(1) Final finish for the fabrication process to replace hot air solder leveling (HASL).
(2) A lead-free alloy to replace the current tin/lead solder in wave machines at assembly.
(3) Solder pastes to replace those currently utilizing tin/lead solder powder.

1. Introduction

In the past few years, there has been an accelerated push in the search for alternative printed circuit board final finish coatings. Typically, HASL has been used as the final finish for well over a decade. However, the increased demands of surface mount coplanarity with finer pitch surface mount pads has been the primary impetus in the search for alternative coatings. Recent lead banning legislation has also gained momentum and has accelerated the need for lead-free final finish alternatives.
The purpose of this study was to determine the the viability of lead-free final finishes with respect to solderability as a function of a number of accelerated aging tests. In this study, both the wettability and solder wicking are measured and compared among the lead-free final finishes and referenced to the lead containing standard: Hot Air Solder Leveling (HASL). The final finishes chosen include the following:

Hot Air Solder Leveling (HASL)
Organic Metal based Immersion Tin
Conventional Immersion Tin
Electroless Silver
Organic Solder Preservative (OSP)
Electroless Nickel / Immersion Gold

The ability of each final finish to address the various assembly and bare board issues is shown in Table 1.

All final finishes, with the exception of Organic Metal based Immersion Tin, were prepared on Ormecon sample test boards or copper clad laminate from a number of printed circuit board shops utilizing a given final final process. The ORMECON CSN final finish was made at Ormecon in Germany according to standard procedures.

Table 1: Overview "Performance characteristics of Surface Finishes" (cf. original paper)
After the etch surface preparation, the sample is immersed for 30 sec in Organic Metal water dispersion PCB 7000, followed after the tin deposition in CSN 7001.

Accelerated aging

A number of different accelerated aging tests were performed on each final finish:

dry aging was done at 155 °C for 4 hours.
steam aging for 8 hours
at 65° C / 85% relative humidity (RH) and
at 85 °C / 85% RH performed at Trace Labs and Ormecon, resp..
unaged samples were subjected to three air reflow cycles.

Solderability tests

The first solderability test was to take a test board containing 16 small thru-holes, 24 medium size holes and 16 larger holes. The test board contains varying sized surface mount pads as well. The board was coated with the final finish and subjected to the various aging tests. Following this, the board was immersed in Florida CirTech’s SS-10 no-clean wave solder flux. The board was then used on a commercial wave solder machine without any components to see if the solder would wick into the holes.
In the second solderability test, a small amount of Florida CirTech NC 611 no clean solder paste was applied onto a copper clad coupon coated with the final finish subjected to the various aging processes. In this case, the solder paste was printed in the form of a circle with a diameter of 0.068 inches. The degree of solder paste spread as the paste reflowed was then measured by photocopying the coupon and cutting out the area with solder paste and weighing the paper with a very accurate scale (0.0001 gram sensitivity).
A novel solder paste spread test was also performed. In this case, 20 milligrams of NC 611 no clean solder paste was printed onto a small circle with the final finish surface. In this case, the circle was connected to a 20 mil. wide 2.1 inch long trace, analogous to a mercury thermometer. The greater the wetting, the greater the paste would be pushed along the trace (higher temperature analogue).
Additionally, test coupons with the same final finishes were prepared and subjected to identical aging. The thickness of each layer: surface oxides, surface tin and copper-tin intermetallic species were measured by Selective Electrochemical Reduction (and Oxidation, resp.) Analysis. These measurements were then correlated with wetting balance measurements where the wetting force and wetting angle were both experimentally determined.

Discussion of results

The results of hole filling are shown in Table 2 and 5. A hole was considered filled if the hole was more than 50% filled. In addition to hole filling, the test boards contained two very large surface mount pads. In the case of steam aging, the OSP and Ni/Au final finishes exhibited solder dewetting on these planes. This is consistent with the low hole filling for these two final finishes with steam aging. In the case of dry aging (155 °C for 4 hours), the OSP and Silver final finishes dewetted. While the Silver still showed respectable solder wicking (95%) hole fill, the OSP was unacceptable at 11%. In all other accelerated aging tests, all final finishes showed acceptable wetting on the surface mount pads. Finally, the average of all six process including the virgin final finishes was taken to give a better picture of the performance of each final finish with respect to hole fill as a function of aging.
The second array of tests concerned the degree of solder paste spread on a bare copper clad laminate coated with each final finish. The solder paste spread was measured by weight and normalized to the weight of paper corresponding to a 0.068 inch diameter circle. The solder spread results are shown in Table 3. In this case, numbers exceeding 1 would correspond to positive wetting and any number less than one would be negative wetting or dewetting. A zero indicates complete dewetting as the solder paste minimizes it’s surface tension by coalescing into a ball. Again, one can see that poor wetting for Ni/Au is seen for steam aged boards and with OSP for dry aging. This is consistent with the hole fill and dewetting results shown in table 2 and 5. Again, an average was taken to illustrate the general wetting of each final finish averaged over all aged processes.
The solder wetting were also measured by a new technique discussed above where the reflow solder paste pushes the excess paste along a connected 20 mil trace of 2.1 inches in length. The results of this tests are shown in Table 4. In this case, the 2.1 inch length was multiplied by a factor of 1000 for simplification. One can see that HASL seems to push more paste along the trace than any of the other alternative finishes. ORMECON is a respectable second and the other finishes are well below this. Very little solder wetting is seen for the OSP and Silver coatings.
Finally, wetting force and wetting angle measurements performed are shown in Table 5. In this test, one can see that the wetting force, which measures instantaneous solder paste wetting at reflow, is satisfactory for most cases. However, in the case of dry aging, both OSP and Silver have no wetting force. Again, this is consistent with the poor hole filling and dewetting results shown in table 2 and modest or poor solder paste wetting results listed in tables 3 and 4. Solder force wetting numbers between 2 and 4 are probably not as significant as the solder wetting angles results discussed next.
In the solder wetting angles, a higher angle is indicative of a greater degree of dewetting. The higher the angle, the greater the interfacial tension. One can see that in general, a greater solder wetting force is correlated with a lower solder wetting angle. In addition, one sees a fairly high solder wetting angle as the final finishes are subjected to aging. There are some other fairly general trends one can observe. Again, dry aging is particularly difficult for the OSP and Silver finishes (>120 degrees). HASL has a fairly consistent wetting angle regardless of aging and the immersion tins are also relatively low. Ni/Au has an increasing wetting angle as this finish is subjected to aging with steam aging in the area of 100 degrees.

Table 2 Hole filling as a function of aging and coating using no clean flux.

 
Table 3 Solder paste spread as a function of aging and coating using no clean solder paste

Table 4 Solder wetting as a function of aging and coating using no clean solder paste

Comparison of different Immersion Tin finishes

Additional to the aging and wetting tests shown here, we have performed an extensive study regarding the aging kinetics (under dry conditions). For this purpose, aging was made at various different temperatures (between 50 and 170 °) for different time. After this, the tin and intermetallic layer thickness was electrochemically (oxidatively) measured, as well as the oxide layers by electrochemical reduction.
As can be seen in table 5, the conventional tin has quickly lost almost all of its tin layer, and has grown significant amount of oxides. But as the tin chemistry is different from ours, we have decided to compare here ORMECON CSN deposited with and without pretreatment by the Organic Metal, in order to be able to directly determine the effect of the use of the Organic Metal, which was earlier found to catalyze the tin deposition by a clear one-electron process.
As the temperature and time conditions covered a broad enough range also close enough to the real aging condition (at room temperature), we have been able to determine the activation energy for the tin-into-copper diffusion and for the oxide development, and to interpolate the kinetics for room temperature.
The surprising result is, that the velocity at which tin is lost or oxides are grown, is not just only more or less higher for the tin deposited without the Organic Metal (as one would expect from aging results at 155 °C). But in contrast, the lower the temperature, the slower diffuses the tin which was deposited with the help of the Organic Metal, and the slower also it gets oxidized. (Table 6 and fig. 1) It can be seen, that at room temperature, over one year only 0,29µm tin is being lost compared to 1,46µm tin when deposited without the Organic Metal (at 155°/4 h, the respective values are 0,60 and 0,68).
After one year, only 0,73 nm SnO and 0,03 nm SnO₂ are formed, when tin was deposited catalyzed by the OM; however, without OM, 17,9 nm and 0,56 nm of these oxides are formed, parallel with the decrease of solderability. (The respective values at 155°/4 h are 2,36 / 0,91 nm and 2,4 / 1,2 nm.)
Microscopic and SEM studies show, that the OM process results in a highly crystalline tin surface, whereas the conventional procedure leads to more or less completely amorphous tin layers. Here we see the reason for a quicker diffusion and oxidation, resulting in a quicker loss of solderability. 

Table 5 Analytical results of various surface finishes

Nevertheless, we have not been content with some aspects of the ORMECON CSN found in the humid aging. Compared to HASL and Ni/Au, the wetting force and even more the wetting velocity is not as high after humid aging for the OM catalyzed immersion tin.
As a result of a recent development project, a new modified process was designed leading to a finish which shows no decrease in the dry aging results, but a significant increase of the wettability and solderability after humid aging. The wetting balance shows, that wetting occurs almost instantaneously, at least after 1 or 1.5 sec. Sometimes, wetting was so quick that the wetting balance apparatus did not reckognise it.
Also solderability under commercial technical conditions has been improved to be now 100% reliable.First samples in technical scale are under test in facilities of various selected customers in Europe, Asia and US.

Table 7 Comparison of results for all surface finishes

"rating" is the area of the Decaeder formed by the length of 10 axes, having an angle of 36° between each other, representing the values in each different column of the table; the values "wave solder" are generated by dividing the % value by 25

Summary of results

A matrix of final finishes solderability as a function of the various aging processes has been studied. While the results are not one hundred percent internally consistent, the trends are unmistakable. Overall, HASL performs very well under dry and humid aging conditions. Ni/Au also does well except for steam aging. On the average tin also performs quite well, especially when deposited with the help of the Organic Metal. A new OM / tin process leads to finish characteristics matching those of HASL. OSP and Silver, while performing well under humid conditions, are both quite susceptible to poor solderability when subjected to dry aging. The results are comprised in table 7.