Fluorinated porous frameworks enable robust anode-less sodium metal batteries

Abstract

Sodium metal batteries hold great promise for energy-dense and low-cost energy storage technology but are severely impeded by catastrophic dendrite issue. State-of-the-art strategies including sodiophilic seeding/hosting interphase design manifest great success on dendrite suppression, while neglecting unavoidable interphase-depleted Na⁺ before plating, which poses excessive Na use, sacrificed output voltage and ultimately reduced energy density. We here demonstrate that elaborate-designed fluorinated porous framework could simultaneously realize superior sodiophilicity yet negligible interphase-consumed Na⁺ for dendrite-free and durable Na batteries. As elucidated by physicochemical and theoretical characterizations, well-defined fluorinated edges on porous channels are responsible for both high affinities ensuring uniform deposition and low reactivity rendering superior Na⁺ utilization for plating. Accordingly, synergistic performance enhancement is achieved with stable 400 cycles and superior plateau to sloping capacity ratio in anode-free batteries. Proof-of-concept pouch cells deliver an energy density of 325 Watt-hours per kilogram and robust 300 cycles under anode-less condition, opening an avenue with great extendibility for the practical deployment of metal batteries.

Anode-less Na metal battery with high energy density and long life is simultaneously achieved by fluorinated porous frameworks.

INTRODUCTION

Sodium-based batteries have been regarded as promising candidates for “beyond lithium-ion” technologies by virtue of similar properties to Li but more natural abundance and low cost (1, 2). In this regard, Na metal is undoubtedly the ultimate anode material choice due to its low redox potential (−2.714 V versus standard hydrogen electrode) and high specific capacity (1165 mAh g⁻¹) (3, 4), far exceeding the current close-commercial hard carbon anodes (e.g., 300 mAh g⁻¹). Unfortunately, the practical implementation is critically hindered by the notorious uncontrolled Na dendrite growth and infinite dimensional changes, thus leading to low active Na⁺ utilization, inferior reversibility and cyclic stability, and even serious safety hazards. Several effective solutions have been devoted to stabilizing alkaline metal anode, such as constructing nucleation layers (5), building artificial solid-electrolyte interphase (SEI) (6–9), optimizing electrolytes (10–15), exploring solid-state/gel polymer electrolytes (16, 17), and designing porous hosts (18–21). Among these, designing sodiophilic seeding/hosting interphases is of pivotal significance in sequestering dendrite-induced catastrophic issues because conductive Na could deposit randomly with unpredictable shape and is “hostless” during plating and stripping (Fig. 1A) (22, 23).

Fig. 1. Na plating behavior under limited Na⁺ source conditions.

Schematic diagrams for Na plating on (A) bare Cu, (B) conventional sodiophilic seeding/hosting interphases (here shows porous triazine framework as an example), and (C) sodiophilic yet low Na⁺ depletion enabled by fluorinated porous triazine frameworks. (D) Corresponding discharge profiles in anode-free full cells with covalent triazine framework (CTF) and fluorinated CTF (FCTF) seeding/hosting interphases, and the inset showing the scheme of Na⁺ from cathode materials distribution by interphase and plating. The Na⁺ utilization for plating is 75% for FCTF-based full cell and 0% for CTF based in anode-free configurations.

In this context, plating/stripping reversibility and durability have been improved with suppressed dendrite formation, which, however, mainly relies on using a large excess of Na sources, as exemplified by low depth of discharge (DOD; e.g., <10%) (24, 25) in symmetrical cells and high negative to positive capacity ratio (N/P; e.g., >50) (16, 26) in full cells. This is because excess Na inventory can continuously compensate for the irreversible Na⁺ depletion, mainly from unceasing SEI formation/cracking and inactive or “dead Na”, posed by ineluctable Na dendrite, resulting in artificially improved performances but inapplicability in realistic high-energy batteries. A practical alkaline metal battery needs a DOD over 90% in symmetrical cells and N/P ratio less than 2 in full cells (19). Under these limited Na conditions, dendrite growth with capacity and stability decay is aggravated due to the incapability for replenishing Na⁺ consumption. To date, Na batteries with high-energy density and long-term operating (e.g., >200 cycles) under low N/P ratios (e.g., <2) have scarcely been explored, especially in pouch cells. Consequently, boosting the plating/stripping stability with mitigated dendrite under high DOD and low N/P ratio is a key fundamental advance toward practical deployment.

Now, tremendous effort has been devoted to constructing state-of-the-art seeding/hosting interphases, which mainly include (i) introducing intrinsic defects or heteroatom doping in carbons (27–29), (ii) incorporating the sodiophilic alloying metals (5, 30, 31), and (iii) constructing three-dimensional alloying metal-based porous skeletons (24, 32). These interphases can alleviate the dendrite problems but nevertheless deplete a large amount of Na⁺ before plating via reactions such as adsorption, defect binding, intercalation, alloying, or conversion reactions (33–38), posing several intractable issues especially under limited Na sources for realistic applications. First, substantial Na⁺ consumption in current sodiophilic interphases gives rise to the low ratio of Na participated for plating/stripping, namely, high N/P ratio. A survey of references shows that the average interphase-depleted Na⁺ is around 0.6 mAh cm⁻² (fig. S1A), while the cathode areal capacity is less than 1 mAh cm⁻² (fig. S1, B and C). That is to say, more than half of Na⁺ is consumed by the interphase rather than participating in plating/stripping if in an anode-free cell (N/P = 0), in which the cathode is the sole Na source (10, 32, 39). Second, interphase-depleted Na⁺ generally occurs at higher potential than that of Na⁺ plating in anode, which would be detrimental to the output voltage of full cells especially at low N/P ratio even down to 0 (Fig. 1, B and D). The high N/P ratio and decreased output voltage together contribute to the less-than-perfect energy density of Na batteries. Last but not least, the reaction-induced sodiophilicity of current seeding/hosting interphases also leads to poor stability due to volume variations such as alloying reactions [e.g., 420% of Na-Sn alloying (40)]. These escalate the challenges to construct comprehensive interphases under low N/P ratio or even anode-free prototype to achieve durable and highly efficient Na plating/stripping, which essentially demands revolutionary interphases design aimed at transcending the limitations associated with intractable paradox properties, i.e., both ultralow Na⁺ storage capacity while maintaining superior sodiophilicity.

Here, we propose a precisely molecular-designed porous triazine framework with integrated F functionality (fluorinated covalent triazine framework, FCTF), capable of delivering excellent sodiophilicity yet low Na⁺ consumption. The well-defined and adjustable F substituent group on highly ordered layered channels and periodic networks serves as uniform sodiophilic seeding sites for decreasing nucleation overpotential and achieving crack-free Na deposition and meanwhile reduces the reactivity of Na⁺ for inevitable Na⁺ depletion (Fig. 1C), as synergistically corroborated by the physicochemical measurements and density functional theory (DFT) computations. Consequently, dendrite-free deposition is achieved using FCTF interphase, exhibiting low nucleation overpotential of 9 mV, high average Coulombic efficiency (CE) of 99.6% during 400 cycles in half cells, and remarkable interfacial stability even under 95% DOD in symmetric cells. Together with the low Na⁺ depletion, FCTF enables the anode-free full cells (without any Na metal predeposition) with capacity decay of 0.11% per cycle (twice less than that of CTF) over 400 cycles and exceptional high-plateau to low-sloping capacity contribution especially with increasing cathode areal capacity, thus leading to enhanced average cell output voltage (Fig. 1D). Proof-of-concept pouch cells with a low N/P ratio of 1.5 could realize an energy density up to 325 Wh kg⁻¹ based on active materials and interphases, and robust cycling with 94.7% retention was attained after 300 cycles, hitting one of record values in pouch cells. While the application of porous crystalline framework in Na anodes is still in its infancy, this work also validates its feasibility and great extendibility to Li anodes, propelling the practical implementation of energy-dense alkaline metal batteries.

RESULTS

Preparation and characterization of the FCTF

Unlike the harsh conditions using ZnCl₂ and high-temperature ionothermal approach, FCTFs and CTF were mildly synthesized using a superacid-catalyzed strategy to well retain the structural integration (fig. S2) (41). The F functionality decorated in triazine frameworks was realized by trimerization of 1,4-dicyanotetrafluorobenzene (DCFB) and 1,4-dicyanobenzene (DCB), and the resulting samples were denoted as FCTF-x/y, in which x and y represent the molar ratio of DCFB and DCB, respectively. FCTF refers to FCTF-1/1 unless otherwise mentioned. The skeletons are in the form of hexagons in each layer (Fig. 2A), with eclipsed stacking between layers (41, 42), which can be revealed by strong diffraction peaks in x-ray diffraction (XRD) patterns (Fig. 2B). The lamellar structure can be observed in transmission electron microscopy (TEM) and high-angle annular dark field scanning TEM (Fig. 2, C and E), and they are further stacked into large flake morphologies (fig. S3). High-resolution TEM (HRTEM) images further confirm that the gauged interlayer distance is 0.34 nm by clear lattice fringes labeled to the interlayer reflection (001) (Fig. 2D and fig. S4), in agreement with XRD results (Fig. 2B). Then, formed nanochannel array results in nanoporous structure with specific Brunauer-Emmett-Teller (BET) surface areas of 401 and 528 m² g⁻¹ for FCTF and CTF, respectively (fig. S5). A pore size of around 1.3 nm observed for these materials is in agreement with the optimized model (fig. S5C). With increasing DCFB dosage, the surface area and crystallinity decrease monotonously due to the incorporation of relatively large-size F (figs. S5 and S6, and table S1). The presence of F was further revealed by the elemental mapping, showing the uniform distribution of F together with C and N across the entire selected area (Fig. 2E). The chemical bonding configuration of F was further corroborated by the Fourier-transform infrared (FTIR) and x-ray photoelectron spectroscopy (XPS) spectra. Apart from the presence of three peaks at around 803, 1349, and 1510 cm⁻¹ attributed to the triazine ring in both CTF and FCTF, an additional peak at 988 cm⁻¹ appears exclusively in FCTF, corresponding to the C-F stretching (Fig. 2, F and G). Likewise, the presence of C-F is aslo verified in the deconvoluted C 1s spectrum at 290.4 eV in FCTF, besides the C═C (284.8 eV) and C═N (287.1 eV) peaks in both samples (Fig. 2H and fig. S7). Meanwhile, the F 1s spectrum shows a strong peak at 687.8 eV (fig. S7C), and the corresponding F content of FCTF is estimated to be 8.17 atomic % from XPS (table S2), slightly lower than the theoretical value (43, 44). Upon increasing the ratio of DCFB, the F peak in XPS and the C-F peak in FTIR are gradually strengthened (figs. S8 and S9), and the F content increases accordingly (table S2). Traces of oxygen observed in XPS spectra and elemental mapping might be related to the edges or defects in the framework (figs. S8 and S10). From these results, it can be confirmed the successful construction of well-defined triazine frameworks with tailored F functionality.

Fig. 2. Physicochemical characterization of FCTF.

(A) Schematic representation of the structure of FCTF (left) and CTF (right). (B) XRD patterns and the calculated pattern (black) of FCTF and CTF. HRTEM images of FCTF with (C) low and (D) high magnifications. (E) High-angle annular dark field scanning TEM image of FCTF and corresponding element mappings of C, F, and N in energy-dispersive x-ray spectroscopy. (F) FTIR spectra of DCFB, FCTF, and CTF. (G) Enlarged curves of FTIR spectra. (H) XPS spectra of FCTF and CTF in C 1s spectra.

Sodiophilicity with low-barrier nucleation

The incorporation of F in the polymeric frameworks could result in strong electronegativity (43, 45), as revealed by the high electron density distribution at F sites (fig. S11), and thus it is anticipated to exhibit high affinity and act as preferential nucleation sites for homogeneous deposition (46, 47). To demonstrate the effectiveness of F, the adsorption energy (E_ad) of fragments, namely, benzene and tetrafluorobenzene as model compounds, was firstly calculated. Obviously, tetrafluorobenzene exhibits much higher affinity with E_ad of −5.57 eV, far exceeding benzene rings with E_ad of −0.72 eV (Fig. 3A). Furthermore, the E_ad of Na atom on the monolayer periodic structure was calculated with different fluorinated edges. CTF shows the E_ad value of −2.50 eV, while the value decreases to −3.30 eV with two para-fluorinated edges (Fig. 3, B and C). With three fluorinated edges, the E_ad further reduces to less than −3.80 eV regardless of its relatively spatial distribution (Fig. 3C and fig. S12). These results theoretically confirm the high sodiophilicity navigated by the integration of F in porous frameworks. Further experimental corroboration was carried out by the Na nucleation overpotential, which is defined by the gap between the shrill tip potential and the subsequent steady overpotential. The lower the value, the easier for homogeneous nucleation with less barrier. As expected, the nucleation overpotential of FCTF (9 mV) is lower than those of CTF (17 mV) and bare Cu foil (36 mV) at 1 mA cm⁻² (Fig. 3D), demonstrating advantageous Na growth on the F-containing periodic framework. The overpotential decreases by elevating the F content from CTF, FCTF-1/3, to FCTF (fig. S13), in line with the above affinity from DFT calculation (Fig. 3C). As the current density increases, FCTF always retains constantly lower overpotentials (e.g., 10 mV at 4 mA cm⁻²) compared with CTF (e.g., 30 mV at 4 mA cm⁻²), indicative of fast kinetics and high-rate properties (Fig. 3E and fig. S14). These results are also lower than the previously reported interphases at comparable current densities (e.g., 11 to 40 mV at 4 mA cm⁻²; table S3). Generally, the overpotential tends to increase sharply as current density increases due to restricted Na⁺ responses under high-rate operation. However, note that the overpotential of FCTF is almost independent of the current density (e.g., 10 mV from 2 to 4 mA cm⁻²), which is rare in previous reported literatures (table S3). It is anticipated to associate with the unimpeded charge transfer kinetics including the electrons and ions. Considering the same ratio of conductive additives in these interphases, the electron conduction can be excluded. Thus, the ionic behavior is responsible for the fast kinetics. To this end, FCTF and CTF were coated on the commercial separator to measure the ionic conductivity (fig. S15). FCTF-coated separator demonstrates the ionic conductivity of 1.28 × 10⁻³ S cm⁻¹, 4.6 times higher than that of the CTF-coated separator, which could be responsible for the high-rate property and help for homogenizing ion flux (46).

Fig. 3. Theoretical investigation of Na nucleation and electrochemical performance of Na anodes.

Adsorption energy of Na atom with (A) the fragment of structure and (B) the monolayer periodic structure. (C) Comparison of E_ad of Na with CTF and FCTFs. (D) Voltage profiles of galvanostatic Na deposition on different substrates at 1 mA cm⁻². (E) Comparison of nucleation overpotential on FCTF and CTF under various current densities. (F) CEs of FCTF, CTF, and Cu foil at 2 mA cm⁻² and 1 mAh cm⁻². (G) Rate performances of FCTF, CTF, and Cu from 1 to 4 mA cm⁻² and return to 2 mA cm⁻² at 2 mAh cm⁻². (H) Voltage profile of FCTF at 1 mA cm⁻² and 10 mAh cm⁻².

High plating/stripping reversibility

The low nucleation barrier is beneficial for the subsequent repeated plating/stripping, as evaluated by the CE. The half cells with FCTF-coated Cu deliver stable CEs of 99.8% over 1200 cycles at 1 mA cm⁻², 99.6% for 400 cycles at 2 mA cm⁻², and 99.2% for 200 cycles at 3 mA cm⁻² (Fig. 3F and fig. S16). The corresponding potential curves keep stable over 1000 cycles and show low voltage polarization ranging from 7 to 14 mV (fig. S17), suggesting the high stability of F anchoring sites. The half cells with CTF interphase and bare Cu foils exhibit fluctuating CEs with lower average values of 97.9 and 96.9% for 100 cycles at 2 mA cm⁻² (Fig. 3F), respectively, despite the relatively stable cycling at low current density of Cu foil in ether electrolyte (fig. S16A) (12, 13). With steadily enhancing the current density to 4 mA cm⁻², FCTF interphase also presents robust CE up to around 99.5%, whereas serious fluctuation occurred in CTF interphase and bare Cu foil especially at higher current density (Fig. 3G). In addition, the CE remains high and stable by increasing the F content. However, excessive increase of the F in framework reduces CE and increases nucleation overpotential (FCTF-3/1; fig. S18), which is probably due to the low porosity and poor crystallinity, as revealed by the BET and XRD results (figs. S5A and S6). These results manifest the regulation of the Na anode reversibility in a molecular fashion via using the porous framework. Even at higher areal capacity (10 mAh cm⁻²), FCTF interphase still exhibits stable average CE close to 100% over 3700 hours (Fig. 3H and fig. S19), whereas the bare Cu foil delivers a fluctuating voltage profile and deteriorated CE after 200 hours (fig. S20), indicative of dendrite-free deposition with high interfacial stability of FCTF. Ex situ scanning electron microscopy (SEM) test was further implemented to trace the morphology with deposited metallic Na. The process goes through sodiation (a1 and b1 points), nucleation (lowest tip potential), and deposition (a2, a3 and b2, b3 points), as indicated by the discharge curves (Fig. 4A). The morphology and electrode thickness remain similar to that of pristine FCTF interphase upon framework sodiation and partial deposition at a2 point of FCTF (figs. S21 and S22), indicating that the porous structure and interstitial space could accommodate Na as host. With further deposition to 1 mAh cm⁻² (a3 point), a homogeneous deposited Na layer was observed on the FCTF surface (Fig. 4, a3), while some dendrites appear on CTF interphase and Cu foil (Fig. 4, b3, and fig. S23). Furthermore, the increased thickness of FCTF interphase (4 μm) is lower than that of CTF interphase (8 μm) as well as the theoretically calculated value (8.9 μm) (Fig. 4B and figs. S21 and S22), indicative of a tight Na deposition gradually filling the nanospace in FCTF interphase compared with the large quantity of Na dendrite on Cu foil (Fig. 4C). In this context, masses of dendrites and cracks almost completely cover the CTF interphase and Cu foil with plating up to 10 mAh cm⁻², whereas a dense and flat surface is observed in FCTF interphase (Fig. 4, D to F, and fig. S24), highlighting the critical role of F in navigating dendrite-free deposition. After fully stripping away, the morphology of FCTF interphase is consistent with the pristine (fig. S25) and also similar to the repeated 100 cycles (fig. S26). This dendrite-free and reversible plating/stripping at higher areal capacities must form the stable SEI between the electrode and electrolyte. Thus, the structure of SEI was explored by XPS surface analysis (figs. S27 and S28, and table S4) and in-depth measurements (figs. S29 and S30, and table S5). As intended, the SEI of FCTF interphase after different cycles (50, 100, and 200 cycles) and sputtering times (30, 120, and 210 s) is dominated by NaF constituents, responsible for the stable SEI and thus boosting long-term cycling. TEM, HRTEM, and XRD analysis also reveal the homogenous distributed crystalline NaF in the SEI of FCTF (figs. S31 to S34). This dendrite-free deposition, together with stable NaF-dominated SEI, would relieve “dead Na” generation and prevent continuous SEI formation/breaking, therefore greatly boosting the efficient Na⁺ utilization.

Fig. 4. Morphology characterization of Na deposition and electrochemical performance of symmetric cells.

Discharging curves at the first cycle for (A) FCTF and CTF in half cells at 1 mA cm⁻². Corresponding surface morphologies (a3 and b3) after discharging for a given time. (B) Increased thickness value of CTF, FCTF, and theoretical after plating 60 min. (C) Schematic of Na deposition on FCTF, CTF, and Cu electrodes. SEM images of (D) FCTF, (E) CTF, and (F) Cu foil electrode after plating for 10 mAh cm⁻². (G) Voltage profiles of CTF and FCTF in symmetric cells under 2 mA cm⁻² and 1 mAh cm⁻². The inset in the panel is detailed voltage profiles of selected cycles. (H) Electrochemical impedance spectroscopy of cells with FCTF and CTF electrodes at different cycles under 1 mA cm⁻² and 1 mAh cm⁻². (I) Voltage profile of FCTF in symmetric cell with 95% DOD. (J) Comparison of cycling time, areal capacity, and DOD of different Na anodes, the related references are provided in table S6.

The dendrite-free deposition with high interfacial stability and reversibility is further demonstrated in FCTF/Na symmetric cells with a certain amount of predeposited Na. FCTF/Na displays a low overpotential and extraordinary cycling stability of 2600 hours at 2 mA cm⁻² and 1600 hours at 3 mA cm⁻², whereas Cu/Na and CTF/Na shows higher overpotential and fierce fluctuation (Fig. 4G and figs. S35 to S37). Meanwhile, the cycling performance of FCTF/Na also far exceeds that of symmetric cells using Na foils (fig. S38). In addition, FCTF/Na still operates with a good rate performance even at a high current density of 8 mA cm⁻² (fig. S39). Electrochemical impedance spectroscopy (EIS) displays that FCTF/Na had a quite low and stable resistance after 500 cycles compared with CTF/Na, preliminarily proving the importance of F in boosting interfacial stability and NaF-dominated SEI (Fig. 4H and fig. S40). Figure 4I further explores the interfacial stability under deep cycling condition, which indicates intensive or even completely stripped and fully plated electrode, as manifested by DOD. As mentioned in the introduction and summarized in table S6, the majority of previously reported works adopted thick Na deposition but involved only 30% or less of Na participating in plating/stripping at each cycle (i.e., DOD less than 30%). Consequently, long-term interfacial durability under high DOD close to 100% represents the harshest conditions equal to anode-free configurations in full cell, which escalates the challenges of sophisticated seeding interphase design. Impressively, FCTF/Na symmetric battery exhibits a lower and constant voltage hysteresis and cycling life of over 1300 hours under DOD up to 95% at a large capacity of 9.5 mAh cm⁻² compared with the fierce fluctuation of Cu/Na (Fig. 4I and fig. S41), which is ranked among the top documented values to our knowledge (Fig. 4J and table S6) (24, 25, 27, 30, 31, 35, 48–57). Besides, a similar voltage profile and a low overpotential of 7 mV in FCTF symmetric battery are achieved under an unprecedented DOD of 100% (fig. S42), indicative of possibility for developing anode-free cell in which all the active ions are from the cathode. This reversible Na plating/stripping under limited Na sources further was credited to the dendrite-free Na deposition with robust SEI and relieved “dead Na,” aiding by the sodiophilic interphase with uniform F functionality. FCTF also exhibits a high CE, a low nucleation overpotential, and a stable hysteresis in Li anode (fig. S43).

Investigation of the low Na⁺ depletion in FCTF

These encouraging results manifest the essential role of FCTF in directing stable and reversible cycling, but achieving near-utility metal cation is still a longstanding challenge for Na metal batteries. Generally, high sodiophilic interphase layer leads to substantial high-potential Na⁺ depletion before nucleation. Taking porous carbon interphases as an example, their mass loading is around 1 to 3 mg cm⁻² (equal to Na⁺ depletion of 0.3 to 0.9 mAh cm⁻²; fig. S1A) according to the reported literatures. When assembled with the cathode such as Na₃V₂(PO₄)₃ (NVP), its areal capacity is around 0.2 to 0.6 mAh cm⁻² (fig. S1B). In this context, the assembled cell would be essentially a Na ion battery without any Na plating/stripping if under anode-free configuration. A typical curve of Na⁺ electrochemical reaction could be identified before nucleation in voltage profiles, i.e., the slopping-dominated interphase-depleted Na⁺, which is quantified by the corresponding areal capacity. Note that the FCTF interphase-depleted Na⁺ (0.052 mAh cm⁻²) is close to Cu foil, which is far lower than that of CTF (0.21 mAh cm⁻²) or other reported anodes (0.11 to 1.22 mAh cm⁻²) (5, 18, 25, 55, 56, 58–60) at 1 mA cm⁻² (Fig. 5A). In addition, FCTF exhibits very low values of interphase-depleted Na⁺ no matter the current density is up to 4 mA cm⁻² (fig. S14) (48, 51, 61, 62) or down to 0.05 mA cm⁻² (fig. S44 and table S7) (49, 63–65). Considering the mass loading of interphases also determines the absolute Na⁺ depletion before plating, and thus a coefficient (∆C/∆V) is introduced, defined as the reduced areal capacity to the decreased nucleation overpotential ratio compared with the corresponding control specimen for convincing study. Current studies favoring low nucleation barrier tend to result in high sloping capacity or vice versa, thus generally showing negative of ∆C/∆V. Unexpectedly, the ∆C/∆V values of FCTF are positive at different current densities (Fig. 5B), suggesting that this interphase layer depletes less Na and is thus favorable for plating/stripping (54, 60, 61, 66). Half cells with FCTF or CTF as working electrode and Na as counter electrode were assembled, and the galvanostatic charge and discharge (GCD) profiles show sloping region (Fig. 5C and fig. S45), corresponding to the Na⁺ storage in FCTF or CTF. Notably, FCTF delivers much lower capacity of 75 mAh g⁻¹, whereas CTF exhibits a notably higher specific capacity (Fig. 5D and fig. S46), suggesting low Na⁺ storage in FCTF interphases. As the F content increases, the specific capacity of FCTFs is gradually decreased (fig. S47). The difference in Na⁺ storage behavior is further characterized by cyclic voltammetry (CV), in which the curves are in agreement with GCD profiles. The first cathodic scan of CTF contains three dominant cathodic peaks at 0.85, 0.36, and 0.05 V (Fig. 5E), probably due to the insertion of Na⁺ on the triazine or benzene rings (67). Comparatively, the weakened peaks are detected for the FCTF, indicative of the weak Na⁺ reactivity, highlighting the poor Na⁺ storage capability of FCTF.

Fig. 5. Mechanism elucidation of the low Na⁺ depletion for FCTF.

Comparison of (A) interphase-depleted Na⁺ before nucleation and (B) ∆C/∆V of FCTF with other reported anodes under different current densities. The reduced areal capacity and decreased nucleation overpotential compared with the control sample are defined as ΔC and ΔV, respectively. (C) GCD profiles of CTF and FCTF at 0.1 A g⁻¹ after 100 cycles. (D) Cycling performances of CTF and FCTF at 0.1 A g⁻¹. The initial two cycles at 0.05 A g⁻¹ for activating were removed. (E) CV curves of FCTF and CTF at 0.1 mV s⁻¹ in half cells. Ex situ FTIR spectra recorded for the (F) CTF and (H) FCTF electrodes. The pristine electrode was prepared by disassembling the untested half cells. (G) Schematic illustration of sodiophilicity and Na⁺ reactivity on CTF and FCTF frameworks before deposition.

This distinct Na⁺ storage capability in FCTF/CTF frameworks was further elaborated by ex situ FTIR measurements and theoretic calculation. During the discharged (sodiation) process to 0.01 V, several peaks were assigned to the triazine and benzene rings (e.g., 803, 1349, 1510, and 880 cm⁻¹) of CTF electrode decay progressively, indicating the synchronized sodiation (Fig. 5F) (41). Besides, the blue shift can be observed on the peak of 803 cm⁻¹, possibly due to the inductive effect of the Na⁺ reactions with triazine rings (68). In contrast, only the peak of benzene ring in FCTF gradually weakened upon discharge process, while the triazine (1510 and 803 cm⁻¹) and C—F (986 cm⁻¹) peaks showed negligible changes, implying that the triazine is possibly passivated after introducing F functionality (Fig. 5H). Theoretically, Na⁺ affinity toward the triazine ring was studied via DFT computations. For the introduction of Na⁺ in the center of the triazine ring in CTF, the final position of Na⁺ in the relaxed structure remains unchanged, indicative of the strong affinity with Na⁺ (fig. S48A and movie S1). Notably, the Na⁺ migrates toward the adjacent benzene ring from the triazine ring in FCTF (fig. S48B and movie S2), further confirming that the impaired Na⁺ storage activity of triazine ring is due to steric hindrance effect of F. As reported, after introducing F functionality, the lowest unoccupied molecular orbital (LUMO) energy level falls, probably also responsible for the poor electrochemical performance of half cells (69). Obviously, FCTF displays the lower LUMO of −3.32 eV (fig. S49), corroborating the unfavorable Na⁺ storage capability. These results demonstrate that the Na⁺ activity on triazine ring of FCTF is weakened, the Na⁺ depletion of skeletons is reduced, and the Na utilization for deposition is thus enhanced (Fig. 5G).

Full cell evaluation towards low N/P ratio

To evaluate the potential application of the FCTF, full cells were assembled with NVP and Na₃V₂(PO₄)₂F₃ (NVPF) as the cathode. The FCTF/Na||NVP cell shows a capacity of 97 mAh g⁻¹ after 700 cycles under conventional excess Na condition (N/P = 14.5), while a sharp decay short circuit occurred within 230 cycles of CTF/Na||NVP (fig. S50A). Moreover, the full cell delivers a higher reversible specific capacity from 0.1 to 5 C (fig. S50B). Under hard conditions of low N/P ratio and high cathode loading, FCTF/Na||NVP cell still displays a good capacity retention (N/P of 2 and cathode loading of 7.1 mg cm⁻²; fig. S51). More intriguingly, when further decreasing the N/P ratio of 0, an average CE of up to 99.7% and a capacity decay of 0.11% per cycle over 400 cycles were achieved (Fig. 6A and fig. S52), which is twice less than that of CTF||NVP under the same conditions. On the contrary, the Cu||NVP cell displays a rapid capacity decay over 50 cycles (Fig. 6A). These outstanding results using FCTF interphase outperform previously published results (table S8). More insightful analysis of GCD profiles of full cell further validate the essential role of FCTF interphase (Fig. 6B). FCTF||NVP cell exhibits a high initial CE of 90.5% with a low nucleation overpotential of 5.3 mV (judging from the sharp peak in NVP redox plateau during the initial charge process), whereas the values are 80.5% and 10.8 mV for CTF||NVP, respectively. Likewise, the superior full cell performance can be extended to using NVPF cathode (figs. S53 and S54). Meanwhile, the low-cost Al current collector was also evaluated (14, 15), showing high specific capacity and good retention using FCTF interphase (fig. S55). These results certify that F-contained framework is valid in inhibiting dendrites growth, side reactions, and facilitating life span with great tolerance to different cathodes and current collectors.

Fig. 6. Electrochemical performance of full cells.

(A) Cycling stability of the anode-free FCTF||NVP, CTF||NVP, and Cu||NVP cells with a high NVP loading (7.4 mg cm⁻²) at 2 C. (B) Galvanostatic initial discharge/charge curves of FCTF||NVP and CTF||NVP between 2.5 and 3.8 V at 2 C. The inset displays the enlarged curves. (C) Mechanism of energy storage and reaction process of FCTF/CTF interphases in anode-free configuration. (D) First discharge profiles of the anode-free full cells with the NVP loading of 1.5 mg cm⁻² under 2 C. Because the charging/discharging processes are reversible, the proportions of above steps in charging process are presented more intuitively in the corresponding discharge profiles. (E) Specific capacity ratio of plateau to slope of the anode-free full cells under various cathode areal capacities. (F) Cycling stability of the FCTF/Na||NVP pouch cell (N/P = 1.5) with a high-loading cathode (NVP loading: 6.4 mg cm⁻²) at 2 C, the insert showing the schematic presentation of pouch cell. (G) Rate performance of the FCTF/Na||NVP pouch cell (NVP loading: 5.9 mg cm⁻²) at different current densities.

The effect of alleviated FCTF interphase-depleted Na⁺ on the capacity and output voltage in anode-free full cells was further investigated. The charge-discharge curve can be divided into a dominated plateau region and slopping region (Fig. 6C). The former is attributed to the Na plating/stripping (red curve), and the latter is the sodiation/desodiation of FCTF/CTF interphases with CTF-Na⁺/FCTF-Na⁺ (blue curve). In addition, all the Na⁺ is from cathode in such anode-free configuration. Under a given areal capacity of NVP (0.17 mAh cm⁻², 1.5 mg cm⁻²), it can be clearly seen that the Na⁺ from cathode was completely sodiated by CTF interphase with little Na⁺ for deposition (Fig. 6D), sacrificing the inherent advantages of Na anodes. In contrast, 75% of the Na⁺ source could be used for deposition in FCTF||NVP, further increasing the average discharge voltage by around 0.26 V. The enhanced output voltage and high capacity together lead higher energy density, as confirmed by the larger integral area of FCTF||NVP than that of CTF||NVP and Cu||NVP (Fig. 6D and fig. S56). With increasing the cathode areal capacity from 0.17 mAh cm⁻² (1.5 mg cm⁻²) to 1 mAh cm⁻² (8.5 mg cm⁻²), the ratio of plateau to slopping capacity of FCTF||NVP is always larger than that of CTF||NVP. For example, as the cathode areal capacity increases to 1 mAh cm⁻², the ratio for FCTF increases to 38.7, while that of CTF is only 9.4 (Fig. 6E and fig. S57). To our best knowledge, the Na⁺ consumption in anode interphase on capacity and output voltage of batteries is proposed for the first time, which is of great significance for the energy density.

Furthermore, the single layer pouch cells were assembled to investigate the commercialization prospects in the realistic conditions including low N/P ratio and high cathode loading (70). As proof of concept, 16-cm² FCTF/Na||NVP pouch cell (N/P = 1.5) displays a high initial specific capacity of 115 mAh g⁻¹, a capacity retention over 94.7% after 300 cycles (Fig. 6F), which is one of highest durability under limited Na conditions to the best of our knowledge (10, 71, 72). As displayed in the evaluation of the rate capability, the full cell achieves a reversible capacity of 114 mAh g⁻¹ at 0.5 C and 102 mAh g⁻¹ at 10 C (close to 6 min charging) with a lower polarization voltage, suggesting a fast-charging capability (Fig. 6G and fig. S58). The specific energy and power density were calculated on the basis of the total mass of FCTF, NVP, and deposited Na (table S9) and compared with the previously reported cells by a Ragone plot (fig. S59 and table S10). Encouragingly, a gravimetric energy density of 325 Wh kg⁻¹ and a maximized power output of 2.8 kW kg⁻¹ can be realized, representing a superior value in the reported literatures. The excellent performances of the FCTF/Na||NVP pouch cell are in stark contrast to the fast deterioration of specific capacity in the Cu/Na||NVP pouch cell (fig. S60). The N/P ratio of the FCTF||NVP pouch cell can be further reduced to 0.5 and even to 0 with stable long cycling at 2 C (figs. S61 and S62). A series of visual exhibition tests were carried out under harsh conditions to verify the practical potential of the anode-free pouch cells (figs. S63 and S64). These prominent results demonstrate the potential of porous polymers in the field of energy-dense and practically viable energy storage devices.

DISCUSSION

In summary, we have designed a dual seeding/hosting interphase with superior sodiophilicity and relieved Na⁺ activity based on fluorinated porous frameworks for robust anode-less Na metal batteries. In light of the well-orchestrated molecular structure, highly sodiophilic fluorinated edge channels are uniformly dispersed throughout the framework to achieve homogeneous Na nucleation and negligible Na⁺ depletion simultaneously, as judiciously confirmed by physicochemical and theoretical characterizations. Benefiting from these virtues, FCTF interphases could boost electrochemical performance in the following aspects: (i) Flat Na plating/stripping process and robust SEI can be realized, thus resulting in high average CE (99.6% for 400 cycles under 2 mA cm⁻²) in half cells and high interfacial durability under deep cycling of 95% DOD at a large areal capacity of 9.5 mAh cm⁻² in symmetric batteries. (ii) Together with suppressed interphase-depleted Na⁺, robust cycling of 400 cycles and superior plateau to sloping capacity ratio especially upon increasing cathode areal capacity are achieved in anode-free full cells. (iii) Owing to the molecularly designed interphases, the pouch cell (N/P = 1.5) delivers impressive cycling stability (94.7% capacity retention for 300 cycles), high-energy density (325 Wh kg⁻¹), and a maximized power density (2.8 kW kg⁻¹). This work provides anode-less prototyping for boosting the energy-dense device and sheds light on future practical Na batteries via elaborate interphase design.

MATERIALS AND METHODS

Material preparation

CTFs were synthesized by superacid-catalyzed strategy under neat conditions according to the previous literature (41). A mixture of the monomers (Aladdin Biochemical Technology Co. Ltd.) and trifluoromethanesulfonic acid (CF₃SO₃H; Innochem Science & Technology Co. Ltd.) was added into a pyrex tube successively, cooled down in liquid nitrogen, and flame-sealed under vacuum. Then, the tube was heated at 250°C for 12 hours with a heating rate of 5°C min⁻¹ in a muffle furnace. After the reaction, the pyrex tube was cooled down to room temperature, immersed in liquid nitrogen for 5 min, and then opened in the fume hood. The product was washed three times with water and ethanol to remove the residual solvent, collected by filtration, and dried under vacuum at 60°C for 24 hours. The powder collected was transferred into a crucible and then heated in a tubular furnace under a N₂ flow to 350°C for 2 hours with a heating rate of 5°C min⁻¹. A total of 256 mg of DCB (2 mmol) and 150 mg of CF₃SO₃H (1 mmol) were used for preparation of CTF. A total of 192 of mg DCB (1.5 mmol), 100 mg of DCFB (0.5 mmol), and 150 mg of CF₃SO₃H (1 mmol) were used for preparation of FCTF-1/3. A total of 128 mg of DCB (1 mmol), 200 mg of DCFB (1 mmol), and 150 mg of CF₃SO₃H (1 mmol) were used for preparation of FCTF. A total of 64 mg of DCB (0.5 mmol), 300 mg of DCFB (1.5 mmol), and 150 mg of CF₃SO₃H (1 mmol) were used for preparation of FCTF-3/1. CTF was obtained as a greenish powder, whereas the other specimens were obtained as a brown powder.

Physicochemical characterizations

The morphologies of samples were characterized through field-emission scanning electron microscope (NANOSEM450, FEI) and TEM (Talos F200X, FEI). The XRD (Bruker D8, Cu Kα, λ = 1.54056 Å) was used for crystal phase analysis. The FTIR spectra (Bruker TENSOR II infrared spectrometer) were used to analyze functional groups. XPS used for solid electrolyte interface layer analysis was performed by an AXIS Supra x-ray photoelectron spectrometer with a monochromatic Al Kα x-ray source. The N₂ adsorption-desorption measurements were conducted using an automated gas sorption analyzer (Micromeritics ASAP 2020), and the BET surface area was based on BET theory.

For the ex situ XPS and SEM characterizations, the electrodes were disassembled and washed in diglyme repeatedly followed by a drying processing in an argon-filled glove box. Afterward, the electrode was sealed in an Ar-filled tube and transferred into the vacuum chamber of the corresponding equipment as quickly as possible to avoid air exposure.

Cell assembling and electrochemical measurements

Electrochemical measurement of Na plating/stripping in half cell, symmetrical cell and full cell were assembled in CR2025-type coin cells with 1 M NaPF₆ in diglyme electrolyte in Ar-filled glovebox with O₂ and H₂O content below 0.1 parts per million. The electrochemical measurements were carried out on CT2001A model battery test system (LAND, Wuhan LAND Electronic Co. Ltd., China) at room temperature.

Preparation of FCTF or CTF interphases

For evaluating the Na plating/stripping behavior using FCTF or CTF interphase, they were prepared by the slurry coating method with 70 weight % (wt %) FCTF (or CTF), 20 wt % conductive carbon (super P), and 10 wt % polyvinylidene fluoride (PVDF) binder on Cu current collector. The incorporation of 20 wt % conductive carbon helps to enhance the electrical conductivity and achieves the Na deposition. Then, the foil was punched into discs with a diameter of 12 mm. The mass loading of these coatings is 0.5 to 0.7 mg cm⁻², and the thickness is around 20 μm. The coating process is the same when Al foil was used as a current collector.

Assembling of half cells for Na plating/stripping

The behavior of the Na plating and stripping was conducted using the half cells, in which FCTF-coated Cu serves as the working electrode and Na foil (10 mm) as the counter electrode. The polypropylene membrane (Celgard 2500) was used as the separator, and the electrolyte is about 35 μl. The assembled half cells were subjected with five precycling between 0.01 and 1 V (versus Na⁺/Na) under 0.1 mA cm⁻² to stabilize the SEI and remove surface contaminations. Afterward, a certain amount of Na (e.g., 1, 2, and 10 mAh cm⁻²) was deposited onto the FCTF-coated Cu current collector under the galvanostatic conditions with controlled time and then stripped to 1 V for each cycle. The CE was defined as the ratio of the charge (stripping) capacity to discharge (plating) capacity.

Assembling of symmetrical cells for Na plating/stripping

A certain amount of Na was first predeposited at 1 mA cm⁻² onto the above FCTF-coated Cu (or CTF-coated Cu and bare Cu foil) in half cells, which were then disassembled in glove box to obtain the predeposited Na loading electrode (denoted as FCTF/Na, CTF/Na and Cu/Na). The symmetric cells were assembled using the above two identical predeposited Na electrodes for the plating/stripping measurement. The amount of predeposited Na is fixed at 2 mAh cm⁻² with charge/discharge capacity of 2 mAh cm⁻², and it is fixed at 4 mAh cm⁻² with charge/discharge capacity of 1 mAh cm⁻², whereas it is fixed at 10 mAh cm⁻² with plating/stripping capacity of 9.5 mAh cm⁻². EIS analysis with symmetrical cells was carried out using an electrochemical work station (Bio-Logic VMP3) by applying an ac amplitude of 5 mV over the frequency range from 10 Hz to 1 MHz.

Assembling of coin-type full cells and pouch cells

An NVP or NVPF slurry was prepared by mixing the NVP or NVPF (MTI corporation), Super P, and PVDF with a weight ratio of 80:10:10 in N-methyl-2-pyrrolidone solvent. The obtained slurry was coated onto the Al foil and dried in vacuum oven at 80°C for 8 hours. The mass loading of NVP or NVPF was ~2.5 mg cm⁻². The limited or zero amount of Na was deposited on the above FCTF-coated Cu (or CTF-coated Cu and bare Cu foil) and served as the anode in the full cells. For the anode-free full cells, the modified current collector was activated five cycles under 0.1 mA cm⁻² to obtain stable SEI layer before the cell assembling. The full cells were assembled with 120 μl of electrolyte and glass fiber membrane (Whatman) and then tested with a voltage window of 2.5 to 3.8 V for NVP (2.5 to 4.3 V for NVPF). The preparation process is the same when Al foil was used as anode current collector.

For single-layer pouch cells, the NVP cathode and the FCTF-coated Cu current collector (or bare Cu) were cut into rectangles of 4 cm by 4 cm. The limited or zero amount of Na was predeposited on the FCTF-coated Cu (or bare Cu foil) electrode as the anode. The separator is a glass fiber membrane. The pouch cells were assembled and sealed in an Al plastic film packaging and conducted for charge/discharge cycling with a voltage range of 2.5 to 3.8 V.

Ionic conductivity measurement

FCTF (or CTF)–coated separator was prepared by the slurry coating with 90 wt % FCTF (or CTF) and 10 wt % binder PVDF on the polypropylene and assembled between two stainless steel plates in coin cells. The EIS data were collected with a frequency range from 0.01 to 100 kHz at Bio-Logic VMP3.

Na⁺ storage capacity measurement

The Na ion batteries were assembled in the form of CR2032-type coins using the glass fiber membrane as the separator, the above FCTF-coated Cu (or CTF-coated Cu and bare Cu foil) electrode, and Na metal as the working electrode and counter electrode, respectively. The electrolyte amount is 120 μl. Electrochemical experiments were performed at the voltage range of 0 to 2 V. The capacity was calculated on the basis of the mass of active materials.

DFT calculations

The calculations for adsorption between Na atom/ion and the molecule or periodic structure of CTF/FCTF were performed by using CASTEP based on DFT (73). A 15-Å vacuum space was adopted for the slab model to avoid interactions between neighboring layers. For electronic exchange-correlation interactions, the generalized gradient approximation (GGA) in the form of the Perdew-Burke-Ernzerhof (PBE) functional method was used. The energy cutoff was set as 598.7 eV, and the Monkhorst-Pack k-point grid was set as 1 × 2 × 1. Specially, the Monkhorst-Pack k-point grid of the Na atom related calculation was set as 2 × 2 × 2. The convergence conditions for geometry optimization were as follows: 1.0 × 10⁻⁵ eV/atom for energy, 0.03 eV/Å for force, 0.05 GPa for stress, and 0.001 Å for displacement. E_ad is defined as E_ad = E_slab+Na − E_slab − E_Na, where E_slab+Na, E_slab, and E_Na are the energy of the total system, adsorption substrate, and Na, respectively. The calculations for LUMO and charge density of CTF/FCTF were performed by using the Vienna Ab initio Simulation Package, which is based on DFT (74). For electronic exchange-correlation interactions, GGA in the form of the PBE functional method was used. The plane-wave basis set cutoff energy was set as 400 eV, and the Gamma point was set as k-point. During the calculation process, 1 × 10⁻⁵ eV/atom was set for energy convergence.

Method for XRD pattern simulation

The extended structure of CTF and FCTF was modeled on the basis of a hexagonal lattice and carried out the geometry optimization by the Materials Studio software package. Specifically, the pseudo-Voigt profile function was used for profile fitting, and the Berar-Baldinozzi function was used for asymmetry correction. The lattice parameters for the optimized unit cell are a = b = 29.39 Å and c = 3.4 Å.

Supplementary Materials

This PDF file includes:

Figs. S1 to S64

Tables S1 to S10

Legends for movies S1 and S2

References

Other Supplementary Material for this manuscript includes the following:

Movies S1 and S2