Contact Us
  • Home
  • BLOG
  • Why Does ENIG Turn Black After Reflow?

Why Does ENIG Turn Black After Reflow?

Black pad is corroded nickel. The gold is already gone into the solder, and what is left at the joint interface is a phosphorus rich nickel surface that was attacked in the immersion gold bath. The joint fails on that surface, not in the solder. If a pad turns dark only after reflow, the corrosion was there before reflow; the reflow just removed the gold that was covering it.

Two numbers carry most of the risk: nickel at 2.5 to 5 um, which is the 100 to 200 micro inch window we publish, and gold at 0.05 to 0.08 um, our 2 to 3 micro inch spec. The gold is not a barrier layer. It is a thin displacement layer that has to dissolve completely into the joint in the first seconds of the reflow. When the bath is too active it dissolves nickel faster than it deposits gold, and that is the defect in one sentence.

What happens inside the gold bath

Immersion gold is a displacement reaction. Nickel metal reduces gold ions and goes into solution. There is no reducer and no external current; the nickel is its own electron source. Raise the temperature, run the pH low, run gold concentration high or keep a strong complexing agent in the bath, and nickel keeps dissolving after the surface is visually covered. It goes at the grain boundaries first. The phosphorus that was alloyed in the nickel stays behind as a layer that can reach 10 to 14% P, nanometres to tens of nanometres thick.

That layer is the failure. Tin cannot build a normal intermetallic on it, so the joint ends up as solder sitting on a brittle Ni-Sn-P crust. The fracture is flat and dull with no ductile dimples, which is why the same joint passes a pull test on the bench and cracks in a drop or vibration test three months later.

What you seeWhat it isHow to confirm
Dark pad after reflow, often one corner of a panelhyper-corrosion of nickel under the goldtape pull on a virgin pad, EDS on the fracture
Dark ring at the pad edgethin gold, nickel exposedXRF map across the pad
Whole lot dull after storagetarnish or oxide, not black padwetting test, solderability per J-STD-002
Joint cracks at low load, flat surfaceNi-Sn-P brittle layerEDS shows Ni and P, no Sn-Cu

Why thicker gold makes it worse

Buyers ask for 0.15 to 0.2 um of gold as insurance. In an immersion bath that gold takes longer to deposit, so the panel spends longer in the chemistry that is attacking the nickel. The extra gold also does not leave the joint. On a 0.5 mm round pad with a 0.15 mm paste deposit, the joint metal is about 2.8e-4 g. Gold at 0.05 um is 0.09% of that weight, at 0.15 um it is 0.27%, at 0.20 um it is 0.36%. The embrittlement threshold commonly quoted for gold in tin based joints is about 0.3%. The published 2 to 3 micro inch window has margin; 8 micro inches has none.

So the two goals pull in opposite directions. Thick gold protects the nickel from the air and embrittles the joint. Thin gold keeps the joint ductile and gives the nickel less protection in storage. The window exists because both failures are real, and a supplier that quotes one number without the other is quoting half a finish.

Three things that get read wrong

Judging by colour. Black pad can be invisible while the gold is intact, and a dark pad can be nickel oxide from storage, or gold that is too thin, or a stain from a HASL pot. Colour is a hint. The verdict comes from removing the gold with tape and looking at the nickel underneath, then from a fracture surface under SEM.

Trying to rework it. Scraping the pad or reflowing the joint again exposes corroded nickel to air and flux and does not restore the metal. The pad has to be stripped and re-plated, which on a finished assembly means scrap. That is the practical reason to catch it at incoming inspection on the bare board and not after assembly.

A cross-section with no stress and no chemistry. A section proves thickness. It does not prove adhesion, and it does not prove the nickel was sound. IPC-4552 asks for phosphorus content on the nickel by EDXRF, and most customer specs add a wetting test plus joint strength after one to three reflow passes. A report that shows only a nickel thickness is a report about geometry.

What SUPERB PCB holds on an ENIG build

Nickel 100 to 200 micro inches and gold 2 to 3 micro inches, both verified by XRF per lot. P content measured on coupons to IPC-4552, bath make up and stabiliser logged per shift, gold drag out controlled, solderability by wetting test before packing. Line and space to 0.05 mm, 68 layers on the high layer line, impedance to plus or minus 5% with plus or minus 2% on critical nets, and a TDR coupon on every panel. The finish window and the alternatives are on the PCB capability page, the RF side is on RF PCB, and the choice above 10 GHz is in which surface finish to use above 10 GHz. SUPERB Tech reports P content with the lot, because the number that predicts black pad is not on the drawing.

What to send with the question

Both metal thicknesses from your finish spec, the number of reflow passes the assembly sees including rework, whether wire bonding follows the finish, the pad size and pitch, and whether the dark pads sit on one panel corner or across the whole lot. One corner points at current distribution and bath geometry. Across the lot points at chemistry and gold concentration. Those two answers lead to different corrective actions, and neither can be chosen from a photograph.

Need a quote or a review of your stack against these limits? pcba@superb-tech.com · WhatsApp +86 133 9608 1443