Open Circuit Voltage Of A Solar Cell — Formula, Derivation, And Cold-Morning String Sizing (2026)
The open-circuit voltage of a solar cell is given by Voc = (n·kT/q) · ln(I_L / I₀ + 1), derived by setting the current in the Shockley diode equation to zero. It is the highest voltage a solar cell can produce — the voltage you measure across the terminals when no current flows. Under STC (25 °C), a single c-Si cell measures 0.65–0.75 V; a 60-cell module measures 38–46 V; a 72-cell module measures 46–55 V. Voc decreases at about −2.2 mV/°C per cell with rising temperature, and increases by the same amount in cold mornings — which is why cold-climate string sizing is the most common reason inverters get fried in residential PV. This guide derives the equation from first principles, explains the physics of every term, and walks through cold-morning string sizing the way the NEC requires it.
I built a 6 kW PV array on my own house in Slovenia in 2024. Slovenian winter mornings hit −10 to −15 °C cell temperature on a clear sunrise, and the first thing the installer asked for was the temperature-corrected Voc for my chosen string length. Get this number wrong and the inverter MPPT input dies on the first cold sunny day. This article is the math I had to do.
What Voc Actually Is
The open-circuit voltage (Voc) is the voltage you measure across the terminals of a solar cell when no current is flowing — that is, when the external circuit is open. It is the maximum voltage the cell can produce. As soon as you connect a load, the cell voltage drops below Voc; the more current you pull, the further it drops, until at short circuit the voltage is zero and the current is the maximum (Isc).
The Voc point sits at the right end of the I-V curve, where the curve crosses the voltage axis. Pmax — the maximum power point — sits inside the curve at a voltage Vmp that is always 80–90 % of Voc, where the product of voltage and current peaks.
The Voc Equation — Derivation From The Shockley Diode Equation
A solar cell is a large-area p-n junction — a diode — illuminated by sunlight. The current-voltage relationship for an illuminated diode, derived by William Shockley in 1949, is:
I = I₀ · (exp(qV / nkT) − 1) − I_L
Where:
- I = current flowing out of the cell (A)
- I₀ = dark saturation current (A) — current that would flow through the diode in reverse bias
- q = elementary charge = 1.602 × 10⁻¹⁹ C
- V = voltage across the cell terminals (V)
- n = ideality factor (dimensionless, 1.0–2.0 for real cells)
- k = Boltzmann's constant = 1.381 × 10⁻²³ J/K
- T = absolute temperature (K)
- I_L = light-generated current (A)
At open circuit, no current flows: I = 0. Substitute and solve for V:
0 = I₀·(exp(qV / nkT) − 1) − I_L
I_L = I₀·(exp(qV / nkT) − 1)
I_L / I₀ + 1 = exp(qV / nkT)
qV / nkT = ln(I_L / I₀ + 1)
Solving for V at this open-circuit condition gives Voc:
Voc = (n · k · T / q) · ln(I_L / I₀ + 1)
This is the open-circuit voltage formula every solar physics textbook prints. It is not arbitrary — it falls directly out of forcing the diode equation to zero current.
For real cells where I_L is many orders of magnitude larger than I₀, the "+1" inside the logarithm is negligible, and the equation simplifies to:
Voc ≈ (n · k · T / q) · ln(I_L / I₀)
This is the form used in cell-design software.
What Each Term Means Physically
| Term | Symbol | Physical meaning | Typical value (2026 c-Si) |
|---|---|---|---|
| Thermal voltage | k·T/q | The thermal energy per unit charge — sets the "natural" voltage scale of any diode | 25.85 mV at 25 °C |
| Ideality factor | n | How closely the diode follows ideal Shockley behavior; n = 1 ideal, n = 2 recombination-limited | 1.0–1.3 |
| Light-generated current | I_L | The photocurrent produced by absorbed photons; proportional to irradiance and panel area | 9–14 A per cell |
| Dark saturation current | I₀ | The minority-carrier reverse current; set by bandgap, doping, and recombination | 10⁻¹⁰ to 10⁻¹² A |
The sensitivity is dramatic. Voc depends logarithmically on I_L/I₀, so doubling I_L (more sunlight) only adds 0.026 × ln(2) ≈ 18 mV of Voc. But cutting I₀ by a factor of 10 (better passivation) adds 0.026 × ln(10) ≈ 60 mV of Voc. This is why surface passivation is the dominant lever in modern cell design — it goes after I₀.
Thermal Voltage kT/q
At T = 298.15 K (25 °C), the thermal voltage is kT/q = 25.85 mV. This is the unit on which the entire diode equation is built. It is the same number that appears in transistor physics, in the Boltzmann distribution of carriers across a band, and anywhere a thermal-equilibrium charge distribution matters. It is not a property of silicon — it is a property of temperature.
This is also why Voc has temperature dependence (next section): kT/q increases with T, but I₀ increases much faster with T, and the I₀ term dominates.
Ideality Factor n
The ideality factor n is a real-world fudge that captures how a diode deviates from the ideal Shockley equation. The two end-points are:
- n = 1 — The ideal Shockley diode. Current is dominated by diffusion of minority carriers across the p-n junction. This is what a perfect, defect-free silicon cell would have.
- n = 2 — A diode whose current is dominated by recombination in the depletion region. This happens in damaged cells, low-quality material, or at low forward bias.
Real silicon cells live between these two regimes, with effective n typically 1.0 at high illumination (where diffusion dominates) and rising toward 1.3–1.5 at low illumination. Modern cell models — the two-diode model — fit two diodes in parallel, one with n = 1 and one with n = 2, to capture both regimes.
A lot of confusion about "ideality factor between 0 and 1" in older textbooks comes from mixing it up with quantum efficiency or other dimensionless cell parameters. The ideality factor is between 1 and 2 by definition. If you compute a value below 1, your fit is wrong.
Light-Generated Current I_L
I_L is the photocurrent — the current that would flow if the cell were short-circuited. Numerically it equals Isc on the I-V curve (within a few percent — they differ slightly because of internal series resistance). It is proportional to:
- Plane-of-array irradiance (linear)
- Cell area (linear)
- Cell quantum efficiency (which determines how many photons become carriers)
- Optical properties of the front coating, busbars, and texturing
For a modern 410 W LONGi Hi-MO 6 cell, Isc = 13.85 A at 1,000 W/m², so I_L ≈ 13.9 A.
Dark Saturation Current I₀
I₀ is the most important parameter in the equation. It is the current that would flow through the diode in reverse bias if there were no other limiting effects. Physically, it is set by:
- Bandgap energy Eg — I₀ scales with exp(−Eg/kT), so wider bandgaps → smaller I₀ → higher Voc
- Doping concentrations — heavier doping reduces I₀ (up to a point — Auger recombination eventually wins)
- Minority carrier lifetime — longer lifetime → smaller I₀ → higher Voc
- Surface recombination velocity — better surface passivation → smaller I₀
This is why HJT cells (with their amorphous-silicon passivation layers) and TOPCon cells (with their tunnel-oxide passivated contacts) achieve higher Voc than older Al-BSF and PERC architectures: their I₀ is 5–10× lower because their surface recombination is dramatically suppressed.
Worked Example — Computing Voc From First Principles
Let's compute the Voc of a single high-quality c-Si cell with realistic 2024 numbers:
- Ideality factor n = 1.1
- Light-generated current I_L = 10.5 A
- Dark saturation current I₀ = 1.0 × 10⁻¹⁰ A
- Cell temperature T = 298.15 K (25 °C)
Voc = (n · kT / q) · ln(I_L / I₀ + 1)
= (1.1 × 0.02585) · ln(10.5 / 1e-10 + 1)
= 0.02844 · ln(1.05 × 10¹¹)
= 0.02844 · 25.38
= 0.722 V
So this cell has a Voc of 722 mV. That is right in the range of a Tier 1 n-type cell in 2026.
Notice how the logarithm dominates: ln(1.05 × 10¹¹) = 25.38 — a huge number. This is why doubling I_L only adds ~18 mV of Voc, while improving I₀ by a factor of 10 adds ~60 mV. The equation rewards passivation, not photocurrent.
Now, a 60-cell module wires 60 of these cells in series, so its module-level Voc would be 60 × 0.722 = 43.3 V (before tolerances and small mismatch losses). A 72-cell module wires 72 in series → 52.0 V. These are the per-string Voc numbers you actually see on datasheets.
Voc Temperature Dependence — Where −2.2 mV/°C Comes From
The single most important practical consequence of the diode equation is that Voc decreases as temperature increases — and increases as temperature decreases.
The derivation is in any semiconductor physics textbook (Sze & Ng is the canonical reference). The result for silicon, computed from differentiating the Voc equation with respect to temperature, is:
dVoc/dT ≈ −(Vg₀ − Voc + γkT/q) / T
Where Vg₀ is the bandgap at 0 K (~1.2 V for silicon) and γ is a constant near 3 from the temperature dependence of I₀. Plug in numbers and you get:
dVoc/dT ≈ −2.0 to −2.3 mV/°C per silicon cell
In percentage terms, βVoc ≈ −0.27 % to −0.32 %/°C of the room-temperature Voc.
Manufacturers measure βVoc empirically and print it on the datasheet. The typical 2026 numbers:
| Module | Voc (STC) | βVoc | dVoc/dT |
|---|---|---|---|
| LONGi Hi-MO 6 LR5-54HTH 410W (HPBC) | 37.50 V | −0.25 %/°C | −94 mV/°C (per module) |
| REC Alpha Pure-R 430W (HJT) | 39.40 V | −0.24 %/°C | −95 mV/°C |
| Trina Vertex S+ 440W (TOPCon) | 38.80 V | −0.25 %/°C | −97 mV/°C |
| JinkoSolar Tiger Neo 72HL4-V 580W (TOPCon) | 50.30 V | −0.25 %/°C | −126 mV/°C |
| Maxeon 7 440W (IBC) | 41.50 V | −0.22 %/°C | −91 mV/°C |
| Legacy PERC 400W | 49.50 V | −0.28 %/°C | −139 mV/°C |
Two patterns:
- n-type cells (HJT, TOPCon, IBC) have lower βVoc than older p-type PERC. HJT in particular sits at −0.22 to −0.24 %/°C.
- The absolute mV/°C is similar across cells because higher Voc cells lose proportionally less per degree.
Cold-Morning String Sizing — Why This Matters For Inverters
Here is the practical reason the Voc equation matters: at sub-freezing temperatures, Voc rises, and the string voltage you feed your inverter rises with it. If the inverter's maximum DC input voltage is exceeded, the MPPT input dies. This is the most common cold-climate failure in residential PV.
The NEC 2023 Article 690.7 requires installers to compute the maximum string Voc at the lowest expected ambient temperature for the install site, using either:
- The manufacturer's βVoc temperature coefficient (preferred — accurate)
- The NEC's prescribed temperature correction factor table (conservative — used when βVoc is unavailable)
The cold-corrected Voc is:
Voc_cold = Voc_STC × (1 + βVoc × (T_cold − 25))
where T_cold is the lowest expected cell temperature (in °C) and βVoc is the open-circuit voltage temperature coefficient (in fractional form, i.e. −0.0025 not −0.25 %/°C).
Worked Example — A 12-Panel String In Slovenia
My 6 kW array uses 14 LONGi Hi-MO 6 LR5-54HTH 410W panels in two strings of 7. Local lowest expected ambient is around −15 °C. On a clear sunny morning at −15 °C, cell temperature equals ambient (no thermal load yet, no irradiance heating). NEC requires sizing against that temperature.
Per-panel Voc_cold = 37.50 × (1 + (−0.0025) × (−15 − 25))
= 37.50 × (1 + (−0.0025) × (−40))
= 37.50 × (1 + 0.10)
= 37.50 × 1.10
= 41.25 V
A single panel that nameplates at 37.5 V will hit 41.25 V on a cold morning — about 10 % higher than STC.
For a string of 7 panels in series:
String Voc_cold = 7 × 41.25 = 288.75 V
My inverter's maximum DC input is 600 V, so 289 V is safe with comfortable headroom. If I had wanted a string of 14 panels in series (one big string), I would have hit 14 × 41.25 = 577.5 V — still under 600 V but with very little margin. NEC inspectors typically want at least 50 V of headroom on top of the cold-corrected Voc.
What If I Used An Older PERC Panel?
A legacy 400 W PERC panel (Voc 49.5 V, βVoc = −0.32 %/°C) at −15 °C:
Per-panel Voc_cold = 49.5 × (1 + (−0.0032) × (−40)) = 49.5 × 1.128 = 55.85 V
Same string of 7 = 390.95 V. A string of 14 = 781.9 V — blows past a 600 V inverter input. This is exactly why older PERC strings in cold climates often had to be limited to 10 or 11 panels in series, while modern n-type strings can run 12–14 panels in series safely.
The Voc equation isn't a textbook curiosity. It is the equation that determines how many panels you can string together in your climate.
Common Misreadings
- "Voc is the operating voltage." No — Voc is the no-load voltage. The operating voltage at maximum power is Vmp, which is ~80–90 % of Voc. A cell can never deliver power at its Voc point; the current is zero by definition.
- "Ideality factor is between 0 and 1." No — it is between 1 and 2 by definition (1 = ideal Shockley diffusion, 2 = recombination-limited). A computed value below 1 means your fit is broken.
- "More sunlight → much more Voc." No — Voc depends logarithmically on photocurrent. Doubling irradiance only adds ~18 mV per cell. Voc is set primarily by I₀, which is set by passivation.
- "Voc rises in summer because it's hotter." Backwards — Voc falls in summer. It rises in winter. The temperature coefficient is negative.
- "My cold-morning string is fine because the panels rated below the inverter max." Always size against the temperature-corrected Voc, not the STC nameplate. NEC 690.7 requires it, and it is the single most common reason inverter MPPT inputs die.
- "All panels have the same Voc temperature coefficient." βVoc varies meaningfully between technologies — HJT/IBC cells are about −0.22 to −0.24 %/°C, modern TOPCon is about −0.25 to −0.26 %/°C, older PERC is −0.30 to −0.32 %/°C. In cold climates the difference is the difference between fitting 14 panels in a string and being limited to 11.
Bottom Line
The open-circuit voltage of a solar cell — Voc = (n·kT/q)·ln(I_L/I₀ + 1) — is one of the cleanest derivations in semiconductor physics. It comes directly from the Shockley diode equation by forcing current to zero, and it cleanly exposes the levers cell engineers actually pull: ideality factor, photocurrent, and (above all) dark saturation current.
In the field, the equation matters because it tells you how Voc changes with temperature, and that determines how many panels you can put in a string before your inverter's MPPT input voltage limit gets exceeded on a cold winter morning. Get this number wrong in a cold climate and you damage inverters; get it right and your strings deliver maximum DC bus voltage at exactly the right operating point.
Keep Reading
If you found this useful, these guides go deeper into related topics:
- Solar Panel Output Voltage Explained
- STC In Solar Panels — The Foundation Of Every Datasheet
- STC vs NOCT (NMOT) — Temperature Math And Modern Datasheet Comparison
- NMOT In Solar — The Faiman Thermal Model Explained
- How To Calculate Solar Panel Efficiency
- How Many Amps Does A 100 Watt Solar Panel Produce
- Standard Solar Panel Sizes And Wattages
- Average Peak Sun Hours By State
- Solar Panel Calculator — Full Energy Estimate
Frequently Asked Questions
What is the formula for open-circuit voltage of a solar cell?
What is open-circuit voltage in a solar panel?
What is the ideality factor in the diode equation?
Why does Voc decrease as temperature increases?
What is the dark saturation current I₀?
How does cold weather affect string voltage?
How do I calculate cold-morning string Voc for inverter sizing?
What is a typical Voc for a 2026 residential solar panel?
Why does the Voc equation collapse to Voc ≈ (kT/q) · ln(I_L / I₀) for real cells?
Sources
- Shockley, W. (1949) — The Theory of p-n Junctions in Semiconductors and p-n Junction Transistors. Bell System Technical Journal 28, 435–489
- PVEducation — Open-Circuit Voltage
- PVEducation — The Diode Equation
- PVEducation — Effect Of Temperature
- PVEducation — Ideality Factor
- Sze, S.M. & Ng, K.K. (2006) — Physics of Semiconductor Devices, 3rd Edition. Wiley-Interscience
- Green, M.A. (1981) — Solar cell fill factors: General graph and empirical expressions. Solid-State Electronics 24, 788–789
- NEC 2023 Article 690.7 — Maximum Voltage (Source-Circuit Sizing)
- LONGi Hi-MO 6 LR5-54HTH datasheet — temperature coefficients (2024)
- REC Alpha Pure-R 430W datasheet — temperature coefficients (2024)
- JinkoSolar Tiger Neo 72HL4-(V) datasheet — temperature coefficients (2024)