Results

All the experimental results are provided in an Excel file containing different worksheets corresponding to each battery with distinct operating regimes, i.e., temperature and C-rate. In each of the tabs, there are three columns for every charge or discharge step that include the time [s], current [mA], and voltage [V] recording.

Protocol for reporting

The report is expected to follow specific structured guidelines to ensure meaningful interpretation of the data and reporting, given that only basic knowledge of electrochemistry is expected and that this opportunity is a first practical experience with applied electrochemistry in general, and more specifically with batteries. The final report is generally due within 2 weeks and given that English is often not the native language of the students, submissions in the English (US style) language are encouraged to practice academic writing, as well. The report includes the common sections of abstract, introduction, methodology, results and discussion, conclusions, and references. The guidelines instruct students to include six core graphs in the results section, where each of them is accompanied by brief discussion and responses to the guided questions. These core plots are chosen to reinforce the concepts of capacity retention, rate performance, overpotential, solid electrolyte interphase formation, and cell degradation over cycling.

Data analysis

Each of the six graphs focuses on a different aspect to guide toward adapting theoretical knowledge of basic electrochemistry into applied electrochemistry in the context of batteries, both qualitatively and quantitatively. Special emphasis is placed on aspects such as interpreting charge and discharge profiles, quantifying electrode capacity, understanding the difference between theoretical and practical capacity, comprehending the impact of C-rate and temperature on polarization losses and cell efficiency, and tracking degradation over cycling with analysis of possible causes.

The first graph, shown in Fig. 4a, depicts the charging and discharging profile of LCO at any fixed rate and temperature. The first and second cycles are the ones which are compared in this graph. The purpose of this section’s discussion is to address the irreversible capacity during the formation of the SEI, and the difference in the practical discharge capacity in comparison to the theoretical one. The first and most important step is to identify the special features related to the charge and discharge curves. The response of the voltage to current application is non-linear as the unique design of the electrochemical cell does not follow a purely resistor-like behavior. The processes in the cell are governed by polarization attributed to activation and mass-transport, and phase transitions. Possible phase transitions are oxidation of the transition metal of the cathode oxides or changing the stages of lithiation of graphite. In addition, the first cycle is characterized by a series of reductive electrolyte side reactions to form the SEI, exhibited as gradual change in voltage as the chemical potential of the surface and its resistance are changing over time.

Fig. 4

Six core graphs required for quality reporting: a charging profiles of LCO at any fixed temperature and C-rate, at 1st and 2nd cycles; b charging profiles of LCO at 1C rate for different temperatures, at 1st cycle; c charging profiles of LCO at room temperature for different C-rates, at 1st cycle; d charging profiles of LCO and LMO at room temperature for 1-C rate for any cycle besides the first; e discharge capacity and CE over cycling for LCO at any temperature and C-rate; f CE of LCO and LMO over cycling at room temperature for 1-C rate

Next, the gap in charge capacity and the subsequent discharge capacity should be addressed, or in other words the irreversible capacity. During the first cycles of the cell, the irreversible capacity mostly originates from the SEI formation, which in this case involves some contributions from the parasitic reductive side reactions, which do not take place in commercial Li-ion cells, e.g., water reduction.

The next step is to acknowledge the change in the gap between the charging and discharging curves and to associate it with polarization losses. Notably, while the discharge curve remains without any notable change, the charging curve majorly alters. The students must understand that the processes of intercalation and deintercalation are asymmetrical in the view of polarization losses. For instance, each process imposes different phase transitions requiring different energy for overcoming the activation barrier, different kinetic barriers for ion transport related to the diffusional path, different mechanical aspects such as stress and strain and even the effect of different electrostatic resistance at different states of charge.

The second graph (Fig. 4b) depicts the charging and discharging profiles of LCO at three different temperatures, at any given C-rate, and for the first cycle. The discussion in this section is intended to focus on the impact of temperature on the overpotential and practical capacity. Another emphasis here is on the influence of kinetics in comparison to recording rate; when the former is slower than the latter, unique features emerge, such as plateaus attributed to phase transitions imposed by the slower kinetics at low temperatures. One is expected to explain the lower discharge capacity at lower temperatures in terms of overpotential and ion transport with a focus on analyzing it through three polarization loss sources, i.e., the ohmic resistance, the activation polarization, and the concentration polarization.

Ohmic resistance is lower at higher temperatures. The electrolyte’s viscosity decreases, thus its conductivity is enhanced, as ionic transport via diffusion is faster. While the electrolyte resistance is not the only contributor to ohmic resistance, it is the bottleneck, as the resistance of conductive solids is generally considered lower. The activation polarization subsides with increasing temperature, as thermal energy assists in overcoming the energy barrier for the electrochemical reaction. In addition, these conditions also contribute to a lower viscosity of the electrolyte and, thus, faster mass transport, which minimizes the concentration gradient in the solution at the interface. This is why the risk for metallic Li plating is increased at low temperatures, in which the concentration polarization is dominant. Nonetheless, though higher temperatures may be beneficial in terms of capacity, they present challenges during long cycling, while these effects are seen over a longer term than herein. Higher temperature provokes particle cracking and high defect density [60].

The third graph presented in Fig. 4c shows the rate capability of LCO-based cells at 0.5 C, 1 C, and 2 C. Except for the first cycle, students must plot the charge and discharge profiles for varied rates at a given temperature. Once again, it displays the stepwise character of the profiles, imposed by the lowest C-rate. The trend of decreasing discharge capacity with increasing C-rates can be seen, which can be explained in terms of overpotential and ion transport. The focus should be on three main sources of polarization losses to analyze the results: ohmic resistance, activation polarization, and concentration polarization. The ohmic resistance is practically related to all sources of resistance in the cell, from all components, the electrodes, separator, and the electrolyte, noting that the relative resistance of the separator and the electrolyte compared to the electrodes is larger. The number of separator layers is actually a compromise between lower ohmic resistance and larger electrolyte intake. As the ohmic resistance depends linearly on the current, increased C-rate would practically mean increased ohmic losses and larger voltage drops. Activation polarization would also require higher overpotential, or driving force, to impose reactions to occur at higher rates. Concentration polarization, on the other hand, is eased when the cells are cycled at lower C-rate. The ion transport kinetics are relatively faster than ion depletion in the electrode–electrolyte interface, maintaining some concentration gradient which is not absolute. It is also important to notice that the double layer capacitance is less significant when cycling measurements are performed in galvanostatic mode, as its contribution is minimal to the total cell capacity compared to the faradaic processes.

The fourth graph compares cells with different cathode chemistry, i.e., LCO versus LMO. Figure 4d depicts charging and discharging curves for LCO- and LMO-based cells at a given temperature, at 1-C rate, and any fixed cycle. Every other variable is fixed in the system, except the identity of the cathode active material. Thus, the overpotential during intercalation and deintercalation is directly related to the lattice structure and Li-ion mobility therein. LCO has a layered structure while LMO has a spinel lattice structure. As the Li-ion cell relies on reversible intercalation and deintercalation processes, the ability of Li ions to diffuse into the lattice is crucial. The diffusional pathways of Li ions in these two distinct lattice structures directly affect the polarization losses during cycling. Lithium-ion transport in LCO is two-dimensional, mainly in the space between the CoO2 layers, while for LMO it is three-dimensional, within the interconnected pathways. This unique structure-related property changes the effective diffusion coefficients, i.e., Li-ion diffusion coefficient in LCO is generally lower than that in LMO [61, 62]. This property depends on Li-ion content; therefore, the diffusion coefficient drops as the vacancies are filled and limit the diffusion path. In LMO electrodes, the three-dimensional diffusion mechanism supports a rather faster transport kinetics, decreasing the losses due to concentration polarization. In this part, the students must designate an application to each cell based on power density against energy density. This is a chance to learn how each of these parameters is defined. Energy density is intuitively imagined as the amount of water in a bucket, while the power density is how fast the water from the bucket can be poured. In a more classical manner, the energy density and the power density are described mathematically in Eqs. 16 and 17 [10], where the nominal voltage is the average operating voltage. LCO has significantly higher discharge capacity, while LMO has slightly less polarization losses, providing LMO with higher nominal voltage. Thus, applications which require higher energy density would benefit from using LCO, whereas applications that require higher power density would preferentially choose LMO over LCO. From a structural point of view, the LCO’s layered structure is more spacious compared to LMO, which depends on its tetrahedral (8a) sites for active sites for Li ion intercalation and octahedral sites (16c) for hopping, limiting the practical hosting capacity [63].

$$}\left[ }}}}}} \right] = \frac}\left[ } \right] \times }\left[ } \right]}}}\left[ }} \right]}}$$

(16)

$$}\left[ }}}}}} \right] = \frac}\left[ } \right] \times }\left[ } \right]}}}\left[ }} \right]}}$$

(17)

The fifth graph (Fig. 4e) shows the discharge capacity of a cell while cycling at a fixed temperature and C-rate. The purpose of this part is to demonstrate the notion of capacity retention and to investigate the various causes of deterioration. The discussion is set to involve an analysis of the most likely failure modes in terms of inspecting the anode, cathode, and electrolyte behavior. Capacity decay is a result of multiple degradation mechanisms, encompassing all cell components, namely the anode, the cathode, the separator [64], and electrolyte deteriorations. One factor which is more general and related to operational aspects is the growing internal resistance due to SEI formation over the cycles and polarization losses. More particularly, the anode and the cathode are exposed to internal stresses due to ongoing intercalation and deintercalation, leading to structural instabilities and damage, such as exfoliation of the anode or lattice distortions in the cathodes. When overcharging, there are risks, such as dendritic growth of metallic Li on the anode, which may lead to loss of active material and short circuiting. In extreme cases, even corrosion of the current collectors may be detectable [65, 66]. When overdischarged, the anode may suffer from current collector corrosion, while the cathode may lose its structural stability, due to transition metal dissolution.

The final sixth graph (Fig. 4f) compares the coulombic efficiency of each cell using different cathodes, i.e., LCO and LMO, at room temperature and at a rate of 1 C. This figure provides insight into the modifications observed between the two active cathode materials in relation to the irreversible capacity resulting from SEI formation and the sensitivity of the active material to air-sourced contamination. This compels one to determine which chemistry provides greater stability throughout cycling. While transition metal dissolution from LCO cathodes is less dominant at voltages below 4.2 V, their dissolution from LMO cathodes is a major challenge when batteries are cycled above 4 V, as they are subjected to Jahn–Teller distortion [36, 67]. Upon overcharging, or full de-lithiation of Li ions from the cathode, the manganese ions are at an oxidation state of + 4. Water contamination leads to oxidative deterioration of the electrolyte and formation of acidic byproducts, e.g., HF, that catalyze reduction of manganese ions to an oxidation state of + 3. These species are unstable, and readily disproportionate to form ions of oxidation states of + 2 and + 4. The produced manganese ions with oxidation state of + 2 are soluble, and deposited over the negative electrode, catalyzing further electrolyte reduction and SEI formation [68]. The formed SEI is also considered to be thicker and therefore increases the cell’s overall impedance.

The “3-Rs” and where things may go wrongReplicability, repeatability, and reproducibility

Ensuring replicability, repeatability, and reproducibility (the 3-Rs) is critical for proving the scientific significance of the experimental results. Repeatability is the ability of the same working team to produce the same outcomes within the limits of precision using specific equipment and procedure. Unfortunately, the short time scheduled for the lab meeting requires some sacrifice of repeatability, as the students only construct six batteries, one cell for each test. It means that the validity of the results is questionable, and that the results may be untrustworthy in terms of randomness, leading to possibly incorrect conclusions being drawn. This issue is resolved by conducting a preliminary review of the results data before sending it to the students, ensuring that no unexpected artifacts are present and that the observed trends align with expectations more or less. If a major deviation is detected, an alternative dataset is provided so that meaningful analysis can still be performed.

Replicability refers to multiple teams achieving the same outcomes within the limits of accuracy, using the same equipment and methodology. According to the data collected over a 7-year period, with about 20 groups participating each year, the experiment can be replicated. This component helps to increase the procedure’s credibility.

Reproducibility refers to the ability for multiple teams to achieve the same outcome, within the limits of accuracy using different equipment or procedures. In this case, even though the methodology is the same and detailed to minimize variation, human and material or equipment issues may still interfere and impose different conditions. The human aspect refers to the students’ assembly precision, expertise gained via hands-on previous lab experience, and level of confidence. When it comes to the materials, the storage conditions are crucial to avoid any potential physical or chemical changes, resulting from exposure to a non-inert atmospheric environment, temperature fluctuations, or mechanical damage. Chemical changes, such as introduction of contamination, moisture entrapment, and oxidative degradation, would affect the acquired capacity, coulombic efficiency, voltage, and cycle life, as the chances for side reactions and resistance increases are high. Physical changes include peeling, denture or scratch-induced point and line defects, and creasing-induced surface heterogeneity, boosting the localized current density. These types of changes can result in a variety of scenarios, including loss of electrical contact, increased interfacial resistance, localized heating, and dendritic Li metal growth, leading to short-circuiting. Other less severe physical deviances are caused by the punching tool. Since these tools are typically utilized by several lab members for various purposes, the likelihood of cross-contamination may increase. Furthermore, given that punching tools are frequently utilized in the lab, they may become inefficient due to a loss of form or sharpness. This results in the production of electrode discs with rough edges such as rollovers or other types of burs, in addition to chipping. These characteristics may be deceiving when estimating theoretical capacity, since the real surface area differs from the geometric one, imposing errors when calculating some experimental parameters. From another perspective, the rough edges can cause unequal pressure or even separator puncturing, further interfering with the cell performance. However, proper storage and handling can help alleviate these issues: this involves pre-cutting the electrode sheets into smaller pieces and keeping these pieces between two pieces of glass or plastic in an airtight container with desiccant or in an inert environment. Moreover, the tools should be examined before use and replaced on a regular basis if they present signs of damage to their integrity.

Intrinsic error sources

This lab experiment is designed to work in an atmospheric environment of a glove box, which most of the students have never worked with. It requires additional safety training which includes an introduction to the main working chamber and antechamber structure, and the inert gas system and its control. This is performed to minimize environmental contaminants, and it involves some practice and training in techniques and reciprocal coordination. Nonetheless, the work in a non-inert environment has its cost as the nonaqueous solvents and Li salts comprising the electrolyte are sensitive: first, there is the risk of electrolyte oxidation due to overcharging, resulting in decomposition into passivating and pore-blocking byproducts, with increased risk of cell swelling due to gas evolution. This problem is solved by limiting the upper voltage cutoff to 4.2 V and by choosing an EC/DMC mixture, which is considered more resistant to oxidation. Second, the electrolyte is susceptible to undesired reduction processes as well. While ideally, under cathodic polarization, an SEI is formed on the anode’s surface, this process is only beneficial if it is confined to an initial short period of time. Other reductive processes may occur in the presence of oxygen, water, and carbon dioxide, resulting in major increases in interfacial resistance, due to formation of Li oxide, Li hydroxide, and Li carbonate (Eqs. S2–S4), all acting as passivating layers. Water contamination is not only detrimental to solvents based on alkyl carbonate but also to the Li salt itself. Electrodes stored in an open package adsorb 100 to 150 ppm of water, while commercial electrolytes have water content of up to 20 ppm, as verified by Karl Fischer titration. For instance, PF6− anions degrade oxidatively into PF5, which is a strong Lewis acid, reacting with water, to produce HF, which is detrimental to the cathode stability (Eq. S5) [69,70,71,72,73]. Moisture can be also detrimental to the current collectors, i.e., aluminum and copper. The resulting corrosion products may increase the contact resistance in the cell. This process is exacerbated in the presence of water residues, as they catalyze the formation of strong acidic byproducts, e.g., HF [74] (Eq. S6). Upon over discharging, copper may experience localized corrosion in the form of pitting, producing soluble copper ions (Eq. S7), which may redeposit and pierce the separator [73]. In addition, gas inclusions unfortunately cannot be excluded in these cases. The steps to minimize these side reactions include prior drying of the electrodes in a vacuum oven and pulling the sensitive components (such as electrodes and electrolytes) from the protective storage (glove box) in small portions and near the execution time. Water contamination in dried electrodes is lowered to less than 50 ppm, and in a properly stored electrolyte, i.e., in the glove box, to below 20 ppm. If possible, disposable glove bags or compact chambers filled with inert gas are recommended. Another alternative that may decrease the sensitivity to air is to replace the LiPF6 salt with LiBF4, which hydrolyzes relatively slowly, but it may lead to a compromise on properties such as conductivity and static dielectric coefficient and may be pricy [75, 76]. Other issues such as self-discharge are insignificant in this case, as the cells are not delayed longer than a minute in charged state at open cell potential.

Common failure routes and diagnostic examples

In practical implementation of the lab, some of the assembled cells display early failure to deliver the required (electrochemical) performance. As a constructive discussion, the diagnosis of the typical failure modes being encountered are presented in this section. These include mechanical assembly errors, e.g., partial or off-axis crimping or insufficient pressure, electrolyte degradation, and internal component corrosion. Some coin cells fail to operate adequately even after only ten cycles while the core issues, associated with mechanical assembly errors, are instantly apparent from a visual standpoint. For example, the coin cell might be faulty owing to poor crimping. Improper crimping can be identified by partly crimped coin cells, off-axis crimping, and unintended casing curvature. Although these mechanical distortions may appear insignificant to an undergraduate student, it is important to emphasize that they are going to have an impact on the electrochemical performance. Here, we will cover some of the major failure modes and the reasons for their appearance as well as how to mitigate them.

Figure 5a illustrates a partial and off-axis crimping. The first refers to improper sealing, which contributes to electrolyte evaporation, electrolyte exposure to the ambient environment, as well as possible leakage, which can lead to corrosion of the casing. Electrochemically, it might appear as poor cycle stability and a significant drop in the discharge capacity due to an increase in internal resistance. The latter indicates nonuniform contact pressure: this condition implies that current density is unevenly distributed, which may exacerbate electrode deterioration and irreversible capacity loss. Figure 5b shows that in certain circumstances, even when the crimping was effective and the casing was sealed, some curvature occurred on the casing due to improper compression. In practice, this implies that the contact between the coin cell’s components may be disrupted, resulting in a significant increase in contact resistance, accompanied by local electrolyte starvation.

Fig. 5

Different failure modes of coin cells: a off-axis partial crimping; b mechanically distorted casing; c green stained separator; d current collector corrosion; e dried cell; f purple stained components; g gel-like formation; h lack of separator; i a cracked electrode; j delaminated electrode, k cell with LMO cathode

When disassembled, coin cells are examined and colored byproducts may reveal additional failure mechanisms. For instance, as shown in Fig. 5c and d, inspection of the components of some of the dismantled coin cells revealed that the separator is colored by a green hue. Once the components have been completely separated and the back side of the current collector unveiled, greenish-blue precipitates are evident. Phenomenologically, this indicates that copper was dissolved and precipitated as salt. Water contamination is extremely likely, since the electrolyte is in contact with the lab’s ambient environment. The formation of a strong acidic byproducts when LiPF6 decomposes, as well as the generation of carbon dioxide via SEI decomposition, may both result in the precipitation of copper carbonate, whose color matches the observation. These chemical changes may affect not only the cell’s internal resistance but also lead to cell swelling, and loss of lithium inventory via continuous destabilization of the SEI, resulting in an increase in the portion of irreversible capacity loss in each cycle, up to total electrolyte decomposition and starvation. The specific dismantled cell in Fig. 5e demonstrates that the separator is dry, supporting the diagnosis.

Another phenomenon involves disassembled cells with a purple stain on the separator (Fig. 5f), but in this case, the electrolyte is not completely consumed and instead, alters consistency and becomes gel-like (Fig. 5g). The proposed analysis for the mechanistic failure route in this case is based on the acceleration of EC ring-opening, which is catalyzed by LCO as a strong Lewis acid in the presence of water contamination. The byproducts lead to exacerbation of cobalt site activation on the LCO electrode surface for further ring-opening, which serves as a platform for polymerization, leading to the formation of gel. Meanwhile, excessive leaching leads to the formation of organometallic complexes even though the overcharging is limited by setting a cutoff of 4.2 V [77, 78].

Other typical faults that might be identified throughout disassembly include a lack of critical components or defective electrodes. For example, even when students are instructed to arrange all components prior to assembly, some continue to assemble without sufficient ordering of the workstation, resulting in the absence of some critical components. Figure 5h demonstrates this exact instance. Since a separator was absent during cell assembly, both electrodes are directly contacting, resulting in a short circuit.

In other circumstances, the electrodes are flawed or deteriorated due to repeated cycling. Figure 5i and j depict how electrodes deteriorate via adhesion loss and cracking. During cycling, shear stress is generated by lattice deformation and phase changes. These stresses produce swelling, weakening particles and binder bonding and can cause delamination of the active material from the current collector.

Furthermore, as shown in Fig. 5k, brownish staining of the separator in cells based on LMO cathodes is frequently observed when cells are dismantled. In the presence of water traces, enhancing the electrolyte acidification, which in combination with high oxidation potential, leads to conditions favoring disproportionation of Mn3+ to Mn2+ and Mn4+[79,80,81]. Mn2+ is soluble and migrates through the separator to the anode side. When the ions encounter oxidizing electrolyte radical fragments in the separator, they oxidize and precipitate as brownish mixed valence oxides (MnOx).