- Email: [email protected]
Highly conductive, flexible polymer electrolyte membrane based on poly(ethylene glycol) diacrylate-co-thiosiloxane network
Solid State Ionics 322 (2018) 61–68
Contents lists available at ScienceDirect
Solid State Ionics journal homepage: www.elsevier.com/locate/ssi
Highly conductive, flexible polymer electrolyte membrane based on poly (ethylene glycol) diacrylate-co-thiosiloxane network Camilo Piedrahitaa, Victor Kusumab, Hunaid B. Nulwalac, Thein Kyua,
T
⁎
a
Department of Polymer Engineering, University of Akron, OH 44325, USA AECOM Corporation, 626 Cochrans Mill Rd, Pittsburgh, PA 15236, USA c Department of Chemistry, Carnegie Mellon University, Pittsburgh, PA 15213, USA b
A R T I C LE I N FO
A B S T R A C T
Keywords: Thiol-ene reaction Thiosiloxane Poly(ethylene glycol) diacrylate Solid polymer electrolyte membrane
An amorphous co-network consisting of poly(ethylene glycol) diacrylate (PEGDA) and thiosiloxane was prepared by means of thiol-ene “click” reaction to impart high ion conduction via enhanced segmental mobility of the conetwork. By varying PEGDA-thiosiloxane ratios, several polymer electrolyte membranes (PEM) have been fabricated to control the glass transition temperature (Tg), while keeping the equal molar ratio of succinonitrile (SCN) plasticizer and bis(trifluoromethane sulfonyl)imide (LiTFSI) salt for solid-state lithium ion batteries. It was found that the addition of thiosiloxane into the PEM network not only results in improved thermal stability and greater extensibility, but also reduces Tg leading to higher ionic conductivity that reaches the level of superionic conductor at room temperature, which further increases to 10−2 Scm−1 at high battery operating temperatures (60–70 °C). To evaluate their electrochemical performance, half-cells comprising of LiFePO4/PEM/Li foil were tested by means of cyclic voltammetry. An initial discharge capacity of 148 mAhg−1 was obtained at a current rate of 0.2 C with capacity retention of about 89%.
1. Introduction Nowadays, lithium-ion batteries have been commonly used in our daily life to power portable electronics, wearable devices, and emerging technologies including high power, high energy all solid-state battery for electric vehicles. The current Li-ion battery technology is designed based on organic liquid electrolytes such as mixtures of cyclic ethylene carbonate (EC) with volatile diethylene carbonate (DEC) or dimethyl carbonate (DMC) solvents to afford high Li-ion conduction [1]. However, because of their poor thermal stability and volatility coupled with low flash points, these liquid electrolyte solutions can easily catch fire during battery operation [2–4]. To overcome the aforementioned flammability of liquid electrolytes, solid-state polymer electrolyte membranes (PEM) have been developed in order to alleviate fire hazard of lithium ion battery [5]. The first solid electrolyte was reported by Wright in 1975 based on a mixture of poly(ethylene oxide) (PEO) and alkaline salt that showed an ionic conductivity of 10−7 Scm−1 at room temperature [6]. One major drawback of the early generation of PEO based solid polymer electrolytes (SPE) has been their low ionic conductivities at room temperature, i.e., 10−9–10−7 Scm−1, which is several orders of magnitude lower than that of liquid electrolyte counterparts (10−3–10−2 Scm−1) [7,8].
⁎
Corresponding author. E-mail address: [email protected] (T. Kyu).
https://doi.org/10.1016/j.ssi.2018.05.006 Received 17 April 2018; Received in revised form 2 May 2018; Accepted 6 May 2018 0167-2738/ © 2018 Elsevier B.V. All rights reserved.
It was later in 1979, Armand et al. [9] developed lithium-ion batteries by mixing succinonitrile (SCN) plastic crystals and various lithium salts that have improved high ionic conductivity to the range of 10−4–10−3 Scm−1, but such electrolytes had a paste-like appearance with poor mechanical integrity. The solid-state PEM needs not only to be mechanically sturdy, but also room-temperature ionic conductivity, thermal and electrochemical stability must be improved. The simultaneous improvement of these properties is hard to come by because ionic conductivity and mechanical strength have diametrically opposing effects [10]. PEMs hitherto developed were primarily based on functional polymers and lithium salts such as polyphosphazene [11], olygoethers [12] and PEO-modified polysiloxanes [13]. The manner in which these PEMs were made was by first dissolving lithium salt into a polymer precursor (i.e., prepolymer or macromonomer) and then photo-cured via UV irradiation to obtain a self-standing solid film. PEO-modified polysiloxane is an interesting approach in the development of novel polymer matrix. Neat polysiloxanes are characterized by their good mechanical strength, chemical, thermal stability, and low glass transition temperatures, but poor conductivity [14,15]. However, siloxane modification with PEO can afford high ion conduction in conjunction with good physical properties of siloxane, making these
Solid State Ionics 322 (2018) 61–68
C. Piedrahita et al.
imide (LiTFSI) (purity 99.9%), succinonitrile (purity > 99%) (SCN) and 2-butanone (purity ≥90%) were purchased from Sigma-Aldrich. Poly(vinylidene fluoride) (PVDF) having Mw of 534,000 (SigmaAldrich) was dissolved in anhydrous (99.5%) 1-methyl-2-pyrrolidinone (NMP) (Sigma-Aldrich) solvent. Electrode material, viz., lithium iron phosphate (LiFePO4) obtained from MTI Corp., was blended with acetylene black (AB) (also from MTI) to improve electron conductivity of the cathode. In the cathode fabrication, the solution mixture consisting of PVDF (10 wt%), acetylene black (10 wt%), and active electrode material (LiFePO4) (80 wt%) in NMP was used as electrode binder in the slurry form.
polymer matrices highly desirable for use in all solid-state lithium ion batteries. However, neat polysiloxane derivatives are difficult to be dissolved in PEO matrix. One common approach is to modify these siloxanes by covalently linking with polar polyether groups to improve solubility in the PEO matrix. In 1988, Khann synthesized comb-like polysiloxane having oligo(oxyethylene) side branches and subsequently incorporated them in the PEM [16]. The ionic conductivity values of these PEMs were improved to the acceptable conductivity range of the order of 10−4 Scm−1, albeit still low. Wang et al. [17] synthesized siloxane polymer and then mixed it with lithium bis(trifluoromethyl sulfonyl)imide (LiTFSi) salt. However, the highest ionic conductivity value thus obtained was on the order of 10−4 Scm−1 at room temperature. Similarly, Zhang et al. [18] prepared a solid PEM via hydrosilylation of polymethylhydrosiloxane (PMHS) and subsequently mixed with LiTFSI salt. They achieved a similar order of magnitude in ionic conductivity (i.e., 10−4 Scm−1) with their crosslinked siloxane-based solid PEM network having a low Tg of approximately −50 °C. Burjanadze et al. [19] investigated the electrochemical stability of crosslinked salt-in-polysiloxane membranes containing two lithium salts, i.e., boron based anions, viz., LiDFOB (lithium difluoro(oxalato)borate) and LiBOB (lithium bis(oxalato)borate). The ionic conductivity values were relatively low that only reached the order of 10−5 Scm−1 at room temperature, but with an enhanced electrochemical stability up to 4.7 V against Li/Li+. One interesting approach in fabricating the siloxane-PEO compounds without performing tedious chemical synthesis is to conduct crosslinking reactions between polyethylene glycol derivatives such as poly(ethylene glycol) diacrylate (PEGDA) and thiol (SH) containing compound such as thiosiloxane via the thiol-ene ‘click’ reaction between thiol (SH) of siloxane and the C]C double bonds of functional acrylate groups of the polyether backbone [20]. This click reaction is reported to be fast, easy to execute, and more importantly provides high yield. In fact, such chemical route has been already demonstrated by the co-authors while they were at DOE [21,22], who reported the achievement of completely amorphous crosslinked co-networks consisting of PEGDA and thiosiloxane. These thiosiloxane/PEGDA co-networks have high selectivity and separation of carbon dioxide (CO2) from nitrogen (N2) or methane (CH4) with potential applications in control of greenhouse gas emission [23]. The presence of both polar and non-polar moieties in the thiosiloxane/PEGDA co-network has been attributed to be the key factor in the enhancement of gas permeability and selectivity. In this article, the aforementioned thiol-ene ‘click’ crosslinking reaction was employed to afford an amorphous thiosiloxane/PEGDA conetwork for use as a polymer matrix in the polymer electrolyte membrane for solid-state lithium ion battery. Firstly, the miscibility of a binary thiosiloxane/PEGDA mixture was determined in order to provide guidance to the formation of isotropic amorphous polymeric conetwork. Secondly, various thiosiloxane/PEGDA ratios were further mixed with lithium salt (LiTFSI) and succinonitrile (SCN) plasticizer to produce a completely amorphous, solid-state polymer electrolyte membrane (PEM). Mechanical and ion conduction properties of the above thiosiloxane/PEGDA membranes were evaluated by means of tensile tests and AC impedance spectroscopy, respectively. Cyclic voltammetry measurements were also carried out in order to evaluate electrochemical performance and discuss potential utilization of siloxane-PEGDA PEMs as solid electrolytes for all solid-state lithium ion batteries.
2.1. Binary thiosiloxane-PEGDA system Various thiosiloxane/PEGDA binary mixtures were prepared in accordance with the predetermined weight ratios thiosiloxane:PEGDA 10:90, 15:85 and 25:75. These binary components were fully dissolved upon mixing in the liquid state. By adding the indicated amount of DMPA photoinitiator, the thiol (SH) group of thiosiloxane was activated and subsequently reacted with acrylate groups of PEGDA via the socalled thiol-ene ‘click’ reaction. Note that the photo-initiator (DMPA) amount was kept at 3 wt% with respect to the total prepolymer amount. This thiol-ene reaction produced a crosslinked network between nonpolar siloxane and polar PEGDA. Upon photo-crosslinking, the binary blend membrane remained homogeneous and transparent, suggestive of the formation of an isotropic co-network. 2.2. Pseudo-ternary system – thiosiloxane-PEGDA/SCN/LiTFSI Binary mixtures of thiosiloxane/PEGDA and SCN/LiTFSI were prepared separately inside a glovebox under Argon atmosphere. Thiosiloxane/PEGDA blends in the indicated weight ratios were found to be completely miscible. However, the SCN/LiTFSI mixture had to be heated above the crystal melting temperature of SCN (i.e., 60 °C) for several minutes in order to completely dissolve and form a clear melt mixture. Subsequently, these individual binary mixtures were further mixed with thiosiloxane/PEGDA in their melt state. The initial quaternary (thiosiloxane:PEGDA/SCN/LiTFSI) liquid mixture exhibited a cloudy appearance, but it turned transparent upon adding 2-propanone as a common solvent. After exposure to UV irradiation for 5 min, copolymerization occurred between thiosiloxane and PEGDA via the ‘click’ reaction and formed a thiosiloxane-PEGDA copolymer. Upon removal of solvent, a clear, solid polymer electrolyte membrane (PEM) consisting of thiosiloxane-PEGDA copolymer, SCN and LiTFSI (viz., pseudo-ternary constituents) was obtained. 2.3. Polymer electrolyte membrane The polymer electrolyte membranes (PEM) compositions are given in weight ratios, e.g., 20/40/40 (thiosiloxane:PEGDA)/SCN/LiTFSI. In the PEM formulation, the copolymer ratio of thiosiloxane:PEGDA was varied as 2:18(10/90)/40/40, 3:17(15/75)/40/40 and 5:15(25/75)/ 40/40, while fixing their total amount at 20 wt%. Also the weight percent for SCN/LiTFSI was kept as constant (at 40 wt% each) such that the only variable is the copolymer ratio. The amount of photo-initiator (DMPA) was kept at 3 wt% with respect to the total prepolymer amount. It should be pointed out that the above PEM mixture belongs to the isotropic region of their ternary phase diagram (figure not shown). This transparent mixture was spread within a spacer (10 mm × 10 mm frame and 0.3 mm thick). Subsequently, it was photocrosslinked under uniform UV irradiation at 350 nm with an intensity 5 mW cm−2 for 2 min. After curing, solid films thus obtained were subsequently dried for 60 min at 80 °C and then kept inside a glovebox overnight under argon atmosphere. These fabricated PEM film samples were characterized by using various characterization techniques such as Fourier transformed infrared (FTIR), Raman, wide angle X-ray
2. Materials and methods (Mercaptopropyl) methyl siloxane prepolymer (thiosiloxane), Mw = 5000 was purchased from Gelest. Poly(ethylene glycol) diacrylate (PEGDA), Mw = 700 g/mol was bought from Sigma-Aldrich. Photo-initiator Irgacure®651 (2,2-dimethoxy-2-phenylacetophenone, DMPA) was obtained from CIBA. Lithium bis(trifluoromethyl sulfonyl) 62
Solid State Ionics 322 (2018) 61–68
C. Piedrahita et al.
Fig. 1. FTIR spectra for (a) binary thiosiloxane-PEGDA system (15:85 ratio by weight) (left), showing the reduction of stretching vibration (va) mode of the C]C bonds before (grey line) and after (black line) UV curing. Enlarged area shows the peak of 1627 cm−1 corresponding to the stretching of C]C bond and (b) (thiosiloxane:PEGDA)/SCN/ LiTFSI (left) displaying reduction of the peak at 1627 cm−1 for the equivalent vibrational mode before (grey line) and after (black line) UV exposure.
polished stain-steel electrodes having an area of 10 mm × 10 mm with a thickness of 1 mm. To evaluate the electrochemical stability of the PEM, cyclic voltammetry (CV) tests were performed at room temperature (25 °C) by using a SI 1260 Impedance/Gain Phase Analyzer and SI 1287 Electrochemical Interface (Solartron Analytical Inc.). A coin cell was prepared in the stain steel (SS)/PEM/Li assembly in which SS represented the working electrode, whereas the Li foil (Alfa Aesar) acted as both counter- and reference-electrodes. A linear sweep voltammetry was performed between 1–6 V at a sweep rate of 1 mV/s. The cyclic voltammetry measurement was carried out at a scan rate of the 0.2 mV/ s in the voltage range of −0.5 to 4.5 V. The thickness of PEM films ranged from 150 μm to 250 μm. For the cathode preparation, a slurry was made in NMP at a weight ratio of 80/10/10 corresponding to active electrode material/carbon black (CB)/poly(vinylidene fluoride) (PVDF) and further spread on Al foil (current collector). The coated electrodes were first dried on a hotplate at 80 °C for 2 h and then vacuumed dried in oven at 110 °C overnight. LiFePO4 loading was in the range of 0.5–1.3 mg. Further assembling of the coin cell was performed inside a glovebox under Argon gas environment. Cyclic voltammetry test was performed in halfcell configuration by sandwiching the PEM between LiFePO4-cathode and Li-foil (i.e., LiFePO4/PEM/Li foil) at a scan rate of the 0.2 mV/s within the voltage range of 2.5 to 4.2 V. Galvanostatic charge/discharge cycling test was undertaken on the half-cell under the same configuration within the established stable voltage of the electrode utilizing a MTI 8-channel battery cycler (MTI Corp.) at a constant current rate of 0.2 C.
diffraction (WAXD) along with AC impedance spectroscopy and cyclic voltammetry.
2.4. Membrane characterization Differential scanning calorimetry (DSC) (TA Instruments, Model Q200) was carried out to determine glass transition temperature of various binary blends as well as their PEMs containing specified amounts of SCN plasticizer and LiTFSI salt. Recommended amount of 5 mg the blends were placed in aluminum pans and hermetically sealed using aluminum lids in a glovebox under nitrogen to avoid moisture absorption. All DSC scans were acquired at a ramp rate of 10 °C/min, unless indicated otherwise. Fourier Transformation Infrared (FTIR) spectroscopy was performed in attenuated total reflection (ATR) mode by utilizing a Thermo Fisher Scientific instrument (Model Nicolet iS50 Spectrometer). Prior to FTIR investigation, dried samples were stored overnight inside a glovebox under argon environment and sealed in vials to prevent any exposure to air and/or possible contamination before analysis. By placing the PEM on the ZnSe crystal, the IR spectra were acquired in the spectral range of 450 to 4000 cm−1 with an average of 32 scans and a spectral resolution of 4 cm−1. RAMAN spectroscopy was carried out using a Micro Raman Spectrometer (LabRam HR, Horiba equipped with an Nd:YAG laser at 532 nm and 60 mW) for an exposure time of 5 min. After UV curing, (thiosiloxane:PEGDA)/SCN/LiTFSI PEM film exhibits transparent appearance suggestive of a homogeneous mixture although by no means a proof. To verify the absence of any crystals in the PEM, WAXD 2θ-scans were acquired on a D/MAX RAPID II microdiffractometer. The x-ray (Cu-Kα radiation at the wavelength of 1.5417 Å) was collected over a scattering angle range of 0 < 2θ < 45. Thermogravimetric Analysis (TGA) experiment was carried out on dried, cured PEM films (weighing recommended amount of approximately 10 mg) by using a TA Instrument Model Q50 within a temperature range of 25 °C to 300 °C at a ramped rate of 10 °C/min. Differential scanning calorimetry (DSC) was performed on a TA Instrument model Q2000. To determine the glass transition of the PEM, dried samples weighing about 100 mg were sealed in an aluminum pan and subsequently DSC scans were performed within a temperature range from −130 and 30 °C at a ramp rate of 5 °C/min. Tensile tests of the PEM films were performed by mean of dynamic mechanical analyzer (DMA, TA Q800, TA Instruments, Inc.) Film samples having the dimensions of 10.0 mm × 5.0 mm × 0.6 mm (length × width × thickness) were stretched at 25 °C in air until break using a force rate of 0.3 N/min. Ionic conductivity measurements were carried out by means of an AC impedance spectrometer (HP4192A, Hewlett-Packard) interfaced to a standard PC computer by scanning from 13 MHz to 5 Hz at an applied voltage of 1 V in the temperature range of 25–130 °C. In the AC impedance measurements, the dried PEM films were placed between two
3. Results and discussion 3.1. FTIR characterization of thiol-ene reaction in thiosiloxane/PEGDA conetwork Various thiosiloxane-PEGDA mixtures were prepared by mixing in the liquid state at ambient temperature. Fig. 1a shows the IR spectra of a thiosiloxane/PEGDA mixture (15:85 ratio by weight) acquired before and after UV exposure for a few mins. The rate of photopolymerization of the curing reaction is so fast that it requires only 1 min of UV irradiation. Of particular importance is that all signals corresponding to the C]C bonds of acrylate groups, located at 1627 cm−1 (stretching) which virtually disappeared due to the thiol-ene reaction between the acrylate double bonds and thiol group. The estimated conversion of the reaction based on the areas under the above IR bands was approximately 93%. A similar observation of reaction conversion was made for other compositions, viz., 5:95 and 10:90 ratios by weight of the thiosiloxane/PEGDA mixtures (data not shown). When additional components of the PEM such as SCN plasticizer and LiTFSI salt are incorporated into the binary thiosiloxane-PEGDA 63
Solid State Ionics 322 (2018) 61–68
C. Piedrahita et al.
DMPA
+
UV
Fig. 2. Proposed reaction scheme of thiosiloxane and PEGDA forming a co-network.
2572 cm−1 in the (3:17)/40/40 and (5:15)/40/40 compositions implies that the SH bonds are either fully consumed or the signal falls below its detection limit. According to the IR experiment, the conversion of thioene reaction is about 90%, and thus the lack of Raman peak at 2572 cm−1 in the 15:85 (3:17) and 25:75 (5:15) may simply suggest that sizable portions of the SeH groups were reacted with the acrylate double bonds and only an undetectable level of thiosiloxane may be left unreacted. Moreover, when the concentration of the 25:75 (thiosiloxane:PEGDA) exceeds to the exact molar ratio 1:4 (20:80 thiol:acrylate), there is a possibility of excess thiosiloxane molecules not involved in the reaction and thus the complete disappearance of SeH signal simply implies the limitation of the Raman approach [21,23]. It is reasonable to infer that the enhanced molecular mobility afforded by the incorporated plasticizer raises the greater probability of SeH groups to encounter with the acrylate double bonds, thereby increasing the conversion of the thiol-ene reaction as manifested by the virtual depletion of the SeH Raman peak (Fig. 3c).
blend, the characteristic peaks representing eCH2]CH2 bonds at 956 cm−1 and 798 cm−1 showed a slight reduction in their respective areas. However, the signal located at 1631 cm−1 corresponding to the C]C stretching decreases drastically (i.e., the conversion of approximately 90%) after the UV exposure, implying high consumption of these bonds either via “ene-ene” or “thiol-ene” reaction as proposed in Fig. 2. Upon addition of SCN plasticizer into the PEM system, the segmental motion of the network chains is expedited, thereby raising the probability of collision among these active groups to react with each other. It may be emphasized that once the thiol-ene reaction occurs; the newly formed crosslinked junctions would restrict the segmental motions of the network chains. The reduced segmental mobility in turn might prevent the complete conversion of the thiol-ene reaction, as manifested by the residual minor broad peak at 1631 cm−1 after curing (please see the zoom-in inset of Fig. 1b). 3.2. Characterization of thiol-ene reaction in thiosiloxane-PEGDA by RAMAN spectrometry
3.3. WAXD investigation of PEM networks
RAMAN spectroscopic investigation was performed to determine the consumption of thiol (SH) after the “thiol-ene” reaction. The signal from SeH bond located at 2572 cm−1 was detected for the neat thiosiloxane (Fig. 3a) as well as for pseudo-ternary PEM mixtures containing higher thiol concentrations of 3:17 (i.e., 15:85 15% thiosiloxane: 85% PEGDA) and (5:15) (25:75 25% thiosiloxane: 75% PEGDA) (Fig. 3b). In the case of 2:18 (10:90 10% thiosiloxane: 90% PEGDA) system, there is no discernible peak, suggesting that the signal may be below the minimum detectable level of SeH bond by this RAMAN technique. However, upon UV-curing, the observed bands of the 15:85 and 25:75 (thiosiloxane:PEGDA) (i.e., higher contents of SeH) seemingly disappeared, indicating that a high percentage of SeH groups was consumed in the photoreaction (Fig. 3c). It should be cautioned that the absence of the RAMAN peak at
It should be pointed out that the starting PEGDA precursor and LiTFSI are crystalline, whereas SCN exhibits the plastic crystalline phase at room temperature. Although the photo-cured PEM shows a single Tg, it is essential to perform wide angle x-ray diffraction (WAXD) investigation to determine the internal structure of the PEM by varying the thiosiloxane:PEGDA ratios, while fixing the concentration of SCN plasticizer and LiTFSI salt at 40% each. Fig. 4 exhibits the Bragg's scans of the PEMs (thiosiloxane:PEGDA)/ SCN/LiTFSI at three different compositions. The scattering profiles revealed two broad scattering peaks at 2θ of 11.9 and 19.6, corresponding to the radial distribution functions arising from the amorphous thiosiloxane - PEGDA co-network. The lack of crystalline peaks of the polymer matrix (PEGDA) is consistent with those reported by Echeverri et al. [25] for various ternary PEGDA/SCN/LiTFSI mixtures
Fig. 3. (a) RAMAN spectra for neat thiosiloxane, (b) thiosiloxane-PEGDA mixtures before crosslinking and (c) crosslinked thiosiloxane-PEGDA co-network based PEMs as a function of co-network ratios, showing the complete disappearance of the SeH peak suggestive of the occurrence of thiol-ene reaction during the conetwork formation. 64
Solid State Ionics 322 (2018) 61–68
C. Piedrahita et al.
thiosiloxane compositions. In the thermograms, there is no appreciable weight loss (< 5%) at 110 °C, indicating that there is little or no residual solvent or moisture in the PEMs. That is to say, the initial decomposition temperature (IDT) corresponding to an initial weight loss of < 5% was noticed only in the vicinity of 110–115 °C, implying that the PEMs may be thermally stable at such high temperature environments without losing their ingredients of the PEM networks. Fig. 6a shows the DSC thermograms of the photo-crosslinked thiosiloxane-PEGDA as a function of composition in weight ratios. Neat thiosiloxane (denoted by 100:0), which is known to be hydrophobic, exhibits a very low glass transition temperature (Tg) of −95 °C in the DSC thermogram (i.e., the mid-point of the inflexion of the DSC curve). On the other hand, the neat PEGDA (denoted by 0:100), which is hydrophilic, reveals a distinct Tg at −39 °C (Fig. 6a). With increasing PEGDA, the single Tg moves gradually with increasing thiosiloxane to the indicated temperatures intermediate between those of the neat constituents. At 25 wt% loading of thiosiloxane to PEGDA co-network, the Tg has declined to −50 °C (Fig. 6a). The revelation of a single glass transition in combination with the trend of the gradual movement of the single Tg with composition implies the formation of amorphous conetworks, albeit showing slight negative departure from the estimated linear relationship. This finding is consistent with the amorphous character of the WAXD scans reported earlier in Fig. 4. Fig. 6b displays thiosiloxane-PEGDA co-network as PEM components. These exhibits the Tg values around −83 °C ( ± 1 °C). Inclusion of the sizable amount of a solid plasticizer/ionizer such as SCN further promotes the segmental motion of the thiosiloxane-PEGDA co-network. These Tg values were found to be consistent with at those reported by He, et al. [26] for the ternary system PEGDA/SCN/LiTFSI 20/40/40 (i.e., Tg of approximately −82 °C), which also exhibited a high ionic conductivity on the order of 8.4 × 10−4S/cm at room temperature.
Fig. 4. WAXD 2θ - scans for thiosiloxane-PEGDA based PEM's at various compositions fixing the SCN plasticizer and LiTFSI salt concentration at 40 wt%. The PEM is transparent and shows no discernible crystalline peaks, suggestive of amorphous character.
corresponding to the isotropic region of their ternary phase diagrams. The resulting film remains transparent without any indication of phase separation during photo-crosslinking, suggestive of a completely crosslinked co-network (Fig. 4, please see the inset picture). The absence of crystalline WAXD peaks in conjunction with the observed optical clarity and a single Tg of the PEM imply the completely amorphous character of the present PEM. This finding is consistent with the single Tg with composition trend obtained from the DSC investigation to be discussed in a latter session. These observations were also in agreement with that reported by Kusuma et al. [23] for the binary thiosiloxane-PEGDA network.
3.5. Mechanical properties of PEMs 3.4. Thermal characterization of PEMs
Fig. 7 depicts the stress-strain behaviors for the PEMs as a function of PEM matrix compositions while keeping the constant ratio of SCN plasticizer and LiTFSI salt at 40 wt% each. The neat PEGDA based PEM, denoted by (0:20)/40/40 (thiosiloxane: PEGDA)/SCN/LiTFSI, exhibits a brittle character, having a high tensile modulus, but the lowest elongation-at-break of around 5%. With increasing thiosiloxane to 2:18 ratio, the PEM (2:18)/40/40 (thiosiloxane:PEGDA)/SCN/LiTFSI PEM shows an improved elongation to 17%. The (3:17)/40/40 composition shows the highest value of approximately 20% elongation to that of (2:18)/40/40 (thiosiloxane:PEGDA)/SCN/LiTFSI. With further increase of the thiosiloxane ratio to 5:15 in the PEMs both tensile strength and modulus have declined noticeably, which may be attributed to the excess thiosiloxane, which may be left unreacted above the molar ratio of 1:4 (4:16). It is reasonable to infer that the inclusion of thiosiloxane increases the segmental motion of the matrix as manifested by a lower glass transition temperature (Tg) of the final co-network, thereby increasing the network extensibility. This improved chain flexibility is anticipated to improve the PEM performance eventually, since it can alleviate the well-known problem of electrode cracking driven by expansion and contraction of the PEM/ binder during battery operation. In our recent study, the attachment of dangling side chains to the PEM network provides greater mobility of the network chains resulting in not only lowering glass transition temperatures, but also increasing ionic conductivity considerably. More importantly, the enhanced chain mobility at the PEM/electrode interface further improves the capacity retention. Hence, it is important to investigate the electrochemical performance of the present thiosiloxane-PEGDA PEM.
Li-ion batteries are generally known to generate heat during battery operation; e.g., the temperature of some electronic devices can reach 50–70 °C. Hence, thermal management is of paramount importance to provide thermally stable environment to the PEM. Silicon-based compounds, such as thiosiloxane are known for their good thermal stability and hydrophobicity. Once the prepolymer thiosiloxane/PEGDA mixture was co-polymerized via thiol-ene reaction, the resulting co-network becomes thermally stable up to 300 °C [23]. However, in the PEM formulation, plasticizers such as SCN and/or EC and LiTFSI salt were compounded into the thiosiloxane/PEGDA mixtures and thus thermal stability of such multicomponent PEM should be examined. Fig. 5 exhibits the TGA thermograms for PEM's at various
3.6. Ionic conductivity Fig. 5. TGA thermograms for PEM containing various thiosiloxane contents.
Fig. 8 exhibits the plot of the ionic conductivity versus the 65
Solid State Ionics 322 (2018) 61–68
C. Piedrahita et al.
Fig. 6. DSC thermograms of UV-cured (a) thiosiloxane-PEGDA and (b) (thiosiloxane-PEGDA)/SCN/LiTFSI PEM as a function of compositions.
expression; σ = l/(AZ) where l represents the thickness, A the area of the sample and Z total impedance. The PEM containing neat PEGDA denoted by (0:20)/40/40 shows high room temperature ionic conductivities of 9.3 × 10−4 S/cm at room temperature, whereas the neat thiosiloxane (20:0) mixture exhibits a liquid-like appearance and a slightly lower value of 8.4 × 10−4 S/cm. Of particular interest is that all intermediate copolymer compositions exhibited higher values relative to the neat PEGDA or thiosiloxane networks. That is to say, the ionic conductivities for the PEM having (thiosiloxane:PEGDA) 2:18 and 3:17 ratios improves due to higher flexibility (or elongation-at-break) to the values of 2.17 × 10−3 S/cm and 2.48 × 10−3 S/cm at 25 °C, respectively. With further increase of thiosiloxane:PEGDA ratio to 5:15, the ionic conductivity has declined to 1.07 × 10−3 S/cm due to the excess thiosiloxane which were presumably left unreacted above the molar ratio of 1:4. This finding is consistent with that of the stress-strain behavior reported earlier (see Fig. 7). Of particular interest is that some thiosiloxane-modified PEMs such as 2:18 and 3:17 can reach the superionic conductivity level on the order 10−3 S/cm at room temperature. It may be hypothesized that lithium ion transport is governed by two opposing effects; the ion-dipole complexation (or coordination bonds between lithium cations and ether oxygen) which retards the segmental motion and plasticization of the polyethylene glycol network (as manifested by lower Tg) as well as dissociation of lithium-ether oxygen complexes by SCN [27–29]. Upon heating above 70 °C some compositions can reach the order of 10−2 S/ cm, which is comparable in magnitude to the organic liquid electrolyte counterpart, although most of liquid electrolyte batteries would not survive at such high operating temperatures above 60 °C. It can be noticed that the temperature dependence of the ionic conductivity for thiosiloxane-modified PEMs exhibit departures from the Arrhenius behavior (i.e., non-linear slope). Ion conduction is governed by the mobility of cations coupled to the segmental motion of the polymer chains and thus it may be better represented by VogelTammann-Fulcher (VTF) equation [10,30]. All PEMs exhibited curvatures, which may be fitted with the respective VTF parameters (Table 1), i.e., the reference temperature (To) was taken as To = Tg − 50 °C.
Fig. 7. Strain-stress curves for thiosiloxane-PEGDA PEM's. Increasing thiosiloxane content increases the elongation-at-break of the PEM up to 3:17 ratio of the copolymer, then it falls off the molar ratio of 1:4, i.e., 5:15 ratio.
Fig. 8. Variation of ionic conductivity versus reciprocal temperature for thiosiloxane PEM's at various thiosiloxane concentrations showing curvatures suggestive of non-Arrhenius character. Note that (0:20)/40/40 and (20:0)/40/ 40 represent pure PEGDA and pure thiosiloxane PEMs. The solid curves were fitted in accordance with VTF equation.
Table 1 VTF parameters; room temperature ionic Conductivity (σRT) and activation energy (B) for various (thiosiloxane:PEGDA)/SCN/LiTFSI compositions.
reciprocal absolute temperature for various PEM compositions. Ionic conductivity values were calculated according to the following 66
(Thiosiloxane:PEGDA)/SCN/LiTFSI (wt%)
σRT (S/cm)
B (kJ/mol)
R2
(2:18)/40/40 (3:17)/40/40 (5:15)/40/40
1.81 × 10−3 2.10 × 10−3 1.07 × 10−3
5.9 6.6 6.7
0.998 0.999 0.999
Solid State Ionics 322 (2018) 61–68
C. Piedrahita et al.
Fig. 9. a) Linear sweep voltammetry performed between 1 and 6 V for SS/thiosiloxane-PEG PEM/Li foil. b) Voltammogram for thiosiloxane-PEG PEM between −0.5 to 4.2 V. Zoom area from 0.3–1.4 V evidencing appearance of peaks that come from side redox reaction within the PEM.
3.7. Electrochemical performance of PEMs
3.8. Charge-discharge test for PEM
Fig. 9a shows the linear sweep voltammetry conducted in the range of 1–6 V with the aim of detecting the maximum voltage at which the PEM can be stable and also to detect the voltage range whereby irreversible side oxidative reactions may be taking place. These reactions can affect its capability to undergo through reversible reactions and will affect the performance of the PEM. Samples were assembled in a Limetal/PEM/stain-steel (SS) configuration. (thiosiloxane:PEGDA)/SCN/ LiTFSI (3:17)/40/40 PEM appeared stable up to 4.8 V versus Li/Li+ as indicated by an arrow in Fig. 9a. Similar electrochemical stability limit was found for other compositions, and thus the upper limit voltage was set at 4.2 V, which is below the onset voltage versus current to prevent any irreversible reaction. The lower limit was set at 0.5 V, i.e., above the plating and stripping (reduction and oxidation) potential of the Li-ions. Voltammograms reveal the presence of two redox reaction peaks at −0.3 V and 0.5 V peaks, corresponding to the reduction (cathodic) and oxidation (anodic) associated with Li stripping/plating (Fig. 9b). A small reduction peak can be discerned at approximately 0.55 V which is accompanied by an oxidation signal around 1.02 V. These signals are believed to arise from the redox reaction of the PEM containing LiTFSI salt [31]. The intensity of the peaks located at −0.5 V and 0.25 V (related to the concentration of Li+ ions) remains virtually invariant suggestive of reversible redox reactions. The same experimental conditions were adopted for other compositions (2:18 and 5:15 thiosiloxane:PEG) that showed a similar trend (data not shown).
Fig. 10a exhibits the cyclic voltammetry of the LiFePO4/PEM/Li half-cells performed at ambient temperature (25 °C) at a current rate of 0.2 mV/s. Generally, LiFePO4 has been used as a cathode material due to its high voltage operation range (i.e., 2.5–4.2 V). It can be seen in the voltammograms that the cathodic peak for LiFePO4 is located at 3.7 V vs Li/Li+ (corresponding to the Li-ions lithiation (i.e., reduction)), while the andic peak was observed at 3.3 V vs Li/Li+ which represents delithiation (i.e., oxidation) of Li-ion. It can be noticed that the peaks from the second and third cycles remain invariant with increasing number of cycles, indicating excellent reversibility of the redox reactions at the PEM/electrode interface. Fig. 10b exhibits the galvanostatic charge/discharge cycling test of the same half-cell at ambient temperature (25 °C). The current rate was set at 0.2 C (i.e., it takes 5 h to complete 1 cycle). Specific charge/discharge capacities are plotted with respect to the number of cycles. It can be seen that the half-cell battery (thiosiloxane:PEGDA)/SCN/LiTFSI (3:17)/40/40) exhibits a reasonable rechargeable characteristic. That is to say, the observed initial capacity value was 148 mAh g−1, but it gradually dropped to around 132 mAh g−1, i.e., 89% retention. Moreover, the Coulombic efficiency is almost 100% for at least 50 cycles tested indicating reasonable charge/discharge cyclic stable behavior. According to a previous paper [31], the post-mortem analysis revealed no noticeable changes in the interface structure during cycling at room temperature, but some structural changes occurred in the LiFePO4/ PEGDA-based PEM/Li half-cell after 50 cycling at an elevated temperature of 60 °C. With the lithium bis(oxalato)borate LiBOB modification, such structural change has been alleviated [31].
Fig. 10. (a) Voltammogram for half-cell LiFePO4/PEM/Li the voltages range operation 2.5 V to 4.2 V. The ramp rate was 0.5 mV/s. (b) Capacity retention behavior of PEM (thiosiloxane:PEGDA)/SCN/LiTFSI (3:17)/40/40 in a half-cell charge/discharge test versus number of cycles at 0.2C for 50 cycles. 67
Solid State Ionics 322 (2018) 61–68
C. Piedrahita et al.
4. Conclusions
[9] M.B. Armand, J.R. MacCallum, C.A. Vincent (Eds.), Polymer Electrolyte Reviews, 1 Elsevier Appl. Sci, London, 1987, pp. 1–22. [10] R. He, M. Echeverri, D. Ward, Y. Zhu, T. Kyu, Highly conductive solvent-free polymer electrolyte membrane for lithium-ion batteries: effect of prepolymer molecular weight, J. Membr. Sci. 498 (2016) 208–217. [11] G. Nazri, D. MacArthur, J. Ogara, Polyphosphazene electrolytes for lithium batteries, Chem. Mater. 1 (3) (1989) 370–374. [12] M. Watanabe, Effects of network structures and incorporated salt species on electrochemical properties of polyether-based polymer electrolytes, Solid State Ionics 79 (1995) 306–312. [13] J. Smid, D. Fish, I. Khan, E. Wu, G. Zhou, Polysiloxane-based polymer—electrolyte complexes, Adv. Chem. (1989) 113–123. [14] N. Grassie, K. Francey, I. Macfarlane, The thermal degradation of polysiloxanes—part 4: poly(dimethyl/diphenyl siloxane), Polym. Degrad. Stab. 2 (1) (1980) 67–83. [15] Y. Abe, T. Gunji, Oligo- and polysiloxanes, Prog. Polym. Sci. 29 (3) (2004) 149–182. [16] I. Khan, Y. Yuan, D. Fish, E. Wu, J. Smid, Comblike polysiloxanes with oligo(oxyethylene) side chains. Synthesis and properties, Macromolecules 21 (9) (1988) 2684–2689. [17] F. Wang, C. Hu, S. Lo, Y. Wang, C. Wan, The investigation of electrochemical properties and ionic motion of functionalized copolymer electrolytes based on polysiloxane, Solid State Ionics 180 (4–5) (2009) 405–411. [18] Z. Zhang, D. Sherlock, R. West, R. West, K. Amine, L. Lyons, Cross-linked network polymer electrolytes based on a polysiloxane backbone with oligo(oxyethylene) side chains: synthesis and conductivity, Macromolecules 36 (24) (2003) 9176–9180. [19] M. Burjanadze, Y. Karatas, N. Kaskhedikar, L. Kogel, S. Kloss, A. Gentschev, M. Hiller, R. Müller, R. Stolina, P. Vettikuzha, H. Wiemhöfer, Salt-in-polymer electrolytes for lithium ion batteries based on organo-functionalized polyphosphazenes and polysiloxanes, Z. Phys. Chem. 224 (10−12) (2010) 1439–1473. [20] C. Hoyle, A. Lowe, C. Bowman, Thiol-click chemistry: a multifaceted toolbox for small molecule and polymer synthesis, Chem. Soc. Rev. 39 (4) (2010) 1355. [21] L. Kwisnek, S. Heinz, J. Wiggins, S. Nazarenko, Multifunctional thiols as additives in UV-cured PEG-diacrylate membranes for CO2 separation, J. Membr. Sci. 369 (1–2) (2011) 429–436. [22] M. Mellott, K. Searcy, M. Pishko, Release of protein from highly cross-linked hydrogels of poly(ethylene glycol) diacrylate fabricated by UV polymerization, Biomaterials 22 (9) (2001) 929–941. [23] V. Kusuma, E. Roth, W. Clafshenkel, S. Klara, X. Zhou, S. Venna, E. Albenze, D. Luebke, M. Mauter, R. Koepsel, A. Russell, D. Hopkinson, H. Nulwala, Crosslinked poly(ethylene oxide) containing siloxanes fabricated through thiol-ene photochemistry, J. Polym. Sci. A: Pol. Chem 53 (13) (2015) 1548–1557. [25] M. Echeverri, N. Kim, T. Kyu, Ionic conductivity in relation to ternary phase diagram of poly(ethylene oxide), succinonitrile, and lithium Bis(trifluoromethane) sulfonyl imide blends, Macromolecules 45 (15) (2012) 6068–6077. [26] R. He, T. Kyu, Effect of plasticization on ionic conductivity enhancement in relation to glass transition temperature of crosslinked polymer electrolyte membranes, Macromolecules 49 (15) (2016) 5637–5648. [27] E. Rietman, M. Kaplan, R. Cava, Lithium ion-poly (ethylene oxide) complexes. I. Effect of anion on conductivity, Solid State Ionics 17 (1) (1985) 67–73. [28] R. Frech, W. Huang, Conformational changes in diethylene glycol dimethyl ether and poly(ethylene oxide) induced by lithium ion complexation, Macromolecules 28 (4) (1995) 1246–1251. [29] V. Aravindan, J. Gnanaraj, S. Madhavi, H. Liu, Lithium-ion conducting electrolyte salts for lithium batteries, Chem. Eur. J. 17 (51) (2011) 14326–14346. [30] K.M. Diederichsen, H.G. Buss, B.D. McCloskey, The compensation effect in the Vogel–Tammann–Fulcher (VTF) equation for polymer-based electrolytes, Macromolecules 50 (2017) 3831–3840. [31] R. He, F. Peng, W.E. Dunn, T. Kyu, Chemical and electrochemical stability enhancement of lithium bis(oxalato)borate (LiBOB)-modified solid polymer electrolyte membrane in lithium ion half-cells, Electrochim. Acta 246 (2017) 123–134.
In this work, we have incorporated thiosiloxane as a matrix cocomponent of the PEM with the aim of enhancing the polymer segmental motion and hence, the ionic conductivity. Thiosiloxane and PEGDA mixtures were crosslinked by UV irradiation via the “thio-ene” reaction between the thiol functionality and the double bonds of the PEGDA precursor, forming a complete homogeneous self-standing PEM. The thiosiloxane modified PEM film exhibits higher extension-at-break in comparison to its counterpart, i.e., the PEM containing neat PEGDA700/SCN/LiTFSI 20/40/40. As evidenced in FTIR and Raman spectroscopy, sizable portions of the thiol (SH) groups were consumed in the thiol-ene reaction suggestive of high conversion in the co-network. With increasing thiosiloxane, the Tg of the PEM was declined, leading to higher ionic conductivity. Thiosiloxane:PEGDA based PEMs are thermally stable exhibiting good anodic stability up to 4.8 V versus Li/Li+. When tested in half-cells using LiFePO4 as cathode against Li/ Li+ electrode, the specific discharge capacity showed the initial capacity value of 148 mAh g−1 with reasonably good capacity retention (i.e., 89%). The achievement of improved thermal, ion conduction, and stable electrochemical performance implies that the present solid-state PEM based on thiosiloxane modified PEGDA co-network has potential for implementing all solid-state Li-ion batteries. Acknowledgment Support of this work by the U.S. National Science Foundation, DMRPolymers, Grant # 1502543 is gratefully acknowledged. References [1] J. Goodenough, K. Park, The li-ion rechargeable battery: a perspective, J. Am. Chem. Soc. 135 (2013) 1167–1176. [2] D. Abraham, E. Roth, R. Kostecki, K. McCarthy, S. MacLaren, D. Doughty, Diagnostic examination of thermally abused high-power lithium-ion cells, J. Power Sources 161 (2006) 648–657. [3] G. Eshetu, S. Grugeon, S. Laruelle, S. Boyanov, A. Lecocq, J. Bertrand, G. Marlair, In-depth safety-focused analysis of solvents used in electrolytes for large scale lithium ion batteries, Phys. Chem. Chem. Phys. 15 (23) (2013) 9145. [4] G. Eshetu, J. Bertrand, A. Lecocq, S. Grugeon, S. Laruelle, M. Armand, G. Marlair, Fire behavior of carbonates-based electrolytes used in Li-ion rechargeable batteries with a focus on the role of the LiPF6 and LiFSI salts, J. Power Sources 269 (2014) 804–811. [5] M. Nakayama, S. Wada, S. Kuroki, M. Nogami, Factors affecting cyclic durability of all-solid-state lithium polymer batteries using poly(ethylene oxide)-based solid polymer electrolytes, Energy Environ. Sci. 3 (12) (2010) 1995. [6] P. Wright, Electrical conductivity in ionic complexes of poly(ethylene oxide), Polym. J. 7 (5) (1975) 319–327. [7] C. Wang, X. Zhang, A. Appleby, Solvent-free composite PEO-ceramic fiber/mat electrolytes for Lithium secondary cells, J. Electrochem. Soc. 152 (1) (2005) A205. [8] W. Liu, N. Liu, J. Sun, P. Hsu, Y. Li, H. Lee, Y. Cui, Ionic conductivity enhancement of polymer electrolytes with ceramic nanowire fillers, ACS Nano 15 (4) (2015) 2740–2745.
68
Contents lists available at ScienceDirect
Solid State Ionics journal homepage: www.elsevier.com/locate/ssi
Highly conductive, flexible polymer electrolyte membrane based on poly (ethylene glycol) diacrylate-co-thiosiloxane network Camilo Piedrahitaa, Victor Kusumab, Hunaid B. Nulwalac, Thein Kyua,
T
⁎
a
Department of Polymer Engineering, University of Akron, OH 44325, USA AECOM Corporation, 626 Cochrans Mill Rd, Pittsburgh, PA 15236, USA c Department of Chemistry, Carnegie Mellon University, Pittsburgh, PA 15213, USA b
A R T I C LE I N FO
A B S T R A C T
Keywords: Thiol-ene reaction Thiosiloxane Poly(ethylene glycol) diacrylate Solid polymer electrolyte membrane
An amorphous co-network consisting of poly(ethylene glycol) diacrylate (PEGDA) and thiosiloxane was prepared by means of thiol-ene “click” reaction to impart high ion conduction via enhanced segmental mobility of the conetwork. By varying PEGDA-thiosiloxane ratios, several polymer electrolyte membranes (PEM) have been fabricated to control the glass transition temperature (Tg), while keeping the equal molar ratio of succinonitrile (SCN) plasticizer and bis(trifluoromethane sulfonyl)imide (LiTFSI) salt for solid-state lithium ion batteries. It was found that the addition of thiosiloxane into the PEM network not only results in improved thermal stability and greater extensibility, but also reduces Tg leading to higher ionic conductivity that reaches the level of superionic conductor at room temperature, which further increases to 10−2 Scm−1 at high battery operating temperatures (60–70 °C). To evaluate their electrochemical performance, half-cells comprising of LiFePO4/PEM/Li foil were tested by means of cyclic voltammetry. An initial discharge capacity of 148 mAhg−1 was obtained at a current rate of 0.2 C with capacity retention of about 89%.
1. Introduction Nowadays, lithium-ion batteries have been commonly used in our daily life to power portable electronics, wearable devices, and emerging technologies including high power, high energy all solid-state battery for electric vehicles. The current Li-ion battery technology is designed based on organic liquid electrolytes such as mixtures of cyclic ethylene carbonate (EC) with volatile diethylene carbonate (DEC) or dimethyl carbonate (DMC) solvents to afford high Li-ion conduction [1]. However, because of their poor thermal stability and volatility coupled with low flash points, these liquid electrolyte solutions can easily catch fire during battery operation [2–4]. To overcome the aforementioned flammability of liquid electrolytes, solid-state polymer electrolyte membranes (PEM) have been developed in order to alleviate fire hazard of lithium ion battery [5]. The first solid electrolyte was reported by Wright in 1975 based on a mixture of poly(ethylene oxide) (PEO) and alkaline salt that showed an ionic conductivity of 10−7 Scm−1 at room temperature [6]. One major drawback of the early generation of PEO based solid polymer electrolytes (SPE) has been their low ionic conductivities at room temperature, i.e., 10−9–10−7 Scm−1, which is several orders of magnitude lower than that of liquid electrolyte counterparts (10−3–10−2 Scm−1) [7,8].
⁎
Corresponding author. E-mail address: [email protected] (T. Kyu).
https://doi.org/10.1016/j.ssi.2018.05.006 Received 17 April 2018; Received in revised form 2 May 2018; Accepted 6 May 2018 0167-2738/ © 2018 Elsevier B.V. All rights reserved.
It was later in 1979, Armand et al. [9] developed lithium-ion batteries by mixing succinonitrile (SCN) plastic crystals and various lithium salts that have improved high ionic conductivity to the range of 10−4–10−3 Scm−1, but such electrolytes had a paste-like appearance with poor mechanical integrity. The solid-state PEM needs not only to be mechanically sturdy, but also room-temperature ionic conductivity, thermal and electrochemical stability must be improved. The simultaneous improvement of these properties is hard to come by because ionic conductivity and mechanical strength have diametrically opposing effects [10]. PEMs hitherto developed were primarily based on functional polymers and lithium salts such as polyphosphazene [11], olygoethers [12] and PEO-modified polysiloxanes [13]. The manner in which these PEMs were made was by first dissolving lithium salt into a polymer precursor (i.e., prepolymer or macromonomer) and then photo-cured via UV irradiation to obtain a self-standing solid film. PEO-modified polysiloxane is an interesting approach in the development of novel polymer matrix. Neat polysiloxanes are characterized by their good mechanical strength, chemical, thermal stability, and low glass transition temperatures, but poor conductivity [14,15]. However, siloxane modification with PEO can afford high ion conduction in conjunction with good physical properties of siloxane, making these
Solid State Ionics 322 (2018) 61–68
C. Piedrahita et al.
imide (LiTFSI) (purity 99.9%), succinonitrile (purity > 99%) (SCN) and 2-butanone (purity ≥90%) were purchased from Sigma-Aldrich. Poly(vinylidene fluoride) (PVDF) having Mw of 534,000 (SigmaAldrich) was dissolved in anhydrous (99.5%) 1-methyl-2-pyrrolidinone (NMP) (Sigma-Aldrich) solvent. Electrode material, viz., lithium iron phosphate (LiFePO4) obtained from MTI Corp., was blended with acetylene black (AB) (also from MTI) to improve electron conductivity of the cathode. In the cathode fabrication, the solution mixture consisting of PVDF (10 wt%), acetylene black (10 wt%), and active electrode material (LiFePO4) (80 wt%) in NMP was used as electrode binder in the slurry form.
polymer matrices highly desirable for use in all solid-state lithium ion batteries. However, neat polysiloxane derivatives are difficult to be dissolved in PEO matrix. One common approach is to modify these siloxanes by covalently linking with polar polyether groups to improve solubility in the PEO matrix. In 1988, Khann synthesized comb-like polysiloxane having oligo(oxyethylene) side branches and subsequently incorporated them in the PEM [16]. The ionic conductivity values of these PEMs were improved to the acceptable conductivity range of the order of 10−4 Scm−1, albeit still low. Wang et al. [17] synthesized siloxane polymer and then mixed it with lithium bis(trifluoromethyl sulfonyl)imide (LiTFSi) salt. However, the highest ionic conductivity value thus obtained was on the order of 10−4 Scm−1 at room temperature. Similarly, Zhang et al. [18] prepared a solid PEM via hydrosilylation of polymethylhydrosiloxane (PMHS) and subsequently mixed with LiTFSI salt. They achieved a similar order of magnitude in ionic conductivity (i.e., 10−4 Scm−1) with their crosslinked siloxane-based solid PEM network having a low Tg of approximately −50 °C. Burjanadze et al. [19] investigated the electrochemical stability of crosslinked salt-in-polysiloxane membranes containing two lithium salts, i.e., boron based anions, viz., LiDFOB (lithium difluoro(oxalato)borate) and LiBOB (lithium bis(oxalato)borate). The ionic conductivity values were relatively low that only reached the order of 10−5 Scm−1 at room temperature, but with an enhanced electrochemical stability up to 4.7 V against Li/Li+. One interesting approach in fabricating the siloxane-PEO compounds without performing tedious chemical synthesis is to conduct crosslinking reactions between polyethylene glycol derivatives such as poly(ethylene glycol) diacrylate (PEGDA) and thiol (SH) containing compound such as thiosiloxane via the thiol-ene ‘click’ reaction between thiol (SH) of siloxane and the C]C double bonds of functional acrylate groups of the polyether backbone [20]. This click reaction is reported to be fast, easy to execute, and more importantly provides high yield. In fact, such chemical route has been already demonstrated by the co-authors while they were at DOE [21,22], who reported the achievement of completely amorphous crosslinked co-networks consisting of PEGDA and thiosiloxane. These thiosiloxane/PEGDA co-networks have high selectivity and separation of carbon dioxide (CO2) from nitrogen (N2) or methane (CH4) with potential applications in control of greenhouse gas emission [23]. The presence of both polar and non-polar moieties in the thiosiloxane/PEGDA co-network has been attributed to be the key factor in the enhancement of gas permeability and selectivity. In this article, the aforementioned thiol-ene ‘click’ crosslinking reaction was employed to afford an amorphous thiosiloxane/PEGDA conetwork for use as a polymer matrix in the polymer electrolyte membrane for solid-state lithium ion battery. Firstly, the miscibility of a binary thiosiloxane/PEGDA mixture was determined in order to provide guidance to the formation of isotropic amorphous polymeric conetwork. Secondly, various thiosiloxane/PEGDA ratios were further mixed with lithium salt (LiTFSI) and succinonitrile (SCN) plasticizer to produce a completely amorphous, solid-state polymer electrolyte membrane (PEM). Mechanical and ion conduction properties of the above thiosiloxane/PEGDA membranes were evaluated by means of tensile tests and AC impedance spectroscopy, respectively. Cyclic voltammetry measurements were also carried out in order to evaluate electrochemical performance and discuss potential utilization of siloxane-PEGDA PEMs as solid electrolytes for all solid-state lithium ion batteries.
2.1. Binary thiosiloxane-PEGDA system Various thiosiloxane/PEGDA binary mixtures were prepared in accordance with the predetermined weight ratios thiosiloxane:PEGDA 10:90, 15:85 and 25:75. These binary components were fully dissolved upon mixing in the liquid state. By adding the indicated amount of DMPA photoinitiator, the thiol (SH) group of thiosiloxane was activated and subsequently reacted with acrylate groups of PEGDA via the socalled thiol-ene ‘click’ reaction. Note that the photo-initiator (DMPA) amount was kept at 3 wt% with respect to the total prepolymer amount. This thiol-ene reaction produced a crosslinked network between nonpolar siloxane and polar PEGDA. Upon photo-crosslinking, the binary blend membrane remained homogeneous and transparent, suggestive of the formation of an isotropic co-network. 2.2. Pseudo-ternary system – thiosiloxane-PEGDA/SCN/LiTFSI Binary mixtures of thiosiloxane/PEGDA and SCN/LiTFSI were prepared separately inside a glovebox under Argon atmosphere. Thiosiloxane/PEGDA blends in the indicated weight ratios were found to be completely miscible. However, the SCN/LiTFSI mixture had to be heated above the crystal melting temperature of SCN (i.e., 60 °C) for several minutes in order to completely dissolve and form a clear melt mixture. Subsequently, these individual binary mixtures were further mixed with thiosiloxane/PEGDA in their melt state. The initial quaternary (thiosiloxane:PEGDA/SCN/LiTFSI) liquid mixture exhibited a cloudy appearance, but it turned transparent upon adding 2-propanone as a common solvent. After exposure to UV irradiation for 5 min, copolymerization occurred between thiosiloxane and PEGDA via the ‘click’ reaction and formed a thiosiloxane-PEGDA copolymer. Upon removal of solvent, a clear, solid polymer electrolyte membrane (PEM) consisting of thiosiloxane-PEGDA copolymer, SCN and LiTFSI (viz., pseudo-ternary constituents) was obtained. 2.3. Polymer electrolyte membrane The polymer electrolyte membranes (PEM) compositions are given in weight ratios, e.g., 20/40/40 (thiosiloxane:PEGDA)/SCN/LiTFSI. In the PEM formulation, the copolymer ratio of thiosiloxane:PEGDA was varied as 2:18(10/90)/40/40, 3:17(15/75)/40/40 and 5:15(25/75)/ 40/40, while fixing their total amount at 20 wt%. Also the weight percent for SCN/LiTFSI was kept as constant (at 40 wt% each) such that the only variable is the copolymer ratio. The amount of photo-initiator (DMPA) was kept at 3 wt% with respect to the total prepolymer amount. It should be pointed out that the above PEM mixture belongs to the isotropic region of their ternary phase diagram (figure not shown). This transparent mixture was spread within a spacer (10 mm × 10 mm frame and 0.3 mm thick). Subsequently, it was photocrosslinked under uniform UV irradiation at 350 nm with an intensity 5 mW cm−2 for 2 min. After curing, solid films thus obtained were subsequently dried for 60 min at 80 °C and then kept inside a glovebox overnight under argon atmosphere. These fabricated PEM film samples were characterized by using various characterization techniques such as Fourier transformed infrared (FTIR), Raman, wide angle X-ray
2. Materials and methods (Mercaptopropyl) methyl siloxane prepolymer (thiosiloxane), Mw = 5000 was purchased from Gelest. Poly(ethylene glycol) diacrylate (PEGDA), Mw = 700 g/mol was bought from Sigma-Aldrich. Photo-initiator Irgacure®651 (2,2-dimethoxy-2-phenylacetophenone, DMPA) was obtained from CIBA. Lithium bis(trifluoromethyl sulfonyl) 62
Solid State Ionics 322 (2018) 61–68
C. Piedrahita et al.
Fig. 1. FTIR spectra for (a) binary thiosiloxane-PEGDA system (15:85 ratio by weight) (left), showing the reduction of stretching vibration (va) mode of the C]C bonds before (grey line) and after (black line) UV curing. Enlarged area shows the peak of 1627 cm−1 corresponding to the stretching of C]C bond and (b) (thiosiloxane:PEGDA)/SCN/ LiTFSI (left) displaying reduction of the peak at 1627 cm−1 for the equivalent vibrational mode before (grey line) and after (black line) UV exposure.
polished stain-steel electrodes having an area of 10 mm × 10 mm with a thickness of 1 mm. To evaluate the electrochemical stability of the PEM, cyclic voltammetry (CV) tests were performed at room temperature (25 °C) by using a SI 1260 Impedance/Gain Phase Analyzer and SI 1287 Electrochemical Interface (Solartron Analytical Inc.). A coin cell was prepared in the stain steel (SS)/PEM/Li assembly in which SS represented the working electrode, whereas the Li foil (Alfa Aesar) acted as both counter- and reference-electrodes. A linear sweep voltammetry was performed between 1–6 V at a sweep rate of 1 mV/s. The cyclic voltammetry measurement was carried out at a scan rate of the 0.2 mV/ s in the voltage range of −0.5 to 4.5 V. The thickness of PEM films ranged from 150 μm to 250 μm. For the cathode preparation, a slurry was made in NMP at a weight ratio of 80/10/10 corresponding to active electrode material/carbon black (CB)/poly(vinylidene fluoride) (PVDF) and further spread on Al foil (current collector). The coated electrodes were first dried on a hotplate at 80 °C for 2 h and then vacuumed dried in oven at 110 °C overnight. LiFePO4 loading was in the range of 0.5–1.3 mg. Further assembling of the coin cell was performed inside a glovebox under Argon gas environment. Cyclic voltammetry test was performed in halfcell configuration by sandwiching the PEM between LiFePO4-cathode and Li-foil (i.e., LiFePO4/PEM/Li foil) at a scan rate of the 0.2 mV/s within the voltage range of 2.5 to 4.2 V. Galvanostatic charge/discharge cycling test was undertaken on the half-cell under the same configuration within the established stable voltage of the electrode utilizing a MTI 8-channel battery cycler (MTI Corp.) at a constant current rate of 0.2 C.
diffraction (WAXD) along with AC impedance spectroscopy and cyclic voltammetry.
2.4. Membrane characterization Differential scanning calorimetry (DSC) (TA Instruments, Model Q200) was carried out to determine glass transition temperature of various binary blends as well as their PEMs containing specified amounts of SCN plasticizer and LiTFSI salt. Recommended amount of 5 mg the blends were placed in aluminum pans and hermetically sealed using aluminum lids in a glovebox under nitrogen to avoid moisture absorption. All DSC scans were acquired at a ramp rate of 10 °C/min, unless indicated otherwise. Fourier Transformation Infrared (FTIR) spectroscopy was performed in attenuated total reflection (ATR) mode by utilizing a Thermo Fisher Scientific instrument (Model Nicolet iS50 Spectrometer). Prior to FTIR investigation, dried samples were stored overnight inside a glovebox under argon environment and sealed in vials to prevent any exposure to air and/or possible contamination before analysis. By placing the PEM on the ZnSe crystal, the IR spectra were acquired in the spectral range of 450 to 4000 cm−1 with an average of 32 scans and a spectral resolution of 4 cm−1. RAMAN spectroscopy was carried out using a Micro Raman Spectrometer (LabRam HR, Horiba equipped with an Nd:YAG laser at 532 nm and 60 mW) for an exposure time of 5 min. After UV curing, (thiosiloxane:PEGDA)/SCN/LiTFSI PEM film exhibits transparent appearance suggestive of a homogeneous mixture although by no means a proof. To verify the absence of any crystals in the PEM, WAXD 2θ-scans were acquired on a D/MAX RAPID II microdiffractometer. The x-ray (Cu-Kα radiation at the wavelength of 1.5417 Å) was collected over a scattering angle range of 0 < 2θ < 45. Thermogravimetric Analysis (TGA) experiment was carried out on dried, cured PEM films (weighing recommended amount of approximately 10 mg) by using a TA Instrument Model Q50 within a temperature range of 25 °C to 300 °C at a ramped rate of 10 °C/min. Differential scanning calorimetry (DSC) was performed on a TA Instrument model Q2000. To determine the glass transition of the PEM, dried samples weighing about 100 mg were sealed in an aluminum pan and subsequently DSC scans were performed within a temperature range from −130 and 30 °C at a ramp rate of 5 °C/min. Tensile tests of the PEM films were performed by mean of dynamic mechanical analyzer (DMA, TA Q800, TA Instruments, Inc.) Film samples having the dimensions of 10.0 mm × 5.0 mm × 0.6 mm (length × width × thickness) were stretched at 25 °C in air until break using a force rate of 0.3 N/min. Ionic conductivity measurements were carried out by means of an AC impedance spectrometer (HP4192A, Hewlett-Packard) interfaced to a standard PC computer by scanning from 13 MHz to 5 Hz at an applied voltage of 1 V in the temperature range of 25–130 °C. In the AC impedance measurements, the dried PEM films were placed between two
3. Results and discussion 3.1. FTIR characterization of thiol-ene reaction in thiosiloxane/PEGDA conetwork Various thiosiloxane-PEGDA mixtures were prepared by mixing in the liquid state at ambient temperature. Fig. 1a shows the IR spectra of a thiosiloxane/PEGDA mixture (15:85 ratio by weight) acquired before and after UV exposure for a few mins. The rate of photopolymerization of the curing reaction is so fast that it requires only 1 min of UV irradiation. Of particular importance is that all signals corresponding to the C]C bonds of acrylate groups, located at 1627 cm−1 (stretching) which virtually disappeared due to the thiol-ene reaction between the acrylate double bonds and thiol group. The estimated conversion of the reaction based on the areas under the above IR bands was approximately 93%. A similar observation of reaction conversion was made for other compositions, viz., 5:95 and 10:90 ratios by weight of the thiosiloxane/PEGDA mixtures (data not shown). When additional components of the PEM such as SCN plasticizer and LiTFSI salt are incorporated into the binary thiosiloxane-PEGDA 63
Solid State Ionics 322 (2018) 61–68
C. Piedrahita et al.
DMPA
+
UV
Fig. 2. Proposed reaction scheme of thiosiloxane and PEGDA forming a co-network.
2572 cm−1 in the (3:17)/40/40 and (5:15)/40/40 compositions implies that the SH bonds are either fully consumed or the signal falls below its detection limit. According to the IR experiment, the conversion of thioene reaction is about 90%, and thus the lack of Raman peak at 2572 cm−1 in the 15:85 (3:17) and 25:75 (5:15) may simply suggest that sizable portions of the SeH groups were reacted with the acrylate double bonds and only an undetectable level of thiosiloxane may be left unreacted. Moreover, when the concentration of the 25:75 (thiosiloxane:PEGDA) exceeds to the exact molar ratio 1:4 (20:80 thiol:acrylate), there is a possibility of excess thiosiloxane molecules not involved in the reaction and thus the complete disappearance of SeH signal simply implies the limitation of the Raman approach [21,23]. It is reasonable to infer that the enhanced molecular mobility afforded by the incorporated plasticizer raises the greater probability of SeH groups to encounter with the acrylate double bonds, thereby increasing the conversion of the thiol-ene reaction as manifested by the virtual depletion of the SeH Raman peak (Fig. 3c).
blend, the characteristic peaks representing eCH2]CH2 bonds at 956 cm−1 and 798 cm−1 showed a slight reduction in their respective areas. However, the signal located at 1631 cm−1 corresponding to the C]C stretching decreases drastically (i.e., the conversion of approximately 90%) after the UV exposure, implying high consumption of these bonds either via “ene-ene” or “thiol-ene” reaction as proposed in Fig. 2. Upon addition of SCN plasticizer into the PEM system, the segmental motion of the network chains is expedited, thereby raising the probability of collision among these active groups to react with each other. It may be emphasized that once the thiol-ene reaction occurs; the newly formed crosslinked junctions would restrict the segmental motions of the network chains. The reduced segmental mobility in turn might prevent the complete conversion of the thiol-ene reaction, as manifested by the residual minor broad peak at 1631 cm−1 after curing (please see the zoom-in inset of Fig. 1b). 3.2. Characterization of thiol-ene reaction in thiosiloxane-PEGDA by RAMAN spectrometry
3.3. WAXD investigation of PEM networks
RAMAN spectroscopic investigation was performed to determine the consumption of thiol (SH) after the “thiol-ene” reaction. The signal from SeH bond located at 2572 cm−1 was detected for the neat thiosiloxane (Fig. 3a) as well as for pseudo-ternary PEM mixtures containing higher thiol concentrations of 3:17 (i.e., 15:85 15% thiosiloxane: 85% PEGDA) and (5:15) (25:75 25% thiosiloxane: 75% PEGDA) (Fig. 3b). In the case of 2:18 (10:90 10% thiosiloxane: 90% PEGDA) system, there is no discernible peak, suggesting that the signal may be below the minimum detectable level of SeH bond by this RAMAN technique. However, upon UV-curing, the observed bands of the 15:85 and 25:75 (thiosiloxane:PEGDA) (i.e., higher contents of SeH) seemingly disappeared, indicating that a high percentage of SeH groups was consumed in the photoreaction (Fig. 3c). It should be cautioned that the absence of the RAMAN peak at
It should be pointed out that the starting PEGDA precursor and LiTFSI are crystalline, whereas SCN exhibits the plastic crystalline phase at room temperature. Although the photo-cured PEM shows a single Tg, it is essential to perform wide angle x-ray diffraction (WAXD) investigation to determine the internal structure of the PEM by varying the thiosiloxane:PEGDA ratios, while fixing the concentration of SCN plasticizer and LiTFSI salt at 40% each. Fig. 4 exhibits the Bragg's scans of the PEMs (thiosiloxane:PEGDA)/ SCN/LiTFSI at three different compositions. The scattering profiles revealed two broad scattering peaks at 2θ of 11.9 and 19.6, corresponding to the radial distribution functions arising from the amorphous thiosiloxane - PEGDA co-network. The lack of crystalline peaks of the polymer matrix (PEGDA) is consistent with those reported by Echeverri et al. [25] for various ternary PEGDA/SCN/LiTFSI mixtures
Fig. 3. (a) RAMAN spectra for neat thiosiloxane, (b) thiosiloxane-PEGDA mixtures before crosslinking and (c) crosslinked thiosiloxane-PEGDA co-network based PEMs as a function of co-network ratios, showing the complete disappearance of the SeH peak suggestive of the occurrence of thiol-ene reaction during the conetwork formation. 64
Solid State Ionics 322 (2018) 61–68
C. Piedrahita et al.
thiosiloxane compositions. In the thermograms, there is no appreciable weight loss (< 5%) at 110 °C, indicating that there is little or no residual solvent or moisture in the PEMs. That is to say, the initial decomposition temperature (IDT) corresponding to an initial weight loss of < 5% was noticed only in the vicinity of 110–115 °C, implying that the PEMs may be thermally stable at such high temperature environments without losing their ingredients of the PEM networks. Fig. 6a shows the DSC thermograms of the photo-crosslinked thiosiloxane-PEGDA as a function of composition in weight ratios. Neat thiosiloxane (denoted by 100:0), which is known to be hydrophobic, exhibits a very low glass transition temperature (Tg) of −95 °C in the DSC thermogram (i.e., the mid-point of the inflexion of the DSC curve). On the other hand, the neat PEGDA (denoted by 0:100), which is hydrophilic, reveals a distinct Tg at −39 °C (Fig. 6a). With increasing PEGDA, the single Tg moves gradually with increasing thiosiloxane to the indicated temperatures intermediate between those of the neat constituents. At 25 wt% loading of thiosiloxane to PEGDA co-network, the Tg has declined to −50 °C (Fig. 6a). The revelation of a single glass transition in combination with the trend of the gradual movement of the single Tg with composition implies the formation of amorphous conetworks, albeit showing slight negative departure from the estimated linear relationship. This finding is consistent with the amorphous character of the WAXD scans reported earlier in Fig. 4. Fig. 6b displays thiosiloxane-PEGDA co-network as PEM components. These exhibits the Tg values around −83 °C ( ± 1 °C). Inclusion of the sizable amount of a solid plasticizer/ionizer such as SCN further promotes the segmental motion of the thiosiloxane-PEGDA co-network. These Tg values were found to be consistent with at those reported by He, et al. [26] for the ternary system PEGDA/SCN/LiTFSI 20/40/40 (i.e., Tg of approximately −82 °C), which also exhibited a high ionic conductivity on the order of 8.4 × 10−4S/cm at room temperature.
Fig. 4. WAXD 2θ - scans for thiosiloxane-PEGDA based PEM's at various compositions fixing the SCN plasticizer and LiTFSI salt concentration at 40 wt%. The PEM is transparent and shows no discernible crystalline peaks, suggestive of amorphous character.
corresponding to the isotropic region of their ternary phase diagrams. The resulting film remains transparent without any indication of phase separation during photo-crosslinking, suggestive of a completely crosslinked co-network (Fig. 4, please see the inset picture). The absence of crystalline WAXD peaks in conjunction with the observed optical clarity and a single Tg of the PEM imply the completely amorphous character of the present PEM. This finding is consistent with the single Tg with composition trend obtained from the DSC investigation to be discussed in a latter session. These observations were also in agreement with that reported by Kusuma et al. [23] for the binary thiosiloxane-PEGDA network.
3.5. Mechanical properties of PEMs 3.4. Thermal characterization of PEMs
Fig. 7 depicts the stress-strain behaviors for the PEMs as a function of PEM matrix compositions while keeping the constant ratio of SCN plasticizer and LiTFSI salt at 40 wt% each. The neat PEGDA based PEM, denoted by (0:20)/40/40 (thiosiloxane: PEGDA)/SCN/LiTFSI, exhibits a brittle character, having a high tensile modulus, but the lowest elongation-at-break of around 5%. With increasing thiosiloxane to 2:18 ratio, the PEM (2:18)/40/40 (thiosiloxane:PEGDA)/SCN/LiTFSI PEM shows an improved elongation to 17%. The (3:17)/40/40 composition shows the highest value of approximately 20% elongation to that of (2:18)/40/40 (thiosiloxane:PEGDA)/SCN/LiTFSI. With further increase of the thiosiloxane ratio to 5:15 in the PEMs both tensile strength and modulus have declined noticeably, which may be attributed to the excess thiosiloxane, which may be left unreacted above the molar ratio of 1:4 (4:16). It is reasonable to infer that the inclusion of thiosiloxane increases the segmental motion of the matrix as manifested by a lower glass transition temperature (Tg) of the final co-network, thereby increasing the network extensibility. This improved chain flexibility is anticipated to improve the PEM performance eventually, since it can alleviate the well-known problem of electrode cracking driven by expansion and contraction of the PEM/ binder during battery operation. In our recent study, the attachment of dangling side chains to the PEM network provides greater mobility of the network chains resulting in not only lowering glass transition temperatures, but also increasing ionic conductivity considerably. More importantly, the enhanced chain mobility at the PEM/electrode interface further improves the capacity retention. Hence, it is important to investigate the electrochemical performance of the present thiosiloxane-PEGDA PEM.
Li-ion batteries are generally known to generate heat during battery operation; e.g., the temperature of some electronic devices can reach 50–70 °C. Hence, thermal management is of paramount importance to provide thermally stable environment to the PEM. Silicon-based compounds, such as thiosiloxane are known for their good thermal stability and hydrophobicity. Once the prepolymer thiosiloxane/PEGDA mixture was co-polymerized via thiol-ene reaction, the resulting co-network becomes thermally stable up to 300 °C [23]. However, in the PEM formulation, plasticizers such as SCN and/or EC and LiTFSI salt were compounded into the thiosiloxane/PEGDA mixtures and thus thermal stability of such multicomponent PEM should be examined. Fig. 5 exhibits the TGA thermograms for PEM's at various
3.6. Ionic conductivity Fig. 5. TGA thermograms for PEM containing various thiosiloxane contents.
Fig. 8 exhibits the plot of the ionic conductivity versus the 65
Solid State Ionics 322 (2018) 61–68
C. Piedrahita et al.
Fig. 6. DSC thermograms of UV-cured (a) thiosiloxane-PEGDA and (b) (thiosiloxane-PEGDA)/SCN/LiTFSI PEM as a function of compositions.
expression; σ = l/(AZ) where l represents the thickness, A the area of the sample and Z total impedance. The PEM containing neat PEGDA denoted by (0:20)/40/40 shows high room temperature ionic conductivities of 9.3 × 10−4 S/cm at room temperature, whereas the neat thiosiloxane (20:0) mixture exhibits a liquid-like appearance and a slightly lower value of 8.4 × 10−4 S/cm. Of particular interest is that all intermediate copolymer compositions exhibited higher values relative to the neat PEGDA or thiosiloxane networks. That is to say, the ionic conductivities for the PEM having (thiosiloxane:PEGDA) 2:18 and 3:17 ratios improves due to higher flexibility (or elongation-at-break) to the values of 2.17 × 10−3 S/cm and 2.48 × 10−3 S/cm at 25 °C, respectively. With further increase of thiosiloxane:PEGDA ratio to 5:15, the ionic conductivity has declined to 1.07 × 10−3 S/cm due to the excess thiosiloxane which were presumably left unreacted above the molar ratio of 1:4. This finding is consistent with that of the stress-strain behavior reported earlier (see Fig. 7). Of particular interest is that some thiosiloxane-modified PEMs such as 2:18 and 3:17 can reach the superionic conductivity level on the order 10−3 S/cm at room temperature. It may be hypothesized that lithium ion transport is governed by two opposing effects; the ion-dipole complexation (or coordination bonds between lithium cations and ether oxygen) which retards the segmental motion and plasticization of the polyethylene glycol network (as manifested by lower Tg) as well as dissociation of lithium-ether oxygen complexes by SCN [27–29]. Upon heating above 70 °C some compositions can reach the order of 10−2 S/ cm, which is comparable in magnitude to the organic liquid electrolyte counterpart, although most of liquid electrolyte batteries would not survive at such high operating temperatures above 60 °C. It can be noticed that the temperature dependence of the ionic conductivity for thiosiloxane-modified PEMs exhibit departures from the Arrhenius behavior (i.e., non-linear slope). Ion conduction is governed by the mobility of cations coupled to the segmental motion of the polymer chains and thus it may be better represented by VogelTammann-Fulcher (VTF) equation [10,30]. All PEMs exhibited curvatures, which may be fitted with the respective VTF parameters (Table 1), i.e., the reference temperature (To) was taken as To = Tg − 50 °C.
Fig. 7. Strain-stress curves for thiosiloxane-PEGDA PEM's. Increasing thiosiloxane content increases the elongation-at-break of the PEM up to 3:17 ratio of the copolymer, then it falls off the molar ratio of 1:4, i.e., 5:15 ratio.
Fig. 8. Variation of ionic conductivity versus reciprocal temperature for thiosiloxane PEM's at various thiosiloxane concentrations showing curvatures suggestive of non-Arrhenius character. Note that (0:20)/40/40 and (20:0)/40/ 40 represent pure PEGDA and pure thiosiloxane PEMs. The solid curves were fitted in accordance with VTF equation.
Table 1 VTF parameters; room temperature ionic Conductivity (σRT) and activation energy (B) for various (thiosiloxane:PEGDA)/SCN/LiTFSI compositions.
reciprocal absolute temperature for various PEM compositions. Ionic conductivity values were calculated according to the following 66
(Thiosiloxane:PEGDA)/SCN/LiTFSI (wt%)
σRT (S/cm)
B (kJ/mol)
R2
(2:18)/40/40 (3:17)/40/40 (5:15)/40/40
1.81 × 10−3 2.10 × 10−3 1.07 × 10−3
5.9 6.6 6.7
0.998 0.999 0.999
Solid State Ionics 322 (2018) 61–68
C. Piedrahita et al.
Fig. 9. a) Linear sweep voltammetry performed between 1 and 6 V for SS/thiosiloxane-PEG PEM/Li foil. b) Voltammogram for thiosiloxane-PEG PEM between −0.5 to 4.2 V. Zoom area from 0.3–1.4 V evidencing appearance of peaks that come from side redox reaction within the PEM.
3.7. Electrochemical performance of PEMs
3.8. Charge-discharge test for PEM
Fig. 9a shows the linear sweep voltammetry conducted in the range of 1–6 V with the aim of detecting the maximum voltage at which the PEM can be stable and also to detect the voltage range whereby irreversible side oxidative reactions may be taking place. These reactions can affect its capability to undergo through reversible reactions and will affect the performance of the PEM. Samples were assembled in a Limetal/PEM/stain-steel (SS) configuration. (thiosiloxane:PEGDA)/SCN/ LiTFSI (3:17)/40/40 PEM appeared stable up to 4.8 V versus Li/Li+ as indicated by an arrow in Fig. 9a. Similar electrochemical stability limit was found for other compositions, and thus the upper limit voltage was set at 4.2 V, which is below the onset voltage versus current to prevent any irreversible reaction. The lower limit was set at 0.5 V, i.e., above the plating and stripping (reduction and oxidation) potential of the Li-ions. Voltammograms reveal the presence of two redox reaction peaks at −0.3 V and 0.5 V peaks, corresponding to the reduction (cathodic) and oxidation (anodic) associated with Li stripping/plating (Fig. 9b). A small reduction peak can be discerned at approximately 0.55 V which is accompanied by an oxidation signal around 1.02 V. These signals are believed to arise from the redox reaction of the PEM containing LiTFSI salt [31]. The intensity of the peaks located at −0.5 V and 0.25 V (related to the concentration of Li+ ions) remains virtually invariant suggestive of reversible redox reactions. The same experimental conditions were adopted for other compositions (2:18 and 5:15 thiosiloxane:PEG) that showed a similar trend (data not shown).
Fig. 10a exhibits the cyclic voltammetry of the LiFePO4/PEM/Li half-cells performed at ambient temperature (25 °C) at a current rate of 0.2 mV/s. Generally, LiFePO4 has been used as a cathode material due to its high voltage operation range (i.e., 2.5–4.2 V). It can be seen in the voltammograms that the cathodic peak for LiFePO4 is located at 3.7 V vs Li/Li+ (corresponding to the Li-ions lithiation (i.e., reduction)), while the andic peak was observed at 3.3 V vs Li/Li+ which represents delithiation (i.e., oxidation) of Li-ion. It can be noticed that the peaks from the second and third cycles remain invariant with increasing number of cycles, indicating excellent reversibility of the redox reactions at the PEM/electrode interface. Fig. 10b exhibits the galvanostatic charge/discharge cycling test of the same half-cell at ambient temperature (25 °C). The current rate was set at 0.2 C (i.e., it takes 5 h to complete 1 cycle). Specific charge/discharge capacities are plotted with respect to the number of cycles. It can be seen that the half-cell battery (thiosiloxane:PEGDA)/SCN/LiTFSI (3:17)/40/40) exhibits a reasonable rechargeable characteristic. That is to say, the observed initial capacity value was 148 mAh g−1, but it gradually dropped to around 132 mAh g−1, i.e., 89% retention. Moreover, the Coulombic efficiency is almost 100% for at least 50 cycles tested indicating reasonable charge/discharge cyclic stable behavior. According to a previous paper [31], the post-mortem analysis revealed no noticeable changes in the interface structure during cycling at room temperature, but some structural changes occurred in the LiFePO4/ PEGDA-based PEM/Li half-cell after 50 cycling at an elevated temperature of 60 °C. With the lithium bis(oxalato)borate LiBOB modification, such structural change has been alleviated [31].
Fig. 10. (a) Voltammogram for half-cell LiFePO4/PEM/Li the voltages range operation 2.5 V to 4.2 V. The ramp rate was 0.5 mV/s. (b) Capacity retention behavior of PEM (thiosiloxane:PEGDA)/SCN/LiTFSI (3:17)/40/40 in a half-cell charge/discharge test versus number of cycles at 0.2C for 50 cycles. 67
Solid State Ionics 322 (2018) 61–68
C. Piedrahita et al.
4. Conclusions
[9] M.B. Armand, J.R. MacCallum, C.A. Vincent (Eds.), Polymer Electrolyte Reviews, 1 Elsevier Appl. Sci, London, 1987, pp. 1–22. [10] R. He, M. Echeverri, D. Ward, Y. Zhu, T. Kyu, Highly conductive solvent-free polymer electrolyte membrane for lithium-ion batteries: effect of prepolymer molecular weight, J. Membr. Sci. 498 (2016) 208–217. [11] G. Nazri, D. MacArthur, J. Ogara, Polyphosphazene electrolytes for lithium batteries, Chem. Mater. 1 (3) (1989) 370–374. [12] M. Watanabe, Effects of network structures and incorporated salt species on electrochemical properties of polyether-based polymer electrolytes, Solid State Ionics 79 (1995) 306–312. [13] J. Smid, D. Fish, I. Khan, E. Wu, G. Zhou, Polysiloxane-based polymer—electrolyte complexes, Adv. Chem. (1989) 113–123. [14] N. Grassie, K. Francey, I. Macfarlane, The thermal degradation of polysiloxanes—part 4: poly(dimethyl/diphenyl siloxane), Polym. Degrad. Stab. 2 (1) (1980) 67–83. [15] Y. Abe, T. Gunji, Oligo- and polysiloxanes, Prog. Polym. Sci. 29 (3) (2004) 149–182. [16] I. Khan, Y. Yuan, D. Fish, E. Wu, J. Smid, Comblike polysiloxanes with oligo(oxyethylene) side chains. Synthesis and properties, Macromolecules 21 (9) (1988) 2684–2689. [17] F. Wang, C. Hu, S. Lo, Y. Wang, C. Wan, The investigation of electrochemical properties and ionic motion of functionalized copolymer electrolytes based on polysiloxane, Solid State Ionics 180 (4–5) (2009) 405–411. [18] Z. Zhang, D. Sherlock, R. West, R. West, K. Amine, L. Lyons, Cross-linked network polymer electrolytes based on a polysiloxane backbone with oligo(oxyethylene) side chains: synthesis and conductivity, Macromolecules 36 (24) (2003) 9176–9180. [19] M. Burjanadze, Y. Karatas, N. Kaskhedikar, L. Kogel, S. Kloss, A. Gentschev, M. Hiller, R. Müller, R. Stolina, P. Vettikuzha, H. Wiemhöfer, Salt-in-polymer electrolytes for lithium ion batteries based on organo-functionalized polyphosphazenes and polysiloxanes, Z. Phys. Chem. 224 (10−12) (2010) 1439–1473. [20] C. Hoyle, A. Lowe, C. Bowman, Thiol-click chemistry: a multifaceted toolbox for small molecule and polymer synthesis, Chem. Soc. Rev. 39 (4) (2010) 1355. [21] L. Kwisnek, S. Heinz, J. Wiggins, S. Nazarenko, Multifunctional thiols as additives in UV-cured PEG-diacrylate membranes for CO2 separation, J. Membr. Sci. 369 (1–2) (2011) 429–436. [22] M. Mellott, K. Searcy, M. Pishko, Release of protein from highly cross-linked hydrogels of poly(ethylene glycol) diacrylate fabricated by UV polymerization, Biomaterials 22 (9) (2001) 929–941. [23] V. Kusuma, E. Roth, W. Clafshenkel, S. Klara, X. Zhou, S. Venna, E. Albenze, D. Luebke, M. Mauter, R. Koepsel, A. Russell, D. Hopkinson, H. Nulwala, Crosslinked poly(ethylene oxide) containing siloxanes fabricated through thiol-ene photochemistry, J. Polym. Sci. A: Pol. Chem 53 (13) (2015) 1548–1557. [25] M. Echeverri, N. Kim, T. Kyu, Ionic conductivity in relation to ternary phase diagram of poly(ethylene oxide), succinonitrile, and lithium Bis(trifluoromethane) sulfonyl imide blends, Macromolecules 45 (15) (2012) 6068–6077. [26] R. He, T. Kyu, Effect of plasticization on ionic conductivity enhancement in relation to glass transition temperature of crosslinked polymer electrolyte membranes, Macromolecules 49 (15) (2016) 5637–5648. [27] E. Rietman, M. Kaplan, R. Cava, Lithium ion-poly (ethylene oxide) complexes. I. Effect of anion on conductivity, Solid State Ionics 17 (1) (1985) 67–73. [28] R. Frech, W. Huang, Conformational changes in diethylene glycol dimethyl ether and poly(ethylene oxide) induced by lithium ion complexation, Macromolecules 28 (4) (1995) 1246–1251. [29] V. Aravindan, J. Gnanaraj, S. Madhavi, H. Liu, Lithium-ion conducting electrolyte salts for lithium batteries, Chem. Eur. J. 17 (51) (2011) 14326–14346. [30] K.M. Diederichsen, H.G. Buss, B.D. McCloskey, The compensation effect in the Vogel–Tammann–Fulcher (VTF) equation for polymer-based electrolytes, Macromolecules 50 (2017) 3831–3840. [31] R. He, F. Peng, W.E. Dunn, T. Kyu, Chemical and electrochemical stability enhancement of lithium bis(oxalato)borate (LiBOB)-modified solid polymer electrolyte membrane in lithium ion half-cells, Electrochim. Acta 246 (2017) 123–134.
In this work, we have incorporated thiosiloxane as a matrix cocomponent of the PEM with the aim of enhancing the polymer segmental motion and hence, the ionic conductivity. Thiosiloxane and PEGDA mixtures were crosslinked by UV irradiation via the “thio-ene” reaction between the thiol functionality and the double bonds of the PEGDA precursor, forming a complete homogeneous self-standing PEM. The thiosiloxane modified PEM film exhibits higher extension-at-break in comparison to its counterpart, i.e., the PEM containing neat PEGDA700/SCN/LiTFSI 20/40/40. As evidenced in FTIR and Raman spectroscopy, sizable portions of the thiol (SH) groups were consumed in the thiol-ene reaction suggestive of high conversion in the co-network. With increasing thiosiloxane, the Tg of the PEM was declined, leading to higher ionic conductivity. Thiosiloxane:PEGDA based PEMs are thermally stable exhibiting good anodic stability up to 4.8 V versus Li/Li+. When tested in half-cells using LiFePO4 as cathode against Li/ Li+ electrode, the specific discharge capacity showed the initial capacity value of 148 mAh g−1 with reasonably good capacity retention (i.e., 89%). The achievement of improved thermal, ion conduction, and stable electrochemical performance implies that the present solid-state PEM based on thiosiloxane modified PEGDA co-network has potential for implementing all solid-state Li-ion batteries. Acknowledgment Support of this work by the U.S. National Science Foundation, DMRPolymers, Grant # 1502543 is gratefully acknowledged. References [1] J. Goodenough, K. Park, The li-ion rechargeable battery: a perspective, J. Am. Chem. Soc. 135 (2013) 1167–1176. [2] D. Abraham, E. Roth, R. Kostecki, K. McCarthy, S. MacLaren, D. Doughty, Diagnostic examination of thermally abused high-power lithium-ion cells, J. Power Sources 161 (2006) 648–657. [3] G. Eshetu, S. Grugeon, S. Laruelle, S. Boyanov, A. Lecocq, J. Bertrand, G. Marlair, In-depth safety-focused analysis of solvents used in electrolytes for large scale lithium ion batteries, Phys. Chem. Chem. Phys. 15 (23) (2013) 9145. [4] G. Eshetu, J. Bertrand, A. Lecocq, S. Grugeon, S. Laruelle, M. Armand, G. Marlair, Fire behavior of carbonates-based electrolytes used in Li-ion rechargeable batteries with a focus on the role of the LiPF6 and LiFSI salts, J. Power Sources 269 (2014) 804–811. [5] M. Nakayama, S. Wada, S. Kuroki, M. Nogami, Factors affecting cyclic durability of all-solid-state lithium polymer batteries using poly(ethylene oxide)-based solid polymer electrolytes, Energy Environ. Sci. 3 (12) (2010) 1995. [6] P. Wright, Electrical conductivity in ionic complexes of poly(ethylene oxide), Polym. J. 7 (5) (1975) 319–327. [7] C. Wang, X. Zhang, A. Appleby, Solvent-free composite PEO-ceramic fiber/mat electrolytes for Lithium secondary cells, J. Electrochem. Soc. 152 (1) (2005) A205. [8] W. Liu, N. Liu, J. Sun, P. Hsu, Y. Li, H. Lee, Y. Cui, Ionic conductivity enhancement of polymer electrolytes with ceramic nanowire fillers, ACS Nano 15 (4) (2015) 2740–2745.
68