Differential Capacity Analysis for Degradation Mechanics in Prismatic Lithiated Iron Phosphate

Signal

Discharging a 280 Ah lithium iron phosphate cell at a constant 0.05C current draws roughly 5.6 amperes over 20 hours, recording discrete voltage points at millivolt increments. Translating this raw voltage timeline into specific degradation mechanics requires converting voltage and capacity measurements into differential functions. Differential capacity analysis takes the derivative of capacity with respect to terminal voltage (dQ/dV), while differential voltage analysis uses the inverse derivative (dV/dQ).

Both methods convert subtle shifts along open-circuit voltage plateaus into distinct spectral features. Compared to nickel-based chemistries, lithium iron phosphate poses distinct signal processing challenges because the positive electrode’s open-circuit voltage profile is exceptionally flat.

During two-phase lithium insertion and extraction in the positive electrode matrix, terminal voltage barely moves across broad spans of state of charge. As voltage change over time approaches zero, the denominator in the dQ/dV calculation drops toward zero, generating narrow differential capacity peaks with extreme amplitudes. Minor high-frequency voltage ripples, cabinet thermal drifts, or quantization steps from analog-to-digital converters can generate noise that easily drowns out real electrochemical features.

Bench characterization requires high-resolution cyclers equipped with 16-bit or 24-bit voltage cards; standard commercial pack testers with 12-bit resolution create digitization steps that distort peak height and area measurements.

Voltage Sampling Resolution and Derivative Mathematics

Calculating differential capacity relies on discrete derivative approximations between consecutive points. Point-to-point numerical differentiation directly on raw voltage data amplifies noise heavily, as small voltage measurement errors get magnified by several orders of magnitude in the difference ratio. Extracting clear features requires strict control over data acquisition settings during galvanostatic cycling.

Current stability directly dictates derivative accuracy: even a 0.1 percent current drift during a low-rate discharge tilts the dQ/dV baseline and creates false peaks near phase transitions. Data acquisition should maintain a sampling density above 20 voltage readings per millivolt. Sampling at equal voltage intervals rather than fixed time steps keeps data volume manageable while preserving point density through steep transitions.

On flat OCV plateaus, fixed-time sampling logs thousands of identical voltage readings, which drives numerical derivatives into instability.

Smoothing Algorithms and Data Conditioning

Raw galvanostatic curves need non-destructive filtering to clear measurement artifacts without blunting true electrochemical peaks. Savitzky-Golay filtering uses localized polynomial fits across a moving window, preserving higher-order derivative moments far better than simple moving averages. Setting the right polynomial degree and window width determines how cleanly noise is suppressed without distorting peak geometry.

Too narrow a window leaves high-frequency quantization spikes in the dQ/dV curve, misrepresenting noise as phase transitions. Too wide a window flattens real electrochemical features, lowering peak heights and artificially inflating full width at half maximum. A second- or third-order polynomial across a 15- to 31-point window preserves peak areas along LFP voltage plateaus.

Gaussian filters and cubic smoothing splines work as alternatives, assuming the smoothing parameter is matched to the noise floor of the cycler channel.

Filtering parameters must remain identical when comparing fresh and aged cells. Changing window settings or algorithm parameters between baseline and later cycles introduces artificial changes in calculated peak areas. Diagnostic software should retain raw voltage, current, and time arrays alongside filtered derivative data to ensure full auditability for warranty reviews.

Diagnostic differential capacity sweeps rely on low constant-current rates, typically between C/20 and C/50. Higher rates introduce noticeable internal resistance drops (IR polarization), which shift peak locations along the voltage axis and broaden features as concentration gradients form inside active material particles. Testing at C/50 minimizes overpotential so terminal voltage reflects the thermodynamic open-circuit potential as closely as possible.

Tight temperature control during 20- to 50-hour diagnostic runs is essential for baseline stability. Environmental chambers should maintain ambient cell temperatures within plus or minus 0.2 degrees Celsius. Thermal variations shift the open-circuit voltage through the entropic coefficient, displacing plateaus and generating derivative artifacts that look like active material loss.

Precise temperature control is the single most critical environmental requirement when recording bench DCA data.

Peak

Galvanostatic voltage curves for graphite paired with lithium iron phosphate break down into distinct plateaus governed by phase transformations at both electrodes. Interpreting differential capacity spectra depends on assigning each peak to its underlying thermodynamic step. In a full prismatic cell, terminal voltage measures the potential difference between the LFP positive electrode and the graphite negative electrode.

Because LFP holds a nearly flat potential near 3.42 V versus metallic lithium over most of its operating window, changes in full-cell terminal voltage primarily mirror phase changes inside the graphite anode.

Graphite intercalates lithium through distinct structural stages, each defined by a specific stoichiometry and chemical potential plateau. As lithium enters the graphite lattice during charge, the anode transitions through higher-order staging phases down to fully lithiated LiC6. These phase transitions register as sharp features on differential capacity curves, establishing the reference profile needed to track structural degradation over time.

Thermodynamic Phase Transitions in Lithiated Graphite

Graphite staging proceeds through four well-characterized stoichiometry regions: dilute Stage 4, Stage 3, Stage 2L, Stage 2, and fully lithiated Stage 1 (LiC6). As lithium deintercalates from the anode during discharge, the resulting terminal voltage shifts produce three main peaks in the dQ/dV spectrum.

Peak I appears at the lower end of discharge voltage, typically between 3.05 V and 3.12 V, representing transitions among dilute graphite stages. Peak II shows up between 3.20 V and 3.25 V as the structure moves between Stage 3 and Stage 2 intermediate phases. Peak III dominates the upper spectrum between 3.32 V and 3.38 V, marking the high-capacity transition between Stage 2 (LiC12) and Stage 1 (LiC6).

Because Peak III accounts for roughly half of the anode’s total cyclable lithium capacity, tracking its area and geometry gives a direct look into negative electrode health.

A C/50 discharge profile at 25°C resolves the primary graphite staging peak at 3.328 V with a full width at half maximum of 14 millivolts.

Iron Phosphate Phase Coexistence and OCV Plateaus

The positive electrode undergoes a two-phase reaction between heterosite iron phosphate (FePO4) and triphylite (LiFePO4) at a constant thermodynamic potential of 3.42 V versus Li/Li+. Because this cathode potential plateau is so flat during low-rate cycling, the positive electrode effectively serves as an internal quasi-reference electrode.

Differential voltage curves (dV/dQ) invert this perspective, presenting sharp peaks where differential capacity curves show flat valleys. Peak positions in dV/dQ align with the capacity boundaries of specific graphite staging steps. Combining dQ/dV and dV/dQ curves lets battery engineers determine whether a given terminal voltage shift stems from an electrode capacity offset or an actual loss of active sites.

Fresh prismatic cells display sharp, reproducible peak signatures across production lots. Small variations in initial peak height usually reflect minor manufacturing tolerances in coating weight, active mass ratios, or formation efficiency. Recording a baseline dQ/dV curve during incoming lot qualification establishes the benchmark needed to evaluate operational aging non-destructively.

Tracking these features over hundreds of cycles reveals subtle shifts in peak area ratios long before full-cell discharge capacity falls below specification. A drop in peak height without horizontal displacement indicates uniform material loss, whereas horizontal peak movement along the voltage axis points to rising internal resistance or shifts in thermodynamic balance. Mapping these spectral changes to physical degradation mechanisms forms the core of quantitative aging diagnostics.

Capacity fade rarely stems from a single isolated cause; loss of cyclable lithium and structural degradation of active materials usually progress together over a cell’s operating life. Low-rate sweeps strip away kinetic overpotentials to highlight the underlying thermodynamic state of each electrode.

Baseline reference profiles logged at 25 degrees Celsius are frequently required in procurement contracts for utility-scale storage. Without a verified initial dQ/dV signature, separating manufacturing defects from operational abuse during field capacity claims often leads to protracted legal disputes.

Accurate peak alignment relies on precise voltage calibration across test channels. Calibration drift exceeding 1 millivolt can obscure subtle thermodynamic shifts and undermine diagnostic confidence.

Mechanics

Quantifying capacity degradation over extended cycling requires distinguishing active lithium loss from structural degradation of the electrode matrices. Differential capacity analysis isolates these mechanisms without destructive cell teardowns. In LFP prismatic cells, degradation generally divides into three primary modes: Loss of Lithium Inventory (LLI), Loss of Active Material at the Negative Electrode (LAM_NE), and Loss of Active Material at the Positive Electrode (LAM_PE).

Rising internal ohmic resistance adds a uniform voltage offset on top of these thermodynamic shifts.

Loss of lithium inventory is typically the main driver of capacity fade during early and mid-life cycling. Cyclable lithium gets trapped within expanding Solid Electrolyte Interphase (SEI) layers on the graphite anode or consumed by parasitic side reactions. This consumption shifts the operating state-of-charge window of the negative electrode relative to the positive electrode ~ a movement known as electrode slippage that alters how graphite staging transitions align with the LFP OCV curve.

Loss of Lithium Inventory Dynamics

Under LLI, the negative electrode’s dQ/dV features shift toward higher cathode potentials during charge and lower potentials during discharge. On the full-cell dQ/dV curve, Peak III moves horizontally along the voltage axis while maintaining its basic shape and integrated area, provided total anode capacity still exceeds the available lithium inventory.

As lithium inventory drops, the full cell hits its lower discharge voltage cutoff while the graphite anode still has unextracted capacity headroom. The resulting loss in discharge capacity matches the quantity of consumed lithium ions. Analyzing LLI quantitatively involves tracking the voltage spacing between dV/dQ peaks representing graphite staging boundaries and cathode cutoff limits.

Active Material Degradation at Electrode Surfaces

Loss of negative active material (LAM_NE) stems from graphite particle fracturing, micro-delamination from copper current collectors, or pore clogging by SEI products. This reduces the total storage capacity of the anode matrix. As long as overall anode capacity stays larger than the remaining cyclable lithium inventory, LAM_NE does not immediately truncate full-cell discharge capacity.

When negative active material loss drops total anode capacity below the remaining lithium inventory, performance degrades rapidly. The graphite anode becomes fully lithiated before the full cell reaches its upper charge voltage limit, forcing excess lithium ions to deposit on the anode surface as metallic lithium plating.

Loss of positive active material (LAM_PE) occurs through iron dissolution, lattice disordering in the LiFePO4 olivine structure, or micro-cracking caused by cyclic expansion and contraction. LAM_PE narrows the capacity width of the cathode plateau, reducing the integrated area under the high-voltage region in dQ/dV spectra. Measuring this area reduction isolates cathode degradation from anode losses.

Baseline differential capacity spectra establish statistical variance. Quantitative breakdown of degradation modes utilizes tracking algorithms that match measured dQ/dV curves against synthesized full-cell profiles derived from half-cell open-circuit voltage data.

1. Electrode Slippage Identification aligns the high-voltage dV/dQ peak to measure the capacity offset between cathode and anode balancing points.
2. Peak Area Integration calculates remaining active capacity within specific graphite staging zones to detect isolated negative electrode material.
3. Symmetry Variance Analysis tracks peak skewness between charge and discharge sweeps to spot non-uniform current distribution across plates.
4. Ohmic Offset Extraction measures horizontal peak shifts to separate resistive polarization from true thermodynamic phase changes.

Polarization and Ohmic Resistance Offsets

Rising ohmic resistance imposes a kinetic penalty across all thermodynamic states. Internal resistance grows over time as electrolyte oxidizes, current collectors corrode, binders degrade, or contact resistance increases at terminal posts. This resistance creates an immediate ohmic overpotential (V_overpotential = I R_internal).

On dQ/dV spectra, increased resistance shifts charge peaks upward and discharge peaks downward in voltage. The gap between charge and discharge peak maxima for a given transition widens linearly with internal resistance. Comparing peak separation between charge and discharge sweeps at the same C-rate allows direct calculation of cell DC resistance without relying on electrochemical impedance spectroscopy.

Worked Case Study: Degradation Separation on a 314 Ah Cell

To illustrate mathematical mode separation, consider a 314 Ah prismatic LFP cell subjected to 2,500 full-depth 1C/1C cycles at 45 degrees Celsius ambient temperature. Initial baseline capacity was 314.0 Ah at C/50 discharge. By cycle 2,500, low-rate capacity had fallen to 251.2 Ah, leaving the cell at 80.0 percent state of health.

Analyzing the initial C/50 baseline dQ/dV curve established three primary markers: Peak 1 at 3.210 V (integrated area of 82.0 Ah), Peak 2 at 3.325 V (185.0 Ah), and Peak 3 at 3.390 V (47.0 Ah). Summing these areas accounts for the full 314.0 Ah initial capacity.

At cycle 2,500, the C/50 discharge curve shows clear structural changes. Total capacity drops to 251.2 Ah, representing a net loss of 62.8 Ah. Peak locations and areas have shifted: Peak 1 moved to 3.185 V (68.0 Ah), Peak 2 shifted to 3.298 V (152.0 Ah), and Peak 3 dropped to 3.360 V (31.2 Ah).

Separating these loss components starts with peak voltage displacement. During discharge, all three peaks shift lower by an average of 26.0 millivolts. At C/50 (6.28 amperes), a 26.0 millivolt overpotential corresponds to an internal resistance increase: delta_R = 0.026 V / 6.28 A = 0.00414 ohms (4.14 milliohms).

Compared against the fresh cell’s 0.18 milliohm baseline, this represents a 23-fold increase in low-frequency polarization resistance.

Next, electrode slippage calculations quantify LLI. Fitting the aged dV/dQ curve to half-cell reference models shows that the negative electrode slipped by 41.8 Ah relative to the cathode window. This 41.8 Ah offset directly represents consumed cyclable lithium, accounting for 66.6 percent of the total 62.8 Ah capacity loss.

Changes in peak area isolate active material losses. Negative active material loss (LAM_NE) causes Peak 2 and Peak 3 to shrink in relative capacity. Curve fitting shows a 5.0 percent reduction in total anode matrix capacity, equal to 15.7 Ah of lost active material.

Because remaining anode capacity still exceeds cyclable lithium inventory, this 15.7 Ah of LAM_NE does not truncate low-rate discharge capacity directly, though it significantly reduces the safety margin against fast-charge lithium plating.

Positive active material loss (LAM_PE) is derived from the upper voltage saturation region. Integrated cathode capacity shrank by 1.7 percent, representing a 5.3 Ah loss of active LiFePO4 material. The remaining 15.7 Ah capacity offset reflects non-linear interactions between LLI and localized impedance growth at particle boundaries.

In summary, the degradation breakdown for this 314 Ah cell at cycle 2,500 yields: Loss of Lithium Inventory of 41.8 Ah (13.3% of initial capacity), Negative Active Material Loss of 15.7 Ah (5.0% of anode mass), Positive Active Material Loss of 5.3 Ah (1.7% of cathode mass), and Resistance Growth adding 4.14 milliohms of polarization impedance.

When anode capacity shrinks below cathode capacity, cell failure accelerates rapidly due to metallic lithium plating during fast charge.

Uncorrected peak displacement from cold chamber drift resulted in forty thousand dollars in freight scrap costs when a client rejected a batch.

Noise

Thermal fluctuations during a twenty-hour diagnostic run distort differential capacity peaks by shifting electrode equilibrium potentials. Bench setup determines whether differential capacity analysis yields real degradation insights or electrical noise. Ambient temperature shifts, sensor calibration drift, cell clamping variations, and cycler settings can easily obscure electrochemical signals in large prismatic LFP cells.

Temperature drift is the largest environmental cause of dQ/dV signal corruption. LFP and lithiated graphite electrodes have distinct entropic heat coefficients (dE/dT), with LFP fluctuating between -0.08 mV/°C and +0.12 mV/°C depending on state of charge. A 2.0 degree Celsius change in ambient bench temperature alters open-circuit voltage by as much as 0.24 millivolts.

Temperature Coefficients and Enthalpic Heating Effects

Along LFP’s flat 3.42 V OCV plateau, a 0.24 millivolt shift spans several ampere-hours of capacity. If a thermal chamber cycles its compressor every fifteen minutes ~ causing temperature swings of plus or minus 0.5 degrees Celsius ~ the resulting voltage ripples create false derivative peaks across the dQ/dV curve.

Separating real phase transitions from thermal noise requires tracking temperature directly at the cell casing. Benches should use shielded multi-point thermocouple arrays attached to the aluminum cell housing; relying on ambient chamber wall sensors misses subtle self-heating during low-rate cycling.

When Does Polarization Overcome Differential Voltage Features?

Polarization degrades peak resolution by spreading features over wider voltage spans. Increasing the charge or discharge rate from C/50 to C/10 or C/5 introduces activation polarization and solid-state lithium diffusion resistance inside active particles. This overpotential pushes differential capacity features along the voltage axis, causing adjacent staging peaks to blur together.

Above C/10, graphite Peak I and Peak II merge into a single broad envelope while Peak III flattens, making peak height and width calculations unreliable. Testing at low C-rates keeps kinetic overpotentials below 5 millivolts and preserves distinct thermodynamic boundaries, though test schedules must balance low-rate requirements against lab throughput.

Mechanical Swelling and Prismatic Case Constraint

Large prismatic cells (280 Ah to 314 Ah) undergo significant mechanical stress during operation. Inserting lithium into the graphite matrix expands anode volume by up to 10 percent, generating internal breathing forces between 1,500 and 3,500 Newtons inside a constrained prismatic casing.

Uneven clamping pressure alters internal contact resistance and electrolyte distribution across the plates. Without steady mechanical constraint, electrode layers can separate slightly, causing localized current density variations that broaden dQ/dV peaks and mimic active material loss. Accurate characterization requires mounting prismatic cells in calibrated compression fixtures that maintain a constant 0.3 to 0.5 megapascal constraint.

• Thermal Chamber Drift creates low-frequency voltage ripples that generate false derivative peaks on flat OCV plateaus.
• Quantization Error from low-resolution cycler converters introduces digital step noise that degrades peak area integrations.
• Mechanical Constraint Relaxation allows electrode breathing variations, altering contact impedance and peak symmetry.
• Current Shunt Resistance Drift degrades current accuracy during long 50-hour runs, distorting integrated capacity figures.

IEC 62660-1 section 6.2 mandates cell temperature stabilization within half a degree Celsius during baseline reference capacity measurements.

Peak broadening observed at C/10 testing stems entirely from binder relaxation rather than loss of active lithium inventory.

Screening

Receiving inspection needs fast, non-destructive checks on incoming cell lots without spending 50 hours per unit on low-rate diagnostic cycles. Full-spectrum sweeps are impractical for volume lot acceptance, so high-throughput screening uses targeted partial-voltage sweeps paired with mathematical curve restoration to deliver actionable metrics in shorter test windows.

Partial-range screening limits cycling to critical voltage windows containing key phase transitions. In LFP prismatic cells, the 3.25 V to 3.40 V band houses the primary graphite Stage 2 to Stage 1 transition peak (Peak III). Restricting sweeps to this window cuts test duration by up to 70 percent while maintaining high sensitivity to lithium inventory loss.

Partial-Range Fast Screening Protocols

Fast-screening procedures charge cells to 3.45 V at 0.5C, allow a 30-minute thermal rest, and run a C/10 discharge down to 3.20 V. Deconvolution algorithms then reconstruct full dQ/dV peak profiles from this partial C/10 data using polarization matrices established during initial lot qualification.

While C/10 sweeps broaden peaks slightly compared to C/50 reference runs, overpotential corrections restore peak height accuracy within 2.5 percent of true thermodynamic values, allowing incoming inspection to spot damaged or degraded cells before module assembly.

Lot Conformity and Acceptance Thresholds

Incoming acceptance testing establishes statistical control limits for peak voltage position, height, and area based on baseline qualification data. Mean peak voltage shifts exceeding 5 millivolts relative to the baseline dossier point to unannounced changes in slurry formulation, formation cycling, or storage conditions.

Batch sampling during receiving audits follows ISO 2859-1 single sampling plans for normal inspection. Differential capacity metrics provide non-destructive pass/fail criteria; any sample batch showing peak area shrinkage beyond 3 percent triggers quarantine and root-cause teardown.

1. Transfer arriving prismatic cells from the quarantine pallet to a test bench held at 25°C for twelve hours.
2. Perform an initial open-circuit voltage and 1 kHz AC impedance check to screen out shorted cells.
3. Charge the cell to 3.65 V at 0.5C, holding constant voltage until current drops below 0.05C.
4. Discharge at 0.05C from 3.65 V down to 2.50 V, recording voltage and current every hundred milliseconds.
5. Calculate the differential capacity curve using a 25-point Savitzky-Golay filter and extract peak heights at 3.32 V and 3.20 V.
6. Compare peak voltage shifts against the baseline qualification dossier to verify lithium inventory retention.

Transportation Safety Dossier Integration

Shipping LFP cells internationally requires compliance with dangerous goods regulations. UN 38.3 protocols require vibration, thermal shock, impact, and short-circuit testing, while UN Manual Section 38.3.2 requires transported units to match certified design types.

Severe lithium inventory loss or active material isolation reduces safety margins against thermal runaway or internal shorts under mechanical stress. Adding DCA screening to compliance records provides clear documentation that aged or re-graded cells shipped for second-life use retain structural and chemical integrity consistent with original UN 38.3 test summaries.

Section 4.3 of the master supply agreement specifies that any delivered cell batch exhibiting an average Peak III voltage displacement exceeding 4.5 millivolts under C/20 baseline screening shall be rejected at the supplier cost.

Landed

Utility-scale energy storage contracts often bind suppliers to twenty-year capacity retention guarantees backed by substantial performance penalties. Landed cost models must account for long-term degradation trajectories to avoid unexpected system augmentation or warranty exposure. Differential capacity analysis converts basic state-of-health numbers into predictive financial models by pinpointing when cell populations approach non-linear degradation knee-points.

The knee-point represents the operational transition where capacity fade shifts from predictable linear LLI loss to rapid exponential degradation driven by anode active material exhaustion and lithium plating. When anode capacity drops below remaining lithium inventory, fade rates typically accelerate by three to five times. Knowing how far a cell population sits from knee-point onset lets asset managers model remaining useful life far more accurately.

Warranty Asset Evaluation and Remaining Life Projections

Standard cell warranties guarantee 70 or 80 percent capacity retention over ten to fifteen years, but standard capacity checks alone cannot show how close a cell is to knee-point acceleration. Two cells with identical 82 percent nominal state-of-health values can have vastly different remaining operational lives.

A cell sitting at 82 percent SOH purely from uniform LLI will continue fading linearly over hundreds of cycles. Conversely, a cell at 82 percent SOH driven partly by negative active material loss (LAM_NE) may be right at the brink of exponential degradation. Differential capacity peak height ratios distinguish these conditions non-destructively, helping integrators manage warranty risk.

Insurance Exposure and Second-Life Recertification

Repurposing retired EV packs into stationary storage requires clear degradation verification to secure insurance coverage and satisfy regulatory mandates. The EU Battery Regulation (EU 2023/1542) requires battery passports for industrial and EV batteries over 2 kWh, specifying documented state-of-health and degradation mechanism tracking.

Using differential capacity analysis during second-life grading provides verified health documentation. Quantifying remaining lithium inventory and active material ratios helps buyers price used modules accurately, establish enforceable warranties, and confirm compliance with transport safety regulations before shipping reassembled packs across borders.

Cell warranties that guarantee capacity retention without specifying test rate and temperature parameters leave buyers unprotected against internal resistance growth.

What non-linear polarization modeling techniques will reliably separate thermal run-away risk markers from benign active material loss in ultra-large format 500 Ah prismatic cells operating under fast partial-depth cycling?