Anodic Bleed Under Conformal Coating: Causes and Controls
Anodic bleed is the dark, sometimes glassy track that appears along a conductor on a coated assembly after the board has spent time under bias in humid air. It is a corrosion product rather than a coating defect, and it belongs to the same family as the adhesion failures described in the notes on coating adhesion testing. The defect is usually found during qualification, because an assembly that passes a dry functional test can still bleed after a few hundred hours of damp heat.
The mechanism needs three things at once: a voltage difference between neighbouring conductors, a thin film of water on the surface, and ions dissolved in that film. Remove any one of the three and the track does not form. That is why the countermeasures divide into cleaning, coating and layout rather than into a single fix that can be applied at the end of the line.
What Anodic Bleed Looks Like
The track starts at the positively biased conductor, because the anode is where metal dissolves. It spreads towards the cathodic conductor as a discoloured band, and under a microscope the deposit is a mixture of metal salts and corrosion products rather than a clean metallic dendrite. In many cases the coating is intact over the top of the track, which is what makes the defect confusing at first inspection.
A true dendrite from electrochemical migration is metallic and grows from the cathode towards the anode, while bleed follows the field in the opposite direction. Both can be present on the same coupon, and both are evidence that the surface carried a continuous water film. The distinction matters for the corrective action, because a metallic dendrite and a corrosion track can come from different residue chemistries.
The Electrochemistry Behind It
Under bias the anodic conductor oxidises and the metal ions migrate through the water film in the direction of the field. The ions then hydrolyse and precipitate as oxides and hydroxides, which is why the deposit is dark and bulky instead of bright and metallic. Chlorides and weak organic acids accelerate the process because they raise the conductivity of the film and complex the metal ions so that more of them stay in solution.
Temperature and humidity set the rate. A qualification condition of 85 °C and 85 percent relative humidity with 10 to 50 V of bias produces visible bleed on a poorly cleaned assembly within a few hundred hours, while a correctly processed assembly survives the same test with no measurable change in leakage current. The water film does not need to be thick, and a monolayer is enough to carry the current.
Ionic Residue as the Fuel
The residue that matters is the ionic fraction, not the total mass of flux left on the board. A board can look clean and still carry enough chloride to start the reaction, and a board with a visible film of benign rosin can survive if that film is non-ionic. Extraction testing rather than visual inspection is therefore what decides whether a surface is suitable for coating.
Typical cleanliness specifications limit extracted contamination to about 1.5 ug of sodium chloride equivalent per square centimetre, and chloride-specific limits are tighter. Where the residue is measured with a solvent extract method such as the one described in the notes on solvent extract testing, the result has to be taken before coating, because the coating locks the residue in place and the failure then only appears under bias in the field.
Bias Voltage and Conductor Geometry
A higher bias increases the driving field, but the geometry of the conductors matters just as much. Two tracks separated by 0.2 mm with 50 V between them see a far higher field than the same tracks with 5 V, and the thin water film concentrates the field at the edges of the copper. Sharp corners, exposed copper at a via and bare test points are therefore favoured sites for the first sign of the track.
Where the design allows it, wider spacing between a high voltage net and a sensitive one reduces the risk more cheaply than any process change. Guard traces tied to a defined potential, and slots that break the surface path between two nets, are the layout answers to the same problem. The condition of the exposed copper, including the finish and its oxidation, is discussed in the notes on surface finish selection.
Coating Adhesion and the Wet Path
A conformal coating stops anodic bleed by keeping the water film away from the surface, and it can only do that where it adheres. Any region where the coating lifts, bubbles or creeps back from an edge provides a channel, and the channel concentrates the current into a narrow path. Adhesion failures usually start where the surface was not clean or where the coating was applied over a sharp edge with poor coverage.
Edge coverage is the weakest point of most coating processes. The meniscus at a conductor edge thins the coating, and a sharp corner can leave the tip of the copper exposed. A thickness of 25 to 75 um over the flat areas of a board says nothing about the coverage at the edges, and the edges are where the failure begins.
Cleaning Before Coating
Cleaning is the step that decides whether the coating has anything to bond to. Flux residue, fingerprints, mould release and silicone contamination all reduce adhesion, and the ionic part of the residue supplies the electrolyte as well. The cleaning process has to be verified with a measurement, because the difference between a clean board and a marginally dirty one is not visible under any practical level of magnification.
Flux residue is not uniform, and neither is its effect on the coating. A no-clean assembly that will be coated has to have that residue removed anyway, and the interaction between the residue, the cleaning chemistry and the adhesion of the finished film is set out in the notes on flux residue and coating adhesion. Tightening the coating step without fixing the cleaning step moves the failure rather than removing it.
Cure, Coverage and Thickness
Cure converts the liquid coating into a barrier, and an under-cured film stays permeable and can retain solvent that later forms bubbles. Typical schedules for acrylic and polyurethane coatings run from 30 to 60 minutes at 90 to 110 °C, or shorter at a higher temperature where the components allow it. Solvent that is not driven off becomes a void in the film and a path for water.
Coverage has to be checked at the places the spray or the needle cannot easily reach: under a tall component, beside a connector body, and along the shadowed side of a shield. UV tracer inspection is the practical way to see the film on the line, and it belongs in the first-article check rather than in an occasional audit. A coating that fails at the first article will fail in the field in the same place.
Detection: SIR Testing and Visual Signs
Surface insulation resistance testing is the standard method because it measures the property the defect destroys. A comb pattern coupon is cleaned and coated exactly as the product, then held at 85 °C and 85 percent relative humidity with a bias of 10 to 100 V DC while the resistance is logged. A common acceptance threshold is 100 megohms or more, together with a limit on the fall from the initial reading.
Visual inspection comes second and is best done on the coated board under magnification with the light at a low angle. A track that follows a conductor edge, or that has a colour different from the surrounding surface, is the signature. Confirmation by cross section or by elemental analysis of the deposit is described in the notes on failure analysis methods.
Process Controls and Records
The controls that keep anodic bleed out of the field are ordinary and mostly already present: a verified cleaning stage with a recorded extract result, a coating thickness and coverage check at first article, a cure profile, and a bias test on the finished product. None of them is exotic, and every one of them has to be recorded to be useful.
The records should tie each number to a lot, and they should include the coating batch and the cure profile. When a bleed failure appears in the field, the question is almost always which of the three variables moved: cleanliness, coverage or cure. A record that answers that question in an hour is worth more than a tighter specification that nobody monitors.
FAQ
Can anodic bleed happen without contamination? It needs a water film and dissolved ions, so a clean surface with a continuous coating does not bleed even under bias. In practice the failure requires a residue or an adhesion defect, and the search should start there.
Does a thicker coating always help? No. Coverage at edges and adhesion matter more than average thickness, and a thick film over a dirty surface simply hides the residue that drives the reaction.
Why does the failure appear only after hundreds of hours? The water film has to form, the ions have to accumulate at the field edge, and the corrosion product has to build to a visible size. All three are slow at ordinary ambient humidity.
