(posting on behalf of Gordon Davy)
 
Bob, Denny, or Mike,  since you're subscribed to TechNet, feel free to post this reaction to Zinn's  abstract (and bio). Perhaps if he sees it, he'll be able to respond before his  June 12 presentation.

Gordon Davy
Peoria, AZ

I don't want to deprecate Zinn's work. Learning how to  make copper nanoparticles and keep them from oxidizing or agglomerating is  difficult. But I am concerned with his claim that reducing the copper particle  size reduces the melting temperature to 200°C. It is true that surface  atoms are not as tightly bonded to their neighbors as bulk atoms, so reducing  a substance's particle  size, by increasing the surface-to-volume ratio, does reduce the melting  temperature somewhat below the bulk value. That is the basis of  sintering, which occurs below the bulk melting temperature. Presumably, if the  particles were small enough, no atom would have the full number of  nearest neighbors – it would be all "surface." But the melting temperature for bulk copper is just  short of 1100°C. That's a stretch! 

The particles of  copper in powders sold for sintering are mostly smaller than 44 µm (-325  mesh). The process requires  compacting the powder (plus lubricant) at 4,000 to 8,000 atmospheres pressure,  then and heating at 750-900°C for 5-7 minutes (see  copper-powders.com).

Dr. Zinn's copper is  in the form of "nano-particles,"  so they are presumably smaller than  1 µm. One can  contemplate whether that  roughly 1½ to 2 orders of magnitude size reduction is  sufficient to account for the differences between the above conditions necessary  for sintering and for conventional reflow soldering. A differential  thermal analysis curve would support the  claim that, with or without compaction, his nano-copper  "solder" melts at 200°C. 

Consider this  analysis taken from everyday observation. When snowflakes, which may have  nano-scale features, fall and land, they sometimes sinter, and sometimes (when  temperatures remain below about -15°C) they do not. People refer to that latter  kind, once it has landed, as "powder" snow. Similarly, one cannot skate on very  cold ice because the surface lacks the "liquid-like" layer to lubricate the  blade. (Pressure melting is a minor factor.) 

So  if dropping the temperature of ice by fifteen degrees below its melting  temperature prevents sintering and skating, is it likely that copper sintering  will occur at a temperature nearly nine hundred degrees below its melting  temperature? As for melting,  even if individual particles were to melt at such  a low temperature and join to form a liquid, what would prevent the liquid, now  with dimensions measuring from micrometers to millimeters, from instantly  freezing?

More likely, the copper particles dissolve into the tin or tin-lead plating  on the board lands and component terminations (to form bronze, if the reaction  goes to completion). Even if the nanoparticles are not melting or sintering,  it's a clever idea, but we need to know: 
  * How well these bronze connections, presumably far stiffer than those of a    tin-based solder, survive temperature cycling.    * The microstructure, so we can understand the attachment mechanism.   * The width of the process window.    * Whether the adequacy of the attachment can be judged by its appearance,    and if so, by what criteria.   * Whether using Zinn's solder at 200°C gives benefits large enough to    warrant replacing Pb-free reflow soldering (peak local temperature up to 260°C    by convection, lower by condensation). (For those still using SnPb solder, the    conditions don't seem different enough to warrant consideration.)   * Whether Zinn has tried his "solder"    with non-tin land finishes such as immersion silver, ENIG,    and ENEPIG, and NiPdAu termination finish. Since I suspect that the colloidal copper does    not melt    or sinter at 200°C, I suspect that    it is not going to perform with Ag    or Pd nearly as nicely as with near-molten Sn, and its reaction with Ni would    be even worse. 
  
    * If it doesn't make reliable bonds with      all the finishes likely to be present on an assembly, then      it is not a "drop-in replacement"      for solder. (Yes, the designer can specify a compatible finish for the lands      of the board he designs, but not the termination finish of the      components he chooses. He can specify, say, immersion tin instead of      ENIG, but must accept NiPdAu on some components.)     * Maybe Zinn can add enough nanoparticle Sn      to the formula to provide the necessary wetting to Pd without increasing the      reflow temperature and without introducing a risk of whiskers from the      solder itself.     * The risk of short circuits due to whiskers      growing SAC solder appears to be low, from SnPb solder even lower, and with      Zinn's solder it may be zero. But regardless of the solder used, the      overall risk for an assembly remains high due to      portions of Pb-free Sn termination finish that don't get      solder-coated.
Here are two other ways of dealing with the risk of tin  whiskers:

  1. Mitigation – speculative For  assemblies for which Pb is permitted, if components with a Pb-free Sn tin finish  were dipped in SnPb solder paste (or perhaps a paste of colloidal Sn and Pb) and  heated (before or during assembly soldering), all of the original finish  might be covered with a layer of SnPb, and the assembly would then have a  low risk of short circuits due to tin whiskers. The paste also might  bridge.

  2. Prevention – reduced to practice After  attaching components by conventional SnPb or SAC soldering, a thin  layer of a whisker-impenetrable metal such as nickel can be applied  to the solder and the remaining (uncoated)  Pb-free Sn finish by electroless  deposition. The process requires immersing the entire assembly in  the bath for a minute or two. But because  electroless deposition occurs only on conducting surfaces, insulating surfaces  remain uncoated, and the performance of the assembly is unaffected. (Full disclosure: Bob, Denny, and I are the named  inventors on a patent application for this process – see www.ldfcoatings.com.) 

A few additional  observations:
  * Note the irony, given the meaning of Zinn in his    mother tongue, of him developing a    tin-free solder.  * Note the    nine-year gap (between 1995 and 2004) in his bio.    * The claim of "10-15x    electrical and thermal conductivity improvements" is irrelevant:    what counts is overall conductance, and the reduction in a connection's    conductance due to use of conventional solder (compared to copper) is    insignificant.  * A substance's melting    temperature is a thermodynamic property. Some might    regard the reduction in the melting temperature of colloidal copper as    "significant" or even "dramatic." But it is not "rapid" (a term    that implies a kinetic effect).

______________________________________________________________________
This email has been scanned by the Symantec Email Security.cloud service.
For more information please contact helpdesk at x2960 or [log in to unmask] 
______________________________________________________________________