Fluorinated Cyclic Phosphorus(III)-based Electrolyte Additives for High

5 days ago - The obtained theoretical and experimental findings show that the considered phospholane molecule class enables high-voltage LIB ...
0 downloads 0 Views 1MB Size
Subscriber access provided by OCCIDENTAL COLL

Energy, Environmental, and Catalysis Applications

Fluorinated Cyclic Phosphorus(III)-based Electrolyte Additives for High Voltage Application in Lithium Ion Batteries: Impact of StructureReactivity Relationships on CEI Formation and Cell Performance Natascha von Aspern, Diddo Diddens, Takeshi Kobayashi, Markus Börner, Olesya Stubbmann-Kazakova, Volodymyr Kozel, Gerd-Volker Röschenthaler, Jens Smiatek, Martin Winter, and Isidora Cekic-Laskovic ACS Appl. Mater. Interfaces, Just Accepted Manuscript • DOI: 10.1021/acsami.9b03359 • Publication Date (Web): 09 Apr 2019 Downloaded from http://pubs.acs.org on April 9, 2019

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 40 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Applied Materials & Interfaces

Fluorinated Cyclic Phosphorus(III)-based Electrolyte Additives for High Voltage Application in Lithium Ion Batteries: Impact of StructureReactivity Relationships on CEI Formation and Cell Performance Natascha von Aspern, Diddo Diddens, Takeshi Kobayashi, Markus Börner, Olesya Stubbmann-Kazakova, Volodymyr Kozel, Gerd-Volker Röschenthaler, Jens Smiatek, Martin Winter** and Isidora Cekic-Laskovic*

Natascha von Aspern, Email: [email protected] Dr. Diddo Diddens, Email: [email protected] Dr. Jens Smiatek, Email: [email protected] Prof. Dr. Martin Winter (co-corresponding author), Email: [email protected] Dr. Isidora Cekic-Laskovic (corresponding author), Email: [email protected] Affiliation: Forschungszentrum Jülich GmbH Helmholtz-Institute Münster, Corrensstrasse 46, 48149 Münster, Germany

Dr. Markus Börner, Email: [email protected] Prof. Dr. Martin Winter (co-corresponding author), Email: [email protected] 1 ACS Paragon Plus Environment

ACS Applied Materials & Interfaces 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Affiliation: MEET Battery Research Center, University of Münster, Corrensstraße 46, 48149 Münster, Germany Takeshi Kobayashi, Email: [email protected] Dr. Jens Smiatek, Email: smiatek_@_icp.uni-stuttgart.de Affiliation: Institute for Computational Physics, University of Stuttgart, Allmandring 3, 70569 Stuttgart, Germany Dr. Olesya Stubbmann-Kazakova, Email: [email protected] Dr. Volodymyr Kozel, Email: [email protected] Dr. Prof. Gerd-Volker Röschenthaler, Email: [email protected] Affiliation: Jacobs University Bremen, Department of Life Sciences and Chemistry, Campus Ring 1, 28759 Bremen, Germany

Keywords: Lithium ion battery, Nonaqueous aprotic electrolyte, High voltage, Solid electrolyte interphase (SEI), Cathode electrolyte interphase (CEI), Functional additives, Phospholane molecules

Abstract Two selected and designed fluorinated cyclic phosphorus(III)-based compounds, namely 2(2,2,3,3,3-pentafluoropropoxy)-1,3,2-dioxaphospholane (PFPOEPi) and 2-(2,2,3,3,3-pentafluoropropoxy)-4-(trifluormethyl)-1,3,2-dioxaphospholane (PFPOEPi-1CF3) were synthesized and comprehensively characterized for high voltage application in lithium ion batteries (LIBs). Cyclic voltammetry (CV) and constant current cycling were conducted, followed by post mortem analysis 2 ACS Paragon Plus Environment

Page 2 of 40

Page 3 of 40 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Applied Materials & Interfaces

of the NMC111 electrode surface via scanning electron microscopy (SEM) and x-ray photoelectron spectroscopy (XPS). To support and complement obtained experimental results, density functional theory (DFT) calculations and molecular dynamics (MD) simulations were performed. Theoretical and experimental findings show that the considered phospholane molecule class enables high voltage LIB application, by sacrificial decomposition on the cathode surface and involvement in the formation of a cathode electrode interphase (CEI) via polymerization reaction. In addition, obtained results point out that the introduction of the CF3 group has a significant influence on the formation and dynamics of the CEI as well as on the overall cell performance, as the cell impedance as well as the thickness of the CEI is increased compared to the cells containing PFPOEPi, which results in a decreased cycling performance. This systematic approach allows to understand the structure-reactivity relationship of the newly synthesized compounds and helps to further tailor the vital physicochemical properties of functional electrolyte additives relevant for high voltage LIB application.

1. Introduction One of the crucial challenges for the lithium ion batteries (LIBs) and lithium metal batteries (LMBs) relates to the enhancement of the energy density by identifying alternatives to the stateof-the-art (SOTA) active and inactive materials.1-2 The SOTA nonaqueous aprotic electrolyte consisting of LiPF6 as conducting salt in a mixture of linear and cyclic organic carbonates reveals different disadvantages.3-4 With this in line, electrolyte decomposition stands for the most pronounced one5-8 and plays a crucial role when increasing the energy density of a LIB by means of so-called high voltage cathodes with a cut-off potential  4.3 V vs. Li/Li+.9 Although this

3 ACS Paragon Plus Environment

ACS Applied Materials & Interfaces 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

approach enables an increase in the specific energy, it happens at an expense in cycle life, structural changes of the layered transition metal electrodes and transition metal dissolution as well as Al current collector dissolution and electrolyte degradation.5, 10-17 Moreover, cycling performance can be influenced by intrinsic material phase transition and irreversible specific capacity loss in the initial cycle.18-20 One effective and cost favorable approach refers to the introduction of functional additives which do not very much affect the bulk properties of the base electrolyte, however, they decompose prior to the electrolyte and are involved in formation of a cathode electrolyte interphase (CEI).11, 13, 21 So far many compounds have been reported in literature22-25. Among them, lithium salts, such as lithium bis(oxalato)borate (LiBOB)26, and difluoro-(oxalato)borate (LiDFOB)27, phosphorus containing compounds, such as tris(2,2,3,3,3-pentafluoropropyl) phosphate (5FTPrP)28, tris(1,1,1,3,3,3-hexafluoropropan-2-yl) phosphate (HFiP)28, and tris(1,1,1,3,3,3hexafluoropropan-2-yl) phosphite (THFPP)28, as well as metal salts such as magnesium bis(trifluoromethane-sulfonyl)imide(Mg(TFSI)2)29, tetrakis(trimethyl-siloxy)titanium (TMST)30, lithium tetrakis-(trimethylsiloxy)aluminate (LiTMSA)30, and tris(trimethyl-siloxy)aluminum (TMSA)30. In some cases the hydrogen atom was substituted by a fluorine atom to impact relevant physicochemical properties such as a high electronegativity and a high ionic potential, which measures the polarizing power of the cation. On the one hand, fluorine substitution leads to a lower energy level of the highest occupied molecular orbital (HOMO) and lowest unoccupied molecular orbital (LUMO) of the molecule, which results in a higher oxidation stability and a lower reduction stability.31-33 Due to the lower HOMO energy level, the cycling stability at the cathode is improved.32, 34 On the other hand, the lower LUMO energy level can negatively influence the cycling stability due to the lower reduction stability, which can lead to a formation of fluorine containing compounds on the anode surface and to a formation of an insufficient solid electrolyte

4 ACS Paragon Plus Environment

Page 4 of 40

Page 5 of 40 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Applied Materials & Interfaces

interphase (SEI).35-36 However, there has recently been some criticism on the use of the HOMO/LUMO concept in battery science.37-39 Phospholane molecules with phosphorus in the oxidation state V were firstly reported as decomposition products of the electrolyte during aging processes.40-41 To reveal their impact on the overall electrochemical performance, Gao et al.42 reported use of ethylene ethyl phosphate (EEP) at an amount of 10 wt% as a co-solvent that reduces burning time of the considered electrolyte by 50% and displays better cycling performance compared to the reference electrolyte with a cut-off voltage of 4.3 V in graphite/LiNi1/3Mn 1/3Co 1/3O2 (NMC111) cells. The amount of 1 wt% of phospholane molecules with the phosphorus atom in the oxidation state V for high voltage application in graphite/LiNi0.5Mn0.3Co0.2O2 (NMC532) cells was first reported by Su et al.43. Hereby, the cells with phospholane containing electrolytes displayed less capacity fading compared to the reference electrolyte counterparts, due to the prior decomposition compared to the reference electrolyte and formation of an effective CEI. Furthermore, they postulated a decomposition mechanism of the phospholane molecule via ring opening and polymerization reaction on the NMC532 electrode surface. With this in line, we designed and synthesized two fluorinated cyclic phosphorus(III)-based containing compounds (2-(2,2,3,3,3-pentafluoropropoxy)-1,3,2-dioxaphospholane (PFPOEPi) and

2-(2,2,3,3,3-pentafluorpropoxy)-4-(trifluormethyl)-1,3,2-dioxaphospholane

(PFPOEPi-

1CF3)). Their structural formulae are depicted in Scheme 1. The cyclic phosphorus compounds were selected on the basis of their ability to polymerize via ring opening reaction on the cathode surface to prevent further electrolyte decomposition.43 The

pentafluoropropyl group was

considered in our previous work related to fluorinated alkylphosphates and it was shown that this group can enhance the cycling stability in regard to the high voltage performance.28 Within this 5 ACS Paragon Plus Environment

ACS Applied Materials & Interfaces 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 6 of 40

work, the role as well as the decomposition mechanism of the new electrolyte additives for high voltage application was elucidated. Furthermore, the CF3 group as an electron withdrawing group (EWG) attached to the ring structure of the phospholane molecule and its impact on CEI formation and overall cell performance was systematically investigated to elaborate structure-reactivity relationships as well as the CEI formation and overall cell performance.

F

F

F F

F

O

O P

F F

O

O P

O

F

F F

O

F F F PFPOEPi

PFPOEPi-1CF3

Scheme 1. Chemical structure of the investigated high voltage electrolyte additives: 2-(2,2,3,3,3-pentafluoropropoxy)1,3,2-dioxaphospholane (PFPOEPi) and 2-(2,2,3,3,3-pentafluoropropoxy)-4-(trifluoro-methyl)-1,3,2-dioxaphospholane (PFPOEPi-1CF3).

2. Results and discussions As known from literature, introduction of fluorine into a molecule increases its oxidation potential.31-33 In order to quantify this effect, adiabatic reduction and oxidation potentials of all considered electrolyte compounds were calculated at the G4MP2 level (i.e. including a reoptimization of the geometry of the reduced/oxidized molecule37-38). The reduction and oxidation potentials were computed according to Equations 1 and 237-38:

𝐸red = ― 𝐸ox =

𝐺 (𝑀 ― ) ― 𝐺(𝑀) 𝐹

𝐺 (𝑀 + ) ― 𝐺(𝑀) 𝐹

―1.4 𝑉

―1.4 𝑉

6 ACS Paragon Plus Environment

(1) (2)

Page 7 of 40 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Applied Materials & Interfaces

Here, G(M-), G(M+) and G(M) represent computed free energies of reduced, oxidized and neutral form of molecule M, respectively, and F is the Faraday constant (96 485.3329 s A/mol). The shift of 1.4 V was subtracted to convert the values from the absolute potential scale to the scale relative to the Li/Li+ redox pair.37-38 Table 1 compares reduction and oxidation potentials of the two considered additive molecules to the values of the other electrolyte components, i.e. EC, EMC molecules and PF6-species. For the oxidation potentials, calculations on single isolated molecules surrounded by an implicit solvent (see experimental section) were performed in a first step. However, we considered possible degradation reactions upon oxidation to yield more accurate results (see below). For the reduction potentials of EC, EMC and PF6-, it turned out that singlemolecule calculations are too simplistic to yield reasonable estimates, in agreement with previous observations.37-38 Since lithium ions are electroactive species in the reduction process and the presence of a positive charge should significantly favor reduction, a coordinating Li+ was included in the calculations of the reduction potentials for EC, EMC and PF6-, thus mimicking the intermolecular environment within the electrolyte to a certain degree. To make the values listed in Table 1 fully comparable, we included lithium ion in the calculations of PFPOEPi and PFPOEPi1CF3, which coordinates to one of the oxygen atoms of the additive molecule (see SI). Nevertheless, we estimated that lithium ion is essentially solvated by EC, EMC or PF6-, as reflected by the fact that the binding free energies of lithium ions to a single EC, EMC or PF6- species are in the range of -8 kcal/mol, whereas it is only -4 kcal/mol for PFPOEPi and even slightly positive (0.6 kcal/mol) for PFPOEPi-1CF3). Additionally, the comparably low concentration of PFPOEPi and PFPOEPi-1CF3 in the electrolyte formulation as functional additive should clearly favor lithium ion solvation by EC, EMC and PF6-. Table 1. Calculated adiabatic reduction and oxidation potentials of PF6-, EC, EMC, PFPOEPi, and PFPOEPi-1CF3 at the G4MP2 level of theory. Note that for the reduction potentials, a coordinating lithium ion was included in the

7 ACS Paragon Plus Environment

ACS Applied Materials & Interfaces 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 8 of 40

calculations to yield more accurate estimates, whereas the oxidation potentials have been computed for the isolated molecules (values on the left side). The second values (on the right side) for the oxidation potentials were obtained when considering proton abstraction reactions from the oxidized molecule by other electrolyte molecules (see text). Reduction Potential Oxidation Potential Molecule/Species [V] vs. Li/Li+ [V] vs. Li/Li+ LiPF6 (red) / PF6- (ox) 0.21 9.55 / Li+-EC (red) / EC (ox)

0.62

7.03 / 5.74

Li+-EMC (red) / EMC (ox)

0.57

6.94 / 5.77

Li+-PFPOEPi (red) / PFPOEPi (ox)

0.54

5.09 / 5.51

Li+-PFPOEPi-1CF3 / PFPOEPi-1CF3 (ox)

0.76

5.30 / 5.37

The calculated intrinsic oxidation potentials in Table 1 (values on the left side) indicate a significantly lower oxidative stability of both phospholane molecules as compared to EC, EMC and PF6-, which should lead to an earlier decomposition of the functional additives on the cathode surface. Furthermore, introduction of a CF3 group to the phospholane ring results in a slightly increased oxidation potential value. Interestingly, the second oxidation potentials for PFPOEPi and PFPOEPi-1CF3 amount to 4.80 V vs. Li/Li+ and 4.87 V vs. Li/Li+, respectively, suggesting that the additive molecules are likely to undergo double oxidation at the initial stage of the CEI formation. In addition, oxidation potentials of PFPOEPi and PFPOEPi-1CF3 were calculated when considering a simultaneous proton transfer from the oxidized molecule to other molecule/species (mainly EC, EMC and PF6-), which may lower the oxidation potential by up to 2 V for common electrolyte compounds (see Supporting Information).37-38 In the present case, however, the values (5.51 V vs. Li/Li+ for PFPOEPi and 5.37 V vs. Li/Li+ for PFPOEPi-1CF3) are larger than the intrinsic potentials in Table 1. On the contrary, for EC and EMC, we find respective oxidation potentials of 5.74 V vs. Li/Li+ and 5.77 V vs. Li/Li+ when simultaneous proton transfer reactions are considered (see SI), which is significantly lower than the corresponding intrinsic potentials (Table 1). Analogous calculations for proton transfer reactions in oxidized EC/ PF6- clusters from

8 ACS Paragon Plus Environment

Page 9 of 40 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Applied Materials & Interfaces

the literature37-38 yielded effective oxidation potentials in the range of 5.5-6.2 V vs. Li/Li+ in good agreement with our values and the experimental data. The fact that oxidation becomes highly favorable due to proton abstraction for EC and EMC but even slightly less favorable for PFPOEPi and PFPOEPi-1CF3 demonstrates that the additive molecules apparently stabilize the positive charge more effectively than the carbonates. This also suggests that the additive molecules are decomposed via a different reaction mechanism than the conventional electrolyte compounds, in line with their property as CEI formers (see experimental part below). Moreover, for all computed values, the oxidation potentials of PFPOEPi and PFPOEPi-1CF3 are lower than for EC and EMC (Table 1). Thus, it is expected that the additives decompose first when increasing the applied voltage. In case of the reduction potentials, phospholane-Li+ clusters display values comparable to those of LiPF6, Li+-EC and Li+-EMC. However, as mentioned above, EC, EMC and PF6- species are more likely coordinated to a lithium ion than to PFPOEPi and PFPOEPi-1CF3. In comparative calculations for isolated PFPOEPi and PFPOEPi-1CF3 molecules (i.e. without a lithium ion), reduction potentials drop to 0.10 V vs. Li/Li+ and 0.12 V vs. Li/Li+, respectively. This in combination with the fact that clusters of the type LiPF6, Li+-EC and Li+-EMC are much more abundant in the vicinity of the anode points towards a good reductive electrochemical stability of PFPOEPi and PFPOEPi-1CF3, such that the SEI formation appears to be less affected by the presence of the considered additives. To confirm obtained theoretical findings, following experimental measurements were conducted. First, a concentration screening for both considered functional additives, PFPOEPi and PFPOEPi1CF3 was performed in order to select their optimum amount in considered electrolyte formulations and compared to the REF. The best overall performance was achieved with 0.5 wt.%

9 ACS Paragon Plus Environment

ACS Applied Materials & Interfaces 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

of each additive in the REF formulation, and therefore used for further characterization. The conductivities of the three considered electrolyte formulations were compared in the temperature range of -30 to 50°C.

Figure 1. Temperature dependent conductivity values of the three considered electrolyte formulations (1M LiPF6 EC:EMC (1:1 by wt.; denoted as REF), REF + 0.5% PFPOEPi and REF + 0.5% PFPOEPi-1CF3 in the temperature range of -30 to 50°C.

As depicted in Figure 1, addition of phospholane-based additives to the REF, has no influence on the conductivity. In the next step, redox behavior on active material (graphite or NMC111) was evaluated as shown in Figure 2.

10 ACS Paragon Plus Environment

Page 10 of 40

Page 11 of 40 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Applied Materials & Interfaces

Figure 2. First cycle cyclic voltammograms of the a) graphite- and b) NMC111-based cells containing considered electrolyte formulations (1M LiPF6 EC:EMC (1:1 by wt.) (REF), REF + 0.5% PFPOEPi and REF + 0.5% PFPOEPi1CF3).

All considered electrolyte formulations show good compatibility with both graphite and NMC111 electrodes and no additional decomposition peaks related to the phospholane molecules were observed. The obtained results are in a good agreement with the calculated redox potentials (Table 1) as they reveal that phospholane molecules are reductively stable. Their oxidative decomposition should occur beyond 4.6 V vs. Li/Li+, thus making them suitable candidates for application in graphite/NMC111 cells with a cut-off voltage of 4.5 V. However, the obtained CV results of NMC111/Li cells do not indicate decomposition of the phospholane molecules. This leads to the question whether the considered phospholane molecules decompose on the NMC111 electrode surface and if the CEI is chemically or electrochemically induced or a combination of both. For this reason, graphite/NMC111 coin cells were assembled and stored for two weeks at 20 °C. The coin cells were thereafter disassembled and XPS analysis of the NMC111 electrode surfaces was performed. Table 2. CEI composition of the NMC111 electrode of the electrolyte formulations (1M LiPF6 EC:EMC (1:1 by wt.) and REF + 0.5% PFPOEPi) after two weeks of storage at 20 °C 1M LiPF6 EC:EMC (1:1) + 0.5% 1M LiPF6 EC:EMC (1:1) CEI Composition PFPOEPi [%at] [% ] at

LiF

0.60

0.07

R2CO3

0.00

0.05

LiCO3

0.04

0.06

C-O

0.07

0.73

CF3

0.00

0.04

LiPFx

0.15

0.03

LiPOxFy

0.13

0.03

11 ACS Paragon Plus Environment

ACS Applied Materials & Interfaces 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The results listed in Table 2 illustrate a significant increase of the C-O concentration on the NMC111 surface for the PFPOEPi containing cell compared to the REF containing counterpart. Furthermore, a peak at 293 eV28 in the C 1s spectra, which can be attributed to CF3, was observed. Due to the fact that other decomposition components of the CEI have the same or lower concentration for the PFPOEPi containing cells compared to the REF cells, it can be assumed that this increase is driven by decomposition of PFPOEPi. As no current was applied, it can be assumed that this decomposition is chemically driven. To confirm this assumption CV measurements in NMC111/Li cells with three different scan rates were performed and obtained results are depicted in Figure 3.

Figure 3. First cycle cyclic voltammograms of 1M LiPF6 EC:EMC (1:1 by wt.) + 0.5% PFPOEPi with different scan rates.

The shape of the cyclic voltammograms does not change with increased scan rate. Therefore, it can be interpreted that in the potential range from OCP to 4.6 V vs. Li/Li+ no electrochemically induced CEI is formed. The combination of these findings as well as the XPS results leads to the conclusion that the phospholane additives are involved in a chemically induced CEI formation on the NMC111 electrode surface.19, 44 The effectiveness of formed CEI is further evaluated. 12 ACS Paragon Plus Environment

Page 12 of 40

Page 13 of 40 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Applied Materials & Interfaces

Figure 4. Coulombic efficiencies (top), specific discharge capacities (middle) and the capacity retention (bottom) vs. cycle number of the three electrolyte formulations (1M LiPF6 EC:EMC (1:1 by wt.) (REF), REF + 0.5% PFPOEPi and REF + 0.5% PFPOEPi-1CF3) in graphite/NMC111 cells at 20°C with a cut-off voltage of 4.5 V.

The plots of the Coulombic efficiency, specific discharge capacity and the capacity retention of graphite/NMC111 cells are depicted in Figure 4 (long term cycling with 80% state of health (SOH) as end of life criterion can be found in the Figure S12.). After 100 charge/discharge cycles, the following order of the capacity retention values was obtained: REF (83.9%) < REF + 0.5% PFPOEPi-1CF3 (88.3%) < REF + 0.5% PFPOEPi (90.6%). In case of the Coulombic efficiencies, no impact of the phospholane additives can be observed, as considered electrolyte formulations display an average efficiency of 99.9%. The same behavior was observed for the first cycle efficiency (82.2% for REF, 82.4% for REF + 0.5% PFPOEPi and 82.3% for REF + 0.5% PFPOEPi-1CF3). However, the introduction of an EWG (CF3 group) to the ring results in a slightly 13 ACS Paragon Plus Environment

ACS Applied Materials & Interfaces 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

decreased cycling behavior compared to the cell with PFPOEPi. This effect, affected by the presence of CF3 group, may result from the CEI formation/composition and was therefore further investigated by EIS. The obtained EIS spectra of symmetric graphite/graphite and NMC111/NMC111 cells as well as the graphite/NMC111 full cells after formation and after 100 charge/discharge cycles are shown in Figure 5. This approach allows to distinguish whether the impedance growth in the graphite/NMC111 cells originates from the anode or cathode side. Due to the fact that the impedance is strongly dependent on the state of the charge, the cells were charged until 50% of lithiation degree. All obtained results were fitted with the equivalent circuit described by Zhao et al.45. Herein, the first resistance Rel represents the electrolyte resistance and is represented in the Nyquist plot by x-axis interception. The first semicircle in the Nyquist plot describes the interphase resistances (RSEI, RCEI and RInt), which results from the formation of a SEI and CEI on the electrode surface respectively. Rct describes the charge transfer resistance and is assigned to the second semicircle in the Nyquist plot. Firstly, for all three electrolyte formulations it can be observed that the major influence on the impedance growth in the graphite/NMC111 cells is driven by the cathode after formation as well as after 100 charge/discharge cycles, as symmetric NMC111/NMC111 cells show a significant increase in the overall impedance, whereas the impedance change in the symmetric graphite/graphite cells seems to play a minor role.

14 ACS Paragon Plus Environment

Page 14 of 40

Page 15 of 40 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Applied Materials & Interfaces

Figure 5. EIS spectra of the three electrolyte formulations (1M LiPF6 EC:EMC (1:1 by wt.) (REF), REF + 0.5% PFPOEPi and REF + 0.5% PFPOEPi-1CF3) in graphite/graphite (left), graphite/NMC111 (middle) and NMC111/NMC111 (right) cells after formation (top row) as well as after 100 charge/discharge cycles (bottom row) at 20 °C.

In a second step, the influence of each electrolyte on the impedance after formation charge/discharge cycles as well as after 100 charge/discharge cycles for the symmetric NMC111/NCM111 cells was elucidated. Here, the REF displays a slight increase in the first semicircle (RCEI), whereas the second semicircle (Rct) increases significantly, when comparing the 15 ACS Paragon Plus Environment

ACS Applied Materials & Interfaces 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

results after formation and after 100 charge/discharge cycles. For the PFPOEPi containing cells, the RCEI increases more than for the REF, whereas the Rct remains almost the same. In case of the PFPOEPi-1CF3 containing cells, the semicircle corresponding for RCEI stays the same, whereas the Rct increases during cycling. In the third step, the effect of the CF3 group was evaluated. By introducing an EWG, a decrease in the CEI resistance can be observed which does not increase during cycling compared to the PFPOEPi containing cells. In the case of the charge transfer resistance, a significant increase after formation compared to the PFPOEPi containing cells can be observed, which increases during prolonged cycling. By correlating these results with the obtained cycling data it can be concluded that a low overall impedance, like for the PFPOEPi containing cells, is beneficial for improved cycling performance (Figure 4). However, the resistance for the CEI increases during cycling, which is an indication for further electrolyte decomposition. Furthermore, PFPOEPi-1CF3 displays the highest overall impedance from which we would suspect the poorest cycling performance, but an improvement compared to the REF could be observed (Figure 4). This could be due to a constant resistance of the CEI, which does not grow during ongoing cycling. Based on the obtained results it can be concluded that there is no simple correlation of the resistance with the improved cycling performance, but it is clear that the EWG has a significant impact. To further elucidate the effect of the EWG on the decreased cycling performance as well as increased overall resistance, SEM analysis of the graphite and NMC111 electrodes was conducted and the obtained SEM images of NMC111 electrodes after formation and after 100 charge/discharge cycles are depicted in Figure 6.

16 ACS Paragon Plus Environment

Page 16 of 40

Page 17 of 40 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Applied Materials & Interfaces

Figure 6. SEM images of the NMC111 electrodes of the three considered electrolyte formulations (1M LiPF6 EC:EMC (1:1 by wt.) (REF), REF + 0.5% PFPOEPi and REF + 0.5% PFPOEPi-1CF3) after formation and 100 charge/discharge cycles as well as the pristine electrode. The red arrows pointing at decomposition spots at the NMC111 surface.

The SEM images of the NMC111 electrodes with REF after formation show presence of some decomposition products on the surface of the particles, whereas after 100 charge/discharge cycles the electrode particles are completely covered by a layer of decomposition products. In the case of the phospholane containing cells after formation, additional spherical shape decomposition products can be observed, which are marked with red arrows. For the PFPOEPi containing cell, the electrode surface after 100 charge/discharge cycles appears to be unchanged compared to the cathode surface after formation cycles. In the case of the PFPOEPi-1CF3 containing cells, the electrode surface resembles that of the electrode cycled in the REF after 100 charge/discharge 17 ACS Paragon Plus Environment

ACS Applied Materials & Interfaces 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

cycles, by showing a complete coverage of the surface by decomposition products. This is an indication that the formed CEI differs depending on the electrolyte formulation and that the addition of the CF3 group to the phospholane ring influences CEI formation as suspected from the EIS results. To get a deeper insight, the thickness and composition of the CEI were determined by means of XPS analysis with the mathematical approach developed by Niehoff et al.46-47. The corresponding results are shown in Figure 7. Regarding the CEI thickness, PFPOEPi containing cells reveal the overall thinnest CEI followed by the REF-based and PFPOEPi-1CF3 containing cells.

Figure 7. Determined thicknesses and compositions of the CEI obtained in the three different electrolyte formulations (1M LiPF6 EC:EMC (1:1 by wt.) (REF), REF + 0.5% PFPOEP and REF + 0.5% PFPOEPi-1CF3) after formation and after 100 charge/discharge cycles.

In the case of the REF-based and the PFPOEPi containing cells, an increase in the thickness of the CEI can be observed during constant current cycling, whereas for the PFPOEPi-1CF3 containing cell, a slight decrease is noticeable. This means that when adding PFPOEPi-1CF3 that after the formation cycles, no further electrolyte decomposition takes place, whereas for the REF and the PFPOEPi containing electrolyte, an on-going decomposition occurs after the formation cycles. By

18 ACS Paragon Plus Environment

Page 18 of 40

Page 19 of 40 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Applied Materials & Interfaces

taking a look into the composition of the CEI, amount of decomposition products differs significantly. First, the highest LiF content can be observed for the REF (0.39 at%), whereas the phospholane containing cells display a much lower content (PFPOEPi: 0.06 at%, PFPOEPi-1CF3: 0.10 at%) after formation. During constant current cycling, the content of LiF decreases for PFPOEPi (0.02 at%) containing cells decreases even further, whereas for PFPOEPi-1CF3 (0.33 at%)

containing cells the LIF content increases to the same amount as for the REF (0.32 at%). This

is an indication that the introduction of an EWG to the phospholane ring leads to a complete consumption of PFPOEPi-1CF3 during formation and that there is no phospholane molecule left to suppress the LiPF6 decomposition to PF5 and LiF. Moreover, this explains the slight decrease in the cycling performance (Figure 4) compared to the cell cycled with PFPOEPi containing electrolyte as the generated LiF insulates the surface of the electrode.48-49 Second, phospholane containing cells display a CF3 content (binding energy at 293 eV28 in a C 1s spectrum) after formation as well as after 100 charge/discharge cycles. After 100 charge/discharge cycles, the CF3 content of the PFPOEPi-1CF3 containing electrolyte is double the amount as for the PFPOEPi containing cells, which goes along with chemical structure of the phospholane compounds (PFPOEPi one CF3 group, PFPOEPi-1CF3 two CF3 groups) and is an indication that the entire molecule decomposes on the electrode surface. Even though the CV measurements showed no early decomposition of phospholane containing electrolytes compared to the REF electrolyte, the XPS results reveal that the phospholane molecules decompose on the cathode surface, as indicated by the calculated oxidation potentials. In addition, obtained XPS results (Table 2) from the stored cells as well as the CV results with different scan rates (Figure 3) indicate a chemically induced CEI formation by the phospholane molecules, it can be concluded that the formed CEI during cycling is manly chemically driven in case of the phospholane containing electrolyte formulations.

19 ACS Paragon Plus Environment

ACS Applied Materials & Interfaces 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Furthermore, calculated oxidation potentials of a simultaneous proton transfer from the oxidized molecule to another molecule revealed a higher oxidation potential values for the phospholane molecules compared to the intrinsic potentials, indicating a different decomposition mechanism, which was confirmed by XPS data. Moreover, the oxidation values for PFPOEPi-1CF3 are slightly higher compared to PFPOEPi, which should lead to a pronounced PFPOEPi decomposition. However, in the initial cycles it seems that the reactivity of the PFPOEPi-1CF3 molecules with the NMC111 electrode is more favored compared to the PFPOEPi containing electrolyte, as the thickness of the CEI differs significantly (PFPOEPi: 0.38 nm, PFPOEPi-1CF3: 1.44 nm). It seems that oxidation potential cannot be considered as single parameter for identifying considered electrolyte additives as candidates for high voltage application in LIBs. Third, the C-O content can be an indication that the phospholane molecules polymerize on the electrode surface. For the REF, the C-O content, which can be assigned to polyethylene oxide (PEO), increases from 0.32 at% to 0.49 at%, the C-O content for the PFPOEPi containing electrolyte on the NMC111 surface amounts to only 0.18 at%, however the value increases to 0.48 at% during ongoing cycling. Due to the fact that the content of other decomposition products does not increase in the same way, the increased C-O content cannot be solely assigned to the decomposition product PEO. This implies that decomposition of PFPOEPi molecules on the NMC111 electrode surface must be included as well. In case of the PFPOEPi-1CF3 containing electrolyte cells, the C-O content after formation amounts to 0.50 at%, which is more comparable to the REF (0.32 at%). The thickness of PFPOEPi-1CF3 changes only slightly during cycling and the obtained NMC111/NMC111 impedance spectra (Figure 5) reveal that the RCEI does not change during cycling. It can be assumed that the electrolyte decomposition occurs during formation cycles. The high C-O content of 0.50 at% after formation for PFPOEPi-1CF3 (REF: 0.32 at% after formation) can be assigned to the PEO

20 ACS Paragon Plus Environment

Page 20 of 40

Page 21 of 40 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Applied Materials & Interfaces

decomposition product as well as the phospholane decomposition on the NMC111 surface. Based on the obtained results and supported by the work of Kluger et al.50 and Su et al.43, the following decomposition mechanism is postulated:

Scheme 2. Postulated decomposition mechanism of the phospholane based additives on the NMC111 electrode surface.

As known from literature, phosphorus III compounds get oxidized to phosphorus V compounds by scavenging oxygen.51-54 This reaction is considered as the first step of the proposed mechanism. Thereafter, the hydroxyl group present on the NMC111 particle surface55-57 reacts with the phosphorus atom resulting in a ring opening reaction. In the next step, the formed hydroxyl group can react further with a second phospholane molecule and is followed by a polymerization reaction of the phospholane molecules. The reaction energies were calculated at the B3LYP/cc-pVTZ level, in which we however, did not explicitly take the NMC111 surface into account. Rather, we restricted ourselves to the propagation reaction between an opened and a closed PFPOEPi or PFPOEPi-1CF3 molecule (see Scheme 3 for the representative model reactions).

21 ACS Paragon Plus Environment

ACS Applied Materials & Interfaces 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

F3C CF2

O O

P

O

OH OH

O R

+

O

P

O

Page 22 of 40

F2C CF3

F2C O

O

OH

R

P

CF3

F2C

O

O

O

O

P

CF3

O O

R

OH R

R= H PFPOEPi R= CF3 PFPOEPi-1CF3

Scheme 3. Reactions used to compute the polymerization energies summarized in Table 3.

The reaction energies (Table 3), which are negative for both additives, reveal the postulated mechanism to be possible from the thermodynamic point of view. Furthermore, the reaction energy values for the PFPOEPi (-2.070 kcal/mol) and the PFPOEPi-1CF3 (-1.569 kcal/mol) systems are similar. This means, that from a theoretical point of view no significant difference in the polymerization reaction on the NMC111 surface should be expected and that the probability of this reaction is likely. The influence of the cathode surface might be elucidated by further calculations.37,

58-60

The experimental results confirm the theoretically obtained data, as the

obtained XPS results reveal double amount of the CF3 content for the PFPOEPi-1CF3 containing electrolyte compared to the PFPOEPi containing electrolyte on the NMC111 electrode surface after 100 charge/discharge cycles. Furthermore, this is an indication that a ring opening decomposition reaction occurs and not a cleavage of the pentafluoropropyl group. Table 3. Calculated reaction energies of the two considered additives decomposing on the NCM111 surface via ring opening and polymerization. Additive ∆E [kcal/mol] PFPOEPi -2.070 PFPOEPi-1CF3

-1.569

To elucidate the effect of the phospholane molecules on the graphite electrode, SEM analysis was carried out. The obtained SEM images of the graphite anode after formation as well as after 100 charge/discharge cycles of the three different electrolyte formulation, depicted in Figure 8, reveal decomposition products after formation cycles in case of REF, whereas after 100 charge/discharge 22 ACS Paragon Plus Environment

Page 23 of 40 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Applied Materials & Interfaces

cycles the surface is completely covered with decomposition products. In the case of the phospholane containing electrolyte formulations, no significant difference can be observed. However, when comparing with the REF after formation, more decomposition products can be observed on the graphite surface, whereas after 100 charge/discharge cycles, the surface of the graphite electrode cycled with the REF displays more decomposition products.

Figure 8. SEM images of the graphite electrodes with the three considered electrolyte formulations (1M LiPF6 EC:EMC (1:1 by wt.) (REF), REF + 0.5% PFPOEPi and REF + 0.5% PFPOEPi-1CF3) after formation and after 100 charge/discharge cycles as well as of the pristine electrode. The red arrows pointing at decomposition spots at the graphite surface.

For further evaluation of the impact of the electrolyte composition on the SEI, XPS measurements were conducted and the obtained results are depicted in Figure 9.

23 ACS Paragon Plus Environment

ACS Applied Materials & Interfaces 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Figure 9. a) Determined thicknesses of the organic and inorganic parts of the SEI and b) C 1s spectra of the three different electrolyte formulations (1M LiPF6 EC:EMC (1:1 by wt.) (REF), REF + 0.5% PFPOEPi and REF + 0.5% PFPOEPi-1CF3) after formation and after 100 charge/discharge cycles.

The obtained results on the graphite reveal that the thickness of the inorganic part is not influenced by constant current cycling or by the addition of the phospholane additives. However, the organic part is influenced by both the constant current cycling and the addition of the phospholane additives, as it is found to be thinner than in the case of REF, and its thickness is increasing during cycling. The phospholane-based additives display in general a higher thickness of the organic part after formation that does not change during cycling for the PFPOEPi containing cells and increases slightly for the PFPOEPi-1CF3 containing cells. After 100 charge/discharge cycles, electrodes with considered electrolyte formulations reveal similar SEI thickness, REF: 4.0 nm; REF + 0.5% PFPOEPi: 4.4 nm; and REF + 0.5% PFPOEPi-1CF3: 4.0 nm, respectively. In the C 1s spectra (Figure 9b) no significant difference between the three electrolyte formulations are noticeable. Furthermore, no peak at the binding energy of 293 eV28 (CF3 group) can be observed. This is an 24 ACS Paragon Plus Environment

Page 24 of 40

Page 25 of 40

indication that the phospholane molecules do not decompose on the graphite surface and there is no significant influence of the CF3 group. The obtained results go along with the calculated reduction potentials and CV measurement, where no additional decomposition peak could be observed and it was expected that the phospholane molecules do not decompose on the anode surface. Experimentally obtained results concerning considered phospholane molecules on the cathode and anode surface were supported by molecular dynamics simulations with 3 wt% of PFPOEPi and PFPOEPi-1CF3 in a mixture of 1M LiPF6 in EC:EMC (1:1 by wt.) as displayed in Figure 10. A comparable procedure was used by Feng et al.61. Herein, they investigated a small amount of water molecules in ionic liquids and showed that the occurrence probability of water differs with respect to the differently charged surfaces.

40 35

Average mass density [g/L]

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Applied Materials & Interfaces

30

3% PFPOEPi 3% PFPOEPi-1CF3

25 20 15 10 5 0 Anode

Cathode

Figure 10. Average mass density of PFPOEPi and PFPOEPi-1CF3 additives within a layer of thickness d=1 nm at the anode and the cathode side for 3 wt.%. The bars denote standard deviations and the fitting range was adjusted to include non-negligible additive concentrations.

25 ACS Paragon Plus Environment

ACS Applied Materials & Interfaces 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Due to the fact that the results obtained from the molecular dynamics simulations refer to a charged cell and the post mortem analysis such as XPS are conducted in the discharged stage, the definition of cathode and anode is changed for the simulation for reasons of clarity. The simulations point out that the average mass density for the phospholane molecules is higher on the cathode surface than on the anode surface, thus indicating the higher probability of the phospholane molecules on the cathode surface compared to the anode. The simulations are in a good agreement with obtained XPS data, as the experimental data reveal that the phospholane molecules decompose on the cathode surface and not on the anode. With this in line, the determination of the probability of the molecules near the anode or cathode surface seems to be a crucial fact besides the reduction and oxidation potential for predicting the decomposition of the additives on the electrode surface. Based on the obtained results it can be concluded that phospholane molecules are active on the cathode side and indirectly influence the composition and thickness of the SEI during formation. Moreover, the addition of the CF3 group leads to a slight decrease in cycling performance, which mainly effects the cell chemistry on the cathode side, whereas the effect on the anode side can be neglected.

3. Conclusion Two newly designed phospholane molecules were synthesized, systematically characterized and investigated as high voltage functional additives in LIB electrolytes. The main focus was set on the impact of the presence and structure of considered phospholane additives on the overall graphite/NMC111 cell performance. By adding 0.5 wt.% of the phospholane molecules to the reference electrolyte (REF), an advanced cycling performance could be achieved. The post mortem 26 ACS Paragon Plus Environment

Page 26 of 40

Page 27 of 40 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Applied Materials & Interfaces

XPS analysis revealed that phospholane molecules do not decompose at the anode surface, however indirectly affect both the composition and thickness of the formed SEI. The obtained EIS results, however, suggest that the effect on the anode surface plays a minor role, whereas the effectively formed CEI contributes the improved cycling stability. Based on the XPS results obtained for the NMC111 electrode, a decomposition mechanism of the phospholane molecules via ring opening polymerization reaction on the electrode surface could be postulated. The data obtained from the DFT calculations display no difference in the reaction energies between the two phospholane molecules. Based on the XPS data from the stored cells as well as the CV results with different scan rates reveal, formed CEI is mainly chemically driven. Furthermore, detailed analysis of the results obtained by means of EIS and XPS provides closer insight into the considered structure-reactivity relationships. The EIS data reveal a significant increase in the overall impedance for the cells with PFPOEPi-1CF3 containing electrolyte compared to the PFPOEPi containing counterparts. This increase in impedance is driven by the Rct, whereas the RCEI remains the same during constant current cycling (PFPOEPi containing electrolyte cells: RCEI increases, Rct stays the same). The same trend was observed for the CEI, referred to its thickness. The introduction of an EWG leads to the formation of an effective CEI, as its thickness does not change during ongoing cycling, whereas for the cells with PFPOEPi containing electrolyte, the thickness of the CEI increases. Additional influence of the EWG relates to the different LiF content in the CEI in cells with the PFPOEPi and PFPOEPi-1CF3 containing electrolytes. After the formation cycles, both cells with phospholane containing electrolytes display a low content of LiF. However, during cycling, the content decreases for the cells with PFPOEPi containing electrolyte, whereas for the cells with PFPOEPi-1CF3 containing electrolyte increases. These effects are considered

27 ACS Paragon Plus Environment

ACS Applied Materials & Interfaces 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

responsible for the decreased cycling behavior of the cells with PFPOEPi-1CF3 containing electrolyte compared to PFPOEPi containing counterparts. In summary, the addition of phospholane molecules represents an effective way to enable high voltage application for the graphite/NMC111 cell chemistry. However, certain differences may be observed as the addition of a CF3 group leads to a decreased cycling stability due to a significant charge transfer impedance, thicker CEI after formation as well as 100 charge/discharge cycles and a higher LiF content after 100 charge/discharge cycles compared to the cells containing PFPOEPi. With this approach we could show that careful tailoring of the structure of the electrolyte additive may significantly influence the formation and dynamics of the CEI, thus resulting in advanced high voltage cell performance. Furthermore, the comprehensive analysis as well as the synergy between theory and experiments, allow us to further understand the impact of certain functional groups on the overall cell performance and to further tailor the electrolyte additives towards advanced LIB and LMB application.

4. Experimental Section Synthesis of cyclic fluorinated phospholanes Synthesis of sodium 2,2,3,3,3-pentafluoropropan-1-olate To a suspension of sodium hydride (1.6 g, 66.2 mmol, 1.1 Eq) in abs. diethyl ether (50.0 mL) 2,2,3,3,3-pentafluoro-1-propanol (6.0 mL, 60.2 mmol, 1.0 Eq) was added dropwise at -80 °C. The reaction mixture was slowly warmed up to room temperature (RT) and stirred for 18 h. The crude product was purified by filtration, solvent evaporation and a white powder was obtained (yield: 10.2 g, 59.3 mmol, 99%). 1H-NMR (400 MHz, CD3CN) [ppm]: δ = 4.21 (qt, JH-F = 16.5 Hz, 2.1 Hz, 2H, CH2).

13C-NMR

(101 MHz, CD3CN) [ppm]: δ = 122.76 (t, JC-F = 37.2 Hz, 1C, CF2), 28 ACS Paragon Plus Environment

Page 28 of 40

Page 29 of 40 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Applied Materials & Interfaces

119.92 (t, JC-F = 36.5 Hz, 1C, CF3), 66.24 (t, JC-F = 21.8 Hz, 1C, CH2). 19F-NMR (376 MHz, CD3CN) [ppm]: δ = -83.84 (s, 3F, CF3), -128.48 (t, JH-F = 16.5 Hz, 2F, CF2). 19F-NMR (376 MHz, CD3CN) [ppm]: δ = -83.84 (s, 3F, CF3), -128.48 (s, 2F, CF2). Synthesis of 2-(2,2,3,3,3-pentafluoropropoxy)-1,3,2-dioxaphospholane Sodium 2,2,3,3,3-pentafluoropropan-1-olate (7.0 g, 40.7 mmol, 1.1 Eq) was dissolved in abs. THF (100 mL) and 2-chloro-1,3,2-dioxaphospholane (3.2 mL, 36.9 mmol, 1.0 Eq) was added dropwise at 0 °C. The reaction mixture was slowly warmed up to RT and stirred for 18 h. The crude product was purified by distillation (70 °C, 40 mbar) and a clear liquid was obtained (yield: 3.4 g, 12,7 mmol, 34%). 1H-NMR (400 MHz, CDCl3) [ppm]: δ = 4.31 – 4.06 (m, 4H, CH2CH2), 4.03 – 3.89 (m, 2H, CH2CF2). 13C-NMR (101 MHz, CDCl3) [ppm]: δ = 118.73 (tq, JC-F = 34.9 Hz, 284.1 Hz, 1C, CF3), 112.54 (qt, JC-F = 34.3 Hz, 256.0 Hz, 1C, CF2), 64.24 (d, JC-P = 8.7 Hz, 2C, CH2CH2), 59.54 (dt, JC-P = 17.8 Hz, JC-F = 27.5 Hz, 1C, CH2CF2). 19F-NMR (376 MHz, CDCl3) [ppm]: δ = 83.83 (s, 3F, CF3), -124.69 (dt, JH-F = 13.1 Hz, JP-F = 6.2 Hz, 2F, CF2). 19F{H}-NMR (376 MHz, CDCl3) [ppm]: δ = -83.83 (s, 3F, CF3), -124.69 (d, JP-F = 6.2 Hz, 2F, CF2). 31P-NMR (162 MHz, CDCl3) [ppm]: δ = 137.45 (hept., JC-P = 8.3 Hz, 1P, PO3). 31P-NMR (162 MHz, CDCl3) [ppm]: δ = 137.25 (dt, JP-F = 6.3 Hz, 1.5 Hz, 1P, PO3). Mass: [C5H6F5O3P] Calc: 240.00 g/mol, Meas: 241.00 g/mol. IC (Cl-): < LOD. Synthesis of 2-(2,2,3,3,3-pentafluorpropoxy)-4-(trifluormethyl)-1,3,2-dioxaphospho-lane To a mixture of 3,3,3-trifluorpropan-1,2-diol (15.0 g, 120.0 mmol, 1.0 Eq) and NEt3 (23.3 g, 240.0 mmol, 2.0 Eq) a solution of PCl3 (16.4 g, 120.0 mmol, 1.0 Eq) in abs. DCM/Et2O (10:1, 250 mL) at 0 °C was slowly added. Thereafter, the mixture was slowly warmed up to RT for 1 h and the completion of the addition was monitored by

31P-NMR

spectroscopy. The crude product was

filtrated from Et2NHCl and washed with Et2O. A solution of 2,2,3,3,3-pentafluorpropanol (18.0 g, 29 ACS Paragon Plus Environment

ACS Applied Materials & Interfaces 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 30 of 40

120.0 mmol, 1.0 Eq) and NEt3 (12.1 g, 120.0 mmol, 1.0 Eq) was added dropwise to the filtrate at RT and monitored by

31P-NMR

spectroscopy. The crude product was filtrated, the solvents

removed at RT/1000 mbar and the remaining liquid distilled at 27°C/10-2 mbar. A clear liquid was obtained as a diastereomeric mixture (yield: 5.3 g, 16 mmol, 14%). Low yield because of the compounds’ high volatility). 1H-NMR (400 MHz, CDCl3) [ppm]: δ = 4.77 – 4.70 (m, 1H, CHCF3), 4.40 – 4.16 (m, 3H, CHCH2, CH2CF2). 13C-NMR (101 MHz, CDCl3) [ppm]: δ = 123.31 (dq, JC-P = 3.6 Hz, JC-F = 279.6 Hz, 1C, CHCF3), 118.68 (tq, JC-F = 34.8 Hz, 287.0 Hz, 1C, CF2CF3), 112.33 (qt, JC-F = 36.8 Hz, 254.3 Hz, 1C, CF2), 73.09 (dq, JC-P = 9.2 Hz, JC-P = 34.0 Hz, 1C, CH), 63.89 (d, JC-P = 9.4 Hz, 1C, CHCH2), 60.01 (dt, JC-P = 8.4 Hz, JC-F = 27.8 Hz, 1C, CH2CF2). 19F-NMR (376 MHz, CDCl3) [ppm]: δ = -79.61 (d, JH-F = 6.5 Hz, 3F, CHCF3), -83.61 (s, 3F, CF2CF3), 124.57 (dt, JH-F = 12.8 Hz, JP-F = 6.9 Hz, 2F, CF2). 19F{H}-NMR (376 MHz, CDCl3) [ppm]: δ = 79.61 (s, 3F, CHCF3), -83.61 (s, 3F, CF2CF3), -124.57 (d, JP-F = 6.9 Hz, 2F, CF2). 31P-NMR (162 MHz, CDCl3) [ppm]: δ = 143.43 (hept., JC-P = 8.0 Hz, 1P, PO3).

31P{H}ss-NMR

(162 MHz,

CDCl3) [ppm]: δ = 143.43 (dt, JP-F = 7.0 Hz, 1.5 Hz, 1P, PO3).

Electrolyte and cell preparation 1M LiPF6 in ethylene carbonate (EC) and ethyl methyl carbonate (EMC) (1:1 by wt.) was used as a reference electrolyte (REF). 0.5 wt.% of the respective phospholane additives was added to the reference electrolyte. The graphite and NMC111 electrodes, provided by Custom Cells®, were dried for 12 h at 120 °C under vacuum. One layer of Separion® was used as a separator. Swagelok® T-cells (for cyclic voltammetry measurements), 2032-type coin cells (for electrochemical impedance spectroscopy) and house made pouch cells (constant current cycling) where assembled

30 ACS Paragon Plus Environment

Page 31 of 40 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Applied Materials & Interfaces

in either glove box (MBraun, H2O and O2