Plot the V-i polarization curve of a proton-exchange membrane fuel cell (PEMFC) with the activation, ohmic and concentration overpotentials separated. Sweep cell temperature, operating pressure, exchange current density, Tafel slope, internal resistance and limiting current density and watch the operating-point voltage, power density, efficiency and hydrogen consumption update in real time — the same design trade-offs faced by the engineers behind Toyota Mirai and Hyundai NEXO stacks.
Parameters
Cell temperature T
°C
Drives Nernst potential and i₀. Real stacks run at 70–85°C
Operating pressure P
atm
Raises the Nernst term, but compressor power eats it back
Exchange current density i₀
A/cm²
Catalyst activity. Boosted by Pt loading or Pt-alloys
Tafel slope b
V/dec
Slope of the ORR. 60–70 mV/dec for Pt/C
Internal resistance R_ohm
Ω·cm²
Sum of Nafion thickness, hydration and contact losses
Limiting current density i_L
A/cm²
GDL diffusion limit for O₂
Active area A
cm²
MEA area per single cell (Toyota Mirai ≈ 273 cm²)
Results
—
Operating current i (A/cm²)
—
Cell voltage (V)
—
Power density (W/cm²)
—
Cell power (kW)
—
Cell efficiency (%)
—
H₂ consumption (g/h)
—
PEMFC single-cell cross-section — H₂/O₂ inlets and H⁺/e⁻ transport
Left: anode (H₂ → 2H⁺ + 2e⁻). Centre: Nafion membrane (only H⁺ crosses). Right: cathode (½O₂ + 2H⁺ + 2e⁻ → H₂O). Electrons travel through the external circuit and do work on the load.
V-i polarization curve (three losses separated)
Power density vs current density (maximum-power point)
PEMFC polarization equation. b = Tafel slope, i₀ = exchange current density, R_ohm = membrane + catalyst resistance, i_L = limiting current density. Term 2 is activation, term 3 is ohmic, term 4 is concentration loss.
Cell efficiency on HHV basis (1.481 V = H₂ thermoneutral voltage) and hydrogen consumption rate. M_{H₂} = 2.016 g/mol. Stoichiometry directly gives the H₂ flow from cell current.
Designing the PEMFC polarization curve
🙋
"So a fuel cell is just burning hydrogen, right? How is it different from an engine?"
🎓
Actually nothing burns. That's the trick: a PEMFC reacts H₂ and O₂ electrochemically without heat, so it skips the Carnot ceiling. At the anode, H₂ splits into 2H⁺ + 2e⁻. The protons slip through the Nafion membrane, the electrons take the long way around through your external circuit and drive your load, and at the cathode they reunite with O₂ to form water. The theoretical cell voltage is 1.229 V; in practice it's 0.6–0.7 V. Stack a few hundred cells in series to get ~300 V and you've got a Mirai-class drivetrain.
🙋
OK… but at the default settings the operating point sits around 0.78 V — that's already 0.45 V down from 1.229. Where is all that going?
🎓
Good catch. The drop splits three ways. First, activation loss η_act = b·log(i/i₀) = 0.06·log(1/1e-4) ≈ 0.24 V — that's the biggest chunk. The cathode oxygen-reduction reaction (ORR) is sluggish even on Pt. Second, ohmic loss η_ohm = i·R = 1·0.15 = 0.15 V — that's the membrane and catalyst layer resistance, fixed by Nafion thickness and hydration. Third, concentration loss η_conc ≈ 0.018 V — small at i = i_L/2 but it explodes as you approach i_L.
🙋
So I just pile on more Pt to push i₀ up? I keep reading Toyota fought to use less platinum…
🎓
That's the entire R&D war. Pt is ~USD 30/g and a single FCEV used to need ~30 g — so the catalyst alone was a four-figure bill. Going from i₀ = 1e-5 to 1e-4 saves you b·log(10) = 60 mV of activation loss, but doubling Pt mass only buys you ≈√2 more surface area. So the modern playbook is Pt-Co or Pt-Ni alloys, nano-structured Pt particles and core-shell catalysts that lift activity without lifting Pt loading. Mirai 2 halved the Pt versus the first generation at the same output.
🙋
Looking at the power-density chart, peak power is well before i_L. Why don't we drive the cell all the way to i_L?
🎓
Because P = V·i is a fight. As i climbs, V crashes, so just before i_L the voltage is near zero and so is the power — you get a concave hill. The peak typically lands at i ≈ 0.5–0.7·i_L. There's also a bigger picture: high efficiency lives at low current (high V). FCEVs hop between operating points — near peak power (V ≈ 0.6) under acceleration, near peak efficiency (V ≈ 0.7) when cruising. Sitting at i = 0.5·i_L is a reasonable "balance of efficiency and power".
🙋
Last one: is it really true that hotter is better? What happens above 95°C?
🎓
Hotter is better up to a point. Higher T boosts i₀ Arrhenius-style and increases proton conductivity, lowering R_ohm. From 60 → 80°C you typically gain 20% power. Above ~90°C, Nafion dries out, conductivity collapses and R_ohm balloons. So 80°C is the sweet spot, and humidifier + radiator design becomes the heart of an FCEV thermal system. SOFCs (solid oxide) run at 700°C and dodge water problems entirely, but they live in stationary cogeneration, not cars.
Frequently asked questions
A PEM fuel cell loses voltage through (1) activation polarization — caused by the activation energy of the electrode reactions and dominant at low current density, (2) ohmic polarization — IR drop across the membrane, catalyst layers and contact resistances, dominant at intermediate current, and (3) concentration polarization — driven by mass-transport limits and rising sharply near the limiting current density i_L. This tool plots V = E_rev − η_act − η_ohm − η_conc with each term separated, so you can see which design lever (catalyst, membrane, GDL) reduces which loss.
The thermoneutral (HHV) potential is 1.481 V and the Nernst potential at standard conditions is 1.229 V, but activation, ohmic and concentration losses pull the practical cell voltage down to 0.6–0.7 V. Running at higher voltage gives better efficiency but lower current; running at lower voltage gives higher power but lower efficiency. Vehicles like the Toyota Mirai operate near 0.6 V for peak power during acceleration and near 0.7 V for high efficiency during cruise, yielding an average stack efficiency of 50–55%.
i_L is set by how fast O₂/H₂ can diffuse through the gas-diffusion layer (GDL) and catalyst layer to the reaction sites. As i approaches i_L the concentration polarization η_conc = −b·log(1 − i/i_L) diverges logarithmically and voltage collapses; at i = i_L it falls to zero. In practice product water can flood the GDL and lower the effective i_L. This tool sets the operating point to i_op = 0.5·i_L so you can check that you have a comfortable margin from the diffusion limit.
Higher temperature lowers the Nernst potential by about −0.85 mV/K but boosts the exchange current density i₀ and proton conductivity exponentially; net effect is a large gain going from 60 to 80°C. Above 90°C Nafion dries out and the humidifier becomes hard to design. Higher pressure lifts the Nernst term by (RT/4F)·ln(P_O₂·P_H₂²), giving ~30 mV/cell from 1 to 3 atm, but the compressor power offsets that, so automotive PEMFCs usually settle near 1.5–2.5 atm.
Real-world applications
Fuel-cell vehicles (FCEVs): The 2nd-generation Toyota Mirai stacks 330 cells for about 128 kW; the Hyundai NEXO uses 440 cells at 95 kW. A 3-minute hydrogen fill gives ~700 km of range, sidestepping the long charging times and weight of large BEV batteries. The defaults here (A = 250, i_L = 2.0, V ≈ 0.65 V, power density 0.8–1.0 W/cm²) sit squarely inside the typical mass-production envelope for an FCEV stack.
Commercial trucks, buses and logistics: For long-haul trucks, BEV batteries cut too far into payload, so Nikola, Hyundai XCIENT and Toyota–Hino are betting on PEMFC. Total powers of 150–300 kW with multiple stacks in parallel deliver diesel-like duty cycles and minutes-long refuelling. Plug Power has shipped more than 70,000 PEM fuel-cell forklifts to date.
Stationary cogeneration (Ene-Farm): The Japanese residential 700 W PEMFCs (Panasonic, Toshiba) reform city gas to produce hydrogen and deliver combined heat-and-power at ~95% total efficiency. Over 300,000 units are operating in Japan. Lowering A here to ~100–150 cm² and operating at low current density to chase efficiency is the textbook stationary-design choice.
Drones, UAVs and space: PEM stacks have 3–5× the energy density of Li-ion, so long-endurance drones (IFC, H2Fly) use them for hours-long flights. Apollo and Space Shuttle ran on PEMFCs and re-used the product water for crew drinking — small A and an efficiency-biased operating point is the space-grade design pattern.
Common misconceptions and pitfalls
The first trap is thinking "fuel-cell efficiency = cell voltage ÷ 1.229 V". 1.229 V is the Nernst potential (ΔG / 2F, LHV basis); for fuel efficiency you divide by 1.481 V on an HHV basis. This tool uses ηcell = V / 1.481, i.e. the fraction of the hydrogen's higher heating value turned into electricity. On a real system you also lose 10–15% to balance-of-plant (BOP) — compressor, pumps, cooling fans — so this tool approximates systemEfficiency = cellEfficiency × 0.85. Always check whether a "60% efficient" spec is LHV or HHV, stack-only or system-level.
The second pitfall is believing more platinum guarantees better performance. Because i₀ scales with Pt surface area and surface area only grows with √(Pt mass), doubling the loading buys you about b·log(√2) ≈ 9 mV of activation gain. Meanwhile Pt is ~USD 30/g, and halving Mirai-1's ≈ 30 g Pt loading was a non-negotiable cost target for commercial viability. Modern PEMFC design attacks the problem from three angles: lower Pt loading, Pt-Co / Pt-Ni alloys for higher specific activity, and engineered carbon supports that nano-disperse the catalyst. Sweep i₀ from 1e-5 to 1e-4 in this tool: η_act only drops from ~0.30 to ~0.24 V, a stubborn 60 mV.
The third pitfall is ignoring water management. PEMFCs make water at the cathode; under high current density that water can flood the cathode GDL and quietly drag the effective i_L downward. At low load and high temperature the opposite happens — Nafion dries out, R_ohm spikes. The i_L in this tool is a design value; real stacks watch per-cell voltage in real time and modulate humidification, gas flow and temperature in concert. "A nice polarization curve" is not the same as "a working stack" — until you've nailed transient response, sub-zero start (down to −30°C) and freeze-thaw durability, you don't have an FCEV-grade design.
How to Use
Set cell temperature (50–80°C typical for PEMFC operation) and absolute pressure (1–3 bar) using the input sliders.
Adjust exchange current density (10⁻¹² to 10⁻⁶ A/cm²) and Tafel slope (40–120 mV/dec) to reflect your membrane and catalyst properties.
Read the polarization curve plot showing voltage drop across activation, ohmic, and concentration loss regions; identify the optimal current density where power density peaks.
Worked Example
A 5 cm² PEMFC operating at 65°C, 2 bar absolute pressure, with Nafion 117 membrane (exchange current density 1×10⁻⁸ A/cm², Tafel slope 60 mV/dec) yields: at i = 0.8 A/cm² the cell voltage is 0.72 V, power density reaches 576 W/cm² (2880 W total), system efficiency is 48%, and H₂ consumption is 1.2 g/h. The curve shows 0.95 V open-circuit, 0.05 V activation loss, 0.18 V ohmic loss, and 0.10 V concentration loss at this current.
Practical Notes
Higher operating temperature (75–80°C) reduces activation overpotential and ohmic resistance in PEMFC stacks but increases membrane water loss; balance with humidification systems.
Pressure increase (2–3 bar) shifts the open-circuit voltage upward and improves oxygen reduction kinetics; useful for vehicle fuel cells targeting 0.85+ V at rated power.
Exchange current density strongly depends on platinum loading and ionomer distribution; a 0.5 mg/cm² Pt catalyst gives i₀ ≈ 2×10⁻⁸ A/cm², while 1.0 mg/cm² reaches 5×10⁻⁸ A/cm².
Concentration losses dominate above 1.2 A/cm² in most PEMFC; this simulator helps identify maximum useful power output before mass-transport limits performance below 0.6 V.