Solderable Final Finishes: Urban Legends and State of the Art

Estimated reading time: 17 minutes

Editor's Note: This article originally appeared in the May 2012 issue of The PCB Magazine.

Introduction

At one time there was an urban legend that led people to believe that alligators lived in the New York City sewer system. Urban legends also abound with regard to solderable surface finishes, leading well-meaning engineers to make ill-informed decisions as to which surface finish should be used. Some of these legends or misconceptions are:

• OSP (organic solderability preservative) technology is only for consumer-based PCBs;
• ENEPIG (electroless nickel electroless palladium immersion gold) solves all the problems associated with final finishes and complex assembly;
• ENIG (electroless nickel-immersion gold) is prone to black pad; and
• Nothing solders like solder (from that standpoint, true; however…).

Others could be added to this list. Regardless, this article presents a review of the latest in solderable finishes and includes special considerations when using these different finishes.

Board Final Surface Finish

The capability of final assembly determines the surface left on the component-mounting lands of the fabricated board. Today, a number of final finish processes have emerged, driven primarily by the fine pitch of the package to be mounted, and the capability to withstand higher-temperature lead-free assembly. ENEPIG has gained some play into the market, especially for IC packaging substrates. Additional attention has led some people to label ENEPIG as a “universal final finish” capable of soldering and wire bonding. Yet there is also a significant market niche for selective ENIG in combination with OSP. There are other emerging finishes that warrant discussion including direct electroless palladium over copper.

The selection of the solderable surface finish is defined by the needs of the assembly operation and in many respects the environment in which the assembled board is used. And lead-free assembly, with its higher temperatures, places considerable stress on the finish to perform over multiple reflows and through-hole assembly. Additionally, IC packaging requires several procedures that could affect the integrity of the surface finish, including plasma cleaning, wire bonding, mold underfill and cure, and lead-free assembly.

To complicate the situation further, components are becoming available only (or predominantly) in lead-free versions, necessitating entire lead-free assembly.

When selecting a surface finish on printed boards, several factors should be considered:

• Cost to produce the boards, including the overall final finish cost;
• Assembly yield with the finish--factors are planarity, wettability by solder, number of heat excursions allowed before soldering problems develop;
• Shelf life of the coating--especially with newer, no-clean assembly fluxes;
• Fine pitch requirements;
• Compatibility with new lead-free regulations;
• Environment where assembled PCB is used;
• Overall end-user reliability requirements;
• Shock/drop issues;
• Corrosion concerns;
• Wire bonding requirements;
• Process latitude for soldering, touch contacts, wire bonding, flip chip attach, etc.; and 
• Ability to rework if necessary.

Since most of the multilayer PCBs will have surface mount components on both sides, it is critical that the surface finish is able to maintain solderability through the various thermal excursions. Furthermore, it is important to understand that lead-free assembly temperatures have the ability to degrade solderable finishes at a much faster rate than in conventional tin-lead assembly. In light of these issues, this article presents the key attributes of each of the main finishes as well as identifies possible issues with which the bare board fabricator and assembly engineer must be concerned. As a point of reference, only finishes that are used as part of the SMOBC (solder mask over bare copper) compatible process are detailed.

Why Are Final Finishes Necessary?

The value of a surface finish is stated simply this way: It preserves the solderability and in turn the reliability of the component-to-board connection. Over time, unprotected bare copper interacts with oxygen. And oxygen is the enemy of good solderability. Thus, a final finish that is robust and applied to the board with excellent process controls will enhance solderability, and ultimately the performance under various assembly conditions. It is true that the final finish chosen will have an impact on final assembly yields. And it should be stated at the outset that no one surface finish meets all of the criteria listed above. Multiple surface finishes continue to be in use today and it is expected that this will continue for the foreseeable future. Prior to making specific recommendations, this paper examines pertinent information on the specific final finishes. It is important to understand that each solderable surface finish presents its own set of tradeoffs for both the PCB fab engineer and the assembly manufacturing engineer.

Lead-Free Hot Air Leveling

It was only six or seven years ago that hot air leveling (both lead and lead-free) commanded about a 40% share of the global PCB market for surface finishes. Today, both lead and lead-free HASL combined account for less than 25%, as shown in Figure 1. This study by Prismark1 showed that lead-free hot air leveling has grown, while RoHS regulations have lead to a decline in tin-lead based hot air leveling. Regardless, this is a process that remains in use today for some market segments. In particular, lead-free hot air leveling has emerged to meet the requirements of RoHS. Why is hot air leveling still popular today? It is well known that the shelf life of boards finished with HASL (and lead-free HASL) is quite long, unless the copper-tin intermetallic has consumed too much of the free tin in the coating. Thus, the old adage that “Nothing solders like solder,” applies. Now, thanks to a few firms embarking on focused product development programs, lead-free hot air leveling formulations have found a place in the industry. Nihon Superior, to name one, has introduced a lead-free alloy based on tin-nickel-copper2.    

Figure 1: Comparison of surface final finishes used, 2006 and 2011. (Source: Prismark)

Immersion Tin (ImSn)

Immersion tin has gained market share, particularly in Europe. Initial concerns with ImSn are that since the process is a galvanic plating reaction, the tin thickness is self-limiting. This of course can lead to limited shelf life due to the growth of the copper-tin intermetallic. Immersion tin does, however, provide a coplanar surface required for fine pitch assembly.  And one should not be too concerned about tin whiskers, as immersion tin deposits are too thin for this to occur. Yet there are a few drawbacks with the majority of the formulations.

One issue as described above is the self-limiting thickness of the immersion tin deposit. The copper-tin intermetallic will start to grow almost immediately after plating. In addition, each thermal excursion consumes tin and increases the thickness of the copper-tin intermetallic. It should be noted that the intermetallic exists in two phases; the first phase exists closest to the copper surface and is known as the e-phase (copper rich). The second phase, known as the n-phase, is on top of the e-phase and is tin rich. These phase thicknesses continue to grow over time and certainly during thermal excursions. And as more tin is consumed in the formation of the intermetallic, the less solderable the immersion tin deposit becomes. Table 1 illustrates the growth of the intermetallics under lead-free reflow conditions3. The measurements were taken with SERA-sequential electrochemical reduction analysis.

Table 1: Immersion tin thickness in microns (u)-by SERA method.In addition, when the finish is exposed to elevated temperatures and higher levels of humidity, oxide coatings greater than 25 nanometers have been observed4, adversely affecting solderability. From an environmental standpoint, immersion tin formulations contain thiourea, a known carcinogen, as a reducing agent. One final issue is the aggressive nature of ImSn processes on soldermask. Concerns over mask attack and mask lifting are known. It is critical that a good quality soldermask be used in conjunction with immersion tin finishes.

Over the last two years, ImSn process improvements have been made. One notable development is referred to as an organic metal immersion tin process5. One such development includes the use of an organic nanometal to enhance the solderability of immersion tin6.

Immersion Silver (ImAg)

Immersion silver process works similar to ImSn. The silver is deposited from an acidic solution containing silver ions onto the bare copper surface of the circuit board. Since this also is a galvanic reaction, silver is deposited by displacing copper from the board surface. Typical thicknesses are in the range of 6-15 µin. The assembly side of the electronics industry finds that silver provides a flat, solderable finish that is compatible with soldermask and lead-free assembly. However, as more firms began to implement ImAg, two critical issues arose. The first was related to creep corrosion. In this case the finished product is exposed to a harsh environment of sulfur, and a few other corrosive elements (Figure 2).  The PCB finish coating’s resistance to creep corrosion is becoming increasingly important.  The reasons for this are twofold. First, more electronic equipment is being installed in areas less protected from atmospheric contamination. Second, increasing circuit density and decreasing PCB conductor spacing increase the risk of circuit failure due to creep corrosion. It is easy to see that excessive creep corrosion will lead to electrical shorts. Efforts are underway to improve ImAg processes to mitigate creep corrosion.

Figure 2: Creep corrosion. (Source: Dr. Randy Schueller, DfR Solutions)

The second issue to affect widespread adoption of immersion silver is the “champagne bubble effect” also known as planar microvoids along the pad surface (Figure 3)7.

Figure 3: Example of microvoids, a.k.a. champagne bubble effect. (Source: IPC W-27 Handbook, G. Milad, UIC, right, and Dr. Randy Schueller, DfR Solutions, left) 

Another issue that has plagued immersion silver that is cause for concern is soldermask interfacial attack (SMIA) or “trench etch.” Actually, this trench-etch defect is partly responsible for creep corrosion as bare copper is exposed, which allows atmospheric sulphur, along with environmental factors, to form copper sulphide. The creep corrosion reaction often emanates from the soldermask-copper interface as there is some exposed copper that is not covered by the surface finish8,9. A galvanic reaction then drives the reaction at the point where the SMIA took place, as shown in Figure 4.

Figure 4: Mechanism for creep corrosion. Exposed copper at the soldermask-copper interface interacts with sulphur in the atmosphere to produce corrosion by-products.

ENIG

ENIG and ENEPIG should not be lumped into the same bucket; each finish serves its own purpose.ENIG is a fairly robust (albeit costly) finish from a solderability perspective, and as a final finish provides the end user with several benefits:

• Flat coplanar surface;
• ICT testing;
• Press-fit connectors;
• Wire-bondable; and
• Withstands multiple reflow cycles.

Unlike some of the other finishes presented, the intermetallic is formed at the nickel interface with the tin in the solder, as nickel does not dissolve. One concern, however, is that the tin-nickel intermetallic (solder joints formed) is less reliable than the tin-copper intermetallic. Researchers have found that the tin-nickel intermetallic is more fragile and brittle than the copper-tin intermetallic formed with OSP, ImAg, HASL, and ImSn10.  The brittle nature of the tin-nickel intermetallic has led some end users concerned with shock-drop failures to abandon ENIG in favor of other finishes.

A second concern with ENIG is related to the interaction of the immersion gold with the underlying nickel. Since immersion gold is deposited by a displacement mechanism, this intimates that in order for gold to deposit on nickel (as an immersion reaction), some nickel must be removed or displaced by the gold. Depending on many factors that have been described elsewhere, there may be degrees of corrosion into the nickel along grain boundaries that can render the nickel unsolderable. Or at best, these grain boundary attacks interfere with a consistent formation of the intermetallic, leaving the solder joint compromised. There are also degrees of hyper-corrosion that, depending on multiple factors (nickel deposit morphology, phosphorus content, galvanic activity of the gold solution, gold concentration), can vary. Figure 5 shows the five degrees of hyper-corrosion. (Actually, degrees 1-3 are very minor and have not led to a loss of solderability. However, degrees 4 and 5 are quite severe and have led to compromised solder joints.

Note: Just because one may see a corrosion spike in a cross-section does not mean there is the black pad defect. All too frequently, perfectly solderable PWBs are rejected due to an erroneous conclusion that there is black pad. It is best to characterize the issue of hyper-corrosion as it is. Black pad is formed when excessive hyper-corrosion leads to an excessive amount of nickel removal, leaving behind a phosphorous-rich layer on top of the nickel. And one cannot see black pad after nickel plating. Black pad only manifests itself (if at all) after the gold is deposited.

ENEPIG 

If hyper-corrosion and black pad are of concern, ENEPIG is a solution. Here, the gold deposits onto the palladium, not the nickel. There is no hyper-corrosion effect as there is with gold over nickel. ENEPIG is often referred to as the “universal finish,” capable of good solderability and wire bondability. However, one must look at this more expensive finish in the context of the circuit board and its intended use/environment. One area in which ENEPIG has found use is the IC substrate market. ENEPIG can function as one finish for both wire bonding and solder attachment. While it is true that ENIG can perform the same functions, ENEPIG is more robust with respect to gold wire bonding. Typical plating thicknesses for this three-metal-stack over copper are as follows:

• Au Layer: 0.03-0.06 micron;
• Pd Layer: 0.10-0.50  micron; and
• Ni Layer: 3.0-6.0 micron.

The nickel present on the surface benefits from a Pd or Au protective layer to improve solderability by reducing brittleness and oxidation of the solder joint. The basic idea is to achieve improved solderability and wire bonding at reduced palladium and gold thicknesses. For the majority of ENEPIG systems, the palladium is deposited as an electroless reaction. Commercial palladium systems are based on one of two reducing agents, hypophosphite or formate. The former will co-deposit 1-6% phosphorous into the deposit, while the latter is nearly 100% pure palladium.

There is no industry specification for ENEPIG, although one is under development. A key component of any specification is the verification of both solderability and wire bondability at varying palladium thicknesses. Again, lower thicknesses of both palladium and gold will enhance the economic viability of this finish as long as the solderability/wire bonding requirements are met.

Selective Electroless Nickel Immersion Gold (SENIG)

Several years ago SENIG offered the promise of reduced gold cost due to the selective nature of the process. In this case, gold would only be deposited on circuit features that would require wire bonding or act as touch pads or contacts. The remainder of the board would be processed through OSP. This promise has been met for the portable devices that are cost sensitive but require high reliability against shock-drop. The process sequence requires a secondary imaging step after soldermask. The photoresist protects the area on the circuit card from the ENIG processing steps. After the ENIG is plated, the resist is stripped, exposing the remaining circuit features, which are bare copper. Here, the board is processed through the OSP line to complete the selective process.

There are a few critical issues that the engineer must be aware of when working with SENIG. First, it is best to employ a higher phosphorous content nickel (i.e., 9-12%) to provide the corrosion resistance necessary to withstand leaching from the photoresist as well as the acidic preplate chemistry of the OSP.  The second consideration is the ability of the OSP to properly coat the bare copper while leaving the gold deposit virtually untouched. Finally, the photoresist used in the secondary imaging step must be capable of withstanding the aggressive nature of the ENIG chemistry and temperatures.

Direct Electroless Palladium over Copper  (EPd)

The EPd process is not entirely new. EPd actually debuted in the mid-1980s for the automotive industry. However, as palladium prices soared, the process fell out of favor. However, thanks in part to ever-increasing gold prices, direct electroless palladium as a solderable final finish has gained new momentum. In the process, palladium is applied directly over copper and is a true electroless (autocatalytic process). The EPd finish may be considered a temporary coating if soldered to the solder joints, forming CuSn IMCs, or a permanent coating if used as a key pad-surface or edge-connector, or a surface for aluminum wire bonding. Even during aging conditions, there is no evidence of copper diffusion through the palladium deposit. This allows for its use as a switch-pad interface or edge-connector interface where reliable and consistent resistance is a key to reliability.

Solderability results have shown that EPd allows for sufficient solder paste spreadability and PTH flow-up, thus providing for reliable solder joint formation. When compared to other commercially available solderable finishes, EPd compares favorably with respect to ball shear strength, wetting balance testing, and solder joint strength. With a direct deposit of palladium over copper, there are no concerns with hyper-corrosion or black pad11.  

Organic Solderability Preservative (OSP)

OSP is a non-metallic coating designed to protect the underlying copper from oxidation and in turn enhance solderability. OSP is not a wire-bondable coating. However, in terms of surface area of printed circuit boards processed with final finishes, OSP is the most widely used. When applied to the PCB correctly, OSP will enhance solderability for both tin-lead and lead-free alloys through multiple thermal cycles. OSP has evolved through several generations of development and the latest iteration has shown improved resistance to oxidation through 5-7 IR lead-free reflow profiles. This enhanced performance has been accomplished primarily through higher stability of the organic coating at the extended reflow temperatures. An OSP coating composed of substituted (alkyl) benzimidazole compounds, which are still widely used in the market today, begins to decompose around 250°C. This is the typical peak-temperature for lead-free soldering. However, the latest OSP composed of Aryl-phenylimidazole technology provides a higher decomposition temperature (approximately 354°C). The advanced azole molecule that is the primary material of the latest generation OSP exhibits much higher temperature stability than the conventionally substituted benzimidazole chemistries, and is beneficial in promoting solder paste spread and solder through-hole flow-up12.

Another proven benefit of latest generation OSP is the excellent solder joint strength when compared to other surface finishes (Figure 6).

Figure 6: Ball shear strength as measured in Newtons (y-axis).

Test conditions are shown in Table 2. A Dage Series 4000 Ball shear tester was used for all shear tests on BGA test vehicle.

Table 2: Test conditions for BGA shear tests.

Finally, if there has been one drawback to OSP, it is the ability to electrically test through the organic film. Concerns over the OSP film transferring to the test pins, creating false opens, is an issue. However, latest generation OSP has been improved significantly to allow for reliable ICT testing.

Of all of the surface finishes presented in this paper, OSP is the simplest process in terms of chemical control. In addition, OSP remains the lowest cost by far of all the surface finishes.

Summary 

It is safe to assume that the solderable finish market will continue to evolve. It is expected that many iterations of solderable finishes will be introduced in the coming years. For now, engineers have a robust line-up of finishes to select from. Decisions in this regard should focus on the best fit for each PCB's end-use application and environment. Of course there will be adjustments assemblers can make to enhance the performance of the finish in the context of the application. Regardless, no one finish will satisfy all of the requirements for every PCB application. Choose the finish that satisfies the majority of the requirements at the lowest overall cost.  

References:1. Prismark market study .2. Scimeca, T, Sikorcin, G. and Lentz, T., “HASL and Flow: A Lead-free Alternative,” Circuitree Magazine, February, 2008.3. OM Group, internal correspondence. 4. Schueller, R., “Considerations for Selecting A Printed Circuit Board Surface Finish,” SMTA International Conference Proceedings, 2009.5. Arendt, C. Arribas, J. Posdorfer, M. Thun, B. Wessling, OnBoard Technology,12, April, 2006.6. pcb007.iconnect007.7. Mukadam, M., “Planar Microvoids in LF Solder Joints”, SMTA International Conference Proceedings, 2006. 8. C. Xu, W. Reents, J.Franey, J.Yaemsiri and J. Devaney, “Creep Corrosion of OSP and ImAg PWB Finishes”, SMT Magazine, November, 2010.9. Veale, R., “Reliability of PCB Alternate Surface Finishes in a Harsh Industrial Environment”, SMTA International Conference Proceedings, 2005. 10. Chai, TC., Hnin, WY, Wong, EH, et. el., “Board Level Drop Test Reliability of IC Packages,” Institute of Microelectronics.11. Trainor, J., “Electroless Palladium as a PCB Surface Finish,” SMTA International Conference, 2011.12. Carano, M., “The Evolution of Organic Solderability Preservatives (OSP) from a Temporary Protectant to a Leadership Position in Surface Finishing Chemistry, “ Circuit World, 37 (2)., 2011.

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 nearly 30 years. Carano holds nine U.S. patents in topics including plating, metallization processes and PWB fabrication techniques.