The least interesting part of a solar panel is the glass you see. The real work happens in a wafer of silicon about twice the thickness of a human hair, where light becomes moving electrons without a single moving part. No steam. No turbine. No fuel. No moving parts that wear out. That sound you hear from a rooftop array is mostly wind.

A solar panel is a photon-to-electron machine. It takes packets of light, knocks electrons loose inside a semiconductor, and sends them through a circuit before they return home. The device is so reliable that the main failure points are not the cells but the wiring, connectors, mounting clips, and the inverter bolted to the wall. That distinction explains most of what you need to know.

From photon to electron: how a panel actually works

The photovoltaic (PV) effect, first observed by French physicist Edmond Becquerel in 1839, is not a modern invention. Becquerel found that certain materials produce a small electric current when exposed to light. Modern panels industrialise that discovery with silicon, chemical doping, and a built-in electric field.

The photovoltaic effect, without the jargon

Silicon is a semiconductor: it conducts electricity better than glass but worse than copper. In its pure form, silicon's electrons mostly stay put. To make them useful, manufacturers add tiny amounts of other elements. Adding phosphorus creates n-type silicon with extra electrons. Adding boron creates p-type silicon with missing electrons, called holes. When the two layers meet, electrons diffuse across the junction and leave behind a charged region. That region creates an electric field, a one-way gate for electrons.

When a photon with enough energy hits the cell, it knocks an electron free. The electric field at the junction pushes the electron toward the front contacts and the hole toward the back. Connect a wire from front to back and the electron flows through it. That flow is direct current (DC). Sunlight in, electricity out.

How a solar cell is built

Most cells are slices of silicon, either a single crystal (monocrystalline) or multiple crystals fused together (polycrystalline). The slices are treated to reduce reflection so they absorb more light instead of bouncing it away. Thin metal fingers on the front collect electrons; a solid layer on the back completes the circuit. A single cell produces only about 0.6 volts, so dozens of cells are wired in series inside a panel to reach useful voltages. A typical home panel now has 60 or 72 cells, or half-cut cells arranged into smaller groups.

That is the basic unit. The panel adds tempered glass, a polymer encapsulant, a weather-resistant backsheet, and an aluminium frame. It has no moving parts and no fuel.

Why panels lose output in the real world

A panel's nameplate rating comes from standard test conditions: 1,000 watts of irradiance per square metre, a cell temperature of 25°C, and a specific light spectrum. The rating is useful for comparison and almost never matches a rooftop. On a real roof, heat is the quiet thief. Most silicon panels lose between 0.3 and 0.5 percent of output for every degree Celsius above 25°C. On a hot day, cells can sit at 60°C or more, so a panel rated at 400 watts may produce well below that. The U.S. Department of Energy explains how heat, orientation, and shade affect real-world output.

Shade is more complicated than people expect. Because cells are wired in series, a single heavily shaded cell can act as a bottleneck for a whole string. Bypass diodes let current skip blocked groups, but only in pieces. That is why a small shadow from a chimney can cut output far more than the shaded area alone. It is also why microinverters and power optimizers, which manage output at the panel level, have become common on complicated roofs.

The inverter is the part that actually talks to your house

The DC electricity from a panel is not what a refrigerator or grid transformer expects. Alternating current flips direction 50 or 60 times per second depending on the country. The inverter synthesises that wave from DC. A string inverter handles a whole array, usually near the meter. Microinverters sit under each panel and convert DC at the source. Power optimizers are a middle path: they adjust each panel's DC output before sending it to a central inverter. No single design is best for every roof; complex shade favours panel-level electronics, simple sun-drenched arrays do fine with a string inverter.

Inverter lifespan is the most underestimated part of a solar quote. The panels may carry a 25-year performance warranty, but a central inverter often carries 10 to 12 years. If the quote does not include inverter replacement costs, it is incomplete. The same applies to monitoring: any well-designed system should show you daily and hourly production, which makes underperformance detectable before it becomes an annual surprise.

Grid-tied systems do not automatically keep your lights on during a blackout. Safety rules require them to switch off when the grid goes down, so line workers are not electrocuted. If you want backup power, you need a battery or a hybrid inverter with an emergency circuit. A battery changes the economics, adding several thousand dollars and its own replacement timeline. It is a separate purchase decision, not a standard feature.

Here's the catch

The catch is not that solar panels fail. The catch is that the numbers most buyers see are laboratory numbers, not roof numbers. Commercial silicon panels have improved from around 15 percent efficiency in the 1990s to about 20 to 23 percent today. Lab cells have gone much higher. The National Renewable Energy Laboratory efficiency chart tracks experimental cells above 47 percent, but those are four-junction concentrator cells under lenses, not panels you can buy. A product you buy only captures a fraction of what the record chart suggests.

There is also the inverter. The cells make DC, but the grid and your house run on alternating current (AC). The inverter converts the two, and no conversion is perfect. Good inverters run at 95 to 98 percent efficiency. Add heat losses, wiring losses, soiling, and the gradual decline of panels over two or three decades, and a 400-watt nameplate rating is best understood as a starting point, not a promise.

None of this means rooftop solar is a bad deal. It means you should compare actual annual production, not peak wattage. A proper estimate uses local irradiance data, roof orientation, shade, temperature, and a monthly forecast. If a quote promises perpetual peak output, treat it the same way you would treat a car that only lists highway mileage.

Hype and reality, graded

• Myth: Solar panels need direct sunlight. Reality: They produce in diffuse light, but output drops to roughly 10 to 25 percent of full sun under heavy cloud. They still work on overcast days; they just work slower.
• Myth: Panel efficiency keeps doubling. Reality: Commercial silicon has crept upward slowly, not doubled. The big cost declines have come from manufacturing scale and supply-chain learning, not from panels converting twice as much light.
• Myth: Solar is maintenance-free. Reality: The cells are solid-state, but inverters typically need replacement after 10 to 15 years. Wiring, connectors, mounting hardware, and drains need inspection. Dirt and snow can cut output.
• Myth: Solar panels are hard to recycle. Reality: Panels are mostly glass, aluminium, silicon, and small amounts of copper or silver. Recycling systems are young but expanding, and several countries now require manufacturers to take panels back.

What the numbers mean on a roof

Average solar irradiance matters more than panel count. In sunny regions, a 5-kilowatt system might produce around 7,000 to 9,000 kilowatt-hours a year. In a cloudier place, the same system might produce half as much. The International Energy Agency's Solar PV tracker documents how these patterns shape global growth, with solar adding more capacity than any other power source in recent years.

Cost has also shifted. Solar module prices fell by more than 90 percent since 2010, mostly because of manufacturing expansion. The expensive parts of a home installation today are often not the panels but the racking, wiring, inverter, permits, and labour. In many places, solar is now the cheapest way to add new generation capacity. That is why cumulative global solar capacity passed 2 terawatts in 2024, driven less by climate rhetoric than by simple economics.

Gregory Nemet at the University of Wisconsin-Madison has spent a career documenting how solar got cheap through thousands of small manufacturing improvements, not a single eureka moment. Panels are not magic, he argues. They are the result of accumulated learning, and that is exactly why they are so hard to beat.

Photo by American Public Power Association on Unsplash