Second-Life Lithium-Ion Batteries

State-of-Health Assessment, Repurposing Pathways, and the Electrochemical Basis for Remaining Useful Life

Retired electric vehicle battery packs do not reach electrochemical end-of-life when they exit automotive service — they reach the end of their automotive application window, typically defined at 70–80% residual capacity (state-of-health, SoH). The electrochemical capacity remaining in that cell is real, measurable, and exploitable. The engineering challenge of second-life deployment is not whether the energy is there, but whether it can be characterized rapidly enough, at low enough cost, and with sufficient predictive accuracy for remaining useful life (RUL) to make repurposing economically defensible against direct recycling. That is a characterization, logistics, and BMS problem as much as a chemistry problem.

---

Why EV Batteries Exit Service Before Electrochemical Exhaustion

Two distinct mechanisms drive battery pack retirement from automotive service before individual cells reach true end-of-life:

Pack-Level Configuration Failure

EV battery packs are series-parallel assemblies of hundreds to thousands of cells. In a series string, the cell with the lowest capacity sets the usable capacity of the entire string — the pack is balanced to the weakest cell. When one module degrades below the pack's functional threshold, the entire pack's energy and power delivery is compromised disproportionately to the degradation of the individual failed unit. The economics of replacing individual modules vs. retiring the pack depend on pack architecture, but in many first-generation EV designs, module-level replacement was not cost-effective, accelerating pack-level retirement even when the majority of cells retained substantial capacity.

Application-Specific SoH Threshold

EV applications impose a hard SoH floor — typically 70–80% of nameplate capacity — below which driving range loss becomes commercially unacceptable to the end user, even though the cells continue to function electrochemically. This threshold is not defined by cell failure but by application requirement. A pack retired at 75% SoH from an EV retains 75% of its original energy content — fully usable in applications with less stringent energy density, rate, or cycle life demands.

---

Second-Life vs. Direct Recycling — The Economic and Environmental Decision

The decision to pursue second-life repurposing vs. direct recycling is not categorically resolved — it is application- and context-dependent, governed by the following competing factors:

Factor	Favors Second-Life	Favors Direct Recycling
Residual SoH	>70% with predictable degradation	<70% or highly variable cell-to-cell spread
Cell chemistry	LFP (stable, long cycle life)	NMC/NCA (higher material recovery value)
Pack disassembly cost	Low (modular design, accessible cells)	High (welded, glued, complex thermal systems)
Remaining useful life predictability	High (known history, consistent chemistry)	Low (unknown duty cycle, mixed chemistries)
Critical material value	Low (LFP: no Co, low Ni)	High (NMC 811, NCA: significant Co, Ni)
Second-life application demand	High (local grid storage, telecom backup)	Low

Direct recycling routes include:

• Hydrometallurgical processing — acid dissolution of electrode material, selective precipitation of Li, Co, Ni, Mn salts; high recovery rate but energy-intensive
• Pyrometallurgical processing — high-temperature smelting; recovers Co and Ni but loses Li and graphite
• Direct (solid-state) recycling — relithiation and reprocessing of cathode material without full dissolution; lowest energy input, highest retained material value, but requires consistent feed chemistry

---

State-of-Health Assessment — The Central Technical Problem

Deploying retired cells in a second-life application requires a reliable estimate of residual capacity, internal resistance, and — critically — remaining useful life (RUL) under the target duty cycle. RUL prediction is the unresolved technical bottleneck: a cell's future degradation trajectory depends on its cumulative electrochemical history (thermal exposure, depth-of-discharge, C-rate profile, number of cycles), most of which is unavailable or incomplete for cells arriving from field-retired packs.

Degradation State at Retirement

By the time a cell reaches 75–80% SoH, the following degradation mechanisms are likely active to varying degrees:

Degradation Mechanism	Diagnostic Signature	Impact on Second-Life
SEI thickening	Increased low-frequency impedance (EIS); capacity fade	Accelerated further by continued cycling
Lithium inventory loss (LLI)	Shifted OCV-SOC curve; reduced capacity	Permanent; sets hard floor on recoverable capacity
Active material loss (LAM) — anode	Peak shift in dV/dQ; capacity fade	Reduced rate capability
Active material loss (LAM) — cathode	dV/dQ peak broadening/disappearance	Reduced energy density
Increased DC resistance	Elevated pulse resistance; power fade	Limits discharge rate in second-life application
Lithium plating (if present)	Asymmetric dV/dQ; coulombic efficiency drop	Safety risk — requires identification and segregation

---

Six-Step Characterization Protocol for Second-Life Cell Assessment

The protocol below establishes the electrochemical baseline of a retired cell, quantifies its residual performance, and generates the data required for application-matched sorting.

---

Step 1 — Preconditioning

Retired cells arriving from field service may have been stored at unknown SOC for indeterminate periods. Extended storage causes SEI evolution, electrolyte redistribution, and in some cases lithium redistribution between electrodes. Preconditioning re-establishes a known electrochemical reference state.

Protocol: Apply 2–3 formation-rate (C/10 or C/20) charge-discharge cycles between the cell's defined voltage limits at 25°C. Measure Coulombic efficiency (CE) per cycle. Cells exhibiting CE < 98.5% after 3 cycles, or anomalous voltage behaviour (stepped discharge, early voltage cutoff), are flagged for further investigation or rejection.

---

Step 2 — Relaxation

After each charge or discharge in the preconditioning sequence, the cell must rest at open circuit until the terminal voltage stabilizes to within ±1 mV over a defined interval (typically 30–60 minutes minimum, longer for high-impedance aged cells). This allows:

• Li⁺ concentration gradients within electrode particles to equilibrate (solid-state diffusion relaxation)
• Salt concentration gradients in the electrolyte to dissipate
• Double-layer charge to decay

Measurements taken without adequate relaxation time reflect polarization overpotentials rather than equilibrium thermodynamic state, producing systematic errors in OCV-SOC mapping and impedance baselines.

---

Step 3 — Capacity Test

Measure true residual capacity by performing a full discharge at a reference C-rate (typically C/10 or C/20 to minimize kinetic limitations) between defined voltage cutoffs at controlled temperature (25 ± 1°C). Repeat for statistical confidence (minimum 2–3 consecutive cycles with <0.5% capacity variation between cycles).

Critical variables to control and report:

• Temperature (capacity is temperature-dependent; a 10°C variation can shift measured capacity by 3–8%)
• C-rate (higher rates underreport true capacity due to polarization-limited voltage cutoff)
• Voltage cutoff window (must match original cell specification — arbitrary cutoff expansion to recover apparent capacity masks degradation)

The measured capacity as a fraction of nameplate capacity is the primary SoH metric:

$$\text{SoH} = \frac{Q_{\text{measured}}}{Q_{\text{nameplate}}} \times 100\%$$

---

Step 4 — OCV vs. SOC Mapping

Measure the open-circuit voltage as a function of SOC across the full operating range using a low-rate quasi-static protocol (C/20 or slower with intermittent rest steps). The resulting OCV-SOC curve is a thermodynamic fingerprint of the cell's current electrode state.

Diagnostic value:

• Voltage plateau shift or plateau shortening relative to a reference (beginning-of-life) curve indicates lithium inventory loss (LLI) or active material loss (LAM)
• dV/dQ analysis of the OCV-SOC curve resolves individual electrode contributions and identifies which electrode is the primary degradation site
• Curve shape change (loss of distinct plateaus, slope increase) indicates structural degradation of intercalation host materials

This data also calibrates the BMS SOC estimation algorithm for the second-life application — an OCV-SOC curve developed at beginning-of-life will introduce SOC estimation errors in a degraded cell if not updated.

---

Step 5 — DC Pulse Resistance Test (HPPC Protocol)

The Hybrid Pulse Power Characterization (HPPC) test quantifies DC internal resistance as a function of SOC, C-rate, and temperature. The protocol applies charge and discharge current pulses (typically 10–30 seconds) at multiple SOC setpoints across the operating range, with defined rest periods between pulses.

From the voltage response to each pulse, the DC resistance at that SOC/temperature/rate combination is extracted:

$$R_{\text{DC}} = \frac{\Delta V}{\Delta I$$

where ΔV is the instantaneous voltage step at pulse onset (ohmic resistance contribution) and the time-evolving voltage response provides the diffusion/kinetic resistance components.

Second-life relevance:

• DC resistance determines peak power capability — critical for sizing the second-life application (a high-impedance cell cannot deliver the same peak current as a fresh cell even if its capacity is comparable)
• Temperature dependence of DC resistance sets the low-temperature power floor — relevant for outdoor stationary applications
• SOC dependence identifies the operating window within which the cell can deliver required power — high resistance at low SOC may narrow the practical DOD window

---

Step 6 — Electrochemical Impedance Spectroscopy (EIS)

EIS applies a small-amplitude AC voltage perturbation (typically 5–10 mV) across a frequency range (typically 10 mHz to 100 kHz) and measures the complex impedance response. The resulting Nyquist or Bode plot is deconvolved into physical contributions using an equivalent circuit model (ECM).

Principal impedance features and their physical origins:

Impedance Feature	Frequency Range	Physical Origin
High-frequency intercept (R₀)	>10 kHz	Ohmic resistance: electrolyte, contacts, current collector
High-frequency semicircle	1–10 kHz	SEI layer ionic resistance and capacitance
Mid-frequency semicircle	1 Hz–1 kHz	Charge-transfer resistance at electrode-electrolyte interface
Low-frequency tail (Warburg)	<1 Hz	Solid-state Li⁺ diffusion within electrode particles

Second-life diagnostic application:

• Increased SEI semicircle diameter → SEI thickening, lithium inventory consumption
• Increased charge-transfer resistance → surface passivation, cathode structural degradation
• Warburg slope change → altered solid-state diffusivity (particle cracking, binder degradation)
• EIS measured at multiple SOC setpoints and temperatures generates an impedance map used for both RUL prediction and second-life BMS parameterization

EIS is the most information-dense single measurement in this protocol, but it requires controlled temperature and full voltage equilibration (Step 2) to produce interpretable spectra. EIS collected on an unrelaxed cell is not reliable.

---

Estimated Remaining Useful Life — Field Data Context

[REVIEW FLAG — Original text states EV batteries last "8 years (80% of State of Charge)." The 8-year figure is an automotive warranty benchmark, not an electrochemical end-of-life date. Many first-generation EV packs are demonstrating 12–15 year calendar life in moderate climates. The 80% figure is the SoH retirement threshold for the automotive application, not a SOC figure. Terminology corrected below.]

Battery Origin	Typical Automotive Retirement SoH	Estimated Second-Life Duration	Primary Degradation Mechanism in Second Life
EV pack (LFP)	75–80%	8–15 additional years at low C-rate	SEI growth (calendar), minimal cycle degradation
EV pack (NMC)	70–80%	5–10 additional years (application-dependent)	Continued cathode degradation, electrolyte oxidation at high SOC
Laptop / consumer cells	60–75% (user-perceived failure)	2–5 additional years at shallow DOD	Li inventory loss; variable cell-to-cell spread

Second-life duration is strongly application-dependent. A retired EV cell deployed in a low-C-rate, shallow depth-of-discharge grid storage application (e.g., 0.1C, 20–80% SOC window) will degrade far more slowly than the same cell operated at 1C, 0–100% DOD. The second-life application must be matched to the cell's residual capability, not the other way around.

---

Application Matching — Sorting Criteria

Post-characterization, cells are sorted into application tiers based on the metrics established in Steps 3–6:

Application Tier	Minimum SoH	Maximum DC Resistance	Cycle Frequency	Example
Grid frequency regulation	>75%	<130% of BoL	High (daily cycling)	Utility ancillary services
Behind-the-meter storage	>70%	<150% of BoL	Moderate (1 cycle/day)	Commercial/residential ESS
Telecom backup / UPS	>65%	<200% of BoL	Low (float, rare discharge)	Tower backup power
Low-power stationary	>60%	<250% of BoL	Very low	Remote sensor power
Reject / direct recycle	<60%	Any	—	Hydrometallurgical recovery

---

•