Re: [ATM] Dense

[Matt Considine <matt@considine.net>]

I spent some time last summer making some speculum disks in a propane fired
kiln, having gone through Wm Herschel's speculum-construction notes.  I had also
gathered as many descriptions of mixes as I could find.

The upshot seemed to be that you really need to hit the 3:1 (atoms:atoms) ratio
pretty closely.  Also, I suspect that the heating and annealing curves need to
be monitored, judging from a phase diagram.  Otherwise, I ended up with some
intermetallic that fairly quickly turned brassy colored.  Also, it does not
behave on a metal lathe "nicely".  Trying to face the disk on the lathe didn't
result in curling metal - it resulted in very brittle pieces coming off.

Weather and other events intervened to prevent me from going further at that
time, but I hope to resurrect that.  A guy in VT sells small disks, by the way :

www.classicscience.com/speculum/speculum.html

Matt

Quoting Anthony Stillman <atmer@flash.net>:

>   Two thirds copper, one third tin.  What is it?
>    
>   Some time ago in some personal communication the topic of speculum's
> density arouse.  Two years of making ingots and I still hadn't gotten to
> that.  I put a measuring cup on the dinning room table and with some tap
> water and half kilograms of as marked alloy, I arrived at a value startlingly
> close to Herschel’s, 8.9 grams per cubic centimeter.  Then I worked out the
> disappointing uncertainty, +-0.4  
>     
>   Next I embarked on an extended effort to weigh on a triple beam balance
> many individual hundred gram billets in air and while buoyed by water.  I
> made a LOT of measurements.  The calculated densities ranged from 8.77 to
> 8.95+-0.04 g/cm^3.  The rounded average, 8.9.
>    
>   Way back when I began this metal mirror making exercise, I wanted to see
> what methods I might devise on my own.  I have abandoned this approach.  What
> follows are a few highlights of my investigation:
>    
>   There is renewed industrial interest in the Cu-Sn system.  This due in part
> to the demand for ruggedized lead free electronics and the need to understand
> and control the brittle intermetallics(1) that form where copper meets
> solder.  
>    
>   In the 45 year old text "Tin and Its Alloys," bulk speculum is identified
> as Cu3Sn, the epsilon phase in the copper tin system.  In that same book,
> electroplated speculum, used as a hard decorative tarnish resistant coating
> up to the 1960s, is identified as Cu6Sn5, the eta phase. Eta and epsilon are
> both intermetallic compounds.
>    
>   NIST has recently determined the bulk properties of Cu3Sn and Cu6Sn5.  As
> casting could not create sufficiently pure intermetallics samples, they
> devised a means to create pure powders and then heated and compressed them
> into billets of maximum density.  Their results show these species have
> densities of 8.9 and 8.3. (2)  
>    
>   Two other compositions of interest in speculum are the intermetallic delta
> phase(3) and alpha bronze.  Alpha bronze is a solid solution(1), not an
> intermetallic.  It occurs in great quantity in alloy containing less than 14
> percent tin.  Bronze(4) used in casting sculptures is almost entirely the
> alpha phase.
>    
>   Speculum mirror makers from 300 odd years ago note the difficulties
> associated with surface defects (pits) encountered during polishing.  My
> personal experience confirms the pit nightmare.
>    
>   Porosity lowers density, leaves pits, and weakens castings.  Porosity in
> bronze occurs by four means.  Gas entrapment, gas evolution, coring, and
> condensation of Kirkendall voids.  
>    
>   Coring(5) and Kirkendall voids(6) are of little consequence in cast bronze
> containing more than 25% tin.
>    
>   Gas entrapment occurs during melting and pouring.  The latter particularly
> so if it is turbulent. 
>    
>   Gas evolution may result from contamination in the melt and/or from the
> mold dressing.  Of principal interest is water or water vapor in contact with
> the liquid metal.  The water disassociates into its constituents which either
> dissolve into, react with, or escape the melt.  This is called the inverse
> steam reaction.  When bronze solidifies, dissolved hydrogen is rejected. 
> This reacts with dissolved oxygen and oxides to form steam.  The steam then
> creates pressure voids.
>    
>   Stop casting porosity.
>    
>   Hydrogen may be removed from liquid metal by oxidizing it.  This is called
> degassing and may be done using an oxidizing flux.  Flux is used to cover the
> melt and more may be mixed in prior to casting.(7)
>    
>   After degassing the melt must be deoxidized.  Historically, this was done
> by stirring in a small amount of zinc just prior to casting.  Aluminum will
> also work.  However, both these metal's oxides are solid and become
> contaminates.  Alternatively, phosphorus forms oxides which are fluid and
> float on the molten metal.  The down side of using phosphorus,(8). 
>    
>   Early telescope mirror makers added one half to two percent of arsenic
> after melting the speculum ingots.  The resulting cast metal was described as
> denser and "whiter".  Later workers omitted the arsenic.  
>    
>   The addition of arsenic produces a marked increase in the amount of
> alpha+delta constituent.  This hardens bronze and in chill cast degassed
> deoxidized speculum, lowers the density.  Arsenic concentrations in excess of
> one percent do not significantly increase these effects.  Arsenic, like zinc
> and aluminum, though not nearly so effectively, will function as a
> deoxident.
>    
>   Porosity is not the only source of pitting in speculum.  Arsenic is not the
> only adulterant capable of increasing the amount of delta intermetallic
> present in a casting. 
>    
>   Speculum is not a pure intermetallic, but a combination of delta, epsilon,
> and eta along with alpha solid solution.  How much of each and where it is
> found determines the structure and properties of the metal.  This depends
> somewhat on the constituent ratios and greatly on the thermal history.
>    
>   At this point I've done a fair amount of research and am under the illusion
> that I know enough to give a talk.  I've offered to do so at this year's
> RTMC.  I'm still waiting to learn if I will be given the opportunity.  If
> anyone would like to lobby for the talk, please write the RTMC program
> director. 
>    
>   Anthony
>    
>   alea iacta est
>   Literally: The die is cast.  (That’s die as in the singular of dice.)
>    
>    
>   Footnotes:
>   1) A solid solution is a single phase solid formed from a liquid with the
> same average composition.  One metal has dissolved into the other and
> together they have solidified.  Supersaturation occurs when there is more
> metal in solution than there can be for that temperature.  This results in
> the precipitation of intermetallic compounds.  Solid solutions generally have
> the same crystalline structure as the element present in the greatest amount.
>  An intermetallic compound has a crystalline structure different from its
> constituents.  Intermetallics form when the bonding strength of unlike atoms
> is greater than that of like atoms.
>    
>   2) The uncertainty in the NIST measurements is +-0.02g/cm^3.  It is
> reported without a value for uncertainty that Hershel measured a density of
> 8.8 for "ideal" speculum.  The meaningfulness of that number is blurred by
> unknown porosity and trace adulterants.
>    
>   3) Until about 1930 the delta phase was thought to be Cu4Sn.  Strong and
> others believed this to be speculum.  Subsequently delta was determined to be
> Cu31Sn8.  This identification persists in published work to this day. 
> However, currently NIST identifies delta as Cu41Sn11.  Delta to some degree
> is present in all cast bronze, often at the grain boundaries.  Delta is
> metastable below 348C.  To exist in bulk at room temperature this
> intermetallic must be chill cast down to about 200C.  Though extremely slow,
> measurable changes in the composition of high tin bronze continue to occur
> even below 200C.
>    
>   4) Much modern bronze is in fact not bronze at all but brass, a mixture of
> copper and zinc.  Bronze is a difficult metal to cast, work, polish, and
> generally everything else.  Brass can be made to look like bronze and is
> comparatively simple to manipulate.
>    
>   5) Coring occurs as a result of non-equilibrium conditions which exist at
> the freeze boundary.  The principle is used to good effect in a process known
> as zone refining. 
>    
>   6) Cast bronze bearings (high tin bronze) subject to prolong exposure to
> temperatures above 250C will develop Kirkendall voids.
>    
>   7) One such flux contains equal parts of silica (dry sand), borax, and CuO.
>  Flux is expensive and production casting can consume large quantities.  A
> vacuum or electric furnace will keep the hydrogen content of the melt low.  A
> draft, oil or gas furnace are successively less desirable.
>    
>   8) Care must be taken to insure thorough mixing.  Residual un-oxidized
> phosphorus exceeding 0.02 percent can significantly alter the properties of
> bronze.  Elemental phosphorus is particularly difficult to handle.  Red
> phosphorus powder compounded with zinc powder is an alternate de-oxidant.
>    
>   References:
>   "The History of The Telescopes"  King
>   "Reflecting Telescope Optics I"  Wilson
>   "Reflecting Telescope Optics II"  Wilson
>   "Amateur Telescope Making - Book One"  Ingalls
>   "Amateur Telescope Making - Advanced"  Ingalls
>   "Prism and Lens Making"  Twyman
>   "Procedures in Experimental Physics"  Strong
>   "Tin and Its Alloys"  Hedges
>   "Chill Cast Bronze"  Hanson and Pell-Walpole
>   "Structure of Metals through Optical Microscopy"  Tomer
>   "Principles of Solidification"  Chalmers
>   "Intermetallics"  Sauthoff
>   "CRC Handbook of Chemistry and Physics"  Weast
>   http://www.metallurgy.nist.gov/mechanical_properties/solder_paper.html
> 
> _______________________________________________
> ATM mailing list http://www.atmlist.net/
> 




_______________________________________________
ATM mailing list http://www.atmlist.net/