Introduction
Lithium-ion batteries (LiBs) are currently trending in the portable energy market due to their advantageous flexibility in terms of portability via design, high power density, and broad applications. The reach of LiBs covers various types of electronic gadgets and laptops that are used in our daily life as well as state-of-the-art electric and hybrid vehicles [
1–
4]. Despite the implementation of strategic approaches such use of Li alloys/composite materials as anodes [
5–
7], the immense popularity of LiBs is often marred by the safety concerns, especially with regard to flammability issues. The Boeing Dream Liner 787 mishap resulted in apprehension regarding the safety of LiBs [
8]. Even prior to that incident, several incidents of fire-related accidents in laptops and electric vehicles due to faulty LiBs had been reported [
9]. Most commercial LiBs employ electrolytes, such as LiPF
6 salt, in various polyethylene oxide (PEO)-based electrolytes along with linear and cyclic carbonate additives. PEO-based electrolytes pose a risk due to liquefying and even ignition at higher temperatures. EC: DMC is often the preferred additives in commercial lithium-ion batteries despite the imminent danger of ignition at higher temperatures due to thermal runaway. For example, in a laboratory experiment conducted by our group, a commercial electrolyte (i.e., 1 M LiPF
6-EC: DMC) was subjected to a flammability test, as shown in Fig. 1. The sample was ignited, and the flame was extinguished. The flame failed to cease immediately and was completely consumed within 20 s of ignition. The chemistry behind this behavior can be best described by considering the electrolyte solution as two components. The first component is a lithium salt (i.e., LiPF
6) that is known to decompose above 100°C and release gaseous PF
5 along with insoluble LiF. The second component (i.e., EC: DMC) is the flammable component. In a cell set-up, the release of PF
5 gas under pressurized conditions and high temperatures would lead to serious fire hazards [
10].
The use of a fire safety mechanism can be an expensive alternative, which can adversely affect the portability of LiBs. A smarter alternative may involve the use of flame-resistant electrolytes or flame-retardant additives [
11–
15]. Despite their flammable nature, alkyl carbonates are often employed as electrolyte additives in LiBs due to their high polarity and their ability to dissolve numerous lithium salts. The electrochemical reduction of the electrolyte and additives on the electrode surface results in the formation of a solid electrolyte interface (SEI) layer. In the commercial electrolyte used in our studies, LiPF
6 (EC: DMC) the SEI is mainly due to the electrolytic reduction of EC, DMC, and PF
6 anion, with a major contribution from the alkyl carbonates. During the charge-discharge studies of hybrid electrolyte, EC: DMC was used as an electrolyte additive to mimic the formation of the SEI similar to the commercial electrolyte. Thus, we assert that the structural framework of the preliminary SEI would essentially be similar in either of the cases. However, with regard to subsequent SEI during extensive cycling and therefore, the functional aspect is bound to be different, considering the various additional components in the hybrid electrolyte. An in-depth analysis of the SEI formation requires highly sophisticated techniques like
in situ XRD and Raman spectroscopy.
The development of flame-retardant materials for use as either a co-solvent or additive including various types of phosphates has often met with problems related to the decomposition of commercially viable graphite anodes [
16,
17]. Technically, a safety mechanism composed of non-flammable components, such as a redox shuttle additive, flame-retardant additive or positive temperature coefficient device, are conducive to the prevention of any thermal runaway related disasters [
18]. However, the likelihood of designing flame-retardant or flame-resistant electrolyte alternatives may be more plausible than redesigning non-flammable organic additives. An
in situ sol-gel synthetic route has been employed to synthesize ionic liquids and silica composites [
19]. Our research group has recently published a paper reporting the synthesis and ion conductive parameters of a novel borosilicate polymeric network, in which, the structure of organic-inorganic hybrid ion-gel electrolytes has been reinforced with the borosilicate network. The incorporation of boron into the electrolyte matrix helps to improve the ion conductive properties of electrolyte [
20]. Two different lithium salts have been employed in the synthesis of organic-inorganic hybrid ion-gel electrolytes (i.e., lithium bis(trifluoromethylsulfonyl)imide (LiTFSA)) due to their superior ion-conductive and plasticizing properties. LiPF
6, commonly used in various commercial electrolytes, was used as the other salt [
21]. Low porosity and formation of microcracks during electrochemical cycling often question the reliability of solid electrolytes. Considering the brittle framework of silicate/borosilicate framework in the hybrid electrolyte, expansion and crack formation are inevitable phenomena over long-term cycling. However, B-O-Si frameworks are significantly flexible and durable compared with Si-O-Si, which marks the preferable commercialization of borosilicate materials over pristine SiO
2 materials. As an intuitive pre-cautionary measure, the powdered hybrid electrolyte is used for electrochemical studies. It substantially reduces/negates the issues of expansion or crack formation in the electrolyte. Further, it would be noteworthy to mention that the porosity of the hybrid electrolytes is strongly dependent on the lithium salt and alkoxyborane precursor [
22].
The battery performance of commercial 1M LiPF
6- EC: DMC at 0.1 C charging rate is shown in Fig. 2. The highest discharge capacity is found to be 292 mAh/g. Among the hybrids, discharge capacity at a similar charging current is observed to be in the range of 100–160 mAh/g. The discharge capacity is largely dependent on the type of lithium salt, alkoxyborane precursor and its concentration in the sol-gel mixture, and considered as potential limits during the electrochemical experiments [
22].
Numerically, the discharge capacity of the hybrid electrolyte is lower than of that the commercial electrolyte. Nonetheless, the dependency of capacity on that of the lithium salt concentration in the hybrid electrolytes could be a factor. The concentration of lithium salt is 1 mol in the commercial electrolyte while in the hybrids, the initial concentration of lithium salt used in the sol-gel mixture is 1 mmol, i.e., 1000 times lesser than that of the commercial electrolyte.
In the current study, the nonflammable/flame-retardant attributes of the ion-gel electrolytes compared to those of conventional commercial electrolytes have been confirmed using flame tests. Due to the dispersed ionic liquid phase in the networked borosilicate/silicate scaffold, these hybrids provide advantageous thermal-resistant properties compared to other solid-state electrolytes. The borosilicate polymeric matrix endows the novel organic-inorganic hybrid electrolyte with an incombustible nature and mechanical durability, and the ionic liquid provides optimum ionic conductivity and flame resistance. These hybrids exploit the synergistic union of two flame-resistant or flame-retardant materials. In addition, a comparative flame test for the hybrid electrolyte and commercial electrolyte (i.e., 1 M LiPF6-EC: DMC) has also been conducted to confirm the superiority of the non-flammability/flame retardancy of the organic-inorganic hybrid ion-gel electrolyte. In commercial electrolyte solutions, such as 1 M LiPF6-EC: DMC solutions, the carbonate additives/solvent are primarily responsible for the flammability. In addition to flame tests, complementary structural data have been obtained based on differential scanning calorimetric analyses.
Experimental section
Materials
The organic-inorganic hybrids were synthesized using an
in situ sol-gel condensation process in the presence of alkoxyborane and alkoxysilane precursors and ionic liquids. The detailed procedure and compositional aspects were discussed in Ref. [
21]. However, a brief description of the synthesized hybrids is provided in the flowchart, as displayed in Fig. 3.
Flame tests
The flame tests were conducted by subjecting a designated amount of hybrid to flame inside a hood. The time-sequenced study was performed to confirm the durability of the hybrids in the presence of a flame. Furthermore, the corresponding weight loss due to flammability test was determined. The flammability of all of the hybrids was studied by subjecting a certain mass (10–20 mg) of the sample to a constant flame and measuring the changes in the sample mass after the post-flame treatment. The samples were subjected to a continuous flame for 1 min, and this time was monitored using a stopwatch located in the background.
Differential scanning calorimetric analyses
Differential scanning calorimetry (DSC) was performed using an MDSC 2920 instrument (TA instruments) for a single scan within a temperature scan from ‒100°C to 100°C at a heating rate of 3°C per minute. The instrument was equipped with a liquid nitrogen cooling system to regulate the temperature. The sample weights were 10–15 mg, which were measured in hermetically sealed aluminum pans.
X-ray photoelectron spectroscopic analyses
The XPS analysis was conducted using an XPS system (S-probe TM 2803, Fisons Instruments), installed in a UHV (ultrahigh vacuum) chamber with a base pressure of 3 × 10−8 Pa with a monochromatic Al Ka X-ray source.
Results and discussion
All of the hybrids were prepared by
in situ sol-gel condensation reactions of alkoxyborane/alkoxysilane in the presence of a low viscosity ionic liquid (1,3-diallyl imidazolium TFSA) and acatalytic amount of dil. HCl, according to a procedure that was previously reported by our group [
21]. Hybrids A–F contained LiTFSA, and hybrids G–L contained LiPF
6. For further classification, each set of hybrids differed based on the employed alkoxyborane precursor (i.e., either mesityldimethoxyborane (MDMB) or trimethoxyborane (TMB) as the alkoxyborane precursor). The stoichiometric composition of the hybrids was retained, as reported in our previous study. Typically, the flammability tests of the electrolyte or electrolyte additives were conducted using self-extinguishability times (SETs), which referred to the time required for the flame to be self-extinguished from the time of ignition [
23]. Our preliminary investigations indicated that the organic-inorganic hybrid ion-gel electrolytes were self-extinguishable. Therefore, the testing procedure was changed by subjecting the designated mass (10–20 mg) of electrolyte to a flame for a specific time frame (1 min).
Flammability studies of the hybrids
Hybrids A–C (LiTFSA with MDMB precursor)
Compositional details of Hybrids A–C (LiTFSA with MDMB precursor) are given in Table 1. In general, the transparency of the LiTFSA-based hybrids decreased as the concentration of the alkoxyborane precursor increased irrespective of its type (i.e., either MDMB or TMB) [
21]. In hybrid A, the weight loss was determined to be 15.9% of the initial weight. The flame was pink in color, which might be caused by the burning of anon-trapped lithium salt or exposed lithium from within the matrix resulting from progressive decomposition. Hybrid B closely resembled hybrid A in the flammability test. The weight loss (9.77%) was lower than that of hybrid A (with an evident pink flame), which indicated the structural rigidity of the organic-inorganic hybrid ion-gel electrolyte. After a minute of exposure to the flame, the sample did not become carbonized. Although hybrid C exhibited transparent and opaque domains, only the transparent portions of hybrid C were selected for the flame test. The weight loss (13.2%) was greater than that of hybrid B and similar to that of hybrid A. The weight loss in the flammability tests was indicative of the structural rigidity of the hybrid structures. For example, the greater interactions between the organic-inorganic moieties implied a greater structural rigidity, thereby a lesser weight loss and vice versa. Among the present set of hybrids, hybrid B by virtue of its weight loss pattern might be regarded as the most structurally rigid organic-inorganic hybrid ion-gel electrolyte. The corresponding time sequenced flame tests of hybrids A–C are shown in Fig. 4.
Hybrids D–F (LiTFSA with TMB precursor)
Hybrids D–F with TMB as the alkoxyborane precursor and LiTFSA as the lithium salt additive exhibited a weight loss in the 10%–17% range after a minute of exposure to a naked flame. The images from the flame test are shown in Fig. 5. A pink flame was observed though to a lesser extent compared to its predecessors, indicating a greater degree of networking within the electrolyte matrices due to structural flexibility of TMB compared to that of MDMB. Hybrid E (10.3%) exhibited results that were similar to those of hybrid D (11.5%) with a weight loss post flame test close to 11% with a notable morphological similarity. Interestingly, the pink flame was absent in this flame test. Hybrid F, which exhibited a turbid morphology that was different from that of hybrids D and E, exhibited a weight loss of approximately 16.9% after being subjected to the flame test. The sample was charred to a greater extent than the other hybrids. Possibly, in hybrid F, the structural networking was the weakest, leading to a higher weight loss compared to hybrids D and E. The sequenced images of the flame tests of hybrids D–F are demonstrated in Fig. 5.
Therefore, in general, the LiTFSA-based hybrids exhibited excellent flame resistance properties with weight losses of 10%–17%. All of the mesityldimethoxyborane additive samples (i.e., samples A–C) and hybrids with the trimethoxyborane additive, such as D and F, exhibited a pink flame, which was characteristic of a lithium salt. A minute trace of lithium salt that was encased in the electrolyte matrices might have slowly burned during the progressive decomposition of the organic moiety in the hybrids. In addition, the morphological feature of the hybrids was a significant indicator of the flame retardancy of the hybrids.
Hybrids G–L (LiPF6 with MDMB and TMB precursors)
In hybrids G–I, MDMB was employed as the alkoxyborane precursor. This set exhibited a weight loss in the 15%–21% range during the flammability tests. Hybrid G exhibited a weight loss of more than 16.8% after being subjected to a naked flame for 1 min. In addition, considerable charring was observed on the areas exposed to the naked flame. Hybrid H, which had a turbid matrix, oxidized to 15.5% of its original weight. Hybrid I could sustain a time frame of more than a minute, and the corresponding weight loss was higher (21.1%) than those of the other hybrids in this set. The images shown in Fig. 6 indicate a highly oxidized mass in all of the hybrids. This turbid matrix exhibited a huge loss in weight due to the burning of the excess of organic components included in the initial composition of the hybrid.
Hybrids J–L utilized TMB as the alkoxyborane precursor and exhibited predominantly turbid morphologies. Hybrid J suffered a weight loss of more than 33.5% in the flame test. Hybrid K also demonstrated a weight loss that exceeded 25.6% with considerable charring, as depicted in Fig. 7. Although the considerable loss was only observed on the side exposed to the flames, the weight loss was high compared to that of the other LiTFSA-based hybrids. The last hybrid in the set (i.e., hybrid L) exhibited a weight loss of 16.9%, which was a smaller weight loss compared to those of the other counterparts in the set.
The weight losses of all of the organic-inorganic hybrid samples from the flame test are listed in Table 1.
In general, all of the hybrids irrespective of their components exhibited self-extinguishing properties. The hybrids did not sustain a flame when the flame source was removed and were only oxidized due to the presence of a flame. The borosilicate or silicate matrix along with the ionic liquid constituent provided thermal resistance to the organic-inorganic hybrid materials. To highlight this specific advantage of the organic-inorganic hybrid over other conventional electrolytes, a 1 M LiPF
6-EC: DMC solution was also subjected to the previously mentioned flame test. Another common observation was obtained for the LiPF
6-based hybrids, which underwent a greater loss of weight due to the flame test compared to those of the LiTFSA-based hybrids. However, LiPF
6-based hybrids exhibited a turbid morphological profile, which was characteristic of weaker interactions between the organic and inorganic moieties that led to a weaker structural matrix. Except for hybrids with high alkoxyborane concentrations (e.g., C and F), LiTFSA-based hybrids had transparent morphological features that gave rise to robust mechanical features. Regardless of the alkoxyborane concentration, the effective conversion during the sol-gel condensation process determined the strength of the silicate/borosilicate framework. Further, the interactions between the organic and inorganic moieties determined the structural rigidity of the hybrid framework. At a macro level, a transparent texture was indicative of strong organic-inorganic phase interactions, and weaker interactions between the different phases led to a turbid appearance. Therefore, the appearance of the hybrids was indicative of their internal structural features. Based on the flame test, the order of weight loss was directly correlated with the appearance of the matrix (i.e., transparency indicates a smaller weight loss and turbidity indicates a higher weight loss). Beyond the generalized view of these parameters, localized phenomena would still be present and result in variations in weight loss. To stress this fact, our previously reported results was restated regarding the TGA profiles of the hybrids and their relationship with the morphological features as presented in Fig. 8 [
21]. The interactions between the organic and inorganic moieties were stronger for LiTFSA-based hybrids compared to LiPF
6-based hybrids. Strong interactions resulted in a greater structural rigidity, which, in turn, resulted in a homogeneous morphology and one-step degradation. However, weaker interactions (i.e., for LiPF
6-based hybrids) resulted in a greater degree of phase separation, which caused a dual-layer morphological profile with 2-step degradation pathways in the thermogravimetric experiments. However, the structural and morphological interplay observed in the results from the flame tests indicated the existence of potential localized interactions and their effects, which prompted us to perform differential scanning calorimetric (DSC) analyses.
Differential scanning calorimetric analyses
As an extension to the flammability studies, DSC analyses were performed to study the glass transition temperature (
Tg) of the organic-inorganic hybrids. A
Tg value close to –90°C was observed in the DSC traces of the hybrids, which is indicative of the
Tg of only the ionic liquid [
24]. Typically, the
Tg values of the silicate or borosilicate networks were not observed in the experimental range studied.
Interestingly, all of the TFSA-based hybrids exhibited a second Tg value at –50°C to –70°C. Although the Tg values of samples A and C were similar in the range of –50°C, the Tg value of sample B was approximately –70°C. Due to the coordinative interactions between the cations in the ionic liquid and the bulky TFSA anions (which were in excess because LiTFSA was the lithium salt employed in hybrids A–F), localized ion-pair aggregations were possible, which might lead to additional plasticization. In hybrids A–F, where MDMB was the alkoxyborane precursor, these phenomena were most likely localized within the electrolyte matrix due to sterical hindrance of the bulky mesityl group interactions in the borosilicate/silicate network. However, the TMB-based hybrids with LiTFSA (i.e., D and E) exhibited a Tg value in a range close to –60°C, and hybrid F exhibited an anomalous Tg that was not clearly understood. The DSC plots of hybrids A–F are shown in Fig. 9(a). In contrast, hybrids G–L (Fig. 9(b)) (i.e., LiPF6-based hybrids including trimethoxyborane and mesityldimethoxyborane as alkoxyborane precursors) exhibited Tg values close to –90°C, which corresponded to the Tg value of the pure ionic liquid, as shown in Fig. 9(a). Therefore, this result confirmed the lack of efficient interactions between the ionic liquid and the matrices in these systems. Therefore, an additional Tg value was absent from the data for the hybrids where LiPF6 was used as the lithium salt additive.
Based on these results, localized interactions were present in the LiTFSA-based hybrids based on the additional Tg values. This reinforcement due to additional plasticization ensured a robust internal framework, which was reflected by the smaller loss of weight in the flame tests in contrast to the LiPF6-based hybrids.
X-ray photoelectron spectroscopic studies
Finally, to understand the change in the chemical composition of the hybrid after the flame test, X-ray photoelectron spectroscopy was performed on the samples before and after the flame test. XPS survey spectra of the organic-inorganic hybrid before and after the flame test are provided in Fig. 10. A representative sample of the LiTFSA-based hybrid was used for this study. The results from the XPS of the hybrid are given in Fig. 11.
The survey spectrum and elemental composition of the sample (Table 2) indicated that the percentage of oxygen in the hybrid increased substantially after the flame test. In addition, the atomic percentage of the elements associated decreased after the flame test as expected. The Si high-resolution scan revealed broadening of the Si 2p peaks after the flame test, which might be caused by the formation of crystalline Si-O functionalities that were formed by pyrolysis. In addition, the peak intensity indicated the abundant presence of Si-O after pyrolysis of the hybrid.
Conclusions
All the organic-inorganic hybrids were highly self-extinguishable compared to the sacrificial commercial electrolyte (i.e., 1 M LiPF6-EC: DMC) based on the results of the flame tests. The range of weight loss was ca. 10%–30% in the time period studied. Although the LiPF6-based hybrids exhibited a pronounced loss of weight, the LiTFSA-based hybrids exhibited a relatively stable performance. Although a direct connection between the concentration of the alkoxyborane precursor and the weight loss was not possible, the overall experimental results were consistent with the previously published results of the thermogravimetric parameters based on morphological profiles. The LiTFSA-based hybrids exhibited an additional Tg value in addition to the Tg value that was caused by the localized TFSA anions in the presence of bulky mesityl groups, and this additional Tg value was absent for the other LiPF6-based hybrids, revealing the possibility of localized plasticization due to LiTFSA, which was an additional factor that could enhance the structural properties and benefit the non-flammability of the hybrids.
Higher Education Press and Springer-Verlag GmbH Germany, part of Springer Nature
PDF
(1768KB)
