Microstructural characterization and multiscale ionic conductivity in lithium and sodium-based solid state electrolytes

In order to meet the increasing need for energy storage systems in consumer devices,
electric vehicles and stationary energy storage devices, the existing battery technologies
are constantly being further developed. An important criterion for the performance of
the battery is its energy density. Even if the cathode accounts mostly for the weight of
a lithium ion battery cell, there is a promising possibility of weight saving on the anode
side by replacing graphite with pure lithium metal. To this end, an electrolyte needs to
be developed that is stable against the potential of the pure metal. Liquid electrolytes
cannot meet this requirement and also involve safety risks as
ammability. Solid-state electrolytes are supposed to enable the use of a lithium-metal electrode.
Additionally, due to the scarcity of resources for lithium, sodium-based technologies are
being developed for use as stationary energy storage devices.
Thiophosphates feature the highest ionic conductivities among all solid electrolytes. In
many cases, amorphous thiophosphates offer even higher conductivities than their crystalline
analogues and their structure can differ, too. ... mehrA structural investigation of amorphous
compounds may not be possible with a standard X-ray analysis. However their
local structure has to be enlightened so that a reproducible synthesis can take place.
The solid electrolyte Na2P2S6 was synthesized via ball milling in an amorphous state
with subsequent crystallization. The structure of the crystalline phase differs markedly
compared to the corresponding amorphous phase. A combination of XRD-PDF analysis
and 23Na/31P MAS NMR spectroscopy measurements indicate that single PS30-4 tetrahedra
and corner-sharing tetrahedra are transformed to edge-sharing-tetrahedra during
crystallization of amorphous Na2P2S6 to crystalline Na2P2S6. Impedance spectroscopy
shows that amorphous Na2P2S6 has a conductivity of 5.710-8 S cm-1 which is three
orders of magnitude higher than crystalline Na2P2S6 (2.6 10-11 S cm-1). The higher
conductivity can also be recovered by ball milling crystalline Na2P2S6, inducing a reamorphization.
Lithium Lanthanum Zirconium oxide (LLZO) exhibits a high ionic conductivity and
stability against lithium metal and is therefore a promising candidate as solid state electrolyte.
Yet, specifcations on its conductivity are often not reliable and spread widely.
Attempts are made to attribute the differences in reported conductivities to the different
substituents, sintering times or surface passivations. A microstructural comparison of
four differently substituted samples is performed to elucidate the reasons for the different
conductivities.
X-ray diffraction revealed that commercial LLZO samples crystallize in the hydrogarnet
structure (space group No. 220), which is described for the first time with a substituent
on Zr and La sites. Ball milling of Al3+, Nb5+, Ta5+ and W6+ substituted LLZO results
in a phase transformation from the garnet structure into the hydrogarnet structure
with a lower symmetry. The distribution of lithium ions in the hydrogarnet structure
differs from that in the garnet structure which was investigated with 6Li MAS NMR and
neutron diffraction. A targeted conversion of the hydrogarnet structure into the garnet
structure is proved by calcining the material at 1100 °C for 10 h. With high-temperature
X-ray diffraction, an low thermal expansion of the hydrogarnet unit cell is observed in
comparison to the greater expansion of the garnet unit cell. The ionic mobility of Li
ions in the hydrogarnet structure is examined by means of NMR, in particular line shape
analysis, relaxometry and pulsed-field gradient NMR. This combination of techniques
shows that the mobility of lithium is significantly reduced on small length scales. In
combination with the structural analysis, this can be traced back to the high occupancy
of the Li3 position in the hydogarnet structure, blocking long-range lithium diffusion.
However, it was not possible to access the long-range mobility of Li in the hydrogarnet
structure (at 25 °C).
Therefore, the contribution of the ceramic component to the total ionic conductivity
of polymer composite electrolytes is evaluated. The question whether the long-range
lithium mobility in the hydrogarnet structure is lower compared to the garnet structure
is assessed without the necessity for sintering the LLZO to pellets. Impedance spectroscopy
shows a conductivity of 1.2 10-6 S cm-1 for a composite electrolyte with a
hydrogarnet structure and 3.4 10-6 S cm-1 for a composite electrolyte with a garnet
structure. The higher Li mobility of the garnet-based composite electrolyte compared
to the hydrogarnet-based electrolyte is verified with PFG-NMR measurements of the
diffusion coeffcient: 6.1 10-14 m2 s-1 (garnet), resp. 1.1 10-14 m2 s-1 (hydrogarnet).
The measured activation energy of dffusion is also higher in the hydrogarnet composite.
The conductivity results measured with impedance spectroscopy are compared with
commercial composite electrolytes; a SiO2-ceramic-polymer and a purely polymer-based
electrolyte.
The next step from optimizing a solid state electrolyte in terms of ionic conductivity is
to look at its compatibility with the electrodes, here the cathode. It is tested whether
ball milling of LLZO with the established cathode material Lithium Nickel Cobalt Manganese
oxide (NCM) results in a good contact of the two materials and consequently a
low Li ion diffusion barrier.
The interface between LLZO and NCM is investigated by X-ray diffraction, 6Li MAS
NMR and transmission electron microscopy. A model system consisting of LLZO and
NCM is characterized with impedance spectroscopy for a lithium diffusion barrier sandwiched
between an auxiliary electrolyte in order to separate the ionic conductivity from
the electrical. An evaluation of only the ionic conductivity apart from the electrical
conductivity is not possible due to the high electrical conductivity of the auxiliary electrolyte.
The electrolyte Lithium Aluminum Titanium Phosphate (LATP) is examined crystallographically
against the background of upscaling of the synthesis. If the process is
upscaled, local inhomogeneities of the educts can be expected in a way that varying
educt contents have an effect on the product. This applies especially to phosphoric acid,
the concentration of which cannot be specified precisely due to its hygroscopy. A Rietveld
refinement analysis against X-ray diffraction data of LATP with varying phosphoric acid
content during synthesis is performed. An excess of phosphoric acid leads to the formation
of AlPO4, which impedes the ionic conductivity. Insufficient phosphoric acid causes
the formation LiTiOPO4. TiO2 is formed in this material after a second sintering step.
The findings in this work contribute the understanding of structural changes in solid
electrolytes during processing and thus contribute to the improvement of future solidstate
batteries.