Problems hindering the commercialization of high-energy silicon lithium-ion battery laboratory innovations
The energy density of traditional lithium-ion batteries has reached its limits and cannot meet the rapidly growing demand for long-range electric vehicles. Due to the high weight capacity of silicon (3592 mAh g-1), there have been many attempts to introduce silicon-based materials into traditional graphite anodes to improve energy density. However, mixing silicon-based materials (SiOx and Si-C composites) with graphite has had limited industrial success. Due to the degradation of performance due to the inherent problems of silicon anodes (figure), the use of silicon in commercial anodes is still very limited (less than 5 wt%).
Figure 1: Specifications for a commercial silicon-containing 18650 battery. The silicon content and specific capacity of the anodes used in commercial 18650 silicon-containing cells by battery companies. Si-C and SiOx-graphite composites are the two main silicon-based anodes used in industry.
Most research on silicon-based anodes has focused on improving cycling performance under relatively mild test conditions. Actual test conditions must include active materials with high load masses, inactive ingredients with low mass and volume fraction (such as binders and carbon additives), dense electrodes with low porosity, and poor electrolytes. In addition, calendar aging, storage conditions, stacking pressure, safety, and cost-effectiveness are also important factors in practical applications, but these factors have rarely been considered in previous studies.
Despite increasing collaboration between academia and industry to develop silicon-based anodes, there is still no comprehensive set of practical performance metrics to evaluate silicon-containing cells. In addition, previous reviews have focused on advances in materials development, and few reports have discussed the challenges of silicon-containing cells in detail from a practical perspective.
In academia, energy density is usually calculated only from active materials, while in industry, energy density must be considered at the battery level. Values obtained without taking into account actual conditions, such as electrode expansion and inactive components, are artificially overestimated. In addition to overestimation, there are important factors that affect the practical application of silicon anodes, such as discharge cut-off voltage and reduced calendar life, which are easily overlooked by many researchers.
The following will consider the factors hindering the practical development of silicon-based batteries from various aspects of anodes. First, the importance of electrode expansion in actual battery design is discussed, especially its influence on battery energy density. The relationship between the cut-off voltage and the silicon content in the silicon-based anode is then highlighted. The calendar life, safety and cost-effectiveness of silicon-containing cells, as well as comparison with graphite cells, are also considered. With all these factors in mind, a general test protocol is recommended to evaluate the commercial viability of silicon-based anodes.
Effect of electrode expansion on battery energy density
The energy density of the battery is maximized when the active ingredients occupy the limited space of the battery. However, in commercial battery designs, any free space inside a hard case or soft bag is usually used to accommodate the expansion of the battery (figure). Since the electrode expansion rate of conventional graphite anodes is less than 20%, graphite batteries require relatively small free space to accommodate the expansion during battery operation. Due to the low expansion and small free space of graphite batteries, despite their low theoretical specific capacity, they can still achieve a reasonable energy density. However, due to the large expansion of the electrode, silicon-containing cells inevitably require more free space than graphite batteries. This will reduce the energy density and reduce the advantage of the high specific capacity of silicon-based anodes. Most studies, while acknowledging the expansion of the electrodes, ignore the relationship between energy density and the free space inside the battery.
Figure 2: Volume expansion of silicon anodes and related issues. Schematic diagram of a laminated cell with edge space representing free space. b, cross-section of a laminated cell. c, Relationship between electrode expansion and lithiation capacity of silica-containing graphite electrodes (30 wt% silicon nanoparticles). Two different electrode porosities. d-i, electrode expansion (d, g) and estimated cell bulk density (e, h) and weight density (f, i) and specific capacity as a function of silicon oxide (d-f) or silicon content (g-i) in the anode. The electrode porosity value (%) is given in d and g. Data on electrode expansion, porosity, and specific capacity come from the literature. j, schematic showing the change in electrode thickness caused by reversible and irreversible electrode expansion during cycling. m, the average cell pressure of graphite cells and silicon hybrid graphite cells in the process. n, real-time electrode expansion behavior of silicon graphite anode during cycling in whole cells.
In addition to free space, electrode density or electrode porosity should also be considered. Research by Prado et al. showed that the expansion of the silica graphite (30 wt% silicon content) electrode (figure) has a large relationship with the electrode density. In general, a silicone-containing electrode with a lower bulk density (higher porosity of the electrode) has less expansion because it has more space to accommodate the volume expansion. However, higher electrode porosity requires not only thicker electrodes (to maintain the same load), but also more electrolyte to fill the pores, which reduces battery energy. It is reported that the electrode density of commercial graphite or silicon-containing anodes is generally about 1.6 g cm-3, and the porosity of the electrode is about 30%. However, many studies on silicon-based anodes employ higher electrode porosity to reduce volume changes that can severely reduce energy density.
To clarify the effect of electrode expansion on energy density, we estimated the energy density of commercial cells consisting of SiOx or Si nanoparticles and graphite with different Si content. As an illustration, we chose a common nickel-manganese-cobalt cathode (NMC811) to construct a pouch full battery. Pouch batteries typically measure 102 mm × 354 mm × 9.5 mm and have a capacity of more than 50 Ah.
First, we confirmed the anode electrode expansion as a function of silicon content based on information in the literature on electrode expansion and porosity (figure). Silicon hybrid electrodes have a higher specific capacity than silicon oxide hybrid electrodes, but also higher porosity. Therefore, the two electrodes have similar expansion.
Then, a plot of volumetric energy density vs. Si and SiOx content is plotted (figure). Batteries with a silicon oxide content of more than 10 wt% have a higher energy density than commercial graphite batteries. However, it is worth noting that the energy density of silicon-containing cells containing SiOx in more than 20 wt% in the anode has decreased. This is because higher electrode expansion requires more free space, which reduces the number of stacks in the limited space of the pouch battery, resulting in lower battery capacity. The energy density of a cell containing 60 wt% silicon oxide in the anode is close to that of a graphite battery, although its specific capacity is more than three times higher than that of a graphite battery. Batteries with a silica content of more than 70% have a lower energy density than graphite batteries. A similar trend is observed in silicon-containing cells; However, due to the high porosity and expansion of the electrodes, their energy density is lower than that of graphite batteries.
The gravitational energy density (Wh kg-1) of each cell was also calculated to observe the effect of adding Si to the anode (figure). Silicon oxide-containing cells show a higher gravitational energy density than graphite cells, and gravity energy density increases with increasing silicon content. The gravitational energy of silicon-containing cells is also on the rise, although their energy density is lower than that of graphite batteries. However, it should be noted that both silicon oxide cells and silicon-containing cells have a small increase in gravity energy density. As mentioned above, to maintain a fixed capacity ratio (N/P) of the anode and cathode, the anode load must decrease significantly as the silicon or silica content increases. Increasing the number of electrode stacks can increase the energy density. However, due to the expansion of the anode, it is not possible to add many electrode stacks, thus limiting the increase in gravitational energy density. For example, although the specific capacities of anodes containing 50 wt% silicon (or SiOx) and 100 wt% silicon (or SiOx) vary greatly, their severe expansion means that the maximum number of electrode stacks is the same. In addition, due to the low initial coulomb efficiency of the silicon anode, the area capacity of the battery decreases as the silicon content increases, which also hinders the increase in gravity energy density.
While these results may differ from batteries in real-world applications, they show the importance of electrode expansion and porosity to battery energy density. In addition, these estimates also show that the greater the specific capacity of the anode, the higher the energy of the battery. Therefore, we encourage researchers to carefully evaluate the performance of silicon-based anodes considering actual battery conditions.
Observing the change in anode thickness during repeated cycles is a good indication that the electrode expansion has stabilized. After an initial expansion of 15%, the graphite exhibits a steady thickness change, and the overall thickness does not increase further over a long period of cycling. However, unlike graphite, silicon-containing cells gradually increase electrode expansion during repeated cycles, even if the electrode expansion is large during the initial cycle, because the silicon-containing electrode expands irreversibly after full discharge. This irreversible expansion is caused by morphological changes in silicon, accumulation of solid-electrolyte interface layers (SEIs), detachment of particles in the electrodes, and delamination of current collectors (figure). Therefore, even if the silicon-based anode exhibits only a slight expansion during the initial cycle, gradual expansion of the electrode may occur during a long cycle. Prado et al. demonstrated the irreversible expansion behavior of a silica graphite electrode containing 30 wt% Si (figure): the silica graphite electrode expands with circulation, and the irreversible expansion also expands with cycling, resulting in permanent expansion of thickness after complete detitanium.
In practical battery applications, a certain superimposed pressure is usually applied to the module. External pressure inhibits irreversible expansion caused by interparticle disengagement and electrode layering, thereby reducing the effect of electrode expansion on energy density. However, due to the increasing expansion during cycling, external pressure will bring extremely high compressive stress to silicon-containing cells. External pressure and high stress in the cell can reduce the porosity of the electrode, increasing the pore resistance during cycling, resulting in capacity decay. Louli et al. performed operational pressure measurements to determine the effect of irreversible expansion on the performance of various silicon-containing cells. Their study showed that graphite batteries maintain an almost constant pressure when discharged, while silicon-containing cells continue to rise during cycling (figure), which can cause fatal damage to the battery. Therefore, the electrode thickness must be monitored to determine if the silicon-based anode has reached a stable expansion point over several cycles (figure). If the electrode expansion cannot be stabilized, it should not be used as an anode in an actual battery. To ensure the stability of the battery, the electrode expansion of the stability point should be considered when designing silicon-containing cells, rather than the first cycle. Otherwise, the battery may bulge or burst during cycling due to excessive expansion of the electrodes.
In addition to irreversible expansion, the anode may also exhibit in-plane (x-direction) expansion and inter-plane (z-direction) expansion. Pietsch et al. report that graphite anodes exhibit a small amount of in-plane expansion during charging because the particle shape and expansion direction allow the electrodes to expand to all directions. Compared to through-plane expansion, graphite's intraplane expansion is negligible. However, for silicon-based anodes, in-plane expansion can be severe due to the large anisotropic expansion of silicon. In-plane expansion can cause fatal problems, including electrode delamination and contact short-circuits between the anode and cathode. Although this in-plane bloat is important, it has been ignored due to the lack of academic awareness of the issue.
Cut-off voltage and its relationship to silicon content
The figure shows the delithiation voltage curve of an anode prepared with graphite, silicon-coated graphite (5 wt% Si) and silicon-coated carbon (35 wt% Si). Compared to graphite anodes, the voltage curve of silicon-containing anodes is more skewed because silicon has a relatively high operating potential and amorphous nature. This slope varies depending on the properties of the silicon material. As shown, most of the total capacity of graphite is provided in the low-voltage region (below 0.3 V), while silicon-containing anodes have a large capacity contribution below 1.5 V. The inclined voltage curve of silicon-based anodes means that a higher dequartzation voltage cutoff point is required to use its maximum capacity compared to graphite. This also means that in a full-cell configuration, silicon-containing batteries require a lower cut-off voltage than graphite batteries to achieve their full capacity.
Figure 3: The cut-off voltage decreases after silicon is added to the graphite anode. a. Normalized delithiumization voltage curves of graphite, silicon-coated graphite (5 wt% silicon) and silicon-coated carbon (35 wt% silicon) anodes with specific capacities of approximately 360, 510 and 1,250 mAh g-1. b. The anode cut-off voltage of various full batteries with different full battery cut-off voltages and the corresponding silicon content in the anode. c. The voltage curves of the whole battery with different silicon content are estimated by subtracting various anode voltage curves from the NMC622 cathode voltage curve. Enlarged view of the voltage curve in d, c showing the discharge cut-off voltage required to obtain 99% usable battery capacity.
The figure shows the anode cutoff voltage and the corresponding full-cell cut-off voltage from previous studies of silicon-containing batteries. Each cut-off voltage of the silicon-based anode is measured by a full-cell test in a three-electrode configuration. At a full battery cut-off voltage of 3.0 V, the end-of-discharge voltage for an anode containing approximately 15 wt% Si is approximately 0.68 V. With a full battery cut-off voltage of 2.8 V, the anode has a termination voltage of 0.83 V. These voltages (0.68 V and 0.83 V) are too small to use the full capacity of silicon-based anodes. For a 2.5 V cut-off voltage, the end-of-discharge voltage for the anode is approximately 1.12 V for silicon content greater than 15 wt% and 1.09 V for silicon content less than 5 wt%. Regardless of the silicon content, the end-of-discharge voltage is similar in all cases. As shown in the figure, since the higher the silicon content, the more inclined the voltage plateau, so the capacity of a cell with a lower silicon content is closer to its maximum than a cell with a higher silicon content. For silicon-dominated cells (silicon content greater than 85 wt%), the end-of-discharge voltage is approximately 0.91 V at a cut-off voltage of 2.3 V. This relatively low discharge voltage is not sufficient to fully utilize the capacity of the silicon-dominated anode (figure). These results show that silicon-containing batteries require a lower full-cell cut-off voltage to reach their maximum capacity compared to graphite batteries. In addition, it should be noted that these cutoff voltage behaviors depend on the battery design, including N/P ratio, cell size, and target energy density. Even with electrodes consisting of the same material, the stock of active lithium ions varies depending on the battery design, resulting in different cut-off voltages.
We calculate the full cell discharge voltage curve for cells with different silicon content by subtracting the voltage curves of the various anodes previously reported in the cathode from the voltage curve of the lithium nickel-cobalt-manganate (LiNi0.6Mn0.2Co0.2O2, NMC622) cathode (Figure). As the silicon content in the anode increases, the slope of the voltage curve becomes more pronounced. To verify the effect of adding Si on voltage characteristics, we confirm the discharge cut-off voltage required to reach 99.0% of theoretical battery capacity (figure). With more Si added to the electrode, the discharge cut-off voltage drops from 3.13 V to 2.49 V. In addition, we calculated the average voltage for each cell using the voltage curve shown in the figure (figure). As the Si content in the anode increases, the average voltage of the battery gradually decreases from 3.68 V to 3.39 V, with a battery containing 35 wt% Si in the anode having an 8% lower voltage than a graphite battery. The reduction in voltage not only reduces the energy of the battery, but also diminishes the benefits of using high levels of silicon in the anode. In addition, a low discharge cut-off voltage increases the anode potential, which induces side reactions such as copper dissolution in the anode current collector. Copper dissolution can cause serious problems, including the formation of additional SEIs, increased battery resistance, uneven current density, and safety issues. Therefore, the cut-off voltage of silicon-containing cells is critical because it directly affects battery performance. However, most studies related to the development of silicon materials do not consider this issue in detail.
The importance of calendar longevity and its assessment
Calendar aging is a time-varying performance decay and is an intrinsic behavior that must be considered to estimate the long-term health of silicon-containing batteries. However, most studies have focused on strategies to improve the charge-discharge cycle life of silicon cells, and only a few studies have involved calendar lifetimes of silicon-containing cells. At storage conditions of 20-40 °C, conventional graphite batteries have a calendar life of more than 15 years (the time when the battery capacity reaches 80%), while silicon-containing cells typically have a calendar life of less than 1 to 2 years, even with low silicon content. The U.S. Alliance for Advanced Batteries (USABC) believes that the calendar life of electric vehicle batteries is preferably more than 5 years, and this standard should be extended for batteries that require longer life, such as large-scale energy storage. While the electrochemical cycling characteristics and energy densities of state-of-the-art silicon-containing cells are rapidly approaching the target values set by the U.S. Department of Energy (DOE), there is still a significant gap between the calendar lifetime of silicon-containing cells and the target values for DOEs. This gap means that strategies that increase cycle life may not significantly improve attenuation over time. Short calendar life is one of the main obstacles to the practical application of silicon-containing batteries.
In general, the cycle life aging of silicon-containing cells is mainly due to undesirable chemical reactions between the newly exposed anode surface (caused by SEI rupture) and the electrolyte, ongoing electrolyte decomposition, and mechanical degradation. Calendar life aging is related to the stability of the surface passivation (SEI) layer and the silicon under it over the long term. Unfortunately, the passivation layer on the silicon surface is not stable, because at the same potential, the chemical reactivity of silicon graphitized is higher than that of graphite. Charged silicon tends to reduce and break down binders, lithium salts, and organic solvents in the electrolyte, resulting in the reconstruction of the SEI layer and loss of recycled lithium content. The loss of lithium stocks in silicon leads to self-discharge and contraction of lithium particles, resulting in mechanical destruction of the SEI layer. Self-discharge may also create an additional SEI layer on the surface of newly exposed silicon, leading to further loss of recycled lithium. In addition, silicon-containing batteries are highly susceptible to hydrofluoric acid (HF) produced during hydrolysis of common lithium salts such as LiPF6. Traditional graphite batteries usually have a certain tolerance to hydrofluoric acid and will not have serious battery degradation. However, silicon is etched with hydrofluoric acid, resulting in the loss of active silicon and the production of H2O, gases, and soluble silicon components (SiF4 and H2SiF6). These by-products accelerate the generation of high frequencies, clog the pore network of the anode and cathode, and increase the impedance by forming a passivation layer on the surface of the cathode. Figure depicts the mechanism of calendar aging. These degradations accelerate when the battery is in a higher state of charge or storage temperature is high. In this case, pre-lithiation of silicon-containing anodes may improve their cycle life, but at the same time may lead to more severe calendar aging due to the high reactivity of silicon lithide.
Figure 4: Mechanism and analysis of calendar lifetime aging of silicon anodes. a. Schematic diagram of various degradation phenomena of silicon-containing anodes during calendar aging. The damaged beige sphere represents the aged silicon particles, and the gray layer is SEI. b. Voltage hold test protocol developed by the U.S. Department of Energy's Silicon Alliance program. c, Combined parasitic capacity of pure graphite cells and silicon-containing graphite cells (15 wt% silicon content) during calendar aging with or without FEC added to the electrolyte. e, HF produced in electrolyte with or without FEC during 30 days of thermal aging at 60 °C.
To evaluate calendar aging, long-term experiments are required, which may last more than one year, such as the established test procedures. This long-term testing is actually difficult to perform in a laboratory environment. As an alternative, potentiostat-based methods have been developed to measure the lithium stock loss of the SEI layer in silicon-containing cells in shorter-duration experiments at the interface pair. Figure shows the voltage hold test curve established by the DOE Silicon Alliance project. However, it should be noted that voltage retention cannot replace the traditional calendar aging long-term test. Schulze et al. discuss the principles and methods of potentiostatic holding testing as a calendar life acceleration predictor for various anodes, including graphite and silicon anodes. The current responses they measured included reversible and irreversible capacity, which means continuous charging of the battery and irreversible loss of lithium (i.e., parasitic reactions such as the formation of SEI). Kalaga et al. evaluated the calendar lifetime of a 15 wt% silicon anode cell and compared it to a graphite battery. The study clearly shows that the addition of silicon significantly increases the combined side reaction current compared to silicone-free graphite batteries (Figure). In addition, the addition of fluoroethylene carbonate (FEC) appears to inhibit the side reactions of both batteries during voltage retention. However, FEC-containing silicon-containing cells lose almost twice as much capacity as FEC-free silicon-containing cells over time (figure). This opposite result of the side reaction current during voltage retention and the capacity loss after calendar aging can be explained by the oscillating particle size of silicon. During the voltage holding process, the silicon particles remain intact without any substantial changes, and the SEI layer is relatively stable to mechanical damage (such as silicon particle shrinkage), which can effectively prevent high-frequency erosion of silicon. However, when the battery is discharged after the voltage holding step, the SEI layer collapses due to the shrinkage of the silicon particles. The accumulated high frequencies can penetrate the SEI layer and react with the silicon, resulting in severe degradation. FEC is known to produce HF by dehydrofluorination at high temperatures. Shin et al. confirmed that batteries containing 5 wt% FEC in the electrolyte produced more HF after 30 days of thermal aging at 60 °C than batteries without FEC (figure). This suggests that FEC additives lead to higher capacity loss after silicon-containing cells age over time at high temperatures. Considering that FECs are often used as additives to improve the cycle life of silicon cells, stability issues affecting the calendar life of FEC-containing cells pose another serious challenge to the design of silicon-containing cells.
Security issues
The use of silicon-based materials in anodes can create safety concerns because large volume changes and chemical instabilities of silicon materials during the (de-lithium) process can lead to cell bursting and thermal runaway. As mentioned above, gassing of silicon-based anodes may be more severe than traditional graphite anodes because the high chemical reactivity of lithium carbide produces an unstable passivation layer. Typically, the gas produced by the initial cycle is discharged through the degassing process. However, due to the adverse decomposition reaction caused by the inherent reactivity of silicon, gas is constantly generated during battery operation. Seitzinger et al. investigated the chemical reactivity and venting behavior of silicon anodes in relation to lithium salts and silicon surface chemistry (figure). They found that all silicon anodes terminating Si (PECVD, Nanoamor and S5505) produce various gases due to the dissolution of Si and the decomposition of electrolytes and lithium salts. When a battery has a closed case, violent gas evolution can cause the cell to bulge or burst, exposing the active electrode and flammable carbonate electrolyte to air (figure). These gas behaviors can cause thermal runaway of the battery. In addition to gas evolution, electrode expansion during cycling can cause electrode deformation or bulging, causing fatal safety issues such as internal short circuits (figure).
Figure 5: Safety concerns posed by silicon-based anodes. a, the volume of gas produced per unit surface area of silicon-containing materials over a 14-hour period. b, Schematic diagram of silicon reacting with electrolyte to produce gas evolution leading to bulging cells. c, Computed tomography measurements of cylindrical cells containing silicon anodes during cycling. Illustration: Electrode deformation due to expansion. d, SEM images of silica fumes with particle sizes of 15, 5, and 0.03-0.050 μm, respectively. e, corresponding DSC curve in the fully actinized state of the electrode prepared using silica fume S1, S2, or S3. f, Flow chart of thermal runaway mechanism of silicon cell.
Although silicon materials generally exhibit good thermal stability comparable to graphite, due to the exothermic nature of SEI decomposition, excessive SEI on the surface of the silicon anode during repeated cycling leads to higher calorific value. Park and Lee used differential scanning calorimetry (DSC; Fig. ) analyzes the thermal stability of silicon graphitide in relation to silicon particle size. The results of the analysis show that electrodes prepared with smaller silicon particles tend to generate more heat at an earlier stage (heating to 300 °C) and lower heating temperatures due to the accumulation of SEI layers over a larger silicon surface area (figure). As we all know, reducing the particle size of silicon is a common way to solve the inherent problems caused by the volume change of silicon and improve its cycle stability. However, it should be noted that this approach also raises safety concerns due to the larger surface area of smaller silicon particles (i.e., large amounts of SEI), as abnormal heating can trigger pre-existing SEI decomposition, which in turn generates more heat, eventually causing the polymer separator to melt, resulting in an internal short circuit. This highlights the need to carefully consider the trade-off between improved performance and safety concerns when developing silicon-based anodes. The figure summarizes the sequence of events caused by the inherent problems of silicon-based anodes, and the various safety issues that result from them.
Cost-effectiveness of adding silicon batteries
As shown in the figure, adding an appropriate amount of silicon to the graphite anode can improve the energy density of the battery. However, the increase in energy density does not guarantee the cost-effectiveness of silicon-containing cells on the $Wh-1 metric. There is no doubt that performance factors such as energy density, cycle life, and calendar life are important for the practical application of silicon cells, but cost-effectiveness is also an important consideration for the commercial viability of silicon cells. It is well known that the cost of traditional graphite anodes is only a small part (<15%) of the total cost of the battery due to the moderate price, while the typical nickel substrate oxide cathode accounts for more than 50% of the total cost of the battery. Therefore, the cost of graphite batteries is mainly concentrated in the cathode. In the case of silicon, the raw material cost of bulk silicon (<5$kg-1) is lower than the raw material cost of commercial graphite (<15$kg-1). However, silicon materials used in anodes require complex and expensive synthetic methods to alleviate the inherent problems primarily associated with volume variation. These multi-stage and expensive methods greatly increase the overall cost of battery production.
Silicon oxide is widely used in commercial silicon anodes and can be synthesized by proven technologies. Greenwood et al. recently reported a detailed bottom-up assessment of the performance and cost of soft-contained silicon cells. By measuring the battery-scale energy density of a whole cell containing 5 wt% SiOx in graphite-based anodes and various cathodes, they found that the introduction of a small amount of SiOx increased the energy density by 7.62 Wh L-1, even after accounting for electrode expansion (figure). Depending on the battery design, increasing the amount of silica does not necessarily result in an increase in energy, which is consistent with the results shown in the figure. In terms of cost-effectiveness, adding 5 wt% silica is not more advantageous than graphite batteries when the price of silica is 60 $/kg (figure). According to Greenwood et al. modeling, for every 1 wt% increase in silica, the anode cost increases by approximately 0.53 $ kg-1. If more silica is used to obtain higher energy, this cost-effectiveness will be worse. The figure shows the target cost of silica as a function of the amount of silicon added to achieve a flat cost of graphite. These costs are simulated under optimistic cathode capacity conditions, and depending on the battery design, the values may differ from the actual battery. Meanwhile, Schmuch et al. estimate that adding a Si-C composite (20 wt% Si) is more cost-effective than conventional graphite anode cells (Figure). This different cost trend is due to the fact that the material cost of Si-C composites (25$kg-1) is much lower than that used in the study by Greenwood et al. (60$kg-1). The low price of silicon indicates that the addition of cheap silicon with high specific volume is beneficial to reduce battery costs. However, while silicon's raw material costs are typically low, the fabrication of silicon anodes may require additional processes such as nanoengineering and building composites with carbon or inactive matrices to overcome silicon's inherent problems such as high volume expansion and low conductivity, thereby raising the total cost. Alternative approaches to solving these problems often require the use of inactive ingredients such as functional binders, electrolytes, additives, and expensive conductive carbon, which are not present in graphite batteries, which also increases material costs (figure). In addition, auxiliary equipment can also increase costs, such as surface pressure devices that balance changes in battery thickness during actual battery operation. For these reasons, while it is difficult to accurately calculate the cost of newly developed silicon materials in the lab, we encourage battery researchers to carefully consider the cost of the raw materials used, the technological maturity of the synthesis methods, and the large-scale production of the materials.
Figure 6: Cost of silicon-containing cells. a. Energy density (yellow), specific energy (blue) and battery energy (pink) of various full batteries. The anode containing 5 wt% SiO is paired with the NCM111, NCM523, NCM622, W-NCM811, W-NCM955, and W-LNO cathodes. Tungsten-doped cathodes are denoted by the prefix W. b, the effect of adding SiO to the anode on battery cost. c, the silica anode achieves the maximum cost of silicon oxide as a function of silicon oxide at the same cost as the battery cost using graphite-only. d, Cost estimation of batteries with different anode and cathode composition at the electrode stack level. e, Battery cost estimation diagram, based on the high manufacturing cost of silicon, the addition of expensive inactive materials such as conductive agents, binders, electrolytes and additives, and auxiliary equipment.
Test the practical feasibility of silicon anodes
To encourage battery researchers to consider the above factors, we propose a test protocol to estimate the practical feasibility of newly developed silicon anodes at laboratory scale (Figure). First, for the synthesis of silicon anodes, we recommend that researchers carefully consider the cost of raw materials and the technical maturity of the synthesis methods used. Even if silicon anodes show excellent electrochemical properties, their industrial applications can be difficult if the material or processing cost is too high.
Figure 7: Test protocol for silicon-based anodes.
Test protocol flowchart for evaluating silicon-based anodes. The gray box in the Electrode Engineering box shows the specifications of commercial graphite batteries. This specification needs to be considered when developing viable silicon-based anodes with high energy density. The plots and schematics in the "Expansion Measurement" and "Energy Density Estimation" boxes show that when estimating energy density, free space within the cell size should be considered by measuring the change in electrode thickness due to reversible and irreversible expansion until the thickness change tends to stabilize. The gray box in the Full Battery Evaluation box shows the amount of electrolyte (≤2 g/ah) and N/P ratio (≤1.1) used for battery evaluation in commercial battery designs. This industrial evaluation condition should be considered when evaluating practical silicon-containing cells in the laboratory.
Second, electrode specifications, including composition, area capacity, and porosity, should be as competitive as those of commercial graphite anodes. The use of silicon anodes may require more inactive ingredients such as binders and conductive carbon than graphite, as the volume variation and low conductivity of silicon anodes can reduce the energy density of the battery. Higher energy density is also required for higher area capacity, which can be achieved by minimizing inactive battery components such as current collector. As shown, electrode porosity also affects battery energy density. Appropriate electrode porosity values must be set to maximize battery energy density. In the half-cell test, it is difficult to determine whether the cycle stability of the silicon anode is good because the half-cell contains excess lithium and electrolyte. Therefore, we encourage researchers to use only half-cell test results to obtain basic performance information for silicon anodes, such as reversible capacity, initial coulomb efficiency, and short-term cycle life.
The third step is to measure the expansion of the electrode during a long cycle to determine its stable value, which can then be used to calculate the space margin inside the battery sleeve. Free space is the sum of the increments in electrode thickness produced by reversible and irreversible expansion until the expansion stabilizes. We recommend measuring the in-situ thickness with a battery dilator or micrometer after disassembling the battery in a fully lithiated state, rather than measuring the expansion with a cross-sectional scanning electron microscope (SEM). Electrode thickness observed with a scanning electron microscope can easily be mismeasured due to sample damage or local observation area.
Fourth, in order to estimate the energy density of the battery, it is necessary to provide information about the electrode density, electrode and diaphragm porosity, and the average mass load. As we discussed earlier, since silicon-containing cells can undergo severe solvation, the maximum electrode thickness after solvation should be considered when determining the number of electrode stacks. Note that each battery company can change these battery design parameters. To demonstrate the practical feasibility of silicon materials, we also encourage comparison of the energy density of silicon cells with the energy density of state-of-the-art commercial graphite cells with high loading, low inactive content and low electrode porosity.
Finally, a full battery test is required to evaluate cycle performance and calendar life aging. We recommend full battery testing with a small amount of electrolyte (approximately less than ≲2 g AH-1) for reliable results. In commercial battery designs, only a minimal amount of electrolyte is injected into the battery to fill the pores of the electrodes and separators. Excess electrolyte may produce seemingly good performance, but a large amount of electrolyte will not only reduce battery energy, but also take up a lot of volume in a battery with limited space. In addition to the amount of electrolyte, it is important to use an appropriate N/P ratio (N/P ratio of ≤ 1.1 for commercial graphite batteries), minimize the size difference between electrodes, use reasonable current density during cycling, and establish a wide enough voltage cut-off range to fully utilize battery capacity. Higher N/P ratios and large deviations in electrode size reduce energy density because less active material is used. As we discussed earlier, a suitable cut-off voltage range needs to be determined for a given silicon content to fully utilize the silicon's capacity (figure). We recommend investigating whether the experimental capacity values are close to the expected capacity calculated from the full battery design. Performance targets set by the U.S. Department of Energy call for a battery life (80% capacity retention rate) of more than 800 cycles and a specific battery energy of more than 325 Wh kg-1. We would like to remind researchers to pay attention to cycle life and energy density targets when conducting full-cell evaluations.
In terms of calendar life aging, experiments of up to a year may be required to obtain meaningful data. This long-term testing is difficult to perform in laboratory-scale experiments. There are several reliable alternatives to estimating calendar aging based on potentiostat retention, which requires shorter interface pairs when testing. We recommend that researchers carefully consider the voltage hold test conditions set by the DOE Silicon Alliance project, as well as the considerations reported by Schulze et al. and Kalaga et al. First, the cathode electrode must be designed with sufficient active lithium ions (e.g., N/P ratio less than 1) with negligible potential changes to prevent depletion of active lithium ion stockpiles to obtain a reliable anode potential profile. Therefore, the lithium iron phosphate cathode is highly recommended as an ideal positive electrode for voltage holding testing because it has a flat platform. Second, the qualitative voltage holding test will only be accurate if sufficient hold-up time is applied to relax reversible lithiation and the hold-hold test is dominated by a parasitic process. Finally, the choice of voltage range is also important, as each tested silicon anode may have a different operating potential.
conclusion
This work proves that for silicon anodes with specific specifications, the seemingly good performance parameters obtained do not guarantee their practical viability unless thorough testing is carried out to demonstrate that the anode meets all the requirements discussed in our test protocol. For example, certain strategies can improve the cycle life of silicon anodes, but they may not have the same benefits for calendar life. Therefore, it is necessary to consider all practical requirements when developing silicon anodes. We hope that this report will provide guidance for the development and evaluation of silicon anodes with practical performance specifications. We also hope that the report will have an impact on the development of other next-generation battery anodes, such as various alloy or conversion anodes and alkali metal (lithium, nickel and potassium) anodes, which face similar challenges to silicon anodes.
Kim, N., Kim, Y., Sung, J. et al. Issues impeding the commercialization of laboratory innovations for energy-dense Si-containing lithium-ion batteries. Nat Energy (2023). https://doi.org/10.1038/s41560-023-01333-5
Source: Electrochemical Energy
Note: Most of the articles reprinted on this site are collected on the Internet, and the copyright of the article belongs to the original author and the original source. The views in the article are only for sharing and communication, if it involves copyright and other issues, please let us know, I will deal with it in time!