Re: powder packing density!!!

From: Klas Boivie <>
Date: Tue Feb 24 2004 - 13:31:05 EET

Hmmm, particle packing densities....

The articles regarding the packing of ellipsoids can be read in
"Science" vol. 303, feb. 13 2004 pp 968-969 and pp 990-993, for those interested.

The suggested explanation for the close-to-theoretical particle packing
densities is the possibility for the ellipsoid particles rotate in the x-y
plane out of positions where a sphere would be in a locked position. This
possibility for orientation comes in handy when particles first randomly
are poured in a container and then vibrated for further densification.
However, the resarch was carried out on M&M candies, which have a size
range far above what would be practical for most SFF/RP applications, and
not the least important, has a nice smooth low friction surface. (Though
one might wonder if this is an explanation why there always seems to be
room for more sweets. ;-))

Particle packing properties are not only dependent on particle shape and
size range interval, another issue of significance, especially for the
small sized particles and particles of a light weight is the inner friction
from Van der Waals and electrostatic forces in the powder mass. The effect
of these forces on each individual particle is dependent on the
surface/mass ratio, which means that the more surface you've got in your
powder the more friction and the more adverse effect to interparticle
motion and high density particle packing packing arrangement. For most
shperical metal powders this effect can start to influence powder behaviour
in the size range interval just below -45 micron, for coarser shaped
particles the effect of inner friction in the powder mass may be an issue
already at 100 micron. In the interest of precision, size ranges above 100
micron would hardly be practical for most powder based SFF/RP eqiuipment.
The smooth ellipsoid shape, which apparently packs densely in the form of
candies, has a larger surface/volume ratio than the spherical shape and
could therefore be expected to be more sensitive of the Van der Walls and
electrostatic forces. Thus one could expect to see the effect of these at
larger size ranges compared to spherical particles. ...-and then of course
there is the small problem of finding smooth ellipsoid particles in a
SFF/RP suitable size range....

Some years ago I did some research regarding the practical limits of metal
powder packing in a DTM SLS Sinterstation using size range composition of
spherical particles. Neil is correct regarding the theoretical size range
ratio for intersitial packing of 1:7:49:....., which has been used by
McGeary to compose and vibrate a 4 size range component blend in a
cylindrical container to around 95% density. However he started with a
largest particle size of about 12 mm. For my own research I used an upper
practical size limit of particles at 100 micron. When you work at these
size ranges there are some other practicalities to deal with. To begin with
you will have to accept that all your different powders will have a size
range interval and these are usually wider spread relative to the particle
size for smaller the particles. Even with the first of your components
will a size range interval have a negative influence on particle
arrangement, and of course the combination of two or more different size
range intervals will add to that adverse effect on the densification.
However, with my set up the highest density reached in the Sinterstation
was 68.6% with a bimodal blend but in theory an ideal bimodal composition
could yield 86% density. A three component blend was tried, but with a
particle size range of 1-4 micron the inner friction made the powder mass
behave similar to wet sand and could not be spread out in stable, smooth
layers. Instead the powder in the building container clogged together and
started to move along during the spreading of new layers. Another issue
when working with really small sized particles is their tendency to raise a
suspension in the air. After loading or cleaning the machine you will find
yourself working in an air-powder suspension, and since in my case the
powder was carbonyl iron, this is an environment that might not be harmful
to humans, but is not to recomend for electronic equipment.

Regards, especially for those who have taken the time to read this far...


who, by the way, is in the process of completing a thesis including the
above mentioned research. The thesis will be published and made available
for those interested during this spring.

At 15:21 2004-02-20 +0000, you wrote:
>Dear Jean-Marc Boechat,
>In response to your email.
>Perfect spheres may have a maximum packing density of 0.74 in a perfect=20
>However, I did mention in the original mail that these are randomly jumbled =
>particles in a container. This will affect the packing density and also=20
>makes the experiments more useful for correlation with manufacturing=20
>technologies. The article I read mentioned a series of experiments, which=20
>are published in a scientific journal which confirm this.
>Toby Gill
>--On 20 February 2004 16:04 +0100 Jean-Marc Bo=E9chat=20
><> wrote:
> > Hello list and Toby:
> >
> > I have a problem with your statement that it packs more densely than
> > perfect spheres and you said it approaches 0.74. In fact if you calculate
> > the max packing density of perfect spheres, having the same diameter, it
> > is 0.74. So anything under this is not really interesting to me.
> > Could you elaborate on your statement?
> >
> > I think if you want to improve the packing density the easiest way is to
> > use bi-modal particle distribution.
> >
> > Just my 2 cents worth! :)
> >
> > J.-M. Bo=E9chat
> >
> > Toby Gill wrote:
> >
> >> Thought this may be of interest to some,
> >>
> >> It has just been published that M& M sweets pack together more densely
> >> then perfect spheres when randomly jumbled in a container.
> >> It was also discovered that by stretching the M&M shape (so elliptical,
> >> like an almond) in computer simulations, a packing density approaching
> >> the packing fraction of 0.74 was achieved.
> >>
> >> This seems to me to be important to anyone involved with producing
> >> components from powders, be it by conventional methods or by rapid
> >> manufacturing techniques. Does anyone have any comments/thoughts on this
> >> development?
> >>
> >> P.S. I have just finished a PhD based on SLS and am currently looking
> >> for employment. Anyone who has any vacancies arising in the near future
> >> please contact me at the email address above.
> >>
> >> Regards
> >>
> >> Toby Gill
> >
> > --
> > Jean-Marc Bo=E9chat email:
> > MIM Systems Ltd
> > phone:+4126 430 08 08 fax: +4126 430 08 09 Swiss Branch office
> > website:
> > P.O. Box
> > CH- 1723 Marly 2
> > Switzerland "when things go wrong, don't go with
> > them!"
> >
> > Please remove the nafof in my e-mail address before use!
> > Disclaimer: "I speak for myself only, it's hard enough!"
> >
> >
Received on Tue Feb 24 14:36:27 2004

This archive was generated by hypermail 2.1.8 : Tue Jan 12 2010 - 19:37:11 EET