Understanding the Role of π-Conjugated Polymers as Binders in Enabling Designs for High-Energy/High-Rate Lithium Metal Batteries

Developing lithium-ion batteries with both high speci ﬁ c energy and high-power capability is a challenging task because of the necessity for meeting con ﬂ icting design requirements. We show that high-energy and high-rate capability can be achieved by using various π -conjugated p-dopable polymers as binders at the cathode and by lowering the mass fraction of all the inactive components of the cell. We report a lithium-metal battery that can deliver 320 Wh kg − 1 at C/2 using a mass-ef ﬁ cient cell design. To this end, three conducting polymers with different ionic and electronic conductivities have been studied; dihexyl-substituted poly(3,4-propylenedioxythiophene) (PProDOT-Hx 2 ), poly(3-hexylthiophene) (P3HT), and a new Random Copolymer (Hex:OE) (80:20) PProDOT. These conducting polymers are compared against a conventional polymer binder, PVDF. We show that under the mass-ef ﬁ cient conditions required for achieving high speci ﬁ c energy and rate capability, the conducting polymers play a crucial role by providing electronic and ionic conductivity, protection against rapid growth of solid electrolyte interphase (SEI), and access to a large electrochemically active surface area. Thus, the use of conducting polymers with appropriate molecular structure as binders opens a viable pathway to maximizing the speci ﬁ c energy and rate capability of lithium-ion battery cathodes.

The continuous development and improvement of high-energy and high-rate lithium-ion batteries (LIBs) is crucial to satisfy the ever-growing demand for lighter portable devices with long run-times. [1][2][3][4] This persistent need has been driving the scientific research towards new materials, compositions, and cell configurations. [5][6][7][8][9][10] To this end, conducting polymers (CPs) are attractive as additives/binders for electrodes in LIBs. These CP binders reduce the electrode resistance, protect the surface of the intercalation materials, and enhance lithium-ion transport, leading to increased rate capability, discharge capacity and cycle life. [11][12][13][14][15] Furthermore, pairing of a high-capacity cathode material with a lithium metal anode, to take advantage of the high theoretical capacity (3860 mAh g −1 ) and low density (0.535 g ml −1 ) of lithium, has been getting more attention. [16][17][18][19] Although there have been several reports demonstrating the benefits of CPs as binder/additives with cathode and anode materials, 15,[20][21][22][23][24][25][26][27] many of these studies employ low areal mass loadings of active material (thin electrodes, 1 to 5 mg cm −2 ), with high mass fractions of carbon additive (3 to 10%) and/or excessive amounts of electrolyte and anode material. For example, previous work from our team has explored the stability and electrochemical properties of dihexyl-substituted poly(3,4-propylenedioxythiophene) (PProDOT-Hx 2 ) 28 and poly(3-hexylthiophene) (P3HT) 13 as an electrode additive. These studies were on cathodes with ample amounts of carbon additive (6 wt%), an excess of electrolyte (5.5 μl mg −1 ), and a low areal mass loading (1.5 and 6 mg cm −2 ). Such "non-limiting" conditions, common in the scientific literature, differ significantly from those used in cells built for practical applications. Specifically, in the practical design of commercial LIBs, the amount of "inactive" materials is curtailed to the extreme, and a mass loading of 10-15 mg cm −2 is commonly used to maximize the specific energy. As a result, translating the improvements showed by "non-limiting" designs to realize a high energy cell with high power density is not often feasible despite the valid claims of a new material or an improved formulation. [29][30][31][32] Improvements directed at high-energy and high rate-capability must focus on "mass-efficient" designs where the mass of inactive materials namely everything other than the mass of the cathode and anode materials are minimized and the electrodes have a high areal mass loading.
In this study, we demonstrate how various π-conjugated pdopable polymers binders/additives can achieve a high specific energy/high-rate lithium metal-based battery at the cell level. Replacing common cathode binders such as polyvinylidene fluoride (PVDF) with CPs allows us to achieve an impressive specific energy (at the cell component level, excluding the cell packaging) of 320 Wh kg −1 at a practical high rate of C/2. Most importantly, we have found that the CPs have a multifunctional role in preserving the specific energy of the cell under "mass-efficient" conditions, compared to the "non-limiting" designs used in scientific studies. We have revisited the use of PProDOT-Hx 2 and P3HT as binders for a mass-efficient cell design, and we also introduce a third new polymer with enhanced ionic conductivity denoted as (Hex:OE) (80: 20) PProDOT. We have utilized extensively electrochemical impedance spectroscopy (EIS), differential capacity (dQ dV −1 vs V) analysis and characterization of electrode morphology to gain insight into the enabling role of the CPs in modifying the underlying chemical and transport phenomena for achieving high specific energy and high rate capability.

Experimental
Monomer and polymer synthesis.-The relevant molecular structures and results of physical characterization for all the polymers are included in the supplementary material.
Synthesis of P3HT (P1).-The polymer P3HT was prepared according to literature procedure. 33 In an oven dried 50 ml threenecked round bottom flask, monomer S1 (247 mg, 1 mmol), K 2 CO 3 (207 mg, 1.5 mmol) and NDA (52 mg, 0.3 mmol) were dissolved in 25 ml dry DMA to give a 0.04 M solution. The solution was then degassed for 10 min. Then 0.25 mol% Pd(OAc) 2 was added, and the reaction mixture was further degassed for 20 min. Then the reaction mixture was immersed into a pre-heated oil bath at 70°C for 48 h. After cooling to room temperature, a minimum amount of CHCl 3 was added to the reaction mixture and precipitated into a cooled 10% NH 4 OH/MeOH solution with rapid stirring. The polymer product was then filtered and purified by Soxhlet extraction with MeOH, hexanes, and CHCl 3 . The chloroform fraction was concentrated by evaporation and precipitated into chilled MeOH with rapid stirring. The polymer was finally filtered and dried overnight under high vacuum. Yield (after Soxhlet extraction): 60%, Mn = 13.5 kg mol −1 , Ð = 2.14. 1  Synthesis of PProDOT-Hx 2 (P2).-The polymer PProDOT-Hx 2 was prepared by modifying literature procedure. 28 In a 15 ml oven dried high-pressure vessel with a stir-bar, capped with an inverted rubber septum and cooled under a stream of nitrogen, was added K 2 CO 3 (133 mg, 0.96 mmol), PivOH (18 mg, 0.18 mmol), S2 (145 mg, 0.3 mmol) and S3 (97 mg, 0.3 mmol) under nitrogen. To the above mixture of reagents, 3 ml of 1:1 mixture of degassed CPME:DMA was added under nitrogen with a syringe followed by 15 min of degassing under nitrogen before finally adding the catalyst, Pd(OAc) 2 (3 mg, 2 mole%) to the reaction mixture. The rubber septum was then quickly replaced with a Teflon screwcap equipped with a rubber o-ring, and the vessel was placed in a preheated oil bath at 140°C for 22 h. After cooling to room temperature, a minimum amount of CHCl 3 was added to dissolve the solids followed by precipitation into a chilled 10% NH 4 OH/MeOH solution with rapid stirring. The polymer product was then purified via Soxhlet extraction using MeOH, hexanes, and CHCl 3 . The chloroform fraction was concentrated in vacuo and precipitated into cooled MeOH with rapid stirring which was subsequently filtered and dried overnight under vacuum at room temperature. The polymer was obtained as a purple solid. Yield Synthesis of (Hex:OE)(80:20) PProDOT random copolymer (P3).-The (Hex:OE) PProDOT random copolymer was prepared by following literature procedure. 28 In a 15 ml oven dried high-pressure vessel with a stir-bar, capped with an inverted rubber septum and cooled under a stream of nitrogen, was added K 2 CO 3 (122 mg, 0.9 mmol), P(t-Bu) 2 MeHBF 4 (15 mg, 0.06 mmol), NDA (31 mg, 0.18 mmol), S2 (145 mg, 0.3 mmol), S3 (58 mg, 0.18 mmol) and S5 (51 mg, 0.12 mmol) under nitrogen. Then 3 ml of 1:1 mixture of degassed CPME:DMA was added to the above mixture of reagents under nitrogen with a syringe followed by degassing for 15 min under nitrogen before finally adding the catalyst, Pd(OAc) 2 (7 mg, 5 mole%) to the reaction mixture. The rubber septum was then quickly replaced with a Teflon screwcap equipped with a rubber o-ring, and the vessel was placed in a pre-heated oil bath at 140°C for 20 h. After cooling to room temperature, the solids were dissolved in minimum amount of CHCl 3 and precipitated into a cooled 10% NH 4 OH/MeOH solution with rapid stirring. The polymer product was then purified using Soxhlet extraction using MeOH, hexanes, and CHCl 3 . The chloroform fraction was concentrated by evaporation and precipitated into cooled MeOH with rapid stirring which was subsequently filtered and dried overnight under vacuum at room temperature. The polymer was obtained as purple solid. Yield Conductivity measurements.-Interdigitated microelectrodes (IDM) were purchased from Metrohm Dropsens (DRP-G-IDEAU5-U20). Each microelectrode is composed of 250 digits with a digit length of 6760 μm and a gap of 5 μm between the digits. Polymer solutions were prepared by dissolving the polymer powder in 1,2-dichlorobenzene (99%, Sigma-Aldrich). The solution was then heated under argon for two hours at 40°C. 5 μl of the prepared solution was spin-coated on the gold IDM at 1000 RPM for 30 seconds to produce a 50 nm polymer film measured by SEM. 28,36 The prepared electrodes were then annealed under vacuum at 110°C for two hours, then transferred to an argon glovebox. All electrochemical tests were performed in an argon glovebox at room temperature. The details of the experimental technique for electrochemical doping and the determination of ionic and electronic conductivity has been previously reported by our group. 36 Briefly, to electrochemically dope the polymers, a 3electrode cell was assembled using the polymer on the IDM as the working electrode, lithium foil as the counter and reference electrodes, and 1 M LiPF 6 in a mixture of 1:1 by volume of ethylene carbonate and diethyl carbonate (EC/DEC) as the electrolyte. EIS of the 3-electrode cell was measured after electrochemical doping at each potential to obtain the ionic conductivity of the polymer films between 100 mHz and 100 kHz at a sinusoidal excitation of ±10 mV peak-to-peak. To determine electronic conductivity, impedance was measured between the two terminals of the gold electrodes at open circuit potential between 100 mHz and 100 kHz at a sinusoidal excitation of ±10 mV peak-to-peak.
Cathode preparation.-Lithium-ion battery cathodes consisting of lithiated nickel cobalt aluminum oxide (LiNi 0.8 Co 0.15 Al 0.05 O 2 (NCA) and π-conjugated polymer were prepared by mixing pristine powders of NCA (NEI corporation), Super P carbon black (MTI), multiwalled carbon nanotubes (CNTs, Cnano, OD × ID × L: 25 nm × 10 nm × 10 μm), and the π-conjugated polymer/1,2-Dichlorobenzene (ODCB) (50 g l −1 ) in a weight ratio of 95:0.5:0.5:4. The cathode powder was gently mixed in a mortar and pestle before being added to the ODCB solution. Control NCA-PVDF electrodes were fabricated by dissolving polyvinylidene fluoride (PVDF, MTI) in N-methyl-2-pyrrolidone (NMP, Sigma-Aldrich) and mixed with NCA, Super P carbon black, and CNTs in the same ratio. The addition of CNTs creates a mesoscale porosity that improves the electrolyte penetration into the electrode structure. [37][38][39] The resulting slurry was coated onto an aluminum foil using the doctor blade technique. 40 The electrode films were vacuum-dried overnight, roll pressed (12″ vise-mount slip roll T10727, Grizzly Industrial) and then punched to 14 mm diameter discs. Areal active material mass loading of the electrodes was approximately 14 mg cm −2 .
Lithium anode preparation.-For the cells under mass-efficient conditions, metallic lithium chips (MTI, D × T:16 mm × 0,6 mm) were polished, hammered and roll-pressed between two layers of pure nitrile sheets to obtain an extremely thin lithium foil (46 μm thick). The foil was carefully punched into 14 mm diameter lithium disks. The thinning process was repeated multiple times until obtaining the desired weight of lithium based on an N/P ratio of ≈3. For the non-limited cells, lithium chips with a weight of ≈40 mg were rendered shiny by scratching the surface with a tweezer and used as is. Lithium handling was carried out in an argon-filled glove box (VAC system 60387, NexGen 2 P, Vacuum Atmospheres Company) with less than 0.5 ppm of moisture and 0.2 ppm of oxygen.
Electrolyte.-All cells utilized commercial carbonate electrolyte consisting of 1.0 M Lithium hexafluorophosphate in 1:1 in ethylene carbonate (EC) and diethyl carbonate (DEC) from Sigma-Aldrich.
Electrode characterization.-The microstructure of the electrodes was examined with a scanning electron microscope (Nova NanoSEM 450 Field Emission SEM). Elemental mapping (EDS) was performed using an energy dispersive X-ray spectrometer JSM-7001F-LV. SEM images for each electrode type in the scale of 100 microns and 500 nm. The macro and meso pores at the surface were characterized using ImageJ (open source software NIH) with a Bandpass filter with a maximum of 500 pixels and a minimum of 6 pixels for each image followed by a particle analysis that included the image edges and bare outlines. Imbibition/Archimedes' method was utilized for estimating the mean effective porosity of the electrodes using the average value of triplicate measurements using isopropyl alcohol as the fluid. The imbibition method to determine the pore volume consists of saturating the dried electrode with a liquid of known density followed by weighing. The Archimedes' method is used to determine the electrode volume by immersing the electrode in a fluid of known density (subtracting the volume of the current collector). The mean effective porosity is then obtained by dividing electrode volume by the pore volume.
Cell assembly.-CR2032 coin-type cells were fabricated with the NCA-π-conjugated polymer electrode as the working electrode, metallic lithium as the counter electrode, and Celgard 2325 (PP/ PE/PP) as the separator. The amount electrolyte and lithium for the mass-efficient cells was calculated according to the N/P and E/C ratio of ≈ 3. N/P was defined as the negative/positive electrode areal capacity, and E/C defined as electrolyte to capacity ratio in g Ah −1 with an electrolyte density of 1.2 g ml −1 . Non-limited cells had an N/P and E/C ratio of approximately 45 (≈40 mg of lithium and 500 μm thick) and 17 (≈50 μl), respectively. Cells were crimped with a pressure-controlled electric crimping machine (MSK-160E, MTI) between 0.9-1.0 metric tons (T). Cell assembly was performed in an argon-filled glove box (VAC system 60387, NexGen 2 P, Vacuum Atmospheres Company) with less than 0.5 ppm of moisture and 0.2 ppm of oxygen.
Electrochemical testing and analysis.-Galvanostatic chargedischarge (GCD) cycling, rate capability tests, and electrochemical impedance spectroscopy (EIS) measurements of all coin cells were conducted using a battery test station BCS-815 series potentiostat/ galvanostat with EIS (BioLogic) at room temperature (23°C). All cells were subjected to two formation cycles at C/10 chargedischarge before additional rate-capability and extended cycling tests. For the extended cycling studies, the cells were charged at a constant rate of C/5 and discharged at C/2. C-rate is based on the reversible capacity of NCA, where 1 C corresponds to 160 mA g −1 .
For the rate capability tests with active areal loading of ≈14 mg cm −2 , these cells were charged at C/5 and discharged in increments of C/5, C/2, 1 C, 2 C, 3 C, 4 C, and C/5. The charging potential cut-off for all cells was at 4.2 V, while the discharge cut-off was at 2.7 V. The EIS response was measured at open circuit voltage in the frequency range of 0.1 Hz to 100 kHz using a sinusoidal excitation amplitude of ±5 mV (peak to peak). The impedance response was analyzed using the Zfit software (EC-Lab/BT-Lab −2018 version). The fitting was executed utilizing the non-linear complex downhill Simplex method with a previous randomized step with 5000 iterations and more than 10000 iterations for the method with a weight based on the |z|. Specific capacity (mAh g −1 ) was based on the weight of the NCA active material as the capacity contribution from the π-conjugated polymer was negligible. Specific energy at the cell-level (Wh kg −1 ) was defined as E cell = E int /W cell , where E int is the calculated energy by integrating the area under the curve of a voltage vs capacity plot for each GCD cycle, and W cell , the total weight of cell components that includes the current collector, cathode, electrolyte, separator, and anode, but excludes the cell casing. Thus, the results presented here can be extended to various types of cell configurations by appropriate inclusion of external packaging mass as per the cell size.

Results and Discussion
Cathodes with new π-conjugated polymers for a high-energy lithium battery.-A comparison of three types of CPs allowed us to understand the benefits of the role of polymer structure in achieving high specific energy and rate capability. Accordingly, in addition to PProDOT-Hx 2 and P3HT, we synthesized a new variety of dihexylPProDOT polymer with 20% oligoether (OE) chains referred to here as (Hex:OE)(80:20) (Figs. S1-S5). 13,28 The new (Hex:OE) (80:20) underwent reversible electrochemical doping in the potential range of 3.2 to 4.5 V vs Li + /Li (Figs. 1a-1c). The slight asymmetry of the peaks in the voltammograms during oxidation and reduction is attributed to the reversible morphological changes that occur in the polymer film due to solvent uptake during cycling 28 When doped, (Hex:OE)(80:20) had a maximum electronic conductivity of ∼10 −2 S cm −1 and ionic conductivity of ∼10 −6 S cm −1 over the potential range of 3.5 V to 4.2 V vs Li + /Li as measured by the methods developed previously in our group. 36 The addition of the OE side group in the dihexylPProDOT structure enhanced the ionic conductivity compared to PProDOT-Hx 2 and P3HT (Fig. 1d). The polymers stayed doped with small changes to the conductivity values over the potential range of 3.3 V to 4.2 V vs Li + /Li that encompasses the potential window of cycling of typical cathode materials such as NCA and NMC622.
We have tested these polymers with NCA as the electrode active material due to NCA's excellent reported stability, 41,42 high theoretical charge capacity (265 mAh g −1 ), 41 high electronic conductivity (∼10 −4 S cm −1 lithiated, ∼10 −3 S cm −1 delithiated), 43 high lithiumion diffusivity (∼2 × 10 −10 cm 2 s −1 ) 43 and commercial availability. 42 The NCA cathode was formulated with 95% of NCA, 1% of carbon, and 4% of the CP with an areal mass loading of approximately 14 mg cm −2 to reflect a commercially-relevant electrode composition that maximized the fraction of active materials. With the goal of increasing the specific energy, the mass of all the additional electrode and cell components was reduced significantly, viz., a negative/positive electrode capacity ratio (N/P) of ≈3 with metallic lithium as the negative electrode, and an electrolyte to cathode capacity ratio (E/C) of ≈3 g (Ah) −1 ( Table I). The properties of these Li-NCA-CP cells with this mass-efficient formulation were compared with cells of the same electrode composition and loading but with excess electrolyte and excess anode material (lithium) denoted as "non-limited" cells. The anode and the electrolyte contributed to just 28% in the mass-efficient design whereas it was 75% in the non-limited type cells (Fig. 2a). The mass-efficient and non-limiting formulations were also assembled with the conventional PVDF binder for comparison.
The specific energy for the Li-NCA-CP cells of the mass-efficient type with Li-NCA-PProDOT-Hx 2 , Li-NCA-(Hex:OE)(80:20), Li-  NCA-P3HT, and Li-NCA-PVDF cells reached 359, 336, 336, and 312 Wh kg −1 , respectively, delivering over 3 times the specific energy of their non-limited counterparts (Fig. 2b). The CPs not only provide a higher mixed (electronic and ionic) conductivity and reversibility 13,15,23 but can also form a protective thin film over the cathode particles. [13][14][15] We have found that the replacement of the conventional PVDF binder with CPs reduced the electrode resistance and the growth of the solid electrolyte interphase (SEI) on the NCA particles. The benefit of the mixed conductivity exhibited by the CP electrodes became evident as the discharge rate is increased; at the C/2 rate, cells of the mass-efficient type with NCA-PProDOT-Hx 2 , NCA-(Hex:OE)(80:20), NCA-P3HT and NCA-PVDF cathodes delivered an impressive initial specific energy of 323, 271, 294, 250 Wh kg −1 respectively and the average capacity decay per cycle after the formation cycles was 1.7, 1.5, 2.2, and 3.1%, respectively. However, for the cells of the non-limited type the specific energy ranged from 90 to 70 Wh kg −1 although the cathodes exhibited higher overall values of specific capacity than the massefficient type (Figs. S6-S9). While the high mass fractions of the inactive components in the non-limited cells ensured good utilization of the material, low electrode resistance and longer cycle life, the specific energy of the cell was reduced in proportion to the mass fraction of the inactive components in the electrodes. Thus, the specific energy of the cell was low because of the large mass of the inactive cell components. Unlike the mass-efficient cells, the capacity reduction with increasing rate of discharge in the nonlimited cells was not steep. Thus, there is a trade-off between the mass fraction of the inactive materials, their role in enhancing the performance, and the achievable specific energy. In this design optimization of the N/P ratio, the active material loading, porosity/ tortuosity of the electrode, electrode thickness, and the electrolyte to capacity (E/C) ratio are the crucial parameters. 29,31,[44][45][46] Furthermore, GCD cycling of solely the CPs (without NCA) under lean electrolyte conditions over 100 cycles in the potential range of 2.7 V to 4.2 V vs Li + /Li exhibited a capacity retention of 98.3, 99.8, and 99.5% for PProDOT-Hx 2 , (Hex:OE)(80:20), and P3HT respectively, confirming the excellent cycling properties and stability of the CPs studied here (Fig. S10).
Understanding the impact of CPs on the cell's internal processes.-We have measured the impedance response for all the non-limited and mass-efficient cells (Fig. 2b) as a function of state of charge (SOC) during charge and discharge for the 1st cycle (after formation) and the 26th cycle. The percentage SOC during charge is defined as the percentage of the charge capacity until the cut-off of 4.2 V vs Li + /Li. The percentage SOC during discharge is defined as the percentage of the discharge capacity delivered up to a cut-off voltage of 2.7 V vs Li + /Li. On the first galvanostatic chargedischarge (GCD) curves (Figs. 3a, 3b) the yellow circles indicate the various SOC values at which the electrochemical impedance (EIS) was measured as a function of frequency for the mass-efficient Li-NCA-PProDOT-Hx 2 and Li-NCA-PVDF cells. We observed a higher overall polarization for the PVDF cell compared to the PProDOT-Hx 2 cell due to the mixed electronic and ionic conductivity that the conducting polymer provides over the insulating PVDF binder. The EIS response was fitted to a Randles-type equivalent electric circuit (EEC) model with an added Voigt element (Fig. 3c). The EEC model is composed of the ohmic resistance of electrode and electrolyte (R 1 ), a Voigt element representing the resistance and capacitance of the SEI layer at the electrodes (R 2 and C 1 ), the charge-transfer resistance (R 3 ) in parallel with the double layer capacitance (C 2 ) in series with a semi-infinite Warburg element (W) representing the diffusion processes associated with the electrochemical intercalation reaction. The corresponding impedance measurements at various SOCs during charge are presented as colored spheres in the two 3D-Nyquist plots (Figs. 3d, 3e) along with the black lines that correspond to the simulation fit of the EEC model. The 3D-Nyquist plots for all mass-efficient and non-limited CPs cells are shown in Figs. S11-S18.
The values of impedance for Li-NCA-PProDOT-Hx 2 cells were 3 to 5 times lower than those for the Li-NCA-PVDF cells. The impedance decreased at higher cell voltage values, for both the mass-efficient and non-limited cells. The impedance elements fitted to the EEC (Fig. 3c) were found to vary with the SOC (Fig. 4). Tables SI and SII show the best fit of the EEC elements at mid-SOC for all the Li-NCA-π-conjugated polymer cells under mass-efficient and non-limited conditions. The ohmic resistance R 1 was in the range of 5 to10 ohms for all the freshly-prepared cells and did not vary significantly with SOC. However, after 26 cycles the massefficient cells exhibited an increase in their ohmic resistance that we attribute to electrolyte insufficiency. The limited amount of electrolyte in these cells is mainly consumed by reaction at the lithium electrode due to the continuous growth of the SEI during repeated cycling. [47][48][49] Thus, by avoiding electrolyte consumption at the anode we can ensure a long-cycle life cell with a mass-efficient design. The value of R 1 following cycling for the mass-efficient Li-NCA-PVDF cell (red dots) was about two times larger than that of the Li-NCA-CP cells, indicating a higher rate of electrolyte consumption in the PVDF cells compared to the mass-efficient cells with CPs. This finding is consistent with the protective role of the CPs coated on the NCA particles forming a more electrochemically-stable SEI, preventing the electrolyte consumption at the cathode. [13][14][15] In stark contrast, in the non-limited cells because of the abundance of electrolyte, we observe only a slight increase in the ohmic resistance after cycling.
The value of R 2 for the NCA-PVDF electrodes (red and pink dots) was 4 to 10 times higher than the cells with CPs, confirming the benefit of formation of a stable SEI over the NCA particles with the CP-based binders (Fig. 4b). However, we observe that after 26 cycles there is a slight increase in R 2 for just the mass-efficient Li-NCA-(Hex:OE)(80:20) cell (olive dots). These results suggest that the differences in electronic conductivity (Fig. 1d) and the effect of degree of swelling of the polymers became evident when the availability of the electrolyte is restricted. Figure 4c shows the changes of R 3 , the polarization resistance associated with the electrochemical reaction kinetics at the electrode. For all the polymer binders R 3 decreased with increasing SOC. Further, the CP containing cells exhibited 70% lower values compared to the PVDF cells. The polarization resistance, R 3 , is related to the exchange current density and electrochemically active surface area by R 3 = RT (nFA o i o ) −1 , where i o is the exchange current density, A o is the active surface area, F the Faraday constant, n is the number of electrons transferred, T is the absolute temperature, and R is the ideal gas constant. Thus, the decrease in R 3 with increasing SOC is consistent with the increase in i o observed with the charging of NCA. 42,50,51 However, after 26 cycles there is a significant increase in R 3 for the PVDF cells indicating that the active area (A o ) of the electrode available for electrochemical intercalation of lithium decreased as a function of cycling, most likely due to the continuous growth of a thicker/insulating SEI layer on the cathode. The lower values of R 3 for the CP cathodes in comparison to PVDF-containing cathodes is attributed to the electronic conductivity of the conjugated polymers that results in an increase of the electrochemically accessible surface area.
We attribute the differences in R 3 observed among the various Li-NCA-CP cells, to the porosity variations of the electrodes. From the SEM studies on the electrodes, we found that all electrodes consisted of macro-spherical agglomerates of NCA with a diameter ranging from 4-10 μm, where each sphere is comprised by smaller NCA particles (100 to 500 nm in size) creating a hierarchical morphology (Figs. 5a-5d). Despite the same mass fraction of the various constituents and the uniform distribution of electrode components, the different CPs gave rise to different values of final porosity and pore sizes. The primary (macro) and secondary (meso) pore distribution of the NCA electrodes, analyzed using the ImageJ software (Figs. S19-S27), showed that NCA-P3HT electrodes had a larger fraction of macro and meso pores compared to the other CPs and PVDF-based electrodes. The effective porosity (as measured by the imbibition/Archimedes' method, Table SIII) for the NCA-P3HT, NCA-PProDOT-Hx 2 , NCA-(Hex:OE)(80: 20), and NCA-PVDF electrodes was 80%, 76%, 68%, and 61% respectively. Using Energydispersive X-ray Fluorescence Spectroscopy (EDS) we confirmed the uniform distribution of the CPs (Figs. S28-S35). Therefore, we conclude that the lower value of R 3 for PProDOT-Hx 2 (blue and turquoise dots) and P3HT (purple and magenta dots) was attributable to the higher porosity of these electrodes (Fig. 4c). During the 1st cycle, CP cells showed C 1 values of 100-1000 times that of PVDF cells. The higher value of C 1 was attributed to the ability of the CP layers to store charge during the polymer doping process, not possible with the PVDF cells (Fig. 4d). The value of C 1 represents the faradaic pseudo-capacitance of the conducting polymer with a value of about 0.01 F cm −2 consistent with the polymer being present only to the extent of 5% of the total electrode mass. Upon cycling the increasing thickness of the SEI on the cathode leads to a decrease of the values of C 1 for all the polymers. Changes to the active surface area for the charge-transfer reactions is reflected in C 2 . During the first cycle PVDF electrodes tended to have higher values of C 2 attributed to higher active area due to a more active SEI formation (Fig. 4e). For all the cells, we found the Warburg coefficient, S 1 , increased with decreasing SOC values (Fig. 4f). S 1 is inversely related to the diffusion coefficient of the lithium ion in the intercalation material. At lower SOCs, when NCA is highly lithiated, only a limited number of lattice sites are available for lithium-ion diffusion. With progressive de-lithiation of the NCA, more pathways become available for diffusion leading to an increase in the apparent diffusion coefficient of the lithium ion. Additionally, towards the end of discharge, the PVDF cells exhibited a higher value of S 1 compared to the CP electrodes. This observation suggests a higher interfacial area of contact for the mass transport of lithium ions in the CP cells compared to the PVDF cells.
We have also characterized the impedance response of LiNi 0.6 Mn 0.2 Co 0.2 O2 (NMC-622) with PProDOT-Hx 2 and PVDF binders (Figs. S36-S38). The data shows that electrodes made with PProDOT-Hx 2 exhibited 6 to 7 times lower charge transfer resistance compared to the electrodes made with PVDF confirming the benefit of utilizing conducting polymers with other lithium-ion intercalating materials.
Phase equilibria and polarization via differential capacity analysis.-To analyze cell degradation (capacity loss) as a function of cycle number, we determined the differential capacity (dQ dV −1 vs V) from the GCD data for all the Li-NCA-CP cells and Li-NCA-PVDF cells. Differential capacity is indicative of the phase equilibria and changes to the internal resistance of NCA during de-lithiation and lithiation. [52][53][54][55] The shift in peak position and changes to peak heights can provide insight into the structure evolution of NCA during lithium ion intercalation/deintercalation. 54,56,57 During charging in the first cycle of formation for Li-NCA-PProDOT-Hx 2 cell showed (Fig. 6a, red line) a peak between 3.55 and 3.65 V vs Li + /Li attributed to the phase coexistence of a hexagonal (H1) and monoclinic (M) lattice (H1 + M), followed by a second peak between 3.95 and 4.05 V associated with the coexistence of M with a second hexagonal (H2) lattice, (M + H2). 53,58 During discharge we observed a peak between 4.2 and 4.12 V attributed to the phase equilibria of a third hexagonal lattice (H3) with H2, (H2 + H3), followed by peaks for the coexistence of M and H2 between 4.05 and 3.85 V, and the (H1 + M) equilibria between 3.85 and 3.55 V. The results of differential capacity analysis indicate that the NCA particles coated with CPs retain their phase composition over cycling, suggesting a protective role of the CPs to mitigate the material transformation of the NCA particles. 51,52,59,60 In situ XRD studies combined with XPS would be necessary to confirm this hypothesis. We observed that as a function of cycling, all cells in Fig. 6 exhibited a significant shifting of the discharge peak for H1 + M phase equilibria to lower potentials that can be correlated with a higher overall impedance (Figs. S11-S18). This shift is most significant with the mass-efficient PVDF cells (Fig. 6d). We found that the non-limited cells maintained their phase transitions over cycling except for the non-limited-PVDF cell that exhibited a notable decrease in the height of the peaks and shift of peak positions consistent with the observed capacity fade and higher impedance of the PVDF containing cells. We also found that the mass-efficient-type cells with PProDOT-Hx 2 (Fig. 5a) and (Hex:OE) (80:20) (Fig. 6b) cells maintained their discharge peaks over cycling. However, with the P3HT cells (Fig. 6c) their charge and discharge peaks became smaller with cycling, consistent with a higher rate of capacity fade and impedance increase (Fig. 4a) for these cells. While all the CPs appear superior to PVDF in ensuring phase stability, PProDOT-Hx 2 and (Hex:OE)(80:20) seem to provide beneficial properties over P3HT. Thus, under lean electrolyte conditions, regions of the P3HT electrode could become devoid of electrolyte due to its higher porosity (Figs. S19-S27 and Table SIII). Such a situation would lead to insufficient electrolyte availability to all parts of the electrode during cycling, and consequently a lower utilization of the active material.
Rate capability under mass-efficient and non-limited conditions.-We studied the discharge specific energy and specific capacity as function of cycle number and discharge rate for the Li-NCA-CP and Li-NCA-PVDF cells under mass-efficient and nonlimited conditions (Figs. 7a, 7b). Figures 7c-7f are the corresponding GCD curves for the mass-efficient cells from Fig. 7a. The GCD curves for the non-limited cells in Fig. 7b are provided in Fig. S39.
For all the mass-efficient cells (Fig. 7a) we found a comparable utilization of the cathode materials during formation. However, as the C-rate is increased the beneficial role of the CPs and the electrode morphology became apparent. Even at C/2 we observed a dramatic decrease in the capacity for the Li-NCA-PVDF cell. Two main factors contribute to this drop in capacity; a lower electrode conductivity, and a more aggressive consumption of the limited amount of electrolyte during the SEI build-up that is consistent with the impedance results (Figs. 4a, 4b). At 1C we observe that the PProDOT-Hx 2 and (Hex:OE)(80:20) cells retain an impressive specific energy of ≈200 Wh kg −1 followed by P3HT (150 Wh kg −1 ) while the PVDF cells stay at a low value of 50 Wh kg −1 . Assuming that all cells have a similar consumption of electrolyte at the lithium electrode, the differences among the CP cells can be attributed to the mixed ionic/electronic conductivity and the morphological differences of the NCA electrode. Specifically, higher ionic conductivity is an advantage at high rates under electrolyte-lean conditions. Thus, the CP binders are particularly suited for the sustaining electronic and ion transport as the electrolyte is consumed, a feature and benefit that is unavailable with PVDF as a binder. While porosity is usually an advantage, in the NCA-P3HT electrode the higher porosity and wider macro pores could be detrimental because of the limited amount of available electrolyte. Consequently, there is an early onset of impedance increase with the mass-efficient P3HT cells especially when discharged at the higher rates (C/2-1C). The impact of re-distribution and consumption of the electrolyte due to the higher porosity of the P3HT cell can also be observed in the higher ohmic resistance after 26 cycles (Fig. 4a), and the peak shifts following cycling in the differential capacity analysis (Fig. 6c). At the 2C rate these effects are amplified further. The cells with (Hex:OE)(80:20) showed the highest utilization and retention of all the cells due to its enhanced ionic conductivity and lower porosity. These limitations were not observed when the cells were assembled with excess of electrolyte (Figs. 7b and 2b). Cells made of the CPs with higher electronic conductivity (PProDOT-Hx 2 and P3HT) retained a higher capacity as rate was increased. Therefore, we have identified that the benefits of low electrode porosity and high polymer ionic conductivity are the critical parameters affecting rate capability under mass-efficient conditions, whereas the electronic conductivity is the main factor under non-limited conditions.

Conclusions
We show that lithium metal batteries with CPs as cathode binders can deliver a high specific energy (at the cell component level excluding packaging) of 320 Wh kg −1 at C/2 and 208 Wh kg −1 at 1C. The high-specific energy/high-rate performance is achieved by reducing the weight of the inactive components enabled by the benefits of higher electronic/ionic conductivity and protection of the intercalating particles by the π-conjugated polymers in the cathode. The benefits of polymer structure and conductivity was exemplified in the behavior of PProDOT-Hx 2 , (Hex:OE)(80: 20), and P3HT at higher discharge rates. The electrode morphology characterization, rate capability testing, EIS analysis, and differential capacity analysis indicated that under the conditions of limited electrolyte availability, polymers with higher-ionic conductivity such as (Hex: OE)(80:20) are preferred for achieving higher specific energy and higher rate. Under these mass-efficient conditions, the electrodes of higher porosity with wide macro pores without the protection of the NCA particles leads to poor cell performance. In contrast, for the cells where the electrolyte content is not limited, the pore size and porosity differences of the electrodes did not have a significant contribution to the rate capability whereas the higher electronic/ionic conductivity were important. The rigorous attention to mass-efficient cell designs and control of parameters such as the N/P and E/C ratios during cell fabrication facilitate the translation of our findings to practical designs. With the progress in the design of less reactive electrolytes and protected lithium anodes, the conducting-polymerbased cathodes provide a new pathway to lithium-ion batteries with high specific energy, high rate capability, and long cycle life.