IN-SITU POLYMERIZED SOLID ELECTROLYTES FOR LITHIUM BATTERIES A Dissertation Presented to the Faculty of the Graduate School of Cornell University in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy by Nyalaliska Wardhianingtyas Utomo August 2024 © 2024 Nyalaliska Wardhianingtyas Utomo ALL RIGHTS RESERVED IN-SITU POLYMERIZED SOLID ELECTROLYTES FOR LITHIUM BATTERIES Nyalaliska Wardhianingtyas Utomo, Ph.D. Cornell University 2024 The world is moving towards a demand for safer Lithium batteries in which tra- ditional volatile liquid electrolytes are replaced by solid-state ones. Methods for fabricating and integrating such solid-state electrolytes (SSEs) in rechargeable batteries have been reported in a large number of literature articles and review papers. SSEs are viewed as a requirement for safe operation of lithium batter- ies that use metallic lithium as the anode, i.e. Lithium-metal batteries (LMB). SSEs are presently challenged by poor room-temperature ionic conductivity in the bulk and by complex interfacial chemistry and poor ion transport through the interfaces that SSEs form with the LMB anode and cathode. Recently, solid polymer electrolyte (SPE) composed of poly(1,3-dioxolane) (polyDOL) formed inside electrochemical cells by ring-opening polymerization (ROP) of the ethers have received attention for their potential to ameliorate poor interfacial contact between the electrodes and solid electrolyte, enabling a more uniform charge transport. ROP of ethers is unfortunately reversible due to the sensitivity depen- dence of the equilibrium constant on the relative difference in ring strain in the monomer and polymer. An undesirable consequence is that the ROP of ethers is typically accompanied by large amounts of residual monomer, decreasing the mechanical strength of SPE as well as increasing the potential for parasitic electrochemical side reactions during battery operations. Research summarized in this thesis investigate two potential solutions to this issue: mechanical rein- forcement of the in-situ formed SPE using particles with affinity for the growing polyether; and co-polymerization of the ether with multifunctional variants that serve as anchors for the growing chains. Research towards the first solution considers the role of nano- and micro- sized fillers on the ROP of 1,3-dioxolane. We find that the particles have di- rect and indirect effects on SPEs. They alter the polymerization kinetics, ionic conduction mechanism, and in the case when nano-sized SiO2 grafted with short chains (aka, PEG-SiO2 hairy nanoparticles (HNPs)) may also introduce co-crystallization of polyDOL with PEG tethers. The co-crystallization creates a uniform structure, increasing room- temperature ionic conductivity to as high as 4 mS/cm. The introduction of micron-sized Li2O, on the other hand, produce rather different effects on particle and electrolyte length scales. At the particle scale, the physical size, hardness, basicity, and gravitational settling tendencies of the particles lead to SPEs with interesting gradient physical properties. The basic Li2O particles for instance retards ROP of DOL in the particle-rich regions of an SPE, producing materials that become progressively more solid-like away from interfaces at which the particle concentration is enlarged by settling. At the electrode scale,Li2O functions an electro-active material and is capable of undergoing reversible redox reactions at sites near the electrode. These pro- cesses contribute lithium inside an otherwise closed electrochemical cell, en- abling extended cycling of lithium batteries in the most challenging anode-free configurations. Research towards the second solution considers the effects of crosslinking poly(DOL) on the reversibility of the polymerization reaction. It is found that as the degree of crosslinking increases, the residual monomer content at equi- librium falls. And, through straightforward manipulations of the cross-linker content alone, it is possible to create SPEs in a variety of soft matter states, with commensurate gradations in mechanical properties. As a demonstration of the potential benefits of the in-situ formed cross-linked polyDOL SPEs formed inside a conventional ethylene carbonate and dimethyl carbonate (EC/DMC) host, we investigated the electrochemical cycling characteristics of the materials in Li-Cu half cells and in full-cell batteries in which either graphite or metallic lithium are paired with a high-Nickel (NCM811) cathode. These studies show that the materials are mechanically robust with a wide electrochemical stability window, and good oxidative stability. BIOGRAPHICAL SKETCH Liska was born in the Indonesian city of heroes, Surabaya. She grew up in Surabaya and Bogor before moving to the United States as an 18 year old to pursue a Bachelor’s degree in Chemical Engineering at Pennsylvania State Uni- versity. She took immense interest in polymers and decided to minor in poly- mer science and engineering. Her passion brought her to researching bio-based polymeric solutions and their rheology under the guidance of Professor Ralph H. Colby. Her experience as an undergraduate researcher since she was a sopho- more taught her a lot about polymer physics research, and she decided to pur- sue a Master’s in Materials Science and Engineering within the same laboratory. She enjoyed research, and this led her to work toward her Ph.D. in Chemical En- gineering under the guidance of Professor Lynden A. Archer. Under his men- toring, she learned a lot about polymer electrolytes and lithium-metal batter- ies. She also discovered hairy nanoparticles and their fundamental properties, strengthening her polymer science foundation. iii This dissertation is dedicated to my parents, brothers, and partner. iv ACKNOWLEDGEMENTS I would like to first and foremost thank my Ph.D. advisor, Professor Lynden A. Archer, for his continuous guidance and mentorship. Our insightful meetings and his inspiring stories are what motivate me. I learned not only of scientific pursuits, but also a plethora of different life skills. I would also like to thank my committee members Professor Yong L. Joo and Professor Geoffrey W. Coates for their support and guidance throughout my Ph.D. I am forever grateful to my undergraduate and Master’s advisor Professor Ralph Colby and mentor Dr. Behzad Nazari for introducing me to the world of polymers. I am thankful for the support my friends have given me. Throughout this journey, I have been walking together with Prince Ochonma, Apoorva Jain, and Ayushi Tripathi. I also thank everyone in the Archer lab, especially Dr. Xiao- tun Liu and Professor Qing Zhao for their mentorship in the first year of my Ph.D. For their companionship and support, I am grateful for Dr. Yue Deng, Dr. Regina Garcia, Dr. Ken Kim, Dr. Arpita Sharma, Ritwick Sinha, and Samuel Baffour. I can be here today only because of the support of my parents, Fandi Utomo and Lucy Dyah Hendrawati. I am inspired to give as much love to my family and friends as they have given me. For being my true best friends, I thank my brothers Rio and Bram, this Ph.D. is for them too. Lastly, I am grateful beyond words for my partner, Vivaldi, without whom I would not know how to keep going. Quoting his favorite philosopher Albert Camus; ”The only way to deal with an unfree world is to become so absolutely free that your very existence is an act of rebellion.” v TABLE OF CONTENTS Biographical Sketch . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iii Dedication . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iv Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . v Table of Contents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . vi List of Tables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . viii List of Figures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ix 1 Introduction 1 1.1 Solid Polymer Electrolytes (SPEs) . . . . . . . . . . . . . . . . . . . 1 1.2 Ring-opening Polymerization (ROP) . . . . . . . . . . . . . . . . . 6 1.3 Mechanical Reinforcements Through Composite Electrolytes . . . 11 1.4 Forward-Favoring ROP . . . . . . . . . . . . . . . . . . . . . . . . 16 1.5 Outline of Dissertation . . . . . . . . . . . . . . . . . . . . . . . . . 18 2 Structure and Evolution of Quasi-Solid-State Hybrid Electrolytes Formed Inside Electrochemical Cells 19 2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19 2.2 Materials and Method . . . . . . . . . . . . . . . . . . . . . . . . . 24 2.2.1 Material Preparation . . . . . . . . . . . . . . . . . . . . . . 24 2.2.2 Characterization Methods . . . . . . . . . . . . . . . . . . . 27 2.3 Results and Discussions . . . . . . . . . . . . . . . . . . . . . . . . 31 2.4 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49 3 Anode-free Lithium Batteries Enabled by Polymer-Particle Hybrid Electrolytes 51 3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51 3.2 Materials and Method . . . . . . . . . . . . . . . . . . . . . . . . . 54 3.2.1 Electrolyte and battery preparations . . . . . . . . . . . . . 54 3.2.2 Material characterizations . . . . . . . . . . . . . . . . . . . 55 3.2.3 Electrochemical testing . . . . . . . . . . . . . . . . . . . . . 57 3.3 Results and Discussions . . . . . . . . . . . . . . . . . . . . . . . . 57 3.4 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74 4 In-situ synthesis of solid polymer electrolytes with high ionic mobility 76 4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76 4.2 Materials and Method . . . . . . . . . . . . . . . . . . . . . . . . . 79 4.2.1 In-situ fabrication of crosslinked SPE and electrochemical cells . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79 4.2.2 Characterization of crosslinked SPE . . . . . . . . . . . . . 79 4.2.3 Electrochemical characterizations . . . . . . . . . . . . . . 80 4.3 Results and Discussions . . . . . . . . . . . . . . . . . . . . . . . . 81 4.4 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91 vi A Chapter 2 appendix 92 B Chapter 3 appendix 106 C Chapter 4 appendix 121 Bibliography 133 vii LIST OF TABLES 1.1 Strain energies (kJ/mol) of cyclic compounds . . . . . . . . . . . 9 A.1 Transition temperatures Tc, Tm, Tg and heat of melting ∆Hm of self-suspended HNPs, poly(DOL), and hybrid systems obtained from DSC curve in Figure A.4 . . . . . . . . . . . . . . . . . . . . . 94 B.1 Three locations of Li2O@Cu were probed through XPS, with the resulting Li2O and Li2O2 percentage listed for both high resolu- tion O- and Li-scan. Survey scan shown in Figure 3.4g presents different atomic percentage of elements of O, C, and Li. . . . . . . 106 viii LIST OF FIGURES 1.1 Lithium coordination with ether oxygen in PEO backbone . . . . 4 1.2 Maxwell model fit for suspension electrolytes . . . . . . . . . . . 14 2.1 Temperature-dependent conductivities of neat and hybrid sam- ples probed through dielectric relaxation spectroscopy (DRS). (a) The conductivities of poly(DOL) (1 mM Al(OTf)3 and 10 mM Al(OTf)3), self-suspended PEG-SiO2 HNPs (ϕc = 15 vol.%), and hybrid system composed of both (ϕc = 2.7 vol.%). All samples include 2 M LiTFSI salt. (b) Temperature-dependent DC con- ductivity of hybrid system (ϕc = 2.7 vol.%) polymerized with 10 mM Al(OTf)3 for the 1st and 5th hour, 24th hour, 29th hour, 48th hour, and 10th day of polymerization. . . . . . . . . . . . . . . . . 35 2.2 Small angle X-ray scattering (SAXS) profiles to determine struc- ture of PEG-SiO2 HNPs/poly(DOL) hybrid material. (a) Inten- sity profile and (b) structure factor of HNPs/poly(DOL) with ϕc = 2.7 vol.% with initiator Al(OTf)3 content varying from 10 to 50 mM. (d) Structure factor analysis gives interparticle distance dp−p (red data points) and distance between tethers (green data points) as described in the diagram (c), as well as structure factor at q = 0. Dashed lines in (d) represent result for self-suspended HNPs with ϕc = 11 vol.%. . . . . . . . . . . . . . . . . . . . . . . . 37 2.3 Time-dependent dynamic shear flow measurements are used to study changes in polymerization kinetics induced by PEG-SiO2 HNPs in a poly(DOL) SPE. (a) Time sweep measurement of DOL monomers polymerized with 10 mM Al(OTf)3 with the addition of: (b) 10 mM LiNO3, (c) PEG-SiO2 HNPs (ϕc = 2.7 vol.%). (d) DOL polymerized with 1 mM Al(OTf)3 with PEG-SiO2 HNPs (ϕc = 2.7 vol.%). The inset in (d) shows hybrid 1 mM initiator system with the addition of 2 M LiTFSI. The inset in (a) repre- sents a strain sweep at 30-minute time-point during polymer- ization. The schematics of polymerization process going from monomers to amorphous polymer to polymer crystals were dis- played in (a). Dashed line in (b) shows data point from (a) as a comparison between poly(DOL) polymerization with and with- out LiNO3. The inset in (c) shows clearer peaks in G” and G’ at early time. All room temperature time sweep measurements were carried out at angular frequency of ω = 10 rad/s and strain of γ = 5%. Strain sweep was measured also at RT and ω = 10 rad/s. Al(OTf)3 contents of 1, 20 – 50 mM for each type of sam- ple are shown in Figures A.5 and A.6. . . . . . . . . . . . . . . . . 41 ix 2.4 Electrochemical performance of HNPs/poly(DOL) hybrid elec- trolyte. (a) Schematic of in-situ polymerization of DOL in the presence of HNPs. (b) Coulombic efficiency (CE) of various poly(DOL) electrolyte and hybrids containing PEG-SiO2 HNPs (ϕc = 2.7%) with and without 0.5 M LiNO3 at 0.1 mA/cm2 in Li//Cu cells (c) Galvanostatic cycling profile for Li//sPAN cell with hybrid electrolyte containing ϕc = 2.7%, 2 M LiTFSI, and 0.5 M LiNO3, with discharge capacity and Coulombic effi- ciency (CE) over cycle shown in (d). Rate performances for the Li//sPAN cells at different c-rates are shown in (e) with 1.0C is current density of 1 mA/cm2. . . . . . . . . . . . . . . . . . . . . . 47 3.1 (a) Schematic illustration of method used to synthesize Li2O/poly(DOL) hybrid electrolytes with attractive gradient properties produced by gravity settling of Li2O. (b) A cell design using a pair of Celgard 3501 separators is used to sequester the particle-rich phase to the region near the Li anode. (c) Results from Fourier transform infrared spectroscopy (FTIR) analysis of the anode and cathode facing sections of the separators. (d) SEM analysis of the structure of the cathode and anode-facing separa- tor surfaces, compared with that of the pristine Celgard material. It is apparent that polymerization of the DOL yields a material that covers the pores on the cathode-facing separator (e). How- ever, the separator on the anode side is seen to still retain the porous structure (f). . . . . . . . . . . . . . . . . . . . . . . . . . . 61 3.2 (a) Normalized ionic conductivity value following Maxwell model for Li2O/DOL suspension electrolytes at various volume fractions. (b) As polymerization of a Li2O/DOL hybrid elec- trolyte (1st layer: 10N7; 2nd layer: poly(DOL) +2 M LiTFSI + 1 mM Al(OTf)3), proceeds with time, the ionic conductiv- ity measured at 30°C first rises and stabilizes at values higher than observed either for the precursor Li2O/DOL suspension or pure poly(DOL) electrolyte with the same salt concentration. (c) Temperature-dependent ionic conductivity values for in-situ- formed Li2O/poly(DOL) hybrid electrolyte, for a Li2O/DOL suspension electrolyte, and for in-situ-formed poly(DOL) elec- trolyte. (d) Temperature-dependent changes in ionic conductiv- ity of an in-situ formed Li2O/poly(DOL) hybrid electrolyte as a function of time following the onset of polymerization. . . . . . 65 x 3.3 Galvanostatic charge and discharge of Li//NCM811 in hybrid Li2O/DOL electrolytes containing (a) 10 vol.% and (b) 50 vol.% Li2O and 2 M LiTFSI. (c) The discharge capacity and Coulombic efficiency (CE) of the cells in (a) and (b) over 120 cycles. Cells were run with Li//NCM811 configuration at C/10 for the first five cycles, followed by 1C for the rest of the cycle. (d) Coulom- bic efficiency obtained in Li//Cu half-cells configuration for dif- ferent Li2O volumetric concentrations. The resulting CE value at each concentration is presented in (e). . . . . . . . . . . . . . . . . 68 3.4 (a) Galvanostatic cycling performance and electrochemical prop- erties of Anode-free Cu//NCM811 cells based on Li2O/poly(DOL) electrolytes for 10 vol.% Li2O. The poly(DOL) was polymerized inside the battery cell using 1 mM Al(OTf)3 and the cells were cycled at a rate of 0.5 mA/cm2 (C/2). For the results in (a) a salt blend consisting of 2 M LiTFSI + 0.5 M LiNO3 (here termed N5) was used. The respective discharge capacities and Coulombic efficiencies (CE) values are reported in (b). Included in (b) are discharge capacity values of DOL + 2 M LiTFSI electrolyte with- out and with the addition of 10 vol.% Li2O, cycled at C/10 for 5 cycles and a comparable C/2 rate for the rest of the cycles. Solid electrolyte interphase (SEI) buildup during the formation step at C/10 is detailed in Fig. S12. CE values are used to predict capac- ity fading shown in the dashed line of (c), while solid data points show actual capacity fading. The difference between CE of this electrolyte and a control DOL + 2 M LiTFSI electrolyte is shown in (d). The corresponding Li//NCM811 cell performance for the same electrolyte as in (a) is shown in Fig. S11. (e) Cyclic voltam- metry (CV) measured in cells in which Li2O particles in a carbon cloth (CC) current collector are paired with Cu-foil in a DOL + 2 M LiTFSI electrolyte. (f) X-ray photoelectron spectroscopy (XPS) of Li2O@CC electrode post-CV and held at oxidation potential for 5 hours indicating existence of oxygen, carbon, and lithium. Two peaks attributed to lithium oxide (Li2O) and lithium perox- ide (Li2O2) are seen in the (g) O scan and (h) Li scan. . . . . . . . 70 4.1 (a) 1HNMR and 13CNMR spectra of 1 M LiPF6 + EC/DMC, polymerized UN05M, and crosslinked CR05M10. (b) FTIR spec- tra of DOL, polyDOL polymerized using 1 mM Al(OTf)3, and crosslinked CR05M10. (c) Polymerization to form branched polymers and crosslinking mechanism of branched polyDOL us- ing TMPTGE. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84 xi 4.2 Residual monomer content of different TMPTGE concentrations ranging from 0% (UN05M) to 10% (CR05M10). Values are ob- tained through 1HNMR peak integration of data shown in Fig- ure C.3. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85 4.3 (a) Temperature-dependent ionic conductivity showing Vogel- Fulcher-Tamman (VFT) behavior and (b) ionic conductivity val- ues at room temperature for different TMPTGE concentrations for CR1My. The y2-axis is complex viscosity of the CR05My sam- ples with the same variation of TMPTGE concentrations. . . . . . 86 4.4 (a) Time-dependent rheology measurement indicating the in- creasing storage (G’) and loss (G”) moduli with polymeriza- tion and crosslinking time for CR1M10. (b) Enhanced strain- dependent mechanical properties of G’ and G” with crosslink- ing, with the grey data points belonging to UN1M and blue to CR1M10. (c) Liquid-to-solid mechanical responses due to crosslinking probed through frequency-dependent rheology measurement. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88 4.5 Galvanostatic stripping and plating for 50 cycles of Li/NCM811 of (a) CR05M10, (b) CR05M20, and (b) CR1M10, as well as their respective Coulombic efficiencies (CE) and areal discharge capacity (d). Measurements employed current density of 0.5 mA/cm2. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89 A.1 Melting point of pure PEG-SiO2 hairy nanoparticles (red curve) and free PEG chains of the same molecular weight Mn = 5 kDa (black curve). Tethering causes the conformation to shift from extended (Tm = 61°C), single-folded (Tm = 59.5°C), and double- folded (Tm = 53.4°C) to only the extended chain conformation (Tm = 61.3°C) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93 A.2 Thermogravimetric curve used to calculate the weight fraction of SiO2 and PEG of the HNPs. The weight fraction of SiO2 wS i is obtained from the remaining weight at 600°C, and the weight fraction of PEG wPEG is the weight loss between 100°C and 600°C. Weight fractions for sample above was found to be wS i = 17 wt.% and wPEG = 83 wt.%, with the calculation of core volume fraction ϕc and grafting density Σ given below. . . . . . . . . . . . . . . . . 93 A.3 Temperature-dependent conductivities of neat and hybrid sam- ples probed through dielectric relaxation spectroscopy (DRS). The conductivities presented are of poly(DOL) (1 mM Al(OTf)3, green open circles and 10 mM Al(OTf)3, blue open stars), hybrid system composed of HNPs (ϕc = 2.7 vol.%) in poly(DOL) with- out (red open diamonds) and with the addition of 0.5 M LiNO3 (blue open stars). All samples include 2 M LiTFSI salt. . . . . . . 95 xii A.4 DSC curves of self-suspended HNPs (ϕc = 15 vol.%), poly(DOL), poly(DOL) with 10 mM LiNO3 addition, and hybrid systems containing ϕc = 2.7 vol.% with and without LiNO3. All poly(DOL) samples were polymerized with 50 mM Al(OTf)3. All samples possess single recrystallization temperature Tc and broad glass transition temperature Tg, with a hybrid sample indi- cating a couple melting points Tm. The two Tm in the hybrid show contribution of the two components: HNPs and poly(DOL), but the single Tc show that upon recrystallization, the two compo- nents co-crystallized. Crystallinity was inferred from both Tc and the heat of melting ∆Hm, with lower value of Tc indicates less crystalline domains and lower value of ∆Hm indicates lower extent of crystallinity in those domains. All key transition tem- peratures are summarized in Table A.1. . . . . . . . . . . . . . . . 96 A.5 Time sweep measurement of DOL monomers polymerized with (a) 1 mM (b) 15 mM (c) 20 mM (d) 30 mM (e) 40 mM and (f) 50 mM Al(OTf)3. All room temperature time sweep measurements were carried out at angular frequency of ω = 10 rad/s and strain of γ = 5%. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97 A.6 Time sweep measurement of DOL monomers polymerized with (a) 20 mM (b) 30 mM (c) 40 mM and (d) 50 mM Al(OTf)3 in the presence of HNPs (ϕc = 2.7 vol.%). The insets show clearer peaks in G” and G’ at early times. All room temperature time sweep measurements were carried out at angular frequency of ω = 10 rad/s and strain of γ = 5%. . . . . . . . . . . . . . . . . . . . . . . 98 A.7 Time sweep measurement of DOL monomers polymerized with (a) 20 mM (b) 30 mM (c) 40 mM and (d) 50 mM Al(OTf)3 in the presence of 10 mM LiNO3. Dashed lines show data points from Figure A.4 as a comparison between poly(DOL) polymer- ization with and without LiNO3. All room temperature time sweep measurements were carried out at angular frequency of ω = 10 rad/s and strain of γ = 5%. The low viscosity upon the addition of LiNO3 for high initiator contents created difficulty in observing the end of induction time and the start of the ini- tiation process. Hence, even though the samples were loaded as soon as there were apparent viscosity changes, the resulting time-dependent data seems to show initiation process began ear- lier than what was inferred. . . . . . . . . . . . . . . . . . . . . . . 99 A.8 Time sweep measurement of DOL monomers polymerized with (a) 20 mM (b) 30 mM (c) 40 mM and (d) 50 mM Al(OTf)3 in the presence of HNPs (ϕc = 2.7 vol.%) and 10 mM LiNO3. All room temperature time sweep measurements were carried out at an- gular frequency of ω = 10 rad/s and strain of γ = 5%. . . . . . . . 100 xiii A.9 (a) Time taken to start obtaining plateau in G” and G’ during polymerization (tp) (b) equilibration time post-crystallization (tc) for PolyDOL samples with Al(OTf)3 content varying from 10 to 50 mM. (c) Time taken reach the first peak in G’ (tpk) and (d) to obtain equilibrium moduli G′′eq and G′eq (teq) in DOL polymeriza- tion in the presence of HNPs (ϕc = 2.7 vol.%). . . . . . . . . . . . . 101 A.10 Equilibrium moduliG′′eq and G′eq for DOL polymerization in the presence of HNPs (ϕc = 2.7 vol.%) (Figures 2c and A.5) normal- ized by the initial moduli values G′′0 and G′0 over Al(OTf)3 con- tent of 1 – 50 mM. The resulting ratios G′′eq/G ′′ 0 and G′eq/G ′ 0 show an increasing value with the initiator content after 10 mM initia- tor content. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102 A.11 Strain-dependent measurement of self-suspended HNPs (ϕc = 18 vol.%) and hybrids (ϕc = 2.7 vol.%) polymerized with Al(OTf)3 content varying from 1 mM to 50 mM. Strain sweeps of hybrids were measured at 70°C with angular frequency ω = 10 rad/s while result for self-suspended HNPs was obtained at 70°C with ω = 0.25 rad/s. The value of loss modulus G” is normalized to loss modulus at zero strain G′′γ→0 to emphasize peak in G” at high strain in self-suspended HNPs, which indicates soft glassy behavior. This peak is minimal in the hybrids, illustrating the loss of soft glassy behavior and the closer-to-polymer nature of the hybrids. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103 A.12 Effects of LiNO3 addition towards the electrochemical perfor- mance of PEG-SiO2 HNPs/PolyDOL electrolyte in Li//Cu cells. (a) Reactivity was observed by comparing chronoamperometry results with potentials descending from 2.0 to 0.2 V of poly(DOL) hybrids (ϕc = 2.7%) without and (b) with the addition of 0.5 M LiNO3. (c) Interfacial stability was observed via electrochemical impedance spectroscopy (EIS) post-chronoamperometry at dif- ferent potentials for the same hybrid electrolyte without and (d) with 0.5 M LiNO3. All electrolytes include 2 M LiTFSI salt and all poly(DOL) electrolytes were polymerized with 1 mM Al(OTf)3. 104 A.13 (a) Galvanostatic cycling profile for Li//LFP cell with hybrid electrolyte containing ϕc = 2.7%, 2 M LiTFSI, and 0.5 M LiNO3, and (b) without PEG-SiO2 HNPs. Both electrolytes were cycled at a rate of 0.2C. The discharge capacity and Coulombic effi- ciency (CE) over cycle for both electrolytes are shown in (c). It is shown that despite larger capacity fade in hybrid electrolyte before stabilizing, hybrid electrolyte possesses higher discharge capacity throughout the cycle range presented as well as a more stable overpotential value. . . . . . . . . . . . . . . . . . . . . . . 105 xiv B.1 (a) Hybrid electrolyte formed outside of electrochemical cell in- dicates solid-like behavior with (b) stratified layers of suspen- sion and polymeric layer. SEM images of polymerized separator at (c) 4 µm and (d) 20 µm scale. Suspension-laden separator is shown in (f) 4 µm and (g) 40 µm scale. (e) Postmortem image of Lithium metal for Li//NCM811 cell configuration with added Li2O and (h) Copper substrate for Cu//NCM811 cell configura- tion with Li2O included after more than 50 cycles. . . . . . . . . . 107 B.2 (a) Particle size distribution measured at minute 1 to 15 with 1- minute interval and (b) the average size at each minute. . . . . . 107 B.3 Differential scanning calorimetry (DSC) curve of poly(DOL) layer near and away from Li2O particle settlement at 10 vol.% particle concentration and its corresponding (b) thermogravi- metric analysis (TGA) curve along with result for 30 vol.% Li2O. 108 B.4 Temperature-dependent ionic conductivity of QSSE with vari- ous Li2O contents in (a) DOL electrolyte with 2 M LiTFSI and (b) ethylene carbonate (EC) with 2 M LiTFSI and 0.5 M LiNO3. Ionic conductivity at 30◦C and activation energy value Ea of (b) DOL electrolyte and (d) EC electrolyte. . . . . . . . . . . . . . . . . . . 109 B.5 (a) Maxwell dependence of normalized conductivity on Li2O volume fractions at temperatures of 10 – 60◦C and (b) the col- lapsed universal plot with data points showing temperature- averaged normalized conductivity values. . . . . . . . . . . . . . 110 B.6 (a) Strain-dependent loss and storage moduli G” and G’ and (b) stress-strain curves of Li2O suspensions of 10, 30, and 50 vol.%. caption . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110 B.7 Morphology of Li-metal deposited on Cu-foil for 10 mAh with electrolyte of (a, b) DOL with 2 M LiTFSI, (c,d) QSSE with 10 vol.% Li2O, (e,f) 30 vol.% Li2O, and (g,h) 50 vol.% Li2O. Scale bars of (a, c, e, g) show 100 m and (b, d, f, h) 10 µm. . . . . . . . . 111 B.8 Electrochemical impedance spectroscopy (EIS) Nyquist plot of suspension electrolytes at volume fractions of 0 – 50 vol.%. . . . . 112 B.9 Discharge capacity and CE of control cases in anode-free Cu//NCM811 configuration of (a) DOL + 2M LiTFSI. (b) Un- polymerized DOL + Li2O (open symbols) and polymerized poly(DOL) without Li2O (closed symbols) can only cycle at C/10 and are close to failure during SEI formation step. . . . . . . . . 113 B.10 Corresponding galvanostatic stripping and plating for elec- trolyte utilized in Figure 3.4a but including Li-metal anode in a Li//NCM811 configuration instead of anode-free. (a) SEI forma- tion step at C/10 with the first and fifth cycle shown for the hy- brid electrolyte. (b) Post-SEI formation step, battery was cycled at C/2 for 100 cycles, with the corresponding discharge capacity and Coulombic efficiency (CE) at all cycle shown in (c). . . . . . . 114 xv B.11 (a) Cyclic voltammogram of control liquid electrolyte in carbon cloth and Cu-foil configuration, without any Li2O particle addi- tion like seen in Figure 3.4e. (b) Correlation of peak current to the square root of scan rate, extracted from Figure 3.4e, is seen to have a linear relationship. (c) At the highest scan rate of 5.0 mV/s, reduction and oxidation peak areas are seen to have the same value. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114 B.12 (a) X-ray diffraction (XRD) analysis of pristine Li2O, pristine Li2O2, and Li2O on the Cu substrate after the third discharge of a Cu//Li2O cell. (b) Raman spectra of Li2O after cycling for more than 100 cycles, compared to spectra for pristine Li2O and Li2O2. The presence of Li2O2 is indentified by red stars above the peaks. 115 B.13 Electrochemical performance for anode-free 10N5 in Cu//NCM811 configuration at different current rates of C/2, 1C (= 1 mA/cm2), 2C, and C/2. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116 B.14 Galvanostatic stripping and plating of hybrid electrolyte con- taining poly(DOL) and (a) 10N7, (c) 50N5, (e) 50N7, and (b, d, f) their corresponding discharge capacity and Coulombic efficiency. 117 B.15 Predicted capacity fade in 50N7 hybrid electrolyte, calculated from the experimental CE at different cycle numbers. This value is compared to the actual experimental value of capacity at each cycle. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 118 B.16 The effects of LiNO3 concentration on interfacial and bulk prop- erties of hybrid electrolytes. (a) Nyquist plot obtained from EIS, (b) transference number measured through the Bruce-Vincent method compared to pure poly(DOL) (blue star) and pure QSSE (green triangle), and (c) ionic conductivity at 30◦C. The role of LiNO3 as an electrolyte additive in LMBs and LIBs has been ex- tensively studied and the main finding of these studies support its role in building a stable SEI. We studied the effect of LiNO3 concentration on the hybrid electrolytes and observed higher in- terfacial resistance indicated by amplified circumference and an additional semicircle in the Nyquist plot (Figure B.16). This sec- ond semicircle grows with increasing LiNO3 content. Increasing LiNO3 concentration from 0.5M (N5) to 0.7M (N7) does not seem to affect transference number and conductivity significantly. . . . 119 B.17 (a) Coulombic efficiency (CE) of suspension electrolytes made up of various SiO2 particles in DOL + 0.5 M LiNO3 + 1 M LiFSI. (b) CE presented as a function of SiO2 volume fraction. . . . . . . 120 xvi C.1 Ex-situ fabricated (a) CR1M10 and (b) CR05M10 bent to show its flexibility. The resulting material of CR1M10 is brittle with a high Tg, as indicated by its (c) DSC curve, compared to the flexible CR05M10 with no Tg detected down to -80°C. (d) DSC curve comparing poly(DOL) made with 100 mM Al(OTf)3 with and without 10% TMPTGE. The inset shows the crystalline poly(DOL) without 10% TMPTGE and amorphous one with TMPTGE. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122 C.2 Temperature-dependent rheology measurement of CR05M at different TMPTGE concentrations of 5, 10, and 20%. . . . . . . . . 123 C.3 1HNMR of CR05M samples with 2%, 4%, and 6% TMPTGE. Monomer concentration was calculated (see Equation C.1) based on this result to obtain values shown in Figure 4.2. The unlabeled peaks belong to EC/DMC. . . . . . . . . . . . . . . . . . . . . . . 124 C.4 1HNMR of methyl ether poly(ethylene glycol) (mPEG) and DOL with ratios of 30:70 mPEG:DOL and 50:50 mPEG:DOL, indicat- ing the ratio of integrals match the added concentrations. . . . . 125 C.5 Room-temperature ionic conductivity of UN05M and CR05M10 compared to 1.0 M LiPF6 samples. . . . . . . . . . . . . . . . . . . 126 C.6 (a) Frequency-dependent rheology measurement of CR05M at TMPTGE concentrations of 5, 10, 20, 30, and 40%. (b) Storage modulus G′ = Ge and mesh size of the network that can be cal- culated using the equation ξ = ( G′NA RT )− 1 3 . . . . . . . . . . . . . . . . 127 C.7 The first five cycles of galvanostatic stripping and plating of CR05M10 in a pouch cell with graphite/NCM811 configuration. The area of both electrodes is 9 cm2. Measurements employed current density of 0.5 mA/cm2. . . . . . . . . . . . . . . . . . . . . 128 C.8 Galvanostatic stripping and plating for 50 cycles of graphite/NCM811 of (a) CR05M5, (b) CR1M5, (d) CR05M10, (e) CR05M20, and (f) CR1M10. (c) Comparison of CE and discharge capacity for CR05M5 and CR1M5. Measurements employed current density of 1 mA/cm2. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129 C.9 Linear sweep voltammetry (LSV) of UN05M and CR05M of var- ious TMPTGE concentrations. . . . . . . . . . . . . . . . . . . . . 130 C.10 Stripping and plating lithium onto copper in a Li/Cu configura- tion for (a) CR05M5 (b) CR05M20 and (c) their CE values at each cycle. CE was determined through the ratio of charging time and 60-minute of discharging. . . . . . . . . . . . . . . . . . . . . . . 130 C.11 The first five cycles of Galvanostatic charging and discharging for polyDOL initiated by 1 mM Al(OTf)3, doped by 2 M LiTFSI, and crosslinked by 5% TMPTGE. . . . . . . . . . . . . . . . . . . . 131 xvii C.12 Room-temperature ionic conductivity of polyDOL with various TMPTGE concentrations. The pink points represent the poly- DOL polymerized with 1 mM Al(OTf)3 with 2 M LiTFSI and green point with 10 mM Al(OTf)3 with 2 M LiTFSI. . . . . . . . . 132 xviii CHAPTER 1 INTRODUCTION 1.1 Solid Polymer Electrolytes (SPEs) Electrochemical batteries have achieved great triumphs in recent decades, with commercially available technology developed after the first cell invented by Count Volta in the 1800s [37]. Commercially available batteries include lead- acid, nickel-cadmium, nickel-metal hydride, and lithium-ion batteries (LIBs). LIBs, in particular, have revolutionized especially the transportation and com- munications industries, such that the Royal Swedish Academy of Sciences awarded the 2019 Nobel Prize in Chemistry to Goodenough, Wittingham, and Yoshino for the development of LIBs. LIBs are typically composed of a graphite anode, polyolefin separators wetted by a liquid electrolyte, and an intercala- tion type cathode such as lithium cobalt oxide (LiCoO2), lithium manganese oxide (LiMnO4), lithium iron phosphate (LFP), and lithium nickel manganese cobalt oxide (LiNixMnyCo1−x−yO2 or NMC). During the discharging of LIBs, Li- ions transport through the electrolyte from the anode to the cathode where they are intercalated. When the battery is charged, the Li-ions are released from the cathode and travel back to the anode where they intercalate into the graphitic carbon host. McKinsey Battery Insights team project that the battery market size of 2.6 TWh and yearly growth of 25% from 2023 to 2030, when it would reach a value of more than USD 400 billion. This rate of growth is pushed by regulatory shifts toward sustainability (such as New York’s 2050 net-zero goal, Europe’s “Fit for 55” program, EU’s ban of internal combustion engines (ICEs), etc.), increasing 1 customer demand in lieu of highly polluting technologies, and original equip- ment manufacturers (OEMs) targets for emission reduction. Despite the immense importance of LIBs worldwide, the practical energy density of currently available commercial batteries have trailed the rate of mar- ket growth. Over the past 150 years, there has only been a 6-fold increase in energy density from the first-generation lead-acid batteries (about 40 Wh/kg) to the present LIBs (about 240 Wh/kg) [37, 83]. This translates to 7 – 8% en- ergy density growth per year, compared to the projected 25% market growth per year expected over the remaining years of this decade. Current LIBs are also nearing the theoretical value of energy density that can be provided by the cathode/anode materials, with the highest energy density achievable in cells in which the graphite anode is paired with NCM-family cathodes with high Nickel content. Substituting the graphite anode with Lithium (Li) metal increases the anode energy density by a factor of nearly 10-fold (the theoretical capacity of Li is 3860 mAh/g, in comparison to 360 mAh/g for graphite), and Li offers the most negative reduction potential (-3.04 V vs the standard hyrogen electrode). These effects are amplified further in Lithium metal batteries (LMBs) in which the intercalation cathode is replaced with a conversion material (e.g., Li-sulfur (Li-S), Li-oxygen (Li-O2)),. For instance, Li-O2 batteries present an energy den- sity of 3505 Wh/kg while it is 2600 Wh/kg for Li-S batteries [37]. A number of fundamental technical problems currently limit wide-spread, commercial use of LMBs [93]. As LMBs are being charged and discharged, lithium metal gets plated and stripped from the anode/cathode. Throughout the plating and stripping of lithium metal, mossy structures loosely termed den- drites form at the surface of Li-metal anode and may grow to connect the anode 2 and cathode. This causes short-circuiting of the batteries. Short-circuiting leads to thermal runaway and a risk of fire and explosion hazard, especially when the reactive Li-metal anode is paired with the commercial flammable electrolytes in- cluding linear carbonates of ethyl methyl carbonate (EMC), dimethyl carbonate (DMC), and diethyl carbonate (DC) with flash point around room temperature (between 16 and 33°C) [85]. The importance of safety in LMB operation re- quires new electrolyte chemistries that are low in flammability, reactivity, and toxicity. Solid-state electrolytes (SSEs) have been proposed as a requirement for safe LMB’s. In their simplest form, SSEs would be designed with high enough me- chanical modulus to prevent dendrite growth normal to the current collector. SSEs are additionally designed to reduce toxicity and flammability, compared to liquid electrolytes present other perhaps obvious advantages. Solid-state ma- terial as an electrolyte was first introduced by Faraday in 1849, where PbF2 was found to be a good F- conductor [64]. In 1973, Fenton, Parker, and Wright estab- lished the first solid polymer electrolyte (SPE) made up of complexes of alkali metal with poly(ethylene oxide) (PEO) [65]. The complexation of PEO and alkali metal ions was later reported to occur through the ether and carbonyl oxygen in polymer backbone, where multiple oxygen sites coordinate with, for instance, lithium ion (Figure 1.1) [176, 181]. This complexation is akin to that seen in the coordination between crown ethers and metals [114, 29]. The segmental motion of the polymers above their glass transition temperature allows for free volume, assisting with the complexation of lithium and the polar ether oxygens for ionic transport. Free volume created by polymers’ segmental motion is essential for ionic 3 Figure 1.1: Lithium coordination with ether oxygen in PEO backbone transport. For semicrystalline polymers such as PEO, ionic conduction hap- pens mostly in amorphous domains [23]. Free volume dictates ionic conduction at the amorphous domains, dynamically generating suitable coordination sites that provide a “bond” or temporary bridges for ions to hop [58]. The importance of segmental motion in ion transport is often reflected by fitting temperature- dependent ionic conductivity σ(T ) to the Vogel-Fulcher-Tammann (VFT) equa- tion (Equation 1.1), such seen in PEO SPE doped with LiClO4 [170]. σ = A exp ( EA R(T − TVF) ) (1.1) Here, A is prefactor, EA is activation energy associated with segmental motion, R is gas constant, T is temperature, and TVF is Vogel-Fulcher temperature (glass transition temperature Tg − 50). Due to its high crystallinity, however, PEO and many other polyethers are known to manifest low conductivity at room temper- ature and often only achieve practical conductivity values required for normal battery operations (i.e., σ ≥ 1 mS/cm) at elevated temperatures as high as 100°C 4 [96]. Much effort has been made in introducing more amorphous domain in SPE in order to raise their room-temperature conductivity values. Amorphous regime can be obtained for instance by introducing plasticizers such as LiTFSI salt [174], ethylene carbonate [191, 31, 139, 39], propylene carbonate [31, 39], dibutyl phthalate, and dioctyl adipate [172]. Crosslinking, branching, and in- troduction of networked segments are also known to have plasticizing effects [48, 193, 77]. Despite the optimized design of SPE, all SSEs are plagued by the issue of poor interfacial contact between the electrolyte and electrodes. While liquid electrolytes can wet electrodes well, SPEs and electrodes are in rigid contact with very high interfacial impedance [36]. It was shown that PEO-LiClO4 SPE in contact with Li-metal anode possesses interfacial resistance of 1000Ω/cm2 even at 70°C [115]. This high interfacial resistance and poor contact between elec- trodes and SPE affects the cycle life of polymer LMBs. The high interfacial impedance affects the charge transfer process at the electrode-electrolyte interface and poor contact creates high energy barrier for the ion transfer between solid electrolyte and electrode particles [36, 190]. Due to this, it was reported that the ionic conductivity of Li-SPE interface is more correlated with the cycle life LMB compared to bulk conductivity or modulus [150]. A similar conclusion has been reached for graphite-SPE cells composed of poly(ethylene oxide-co-2-(2-methoxyethoxy)ethyl glycidyl ether). The poor interfacial contact between the polymer electrolyte and graphite renders charge transfer process a rate-limiting step, which affects the battery performance [57]. The volume expansion of the electrodes during charging/discharging are also 5 found to worsen the electrode/SPE interfacial contact, eventually leading to bat- tery failure[57]. Several methods have been employed to mitigate the interfacial issues in SSEs. The high interfacial impedance was seen to be alleviated by the addition of nanofillers that could create a good interface on the electrode surface. In the case of PEO-LiX SPE, this has been demonstrated by adding a small amount of ferroelectric BaTiO3 nanoparticles [115]. Other composite electrolytes will be discussed more thoroughly later in this dissertation. Another method that has been utilized in ceramic electrolytes is by melting the electrolyte to ensure that electrode particles are wetted by electrolyte melt and a good interfacial con- tact can be established. While this method has been successful for inorganics, SPEs melt viscosity is often high and does not favor penetration and wetting [2]. There have also been manufacturing methods reported to help with SPE interfacial issues, such as spraying, sputtering, and painting the electrolyte onto electrodes [166]. 1.2 Ring-opening Polymerization (ROP) A recent method in mitigating the poor interface between electrodes and SPE is in-situ polymerization of liquid monomers to form SPEs inside battery cells. In one approach, polymerizable liquid precursors such as 1,3-dioxolane and Lewis acid initiator Al(OTf)3 are first introduce to the battery cell to wet all components. Ring-opening polymerization (ROP) converts the liquid DOL to poly(1,3-dioxolane) (poly(DOL)) in timescales of minutes, allowing for an in- built interface to form that facilitates interfacial ion transport in a LMB [219]. 6 Ring-opening polymerization (ROP) is a polymerization process that starts with cyclic or heterocyclic monomers. It can proceed through anionic, cationic, or radical polymerization [144]. ROP is an important industrial process, producing important polymers such as Nylon-6 from the monomer of ϵ-caprolactam [59] and oligomers in combination with ethylene oxide from the monomer of methyl oxirane [210, 28]. Both polymerizations are examples of anionic ROP (AROP), which can be described as the nucleophilic attack of the growing chain end on a heterocyclic monomer molecule on an electron-rich active species. Cationic ROP (CROP), on the other hand, proceeds on electron-deficient active species (cations or a species with partial positive charge), with propagation also involv- ing attack of the nucleophile on electrophilic active centers [132]. CROP, utilized in the in-situ polymerization of poly(DOL), proceeds through three elementary reaction steps: initiation, propagation, (chain transfer), and termination. During initiation, active species are generated through interaction between initiator and monomer molecules. In the polymerization of hetero- cyclics, specifically, species that are formed in the initiation step may differ in structure and reactivity from the propagating species. This is because initia- tion might involve a sequence of at least two reactions, and the second reac- tion might be slow enough that the initiation becomes a rate-determining step [132]. Initiation can proceed with the help of initiators such as: (i) protonic acids (e.g.: triflic acids, fluorosulfonic acids, and perchloric acids), (ii) stable organic salts (e.g.: carbenium salts, oxocarbenium salts, carboxonium salts, and oxo- nium salts), (iii) covalent compounds (e.g.: alkyl bromide, boron or aluminum tristriflates), (iv) Lewis acids (e.g.: SbF5, PF5, BF3, etc.), and (v) others (e.g.: silicon-containing initiators, photoinitiation, and radiation-induced initiation) [132]. 7 Propagation in CROP first involves a growing chain with cationic center at the chain end, which adds to another monomer molecule via a SN1 or SN2 mechanism [30]. Another proposed mechanism for CROP propagation is that after a monomer is activated and carries the cationic center. The growth re- action for this encapsulates an electrophilic attack of the activated monomer on the chain end. For some systems, it was shown that these different mecha- nisms can happen simultaneously [22]. Perhaps the most interesting aspect of CROP propagation is its thermodynamics, which is different than other com- mon polymerization methods and could change quite dramatically with in- creasing monomeric ring size. The conversion of monomer into polymer is possible when the change of free energy ∆G is negative, where it depends on the changes of enthalpy ∆H and entropy ∆S , as well as temperature T : ∆G = ∆H − T∆S (1.2) Both ∆H and ∆S for common polymerization processes are typically nega- tive. However, ROP provides a unique phenomenon where polymerization is entropy-driven for some cyclic monomers. The driving force of polymerization can be conceived as a passage from higher to lower energy form, and in the case of cyclic monomers, this depends on the release of ring strain. Ring strain is dictated by the linkage of the monomers that enforces distor- tion of the preferred bond angles and lengths. If the ring strain energy of cyclic monomer is high, the energy of the ring being closed sits at a higher level than when the ring is opened, prompting the ring-opening of the cyclic monomer. For monomers with high ring strain energies, ∆H is always a relatively large negative value, which will always exceed the −T∆S value (below decompo- 8 sition temperature). For moderately or weakly strained rings, −T∆S value can become equal to ∆H within available temperature range. In this case, ∆G returns as zero and the propagation process of CROP becomes reversible [132]. This re- versibility often results in poor conversion and a significant amount of residual monomeric content, often undesirable in polymerization processes. This is ev- ident in the presence of approximately 20% residual monomeric content in the in-situ polymerized poly(DOL) with the addition of 0.5 Al(OTf)3 [219]. Ring size -CH2- -O- -S- -NH- -O-CH2-O- 3 115 117 77.8 96.1 4 109 79.0 5 25.5 28.0 4.1 30.5 6 0.4 9.2 12.1 7 25.5 8 40.5 9 52.5 Table 1.1: Strain energies (kJ/mol) of cyclic compounds The ring strain energies in kJ/mol is presented in Table 1 [49]. As presented in the table, larger rings including the 5-membered ring of 1,3-dioxolane tend to have lower ring strain energies, presenting more opportunities for reversible propagation. Although 7- and 8-membered lactones and lactams polymeriza- tion is driven by an enthalpic contribution, the ring strain of these monomers is approximately only 6 J/mol [27]. A 3-membered ring oxirane, on the other hand, has a very high strain energy of 117 kJ/mol, inducing an irreversible poly- merization reaction [132]. Although mainly governed by ring strain, enthalpy of polymerization can also be lowered considerably when release of ring strain is compensated by nonbonded interactions appearing in polymer unit. This is usually the case with cyclic monomers possessing substituents, where the enthalpy of polymerization in cyclopentanes shifts from -21.9 kJ/mol with H- 9 substituents at R1 and R2 to -13.4 kJ/mol with CH3- substituents at the same sites [69]. As seen in Table 1, 5-membered 1,3-dioxolane has one of the lowest ring strain energies of 30.5 kJ/mol. Because of the low ring strain, the current state- of-the-art in-situ polymerization of 1,3-dioxolane (DOL) is reversible, leaving a significant amount of residual monomeric content in the electrochemical cell. Not only does the residual DOL have a low oxidative stability [219], but it also lowers the mechanical strength of SPE as well as hinders the application of the SPE in Lithium-Sulfur (Li-S) batteries. The unpolymerized fraction of DOL dissolves intermediate cathode reaction products such as lithium polysulfides (Li2S x, 2 < x ≤ 7), resulting in a rapid loss in cathode capacity [184]. In fact, residual monomers are often undesirable in many areas, including dental fillings [81, 164, 15] and food packaging [54]. Many studies have been dedicated to lowering the residual methylmethacrylate (MMA) monomer con- centration in poly(methyl methacrylate) (PMMA) dental fillings [164, 15]. There have also been regulatory activities over the years associated with potential mi- gration of residual monomers from a number of commodity food packages into food items (FDA Docket No. 76N-0070, September 23, 1977). Moreover, it was reported that residual monomers hinder fundamental studies of polyamides (PA) degradation, as they obscure the measure of degraded portion of the im- portant biodegradable polymers. Instead of the interaction with the PA, the increase carbon evolution, oxygen demand, or biomass might be caused by mi- crobial metabolism of the residual monomers or oligomers [105]. Despite the unsought residual monomers, they seem to be plaguing many polymeric systems. Methods that are typically employed in order to reduce 10 the residual monomeric fraction include: (1) employing higher reaction tem- peratures, (2) combining different initiators, (3) increasing solubility of initia- tor in monomer, (4) employing “scavenger” monomers that are highly reac- tive with target monomers, (5) devolatilization (removal of monomer through vacuum and heat), and (6) utilizing ion exchange resin to capture monomers and oligomers [14]. Most industrial processes opt for post-polymerization de- volatilization involving vacuum, heat, and even ionizing radiation [54] to re- move residual monomers, which can be energy intensive. Two methods are to be detailed in this dissertation in order to tackle the re- versibility of 1,3-dioxolane inside electrochemical cell. To address the lowered mechanical strength of the reversible SPE, composite electrolytes are first uti- lized; optimizing not only mechanical properties, but also materials’ microstruc- ture and architecture that improves ionic conductivity. The later portion of this dissertation focuses on enabling a more reliable ROP through crosslinking of poly(DOL), which promotes an effect similar to surface-anchoring of free rad- icals in emulsion polymerization process that is known to produce polymers with high conversions [14]. 1.3 Mechanical Reinforcements Through Composite Elec- trolytes Most of this dissertation encapsulates works of composite electrolytes cre- ated from grafted and non-grafted nano- and micro-particles. Composite elec- trolytes have been used in many electrochemical systems for various reasons such as improving transference number [161], electrochemical stability window 11 [135, 41, 42, 12], interfacial resistance [202, 16], ionic conductivity [51], and for particles’ plasticizing effects on otherwise crystalline polymers [71]. The fillers added into SPE can be broadly classified into passive and active fillers. While passive fillers do not interact with ionic species inside the electrolyte, active fillers can enhance ionic conductivity even further by interacting with ionic species in some way. Some examples of passive fillers are polymer-grafted SiO2 and ZrO2 nanoparticles [161, 41, 135], γ-LiAlO2 [12], TiO2 [111], Al2O3 [196], Li1.3Al0.3Ti1.7(PO4)3 (a combination of Al2O3, TiO2, and Li2CO3) [127], and un- tethered ZrO2 [51, 111]. Despite not actively participating in conducting ionic species, these fillers offer interesting benefits within their respective SPE sys- tems. For instance, TiO2 and ZrO2 in PEO/LiBF4 complex depress the melting temperature of the polymeric complex and retard the kinetics of PEO crystalliza- tion, resulting in conductivity relaxation below the melting temperature of PEO [111]. The addition of α-Al2O3 into PEO and poly(N,N-dimethylacrylamide) (PNNDMAA), on the other hand, arouses the acid-base interaction involving polyether oxygens, filler acid/base centers, and alkali metal cations. The fillers change the fraction of available oxygen sites, resulting in the formation of ionic aggregates, enhancing ionic conductivity as filler molecules are in the vicinity of Li+ coordination sphere [196]. This has also been reported for many other ox- ide particles, where transport happens through complexation of salts with the particles’ surface through acid-base interaction. In many cases, oxide particles provide charged layer for ion transport [25]. Active fillers can encompass a broad range of particles, including NASI- CON [155], LISICON [10], garnet-type [94], perovskite-type [89], and ionically 12 conductive sulfide glass additives [97]. All of these active fillers serve as an enhancement of ionic conductivity as they conduct ions well. However, they also serve as crosslinking centers for polymer chain segments [155], which in- hibits crystallization as also seen in passive filler of TiO2 and ZrO2 [111]. This “crosslinking” destabilizes the coordination around cations, easing migration of Li+ ions from site to site in the vicinity of the fillers [155]. Like what is observed in addition of α-Al2O3 into PEO/PNNDMAA [196], the addition of active NA- SICON fillers also enhances Lewis acid-base interactions between the ceramic surface and anions, this interaction competes with anion-cation interaction, pro- moting better salt dissociation via ‘ion-ceramic complex’ formation [155]. Conductivity in composite materials have been studied since the 1800s, when J.C. Maxwell evaluated thermal conductivity of heterogeneous material (particles dispersed in a suspending medium) [133]. According to his findings, it is possible for particles to behave as insulators and not participate in conduct- ing heat (and in our case, ions). Maxwell correlated particle conductivity σp and medium conductivity σ0 through a coefficient α: α = σ0 − σp 2σ0 + σp (1.3) If particle conductivity is negligible, α = 1/2. He further developed the corre- lation between overall conductivity of composite material σ and σ0 employing the coefficient α and core volume fraction ϕ: σ σ0 = 1 − 2αϕ 1 + αϕ (1.4) If particles are insulating, and α = 1/2, then: 13 Figure 1.2: Maxwell model fit for suspension electrolytes σ σ0 = 2(1 − ϕ) 2 + ϕ (1.5) The slope of the σ normalized to σ0 against the bulk particle volume fraction term results in a slope of 2. This correlation has been previously observed in nanoporous composite electrolyte employing oligomer-suspended SiO2-PEG nanoparticles [160]. It is an overall helpful correlation in composite electrolytes to determine whether or not particles are participating in conducting ions. Out of the plethora of composite materials that are utilized as electrolytes, the first part of this dissertation discusses particularly the addition of hairy nanoparticles into an in-situ polymerized solid electrolyte. Hairy nanoparti- cles (HNPs) are a class of material made by grafting polymeric chains onto nanoparticle cores. While composite polymeric materials have found their place in many applications, ranging from biomedical applications [87] to soft actua- tors [32], their applicability tends to be limited by irreversible agglomeration of nanoparticles in liquid or polymeric matrices as well as lack of macrostructural stability. A method commonly employed to get past this issue is by creating 14 hairy nanoparticles to ensure stable and uniform dispersion of nanoparticles in their host matrices. Previous experimental [66, 40, 125] and theoretical [211, 45] works suggest that uniform distribution could be achieved in self-suspended HNPs. As particles are grafted with polymeric chains, the core particles impose a strong space-filling constraint, which exerts constraints on the polymer chain conformation and in turn induces long-range inter-core correlations mediated by the polymeric chains [211, 45, 44]. This results in the suppression of long- wavelength density fluctuations and causes cores to uniformly fill space. Specif- ically, upon inspecting the structure factor S (q) deduced from small-angle X- ray scattering (SAXS), self-suspended grafted nanoparticles predicts a S q→0 = 0 [211, 45], which is an expected value for a single-component incompressible molecular fluid and not for a suspension of hard spheres. Various chemistries of particles and grafts have been explored in different fields. For instance, metal nanoparticles such as platinum, gold, and rhodium show liquid-like behavior when grafted with ionic liquids [192]. Other parti- cles such as silica [40, 125, 101, 4], maghemite [70], and molybdenum disul- phide [78] have also been explored with grafts such as poly(ethylene oxide) (PEO), poly(caprolactone), and poly(ethylene glycol) (PEG). Nanoparticles that are grafted can also be made out of polymeric materials such as polystyrene (PS) grafted by another polymer of poly(methyl acrylate) (PMA) [103]. Other examples include magnetic nanoparticles iron oxide grafted by polyvinyl-based polymers [227], poly(dimethyl siloxane) (PDMS) [80], PS [189], and polysilox- ane copolymers [108]. There is also a library of polymer matrices that these HNPs can be dispersed in, creating a very vast combination of nanoparticles, 15 grafting polymers, and host polymers. Some examples of polymeric hosts that have previously been employed include PEG [130], PMMA [130], PEO [92], and methyl ether PEG [126]. 1.4 Forward-Favoring ROP Although composite electrolytes made up of bare and hairy particles can improve mechanical properties and conductivity, there is no evidence that they promote a more reliable ROP. Hence, uncontrolled amounts of residual monomers still co-exist with polymer at equilibrium. The latter part of this dis- sertation discusses a method to alleviate this issue using cross-linking. Learning from the emulsion polymerization that yields very high conversions of PS and PMMA [14], the surface-anchoring of free radicals can be utilized as a tool in reducing residual monomeric content. Copolymerization of styrene and acry- lonitrile grafted onto polybutadiene seeds is seen to result in high conversions, and increasing conversion is seen with increasing grafting density [1]. Grafting is a way to reduce monomer and propagating polymer’s entropy, restricting the polymer from depolymerizing. As we have previously learned in the discussion of ROP, in the polymeriza- tion of 5-membered and larger rings, enthalpy change is not the dominant fac- tor in determining polymerization direction; entropic factors can become large enough to dominate [132]. Entropic contribution in monomers can come from the translation, rotation, and torsion of monomeric molecules. In many cases, the positive rotational and torsional entropies can be much larger than transla- tional entropy, inducing polymerization of large macrocycles instead of linear 16 polymers [185, 136]. Reducing the free volume of polymer chains by increas- ing grafting density is known to arrest chain dynamics and in turn reduce en- tropic contribution coming from rotation, torsion, and translation [145]. Simi- larly, crosslinking polymer chains have also been reported in higher conversion, characterized by increasing moduli with increasing crosslinking density [165]. This is the method that will be employed in reducing the residual monomeric content in ROP. Entropy loss due to binding has been observed in many systems, especially in biological systems. Loss of translational and rotational entropy was seen in small molecules binding onto proteins, losing a rigid-body entropy of 15- 20 kJ/mol [140]. The association reaction of two molecules to form a single complex must overcome a large entropic barrier because of the loss of trans- lation and rotational degree of freedom. Dimerization of insulin (2In to In2), for instance, loses entropy equal to losing the entropy of one monomer [180]. Moreover, this binding can be used in order to prevent depolymerization. Sta- bilization of microtubules and prevention of its depolymerization was done by the binding of guanosine-5’-triphosphate (GTP) to beta-tubulin [137]. The GTP cap stabilizes the end of microtubule, allowing new alpha-beta heterodimers to be added. Similarly, actin can be used in binding immunoglobin (Ig), demon- strating stabilization of actin filaments and preventing depolymerization [13]. This is similar to the stabilization of monomers onto a grafting site in order to monomers to be available for propagation. 17 1.5 Outline of Dissertation This dissertation focuses on strengthening in-situ polymerized SPE by mechan- ical reinforcement as well as controlling the residual monomer content. Chap- ter 2 starts with the role of nano-sized PEG-SiO2 dispersion in in-situ poly- merized poly(DOL). The resulting increase in mechanical strength and room- temperature ionic conductivity is discussed in context of co-crystallization of PEG tethers with poly(DOL). The altered kinetics of polymerization in the hy- brid SPE is evaluated through time-dependent rheology measurement, and the hybrid SPE’s functionality is assessed in a system of Li-S battery. Chapter 3 outlines the utilization of Li2O microparticles not only as a me- chanical reinforcement for in-situ polymerized poly(DOL), but also as electro- active materials. Multi-functional hybrid SPE with gradient property resulting from gravitational settling and basicity of Li2O is fabricated in-situ, with the resulting hybrid SPE possessing good mechanical properties and ionic conduc- tivity. The electro-active Li2O particles enable the application of hybrid SPE in an anode-free configuration through the redox reaction between Li2O and Li2O2. Finally, chapter 4 implements the crosslinking of poly(DOL) as a method of controlling the amount of residual DOL monomer content. A correlation be- tween crosslinking density with residual monomer content is developed and further discussed. The resulting crosslinked poly(DOL) not only possesses min- imal residual monomer content, but is also functional in a Lithium metal and ion batteries paired with NCM811 cathode. 18 CHAPTER 2 STRUCTURE AND EVOLUTION OF QUASI-SOLID-STATE HYBRID ELECTROLYTES FORMED INSIDE ELECTROCHEMICAL CELLS 2.1 Introduction Solid-state rechargeable batteries have in recent years drawn significant atten- tion from researchers, as well as from investors —globally. An important aspect of their promise is that such batteries remove fundamental safety and perfor- mance barriers to high-energy, low-cost storage of electrical energy in cells that employ Li metal as anode [175, 82, 131]. Limited choices of electrolyte mate- rials able to simultaneously meet the mechanical and electrochemical require- ments of such cells slowed early progress towards cost-competitive, practical solid-state batteries by at least three decades [131]. This progress has notice- ably quickened over the last decade as a number of solid-state electrolyte op- tions, including inorganic solids (e.g. ceramics), organic polymers, and organic- inorganic hybrids, have emerged [131]. Solid state electrolytes (SSEs) based on inorganic glasses are conventionally thought to be attractive because they pro- vide enhanced safety due to their non-volatile nature, slower chemical reactions with the metallic Li electrode, as well as for their fundamental ability to me- chanically retard non-planar/dendritic growth of the Li anode during repeated cycles of charge and discharge [61, 116]. While significant progress in synthesis and fabrication of versions of these materials with high room-temperature ionic conductivity and good mechanical properties is a source for optimism about the future of solid-state batteries, poor/heterogeneous wetting of the many explicit and implicit interfaces at and in the battery electrodes, and recent observations 19 that Li dendrites may selectively proliferate into microdefect networks in some solid ceramic electrolytes to short-circuit battery cells [73, 95] continue to pose fundamental barriers to progress. An emerging practice is to fabricate the electrodes of solid-state battery cells with in-built ionic conducting pathways (i.e., fabricate the cell using solid-state catholyte and anolyte, in addition to the solid-state electrolyte), to enable uni- form ion transport. This strategy while satisfying a perhaps obvious require- ment for maintaining good access to the electrochemically active materials in a solid-state battery electrode introduce as obvious shortcomings, mostly because it lowers the volumetric and specific capacity of the electrodes and, by adding new fabrication steps, may also increase battery cost. Solid polymer electrolytes (SPEs) can be tailored to respond to imposed stress and thermal fields to over- come the wettability challenges faced by SSEs. They may also be formulated to achieve low volatility, high chemical and electrochemical stability in contact with Li metal, while at the same time maintaining good-enough mechanical properties to suppress non-planar/dendritic growth of the Li metal electrode [116, 215, 7].Unfortunately, the most successful SPE (poly(ethylene oxide), PEO) is a semi-crystalline polymer, which crystallizes at temperatures in the range of 60 – 65°C. SPEs based on PEO polymer are therefore only able to achieve prac- tically relevant ionic conductivity and meaningful ability to respond to stress or thermal fields to infiltrate the pores of the battery electrodes at temperatures somewhat above those targeted in the majority of applications [50]. As a first step towards a solution, we recently reported a process whereby liquid electrolyte precursors based on 1,3-dioxolane (DOL) may be transformed to produce a solid-state ion-conducting material inside an electrochemical cell. 20 The transformation is produced by initiating ring-opening polymerization of the liquid DOL using a Lewis acid salt (AlCl3, Al(OTf)3, etc.) dissolved in the electrolyte [219, 225, 123, 146]. The approach has been shown to be ef- fective in promoting reversible stripping and plating of Li by simultaneously limiting chemical and electrochemical side reactions with electrolyte compo- nents and promoting formation of stable solid electrolyte interfaces with Li. The solid-state electrolytes thus formed sustain stable cycling of Li//LFP full cells, provided suitable steps were taken to slow oxidative degradation of the poly(DOL) at the cathode and to limit corrosion of the cathode current collec- tor by the Lewis acid initiator [219]. A drawback nonetheless is that the ring- opening polymerization reaction is reversible, meaning that a broad distribu- tion of macromolecular species is present in the battery cell at any given time [219, 225, 123, 146], which while beneficial for achieving high room-temperature ionic conductivity, makes it impossible to create sufficiently high molecular weight polymers to achieve electrolytes with the mechanical characteristics re- quired for a solid-state battery. Herein, we report on the synthesis, structure, thermal properties, and elec- trochemical characteristics of a family of hybrid SPEs formed inside an elec- trochemical cell. Created by Al(OTf)3-initiated ring-opening polymerization of 1,3-dioxolane (DOL)/PEG-SiO2 nanoparticle dispersions containing LiTFSI- LiNO3 salt mixtures, the hybrid SPEs preserve the favorable features of the first-generation, in-situ formed poly(DOL) SPEs but also offer enhanced ion- transport, thermal and mechanical properties, which we attribute to the PEG- SiO2 nanoparticle component. Significantly, the beneficial effects of the PEG- SiO2 particles are already apparent at SiO2 volume fractions as low as 3%, where the particles result in faster development of poly(DOL) mechanical mod- 21 ulus and a more than doubling of the room-temperature ionic conductivity of poly(DOL). Additionally, the PEG-SiO2 nanoparticles produce dramatically lower activation energy barriers for ion transport — relative to electrolytes com- posed either of the pure poly(DOL) or pure PEG-SiO2 materials. Small-angle X-ray scattering, thermal analysis, mechanical characterizations were used to understand the structural, thermal, and mechanical factors responsible for these improvements. By means of electrochemical analysis in Li//Cu half cells and Li//SPAN (Sulfur/polyacrylonitrile) full cells, we report further, that the hy- brid SPEs facilitate stable electrochemical cycling of batteries that employ Li metal anode. Polymer-ceramic hybrid materials provide a well-researched route for achieving free-standing SPEs with high mechanical strength and improved room temperature ionic conductivity (σRT ) [116, 160, 214, 47]. While the first of these features is intuitive, the second is not —it is believed to originate from suppression of crystallization of the polymer component, which promotes wetting and ion transport. The notoriously poor colloidal stability of small inorganic particles in polymers render most hybrid SPE designs impractical [160, 26, 24, 109, 110]. SiO2 nanoparticles tethered with short (< 10 kDa molecu- lar weight) PEG chains have been observed to form stable dispersions in poly- mers due to specific interactions between the dispersing medium and the teth- ered chains, enhanced particle curvature, and the fact that the space-filling con- straint between tethers provides a strong thermodynamic incentive for polymer molecules not attached to the nanoparticles to fill the space between the cores [168, 3, 102, 129]. Nanoparticles composed of high dielectric constant SiO2 nanocores (ϵ = 3.6; 22 dSiO2 ≈ (10 ± 2) nm) densely grafted with 5 kg/mol PEG chains were selected as candidates for the in-situ formed hybrid SPEs of interest in the present study. This choice was motivated by a number of considerations. The most important is that prior to ring-opening polymerization of DOL, it is essential to preserve simple, liquid-like flow properties in the DOL/PEG-SiO2 hybrid liquid elec- trolytes to facilitate complete wetting of all interfaces inside a battery cell. A col- loidally stable, unform dispersion of nanoparticles in a liquid host is a require- ment for achieving this goal. Additionally, the PEG chains may serve at least three beneficial purposes. First, a densely grafted layer of PEG chains on each SiO2 nanoparticle will enhance particle dispersion both by mechanisms outlined above and by the more obvious fact that the tethered chains will provide a steric barrier to prevent formation of particle aggregates. Second, by co-crystallizing with the host SPE molecules, the PEG chains will introduce disorder in crys- talline domains of the host, reducing crystallite size and potentially lowering the SPE’s melting temperature; we will show later that this is consistent with what is observed in our experiments. Third, the tethered PEG oligomers are themselves capable of conducting Li ions. It means that the hybrid SPEs will present both bulk and interfacial pathways for ions to move in the electrolyte, which could provide synergistic enhancement in ionic conductivity. We note further that the SiO2 cores impart other beneficial attributes. SiO2 particles are known, for example, for their ability as additives in polymeric systems as they improve mechanical strength [9, 162, 107], electrical and dielectric properties [183, 106, 197], and flame resistance [31,32]. As a final consideration, we point to studies on PEG-SiO2 hairy nanoparticles (HNPs) dispersed in poly(methyl methacrylate) (PMMA) which showed that the negative Flory-Huggins param- eter, χ< 0, between PEG and PMMA leads to very large enhancements in col- 23 loidal stability relative to what is achieved from the curvature and space filling effects discussed earlier [129]. Although the χ parameter for PEG/poly(DOL) mixtures has not been reported, results from thermal and small-angle scattering measurements in a chemically similar system, poly(ethylene glycol) dimethyl ether (mPEGm)/PEG-SiO2 [125, 126], reveal propensity for co-crystallization between the two chains [147]. 2.2 Materials and Method 2.2.1 Material Preparation Synthesis of PEG-tethered SiO2 nanoparticles Silica nanoparticles (LUDOX SM30, 10 ± 2 nm), poly(ethylene glycol) monomethyl ether (mPEG-OH) with Mn = 5 kDa, 3-(triethoxysilyl)propyl iso- cyanate, 1,4-diazabicyclo [2.2.2.]-octane (DABCO), and anhydrous dichloromethane were purchased from Sigma-Aldrich. PEG-tethered silica nanoparticles were synthesized according to a previously reported method.[34,35] mPEG-OH (10.0 g, 2.00 mmol), 3-(triethoxysilyl)propyl isocyanate (0.495 g, 2.00 mmol), and DABCO (0.336 g, 3.00 mmol) were dissolved in anhydrous dichloromethane (10.0 mL) and reacted at 50°C for 48 hours. The resulting mixture was pre- cipitated into excess hexanes. mPEG-silane was isolated by decantation before drying in vacuo at room temperature. The resultant dry, pure mPEG-silane was stored at 2 – 8°C prior to usage. The silica nanoparticles (1.82 mL) were diluted in excess deionized water 24 (400 mL) and a solution of mPEG-silane (2.00 g) in deionized water (50.0 mL) was added dropwise. The mixture was reacted at 70°C for 48 hours. The reacted mixture was dried partially in a convection oven at 45°C for 24 hours before fur- ther purification by repeated centrifugation in a 1:4 (v/v) chloroform/hexanes at 8500 rpm. Self-suspended hairy nanoparticles were dried further in vacuo at room temperature before storage under argon atmosphere. Differential scan- ning calorimetry (DSC) shows that self-suspended HNPs forced PEG chains to only adopt extended conformation due to space-filling constraint [102] (Figure A.1). The grafting density and core volume fraction can be varied by adjusting the ratio of SiO2 nanoparticles to the mPEG-silane. Thermogravimetric analysis (TGA) was employed to determine the weight fraction of the undegraded inor- ganic content, which can be translated to the core volume fraction of SiO2 and grafting density (Figure A.2 and Equations A.1,A.2). Dispersion of HNPs in poly(DOL) 1,3-dioxolane (DOL) and Aluminum trifluoromethanesulfonate (Al(OTf)3, 99.9% metals basis) were purchased from Sigma-Aldrich, and Lithium nitrate (LiNO3, 99.9% metals basis) was purchased from Chem-Impex Int’l, Inc. Due to aggregation of self-suspended HNPs and hydrophilicity of DOL, room- temperature ultrasonification of HNPs in DOL was not an accessible method. HNPs were dispersed in deionized water (15.0 mL) and dispersion was freeze- dried for 72 hours to ensure minimal water content. The resulting HNPs were no longer aggregated and can be easily dispersed in PolyDOL precursors through mechanical shaking. The unaggregated, flake-like HNPs were dried in vacuo at room temperature for at least 24 hours before stored under argon 25 atmosphere. HNPs (0.10 g) and Al(OTf)3 initiator were added to DOL (1.00 mL) and DOL was let polymerized in presence of HNPs. Various initiator con- tent was employed for X-Ray scattering and rheology measurements, ranging from 1 mM (1.89 mg in 4.00 mL DOL) to 50 mM (23.7 mg in 1 mL DOL). For time-dependent measurement involving LiNO3, LiNO3 (0.69 mg, 10 mM) was similarly added to DOL (1.00 mL) precursors. Core volume fraction for 0.10 g HNPs in 1.00 mL poly(DOL) was evaluated through TGA in Figure A.2. Electrolyte preparation All electrolytes were prepared in argon glove box (Inert Inc.) with O2 and H2O content lower than 0.5 ppm. Lithium bis(trifluoromethanesulfonyl)imide (LiTFSI, puriss., ≥ 99.0%) was purchased from Sigma-Aldrich. All electrolytes containing HNPs were further dried chemically with Li-strips, reacted in elec- trolyte solution at 60°C for 1 hour. All electrochemical tests utilized coin 2032 cells with Lithium foil as the anode and Celgard 3501 as the separator. Half cells were assembled with copper foil as the cathode and full cells were assem- bled with sulfur/polyacrylonitrile (sPAN) and lithium iron phosphate (LFP). LFP cathodes were prepared by a procedure previously outlined [219] with LFP loading of 4 mg/cm2 and the preparation of sPAN cathodes was previously done [194] with Sulfur loading of 0.5 mg/cm2. sPAN cathodes were coated by dropping ≈ 20 µL diluted 1:10 Nafion:ethanol solution onto 3/5” sPAN and vac- uum drying at 60°C. 26 2.2.2 Characterization Methods Small-angle X-Ray scattering (SAXS) SAXS measurements were conducted using Anton Paar SAXSess bench-top X- ray scattering system. The system employed a line collimated beam at 0.1542 nm and a block camera setup for data collection. All samples were measured at 70°C, which is above the melting point of samples. The experimental line collimated intensity Iexp(q) is a function of point collimated intensity I0(q) [75, 169] such that, Iexp(q) = ∫ ∞ −∞ ∫ ∞ −∞ Wx(x) Wy(y) I0  √( λaq 2π − y )2 + x2  dx dy (2.1) Wx(x) and Wy(y) are the horizontal and vertical X-ray beam, x and y are the horizontal and vertical dimensions, q is the scattering wave vector, and a is the sample-to-detector distance. Generalized indirect Fourier transformation (GIFT) was utilized to desmear Iexp(q) in obtaining I0(q). Particle scattering intensity Iparticle is defined in terms of the scattering intensity of particles in suspending medium Iparticle/medium, scattering intensity of suspending medium Imedium, and volume fraction of suspending medium ϕmedium [117]. Iparticle = Iparticle/medium − ϕmedium Imedium (2.2) For spherical particles [41,40], scattering intensity of particles is Iparticle(q) = ϕc ∆ρ 2 e VP(q) S (q) (2.3) 27 where ϕc is the core volume fraction, V is the volume of a single particle, ∆ρe is the electron density contrast, P(q) is the form factor, and S (q) is the struc- ture factor. The interparticle correlations vanish at the dilute limit, causing the structure factor to be S (q)→ 1 and the form factor P(q) can be obtained directly from the scattering intensity. Bare charge-stabilized SiO2 nanoparticles (LUDOX SM30) was measured in a diluted aqueous suspension to obtain the P(q). S (q) was then obtained through GIFT method involving the Percus-Yevick closure relation that utilized a hard sphere structure factor model. Thermogravimetric analysis (TGA) The inorganic SiO2 content of self-suspended PEG-SiO2 hairy nanoparticles and their suspensions was estimated using thermogravimetric analysis (TGA) (TA Instruments Q500). TGA was performed under a nitrogen atmosphere from 20°C to 600°C at 10°C/min ramping rate. The self-suspended HNPs used for SAXS measurement were found to have 17 wt.% inorganic weight fraction, translating to 11 vol.% core volume fraction. The hybrid used in SAXS were found to have 4.2 wt.% inorganic weight fraction, translating to 2.7 vol.% core volume fraction. A sample calculation for the former is provided under Equa- tions A.1,A.2 along with the estimation of tether density. Mechanical rheology measurements Oscillatory shear measurements were performed using a strain-controlled ARES-LS rheometer (Rheometric Scientific) with a cone and plate geometry (10 mm, 4° cone angle) or otherwise stated. Time-dependent measurements were 28 carried out at room temperature with angular frequency of ω = 10 rad/s and strain of γ = 5%. Strain-dependent measurements were measured at 70°C with ω = 10 rad/s. Self-suspended HNPs were also measured at 70°C, above their melting point, with ω = 0.25 rad/s. Reaction components were mixed inside glove box and loaded on the rheometer only after induction time ends, which is signaled by an apparent change in viscosity. Dielectric relaxation spectroscopy (DRS) DC conductivity measurement was carried out using a Novocontrol broadband dielectric/impedance spectrometer. The same coin 2032 cells without electrodes were utilized, with Teflon ring containing the dried electrolytes. Conductiv- ity was measured at room temperature before temperature was increased from 30°C to 70°C, and measurement with frequency of 107 – 1 Hz was done every 10°C. The conductivity value was taken to be the value at the frequency at which the maximum of tan(δ) was seen. Arrhenius or Volger-Fulcher-Tammann (VFT) equations were used to fit the conductivity of the electrolytes, depending on the case as shown in results and discussions. Differential scanning calorimetry (DSC) DSC (TA Instruments Q2000) was adopted to evaluate thermal transitions of the self-suspended HNPs and hybrid samples. Thermal transitions were measured under nitrogen flow at a fixed ramp rate of 5°C/min. Hybrid materials were first heated to 70°C to remove thermal history before brought down to -100°C and heated back up to 70°C. Thermal history erasure of self-suspended HNPs 29 was done by first heating to 100°C, followed by data collection in the second cycle with cooling to -150°C and subsequent reheating to 100°C. Measurements during the second heating cycle were used to evaluate the melting temperature Tm, recrystallization temperature Tc, and heat of melting ∆Hm that translates to crystallinity. Electrochemical Measurements Galvanostatic lithium stripping/plating tests were operated using Neware CT- 3008 battery tester at room temperature, with current density of 1 mAh/cm2 for asymmetric Li//Cu cells. Galvanostatic discharge/charge tests for Li//sPAN cells were done under also done under room temperature with current density of 0.1 mAh/cm2. Li//LFP cells were charged and discharged at 0.1 mAh/cm2 for the first cycle and 0.2 mAh/cm2 for the following cycles. Chronoamperome- try profiles and linear sweep voltammetry were obtained using CH 600E electro- chemical workstation at room temperature. Potentials were held at 2.0, 1.8, 1.6, . . . , 0.2 V for 500 seconds in chronoamperometry measurement. Plateaus were observed by the end of the 500-s mark. Potentials were varied from 2.5 V to 5.5 V with 1 mV/s scan rate on Li//SS (stainless steel) cell in linear sweep voltam- metry measurement. Electrochemical impedance spectrometry (EIS) measure- ments were performed using a Solartron Frequency Response Analyzer (Model 1252) with frequencies ranging from 50 kHz to 10 mHz and at an amplitude of 10 mV. 30 Scanning Electron Microscopy (SEM) Scanning electron microscopy (SEM) was done on Zeiss Gemini 500 scanning electron microscope with energy-dispersive X-ray spectroscopy (EDS) installed. The surface of characterized Lithium was plated on copper foil for 10 mAh/cm2. UV-Vis Spectroscopy UV-Vis spectra were obtained through the use of Molecular Devices SpectraMax M2. sPAN and Nafion-coated sPAN were immersed in DOL electrolyte contain- ing 2 M LiTFSI and 0.5 M LiNO3 for 10 days before spectroscopy measurement. 2.3 Results and Discussions Figure 2.1 reports the ionic conductivity of SPEs composed of poly(DOL), PEG- SiO2, and poly(DOL)/ PEG-SiO2 hybrids as a function of temperature. In every case the electrolytes were doped with 2 M LiTFSI salt and conductivity mea- sured through dielectric relaxation spectroscopy (DRS) in coin 2032 cells. The main finding is that the room-temperature conductivity increases by 1.5 mS/cm (i.e., is more than doubled) upon introduction of PEG-SiO2 at a relatively moder- ate SiO2 core particle loading, ϕc = 2.7 vol.%. High room-temperature (RT) ionic conductivity σRT values have previously been reported for poly(DOL) [219], our observation is that it is possible to further enhance these values by adding PEG- SiO2 nanoparticles to the materials. We wondered whether the improvements in room temperature ionic con- 31 ductivity for the poly(DOL) SPEs are a consequence of changes (e.g., slow- down) in the kinetics of ring-opening polymerization of DOL in the presence of the PEG-SiO2 nanoparticles. We evaluated how the ionic conductivity for the hybrid SPEs evolve during the polymerization reaction for a fixed concentration (10 mM) of the Al(OTf)3 initiator and fixed value of ϕc = 2.7 vol.%. In parallel we also studied how the dynamic elastic/storage, G’, and loss, G”, moduli of the hybrid SPE’s evolve with time. The results reported in Figure 2.1b show that within 1 hour of initiation, both moduli increase by nearly four orders of mag- nitude and reach stable steady-state values, indicating that at 10 mM Al(OTf)3 the polymerization reaction progresses quite efficiently, at least in comparison to the pure poly(DOL) case reported later in Figure 2.3a. This behavior is like what is observed for σRT , which reaches a value of approximately 0.36 mS/cm after 1 hour and maintains conductivities close to that value (e.g., 0.32 mS/cm) after ten days. More remarkable, however, is how the temperature-dependent ionic con- ductivity σ(T ) evolves with time. Specifically, we observe a gradual transition from Volger-Fulcher-Tammann (VFT) to Arrhenius behavior as time increases. To understand this behavior, we note first that the VFT logσ = logσ0 − B kB(T−TVF ) and Arrhenius logσ = logσ0 − Ea kBT models for ionic transport in SPEs differ pri- marily in the assumed mechanisms by which ions move in materials. The differ- ences are reflected in the activation energy, Ea, and apparent activation energy, B, barriers for the transport. Here σ0 is a preexponential factor, T is the abso- lute temperature, kB the Boltzmann constant, TVF ≈ Tg − 50, is termed the Vogel- Fulcher temperature, Tg being the glass transition temperature of the SPE, which is determined to be 183 K for poly(DOL) with 10 mM Al(OTf)3. It is apparent that the two models approximately converge when the thermodynamic temper- 32 ature is recast as the temperature distance from the glass transition. Thus, while the VFT model assumes that ions in a SPE move by a combination of normal diffusion through the medium and by coupled glassy motions of polymer chain segments, the Arrhenius model imagines that diffusion alone governs ion trans- port. This can be understood in physical terms by noting that when a material is cooled, density increases and the free volume decreases. Consequently, segmen- tal motion is arrested, and ions can only move via diffusion [187, 72, 177]. The transition from VFT to Arrhenius within 24 hours of polymerization can then be understood to reflect gradual arrest of the segmental motions such that the coupled ion and chain dynamics characteristic of VFT changes to a dominantly ion-hopping/diffusion- based transport process. Although both Arrhenius and VFT models can fit the σ(T ) curves well at all times, the VFT model fits the data best for the first 5 hours of polymerization, and the Arrhenius expression is best employed thereafter. In the first 5 hours, Ea is 0.71 ± 0.01 kJ/mol and is analyzed best using VFT fit. There is an increase in B ≈ Ea post 5 hour of polymerization along with a shift from VFT to Arrhenius- like behavior, with a constant B of 0.38 ± 0.021 kJ/mol observed over a period of 10 days, post the fifth hour. Curiously, decreasing the initiator content by ten- fold causes the conductivity value to increase by ten-fold, as seen in the case of 1 mM Al(OTf)3 poly(DOL) with σRT = 3.75 mS/cm. As will be observed in the mechanical responses, lowering initiator content increases the molecular weight of the resulting poly(DOL), which in turn makes chains more flexible and less crystalline. Previous studies reported that the addition of 1 mM Al(OTf)3 re- sulted in poly(DOL) with Mn of 8.5 kg/mol, Mw of 15 kg/mol (Polydispersity, PDI = Mw/Mn ≈ 1.8), and DOL/poly(DOL) ratio of 14.% [219]. The remaining unreacted DOL fraction in the system has been reported to further enhance the 33 room-temperature ionic conductivity. Compared to poly(DOL) at the same 1 mM initiator concentration and self-suspended PEG-SiO2 (ϕc = 15 vol.%), the hybrid SPE with ϕc = 2.7 vol.% possesses the highest RT conductivity σRT = 3.75 mS/cm as well as the lowest activation energy Ea = 4.6 kJ/mol and maintain the highest σRT out of the three electrolyte systems within the experimental 50 K window. Addition of LiNO3 expectedly increases σRT to 4.55 mS/cm and de- creases Ea to 4.2 kJ/mol (Figure A.3) due to higher mobile DOL fraction, as re- flected in the lower moduli G” and G’ (Figure 2.3d). In contrast, self-suspended PEG-SiO2 possess extremely low σRT of 0.2 nS/cm and significantly higher Ea of 32 kJ/mol. It has been reported that self-suspended PEG-SiO2 behave as so-called soft glassy solids.[22,24,34,35,40] Two of the reported consequences include the change in PEG tethers to stretched conformation[24,34] as well as the caging phenomenon that is seen in strain-dependent mechanical property measurements (Figure A.11) [168, 102, 126, 6]. We used small-angle X-ray scattering (SAXS) to study the dispersion of PEG- SiO2 nanoparticles in the hybrid poly(DOL)/PEG-SiO2 SPEs. Figure 2.2 reports our findings for HSPEs at different Al(OTf)3 initiator concentrations. The re- sults in Figure 2.2a shows that the scattered intensity I(q) exhibits a plateau in the low q region and a scaling of q−4 in the high q region, for all initia- tor contents. Both characteristics are known features of scattering from dis- persions well-distributed and unaggregated spheres [125, 75]. Further insight on the dispersion state can be obtained by analyzing the structure factor S (q). Based on previous experimental and theoretical studies [212, 211, 41], the peak in S (q) at the lowest q = q1 value reflects the repulsive interactions between hairy nanoparticles, while the peak at the next lowest q = q2 reflects entropic attractions between PEG tethers. The interparticle distance is then defined as 34 Figure 2.1: Temperature-dependent conductivities of neat and hybrid samples probed through dielectric relaxation spectroscopy (DRS). (a) The conductivities of poly(DOL) (1 mM Al(OTf)3 and 10 mM Al(OTf)3), self-suspended PEG-SiO2 HNPs (ϕc = 15 vol.%), and hybrid system composed of both (ϕc = 2.7 vol.%). All samples include 2 M LiTFSI salt. (b) Temperature- dependent DC conductivity of hybrid system (ϕc = 2.7 vol.%) polymerized with 10 mM Al(OTf)3 for the 1st and 5th hour, 24th hour, 29th hour, 48th hour, and 10th day of polymerization. 35 d(p−p) = 2π q1 while distance between tethers can be estimated as 2π q2 . Figure 2.2c indicates that both distances increase with increasing initiator content. This trend is typically observed with the dilution of homogeneous par- ticle suspension, with larger interparticle distances caused by less correlated cores.[35] The entropic attraction between tethers naturally decreases with di- lution as it weakens the space-filling constraint of the tethers [126]. The in- creasing initiator concentration possesses similar effect as dilution, with both 2π q1 and 2π q2 increases with Al(OTf)3 content. Radical polymerization including the ring-opening polymerization of DOL monomers is known to have the result- ing polymer chain numbers depending significantly on initiator content. With higher initiator content, more radical sites are introduced during initiation pro- cess and during polymerization, higher termination rate at lower chain length is typically achieved [149, 86]. This creates a greater number of shorter chains distributed throughout the system. These chains can behave similarly to sol- vent molecules surrounding the HNPs, thus increasing Al(OTf)3 content has a similar effect to dilution. The polymerization process and kinetics of this hybrid system is further discussed through time sweep rheology measurement. The in- terparticle distance of self-suspended HNPs with ϕc = 11 vol.% is also presented in Figure 2.2c. It indicates that at the solvent-less, self-suspended form, the in- terparticle distance is d(p−p) = 19 nm. This value is in a good agreement with estimated value from random close-packing model (d(p−p) = 18 nm, Equations A.3,A.4). The Structure factor in the limit of zero wave vector S (0) is related to the isothermal compressibility of the material, which in the case of the hybrid SPEs reflects the ease with which long-range fluctuations in particle density can grow 36 Figure 2.2: Small angle X-ray scattering (SAXS) profiles to determine structure of PEG-SiO2 HNPs/poly(DOL) hybrid material. (a) Intensity profile and (b) structure factor of HNPs/poly(DOL) with ϕc = 2.7 vol.% with initiator Al(OTf)3 content varying from 10 to 50 mM. (d) Structure factor analysis gives interparticle distance dp−p (red data points) and distance between tethers (green data points) as described in the diagram (c), as well as structure factor at q = 0. Dashed lines in (d) represent result for self-suspended HNPs with ϕc = 11 vol.%. [168, 43]. The results reported in Figure 2.2d indicate that density fluctua- tions initially decrease with increasing Al(OTf)3 concentration, but thereafter are largely insensitive to the initiator concentration. It was previously observed that the largest effect on S (0) comes from two main sources: (i) poor dispersion of hairy nanoparticles in a host material [49]; and (ii) polydispersity in grafting density [212] of the hairy nanoparticles, i.e., particles are composed of different 37 number of tethers and this difference has a small but observable effect on the long-range density fluctuations. As the PEG-SiO2 particle chemistry and con- centration of DOL remains essentially unchanged with the changing initiator content, our results indicate that the SiO2 nanocore dispersion in the poly(DOL) host remains largely unchanged. This result is consistent with the gross behav- iors observed in I(q) as well as S (q) and confirm the anticipated benefits of con- structing hybrid SPEs using polymerizable DOL and PEG-SiO2 hairy nanopar- ticles as the structural building blocks. Considering the beneficial effects of the PEG-SiO2 nanoparticles on room temperature ionic conductivity of the hybrid SPEs, an open question is whether nanoparticles achieve this effect by interfering in some way with the ring- opening polymerization of DOL. We investigated the polymerization reaction kinetics of 1,3-dioxolane (DOL) at room temperature by measuring the time- dependent evolution of the mechanical loss, G”, and storage, G’, moduli (Figure 2.3a). Similar so-called “time sweep” mechanical measurements have been used to observe processes such as polymer degradation, crosslinking, and phase sep- aration [188, 134, 208] that lead to time-dependent changes in polymer molecu- lar weight and/or structure. A similar approach was also reported in our pre- vious work for interrogating time-dependent growth of poly(DOL) when DOL is exposed to Al(OTf)3 [219]. As DOL monomers propagate to create longer poly(DOL) chains, both the elasticity, as measured by G’, and the fluidity, as measured by G”, of the hybrid SPEs are expected to change as a function of time. The results reported in Figure 2.3 indicate that both G” and G’ rise rapidly at first and plateau after approximately 20 minutes, with the G” plateau typically larger than the plateau in G’. This observation indicates that the fully polymer- ized SPEs are likely composed of low-molar-mass, unentangled polymer chains. 38 It is also considered favorable for an electrolyte because viscous rather than elas- tic properties of the materials dominate, creating a structure with relatively low resistance to ion motion and deformation [158, 171, 67]. The inset to Figure 2.3a reports the effect of shear strain, γ, on G” and G’ after the plateau is observed, at the 30-minute time point during the polymerization reaction. The results show that irrespective of the imposed strain, the hybrid SPE response is dominated by viscous stresses that become less strain dependent as shear strain rises. This leads to so-called shear-thinning behavior, which is a commonly seen character- istic of amorphous, linear polymers [158]. It is known, however, that ether-based polymers like poly(DOL) are semi- crystalline materials at room temperature [50]. Crystallization creates discrete domains in the materials in which many individual polymer chains are local- ized. This yields entanglement-like effects and elasticity in a macroscopic mate- rial. Evidence of both behaviors are seen in long-time moduli measurements re- ported in Figures 3a and 3b where a cross-over from viscous to elastic dominant behavior is observed, which is followed by another plateau regime where G’ > G”. In this latter regime moduli values reach 105 Pa and are comparable to those expected in melts of very high molecular weight polyethers [195]. The second- stage moduli growth, and enhanced elasticity coincides with observations of a recrystallization peak in differential scanning calorimetry (DSC) measurements (Figure A.4), consistent with its connection to crystallization of the poly(DOL). As higher initiator content is often correlated to faster polymerization kinet- ics and shorter polymer chains [86], the characteristic times for appearance of the first plateau would be expected to decrease with increasing Al(OTf)3 con- centration, which is precisely what is observed (Figure A.9). We define two 39 characteristic times for the process. The first corresponds to the time (tp) at which the first plateau is formed, and is considered a feature of the amorphous poly(DOL) polymer; the second time-scale (tc) at which the second, elasticity- dominant, plateau is observed is attributable to formation of semicrystalline domains in the polymer. Indeed, while shorter polymer chains tend to crystal- lize more readily, they are known to possess lower mechanical strength (Figure A.5) [67]. The results reported in Figure 2.3c indicate that hairy PEG-SiO2 HNPs have profound effects on the evolution of both G” and G’ during polymerization of DOL. Specifically, as seen in the inset to Figure 2.3c, peaks in G” and G’ are observed at early times before either modulus reaches the first plateau. The number of peaks is reduced and the spacing between peaks lowered with in- creasing Al(OTf)3 concentration (Figure A.6), such that at 50 mM Al(OTf)3 only a single modulus peak is observed. It is conjectured that the strong interaction between PEG tethers and poly(DOL) is responsible for both observations. In the first place (i.e., before the formed poly(DOL) crystallizes), such interactions retard poly(DOL) chain growth by extending the observational window for the reversible polymerization – depolymerization reaction. This effect should dis- appear as the initiation rate for the ring-opening reaction becomes higher at higher Al(OTf)3 concentration, which is consistent with what we observe. The time at which the first G’ peak (tpk) is observed is reported in Figure A.9. The results show that tpk manifests an appreciable change with initiator content, which is again in line with our hypothesis that interactions of the PEG tethers and the poly(DOL) hinders polymer chain growth in the amorphous regime. Previous studies of radical polymerization in the presence of inorganic 40 Figure 2.3: Time-dependent dynamic shear flow measurements are used to study changes in polymerization kinetics induced by PEG- SiO2 HNPs in a poly(DOL) SPE. (a) Time sweep measurement of DOL monomers polymerized with 10 mM Al(OTf)3 with the addition of: (b) 10 mM LiNO3, (c) PEG-SiO2 HNPs (ϕc = 2.7 vol.%). (d) DOL polymerized with 1 mM Al(OTf)3 with PEG- SiO2 HNPs (ϕc = 2.7 vol.%). The inset in (d) shows hybrid 1 mM initiator system with the addition of 2 M LiTFSI. The inset in (a) represents a strain sweep at 30-minute time-point dur- ing polymerization. The schematics of polymerization process going from monomers to amorphous polymer to polymer crys- tals were displayed in (a). Dashed line in (b) shows data point from (a) as a comparison between poly(DOL) polymerization with and without LiNO3. The inset in (c) shows clearer peaks in G” and G’ at early time. All room temperature time sweep measurements were carried out at angular frequency of ω = 10 rad/s and strain of γ = 5%. Strain sweep was measured also at RT and ω = 10 rad/s. Al(OTf)3 contents of 1, 20 – 50 mM for each type of sample are shown in Figures A.5 and A.6. 41 fillers [159, 120, 153], show that there is normally a decrease of polymeriza- tion rate in the presence of filler particles, attributed to interactions between fillers and monomers/polymers. Results reported in previous study [159] re- veal, further, that fractions of macroradicals are immobilized in the interface re- gion. The reduced compatibility between filler particles and growing macrorad- icals produces polymer chains with higher average molecular weight. This find- ing is consistent with our observation of increasing G” and G’ with increasing Al(OTf)3 concentration as opposed to the opposite trend for poly(DOL) with- out PEG-SiO2 nanoparticles. Other studies[120, 153] made similar observations, noting that polymers adsorption on ceramic filler particles can drive sponta- neous de-wetting of the particles by the suspending medium, producing phase separation near the filler surface. Normally an increase in initiator concentration would lead to enhanced polymerization rate along with decreasing mechanical strength [149, 86, 90, 76]. For the hybrid SPEs, however, we find more complex behaviors. In Figure A.10, equilibrium moduli G′′eq and G′eq, normalized to initial moduli G′′0 and G′0, re- veal that G′′eq/G ′′ 0 and G′eq/G ′ 0 both increase with Al(OTf)3 concentration. There is also a more rapid increase in G’ compared to G”; the difference becomes larger as the Al(OTf)3 concentration increases. As the viscous portion of bulk ma- terial is reflected in the loss modulus, G”, we infer that there is a gradual ar- rest to the viscous dynamics in the system as initiator content is raised. This agrees with the observation of higher crystallinity in poly(DOL) synthesized at higher Al(OTf)3 concentration, as well as with the lack the entanglements of the relatively shorter polymer chains.[54,55] Increasing Al(OTf)3 concentration also causes a longer equilibrium time (teq) (Figure A.9). For the SPE with 1 mM added Al(OTf)3, however, the normalized moduli are higher in comparison to 42 the 10 mM case. It is conjectured that as polymerization proceeds, Mw in the orders of 15 kg/mol is reached [219]. This value is larger than the predicted en- tanglement molecular weight of poly(DOL) (Me ≈ 1.2 kg/mol [219]), resulting in an entangled poly(DOL) that expectedly has larger moduli than the shorter, unentangled poly(DOL) chains. PEG chains tethered to particles facilitate mixing with host polymers such as PMMA, which exhibit strong enthalpic interactions with PEG [130]. These interactions have been reported to lead to slower dynamics of the host ma- terial, increasing the average relaxation time of PEG-SiO2/PMMA [130] com- posites. They also alter the melting point and recrystallization temperature of PEG-SiO2/mPEG due to co-crystallization of mPEG hosts and the tethered PEG chains [126, 147]. Co-crystallization is observed in our PEG-SiO2/poly(DOL) hybrids as a single recrystallization peak in DSC measurements reported in Fig- ure A.4. The results show further that both the crystallization and melting peaks are shifted to lower temperatures than seen in either pure poly(DOL) or the self- suspended PEG-SiO2, indicating the inhibition of crystallization is mutual and strong. These findings confirm that the tethered chains interact strongly with the poly(DOL), providing a straightforward explanation for the uniform distri- bution of particles observed from the SAXS I(q) profile (Figure 2.2a). The role of Lithium nitrate (LiNO3) as an electrolyte additive for stabiliz- ing the interfaces formed between ether-based electrolytes, including DOL and dimethylether (DME), and metallic Li anodes is well-known [205, 213, 79], and extensively studied, particularly in the context of Lithium-Sulfur (Li-S) batteries [205, 213, 79]. LiNO3 is also thought to be advantageous in Li-S batteries because it inhibits lithium polysulfide (LiPS) shuttling [206, 217], which improves the 43 cell-level cycling efficiency in ether-based electrolytes [118, 201]. Although the source of these improvements is conventionally thought to be the unique chem- istry and physical properties of the interfacial materials phases (interphases) ether-LiNO3-LiPS mixtures form upon chemical and electrochemical reduction at a Li electrode, our recent study indicates that LiNO3 may also produce strong coupling between ether oxygens to produce polymer-like bulk material behav- ior in liquid ethers [219, 220]. Paired with LiFSI, infrared spectroscopy in fact shows that cyclic DOL molecules are highly strained in the presence of LiNO3 and appear less prone to undergo ring-opening polymerization. Figure 2.3b compares the time-dependent growth of G” and G’ in Al(OTf)3-initiated poly- merization of DOL in systems with (data points) and without (dashed line) LiNO3. We find that the greatest effects of LiNO3 is in increasing the induc- tion time before the moduli begin to grow. Once the growth starts, LiNO3 does not appear to have any significant effect on the polymerization reaction kinet- ics. The plateau G” and G’values are also largely insensitive to LiNO3, imply- ing that the poly(DOL) structure in both the amorphous and crystalline state are unaffected by LiNO3. Our results therefore imply that LiNO3 is perhaps better thought of as a retardant than an inhibitor for the ring-opening polymer- ization of DOL. Importantly, we show in Figures A.7 and A.8 that these findings hold true for all measured Al(OTf)3 concentrations. Retarders typically produce slightly larger rate of secondary radical generation, compared to primary radi- cal generation (kR2/kR1 < 10)), resulting in secondary radicals with chemical and electronic structure, polarity, and stereochemical properties different than those of the propagating radicals.[66] This also means that they have different reac- tivity, which may alter macrokinetics depending on the rate of reaction between the secondary radicals and monomer molecules, features that will require fur- 44 ther study. Figure 2.3d reports the time-dependent moduli during polymerization of DOL in electrolytes containing both LiNO3 and PEG-SiO2 hairy nanoparti- cles. At low Al(OTf)3 concentration (i.e., 1 mM), the depolymerization reaction dominates over ring-opening polymerization, and the moduli are low. As the Al(OTf)3 concentration is increased, the forward polymerization begins to dom- inate and an equilibrium elastic modulus, G’ of 103 Pa is achieved at 50 mM Al(OTf)3. We employ hybrid poly(DOL) synthesized with 1 mM Al(OTf)3 ini- tiator for all electrochemical studies to follow. This initiator concentration is twice that used by Zhao, et al. [219] and is deliberately chosen to achieve larger degrees of ring-opening polymerization. We will show later that LiNO3 plays a crucial role as a salt additive in such electrolytes, allowing Li to achieve highly reversible cycling when poly(DOL)/PEG-SiO2 hybrid SPEs are employed as electrolytes in battery cells. Results in Figure 2.4a indicate that the coulombic efficiency (CE) of hairy nanoparticles-containing hybrid electrolyte (orange circles) in Li//Cu electro- chemical cell is improved markedly to values as high as 99% upon addition of LiNO3 to these electrolytes. A similar effect is seen for liquid DOL electrolyte (black circles), where addition of LiNO3 also produces an increase of the CE to values as high as 97%. In the case of DOL, it was reported [220] that CE values as high as 99% could be achieved in dual-salt electrolytes containing both LiTFSI and LiNO3. This effect was also observed for other electrolytes such as ethy- lene carbonate (EC), dimethyl carbonate (DMC), and dimethoxyethane (DME) [219]. Polymerization from DOL to poly(DOL) reportedly also increases the CE.[11] The addition of LiNO3 has been reported to lower DOL reactivity over 45 a wide potential range, which is likely also partially responsible for the large im- provements in CE observed. Chronoamperometry measurements using HSPEs with ϕc = 2.7%, with and without the addition of 0.5 M LiNO3 (Figure A.12), support these conclusions. In these experiments we held the voltage fixed at progressively lower values, approaching the Li electrode reduction potential. The measured leakage current at the intermediate voltages reflect reduction of electroactive species in the electrolyte, including the PEG chains tethered to the HNPs our case [18]. The results show that the leakage currents from the in-situ formed poly(DOL) hybrid electrolytes containing LiNO3 are generally lower. Previous works highlight the role of LiNO3 additive in creating a stable layer that limit contact between electrolyte and Li-metal [41, 206, 118, 79]. This pas- sivation layer allows for a more homogenenous SEI [223, 201]. To evaluate the first effect, electrochemical impedance spectroscopy (EIS) was performed for both systems in Li//Cu electrochemical cell upon holding potential for 1000 s at 0.2 V (Figure A.12). It is seen that interfacial impedance of HNPs/PolyDOL with LiNO3 is lower compared to the hybrid system without LiNO3, indicated by the smaller semicircle in the imaginary impedance Im(Z) – real impedance Re(Z) plot. This Nyquist plot shows a single semicircle for both systems, and the bulk resistance can be fitted in series with a charge-transfer resistance that is in parallel with a double-layer capacitance [33]. This results in interfacial resis- tance of 107 Ω for hybrid without LiNO3 and a lower value of 76.5 Ω for hybrid electrolytes the contain LiNO3. When the impedance is evaluated after different potential steps in chronoamperometry (Figure A.12), hybrid electrolyte contain- ing LiNO3 is seen to possess high interfacial stability with a uniform value of impedance while the electrolyte without shows increasing value of interfacial impedance with more reduction at lower potential. There is an increase in the 46 Figure 2.4: Electrochemical performance of HNPs/poly(DOL) hybrid electrolyte. (a) Schematic of in-situ polymerization of DOL in the presence of HNPs. (b) Coulombic efficiency (CE) of various poly(DOL) electrolyte and hybrids containing PEG-SiO2 HNPs (ϕc = 2.7%) with and without 0.5 M LiNO3 at 0.1 mA/cm2 in Li//Cu cells (c) Galvanostatic cycling profile for Li//sPAN cell with hybrid electrolyte containing ϕc = 2.7%, 2 M LiTFSI, and 0.5 M LiNO3, with discharge capacity and Coulombic effi- ciency (CE) over cycle shown in (d). Rate performances for the Li//sPAN cells at different c-rates are shown in (e) with 1.0C is current density of 1 mA/cm2. 47 interfacial impedance of HSPEs compared to poly(DOL) SPEs in Li//Li electro- chemical cell due to the addition of PEG-SiO2 HNPs (Figure A.13). It is expected that HNPs add to the thickness and composition of the SEI, thus adding to the interfacial resistance. Results from scanning electron microscopy (SEM) paired with energy-dispersive X-ray spectrosopy (EDS) on lithium plated on copper foil at 10 mAh/cm2 indicates the presence of well-distributed elemental silicon without any large aggregates observed (Figure A.14). As an initial proof of concept and to illustrate the potential practical benefits of our HSPEs we fabricated Li-S batteries composed of Sulfur/polyacrylonitrile (sPAN) cathode (Figure 2.4c), metallic Li foil as anode, and the HSPE (ϕc = 2.7%) with and without 0.5 M LiNO3 as electrolyte. The results show that, but for the first cycle, the CE for the HSPEs is high and stable. After an initial period of decay, the discharge capacity (Figure 2.4c) also shows good stability at constant current density, as well as good responsiveness to changes in current density (Figure 2.4d) in the range 0.1 mA/cm2 (=0.1C) to 1 mA/cm2 (=1.0C). Notably, when the C-rate is reduced to its original value of 0.1C following higher rate cycling, a good capacity retention of 90% is seen compared to the steady value of the first few cycles. Previous Li//sPAN results for DOL/DME electrolyte re- sulted in more than 60% capacity loss within the first 50 cycles [194]. A decay in capacity was also observed in Li//sPAN for HSPEs due to DOL electrolyte reaction with sPAN, which produces lithium polysulfides that are observable under UV-Vis spectroscopy [194]. The amount of polysulfide products could be reduced by coating sPAN cathode with thin Nafion coating (Figure A.15). As a second example, we also investigated the electrolytes in Li//LFP cells (Figure A.16). It is shown that despite larger capacity fade in HSPEs before stabiliz- ing, hybrid electrolyte possesses higher discharge capacity throughout the cycle 48 range presented as well as a more stable overpotential value. Stable plating and stripping for symmetric Li//Li cell for > 250 hours and electrochemical stability up to 4.75 V in Li//SS cell were also observed (Figure A.17). 2.4 Conclusions It is reported that ring-opening polymerization of DOL containing PEG-SiO2 hairy nanoparticles can be used to synthesize hybrid solid-state poly(DOL) elec- trolytes inside a battery cell at various initiator contents. The PEG-SiO2 struc- tures hinder crystallization of poly(DOL), which leads to a substantial increase in room-temperature ionic conductivity σDC of the hybrid electrolytes, relative to the neat (particle-free) poly(DOL) SPE. The effect is synergistic in that the hy- brid electrolyte also manifests a dramatically higher σDC — from nS/cm-scale to mS/cm scale values, in comparison to SPEs composed entirely of the PEG-SiO2. The structure of dispersion was studied through small-angle X-ray scattering (SAXS), and it is observed that the PEG-SiO2 particles are well dispersed in their poly(DOL) host. Analysis of the structure factor S (q) deduced from SAXS in- dicates that both the distances between nanoparticles and PEG tethers increase with increasing initiator content. Increasing initiator content is thus seen to have an effect analogous to dilution of nanoparticles in a suspending host. SAXS also reveals a decrease in long-range density fluctuations, as reflected in the decreas- ing value of S (0) with increasing initiator content. Time-dependent mechanical shear analysis indicate that the SiO2-PEG particles alter polymerization kinetics and that this effect is enhanced by addition of LiNO3 salt additive. These mea- surements also reveal a competition between polymerization-depolymerization processes, which manifests as peaks in moduli G” and G’ at early times. Even- 49 tually, at high initiator content, the maxima coalesce, consistent with a poly- merization process in which the forward ring-opening reaction dominates the reverse reaction, resulting in higher molecular weight poly(DOL). Inclusion of LiNO3 retards ring-opening polymerization and manifests as a longer induction time, but otherwise has no effect on the polymerization process. The enhancement in room-temperature ionic conductivity relative to SPEs composed of self-suspended PEG-SiO2 is attributed to the shift from arrested soft-glassy dynamics to polymer-like behavior. Hybrid SPEs with ϕc = 2.7 vol.% are found to exhibit lower energy Ea = 4.6 kJ/mol, compared to poly(DOL) at the same initiator content and self-suspended PEG-SiO2 HNPs. Addition of LiNO3 increases σRT to 4.55 mS/cm and decreases Ea to 4.2 kJ/mol. The higher frac- tion of mobile ion carriers is thought to be the reason to this high σDC and low Ea. Evaluation of the in situ-formed PEG-SiO2 HNPs-poly(DOL) electrolytes in Li//Cu half cells, revealed high Coulombic efficiency (CE) and excellent Li re- versibility, particularly in electrolytes where LiNO3 is used as an additive. The resulting hybrid PolyDOL electrolyte was finally evaluated in Li//SPAN full cells and demonstrated to support enhanced battery cycling. 50 CHAPTER 3 ANODE-FREE LITHIUM BATTERIES ENABLED BY POLYMER-PARTICLE HYBRID ELECTROLYTES 3.1 Introduction The growing interest in Lithium metal as a replacement anode material for the commonly employed Lithium-infused graphite materials used in state-of-the art Lithium-ion batteries is reflected in the increasing number of studies and lit- erature reviews [121, 37, 122, 74, 119]. It is understood that a key driver of this interest is the nearly 10-fold increase in theoretical specific capacity of the an- ode achieved by replacing the graphite host with lithium metal [119]. The com- bination of the high chemical activity and low reduction potential of Li-metal introduces multiple technical barriers that have been thoroughly discussed in several recent reviews [122, 74, 119]. Solid-state electrolytes (SSEs) are generally thought to provide a straight- forward strategy towards lithium metal batteries that are safer and less prone to run-away thermal events associated with non-planar, mossy Li deposition during battery recharge [131, 221]. However, these benefits are typically accom- panied by sacrificed room-temperature ionic conductivity and poor electrode- electrolyte contact. Significant efforts have been concentrated on overcoming these challenges using solid electrolyte that emerge from wettable precursors [219, 184], and through regulation of ionic pathways throughout SSE’s crys- tal structure design [184, 207, 209]. Employing Li-metal anode also requires a solid-electrolyte interphase (SEI) capable of maintaining chemical and mechan- ical integrity at the high reduction potentials at which Li plates during battery 51 charging, and which protect the freshly deposited Li from continuous loss due to parasitic chemical reactions with electrolyte components [221, 173, 186]. Inorganic-organic hybrid electrolytes are of particular interest due to their ability to separate the mechanical, ion-transport, and interfacial functions of an electrolyte using discrete ingredients with properties optimized for these functions. A consequence is that such electrolytes are unique in their ability to provide combinations of high, liquid-like ionic conductivity and solid-like me- chanical strength —comparable to SSEs [21]. Typically, hybrid electrolytes are created by suspending electrochemically inert particles (typically metal oxides) in an ion-conducting liquid or plasticized polymer host [11, 63, 62]. Among the advantages of such materials are their ease of fabrication, straightforward compatibility with normal battery manufacturing methods, and versatility in the chemistry of the suspended particles and suspending electrolyte that can be used. Oxide particles suspended in a liquid electrolyte adsorb ionic species, cre- ating a space-charge layer on their surface, which has been thought to facilitate ion-pair dissociation and transport in the electrolyte [128]. Al2O3 particles in solid LiI, for instance, have been reported to adsorb Li+ on their surface, result- ing in an increased vacancy concentration in lithium sublattice that increases the overall conductivity. SiO2 introduced to PbF2 has likewise been argued to adsorb F-, creating fluoride vacancies. These advantages are countered by the generally high interfacial impedances that occur if the particles segregate to the electrode and by the potential for irreversible loss of Li ions that either bind too strongly or which chemically react with the oxide particles. The overwhelming majority of studies of suspension electrolytes focus on systems containing electrochemically inert, nanosized particles in the dilute 52 concentration regime [99, 52, 20, 100]. Few consider systems composed of larger microspheres capable of spontaneously settling under gravity. Recently Kim, et al. [99] reported that electrolytes created by addition of nm-sized Li2O particles in aprotic liquids containing fluorinated additives sustain reversible cycling of anode-free Cu//NCM811 batteries for at least 30 cycles. By means of Cryo- STEEM energy loss spectroscopy the authors conclusively showed that the SEI formed at the anode surface is enriched in Li2O. And, by means of Density Func- tional calculations, they further showed that the Li2O particles are surrounded by a solvation layer enriched with F- ions [99]. Recently we reported that liquid electrolytes based on 1,3-dioxolane (DOL) undergo an Al(OTf)3 Lewis acid-initiated ring-opening polymerization inside a battery cell to create solid polymer electrolytes (SPEs) with room tempera- ture ionic conductivities exceeding 1 mS/cm at Al(OTf)3 concentrations below 1 mM [219]. Suspension electrolytes created by dispersing poly(ethylene gly- col) (PEG)-grafted SiO2 nanoparticles in DOL were subsequently reported to undergo a similar ring-opening polymerization to create hybrid solid polymer electrolytes (HSPEs) with even higher room temperature ionic conductivity ≈ 4 mS/cm [184]. It was theorized that interactions between and co-crystallization of the tethered PEO oligomers and poly(DOL) provided more ether oxygen do- mains to complex with Li+, which facilitated ion transport. Here, we wish to expand on this concept by creating HSPEs in which electrochemically active particles, such as Li2O, are used to regulate both the physico-chemical proper- ties of the in-situ formed solid-state, poly(DOL) electrolytes in bulk and in the solid-electrolyte interphase. We show that such electrolytes enable extended cy- cling of anode free Cu//NCM811 with low interfacial resistances and effective values of the Columbic efficiencies computed by fitting the cycling data to a 53 power-law decay function markedly larger than observed in either the unpoly- merized or polymerized electrolyte in the absence of the Li2O particles. The electrolytes also manifest interesting gradient physical and transport properties (i.e., they are solid-like polymers far from the anode surface and are liquid-like near the anode). 3.2 Materials and Method 3.2.1 Electrolyte and battery preparations To fundamentally understand the source of the enhanced Coulombic Efficiency extended cycling of anode-free Lithium Battery Cells, we investigated thermal, transport, and electrochemical properties of Li2O/poly(DOL) and Li2O/DOL electrolytes, as well as their analogous Li2O-free electrolyte controls. Acknowl- edging the importance of cathode electrolyte interphase (CEI) for high-voltage batteries [8, 216] and anodic solid electrolyte interphase (SEI) for batteries with better cycling stability [221, 173, 186, 151], hybrid solid-state electrolytes formed by ring-opening polymerization of Li2O/DOL suspension electrolyte using a Lewis Acid initiator is of particular interest. To fabricate such electrolytes in Li//NCM811 and so-called anode free, Cu//NCM811, cells we first created Li2O/DOL suspensions (see Figure 3.1a). The Li2O particles (100 mesh, ravg ≤ 150µm) was purchased from BeanTown Chemical and used as obtained. Elec- trolytes suitable for electrochemical studies were created by adding a combina- tion of two salts (LiTFSI and LiNO3) known from previous studies [118] to work in tandem to enable formation of good interphases on Li; the studies reported 54 herein used electrolytes comprised either of 2 M LiTFSI + 0.5 M LiNO3 (N5) or 2 M LiTFSI + 0.7 M LiNO3 (N7). Different volume fractions of added Li2O were utilized, and the hybrid electrolytes with x vol.% are denoted as either xN5 or xN7 (10 vol.% in N7 electrolyte is denoted as 10N7). A similar strategy was used to create Li2O/Ethylene Carbonate (EC) electrolytes for comparison stud- ies. DOL, EC, and LiTFSI were purchased from Sigma-Aldrich and LiNO3 was purchased from Chem-Impex Int’l. Suspension electrolytes with Li2O particles in either N5 or N7 were first dropped onto Cu-foil and a Celgard 3501 separator was placed on top of the suspension. Another Celgard 3501 separator was used and DOL with 2 M LiTFSI was added. To create solid-state Li2O/poly(DOL) hybrid electrolytes we introduced 1 mM of the Lewis Acid initiator, Al(OTf)3 to the top layer of DOL electrolyte. Poly(DOL) utilized throughout this study was polymerized by 2 M LiTFSI and 1 mM Al(OTf)3. By varying the concentration of Li2O from 10 to 50 vol.%, hybrid poly(DOL) electrolytes with a range of phys- ical and electrochemical properties were facilely created. The hybrid electrolyte on Cu-foil was paired with Nickel cobalt manganese oxide (NCM811) cathode obtained from NEI Corporation, and coin 2302-type cells were assembled. All cells were rested for 10 hours before any electrochemical testing as polymeriza- tion concludes over time, outlined by previous kinetics studies (9). 3.2.2 Material characterizations FTIR spectra were characterized using a Thermo Scientific spectrometer in the attenuated total reflection (ATR) mode. The porous morphology of Celgard sep- arators was imaged by field emission Zeiss Gemini 500 scanning electron micro- scope (SEM). DC conductivity measurement was carried out using Novocon- 55 trol broadband dielectric/impedance spectrometer in the same coin 2032 cells without electrodes and Teflon ring instead of separators. Oscillatory shear mea- surements were performed using strain-controlled ARES-LS rheometer (Rheo- metric Scientific) with a cone and plate geometry (10 mm, 4° cone angle and 25mm, 1° cone angle). Thermogravimetric analysis (TGA) was conducted using TA Instruments Q500 under nitrogen atmosphere at 10°C/min ramping rate. TA Instruments Differential Scanning Calorimetry (DSC) Auto 2500 was uti- lized to evaluate thermal transitions under nitrogen flow at 10°C/min ramping rate. XRD tests were conducted on a Bruker D8 Discover powder diffractome- ter using Cu Kα radiation of approximate wavelength λ = 1.54 Å. X-ray photo- electron spectroscopy (XPS) was performed using Thermo Scientific Nexsa G2 X-Ray Photoelectron Spectrometer, with operating pressure of approximately 10−10 Torr, monochromatic Al Kα x-rays at 1486.6 eV, with 400 µm diameter anal- ysis spot. A flood gun was used for charge neutralization of non-conductive samples; all samples were charge corrected using adventitious carbon binding energy (284.8 eV). Raman spectra were collected using a WITec-Alpha 300R con- focal Raman microscope. A 532 nm green laser was used, a grating of 1200 l / mm (± 1 cm−1), 15 accumulations, and 30 s integration time. Particle size anal- ysis was done by Anton Paar particle size analyzer (PSA 1190) utilizing laser diffraction in liquid dispersion of Li2O/DOL with concentration of ≈ 5 vol.%. Stirring and sonication were done within the first minute of measurement and stopped to measure particle size over time every minute for 15 minutes. 56 3.2.3 Electrochemical testing Galvanostatic stripping/plating tests were performed using Neware CT-3008 battery tester at room temperature. Electrochemical impedance spectroscopy (EIS) measurements were performed by Solartron Frequency Response Ana- lyzer (Model 1252) with frequencies ranging from 50kHz to 10mHz and at an amplitude of 10 mV. Cyclic voltammetry (CV) was performed on BioLogic SP- 200 potentiostat. 3.3 Results and Discussions A simple force balance (Equation 3.1) comparing the gravitational, buoyancy, and thermal force, kT/r, on particles of radius, r, and density, ρs, dispersed in a liquid of density ρ f , indicates that particles with radii above a critical value, r ≥ rc will settle in a suspension to create a two-phase material in which a sub- stantially particle-free liquid phase coexists with a porous, granular material with solvent in its pores. Here k is Boltzmann constant (1.38 × 10−23 m2kg/s2K), g the gravitational acceleration (9.8 m/s2), and T is temperature (303 K). Taking the density of Li2O as ρs ≈ 2 g/mL at T = 303K, we find that rc = 0.62 µm for the Li2O/DOL (ρ f = 1.32 g/mL) suspensions and rc = 0.57 µm for the Li2O/EC (ρ f = 1.06 g/mL) suspensions. Thus, a Li2O/DOL or Li2O/EC suspension electrolyte composed of 100-mesh Li2O microspheres will exist as a two-phase system — a supernatant liquid containing Li2O particles with sizes less than around 600 nm and a precipitated porous bed formed by gravitational settling of larger parti- cles. 57 rc = ( 3kT 4πg(ρs − ρ f ) ) 1 4 (3.1) Results reported in Figure B.1 and B.3 show that Lewis-acid initiated poly- merization of Li2O/DOL suspensions produces a two-phase material in which an Li2O-lean poly(DOL) layer coexists with a Li2O-rich liquid DOL layer. In- terestingly, we find that if the suspensions are continuously agitated to pre- vent settling of Li2O particles, the ring-opening polymerization of DOL is com- pletely arrested. A straightforward explanation of these observations is possi- ble. Lithium oxide (Li2O) particles are Lewis bases with pH of 10 in Li2O/DOL suspensions. We note further that Li2O particles are known to adsorb anionic species on their surfaces [99]. We conclude that the Li2O particles neutralize the Lewis acid Al(OTf)3 initiator needed for ring-opening polymerization of DOL. In the absence of mixing, the particle concentration is highest in the gravity preferred direction, which means that DOL polymerization would occur at a progressively lower rate as the concentration of particles rises. In the synthesis cell configuration illustrated in Figure B.1a, this would result in an essentially liquid DOL electrolyte near the base of the cell. The liquid DOL electrolyte is in equilibrium with a hybrid Li2O/poly(DOL) electrolyte with a reduced con- centration of Li2O particles that are too small to settle on the timescale of the polymerization reaction, which is consistent with what is observed. Analysis of the particle size distribution using an Anton Paar Particle Size Analyzer (PSA) indicates that immediately after mixing the average suspended particle size is 18 ± 0.6 µm (Figure B.2). The particles are evidently large-enough to spontaneously settle. Indeed, after a period of approximately 15 mins fol- lowing mixing, gravity-driven settling is observed and a moderately narrower 58 particle size distribution is seen in the supernatant, with particle size plateau- ing at 25 ± 0.5 µm by the 15th minute. This is also accompanied by a new peak in the distribution associated with much smaller particles with sizes between 1.5 – 2.0 µm. SEM analysis of the settled particle phase reveal that some of the Li2O particles have sizes as large as 80 µm (Figure B.1g). Figure B.3 illus- trates the gradient property produced by particle settling. The resultant hybrid electrolyte is composed primarily of poly(DOL) with melting points at Tm = 31 ◦C and 49 ◦C as well as recrystallization temperature of Tc = -15 ◦C. On the other hand, Poly(DOL) formed near the settled particles lacks an obvious melt- ing point and its glass transition temperature is shifted to a lower value. These observations imply that whereas a lower molecular weight, amorphous poly- mer is formed by ring-opening polymerization of DOL in the particle-rich sedi- ment, a semicrystalline poly(DOL) material is formed when the polymerization occurs in the Li2O-particle lean supernatant. Thermogravimetric analysis in fact suggests that a more liquid-like organic-rich phase results from polymerization of DOL in the particle-rich phase. To understand how such a two-phase electrolyte might influence reversibil- ity of a Li//NCM811 battery cell or, more challengingly, a Cu//NCM811 anode-free battery cell, we created coin cells with the configurations illus- trated in Figure 3.1c. The cells were designed with two separators (Celgard 3501, which is manufactured to have micron-sized pores) as illustrated, and the Li/Cu anode oriented in the gravity assisted direction. In assembling the cells, Li2O/DOL suspension was dropped on the anode side and both separators ap- plied in sequence. A DOL electrolyte containing 1 mM Al(OTf)3 was thereafter applied to the cathode side and the cathode installed to complete the cell assem- bly. 59 In-situ ring-opening polymerization of Li2O/DOL electrolytes in such cells would produce materials with distinct characteristics in the separator near the particle-laden anode versus near the particle-lean cathode (Figure 3.1c). Anal- ysis by Fourier transform infrared spectroscopy (FTIR) and scanning electron microscopy (SEM) largely confirm these expectations. FTIR analysis of the sep- arator near the cathode reveals a strong poly(DOL) peak at ≈ 1000 cm−1 (8, 9), with no evidence of Li2O particles (Figure 3.1b). The opposite is seen for the separator near the anode, which is rich in Li2O particles. SEM images reveal striking morphological differences between the two separators. As polymer precursor wets the separator and polymerization happens, the cathode facing separator is seen to lose its porous microstructure as poly(DOL) forms within the pores (Figure 3.1e). In contrast the anode-facing separator retains its porous structure (Figure 3.1f). Thus, we are able to confirm that the poly(DOL) elec- trolyte is formed primarily at the battery cathode. At the highly reducing potentials at which Li+ is plated at a battery anode, all electrolyte components would be expected to degrade in time. The forma- tion of a well-formed SEI containing inorganic and organic components at the anode is a well-practiced strategy for passivating the electrode to prevent con- tinuous electrolyte degradation [151]. Li2O is an SEI component that have been reported in some systems [152, 138, 60], hence suspension electrolyte made up of Li2O suspension is a pathway to create an ideal SEI that protects the anode and, as we will demonstrate, possesses ionic conduction mechanism closely re- lated to liquid electrolytes. At the other end of the battery, highly oxidized cathode surface at high potentials tend to arouse interfacial irreversible reac- tion between cathode and electrolyte, forming the so-called cathode-electrolyte interphase (CEI) that could lead to capacity loss [8, 216]. It is then crucial to sup- 60 Figure 3.1: (a) Schematic illustration of method used to synthesize Li2O/poly(DOL) hybrid electrolytes with attractive gradient properties produced by gravity settling of Li2O. (b) A cell de- sign using a pair of Celgard 3501 separators is used to sequester the particle-rich phase to the region near the Li anode. (c) Re- sults from Fourier transform infrared spectroscopy (FTIR) anal- ysis of the anode and cathode facing sections of the separators. (d) SEM analysis of the structure of the cathode and anode- facing separator surfaces, compared with that of the pristine Celgard material. It is apparent that polymerization of the DOL yields a material that covers the pores on the cathode-facing separator (e). However, the separator on the anode side is seen to still retain the porous structure (f). 61 press side reactions of liquid DOL, which has low oxidative stability, to facilitate use in LMBs based on high-voltage cathodes. This can be done by polymerizing DOL into poly(DOL) near the cathode. The stratification of these two layers, all formed in-situ, is an overall pathway in creating more reliable SSEs that fit the anode/cathode duality of batteries. Oxide particles are conventionally thought to help ion dissociation by ad- sorbing ionic species, and hence create a space-charge layer that serves as an ionic conduction pathway [21], here we see no such effects. The strong Lewis basicity of the Li2O particles might then be expected to exert a large influence on ion transport by facilitating ion pair dissociation. To elucidate the role played by Li2O on the ion-transport properties, we employed a simple, theoretical frame- work attributed to Maxwell [133] to describe ionic conductivity data obtained in Li2O/DOL suspensions electrolytes. The Maxwell analysis has previously been applied to quantitatively explain ion transport in suspension electrolytes under electrochemical driving forces [160]. The conductivity σ of a suspension of par- ticles with conductivity, σp, is related to that of the suspending medium, σ0, and particle concentration, ϕ, through (Equation 3.2). The coefficient α is a function of the suspending medium, σ0, and particle, σp, conductivity (Equation 3.3). For perfectly insulating particles, σp ≪ σ0, α = 1/2 and σ/σ0 is a function of ϕ only (Equation 3.4). σ/σ0 = 1 − 2αϕ 1 + αϕ (3.2) α = σ0 − σp 2σ0 + σp (3.3) 62 σ/σ0 = 2(1 − ϕ) 2 + ϕ (3.4) We measured the temperature-dependent ionic conductivity of Li2O/DOL electrolytes as a function of ϕ (see Figure B.3a) and compare the experimen- tal results with theoretical predictions based on Equation 3.4 in Figure 3.2a. It is seen that up to a particle concentration of approximately 40 vol.%, ion trans- port in Li2O/DOL electrolytes are in nearly perfect agreement with the Maxwell model, revealing that the Li2O particles behave as perfect insulators. We won- dered whether these observations are a consequence of the low dielectric con- stant (ϵDOL = 19) of DOL and performed similar experiments using suspension electrolytes in which the same Li2O particles were dispersed in a solvent, ethy- lene carbonate (EC), with a substantially higher dielectric constant (ϵEC = 89.8). The results reported in Figure B.5 again clearly show agreement with Equa- tion 3 for particle concentrations up to 40 vol.%. Additionally, we note that notwithstanding the decrease in ionic conductivity with increasing Li2O parti- cle volume fraction, no significant changes in the activation energy Ea for ion transport are observed (see Figure B.4), especially for suspensions with particle volume fractions 10vol.%. This result confirms that the ion conduction mecha- nism [17, 154, 203] is essentially unaffected by the Li2O particles and implies that the ions move primarily via a network formed by the liquid DOL electrolyte. Like many particulate additives, Li2O addition offers a degree of mechanical reinforcement. Figure B.6 shows results from strain-dependent measurement of Li2O suspensions in DOL containing 1 M LiTFSI, where strain was varied from 0.025 – 10% at a constant angular frequency. Storage G’ and loss G” moduli reflect the solid-like and viscous-like contributions within the material, and both 63 increase with increasing Li2O particle concentration. G’ shows value as high as 100 Pa for 10 vol.%, comparable to those of low molecular weight polymers or oligomers used as polymer electrolytes [112]. Increasing Li2O concentration to 30 vol.% causes G’ to have a value of 1000 Pa. As suspension reaches 50 vol.%, G’ of 4 × 105 Pa is observed. This very high modulus is comparable to values typically observed in crystalline polymers or soft materials close to their glass/jamming transition [88]. The increase in mechanical strength is also accompanied by changes in the stress-strain curve. Materials that manifest a yield point typically first display a linear stress-strain response before yielding alters the slope of the stress-strain curve [169]. Many suspensions are known to have yield stress, and the denser the suspensions, the more likely yielding occurs as it takes a certain degree of deformation for particles to move and get out of their “cages”. Suspension with 10 vol.% Li2O does not exhibit any yield stress, but increasing particle concen- tration causes yielding behavior in the suspension. The yield stress value τy, taken from the point where deviation from linearity in the stress-strain curve happens, increases with increasing particle concentration and eventually reach- ing τy ≈ 100 Pa at 50 vol.% Li2O content. The more tortuous pathway due to particle aggregation likely causes deviation from Maxwell prediction and ex- plains the declining conductivity value with increasing volume fraction (Figure 3.2a). Hybrid electrolytes created by polymerizing DOL containing Li2O particles utilizing the method shown in Figure 3.1a, manifest higher ionic conductivity values than both precursor materials (Figure 3.2b and c). The hybrid electrolyte for example exhibits an ionic conductivity value of σ = 4.3 mS/cm at 30◦ C 64 Figure 3.2: (a) Normalized ionic conductivity value following Maxwell model for Li2O/DOL suspension electrolytes at various vol- ume fractions. (b) As polymerization of a Li2O/DOL hy- brid electrolyte (1st layer: 10N7; 2nd layer: poly(DOL) +2 M LiTFSI + 1 mM Al(OTf)3), proceeds with time, the ionic con- ductivity measured at 30°C first rises and stabilizes at val- ues higher than observed either for the precursor Li2O/DOL suspension or pure poly(DOL) electrolyte with the same salt concentration. (c) Temperature-dependent ionic conductivity values for in-situ-formed Li2O/poly(DOL) hybrid electrolyte, for a Li2O/DOL suspension electrolyte, and for in-situ-formed poly(DOL) electrolyte. (d) Temperature-dependent changes in ionic conductivity of an in-situ formed Li2O/poly(DOL) hy- brid electrolyte as a function of time following the onset of polymerization. while pure poly(DOL) and pure suspension have σ = 2.4 and 3.2 mS/cm, re- spectively. Remarkably, the enhanced conductivity values (σ ≥ 1 mS/cm) re- main even at temperatures as low as -30◦C. It is interesting that such high low- temperature ionic conductivity values are unattainable for the semicrystalline poly(DOL) electrolyte. Furthermore, the hybrid electrolytes manifest the lowest activation energy Ea of 4.2 ± 0.2 kJ/mol compared to pure poly(DOL) (Ea = 6.9 65 ± 0.2 kJ/mol) and pure suspension (Ea = 4.6 ± 0.3 kJ/mol). Ea values are cal- culated through Arrhenius equation σ = Ae− Ea RT with A being a pre-exponential factor, R universal gas constant, and T temperature. The hybrid electrolyte conductivity is seen to evolve with polymerization time, with conductivity value first increasing and ultimately plateauing after around five hours of polymerization. Temperature-dependent conductivity at the 1st, 5th, 10th, 24th, and 96th hour is shown in Figure 3.2d. A key finding is that the discontinuous transition in conductivity seen in the 1st hour, gaping below and above 30◦C, indicated by two shaded regimes in the figure largely eases as polymerization progresses with time, as seen in the continuous trend at 96th hour. We have previously reported a shift from a Vogel-Fulcher-Tamman (VFT)- like behavior of amorphous poly(DOL) hybrid electrolyte to an Arrhenius one as crystallization of poly(DOL) takes place [184]. As crystallization and associ- ated chain reconfiguration in semicrystalline polymers typically produce large changes in the conduction mechanism near the melting point or glass transition temperature Tg [154]. Discontinuous changes in ionic conductivity have been reported extensively to accompany these thermal transitions in a range of poly- mers [178, 38, 53]. Stratification of polydisperse microsphere suspension has been extensively examined [141, 142]; larger particles settle, and smaller ones get suspended in liquids. This is also what we observed through TGA (see Fig- ure B.3b), with a concentration gradient increasing towards the plane of highest gravitational force. The stratification process is a time-dependent process, with the system shifting from a non-equilibrium state to an equilibrium, stratified one [142]. To evaluate the electrochemical features of the hybrid electrolytes we stud- 66 ied Li2O/DOL suspension electrolytes containing 10 vol.% (Figure 3.3a) and 50 vol.% Li2O (Figure 3.3b) and 2 M LiTFSI in Li//NCM811 full and Li//Cu half cells. A fixed current density of 1 mA/cm2 was used for these experiments. The results reveal that at either of the Li2O concentrations studied, the cells cycle stably, reaching a discharge capacity of 1.45 mAh/cm2 by 100 cycles (Fig- ure 3.3c). Prior to cycling cells were subjected to a solid electrolyte interphase (SEI) buildup process at a low current density of C/10 for the first five cycles. The results are also interesting because DOL is not commonly used as a stand- alone electrolyte solvent in LMBs due to their poor oxidative stability (¡ 4 V) and chemical instability with the most used lithium salts, like LiPF6 [222]. Results in Figure 3.3d and 3e reveal that notwithstanding the uniform sta- ble cycling observed, the Coulombic efficiency (CE) and morphology of plated Lithium metal varies with Li2O volume fraction (Figure B.7). SEM analysis of the Cu electrode harvested from the Li//Cu half cells after depositing 10 mAh/cm2 Li on the copper reveal that the relationship between electrodeposit morphology and Li2O concentration is non-monotonic. Specifically, the rough- est Li deposits are observed for the pure DOL electrolyte, followed by elec- trolytes containing 50 vol.%, 30 vol.%, and 10 vol.% Li2O. The trend is not the same for the CE measured in Li//Cu cells. These measurements show that the CE rises from 98.0% for cells containing the pure DOL electrolyte to 98.8% for cells containing suspension electrolytes with 10 vol.% Li2O (Figure 3.3d). CE further increases to 99.2% for electrolytes containing 20 vol.% Li2O and changes negligibly thereafter with increasing Li2O concentration. Nyquist plots reported in (Figure B.8) indicate that both the coarser Lithium deposits and plateauing CE values coincide with an increase in interfacial resistance at higher Li2O particle concentrations. 67 Figure 3.3: Galvanostatic charge and discharge of Li//NCM811 in hybrid Li2O/DOL electrolytes containing (a) 10 vol.% and (b) 50 vol.% Li2O and 2 M LiTFSI. (c) The discharge capacity and Coulom- bic efficiency (CE) of the cells in (a) and (b) over 120 cycles. Cells were run with Li//NCM811 configuration at C/10 for the first five cycles, followed by 1C for the rest of the cycle. (d) Coulombic efficiency obtained in Li//Cu half-cells configura- tion for different Li2O volumetric concentrations. The resulting CE value at each concentration is presented in (e). Anode-free lithium cells provide a particularly challenging testbed for eval- uating the attributes of any hybrid or solid-state electrolyte. Because Li metal is not present during cell assembly, an anode-free LMB is also viewed as attrac- tive from a practical point of view because it removes safety issues involved in handling Lithium metal in a manufacturing setting [182, 156, 179] and en- hances the battery’s energy density as the electrode weight is reduced. Reliable anode-free LMBs are presently hindered by the low Coulombic efficiency (CE) due to interfacial reactivity, transport associated with the nonuniform deposi- tion of Lithium as well as its parasitic chemical and electrochemical reactions 68 with electrolyte components [182]. Figure 3.4a reports the galvanostatic cycling of Cu//NCM811 cells containing hybrid electrolytes created by ring-opening polymerization of a 10% Li2O/DOL suspension containing 1 mM Al(OTf)3 + 2 M LiTFSI + 0.5 M LiNO3. This electrolyte is identified as 10N5 in what follows exhibits a nearly constant CE value of 97% and a first-cycle discharge capacity of approximately 1.4 mAh/cm2 at a cycling rate of C/2. As illustrated Figure 3.4b the discharge capacity of the cells decreases gradually with cycling, but the rate of decrease is significantly lower than observed in previous studies of anode- free LMBs, and markedly lower than for anode-free cells based on the control electrolytes (i.e., DOL and Li2O/DOL suspension electrolytes) (see also, Figure B.9), underscoring the benefits of polymerization of the DOL. As illustrated in Figure 3.4c, the results reported in Figure 3.4b are remarkable for another rea- son. Specifically, for an anode-free battery cell that runs with a nearly constant CE, the discharge capacity at the nth cycle is related to the value at the 1st cycle (n = 1), as a simple power-law of CE: CEn = Capacityn Capacityn=1 (3.5) The dashed line in Figure 3.4c was obtained using Equation 3.5, with the value estimated from the experimentally measured discharge and charge ca- pacity per cycle (CE ≈ 0.97) inserted in the expression. A plot of log ( Capacityn Capacityn=1 ) against n linearizes Equation 3.4, with the slope equal to log CE. Comparing this line to the actual discharge capacity measured for the anode-free cells, a large discrepancy is observed. Namely, the anode-free cells containing the hybrid electrolyte show more efficient utilization of the Li stored in the cathode. We fit- ted the experimentally determined capacities to the linearized version of Equa- 69 Figure 3.4: (a) Galvanostatic cycling performance and electrochemi- cal properties of Anode-free Cu//NCM811 cells based on Li2O/poly(DOL) electrolytes for 10 vol.% Li2O. The poly(DOL) was polymerized inside the battery cell using 1 mM Al(OTf)3 and the cells were cycled at a rate of 0.5 mA/cm2 (C/2). For the results in (a) a salt blend consisting of 2 M LiTFSI + 0.5 M LiNO3 (here termed N5) was used. The respective discharge capacities and Coulombic efficiencies (CE) values are reported in (b). Included in (b) are discharge capacity values of DOL + 2 M LiTFSI electrolyte without and with the addition of 10 vol.% Li2O, cycled at C/10 for 5 cycles and a comparable C/2 rate for the rest of the cycles. Solid electrolyte interphase (SEI) buildup during the formation step at C/10 is detailed in Fig. S12. CE values are used to predict capacity fading shown in the dashed line of (c), while solid data points show actual capac- ity fading. The difference between CE of this electrolyte and a control DOL + 2 M LiTFSI electrolyte is shown in (d). The cor- responding Li//NCM811 cell performance for the same elec- trolyte as in (a) is shown in Fig. S11. (e) Cyclic voltammetry (CV) measured in cells in which Li2O particles in a carbon cloth (CC) current collector are paired with Cu-foil in a DOL + 2 M LiTFSI electrolyte. (f) X-ray photoelectron spectroscopy (XPS) of Li2O@CC electrode post-CV and held at oxidation potential for 5 hours indicating existence of oxygen, carbon, and lithium. Two peaks attributed to lithium oxide (Li2O) and lithium per- oxide (Li2O2) are seen in the (g) O scan and (h) Li scan. 70 tion 3.5 and extracted the CE value required to obtain the best fit. It is apparent that the effective CE ≈ 0.994 of the cells is much closer to unity. We estimated the apparent increment in CE at each cycle by taking the nth root difference of the discharge capacity measured in the anode-free cells cycled in the control elec- trolyte from those measured in the10N5 hybrid electrolyte. This analysis reveals a relatively stable increment of around 15% at each cycle (Figure 3.4d). We hypothesized that in addition to the role in reinforcing mechanical prop- erties of the hybrid electrolytes, in retarding polymerization of DOL, and in producing gradient properties in the hybrid electrolytes, the Li2O particles con- tribute some amount of Li to compensate for losses each cycle. As partial confir- mation of the hypothesis, Figure B.10 reports results from galvanostatic cycling of a Li//NCM811 cell that utilizes the 10N5 electrolyte under the same condi- tions used for the measurements reported in Figure 3.4a. The presence of excess Li in the cells clearly improves their cycling. As a more direct test of the hy- pothesis, we performed XRD and Raman spectroscopic analysis of Cu anodes harvested from the cycled cells. The results reported in Figure B.12 support formation of Li2O2 in the electrodes, but the signals are relatively weak. To assess the role of the Li2O in this process and to evaluate the reversibility of the formed Li2O2 under our cell running conditions, we performed extensive cyclic voltammetry measurements using Li2O//Cu cells (see Figure 3.4e.). The Li2O electrodes used for these studies were created by sandwiching Li2O par- ticles between carbon cloth (CC) as a current collector and using liquid DOL + 2 M LiTFSI as the electrolyte. Three peaks are evident in the cyclic voltammo- grams, with peak (3) attributed to decomposition of the liquid electrolyte. CV results for control CC//Cu cells employing the same electrolyte show an irre- versible reduction peak at ≈ -3.0 V, and no observable peaks (1) and (2) (Figure 71 B.11a). These peaks are then attributed to the oxidation (1) and reduction (2) of Li2O. The voltammograms also indicate that the peaks currents increase ap- proximately as the square root of scan rate (Figure B.11a) indicating that the re- actions are transport limited. Our findings are consistent with previous reports which show that Li2O undergoes an oxidation reaction during discharging to form Li2O2, a prominent reaction in Lithium-air battery systems [124, 157]. Additional support comes from XPS analysis of the Li2O electrodes used for the CV experiments. Results reported in Figure 3.4f – h were obtained by holding the Li2O electrodes at an oxidizing potential of approximately 1.5V (vs Li+/Li) for 5 hours. The high-resolution O 1s spectra were deconvoluted into two primary peaks centered at 531 ± 0.4 eV and 529 ± 0.3 eV, which indicate oxide (Li2O) and peroxide (Li2O2) bonding modalities respectively [200, 148]. An average of three measurements (Table B.1) revealed a 2:3 ratio of peroxide to oxide on the surface (up-to 10 nm scanning depth). These observations were further corroborated by high-resolution Li 1s scans – deconvoluted into two peaks centered at 53.3 ± 0.2 eV for Li2O, and 54.4 ± 0.4eV for Li2O2 – which also indicate a 2:3 distribution. E = E0 + RT nF ln ( [Ox] [Red] ) (3.6) |Ep − E(p/2)| = 2.2 RT nF = 57 n mV (3.7) The 2Li2O → Li2O2 + 2Li+ + 2e- reaction is also seen to be reversible in the DOL electrolyte. The Nernst equation (Equation 3.5) can also be utilized to see the reversibility of the redox reaction. It relates the potential of an electrochem- 72 ical cell E to the standard potential of a species E0, as well as the relative activ- ities of oxidized/reduced analyte at equilibrium. R is universal gas constant, F is Faraday’s constant, n is the number of electrons, and T is temperature. Estimating E0 as half-peak potential Ep/2 and taking E as the potential at the peak current Ep, it is straightforward to conclude that reversible redox reactions should have a value of 57 mV at T = 25◦C (Equation 3.6) [18]. The results in Figure B.11c reveal that for peaks (1) and (2), these values are 39 mV and 43 mV, respectively. The charge provided by the redox reaction, quantified by the area under the i – V curve, is shown to be approximately equal (Figure B.11c) indicative of a reversible reaction. To evaluate the longer-term, electrochemical fitness of the anode free cells, we subjected cells cycled at a fixed rate of C/2 for 100 cycles, to varying rates: 1C and 2C before returning it to the original current rate to C/2. The results reported in Figure B.13 show that the cells remain stable and that more than 92% of the capacity measured at the 100th cycle is recovered at the 140th cycle, when the C/2 rate is restored. We also investigated the combined effects of Li2O particle concentration and LiNO3 concentration on galvanostatic cycling of the anode-free Cu//NCM811 cells. The results reported in Figure B.14 show that a higher concentration of Li2O or LiNO3 increases the CE, but lowers the dis- charge capacity, which is accompanied by a higher overpotential. Consistent with the results reported in Figure 3.4 for hybrid electrolytes with 10% Li2O, results reported in Figure B.15 show that a best fit line to the linearized form of Eq. 4 again shows effective CE is increased when Li2O particles are present in the electrolyte. Contrary to what is observed in Li2O (Figure 3.3a), suspen- sions created by SiO2 in DOL electrolytes show lower CE with the addition of particles (Figure B.17). As SiO2 provides no Li through any redox reaction, this 73 further highlights the role of Li2O electro-active particles in compensating lost Li throughout battery cycling. 3.4 Conclusions Suspension electrolytes made up of micron-sized Li2O particles in DOL undergo Al(OTf)3 Lewis acid-initiated ring-opening polymerization inside a battery cell to create hybrid solid-state electrolytes (SPEs) with room temperature ionic con- ductivities exceeding 1 mS/cm at Al(OTf)3 and beneficial gradient properties. The gradient properties are thought to arise from neutralization of the Lewis- acid initiator by the basic Li2O particles. Gravitational settling of the Li2O parti- cles is shown to drive macrophase separation, which in-turn results in gradients in polymerization inside the cell. Likewise, retardation of ring-opening near the surface of Li2O particles creates a region of limited or no polymerization near the Li2O particle surfaces, yielding solid-state materials with room tempera- ture ionic conductivity exceeding 4 mS/cm. This behavior contrasts markedly with observations in Li2O/DOL suspension electrolytes, which manifest simple Maxwellian conductivity in which particles act as simple insulating inclusions in the liquid DOL host. Used as electrolytes in conventional Li//NCM81, as well as anode-free Cu//NCM811 cells, the hybrid electrolytes enable extended long-term cycling. Curiously, we find that the cycling achieved in the anode- free cells is substantially better than expected from the nominal Coulombic Ef- ficiencies (CE) deduced from their charge/discharge characteristic. We fit the cycling profiles to a power law and estimate the effective CE in the cells to be 99% or higher, indicating that cycling benefits from a source of Li other than the NCM811 cathode. By means of cyclic voltammetry and spectroscopic analysis, 74 we show that reversible redox reaction of the Li2O particles localized near the anode contribute a small amount of Li each cycle that compensates to an extent for the normal parasitic losses, extending the cycle life of the anode free cells to levels heretofore unseen in the literature. 75 CHAPTER 4 IN-SITU SYNTHESIS OF SOLID POLYMER ELECTROLYTES WITH HIGH IONIC MOBILITY 4.1 Introduction Lithium-ion migration and diffusion across interfaces formed at battery elec- trodes and solid-state electrolytes (SSEs) is essential for reversible storage of electrical energy [131]. A prominent issue impeding wide-spread adoption of battery technologies based on SSEs is the poor interfacial contact and hence poor ion transport across the electrode/electrolyte interface [84, 95, 221]. Recently we reported that in-situ polymerization of liquid electrolytes inside electrochemical transforms the liquid precursor to a solid-state polymer (SPE) electrolyte with good interfacial contact with all electrode components [219], removing barri- ers to interfacial charge transport at both the implicit and explicit electrode- electrolyte interfaces formed between the SPE and the cathode and anode, re- spectively, of lithium metal batteries. Based on ring-opening polymerization (ROP) of 1,3-dioxolane (DOL) monomer initiated by Lewis acid salts (e.g., alu- minum triflate Al(CF3SO3)3 and Al(OTF)3) introduced to the liquid electrolyte, the poly(1,3-dioxolane) (polyDOL) SPE formed by this in-situ process is fun- damentally challenged by the spontaneous reversibility of the ROP reaction. ROP of other larger ring heterocyclic monomers, such as 1,3,6-trioxocane (7- membered, and higher order rings) and 1,3-dioxolane (5-membered ring) like- wise suffer from reversibility of the chain propagation reaction [132], which is now understood to be a consequence of the lack of a significant energy differ- ence between the ring closed and the ring opened state of the monomers. At 76 equilibrium, the reversibility of the ROP of DOL leads to a residual monomer fraction as high as 30 wt.%, enhancing ion transport in the bulk and at inter- faces, relative to what would be expected for a pure polyDOL SPE [219]. The presence of such large DOL monomer concentrations in a “solid-state” electro- chemical cell is problematic, however, because it lowers the oxidative stability of the electrolyte. Additionally, when used as the electrolyte battery cells that uti- lize low-cost, ether-soluble cathodes (e.g., Lithium//Sulfur batteries), the un- polymerized DOL fraction has been reported to dissolve intermediate cathode reaction products such as lithium polysulfide (Li2Sx, 2 < x ≤ 7), causing rapid loss of cathode capacity and a continuous charging process termed polysulfide shuttling, when the dissolved Li2Sx diffuses from the cathode to the Li anode and is electro-reduced [184]. Surface-anchoring of polymer chains during the propagation step provides a straightforward process for reducing depolymerization kinetics, hence limit- ing the fraction of residual monomer at equilibrium. The approach has been used to anchor free radicals in emulsion polymerization of styrene and methyl methacrylate and reported to yield very high monomer conversions [14]. In emulsion graft copolymerization, higher conversion of monomer to polymer has likewise been associated with increasing polymer grafting density [1]. We note that studies of polymerization of 5-membered and larger rings have estab- lished that the enthalpy change associated with the forward (polymerization) reaction is of lesser consequence in determining the polymer fraction at equilib- rium, in comparison to changes associated with entropic factors [132]. The pos- itive rotational and torsional entropies have for instance been shown to become much larger than the translational entropy in larger rings, inducing polymeriza- tion of large macrocycles [185, 136]. Reducing free volume of polymer chains by 77 increasing grafting density ihas been argued to reduce entropic contributions associated with chain rotation, torsion, and translation of chain segments in polymer brushes in good solvents [145]. Finally, crosslinking of polymer chains have been reported to increase conversion, characterized by increasing moduli with increasing crosslinking density [165]. Here we investigate the reversibility of ROP of DOL in the presence of multifunctional ethers able to copolymerize with or cross-link active polyDOL chains during the chain propagation reaction. We find that in the presence of such species, ROP of DOL performed in conventional carbonate electrolytes, e.g. EC/DMC as a solvent, yields solid-state branched/cross-linked SPEs with substantially reduced residual monomer fractions at equilibrium. The SPEs are reported further to manifest liquid-like ionic conductivity, temperature- dependent ionic conductivity following Vogel-Fulcher-Tammann (VFT) behav- ior, and high-enough oxidative stability to support stable cycling of solid-state Lithium-metal and Lithium-ion batteries based on high-Ni NCM811 cathodes. Finally, the ionic conductivity of the crosslinked SPEs is reported to decrease as a nearly exponential function of the multi-functional ethers, consistent with expectations for arrest of chain dynamics produced by shorter chain segments between crosslinks. 78 4.2 Materials and Method 4.2.1 In-situ fabrication of crosslinked SPE and electrochemical cells All chemicals were obtained from Sigma-Aldrich unless otherwise specified. 100 µL solution of 1.0 M or 2.0 M LiPF6 in 1:1 ethylene carbonate : dimethyl car- bonate (EC/DMC) was dropped onto a Celgard 3501 separator that had been assembled to be sandwiched between graphite or Lithium anode and NCM811 cathode. An equal amount of DOL monomer was added to the EC/DMC + LiPF6 solution to create UN05M or UN1M, with UN05M indicating a final LiPF6 concentration of 0.5 M. In order to create CRxMy star/crosslinked polyDOL, y% of trimethylolpropane triglycidyl ether (TMPTGE) was added. Coin 2032 cells were employed to create Li/NCM811, graphite/NCM811, and Li/Cu cells. The pouch cell was assembled by employing 9 cm2 (3 cm x 3 cm) graphite and NCM811, with the electrolyte also fabricated in-situ (inside the pouch cell). All electrochemical cells were prepared inside an argon glove box (Inert Inc.) with O2 and H2O content lower than 0.5 ppm. 4.2.2 Characterization of crosslinked SPE All mechanical rheology measurements were performed using a stress- controlled MCR501 (Anton Paar) with a parallel plate geometry (diameter of 25 mm). All measurements were carried out at T = 25◦C, with employed an- gular frequency of ω = 10 rad/s and strain of γ = 1% for time-dependent mea- 79 surements and γ = 10% for frequency-dependent measurements. Differential scanning calorimetry (TA Instruments Q2500 DSC) was adopted to evaluate the thermal transitions of the electrolytes. Thermal transitions were measured un- der nitrogen flow at a fixed ramp of 10◦C/min. Bruker Vertex V80V Vacuum Fourier transform infrared spectroscopy (FTIR) was utilized to evaluate chem- ical bonds in the electrolyte systems. The attenuated total reflectance (ATR) mode was used throughout the measurement. Bruker AV 500 nuclear mag- netic resonance (1HNMR and 13CNMR) were used to identify the hydrogen and carbon species within the electrolytes. Prior to NMR measurements, elec- trolytes were dissolved in dimethyl sulfoxide-d6 (DMSO-d6) solvent. DC con- ductivity measurement was carried out using Novocontrol broadband dielec- tric/impedance spectrometer. The same coin 2032 cells without electrodes were utilized, with Teflon ring containing the electrolytes. 4.2.3 Electrochemical characterizations Galvanostatic lithium stripping/plating tests were done using Neware CT-3008 battery tester at room temperature, with a current density of 1 mA/cm2 for graphite/NCM811 coin cells, 0.5 mA/cm2 for graphite/NCM811 pouch cell, and 0.5 mA/cm2 for Li/NCM811 coin cells. Electrochemical impedance spec- troscopy (EIS) measurements were performed using Solartron Frequency Re- sponse Analyzer (Model 1252) with frequencies ranging from 50 kHz to 10 mHz and at an amplitude of 10 mV. Linear sweep voltammetry (LSV) was performed on BioLogic SP-200 potentiostat. 80 4.3 Results and Discussions Poly(1,3-dioxolane) (polyDOL) has previously been reported to be produced by a Lewis acid assisted polymerization of DOL using Al(OTf)3 as initia- tor [219, 184, 146]. Dissociation of dissolved salts like LiDFOB and LiPF6 [218, 98, 204] used in electrolytes (e.g., LiPF6 dissociates into LiF and PF5 anion), could also ring open DOL monomer to produce (polyDOL) [98]. Both methods yield polymers with high residual monomer content due to the reversibility of the ring-opening polymerization reaction [219, 132, 146]. The addition of 0.5 mM Al(OTf)3, for example, results in approximately 20% residual monomer content [219]. One method to alleviate this issue, borrowed from the concept of the well-known grafting polymerization that results in high conversion, is to introduce a crosslinker. The addition of initiation sites provided by an anchor point, which may arise either from grafting or crosslinking, has been reported to yield polymers with high conversion, but relatively low molecular weight [14]. To synthesize the SPEs used in the study, a solution of LiPF6 in ethylene car- bonate and dimethyl carbonate (EC/DMC) was first mixed with DOL monomer and left to stand. By varying the concentration, x, of LiPF6 it is possible to create un-crosslinked polyDOL in the form (UNxM). Addition of Trimethylol- propane triglycidyl ether (TMPTGE) crosslinker to the synthesis mixture yields branched/crosslinked polyDOL of the form (CRxMy, with y being the percent- age of TMPTGE) (Figure 4.1c). Depending on how much TMPTGE is added, the epoxide integrated into the growing polyDOL chain could undergo ring- opening, producing a crosslinked polymer (Figure 4.1c). The resulting SPEs formed by these reactions can be classified as either a 81 randomly branched or crosslinked with variable fractions of dangling chains by probing their thermal and/or mechanical properties. Crosslinked polymers are notably impossible to remold [143, 226], which is evidenced by an absence of a melting or softening transition. Differential scanning calorimetry (DSC) traces in Figure C.1 report the thermal properties of the SPE materials studied. At tem- peratures up to 100°C, CR05M20 shows no signs of softening as apparent from the weak temperature dependence of the storage moduli G’ (Figure C.2). Obser- vation of the materials post rheological testing at high temperature confirm that it remains solid. These findings are different from what is seen for CR05M5 and CR05M10, where both materials undergo an obvious softening transition at tem- peratures at around 75°C; ultimately transitioning to a liquid-like state at higher temperatures. Crosslinking effects are also observed in poly(DOL) polymerized with 100 mM Al(OTf)3, reducing its crystallinity seen through DSC (Figure C.1). To evaluate the reversibility of the resultant polyDOL-based SPEs, 1H Nu- clear Magnetic Resonance Spectroscopy (1HNMR) analysis was used to deter- mine the residual monomer content in the various materials. These experiments reveal that UN05M has a residual monomeric content as high as 77% (see Ap- pendix C), defining the material as a soft polymer. Addition of 10% TMPTGE to the polyDOL produces a greater than five-fold reduction in the residual monomer content (Figure 4.1a). The 1HNMR results for UN05M, CR05M2, CR05M4, CR05M6, and CR05M10 shows that one can systematically reduce the residual monomer content in the SPE by simply varying the concentra- tion TMPTGE as illustrated in the (Figure 4.1a, Figure 4.1, and Figure C.3). In networked polymers, unreacted functional groups associated with fraction of soluble polymer chains are known to arise due to dangling chains result- ing from random insertion of crosslinks. These chains have been reported to 82 profoundly affect physical properties of the materials both in the bulk and at interfaces [113, 19]. These challenges are compounded by the fact that studies of poly(dimethyl siloxane) crosslinking reactions show that even samples with complete reaction turnover contain vinyl groups that do not participate in the crosslinking reaction [34]. Results from the same study indicate that the fraction of soluble polymer chains, termed the sol fraction, also decreases with increas- ing crosslinking density. At low crosslinking densities, the molecular weight distribution tends to be broad, and complex structures are present [34]. Increas- ing crosslinking density decreases the size of these chains and the presence of complex structures is replaced by more linear ones. In order to more concretely estimate the residual monomer content, we per- formed control NMR analysis of mPEG/DOL mixtures with known concentra- tion of free DOL. The resulting DOL ratio deduced from the controls agree well with the ratio in the mPEG/DOL mixtures employed in the analysis (Figure C.4). Fourier transform infrared spectroscopy (FTIR) performed on CR05M10 indicate that the material shows that the signature peak is consistent with that observed for polyDOL polymerized with Al(OTf)3 (grey arrows), with an un- ostentatious signature peak belonging to DOL monomer (pink arrow) (Figure 4.1b). Taken altogether our results for the control mPEG/DOL systems and for the CRxMy materials confirm that the presence of the crosslinker enhances the conversion of DOL from monomer to polymer, which can be understood as the arrest to the reversibility of the ROP of DOL. Temperature-dependent ionic conductivity of the crosslinked and uncross- linked polyDOL SPEs are reported in Figure 4.3a. The results show that the val- ues are lower for the former materials, but the room-temperature conductivity 83 Figure 4.1: (a) 1HNMR and 13CNMR spectra of 1 M LiPF6 + EC/DMC, polymerized UN05M, and crosslinked CR05M10. (b) FTIR spectra of DOL, polyDOL polymerized using 1 mM Al(OTf)3, and crosslinked CR05M10. (c) Polymerization to form branched polymers and crosslinking mechanism of branched polyDOL using TMPTGE. 84 Figure 4.2: Residual monomer content of different TMPTGE concentra- tions ranging from 0% (UN05M) to 10% (CR05M10). Values are obtained through 1HNMR peak integration of data shown in Figure C.3. values are still quite high (σ ≈ 6 mS/cm for CR1M5 compared to σ ≈ 9 mS/cm for UN1M) (Figure 4.3b). The ionic conductivity is also seen to follow the Vogel- Fulcher-Tamman (VFT) equation even after crosslinking (σ = A exp ( B T−T0 ) , with A and B constants, T temperature and T0 Vogel temperature of glass transition temperature Tg − 50) (Figure 4.3.a). The VFT temperature-dependence, paired with the high, liquid-like conductivity value, is consistent with our intuition that ion transport happens primarily in the liquid, unpolymerized EC/DMC mobile carrier. In many cases an ion-hopping mechanism reflected in the Ar- 85 Figure 4.3: (a) Temperature-dependent ionic conductivity showing Vogel- Fulcher-Tamman (VFT) behavior and (b) ionic conductivity values at room temperature for different TMPTGE concen- trations for CR1My. The y2-axis is complex viscosity of the CR05My samples with the same variation of TMPTGE concen- trations. rhenius temperature dependence has been reported for SPEs [219, 184, 104, 167]. The arrest of polymer dynamics and the lack of mobile carrier often also extends to lower ionic conductivity values, which is the case of highly crystalline poly- DOL. PolyDOL polymerized using high concentration of 5 mM Al(OTf)3 could have an ionic conductivity of as low as 0.02 mS/cm [219]. Results reported in Figure 4.3b reveal an exponential decay in ionic conduc- tivity with TMPTGE concentration. This finding is consistent with the expected exponential increase in electrolyte viscosity (right axis of Figure 4.3b), viewed either from a branched melt or solution perspective [158, 56]. Decreasing LiPF6 concentration in the cross-linked SPEs also lowers ionic conductivity (Figure C.5). The as-synthesized branched/crosslinked PolyDOL not only manifests 86 higher conversion with lower residual monomer content, but also exhibits en- hanced mechanical strength. Time-dependent rheology measurements (Figure 4.4a) for instance suggest that a stable crosslinked structure is achieved after 8 hours. We note that a consequence of the fast onset of polymerization is the instantaneous rise of, G’ and it dominates over G” — well before crosslink- ing starts. As crosslinking progresses, both moduli increase, indicative that a new structure is being established before equilibration. Post-equilibration of crosslinked structure, CR1M10 (cyan data points) reaches an elastic modulus G’ of 104 Pa while UN1M’s (grey data points) sits at slightly above 0.1 Pa (Fig- ure 4.4b). CR1M10 G’ is also higher than what is previously reported polyDOL polymerized with 0.5 mM Al(OTf)3 (G’ ≈ 103 Pa) [219] and when polyDOL is turned into a composite by the dispersion of hairy nanoparticles with the same polymerizing initiator at a concentration of 1 mM (G’ ≈ 103 Pa) [184]. More in-depth understanding of the SPEs can be obtained by probing their dynamic rheological properties at descending angular frequency (Figure 4.4c). Crosslinked and highly entangled branched polymers typically exhibit nearly frequency-independent G’ values over a wide range. Viscous-dominated mate- rials like UN1M (G” > G’) in contrast manifest increasing moduli with increas- ing angular frequency – a trend commonly seen in viscoelastic liquids. On this basis, we characterize UN1M as a soft glassy material [68, 5, 224]. The value of G’ in the plateau (frequency-independent) can be correlated with mesh size of a network through the equation ξ = ( G ′NA RT )− 1 3 [5, 91, 55, 35]. Mesh size ξ depends not only on G’, but also on Avogadro’s number NA, gas constant R, and tempera- ture T . As the concentration of TMPTGE is raised in CR05M system, an increase in G’ is observed, which translates to a decrease in mesh size ξ (Figure C.6). The smaller mesh size with higher crosslinking density has been observed many 87 Figure 4.4: (a) Time-dependent rheology measurement indicating the in- creasing storage (G’) and loss (G”) moduli with polymeriza- tion and crosslinking time for CR1M10. (b) Enhanced strain- dependent mechanical properties of G’ and G” with crosslink- ing, with the grey data points belonging to UN1M and blue to CR1M10. (c) Liquid-to-solid mechanical responses due to crosslinking probed through frequency-dependent rheology measurement. times in crosslinked and gel systems [199, 198, 46, 163]. As crosslinking density increases, there is a higher chance for crosslinking of shorter chains compared to propagation to create longer polymer chains. Shorter polymer chain with div- ing ξ turns to increase relaxation time of the chains, increasing viscosity as also seen in Figure 4.3b [163]. We characterized the electrochemical properties of the cross-linked SPEs using linear sweep voltammetry (LSV). Results reported in Figure C.9 show that whether un-crosslinked or crosslinked, the SPEs are stable up to approx- imately 4.5 V. To illustrate the potential benefits of the materials in battery cells, we performed galvanostatic cycling studies of Cu/Li half cells, as well as graphite/NMC811 Li-ion and of Li/NMC811 Li-metal batteries at a current density of 1mA/cm2 and 0.5 mA/cm2. The effectiveness of the electrolyte in conducting ions inside the batteries can be evaluated through Coulombic ef- 88 Figure 4.5: Galvanostatic stripping and plating for 50 cycles of Li/NCM811 of (a) CR05M10, (b) CR05M20, and (b) CR1M10, as well as their respective Coulombic efficiencies (CE) and areal discharge capacity (d). Measurements employed current density of 0.5 mA/cm2. ficiency (CE) by plating and stripping lithium to and from copper foil, which is shown in Figure C.10. By employing the ratio of charging time to 60-minute charging in Li-Cu configuration, it was found that the CE of CR05M5, CR05M10, and CR05M20 approaches 85%. Figure 4.5 illustrates the first 50 cycles of galvanostatic plating and strip- ping of lithium-metal batteries composed of NCM811 cathode, Li-metal anode, CR05M10 (Figure 4.5a), CR05M20 (Figure 4.5b), and CR1M10 (Figure 4.5c) elec- trolytes. By the 50th cycle, CR05M5 reaches a stable discharge capacity of 0.84 89 mAh/cm2, while CR05M20 and CR1M10 reach 0.32 and 0.99 mAh/cm2, respec- tively (Figure 4.5d). The functionality of this electrolyte is also tested with a pouch cell with 9 cm2 electrodes of graphite and NCM811 (Figure C.7). As the ionic conductivity falls with the addition of TMPTGE crosslinker, so does the electrochemical performance in lithium-ion battery. At 20% TMPTGE concen- tration, CR05M20 reaches the lowest discharge capacity, as well as experiences the harshest capacity fading over cycles. This trend holds true for Li-ion batter- ies comprised of NCM811 cathode and graphite anode (Figure C.8). In both cell configurations, CR1M10, which has higher ionic conductivity than its CR05M counterparts, remains the electrolyte with the least amounts of capacity fading. Despite the minimized residual monomer content, there is still a concern of the EC/DMC mobile species’ volatility. In lieu of LiPF6 in EC/DMC, the initiator of Al(OTf)3 could be utilized to polymerize DOL before crosslinking the polyDOL with TMPTGE. The resulting SPE containing 1 mM Al(OTf)3, 2 M LiTFSI, and 5% TMPTGE is presented in a preliminary 5-cycle Galvanostatic charging and discharging (Figure C.11). Due to the lack of mobile species, room- temperature ionic conductivity of this SPE is lower than when DOL is polymer- ized using LiPF6 in EC/DMC. However, ionic conductivity values of both SPEs copolymerized with 10% and 30% TMPTGE still exceed 1 mS/cm (Figure C.12). Increasing initiator content to 10 mM Al(OTf)3 decreases the ionic conductivity value further, as observed previously by [184]. 90 4.4 Conclusions Despite its benefits in creating good interfacial contact between electrodes and electrolyte when polymerized in-situ, PolyDOL polymerization is still plagued with inherent reversibility. We demonstrate the suppression of DOL polymer- ization reversibility inside electrochemical cells through the use of TMPTGE crosslinker. The addition of 10% TMPTGE into a system of UN05M could dra- matically shift the residual monomer content from 77% to 13.5%, calculated through 1HNMR peak integrations. The addition of 20% TMPTGE concentra- tion is shown to create a crosslinked thermoset. Below 20% TMPTGE concen- tration, entangled branched polymers are generated. Furthermore, a significant increase in modulus obtained from mechanical rheology measurements is also observed. UN1M, a soft polymeric material, could be turned into CR1M10, a material with elastic modulus G’ of 104 Pa. The addition of crosslinker also reduces the mesh size of the crosslinking/entanglement network, which pro- motes longer relaxation time of polymer chain and an exponentially higher vis- cosity. This exponential increase in viscosity arrests chain motion, which in turn exponentially decreases ionic conductivity. Ionic conduction mechanism is ob- served to happen in the liquid, unpolymerized EC/DMC caged within the 3D network. Due to the lower ionic conductivity, electrochemical performance of Li-ion and Li-metal batteries comprised of higher TMPTGE concentrations de- cline. The best performing electrolyte is seen to be CR05M5, which reaches a discharge capacity of 0.625 mAh/cm2 with CE higher than 99% by the 50th cy- cle. Utilizing the liquid electrolyte of LiPF6 + EC/DMC encapsulated within branched/crosslinked polymer network, we were able to achieve a solid-state, in-situ fabricated electrolyte with high ionic conductivity. 91 APPENDIX A CHAPTER 2 APPENDIX SiO2 core volume fraction and tether density from thermogravimetric analy- sis (TGA) The core volume fraction is calculated from wS i, wPEG, and densities of SiO2 ρS i and PEG ρPEG such that ϕc = wSi ρSi wSi ρSi + wPEG ρPEG = 17% 1.9 g/cm3 17% 1.9 g/cm3 + 83% 1.2 g/cm3 = 0.114 (A.1) The grafting density of chains Σ can be calculated as Σ = wPEG Mn NA (4πR2 0wSi) ( 4 3πρSiR3 0) = wPEG NA R0 ρSi 3 Mn wSi = (83%) (6.0x1023 mol−1) (5.0 nm) (1.9 g/cm3) 3.0 (5.0x103 g/mol) (17%) = 1.9 chains/nm2 (A.2) Random close-packing estimation of interparticle distance Simple geometric model proposed by Liu, et al. [125] stated that HNPs have an effective radius Re f f that is a function of core volume fraction ϕc and core radius R0 = 5 nm 92 Figure A.1: Melting point of pure PEG-SiO2 hairy nanoparticles (red curve) and free PEG chains of the same molecular weight Mn = 5 kDa (black curve). Tethering causes the conformation to shift from extended (Tm = 61°C), single-folded (Tm = 59.5°C), and double-folded (Tm = 53.4°C) to only the extended chain conformation (Tm = 61.3°C) Figure A.2: Thermogravimetric curve used to calculate the weight fraction of SiO2 and PEG of the HNPs. The weight fraction of SiO2 wS i is obtained from the remaining weight at 600°C, and the weight fraction of PEG wPEG is the weight loss between 100°C and 600°C. Weight fractions for sample above was found to be wS i = 17 wt.% and wPEG = 83 wt.%, with the calculation of core volume fraction ϕc and grafting density Σ given below. 93 Reff = ( ϕc 0.64 )− 1 3 R0 (A.3) where 0.64 is the volume efficiency of random close packing. The average inter- particle distance d(p−p) is then defined as dp-p = 2Reff = 8.6 × ϕ− 1 3 c (A.4) Calculating this with the self-suspended HNPs with ϕc = 11 vol.%, the random close-packing estimation is dp-p = 18nm (A.5) Table A.1: Transition temperatures Tc, Tm, Tg and heat of melting ∆Hm of self-suspended HNPs, poly(DOL), and hybrid systems ob- tained from DSC curve in Figure A.4 Sample Tc (°C) Tm (°C) Tg (°C) ∆Hm (J/g) Self-suspended HNPs (ϕc = 15 vol.%) 31 52 -29 90 Poly(DOL) -9.6 32 -28 110 Poly(DOL) + 10 mM LiNO3 6.7 48 -66 99 Hybrid (ϕc = 2.7 vol.%) 3.1 33, 46 -30 41, 61 Hybrid (ϕc = 2.7 vol.%) + 10 mM LiNO3 -6.5 27 -30 122 94 Figure A.3: Temperature-dependent conductivities of neat and hybrid samples probed through dielectric relaxation spectroscopy (DRS). The conductivities presented are of poly(DOL) (1 mM Al(OTf)3, green open circles and 10 mM Al(OTf)3, blue open stars), hybrid system composed of HNPs (ϕc = 2.7 vol.%) in poly(DOL) without (red open diamonds) and with the addi- tion of 0.5 M LiNO3 (blue open stars). All samples include 2 M LiTFSI salt. 95 Figure A.4: DSC curves of self-suspended HNPs (ϕc = 15 vol.%), poly(DOL), poly(DOL) with 10 mM LiNO3 addition, and hy- brid systems containing ϕc = 2.7 vol.% with and without LiNO3. All poly(DOL) samples were polymerized with 50 mM Al(OTf)3. All samples possess single recrystallization temper- ature Tc and broad glass transition temperature Tg, with a hy- brid sample indicating a couple melting points Tm. The two Tm in the hybrid show contribution of the two components: HNPs and poly(DOL), but the single Tc show that upon recrystalliza- tion, the two components co-crystallized. Crystallinity was in- ferred from both Tc and the heat of melting ∆Hm, with lower value of Tc indicates less crystalline domains and lower value of ∆Hm indicates lower extent of crystallinity in those domains. All key transition temperatures are summarized in Table A.1. 96 Figure A.5: Time sweep measurement of DOL monomers polymerized with (a) 1 mM (b) 15 mM (c) 20 mM (d) 30 mM (e) 40 mM and (f) 50 mM Al(OTf)3. All room temperature time sweep measurements were carried out at angular frequency of ω = 10 rad/s and strain of γ = 5%. 97 Figure A.6: Time sweep measurement of DOL monomers polymerized with (a) 20 mM (b) 30 mM (c) 40 mM and (d) 50 mM Al(OTf)3 in the presence of HNPs (ϕc = 2.7 vol.%). The insets show clearer peaks in G” and G’ at early times. All room temper- ature time sweep measurements were carried out at angular frequency of ω = 10 rad/s and strain of γ = 5%. 98 Figure A.7: Time sweep measurement of DOL monomers polymerized with (a) 20 mM (b) 30 mM (c) 40 mM and (d) 50 mM Al(OTf)3 in the presence of 10 mM LiNO3. Dashed lines show data points from Figure A.4 as a comparison between poly(DOL) polymerization with and without LiNO3. All room tempera- ture time sweep measurements were carried out at angular fre- quency of ω = 10 rad/s and strain of γ = 5%. The low viscosity upon the addition of LiNO3 for high initiator contents created difficulty in observing the end of induction time and the start of the initiation process. Hence, even though the samples were loaded as soon as there were apparent viscosity changes, the resulting time-dependent data seems to show initiation pro- cess began earlier than what was inferred. 99 Figure A.8: Time sweep measurement of DOL monomers polymerized with (a) 20 mM (b) 30 mM (c) 40 mM and (d) 50 mM Al(OTf)3 in the presence of HNPs (ϕc = 2.7 vol.%) and 10 mM LiNO3. All room temperature time sweep measurements were carried out at angular frequency of ω = 10 rad/s and strain of γ = 5%. 100 Figure A.9: (a) Time taken to start obtaining plateau in G” and G’ during polymerization (tp) (b) equilibration time post-crystallization (tc) for PolyDOL samples with Al(OTf)3 content varying from 10 to 50 mM. (c) Time taken reach the first peak in G’ (tpk) and (d) to obtain equilibrium moduli G′′eq and G′eq (teq) in DOL polymerization in the presence of HNPs (ϕc = 2.7 vol.%). 101 Figure A.10: Equilibrium moduliG′′eq and G′eq for DOL polymerization in the presence of HNPs (ϕc = 2.7 vol.%) (Figures 2c and A.5) normalized by the initial moduli values G′′0 and G′0 over Al(OTf)3 content of 1 – 50 mM. The resulting ratios G′′eq/G ′′ 0 and G′eq/G ′ 0 show an increasing value with the initiator con- tent after 10 mM initiator content. 102 Figure A.11: Strain-dependent measurement of self-suspended HNPs (ϕc = 18 vol.%) and hybrids (ϕc = 2.7 vol.%) polymerized with Al(OTf)3 content varying from 1 mM to 50 mM. Strain sweeps of hybrids were measured at 70°C with angular frequency ω = 10 rad/s while result for self-suspended HNPs was ob- tained at 70°C withω = 0.25 rad/s. The value of loss modulus G” is normalized to loss modulus at zero strain G′′γ→0 to em- phasize peak in G” at high strain in self-suspended HNPs, which indicates soft glassy behavior. This peak is minimal in the hybrids, illustrating the loss of soft glassy behavior and the closer-to-polymer nature of the hybrids. 103 Figure A.12: Effects of LiNO3 addition towards the electrochemical perfor- mance of PEG-SiO2 HNPs/PolyDOL electrolyte in Li//Cu cells. (a) Reactivity was observed by comparing chronoam- perometry results with potentials descending from 2.0 to 0.2 V of poly(DOL) hybrids (ϕc = 2.7%) without and (b) with the addition of 0.5 M LiNO3. (c) Interfacial stability was observed via electrochemical impedance spectroscopy (EIS) post-chronoamperometry at different potentials for the same hybrid electrolyte without and (d) with 0.5 M LiNO3. All electrolytes include 2 M LiTFSI salt and all poly(DOL) elec- trolytes were polymerized with 1 mM Al(OTf)3. 104 Figure A.13: (a) Galvanostatic cycling profile for Li//LFP cell with hy- brid electrolyte containing ϕc = 2.7%, 2 M LiTFSI, and 0.5 M LiNO3, and (b) without PEG-SiO2 HNPs. Both electrolytes were cycled at a rate of 0.2C. The discharge capacity and Coulombic efficiency (CE) over cycle for both electrolytes are shown in (c). It is shown that despite larger capacity fade in hybrid electrolyte before stabilizing, hybrid electrolyte pos- sesses higher discharge capacity throughout the cycle range presented as well as a more stable overpotential value. 105 APPENDIX B CHAPTER 3 APPENDIX Table B.1: Three locations of Li2O@Cu were probed through XPS, with the resulting Li2O and Li2O2 percentage listed for both high resolu- tion O- and Li-scan. Survey scan shown in Figure 3.4g presents different atomic percentage of elements of O, C, and Li. High-resolution O-scan High-resolution Li-scan Li2O Li2O2 Li2O Li2O2 Spot 1 7.06 92.94 31.82 68.18 Spot 2 69.6 30.4 51.82 48.18 Spot 3 31.45 68.55 63.14 36.86 Mean 36.04 63.96 48.93 51.07 106 Figure B.1: (a) Hybrid electrolyte formed outside of electrochemical cell indicates solid-like behavior with (b) stratified layers of sus- pension and polymeric layer. SEM images of polymerized sep- arator at (c) 4 µm and (d) 20 µm scale. Suspension-laden sepa- rator is shown in (f) 4 µm and (g) 40 µm scale. (e) Postmortem image of Lithium metal for Li//NCM811 cell configuration with added Li2O and (h) Copper substrate for Cu//NCM811 cell configuration with Li2O included after more than 50 cycles. Figure B.2: (a) Particle size distribution measured at minute 1 to 15 with 1-minute interval and (b) the average size at each minute. 107 Figure B.3: Differential scanning calorimetry (DSC) curve of poly(DOL) layer near and away from Li2O particle settlement at 10 vol.% particle concentration and its corresponding (b) thermogravi- metric analysis (TGA) curve along with result for 30 vol.% Li2O. 108 Figure B.4: Temperature-dependent ionic conductivity of QSSE with vari- ous Li2O contents in (a) DOL electrolyte with 2 M LiTFSI and (b) ethylene carbonate (EC) with 2 M LiTFSI and 0.5 M LiNO3. Ionic conductivity at 30◦C and activation energy value Ea of (b) DOL electrolyte and (d) EC electrolyte. 109 Figure B.5: (a) Maxwell dependence of normalized conductivity on Li2O volume fractions at temperatures of 10 – 60◦C and (b) the col- lapsed universal plot with data points showing temperature- averaged normalized conductivity values. Figure B.6: (a) Strain-dependent loss and storage moduli G” and G’ and (b) stress-strain curves of Li2O suspensions of 10, 30, and 50 vol.%. caption 110 Figure B.7: Morphology of Li-metal deposited on Cu-foil for 10 mAh with electrolyte of (a, b) DOL with 2 M LiTFSI, (c,d) QSSE with 10 vol.% Li2O, (e,f) 30 vol.% Li2O, and (g,h) 50 vol.% Li2O. Scale bars of (a, c, e, g) show 100 m and (b, d, f, h) 10 µm. 111 Figure B.8: Electrochemical impedance spectroscopy (EIS) Nyquist plot of suspension electrolytes at volume fractions of 0 – 50 vol.%. 112 Figure B.9: Discharge capacity and CE of control cases in anode-free Cu//NCM811 configuration of (a) DOL + 2M LiTFSI. (b) Un- polymerized DOL + Li2O (open symbols) and polymerized poly(DOL) without Li2O (closed symbols) can only cycle at C/10 and are close to failure during SEI formation step. 113 Figure B.10: Corresponding galvanostatic stripping and plating for elec- trolyte utilized in Figure 3.4a but including Li-metal anode in a Li//NCM811 configuration instead of anode-free. (a) SEI formation step at C/10 with the first and fifth cycle shown for the hybrid electrolyte. (b) Post-SEI formation step, bat- tery was cycled at C/2 for 100 cycles, with the corresponding discharge capacity and Coulombic efficiency (CE) at all cycle shown in (c). Figure B.11: (a) Cyclic voltammogram of control liquid electrolyte in car- bon cloth and Cu-foil configuration, without any Li2O parti- cle addition like seen in Figure 3.4e. (b) Correlation of peak current to the square root of scan rate, extracted from Figure 3.4e, is seen to have a linear relationship. (c) At the highest scan rate of 5.0 mV/s, reduction and oxidation peak areas are seen to have the same value. 114 Figure B.12: (a) X-ray diffraction (XRD) analysis of pristine Li2O, pristine Li2O2, and Li2O on the Cu substrate after the third discharge of a Cu//Li2O cell. (b) Raman spectra of Li2O after cycling for more than 100 cycles, compared to spectra for pristine Li2O and Li2O2. The presence of Li2O2 is indentified by red stars above the peaks. 115 Figure B.13: Electrochemical performance for anode-free 10N5 in Cu//NCM811 configuration at different current rates of C/2, 1C (= 1 mA/cm2), 2C, and C/2. 116 Figure B.14: Galvanostatic stripping and plating of hybrid electrolyte con- taining poly(DOL) and (a) 10N7, (c) 50N5, (e) 50N7, and (b, d, f) their corresponding discharge capacity and Coulombic efficiency. 117 Figure B.15: Predicted capacity fade in 50N7 hybrid electrolyte, calculated from the experimental CE at different cycle numbers. This value is compared to the actual experimental value of capac- ity at each cycle. 118 Figure B.16: The effects of LiNO3 concentration on interfacial and bulk properties of hybrid electrolytes. (a) Nyquist plot obtained from EIS, (b) transference number measured through the Bruce-Vincent method compared to pure poly(DOL) (blue star) and pure QSSE (green triangle), and (c) ionic conduc- tivity at 30◦C. The role of LiNO3 as an electrolyte additive in LMBs and LIBs has been extensively studied and the main finding of these studies support its role in building a stable SEI. We studied the effect of LiNO3 concentration on the hy- brid electrolytes and observed higher interfacial resistance in- dicated by amplified circumference and an additional semi- circle in the Nyquist plot (Figure B.16). This second semicir- cle grows with increasing LiNO3 content. Increasing LiNO3 concentration from 0.5M (N5) to 0.7M (N7) does not seem to affect transference number and conductivity significantly. 119 Figure B.17: (a) Coulombic efficiency (CE) of suspension electrolytes made up of various SiO2 particles in DOL + 0.5 M LiNO3 + 1 M LiFSI. (b) CE presented as a function of SiO2 volume fraction. 120 APPENDIX C CHAPTER 4 APPENDIX [m] = Normalized area of monomer Normalized area of monomer + normalized area of polymer (C.1) For CR05M10 in Figure 4.1a of main text: [m] = 0.119 + 0.114 (0.119 + 0.114) + (1 + 0.488) = 13.5% (C.2) 121 Figure C.1: Ex-situ fabricated (a) CR1M10 and (b) CR05M10 bent to show its flexibility. The resulting material of CR1M10 is brittle with a high Tg, as indicated by its (c) DSC curve, compared to the flexible CR05M10 with no Tg detected down to -80°C. (d) DSC curve comparing poly(DOL) made with 100 mM Al(OTf)3 with and without 10% TMPTGE. The inset shows the crystalline poly(DOL) without 10% TMPTGE and amorphous one with TMPTGE. 122 Figure C.2: Temperature-dependent rheology measurement of CR05M at different TMPTGE concentrations of 5, 10, and 20%. 123 Figure C.3: 1HNMR of CR05M samples with 2%, 4%, and 6% TMPTGE. Monomer concentration was calculated (see Equation C.1) based on this result to obtain values shown in Figure 4.2. The unlabeled peaks belong to EC/DMC. 124 Figure C.4: 1HNMR of methyl ether poly(ethylene glycol) (mPEG) and DOL with ratios of 30:70 mPEG:DOL and 50:50 mPEG:DOL, indicating the ratio of integrals match the added concentra- tions. 125 Figure C.5: Room-temperature ionic conductivity of UN05M and CR05M10 compared to 1.0 M LiPF6 samples. 126 Figure C.6: (a) Frequency-dependent rheology measurement of CR05M at TMPTGE concentrations of 5, 10, 20, 30, and 40%. (b) Storage modulus G′ = Ge and mesh size of the network that can be calculated using the equation ξ = ( G′NA RT )− 1 3 . 127 Figure C.7: The first five cycles of galvanostatic stripping and plating of CR05M10 in a pouch cell with graphite/NCM811 configura- tion. The area of both electrodes is 9 cm2. Measurements em- ployed current density of 0.5 mA/cm2. 128 Figure C.8: Galvanostatic stripping and plating for 50 cycles of graphite/NCM811 of (a) CR05M5, (b) CR1M5, (d) CR05M10, (e) CR05M20, and (f) CR1M10. (c) Comparison of CE and discharge capacity for CR05M5 and CR1M5. Measurements employed current density of 1 mA/cm2. 129 Figure C.9: Linear sweep voltammetry (LSV) of UN05M and CR05M of various TMPTGE concentrations. Figure C.10: Stripping and plating lithium onto copper in a Li/Cu config- uration for (a) CR05M5 (b) CR05M20 and (c) their CE values at each cycle. CE was determined through the ratio of charg- ing time and 60-minute of discharging. 130 Figure C.11: The first five cycles of Galvanostatic charging and discharg- ing for polyDOL initiated by 1 mM Al(OTf)3, doped by 2 M LiTFSI, and crosslinked by 5% TMPTGE. 131 Figure C.12: Room-temperature ionic conductivity of polyDOL with var- ious TMPTGE concentrations. The pink points represent the polyDOL polymerized with 1 mM Al(OTf)3 with 2 M LiTFSI and green point with 10 mM Al(OTf)3 with 2 M LiTFSI. 132 BIBLIOGRAPHY [1] F. Abbasi, A.M. Agah, and E. Mehravar. Study on the effective pro- cess parameters influencing styrene and acrylonitrile grafting onto seeded polybutadiene latex. Journal of Applied Polymer Science, 119(3):1752–1761, 2011. [2] A. Aboulaich, R. Bouchet, G. Delaizir, V. Seznec, L. Tortet, M. Morcrette, P. Rozier, J.-M. Tarascon, V. Viallet, and M. Dollé. A new approach to de- velop safe all-inorganic monolithic li-ion batteries. Advanced Energy Mate- rials, 1(2):179–183, 2011. [3] P. Agarwal, S.A. Kim, and L.A. Archer. Crowded, confined, and frus- trated: Dynamics of molecules tethered to nanoparticles. Phys. Rev. Lett., 109:258301, Dec 2012. [4] P. Agarwal, H. Qi, and L.A. Archer. The ages in a self-suspended nanopar- ticle liquid. Nano Letters, 10(1):111–115, 2010. [5] P. Agarwal, S. Srivastava, and L.A. Archer. Thermal jamming of a colloidal glass. Phys. Rev. Lett., 107:268302, Dec 2011. [6] A. Agrawal, H.-Y. Yu, A. Sagar, S. Choudhury, and L.A. Archer. Molec- ular origins of temperature-induced jamming in self-suspended hairy nanoparticles. Macromolecules, 49(22):8738–8747, 2016. [7] R.C. Agrawal and G.P. Pandey. Solid polymer electrolytes: materials de- signing and all-solid-state battery applications: an overview. Journal of Physics D: Applied Physics, 41(22):223001, oct 2008. [8] S.J. An, J. Li, C. Daniel, D. Mohanty, S. Nagpure, and D.L. Wood. The state of understanding of the lithium-ion-battery graphite solid electrolyte interphase (sei) and its relationship to formation cycling. Carbon, 105:52– 76, 2016. [9] S. Ansari and E.P. Giannelis. Functionalized graphene sheet—poly(vinylidene fluoride) conductive nanocomposites. Jour- nal of Polymer Science Part B: Polymer Physics, 47(9):888–897, 2009. [10] H. Aono, E. Sugimoto, Y. Sadaoka, N. Imanaka, and G. Adachi. Ionic conductivity and sinterability of lithium titanium phosphate system. Solid State Ionics, 40-41:38–42, 1990. 133 [11] G.B. Appetecchi, Y. Aihara, and B. Scrosati. Investigation of swelling phe- nomena in peo-based polymer electrolytes: Ii. chemical and electrochem- ical characterization. Solid State Ionics, 170(1):63–72, 2004. [12] G.B. Appetecchi, G. Dautzenberg, and B. Scrosati. 143(1):6, jan 1996. [13] A.R. Arachchige, J. Tran, A. Michaelis, Z. Zhao, H. Dammann, E. Geis- brecht, and M.R. Beck. Exploring the impact of myopalladin ig3 domain mutations on actin interactions in cardiomyopathy. Biophysical Journal, 123(3):536A, 2024. [14] P.H.H. Araújo, C. Sayer, R. Giudici, and J.G.R. Poço. Techniques for reduc- ing residual monomer content in polymers: A review. Polymer Engineering & Science, 42(7):1442–1468, 2002. [15] E.A. Ayaz, R. Durkan, A. Koroglu, and B. Bagis. Comparative effect of different polymerization techniques on residual monomer and hardness properties of pmma-based denture resins. Journal of Applied Biomaterials & Functional Materials, 12(3):228–233, 2014. [16] A. Bac, M. Ciosek, M. Bukat, M. Marczewski, H. Marczewska, and W. Wieczorek. The effect of type of the inorganic filler and dopant salt concentration on the peo–liclo4 based composite electrolyte–lithium elec- trode interfacial resistivity. Journal of Power Sources, 159(1):405–411, 2006. [17] J.C. Bachman, S. Muy, A. Grimaud, H.-H. Chang, N. Pour, S.F. Lux, O. Paschos, F. Maglia, S. Lupart, P. Lamp, L. Giordano, and Y. Shao-Horn. Inorganic solid-state electrolytes for lithium batteries: Mechanisms and properties governing ion conduction. Chemical Reviews, 116(1):140–162, 2016. [18] A.J. Bard and L.R. Faulkner. Electrochemical Methods: Fundamentals and Applications. Wiley, 2001. [19] A. Batra, C. Cohen, and L.A. Archer. Stress relaxation of end-linked poly- dimethylsiloxane elastomers with long pendent chains. Macromolecules, 38(16):7174–7180, 2005. [20] A.J. Battacharyya, J. Maier, R. Bock, and F.F. Lange. New class of soft matter electrolytes obtained via heterogeneous doping: Percolation effects in “soggy sand” electrolytes. Solid State Ionics, 177(26):2565–2568, 2006. Solid State Ionics 15: Proceedings of the 15th International Conference on Solid State Ionics, Part II. 134 [21] A.J. Battacharyya, M. Patel, and S.K. Das. Soft matter lithium salt elec- trolytes: ion conduction and application to rechargeable batteries. Monat- shefte für Chemie - Chemical Monthly, 140:1001–1010, 2009. [22] M. Bednarek, P. Kubisa, and S. Penczek. Coexistence of activated monomer and active chain end mechanisms in cationic copolymerization of tetrahydrofuran with ethylene oxide. Macromolecules, 32(16):5257–5263, 1999. [23] C. Berthier. Microscopic investigation of ionic conductivity in alkali metal salts-poly(ethylene oxide) adducts. Solid State Ionics, 11(1):91–95, 1983. [24] G. Beyer. Flame retardancy of thermoplastic polyurethane and polyvinyl chloride by organoclays. Journal of Fire Sciences, 25(1):65–78, 2007. [25] A.J. Bhattacharyya. Ion transport in liquid salt solutions with oxide dis- persions: “soggy sand” electrolytes. The Journal of Physical Chemistry Let- ters, 3(6):744–750, 2012. [26] M.R. Bockstaller, R.A. Mickiewicz, and E.L. Thomas. Block copolymer nanocomposites: Perspectives for tailored functional materials. Advanced Materials, 17(11):1331–1349, 2005. [27] J. Brandrup and E.H. Immergut. Polymer Handbook. Second Edition. Wiley, 1975. [28] W. Braune and J. Okuda. An efficient method for controlled propylene ox- ide polymerization: The significance of bimetallic activation in aluminum lewis acids. Angewandte Chemie International Edition, 42(1):64–68. [29] P.G. Bruce, B.F.G. Johnson, and M. Schröder. Coordination chemistry in the solid state. Philosophical Transactions of the Royal Society of London. Series A: Mathematical, Physical and Engineering Sciences, 354(1706):415–436, 1996. [30] D.J. Brunelle. Ring-Opening Polymerization: Mechanisms, Catalysis, Struc- ture, Utility. Carl Hanser Verlag GmbH Co, 1992. [31] E. Cazzanelli, G. Mariotto, G.B. Appetecchi, F. Croce, and B. Scrosati. Study of ion-molecule interaction in poly(methylmethacrylate) based gel electrolytes by raman spectroscopy. Electrochimica Acta, 40(13):2379–2382, 1995. International symposium on polymer electrolytes. 135 [32] L. Cestarollo, S. Smolenski, and A. El-Ghazaly. Nanoparticle-based mag- netorheological elastomers with enhanced mechanical deflection for hap- tic displays. ACS Applied Materials and Interfaces, 14:19002–19011, 2022. [33] B.-Y. Chang and S.-M. Park. Electrochemical impedance spectroscopy. An- nual Review of Analytical Chemistry, 3(1):207–229, 2010. [34] W. Chassé, M. Lang, J.-U. Sommer, and K. Saalwächter. Correction to cross-link density estimation of pdms networks with precise considera- tion of networks defects. Macromolecules, 48(4):1267–1268, 2015. [35] O. Chaudhuri, S.H. Parekh, and D.A. Fletcher. Reversible stress softening of actin networks. Nature, 445:295–298, 2007. [36] X. Chen, W. He, L.-X. Ding, S. Wang, and H. Wang. Enhancing inter- facial contact in all solid state batteries with a cathode-supported solid electrolyte membrane framework. Energy Environ. Sci., 12:938–944, 2019. [37] X.-B. Cheng, R. Zhang, C.-Z. Zhao, and Q. Zhang. Toward safe lithium metal anode in rechargeable batteries: A review. Chemical Reviews, 117(15):10403–10473, 2017. [38] R. Chiba, T. Ishii, and F. Yoshimura. Temperature dependence of ionic conductivity in (1 − x)zro2(x − y)sc2o3 − yyb2o3 electrolyte material. Solid State Ionics, 91(3):249–256, 1996. [39] S. Chintapalli and R. Frech. Effect of plasticizers on high molecular weight peo-licf3so3 complexes. Solid State Ionics, 86-88:341–346, 1996. Proceed- ings of the 10th International Conference on Solid State Ionics. [40] S. Choudhury, A. Agrawal, S. Kim, and L.A. Archer. Self-suspended sus- pensions of covalently grafted hairy nanoparticles. Langmuir, 31(10):3222– 3231, 2015. [41] S. Choudhury, A. Agrawal, S. Wei, E. Jeng, and L.A. Archer. Hybrid hairy nanoparticle electrolytes stabilizing lithium metal batteries. Chemistry of Materials, 28(7):2147–2157, 2016. [42] S. Choudhury, S. Stalin, Y. Deng, and L.A. Archer. Soft colloidal glasses as solid-state electrolytes. Chemistry of Materials, 30(17):5996–6004, 2018. 136 [43] A. Chremos and J.F. Douglas. Particle localization and hyperunifor- mity of polymer-grafted nanoparticle materials. Annalen der Physik, 529(5):1600342, 2017. [44] A. Chremos and A.Z. Panagiotopoulos. Structural transitions of solvent- free oligomer-grafted nanoparticles. Phys. Rev. Lett., 107:105503, Sep 2011. [45] A. Chremos, A.Z. Panagiotopoulos, H.-Y. Yu, and D.L. Koch. Structure of solvent-free grafted nanoparticles: Molecular dynamics and density- functional theory. The Journal of Chemical Physics, 135(11):114901, 09 2011. [46] A. Chyzy, M. Tomczykowa, and M.E. Plonska-Brzezinska. Hydrogels as potential nano-, micro- and macro-scale systems for controlled drug de- livery. Materials, 13(1). [47] B. Commarieu, A. Paolella, J.-C. Daigle, and K. Zaghib. Toward high lithium conduction in solid polymer and polymer–ceramic batteries. Cur- rent Opinion in Electrochemistry, 9:56–63, 2018. [48] J.M.G. Cowie and K. Sadaghianizadeh. Effect of side chain length and crosslinking on the ac conductivity of oligo (ethyleneoxide) comb-branch polymer-salt mixtures. Solid State Ionics, 42(3):243–249, 1990. [49] J.D. Cox. A bond energy scheme—ii: Strain and conjugation energies in cyclic compounds. Tetrahedron, 19(7):1175–1184, 1963. [50] F. Croce, G.B. Appetecchi, L. Persi, and B. Scrosati. Nanocomposite poly- mer electrolytes for lithium batteries. Nature, 394:456–458, 1998. [51] F. Croce, L. Persi, F. Ronci, and B. Scrosati. Nanocomposite polymer elec- trolytes and their impact on the lithium battery technology. Solid State Ionics, 135(1):47–52, 2000. Proceedings of the 12th International Confer- ence on Solid State. [52] S.K. Das, S.S. Mandal, and A.J. Bhattacharyya. Ionic conductivity, me- chanical strength and li-ion battery performance of mono-functional and bi-functional (“janus”) “soggy sand” electrolytes. Energy Environ. Sci., 4:1391–1399, 2011. [53] H.-J. Deiseroth, J. Maier, K. Weichert, V. Nickel, S.-T. Kong, and C. Reiner. Li7ps6 and li6ps5x (x: Cl, br, i): Possible three-dimensional diffusion path- ways for lithium ions and temperature dependence of the ionic conductiv- 137 ity by impedance measurements. Zeitschrift für anorganische und allgemeine Chemie, 637(10):1287–1294, 2011. [54] R.L. Derbyshire. The lomor process: A solution for residual monomers? Radiation Physics and Chemistry, 14(3):333–342, 1979. [55] M. Djabourov, Lechaire J, P, and F. Gaill. Structure and rheology of gelatin and collagen gels. Biorheology, 30 3-4:191–205, 1993. [56] M. Doi and S.F. Edwards. The Theory of Polymer Dynamics. Clarendon Press, 1986. [57] T. Doi, Y. Iriyama, T. Abe, and Z. Ogumi. Lithium-ion transfer at an electrolyte/heat-treated nongraphitizable carbon electrode interface. Jour- nal of The Electrochemical Society, 152(8):A1521, jun 2005. [58] S.D. Druger, M.A. Ratner, and A. Nitzan. Polymeric solid electrolytes: Dynamic bond percolation and free volume models for diffusion. Solid State Ionics, 9-10:1115–1120, 1983. [59] A. Duda and A. Kowalski. Thermodynamics and Kinetics of Ring-Opening Polymerization, chapter 1, pages 1–51. John Wiley Sons, Ltd, 2009. [60] K. Edström, M. Herstedt, and D.P. Abraham. A new look at the solid elec- trolyte interphase on graphite anodes in li-ion batteries. Journal of Power Sources, 153(2):380–384, 2006. Selected papers presented at the 2004 Meet- ing of the International Battery Association. [61] T. Famprikis, P. Canepa, J.A. Dawson, M.S. Islam, and C. Masquelier. Lithium battery chemistries enabled by solid-state electrolytes. Nature Materials, 18:1278–1291, 2019. [62] L.-Z. Fan, Y.-S. Hu, A.J. Bhattacharyya, and J. Maier. Succinonitrile as a versatile additive for polymer electrolytes. Advanced Functional Materials, 17(15):2800–2807, 2007. [63] L.-Z. Fan, X.-L. Wang, F. Long, and X. Wang. Enhanced ionic conductivi- ties in composite polymer electrolytes by using succinonitrile as a plasti- cizer. Solid State Ionics, 179(27):1772–1775, 2008. Solid State Ionics 16: Pro- ceedings of the 16th International Conference on Solid State Ionics (SSI- 16), Part II. 138 [64] M. Faraday. Experimental Researches in Electricity. R.Taylor and W.Francis, 1849. [65] D.E. Fenton, J.M. Parker, and P.V. Wright. Complexes of alkali metal ions with poly(ethylene oxide). Polymer, 14(11):589, 1973. [66] N.J. Fernandes, H. Koerner, E.P. Giannelis, and R.A. Vaia. Hairy nanopar- ticle assemblies as one-component functional polymer nanocomposites: opportunities and challenges. MRS Communications, 3, 2013. [67] J.D. Ferry. Viscoelastic Properties of Polymers. John Wiley Sons, Inc., 1961. [68] J.D. Ferry. Viscoelastic Properties of Polymers. Wiley, 1980. [69] H. L. Finke, D. W. Scott, M. E. Gross, J. F. Messerly, and G. Waddington. Cycloheptane, cycloöctane and 1,3,5-cycloheptatriene. low temperature thermal properties, vapor pressure and derived chemical thermodynamic properties. Journal of the American Chemical Society, 78(21):5469–5476, 1956. [70] C. Flesch, C. Delaite, P. Dumas, E. Bourgeat-Lami, and E. Duguet. Graft- ing of poly(-caprolactone) onto maghemite nanoparticles. Journal of Poly- mer Science Part A: Polymer Chemistry, 42(23):6011–6020, 2004. [71] X. Fu, D. Yu, J. Zhou, S. Li, X. Gao, Y. Han, P. Qi, X. Feng, and B. Wang. Inorganic and organic hybrid solid electrolytes for lithium-ion batteries. CrystEngComm, 18:4236–4258, 2016. [72] G.S. Fulcher. Analysis of recent measurements of the viscosity of glasses. Journal of the American Ceramic Society, 8(6):339–355, 1925. [73] R. Garcia-Mendez, J.G. Smith, J.C. Neuefeind, D.J. Siegel, and J. Sakamoto. Correlating macro and atomic structure with elastic properties and ionic transport of glassy li2s-p2s5 (lps) solid electrolyte for solid-state li metal batteries. Advanced Energy Materials, 10(19):2000335, 2020. [74] Z.A. Ghazi, Z. Sun, C. Sun, F. Qi, B. An, F. Li, and H.-M. Cheng. Key aspects of lithium metal anodes for lithium metal batteries. Small, 15(32):1900687, 2019. [75] O. Glatter and O. Kratky. Small angle x-ray scattering. Academic Press, Inc., 1982. 139 [76] A. Goto and T. Fukuda. Effects of radical initiator on polymerization rate and polydispersity in nitroxide-controlled free radical polymeriza- tion. Macromolecules, 30(15):4272–4277, 1997. [77] M.S. Grewal, M. Tanaka, and H. Kawakami. Free-standing polydimethylsiloxane-based cross-linked network solid polymer elec- trolytes for future lithium ion battery applications. Electrochimica Acta, 307:148–156, 2019. [78] S.-Y. Gu, X.-F. Gao, and Y.-H. Zhang. Synthesis and characterization of solvent-free ionic molybdenum disulphide (mos2) nanofluids. Materials Chemistry and Physics, 149-150:587–593, 2015. [79] J. Guo, Z. Wen, M. Wu, J. Jin, and Y. Liu. Vinylene carbonate–lino3: A hy- brid additive in carbonic ester electrolytes for sei modification on li metal anode. Electrochemistry Communications, 51:59–63, 2015. [80] S. Hanosh, B.S. Lekshmi, S.N. Varanakkottu, and S.D. George. Simple fab- rication of pdms-grafted hydrophobic iron oxide particles for oil–water separation and stimuli-responsive liquid marbles. ACS Applied Polymer Materials, 5(5):3507–3514, 2023. [81] A. Harrison and R. Huggett. Effect of the curing cycle on residual monomer levels of acrylic resin denture base polymers. Journal of Den- tistry, 20(6):370–374, 1992. [82] K.B. Hatzell, X.C. Chen, C.L. Cobb, N.P. Dasgupta, M.B. Dixit, L.E. Mar- bella, M.T. McDowell, P.P. Mukherjee, A. Verma, V. Viswanathan, A.S. Westover, and W.G. Zeier. Challenges in lithium metal anodes for solid- state batteries. ACS Energy Letters, 5(3):922–934, 2020. [83] P. He, T. Zhang, J. Jiang, and H. Zhou. Lithium–air batteries with hybrid electrolytes. The Journal of Physical Chemistry Letters, 7(7):1267–1280, 2016. [84] E.G. Herbert, W.E. Tenhaeff, N.J. Dudney, and G.M. Pharr. Mechanical characterization of lipon films using nanoindentation. Thin Solid Films, 520(1):413–418, 2011. [85] S. Hess, M. Wohlfahrt-Mehrens, and M. Wachtler. Flammability of li- ion battery electrolytes: Flash point and self-extinguishing time measure- ments. Journal of The Electrochemical Society, 162(2):A3084, 2015. 140 [86] P.C. Hiemenz and T.P. Lodge. Polymer Chemistry. CRC Press, 2007. [87] Z. Hou, Y. Liu, J. Xu, and J. Zhu. Surface engineering of magnetic iron ox- ide nanoparticles by polymer grafting: synthesis progress and biomedical applications. Nanoscale, 12:14957–14975, 2020. [88] S. Humbert, O. Lame, R. Séguéla, and G. Vigier. A re-examination of the elastic modulus dependence on crystallinity in semi-crystalline polymers. Polymer, 52(21):4899–4909, 2011. [89] Y. Inaguma, C. Liquan, M. Itoh, T. Nakamura, T. Uchida, H. Ikuta, and M. Wakihara. High ionic conductivity in lithium lanthanum titanate. Solid State Communications, 86(10):689–693, 1993. [90] K. Ito. Initiator concentration dependence of the autoacceleration of poly- merization rate. Journal of Polymer Science: Polymer Chemistry Edition, 13(2):401–413, 1975. [91] H.M. Jaeger and S.R. Nagel. Physics of the granular state. Science, 255(5051):1523–1531, 1992. [92] A.M. Jimenez, A.A. Krauskopf, R.A. Pérez-Camargo, D. Zhao, J. Pribyl, J. Jestin, B.C. Benicewicz, A.J. Müller, and S.K. Kumar. Effects of hairy nanoparticles on polymer crystallization kinetics. Macromolecules, 52(23):9186–9198, 2019. [93] J.M.Tarascon and M. Armand. Issues and challenges facing rechargeable lithium batteries. Nature, 414:359–367, 2001. [94] H.M. Kasper. Series of rare earth garnets ln3+3m2li+3o12 (m=te, w). In- organic Chemistry, 8(4):1000–1002, 1969. [95] E. Kazyak, R. Garcia-Mendez, W.S. LePage, A. Sharafi, A.L. Davis, A.J. Sanchez, K.-H. Chen, C. Haslam, J. Sakamoto, and N.P. Dasgupta. Li pen- etration in ceramic solid electrolytes: Operando microscopy analysis of morphology, propagation, and reversibility. Matter, 2(4):1025–1048, 2020. [96] I. Kelly, J.R. Owen, and B.C.H. Steele. Mixed polyether lithium-ion con- ductors. Journal of Electroanalytical Chemistry and Interfacial Electrochem- istry, 168(1):467–478, 1984. 141 [97] J.H. Kennedy, S. Sahami, S.W. Shea, and Z. Zhang. Preparation and con- ductivity measurements of sis2li2s glasses doped with libr and licl. Solid State Ionics, 18-19:368–371, 1986. [98] K. Khan, Z. Tu, Q. Zhao, C. Zhao, and L.A. Archer. Synthesis and prop- erties of poly-ether/ethylene carbonate electrolytes with high oxidative stability. Chemistry of Materials, 31(20):8466–8472, 2019. [99] M.S. Kim, Z. Zhang, P.E. Rudnicki, Z. Yu, J. Wang, H. Wang, S.T. Oy- akhire, Y. Chen, S.C. Kim, W. Zhang, D.T. Boyle, X. Kong, R. Xu, Z. Huang, W. Huang, S.F. Bent, L.-W. Wang, J. Qin, Z. Bao, and Y. Cui. Suspension electrolyte with modified li+ solvation environment for lithium metal bat- teries. Nature Materials, 21:445–454, 2022. [100] M.S. Kim, Z. Zhang, J. Wang, S.T. Oyakhire, andS.C. Kim, Z. Yu, Y. Chen, D.T. Boyle, Y. Ye, Z. Huang, W. Zhang, R. Xu, P. Sayavong, S.F. Bent, J. Qin, Z. Bao, and Y. Cui. Revealing the multifunctions of li3n in the suspension electrolyte for lithium metal batteries. ACS Nano, 17(3):3168–3180, 2023. [101] S. Kim, R. Mangal, and L.A. Archer. Relaxation dynamics of nanoparticle- tethered polymer chains. Macromolecules, 48(17):6280–6293, 2015. [102] S.A. Kim and L.A. Archer. Hierarchical structure in semicrystalline poly- mers tethered to nanospheres. Macromolecules, 47(2):687–694, 2014. [103] Y.-G. Kim, M. Wagner, and H. Thérien-Aubin. Dynamics of soft and hairy polymer nanoparticles in a suspension by nmr relaxation. Macromolecules, 53(3):844–851, 2020. [104] A. Kisliuk, V. Bocharova, I. Popov, C. Gainaru, and A.P. Sokolov. Fun- damental parameters governing ion conductivity in polymer electrolytes. Electrochimica Acta, 299:191–196, 2019. [105] F. Klaeger, A.S. Tagg, S. Otto, M. Bienmüller, I. Sartorius, and M. Labrenz. Residual monomer content affects the interpretation of plastic degrada- tion. Scientific Reports, 9(2120), 2019. [106] H. Koerner, W. Liu, M. Alexander, P. Mirau, H. Dowty, and R.A. Vaia. Deformation–morphology correlations in electrically conductive carbon nanotube—thermoplastic polyurethane nanocomposites. Polymer, 46(12):4405–4420, 2005. In Honor of James E. Mark. 142 [107] H. Koerner, G. Price, N.A. Pearce, M. Alexander, and R.A. Vaia. Remotely actuated polymer nanocomposites—stress-recovery of carbon-nanotube- filled thermoplastic elastomers. Nature Materials, 3:115–120, 2004. [108] S. Krajangpan, H. Kalita, B.J. Chisholm, and A.N. Bezbaruah. Iron nanoparticles coated with amphiphilic polysiloxane graft copolymers: Dispersibility and contaminant treatability. Environmental Science & Tech- nology, 46(18):10130–10136, 2012. [109] R. Krishnamoorti. Strategies for dispersing nanoparticles in polymers. MRS Bulletin, 32:341–347, 2007. [110] R. Krishnamoorti, R.A. Vaia, and E.P. Giannelis. Structure and dynam- ics of polymer-layered silicate nanocomposites. Chemistry of Materials, 8(8):1728–1734, 1996. [111] B. Kumar and L.G. Scanlon. Polymer–ceramic composite electrolytes: conductivity and thermal history effects. Solid State Ionics, 124(3):239–254, 1999. [112] S.K. Kundu, M. Yoshida, and M. Shibayama. Effect of salt content on the rheological properties of hydrogel based on oligomeric electrolyte. The Jounal of Physical Chemistry B, 114(4):1541–1547, 2010. [113] L.J.T. Landherr, C. Cohen, and L.A. Archer. Effect of pendent chains on the interfacial properties of thin polydimethylsiloxane (pdms) networks. Langmuir, 27(10):5944–5952, 2011. PMID: 21495649. [114] J.-M. Lehn. Supramolecular Chemistry: Concepts and Perspectives. Wiley, 1995. [115] Q. Li, H.Y. Sun, Y. Takeda, N. Imanishi, J. Yang, and O. Yamamoto. In- terface properties between a lithium metal electrode and a poly(ethylene oxide) based composite polymer electrolyte. Journal of Power Sources, 94(2):201–205, 2001. Special Issue: Interfacial Phenomena in Batteries. [116] S. Li, S.-Q. Zhang, L. Shen, Q. Liu, J.-B. Ma, W. Lv, Y.-B. He, and Q.-H. Yang. Progress and perspective of ceramic/polymer composite solid elec- trolytes for lithium batteries. Advanced Science, 7(5):1903088, 2020. [117] T. Li, A.J. Senesi, and B. Lee. Small angle x-ray scattering for nanoparticle research. Chemical Reviews, 116(18):11128–11180, 2016. 143 [118] W. Li, H. Yao, K. Yan, G. Zheng, Z. Liang, Y.-M. Chiang, and Y. Cui. The synergetic effect of lithium polysulfide and lithium nitrate to prevent lithium dendrite growth. Nature Communications, 6:7436, 2015. [119] D. Lin, Y. Liu, and Y. Cui. Reviving the lithium metal anode for high- energy batteries. Nature Nanotechnology, 12:194–206, 2017. [120] Y.S. Lipatov, L.V. Kosyanchuk, A.E. Nesterov, and O.I. Antonenko. Filler effect on polymerization kinetics and phase separation in polymer blends formed in situ. Polymer International, 52(5):664–669, 2003. [121] B. Liu, J.-G. Zhang, and W. Xu. Advancing lithium metal batteries. Joule, 2(5):833–845, 2018. [122] J. Liu, Z. Bao, Y. Cui, E.J. Dufek, J.B. Goodenough, P. Khalifah, Q. Li, B.Y. Liaw, P. Liu, A. Manthiram, Y.S. Meng, V.R. Subramanian, M.F. Toney, V.V. Viswanathan, M.S. Whittingham, J. Xiao, W. Xu, J. Yang, X.-Q. Yang, and J.-G. Zhang. Pathways for practical high-energy long-cycling lithium metal batteries. Nature Energy, 4:180–186, 2019. [123] T. Liu, J. Zhang, W. Han, J. Zhang, G. Ding, S. Dong, and G. Cui. Re- view—in situ polymerization for integration and interfacial protection to- wards solid state lithium batteries. Journal of The Electrochemical Society, 167(7):070527, feb 2020. [124] T. Liu, S. Zhao, Q. Xiong, J. Yu, J. Wang, G. Huang, M. Ni, and X. Zhang. Reversible discharge products in li–air batteries. Advanced Materials, 35(20):2208925, 2023. [125] X. Liu, B.A. Abel, Q. Zhao, S. Li, S. Choudhury, J. Zheng, and L.A. Archer. Microscopic origins of caging and equilibration of self-suspended hairy nanoparticles. Macromolecules, 52(21):8187–8196, 2019. [126] X. Liu, N.W. Utomo, Q. Zhao, J. Zheng, D. Zhang, and L.A. Archer. Effects of geometric confinement on caging and dynamics of polymer-tethered nanoparticle suspensions. Macromolecules, 54(1):426–439, 2021. [127] D.R. MacFarlane, P.J. Newman, K.M. Nairn, and M. Forsyth. Lithium- ion conducting ceramic/polyether composites. Electrochimica Acta, 43(10):1333–1337, 1998. 144 [128] J. Maier. Ionic conduction in space charge regions. Progress in Solid State Chemistry, 23(3):171–263, 1995. [129] R. Mangal, P. Nath, M. Tikekar, and L.A. Archer. Enthalpy-driven stabilization of dispersions of polymer-grafted nanoparticles in high- molecular-weight polymer melts. Langmuir, 32(41):10621–10631, 2016. [130] R. Mangal, S. Srivastava, and L.A. Archer. Phase stability and dynamics of entangled polymer–nanoparticle composites. Nature Communications, 6(1), 2015. [131] A. Manthiram, X. Yu, and S. Wang. Lithium battery chemistries enabled by solid-state electrolytes. Nature Review Materials, 2(16103), 2017. [132] K. Matyjaszewski. Cationic Polymerizations Mechanisms, Synthesis, and Ap- plications. Marcel Dekker, Inc., 1996. [133] J.C. Maxwell. Treatise on Electricity and Magnetism. Dover Publications, 1873. [134] Q. Meng, M.-C. Heuzey, and P.J. Carreau. Control of thermal degradation of polylactide/clay nanocomposites during melt processing by chain ex- tension reaction. Polymer Degradation and Stability, 97(10):2010–2020, 2012. 3rd International Conference on Biodegradable and Biobased Polymers (BIOPOL-2011) - Strasbourg 2011. [135] S.S. Moganty, N. Jayaprakash, J.L. Nugent, J. Shen, and L.A. Archer. Ionic- liquid-tethered nanoparticles: Hybrid electrolytes. Angewandte Chemie In- ternational Edition, 49(48):9158–9161, 2010. [136] S. Monfette and D.E. Fogg. Equilibrium ring-closing metathesis. Chemical Reviews, 109(8):3783–3816, 2009. [137] S.L. Mooberry. Microtubules as a target for anticancer drugs. Clinical Advances in Hematology Oncology, 9, 2011. [138] K. Morigaki and A. Ohta. Analysis of the surface of lithium in organic electrolyte by atomic force microscopy, fourier transform infrared spec- troscopy and scanning auger electron microscopy. Journal of Power Sources, 76(2):159–166, 1998. 145 [139] M.Z.A. Munshi and B.B. Owens. Performance of polymer electrolyte cells at +25 to +100°c. Solid State Ionics, 26(1):41–46, 1988. [140] C.W. Murray and M.L. Verdonk. The consequences of translational and rotational entropy lost by small molecules on binding to proteins. Journal of Computer-Aided Molecular Design, 16:741–753, 2002. [141] N.-Q. Nguyen and A.J.C. Ladd. Microstructure in a settling suspension of hard spheres. Phys. Rev. E, 69:050401, May 2004. [142] N.-Q. Nguyen and A.J.C. Ladd. Sedimentation of hard-sphere suspen- sions at low reynolds number. Journal of Fluid Mechanics, 525:73–104, 2005. [143] L.E. Nielsen. Cross-linking–effect on physical properties of polymers. Journal of Macromolecular Science, Part C, 3(1):69–103, 1969. [144] O. Nuyken and S.D. Pask. Ring-opening polymerization—an introduc- tory review. Polymers, 5(2):361–403, 2013. [145] K. Ohno, T. Sakamoto, T. Minagawa, and Y. Okabe. Entropy of polymer brushes in good solvents: A monte carlo study. Macromolecules, 40(3):723– 730, 2007. [146] M. Okada, Y. Yamashita, and Y. Ishii. Polymerization of 1,3-dioxolane. Die Makromolekulare Chemie, 80(1):196–207, 1964. [147] O. Olabisi, L.M. Robseon, and M.T. Shaw. Polymer-Polymer Miscibility. Academic Press, Inc., 1979. [148] S.-K. Otto, Y. Moryson, T. Krauskopf, K. Peppler, J. Sann, J. Janek, and A. Henss. In-depth characterization of lithium-metal surfaces with xps and tof-sims: Toward better understanding of the passivation layer. Chem- istry of Materials, 33(3):859–867, 2021. [149] P.C. Painter and M.M. Coleman. Essentials of Polymer Science and Engineer- ing. DEStech Publications, Inc., 2009. [150] Q. Pan, D. Barbash, D.M. Smith, H. Qi, S.E. Gleeson, and C.Y. Li. Cor- relating electrode–electrolyte interface and battery performance in hybrid solid polymer electrolyte-based lithium metal batteries. Advanced Energy Materials, 7(22):1701231. 146 [151] E. Peled. The electrochemical behavior of alkali and alkaline earth metals in nonaqueous battery systems—the solid electrolyte interphase model. Journal of The Electrochemical Society, 126(12):2047, dec 1979. [152] E. Peled, D. Bar Tow, A. Merson, A. Gladkich, L. Burstein, and D. Golod- nitsky. Composition, depth profiles and lateral distribution of materials in the sei built on hopg-tof sims and xps studies. Journal of Power Sources, 97- 98:52–57, 2001. Proceedings of the 10th International Meeting on Lithium Batteries. [153] M. Potoczek. A catalytic effect of alumina grains onto polymerization rate of methacrylamide-based gelcasting system. Ceramics International, 32(7):739–744, 2006. [154] J. Przyluski, M. Siekierski, and W. Wieczorek. Effective medium theory in studies of conductivity of composite polymeric electrolytes. Electrochimica Acta, 40(13):2101–2108, 1995. International symposium on polymer elec- trolytes. [155] J. Płcharski and W. Weiczorek. Peo based composite solid electrolyte con- taining nasicon. Solid State Ionics, 28-30:979–982, 1988. [156] J. Qian, B.D. Adams, J. Zheng, W. Xu, W.A. Henderson, J. Wang, M.E. Bowden, S. Xu, J. Hu, and J.-G. Zhang. Anode-free rechargeable lithium metal batteries. Advanced Functional Materials, 26(39):7094–7102, 2016. [157] Y. Qiao, K. Jiang, H. Deng, and H. Zhou. A high-energy-density and long- life lithium-ion battery via reversible oxide–peroxide conversion. Nature Catalysis, 2:1035–1044, 2019. [158] M. Rubinstein and R.H. Colby. Polymer Physics. Oxford University Press, 2003. [159] M. Sadej and E. Andrzejewska. Silica/aluminum oxide hybrid as a filler for photocurable composites. Progress in Organic Coatings, 94:1–8, 2016. [160] J.L. Schaefer, S.S. Moganty, D.A. Yanga, and L.A. Archer. Nanoporous hybrid electrolytes. J. Mater. Chem., 21:10094–10101, 2011. [161] J.L. Schaefer, D.A. Yanga, and L.A. Archer. High lithium transference number electrolytes via creation of 3-dimensional, charged, nanoporous 147 networks from dense functionalized nanoparticle composites. Chemistry of Materials, 25(6):834–839, 2013. [162] D.F. Schmidt, F. Clément, and E.P. Giannelis. On the origins of silicate dis- persion in polysiloxane/layered-silicate nanocomposites. Advanced Func- tional Materials, 16(3):417–425, 2006. [163] G.S. Sheridan and C.M. Evans. Understanding the roles of mesh size, tg, and segmental dynamics on probe diffusion in dense polymer networks. Macromolecules, 54(23):11198–11208, 2021. [164] J.S. Shim and D.C. Watts. Residual monomer concentrations in denture- base acrylic resin after an additional, soft-liner, heat-cure cycle. Dental Materials, 15(4):296–300, 1999. [165] A. Shokuhfar and B. Arab. The effect of cross-linking density on the me- chanical properties and structure of the epoxy polymers: molecular dy- namics simulation. Journal of Molecular Modeling, 19:3719–3731, 2013. [166] N. Singh, C. Galande, A. Miranda, A. Mathkar, W. Gao, A.L.M Reddy, A. Vlad, and P.M. Ajayan. Paintable battery. Scientific Reports, 2(1):481, 2012. [167] Z. Song, F. Chen, M. Martinez-Ibañez, W. Feng, M. Forsyth, Z. Zhou, M. Armand, and H. Zhang. Fa reflection on polymer electrolytes for solid- state lithium metal batteries. Nature Communications, 14:4884, 2023. [168] S. Srivastava, P. Agarwal, and L.A. Archer. Tethered nanoparti- cle–polymer composites: Phase stability and curvature. Langmuir, 28(15):6276–6281, 2012. [169] S. Srivastava, J.H. Shin, and L.A. Archer. Structure and rheology of nanoparticle–polymer suspensions. Soft Matter, 8:4097–4108, 2012. [170] Z. Stoeva, C.T. Imrie, and M.D. Ingram. Effect of pressure on ion trans- port in amorphous and semi-crystalline polymer electrolytes. Phys. Chem. Chem. Phys., 5:395–399, 200. [171] G.R. Strobl. The Physics of Polymers. Springer, 1997. [172] M. Sukeshini, A. Nishimoto, and M. Watanabe. Transport and electro- chemical characterization of plasticized poly(vinyl chloride) solid elec- 148 trolytes. Solid State Ionics, 86-88:385–393, 1996. Proceedings of the 10th International Conference on Solid State Ionics. [173] L. Suo, D. Oh, Y. Lin, Z. Zhuo, O. Borodin, T. Gao, F. Wang, A. Kushima, Z. Wang, H.-C. Kim, Y. Qi, W. Yang, F. Pan, J. Li, K. Xu, and C. Wang. How solid-electrolyte interphase forms in aqueous electrolytes. Journal of the American Chemical Society, 139(51):18670–18680, 2017. [174] S. Sylla, Y. Sanchez, and M. Armand. Electrochemical study of lin- ear and crosslinked poe-based polymer electrolytes. Electrochimica Acta, 37(9):1699–1701, 1992. [175] K. Takada. Progress and prospective of solid-state lithium batteries. Acta Materialia, 61(3):759–770, 2013. The Diamond Jubilee Issue. [176] T. Takahashi. High Conductivity Solid Ionic Conductors: Recent Trends and Applications. World Scientific Publishing Company, 1989. [177] G. Tammann and W. Hesse. Die abhängigkeit der viscosität von der tem- peratur bie unterkühlten flüssigkeiten. Zeitschrift für anorganische und all- gemeine Chemie, 156(1):245–257, 1926. [178] A.A. Teran, S.A. Mullin, D.T. Hallinan, and N.P. Balsara. Discontinuous changes in ionic conductivity of a block copolymer electrolyte through an order–disorder transition. ACS Macro Letters, 1(2):305–309, 2012. [179] Y. Tian, Y. An, C. Wei, H. Jiang, S. Xiong, J. Feng, and Y. Qian. Recently advances and perspectives of anode-free rechargeable batteries. Nano En- ergy, 78:105344, 2020. [180] B. Tidor and M. Karplus. The contribution of vibrational entropy to molec- ular association: The dimerization of insulin. Journal of Molecular Biology, 238(3):405–414, 1994. [181] Y. Tominaga, K. Yamazaki, and V. Nanthana. Effect of anions on lithium ion conduction in poly(ethylene carbonate)-based polymer electrolytes. Journal of The Electrochemical Society, 162(2):A3133, 2015. [182] Z. Tong, B. Bazri, S.-F. Hu, and R.-S. Liu. Interfacial chemistry in anode- free batteries: challenges and strategies. J. Mater. Chem. A, 9:7396–7406, 2021. 149 [183] A. Usuki, Y. Kojima, M. Kawasumi, A. Okada, Y. Fukushima, T. Kurauchi, and O. Kamigaito. Synthesis of nylon 6-clay hybrid. Journal of Materials Research, 8(5):1179–1184, 1993. [184] N.W. Utomo, Y. Deng, Q. Zhao, X. Liu, and L.A. Archer. Structure and evolution of quasi-solid-state hybrid electrolytes formed inside electro- chemical cells. Advanced Materials, 34(32):2110333, 2022. [185] C. Vasile. Handbook of Polyolefins. CRC Press, 2000. [186] P. Verma, P. Maire, and P. Novak. A review of the features and analyses of the solid electrolyte interphase in li-ion batteries. Electrochimica Acta, 55(22):6332–6341, 2010. [187] D.H. Vogel. Das temperaturabhaengigkeitsgesetz der viskositaet von fluessigkeiten. Physikalische Zeitschrift, 22:645, 1921. [188] M. Wang, H.H. Winter, and G.K. Auernhammer. Time and frequency de- pendent rheology of reactive silica gels. Journal of Colloid and Interface Sci- ence, 413:159–166, 2014. [189] Y. Wang, X. Teng, J.-S. Wang, and H. Yang. Solvent-free atom transfer radical polymerization in the synthesis of fe2o3@polystyrene coreshell nanoparticles. Nano Letters, 3(6):789–793, 2003. [190] Y. Wang and W.-H. Zhong. Development of electrolytes towards achiev- ing safe and high-performance energy-storage devices: A review. Chem- ElectroChem, 2(1):22–36, 2015. [191] Z. Wang, B. Huang, H. Huang, L. Chen, R. Xue, and F. Wang. Investigation of the position of li+ ions in a polyacrylonitrile-based electrolyte by raman and infrared spectroscopy. Electrochimica Acta, 41(9):1443–1446, 1996. [192] S.C. Warren, M.J. Banholzer, L.S. Slaughter, E.P. Giannelis, F.J. DiSalvo, and U.B. Wiesner. Generalized route to metal nanoparticles with liquid behavior. Journal of the American Chemical Society, 128(37):12074–12075, 2006. [193] M. Watanabe, S. Nagano, K. Sanui, and N. Ogata. Ionic conductivity of network polymers from poly(ethylene oxide) containing lithium perchlo- rate. Polymer Journal, 18(11):809–817, 1986. 150 [194] S. Wei, K.E. Hendrickson L. Ma, Z. Tu, and L.A. Archer. Metal–sulfur battery cathodes based on pan–sulfur composites. Journal of the American Chemical Society, 137(37):12143–12152, 2015. [195] R.E. Wetton and G. Allen. The dynamic mechanical properties of some polyethers. Polymer, 7(7):331–365, 1966. [196] W. Wieczorek, D. Raducha, A. Zalewska, and J.R. Stevens. Effect of salt concentration on the conductivity of peo-based composite polymeric elec- trolytes. The Journal of Physical Chemistry B, 102(44):8725–8731, 1998. [197] K.I. Winey, T. Kashiwagi, and M. Mu. Improving electrical conductivity and thermal properties of polymers by the addition of carbon nanotubes as fillers. MRS Bulletin, 32:348–353, 2007. [198] M.A. Wisniewska, J.G. Seland, and W. Wang. Determining the scaling of gel mesh size with changing crosslinker concentration using dynamic swelling, rheometry, and pgse nmr spectroscopy. Journal of Applied Poly- mer Science, 135(45):46695, 2018. [199] R.S.H. Wong, M. Ashton, and K. Dodou. Effect of crosslinking agent con- centration on the properties of unmedicated hydrogels. Pharmaceutics, 7:305–319, 2015. [200] K.N. Wood and G. Teeter. Xps on li-battery-related compounds: Analysis of inorganic sei phases and a methodology for charge correction. ACS Applied Energy Materials, 1(9):4493–4504, 2018. [201] H. Wu, Y. Cao, L. Geng, and C. Wang. In situ formation of stable interfacial coating for high performance lithium metal anodes. Chemistry of Materials, 29(8):3572–3579, 2017. [202] N. Wu, P.-H. Chien, Y. Li, A. Dolocan, H. Xu, B. Xu, N.S. Grundish, H. Jin, Y.-Y. Hu, and J.B. Goodenough. Fast li+ conduction mechanism and in- terfacial chemistry of a nasicon/polymer composite electrolyte. Journal of the American Chemical Society, 142(5):2497–2505, 2020. [203] W. Wu. Baseline studies of the clay minerals society source clays: Colloid and surface phenomena. Clays and Clay Minerals, 49:446–452, 2001. [204] M. Xie, Y. Wu, Y. Liu, P.P. Yu, R. Jia, W.A. Goddard, and T. Cheng. Pathway 151 of in situ polymerization of 1,3-dioxolane in lipf6 electrolyte on li metal anode. Materials Today Energy, 21:100730, 2021. [205] S. Xiong, K. Xie, Y. Diao, and X. Hong. Properties of surface film on lithium anode with lino3 as lithium salt in electrolyte solution for lithium–sulfur batteries. Electrochimica Acta, 83:78–86, 2012. [206] C. Yan, Y.-X. Yao, X. Chen, X.-B. Cheng, X.-Q. Zhang, J.-Q. Huang, and Q. Zhang. Lithium nitrate solvation chemistry in carbonate electrolyte sustains high-voltage lithium metal batteries. Angewandte Chemie Interna- tional Edition, 57(43):14055–14059, 2018. [207] Y. Yan, R.-S. Kühnel, A. Remhof, L. Duchêne, E.C. Reyes, D. Rentsch, Z. Łodziana, and C Battaglia. A lithium amide-borohydride solid-state electrolyte with lithium-ion conductivities comparable to liquid elec- trolytes. Advanced Energy Materials, 7(19):1700294, 2017. [208] Y. Yang and L.J. Kaufman. Rheology and confocal reflectance microscopy as probes of mechanical properties and structure during collagen and col- lagen/hyaluronan self-assembly. Biophysics Journal, 96(4):1566–1585, 2009. [209] C. Yu, Y. Li, M. Willans, Y. Zhao, K.R. Adair, F. Zhao, W. Li, S. Deng, J. Liang, M.N. Banis, R. Li, H. Huang, L. Zhang, R. Yang, S. Lu, Y. Huang, and X. Sun. Superionic conductivity in lithium argyrodite solid-state elec- trolyte by controlled cl-doping. Nano Energy, 69:104396, 2020. [210] G.-E. Yu, F. Heatley, C. Booth, and T.G. Blease. Anionic polymerization of propylene oxide: Isomerization of allyl ether to propenyl ether end groups. Journal of Polymer Science Part A: Polymer Chemistry, 32(6):1131– 1135. [211] H.-Y. Yu and D.L. Koch. Structure of solvent-free nanoparticleorganic hy- brid materials. Langmuir, 26(22):16801–16811, 2010. [212] H.-Y. Yu, S. Srivastava, L.A. Archer, and D.L. Koch. Structure factor of blends of solvent-free nanoparticle–organic hybrid materials: density- functional theory and small angle x-ray scattering. Soft Matter, 10:9120– 9135, 2014. [213] X. Yu and A. Manthiram. Electrode–electrolyte interfaces in lithium–sulfur batteries with liquid or inorganic solid electrolytes. Ac- counts of Chemical Research, 50(11):2653–2660, 2017. 152 [214] X. Yu and A. Manthiram. A review of composite polymer-ceramic elec- trolytes for lithium batteries. Energy Storage Materials, 34:282–300, 2021. [215] H. Zhang, C. Li, M. Piszcz, E. Coya, T. Rojo, L.M. Rodriguez-Martinez, M. Armand, and Z. Zhou. Single lithium-ion conducting solid polymer electrolytes: advances and perspectives. Chem. Soc. Rev., 46:797–815, 2017. [216] J.-N. Zhang, Q. Li, Y. Wang, J. Zheng, X. Yu, and H. Li. Dynamic evolution of cathode electrolyte interphase (cei) on high voltage licoo2 cathode and its interaction with li anode. Energy Storage Materials, 14:1–7, 2018. [217] S.S. Zhang. Role of lino3 in rechargeable lithium/sulfur battery. Elec- trochimica Acta, 70:344–348, 2012. [218] Q. Zhao, X. Liu, S. Stalin, and L.A. Archer. In-built polymer-in-solvent and solvent-in-polymer electrolytes for high-voltage lithium metal batteries. Cell Reports Physical Science, 1(8):100146, 2020. [219] Q. Zhao, X. Liu, S. Stalin, K. Khan, and L.A. Archer. Solid-state polymer electrolytes with in-built fast interfacial transport for secondary lithium batteries. Nature Energy, 4, 2019. [220] Q. Zhao, X. Liu, J. Zheng, Y. Deng, A. Warren, Q. Zhang, and L.A. Archer. Designing electrolytes with polymerlike glass-forming properties and fast ion transport at low temperatures. Proceedings of the National Academy of Sciences, 117(42):26053–26060, 2020. [221] Q. Zhao, S. Stalin, C.-Z. Zhao, and L.A. Archer. Designing solid-state elec- trolytes for safe, energy-dense batteries. Nature Reviews Materials, 5:229– 252, 2020. [222] Y. Zhao, T. Zhou, D. Baster, M. El Kazzi, J.W. Choi, and A. Coskun. Tar- geted functionalization of cyclic ether solvents for controlled reactivity in high-voltage lithium metal batteries. ACS Energy Letters, 8(7):3180–3187, 2023. [223] J. Zheng, M.S. Kim, Z. Tu, S. Choudhury, T. Tang, and L.A. Archer. Reg- ulating electrodeposition morphology of lithium: towards commercially relevant secondary li metal batteries. Chem. Soc. Rev., 49:2701–2750, 2020. [224] Q. Zheng, L. Ma, R. Khurana, L.A. Archer, and G.W. Coates. Struc- ture–property study of cross-linked hydrocarbon/poly(ethylene oxide) 153 electrolytes with superior conductivity and dendrite resistance. Chem. Sci., 7:6832–6838, 2016. [225] Q. Zheng, D.M. Pesko, B.M. Savoie, K. Timachova, A.L. Hasan, M.C. Smith, T.F. Miller, G.W. Coates, and N.P. Balsara. Optimizing ion trans- port in polyether-based electrolytes for lithium batteries. Macromolecules, 51(8):2847–2858, 2018. [226] X. Zheng, Y. Guo, J.F. Douglas, and W. Xia. Understanding the role of cross-link density in the segmental dynamics and elastic properties of cross-linked thermosets. The Journal of Chemical Physics, 157(6):064901, 08 2022. [227] L. Zhou, B. He, and J. Huang. One-step synthesis of robust amine- and vinyl-capped magnetic iron oxide nanoparticles for polymer grafting, dye adsorption, and catalysis. ACS Applied Materials & Interfaces, 5(17):8678– 8685, 2013. 154