|
BOOK REVIEW
The Pursuit of Perfect
Packing
(2nd edition) by Thomas Aste and Denis
Weaire
I enjoyed
reading this expanded and updated version of the original book
that I had read some years ago. No prior knowledge is required
to enjoy the 200 pages over a weekend. I often think about
imperfect packing when I throw clothes etc in a bag on a Sunday
night before a trip – this book explores the antithesis – what
is best way to pack objects of various shapes together in order
to minimize the volume. We have all seen oranges close packed in
a supermarket, where staggered hexagonal layers
are placed one above the other to achieve a packing fraction of
QUOTE =
0.74; a result that was first conjectured by Kepler in 1607 to
be optimal, but only formally proved by Hales in 1998. It was a
hard proof because there is a more efficient local packing where
12 spheres are positioned around a central sphere in an
icosahedral arrangement, but this can’t be tiled to fill all
space. The situation with disc packing in two dimensions is more
straightforward as the best local packing where 6 discs are
placed uniquely around a central disc, can be tiled to give an
optimal packing fraction of QUOTE =
0.91. This has some importance to the Irish economy (see Figure
1) which is of importance to Denis Weaire who is a professor at
Trinity College, Dublin and with whom I have had the pleasure of
co-authoring 11 papers over the years, starting when we were
both assistant professors at Yale.
The book is written in a loose conversational style
that holds your attention, although you have to pause to fill in
some of the arguments as you read through the assorted packings
including soap bubbles, sands and grains, the Giant’s causeway,
Voronoi polyhedral, squashed peas and pomegranate pips,
honeycombs, concrete, M & M treats, Bucky balls etc. Denis never
misses the chance to include an Irish reference, and why not. I
did not know that Boris Nickolaevich Delauney was Irish! His
surname comes from his Irish ancestor, Deloney, a mercenary in
Russia with Napoleon. Delauney is the French transliteration of
the name and he found a way to cover space starting with any set
of points now widely used in finite element analysis etc and
called Delauney triangulation.
The book contains
something for everyone – did you know that in a close |
packing in 8
dimensions, each hypersphere is surrounded by 240 other
hyperspheres or that in 24 dimensions this number goes up to
196,560. This is known as the kissing problem, and was first
discussed by Isaac Newton and David Gregory in 1694. Should you
care? Well perhaps yes, because this can be used to help
optimize computer codes in the presence of noise. At the other
extreme the largest known loosest packing of spheres in 3d that
is stable was discovered by Mike O’Keeffe (who is an Emeritus
Chemistry Professor at ASU and of Irish ancestry) in 1996 with a
packing density of only 0.1033.
This book contains a number of unsolved problems
which are almost certainly hard, but new solutions in this area
keep coming along – a recent example being by Paul Chaikin who
showed experimentally (by purchasing a few hundred tetrahedral
dice) that regular tetrahedra can be packed to a much higher
density than previously suspected at 0.75 and remarkably that
this optimal packing seems to be disordered. A neat problem here
for someone!
The highlight of the book is the story of the
discovery of the Weaire-Phelan structure, recently used for the
swimming and diving events at the 2008 Olympics. Lord Kelvin
posed the problem in 1887 “What partitioning of space into
equal volumes minimizes their surface area?” Kelvin’s
solution was to fill space with tetrakaidecahedra which is
perhaps more familiar as the convex Voronoi polyhedron
(Wigner-Seitz cell) of the body-centered-cubic lattice. In 1993,
Weaire and Phelan 
discovered a more efficient packing using two kinds of polyhedra
– beating Kelvin by 0.3%, but a win is a win! It has not been
proved that the Weaire-Phelan structure is optimal, but it is
generally believed to be likely. This discovery was a triumph
for interdisciplinary research and was based on a known
clathrate structure – it would have been very hard for a
mathematician to come up with this. This was the structure used
for the water cube, with the advantage that there are only two
lengths of struts which meet four at a time at rigid joints. The
architects used 90 kilometers of steel beam to make the struts
and also cut through the structure at an angle to be
pleasing esthetically as shown in Figure 2 just in case you
missed seeing this on TV last summer.
Michael Thorpe is ASU Foundation Professor of Physics, Chemistry
& Biochemistry. For more information, click
HERE. |