Is This Your Problem?
Thermal cycling and physical stress crack silicon wafers and cut output over time. How Tucson owners detect micro-cracking and what it costs.
Get Matched With a Local ProThermal cycling and physical stress cause microscopic cracks across silicon wafers, reducing power production over time
Micro-cracking is the most deceptive failure mode on a solar array because it is invisible from every angle an owner can actually look from. There is no broken glass, no error code, no obvious stain. The cell has fractured along a line too fine to see, part of the wafer is now electrically isolated, and the module quietly produces less than its neighbours. In a climate with large day-to-night temperature swings, the conditions that propagate those cracks are present every single day of the array's life.
Why a cracked wafer keeps working, badly
Silicon solar wafers are thin, and the finished cell is a brittle crystalline sheet laminated between glass and a polymer back sheet. A micro-crack is a fracture that does not sever the module: current can no longer flow across the crack line efficiently, but it finds paths around it, so the cell still conducts — at reduced efficiency. Because cells are wired in series, the damaged cell now limits the current of the whole string, and the module's output drops by more than the physical area lost would suggest.
The desert makes the propagation worse. Every day the module heats in the sun and cools at night, and the laminate, glass, silicon and back sheet expand and contract at different rates. A latent crack grows a little with each cycle. That is the thermal-cycling mechanism in plain terms, and a Tucson roof provides thousands of these cycles over a 25-year asset life. It is the same mechanical reality that shows up as electrical damage elsewhere in the system; a related failure path is described under bird droppings and hotspot precursors, since a crack and a shaded cell can both cause a hot cell.
Physical stress is the other half. Cracks are commonly introduced long before an owner ever sees them: handling and stacking at the factory, freight, being set down on a corner during delivery, installers stepping on modules or kneeling on the glass, over-torqued clamps at the wrong points on the frame, flexing a module during lifting, and wind or hail loads after commissioning. A module can leave the factory with a hairline crack that survives the install and only becomes a measurable problem after two or three seasons of thermal cycling widen it.
| Cause | When it happens | What it does to the wafer | Why it stays hidden |
|---|---|---|---|
| Manufacturing and cell handling | Before delivery | Latent cracks in the wafer before lamination | Not visible through glass; may not affect output for years |
| Transport and delivery | Before installation | Corner and edge cracks from stacking or impacts | Packaging looks intact; damage is inside the laminate |
| Installation handling | During the install | Cracks from stepping on modules, kneeling, or flexing during lifting | The installer rarely sees a consequence on the day, so it is not reported |
| Clamp placement and torque | During the install | Stress concentrated where the frame meets the glass | Passes a normal commissioning check and shows up later |
| Thermal cycling, day and night | Every year of service | Slow crack propagation through repeated expansion and contraction | Gradual, so it looks like ordinary degradation |
| Wind, hail and storm loading | Storm events, especially monsoon | New cracks or sudden extension of existing ones | Damage is often internal; the glass may not shatter |
| Thermal stress from hot cells | Hot, high-irradiance afternoons | Further mechanical and electrical degradation around a damaged cell | Compounds quietly with every subsequent hot day |
The dominant pattern locally is thermal cycling, because the day-to-night temperature swing repeats daily. Physical causes usually set the crack; thermal cycling is what makes it matter.
What micro-cracking costs the homeowner
Micro-cracking cost is a stacking problem. Modules experience 1.0%–2.0% first-year power loss followed by 0.25%–0.55% annual output degradation over a 25-year asset lifecycle. That is the baseline every array carries. Cracking adds a second, unplanned loss on top of the baseline, and because it usually affects only some modules, it also creates mismatch: the cracked module drags its string, and the loss is larger than the module's own shortfall.
Worked example, illustrative and easy to redo: a Tucson array producing 14,000 kWh a year at a blended 16¢ per kWh. If cracking costs an extra 5% of production above normal degradation, that is 700 kWh a year, about $112, and roughly $2,800 of undiscounted value across the array's life. A single module replacement, by contrast, carries the diagnosis, the part, the labor and the re-flashing at the attachment points. The economics favour finding cracks early, but detection is the part owners skip, because a commissioned array looks finished and nobody wants to hear that the baseline should have been measured on day one.
| Loss component | Stated or illustrative band | kWh per year at 14,000 kWh | Value per year at 16¢ | Value over 25 years |
|---|---|---|---|---|
| First-year power loss (normal) | 1.0%–2.0% | 140–280 | $22–$45 | $560–$1,120 |
| Ongoing annual degradation (normal) | 0.25%–0.55% per year | 35–77 in the first year | $6–$12 | Compounds across the life |
| Additional loss from micro-cracking (illustrative) | +2% | 280 | $45 | $1,120 |
| Additional loss from micro-cracking (illustrative) | +5% | 700 | $112 | $2,800 |
| Additional loss from micro-cracking (illustrative) | +8% | 1,120 | $179 | $4,480 |
The first two rows use the stated first-year and annual degradation figures. The cracking rows are planning bands added on top of the normal baseline, not claims about any specific array.
How to detect micro-cracking
You cannot see micro-cracks by eye, and you cannot find them with string-level monitoring alone. What you can do is look for their consequences and use the right tools when the pattern points at them.
- Watch for module-to-module divergence. With optimisers or microinverters, a panel that drifts below its neighbours over months is a suspect. Without them, compare string currents; a string consistently lower than its peers, after accounting for shading, is the clue.
- Look for the visible residue. Snail trails — silvery discoloured lines on the cell surface — are a common visible companion to cracked cells, and darker grid lines or a patch of discolouration around a cell are worth photographing.
- Get electroluminescence imaging. Electroluminescence imaging is the definitive check: the module is energised in darkness and cracks appear as dark lines in the image. It is also the right way to establish a commissioning baseline on a new array, so later damage can be proven rather than argued about.
- Trace a current-voltage curve. A curve trace quantifies how much the module has actually lost and separates a cracked module from a soiled one.
- Use thermal imaging to find the consequences. A crack can create a hot cell, so thermal imaging is a practical screening method that points you at the modules worth testing properly.
- Re-check after storms. Monsoon wind and hail are the events that extend existing cracks. A post-storm inspection and a re-check of the electrical data is the cheapest way to catch damage early.
| Method | What it detects | Strength | Limitation |
|---|---|---|---|
| String-level monitoring | Aggregate output and trend | Already installed on most systems; free to read | Cannot localise damage to a module or a cell |
| Module-level monitoring | Which panel is underperforming | Narrows the problem to one unit quickly | Does not say whether the cause is a crack, a diode or soiling |
| Physical inspection for snail trails and discolouration | Visible signs of cell damage | No equipment needed; do it at every cleaning | Many cracked cells show no visible sign at all |
| Thermal imaging | Hot cells and anomalies on a sunny, still day | Screening across a whole array in one visit | Shows heat, not the crack itself |
| Current-voltage curve trace | How much output the module has lost | Quantifies severity before you replace anything | Per-module or per-string work; requires access and equipment |
| Electroluminescence imaging | The cracks themselves | Definitive evidence; works as a commissioning baseline | Needs darkness and a power supply; a specialist service |
For warranty purposes, an electroluminescence baseline taken at commissioning is the difference between a documented claim and an argument about what the array looked like when it was new.
What to do about it
Prevention happens in two windows. The first is during design and installation: modules should be handled and stored per the manufacturer's instructions, never walked on or knelt on, lifted without flexing, clamped only at the frame locations the manufacturer specifies and torqued to the specified value, and racked so that wind loading is carried by the structure rather than the module. Ask what the installer's handling and clamping practice is, and ask for the clamp positions to be documented. It is a reasonable question and a revealing one.
The second window is commissioning. Have a baseline recorded while the array is new, and if electroluminescence imaging is available, use it then. An array with no baseline is an array where every later defect is unprovable, and a warranty conversation with no evidence is a conversation you usually lose. Record serial numbers at the same time, mapped to physical positions on the roof, so a single underperforming position can be traced to a specific unit later.
Once a crack is confirmed, the practical response is targeted replacement rather than a rebuild. Replace the affected module with a matched unit, and re-check the string afterwards, because the string's output should improve by more than the module's own rating. If several modules on the same roof area are affected, look for a shared cause — a handling pattern during the original install, a clamp position, a hail path, or a section of roof that runs hottest — and fix that before replacing parts that will crack again.
Finally, keep the loss in perspective. The stated baseline degradation of 1.0%–2.0% in the first year and 0.25%–0.55% annually is normal and expected; not every dip is a crack. The reason to test is that cracking sits on top of that baseline, and the reason to test early is that the difference between a documented 2% loss and an undocumented 8% loss is thousands of dollars over the array's life.
Related Tucson problems
If the yield decline came on gradually and no crack can be confirmed, rule out the cheaper explanations first: soiling from dust and pollen and cell efficiency losses from high heat. If a module is discoloured or a hot cell has been found, read bird droppings and hotspot precursors. The full local index is at common solar problems in Tucson, and a step change rather than a drift is covered at system underperforming.
Next Step
Every case above resolves the same way in practice: someone qualified looks at the actual array, measures what is really happening, and only then prices a fix. Call (520) 593-0496 or request a free match with up to three local pros who work on Tucson roofs every week — no cost, no obligation.
