State of Health (SOH)

As our world becomes increasingly dependent on battery-powered technology, from electric vehicles to grid-scale energy storage, understanding a battery's longevity is no longer an academic curiosity—it's a critical economic and engineering imperative. The key to this understanding is a concept known as the State of Health (SOH). However, SOH is often oversimplified or confused with the more familiar State of Charge (SOC). This article aims to demystify SOH, providing a clear and comprehensive guide to what it truly represents and why it matters. The following chapters will navigate from the fundamental principles to real-world applications. First, in "Principles and Mechanisms," we will dissect the dual nature of SOH, exploring how both capacity fade and resistance growth define a battery's health, and examine the clever techniques used to measure these hidden properties. Subsequently, "Applications and Interdisciplinary Connections" will reveal how SOH serves as a crucial tool in engineering design, real-time control systems, and complex economic decisions, bridging the gap from electrochemistry to sustainable technology management.

Imagine you have a trusty water bottle that you take with you everywhere. When it was new, it held a full liter of water. After a few years of bumps and scrapes, perhaps it has a small dent, and now it only holds a little less, say, 900 milliliters. It still works, but its capacity has diminished. In a nutshell, this is the most basic idea behind a battery's State of Health (SOH). It's a measure of "how good" a battery is now, compared to when it was brand new.

The most common way to think about SOH is just like with our water bottle: we measure the battery's current maximum capacity and compare it to its original, rated capacity. If a new electric scooter battery could run for 10 hours at a certain current, but after a year it can only run for 8.8 hours under a similar test, we can say its capacity has faded. We would define its capacity-based SOH as the ratio of its current capacity to its original capacity. In this case, the SOH would be $0.88$, or $88\%$.

This seems simple enough. But a battery is more than just a storage tank for charge. It also has to deliver that charge on demand. Let's return to our water bottle. What if, besides the dent that reduced its volume, the nozzle has become partially clogged with mineral deposits over time? Even with a full bottle, you can't get the water out as quickly as you used to. The flow rate is restricted.

This is the second crucial aspect of a battery's health: its internal resistance​. Every battery has some internal resistance, a bit like electrical friction, which opposes the flow of current. When the battery is new, this resistance is very low. As it ages, chemical changes inside the battery cause this resistance to increase—the "nozzle" gets clogged. This rise in resistance limits the battery's ability to deliver high power.

We can, therefore, define a second kind of SOH, one based on power capability. Since higher resistance means lower power capability, we define this resistance-based SOH inversely. A healthy battery has low resistance, so we put the new, low resistance value in the numerator:

A battery whose resistance has doubled would have a resistance-based SOH of $0.5$, or $50\%$. So, we see that a battery's health isn't just one number; it's at least two. One tells us about energy (how much charge it can store, i.e., capacity) and the other tells us about power (how fast it can deliver that charge, i.e., resistance).

This distinction between energy and power capability is not just an academic exercise; it's profoundly important for real-world applications. Imagine two old electric vehicle batteries being considered for a "second life" in a stationary energy storage system.

One battery has a capacity SOH of $0.85$ but a resistance SOH of only $0.60$. This means it still holds a good amount of energy, but its internal resistance has increased significantly. It would be a poor choice for an application that needs to deliver huge bursts of power, like stabilizing the grid during a sudden surge in demand. The high internal resistance would cause the voltage to sag dramatically under a high current load, making it unable to meet the power requirement. However, it might be perfectly suitable for a home energy system, slowly charging from solar panels during the day and discharging slowly to power lights and appliances at night.

Now consider the second battery, with a capacity SOH of only $0.70$ but an excellent resistance SOH of $0.85$. This battery can't store as much total energy, but it can deliver what it has very quickly and efficiently. It would be unsuitable for long-duration energy backup but could be ideal for that grid stabilization task, where providing high power for short periods is the primary job. "Health," it turns out, is not an absolute measure; it's relative to the intended job.

Given that a battery's health has these two major facets, can we combine them into a single, unified SOH score? This is a question that engineers grapple with constantly.

A simple approach might be a weighted average of the two SOH values. But what should the weights be? As we've just seen, the importance of power versus energy depends entirely on the application. The brilliant insight is to make the weights themselves dependent on the application's needs. For a task that is "power-intensive," we would assign a larger weight to the resistance-based SOH. For an "energy-intensive" task, we would give more weight to the capacity-based SOH. A single, universal SOH number is less meaningful than a task-specific "utility score."

However, there's a more robust way to think about a single health metric, especially when it comes to deciding when a battery has reached its End-of-Life (EOL). For most applications, a battery is considered "dead" if either its capacity drops below a certain threshold (e.g., $80\%$ of new) or its internal resistance rises above a certain threshold (e.g., grows by $50\%$). Notice the critical word: or​.

You cannot compensate for a dangerously high internal resistance with a slightly better capacity. A high resistance not only limits power but also generates excessive heat ($P_{\text{loss}} = I^2 R$), which can be a safety hazard. This "no compensation" rule is a fundamental principle of battery safety and management.

How can we capture this logic in a single equation? A simple sum is no good, as it would allow for compensation. The elegant solution is to use the maximum function. We can define a normalized "damage" from capacity loss and a normalized "damage" from resistance increase. The overall EOL indicator is the larger of these two damage values. For instance, if EOL is defined as $20\%$ capacity loss or $30\%$ resistance growth, the indicator $s$ could be:

The battery reaches its EOL when this indicator $s$ reaches or exceeds $1$. If the capacity is fine but the resistance has grown by $30\%$, the second term becomes $1$, the max is $1$, and the battery is retired. This mathematical formulation perfectly captures the non-negotiable nature of the EOL criteria.

So far, we've talked about SOH as if capacity and resistance are numbers we can just read off a screen. But how do we actually measure them, especially in a device that's in use? We can't see the health directly; it is a latent variable​, a hidden property that we must infer from clues. This is where the detective work of battery engineering begins.

First, we must be careful not to be fooled by temporary conditions. A battery's performance is highly dependent on temperature. When a battery is cold, the chemical reactions inside slow down, and the electrolyte becomes more viscous. The result? Its available capacity temporarily decreases, and its internal resistance temporarily increases. This doesn't mean the battery has permanently aged overnight. To get a true measure of health, we must correct for temperature effects. We can use mathematical models—often a linear correction for capacity and a more complex Arrhenius model for resistance—to estimate what the capacity and resistance would be at a standardized reference temperature, say $25\,^{\circ}\mathrm{C}$. Only by comparing values at the same reference temperature can we track true, irreversible degradation.

With temperature effects accounted for, we can use more sophisticated techniques to probe the battery's internal state. One of the most powerful is Incremental Capacity Analysis (ICA). If you very slowly charge a battery and plot the tiny amount of charge you add ($dQ$) for each tiny step in voltage ($dV$), you get a $dQ/dV$ curve. This curve is far from flat; it has distinct peaks and valleys that act like a fingerprint or an EKG for the battery. These features correspond to specific chemical and physical transitions happening at the electrodes.

Scientists have discovered that as a battery ages, particularly from the irreversible loss of lithium ions to side reactions, these peaks in the $dQ/dV$ curve shift their position along the voltage axis. Remarkably, for certain peaks, the amount of the voltage shift, $\Delta V$, is directly proportional to the amount of capacity lost, $\Delta Q$. A simple relationship, $\Delta Q \approx H \cdot \Delta V$ (where $H$ is the height of the peak), allows us to estimate the lost capacity just by tracking the voltage of this fingerprint feature. It’s a wonderfully clever, non-invasive way to diagnose a specific disease mechanism inside the battery without ever opening it up.

Finally, let's clear up a common and crucial point of confusion. State of Health (SOH) must not be mistaken for State of Charge (SOC).

Think of it this way:

• SOH is the size of your fuel tank. A new battery has a 10-gallon tank (SOH = $100\%$). An old battery has an 8-gallon tank (SOH = $80\%$).
• SOC is the fuel gauge. It tells you how full your current tank is, from $0\%$ to $100\%$.

When your brand-new battery (10-gallon tank) is fully charged, its SOC is $100\%$, and it holds 10 gallons. When your old battery (8-gallon tank) is fully charged, its SOC is also $100\%$, but it only holds 8 gallons. The "100%" on the fuel gauge is relative to the current size of the tank.

This distinction is vital for any battery-powered device. The Battery Management System (BMS) needs to know the SOH to understand what "100%" SOC even means. Without an accurate SOH estimate, the range prediction on an electric car or the remaining runtime on your laptop would be wildly inaccurate. The journey of understanding a battery's health is a journey from simple analogies to the deep complexities of its inner electrochemical world, a journey essential for building the reliable, long-lasting energy storage of our future.

Having peered into the microscopic world to understand the principles and mechanisms of a battery's State of Health (SOH), we now zoom out to see where this powerful concept truly comes to life. SOH is far from an abstract academic score; it is a vital, practical number that acts as a compass for engineers, a crystal ball for economists, and a set of operating instructions for the intelligent systems that manage our energy future. It is the bridge between the electrochemistry of a single cell and the complex decisions that shape our technological world. In this journey, we will see how the simple idea of "health" helps us design better products, operate them safely, predict their future, and build a more sustainable economy.

At its most fundamental level, SOH is an engineer's tool for characterizing and predicting behavior. Imagine you are designing a new wireless power tool. You need to guarantee it will perform for, say, three years of typical use. How do you do that? You can't just wait three years. Instead, you perform accelerated aging tests, cycling batteries under various conditions and meticulously measuring how their properties change. You'll notice two primary symptoms of aging: the total energy they can store shrinks (capacity fade), and it becomes harder to get power out of them quickly (internal resistance growth).

A sophisticated definition of SOH combines both of these effects. The health isn't just about how long the battery lasts, but how well it performs throughout its life. For instance, we can define SOH as a product of its remaining capacity and its remaining power capability. Since the maximum power a battery can deliver is inversely related to its internal resistance—a consequence of the simple and beautiful Ohm's Law—a rise in resistance directly hurts its health from a power perspective. By fitting the experimental data to physically-motivated aging models, such as those where degradation scales with the square root of the number of cycles, engineers can create a formula that predicts the SOH for any given number of charge-discharge cycles. With this formula, predicting the battery's end-of-life becomes a straightforward calculation, allowing for a robust and reliable design.

This raises a more subtle and crucial point: the "health" of a battery is not absolute. It depends entirely on the job we ask it to do. A battery cell that can no longer power a demanding electric vehicle might be perfectly healthy for a less-strenuous second life in a home energy storage system. This idea of an application-specific SOH is central to the economics of battery recycling and reuse. To screen a used battery for a second-life task, an engineer might define a new SOH metric. This metric could be the more restrictive of two conditions: its remaining energy capacity, and its ability to deliver a required current pulse without the voltage sagging below a critical threshold. A cell could have 80% of its original capacity but fail the test because its increased internal resistance causes too large a voltage drop at high currents. For a high-power application, its SOH would be deemed too low. For a low-power application with no such current pulses, its SOH, judged purely by capacity, might be a perfectly acceptable 80%.

This concept of heterogeneity becomes even more critical when we assemble many cells into a large battery pack, as in an electric car. If you connect cells in series, they all must carry the same current. If one cell has a lower SOH—and thus a lower capacity—it will be the first to run empty during discharge. A smart Battery Management System (BMS) will detect this single cell hitting its low-voltage limit and shut down the entire pack to protect it. The result? The capacity of a billion-dollar electric bus's battery pack is dictated by its single weakest cell. This "weakest link" phenomenon is a major challenge, especially when using recycled cells with varying histories. Engineers have devised clever mitigation strategies, from carefully sorting cells into "bins" of similar SOH before assembly to designing active topologies where a depleted cell can be electronically bypassed, allowing the rest of the pack to continue operating. SOH is therefore not just a property of a cell, but a key consideration in the architecture of the entire system.

So far, we have spoken of SOH as something we measure in a lab or plan for in a design. But what about a battery that is already out in the world, working inside your phone or car? Its SOH is constantly, albeit slowly, changing. How do we track it in real time? The answer lies in the silent, tireless work of the Battery Management System (BMS)—the battery's electronic brain.

Modern BMSs run a "digital twin," a sophisticated software model that mirrors the physical battery's state in real time. This model is continuously updated using data streaming from sensors that measure the battery's voltage, current, and temperature. Using powerful algorithms borrowed from control theory, like the Extended Kalman Filter (EKF), the BMS can perform a remarkable feat of inference. It can simultaneously estimate the battery's "fast" states, like its State of Charge (SOC), and its "slow" SOH parameters, like its true capacity and internal resistance. This is done by augmenting the state vector to include these slowly changing health parameters and tracking how they evolve.

This process reveals a beautiful separation of time scales. The SOC can change from 100% to 0% in a matter of hours. The SOH, on the other hand, might drift from 100% to 80% over many years. The BMS must manage both. It uses a fast estimation loop, running many times per second, to keep track of the SOC. On a much slower timescale, perhaps using data aggregated over days or weeks, a prognostics module updates the SOH parameters. It might use a physical model of degradation that accounts for the stresses the battery has experienced. The estimated SOH is then fed back into the fast loop, because a battery's voltage response (which the SOC estimator relies on) changes as it ages. This intricate dance between fast state estimation and slow parameter prognostics is the heart of a modern Cyber-Physical System for battery management. The SOH is not just a number to be reported; it is an active parameter that allows the BMS to adapt its control strategies, ensuring the battery is operated safely and efficiently throughout its life.

Once we can model and predict SOH, a whole new world of strategic decision-making opens up. SOH becomes the key variable in optimizing performance, managing risk, and making profound economic choices.

For a battery designer, SOH models provide a kind of crystal ball. By creating a mathematical model that links degradation to stress factors like temperature and charging speed (C-rate), we can perform a sensitivity analysis​. By calculating the partial derivatives of SOH with respect to these factors, we can ask precise questions: "How much faster will the battery degrade if we allow it to operate 5 degrees hotter?" or "What is the SOH penalty for cutting the charging time in half?" This allows designers to navigate the fundamental trade-off between performance and longevity, tuning a battery's operating conditions to meet the specific needs of an application.

This predictive power scales up from a single battery to an entire fleet. Imagine managing thousands of electric delivery vans. Some will be driven gently, others harshly. Some will be fast-charged frequently, breaching recommended current limits. We can build a probabilistic model that combines a baseline degradation rate with the expected impact of these random "shock" events. This allows us to compute the expected SOH trajectory for the entire fleet, forecasting what percentage of batteries will need replacement in one, two, or five years. This moves SOH from a deterministic property of one device to a statistical tool for large-scale risk management.

The implications are even more profound as batteries become active players in our energy grid. In Vehicle-to-Grid (V2G) systems, an EV owner can sell power back to the grid during peak demand. To do this safely and profitably, the grid aggregator needs to know three things about the battery: its State of Charge (how much energy is available?), its State of Health (how much stress can the battery handle?), and its State of Power (how much power can it deliver right now​?). SOH is the long-term context for the short-term decisions. A battery with a lower SOH will have a higher internal resistance, limiting its power capability and making aggressive V2G dispatch riskier and less efficient.

Finally, SOH provides the quantitative foundation for the most critical economic decisions in a battery's life cycle. For large, expensive energy storage assets, we can use SOH in sophisticated optimization models to decide the most economical time for replacement. This involves a mixed-integer program where a binary decision variable—to replace or not to replace—is triggered when the SOH is projected to fall below a minimum threshold. The model then weighs the high upfront cost of replacement against the ongoing costs and reduced performance of operating an aging asset, finding the strategy that minimizes the total cost of ownership over decades.

This culminates in the ultimate question at the end of a battery's first life: should it be repurposed or recycled? This is not a question of sentiment, but of economics. By building a Net Present Value (NPV) model, we can rigorously compare the two options. The "reuse" path involves an upfront refurbishment cost but promises a future stream of revenue, a stream that is both enabled and limited by the battery's SOH. However, this path also carries risk—the battery might fail earlier than expected, incurring a penalty. The "recycle" path offers a smaller but certain immediate payout. By equating the NPV of these two paths, we can solve for the threshold SOH at which the decision flips. This calculation, which beautifully links electrochemistry to finance, tells us precisely when a battery's future potential is no longer worth the risk, providing a rational basis for building a circular economy.

From the engineer's bench to the financier's spreadsheet, the State of Health is the unifying concept that allows us to understand, manage, and optimize the technology that powers our world. It is a testament to how a deep understanding of fundamental science can illuminate the path to smarter engineering and a more sustainable future.