Underetch and Overetch: Holding Trace Width in Etching

Underetch is incomplete removal of copper that should have been dissolved, while overetch is removal of copper that should have been protected. Both change the finished trace width, and both are symptoms of a process that has moved away from its balanced condition.

The two are not simply opposite settings of one control. Underetch usually points to an exhausted bath, insufficient dwell or a resist pattern that prevents the etchant from reaching the copper, while overetch points to an aggressive bath or a resist that has lost its adhesion.

What Underetch and Overetch Look Like

Underetch leaves copper slivers between traces, a rough surface along the channel that should be clear, and a residue that can cause a short after assembly. It is often local, appearing where the pattern is densest and the etchant exchange is poorest.

Overetch shows as traces that are narrower than the artwork, a sidewall that is heavily undercut and, in severe cases, traces with a triangular cross section that are easily damaged during handling. The measurement that catches it is trace width, read against the nominal value.

Etch Factor and Sidewall Shape

Etch factor is the ratio of the etch depth to the lateral undercut, and it describes how well the process holds the dimensions defined by the resist. A higher etch factor means less undercut for the same depth and therefore a trace closer to the artwork width.

The shape of the sidewall follows the factor. An ideal etch would leave vertical walls, while a real one leaves a slope, and the slope determines both the width at the top of the trace and the width at the base that carries the current. Both numbers matter, which is why the measurement position must be stated. In practice the width at the base of the trace is the one that carries current and the one a specification should control, while the top width is closer to what optical inspection sees.

Resist Adhesion and Undercut

The etchant attacks laterally beneath the resist wherever the resist is not bonded perfectly. Adhesion depends on surface preparation, lamination parameters and the state of the resist itself, and any weakness there increases undercut even when the bath is ideal.

Overhanging resist that has lost adhesion is the mechanism behind some underetch cases as well, because etchant trapped beneath a lifted edge keeps working after the panel leaves the chamber. Rinsing and neutralising immediately after etching limits that continuation, and the rinse must be effective inside the channels as well as on the surface. Spray pressure and nozzle condition both affect how well the rinse reaches into a narrow channel, and a rinse that is adequate on an open pattern can be inadequate in a dense one.

Bath Concentration and Temperature

Etch rate is controlled by the chemistry and the temperature together. A bath that is at the low end of its copper range and the high end of its temperature etches quickly and produces more undercut, while one that is cool and copper rich etches slowly and may not clear dense patterns.

Alkaline and cupric systems behave differently, so the numbers that matter differ by chemistry. Ammoniacal and cupric systems also differ in how they respond to dissolved copper, so a correction that suits one may push the other out of its window. What is common to both is that the analysis, the temperature and the etch rate have to be recorded together, and that a correction made to one of them changes the other two in effect. Bath control for these systems is covered under alkaline etch bath control.

Conveyor Speed and Dwell

In a conveyorised etcher the dwell time is set by the conveyor speed and the chamber length, and it is the parameter that operators change most often. Reducing the speed to clear an underetched area increases the exposure of every trace on the panel and pushes the process toward overetch.

For that reason, dwell should be adjusted only against measured trace width rather than against a visual impression of the channel. Where the solution exchange is the limiting factor, improving the spray pattern or the nozzle condition delivers a cleaner result than lengthening the dwell.

Compensation in the Artwork

Because etching always removes some copper sideways, artwork is normally compensated so that the finished trace matches the design. The compensation is a single figure applied at CAM, and it is only correct for the process that produced the first article.

A change to the chemistry, the resist type, the panel thickness or the conveyor speed invalidates that figure. Where compensation is not revisited after such a change, the result is a batch of traces that are consistently wider or narrower than the design intended.

Measurement and Sampling

Trace width should be measured on a coupon that runs with the panel, and on the product itself by optical measurement after etching. Coupon measurement is easier and more repeatable, while product measurement catches local effects that a coupon in a quiet area of the panel will not see.

Sampling should cover the densest and the sparsest areas, since etch rate varies with pattern density. A panel that measures correctly in the open area and underetches in a dense field is showing a solution exchange problem rather than a chemistry problem.

Effects on Impedance and Current

For controlled impedance designs the finished width matters as much as the thickness, because the characteristic impedance is a function of both. A trace that is narrower than the design raises the impedance and the loss, and the effect is visible in the signal performance of the finished board.

For power traces the consequence is thermal rather than electrical. A trace that is narrower than the design has a smaller cross section, so the temperature rise at a given current is greater, and the margin to the allowed temperature is reduced. Thickness targets from the plating thickness guide are calculated on the assumption that width is also held.

Control Plan and Records

The control plan should link three numbers: the analysis result, the conveyor speed and the measured trace width. When all three are recorded on the same sheet, a drift in width can be attributed to a cause within minutes rather than by trial and error.

The records should also note any change of resist or etchant source, because those changes alter the etch factor without changing any setting that the operator would notice. A batch that measures narrow after such a change is showing the compensation figure needs updating, not that the machine has gone wrong.

FAQ

What is the difference between underetch and overetch? Underetch leaves copper that should have been removed, while overetch removes copper that should have been protected. Underetch causes shorts and residue; overetch causes narrow traces and weak sidewalls.

What is a normal etch factor? Values in the region of 1.5 to 3 are typical for conveyorised etching, depending on the chemistry and the resist. A higher figure means less lateral undercut for the same depth.

How is trace width kept on target? By compensating the artwork for the etch factor of the actual process and by measuring finished width on coupons and product, then adjusting the bath and dwell against those measurements rather than by eye.

}