Multifunctional structural lithium-ion battery for electric vehicles

Abstract

An innovative concept for a multifunctional structural battery using lithium-ion battery materials as load bearing elements in a sandwich panel construction has been demonstrated. The structural battery prototype has exhibited an initial capacity of 17.85 Ah, an energy density of 248 Wh L⁻¹, a specific energy of 102 Wh kg⁻¹, and a capacity retention of 85.8% after 190 charge–discharge cycles at ~C/3 rate and eight mechanical loading cycles (upto 1060 N). The mechanical stiffness in three-point bend tests follows expectations based on sandwich beam theory, proving that the battery materials are sharing in the load-carrying function of the sandwich panel. While areas for improvement of the fabrication and performance of the prototype still exist, the results of the current investigation demonstrate the promising potential of the proposed structural battery concept for the efficient use of space and mass in an electric vehicle.

Keywords

Multifunctional structural battery lithium-ion battery sandwich panel electric vehicles

Introduction

Multifunctional structural batteries are capable of storing energy while fulfilling a structural role in various applications. Some target applications include satellites (Roberts and Aglietti, 2008; Wang et al., 2014), spacecraft (Roberts and Aglietti, 2010), unmanned air vehicles (Thomas and Qidwai, 2005), and marine systems (Thomas et al., 2013). Lithium-ion batteries are commonly used in such applications by virtue of their high energy density, long cycle life, and environmental friendliness with zero emissions. Previous investigations have used two general approaches for developing structural batteries. One approach incorporates conventional lithium-ion batteries into composite structures (Pereira et al., 2008, 2009; Roberts and Aglietti, 2008, 2010; Thomas et al., 2013; Thomas and Qidwai, 2005; Wang et al., 2014), while the second develops structural electrodes (Ekstedt et al., 2010; Evanoff et al., 2012; Jacques et al., 2012; Kim and Sastry, 2012; Liu et al., 2009; Mullenax et al., 2012), structural separators (Liu et al., 2009; Mullenax et al., 2012), and even structural electrolyte (Ding et al., 2013; Snyder et al., 2009; Willgert et al., 2013) into lithium-ion batteries. These investigations have not produced a satisfactory combination of high energy storage (comparable to a conventional lithium-ion battery) and sufficient mechanical performance. Their energy deficiency is usually an unwanted compromise to incorporate additional functionalities (Liu et al., 2009; Mullenax et al., 2012; Pereira et al., 2008, 2009; Roberts and Aglietti, 2008, 2010; Thomas et al., 2013; Vlad et al., 2015; Wang et al., 2014). However, both energy density and structural performance are needed for structural batteries in electric vehicles (EVs).

The driving range of EVs is very sensitive to the mass of the vehicle. The densest components in a battery, the metal current collectors, are dead weight from a structural perspective. The current collectors are made from excellent structural materials, however, and if they can be used as load bearing elements without reducing the electrochemical performance, significant mass reduction is possible. This is the first battery design to use the current collectors as structural elements in a multifunctional composite.

The innovation embodied in this research is to combine the mechanically efficient attributes of sandwich panels with the electrochemically efficient attributes of state-of-the-art lithium-ion batteries, as disclosed in our recent patent and publication (Rahn et al., 2014; Singh et al., 2015). The mechanical efficiency of sandwich panels results from bonding a thin pair of stiff, strong face skins (such as aluminum) to a lightweight, thick core material (such as foam), as shown in Figure 1(a). The core separates the skins and minimizes transverse shear deformation, resulting in a structure with high bending strength and high rigidity per unit mass. In our sandwich structural battery, transversely stacked battery materials serve as the core, as illustrated in Figure 1(b), and must be bonded to the top and bottom face skins to behave as a single solid composite structure. Thus, the face skins carry the in-plane and bending loads, while the battery core carries the transverse shear and compression loads in addition to storing energy. Hermetic seals in the form of aluminum C-channels in the longitudinal direction and polyetherimide (PEI) end plugs in the transverse direction share in some of the structural load-carrying capacity while containing the liquid electrolyte. Consequently, the energy storage components (i.e. the battery core) have a structural contribution, and the structural components (e.g. face skins and C-channels) are also utilized for the battery enclosure. On a system level, this design has the potential to reduce weight and volume and therefore augments the energy density, specific energy, and structural performance of EVs.

Fabrication of the proposed structural battery presents significant challenges that require innovative design features and material selection. The lithium-ion battery core must have state-of-the-art performance, and the various battery and structural components must be bonded together with high mechanical, electrochemical, and hermetic integrity. The battery anode and cathode must be electrically insulated from each other when they are bonded to the face skins, and the adhesive must be resistant to the battery electrolyte.

This article outlines the approach taken to address these challenges and presents the electrochemical and mechanical performance of a prototype structural battery. This work builds on the previous descriptions of the concept (Rahn et al., 2014) and a paper that focused on the mechanical design, fabrication, and testing of the battery (Singh et al., 2015). The expanded scope of the characterization effort in this work includes adhesive testing (i.e. electrolyte uptake and lap shear strength), electrochemical tests (i.e. charge/discharge cycling and electrochemical impedance spectra (EIS)), and mechanical bending tests.

Fabrication of sandwich structural battery

Fabrication of Li-ion battery core

Commercial state-of-the-art anode and cathode materials were used to fabricate the Li-ion battery core. Mesocarbon microbeads (MCMB) was used for the anode coating with a sodium salt of carboxymethyl cellulose (CMC) and styrene-butadiene rubber (SBR) as the water-based binder. Super P Li carbon black (CB) was used as a conductive additive. The double-sided MCMB loading was typically 21.3 mg cm⁻². The anode coating on both sides of Cu foil (18 µm thick) consists of 96.5 wt% MCMB, 0.5 wt% CB, 1.0 wt% CMC, and 2.0 wt% SBR. Polyvinylidene fluoride (PVDF) dissolved in N-methyl-2-pyrrolidone (NMP) was the binder for the NCM (LiNi_1/3Co_1/3Mn_1/3O₂) cathode coating. The weight percentages of NCM, CB, and PVDF were 94%, 3%, and 3%, respectively. The cathode slurry was coated on both sides of Al foil (20 µm thick), and the double-sided NCM loading was typically 44.1 mg cm⁻². After pressing, the anode and cathode were typically 164 and 180 µm thick. Subsequently, the anode and cathode were slit into 15- and 11-mm-wide strips, respectively. The anodes strips were slightly wider because the anode edges are prone to lithium plating. In order to prevent lithium plating, the cell design makes the anode larger than the cathode in capacity as well as in dimensions and in area (Bugga and Smart, 2010).

A separator pouch for the cathode strip was designed to protect the cathode from shorting with the anode and face skins. Figure 2 shows a schematic cross-section of a cathode strip in a separator pouch stacked with an anode strip. Two pieces of Celgard 2320 separators (20 µm thick and 15 mm wide) were heat sealed together on both edges using an Impulse Sealer (H-963 from Uline, USA). Each seal was slightly less than 2 mm wide so that the 11-mm-wide cathode strip could be easily slid into the separator pouch. One end of the separator pouch was also heat sealed, while the other end was left open for insertion of the cathode strip and welding of the current collector to a tab. Equal widths (15 mm) of the separator pouch and anode strip enabled easy stacking with proper registration, as depicted in Figure 2.

The anode strips and pouched cathode strips were stacked alternatively to form the battery stack. Subsequently, the cathode Al current collectors were welded to Al tabs and the anode Cu current collectors were welded to Ni tabs. The Al tabs were welded to an Al bus bar welded to an end of a threaded Al rod as the cathode terminal. The Ni tabs were welded to a Cu bus bar welded to an end of a threaded Cu rod as the anode terminal. Electric connections from current collectors to battery terminals are critically important to achieve minimum contact resistance (e.g. milliohm level) and headspace (e.g. mm level). An ultrasonic welder (MWX1000 from Branson Ultrasonics Corporation, USA) and a resistance welder (HNW-101 from Hohsen Corporation, Japan) were both used to produce satisfactory welding quality.

Fabrication of sandwich structure

The sandwich structure with battery core is illustrated in a cross-sectional view in Figure 3(a). The battery coating layers were stacked transversely in the battery core region. Two longitudinal C-channels confined the battery core. Face skins (1 mm thick) were adhesively bonded to the top and bottom of the battery core. Both the C-channel and the face skin were made of 2024 alloy aluminum. Cross-sectional dimensions of the C-channel are shown in Figure 3(b).

After the aluminum oxide on the C-channels was removed by sandpaper, the C-channels were cleaned with acetone and coated with a thin layer of LoctiteE-120HP Hysol epoxy (Henkel, USA) in order to electrically insulate them from the battery core and face skins. The epoxy was cured at room temperature for 12 h and then at 65°C for 12 h. Subsequently, the epoxy surface on the C-channels was gently abraded with sandpaper and cleaned with acetone. Afterwards, the C-channels were placed in a steel jig to hold the battery core, as shown in Figure 3(c). The jig was clamped to apply a light pressure around 0.5 atm (Cannarella and Arnold, 2014) onto the battery core.

The end plugs (labeled in Figure 3(c)) were made of PEI to insulate the electrical terminals from the other components. Three threaded National Pipe Thread Taper (NPT) holes (1/16 in pipe size) were machined in each end plug for an electrical terminal and two electrolyte filling ports. The NPT threads were tapered to make a fluid-tight seal and to prevent any electrolyte leakage. Each end plug had dimensions of 50 mm ×15 mm × 6.35 mm. The 50-mm dimension matched the total thickness of stacking layers (37.3 mm) and two C-channels (6.35 mm each). The height of 15 mm was same as the battery layers and C-channels.

A layer of E-120HP epoxy (areal weight of 558 g m⁻²) was applied onto the top side of battery core region,C-channels, and end plugs. Then, a layer of Optiveil™ (Technical Fiber Products, USA) E-glass veil of 10 g m⁻² areal weight was placed on top of the epoxy in order to provide uniform epoxy thickness and electrical insulation. A face skin (309 mm × 50 mm × 1 mm) was pressed onto the E-glass veil and held in position during the epoxy curing for 12 h at room temperature and another 12 h at 65°C. Similarly, the second face skin was bonded to the opposite side of the sandwich. Immediately prior to bonding the face skins, the bonding surface of each face skin was abraded with sandpaper to remove aluminum oxide and, subsequently, cleaned with acetone. The epoxy adhesive layers between the battery core and both face skins were 0.9 mm thick, in total. Accordingly, the sandwich structural battery (Figure 3(d)) was 17.9 mm high (i.e. 15 mm height from the battery core, 2 mm from the two face skins, and 0.9 mm from the adhesive layers). The overall structural battery dimensions were 309 mm× 50 mm × 17.9 mm with the geometric volume of 277 mL. After excluding the hollow volume (8.5 mL each) in the two C-channels, the battery volume was 260 mL.

The parameters of the fabricated structural batteries are designed to achieve a balance between mechanical performance (bending stiffness) and electrochemical performance (specific energy and energy density). To store a given amount of energy, it is easier to make a smaller number of wide electrodes than make a larger number of narrow electrodes, due to the labor-intensive and potentially damaging fabrication steps of slitting, pouching, and stacking. Wide electrodes have the potential for shear wrinkling, but this is mitigated by the close contact of the electrodes and separators and the slight compression applied prior to bonding. The design with the highest energy density would have thin side walls, possibly made from pouch material. The C-channel and face skin dimensions for the prototypes, however, are chosen from readily available materials to ensure that the overall stiffness requirement is met.

Electrolyte filling

Prior to electrolyte filling, good sealing along the full perimeter of epoxy bond lines between face skins, C-channels, and end plugs was ensured. If there was any evidence of electrolyte leakage through these bond lines, additional E-120HP epoxy was applied and cured. Subsequently, the structural battery was dried in a vacuum oven at 60°C for at least 1 day and transported into an argon-filled glove box with low moisture and oxygen levels (typically below 1 ppm) for electrolyte filling.

The electrolyte was 1M LiPF₆ in ethylene carbonate (EC)/dimethyl carbonate (DMC)/diethyl carbonate (DEC, 1:1:1 by volume). The electrolyte was introduced to the battery core through filling ports in the end plugs. The electrolyte filling ports were then sealed except for one port which was fitted with a pressure release valve through which electrolyte and gas/vapor could leave the battery core during formation. The battery was stored in the glove box for 1 day to allow for sufficient electrolyte saturation into the electrode coatings before the formation process. The formation was done in an explosion-proof chamber. After the formation process was completed, the pressure release valve was removed and the port was sealed in the glove box. The final mass of the structural battery after formation was 632.2 g.

Experimental section

Adhesive testing

In order to examine the durability of E-120HP epoxy in the battery electrolyte, a conventional gravimetric swelling test was conducted. Four E-120HP epoxy samples (each ∅2 cm with a thickness of 0.2 cm) were cured at room temperature for 48 h and then immersed in the battery electrolyte: 1M LiPF₆ in EC/DMC/DEC (1:1:1 by volume) in the glove box. After 3 months of soaking at room temperature, the mass of the four epoxy samples was measured and the electrolyte uptake was calculated as the mass difference between before and after soaking.

The bonding strength of the E-120HP epoxy with a 2024 aluminum sheet was evaluated by a lap shear strength test. Each adherend was an aluminum sheet of 63.5 mm long, 12.7 mm wide, and 1.0 mm thick. In the shear area (12.7 mm × 12.7 mm), the adhesive thickness was 0.3 mm and was controlled by inserting steel wires of 0.3 mm diameter between the two adherends. The grip area on each adherend was 25.4 mm×12.7 mm. After E-120HP epoxy was applied to the adherends that had been cleaned and cured at room temperature for 48 h, five lap shear specimens were soaked in the battery electrolyte at room temperature in the glove box. After 3 months of soaking, the lap shear strength test was performed using a crosshead speed of 1.27 mm min⁻¹. Also, five lap shear specimens were tested without soaking in the battery electrolyte.

Electrochemical testing

The sandwich structural battery was placed in an explosion-proof chamber and tested at room temperature using an Arbin BT2000 battery tester. Charge and discharge cycles of the battery were performed at a roughly C/3 rate (6.0 A): charging at a constant current (CC) of 6.0 A to 4.2 V followed by a constant voltage (CV) at 4.2 V until a cut-off current of 0.8 A; 30 min rest; discharging at a CC of 6.0 A to 2.5 V; and 30 min rest. After certain cycles, the cycling was stopped; EIS and mechanical testing were conducted before the cycling continued. EIS were measured at 50% depth of discharge (DoD). The battery was discharged at 6.0 A to 50% DoD after being fully charged. The open-circuit voltage (OCV) after 1 h rest at 50% DoD was applied as a DC bias through a Solartron SI 1287 electrochemical interface when impedance spectra were obtained over a frequency range from 50 kHz to 0.005 Hz using a Solartron SI 1255B frequency response analyzer (FRA) with a small amplitude of 5 mV. The EIS after cycles 2, 50, 52, 100, and 150 were measured to examine the impedance change associated with battery degradation.

Mechanical testing

In order to explore the load capacity as well as deflection without structural or electrochemical failure, the structural battery was tested in three-point bend tests according to ASTM D7250/D7250M-06 (2012) (ASTM Standards, 2012). The test setup is shown in Figure 4. Two cylindrical rollers in the steel fixture were two support points underneath the battery and there was an overhang of 25.4 mm beyond each support point. The span length therefore between the two support points was 258 mm. The servo-hydraulic testing machine had a 13-kN load cell and applied loads at the mid-span position of the battery with a crosshead speed of 0.3 mm min⁻¹. The mid-span deflection was recorded using a potentiometer with a sensitivity of 1.27 mm V⁻¹. The bending stiffness, defined as the applied load per unit mid-span deflection, was obtained from the slope of relatively linear higher load segment in the load versus mid-span displacement plots. An electrolyte leak trap made of aluminum foil was placed under the bend test fixture in case of any electrolyte leakage, but no leakage was observed during the bend tests. The three-point bend tests were performed before electrolyte filling (dry tests) and upon completion of 50 charge–discharge cycles (wet tests). Two maximum load levels (slightly above 750 N and 1000 N) were applied for both the dry and wet tests. Each test was loaded twice to look into the repeatability. Accordingly, the structural battery went through totally eight loading cycles in the three-point bend tests.

Results and discussions

Adhesive

The use of a high-performance adhesive in the sandwich structure is critical to maintain the structural integrity and hermetically enclose the battery core. A number of commercial adhesives were screened in terms of resistance to the battery electrolyte, electrical insulation, cure temperature, electrolyte uptake, and bonding strength, and finally E-120HP epoxy was selected for the fabrication of the sandwich structural battery (Singh et al., 2015).

When the adhesive is used to bond the battery core to the face skins, the adhesive must be cured at a temperature that does not thermally damage any component of the battery core, especially the separator. The E-120HP epoxy can be cured at room temperature for 48 h or for 12 h followed by another 12 h at 65°C. Such cure temperature is safe for the Celgard 2320 separator used that has a shrinkage onset temperature of ∼105°C (Arora and Zhang, 2004).

Four cured E-120HP epoxy disk samples were soaked in the battery electrolyte for 3 months at room temperature in the glove box. Their average mass before and after 3 months of soaking was, respectively, 0.98 and 1.28 g. The electrolyte uptake rate was 31% ± 5%. This uptake is high but tolerable (around 7% of the electrolyte amount used in the structural battery since the epoxy amount used was 24% of the electrolyte amount). The reason for this phenomenon is that the Li⁺ solvation energy with EC is much higher than that with DMC. Therefore, the solvation may decrease the affinity between EC and DMC, thus allowing DMC to penetrate into the epoxy structure (Seo, 2016).

A lap shear strength test was performed on the five lap shear specimens of the E-120HP epoxy with 2024 aluminum sheet after soaking in the battery electrolyte for 3 months in the glove box. The shear strength prior to shear failure is obtained from the shear stress–shear strain curves, and the average shear strength is 11.5 MPa with a wide range of 6.7–19.3 MPa, as illustrated in Figure 5. The wide strength range is attributed to variability in bonding surface preparation and electrolyte uptake (by both diffusion to the epoxy and migration to the epoxy–adherend interface). The shear strength of the five lap shear specimens without soaking in the battery electrolyte (pristine specimens) is 16.6 MPa on average with a range between 12.9 and 23.8 MPa. Accordingly, the average lap shear strength degrades 31% after exposure to the battery electrolyte for 3 months. Nevertheless, such significant degradation did not cause any structural damage or leakage during the charge–discharge cycling of structural battery. Based on the lowest of the measured shear strength after soaking (i.e. 6.7 MPa) and sandwich beam theory, adhesive failure in the current configuration of structural battery in three-point bending can be expected at approximately 3300 N. The results show that E-120HP epoxy can fulfill the bonding strength requirement in the sandwich structural battery.

Cycling performance

Charge and discharge curves at approximately C/3 rate (6.0 A) for various cycles at room temperature are shown in Figure 6(a). The discharge capacity decreases quickly in the first 50 cycles (8.9% capacity fade), and the capacity fade slows in the second (2.3%) and third (1.3%) 50 cycles. The initial discharge capacity is 17.85 Ah and defined as the 100% capacity. The capacity retention is 91.1% after 50 cycles, 88.8% after 100 cycles, and 87.5% after 150 cycles (see Figure 6(b)). After 190 cycles, the capacity retention is 85.8%. This sandwich structural battery demonstrates good cycling performance considering its innovative design and hand fabrication. The cycling performance is expected to improve if the design and fabrication are further optimized and automated.

As shown in Figure 6(b), the charge and discharge capacity increases from cycle 50 to cycle 51 when the battery was rested for 42 h to complete the mechanical testing prior to cycle 51. A similar increase was observed when the battery was rested for a long time (from 24 to 40 h) after cycles 100, 150, and 170. Consequently, the increase was most likely caused by a relaxation effect during the rest period rather than from the influence of mechanical testing. Capacity increase after a rest period of 2 h has also been experimentally observed in 18,650 commercial Li-ion cells (Reichert et al., 2013). Sufficiently, long rest in a fully discharged state relaxes the gradients of lithium-ion concentration and potential across the cell and results in better utilization of cyclable lithium in the subsequent charging cycle (Rashid and Gupta, 2015).

Figure 6(c) presents the charge and discharge energies (measured at ∼C/3 rate) versus cycle number. The initial discharge energy is 64.5 Wh, resulting in an energy density of 248 Wh L⁻¹ and a specific energy of 102 Wh kg⁻¹. The specific energy is lower than state-of-the-art lithium-ion batteries with a specific energy of ∼150 Wh kg⁻¹ (Tarascon and Armand, 2001; Thackeray et al., 2012), but may be enhanced by decreasing the weight of inactive materials. For example, the thickness of current collectors may be reduced (i.e. Cu foil from 18 to 9 µm and Al foil from 20 to 15 µm). Also, the C-channel wall thickness could be decreased from 3.175 to 1.5 mm. The achieved energy density and specific energy are comparable, however, to the prismatic Li-ion batteries used in the EVs: 218 Wh L⁻¹ and 109 Wh kg⁻¹ from Lithium Energy Japan used in the Mitsubishi i-Miev as well as 243 Wh L⁻¹ and 132 Wh kg⁻¹ from Samsung (South Korea) used in the Fiat 500e (Anderman, 2014).

The coulombic efficiency slightly decreases with cycle number (see Figure 6(d)). The average coulombic efficiency is 99.3% in the first, 98.8% in the second, and 98.3% in the third 50 cycles. Voltage efficiency (the discharge/charge voltage ratio) decreases with cycle number from 93.7% at cycle 2 to 88.9% at cycle 190. Energy efficiency (the discharge/charge energy ratio) is the product of coulombic efficiency and voltage efficiency and decreases with cycle number from 93.2% at cycle 2 to 86.7% at cycle 190. All the efficiencies decrease with cycling as the battery degrades.

Impedance rise

The EIS at 50% DoD presents an inductive region at high frequency (>100 Hz), a depressed semicircular arc at middle frequency and a sloped diffusion tail at low frequency (<0.1 Hz), as shown in the Nyquist plots after cycles 2, 50, 52, 100, and 150 in Figure 7. The high-frequency intercept of the Z′ axis (as labeled “1” in Figure 7) represents the ohmic resistance, including ionic resistance from the electrolyte and electronic resistance from electrode particles, current collectors, and battery terminals to instrument leads. The semicircular arc width along Z′ axis (as labeled “2” in Figure 7) shows the interfacial resistance, including solid–electrolyte interphase (SEI) layer resistance and charge transfer resistance on the electrode surfaces (Zhang et al., 2011; Zhang and Wang, 2009, 2013).

The ohmic resistance after cycle 2 is 11.5 mΩ. After subtracting the connection resistance from the battery terminals to the instrument leads (4.3 mΩ), the battery ohmic resistance is 7.2 mΩ. This value is higher than that of a typical Li-ion battery of similar size (<4 mΩ) (Pistoia, 2014), which suggests that the multi-tab connections from current collectors to battery terminals introduce a several milliohm resistance and should be improved.

The ohmic resistance increases to 13.4 mΩ after cycle 50, while the interfacial resistance increases from 1.8 mΩ after cycle 2 to 3.5 mΩ after cycle 50. Mechanical testing after cycle 50 resulted in slightly lower ohmic resistance (13.0 mΩ) but unchanged interfacial resistance (3.5 mΩ) after cycle 52. Consequently, the mechanical testing had no adverse effect on the electrochemical performance (i.e. impedance and capacity) of the structural battery. The ohmic resistance increases to 14.4 and 14.9 mΩ after cycles 100 and 150, where the interfacial resistance increases to 4.1 and 4.4 mΩ.

Three-point bend test

The experimental load versus mid-span displacement plots for the three-point bend tests are presented in Figure 8. The load limits for each test are listed in Table 1. Each test was loaded twice and showed good linearity and repeatability. Some nonlinearity in the low-load range can be expected because of “settling-in” of the loading fixtures. The structural battery did not show any visible damage or leakage during the tests.

Table 1.

Theoretical and experimental bending stiffness on the sandwich structural battery.

				Dry test		Wet test
Parameter	Unit	Theoretical		Test 1	Test 2	Test 1	Test 2
Maximum load, P_max	N	750	1000	758	1025	765	1060
Flexural deflection, Δ_flexural	mm	0.344	0.459
Shear deflection, Δ_shear	mm	0.009	0.012
Δ = Δ_flexural + Δ_shear	mm	0.353	0.471	0.414	0.531	0.421	0.569
Δ_flexural/Δ	%	97.5
Δ_shear/Δ	%	2.5
Bending stiffness	N mm⁻¹	2124		1967	1995	1927	1941

The bending stiffness is obtained from the slope of the load versus mid-span displacement plot in a relatively linear range with higher load of each plot, and the results are summarized in Table 1. The wet tests after 50 charge/discharge cycles have a small reduction in the bending stiffness (<3%) with respect to the dry tests before filling the electrolyte. Considering experimental variability, it can be concluded that charge–discharge cycling and exposure to battery electrolyte have little effect on the bending stiffness. However, such a result can be expected due to the relatively small thickness and elastic modulus of adhesives used in sandwich panel construction. That is, as long as the adhesive does not fail, it has little effect on panel stiffness.

The bending stiffness of the sandwich structural battery can be calculated using the method in ASTM D7250/D7250M-06 (2012) (ASTM Standards, 2012) for a sandwich beam with a known face skin modulus. The mid-span deflection (Δ) depends on the load (P) as follows

Δ = \frac{P S^{3}}{48 D} + \frac{PS}{4 U}

(1)

where S is the span length, D is the flexural rigidity, and U is the through-thickness shear rigidity. The first term on the right side of equation (1) is flexural deflection (Δ_flexural), while the second term is shear deflection (Δ_shear).

The calculation of flexural rigidity (D) in equation (2) includes the contributions of two face skins, twoC-channels, and all the Cu current collectors

D = \sum E_{i} I_{i}

(2)

where E_i and I_i are Young’s moduli and area moments of inertia of the constituents, respectively. The Al current collectors are ignored in the flexural stiffness calculation since they are not directly bonded to the face skins. As seen in Table 2, the C-channels and Cu current collectors have significant contribution (i.e. 36%) to the total flexural rigidity, though the ASTM Standards (2012) only consider the flexural rigidity of the facing for a sandwich beam with typical low-modulus foam or honeycomb cores.

Table 2.

Calculated flexural rigidity (D), core shear modulus (G), and through-thickness shear rigidity (U).

Constituent	D (N m²)	Percentage of total D	G (MPa)	U (MN)	Percentage of total U
Two face skins	500	64	–	–	–
Cu current collectors	58	7	1750	1.60	31
Two C-channels	223	29	3940	3.55	69
Total	781	100	5690	5.15	100

The through-thickness shear rigidity (U) is calculated using equation (3)

U = \frac{G {(d - t)}^{2} b}{d - 2 t}

(3)

where d is the total beam thickness, t is the thickness of one of the two identical face skins, and b is the beam width. The core shear modulus G is calculated using a thickness-based rule-of-mixtures across the width of the beam including only the two C-channels and the Cu current collectors. As with the flexural stiffness, the Al current collectors were neglected for the calculation of core shear modulus. The various contributions to the through-thickness shear rigidity are listed in Table 2. It is seen that the Cu current collectors contribute 31% of the total shear rigidity.

Using equation (1), one can compute flexural deflection (Δ_flexural), shear deflection (Δ_shear), and the total mid-span deflection (Δ) at the various peak loads (P_max) used in the experiments. These deflections are listed in Table 1. It is seen that the structural battery is very stiff in shear, as evidenced by only 2.5% of the overall deflection being due to shear.

The theoretical bending stiffness, defined as load per unit mid-span deflection, P_max/Δ_total, is 2124 N mm⁻¹ (Table 1), while the experimental bending stiffness for the dry and wet conditions (1927–1995 N mm⁻¹) is within 10% of theoretical bending stiffness. This result proves that the structural battery behaves as a sandwich beam and the battery core is performing its structural function.

It is important to note that, if the two C-channels are not considered in the above calculations, the bending stiffness is 1528 N mm⁻¹, which is 72% of the 2124 N mm⁻¹ total stiffness including the two C-channels. If only the two face skins are considered, the bending stiffness is only 2 N mm⁻¹, nearly 0.1% of the 2124 N mm⁻¹ total stiffness. Thus, although the Cu current collectors contribute relatively little to the overall stiffness of the structural battery demonstrated in this work, they can potentially contribute significantly in future designs with thinner C-channels.

Potential of structural batteries for EV applications

In the proposed design, the Cu current collectors carry mechanical loads and electrical current. The sandwich panel design fully utilizes the current collectors in a structural function so they can be considered structural components and not battery components. The energy density of the battery is therefore increased because the weight of the Cu current collectors need not be included. In this structural battery prototype, the battery core weighs 380.5 g, and its volume is 166 mL. Accordingly, the specific energy and energy density of the battery core are 170 Wh kg⁻¹ and 389 Wh L⁻¹, respectively. The 18-µm-thick Cu current collectors account for 17.2% of the battery core mass. Thus, the proposed design has the potential to increase the battery’s specific energy by 20.8% to 205 Wh kg⁻¹. From a volumetric perspective, the Cu current collectors take up 4.4% of the battery core space, so removing them would result in a 4.6% increase in energy density to 407 Wh L⁻¹. If both the anode and the cathode current collectors were fully utilized in a structural function, the battery’s specific energy and energy density could be increased by 27.3% to 216 Wh kg⁻¹ and 8.7% to 422 Wh L⁻¹, respectively. The proposed design also obviates the need for packaging that is designed to protect the battery from penetration in the event of an accident. Often, thick and heavy materials are used for this packaging. Finally, mass reduction in EVs has a compounding effect in the overall vehicle design. A lighter vehicle needs smaller batteries, electric motors, and suspension components. To conclude, this design could reduce the weight and cost of EVs.

Conclusion

A 17.85-Ah multifunctional structural battery based on state-of-the-art lithium-ion battery technology and sandwich panel construction was successfully fabricated and experimentally tested, demonstrating an energy density of 248 Wh L⁻¹ and specific energy of 102 Wh kg⁻¹. After 50 charge–discharge cycles at ∼C/3 rate, three-point bend tests upto 1060 N had no adverse effect on the subsequent electrochemical performance (i.e. impedance and capacity) of the structural battery. After 190 cycles at ∼C/3 rate and eight mechanical loading cycles, the capacity retention was 85.8%.

The measured bending stiffness of >1900 N mm⁻¹ showed only a 3% reduction upon filling with electrolyte and undergoing 50 charge–discharge cycles. Ex situ mechanical tests showed a significant (31%) decrease in lap shear strength of the E-120HP epoxy after exposure to the battery electrolyte for 3 months. However, the significant degradation did not cause any structural damage or leakage during the charge–discharge cycling of structural battery. Sandwich beam theory is able to predict the stiffness of the sandwich structural battery within 10% difference from the experimental bending stiffness, and it proves that the structural battery behaves as a sandwich beam and the battery materials are sharing in the load-carrying function of the sandwich panel. The sandwich beam theoretical calculations show that the shear deflection is only 2.5% in the mid-span deflection, indicating that the battery core is very stiff.

The structural battery prototype has demonstrated good electrochemical and mechanical performance. However, there is obviously opportunity to optimize the fabrication process in terms of adhesive durability, bonding integrity, battery coating and stacking, multi-tab connections, and structural weight reduction, and thus enhance its performance including energy density, specific energy, cycle life, and mechanical properties. Nevertheless, the structural battery shows interesting promise for applications where sandwich structures and batteries are used, such as floors of EVs. In such applications, the structural battery may more efficiently store energy on a mass or volume basis, thereby allowing for enhanced range and performance.

Footnotes

Acknowledgements

The authors thank Dr Chao-Yang Wang for valuable discussions and providing battery manufacturing and experimental facilities as well as Devin O’Donnell for his help in the battery fabrication.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by ARPA-E and Dr Ping Liu through the Department of Energy award DE-AR0000383.

References

Anderman

(2014) Extract from the Tesla Battery Report. Available at: http://www.totalbatteryconsulting.com/industry-reports/Tesla-report/Extract-from-the-Tesla-battery-report.pdf (accessed 22 December 2015).

Arora

Zhang

(2004) Battery separators. Chemical Reviews 104(10): 4419–4462.

ASTM Standards (2012) D7250/D7250M-06. Standard practice for determining sandwich beam flexural and shear stiffness. In: ASTM (ed.) Annual Book of ASTM Standards. West Conshohocken, PA: ASTM International, 15(3).

Bugga

Smart

(2010) Lithium plating behavior in lithium-ion cells. ECS Transactions 25(36): 241–252.

Cannarella

Arnold

(2014) Stress evolution and capacity fade in constrained lithium-ion pouch cells. Journal of Power Sources 245: 745–751.

Ding

Tian

Meng

. (2013) Smart multifunctional fluids for lithium ion batteries: enhanced rate performance and intrinsic mechanical protection. Scientific Reports 3: 2485.

Ekstedt

Wysocki

Asp

(2010) Structural batteries made from fiber reinforced composites. Plastics, Rubber and Composites 39(3–5): 148–150.

Evanoff

Benson

Schauer

. (2012) Ultra strong silicon-coated carbon nanotube nonwoven fabric as a multifunctional lithium-ion battery anode. ACS Nano 6(11): 9837–9845.

Jacques

Kjell

Zenkert

. (2012) Impact of electrochemical cycling on the tensile properties of carbon fibres for structural lithium-ion composite batteries. Composites Science and Technology 72(7): 792–798.

10.

Kim

Sastry

(2012) Effects of carbon fiber electrode deformation in multifunctional structural lithium ion batteries. Journal of Intelligent Material Systems and Structures 23(16): 1787–1797.

11.

Liu

Sherman

Jacobsen

(2009) Design and fabrication of multifunctional structural batteries. Journal of Power Sources 189(1): 646–650.

12.

Mullenax

Browning

Huebsch

. (2012) Composite multifunctional lithium-ion batteries. ECS Transactions 41(41): 175–185.

13.

Pereira

Guo

Nieh

. (2008) Embedding thin-film lithium energy cells in structural composites. Composites Science and Technology 68(7): 1935–1941.

14.

Pereira

Guo

Nieh

. (2009) Energy storage structural composites: a review. Journal of Composite Materials 43(5): 549–560.

15.

Pistoia

(ed.) (2014) Lithium-Ion Batteries: Advances and Applications. 1st ed.Warsaw: Elsevier, p. 45.

16.

Rahn

Bakis

Hickner

. (2014) Sandwich panels with battery cores. US Patent application No. 61/931,877, PCT/US2014/067700.

17.

Rashid

Gupta

(2015) Effect of relaxation periods over cycling performance of a li-ion battery. Journal of the Electrochemical Society 162(2): A3145–A3153.

18.

Reichert

Andre

Rösmann

. (2013) Influence of relaxation time on the lifetime of commercial lithium-ion cells. Journal of Power Sources 239: 45–53.

19.

Roberts

Aglietti

(2008) Satellite multi-functional power structure: feasibility and mass savings. Proc IMechE, Part G: J Aerospace Engineering 222(1): 41–51.

20.

Roberts

Aglietti

(2010) Structural performance of a multifunctional spacecraft structure based on plastic lithium-ion batteries. Acta Astronautica 67(3): 424–439.

21.

Seo

(2016) Polymer swelling effects on mechanical and electrochemical properties for battery adhesives and ion exchange membranes. Master’s Thesis, The Pennsylvania State University, University Park, PA.

22.

Singh

Cao

. (2015) Design, manufacture and test of a novel structural battery based on sandwich construction. Journal of Sandwich Structures & Materials 17(6): 666–690.

23.

Snyder

Wetzel

Watson

(2009) Improving multifunctional behavior in structural electrolytes through copolymerization of structure- and conductivity-promoting monomers. Polymer 50(20): 4906–4916.

24.

Tarascon

Armand

(2001) Issues and challenges facing rechargeable lithium batteries. Nature 414(6861): 359–367.

25.

Thackeray

Wolverton

Isaacs

(2012) Electrical energy storage for transportation—approaching the limits of, and going beyond, lithium-ion batteries. Energy & Environmental Science 5(7): 7854–7863.

26.

Thomas

Qidwai

(2005) The design and application of multifunctional structure-battery materials systems. Journal of the Minerals Metals & Materials Society 57(3): 18–24.

27.

Thomas

Qidwai

Pogue

. (2013) Multifunctional structure-battery composites for marine systems. Journal of Composite Materials 47(1): 5–26.

28.

Vlad

Singh

Galande

. (2015) Design considerations for unconventional electrochemical energy storage architectures. Advanced Energy Materials 5(19): 201402115.

29.

Wang

Peng

Zhang

(2014) Mechanical and electrical behavior of a novel satellite multifunctional structural battery. Journal of Scientific & Industrial Research 73: 163–167.

30.

Willgert

Kjell

Lindbergh

. (2013) New structural lithium battery electrolytes using thiol–ene chemistry. Solid State Ionics 236: 22–29.

31.

Zhang

Wang

C-Y

(2009) Cycle-life characterization of automotive lithium-ion batteries with LiNiO₂ cathode. Journal of the Electrochemical Society 156(7): A527–A535.

32.

Zhang

Wang

C-Y

(2013) State of health estimation of LiFePO₄/graphite cells. ECS Transactions 50(31): 275–281.

33.

Zhang

Wang

C-Y

Tang

(2011) Cycling degradation of an automotive LiFePO₄ lithium-ion battery. Journal of Power Sources 196(3): 1513–1520.