How does a solar panel make electricity? From photon to socket

You look at a roof: dark blue, glass-covered rectangles lined up row after row. There are no moving parts, no sound and no fuel. Yet the cables coming out of those panels run the fridge in the house. In a wind turbine or a dam’s turbine we see a shaft spinning; in a solar panel nothing moves. The electricity comes from light itself and from a thin boundary inside the panel. In this article we follow a photon on its journey from the panel’s glass to its cable.

Light is made of particles: the photon

We usually think of light as a wave, but light hands its energy to matter in small packets. These packets are called photons. A photon’s energy depends on the colour of the light: photons of blue light carry more energy than photons of red light. Albert Einstein described this packet-like behaviour of light in 1905, when he explained how light knocks electrons out of a metal surface (the photoelectric effect); his 1921 Nobel Prize in Physics was awarded for this work.

A solar panel is a machine that catches these packets one by one. At noon on a clear day, about 1000 watts of light fall on every square metre of a surface held at right angles to the Sun’s rays. Panel makers measure their panels under this condition (1000 W/m², 25 °C cell temperature) and print the power on the label accordingly.

Silicon and the p-n junction

Most panels are made of silicon. Silicon is the main element of the quartz in sand; it is purified and cut into slices about a fifth of a millimetre thick. In pure silicon the electrons are held tightly by their atoms. Freeing an electron from this bond so that it can move around takes a certain minimum energy; in silicon this is about 1.1 electronvolts. Photons of visible light and of the near infrared clear this threshold.

When a photon frees an electron, it leaves behind the place the electron vacated. This place is called a hole and behaves like a positive charge. In this way the photon creates an electron–hole pair. On its own, though, the pair is useless: after wandering for a while the electron drops back into a hole and the energy is lost as heat. To make electricity, the pair must be separated, with the electron sent one way and the hole the other.

The p-n junction does this. Most of a cell is p-type silicon with a tiny amount of boron added; boron creates an excess of holes. The very thin layer just below the top surface, where the light enters, has phosphorus added instead; this n-type layer has an excess of electrons. Where the two layers meet, some electrons cross from the n side to the p side and some holes cross the other way, and a permanent electric field forms at the boundary. This field works like a one-way slope: when a pair forms near the junction, its electron is pushed into the n layer and its hole into the p layer.

The electrons collect in the n layer and reach the thin silver contact fingers on the top surface. The fingers gather them into wider strips (busbars). The electrons leave through the cable, run a lamp or a motor, come back through the aluminium contact on the back of the cell and join the holes waiting in the p layer. The loop is closed; as long as the light lasts, so does the current.

Half a volt per cell: why are cells connected in series?

Across the terminals of a silicon cell there are about 0.6–0.7 volts with the circuit open, and about 0.5 volts at the point where the cell delivers the most power. This voltage does not depend on the size of the cell; it comes from the electric field at the junction. A bigger cell gives more current, not more voltage. When the light dims, it is mostly the current that drops; the voltage falls only a little.

Half a volt is not enough to run anything. So cells are connected in series: the front (−) of one cell is soldered to the back (+) of the next with thin copper strips. The voltages of cells in series add up, and the same current passes through every cell. A common panel has 60 cells in series: 60 × 0.5 V = about 30 volts. The panels on a roof are usually connected in series as well, to form a string of a few hundred volts. When more current is needed, strings are connected in parallel: in parallel branches the voltage stays the same and the currents add up. We go through these rules step by step in our series and parallel circuit post and in our series circuits and parallel circuits lessons.

Efficiency: how much of the light becomes electricity?

The silicon panels used on homes today turn about 20% of the light power falling on them into electricity. Part of the rest is reflected, and most of it becomes heat. Why not more?

• Photons without enough energy: photons from the far infrared cannot clear the 1.1 electronvolt threshold; they pass through the silicon or only warm it.
• Extra energy turns into heat: the energy a blue photon has above the threshold cannot be used and is lost as heat. Every photon, whatever its colour, creates at most one pair.
• Recombination and resistance: some pairs recombine before the junction can separate them, and the contacts and strips put a small resistance in the way of the current. The silver fingers on top also shade part of the light.

In 1961 William Shockley and Hans-Joachim Queisser calculated the theoretical upper limit for a single-junction cell in sunlight. With today’s standard solar spectrum, this Shockley–Queisser limit is about 33% for silicon’s 1.1 eV band gap. Once silicon’s own intrinsic losses (Auger recombination) are taken into account, the limit is around 29%. The first practical silicon cell, made at Bell Laboratories in 1954, had an efficiency of about 6%. The best silicon cells in the laboratory today are just above 27%; multi-junction cells that stack different materials go even higher but cost much more.

We can estimate a panel’s power with a single product:

P = G · A · η · cos θ

Here G is the irradiance perpendicular to the Sun’s rays (W/m²), A is the total area of the cells (m²), η is the efficiency and θ is the angle between the rays and the line perpendicular to the panel’s surface. In a 60-cell panel each cell is 15.6 cm × 15.6 cm, so the total cell area is about 1.46 m².

Worked example. The Sun is 60° above the horizon. At this height the light takes a slightly longer path through the atmosphere; Meinel’s air-mass formula gives an irradiance perpendicular to the Sun of about G ≈ 918 W/m². If the panel is tilted 30° and faces south, the rays hit its surface straight on (θ = 0°, cos θ = 1):

P ≈ 918 · 1.46 · 0.20 · 1 ≈ 268 W

Since the voltage is 60 × 0.5 V ≈ 30 V, the current is I = P / V ≈ 268 / 30 ≈ 9 A. You can change these numbers yourself in the figure below.

Angle and shade: same panel, different power

Hold a torch straight down at a table, then tilt it: the same light spreads over a larger area and every point looks dimmer. A panel works the same way. When the rays reach the surface at an angle θ, the power falling on each square metre of the panel drops by a factor of cos θ. A 30° offset reduces the power by only 13% (cos 30° ≈ 0.87); a 60° offset halves it (cos 60° = 0.5).

How high the Sun climbs at noon depends on the season and the latitude. In Ankara (about 40° north) the noon Sun rises about 73° above the horizon at the end of June, but only about 27° at the end of December. We explain why in our seasons post and seasons lesson. With the same model:

• Noon in December (Sun at 27°): a flat panel (θ = 63°) makes about 98 W. Tilt the panel to 40° and θ becomes 23°; the power rises to about 199 W, twice as much.
• Noon in June (Sun at 73°): a flat panel makes about 263 W and a panel tilted at 40° about 253 W. The difference is small.

That is why the tilt of fixed panels is usually chosen close to the local latitude, as a good compromise for the whole year. Motorised mounts that follow the Sun through the day produce more, but they cost more and need maintenance.

The shadow of a single leaf. In a series string the same current has to pass through every cell, so the current of the string is limited by the weakest cell. If a leaf falls on a cell, only diffuse light from the sky reaches that cell (a tenth in this article’s model). Without any protection the current of all 60 cells would drop to that level and the panel would lose nine tenths of its power. On top of that, the shaded cell is forced in reverse by the other cells and heats up; this heating is called a hot spot.

To prevent this, the panel’s junction box contains bypass diodes. A 60-cell panel usually has three diodes, each connected in parallel with a substring of 20 cells. If a substring contains a shaded cell, the current passes through the diode instead of entering that substring. The panel loses the voltage of that substring, but the other 40 cells keep working at full current. In the example, the 268 W panel makes about 179 W with one shaded cell thanks to the bypass diode; without the diode it would drop to about 27 W. Shade should not be underestimated: a chimney, an antenna or a layer of dust can have a similar effect.

The inverter: from direct current to the socket

The current coming out of a panel is direct current (DC): the electrons always flow in the same direction. The sockets in a home supply alternating current (AC); in Türkiye it is 230 volts and goes through 50 full cycles a second, so its direction changes 100 times a second. The inverter does the conversion between them. It uses transistors to switch the direct current on and off thousands of times a second and filters these pulses into a wave that keeps time with the grid.

A good inverter does one more job: it keeps searching for the voltage at which the panel gives the most power in the current light. This is called maximum power point tracking (MPPT). When a cloud passes or a shadow moves, the panel’s best operating point changes, and the inverter adjusts the voltage to match. In grid-connected systems the inverter also switches itself off when the grid power goes out, so that technicians working on the line stay safe.

Temperature matters too: silicon cells lose voltage as they warm up. That is why a panel can work more efficiently on a cool, sunny spring day than at noon on a scorching summer day. The calculations in this article leave temperature out.

Summary

• Sunlight is made of photons. In silicon, a photon with enough energy creates an electron–hole pair.
• The electric field at the p-n junction separates the pair: the electron goes to the n layer on top and to the silver fingers, the hole to the p layer below; the electrons travel around the outer circuit and come back.
• A cell gives about 0.5 V; 60 cells connected in series give about 30 V. In series the voltages add up, in parallel the currents.
• Today’s silicon panels turn about 20% of the light into electricity; roughly P = G · A · η · cos θ.
• When the rays hit the panel at a slant, the power drops by a factor of cos θ; because the winter Sun is low, a tilted panel makes much more in winter.
• In a series string a single shaded cell holds back the whole string; bypass diodes route the current around the shaded substring and limit the loss.
• The inverter turns the panel’s direct current into the 230 V, 50 Hz alternating current of the home and tracks the maximum power point.