Nitrogen and Soldering: Reviewing the Issue of Inerting

Estimated reading time: 15 minutes

While accepted in the wavesoldering process, nitrogen use in reflow ovens is frequently questioned despite a wealth of research.

By M. Theriault, P. Blostein and Dr. A. Rahn

Using nitrogen to inert the atmosphere during soldering has become a generally accepted practice in electronics assembly. But whereas there appears to be little doubt in the minds of users that nitrogen "improves" the wavesoldering process and that it actually "pays its way" simply by reducing dross, its application in reflow equipment is still questioned.

This insecurity, or even misinformation, persists despite a wealth of research on nitrogen use - applied to wave as well as reflow soldering. This discrepancy warrants a re-examination of the entire inerting issue. Based on published results, corroborated by secondary findings as well as research, a clear picture emerges as to when nitrogen may be used, why it will benefit the process and where limitations may be found.

Nitrogen Gas

Nitrogen is a common gas found in the atmosphere. Approximately 78 percent of the air we breathe consists of nitrogen. The rest is made up of oxygen, carbon monoxide and carbon dioxide, as well as traces of argon and other gases. The fact that animal life survives in air is an indication that nitrogen is nontoxic. But pure nitrogen does not sustain life - not because of any inherent toxicity, but rather because of a lack of oxygen. Hence, when using nitrogen, precautions must be taken to avoid a situation where nitrogen is present without the required amount of oxygen.

Nitrogen boils at approximately -320.4°F (-195.8°C), meaning liquid nitrogen (as it may be delivered) is very cold. Cold nitrogen also is heavier than air and accumulates in low-lying areas, if given a chance.

The absence of oxygen in pure nitrogen and the fact that it does not readily react with common metals have earned it the designation "inert." It is this property that benefits the soldering process by protecting metal surfaces from oxidation during heating and ensuring proper flux action. Nitrogen is used to displace oxygen because it is the least expensive gas that is inert, or chemically unreactive.

Nitrogen is not manufactured but extracted from air, giving a zero-sum gain as far as the environment is concerned. The extracted nitrogen is returned to the air, and the air is neither depleted nor enriched because of this circular process.

There are a number of different ways of extracting nitrogen from air. The most common is a freezing process usually referred to as "cryogenic." A cooling process is used to liquefy the gas, which, in a reverse-distillation process, purifies itself because of the different boiling points of oxygen and nitrogen. More recent extraction techniques are on-site production and liquid-assist systems that use a special cryogenic process assisted by a small amount of liquid nitrogen. Other on-site generation systems include membrane process and pressure swing absorption.

With these different processes, it is possible to provide nitrogen from a purity of 95 to 99.99999 percent (0.1 parts per million [PPM]), depending on the need of the process. Obviously, the different processes and purities carry different price tags. The price is also affected by the amount consumed, the delivery system chosen, the distance from the manufacturing plant and many other factors. Because most assemblers require a high-purity atmosphere (less than 1,000 PPM), liquid nitrogen or liquid-assist on-site systems are recommended for assembly operations.1

Effects of Nitrogen

In soldering, there are a number of basic phenomena caused by changing the ambient air to nitrogen. By examining these effects, including spreading behavior, wetting force, wetting angle and surface tension, it can be seen that nitrogen provides noticeable benefits for both wave and reflow soldering.

Research examining the spreading behavior of solder under different environmental conditions has shown that spreading starts at lower temperatures as the residual oxygen level (ROL) is reduced. In this study,2 it is found that 63Sn/37Pb solder already spreads at 401°F (205°C) if the ROL is less than 10 PPM, but it needs 404.6°F (207°C) at 100 PPM and 518°F (270°F) at 1,000 PPM. The same pattern holds true for other solders, indicating oxides inhibit spreading.

Comparing wetting force for different fluxes in air and nitrogen reveals that nitrogen coverage can improve the process. Usually measured with a wetting balance, wetting force is an indicator of the quality and shape of the joint. A study3 comparing a number of fluxes found that, with one notable exception (adipic acid), the wetting force increases if measured under nitrogen rather than air. For some fluxes, the gain is substantial.

One measure used to estimate wetting is the assessment of the wetting angle. Small wetting angles imply good wetting behavior and usually indicate a sound joint. When measuring wetting angles for a variety of fluxes, the same report found that wetting angles were much smaller under nitrogen than in air.

The surface tension of the liquid solder plays a major role in the soldering process. It is responsible for the shape of the fillet as it counteracts wetting and gravity. It is practically impossible to measure the solder surface tension under air, as a minute amount of tin oxide in the surface layer will falsify the measurement. Nevertheless, these measurements3 again indicate that the surface tension is higher under nitrogen than under air.

An excellent demonstration of higher surface tension is verified by an experiment carried out in 1996.4 Looking at the amount of bridging left after reflow when solder pads were overprinted (40 percent coverage) using LR and RMA flux activation in the paste, the dependency on nitrogen coverage became apparent. Only under lower ROL values did the surface tension reach a high enough value to break the bridge and to collect all the solder on the pads only. An increase in wetting force may have aided, too.

The same experiment also investigated the relationship between oxygen deprivation and metal powder particle sizes in pastes. The current tendency toward fine pitch and the consequent use of paste with finer grain sizes increases the importance of such questions. Not only did the experiment confirm that nitrogen coverage improves solder spread in all cases, but it also showed that the finer the powder size, the more crucial it is to use an inert cover gas to achieve sufficient solder spread.

A somewhat puzzling result has been recorded5 in a study related to wetting and surface tension. It may perhaps be understood best in the following way: If copper coupons that exhibit varying degrees of solderability are prepared and then tested under different environmental conditions (ROLs from 7 PPM to air), it is found that those coupons that have excellent solderability perform about equally well under all conditions. However, those with lower solderability need less time to solder under low-ROL conditions than under higher ones. That is, the spread of data (i.e., the variance or standard deviation) is greater under air than under better inerted conditions.

These test results are the explanation for the "increased process window" when addressing nitrogen soldering. The process becomes more forgiving under nitrogen than in air. Excellent solderability does not gain much from the absence of oxygen; it seems, however, even a minor deficiency in solderability benefits from inerting. In practical application, this means that if all boards and components always have excellent solderability, nitrogen does not significantly improve soldering parameters. Realistically, however, with varying solderability, the set of process parameters is much broader under nitrogen than it would be in air.

Driving Forces Behind Nitrogen Usage

What broader industry trends favor nitrogen use? One is the search for a superior surface finish on printed circuit boards (PCB). Fine pitch demands better planarity conditions than hot-air leveling (HAL) can offer. The use of low-solids or no-clean fluxes was the answer to the Montreal Protocol and the cleaning brain-teaser, but this does not make the search for a better surface easier. The introduction of ball grid arrays (BGA) and the resurfacing concern for joint quality also play into the surface issue. The complexity of the modern assembly is not only reflected in the number of board layers, but also in the pressing need for multiple soldering processes or clever process modifications such as reflow soldering through-hole components.

It seems that organic solderability protection (OSP) - a collective name for some copper treatments such as Imidazole - also can provide an answer to these problems. Besides good planarity, it provides limited storage capability and certain price advantages. When LR fluxes are used in multiple thermal excursion processes, however, OSP performs best when soldered under an inert atmosphere. There, it may even outperform Ni/Au.6,7

The question of solder joint reliability was underlined in a project supported by the German government8 in which stress experiments yielded some unexpected results. Failure patterns in two components showed clear differences between those joints soldered in air vs. those soldered in nitrogen. Whereas 183 failures of one component were recorded for joints soldered in air, only three were found for those soldered under nitrogen. The ratio of failures was also favorable for the other component under nitrogen (air: 302; nitrogen: 100).

Such findings are corroborated by other research, including a similar study10 and others in which the peel strength of joints was examined.

Wave Soldering Under Nitrogen

One of the primary differences between wave and reflow soldering is the creation of dross. Dross is formed in wave soldering when the molten solder wave comes in contact with the oxygen contained in atmospheric air. Few users recognize that there are two distinctly different types of dross. There is the silvery sludge that covers the surface, which consists mainly of good solder and some tin oxide and may, with reservations, be deemed mostly harmless. At turbulent spots and around the pump shaft, however, one may observe some black powder. This is the second type of dross that may contain lead oxide - a toxic substance. Although lead normally would not oxidize at these temperatures, friction and the resulting powder may increase the surface ratio so much that lead oxidation may occur, posing a major health hazard. This black dross must be treated with the utmost care and reference should be made to local environmental, health and labor legislation. Naturally, the use of an inert atmosphere largely eliminates the creation of both types of dross.

Another distinction is related to flux, which is applied separately in the wavesoldering process. The choice of flux directly impacts solderability issues as well as the resulting cleanliness of the assembly after soldering. There are other issues also related to flux, such as costs and potential problems with volatile organic compounds (VOC). It is well established that the use of nitrogen generally allows the use of milder fluxes and much lower quantities than would be tolerable in air. It is less understood that residues under nitrogen are less objectionable and, where necessary, easier to clean because they are not oxidized.

In wave soldering, the flux application method is often changed from foam fluxing to spray fluxing, once the step from ambient soldering to inert soldering is taken. Controlling the amount of flux applied to the board is essential, as it correlates directly with a decrease in surface insulation resistance (SIR) value after soldering. High SIR values are becoming more important as pitch decreases, particularly with high-frequency circuits.

But spray fluxing under nitrogen also has its monetary rewards: savings in flux and alcohol solvents may be dramatic and may directly justify the switch to spray fluxing. Most wave equipment manufacturers recommend as much as a 60 percent reduction of the flux application when inerting. Various published and unpublished reports have supported flux reductions in spray fluxing when switching from air to inert.

Figure 1. Example of an inert boundary system.

Inerting the wavesoldering system may be done in a number of different ways. Systems are available that protect the solder pot area with a blanket of nitrogen gas bled over the wave crest (inert boundary). These systems can be retrofitted (Figure 1). Such methods enable a significant dross reduction and provide higher quality while minimizing investment and maintaining low operating costs.

Inerting tunnels are offered in different lengths and designs. It is easier to inert the critical region properly if the interface with the environmental air is pushed as far away as possible. Properly designed and maintained tunnel systems translate to lower dross production than short hoods and inert boundary systems. Capital costs of tunnels are, however, higher than localized inerting systems. Whether it is necessary to seal equipment hermetically remains to be considered.

Nitrogen and Reflow

In reflow flux activity, residues and cleanliness also play a major role when deciding whether nitrogen should be used. The same arguments apply that have been used in wave soldering: residues are less objectionable and, if necessary, easier to clean because they are not oxidized. One study9 even indicates that nitrogen reduces residues for some fluxes by as much as 66 percent compared to their application in air. This explains why many users switching from ambient atmosphere to nitrogen notice a reduction of needle-test failures. Failure of the needle-test depends on the amounts of deposited residues and the stickiness of the rosin.

The most important factors in considering a nitrogen process are defect levels and joint reliability. Longitudinal studies10 have provided enough reliable data to confirm that both of these factors are positively affected by nitrogen use. In large-scale manufacturing situations, defect levels were monitored over a two-year period, one year prior to a switch to nitrogen reflow and one year after the switch was completed. With the introduction of nitrogen as the only significant change, the defect level fell from 82 to 37 dpm.

Other operations11 where nitrogen was introduced have shown improvements in first-pass yield between 5 and 7 percent, which translates into a reduction in defect levels of 50 to 60 percent. The fact that not every introduction of nitrogen was equally successful is explained by differences in layout and pitch. Admittedly, bad layout cannot be compensated by simply introducing nitrogen as a cover gas. It seems, however, that those processes with the narrowest pitch benefit the most from nitrogen use.

Lead-free Soldering

After a certain amount of anxiety caused by legislation before the U.S. Congress and Senate in 1995, the question of lead-free soldering had become a non-issue until a push by the Scandinavian Countries introduced a second (and now a third) draft on lead waste in the European Union (EU). This proposal would effectively ban the use of lead for European companies by January 1, 2004. With heavy support from Japan and some cautious reactions in the United States, lead has once more become a major point of concern and discussion.

The problem of soldering without lead is an intricate one. Many different aspects play into it, including toxicity, availability, price, worldwide distribution, wetting ability and reliability.

Much effort has gone into finding a replacement for 63/37, not only because of its lead content (most specialists agree that the danger posed by lead in the solder is rather limited) but because there is a general need for a "better" solder. But as of yet, no one has been successful in finding that ideal combination that replaces (as a drop-in) the solder of 5,000 years ago.

The few potential choices (tin plus some copper; tin plus some silver; tin plus some copper and some silver) all need higher process temperatures and thus may not really work without good inerting.12 The increased amount of tin in the solders not only makes them more expensive (lead is cheap), but heightens the tendency to oxidize, form dross and react with other metals. Although the higher tin content would lead one to believe that wetting could be improved, this is not generally true. Only under nitrogen can similar wetting be achieved.

The question of inerting becomes even more critical when other solders are considered: indium, zinc, bismuth and antimony are just some elements that may be found in future solders. Most such solders show very bad wetting properties in air and some are even marginal under nitrogen. The CERCLA Priority List of Hazardous Substances reveals that some of the reaction products of these metals may actually pose a greater threat to an employee's health than lead. Whether nitrogen can help in this regard must be studied in great detail.13

Conclusion

Nitrogen use in flow and reflow equipment may benefit the process as well as the quality of the end product. In both cases, joint reliability may increase and defect levels may drop to improve first-pass yield. In the case of wave soldering, other benefits may accrue, such as dross reduction, reduced solder pot maintenance, a safer operation and substantial flux savings.

A nitrogen process should not be adopted, however, without fully understanding its benefits and limitations. Nitrogen opens the process window by forgiving the process flaws, but is the increased window necessary? As the pressure to reduce cost in the industry increases, the ultimate decision to use nitrogen should be based on a solid cost-vs.-benefit analysis that goes beyond the unit cost of nitrogen itself. Nitrogen inevitably adds cost to the process, but its overall benefits normally should outweigh the additional adoption expense.

Despite a continuous improvement in flux chemistries, nitrogen is here to stay and its usage will increase. The continuous integration of components, the increased demand for higher quality and reliability assemblies, as well as the introduction of lead-free solder, all increase the vulnerability of the process to oxygen and oxides, enhancing the value of inert soldering.

REFERENCES

1. M. Theriault and P. Blostein, "Reducing the Cost of Inert Soldering," Circuits Assembly, July 1998, p. 46-52.
2. C.C. Dong, et al., "Effects of Atmosphere Composition on Soldering Performance of Lead-free Alternatives," NEPCON West 1997.
3. S.M. Adams, E.K. Chang and M.J. Kirschner, "Atmospheres and Performance During Soldering," NEPCON West 1994.
4. C.C. Dong, et al., "Oxygen Concentration in the Soldering Atmosphere - How Low Must We Go?" NEPCON West 1996.
5. C. Lee, "The Significance of Oxygen Concentration in the Atmosphere," Circuit World, No. 18, 1992.
6. J. Greaves, et al., "An Evaluation of Alternative Surface Finishes by the Circuit Card Assembly and Materials Task Force," NEPCON West 1997.
7. D. Verbockhaven and G. Conor, "OSP Coatings: Influence of a Nitrogen Atmosphere on the Soldering Performance," NEPCON West 1996.
8. M. Ehrlich, G.S. Laneus and G.E. Reichelt, "Investigations about the Reliability of Atmospheric and Inert Wave-soldered 1812 Capacitors at Various Footprint Configurations," Proceedings of NEPCON West 1994.
9. Thibault, Bell and Lemieux, "Reduction of Residues Using Various Atmospheres," Proceedings of NEPCON West 1998.
10. M. Gerhard and H.E. Kiecker, "Quality Improvement by the Implementation of Reflow Soldering with Nitrogen," Productronica Conference, Munich, 1996.
11. C. Boeding, et al., "Measuring the Benefits of Nitrogen for Reflow," NEPCON West 1997.
12. Pratt and Trumble, "Process and Material Characterization for Lead-free Tin/Copper Solder Alloy," SMI Conference, San Jose, Calif., 1998.
13. National Center for Manufacturing Sciences: Lead-free Solder Project, Final Report, August 1997.

For more information, contact Air Liquide America Corp., 2700 Post Oak Blvd., Houston, TX 77056; (800) 820-2522; Web site: www.airliquide.com.