ENIG Plating Process: How Electroless Nickel Immersion Gold Works and Why Black Pad Kills Boards

# ENIG Plating Process: How Electroless Nickel Immersion Gold Works and Why Black Pad Kills Boards

Last March, a customer sent us 500 panels of a 12-layer server board — 0.8mm pitch BGA, 3 mil trace/space, controlled impedance on every signal layer. The boards came back from the fab with ENIG finish, passed incoming visual inspection, and went straight into our SMT line. First-pass yield: 61%. The defect map told the story — every single open was on a BGA pad. We pulled three boards, cross-sectioned the BGA sites, and found it: the nickel layer under the gold had corroded into a black, porous mess. The immersion gold had deposited 3 microinches of Au, right at the low end of the spec, and the nickel underneath was a high-phosphorus deposit that the gold bath had attacked during the immersion step. The solder joints looked fine optically. Under X-ray, they looked fine. But the nickel-gold interface had turned into a brittle, high-resistance barrier that cracked the moment the BGA cooled and pulled. Total loss: $21,000 in boards, plus another $8,400 in assembly labor and components. The root cause? The fab had run the immersion gold bath 15% past its metal turnover limit, and the nickel phosphorus content was at 11.2% — right at the edge where corrosion rate spikes.

That failure is not unusual. ENIG is the most popular high-reliability surface finish in the industry — and also the one most likely to fail catastrophically if the chemistry drifts even slightly out of spec. Understanding how the process works is not academic. It is the difference between a board that survives 10,000 thermal cycles and a board that fails at reflow.

How ENIG Actually Works: The Chemistry Step by Step

ENIG is not a single plating step. It is a sequence of chemical processes that happen in a single production line: electroless nickel, then immersion gold. The "electroless" part is critical — there are no electrodes, no external current. The metal deposits via autocatalytic chemical reduction. If you are coming from our PCB surface finish guide where we compared ENIG against HASL and OSP at a high level, this section goes deeper into what actually happens inside the plating tank.

The process starts after copper patterning and solder mask application. The exposed copper pads — and only the exposed pads — need to be coated. Here is the sequence:

Step 1: Pre-clean and microetch. The copper surface gets cleaned with an acid cleaner, then microetched with a persulfate or peroxide-sulfuric solution to remove 20-40 microinches of copper. This creates a fresh, active surface with the right topography for nickel adhesion. Skip the microetch or under-etch, and the nickel will delaminate in reflow. Over-etch, and you lose pad area and create undercut under the solder mask.

Step 2: Activation (catalyst). The copper surface is activated with a palladium-based catalyst. This is the seed layer that makes electroless nickel possible — the nickel reduction reaction needs a catalytic surface to start. On bare copper, the reaction will not initiate without this catalyst. The palladium deposits as a discontinuous monolayer, typically 0.01-0.05 microinches thick. It is not a structural layer; it is purely catalytic.

Step 3: Electroless nickel deposition. This is the workhorse layer. The board goes into a nickel bath at 80-90°C containing nickel sulfate, sodium hypophosphite (the reducing agent), and complexing agents. The hypophosphite reduces nickel ions to metallic nickel on the catalytic surface, and the nickel itself becomes autocatalytic — once the reaction starts, it keeps going. The deposition rate is typically 15-25 microinches per minute, and the target thickness is 118-236 microinches (3-6 μm) per IPC-4552A. The phosphorus content — a byproduct of the hypophosphite reduction — ends up at 7-11% by weight depending on bath chemistry.

Step 4: Immersion gold deposition. The nickel-plated board goes into a gold bath at 70-85°C containing potassium gold cyanide. This is a displacement reaction, not an autocatalytic one. Gold ions in solution have a higher reduction potential than nickel, so gold deposits on the nickel surface while nickel dissolves into the bath. The reaction is self-limiting: once the nickel surface is fully covered with gold, there is no more nickel exposed to drive the displacement, and the deposition stops. Target thickness: 1.18-2.76 microinches (30-70 nm) per IPC-4552A.

That self-limiting behavior is both ENIG's strength and its danger. It means the gold layer is always thin and uniform — no gold dendrites, no thick deposits bridging fine-pitch pads. But it also means the gold thickness is a direct indicator of how much nickel was dissolved during the immersion step. Too much gold, and you dissolved too much nickel. Too little gold, and the nickel is exposed to oxidation. The window is narrow, and the consequences of missing it are severe.

The Nickel Layer: Phosphorus Content and Why It Matters

The phosphorus content of the electroless nickel layer is not a secondary parameter — it is the single most important variable in ENIG reliability. And it is the one most designers never specify on their fab drawings.

Electroless nickel is not pure nickel. It is a nickel-phosphorus alloy, and the phosphorus content determines the crystal structure, corrosion resistance, hardness, and — critically — how the nickel behaves during the immersion gold step.

Parameter	Low-P Ni (4-6% P)	Mid-P Ni (7-9% P)	High-P Ni (10-12% P)
Crystal Structure	Crystalline	Mixed crystalline/amorphous	Amorphous
Hardness (as-plated)	500-600 HK100	450-550 HK100	400-500 HK100
Internal Stress	Tensile	Near-zero	Compressive
Corrosion Resistance	Moderate	Good	Excellent
Solder Wetting Time (SAC305)	0.8-1.2 sec	1.0-1.5 sec	1.5-2.5 sec
Black Pad Risk	Low	Moderate	High
Typical PCB Use	Rare	Standard ENIG	Avoid for BGA
Bath Stability	Sensitive	Robust	Very robust

The practical implication: mid-phosphorus nickel (7-9% P) is the sweet spot for ENIG on PCBs. It gives you enough phosphorus for partial amorphous structure (which resists through-pore corrosion) without so much phosphorus that the immersion gold step aggressively attacks the grain boundaries. High-phosphorus nickel is great for oil field valves and chemical processing equipment — and terrible for BGA pads, because the immersion gold reaction preferentially attacks the phosphorus-rich grain boundaries, creating the porous, black corroded surface we call "black pad."

Low-phosphorus nickel is rarely used in PCBs because the crystalline structure has more grain boundaries, which means more pathways for corrosion through the nickel to the copper underneath. It solders well, but the long-term corrosion resistance is worse. You sometimes see it on connector contact surfaces where hardness matters more than corrosion resistance, but not on general-purpose PCB pads.

Here is the part most datasheets do not tell you: the phosphorus content is not fixed. It drifts as the bath ages. A fresh mid-P bath starts at around 8% phosphorus. As the bath accumulates turnover cycles (one turnover = the volume of nickel added equals the initial bath volume), the phosphorus content creeps upward. By 4-5 metal turnovers, a mid-P bath can be depositing 10-11% phosphorus nickel — effectively high-P nickel from a bath that was spec'd as mid-P. This is exactly what happened in our server board failure. The fab was running the bath past its recommended life to save cost. The nickel deposit looked fine under visual inspection. But the chemistry had shifted into the danger zone, and the immersion gold step attacked it.

The Gold Layer: How Thin Is Too Thin?

The immersion gold layer in ENIG is not a solderable surface. It is a protective layer. Its job is to prevent the nickel from oxidizing before assembly. During reflow, the gold dissolves into the solder (at roughly 0.1-0.2 μm/sec in Sn-Ag-Cu at 245°C), and the solder bonds directly to the nickel. The gold is consumed. The joint is a Sn-Ni intermetallic, not a Sn-Au joint. This is fundamentally different from solder alloy selection where the bulk solder composition drives reliability — here, the surface finish chemistry determines whether the joint can even form.

This means the gold layer has two failure modes, and they pull in opposite directions:

Too thin (< 1.18 μin / 30 nm): The gold does not fully cover the nickel. Nickel oxide forms at exposed spots, and solder wetting fails at those points. You get dewetting, partial wetting, or complete non-wetting on random pads. This is especially common on large pads where the gold thickness tends to be thinner at the center than at the edges — the immersion reaction starts at the pad edges and works inward, and on pads larger than about 2mm diameter, the center can be 30-40% thinner than the perimeter.

Too thick (> 2.76 μin / 70 nm): The gold layer is thick because the immersion reaction ran long — which means more nickel dissolved. The nickel surface under a thick gold layer is more likely to be corroded. Also, thick gold means more gold dissolved into the solder during reflow, which increases the gold content of the solder joint. Above 3-4 wt% gold in Sn-Ag-Cu solder, the joint becomes brittle due to AuSn₄ intermetallic formation. For a 0.8mm pitch BGA with 0.4mm pads, the gold volume from a 3 μin layer is enough to push a small solder joint past the embrittlement threshold. We measured this directly on a customer's 0.5mm pitch CSP — the gold contribution from a 3.2 μin ENIG layer pushed the joint to 4.8 wt% Au, and drop test failures jumped from 0.2% to 11%.

The IPC-4552A specification calls for 1.18-2.76 microinches (30-70 nm) of gold. That range is tight for a reason. Our recommendation: target 1.5-2.0 μin for most applications. That gives you full nickel coverage without excessive nickel dissolution. If your fab cannot hold that window consistently, find a different fab — this is not a parameter where you want to accept wide tolerance.

Black Pad: The Failure Mode That Made ENIG Infamous

Black pad is not a single defect. It is a class of defects at the nickel-gold interface that cause solder joint failure. The name comes from the appearance of the nickel surface after the gold is stripped — it looks black, rough, and pitted instead of smooth and silver.

The mechanism works like this:

1. During the immersion gold step, the displacement reaction dissolves nickel preferentially at grain boundaries and phosphorus-rich regions. 2. If the nickel has high phosphorus content (>10%), or if the gold bath is aggressive (high temperature, high turnover, low stabilizer), the dissolution is excessive. 3. The dissolved nickel leaves behind a porous, phosphorus-rich surface layer that gold cannot properly cover. 4. The gold deposits over this rough surface, but the coverage is incomplete — there are micro-pores through the gold to the corroded nickel. 5. During storage, air reaches the nickel through these pores and forms nickel oxide. 6. During reflow, the solder cannot wet through the oxide and the corroded nickel structure. The joint either does not form or forms with weak adhesion. 7. Under thermal cycling or mechanical stress, the joint cracks at the nickel-solder interface.

The insidious part: the joint can pass visual inspection, X-ray inspection, and even functional test at room temperature. It fails under thermal cycling, vibration, or drop shock — exactly the conditions that matter in the field. We have seen boards pass 100% ICT at 25°C and then fail 40% at the first thermal cycle from 0°C to 70°C. (Ask me how I know — I spent three days on that root cause analysis before we thought to cross-section the BGA pads.)

Black pad is not detectable by standard incoming inspection. You need cross-sectioning with SEM/EDX analysis to see it. The key indicator is a "mud crack" pattern in the nickel surface under SEM, and elevated phosphorus content at the surface (typically >12% at the interface vs. 7-9% in the bulk nickel). Per the IPC-4552A standard, acceptable nickel-gold interfaces should show no visible corrosion or mud-cracking patterns at 1000x magnification.

Preventing black pad is primarily a fab-side problem, but designers can mitigate it significantly through their fab notes and qualification requirements:

- Specify mid-phosphorus nickel (7-9% P) in your fab notes

• Require gold thickness in the 1.5-2.0 μin range (not just "per IPC-4552A")
• Add a note: "Bath metal turnover shall not exceed 3x for nickel and 2x for gold"
• For Class 3 boards, require cross-section verification of nickel-gold interface on the first article
• If your board has fine-pitch BGAs (≤0.5mm), seriously consider ENEPIG instead

ENEPIG: When You Need the Palladium Safety Net

ENEPIG (Electroless Nickel Electroless Palladium Immersion Gold) adds a palladium layer between the nickel and the gold. The palladium acts as a barrier that prevents the immersion gold reaction from attacking the nickel. It is the fix for black pad — but it comes at a cost that you need to evaluate against your board value.

Parameter	ENIG	ENEPIG	Electroplated Hard Gold
Nickel Thickness	3-6 μm	3-6 μm	3-6 μm
Palladium Thickness	N/A	0.05-0.15 μm	N/A
Gold Thickness	0.03-0.07 μm	0.025-0.05 μm	0.6-1.3 μm
Black Pad Risk	Moderate to High	Very Low	N/A (no immersion step)
Wire Bondability (Au)	Unreliable	Yes	Yes
Wire Bondability (Al)	No	Yes	No (Kirkendall voids)
Contact Resistance	0.1-0.3 mΩ	0.05-0.2 mΩ	0.05-0.15 mΩ
Cost Premium (vs ENIG)	Baseline	+15-25%	+30-50%
Shelf Life	12 months	24+ months	24+ months
Max Reflow Cycles	2-3	3-5	2-3
IPC Specification	IPC-4552A	IPC-4556	IPC-4553
Typical Use	General SMT	Fine-pitch BGA, wire bond	Edge connectors, keypads

The palladium layer in ENEPIG is deposited electrolessly (not by immersion), which means it builds up by autocatalytic reduction — the same mechanism as the nickel layer. The reaction does not dissolve nickel. The gold then deposits by immersion on the palladium, and since palladium is more noble than nickel, the displacement reaction is much milder. Less nickel dissolution, no grain boundary attack, no black pad.

The practical tradeoff: ENEPIG costs 15-25% more than ENIG on a typical board. For a $5 board, that is $0.75-1.25 extra. For a $50 board, it is $7.50-12.50. On a high-value board with BGAs — especially fine-pitch BGAs where the solder volume per joint is small and the gold concentration in the joint matters more — that premium is cheap insurance against a $21,000 field failure. On a $5 IoT board with 0805 passives and a QFP, it is unnecessary cost.

ENEPIG also enables gold and aluminum wire bonding directly on the pads, which ENIG cannot do reliably. The gold is too thin for thermosonic gold wire bonding, and the nickel under aluminum wire bonds creates Kirkendall voids at the Ni-Al interface. If your board needs wire bonding — and many RF modules and power devices do — ENEPIG or thick electroplated gold is your only realistic option.

For most standard SMT assemblies without fine-pitch BGAs or wire bonding requirements, ENIG with properly controlled chemistry is sufficient. ENEPIG is overkill for a 4-layer IoT board with 0.5mm pitch QFPs and 0805 passives. But for a 14-layer telecom board with 0.4mm pitch BGAs, ENEPIG is the right call. The decision framework is simple: if the cost of a single field failure exceeds the total ENEPIG premium on the entire production run, use ENEPIG.

Common Mistakes in Specifying ENIG

1. Specifying "ENIG per IPC-4552A" without thickness targets. IPC-4552A gives ranges, not targets. If you just say "per IPC-4552A," the fab can run gold at the low end of the range to save cost, and you get marginal nickel coverage. Always specify your target: "ENIG, Ni 3-6 μm, Au 1.5-2.0 μin per IPC-4552A." The cost of specifying the target is zero. The cost of not specifying it can be a field failure campaign that wipes out your margin on the entire project.

2. Not specifying phosphorus content. If your fab notes do not mention phosphorus, the fab will run whatever bath chemistry they have. If that bath is near end-of-life, you get high-P nickel and black pad risk. Add "Electroless nickel phosphorus content: 7-9 wt%" to your fab notes. Some fabs will push back — that is a signal to find a different fab for Class 3 boards.

3. Using ENIG for boards that need wire bonding. We have seen this twice in the past year: a customer designed an RF module with gold wire bonds to the PCB, specified ENIG, and then wondered why wire pull strength was 2-3 grams instead of the expected 6-8 grams. ENIG gold is too thin for thermosonic gold wire bonding. You need either ENEPIG or electroplated soft gold (25-50 μin) for wire bonding. The ENIG gold dissolves into the bond before a reliable intermetallic can form.

4. Assuming thicker gold is better. More gold means the immersion reaction ran longer, which means more nickel dissolution. A gold thickness of 4-5 μin is a red flag, not a quality indicator. It means your nickel surface is likely corroded. If your fab reports gold thickness above 3 μin on an ENIG board, ask for a cross-section of the nickel-gold interface before you run assembly. We once had a customer who was happy their fab was delivering "thick gold" at 4.5 μin — right up until the BGA rework yield dropped to 30% because the nickel underneath was destroyed.

5. Not qualifying the fab's ENIG process. Different fabs run different chemistry, different bath management protocols, and different quality control. We have seen gold thickness vary by 40% between two fabs using the "same" ENIG spec on the same board design. Before committing a high-value board to a new fab, always order a first article with cross-section verification of the nickel-gold interface. The cost is $200-400 for the analysis. The cost of skipping it can be the entire assembly batch plus components.

ENIG Process Control Checklist

Before you send your next ENIG board to fab, run through this list:

1. Specify nickel thickness as 3-6 μm and gold thickness as 1.5-2.0 μin in your fab notes — do not rely on "per IPC-4552A" alone.

2. Add phosphorus content requirement: 7-9 wt% for the electroless nickel layer. This is the single most effective black pad prevention measure you can specify.

3. Add bath management note: "Nickel bath metal turnover ≤3x, gold bath metal turnover ≤2x." This limits bath aging, which limits phosphorus drift.

4. For boards with BGA pitch ≤0.5mm, specify ENEPIG instead of ENIG. The palladium barrier eliminates black pad risk on the joints where it matters most.

5. Require first-article cross-section with SEM/EDX for any new fab or any Class 3 board. Verify nickel-gold interface integrity, nickel thickness, gold thickness, and phosphorus content.

6. If your board needs wire bonding, use ENEPIG or electroplated soft gold. ENIG gold is too thin for reliable wire bonds of any type.

7. Flag any gold thickness report above 3 μin as a quality alert. Request cross-section verification of the nickel-gold interface before proceeding to assembly.

8. Set maximum shelf life at 12 months for ENIG boards. After 12 months, nickel oxidation through micro-pores in the gold can degrade solderability. ENEPIG extends this to 24+ months.

FAQ

Q: What is the minimum gold thickness for ENIG that still provides reliable solderability?

Per IPC-4552A, the minimum gold thickness is 1.18 microinches (30 nm). However, in practice, we recommend a minimum of 1.5 microinches because 1.18 μin provides marginal nickel coverage on large pads where the immersion reaction is self-limiting. Below 1.5 μin, you risk nickel oxide formation at exposed spots, causing dewetting during reflow.

Q: How can I detect black pad without destructive cross-sectioning?

You cannot reliably detect black pad non-destructively. Visual inspection, X-ray, and ICT all pass boards with black pad. The only reliable method is cross-sectioning with SEM/EDX analysis, which costs $200-400 per sample. Some fabs offer "gold strip and visual inspection" as a quick check — stripping the gold and looking for dark patches on the nickel — but this is qualitative and misses mild cases.

Q: ENIG vs ENEPIG: which should I choose for a board with 0.4mm pitch BGA?

For 0.4mm pitch BGAs, ENEPIG is the safer choice. The solder volume per joint is small (approximately 0.15-0.20 mm³ for a 0.3mm pad), so any gold embrittlement or nickel corrosion has a proportionally larger effect. ENEPIG's palladium barrier eliminates the black pad mechanism entirely. The 15-25% cost premium on the board finish is negligible compared to the risk of field failures on fine-pitch BGA joints.

Q: What causes ENIG gold thickness to vary across the same board?

Immersion gold deposits from the edges of each pad inward, so large pads tend to have thinner gold at the center. Also, bath agitation and board orientation in the plating tank create thickness gradients — pads near the top of the rack can be 10-20% thinner than those near the bottom. IPC-4552A allows ±0.4 μin variation across the board. If your fab reports uniformity worse than this, their bath control needs attention.

Q: Can I use ENIG finish for aluminum wire bonding?

No. ENIG is not suitable for aluminum wire bonding. The thin immersion gold layer (1.5-2.0 μin) does not provide sufficient barrier or bonding surface for Al wire. Aluminum wire bonds to ENIG form Kirkendall voids at the nickel-aluminum interface due to interdiffusion, reducing pull strength to 2-3 grams (vs. 6-8 grams required per MIL-STD-883). Use ENEPIG with 0.05-0.15 μm palladium, or electroplated nickel/palladium/gold for Al wire bonding applications per IPC-4556.

Q: How many reflow cycles can ENIG withstand before solderability degrades?

ENIG can typically withstand 2-3 reflow cycles before solderability degrades noticeably. Each reflow cycle dissolves gold and exposes fresh nickel, which then re-oxidizes during cooling. After 3 cycles, the nickel surface may have sufficient oxide to cause wetting defects on subsequent assembly passes. ENEPIG extends this to 3-5 cycles because the palladium barrier prevents nickel oxidation. If your assembly process requires more than 3 reflow passes, specify ENEPIG.

Q: What is the cost difference between ENIG and ENEPIG for a typical PCB?

ENEPIG typically costs 15-25% more than ENIG for the same board. For example, a $10 board with ENIG would cost approximately $11.50-12.50 with ENEPIG. The premium comes from the additional palladium bath step and the palladium material itself (palladium traded at approximately $1,000-1,100 per troy ounce as of early 2026). For high-value boards with fine-pitch BGAs, this premium is justified by the elimination of black pad risk.

---

Need expert consultation?

Request a Free Quote

FAQ

Parameter	ENIG	HASL (Lead-Free)	OSP	Immersion Tin
Typical Thickness	3-5 µin Au / 120-240 µin Ni	40-400 µin	8-20 µin	20-40 µin
Shelf Life	12+ months	12+ months	6-12 months	6-12 months
Surface Flatness	Flat (Coplanar)	Uneven	Flat (Coplanar)	Flat (Coplanar)
Wire Bondable	Yes	No	No	No
Black Pad Risk	High	None	None	None
Max Reflow Cycles	3-4	2-3	1-2	3-4

Q: What causes black pad in ENIG finish?

Black pad is caused by corrosion of the nickel layer under the gold during the immersion gold step, often triggered by excessive phosphorus content in the nickel deposit (typically above 10%) or running the gold bath past its metal turnover limit, which can spike the corrosion rate and create a brittle, high-resistance interface.

Q: How thick is the gold layer in ENIG plating?

The immersion gold layer in an ENIG finish is typically between 3 to 5 microinches (0.07 to 0.12 microns), which is just enough to protect the underlying nickel from oxidation while remaining thin enough to dissolve into the solder joint during reflow.

Q: What is the difference between electroless nickel and immersion gold?

Electroless nickel is deposited via autocatalytic chemical reduction to a thickness of 120 to 240 microinches, while immersion gold is deposited through a displacement reaction where gold ions replace nickel atoms on the surface, stopping at a much thinner 3 to 5 microinches once the nickel is fully covered.

Q: How much copper is removed during ENIG microetch?

The microetch step in the ENIG process removes 20 to 40 microinches of copper to create a fresh, active surface with the proper topography for nickel adhesion; skipping this or under-etching can cause nickel delamination during reflow.

Q: Why do BGA solder joints fail with black pad?

BGA solder joints fail because the corroded, porous nickel layer under the gold creates a brittle, high-resistance barrier that cracks under thermal or mechanical stress; this can drop first-pass assembly yields to as low as 61% even when joints pass visual and X-ray inspection.

Q: What phosphorus content is recommended for electroless nickel in ENIG?

For ENIG plating, a mid-phosphorus nickel deposit (around 7% to 9% phosphorus) is generally recommended to balance corrosion resistance and solderability, whereas high-phosphorus deposits reaching 11.2% or more sit at the edge where the corrosion rate spikes and black pad risk increases dramatically.