Jim Hisert’s Tech Support Blog
Recent Entries
-
Ultra Low Res Flip Chip Flux
This is just an ultra low resolution picture of flip chip attachment flux, however ultra low residue flip chip attachment flux is clearly a great solution for no-clean applications which involve subsequent capillary underfilling (pun intended).
There are low residue fluxes designed for dipping, pin transfer, printing, spinning, or spraying. But low residue fluxes do have a unique limitation. Because the solids content of the flux is low and it offers less protection from oxygen at high temperatures, low residue fluxes generally need to be used in an inert atmosphere. And now the good news: most of us are already using <50ppm O2 in our flip chip processing furnaces anyway. It looks like in this case less really is more.
-
What’s Acceptable?
Unacceptable voiding under a die
With less voiding under a power die, more heat can be dissipated through the solder into the substrate, heatsink or lead-frame. Decreasing the amount of voids under a die is not only suggested for maximum solder joint strength, it is mandatory for sufficient thermal transfer. There are several tricks to lower voiding, especially when using a solder paste. One of the most pronounced ways to lower voiding is though iterative profiling.
But what is acceptable voiding? Some die-attach processes limit voiding to 25% of the solder area under the die – although many people accept nothing more than 10%. With proper profiling and a good solder paste, voiding can be limited to <5% in almost any application.
-
Flat Sectioning
First flat section (just below the solder, looking up through the copper traces)
Intermediate flat section (notice solder joint geometry)
Last flat section (near the component, looking at the bottom of the part)
Last flat section (near the component, looking at the bottom of the part)
Most of us think instantly of cross-sectioning when sample examination is mentioned. ‘Flat’ or ‘lap’ sectioning can be a great alternative in certain situations, and is usually a good supplement to cross-section views.
Flat or lap sections are essentially cross-sections that are polished in the same plane as a BGA or die, and are used to view all interconnections in relationship to each-other. As shown in the picture here, a flat section can also reveal oddly shaped solder joints that may have gone unnoticed otherwise. I like to use this technique when I suspect problems with solder mask openings. Often, you can clearly see when solder wets a trace instead of being contained on a pad properly. Flat sectioning is a notable technique, and perhaps another trick you can use to learn more about solder joint geometry in your own application.
-
Why Color A Flux?
Some flux height sensing equipment for flux reservoirs require a dye to be added in order to discern the level of a given flux pool. Colored flux can also help with flux deposit inspection, since clear flux deposits are invisible to most optical inspection equipment. These are both great reasons to use a flux that has a distinct color, but my favorite reason seems at first to be a disadvantage to a dyed material.
When you clean up after using a dyed flux, most people mention that it ‘gets all over’ the equipment, and cleaning takes a bit longer. The cleanup is not actually any more difficult, a flux isn’t going to magically migrate to new parts of your machine – you just notice it more than a clear flux (which makes you wonder if you were really cleaning your equipment effectively before). So the real advantage of using a colored flux may be that it teaches your operators how to completely clean flux from equipment – a skill that I think many technicians can improve upon.
-
Flashback - Contamination
I think the most useful thing I learned from chemistry in college was cleanliness. The first day of chem. lab we were asked to fill our beakers with a clear chemical (I forget what it was) and then rinse each beaker until we thought it was clean. Some of us flushed our beakers 5 times or more – as expected most of the students rinsed only twice. The impromptu test was a drop of liquid administered by the professor that turned red in the presence of the material we were cleaning from the beakers. I have never forgot that lesson, or the threat of contamination in advanced materials testing.
Tomorrow I’ll explain what reminded me of this.
-
Communication
As an engineer in the semiconductor industry, there is always the possibility for you to travel to another country to troubleshoot a problem on a line, participate in a training program, attend a trade show, or even help start-up a new office. I’ve found that a small translation guide is one of the best investments you can make on your next visit to a country where your first language isn’t the common language.
For around seven dollars and an open mind, you can keep yourself entertained on a flight, pick up a few phrases to get you by when you arrive at said country, and possibly inspire you to take on the task (gulp!) of learning a new language. I found the little book pictured above. It fits in a pocket easily, and it has all the essential phrases in the first few pages. If you know someone who speaks the language well, I recommend trying your newfound skills out on them first. They can tell you if you are making a rooky mistake, and they can explain the style of the language much better than a book can.
This all pays off the first time you are able to greet someone in their native language, thank someone that expects you to only know English, or successfully carry out a short interaction (like purchasing something at a convenience store without using any English). You will also truly respect how much work your overseas partners do to be able to communicate with you in your language!
-
Intermetallic Growth
Time is precious, and some alloys take their precious time to form intermetallics. For lower melting point alloys, quick wetting doesn’t mean an appropriate amount of intermetallics have formed to bond the solder optimally. Time above liquidus may need to be extended as long as 5 minutes to allow very low temperature alloys a chance to dissolve a surface finish and form intermetallics.
-
The Infamous Cell Phone
Unquestionably one of the largest technology drivers in the last decade, the cell phone continues to challenge our industry by never being quite good enough to leave well enough alone. My two cents: why not include a small LED flashlight? It could be used for finding your way in the dark and it might be able to double as a flash for that digital camera we have integrated into the modern cell phone.
-
How Low Can You Go?
Due to the limitations of flux, there is bottom end temperature where traditional solder/flux bonding ends. The lowest temperature I have experienced is 100 deg C (peak). This was only achieved with a certain flux / alloy / surface finish / temperature / time combination.
If your question is “how low of a temperature can achieve a solder bond”, there you go – kind of… The exception is cold welding at room temperature. This is achieved with pure indium.
-
Why Solder To Steel When You Can Solder To Copper?
Okay, this has no use for semiconductor packaging that I know of, but it is a neat trick that I discovered in the lab a couple days ago. If you clean a piece of steel with HBF4 that has copper dissolved in it, the copper will lightly ‘plate’ onto the steel and passivate it. Copper is much easier to solder to, and typical fluxes should work (otherwise you need a very specialized iron flux). This opens up a world of possibilities for those rusty old steel surfaces.
-
The Speed of Simplicity
With only 24 hours to answer a customer regarding a strange flux / alloy / surface finish / temperature compatibility, we had no time to have an exotic alloy made into precision spheres. Instead of admitting defeat, we found an ingot of the alloy and shaved small chunks off of it. The chunks of alloy were enough for “proof of concept” and gave us confidence and a clear direction to go for further testing.
-
Corrosion – Part 4, Learn More
Here is a pretty good list of resources on the topic if you’re interested:
1. Chongchen Xu. “Corrosion in Microelectronics“ 2003
2. Charles A. Harper. Electronic Packaging and Interconnection Handbook. Third
Edition. McGRAW-HILL, 2000.
3. Electronic materials handbook, vol.1, ASM international, 1989.
4. A.J. Griffin, S.E. Hernandez and F.R. Brotzen. “A galvanic series for thin-film
metallizations and barrier layers commonly used in the microelectronics
industry”, Journal of electrochemistry society, 141(3), 807-809 (1994).
5. Deny A. Jones. “Principles and prevention of corrosion”. Prentice Hall, Upper
Saddle River, second edition, 1996.
6. W.M. Paulson and R.P. Lorigan. “The effect of impurities on the corrosion of
aluminum metallization”, Proceedings of the IEEE Reliability Physics
Symposium, IEEE, 42-47 (1976).
7. Andrew H. Rawwicz. “Stress induced corrosion of wire micro-joints in
microelectronics – a quantitative model”, Microelectronics and Reliability, 34(5)
875-882 (1994).
8. M. Hansen and K. Anderko. Constitution of binary alloys. Mcgraw-Hill, 1958.
22
9. W.W.Binger, E.H. Hollingsworth and D.O. Sprowls. Aluminum, vol.1, American
Society of Metals, 1967.
10. Simon Thomas and Howard M. Berg. “Micro-corrosion of Al-Au bonding pads”,
IEEE transactions on components, hybrids and manufacturing technology, 10(2),
252-257(1987).
11. Joseph Fauty, Steve Strouse, Jay Yoder, Carlos Acuna, and Phil Evard. “Al-Cu
metal bond pad corrosion during wafer saw”, The International Journal of
Microcircuits and Electronic Packaging, 24(1), 19-29(2001).
12. Fu-gin Chen and A. Jean Osteraas. “Electrochemical dendrite formation during
corrosion of connector leads”, Proceedings of ASM’s third conference on
electronic packaging: materials and processes and corrosion in microelectronics,
175-179(1987).
13. Kim A. Berry. “Corrosion resistance of military microelectronics packages at the
lead-glass interface”, Proceedings of ASM’s third conference on electronic
packaging: materials and processes and corrosion in microelectronics, 55-
61(1987).
14. M.J. Elkind and H.E. Hughes. “Prevention of Stress-Corrosion failure in ironnick-
cobalt alloy semiconductor device leads”, Proceedings of Failure in
Electronics Symposium, 5, 477-495 (1966).
15. K.W. Rosengarth. “Corrosion Protection for semiconductor packaging”, Solid
State Technology, 27, 1991-196 (1984).
16. Vlasta Brusic, Doreen D. Dimilia and Robert D. Maclnnes. “Corrosion of lead, tin
and their alloys”, Proceedings of ASM International 3rd electronic materials and
processing, 261-268(1990).
17. Michel Pecht. “A model for moisture induced corrosion failures in
microelectronic packages”, IEEE transactions on components, hybrids and
manufacturing technology, 13(2), 383-389(1990). -
Corrosion – Part 3, Sources and Prevention
Corrosion is caused by an electrochemical reaction between two dissimilar metals, one acting as an anode – one as a cathode. Halide ions such as fluoride, bromide, and chloride are the usual culprits for electrolytic mobility. Although they are commonly found in fluxes, I was surprised to find they are also used in etching, plating, sawing, and encapsulation. When used in flux, the chemistry is designed to either entrap the halide so it is not harmful, or ease its cleaning so it can be fully removed (such as in a water soluble flux). On low standoff flip chips, water soluble flux clean-ability is of utmost importance during formulation.
All hope is not lost, and you can still assure high reliability by using a no-clean flux appropriately, or using a water soluble flux and verifying your water washing process. Large chips with low standoffs (50 to 100 microns) require precision water jet setup to help the water reach the center of the chip-to-substrate interface.
-
Corrosion – Part 2, How It Affects Us
Corrosion leads to field failures. I could probably stop right there – that’s bad enough.
To flesh that out a little more, corrosion can lower thermal and electrical conductivity, cause opens, it can cause dendrites – which cause shorts, and it can reduce the solder-ability of bumps in later processes. Tomorrow we’ll discuss where the corrosion comes from, and how to prevent it.
-
Corrosion – Part 1, What It Is
This is the first of a 4 part series on corrosion to cover 1) What corrosion is, 2) How it affects us, 3) Sources of corrosion and prevention, and 4) How to learn more.
Corrosion comes in 3 basic varieties. Pitting, stress cracking, and galvanic corrosion are forms of deterioration that change the properties of a pair of metals through an electrochemical reaction. Rust can be seen all around us in everyday life, although a very similar corrosion can happen if we aren’t careful how we build semiconductor packages. Check out tomorrows post to learn what corrosion can do if left unchecked.
-
Cross Sectioning With Robert Belvick of ORS
Recently, I asked Robert Belvick from Onieda Research Services to share some of his experience cross sectioning microelectronic components.
Robert: Even if you have cross-sectioned before, there is always room for improvement. Here are some more advanced cross-sectioning tips to yield better results.
1) Don’t be afraid to experiment with different polishes or techniques. Some people have used the same thing for so long, they forget that new products are released frequently that may better suite your application.
2) Be sure to keep track of what you’re doing. It’s not uncommon to be distracted by talking to a co-worker and before you know it, you’ve gone through your area of interest!
3) Be sure to switch to fine grit paper before you reach the area of interest, or you may end up going through it trying to clean it up.
4) Don’t hesitate to change your polishing cloth. If some contaminate gets in it, it won’t come out easily. Run your clean fingers over the cloth to check for foreign material before you use it. Even a piece of debris from a sample you just polished can cause scratches in the next one.
5) Don’t over polish! You can go home and go nuts polishing your chrome rims and they’ll look great, but over polishing a cross-section is quite easy and results in “rounding”. When rounded, some parts of your image may be more in focus than others, because you’ve actually polished away part of the softer material, leaving behind the harder stuff.
6) This is more preference, but I always wear gloves while cross-sectioning. It’s not necessary and some may like to do without, but it’s far too easy to accidentally grind your fingertips if you ask me! -
Economic Stimulus
Soon, many of us in the US will be receiving a check from Uncle Sam to spend – thus stimulating the economy. While awaiting this somewhat unplanned payment, one can only wonder what to spend the money on?
What if your boss gave you a free pass to purchase one piece of new equipment for your lab or production line without needing any justification? Chances are, the equipment you’d get would make your life easier, and your time at work more fruitful. In the same way, I hope the consumable materials you choose can do the same.
-
Gage R&R – An Interview With Mario Scalzo
You may know Mario from his “Tech Support Blog” on our web site. Today we chatted about the importance of gage R&R.
Jim: We know measurements are only as good as the equipment you use to obtain them. How should our customers insure the integrity of their measurement equipment?
Mario: Through internal calibration by Quality, external calibration by an accredited lab or by the [equipment] manufacturer, testing against a material of “known” properties, and yearly confirmation by an outside lab. Also, many equipment manufacturers offer a calibration program, like a Service Contract for your automobile.Jim: Is there a set procedure for gage R&R, is there a standard that should be referenced?
Mario: GR&R is a test of your measurement system “drift” if you can think of it like that. This is done by measuring the SAME part, or set of parts, over and over again. There is no “set” procedure, but the more measurements, the more accurate the gage.
Also, there are 2 types of gage, to test the equipment and to test the process.
By having the same person measure and re-measure over and over, you are testing the machine. By having 2 or more people measure the same part, you can test the reproducibility of the process.Jim: I cringe when I see expensive equipment misused, what percentage of paste inspection systems do you estimate are properly verified?
Mario: I know what we have gone through with our own paste inspection machine, and they are pretty accurate right out of the box, and are usually set-up by the manufacturer. The problem lies with the way it’s used. Paste inspection is designed to be used to maintain control of the process. Which means that within the natural variation, some piece will “fall out”, or not pass inspection. The way that they are being used is completely different.Paste inspection is an expensive waste if:
• The process is out of control.
• The inspection limits are set too wide, commonly to allow increased through-put
• Nothing fails
• Is being used as a “gate” (Data not tracked)From what I have seen, very few companies are using it properly.
-
Open Invitation
As an open invitation, feel free to send me your thoughts / ideas / concerns for future entries. From anonymous opinions to a full scale DoE, we will field them all! Your ideas have already been helpful, so keep them coming…
Jim Hisert’s e-mail -
Package-on-Package Solder Paste Viscosity
If you are new to Package-on-Package solder paste, expect the material to be lower viscosity then you are used to for printing or even dispensing applications. The PoP dipping process demands maximum solder transfer, which is provided by a different rheology than stencil printing paste. Pictured here is a new Package-on-Package solder paste. You can actually pour it out of the jar into a flux reservoir!
-
Safety Net
It’s always good to be cautious when dealing with high-reliability devices.
No-clean ball attach and package-on-package fluxes are usually designed to reach a certain temperature before the post-reflow flux residue acts as an encasement for the unused activators. Below that temperature, most fluxes fail SIR testing due to electromigration. A unique flux (PoP Flux 30b) passes SIR in both the normal reflowed, and less common un-reflowed state. This is important to note, since the flux can be used in low temp applications with peak temperatures of only 100 degrees C. Most fluxes start working around 150 degrees C and need a peak temperature around 200 degrees C to pass IPC SIR testing.
I think I’ve found my new favorite low temperature flux.
-
What’s Your Number?
There are a lot of things to consider when choosing how many samples you need for testing. Please consider these factors when deciding your sample size for data collection.
1) Statistics – you can’t logically argue with it…
2) Available parts and materials
3) Cost of samples
4) Time allowance for the testing
5) Other peoples time (line time)
6) Industry standards
7) Your confidence level
8) Available equipment and capacity
9) Spare samples, to compensate for lost/damaged samples
10) Make sure you at very least have three data points -
Would You Trust These Parts?
After finding a batch of Package-on-Package components with a ~20% defect rate (that’s not a typo), I asked the vendor to have the lot of components replaced.
At first replacing the whole order was met with opposition, there were still some components unused that met visual inspection. But how do you use these components from a faulty lot without suspecting possible problems? Just because the components have spheres on them, it does not mean they have been properly attached.
These parts were eventually replaced, but should serve as a lesson. Do not pick out good parts from a suspect lot – it’s just not worth the questionable results you could have during testing.
-
Ninjas and Flux
Ninja are known for their ability to combine two very different properties – power and stealth. Flux was never very good at this though, we usually have to choose between halide-free fluxes and appropriately active fluxes. In the next couple months you should be on the lookout for our new ninja-like halide-free ball-attach and flip chip fluxes that are powerful in environmentally delicate applications. Oh, and like a ninja, the water soluble version will disappear without a trace…
Image courtesy of Kirkhamsebooks.com
-
Hand-Picked Designer Materials
It’s great when you find that perfect product, the product that makes life easy. Sometimes that perfect material has already been formulated – it may have been created ahead of its time!
A rich history of flux and alloy development keeps us on top of the game. As an application engineer, I constantly learn about advancements that happened long ago and never found a market. Even if these exact products never reach the semiconductor market, they provide a base of knowledge and a starting point to develop new advanced materials with a shorter cycle time, and more confidence.
Site Feeds
Subscribe to Jim Hisert’s Blog:
Site feeds allow you to “subscribe” to your favorite web sites. There’s no need to fire up your web browser just to see if a site has been updated — the RSS feed will notify you automatically.
To use a site feed, you must either have a web browser that supports RSS or Atom (such as Mozilla Firefox) or a separate “News Reader” (FeedDemon).
Once you have obtained a feed reader, consult its documentation to learn how to use it with this site’s feed.
Email Jim Hisert
Want to read more? Browse the archive of past entries.
