Estimated reading time: 4 minutes

Trouble in Your Tank: More ENIG Process Issues & Defects

Editor's Note: This column originally appeared in the August 2011 issue of The PCB Magazine.

Nickel Peeling from the Copper Surface

This, unfortunately, is an all-too-common problem related to any plating process (Figure 1).

Figure 1: Electroless nickel plating is initiated but loses adhesion to the copper.

First and foremost, the copper surface must be free of residues and have sufficient cleanliness to provide an optimum surface for adhesion. One critical recommendation is to ensure sufficient copper is removed during the micro-etch step prior to the activator (palladium catalyst). Whether potassium monopersulfate or sodium persulfate is utilized as the etch of choice, ensure that 50-80 microinches of copper is removed by the etch. In addition, rebatch the micro-etch when the copper concentration in the micro-etch process tank reaches 10 grams/liter. As the copper concentration increases in the micro-etch solution due to continued use, the etch rate will decrease. If one is not cognizant of this fact (with persulfate-based micro-etches), the amount of copper removal will be insufficient to ensure good plating adhesion.

However, process engineers must also guard against the galvanic effect. This situation arises when two dissimilar metals are in contact with one another (such as nickel and copper with selective ENIG processing, basically copper-nickel gold interface). The issue is exacerbated by the type of micro-etch used. In addition, internal testing has shown that feature size, such as large area gold pads directly connected to copper traces, leads to excessive etching of the copper at the interface. One may compare this to a battery cell being set up.

Figure 2 gives a close-up view of the galvanic etch on selective ENIG circuit boards.

Figure 2.:Galvanic reaction leading to excessive copper removal prior to the deposition of a selective finish--typically selective ENIG followed by OSP for the copper pads that are exposed.

There are a number of possible factors with respect to the chemistry and operational parameters of the micro-etch. An internal study looked at micro-etch basic compositions (sodium persulfate, hydrogen peroxide-sulfuric acid and potassium persulfate). The other factors were kept constant. The results of numerous controlled tests with the same design of test vehicle showed that sodium persulfate yielded the greatest amount of copper removal. The stabilized hydrogen peroxide gave minimal copper removal in accordance with the desired overall etch rate (Figure 3). 

Figure 3: Potassium persulfate with the galvanic effect on the left, with the same area of the test vehicle and no evidence of galvanic effect on the right, with stabilized hydrogen peroxide sulfuric acid etchant.

Immersion Gold Deposit Peeling from Nickel

It is obviously quite disconcerting to find that, after all of the care and controls the engineering staff puts in place to ensure a quality ENIG process, the gold deposit peels from the nickel surface (Figure 4).

Figure 4: Gold adhesion failure to the electroless nickel deposit.

There are several possible causes for the defect:  

• Contaminated nickel drag-out tank. Dump and remake with every new EN bath make-up, or when Ni > 1,000 ppm.
• Poor rinsing quality after Ni: Could be too short, too low of water flow or temperature of rinse too low in temperature.
• Too long of rinse after Ni: should be less than three minutes; may passivate Ni deposit. To save the product, strip the gold, reactivate the nickel and then process again through the immersion gold.

Soldermask Breakdown

There is no argument that the electroless nickel-immersion gold process is quite aggressive on liquid photoimageable soldermasks--even the best of them. The elevated operating temperatures and long dwell times in the electroless nickel plating solution are often blamed for causing soldermask breakdown (Figure 5).

Figure 5: An example of soldermask breakdown.

Ideally, the final product must maintain the color of the soldermask and provide straight sidewalls. There should be minimal to no undercut. Excellent adhesion of the mask to the surface is necessary to minimize or prevent additional breakdown. Additionally, it is worth mentioning that precleaning in the ENIG is also quite aggressive. It is clear the operating window is quite narrow to ensure excellent soldermask adhesion. Therefore one may find fault with the soldermask process itself, or in regards to the plating parameters. As an example, if the soldermask is undercured, underdeveloped and under-exposed, the mask is more vulnerable to breakdown during subsequent wet processing such as the ENIG process. On the contrary, over curing will embrittle the resist, thus making the film more susceptible to breakdown.

Figure 6 shows a close-up view of the mask lifting or peeling near the interface of the ENIG deposit and the soldermask. It is important to note that the applied mask must be of sufficient thickness to withstand the aggressive processing steps during ENIG processing. Monitor the wet thickness of the soldermask as well as wet weights of the mask prior to exposure to ensure sufficient mask thickness. Certainly, final dry and cured mask thickness will depend on additional factors such as mask thixotropic properties, rheology and solvent content. These factors will be discussed in a future Trouble in Your Tank column.  

Figure 6: Close-up view of the mask lifting or peeling near the interface of the ENIG deposit and the soldermask.

Michael Carano is with OMG Electronic Chemicals (formerly Electrochemicals), a developer and provider of processes and materials for the electronics industry supply chain. He has been involved in the PWB, general metal finishing photovoltaic industries for over 29 years. Carano holds nine U.S. patents in topics including plating, metallization processes and PWB fabrication techniques.