In situ strain dependent electrochemical characterization of a stretchable-sliding battery

Flexible and stretchable energy storage devices have attracted significant attention in recent years. These multi-functional batteries can offer mechanical and electrochemical functionalities that address a wide range of applications such as smart suits, stretchable electronics and biomedical implants. Among the energy storage devices, lithium ion batteries are most desirable for portable device applications due to their higher energy density and efficiency. As technology is increasingly integrated into our daily lives, the ability to meet the mobile energy demands without a heavy and rigid battery is critical, and thin-film flexible and stretchable batteries offer suitable practical solutions. Moreover, conventional batteries containing flammable organic liquid electrolyte are prone to leakage and thermal runaway that can lead to catastrophic failures.^1–8 Thus, they are less suitable for special applications such as wearable, implantable, and deformable electronics where the mechanical and thermal stability requirements are critical. Solid electrolytes including polymers and ceramics offer more desirable properties such as higher stability and safety. Ceramic electrolytes exhibit high ion conductivities, but are not suitable for stretchable applications.^8–12 In contrast, solid polymer electrolytes are mechanically flexible and stretchable, but provide lower ion conductivities. Several methods have been employed to improve the ionic conductivity of solid polymer electrolytes^7,13–23 including the addition of nanosized fillers, polymer blending, and plasticization.

While there is an increasing amount of work in developing flexible lithium ion batteries,^24–26 research on stretchable batteries is an emerging field. In 2013, Xu et al., developed a stretchable battery consisting of individual cells electrically connected using spring-like stretchable wires.²⁷ The battery exhibits an initial area capacity of 1.1mAh cm^-2 and is capable of achieving an impressive large strain upwards of 300%. The battery experiences rapid capacity fading over the first twenty cycles which may be attributed to the electrolyte and/or the encapsulation. In 2015, Song et al., created a kirigami-inspired stretchable lithium ion battery.²⁸ Based on its folding design, the battery can stretch 150% and exhibits a stable capacity at both stretched and unstretched states up to 20 cycles. However, from cycles 20-100 at its fully stretched state, the battery experiences a 30% capacity drop. The stretching mechanism of these batteries is based on the design. The battery cells are stationary, while the electrical connections are free to deform. In 2016, Kammoun et al.²⁹ fabricated a thin-film stretchable battery based on polymer electrolyte in a spiral configuration capable of large out-of-plane stretching up to 1300%. The battery shows high voltage retention over 9000 stretching cycles, and maintains stable capacity over 100 charge/discharge cycles.

Here, we introduce a novel stretchable-sliding battery design where the electrodes can slide and the solid polymer electrolyte film can stretch, as shown in Figure 1. The battery contains a single polymer electrolyte film sandwiched between two sliding layered electrodes on each side (Figures 1a–b). The layered electrode configuration allows the active material surfaces to slide and increase the exposed area as the electrolyte film is stretched (Figures 1c–d). Additionally, the conductive nature of the electrodes should guarantee a non-disrupted electric path across the layered electrode configuration. Figure 1e shows the scanning electron microscopy (SEM) cross-sectional image of the layered stretchable-sliding battery. The solid polymer electrolyte in this battery is based on a blend of relatively high (600k) and low (100k) molecular weight polyethylene oxide (PEO). To develop an electrolyte material compatible with stretchable battery application, molecular weight blending was investigated with respect to its mechanical and electrochemical properties. The electrochemical testing of the stretchable-sliding battery shows capacity enhancement as the battery undergoes tensile strain.

Figure 2 shows the ionic conductivities and mechanical properties of 100k and 600k molecular weights PEO solid electrolyte at various blending contents. The ion conduction in solid polymer electrolytes is mainly facilitated by the thermally driven chain segmental motion of the polymer host, which can depend on the molecular weight of the polymer matrix.^30–41 Results in this study show an order of magnitude difference in ionic conductivity of 600k and 100k PEO electrolyte, which agrees well with previously reported values.^7,37 Higher ionic conductivity in 100k PEO generally stems from higher degree of amorphicity and chain motion in the lower molecular weight polymer.

The mechanical properties of the electrolyte film were determined to assess its compatibility with flexible and stretchable batteries. The molecular weight blends of the polymer films contain 0%, 12.5%, 25%, 37.5%, and 50% 100k PEO with the remaining PEO content consisting of 600k. Figures 2a–d compare the elastic modulus, yield strength, percent elongation, and the ultimate tensile strength of the PEO molecular weight blended electrolyte films. The strength of the molecular weight blends decreases with the increasing content of 100k PEO. Higher molecular weight polymers tend to be stronger and mechanically more stable due to the longer polymer chains.

The effect of graphene oxide (GO) nanosheets on the mechanical and electrochemical properties of the molecular weight polymer blends was also investigated. The addition of GO nanosheets to the PEO has been shown to enhance both mechanical and electrochemical properties.^7,39 In this study, we decided to add 1% GO as this composition established optimal improvement as shown by Yuan et al.⁷  Figures 2a–d compare the mechanical properties of the 1 wt.% GO filled and unfilled PEO blends, showing a general enhancement of the elastic modulus, yield strength, and ultimate tensile strength in the nanocomposite blend samples. However, the degree of elongation is found to be lower at the higher 600k Mw content with the addition of 1wt.% GO, and increases as the 600k Mw content decreases. This suggests that the addition of GO at high content 600k Mw has restrictive effects on the polymer elongation. The mechanical properties of both the filled and unfilled PEO molecular weight polymer blends are summarized in Table S1. In addition to the mechanical characterization, a thermo-gravimetric analysis (TGA) of the PEO molecular weight blends (Figure S1) shows the 25% blend PEO exhibiting the highest thermal decomposition temperature of 378 °C.

The ion conductivities of the Mw blended PEO samples are found to increase with the weight percentage of 100k PEO (Figure 2e). Lower molecular weight polymers exhibit higher ion conductivities due to shorter molecular chain lengths and increased chain mobility. The effect of GO nanosheets on ionic conductivity of the Mw blended PEO appears to increase with higher 100k Mw PEO content, as shown in Figure 2e, and is most pronounced in the pure 100k Mw PEO. Nanofillers can be added to the polymer electrolytes to open tangled chain conformations and increase free volume to better facilitate ion transport.¹⁶ Previous studies have reported reduced glass transition temperature of polymer electrolyte upon the addition of nanofillers generally indicative of increased free volume and improved chain mobility.^7,39 Scanning electron microscopy (SEM) was used to qualitatively evaluate the polymer electrolyte structure. The images in Figure 2f show the PEO crystals, changing from large to small sizes, as the content of 600k Mw PEO decreases. The pure 600k Mw sample contains large and tightly packed polymer crystallites with clearly defined crystal edges, whereas the pure 100k Mw sample contains more loosely packed crystallites with no clearly defined edges.

In the previous work by Ardebili and co-workers,³⁷ an in situ measurement technique was developed to determine the ion conductivity of the polymer electrolyte film under tensile deformation. In this study, a similar in situ ionic conductivity test was performed on 25% 100k Mw PEO electrolyte film (Figure 3). The normalized ionic conductivity with respect to σ₀ (ion conductivity of the unstrained polymer) is plotted versus the tensile strain % in Figure 3a. The slope of the linear trend in this plot is referred to as the coefficient of strain-dependent ionic conductivity enhancement (CSDICE) and it represents the rate of ionic conductivity change with strain. For the 25% blend, the average fitted CSDICE value is found to be about 28. This data confirms that the ion conductivity of the solid polymer electrolyte increases with tensile strain. This increase is attributed mainly to the disentanglement of the polymer chains during stretching, and the consequent reduction in tortuosity and increase in chain segmental mobility that can lead to higher ion conduction in the polymer.¹⁶ An in-depth theoretical analysis of the strain-dependent ion conductivity phenomenon has been provided by Berg et al.³⁸ based on multi-physics equations derived from the thermodynamic laws showing dependency of the chemical potential on stress and concentration gradients within the polymer electrolyte.

The 25% 100k Mw PEO blended electrolyte film was introduced into a stretchable-sliding battery with the sliding electrode design as illustrated in Figure 1. In this design, two conventional LiFePO₄ electrodes are layered with a slight offset to make the battery cathode. By layering the electrodes, they can slide, and the active material surface can expand as the battery is strained, demonstrated in Figure 1c. The battery anode is configured in a similar fashion, consisting of two conventional graphite electrodes. The entire cell was encapsulated in VHB (3M) adhesive tape. With this sliding electrode design, the polymer electrolyte film can stretch as the battery is subjected to tensile loading. Additionally, due to the elasticity and the adhesive nature of VHB encapsulation, upon tension release, the electrodes can return to their initial positions, if the entire battery composite is still in elastic region, or otherwise to the most relaxed state.

The electrochemical performance and impedance spectra of the stretchable-sliding battery are shown in Figures 4a–c, evaluated at different strains inside the dry glove box. The battery was placed in a horizontal tensile machine as depicted in Figures 4d–e, where it was charged and discharged at 0.1 mA for 20 cycles followed by cyclic voltammetry and impedance spectroscopy at strain increments of 0%, 2%, 5%, and 10%. In the unstretched state, the battery exhibits an initial capacity of 0.1 mAh cm^-2, which steadily rises to 0.18 mAh cm^-2 during the first 20 charge-discharge cycles, seen in Figure 4a. The capacity enhancement in the first 20 cycles of the unstrained battery can be attributed to the increased availability of activation sites caused by the deformation and volume changes of the electrode active materials as they are repeatedly lithiated and delithiated.⁴¹ The battery is then subjected to tensile deformation. The area capacity (mAh cm^-2) of the battery is re-calculated for each tensile strain % using the increased area of the strained battery. After stretching, the battery exhibits a noticeable initial increase in capacity with each strain increment, as the polymer electrolyte is strained elastically to 2% and 5%. The initial rise in capacity with tensile strain can be attributed to a combination of factors, including (i) increased ionic conductivity of the stretched polymer electrolyte, (ii) reduced thickness of the electrolyte during stretching, (iii) increased exposed active electrode surface, and (iv) interface passivation disruption during sliding/ stretching.

After the initial capacity increase with small strains (2% and 5%), capacity fading is observed with larger strain (10%), observed in Figure 4a. This indicates degradation of the battery active components during stretching and sliding. It is noted that the columbic efficiency of the battery in the unstrained state (zero strain) is relatively lower than expected as shown in Figure 4a. The initial lower columbic efficiency of the stretchable-sliding battery can be attributed to the layered electrode design and the low-pressure flexible/stretchable encapsulation. Several specific factors can be identified including the (i) low accessibility of the upper electrode layer due to the overlap of the sliding electrodes, (ii) lower ionic and electrical conduction in the overlapped electrode layer, and (iii) low interface contact pressure between the sliding and stretching layers in the dynamic battery cell as compared to the conventional high-pressure packaged batteries such as coin cells. As the battery is stretched, the columbic efficiency improves. This suggests that the tensile stress and strain of the stretchable battery counters some or all of the aforementioned adverse effects of the unstrained state on columbic efficiency.

The battery cyclic voltammetry (CV) and impedance spectroscopy are also provided in Figures 4b and 4c, respectively. The CV curves shown in Figure 4b suggest that the electrochemical performance of the battery slightly degrades with 20 cycles at each strain, especially at larger strains (10%). Also, the battery shows increased impedance after 20 charge-discharge cycles per each strain increment, as shown in Figure 4c indicating possible electrode and interfacial degradations due to both cycling and sliding mechanisms. The reversibility of the electrochemical performance under elastic tensile strain was also investigated (Figure S2). The total discharge capacity of the stretchable-sliding battery was determined at various states: initially under no tension (no strain), then stretched, followed by the removal of the tensile strain. The results show that the electrochemical performance of the sliding battery is reversible, provided that the strain to the entire battery is within its elastic limit, which is mainly governed by the elastic limit of the polymer electrolyte.

The mechanical testing of the stretchable-sliding battery and its components are shown in Figure 5. The stretchable-sliding battery is composed of individual layers, each with different mechanical properties. The electrode is made of a current collector metallic foil with a thin composite active layer on the surface; the electrolyte is a viscoelastic solid polymer, and the encapsulation is an elastic adhesive. Figures 5a–b show the stress-strain curves of the stretchable-sliding battery and the 25%100k Mw PEO electrolyte. As the battery is subjected to tensile deformation, the polymer electrolyte and VHB encapsulation begin to stretch while the electrodes slide. At about 6% tensile strain in the battery, a sudden sharp drop in the stress is observed, which indicates potential slipping mechanism of the electrodes. At this point, the electrodes are no longer overlapping, as it can be observed in Figure 5c photo image marked A. Point B on the battery stress-strain curve denotes the complete tearing failure of the polymer electrolyte. Partial tearing of the VHB encapsulation in the battery occurs at B and C, followed by complete failure at D. The first rise in the battery stress-strain curve produces an elastic modulus of 15.1 MPa, yield strength of 0.443 MPa, and an ultimate tensile strength of 0.694 MPa. These values are close to those of the 25% 100k Mw PEO electrolyte, 13.4 MPa, 0.574 MPa, and 0.807 MPa, respectively, and support the assumption that the polymer electrolyte is carrying the main tensile load in the battery.

Figure 5d presents the tensile test stress-strain curve for the VHB material and Figure 5e shows the photo images of the VHB under tensile testing. The elastic modulus, ultimate tensile strength, and % elongation of the VHB material are found to be 0.023 MPa, 0.303 MPa, and 794 respectively. Figure 5f shows the peeling stress-strain curve from the T-peel test, where the adhesion strength is found to be 8.19 kPa. Figure 5g presents the corresponding photo images of the VHB adhesion test.

A stretchable-sliding lithium ion battery was fabricated using a molecular weight blended solid polymer electrolyte, and LiFePO₄ and graphite based sliding electrodes. The 25% polyethylene oxide (PEO) blend was the chosen electrolyte for the stretchable battery for its balance of mechanical and electrochemical properties. The stretchable battery was cycled at 0%, 2%, 5%, and 10% strain increments and exhibits a near 30% capacity enhancement at lower strains. The coulombic efficiency was improved upon stretching, which can be attributed to better ionic conduction across the battery, though the contact pressure among the components remained low due to the packaging nature. Moreover, the battery shows reversibility of electrochemical performance as it is stretched and unstretched within the elastic region of the polymer electrolyte. The tensile stress-strain plot of the entire stretchable battery until mechanical failure provides insight into the mechanical behavior of the layered components. Material properties can be fine-tuned further in future studies to obtain optimum electrochemical and mechanical performance for specific application requirements. Also, further characterizations of electrode-electrolyte interface properties can offer more insight into the sliding and stretching process. This stretchable-sliding battery design provides a platform for in situ characterizations of solid polymer electrolyte films subjected to tensile deformation inside an active electrochemical cell.

The thin-film polyethylene oxide electrolyte samples were prepared using a solution-cast method. 3g of PEO (Aldrich), with varying contents of 100k and 600k molecular weights, and 0.45g of LiClO₄ (99.99%, Aldrich) were dispersed in anhydrous acetonitrile (99.9%, Sigma-Aldrich). The nanocomposite PEO films included 0.042g of graphene oxide (GO) powder (Graphene Supermarket) with layer dimensions of approximately 0.5-5 microns in length and width and 1.1±0.2nm in thickness. The electrolyte solutions were stirred at room temperature for 24 hours, then dried at 50 °C under a slight vacuum. The films were then transferred to the argon-filled glove box and allowed to dry for an additional 96 hours prior to use to ensure solvent removal from the films.

The molecular weight blend electrolyte samples were loaded into a MARK-10 ESM301L motorized test stand in ambient room conditions and strained at 10 mm/min. The yield strength, ultimate tensile strength, and Young’s modulus were determined for each molecular weight blend. The ion conductivities of the electrolyte films were calculated using the obtained bulk resistance values via electrochemical impedance spectroscopy. The polymer samples were placed between two stainless steel electrodes connected to an Autolab multichannel potentiostat fitted with an FRA module. The bulk resistances of the samples were measured in the range of 1MHz-10Hz.

The in situ impedance spectroscopy experimental setup is illustrated in Figure 3. The polymer electrolyte sample was loaded into the MARK-10 ESM301L motorized test stand in ambient room conditions. Aluminum foil electrodes were placed on both sides of the electrolyte and clamped together using electrically insulating plastic pins in order to minimize the contact resistance between the electrode and the electrolyte. The aluminum foil electrodes were constructed in our lab by adhering a layer of aluminum foil to wooden columns using a double-sided adhesive. Electrical tape was then used to mark off two identically sized aluminum foil rectangles. The aluminum surfaces acted as the electrodes. Copper leads connected the aluminum foil electrodes to the Autolab potentiostat for impedance spectroscopy. The electrolyte samples were stretched in 0.5mm increments and impedance spectroscopy was performed at each increment in order to quantify the effect of strain on through-thickness ionic conductivity.

A thermogravimetric analysis (TGA) of the 0%, 25%, 100%, and 100%+GO electrolyte samples was performed using TA Instruments Model Q50 TGA under nitrogen atmosphere. Weight was measured as the samples were heated from room temperature to 750 °C. Further discussion of the TGA is provided in the supplementary material.

The battery is composed of two LiFePO₄ electrodes and two graphite electrodes separated by a thin PEO-based electrolyte film, as seen in Figure 1. Each electrode measures 1cm in width and 2cm in length. A small amount of liquid electrolyte (10 wt. %) is added to each electrode to enhance the electrode-electrolyte interface contact. Copper tape is used as electrical leads to connect the battery to the Autolab multichannel potentiostat battery tester. The battery is encapsulated in a VHB double-sided adhesive tape manufactured by 3M. The encapsulated battery is then placed in a laminating sheet and run through the laminator twice. This process further improves the electrode-electrolyte interface contact.

The encapsulated stretchable-sliding battery is placed in a manual, horizontal tensile test stand (MARK-10) in the glove box and compressed using two clips. Electrical wires connected the copper leads to the Autolab multichannel potentiostat battery tester. The battery is charged and discharged between 2V - 4.2V at a constant current of 0.1 mA. At each strain increment, the battery undergoes 20 charge/discharge cycles and its capacity and columbic efficiencies are evaluated. The area capacity (capacity divided by the area) is based on the re-calculated area at each strain. Cyclic voltammetry and impedance spectroscopy of the battery are conducted at the end of 20 cycles per each strain increment. The impedance spectroscopy of the battery is performed using the Frequency Response Analyzer (FRA) module of the Autolab potentiostat with frequency ranging from 1MHz-10Hz.

The battery is fabricated following the previously mentioned procedure. The copper electrical leads were replaced by nickel and aluminum tabs for negative and positive terminals, respectively. This replacement was required for Aluminum encapsulation which sealed the package under 100 kPa pressure, which effectively omitted the lamination step. The Al-encapsulated battery was left to rest inside the glovebox overnight. Prior to electrochemical evaluation, the Al encapsulation is peeled off. Then, the battery is mounted onto MARK-10 stretching apparatus in the same manner as above, and compressed using two clips. Electrochemical characterization is carried out at the same conditions. The total discharge capacity is allowed to reach steady state to establish a baseline value. The battery is slowly stretched to 1.4% strain before the clips are reapplied. After the discharge capacity stabilizes, the battery is slowly pushed back to the original state. The discharge capacity is monitored until it reaches the baseline value.

The stretchable-sliding battery was loaded into a MARK-10 ESM301L motorized test stand in ambient room conditions. The stress-strain curves were obtained as the battery was mechanically tested under tension until failure. The elastic modulus, ultimate tensile strength, and % elongation were obtained for both the battery and the VHB material. T-peel test (ASTM D1876) was also performed to determine the adhesion strength.

The supplementary material includes a table of mechanical properties of the polymer electrolytes of different blends, thermogravimetric analysis plot, and the reversibility of discharge capacity upon strain removal in the sliding-stretchable battery.

We gratefully acknowledge the National Science Foundation (NSF CAREER: CMMI-1254477) for funding this research. We also thank TcSUH for the financial support.