Chris Funch Advancement Independent Proposal 4/5/17 1 Separating the Effects of Mg 2+ Content, Structural Changes, and Grain Properties on the Ionic Conductivity of Solid Magnesium Electrolytes Abstract Solid state rechargeable batteries can provide increased energy density for powering small- scale electronics. 1,2 Metal anodes of Mg ion batteries have a theoretical capacity of 3830 mAh cm - 3 compared to the theoretical 2060 and 1140 for Li and Na respectively. 1,3 However, use of metallic anodes in liquid electrolyte batteries to date has been limited by chemical incompatibilities between liquid electrolytes and electrodes. 4,5 Solid electrolytes present one solution to that problem but have low ionic conductivities (< 10 -4 S cm -1 at room temp versus 10 -3 –10 -1 for liquid electrolytes) limiting power output. 2,6 Aliovalent doping of lithium- and sodium-based solid electrolytes can increase the ionic conductivity of those materials through the generation of mobile ion vacancies or increased ion content. 7–11 It is proposed to study the effects of aliovalent doping on the ionic conductivity and activation energy of MgZr4(PO4)6 and (MgxHf1-x)4/(4-2x)Nb(PO4)3, both with the sodium super ionic conductor structure. I hypothesize substitution of tetravalent species (Zr 4+ or Hf 4+ ) via doping with trivalent ions (Sc 3+ , Y 3+ , and Al 3+ ) of different ionic radius will increase magnesium content, modify lattice parameters of the crystal phase, and influence grain size or grain boundary composition, which correlate with conductivity. Solid solutions with varying compositions of Sc 3+ , Y 3+ , and Al 3+ will be synthesized for characterization of their conductivity and structural properties via impedance spectroscopy, x-ray diffraction, and scanning electron microscopy with energy dispersive spectroscopy. This will aid in the separation of the parameters that can increase conductivity, guiding future research for further improvements. Understanding how tuning the structure and mobile ion concentration can maximize conductivity in magnesium ion batteries will help build the next generation of safe and cost-effective batteries.
12
Embed
Chris Funch Advancement Independent Proposal 4/5/17 ...
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Chris Funch Advancement Independent Proposal 4/5/17
1
Separating the Effects of Mg2+ Content, Structural Changes, and Grain
Properties on the Ionic Conductivity of Solid Magnesium Electrolytes
Abstract
Solid state rechargeable batteries can provide increased energy density for powering small-
scale electronics.1,2 Metal anodes of Mg ion batteries have a theoretical capacity of 3830 mAh cm-
3 compared to the theoretical 2060 and 1140 for Li and Na respectively.1,3 However, use of metallic
anodes in liquid electrolyte batteries to date has been limited by chemical incompatibilities
between liquid electrolytes and electrodes.4,5 Solid electrolytes present one solution to that problem
but have low ionic conductivities (< 10-4 S cm-1 at room temp versus 10-3–10-1 for liquid
electrolytes) limiting power output.2,6 Aliovalent doping of lithium- and sodium-based solid
electrolytes can increase the ionic conductivity of those materials through the generation of mobile
ion vacancies or increased ion content.7–11 It is proposed to study the effects of aliovalent doping
on the ionic conductivity and activation energy of MgZr4(PO4)6 and (MgxHf1-x)4/(4-2x)Nb(PO4)3,
both with the sodium super ionic conductor structure. I hypothesize substitution of tetravalent
species (Zr4+ or Hf4+) via doping with trivalent ions (Sc3+, Y3+, and Al3+) of different ionic radius
will increase magnesium content, modify lattice parameters of the crystal phase, and influence
grain size or grain boundary composition, which correlate with conductivity. Solid solutions with
varying compositions of Sc3+, Y3+, and Al3+ will be synthesized for characterization of their
conductivity and structural properties via impedance spectroscopy, x-ray diffraction, and scanning
electron microscopy with energy dispersive spectroscopy. This will aid in the separation of the
parameters that can increase conductivity, guiding future research for further improvements.
Understanding how tuning the structure and mobile ion concentration can maximize conductivity
in magnesium ion batteries will help build the next generation of safe and cost-effective batteries.
Chris Funch Advancement Independent Proposal 4/5/17
2
1.0 Introduction
Battery technology has become
increasingly necessary as its utilization to
supply stored energy from the mega to the
micro scale (e.g. grid-level storage, electric
vehicles (EVs), portable user electronics, the
Internet of Things (IoT), and microsensors) has
expanded. While transistors have been able to
decrease in size and power utilization by
several orders of magnitude in the past three
decades, battery technology has only improved
by a factor of ~2.5 (measuring energy density,
Figure 1).12 As such, batteries are currently the
limiting factor for further decreasing the size of
small scale electronics compared to EVs or grid storage where size is less of a restriction.
Currently, lithium-ion batteries (LIBs) have the most widespread use and are extensively
researched due to their high power and energy densities (gravimetric and volumetric) compared to
other conventional chemistries (Figure 1).13,14 In an effort to optimize a battery’s general
performance criteria (i.e. high specific energy, high specific power, low cost, long life, wide
stability window, toxicity, and safety) for
specific applications there is also ongoing
research into solid state LIBs.5,15 All-solid-
state batteries can offer improved safety and
increased capacity, primarily due to the
removal of both organic liquid electrolytes
and filler materials necessary to contain the
electrolyte or physically separate the
electrodes (Figure 2).2 They can also offer
improved thermal stability, reduced self-
discharge rates, and a wider operational
window under different environmental
conditions.1 Similar to liquid electrolyte
batteries, there have been numerous
investigations into solid-state batteries, but no
substantial progress has been realized due to
their reduced ionic conductivity and
subsequently lower power output.
Figure 2: Progress of battery energy density over the past few
decades. Overcoming challenges with lithium ion batteries could
limit its ability to keep up with demand. Magnesium ion batteries
could offer improved energy density and are even better on a
volumetric comparison. Adapted from ref 40.
Figure 1: (left) Schematic of rechargeable liquid LIB. Due to safety
concerns, the Li metal anode is often replaced by graphite, which
lowers the overall capacity. A separator is needed to keep the two
electrodes from contacting each other and shorting the battery.
Additional filler material is needed to encapsulate the liquid
electrolyte. Adapted from ref 14. (right) Schematic of an all solid-
state LIB. Here. A lithium phosphorous oxynitride (LiPON) is a
solid electrolyte. It serves the dual purpose as electrolyte and
separator between electrodes. Overall cell volume is reduced
relative to liquid counterpart, enabling more cells in same battery
volume.
Chris Funch Advancement Independent Proposal 4/5/17
3
There is potential for magnesium ion batteries
(MIBs) to compete with LIBs on both a capacity and
cost standpoint for small scale devices (Figure 3).5,16
The gravimetric charge capacity of magnesium may
not be as good as lithium for devices requiring low
weight (e.g. EVs) but its divalent nature offers almost
twice the volumetric charge capacity of lithium (and
more than five times that of the more common graphite
anode), which is advantageous for smaller devices.
However, even the higher conductivity materials
investigated for MIBs are not conductive enough to
meet some of the power requirements for low power
(<1mW) devices at room temperature. Investigations
into doping and structure for LIBs, and similar
monovalent ions, has seen recent advances.17–20 While
some conclusions may be applied to MIB materials,
low ionic conductivity is still limiting MIB usage.
In order to realize the true potential of MIBs, a more fundamental understanding of the
factors influencing divalent ionic conductivity in the solid electrolyte is needed. I have three
hypotheses to improve this component: (1) If the conductivity is limited by the concentration of
mobile ions, then substitution of a tetravalent with a trivalent cation will increase the concentration
of mobile, divalent cations (due to the need for charge neutrality) and increase conductivity. (2) If
the conductivity is limited by the size of the migration pathway (i.e. “bottleneck”) then substitution
of a larger ionic radius cation will increase the lattice parameter, lowering the activation energy
for mobile ion hopping to occur in the bottleneck, and increasing the conductivity. (3) If the
increased composition of dopant atom, or subsequently Mg2+ ions, either increases the average
grain size or increases the conductivity of the grain boundaries, then the overall conductivity
should also increase. These hypotheses will be tested by the doping of aliovalent species (Sc3+,
Y3+, and Al3+) in two systems with the sodium super ionic conductor (NASICON) structure
(general composition NaZr2(SiO4)x(PO4)3-x) : (MgZr4(PO4)6 and (MgxHf1-x)4/(4-2x)Nb(PO4)3). Ionic
conductivity, structure, grain size and grain composition of these materials will be determined with
impedance spectroscopy (IS), x-ray diffraction (XRD), scanning electron microscopy (SEM), and
energy dispersive spectroscopy (EDS) techniques, respectively. The synthesis and characterization
of these systems will help separate the impact of each component on bulk conductivity and provide
an improved understanding of divalent ionic crystals and the key parameters influencing their ionic
conductivity.
2.0 Background: Magnesium ionic conductors
The design of higher conducting Mg2+ systems often mimic those of Li+ or Na+. This
includes systems with the lithium phosphorous oxy-nitride (LIPON)21 and NASICON
structures.19,22 Of interest to this proposal are the two systems with the NASICON structure.
In one of the earlier reports of a Mg ionic conductor, Ikeda et al. synthesized different
compositions of MgxZr(3y-2x)/4(PO4)y to best match the structure of NaZr2(PO4)3.19 They identified
the composition MgZr4(PO4)6 best matched the target structure with XRD. They do not elaborate
Figure 3: In addition to being ~200 time more
abundant than Li metal, Mg also offers two times the
volumetric capacity compared to Li, and five times that
of the graphite anode. It also has a similarly low
reduction potential, which is necessary to deliver higher
power output. Data adapted from refs 1 and 40.
Chris Funch Advancement Independent Proposal 4/5/17
4
on how the other structures differ from the NaZr2(PO4)3 “target.” It is difficult to draw any
conclusions with their other compositions since their bulk structure, overall composition, and Mg2+
content are not the same. There has been no exploration of the impact of doping with aliovalent
ions on the structure or concentration of mobile ions for this system.
Tamura et al. investigated the (MgxHf1-x)4/(4-
2x)Nb(PO4)3 system.22 Their primary focus was also on
structure. For 0.05 ≤ x ≤ 0.3, the collected XRD patterns
showed the transition from a NASICON structure to a
different phase around x ≤ 0.1. This transition at x = 0.1
resulted in the distortion of Zr2P3O18 “lantern” units, which
are ordered in the NASICON structure (Figure 4). They
suggest the smaller ionic radius of Mg2+ compared to Hf4+
influences this distortion. They did not investigate any
compositions with additional dopants, besides their
substitution of Mg2+ for Hf4+. Additional dopants could
enable increased concentration of Mg ions to charge
balance the system, but this could also come at the expense
of further structural distortions.
Grain size and phase segregation at
boundaries is also something that has not been
reported for either of these systems. The
impact of dopants, and other synthesis
methods on grain morphology have been
investigated for similar monovalent systems
and show an increase in conductivity.23–26
While these solid electrolytes highlight
improvement in the production and
understanding of Mg2+ conductivity, there is
opportunity to further both these endeavors
through aliovalent doping of these materials.
These undoped systems still have greatly
reduced conductivities relative to monovalent
compounds with similar structures (Figure 5).
The decreased conductivity is often attributed
to the increased electrostatics of divalent
versus monovalent ions. However, it is not
clear what factors can counter that force most effectively to improve overall conductivity. This
work will separate the factors that could play a role in conductivity through purposeful selection
of dopants and a systematic investigation of each potential factor, enabling a greater understanding
of how to improve conductivity.
Figure 5: Conductivities of the two Mg systems in this proposal
along with two monovalent systems with the same bulk structure.
The increased electrostatics for the divalent species is one primary
cause for the reduced conductivity. While the monovalent systems
are in a viable range for battery usage (room temperature
conductivity above 10-4 S cm-1) improvement is still needed for
divalent conductors. Data adapted from refs 19, 22, 41, and 42.
Figure 4: Representation of the NASICON
Structure. Alternating octahedral and
tetrahedral bodies form the framework for
Mg2+ to migrate through. The migration
pathway is highlighted by the black arrows.
Adapted from ref 28.
Chris Funch Advancement Independent Proposal 4/5/17
5
3.0 Project description
Goal: To understand role of Mg2+ content, structural changes, and grain properties in the ionic
conductivity and activation energy of Mg-based solid electrolytes.
Research Objectives:
Objective 1 – Determine role of Mg2+ composition on ionic conductivity of of Mg(x+2)/2Zr4-
xMx(PO4)6 and Mgx/2Hf1-xMxNb(PO4)3 where Mx is Sc3+. These systems will be referred to as
MgZPO-M and MgHNPO-M, respectively, with M being the dopant atom.
Objective 2 – Determine role of lattice parameter and migration channel dimensions on ionic
conductivity and activation energy of systems identified above where Mx is Sc3+, Y3+, or Al3+.
Objective 3 – Determine impact of grain size and phase segregation at grain boundaries on the
conductivity and activation energy of the systems identified above.
4.0 Approach
4.1 Objective 1
4.1.1 Background
At low temperatures, a pure stoichiometric ionic solid has ions in all of the
crystallographically equivalent sites. These “normal” sites serve as the lowest energy positions for
the ions. In a solid electrolyte, ionic conductivity is only possible due to the presence of defects
where an ion is either absent from its normal site or exists in an interstitial site or lattice disorder.
The difference between the highest normal-site energy and the lowest interstitial-site energy serves
as the activation energy for ion migration, similar to the electronic energy gap of an insulator or
semiconductor. Defects can be classified as intrinsic or extrinsic. Intrinsic defects exist in
thermodynamic equilibrium with the lattice. These types of defects include vacancy and interstitial
for the ions in the pure crystal. As the crystal is heated close to its melting point, the rates of ion
motion are high enough to establish an equilibrium between their normal sites and a defect site.
This type of lattice disorder is reversible. In general, cations are more likely to form interstitial
defects than their anion counterparts due to their smaller ionic radii. Extrinsic defects can form
from the introduction of aliovalent (doping) ions into a lattice, while grain boundaries and
dislocations are another type of extrinsic defect. Extrinsic defects are much more dependent on the
preparation and thermal history of the crystal and are not in an equilibrium state. These defects
generate irreversible disorder and are the type under investigation in this study.
In addition to defects there can also be other mobile species besides Mg2+ in solid
electrolytes. The transport, or transference number (ti), is the ratio of the conductivity of the
working ion to the sum of all ionic and electronic conductivities.
𝑡𝑖 ≡ 𝜎𝑖 𝜎⁄ (1)
Here, σ is the sum of all ionic conductivities and any electrical conductivity. Ideally, the value of
ti is unity, since there should be no electronic conductivity and only the working ion should be able
to move through the electrolyte. In liquid electrolytes, both cations and anions are mobile, which
results in significantly lower transference numbers. For instance, common organic electrolytes for
LIBs have values between 0.2 - 0.5, which means a solid electrolyte can be 5 - 2 times lower in
conductivity, respectively, and have equivalent conductivity of the working ion.17 Thus, it is
necessary to ensure Mg2+ is the only mobile ion in these systems. A DC bias was applied across a
sample of the MgZPO electrolyte and composition analysis (e.g. electron probe micro analysis)
Chris Funch Advancement Independent Proposal 4/5/17
6
revealed a buildup of only Mg2+ at the cathode.19 Similarly, IS tests under different ambient air
conditions (e.g. O2 rich vs O2 free) eliminated O2- as a mobile ion in MgHNPO.22 Based these
investigations into additional mobile species (e.g. Zr4+, P4+, O2-, e-, or h+) in both parent systems it
is assumed that Mg2+ is the only mobile species.19,22,27
4.1.2 Experimental Method
There is no knowledge on the possible change in conductivity of MgZPO or MgHNPO
systems when doped with aliovalent cations. The intrinsic conductivity of both these materials has
been reported to be ~1x10-4 S cm-1 at 600 °C,19,22 but it is not known whether doping could increase
conductivity or reduce activation energy. The first objective is a preliminary demonstration on the
effects of doping on the conductivity of Mg ion conductors. This will be done by synthesizing both
previously reported compositions along with the doped compositions, listed in Table 1, and
measuring their conductivities over a range of temperatures.
A solution-assisted solid-state reaction (SASSR), described by Ma et al. will be used to
synthesize the compounds of interest.28 Briefly, corresponding amounts of Mg(NO3)2, ZrO(NO3)2,
HfCl4, NbCl5, Sc2O3, Y2O3, and Al(NO3)3 solids are dissolved in either deionized water or HNO3
and mixed into one solution. A calculated amount of NH4H2PO4 is then added while stirring. The
mixture is dried and then calcined at 800°C. The powder is then milled in ethanol with zirconia
balls before being pressed into a pellet and sintered at ~1300 °C. These are the same, or similar,
precursors used in the co-precipitation methods used for the original compounds and should
provide reasonable comparisons. While compounds with Sc3+, Y3+, and Al3+ will all be synthesized
in this proposal, only Sc3+ will be used in Objective 1 assuming it will not have a significant change
in structure due to similar ionic radii between Sc3+, Zr4+, and Hf4+ (Figure 7). This will isolate any
changes in conductivity due to the expected change in Mg2+ content.
Before conductivity measurements are made, initial XRD patterns for the parent systems
and Sc3+ doped systems will be compared to confirm they have the same bulk structure. Once
confirmed, temperature dependent IS will be used to measure the conductivity, and extract
activation energy, for the prepared samples. Ion blocking electrodes (e.g. platinum or gold) are
applied to both sides of an electrolyte sample (Figure 6). At a given temperature, an AC bias is
applied to the sample over a range of frequencies. An equivalent circuit for the electrolyte enables
the calculation of both the grain and grain boundary resistance. The resistance is inversely related