Etch Factor Control In PCB Fabrication

The etch factor describes the shape of the copper that remains after a trace has been etched. A perfect etch would leave a rectangular cross section with vertical walls; a real etch leaves a trapezoid, because the etchant attacks the sides of the trace as well as the exposed copper at the bottom. The etch factor is the ratio of the copper thickness to the amount of sideways attack.

This article explains why the profile matters, how it forms, and how the artwork is compensated for it.

What The Etch Factor Describes

During etching, the resist covers the top of the trace and the exposed copper is dissolved away. The etchant reaches the sides of the trace underneath the resist and removes material there too, a process called undercut. The result is a trace whose top is the width defined by the resist and whose bottom is narrower, with a sidewall at an angle.

The etch factor is the thickness divided by the undercut per side, so a high etch factor means nearly vertical walls and a low one means a strongly tapered trace. The figure depends on the chemistry, the copper thickness, the pattern density and the equipment. It is measured on a section, and it is one of the parameters a fabricator uses to predict the finished width from the artwork. The way the artwork and the process interact in general is described under PCB design and fabrication.

How The Profile Forms

The undercut develops because the etchant is not renewed evenly. In a wide open area the etchant circulates freely and the copper is removed quickly. In a narrow space between two traces the etchant becomes depleted and the reaction products accumulate, so the etch is slower and the undercut is smaller. That is why two traces of the same drawn width on the same panel can finish at different widths if the copper density around them differs.

The direction of the spray and the pattern of the equipment add a further variation. A conveyorised etcher sprays from above and below, and the flow is not identical across the panel, so the leading edge and the trailing edge of a panel can differ. The variation is normally measured by etching a test pattern with lines of several widths and densities and measuring the finished width of each.

Copper Thickness And Its Effect

Thicker copper takes longer to etch, and the longer the etch runs, the more undercut develops at the top. A one ounce foil therefore etches to a more rectangular profile than a two ounce foil, and a heavy copper board has a strongly tapered trace unless the process is adjusted. This is one reason why fine lines and heavy copper do not combine well: the etch that removes two ounce copper reliably also removes the sides of the fine lines.

The finish that is applied afterwards does not change the profile, but it does change the surface. A trace with a tapered wall has a different cross sectional area from one with vertical walls of the same top width, which affects its resistance and its current carrying capacity. The calculation of that capacity is described under trace width and current calculation, and the area used in it should be the actual area rather than the drawn rectangle.

Compensating The Artwork

Because the finished trace is narrower than the artwork, the artwork is normally compensated by the fabricator, who adds a small amount to every trace width so that the finished line meets the design. The compensation depends on the measured etch factor for the specific product, so it changes with the copper thickness, the line width and the density of the pattern.

Where a design specifies a finished width, the fabricator can apply the compensation. Where the design specifies only the artwork width, the finished trace may not be what the designer intended, particularly at fine lines where the same compensation that is correct for one width is wrong for another. This is why a controlled impedance design states the target impedance and the tolerance rather than the width alone, and lets the fabricator adjust the width to reach it. The pad dimensions that go with the traces are treated under PCB pad design standards.

Process Control

The controls are the chemistry, the temperature, the conveyor speed and the spray pressure, and all four are monitored. The concentration of the etchant is maintained by dosing, because the copper it dissolves accumulates and the etch rate changes with it. The temperature is controlled because the reaction rate is strongly temperature dependent. The speed and pressure set the dwell time and the flow, and they are the parameters that are adjusted most often to compensate for a change in the others.

The verification is a test pattern etched with the production panels and measured in a laboratory. The pattern includes lines of the widths and densities used on the board, so that the compensation for each can be checked. A change in the measured width without a change in the settings usually means a change in the chemistry or in the panel itself, and it is worth investigating before the compensation is adjusted to hide it.

Consequences For The Design

Three design decisions depend on the etch behaviour. The first is the minimum line width, which is limited by the copper thickness the design requires. The second is the spacing, because a narrow gap between two traces is the region where the undercut is smallest and the width is largest, which is the opposite of what the usual tolerance reasoning would predict. The third is the tolerance to be specified on the finished width, which should reflect what the process can hold rather than what would be ideal.

Where the design has a wide range of line widths on the same layer, the tolerance is set by the worst case, which is usually the widest line in a dense area. Splitting the design so that fine lines and heavy copper are on different layers, or so that the fine lines are in a region of uniform density, is a layout measure that improves the etch result without changing the process.

One further consequence is worth stating. Because the etch attacks the sides of a trace, a conductor drawn to the minimum width and then etched may finish below the minimum, and the defect appears as a trace that fails a width check while everything around it passes. Specifying a finished minimum rather than a drawn one, and asking the fabricator to confirm what it can hold on the actual copper weight, removes that risk before the panel is ever etched.

FAQ

Is a higher etch factor better? A high etch factor means a more rectangular trace, which is closer to the design intent. It is achieved with thinner copper, a well controlled chemistry and a shorter etch, and it is limited by the thickness the design needs.

Why does the finished width vary across a panel? Because the etch rate varies with the local copper density, the spray pattern and the position on the panel. Etching a test pattern with the product quantifies the variation rather than leaving it unknown.

Does the etch factor affect impedance? It does, because the cross section and therefore the capacitance and inductance per unit length depend on the wall profile. Controlled impedance designs are compensated for the measured profile rather than for a nominal rectangle.