Page 129 - Read Online

Page 6 of 14 Shi et al. Energy Mater 2023;3:300036 https://dx.doi.org/10.20517/energymater.2023.27

extrusion of a membrane with the liquid phase. PC was used instead of EC to avoid any crystallization down
to the melting point of PC (-48.5 °C), even at high solvent contents. In fact, herein, a PC content of up to
70 wt% was used in combination with a polymer with shorter hydrophobic blocks than previously reported
(5,000 vs. 15,000 M ) for improving the polymer electrolyte processability.
w

SANS measurements were performed to assess the phase separation. The results are shown in Figure 1A for
the PTFSI-based polymer with 30 wt% of PC with increasing ionic block length: PTFSI-5/5-30 [i.e., with
ionomeric and (respectively) non-ionic blocks of 5,000 g mol and 30 wt% PC], PTFSI-10/5-30, and PTFSI-
-1
15/5-30 to compare the different block copolymers. On the other hand, the effect of increasing solvent
fractions is shown for PTFSI-10/5. The phase separation is overall less marked than what was obtained with
DMSO-cast membranes but similar to EC-membranes obtained using DMAc as the casting solvent. The
comparison of the different polymers shows that the PTFSI-15/5-30 phase separation is less marked than for
the membranes with shorter ionic blocks (PTFSI-5/5-30 and PTFSI-10/5-30). The phase separation of the
PTFSI-10/5 electrolytes is well maintained with increasing PC content, and all electrolytes exhibit a
hydrophobic domain size of 13 nm.

The conductivity of the PTFSI-10/5-70 electrolyte shown in Figure 1B is ca. ten times lower than that of the
liquid reference (1M LiTFSI in PC). It must be noted, however, that the liquid electrolyte requires the use of
a separator to be used in battery cells that would induce a decrease by a factor of 10 of the conductance of
[29]
the (electrolyte + separator) assembly . Besides, the conductivity is linked to anions and cations for the
liquid electrolyte. Probably due to poorer phase separation of hydrophobic domains and thus a more
tortuous ionic pathway, in this case, the ionic conductivities are somewhat lower than those previously
reported for EC-plasticized membranes obtained via DMSO casting and subsequent swelling. The PTFSI-
-4
-1
-1
-5
10/5-70 membrane, however, still reaches 2.15 × 10 S cm at 20 °C and 8.98 × 10 S cm at 0 °C, which, for
a single-ion conductor that is able to maintain a fast lithium transport at steady-state is, in principle,
sufficient for the application .
[30]
In fact, even the electrolyte with the highest PC content (PTFSI-10/5-70) exhibits a t of 1 as no diffusion
+
+
was detected for the anionic moiety, even at 90 °C (see Table 1). A Li self-diffusion coefficient of
2 -1
-10
0.31 × 10 m s was determined at 30 °C, which is more than 50 times higher than a reinforced PEO/ionic
liquid (IL)/LiTFSI plasticized polymer electrolytes (0.6 × 10 m s at 25 °C ). Moreover, it is similar, at
-12
2 -1
[31]
70 °C, to what was measured for a PTFSI-15/15 polymer with 30 wt% of EC obtained using a DMSO cast
membrane (0.95 vs. 0.96 × 10 m s ).
-10
2 -1
-1
-3
With a lithium content of 1.15 × 10 mol g in the dry PTFSI-10/5 polymer, if we postulate a density of 1.2
for the final electrolyte, the lithium content in the plasticized electrolyte corresponds to ca. 0.29 mol L . This
-1
is relatively low compared to the most conductive concentrations in most organic lithium conducting
electrolytes [e.g., PEO/IL/LiTFSI (plasticized or not) electrolytes or LiTFSI in PC]. Hence, it is likely that
increasing the grafted lithium salt concentration would lead to improved conductivity as an effect of
increased charge carrier concentration. Noticeably, the PTFSI-10/5-50 electrolyte exhibits a conductivity
close to that of the PTFSI-10/5-70 electrolyte at high temperatures, indicating that the higher charge density,
resulting from the lower PC content, is favorable at higher temperatures (where ionic mobility is higher). If
we calculate the theoretical conductivity from the Nernst-Einstein equation (Equation 3) and D , we reach a
Li
-4
molar conductivity of 1.14 S cm mol and a Li conductivity of 3.3 × 10 S cm at 30 °C, which is only 12%
-1
-1
+
2
higher than the experimentally determined value of 2.9 × 10 S cm . This is in excellent agreement with
-1
4
single-ion conductivity and indicates that the ionic functions are well dissociated. The small difference can
be easily accounted for (besides the experimental errors and the hypothesis on density) by the presence of
the non-conductive phase.