the dendrite tip can be clearly observed in the 1D
extracted lines. The overall kinetics of growth are dictated by the
energetics of the electrode/electrolyte interface and the Mg
2+
concentration gradient, which in turn is determined by the
surface tension and electrostatic potential. Fig. S11C (ESI)
indicates that both concentration and electric potential gradients
are larger in the vicinity of the tip, which in turn increases the
local overpotential and results in faster growth. Fig. S11D (ESI)
depicts the Butler–Volmer kinetics under three different sym-
metry factors. A value of a= 0.5 has been used in this study based
on values are reported in the literature for analogous Mg electro-
lyte complexes.
86
The results indicate that the velocity of the
deposition interface follows a highly nonlinear behavior, as is
indeed observed in Videos S1–S5 (ESI).
Experimental
Electrodeposition conditions and videomicroscopy
Symmetric cells were assembled in an argon-filled glove box
(o0.1 ppm O
2
) within three-neck round bottom flasks with two
electrode leads run through two of the rubber septa with a
separation of 5.715 cm. Both leads held Mg ribbon electrodes
(Alfa Aesar, purity of 99.5%) creating symmetric cells. MeMgCl
solutions (3 M in anhydrous tetrahydrofuran (THF), Alfa Aesar)
were diluted using anhydrous THF (DriSolv. EMD Millipore Co.,
purity of Z99.9%). Ligand effects were evaluated through the
addition of oleylamine (0.121 M, Sigma Aldrich) or dodeca-
nethiol (0.0626 M, 0.125 M, or 0.188 M, Sigma Aldrich).
Electrodeposition was performed under Schlenk conditions in
an Ar atmosphere using a programmable power supply
(FB1000, Fisher Scientific) and applying a constant current. A
videomicroscope (Plugable Technologies) was used to monitor
the reactions.
Structural characterization of deposits
Deposits easily were separated from the Mg substrate through
gentle washing with THF. Powder X-ray diffraction (XRD) was
performed in Bragg–Brentano geometry using a Bruker D8-
Focus diffractometer (Cu Ka:l= 1.5418 Å; 40 kV voltage;
25 mA current). X-ray photoelectron spectra (XPS) were obtained
using an Omicron DAR 400 XPS/UPS system with a 128-channel
micro-channel plate Argus detector using a Mg KaX-ray source
(1253.6 eV). A CN10 electron flood source was utilized to reduce
charging. High-resolution scans were collected in constant
analyzer energy (CAE) mode with a 100 eV pass energy and a
step size of 0.05 eV. Spectral line shapes were fit using the
Marquart–Levenberg algorithm for mixed Gaussian–Lorentzian
(7 : 3) line shapes. All spectra were aligned to the C 1s line of
adventitious carbon at 284.8 eV.
Electron microscopy
Scanning electron microscopy images were obtained using a
JEOL JSM-7500F operating at an accelerating voltage of 10 kV,
emission current of 5 mA, and a probe current of 10 mA. Cross-
sectional TEM samples of Mg dendrites were prepared using a
FEI Helios Nanolab 460F1 Dual-Focused Ion Beam (FIB). The
crystal structure and the growth direction of the Mg dendrites
were identified using electron backscatter diffraction (EBSD,
Tescan FERA-3 scanning electron microscope (SEM) with an
accelerating voltage of 20 kV) and bright-field transmission
electron microscopy (TEM, FEI Tecnai G2 F20 Super-Twin
FE-TEM operated at 200 kV).
3D X-ray tomography
Soft X-ray microscopy images were recorded at the SM (101D-1)
beamline of the Canadian Light Source (CLS). The sample was
mounted on a computer-controlled (x, y, y) tilt-stage, which
facilitates spectrotomographic measurements. Tomography
data was acquired at the Mg K-edge from +701to 351in
increments of 51. Data analysis was performed using TomoJ, a
plug-in to the image analysis software, ImageJ.
87
The images
were first aligned using Fourier cross-correlation methods,
then further refined using 3D landmarks. In the latter, an
algorithm locates regions that can be tracked within the series
without the aid of fiducial markers.
88
Conversion to optical
density was carried out using aXis2000 (http://unicorn.mcmas
ter.ca/aXis2000.html). A 3D reconstruction was performed on
the aligned tilt-series using an algebraic reconstruction techni-
que (ART), accessible through TomoJ.
89
A total of 10 iterations
were carried out with a relaxation coefficient of 0.08.
Nanomechanical characterization of deposits
Strips of pristine Mg substrate (never used for electrodeposi-
tion) as well as the 0.5 and 2 M electrodeposits were cast into
separate epoxy stubs. These embedded samples were consecu-
tively mechanically polished using 9, 3, 1, and 0.05 mmdiamond
suspensions. After polishing, the elastic modulus and hardness of
the samples were measured using a Nanomechanics iMicro
indenter equipped with an InForce 50 actuator and a diamond
Berkovich tip. The standard approach of Oliver and Pharr was
used to estimate the elastic modulus and hardness.
90
Indentation
implemented a test with constant :
P/P= 0.21 s
1
, with contin-
uous stiffness oscillation of 2 nm. Twelve indentation tests were
used for each sample as the basis for the reported mechanical
measurements.
Model formulation
Electrolyte diffusion limitations. The amount of dendritic
magnesium from experiments can be directly correlated to the
time between onset of Sand’s time limitation and end of
experimental runtime. Consequently, the electrolyte diffusion
coefficient and symmetric cell system limiting current densities
can be evaluated to explain the formation of magnesium
dendrites. Cationic transference numbers reported in the
literature for EtMgCl in THF, ranging from 0.058 at 0.25 M to
0.018 at 0.4 M, have been used to develop the model.
91
Low
mobilities of dimeric species and ion–ion interactions at high
concentrations are thought to be the origin of the diminution
of the transference number at high concentrations.
Materials Horizons Communication
Published on 25 November 2019. Downloaded by Texas A & M University on 7/24/2020 4:50:20 PM.
View Article Online
852 |Mater. Horiz., 2020, 7, 843--854 This journal is ©The Royal Society of Chemistry 2020
Numerical integration of the phase field model. Formula-
tion of the phase field model is described in the ESI.A metallic
Mg electrode in contact with a 1 M MeMgCl solution in THF
was selected as the reference state. For the equilibrium numerical
simulations, the Mg electrode was located at the bottom of the
simulation cell and an artificial nucleation event was introduced.
The partial molar volumes of Mg
2+
,MeMgCl,andTHFare
approximated to be the same. Eqn (S10)–(S12) (ESI)weresolved
using a finite difference solver in a uniform grid with equal mesh
size using a parallel in-house Fortran code. Boundary conditions
used for eqn (S7)–(S9) are listed in Table S1 (ESI). Only half of the
cell was considered in order to reduce the computational cost;
the domain cell size was set at 300 500.
Conclusions
The promise and excitement of magnesium batteries derives in
large portion from the idea that they are immune to dendrite
formation. Whilst considerable effort has been invested in the
design of novel electrolytes and cathode materials, multivariate
studies of Mg electrodeposition are scarce particularly under
conditions emulative of high local concentration and potential
gradients. Galvanostatic electrodeposition of metallic Mg from
Grignard reagents in symmetric cells reveals a complex phase
map with varying morphologies of plated deposits including
fractal aggregates and highly anisotropic dendrites with singular
growth fronts. Based on electron microscopy, X-ray tomography,
and optical tomography observations, the deposits are highly
faceted primarily zerovalent magnesium with some surface passi-
vation. The growth morphologies have been examined as a
function of current density, concentration, and added coordinat-
ing ligands. Increase of the current density amplifies the extent of
branching, indicating an increase in the electrochemical reaction
rate; increases in concentration induce a transition from a fractal
to a dendritic growth regime. Remarkably, the dendrites show
extended single crystalline domains along the h11%
20igrowth
direction. At lower concentrations, smaller grains comprising
agglomerated thin hexagonal platelets are observed. In contrast,
at higher concentrations more spherical deposits with faceted
hexagonal surficial features are seen. At the highest concentra-
tions, canonical dendritic deposits with a strongly anisotropic
growth direction are observed. Addition of coordinating ligands
greatlyaltersthegrowthmechanisms suppressing dendrite
growth and instead stabilizing single-crystalline high-aspect-
ratio nanowires by altering the extent of supersaturation and
the nature of the electrode/electrolyte interface.
Dendritic electrodeposition is a result of electrolyte trans-
port limitations, with surface self-diffusion rates dictating
morphological variation from needle-like to fractal-like morpho-
logies. Synergistic analytical and phase-field modeling further
establish the proclivity of Mg to form dendrites at high current
densities; variations in electrolyte diffusivity variation with
concentrations have further been delineated. Whilst data on
long-term cycling performance ofMgfullcellsisscarceandit
remains to be observed the extent to which dendrite formation
will emerge as a limitation, it is worth noting that electrochemical
reaction rates can readily surpass self-diffusion rates as a result of
local inhomogeneities; as such, the results herein are expected to
be relevant to systems even wherein averages current densities are
substantially lower. The hardness of Mg dendrites delineated
here, with shear moduli approaching 10 GPa, is substantially
greater than Li dendrites, and further suggests the need for
cautioninthedesignofseparators.
Conflicts of interest
There are no conflicts to declare.
Acknowledgements
R. D. acknowledges the Data-Enabled Discovery and Design
of Energy Materials program funded under NSF award DGE-
1545403. D. S. and C. D. F. would like to acknowledge the
support of the National Science Foundation Graduate Research
Fellowship under grant No. DGE-1746932 and DGE-1746932
respectively. This work was supported by the National Science
Foundation under DMR 1809866. 3D tomography using STXM
was performed at the Canadian Light Source, which is sup-
ported by the Natural Sciences and Engineering Research
Council of Canada, the National Research Council Canada,
the Canadian Institutes of Health Research, the Province of
Saskatchewan, Western Economic Diversification Canada, and
the University of Saskatchewan. Authors also acknowledge the
Materials Characterization Facility at Texas A&M for work using
SEM and XPS.
References
1 N. Nitta, F. Wu, J. T. Lee and G. Yushin, Mater. Today, 2015,
18, 252–264.
2 E. A. Olivetti, G. Ceder, G. G. Gaustad and X. Fu, Joule, 2017,
1, 229–243.
3 J. W. Choi and D. Aurbach, Nat. Rev. Mater., 2016, 1, 16013.
4 B. Liu, J.-G. Zhang and W. Xu, Joule, 2018, 2, 833–845.
5 D. Aurbach, E. Zinigrad, Y. Cohen and H. Teller, Solid State
Ionics, 2002, 148, 405–416.
6 K. J. Harry, D. T. Hallinan, D. Y. Parkinson, A. A. MacDowell
and N. P. Balsara, Nat. Mater., 2014, 13, 69–73.
7M.Ja
¨ckle, K. Helmbrecht, M. Smits, D. Stottmeister and
A. Groß, Energy Environ. Sci., 2018, 11, 3400–3407.
8 G. Henkelman, B. P. Uberuaga and H. Jo
´nsson, J. Chem.
Phys., 2000, 113, 9901–9904.
9 L. C. Merrill and J. L. Schaefer, Front. Chem., 2019, 7, 194.
10 D. Aurbach, Y. Cohen and M. Moshkovich, Electrochem.
Solid-State Lett., 2001, 4, A113.
11 J. O. Besenhard and M. Winter, ChemPhysChem, 2002, 3,
155–159.
12 C. B. Bucur, T. Gregory, A. G. Oliver and J. Muldoon, J. Phys.
Chem. Lett., 2015, 6, 3578–3591.
13 C. Ling, D. Banerjee and M. Matsui, Electrochim. Acta, 2012,
76, 270–274.
Communication Materials Horizons
Published on 25 November 2019. Downloaded by Texas A & M University on 7/24/2020 4:50:20 PM.
View Article Online
This journal is ©The Royal Society of Chemistry 2020 Mater. Horiz., 2020, 7,843--854 | 853
14 M. Matsui, J. Power Sources, 2011, 196, 7048–7055.
15 J. Muldoon, C. B. Bucur and T. Gregory, Chem. Rev., 2014,
114, 11683–11720.
16 Q. Fu, A. Sarapulova, V. Trouillet, L. Zhu, F. Fauth, S. Mangold,
E. Welter, S. Indris, M. Knapp, S. Dsoke, N. Bramnik and
H. Ehrenberg, J. Am. Chem. Soc., 2019, 141, 2305–2315.
17 A. Parija, Y. Liang, J. L. Andrews, L. R. De Jesus,
D. Prendergast and S. Banerjee, Chem. Mater., 2016, 28,
5611–5620.
18 B. Zhou, H. Shi, R. Cao, X. Zhang and Z. Jiang, Phys. Chem.
Chem. Phys., 2014, 16, 18578–18585.
19 E. Levi, Y. Gofer and D. Aurbach, Chem. Mater., 2010, 22,
860–868.
20 H. D. Yoo, Y. Liang, H. Dong, J. Lin, H. Wang, Y. Liu, L. Ma,
T. Wu, Y. Li, Q. Ru, Y. Jing, Q. An, W. Zhou, J. Guo, J. Lu,
S. T. Pantelides, X. Qian and Y. Yao, Nat. Commun., 2017,
8, 339.
21 D. Aurbach, Y. Gofer, A. Schechter, O. Chusid, H. Gizbar,
Y. Cohen, M. Moshkovich and R. Turgeman, J. Power
Sources, 2001, 97–98, 269–273.
22 J. Muldoon, C. B. Bucur and T. Gregory, Angew. Chem., Int.
Ed., 2017, 56, 12064–12084.
23 T. Gao, S. Hou, K. Huynh, F. Wang, N. Eidson, X. Fan,
F. Han, C. Luo, M. Mao, X. Li and C. Wang, ACS Appl. Mater.
Interfaces, 2018, 10, 14767–14776.
24 Q. S. Zhao, Y. N. Nuli, Y. S. Guo, J. Yang and J. L. Wang,
Electrochim. Acta, 2011, 56, 6530–6535.
25 S. DeWitt, N. Hahn, K. Zavadil and K. Thornton,
J. Electrochem. Soc., 2016, 163, A513–A521.
26 M. Ja
¨ckle and A. Groß, J. Chem. Phys., 2014, 141, 174710.
27 J. Tian, D. Cao, X. Zhou, J. Hu, M. Huang and C. Li, ACS
Nano, 2018, 12, 3424–3435.
28 Y. Zhang, J. Xie, Y. Han and C. Li, Adv. Funct. Mater., 2015,
25, 7300–7308.
29 P. G. Shewmon, Trans. Metall. Soc. AIME, 1956, 206, 918–922.
30 J. Combronde and G. Brebec, Acta Metall., 1971, 19, 1393–1399.
31 S. Ganeshan, L. G. Hector and Z. K. Liu, Comput. Mater. Sci.,
2010, 50, 301–307.
32 L. J. Zhang, T. I. Spiridonova, S. E. Kulkova, R. Yang and
Q. M. Hu, Comput. Mater. Sci., 2017, 128, 236–242.
33 A. Parija, G. R. Waetzig, J. L. Andrews and S. Banerjee,
J. Phys. Chem. C, 2018, 122, 25709–25728.
34 T. D. Gregory, R. J. Hoffman and R. C. Winterton,
J. Electrochem. Soc., 1990, 137, 775–780.
35 R. Davidson, A. Verma, D. Santos, F. Hao, C. Fincher,
S. Xiang, J. Van Buskirk, K. Xie, M. Pharr, P. P. Mukherjee,
S. Banerjee, J. Van Buskirk, K. Xie, M. Pharr, P. P. Mukherjee
and S. Banerjee, ACS Energy Lett., 2019, 4, 375–376.
36 J. Bitenc, K. Pirnat, E. Z
ˇagar, A. Randon-Vitanova and
R. Dominko, J. Power Sources, 2019, 430, 90–94.
37 C. Liebenow, J. Appl. Electrochem., 1997, 27, 221–225.
38 M. S. Ding, T. Diemant, R. J. Behm, S. Passerini and
G. A. Giffin, J. Electrochem. Soc., 2018, 165, A1983–A1990.
39 H. Brune, Surf. Sci. Rep., 1998, 31, 125–229.
40 J. W. Evans, P. A. Thiel and M. C. Bartelt, Surf. Sci. Rep.,
2006, 61, 1–128.
41 A. Groß, Theoretical Surface Science – A Microscopic Perspec-
tive, Springer, Berlin, 2nd edn, 2009.
42 K. Pettersen, O. Lohne and N. Ryum, Metall. Trans. A, 1990,
21, 221–230.
43 D. Dube
´, A. Couture, Y. Carbonneau, M. Fiset, R. Angers and
R. Tremblay, Int. J. Cast Met. Res., 1998, 11, 139–144.
44 K. N. Wood, E. Kazyak, A. F. Chadwick, K. H. Chen,
J. G. Zhang, K. Thornton and N. P. Dasgupta, ACS Cent.
Sci., 2016, 2, 790–801.
45 K. Nishikawa, T. Mori, T. Nishida, Y. Fukunaka, M. Rosso
and T. Homma, J. Electrochem. Soc., 2010, 157, A1212.
46 W.-S. Kim and W.-Y. Yoon, Electrochim. Acta, 2004, 50,
541–545.
47 A. J. Ilott, M. Mohammadi, H. J. Chang, C. P. Grey and
A. Jerschow, Proc.Natl.Acad.Sci.U.S.A., 2016, 113, 10779–10784.
48 P. Bai, J. Li, F. R. Brushett and M. Z. Bazant, Energy Environ.
Sci., 2016, 9, 3221–3229.
49 J. Y. Song, H. H. Lee, Y. Y. Wang and C. C. Wan, J. Power
Sources, 2002, 111, 255–267.
50 G. Bieker, M. Winter and P. Bieker, Phys. Chem. Chem. Phys.,
2015, 8670, 8670–8679.
51 J. C. Burns, L. J. Krause, D.-B. Le, L. D. Jensen, A. J. Smith,
D. Xiong and J. R. Dahn, J. Electrochem. Soc., 2011, 158, A1417.
52 F. Hao, A. Verma and P. P. Mukherjee, ACS Appl. Mater.
Interfaces, 2018, 10, 26320–26327.
53 L.Viyannalage, V. Lee, R. V. Dennis, D. Kapoor, C. D. Haines
and S. Banerjee, Chem. Commun., 2012, 48, 5169.
54 D. Aurbach, Electrochem. Solid-State Lett., 1999, 3, 31.
55 C. Fotea, J. Callaway and M. R. Alexander, Surf. Interface
Anal., 2006, 38, 1363–1371.
56 H. Kuwata, M. Matsui and N. Imanishi, J. Electrochem. Soc.,
2017, 164, A3229–A3236.
57 Y. Gofer, R. Turgeman, H. Cohen and D. Aurbach, Langmuir,
2003, 19, 2344–2348.
58 R. M. Bradley and R. C. Ball, Nature, 1984, 309, 225–229.
59 V. Fleury, M. Rosso, J.-N. N. Chazalviel and B. Sapoval, Phys.
Rev. A: At., Mol., Opt. Phys., 1991, 44, 6693–6705.
60 Y. Sawada, A. Dougherty and J. P. Gollub, Phys. Rev. Lett.,
1986, 56, 1260–1263.
61 D.Grier, E. Ben-Jacob, R. Clarke and L. M. Sander, Phys. Rev.
Lett., 1986, 56, 1264–1267.
62 R. F. Voss, Phys. Rev. B: Condens. Matter Mater. Phys., 1984,
30, 334–337.
63 P. Meakin, Phys. Rev. Lett., 1983, 51, 1119–1122.
64 A. Ghosh, R. Batabyal, G. P. Das and B. N. Dev, AIP Adv.,
2016, 6, 015301.
65 L. O. Valo
´en and J. N. Reimers, J. Electrochem. Soc., 2005,
152, A882.
66 D. P. Barkey, Advances in Electrochemical Science and
Engineering, Wiley-VCH Verlag GmbH, Weinheim, FRG,
2001, vol. 7, pp. 151–191.
67 O. Zik and E. Moses, Phys. Rev. E: Stat. Phys., Plasmas, Fluids,
Relat. Interdiscip. Top., 1996, 53, 1760–1764.
68 E. Ben-Jacob and P. Garik, Nature, 1990, 343, 523–530.
69 W. Shao and G. Zangari, J. Phys. Chem. C, 2009, 113,
10097–10102.
Materials Horizons Communication
Published on 25 November 2019. Downloaded by Texas A & M University on 7/24/2020 4:50:20 PM.
View Article Online
854 |Mater. Horiz., 2020, 7, 843--854 This journal is ©The Royal Society of Chemistry 2 020
70 K. R. Kort and S. Banerjee, Small, 2015, 11, 329–334.
71 J. Cho, H. Jin, D. G. Sellers, F. David Watson, D. Hee Son
and S. Banerjee, J. Mater. Chem. C, 2017, 5, 8810.
72 M. P. Staiger, A. M. Pietak, J. Huadmai and G. Dias, Bioma-
terials, 2006, 27, 1728–1734.
73 ASM specialty handbook: magnesium and magnesium alloys,
ed. M. Avedesian and H. Baker, ASM International, 1999.
74 W. C. Oliver and G. M. Pharr, J. Mater. Res., 2004, 19, 3–20.
75 M. Rashad, F. Pan, A. Tang, M. Asif and M. Aamir, J. Alloys
Compd., 2014, 603, 111–118.
76 E. Mostaed, M. Vedani, M. Hashempour and M. Bestetti,
Biomatter, 2014, 4, e28283.
77 C. Monroe and J. Newman, J. Electrochem. Soc., 2005, 152,
A396–A404.
78 Metals Handbook Desk Edition, ed. J. R. Davis, Materials
Park, OH, 2nd edn, 2003.
79 C. T. Love, J. Electrochem. Energy Convers. Storage, 2016, 13, 031004.
80 W. S. LePage, Y. Chen, E. Kazyak, K.-H. Chen, A. J. Sanchez,
A. Poli, E. M. Arruda, M. D. Thouless and N. P. Dasgupta,
J. Electrochem. Soc., 2019, 166, A89–A97.
81 A. Masias, N. Felten, R. Garcia-Mendez, J. Wolfenstine and
J. Sakamoto, J. Mater. Sci., 2019, 54, 2585–2600.
82 A. Puzder, A. J. Williamson, N. Zaitseva, G. Galli, L. Manna
and A. P. Alivisatos, Nano Lett., 2004, 4, 2361–2365.
83 S. A. Morin, M. J. Bierman, J. Tong and S. Jin, Science, 2010,
328, 476–480.
84 F. Hao, A. Verma and P. P. Mukherjee, ACS Appl. Mater.
Interfaces, 2018, 10, 26320–26327.
85 F. Hao,A. Verma and P. P. Mukherjee, Energy Storage Mater.,
2019, 20, 1–6.
86 Y. Viestfrid, M. D. Levi, Y. Gofer and D. Aurbach,
J. Electroanal. Chem., 2005, 576, 183–195.
87 C. MessaoudiI, T. Boudier, C. Sorzano and S. Marco, BMC
Bioinf., 2007, 8, 288.
88 C. Sorzano, C. Messaoudi, M. Eibauer, J. Bilbao-Castro,
R. Hegerl, S. Nickell, S. Marco and J. Carazo, BMC Bioinf.,
2009, 10, 124.
89 G. T. Herman, A. Lent and S. W. Rowland, J. Theor. Biol.,
1973, 42, 1–32.
90 W. C. Oliver and G. M. Pharr, J. Mater. Res., 1992, 7,
1564–1583.
91 A. Benmayza, M. Ramanathan, T. S. Arthur, M. Matsui,
F. Mizuno, J. Guo, P.-A. Glans and J. Prakash, J. Phys.
Chem. C, 2013, 117, 26881–26888.
Communication Materials Horizons
Published on 25 November 2019. Downloaded by Texas A & M University on 7/24/2020 4:50:20 PM.
View Article Online
... Mg anodes were, for example, long believed to not exhibit dendrite formation since the on-set current density is higher than that of Li [21,22]. Furthermore, the morphology of the dendrites are different in equivalent electrochemical environments for three promising metal anodes; Li, Na, and Mg [23,24]. ...
... This is indicative of flatter Mg compared to Li and Na constructions, and implies that Mg metal deposits are flatter than Li and Na even if the crystal structure of the surface, onto which the metal is deposited, deviates from the room temperature bulk crystal structure of the metal. This is consistent with the hexagonal platelet shape observed experimentally for Mg deposits [23], and may help explaining why Mg is less prone to fractal deposition and dendrite formation than Li and Na. ...
Article
Full-text available
Dendrite formation occurs on Li, Na, and Mg metal anodes in rechargeable batteries, and is a safety challenge, as well as a limiting factor for increasing energy- and power density. However, the behaviour of the dendrites differs depending on the anode material. In this study, we investigate the local bulk and surface crystal structure of Li, Na, and Mg surfaces to shed light on how differences in the morphology and structure of the anode surface and its metal deposits can explain differences in dendrite formation on Li, Na, and Mg anodes. The local bulk- and surface structure are found using molecular dynamics simulations in combination with the surface adaptive common neighbour analysis, and indicate that Li and Na surfaces are more prone to surface instabilities and formation of protrusions than Mg surfaces, which remain flat and hexagonal close-packed even near room temperature. Additionally, the equilibrium shapes of the Mg deposits obtained from density functional theory assume more flat and hexagonal shapes than the Li and Na deposits. Together, these results shed light on atomic mechanisms that may contribute to the different propensities of Li, Na, and Mg metal anodes to form dendrites.
... The kinds of electrolyte can also affect the shapes of dendrites. For example, fibrous dendrites have been observed in Li, Na, K, and Mg metal anodes in organic electrolytes [55][56][57][58][59][60][61]. Al dendrites in AlCl 3 /1-ethyl-3-methylimidazolium chloride (AlCl 3 /[EMIm]Cl) ionic liquid electrolyte exhibited a mossy shape [62]. ...
Article
Metal anodes (Li, Na, K, Zn, Mg, Ca, Fe, Al, Mn, etc) based on a plating/stripping electrochemical mechanism have attracted great attention in rechargeable batteries because of their low electrochemical potential, high theoretical specific capacity, and superior electronic conductivity. Metal anodes exhibit large potential in constructing high-energy-density rechargeable batteries. However, challenges such as high chemical reactivity, large volume changes, unstable solid electrolyte interphase (SEI), and uneven electrochemical deposition result in a serious of interfacial issues on metal anodes, including corrosion, side reaction, structural instability, and formation of dendrites. In the past several years, a lot of modification strategies based on interfacial engineering have been proposed to improve the interfacial stability of metal anodes. The interfacial engineering on metal anodes is mainly achieved by solid-liquid reaction, solid-solid reaction, solid-gas reaction, and physical decoration. In this review, we summary and analyse these interfacial engineering strategies on metal anodes in detail. Meanwhile, some perspectives and outlooks are put forward. This review can provide some enlightenment for related researchers and promote the development of metal anodes in rechargeable batteries.
Article
Full-text available
Magnesium metal anodes have attracted widespread attention for their high volumetric capacity and natural abundance, but are precluded from practical applications by poor rate capability and limited lifespan due to sluggish ion‐transfer kinetics and uneven deposition behavior. Herein, for the first time a grain‐boundary‐rich triphasic artificial hybrid interphase, consisting of Sb metal, Mg3Sb2 alloy, and MgCl2, is designed on Mg anode surface by a facile solution treatment method, enabling high‐rate and long‐cycle Mg plating/stripping behavior. The triphasic artificial hybrid interphase affords high magnesiophilicity and ionic conductivity to reduce the energy barriers for Mg²⁺ desolvation and deposition. Meanwhile, the abundant grain boundaries redistribute Mg²⁺ flux at the electrode‐electrolyte interface and guide uniform Mg deposition. Accordingly, the as‐designed Mg metal anode achieves ultralong cycling life of 350 h at a high current density of 5 mA cm⁻² and a large areal capacity of 5 mAh cm⁻², outperforming previously reported Mg metal anodes with artificial interphases. Full cells with Mo6 cathode also show extraordinary stability over a long lifespan of 8000 cycles at a high rate of 5 C. The rational artificial interphase design and the understanding of composition‐structure‐function relationships shed deep insights into the development of fast‐charging and long‐cycling Mg metal batteries.
Article
Electrochemical stability is a critical performance parameter for the materials used as electrolytes and electrodes in batteries. Using first-principles electronic structure calculations, we have determined the electrochemical stability windows of magnesium binary and ternary spinel compounds. These materials are candidates for protective coatings, solid electrolytes, and cathodes in Mg batteries, which represent a promising sustainable alternative to Li-ion batteries that still dominate the battery market. Furthermore, we have applied and assessed two different criteria for the chemical stability of the compounds. For the spinel materials, we identify the critical role of the ionic radii of the transition metal for the stability of the compounds. In addition, we determine the ion mobility in these materials using a recently developed descriptor. We thus provide guidelines for the choice of promising solid materials for Mg batteries with improved properties.
Article
Full-text available
Magnesium (Mg) batteries (MBs), as post-lithium-ion batteries, have received great attention in recent years due to their advantages of high energy density, low cost, and safety insurance. However, the formation of passivation layers on the surface of Mg metal anode and the poor compatibility between Mg metal and conventional electrolytes during charge-discharge cycles seriously affect the performance of MBs. The great possibility of generating Mg dendrites has also caused controversy among researchers. Moreover, the regulation of Mg deposition and the enhancement of battery cycle stability is largely limited by interfacial stability between Mg metal anode and electrolyte. In this review, recent advances in interfacial science and engineering of MBs are summarized and discussed. Special attention is given to interfacial chemistry including passivation layer formation, incompatibilities, ion transport, and dendrite growth. Strategies for building stable electrode/interfaces, such as anode designing and electrolyte modification, construction of artificial solid electrolyte interphase (SEI) layers, and development of solid-state electrolytes to improve interfacial contacts and inhibit Mg dendrite and passivation layer formation, are reviewed. Innovative approaches, representative examples, and challenges in developing high-performance anodes are described in detail. Based on the review of these strategies, reference is provided for future research to improve the performance of MBs, especially in terms of interface and anode design.
Article
Electrodeposition in electrochemical cells is one of the leading causes of performance deterioration. The prediction of electrodeposition growth demands a good understanding of the complex physics involved, which can lead to the fabrication of a probabilistic mathematical model. As an alternative, a convolutional Long Short-Term architecture-based image analysis approach is presented herein. This technique can predict the electrodeposition growth of the electrolytes, without prior detailed knowledge of the system. The captured images from the experiments are used to train and test the model. A comparison between the expected output image and predicted image on a pixel level, the percentage mean squared error, absolute percentage error, and pattern density are investigated to assess the model accuracy. The randomness of the electrodeposition growth is outlined by investigating the fractal dimension and the interfacial length of the electrodeposits. The trained model predictions show a significant promise between all the experimentally obtained relevant parameters with the predicted one. It is expected that this deep learning-based approach for predicting random electrodeposition growth will be of immense help for designing and optimizing the relevant experimental scheme in near future without performing multiple experiments.
Article
Despite the substantial efforts aimed at suppressing metallic dendrite growth in Li-metal batteries, Mg-metal dendrite growth has thus far received relatively little attention, and the formation of Mg dendrites has recently been shown to be a critical limitation for the practical advancement of rechargeable Mg-ion batteries. The development of an appropriate anode to efficiently accommodate Mg deposits is thus key to overcome this limitation. Here, we report the unique design of Ag-decorated Cu foam (ACF) consisting of a porous Cu scaffold decorated with magnesiophilic Ag nanoparticles (NPs) on its surface through a facile one-step synthesis process. For the first time, we demonstrate the strong affinity of Ag atoms to the electrochemically deposited Mg; magnesiophilicity is then adopted to design an efficient anode host for Mg-metal batteries and suppress the Mg dendritic formation. As a result, the ACF exhibits a greatly decreased nucleation overpotential with a longer cycle life compared with those of conventional substrates. In the absence of magnesiophilic Ag nano-seeds, non-uniform and top-oriented Mg depositions are observed; in contrast, the ACF helps contribute to an even deposition of the electrochemically formed Mg over the entire active surface, resulting in improved electrochemical performance.
Article
Metallic Mg is an ideal anode material for rechargeable magnesium-ion batteries (RMBs) due to its high volumetric energy density, low safety concern, two-electron redox as well as the large abundance and environmental friendliness of magnesium element. However, challenges such as interfacial passivation, formation of dendrites, and large volume changes of electrode hinder the development and practical application of Mg metal anode. To address these issues, many modification strategies have been proposed in the past several years, such as modifying liquid electrolytes, designing artificial solid electrolyte interphase (SEI), constructing interfacial alloy layer, fabricating 3D scaffolds, utilizing solid electrolytes, developing advanced characterization methods, and deepening theoretical study. Here a detail review is presented to summary and analyze these modification strategies. Meanwhile, some perspectives and outlooks are put forward. This review can inspire related researchers and promote the application of Mg metal anode in rechargeable Mg batteries.
Article
Metallic magnesium is a promising high-capacity anode material for energy storage technologies beyond lithium-ion batteries. However, most reported Mg metal anodes are only cyclable under shallow cycling (≤1 mAh cm-2) and thus poor Mg utilization (<3%) conditions, significantly compromising their energy-dense characteristic. Herein, composite Mg metal anodes with high capacity utilization of 75% are achieved by coating magnesiophilic gold nanoparticles on copper foils for the first time. Benefiting from homogeneous ionic flux and uniform deposition morphology, the Mg-plated Au-Cu electrode exhibits high average Coulombic efficiency of 99.16% over 170 h cycling at 75% Mg utilization. Moreover, the full cell based on Mg-plated Au-Cu anode and Mo6S8 cathode achieves superior capacity retention of 80% after 300 cycles at a low negative/positive ratio of 1.33. This work provides a simple yet effective general strategy to enhance Mg utilization and reversibility, which can be extended to other metal anodes as well.
Article
Full-text available
Mg rechargeable batteries are a highly perspective battery system for the future post Li-ion batteries. However, their practical application seems quite distant due to poor Mg²⁺ insertion into inorganic cathode materials. Organic cathodes offer an opportunity for realization of practical Mg rechargeable batteries due to circumvention of insertion limitations encountered in inorganic hosts. Herein, we investigate the electrochemical performance of polyanthraquinone (PAQ) organic cathode in two Mg electrolytes based on different combinations of the salts, Mg(TFSI)2-2MgCl2 and MgCl2-AlCl3. PAQ in electrolyte containing Mg(TFSI)2-2MgCl2 exhibits good capacity retention and utilization. MgCl2-AlCl3 based electrolyte shows poorer electrochemical performance and a peculiar upshift of the voltage plateau for 300 mV, which is caused by the presence of AlCl3 in the electrolyte. Mg anode morphology after cycling reveals the formation of pits and cracks on the Mg foil anode as well as uneven Mg deposits several μm in size in both electrolytes. The maximum capacity and capacity retention are referenced with polyanthraquinone performance in Li battery and show that at 1C cycling rate 75% of the maximum capacity in Li battery can be obtained in Mg battery. Long-term capacity fade is comparable, demonstrating the potential for practical realization of Mg‒organic battery.
Article
Full-text available
One of the limiting factors in the development of magnesium batteries is the reversibility of magnesium electrodeposition and dissolution at the anode. Often irreversibility is related to impurities and decomposition. Herein we report on the cycling behavior of magnesium metal anodes in different electrolytes, Mg(HMDS)2 – 4 MgCl2 in tetrahydrofuran (THF) and a butyl sulfone/THF mixture. The deposition morphology and anode-electrolyte interface is studied and related to Mg/Mg cell cycling performance. It is found that adding the sulfone caused the formation of a boundary layer at the electrode-electrolyte interface, which, in turn, resulted in a particle-like deposition morphology. This type of deposition has a high surface area, which alters the effective local current density and results in electronically isolated deposits. Extended cycling resulted in magnesium growth through a separator. Electrolyte decomposition is observed with and without the addition of the sulfone, however the addition of the sulfone increased the degree of decomposition.
Article
Full-text available
The lack of a reliable rechargeable lithium metal (Li-metal) anode is a critical bottleneck for next-generation batteries. The unique mechanical properties of lithium influence the dynamic evolution of Li-metal anodes during cycling. While recentmodels have aimed at understanding the coupled electrochemical-mechanical behavior of Li-metal anodes, there is a lack of rigorous experimental data on the bulk mechanical properties of Li. This work provides comprehensive mechanical measurements of Li using a combination of digital-image correlation and tensile testing in inert gas environments. The deformation of Li was measured over a wide range of strain rates and temperatures, and it was fitted to a power-law creep model. Strain hardening was only observed at high strain rates and low temperatures, and creep was the dominant deformation mechanism over a wide range of battery-relevant conditions. To contextualize the role of creep on Li-metal anode behavior, examples are discussed for solid-state batteries, "dead" Li, and protective coatings on Li anodes. This work suggests new research directions and can be used to inform future electrochemical-mechanical models of Li-metal anodes.
Article
Full-text available
We demonstrate the growth of dendritic magnesium deposits with fractal morphologies exhibiting shear moduli in excess of values for polymeric separators upon the galvanostatic electrodeposition of metallic Mg from Grignard reagents in symmetric Mg—Mg cells. Dendritic growth is understood based on the competing influences of reaction rate, electrolyte transport rate, and self-diffusion barrier evaluated using a dimensionless Damköhler ratio as further corroborated by mesoscale simulations.
Article
Full-text available
With the potential to dramatically increase energy density compared to conventional lithium ion technology, lithium metal solid-state batteries (LMSSB) have attracted significant attention. However, little is known about the mechanical properties of Li. The purpose of this study was to characterize the elastic and plastic mechanical properties and creep behavior of Li. Elastic properties were measured using an acoustic technique (pulse-echo). The Young’s modulus, shear modulus, and Poisson’s ratio were determined to be 7.82 GPa, 2.83 GPa, and 0.381, respectively. To characterize the stress–strain behavior of Li in tension and compression, a unique load frame was used inside an inert atmosphere. The yield strength was determined to be between 0.73 and 0.81 MPa. The time-dependent deformation in tension was dramatically different compared to compression. In tension, power law creep was exhibited with a stress exponent of 6.56, suggesting that creep was controlled by dislocation climb. In compression, time-dependent deformation was characterized over a range of stress believed to be germane to LMSSB (0.8–2.4 MPa). At all compressive stresses, significant barreling and a decrease in strain rate with increasing time were observed. The implications of this observation on the charge/discharge behavior of LMSSB will be discussed. We believe the analysis and mechanical properties measured in this work will help in the design and development of LMSSB.
Article
Rechargeable battery chemistries, with high energy densities, are particularly desirable in order to meet the burgeoning demand for energy storage. In this regard, metal electrodes have recently drawn extensive research interest due to the intrinsic energy density boost, while a fundamental study is needed to reveal the underlying mechanisms governing the electrodeposition stability. Here, we explore the mesoscale interactions in nucleation and growth of electrodeposition, with a focus on the competition of ion transport in the electrolyte, electrochemical reactions at the electrolyte-electrode interface, and surface self-diffusion. It is found that lithium (Li)and sodium (Na)metal anodes have the tendency to form dendrites at high local reaction rates, whereas magnesium (Mg)and aluminum (Al)do not, because of the low intrinsic self-diffusion barriers. Nonuniform electrodeposition at low reaction rates is observed, which could be attributed to the spatial inhomogeneities due to separator wetting, solid electrolyte interphase (SEI)formation, and electrode surface roughness. This work provides a fundamental understanding of the mesoscale underpinnings on the electrodeposition stability of various metal electrodes, especially shedding light on pathways toward potentially dendrite-free electrodeposition morphology.
Article
Orthorhombic V2O5 nanowires were successfully synthesized via a hydrothermal method. A cell-configuration system was built utilizing V2O5 as the cathode and 1 M Mg(ClO4)2 electrolyte within acetonitrile, together with MgxMo6S8 (x~2) as the anode to investigate the structural evolution and oxidation state and local structural changes of V2O5. The V2O5 nanowires deliver an initial discharge/charge capacity of 103 mAh g-1/110 mAh g-1 and the highest discharge capacity of 130 mAh g-1 in the 6th cycle at C/20 rate in the cell-configuration system. In operando synchrotron diffraction and in operando X-ray absorption spectroscopy together with ex situ Raman and X-ray photoelectron spectroscopy reveal the reversibility of magnesium insertion/extraction and provide the information on the crystal structure evolution and changes of the oxidation states during cycling.
Article
Dendrite formation is one of the most pressing issues in nowadays battery research. Lithium based batteries are prone to forming short-circuit causing dendrites, while magnesium based batteries are not. Recently it was proposed that the tendency towards dendrite growth is related to the height of the self-diffusion barrier with high barriers leading to rough surface growth which might subsequently cause dendrite formation which was supported by density functional theory calculations for Li, Na and Mg [ J. Chem. Phys. 141, 174710 (2014)]. We now extend this computational study to zinc and aluminum which are also used as battery anode materials, and we additionally consider diffusion barriers that are relevant for three-dimensional growth such as barriers for diffusion accross steps. Our results indicate in agreement with experimental observations that Li dendrite growth is an inherent property of the metal, whereas Zn dendrite growth results from the loss of metallic properties in conventional Zn powder electrodes.
Article
Mechanistic understanding of lithium electrodeposition and morphology evolution is critical for lithium metal anodes. In this study, we deduce that Li deposition morphology evolution is determined by the mesoscale complexations that underlie due to local electrochemical reaction, Li surface self-diffusion, and Li-ion transport in the electrolyte. Li-ion depletion at the reaction front for higher reaction rates primarily accounts for dendritic growth with needle-like or fractal morphology. Large Li self-diffusion barrier, on the other hand, may lead to the formation of porous Li film for lower reaction rates. Enhanced ion transport in the electrolyte contributes to homogeneous deposition, thereby avoiding nucleation for Li dendrite formation. This study also demonstrates that the substrate surface roughness strongly affects dendritic growth localization over the protrusive surface features. A non-dimensional electrochemical Damkohler number is further proposed, which correlates surface diffusion rate and reaction rate, and allows constructing a comprehensive phase map for lithium electrodeposition morphology evolution.
Article
The known crystal structures of solids often correspond to the most thermodynamically stable arrangement of atoms. Yet, oftentimes there exist a richly diverse set of alternative structural arrangements that lie at only slightly higher energies and can be stabilized under specific constraints (temperature, pressure, alloying, point defects). Such metastable phase space holds tremendous opportunities for non-equilibrium structural motifs, distinctive chemical bonding, and ultimately for realization of novel function. In this feature article, we explore the challenges with the prediction, stabilization, and utilization of metastable polymorphs. We review synthetic strategies that allow for trapping of such states of matter under ambient temperature and pressure including topochemical modification of more complex crystal structures; dimensional confinement wherein surface energy differentials can alter bulk phase stabilities; templated growth exploiting structural homologies with molecular precursors; incorporation of dopants; and application of pressure/strain followed by quenching to ambient conditions. These synthetic strategies serve to selectively deposit materials within local minima of the free energy landscape and prevent annealing to the thermodynamic equilibrium. Using two canonical early transition metal oxides, HfO2 and V2O5, as illustrative examples where emerging synthetic strategies have unveiled novel polymorphs, we highlight the tunability of electronic structure, the potential richness of energy landscapes, and the implications for functional properties. For instance, the tetragonal phase of HfO2 is predicted to exhibit an excellent combination of a high dielectric constant and large bandgap, whereas ζ-V2O5 has recently been shown to be an excellent intercalation host for Mg batteries. Despite recent advances, the discipline of metastable periodic solids still remains substantially dependent on empiricisms given current inadequacies in structure prediction and limited knowledge of energy landscapes. The close integration of theory and experiment is imperative to transcend longstanding chemical bottlenecks in the prediction, rationalization, and realization of new chemical compounds outside of global thermodynamic minima.