What Is FR4 Material: The Name Is Not The Spec
FR4 material is a sheet of glass cloth set in epoxy. Most circuit boards are built on it. The name tells you two things. It says the sheet burns slowly. It says the sheet is a composite of glass and resin. It says nothing more.
That last point is the one buyers get wrong. FR4 is not one material with fixed properties. It is a family, and the members differ by a factor of four on the numbers that decide whether a board survives its own assembly.
So a drawing that says “FR4” and stops there has specified almost nothing. This article covers what the sheet is made of and which number predicts survival. It also covers how to write an order a supplier cannot quietly swap.
The Short Answer
The letters stand for flame retardant, grade four. The number is a position in a series of flame-retardant grades, not a thickness, not a layer count and not a resin type.
The sheet is a composite. Woven glass cloth gives it strength. Epoxy resin holds the cloth together and insulates it. A flame retardant is mixed into the resin so the finished board does not keep burning once the flame is removed.
By volume, glass is roughly a quarter to three fifths of the finished sheet. The rest is resin, filler and additives.
The board you hold is that composite, pressed flat, with copper foil bonded to one or both faces. A 1.6 mm sheet can hold up to eight plies of cloth, and the ply count is what sets the thickness.
Then the copper is etched into tracks, holes are drilled, and the assembly is built. All of that happens on the board that assembly turns into a product. FR4 is the raw sheet underneath it.
The finished sheet usually carries a flammability rating of V-0. That rating comes from a test, not from the name. A laminate sold as FR4 without a V-0 listing has not met the test.
What It Is Made Of
Start with the glass. It is E-glass, the electrical grade. Filaments are drawn about five to ten microns across, which is finer than a human hair.
Two hundred to four hundred of those filaments are bundled into a yarn. The yarn is woven on a loom, and the weave is sold by style number.
Four styles cover most boards: 106, 1080, 2116 and 7628. Thin styles go into the core of a multilayer board. Heavy styles go under thick copper, where the weave has to hold the sheet flat during pressing.
Then the resin. This is where two sheets with the same name stop being the same sheet. The base is epoxy. The hardener, the filler and the flame retardant are chosen by the laminate maker, and so is the cure.
Bromine is the traditional flame retardant. Halogen-free grades use phosphorus or nitrogen instead. Both can carry the same FR4 label.
Add copper. One ounce of copper spread over a square foot is about 35 microns thick, which is the weight most boards start from.
Nothing in that list is fixed by the name. Two suppliers can hand you an FR4 sheet that differs in glass style, resin content, cure, flame system, thickness and copper weight. Both are telling the truth.
Tg Is Not A Temperature Limit
This is the single most useful idea in the article, so it gets its own section.
Tg means glass transition temperature. It is the point where the resin stops behaving like a stiff solid and starts behaving like a soft one. It is not a melting point. The sheet does not turn to liquid at Tg, and it does not fail there either.
Standard FR4 sits near 135 °C. Many sheets are quoted between 130 °C and 140 °C. High-Tg grades are built at 170 °C and above.
Now compare that with the oven. Lead-free solder melts at 217 °C to 220 °C, and reflow runs above that. Every lead-free board passes through its own Tg on the way up and again on the way down.
Read that again, because it breaks the usual buying logic. A board with a 135 °C Tg is not “only good to 135 °C”. It is a board that is designed to cross 135 °C twice during the reflow pass that attaches the parts, and to survive it.
What changes at Tg is the rate of expansion. Below Tg, the resin expands about 50 to 70 parts per million for each degree of heat. Above Tg it expands at roughly 250 to 270.
That is a jump of four to five times, in one step, at one temperature. Copper expands at about 17 parts per million the whole way, so it does not follow.
The copper barrel in a plated hole is fixed to the resin at every layer. When the resin lurches outward and the copper does not, the barrel is stretched. Do that enough times and the barrel cracks.
That is the failure Tg is really about. Not a board that melts, but a joint that splits after enough cycles. The product then fails in the field instead of on the line.
Four numbers describe the thermal side of a sheet. They are not interchangeable, and most quotes give you one of them.
| Number | How it is measured | What it tells you | Standard FR4 | High-Tg FR4 |
|---|---|---|---|---|
| Tg | DSC or DMA | Where the resin changes state | 130–140 °C | 170–180 °C |
| Td | TGA, 5 % weight loss | When the resin starts to break apart | 300–310 °C | 340–345 °C |
| T288 | TMA at 288 °C | Minutes before the sheet delaminates | 0–1 min | 14–39 min |
| Z-axis expansion | TMA below and above Tg | How hard the barrel is pulled | 50–70 / 250–270 ppm | 50–60 / 250–260 ppm |
The two expansion figures in that last row are read as two numbers, one below Tg and one above it, in that order. A sheet can have the same room-temperature behaviour as another and still be twice as hard on the holes.
One Name, Many Grades
If the name is not a spec, what is? The answer is a slash sheet under a standard called IPC-4101.
That standard lists base materials and gives each family a number, written as a slash and a figure. A slash sheet fixes the minimums a sheet must reach to carry that label.
For glass-reinforced epoxy the sheets are grouped by heat class. The low group is built around a Tg minimum of 110 °C. The middle group asks for 150 °C. The high group asks for 170 °C, and also raises the decomposition temperature and the delamination time.
This is the table a buyer should keep. It comes from a paper presented at an IPC conference by two Dow chemists. It is the clearest public version of the slash-sheet values I have found.
| Slash sheet | Tg min, °C | Td, °C | T288, min |
|---|---|---|---|
| /121 | 110 | 310 | 5 |
| /99 filled | 150 | 325 | 5 |
| /124 | 150 | 325 | 5 |
| /126 filled | 170 | 340 | 15 |
| /129 | 170 | 340 | 15 |
| /26 | 170 | — | — |
Every sheet here caps moisture uptake at half a percent, apart from /26 at 0.8. All of them carry a z-axis expansion of 60 below Tg and 300 above it, in parts per million per degree. Total expansion between 50 and 260 °C runs from 3 to 4 percent. Within a heat class the filled sheet sits lower, and it drills more easily.
The filled and unfilled versions of the same heat class are a real choice, not a formality. Filler reduces expansion and helps the drill. It also makes the sheet more brittle. The same IPC paper rates standard FR4 as the toughest sheet on its own scale, with the high-heat grades behind it.
So the trade is real in both directions. Push the heat numbers up, and the sheet gets stiffer and more brittle. That is why a high-grade sheet is not automatically the right buy for a simple board.
The Two Numbers That Predict Survival
Tg gets all the attention. Td and T288 are the ones that predict whether a board comes out of the oven intact.
Td is the decomposition temperature. It is measured by heating a sample at a set rate and watching the weight. The figure quoted is the point where five percent of the weight has gone, as vapour.
For standard FR4 that lands near 300 °C. The lead-free slash sheets require 310, 325 or 340 °C depending on the class.
T288 is a different measurement and a more practical one. A sample is held at 288 °C inside a testing machine, and the clock runs until the sheet starts to come apart.
On a standard sheet that takes under a minute. On the high-heat grades it runs from fourteen to thirty-nine minutes, depending on the resin system.
Half an hour of margin sounds generous next to a reflow pass that lasts a few minutes. It is not generous at all, because the damage is additive. A board sees the oven, then a rework pass, then a second rework, then years of thermal cycles in service.
A sheet with one minute of margin spends that margin on the first pass. A sheet with half an hour has room for every pass after it.
This is the number to ask for, because it is the one a supplier cannot fake with a datasheet headline. Anyone quoting a high-heat grade should be able to send the test report for the lot.
When The Signal Cares About The Resin
For most boards the electrical side is quiet. Put a slow signal on FR4 and it arrives.
The two numbers that matter are the dielectric constant, written Dk, and the loss factor, written Df. Dk sets how fast a signal travels and how wide a track must be for a given impedance. Df sets how much of the signal dies on the way.
A common standard sheet is quoted at a Dk of 4.7 and a Df of 0.014, both at one megahertz. That is where the familiar “about 4.5” figure comes from.
Now the part that trips up drawings. Dk is not a property of the name. It is a property of the construction, because glass has a higher Dk than resin.
One laminate maker publishes design values for two cores from the same family. A single 1080 ply with 63 percent resin comes out at 3.73 at one gigahertz. A single 2116 ply with 44 percent resin comes out at 4.17.
Same family, same name on the box, half a point of Dk apart. Half a point is enough to move an impedance target off its mark.
Resin content also shifts with temperature, so a Dk quoted at room temperature is a starting point, not a promise. One write-up from a board shop explains the measurement problem. The same laminate can be quoted at two different Tg values, depending on whether the lab used DSC or DMA. Those two numbers should not be compared.
The practical rule follows. Ask the fabricator which Dk it used in the impedance calculation, and ask for the stackup that goes with it. A datasheet is for shortlisting. It does not define the board.
Above roughly ten gigabits per second, the loss of ordinary FR4 becomes the design problem rather than the layout. Low-loss FR4 grades exist and sit between standard sheet and the high-frequency materials. They lower Dk and Df, and they cost more than the standard sheet.
For radio work at microwave frequencies, or for a board that carries an antenna, the epoxy sheet runs out of answers. That is PTFE or a ceramic-filled laminate, and a different budget.
When FR4 Is The Wrong Board
FR4 wins on cost and on being understood by every factory. Four situations sit outside that win.
The first is heat that has to leave through the board. Epoxy and glass are insulators, and they behave like it. Through the thickness, standard FR4 conducts about 0.29 watts per metre per degree.
A high-Tg sheet does better at about 0.4. An alumina ceramic does better by a factor of forty or more, quoted between 14 and 24 watts per metre per degree.
So an LED or a power device that dumps heat into a pad has nowhere to send it on FR4. The heat spreads along the copper, which is thin, and then sits. That is a metal-clad board or a ceramic board, not a better grade of FR4.
The second is anything that bends. FR4 is rigid by design. A hinge, a folding section or a camera module that moves needs polyimide or a similar film, and the whole build changes with it.
The third is a hot environment. Think of a hood, a gearbox housing or a spot beside a warm casting. There the ambient heat plus the self-heating of the parts can sit near the sheet’s own limit. Then the answer is a high-heat grade with a tested Td, or polyimide if the margin is still too thin.
The fourth is a dense board in damp air with a voltage across it. Glass bundles can act as a path, and metal can creep along them over time. That is a slow failure that shows up months later, and it is one reason toughened sheets are preferred for fine-via work.
None of these four is exotic. Each one is a sentence on a drawing that decides the substrate. Each one is cheaper to write down than to discover after the first build.
What To Write On The Order
A drawing that names the grade is a drawing a factory can quote against. Four lines do most of the work.
One. Name the slash sheet. Write “IPC-4101/126” or “IPC-4101/121”, and put the words “or equivalent” out of the document. Equivalent is the hole through which a cheaper sheet walks in.
Two. Name the laminate maker and the grade. A grade name from a known maker is a thing a clerk can look up. It is also a thing an inspector can find printed on the sheet.
Three. Put the two heat numbers in with their test methods. A Td of 340 °C by TGA at five percent weight loss. A T288 of fifteen minutes. A number without its method cannot be checked.
Four. Fix the rest of the build. Glass style and ply count, finished thickness, copper weight, surface finish, and whether the sheet must be halogen-free.
Then add one line that is not about the sheet at all: the reports. Ask for the lot’s test data with the shipment. A supplier who has the grade can produce the paper.
This is the same discipline that a product specification sheet needs before a factory quotes. The laminate is one line of it, and it is the line most drawings leave out.
How To Catch A Swap
The failure this article is really about is not a bad sheet. It is a good sheet quoted and a cheaper sheet shipped, with the invoice unchanged.
Five checks catch most of it, and none of them needs a laboratory.
The mark. Laminate makers print their name and grade on the sheet edge or on the shipping label. Ask for a photo of the label before the board is cut.
The reports. Ask for the T288 and Td data for the lot. A factory that bought the grade has the data. A factory that bought the cheaper sheet has a datasheet, and a datasheet is not a lot report.
The cross-section. Cut a board and look at the plies under a microscope. Glass style and ply count should match the stackup you agreed. This is a five-minute check during a factory audit that includes the laminate store.
The method behind Tg. A supplier may quote a Tg that moves twenty degrees when you ask which test produced it. Then you are comparing two tests, not two sheets.
The price. The heat grades cost more per square metre than the standard sheet. A quote that matches standard FR4 down to the cent, against a spec that names a 170 °C sheet, is a quote for something else.
One more applies after the first build. Send a batch through the buyer’s own reflow profile, then cross-section a hole. A barrel that has already started to crack tells you what the sheet did in the oven. No document on the invoice can argue with it.
What The Grade Costs
It helps to know where the sheet sits against the alternatives. In one published comparison of substrate materials, FR4 is the reference at a relative cost of one.
BT epoxy comes in around five times that. Polyimide around six and a half. PTFE ranges far wider, from about thirty-two times the FR4 figure to nearly eighty.
Those figures come from an IC-substrate comparison, so they are an index rather than a price list. They still make the point: FR4 is cheap because glass and epoxy are cheap and every plant knows them.
The useful comparison is not one sheet against another. It is the sheet against the failure. The laminate is one line of a board quote. Now think about the other side. A scrapped lot, a cracked barrel found in the field or a week of line time lost to delamination is a different order of money.
That is the argument for paying for the grade when the application asks. It is also the argument against paying when it does not. A simple board at room temperature does not need a 170 °C sheet. A board that will be reworked twice and then run warm does.
Either way, the decision belongs on the drawing. What goes over the finished board is a separate decision with its own cost. The two are often written on the same page and confused with each other.
The acceptance side has its own document. The workmanship standard used at goods-in covers the solder joints and the plating, so it will not save a board whose laminate was chosen badly. That choice is made before the first panel is cut.
Send me the laminate line from your drawing and the quote that came back for it. I will say whether the grade is named or only implied, whether the heat numbers carry their test methods, and where a cheaper sheet could be swapped in without anyone noticing. No cost, no obligation.
Check My GradeQuestions Buyers Ask
Is FR4 the same as FR-4?
Yes, the hyphen is optional. Neither spelling specifies a grade. The letters name a flame-retardant composite family, and the grade still has to be written separately.
What does the number four mean?
It is a position in a series of flame-retardant grades. It is not a thickness, a layer count or a resin type, and a sheet can carry the label with a range of properties.
Is FR4 waterproof?
No. It takes up moisture, and the slash sheets cap that at half a percent. Absorbed water turns to steam in the oven, which is why panels are dried and packed before assembly.
Can FR4 be used at high frequency?
Only up to a point, and only with a low-loss grade. Standard sheet loses too much signal above roughly ten gigabits per second. Radio work at microwave frequencies needs another material.
What is high-Tg FR4?
A sheet cured to raise Tg to 170 °C or above, usually with a phenolic hardener. The useful change is the rise in Td and in T288, not a higher running temperature for the product.
Which thickness should I order?
1.6 mm covers most work. What the factory actually builds is a ply count, so agree the stackup rather than the millimetre figure alone, especially on a controlled-impedance board.
Questions & Comments
We read every one and reply within 24 hoursIf your drawing says FR4 and nothing more, send me the laminate line. I will say which grade the quote is really for.