Copper Foil Surface Profile and High Frequency Signal Loss

At low frequencies the surface of a copper trace is just a surface. At microwave frequencies the current no longer flows evenly through the conductor, it crowds towards the outside, and the texture of the copper becomes part of the electrical path. A foil that looks identical to the eye can change the insertion loss of a line by a measurable amount, which is why roughness has become a design parameter rather than a fabrication detail. The material data sheet is the place where the copper and the dielectric meet, and it is the combination that decides how a line behaves.

Why Surface Roughness Matters at High Frequency

As frequency rises, current concentrates in a thin layer near the surface of the conductor. That layer becomes thinner than the surface texture of most copper foils, so the current has to follow the profile of the foil rather than a smooth plane, and the path it takes is longer than the geometric length of the trace.

The longer path increases resistance and loss, and the effect grows with frequency. At a few gigahertz the difference between a smooth foil and a rough one can be a significant fraction of the total insertion loss, which is why material datasheets for high speed laminates now quote a roughness figure alongside the dielectric properties. The effect is not a fixed penalty either, since it scales with the length of the line and with the frequency of the signal travelling along it.

How Foil is Made and What Profile Means

Electrodeposited foil is grown on a rotating drum, and the side that faces the drum is smooth while the side that faces the plating solution is rough. That rough side is the one bonded to the laminate, because the texture is what gives the adhesion that holds the conductor in place.

Rolled and annealed foil is produced by rolling, and it has a different and generally smoother surface, which suits PTFE based materials. The profile of a foil is described by the roughness of the treated side, and that figure determines both the peel strength and the electrical loss of the finished trace. A supplier will normally quote the roughness of the treated side together with the treatment weight, because the two together define the bond.

Skin Effect and the Current Path

Skin depth falls as the square root of frequency, so the layer carrying most of the current becomes thinner as the signal gets faster. When the skin depth becomes comparable to the roughness, the current follows the hills and the valleys instead of a flat surface.

The extra path length is the mechanism behind the increase in loss. It also changes the effective permittivity seen by the wave, because part of the field sits in the laminate and part in the copper surface region, which is why the measured impedance of a line can differ from the calculated value. The discrepancy grows with roughness and with frequency, and it is one of the reasons a design that simulates cleanly can measure slightly outside its tolerance.

Measuring Roughness

Roughness is measured with a profilometer or by atomic force microscopy, and the result is reported as an average figure and as a peak figure. The two are not interchangeable, because a surface with a few deep peaks behaves differently from one with a uniform fine texture.

The measurement is taken on the treated side of the foil and, on a finished board, on the copper after lamination and etching. The treatment can change during processing, and the figure that matters is the one at the interface with the dielectric. A coupon etched with the production panel is the practical sample, since it has seen the same lamination and etching as the boards it represents.

Foil Types and Their Trade-offs

The available choices run from standard electrodeposited foil through low profile versions to rolled foil and to foils with a very fine treatment for high frequency work. Each step towards a smoother surface reduces the loss and also reduces the adhesion.

The trade is real, because a very smooth foil needs a special treatment or a bonding layer to reach the peel strength that a standard foil achieves easily. The laminate supplier normally specifies which foil has been qualified with which material, and deviating from that combination moves the risk from the supplier to the designer.

Adhesion, Peel Strength and the Other Side of the Argument

Adhesion is what keeps the conductor attached through drilling, plating and thermal cycling, and peel strength is the measure of it. A foil that is too smooth for the treatment applied will delaminate at a pad or a large plane, which is a far more expensive failure than a small loss penalty.

The answer is usually a matched system rather than a single choice: a low profile foil with a treatment designed for it, on a laminate whose resin chemistry bonds to that treatment. Choosing the two separately is how a design ends up with either poor adhesion or unnecessary loss.

Choosing a Foil for a Design

The decision starts from the highest frequency present and the longest line on the board, because those two together determine how much the roughness matters. A design that runs at a few hundred megahertz is unlikely to be affected, while one at ten gigahertz with long lines is dominated by it.

The second input is the loss budget. Where that budget is tight, the roughness figure and the dissipation factor of the laminate are the two largest terms after the geometry, and both should be reviewed together rather than in isolation. Where the two are in conflict, the laminate choice usually settles the question, because it sets both the loss and the range of foils that will bond to it.

What the Fabricator Can Control

The fabricator controls the choice of foil within the material specified, the lamination cycle that sets how the foil bonds, and the etching that defines the trace. Etching removes copper and exposes the treated surface at the sidewalls, which changes the effective roughness of a narrow trace.

Plating adds a layer whose own surface is rough, so a plated trace behaves differently from one formed from foil alone. That difference is one reason the same design can measure differently when the copper weight or the plating process changes. The panel plating and the pattern plating steps each add copper, and the surface they produce is not the same as the original foil surface.

Verification and Documentation

Verification uses a test coupon with a known line length, measured with a vector network analyser to extract insertion loss. Comparing the result with the value calculated from the material data shows whether the board behaves as the design expected.

The documentation should record the laminate, the foil type, the copper thickness and the measured loss, so that a future build can be compared with the one that was qualified. The material properties that accompany the foil are described in the guide to high frequency laminates.

FAQ

Does roughness matter on a digital board? It can, because the harmonics of a fast edge reach high frequencies. A design with rise times in the tens of picoseconds behaves like a high frequency design even when the clock itself is slow.

Can roughness be measured on a finished board? It can be estimated from a cross section or from a coupon, but the measurement is not routine. The usual approach is to control the foil specification and verify the result through insertion loss.

Is a smooth foil always better? No. Smoothness reduces loss but also reduces adhesion, and a board that delaminates is worse than a board with slightly higher loss. The foil and the laminate should be selected as a pair.