Non-woven based battery separators have been developed rapidly in recent years due to their unique advantages in process, cost and performance. In this study, a novel double-crosslinked three-dimensional network coating was constructed by hyperbranched polyethyleneimine (PEI), Polydimethylsiloxane diglycidyl ether (PDMSDGE) and poly(vinylidene fluoride-co-hexafluoropropylene) PVDF-HFP. Then the caoting used to modify tissue paper, which composed of packed cellulose fibers is typical nonwoven with excellent hydrophilic properties required for lithium separators. As we expected, the composite separator shows unique advantages in terms of porosity, electrolyte absorption, mechanical properties and thermal stability. At the same time, the nucleophilic hyperbranched PEI in the coating makes the composite separator in a lithium-rich state in lithium ion battery, so the separator has a certain electric pumping capacity by electrokinetic surface conduction and electroosmosis. As a result, the electrochemical performance and battery performance are excellent too. In addition, we compared the performance of the double-crosslinked coating with two systems of single-crosslinked coating, PEI/PDMSDGE and PEI/PVDF-HFP. The results showed the excellent performance of the composite separator with double-crosslinked coating in terms of mechanical properties retention (95.05%) after electrolyte infiltration and discharge capacity retention (86.89%) after 200 cycles, which given by the stable molecular structure of the three-dimensional double-crosslinked network.
Research Article
Tissue paper‑based composite separator using double-crosslinked polymer electrolyte as coating layer for lithium ion battery with superior ion transport and cycle stability
https://doi.org/10.21203/rs.3.rs-1651422/v1
This work is licensed under a CC BY 4.0 License
You are reading this latest preprint version
Composite separator
Tissue paper
Double-crosslinked network
Electrokinetic effects
Lithium-ion batteries have received widespread attention by their compact size, huge theoretical capacity and no memory effect since its invention. The separator, commonly accepted as the third electrode of lithium-ion battery, provides channels for the diffusion of lithium ion while isolating the positive and negative electrodes (Costa et al. 2019). With the demand of lithium-ion power batteries, separators are also required to improve performance which conventional polyolefin separators have fallen far short of (Lopez et al. 2019).
Among various new separators, nonwoven formed by disordered stacking of fibers show unique advantages in terms of porosity and pore structure, while their fabrication is easy to realize in the industrialization process. Various engineering plastics PPS (Zhu et al. 2019), PEEK(Li et al. 2018), aramid (Zhang et al. 2013) etc. have been selected to develop a rich variety of nonwoven separator systems by meltblown (Wang et al. 2016), electrostatic spinning (Li et al. 2019) and paper making methods (Zhang et al. 2013) etc. However, a single nonwoven applied as a separator directly often exhibits a decrease in capacity due to slight short circuits formed by some oversized pores, and thus further simple modifications such as spraying and dip coating are often required. With the use of various functionalized coatings and the high porosity tone set by the nonwoven itself, the composite separators often performs well in many aspects (Luo et al. 2018). And the development of this type of separator has received a lot of attention in the preparation process of the base nonwoven as well as types of coatings. However, the stability of the functional coating is equally important. The maintenance of the coating integrity during cycling and its bonding with the matrix are important issues which cannot be avoided in the further development of these separators.
The crosslinking system, which is abundantly used in the development of solid-state electrolytes to ensure the stability of electrolyte integrity, can also be applied in the coating design (Lin et al. 2021, Zhu et al. 2019). In this paper, a stable double-crosslinked polymer coating which can maintain the stability of its own structure and function was constructed by hyperbranched polyethylenimine (PEI), Polydimethylsiloxane diglycidyl ether (PDMSDGE) and Poly(vinylidene fluoride-co-hexafluoropropylene) (PVDF-HFP). Then a highly hydrophilic and thermally stable cellulose based tissue paper was selected as the matrix. The matrix and the coating are closely bonded via H-bonding between the abundant amino groups in PEI of coating and the abundant polar functional groups in the cellulose of tissue paper. The selection of the tissue paper matrix allows the separator to have a high hydrophilic, high electrolyte retention, and high thermal stability, while keeping the production cost of the composite separator under control (Deng et al. 2022). Meanwhile the nucleophilic nature of the abundant amino groups of PEI enables the separator to form a lithium-rich state in the battery. Then the lithium-rich channels of composite separator achieve rapid transfer of lithium ions based on the electrokinetic effect. So the efficiency of lithium ion diffusion improved (Li et al. 2018, Li et al. 2019). At the same time, PVDF-HFP gives the separator good dielectric properties and electrochemical stability (Jeschke et al. 2014), PDMS gives the coating good thermal stability and as a flexible segment to reduce the hardening of the coating due to high crosslinked ratio (Zhao et al. 2014). The double-crosslinked system maked the structure and function of the coating stability in charge/discharge of the battery, while achieved the control of coating properties through the regulation of the feed ratio.
Materials
Vinda tissue paper was purchased and employed as a matrix after pressed at 30 MPa. Poly(vinylidene fluoride-co-hexafluoropropylene) (PVDF-HFP, Mn = 100000) Polydimethylsiloxane diglycidyl ether (PDMSDGE, Mn = 800) and hyperbranched polyethylenimine (PEI, Mw = 600) were purchased from Sigma-Aldrich. Acetone and N, N-dimethylformamide (DMF) were supplied by Sinopharm Chemical Reagent Co., Ltd. 1 M lithium hexafluorophosphate (LiPF6) in ethylene carbonate (EC)/dimethyl carbonate (DMC) (1/1, v/v) was provided by Nanjing Mojiesi Energy Technology Co., Ltd as liquid electrolyte. Lithium iron phosphate (LiFePO4) and lithium metal anodes were purchased from Shenzhen Kejing Star Technology Co., Ltd.
Preparation of paper-based composite separator
The fabrication process of composite membrane is illustrated in Fig. 1 (a). Firstly, PDMSDGE and PEI (2:1, n:n) was dissolved in solvent (DMF/acetone = 4/1) at 60 oC and adequately stirred 48 h. Then, a certen PVDF-HFP (PEI : PVDF-HFP control as 1:10, m:m) subsequently added into the solution and keep stirring 0.5 h, which finally could give a double-crosslinked polymer network. As the prepared solution cooled down to room temperature, tissue paper substrate was immersed in and then removed out and dry in the air until the coating was set, afterwards dried at 60 oC for 12 h under vacuum. Finally, composite membrane CS (2:1:10) was successfully constructed. For comparison, another single-crosslinked composite membrane was fabricated by coating PDMSDGE/PEI and PEI/PVDF-HFP on tissue paper with similar procedure, which was named as CS (2:1) and CS (1:10) respectively.
Physical characterization and measurement
Fourier transform infrared spectroscopy (FTIR, BRUKER TENSOR-27) was used to analyze the formation of new chemical structures. A scanning electron microscopy (SEM, JEOL JSM-6510LV) was employed to examine surface morphologies of composite separators after gold sputtering treatment on samples. The porosity of separator was estimated by n-butyl alcohol immersion method and calculated using Eq. 1:
$$Porosity=\frac{{W}_{wet}-{W}_{dry}}{{\rho }_{b}{V}_{p}}\times 100\%$$
1
where Wdry and Wwet are respectively the mass of sample before and after immersed in n-butyl alcohol for 1 h, ρb is the density of n-butyl alcohol, and Vp is the volume of dry sample.
Contact Angle tester (FM40 Easy Drop, KRUSS) was carried out to measure the contact Angle of electrolyte droplet (2 µL) on the surface of the membrane. Macroscopically, the capillary absorption height was also measured after immersing one end of the sample in electrolyte about 2 h. In addition, separators were fully immersed in the electrolyte for 2 h, and when excess electrolyte were absorbed by filter paper, electrolyte uptake could be calculated according to Eq. 2:
$$Electrolyte uptake=\frac{{W}_{wet }-{W}_{dry }}{{W}_{dry }}\times 100\%$$
2
where Wwet and Wdry are respectively the mass before and after separator immersed in the electrolyte.
Mechanical behavior was tested with a universal testing machine (Instron-5967, USA) at room temperature. The strain rate of the machine was set at 10 mm/min and the standard sample was cut at 4 mm in width and 15 mm in length. The ratio of residual strength was tested as the same way after 16 µL of electrolyte infiltration, then the residual strength could be calculated by Eq. 3:
$$Residual strength=\frac{{{\sigma }}_{wet }}{{{\sigma }}_{dry }}\times 100\%$$
3
where σdry and σwet are breaking strength before and after membrane wetted by the electrolyte respectively.
Thermal dimensional stabilities of membranes were identified by heat treatment at 50°C, 100 oC, 150 oC and 200°C for 1 h respectively. Differential scanning calorimetry (DSC, 204 F1 NETZSCH) was employed to analysis thermal properties at the temperature range of 20°C–200°C with a heating rate of 10°C/min. Furthermore, thermogravimetric analysis (TGA, TG 209F1) was carried out at the temperature range of 20°C–800°C with a heating rate of 10°C/min. to further provide information for thermal stability.
Electrochemical analysis
SS/separator/Li cells were assembled to evaluate electrochemical stability window of separator through linear sweep voltammetry (LSV) measurement with the potential scanned from 1.0 V to 6.0 V under a rate of 10 mV/s.
Nyquist plots of symmetrical cells (SS/separator/SS, SS = stainless-steel sheets) were gained by scanning electrochemical impedance spectroscopy (EIS) on an electrochemical workstation (Ivium Stat. h) with an AC voltage of 10 mV amplitude in the frequency range of 0.1 to 106 Hz. The ionic conductivity (δ) could be calculated with Eq. 4:
$$\delta =\frac{d}{RA}$$
4
where d and A respectively represent the thickness and effective testing area of separator, and R is the bulk resistance of separator and obtained from Nyquist plots.
To analyze interfacial resistance between separator and electrode, electrochemical impedance spectroscopy (EIS) of half cells (LiFePO4/separator/Li) were measured on an electrochemical workstation (Ivium Stat.h) through the same way as above. In addition, we combined chronoamperometry and electrochemical impedance spectra (EIS) of Li/separator/Li cell to estimate lithium ion transference number according to Eq. 5:
$${t}_{{Li}^{+}}=\frac{{I}_{ss}\varDelta V-{I}_{0}{I}_{ss}{R}_{0}}{{I}_{0}\varDelta V-{I}_{0}{I}_{ss}{R}_{ss}}$$
5
where I0 and Iss are the initial and steady-state current obtained from chronoamperometry, respectively; R0 and Rss are the initial and steady-state interfacial resistance acquired from EIS, respectively; ΔV is the step potential difference (10 mV).
Moreover, in order to characterize the diffusion coefficient of lithium ion of separator, cycle voltammetry (CV) measurement of half cells (LiFePO4/separator/Li) were measured on an electrochemical workstation (Ivium Stat. h) with the potential scanned from 2.5 V to 4.2 V under a rate of 0.1 mV/s.. Then compare the diffusion coefficient by Eq. 6:
Where Ip is the peak current; n is the number of electron transfers in redox reactions (1 in lithium-ion battery); A is the area of electrode (2.01 cm2); C is the ionic concentration; D is the diffusion coefficient; V is the scan rate (0.1 mV/s)
Rate and cycle performance characterization
Symmetrical cells (Li/separator/Li) were assembled to evaluate stability between the separator and the anode during the cycle by using LAND battery testing system (CT2001A, Wuhan LAND Electronic Co., Ltd, China) at room temperature with a current density of 0.01 mA cm2 and 1 mA cm2 for 2000 h and 500 h respectively
2016-type coin cell was assembled in an argon filled glove box by sandwiching 60 µL liquid electrolyte-soaked separator between a LiFePO4 cathode and a lithium foil anode. LiFePO4 cathodes were prepared by coating NMP-based mixture of acetylene black, polyvinylidene fluoride and LiFePO4 (1:1:8, m:m:m) on aluminum foils in our lab. Cell performance was tested on a LAND battery testing system with the cut-off voltage range set from 2.5 V to 4.2 V. In the tests of rate capabilities, these cells were charged at 0.2 C rate and discharged at various rates ranging from 0.2 C to 2 C and then back to 0.2 C. At each rate the cells were charged/discharged for 10 times. During cycling tests, these cells were cycled at a fixed charge/discharge current density 0.5 C/0.5 C for 200 times at room temperature.
Crosslinked structure fabricated on tissue paper
Fig. 1 (a) describes the procedure of double-crosslinked coating layer constructed on tissue paper. Firstly, precursor solution was prepared by mixing Polydimethylsiloxane diglycidyl ether (PDMSDGE), hyperbranched polyethyleneimine (PEI) and poly(vinylidene fluoride-co-hexafluoropropylene) PVDF-HFP adequately in solvent. The synthesis of double-crosslinked network is confirmed via FT-IR spectroscopy as shown in Fig. 1 (b). About PEI, it exhibits broad bands at 3279 cm−1 assigned to the N-H bonds and a sharp band at 1594 cm−1 due to NH2 bonds, which are the basis for building cross-linked network through chemical reactions (Chen et al. 2019, Lim et al. 2019). We constructed a first-crosslinked network by PDMSDGE and PEI through epoxy ring opening reaction (Huang et al. 2014, Lee et al. 2019, Lu et al. 2017). For PDMSDGE, characteristic bands of epoxide groups at 915cm−1 and C-H bonds in epoxide groups as other peak around 3057 cm−1 observed (Lim et al. 2019). These signals almost disappear in the CS (2:1) which completed the epoxy ring opening reaction. In addition, the broad bands around 3600 cm-1 in CS (2:1) assigned to stretching vibration of the newly formed −OH (Zhao et al. 2014). After that, we constructed a second-crosslinked network by PVDF-HFP and PEI through the same way as our previous study (Zeng et al. 2022, Zhang et al. 2020). The signal of NH2 at 1594 cm−1 still existed in CS (2:1) but disappeared in CS (2:1:10), which means chemical reaction happened between NH2 and PVDF-HFP (Jeschke et al. 2014, Zhou et al. 2014). On the other hand, we can find the evidence of crosslinking reaction from the viscosity of different solutions as seen in Fig. S1. Next, a certain size tissue paper was dipped into the pre-crosslinking coating solution. Then the final composite separators were obtained after experienced vacuum drying at 60 oC. Futhermore, higher stiffness and strength of CS (2:1:10) compared to the single-crosslinked systems CS (2:1) and CS (1:10) shown in Fig. 2 (a) indicated a more strongly crosslinked network then the two single-crosslinked systems had been formed in the coating, which further provides supporting evidence for the occurrence of double-crosslinking reactions in this system.
Morphology and pore structure of composite separator
The separator sandwiched between the two electrodes to physically isolate the positive and negative electrodes, while ensuring the diffusion of lithium ions through the pores with the electrolyte (Lee et al. 2016). From the Table 1, we can see the porosity of CS (2:1:10) (70%) > CS (1:10) (60%) > CS (2:1) (40%) > Celgard 2400 (35%). Furthermore the specific pore structure we can see from the SEM images of Fig. 1 (c)-(h). And the Fig. 1 (c) and (d) represents the morphological picture of the CS (2:1) at different magnifications, without the imagined rich porous structure, in addition to many closed pores. This mainly because the viscosity of PDMSDGE and PEI single-crosslinked solution is too low, which limited by the weight of the molecular. Thus the coating tend to form a dense structure during the drying and shaping process due to the high liquidity (Pan et al. 2017). On the contrary, Fig. 1 (e) and (f) represented the surface morphology of CS (1:10) which also prepared from single-crosslinked network by dip coating method, but we are able to observe a rich pore structure. This is due to the selection of large molecular weight PVDF-HFP in this system which sets large viscosity of the pre-crosslinked solution. This allows the coating to be shaped in a shorter time, thus facilitating the generation of pores. Our previous work found that the semi-crystallinity of the PVDF-HFP can still adversely affect the pore formation of the coating to a certain extent in the 1:10 cross-linked system of PEI and PVDF-HFP (Zeng et al. 2022, Zhang et al. 2020). At the same time, the pore-forming properties can be significantly improved after further doping. Fig. 1 (g) and (h) is the surface morphology map of CS (2:1:10) at different magnifications. Rich surface morphology and the complex pore structure of large pores over small pores as well as the presence of more small pores indicate that the double-crosslinked system is more prone to pore formation, which is also consistent with the findings in our previous work (Zeng et al. 2022). Furthermore from the Fig. 2 (a) also indicates that the complex molecular structure of the double-crosslinked coating makes no affected to the mechanical properties eventhough the crystallinity reduced. Obviously, more small pores and a certain number of large pores can produce a rich pore structure while increasing the porosity of the composite separator, and this can be achieved by constructing a suitable cross-linked network for the PVDF-HFP.
Table 1 Basic properties of Celgard 2400 and composite separators
|
|
Thickness (μm) |
Bulk resistance (Ω) |
Ion conductivity (ms/cm) |
Porosity (%) |
Electrolyte uptake (%) |
|
Celgard 2400 |
24 |
7.8 |
0.16 |
35 |
80 |
|
CS (2:1) |
45 |
7.4 |
0.3 |
40 |
100 |
|
CS (1:10) |
52 |
5.9 |
0.44 |
60 |
210 |
|
CS (2:1:10) |
55 |
4.7 |
0.59 |
70 |
230 |
Mechanical behavior
Certain mechanical properties are the basis to ensure the success of the separator assembly cell and a certain extent to prevent lithium dendrite piercing (Zhong et al. 2021). From Fig. 2 (a) it can be found that pure tissue paper has a very low tensile strength of only 2.9 MPa. After modification with single-crosslinking coating , the composite separator CS (2:1) reached to 5.1 MPa. But due to the limitation of molecular weight of coating solution (PDMSDGE:PEI, 2:1), the strength of the coating itself is not well. So the coating modification can’t enhance the mechanical properties of tissue too much. On the contrary, the breaking strength of CS (1:10) which is also tissue paper modified by single-crosslinked coating (PEI:PVDF-HFP, 1:10) reached to 10.8MPa. Unfortunately, the Young's modulus of the composite separator CS (1:10) was not as high as the CS (2:1) due to the limitation of PVDF-HFP itself. But after the coating modification of our double-crosslinked network, the composite separator CS (2:1:10) inherits the advantages of both the high Young's modulus of CS (2:1) and the high breaking strength of CS (1:10). In addition, the more strong molecular structure of the double-crosslinked network allows the breaking strength of CS (2:1:10) reach to 19.9 MPa, which is even higher than the value of transverse direction of Celgard 2400. There is no doubt that, it’s a new height for the mechanical properties of composite separator based on tissue paper.
The double-crosslinked network structure of coating we designed takes the mechanical properties of the paper-based composite separator to a new level is only one aspect of our original design intent (Zhong et al. 2021). Our other aim is using the stable molecular structure of the double-crosslinked network to maintain the structural and functional of the composite separator’s coating stability during cycling (Xiao et al. 2011). To characterize this performance, we simulate the state of the separator immersed in electrolyte in the cell by applying a drop of electrolyte at the appropriate level to a stretching strip of the composite separator according to the area. The solid lines in Fig. 2 (b) are the original mechanical property curves of each composite separator as in Fig. 2 (a). The dashed lines in Fig. 2 (b) show the mechanical property curves of each composite separator under electrolyte infiltration. It can be seen that the cross-linked coating modified composite separator can still maintain quite good mechanical properties in the wet state, although the tensile strength of CS (1:10) is slightly lower compared to the other two. This is probably due to the loss of the original chain structure of PVDF-HFP, which is easily swollen by solvent molecules (Gor et al. 2015). This is also one of the reasons for the significant decrease in the discharge capacity of the half-cell assembled by CS (1:10) after 140 cycles as shown in the Fig. 5 (a). Meanwhile, the specific values in Fig. 2 (c) of each separator’s tensile strength in dry and wet states show that the stable coating network of the double-crosslinked system CS (2:1:10) can maximize the retention (95.05 %) of tensile strength after absorption of electrolyte, while the other two single-crosslinked systems CS (2:1) and CS (1:10) are only 70.18 % and 65.13 % respectively. The higher retention of the mechanical properties of the double-crosslinked composite separator in the wet state indicates that the double-crosslinked coating can maintain a higher structural stability after solvent penetration, which strongly confirms the stability of the net-like molecular structure of the double-crosslinked system (Zhu et al. 2019). This excellent performance is also one of the reasons why the half-cells in the CS (2:1:10) in Fig. 5 (a) can maintain almost no decline under 200 charge/discharge cycles.
Wettability and thermal analysis
During the operation of the battery, lithium ion diffuse through the channels of separator with electrolyte, so the wettability performance of the separator is crucial (Zhang et al. 2020). Fig. 2 (d) is contact angle conducted by dripping 2 μL electrolyte on the surface of composite membranes and Celgard 2400. We can see that CS (1:10) exhibits a satisfactory contact angle (32o), much smaller than that of Cellgard 2400 (62o), and this value is rapidly reduced to 11o after 20 s, in contrast to Celgard 2400, which is only reduced to 58o. The excellent hydrophilic properties due to the incorporation of hyperbranched PEI and the choice of tissue paper matrix (Zeng et al. 2022). At the same time, the rich pores of the morphology and the swelling of PVDF-HFP on the electrolyte are the reason why the CS (1:10) diffused rapidly within a short time of 20 s. On the contrary, CS (2:1), which also uses hyperbranched PEI and tissue paper matrix, does not expose enough tissue fibers to contribute to the electrolyte uptake due to the less pores of the morphology, and the super hydrophobicity of the PDMSDGE chain segment makes the contact angle of the CS (2:1) far inferior to CS (1:10). And the diffusion of electrolyte within 20 s is also inferior to that of CS (1:10), even compared to commercial separator Celgard 2400. Nevertheless, The CS (2:1:10) composite separator exhibits an excellent contact angle of 49o and a reduction to 26o only after 20 s, despite the addition of PDMSDGE. It is slightly worse than CS (1:10), but the gap is very small. And it has a clear advantage over CS (2:1). On the one hand, this is due to the fact that the amount of PDMSDGE in the CS (2:1:10) system is not large and the PVDF-HFP chain segment is the main component of the coating, on the other hand, the fact that CS (2:1:10) has most abundant pores, as shown in Fig. 1 (d) and (i). By comparison, we can easily find that for composite separator whose matrix is a hydrophilic material, the most important factor affecting the hydrophilic performance of the separator is the abundance of surface pores. In addition to this, the hydrophilic properties of each component of the coating exert different degrees of influence on the composite separator through the amount of its composition. Also the immersion-height photograph at the time of separator immersed in liquid electrolyte 2 h shown in the Fig. S2 demonstrates the Similar laws CS (1:10) (3.2 cm) > CS (2:1:10) (2.5 cm) > CS (2:1) (0.7 cm) > Celgard 2400 (0.3 cm). This test can indicate the hydrophilicity of the separator and the richness of the pore structure inside the separator to some extent. Moreover, the absorption rates which determined by a combination of material hydrophilicity and porosity shown in the Table 1 exhibit a similar pattern to the contact angle CS (2:1:10) (230%) > CS (1:10) (210%) > CS (2:1) (100%) > Celgard 2400 (80%). The reason why the law is slightly different from the contact angle is that the porosity has a greater impact on the rate of liquid absorption, while the swelling of PVDF-HFP will make a large difference in the rate of liquid absorption of separator with and without this component (Zhang et al. 2020). The better hydrophilicity and electrolyte absorption rate are the main reasons why the composite separator in Fig. 3 (b) can exhibit low impedance in cells with blocked electrodes.
If the separator's hydrophilicity and electrolyte absorption rate are defined as the basis of the battery's power performance, then the thermal stability of the sepatator can be defined as the guarantee of the battery's safety performance (Zhang et al. 2021). Fig. 2 (e) shows the DSC results of each composite separator, and there are obvious heat absorption peaks at 88 oC for CS (2:1) and at 134 oC for CS (1:10). This corresponds to the melting of the main chains of PDMSDGE and PVDF-HFP coatings, respectively. However, the peak at 88 oC completely disappeared and the peak at 134 oC largely disappeared too, which comes from the strong binding of the double-crosslinked web-like molecular structure for the free molecular chain ends, resulting in a further reduction of the motility of the molecular chain segments (Wang et al. 2021). The thermal stability of the composite separator is strongly enhanced. Then the existence of a double-crosslinked network is also further confirmed from the other side. Also from the Fig. 2 (f) we can see the dimensional changes of different separator in the macroscopic state at different temperatures. Celgard 2400 separator are completely molten at 200 oC, which is due to the lower melting point (189 oC) of commercial separator’s PP materials. In comparison, each composite separator CS (2:1), CS (1:10) and CS (2:1:10) at least in the 200 oC range are showing no dimensional changes, which of course has a more stable cross-linked network, the most important is the high thermal stability of the tissue paper’s cellulose fiber which maintains dimensional integrity with slight melting of the coating (Wei et al. 2019). The color change of the composite separator with the increase of temperature is due to further crosslink reactions and oxidation of unreacted amine groups in PEI. Moreover TGA curves shown in Fig. S3, the crosslinked network give composite separators with higher solid residue when temperature increased above to 600 oC. By comparison, Celgard 2400 separator leaves no solid basically when temperature rising to 475 oC. Cellulose of tissue paper matrix and cross-linked coating together give the composite separator macroscopic dimensional stability at high temperature which can effectively avoid short circuit of the battery due to shrinkage of the separator, is the guarantee of battery safety performance.
Electrochemical performance
The electrochemical stability window is the basis for the separator to be applied in lithium-ion batteries and remain stable over their operating voltage range (Zhang et al. 2018). The current as the voltage increases by Linear sweep voltammetry (LSV) method based on SS/separator/Li cell can be seen in Fig. 3 (a). It can be found that the value of the voltage at which the anode current can be observed first for each separator is Celgard 2400 (4.2 V) < CS (2:1) (4.4 V) < CS (1:10) (4.5 V) < CS (2:1:10) (4.8 V) respectively. The value of this voltage means that the oxidation decomposition reaction in this system has started under the conditions (Costa et al. 2019), and there is no doubt that higher of this voltage means that the separators’ broader electrochemical window. The broadest electrochemical window of CS (2:1:10) is due to the stable chemical structure of the three-dimensional network of molecules of the double cross-linked coating, which can effectively maintain its structural and functional stability even at high voltages. This excellent performance makes our double-crosslinked coated composite separator capable of handling conventional lithium-ion batteries, even then showing some promise in the high-voltage battery.
The bulk impedance is usually used to constantly measure the number of effective lithium ion channels in the separator (Lopez et al. 2019). And we measured it by electrochemical impedance spectroscopy (EIS) using SS/separator/SS symmetrical cell. The Nyquist plot whose intersection with X-axis is assigned as the bulk resistance of separator. And we can see in Fig. 3 (b) that CS (2:1:10) (4.8 Ω) < CS (1:10) (5.9 Ω) < CS (2:1) (7.4 Ω) < Celgard 2400 (7.8 Ω). The bulk impedance determine the degree of ohmic polarization during battery operation together with interface impedance (Zhao et al. 2020). Through Eq. 3 combined with the thickness and effective area of each separator we can calculat the ionic conductivity shown in Table 1 CS (2:1:10) (0.59 ms/cm) > CS (1:10) (0.44 ms/cm) > CS (2:1) (0.3 ms/cm) > Celgard 2400 (0.16 ms/cm). Higher of the value means that the separator’s lithium ion transfer capability is stronger, which is one of the most important reasons why the composite separator CS (2:1:10) in Fig. 5 (a) can maintain a high value of discharge capacity on the first few cycles.
Lithium ion transference number is an other important parameter determined the degree of concentration polarization (Yuan et al. 2021). In this study, they were estimated by combining chronoamperometry and EIS of Li/separator/Li symmetrical cells as shown in Fig. 3 (c-f) and calculated by Eq. 4. The tLi+ of separators present an increasing trend in the order of Celgard 2400 (0.22) < CS (2:1) (0.39) < CS (1:10) (0.52) < CS (2:1:10) (0.61). An important reason why the lithium ion transference number of the composite separators are much higher than Celgard 2400 is shown in our previous work, that the cross-linked network of the coating has a physical binding effect on the larger ions PF6- which will restrict mobility of anion and give lithium ions more chance to take part in charge transfer (Zeng et al. 2022). Compared to single-crosslinked coating, the network of double-crosslinked coating is undoubtedly more binding, and thus CS (2:1:10) exhibits a larger lithium ion migration number than CS (2:1) and CS (1:10). However, the lithium ion transference number of CS (2:1) is not in the same level as CS (1:10) and CS (2:1:10), probably because the electrolyte absorption rate is too low to make the cross-linked networks playing a vital role about lithium ion transference number. Higher lithium ion migration number can significantly reduce the concentration polarization, which is very important for the cycle stability and high rate charge/discharge capacity of the battery.
C-rate capacity of battery
The C-rate charge/discharge performance is the battery's ability to charge and discharge at different C-rate (Yuan et al. 2021). The polarization at a higher current density will have different influence on the C-rate performance of the battery (Xia et al. 2021). The ohmic polarization and concentration polarization have been discussed in the previous section, while the difference about the lithium ion diffusion coefficient can also affect polarization. As seen in Fig. 4 (c) the current peak Ip of oxidation in CV test present an increasing trend in the order of CS (2:1) (104.2 μA) < Celgard 2400 (163.3 μA) < CS (2:1:10) (189.3 μA) < CS (1:10) (403.3 μA). According to the Randles-Sevcik equation, the value of the current peak (Ip) is positively related to the lithium ion diffusion coefficient (DLi+) when the same tape cell is scanned at the same rate. Then we can compare relative speed about the lithium ion diffusion. The reason for the high DLi+ of the composite separator is shown in the Fig. 4 (a) and (b),which are the internal ion diffusion mechanism of the battery assembled with composite separator and conventional Celgard 2400 respectively. In the composite separator, due to the choice of nucleophilic cross-linking center PEI, the coating of composite separator will form a lithium-rich state after electrolyte wetting (a self-concentrating feature) and formed a electric double layer (EDL) together with the anions adsorbed by the electrostatic force. Meanwhile, the interfacial electric double layer (EDL) promotes the electrokinetic surface conduction and electro-osmosis (that is an electrokinetic pumping feature), similar to a variety of electrokinetic phenomena in porous media under an electric field, enhanced mass transport during Li plating/stripping (Li et al. 2018, Li et al. 2019). In contrast, Fig. 4 (b) represent the battery equipped with conventional Celgard 2400 which have only a normal cation diffusion mode because they cannot form a electric double layer (EDL). The Li-ion electrokinetic self-concentrating and pumping features of our composite separator can synergistically reduce the concentration polarization and overcome the diffusion-limited current. Therefore regulate the Li-ion concentration difference in thedeionization area to enable a uniform Li-ion distribution. But about the CS (2:1) same as the Li-ion tranference number that the theory of electrokinetic phenomena doesn’t work since the electrolyte absorption rate is too low to make the cross-linked coating playing a vital role. However, the low voltage (3.689 V) of the CS (2:1) oxidation Ip is still inherited by CS (2:1:10) (3.69 V), which the Celgard 2400 and CS (1:10) are 3.72 V and 3.82 V respectively. The lower voltage of oxidation Ip means the lower polarization of the battery too. Therefore the C-rate capacity of battery got enhanced and the specifics are shown in Fig. 4 (d). Meanwhile Fig. 4 (e-h) are the details of voltage-capacity curves in the charging and discharging process at different rates of each separators displayed in Fig. 4 (d). As the discharge rate increases from 0.2 C to 2 C, the polarization phenomenon was more aggravated, which will cause the reduction of discharge voltage and shorten discharge time to less than theoretical value, thus leading to attenuated discharge capacity as we can see from Fig. 4 (e-h). In comparison, CS (2:1:10) has the smallest reduction when discharging at a high rate of 2 C, which means that the CS (2:1:10) can make the battery produce only the smallest polarization phenomenon, and the 0.2 C after 2 C discharge rate can return to almost the same small rate as the beginning, which indicating its better reversibility in cell performance. This outstanding cell performance at elevated rate will provide possibility for the application of CS(2:1:10) separator in high power LIB.
Cycle capacity of battery
The cycle capacity of battery is maintain a high discharge capacity with multiple cycles, which is a premise of application (Xia et al. 2021). As shown in Fig. 5(a), 200 charge/discharge cycles were carried out from different half cell assembled by differrnt separator at 0.5 C/0.5 C to evaluate cycling performance. We can find all of the four systems shown almost 100% coulombic efficiency. Moreover, at the first cycle, cell of CS (2:1) demonstrated the lowest discharge capacity (92 mAh/g), fllowed by Celgard 2400 (115 mAh/g), CS (2:1:10) (145 mAh/g) and CS (1:10) (149 mAh/g). This is consistent with the impedance pattern of each cell before the cycle obtained from the impedance of half-cells in Fig. S4, which should be the main reason for the difference in discharge capacity. Interestingly, the CS(2:1) as well as CS(2:1:10) discharge capacity starts to rise slightly from about the sixth cycle, while the different cells of separators follow CS (2:1:10) (148 mAh/g) > CS (1:10) (146 mAh/g) > CS (2:1) (120 mAh/g) > Celgard 2400 (117 mAh/g). This law is consistent with the impedance law of each separator discussed earlier, which is also the main reason for the difference in discharge capacity of different batteries assemblied by different separators. The slight increase in the discharge capacity of CS (2:1) and CS (2:1:10) after several charges and discharges is probably due to the further infiltration of the electrolyte, which allows the lithium ion channels of the separator to be fully opened (Zhao et al. 2020), thus enabling the most efficient transfer of lithium ion. As the cycle going, the resistance increases with the polarization deeping. As we can see in Fig. S4, the resistance growth after cycled and even the lithium dendrites appearring in the insert of Fig. 5 (b) and (c). That makes the discharge capacity decrease rapidly. The 200th discharge capacity of battery with different separators in order of Celgard 2400 (18 mAh/g) < CS (1:10) (85 mAh/g) < CS (2:1) (86 mAh/g) < CS (2:1:10) (125 mAh/g). The excellent cycling performance of CS (2:1:10) as we discussed before, the stable coating structure of the double cross-linking can bring structural and functional stability to the composite separator, the similar we can find about the impedance variation shown in Fig. S4.
To further investigate the reasons for the excellent cycling performance of the double-crosslinked separator CS (2:1:10), we assembled a Li/separator/Li symmetric cell and performed constant current charge/discharge tests to characterize the stability of the composite separator to the lithium metal anode during long-term cycling (Xia et al. 2021), meanwhile Celgard 2400 was also selected as a control group. Fig. 5 (b) was cycled in 0.1 mA/cm2 for 1000 cycles, and 1 h for half cycle. The Fig. 5 (c) was cycled in higher current density 1 mA/cm2 for just 250 cycles, and 1 h for half cycle too. In this test, the beginning of several cycles of higher voltage such as Fig. 5 (b) is due to the voltage hysteresis caused by polarization, shorter the duration is better, such as CS (2:1:10). The higher of the voltage means the higher resistance of the cell such as Celgard 2400 in Fig.5 (b) and (c), meanwhile the depletion of the electrolyte and the growth of some lithium dendrites lead to a sudden change in resistance that can cause a steep increase in voltage such as 180th cycle of Celgard 2400 in Fig. 5 (c). The better cycling stability of CS (2:1:10) compared to the anode is due to the lower impedance and stable structure of the double-crosslinked system coating as we disscussed before. On the other hand, the encapsulation of the electrolyte by the rich cavities of hyperbranched PEI, which enables the separator to absorb the rich electrolyte and not to leak out during the cycling process, cutting down charge transfer impedance and solid electrolyte interface (SEI) film impedance (An et al. 2022). There is no doubt that the stability between the CS (2:1:10) and the metal anode under different current is an important reason for the performance of the battery assembled by CS (2:1:10), and it is also the basis for the CS (2:1:10) to be used in high power lithium-ion batteries.
We used hyperbranched PEI as the crosslinking center, and selected PDMSDGE and PVDF-HFP to construct a structurally stable three-dimensional double-crosslinking network with suitable modulus, mechanical strength and a wide electrochemical stability window through the adjustment of crosslinking ratio. The composite separator with rich porosity, good hydrophilicity, high liquid absorption rate and excellent thermal stability was prepared by coating it on tissue paper matrix. The stable and complex three-dimensional double-crosslinked molecular structure has a good binding effect on the lithium salt anion so that the separator shows a good lithium ion transference number. And the presence of PEI in the cross-linked coating makes the separator in lithium-rich state after assembly in the battery, so that the rate of lithium ion diffusion coefficient through the separator is greatly accelerated by the electrokinetic phenomena. At the same time, the network structure of the double-crosslinked coating can maintain the structure and function stable for a long time. As a result, the half-cell assembled by the composite separator exhibits excellent cycling and C-rate performance. There is no doubt that the three-dimensional double-crosslinked coating has excellent application prospects in nonwoven separators.
Acknowledgments
This work was supported by Hubei Provincial Natural Science Foundation of China (2018CFB267), the Open Project Program of Beijing National Laboratory for Molecular Sciences (BNLMS201821), the Open Project Program of Key Laboratory for New Textile Materials and Applications of Hubei Province (FZXCL202105).
- An Q, Wang H, Zhao G, Wang S, Xu L, Wang H, Guo H (2022) Understanding Dual-Polar Group Functionalized COFs for Accelerating Li‐Ion Transport and Dendrite‐Free Deposition in Lithium Metal Anodes. Energy Environ Mater 0:1–10. https://doi.org/10.1002/eem2.12345
- Chen C, Chen F, Liu L, Zhao J, Wang F (2019) Cross-linked hyperbranched polyethylenimine as an efficient multidimensional binder for silicon anodes in lithium-ion batteries. Electrochim Acta 326:134964. https://doi.org/10.1016/j.electacta.2019.134964
- Chuang YP, Lin YL, Wang CC, Hong JL (2021) Dual Cross-Linked Polymer Networks Derived from the Hyperbranched Poly(ethyleneimine) and Poly(acrylic acid) as Efficient Binders for Silicon Anodes in Lithium-Ion Batteries. ACS Appl Energy Mater 4:1583–1592. https://doi.org/10.1021/acsaem.0c02802
- Costa CM, Lee YH, Kim JH, Lee SY, Lanceros MS (2019) Recent advances on separator membranes for lithium-ion battery applications: From porous membranes to solid electrolytes. Energy Stor Mater 22:346–375. https://doi.org/10.1016/j.ensm.2019.07.024
- Deng J, Cao D, Yang X, Zhang G (2022) Cross-linked cellulose/carboxylated polyimide nanofiber separator for lithium-ion battery application. Chem Eng J 433:133934. https://doi.org/10.1016/j.cej.2021.133934
- Gor GY, Cannarella J, Leng CZ, Vishnyakov A, Arnold CB (2015) Swelling and softening of lithium-ion battery separators in electrolyte solvents. J Power Sources 294:167–172. https://doi.org/10.1016/j.jpowsour.2015.06.028
- Gu L, Zhang M, He J, Ni P (2018) A porous cross-linked gel polymer electrolyte separator for lithium-ion batteries prepared by using zinc oxide nanoparticle as a foaming agent and filler. Electrochim Acta 292:769–778. https://doi.org/10.1016/j.electacta.2018.09.147
- Huang LY, Shih YC, Wang SH, Kuo PL, Teng H (2014) Gel electrolytes based on an ether-abundant polymeric framework for high-rate and long-cycle-life lithium ion batteries. J Mater Chem A 2:10492–10501. https://doi.org/10.1039/c4ta01527d
- Jeschke S, Mutke M, Jiang Z, Alt B, Wiemhofer HD (2014) Study of carbamate-modified disiloxane in porous PVDF-HFP membranes: new electrolytes/separators for lithium-ion batteries. ChemPhysChem 15:1761–1771. https://doi.org/10.1002/cphc.201400065
- Lee JH, Kim S, Cho M, Chanthad C, Lee Y (2019) Crosslinked Gel Polymer Electrolytes for Si Anodes in Li-Ion Batteries. J Electrochem Soc 166:A2755–A2761. https://doi.org/10.1149/2.0311913jes
- Lee Y, Park J, Jeon H, Yeon D, Kim BH, Cho KY, Lee YM (2016) In-depth correlation of separator pore structure and electrochemical performance in lithium-ion batteries. J Power Sources 325:732–738. https://doi.org/10.1016/j.jpowsour.2016.06.094
- Li G, Liu Z, Huang Q, Gao Y, Regula M, Wang D, Wang D (2018) Stable metal battery anodes enabled by polyethylenimine sponge hosts by way of electrokinetic effects. Nat Energy 3:1076–1083. https://doi.org/10.1038/s41560-018-0276-z
- Li G, Liu Z, Wang D, He X, Liu S, Gao Y, Wang D (2019) Electrokinetic Phenomena Enhanced Lithium-Ion Transport in Leaky Film for Stable Lithium Metal Anodes. Adv Energy Mater 9:1900704. https://doi.org/10.1002/aenm.201900704
- Li H, Zhang B, Liu W, Lin B, Ou Q, Wang H, Wang L (2018) Effects of an electrospun fluorinated poly(ether ether ketone) separator on the enhanced safety and electrochemical properties of lithium ion batteries. Electrochim Acta 290:150–164. https://doi.org/10.1016/j.electacta.2018.08.075
- Li Y, Li Q, Tan Z (2019) A review of electrospun nanofiber-based separators for rechargeable lithium-ion batteries. J Power Sources 443:227262. https://doi.org/10.1016/j.jpowsour.2019.227262
- Lim JY, Kang DA, Kim NU, Lee JM, Kim JH (2019) Bicontinuously crosslinked polymer electrolyte membranes with high ion conductivity and mechanical strength. J Membrane Sci 589:117250. https://doi.org/10.1016/j.memsci.2019.117250
- Lopez J, Mackanic DG, Cui Y, Bao Z (2019) Designing polymers for advanced battery chemistries. Nat Rev Mater 4:312–330. https://doi.org/10.1038/s41578-019-0103-6
- Lu Q, He YB, Yu Q, Li B, Kaneti YV, Yao Y, Yang QH (2017) Dendrite-Free, High-Rate, Long-Life Lithium Metal Batteries with a 3D Cross-Linked Network Polymer Electrolyte. Adv Mater 29:1604460. https://doi.org/10.1002/adma.201604460
- Luo D, Chen M, Xu J, Yin X, Wu J, Chen S, Wang H (2018) Polyphenylene sulfide nonwoven-based composite separator with superior heat-resistance and flame retardancy for high power lithium ion battery. Compos Sci Technol 157:119–125. https://doi.org/10.1016/j.compscitech.2018.01.023
- Pan R, Wang Z, Sun R, Lindh J, Edström K, Strømme M, Nyholm L (2017) Thickness difference induced pore structure variations in cellulosic separators for lithium-ion batteries. Cellulose 24:2903–2911. https://doi.org/10.1007/s10570-017-1312-z
- Sheidaei A, Xiao X, Huang X, Hitt J (2011) Mechanical behavior of a battery separator in electrolyte solutions. J Power Sources 196:8728–8734. https://doi.org/10.1016/j.jpowsour.2011.06.026
- Wang H, Zhang Y, Gao H, Jin X, Xie X (2016) Composite melt-blown nonwoven fabrics with large pore size as Li-ion battery separator. Int J Hydrogen Energy 41:324–330. https://doi.org/10.1016/j.ijhydene.2015.09.130
- Wang TC, Tsai CY, Liu YL (2021) Solid Polymer Electrolytes Based on Cross-Linked Polybenzoxazine Possessing Poly(ethylene oxide) Segments Enhancing Cycling Performance of Lithium Metal Batteries. ACS Sustainable Chem Eng 9:6274–6283. https://doi.org/10.1021/acssuschemeng.0c09230
- Wei N, Hu J, Zhang M, He J, Ni P (2019) Cross-linked porous polymer separator using vinyl-modified aluminum oxide nanoparticles as cross-linker for lithium-ion batteries. Electrochim Acta 307:495–502. https://doi.org/10.1016/j.electacta.2019.04.010
- Xia H, Tang Y, Malyi OI, Zhu Z, Zhang Y, Zhang W, Chen X (2021) Deep Cycling for High-Capacity Li-Ion Batteries. Adv Mater 33:e2004998. https://doi.org/10.1002/adma.202004998
- Yuan B, Wen K, Chen D, Liu Y, Dong Y, Feng C, He W (2021) Composite Separators for Robust High Rate Lithium Ion Batteries. Adv Funct Mater 31:2101420. https://doi.org/10.1002/adfm.202101420
- Zeng X, Liu Y, He R, Li T, Hu Y, Wang C, Wang H (2022) Tissue paper-based composite separator using nano-SiO2 hybrid crosslinked polymer electrolyte as coating layer for lithium ion battery with superior security and cycle stability. Cellulose 29:3985–4000. https://doi.org/10.1007/s10570-022-04499-5
- Zhang J, Kong Q, Liu Z, Pang S, Yue L, Yao J, Cui G (2013) A highly safe and inflame retarding aramid lithium ion battery separator by a papermaking process. Solid State Ionics 245–246:49–55. https://doi.org/10.1016/j.ssi.2013.05.016
- Zhang J, Zhu C, Xu J, Wu J, Yin X, Chen S, Li ZC (2020) Enhanced mechanical behavior and electrochemical performance of composite separator by constructing crosslinked polymer electrolyte networks on polyphenylene sulfide nonwoven surface. J Membrane Sci 597:117622. https://doi.org/10.1016/j.memsci.2019.117622
- Zhang X, Sun Q, Zhen C, Niu Y, Han Y, Zeng G, He W (2021) Recent progress in flame-retardant separators for safe lithium-ion batteries. Energy Stor Mater 37:628–647. https://doi.org/10.1016/j.ensm.2021.02.042
- Zhang Z, Zhou M, Yu J, Cai J, Yang Z (2020) Poly(vinylidene fluoride) Modified Commercial Paper as a Separator with Enhanced Thermal Stability and Electrolyte Affinity for Lithium-ion Battery. Energy Environ Mater 4:664–670. https://doi.org/10.1002/eem2.12153
- Zhao H, Deng N, Kang W, Li Z, Wang G, Cheng B (2020) Highly multiscale structural Poly(vinylidene fluoridehexafluoropropylene) / poly-m-phenyleneisophthalamide separator with enhanced interface compatibility and uniform lithium-ion flux distribution for dendrite-proof lithium-metal batteries. Energy Stor Mater 26:334–348. https://doi.org/10.1016/j.ensm.2019.11.005
- Zhao X, Su Y, Li Y, Zhang R, Zhao J, Jiang Z (2014) Engineering amphiphilic membrane surfaces based on PEO and PDMS segments for improved antifouling performances. J Membrane Sci 450:111–123. https://doi.org/10.1016/j.memsci.2013.08.044
- Zhong S, Yuan B, Guang Z, Chen D, Li Q, Dong L, He W (2021) Recent progress in thin separators for upgraded lithium ion batteries. Energy Stor Mater 41:805–841. https://doi.org/10.1016/j.ensm.2021.07.028
- Zhou R, Liu W, Kong J, Zhou D, Ding G, Leong YW, Lu X (2014) Chemically cross-linked ultrathin electrospun poly(vinylidene fluoride-co-hexafluoropropylene) nanofibrous mats as ionic liquid host in electrochromic devices. Polymer 55:1520–1526. https://doi.org/10.1016/j.polymer.2014.01.047
- Zhu C, Zhang J, Xu J, Yin X, Wu J, Chen S, Wang H (2019) Aramid nanofibers/polyphenylene sulfide nonwoven composite separator fabricated through a facile papermaking method for lithium ion battery. J Membrane Sci 588:117169. https://doi.org/10.1016/j.memsci.2019.117169
- Zhu Y, Cao J, Chen H, Yu Q, Li B (2019) High electrochemical stability of a 3D cross-linked network PEO@nano-SiO2 composite polymer electrolyte for lithium metal batteries. J Mater Chem A 7:6832–6839. https://doi.org/10.1039/C9TA00560A