Effect of Antagonistic Binder–Catalyst Interactions on Catalytic Activity in Lithium–Sulfur Batteries

ABSTRACT

Lithium‐sulfur (Li‐S) batteries are a promising next‐generation energy storage solution, as they can reduce reliance on critical transition metals while offering high energy densities. However, their deployment is hindered by low sulfur utilization and the formation/diffusion of lithium polysulfides (LiPSs). While transition‐metal catalysts and polymeric binders have been independently developed to enhance redox kinetics and LiPS adsorption, their mutual compatibility has remained largely unexplored. We show here that binder‐catalyst interactions can significantly impact catalytic performance. Employing TiO₂ as a generic catalyst, the electrochemical performance is shown to depend strongly on the binder environment. TiO₂ paired with lithiated polyacrylic acid (LiPAA) shows benign interactions, resulting in enhanced cycle life. In contrast, pairing TiO₂ with protonated PAA produces antagonistic interactions that hinder Li₂S growth. A mechanistic analysis unveils that the carboxylic H atom in PAA promotes COO⁻ coordination to Ti sites, occupying catalytic centers and suppressing LiPS adsorption, increasing charge transfer and diffusion resistances. This phenomenon is observed across multiple catalysts, indicating that COOH‐functionalized binders may broadly hinder catalytic activity. Overall, this study underscores the need for holistic cathode design and identifies binder‐catalyst compatibility as an important parameter for high‐performance Li‐S batteries.

The binder is shown to interact antagonistically with the catalyst in a sulfur cathode, impeding Li₂S_x adsorption onto catalytic active sites. This hinders electrochemical performance and prevents the effective deployment of catalysts. This study showcases catalyst‐binder compatibility as a key parameter for cathode design in lithium‐sulfur (Li‐S) batteries.

1. Introduction

The continual increase in global energy demand creates a need for innovative energy storage technologies. Rechargeable lithium‐based batteries are already widely utilized due to their high energy efficiency and commercial availability. However, their large‐scale deployment is constrained by their dependence on less abundant, expensive metals, such as cobalt and nickel, posing humanitarian and supply chain challenges. Lithium‐sulfur (Li‐S) batteries are a viable alternative, employing earth‐abundant and inexpensive sulfur as the cathode active material instead [1].  Further, owing to the high capacities of sulfur (1,672 mA h g⁻¹) and lithium metal (3,862 mA h g⁻¹), Li‐S batteries offer a promising pathway to achieve high energy densities beyond 500 W h kg⁻¹, providing potential for applications in transportation and aerospace sectors.

Although Li‐S batteries are promising, their commercial relevance is limited by low sulfur utilization and insufficient cycle life when minimizing deadweight components [2, 3]. Particularly, owing to the insulating nature of sulfur, its conversion kinetics is slow, leading to low utilization. Furthermore, soluble lithium polysulfide (LiPS) intermediates are formed during charge/discharge. These species can diffuse to the anode, causing significant lithium‐metal corrosion and active material loss, limiting the cycle life. To improve sulfur utilization and mitigate the issues of polysulfide generation, researchers have paid major efforts toward developing materials that can chemically adsorb LiPS intermediates, while rapidly converting them to insoluble Li₂S. These materials include polymeric binders and transition‐metal catalysts, which may adsorb LiPS and/or catalyze their conversion.

Polymeric binders are an essential component in the cathode, providing structural integrity for the electrode, especially as sulfur expands ∼ 80% upon conversion to Li₂S. Suitable binders must have sufficient adhesion and cohesion abilities. Furthermore, effective binders contain functional groups that can also adsorb LiPS, retaining them at the cathode, or even accelerating their conversion [4, 5, 6]. Many aqueous binders are suitable for this role. Particularly, polyacrylic acid (PAA) is a common binder employed in Li‐S batteries, owing to its COOH functional groups and its commercial availability; the COOH group enables hydrogen bonding to enhance adhesion, while effectively adsorbing LiPS [7, 8]. Owing to these favorable properties, many binders have also been developed with PAA as an important component [9, 10, 11].

Meanwhile, transition‐metal catalysts can also adsorb LiPS and catalyze their conversion. In the past decade, significant strides have been made in developing catalysts for Li‐S batteries [12, 13, 14]. Researchers have explored a wide range of transition‐metal compound compositions to achieve the optimal binding energy between the catalyst and LiPS, thereby maximizing catalytic activity. Compounds explored include borides, carbides, nitrides, oxides, phosphides, sulfides, and selenides [14, 15, 16, 17, 18, 19]. Notably, researchers have recently utilized the vast amount of data to build machine learning models, identifying CrB₂ as an optimal catalyst with peak catalytic activity [20]. Although the research field has focused significant efforts on catalyst development, few studies have investigated the broader integration of a catalyst into the cathode. A recent study from our group highlighted the need for concurrent electrode‐catalyst design [21].

This work further emphasizes the need for a holistic electrode design, revealing binder‐catalyst compatibility as an important factor to consider. The performance of a TiO₂ catalyst was investigated with an acidic PAA binder compared to a neutral lithiated PAA (LiPAA) binder, and major differences in catalytic activity were observed. When TiO₂ is added to a cathode with PAA, antagonistic interactions arise between the binder and catalyst, thereby hindering the catalytic activity of TiO₂. To our knowledge, this is the first observation of antagonistic binder‐catalyst phenomena. Overall, this study highlights the need to develop binders suitable for catalyst implementation.

2. Results and Discussion

To demonstrate the impact of catalyst and binder interactions, a generic TiO₂ catalyst was selected for testing. TiO₂ is a semiconductor and extensively studied for photocatalytic and energy storage applications; it has also been demonstrated in Li‐S batteries as a compound with good LiPS adsorption abilities and adequate catalytic activity [22, 23]. TiO₂ nanoparticles were synthesized by a hydrothermal process, as reported in the literature (Figures S1 and S2) [24]. Then, cells were assembled with TiO₂ added to the cathode, while employing PAA or LiPAA as a binder. Thereafter, the electrochemical cycling performance was evaluated.

The cycling performance of cells with TiO₂ is shown to be significantly affected by the binder selection. As shown in Figure 1a, when employing a LiPAA binder, the addition of TiO₂ to a sulfur‐ketjenblack electrode (S/KB/TiO₂) enhances the initial capacity and cycling stability. The LiPAA‐S/KB/TiO₂ cell improves the initial capacity to 790 mA h g⁻¹ at a rate of C/2 compared to 768 mA h g⁻¹ for the baseline LiPAA‐S/KB cell. After 200 cycles, a capacity of 578 mA h g⁻¹ is retained for the S/KB/TiO₂‐LiPAA cell with a capacity retention of 73.2% compared to 486 mA h g⁻¹ for the LiPAA‐S/KB cell with retention of 63.2%. Furthermore, the average Coulombic efficiency (CE) over 200 cycles increases to 99.4% for LiPAA‐S/KB/TiO₂ compared to 99.0% for LiPAA‐S/KB, suggesting that TiO₂ can effectively adsorb LiPS and mitigate shuttling (Figure S3a). This showcases TiO₂ as a functional compound that improves the cycle life, consistent with other studies in the literature [22, 23].

FIGURE 1.

Electrochemical cycling performance of Li‐S batteries with TiO₂ catalyst and PAA or LiPAA binder. Long‐term cycling performance at a rate of C/2 (1C = 1000 mA h g⁻¹) for TiO₂‐incorporated cells where the cathode was prepared with an (a) LiPAA and (b) PAA binder. Cell polarization (ΔE_{50% SOC}) at a C/2 rate over various cycles with the (c) LiPAA and (d) PAA binders. Rate performance of TiO₂‐integrated cells with the (e) LiPAA and (f) PAA binders. For (a)–(f), the cycling performance and cell polarization are reported as the average of two to three duplicate cells, and the dashed lines or error bars indicate the standard deviation.

However, when TiO₂ is paired with a PAA binder instead, the cycling performance is hindered (Figure 1b). Whereas a PAA‐S/KB cell achieves a maximum capacity of 785 mA h g⁻¹ at a rate of C/2, the PAA‐S/KB/TiO₂ cell achieves a lower maximum capacity of 606 mA h g⁻¹. After 200 cycles, only 456 mA h g⁻¹ is retained for the PAA‐S/KB/TiO₂ cell compared to 535 mA h g⁻¹ for the baseline PAA‐S/KB cell. These values correspond to capacity retentions with TiO₂ (75.2%) and without TiO₂ (68.1%). Furthermore, the average CE over 200 cycles is comparable between the PAA‐S/KB/TiO₂ (99.1%) and PAA‐S/KB (99.2%) cells (Figure S3b). Thus, these results suggest that catalyst‐binder interactions may hinder the functionality of TiO₂. Interestingly, this phenomenon causes greater standard deviation error in the cycling performance, as PAA binder leads to greater heterogeneity with more aggregated domains of TiO₂ (Figure S4). It is also noted that the cycling performance is less hindered when PAA is used with bulk TiO₂ rather than nano TiO₂ (Figure S5), suggesting that PAA may modify the TiO₂ surface chemistry. Bulk TiO₂ particles are larger, presenting a lower surface area that can interact with PAA (Figure S6).

The charge‐discharge voltage profiles of the plot provide more insight into the performance of TiO₂ (Figures 1c,d and S7). As shown in Figure 1c, the cell polarization at 50% state of charge (ΔE_{50% SOC}) is comparable for the LiPAA‐S/KB/TiO₂ (307 mV) and baseline LiPAA‐S/KB (298 mV) cells during the early cycles (cycle 10). Despite being an oxide compound, TiO₂ provides sufficient catalytic activity to achieve comparable kinetics, while acting as an effective LiPS adsorption agent. In contrast, the ΔE_{50% SOC} at the 10^th cycle increases to 416 mV for PAA‐S/KB/TiO₂ compared to 335 mV for the baseline PAA‐S/KB cell (Figure 1d). Thus, rather than reducing overpotential, the addition of TiO₂ causes greater overpotentials when paired with PAA. With continued cycling, the trend holds: the ΔE_{50% SOC} for the LiPAA‐S/KB/TiO₂ cell becomes smaller than that for LiPAA‐S/KB, whereas the ΔE_{50% SOC} for the PAA‐S/KB/TiO₂ cell remains larger than that for LiPAA‐S/KB. This reveals how binder‐catalyst interactions may affect cell polarization. The polarization during the first C/2 cycle is also noted (Figure S7a,b), as it limits the attainable sulfur utilization. It is observed that with further early cycling, the cell polarization decreases compared to the first C/2 cycle as sulfur is more optimally redeposited (Figure S7c and d). However, the sulfur utilization still does not increase significantly, as the initial deposition may have already limited the quantity of reversible sulfur. With long‐term cycling, the polarization increases again due to long‐term degradation effects (Figure S7e and f).

The rate performance of the various cells further showcases the impact of binder‐catalyst interactions. Overall, LiPAA‐S/KB has improved rate capability compared to PAA‐S/KB (Figure 1e,f), which may be due to the higher ionic conductivity of LiPAA (Figure S8). With catalyst addition, LiPAA‐S/KB/TiO₂ and LiPAA‐S/KB cells achieve similar capacities even up to a 1C rate. Interestingly, LiPAA‐S/KB/TiO₂ produces lower capacities beyond a 1C rate, as the insufficient conductivity of TiO₂ may limit the performance (Figure 1e). In comparison, the capacities of PAA‐S/KB/TiO₂ cells are reduced at all rates compared to PAA‐S/KB baseline cells (Figures 1f, S9, and S10). These results demonstrate that unfavorable binder‐catalyst couplings, such as the pairing of TiO₂‐PAA, increase cell polarization and impede rate performance.

To investigate the underlying factors causing the deviations in cycling performance, electrochemical characterization techniques were employed. Cyclic voltammetry (CV) was conducted on the Li‐S cells with the two binder‐catalyst pairings. While the cathodic peak at ∼ 2.35 V is similar for all cells, deviations are observed in the cathodic peak at ∼ 2.0 V (Figure 2a,b). A Tafel analysis confirms that the cathodic peak at ∼ 2.35 V is negligibly affected by the addition of TiO₂ in both the LiPAA and PAA systems (Figure 2c). As shown in Figure 2a, the peak currents for Li‐PAA‐S/KB and LiPAA‐S/KB/TiO₂ are, respectively, 659 and 657 µA g_s ⁻¹. Since TiO₂ is a semiconducting oxide, it provides sufficient catalytic activity while enhancing cycle life due to its LiPS adsorption abilities. However, when employing a PAA binder, the cathodic peak at ∼ 2 V decreases to 238 µA g_s ⁻¹ with TiO₂ addition, compared to 719 µA g_s ⁻¹ for the PAA‐S/KB baseline (Figure 2b). Further, the Tafel slope at ∼ 2.1 V increases in magnitude to −150 mV dec⁻¹ for PAA‐S/KB/TiO₂ compared to −47 mV dec⁻¹ for PAA‐S/KB, signifying more sluggish kinetics with TiO₂ (Figure 2d). Although TiO₂ achieves comparable Li₂S conversion kinetics when paired with a LiPAA binder, its functionality is hindered when paired with a PAA binder. Thus, the binder may interact with catalytic sites, limiting Li₂S nucleation/growth.

FIGURE 2.

Kinetic evaluation of TiO₂ when paired with LiPAA and PAA. CV plots of TiO₂‐incorporated cells with (a) LiPAA and (b) PAA binders. Tafel slope analysis of the cathodic peaks at (c) ∼ 2.35 V and (d) ∼ 2.10 V in previous CV curves. Chronoamperometry curves for TiO₂ electrodes discharged at 2.05 V when employing (e) LiPAA and (f) PAA binders. (g) SEM images of deposited Li₂S from chronoamperometry experiments.

Chronoamperometry experiments were conducted to elucidate how a TiO₂‐PAA may hinder the nucleation and growth of Li₂S. Cathodes composed of only TiO₂, Super C65, and binder (PAA or LiPAA) were prepared, and cells were assembled with a Li₂S₈ in TEGDME catholyte. After a galvanostatic discharge to 2.06 V, the voltage was stepped to 2.05 V and held. The current response was recorded, providing insight into the nature of Li₂S nucleation/growth. The time to peak (t _p) is reduced from 4.77 x 10³ s for the LiPAA/C electrode to 3.84 x 10³ s for the LiPAA/TiO₂ electrode, revealing more nucleation sites in the LiPAA/TiO₂ electrode (Figure 2e). Further, the integrated capacity, which reflects the amount of Li₂S grown, increases from 217 to 233 mA h g_s ⁻¹ when employing a LiPAA/TiO₂ electrode. This reflects the ability of TiO₂ to adsorb LiPS and facilitate its conversion to Li₂S when paired with a LiPAA binder in a sufficiently conductive electrode. However, when pairing TiO₂ with PAA, the nucleation and growth of Li₂S is hindered instead. The t _p is increased from 6.29 x 10³ s for a PAA/C electrode to 8.60 x 10³ s for a PAA/TiO₂ electrode (Figure 2f). The integrated capacity is further reduced from 240 to only 164 mA h g_s ⁻¹ when employing a PAA/TiO₂ electrode. The pairing of TiO₂ with PAA may reduce the availability of catalytic sites, which inhibits the nucleation and growth of Li₂S. The greater integrated capacity of PAA/C compared to LiPAA/C can be attributed to the enhanced adsorption ability of PAA compared to LiPAA (Figure S11).

The morphology of Li₂S deposited onto these catalyst‐casted electrodes provides more insight into the catalyst‐binder phenomena. After deposition through chronoamperometry, the electrodes were examined by scanning electron microscopy (SEM) (Figures 2 g and S12). The LiPAA/C electrode facilitates the deposition of nano‐sized Li₂S, but produces some non‐uniformities. Growth onto a LiPAA/TiO₂ electrode leads to more dense and uniform deposition, showcasing the ability of TiO₂ to provide catalytic sites to enhance Li₂S growth. In a PAA/C electrode, the Li₂S deposition is moderately uniform, whereas the PAA/TiO₂ electrode produces significant heterogeneities in Li₂S growth. Rather than a planar deposition observed in the other samples, distinct micron‐sized Li₂S particles appear on the PAA/TiO₂ electrode. The reduced availability of nucleation sites inhibits the deposition of Li₂S onto the framework, forcing the growth of Li₂S crystals instead. Overall, antagonistic catalyst‐binder interactions are shown to hinder the nucleation and growth of Li₂S.

Operando electrochemical impedance spectroscopy (EIS) was employed to unveil how these antagonistic catalyst‐binder interactions manifest in the cell. The measured impedance spectra were converted into distributions of relaxation times (DRT) for analysis. Different time constants (τ) signify different resistances in the cell. Charge‐transfer resistances manifest between 10⁻⁴ and 10⁰ s, whereas diffusion resistances for Li₂S_x and Li⁺ arise at, respectively, 5 × 10⁰ and 10¹ s [25, 26].  With a LiPAA binder, the S/KB and S/KB/TiO₂ cells display similar interfacial, charge transfer, and diffusion resistances throughout the first discharge and charge of the cell (Figure S13); though TiO₂ is an oxide, it negligibly impacts cell resistances due to its semiconducting properties. Employing a PAA binder yields different results. During the discharge and charge of a PAA‐S/KB cell, there are minor charge‐transfer resistances in the cell (Figure 3a). Barriers to Li⁺ diffusion pose as the largest barrier, arising during the Li₂S growth phase (second voltage plateau at ∼ 2.1 V). However, in a PAA‐S/KB/TiO₂ cell, charge‐transfer and diffusion resistances significantly increase during the second voltage plateau (Figure 3b). The increase in charge‐transfer resistance suggests that fewer active sites are accessible, hindering the conversion of soluble LiPS species. Notably, the diffusion resistance of Li₂S_x also grows, as the surface LiPS concentrations become more depleted at the remaining active sites, creating larger concentration gradients. It becomes evident that the PAA‐TiO₂ pairing produces antagonistic interactions.

FIGURE 3.

Evolution of resistances during charge/discharge of a cell. DRT analysis of PAA‐based cathodes for the first cycle of (a) baseline S/KB and (b) S/KB/TiO₂ cells.

To gain insight into how PAA may inhibit catalytic activity, surface chemistry between the catalyst and binder was characterized. The catalyst and binder werere mixed at a 4:6 mass ratio. Although researchers typically assume negligible interactions between the two, it is shown that this cannot be broadly assumed. Fourier transform infrared spectroscopy (FTIR) results reveal how a binder structure changes when coupled with TiO₂ (Figure 4a). A PAA binder is composed of distinct vibrational modes due to various functional groups, including O─H stretching (∼2,500 – 3,300 cm⁻¹), C═O stretching (∼ 1,710 cm⁻¹), CH₂ deformation (∼ 1,450 cm⁻¹), C‐O stretching (∼ 1,160 cm⁻¹), and O─H wagging (∼ 920 cm⁻¹) [27, 28]. When PAA is combined with TiO₂, the O─H_stretch and O─H_wag peak intensity decreases, indicating weakening of this covalent bond. This coincides with the formation of a new peak at ∼1,550 cm⁻¹ (COO⁻ _asym), signifying a partial delocalization of electrons and the formation of resonance [29]. Thus, the carboxylic H atom may interact with TiO₂ and weaken its binding to the COO⁻ functional group in PAA.

FIGURE 4.

Characterization of TiO₂‐binder interactions. (a) FTIR data of TiO₂‐PAA and TiO₂‐LiPAA blends and (b) XRD patterns for the (110) and (211) peaks of TiO₂ in the mixtures. XPS (c) O 1s and (d) Ti 2p spectra for TiO₂‐binder mixtures. (e) DFT simulations of binder interactions with TiO₂.

In a LiPAA binder, the FTIR spectrum is dominated by peaks at ∼1,550 cm⁻¹ (COO⁻ _asym) and ∼1,410 (COO⁻ _sym), unveiling that lithiation of the binder produces a resonance‐stabilized COO⁻ group. Minor signals of C═O and C─O bonding indicate the presence of residual, unlithiated PAA. The addition of TiO₂ to LiPAA does not change the vibrational modes of LiPAA, indicating negligible changes to the binder structure.

X‐ray diffraction (XRD) was conducted on the TiO₂‐binder mixtures (9:1 mass ratio) to unveil how catalyst‐binder interactions can alter the catalyst structure. The XRD patterns of rutile TiO₂ are dominated by the (110) and (211) peaks at, respectively, ∼ 27.5° and ∼ 54.4° (Figures 4b and S14). Combining TiO₂ and LiPAA causes minor reductions in peak intensity without a peak shift, demonstrating negligible alteration of the TiO₂ crystal structure. In contrast, the TiO₂‐PAA mixture produces additional shoulder peaks at ∼ 27.1° and ∼ 54.1°. Though the bulk of TiO₂ is unaltered, the surface lattice is distorted by PAA interactions. H interactions with TiO₂ have been shown to increase the lattice spacing of the crystal structure, as H bonds with the O atom of TiO₂ [30].

The chemical coordination in the TiO₂‐binder mixtures was further probed with x‐ray photoelectron spectroscopy (XPS). The chemical environment of PAA is altered when mixed with TiO₂. Pure PAA has distinct C─O─H (O 1s ≈ 533.4 eV) and C═O (O 1s ≈ 532.0 eV) environments in the O 1s spectrum (Figure S15). These peaks are substantially reduced in the TiO₂‐PAA mixture. The O─H bond in PAA weakens enough that the electron‐delocalized COO⁻ peak arises (O 1s ≈ 531.3 eV) (Figure 4c). Although the O binding environments for pure PAA and pure LiPAA are typically different, interactions with a catalyst produce a similar electron‐delocalized COO⁻ functional group. This also suggests significant positive charge transfer from the carboxylic H atom in PAA to the O anion in TiO₂.

Meanwhile, the O‐Ti environment (O 1s ≈ 529.6 eV) is preserved in the catalyst structure with binder addition, but there is a shift to more positive binding energies (0.09 and 0.07 eV when incorporated with PAA and LiPAA, respectively) (Figure 4d). The H atom or Li ion from the binders interacts with the O anion in the TiO₂ structure, leading to positive charge transfer from the binder to the catalyst. Such O─H linkage between TiO₂ and PAA modifies the surface chemistry of TiO₂. Although the Li⁺ in LiPAA may also interact with TiO₂, these exist as Coulombic forces and cause no structural changes to the TiO₂ lattice. Owing to positive charge transfer from the binder to TiO₂ (in the form of H⁺ or Li⁺ interactions), the peaks in the Ti 2p spectra (Ti 2p_3/2 ≈ 458.4 eV) similarly shift to a positive binding energy; the electron density of the O anion in TiO₂ is reallocated to stabilize Li or H, rather than donated to the Ti cation. Interestingly, despite a higher charge density of H⁺ compared to Li⁺, the shifts in binding energy for the Ti─O environment are comparable for PAA and LiPAA. This suggests that COO⁻ in PAA may also minorly coordinate with TiO₂, negating some positive charge transfer.

Computational density functional theory (DFT) results further corroborate these binder‐catalyst interactions. Small chains of LiPAA or PAA are positioned near the TiO₂ lattice, and the system was allowed to reach its lowest‐energy configuration. In the relaxed state, there are observed interactions between the O atoms of TiO₂ with the Li⁺ ions of the LiPAA or the carboxylic H atoms of PAA (Figure 4e). While the COO⁻ group in LiPAA does not coordinate to Ti, the H─TiO₂ linkage in TiO₂‐PAA facilitates COO⁻ binding to the Ti cation. Thus, the COO⁻ group occupies adsorption sites in TiO₂, consistent with the XPS results discussed above; such interactions may significantly alter how sulfur and LiPS adsorb onto TiO₂.

The sulfur adsorption factor (f _sulfur) provides insight into sulfur adsorption onto the TiO₂ active surface. The f _sulfur was measured to be 0.124 and 0.013 for, respectively, the LiPAA/TiO₂ and PAA/TiO₂ cells (Figures S16 and S17). This confirms that sulfur adsorbs onto the TiO₂ active surface and reduces the electric double‐layer capacitance (C_dl) in a LiPAA/TiO₂ cell, consistent with previous studies [31, 32].  In contrast, the negligible f _sulfur for the PAA/TiO₂ cells reveals that sulfur adsorption is disrupted by antagonistic catalyst‐binder interactions where PAA competes with sulfur for adsorption onto the TiO₂ site.

The impact of these binder‐catalyst interactions on LiPS adsorption is subsequently investigated. TiO₂ was coated with a thin layer of binder (100:1 mass ratio) and immersed in an 8 mM Li₂S₆ solution. By itself, this quantity of binder has negligible adsorption abilities (Figure S18). However, even this small quantity of binder can modify the surface chemistry of TiO₂. After immersion, all the LiPS solutions change colors rapidly (Figure S19). After 48 h, the solution with TiO₂‐PAA is nearly transparent compared to the yellow solutions with TiO₂‐LiPAA or TiO₂ (Figure 5a). The PAA‐TiO₂ mixture adsorbs Li₂S₆ more quickly, revealing that the formation of LiPS may further stabilize the H─TiO₂ interactions and promote the protonation of TiO₂. The visual color changes in the solutions were confirmed with UV–visible spectrometry, showing reduced LiPS signals for the TiO₂‐PAA mixture (Figure 5b).

FIGURE 5.

Characterization of Li₂S₆ interactions in TiO₂‐binder mixtures. (a) Optical photo of 8 mM Li₂S₆ solution after immersing 50 mg TiO₂‐binder mixtures for 48 h and (b) corresponding UV‐Visible spectra of the solutions. XPS (c) Ti 2p, (d) O 1s, and (e) S 2p spectra for Li₂S₆‐TiO₂‐binder mixtures. (f) DFT simulations of Li₂S₆‐TiO₂‐binder mixtures. (g) Summary of antagonistic PAA interactions with TiO₂.

The LiPS interactions with the binder‐catalyst blends were further analyzed with XPS. Mixtures composed of Li₂S₆, TiO₂, and binder were prepared for XPS experiments. As shown in the Ti 2p spectra (Figure 5c), the Ti─O peaks shift to lower binding energies with the addition of Li₂S₆ to the catalyst‐binder mixtures (0.16 eV for Li₂S₆‐TiO₂‐LiPAA and 0.18 eV for Li₂S₆‐TiO₂‐PAA). These shifts unveil that electron density from negatively charged groups is donated to the Ti cation in TiO₂. This charge transfer may be due to interactions with the terminal S⁻ atom in Li₂S₆ or from the COO⁻ group in PAA/LiPAA. The ensuing XPS O 1s and S 2p spectra guide the deconvolution of these interactions.

In the O 1s spectra in Figure 5d, the peak for O‐Ti bonding shifts negatively by 0.10 eV when Li₂S₆ is added to the TiO₂‐LiPAA mixture. This reveals that the interactions between the TiO₂ and LiPAA weaken, as LiPAA may also interact with Li₂S₆. Meanwhile, the addition of Li₂S₆ to a TiO₂‐PAA mixture may further intensify binder‐catalyst interactions. Some O anions in TiO₂ become protonated, leading to a new Ti‐O‐H environment (O 1s ≈ 530.9 eV) [33, 34]. In the remaining O anions of TiO₂, there is a negligible shift in binding energy, suggesting that PAA interactions remain. Thus, the addition of Li₂S₆ appears to stabilize the deprotonation of PAA and protonation of TiO₂.

Meanwhile, the O 1s peaks corresponding to the LiPAA and PAA binder (COO⁻, C─O─H, C═O) shift positively, which may be attributed to interactions with the Li⁺ ion in Li₂S₆ or the Ti cation in TiO₂. Notably, The COO⁻ peak shifts by 0.35 eV for the TiO₂‐PAA mixture when Li₂S₆ is added, compared to 0.18 eV for TiO₂‐LiPAA. This reveals that there are strong interactions between the COO⁻ group and other positively charged components in the mixture (Ti sites in TiO₂ or Li cations in LiPS). Additionally, these interactions are more prominent for the COO⁻ group in PAA than in LiPAA.

The S 2p spectra guide the deconvolution of whether the COO⁻ group of the binder interacts with TiO₂ or Li₂S₆. The S atoms in Li₂S₆ exist in two distinct states: bridging S⁰ (S 2p_3/2 ≈ 163.2 eV) or terminal S¹⁻ (S 2p_3/2 ≈ 161.6 eV) (Figure 5e). Binders with O functional groups are known to bind to Li⁺ in Li₂S₆, rendering the S atoms more negatively charged [35, 36, 37].  Meanwhile, LiPS bonds with the catalysts are known to render the S atoms more positively charged, as electron density is donated to the transition‐metal cation [15, 38, 39].  The large negative shift of the S 2p peaks (∼ 0.19 eV) in the TiO₂‐LiPAA system indicates that the COO⁻ binds strongly with the Li⁺ of Li₂S₆, and minimally with the catalyst. In contrast, the smaller negative shift (∼ 0.09 eV) in the TiO₂‐PAA system reveals that while the COO⁻ binds to Li⁺ in Li₂S₆, it may also strongly bind to the Ti active site of TiO₂, anchored by the protonated site. Thus, this shows competing interactions between the binder and LiPS for accessing the active site of TiO₂.

Computational DFT results further elucidate how the molecules coordinate with each other. A simulation was initialized with Li₂S₆, TiO₂, and a small binder chain. Then, the system was allowed to reach its lowest energy configuration. Notably, in the Li₂S₆‐TiO₂‐PAA system, the Li⁺ in Li₂S₆ is shown to stabilize the deprotonation of PAA and protonation of TiO₂ (Figure 5f). Thus, this may explain the enhanced adsorption ability of a PAA‐TiO₂ mixture. Further, there is notable bonding between the COO⁻ group of PAA and the Ti sites in TiO₂. In contrast, in the Li₂S₆‐TiO₂‐LiPAA system, there is no binding observed between the COO⁻ group of LiPAA and the Ti sites.

The schematic in Figure 5g summarizes the antagonistic surface chemistry between PAA and TiO₂. Although sulfur adsorption onto TiO₂ is already impeded by such interactions, LiPS formation further intensifies the antagonistic behavior. The Li⁺ in Li₂S₆ stabilizes the formation of the COO⁻ group and facilitates the protonation of TiO₂ by the carboxylic H atom in PAA. This results in the anchoring of the COO⁻ group near the Ti site. At the position, the COO⁻ group competes with the S¹⁻ atom in LiPS for adsorption onto the Ti site, further deactivating the catalytic site. This causes growth in charge‐transfer resistance and inhibits the conversion to Li₂S.

In all, a mechanistic understanding of catalyst‐binder interactions is proposed in Figure 6. A binder can have benign linkages with a catalyst that does not hinder its catalytic activity. Although Li⁺ ions in LiPAA may still bind the anion of a catalyst, these interactions do not impede the catalyst's ability to adsorb LiPS. Alternatively, catalyst‐binder interactions can be antagonistic, hindering catalytic activity. H‐bonding in PAA facilitates the binding of the COO⁻ functional group to the catalyst's cation, occupying active sites that facilitate LiPS adsorption. As a result, catalytic activity is diminished, leading to significant charge‐transfer and diffusion resistances in the cell.

FIGURE 6.

Summary schematic depicting (left) benign and (right) antagonistic catalyst‐binder interactions.

Consistent with this mechanistic insight, antagonistic catalyst‐binder interactions were also found to hinder cycling performance when employing WS₂ and Cr₃C₂ catalysts, including at higher sulfur loadings (Figures S20–S24); this phenomenon may be broadly applicable, regardless of the specific cation or anion identity in the catalyst. Overall, this builds upon recent insights that solvent molecules can also compete for adsorption onto catalytic sites [26].

These findings highlight that catalyst–binder compatibility is a critical, yet often overlooked, factor in Li–S cathode design. Although COOH‐functionalized binders are widely used for their effective LiPS adsorption, such functional groups may be incompatible with catalyst integration. Further investigations are necessary to determine which functional groups in binders may be detrimental to catalytic activity. In all, these results underscore the importance of holistic electrode design for advancing the development of viable Li─S batteries.

3. Conclusion

For the first time, catalyst‐binder interactions were investigated and demonstrated to significantly impact catalytic performance. When pairing TiO₂ with a LiPAA binder, there are benign catalyst‐binder interactions. TiO₂ provides sufficient catalytic activity and good LiPS adsorption abilities, improving the cycle life of a Li─S cell. However, when pairing TiO₂ with a PAA binder, antagonistic catalyst‐binder interactions are revealed; the addition of TiO₂ leads to significant cell polarization and reduced capacities compared to the baseline cell without TiO₂. Further, Li₂S growth is hindered by significant charge‐transfer and diffusion resistances. The mechanism behind this was elucidated. The carboxylic H atom in PAA facilitates COO⁻ binding to the Ti cation in TiO₂. Thus, PAA competes for adsorption sites in TiO₂, diminishing LiPS access to the sites. Antagonistic PAA‐catalyst interactions were found to be broadly applicable to various catalysts with different cation/anion compositions. Therefore, binders with a COOH functional group may be detrimental to catalytic activity. In all, this study showcases the need to consider holistic electrode designs to enable viable Li─S batteries.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Supporting File 1: anie72160‐sup‐0001‐SuppMat.docx.

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.