Lead-free implementation

Unit 4: Alternative solders

Contents

Working towards Assignment 2

For most of the AMS products, there will be no alternative to replacing lead-containing solder by a lead-free substitute. But which material or materials to choose? As you read this unit, keep in mind the needs of the whole range of AMS’s customers, as well as any limitations in the equipment currently available that might become apparent.

Introduction

In this unit we are looking at the solder material itself, rather than the joints it makes (Unit 5), the implications for board fabrication and assembly practice (Units 6 and 7) or any reliability issues (Unit 8). Our focus is also the soldering operation (hand; wave; reflow), rather than the use of the solder for the finish on PCBs and component terminations, especially flip-chip bumps and the balls on BGAs and CSPs. Inevitably, however, some process issues will intrude – if you come across interesting information, store it away for later use!

Requirements for alternative solders

As we will be doing for board finishes in Unit 7, we are starting by looking at what is required of a satisfactory alternative to tin-lead solder. This is mostly a matter of reflecting on your experience of assemblies and their method of construction, but also a challenge to think laterally about the needs of suppliers of the materials.

Exercise

Use your knowledge of soldering, and the way in which solder is used within electronics, to compile as complete a list as possible of the requirements that we have to take into consideration when selecting an alternative solder.

When you have done this try a search for solder + requirements + "lead-free alternative".

Click here to see our comments and an alternative perspective from Jennie Hwang.

Last revised 25 June 2004

For most board-assembly purposes, it would be desirable to have a ‘drop-in replacement’ for the commonly used eutectic tin-lead alloys. Unfortunately, such a drop-in replacement with this unique combination of properties doesn’t exist! A wide range of alternative solder materials has therefore been explored, if not yet fully evaluated. The only viable alternative found has been to replace the bulk of the lead by tin, and the search for materials becomes what Harrison called “the search for alloying additions to tin to offer the best, or failing that the least worst, match to the characteristics of the tin-lead family as electronic solders”.

[back to top]


Possible alloying elements

Why did Harrison refer to alloying additions to tin? Simply because tin is one of only two elements that will readily wet to the kinds of surfaces presented by electronic components and PCBs.

Pure tin melts at 232°C; a silvery-white metal, it is soft, ductile and malleable, non-toxic and resistant to corrosion. Its light in comparison with copper, though heavier than aluminium and zinc. It readily forms alloys with other materials, the resulting alloys including bronzes and pewter and specialist alloys used for type and bearings. But tin is quite expensive, and melts at too high a temperature. There are also concerns about the stability of pure tin, which can convert to a grey amorphous α-phase at low temperature and about the tin ‘whiskers’ we will meet in Units 7 and 8, which are caused by internal stresses.

So we have to add different elements to the tin to reduce its melting point and deal with some of the other issues. [It goes without saying that lead is by far the most successful such additive, as witnessed by its long history of use!] Use the web search which follows to identify the other constituents present in as wide a range of lead-free solders as possible.

Research topic

What elements other than tin are used in lead-free solders?

Resource Link(s) Notes

interview with
Jennie Hwang (2001)

To view the video a Real media player is required.
Click here to download a free version of Real Player 10.
look at the early slides

Step 1.2 in the Soltec “5 steps to lead-free soldering” (see below)

Try browsing for lead-free solder formulations at any of the suppliers. There is a useful list on the Soldertec links page.

 

show solution

Last revised 27 June 2004

Key information

Soltec

One of the resources we use in this Unit perhaps requires a little explanation as it will be used elsewhere in your studies. This is the incorporation of the Vitronics-Soltec presentation “5 Steps to Lead-free”. First presented in 2001, this has continued relevance as a resource, and you will probably want to browse other areas. The main page is accessed at this link, while specific tasks will be linked to whichever of the steps is relevant to the research task.

An alternative view to the Soltec 5 Steps is the sequence that used to be promoted by Process Sciences:

1 Establish objectives and timetable
2 Select your solder alloys
3 Select your lead/pad finishes
4 Select your solder flux
5 Qualify your parts
6 Qualify your boards
7 Qualify your new process
8 Qualify your final product
9 Document your new process

This is a theme to which we will be returning. It is interesting to note that both PSI and Soltec start with the solder alloy selection, as this has a wide overall impact.

[back to top]


Preferred materials

Whilst different workers have varied in their preferences and work continues to develop to improve solders, over recent years a general consensus has emerged that certain of the alloys should be selected for general use. In our next research exercise we ask you to look at some of the alternatives and the basis on which these decisions were made.

Research topic

Which are the preferred materials for lead-free solders?

Is the material of choice different for different soldering methods?

Does everyone agree?

Resource Link(s) Notes

interview with
Kay Nimmo (1999)

To view the video a Real media player is required.
Click here to download a free version of Real Player 10.

ZWEI Lead-free soldering: Materials, Components, Processes (2000)

Not as dated as it might seem – the issues are still out there! Section 4 relates to solder choices

interview with
Bruce Moloznik (2002)

To view the video a Real media player is required.
Click here to download a free version of Real Player 10.

interview with
Steve Dowds (2002)

To view the video a Real media player is required.
Click here to download a free version of Real Player 10.

Ken Snowden presentation on Lead-Free Solder Materials (2004)

The early slides have information on material selection

show solution

Last revised 27 June 2004.

[back to top]


Solders for reflow soldering

The options for solder paste include tin-silver, a reliable alloy which, because of its use on early projects, has the longest history of any of the lead-free materials. It has good mechanical properties, and indeed thermal fatigue testing has shown improvements over tin-lead alloys. A eutectic, the Sn3.5Ag alloy1 melts at 221°C. Nevertheless, apart from some conservative military users, the emphasis amongst users has swung round to ternary alloys including copper, which offer reduced melting temperature and arguably better characteristics.

1 When a solder formulation is expressed in this way, it means that the first element named forms the balance of the composition. In this case, the alloy contains 96.5% tin and 3.5% silver. Note that these compositions are wt%, rather than volume% and metals vary widely in density. Using values for Sn=7.31 and Ag=10.49, the percentage of the volume that is silver can be seen to be only 2.5%. [Data for other elements can be found at WebElements]

 

Tin-silver-copper

The tin-silver-copper (SnAgCu, hence SAC) ternary eutectic was developed as an improvement over the basic tin silver alloy. The melting temperature is rather lower; a number of different compositions had been claimed to be eutectic or near eutectic in the region 217°C–219°C, whilst work by Handwerker and her colleagues has since shown that the true eutectic is Sn3.5Ag0.9Cu, which melts at 217°C.

Figure 1: Ternary liquidus surface for the Sn-Ag-Cu system

Ternary liquidus surface for the Sn-Ag-Cu systemHandwerker 2002

The shaded area in Figure 1 shows that a range of different compositions near the eutectic have a liquidus temperature that is under 10°C higher than the eutectic temperature.

Key information

From time to time in any discussion on solder materials, you will come across the idea of the melting range, often in association with a phase diagram. As a reminder of what this is all about, you might like to revisit this link, which is a reminder of how metals cool, changing from liquid to solidus in the process, and how this information is presented in a phase diagram.

Of course a phase diagram shows much more than the compositions and temperatures at which materials are solid, liquid, or a mixture of the two phases; it is also an indication of when different solid phases of the same material are stable (like the transition to α-tin that takes place at −13°C in the absence of any inhibitors), and also shows the limit of solid solubility of one element in another at a given temperature – recall the description given of the solidification process of eutectic tin-lead solder, and the formation of lamellae containing alternately lead-rich and tin-rich phases.

The situation is considerable more complicated when there are a number of solid phases, each with a different crystal structure, as is the case with other binary (2-element) systems, but at least the phase diagram look comparable. What becomes much more difficult is to show the same information when the alloy is of more than two constituents. If we look at the phase diagram of SAC, for example, we find that it looks something like Figure 1 above.

We are not trying to turn you into metallurgists, and this area is extremely complex. If you want to follow it further, then there is additional explanation of the meaning of phase diagrams at this link.

The Sn3.8Ag0.7Cu alloy was recommended for general purpose by the Brite-Euram project as performing better in terms of reliability and solderability than tin silver and tin copper. Brite-Euram researchers also recommended the addition of antimony (0.5%), particularly for wave soldering, as this strengthens the alloy and further increases reliability.

If you browse manufacturer’s web sites, you will find that most offer a number of different compositions, partly to meet specific requirements, partly to deal with patent issues, and partly the result of history!

Castin™ is a patented alloy that melts in the range 217–220°C, and has a broadly similar composition (Sn2.5Ag0.8Cu0.5Sb)2. Unfortunately, some sources report concern about antimony on the grounds of its toxicity, but as you will see from other reports on the AIM site, the issue is far from clear-cut.

2 A full report on this is downloadable at http://www.aimsolder.com.au/pdf_misc/AIM CASTIN Booklet.pdf

 

Suggested reading


A comparison of tin-silver-copper lead-free solder alloys
by Karl Seelig and David Suraski

 

[back to top]


Materials with lower melting temperatures

All these materials have melting temperatures 30–40°C higher than eutectic tin-lead. Fortunately, the general conclusion from the IDEALS project3 was that, both for wave soldering and reflow, the temperature does not have to be increased pro rata to the solder melting point, although the increased thermal input still has to be supplied by combining higher temperatures with longer process times.

3 Information on this project, and a number of other interesting articles on lead-free issues are available at http://www.alphametals.com/products/lead_free/tech_art.html.

 

In order to achieve lower melting temperatures, it is necessary to look at more complex alloys, or alloys that include bismuth. Alloys based on tin-silver-bismuth, with additions of copper or germanium, melt in the range 200–210°C. Adding these extra elements improves strength and flexibility. Their solderability has been reported the best of the range of lead-free alloys, but there are two concerns:

Alloys of tin, zinc and bismuth can be made with melting temperatures down to 190°C, very close to the tin-lead eutectic, but the reactivity of the zinc causes a number of problems during manufacture (excessive dross; needs active fluxes; short paste shelf life). As well as the more serious potential for corroding during life. Widespread use of zinc alloys is confined to Japan and to products not requiring the highest reliability.

[back to top]


Solders for wave soldering

For single-sided wave soldering, the principal problem of using silver-containing solders is one of cost. There is, however, a tin-copper eutectic (Sn0.7Cu) that melts at 227ºC and is fairly close in price to tin-lead. The conclusions from the IDEALS project were that “the melting point of SnCu is too high for wave soldering at acceptable temperature levels”, and Biglari and Oddy recommended using eutectic Sn3.8Ag0.7Cu0.25Sb (SACS) for general-purpose wave soldering, and non-eutectic Sn4Bi1Ag2Sb (SBAS). It was also believed that the mechanical properties of Sn0.7Cu were comparatively poor4.

4 Although tests by Nortel Networks indicated that tin-copper performed better than tin-lead under thermal fatigue testing, Biglari and Oddy found that fatigue life for this alloy was 50–60% of that for Sn40Pb.

 

However, it has been reported that Matsushita have produced several million VCRs with a paper-phenolic board wave-soldered with the nickel-stabilised Sn0.7Cu lead-free alloy patented by Nihon Superior Co. The nickel addition has the effect of making it possible to achieve bridge-free wave soldering with a solder bath temperature of around 255°C, which is within the range that paper-phenolic can handle for the 3–4 seconds it takes for any part of the board to pass through chip and laminar waves. As we will see in Unit 7, this material has also been used in both horizontal and vertical systems to produce a lead-free HASL finish.

Clearly, whilst the situation for reflow is fairly well researched, you will have to take account of different process preferences among your assemblers when you move towards lead-free assembly.

Key information

Any solder formulation is subject to some variation in composition because of manufacturing tolerances. These reflect the purity of the starting materials and any contamination in them, as well as the actual formulation process. The situation is obviously more complex when solders contain three, four or even more constituents, or if special processes such as doping and grain-refining are applied.

One of the sources of confusion is that many different alloys have been developed for specialist applications, and there are differences in results from company to company. Summary data on the main materials is provided at this NPL location, but the National Institute of Standards & Technology and Colorado School of Mines maintain a more up-to-date and far more comprehensive database of solder properties.

[Health warning: whilst such basic information lies behind the properties of the alloys you will actually be using, delving too deep might confuse!]

[back to top]


The material in the joint

The formulation of the alloy used in the solder is not the same as the composition of the joint, because the solder reacts with the joint materials (Figure 2). Typically at least some copper from the board foil will dissolve, and any fusible surface finish will disperse in the joint to a greater or lesser extent. So the joint is both contaminated and potentially not homogenous.

Figure 2: Where the lead comes from in a typical solder joint

source: Ken Snowden (STC Optronics)

A further complication is added if the product is wave-soldered, because the bulk solder in the pot tends to pick up contaminants, dissolved from the board passing through, and also changes in composition as dross is removed. Tin being more reactive than other solder constituents, the dross tends to contain higher levels of tin oxide; when the bath is replenished, the composition of the solder added may need to be different from the initial charge in order to restore the intended composition. For example, with a tin-lead bath, additions of pure tin need to be made from time to time, rather than simply adding more eutectic solder. If the tin concentration is allowed to reduce, this produces an alloy with an extended pasty range and inferior performance.

More serious is the pick-up of copper in the solder. Common practice with tin-lead alloys has been to replace solder when the percentage of copper increased to around 0.3%, at which level the joint appearance began to deteriorate; with Sn0.7Cu, adding the same extra copper has the more substantial effect of increasing the melting temperature, as shown by the ‘copper curve’ in Figure 3.

Figure 3: Melting point of Sn-Cu alloys

Melting point of Sn-Cu alloyssource: Martin Allison (Senju)

[back to top]


Alloy variants and patent issues

As is inevitable with such developments, some materials are subject to patent and licence restrictions. It has been mentioned that the joint composition may not be that of the initial solder. Yet a number of patents refer to the joining material, rather than the initial solder used to create the joint. It is for this reason that one can potentially start with material that appear not to infringe any patents, yet end up with joints that might be claimed to infringe the patents! So, what are the patent issues?

Web research

Which solder compositions are subject to patent restrictions?

What impact will this have on a company wanting to make lead-free assemblies?

Suggestion: try "lead-free solder" "patent issues"

show solution

[back to top]


Solder costs

Two factors in the selection of materials for solders are cost and availability, these being closely related. It has already been mentioned that indium alloys, although used for specialist applications needing a low melting point material, have generally been avoided because of high cost and restricted global supply. Even the most common lead-free alloys, those based on tin, have always been expected to be somewhat more expensive than their lead-containing counterparts, simply because lead is a very cheap metal and tin comparatively expensive. However, the level of extra cost is likely to be substantially higher than originally anticipated. To find out why this might be, and the full extent of the price hike, carry out this next exercise.

Exercise

Based on today’s metal prices, how would you expect the price of a typical SAC alloy bar to compare with that for Sn37Pb solder?

And what would be the approximate relative prices of solder pastes made of the two materials?

Also what are the risks for substantial change between now and 2006?

Resource Link(s)

Interview with
Alan Rae (1999)

Ken Snowden presentation on Lead-Free Solder Materials (2004)

How does slide 9 stack up with today’s metal prices?

There are many sources of metal prices, usually differentiating base metals from precious metals. We liked the information and links at the Bankschalter site.

Try browsing for paste and solder costs at any of the suppliers. There is a useful list on the Soldertec links page.

Steve Dowds presentation at Nepcon (2004)

The focus is on testability, but the early slides on material costs are enlightening!

 

show solution

Additional costs are not the whole story – “after all, it’s only money” – but there may also be adverse environmental consequences of the new materials. Surprisingly, this is not something which has attracted much attention in the last 5 years, perhaps because some of the earlier papers on the subject were tainted by being associated with the negative reaction to the pressures to implement lead-free.

There are concerns about the toxicity of the replacement materials, silver, antimony and copper being potentially harmful to human health, and indium and bismuth compounds being suggested as having similarly harmful effects, based on experiments with rats and mice.

There is also the issue of the environmental impact – it seems surprising that there has been little follow-up to Erik Kluizenaar’s comments on IPC's LeadFree Net in December 1999 that “a proper life cycle analysis of tin-lead solders and the lead-free alternatives has not yet been carried out”. He points out that there are a number of stages in the life cycle of the materials from the mining of the metals through to the disposal of the waste. The impact of the mining of materials such as tin, copper and iron is less than for zinc and lead, but some noble metals have a huge environmental impact. Similarly with waste, where the degree of hazard is much higher for silver than with other materials because of its effect on aqueous eco-systems. There is also a different environmental impact during the soldering phase, where lead-free materials require the expenditure of more energy. Although this aspect currently has a low priority, there is no doubt that the parallel activity under the WEEE Directive to take back and recycle materials makes good environmental sense in the longer term.

[back to top]


Some alternatives

Given that silver, bismuth and antimony have all been slated for adverse health hazards, are there any systems that contain none of these elements? Jenny Hwang promotes combinations of tin, copper, gallium and indium. Metallurgical reactions between the minor elements and the tin are considered to be key in determining the melting temperature and the way that the silver solidifies, and this latter determines the mechanical properties.

Quote

Jennie Hwang

“The 13-year steady, sustained study [reported in her book, Environmentally-friendly Electronics: Lead- free Technology] of lead-free solder alloys has revealed that the ternary alloys cannot reach below 215ºC, and only quaternary alloys can reach below 215ºC.”

Surface Mount Technology, November 2003

Specific compositions (all patented) contain 92.8–93.0% tin, 0.5%–0.7% copper, in combination with the optimal 6% indium and 0.5% of gallium to enhance fatigue resistance. All these compositions have a narrow pasty range (5°C) and solidus temperatures around 209°C. the mechanical performance, in terms of fatigue resistance, is impressive being stronger and having an extended fatigue life. However, one has to be careful about which comparisons are being made.

The melting temperature of the SAC alloy, at around 217°C is still regarded by many users as being too high, particularly if there are sensitive components. Can one add anything to the system to depress the melting temperature? After all, we know that adding silver reduced the melting point of tin, and adding copper further depresses the melting temperature. Bismuth is an attractive candidate, despite the reservations mentioned earlier, and a number of SAC + bismuth alloys have been patented. Jenny Hwang suggests that the optimal composition is 93.3 Sn, 3.1Ag, 3.1 Bi, 0.5Cu, as this has the finest microstructure. Its melting range is 209°C to 212°C, so it has both a small pasty range and a liquidus temperature 5/9° lower than SAC and SA eutectics.

Bismuth is not the only element that can be added to depress the melting point of SAC solder. For example, every percent of indium up to around 12% suppresses the melting temperature by about 1.8°C. As examples, Sn4.1Ag0.5Cu8In melts at 195°–201°C, and Sn4.1Ag0.5Cu12In melts at 185–195°C. As with the bismuth alloys, there are patent implications.

An advantage of adding indium is to enhance the fatigue life, and Sn3Ag0.5Cu8In is described as offering the best balance of properties with high fatigue resistance. Its melting temperatures (196–202°C) have a narrow pasty range and acceptable wetting characteristics. But of course the material is expensive. There is also some concern that, with a high indium content, there is the possibility for the growth of a tin-indium binary eutectic that melts at 117°C, a similar problem to that afflicting bismuth alloys.

Ignoring concerns about reliability, which may be more or less true, the major reason affecting the uptake of these more complex alloys is actually the substantially higher cost of the raw materials, and the patent situation surrounding them. Nevertheless, attempts to make a ‘next generation’ alloy continue to absorb substantial coverage in the media. Perhaps more promising in the longer term will be the work being carried out by EFSOT (Next Generation Environment-Friendly Soldering Technologies), an IMS project with participation from Japan, Korea and the European Union, which is believed to be focusing on improvements to the way in which the materials are formulated and combined, and to their resulting microstructure, rather than on changing the basic alloys themselves.

Solders for other applications

High-melting alloys are important within the electronics industry, both for die attach and solder balls and columns for BGA devices; in the former case, solder is used instead of silver-filled epoxy. It has thermal and electrical conductivity an order of magnitude higher; in the latter case to prevent the collapse of the ball/column.

The most commonly used alloys have been lead-based with the addition of 10% silver (melting point 302°C) or 5% tin (melting point 320°C). The only lead-free material with any track record is the so called ‘J alloy’ (Sn25Ag10Sb). This has a melting point of 365°C, but is not ductile and therefore prone to thermal fatigue failure. J alloy was introduced as a cheaper alternative to eutectic bonding of a silicon die to a gold-plated header, which produces a silicon-gold eutectic melting at 379°C that acts as a solder, but requires expensive heavy gold plating. Although using sophisticated techniques of manufacture to ensure a fine grain structure, the results from this alloy are disappointing.

Lack of any real solution for high-temperature solders are closely intertwined with ongoing discussions about exemptions to the RoHS Directive for military and other critical applications, and to the current exemption of power devices.

[back to top]


Keeping up-to-date

Keeping abreast with developments in soldering is quite challenging, and you have to be aware that different workers report from different viewpoints and have different agendas. It is constructive to search for papers on lead-free soldering in combination with selected names.

In the next activity we suggest some key workers whose papers have been influential. Before proceeding to Unit 5, take a good look at their subjects of interest, browse a few of the articles, and see if you can reach any conclusions as to the ‘spin’ on the subject. In our comments we have identified some of the seminal papers that may perhaps give you information at the right level, without too much metallurgy!

Web research

Try a search for "lead-free" in combination with some of the key workers in this field:

"Tom Baggio"
"Peter Biocca"
"Carol Handwerker"
"Jennie Hwang"
"Ning-Cheng Lee"

The list is alphabetical, but also in increasing numbers of hits. You might have to be more selective when looking at the last worker’s papers!

show comment

It goes without saying that other good ways of keeping up-to-date are to read magazines such as Advanced Packaging, Circuits Assembly and Printed Circuit Design and Manufacture (there is a more complete list at this link) and to go to seminars such as those run by the SMART Group.