1. Limited time only! Sign up for a free 30min personal tutor trial with Chegg Tutors
    Dismiss Notice
Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

A Proof for Close Packing of Congruent Identical Spheres

  1. Jun 12, 2016 #1
    I developed two algorithms for calculating the density of close packed congruent identical spheres in two different arrangements:
    • A tetrahedron with four equilateral triangular faces, and
    • A square pyramid with a square base and four equilateral triangular faces, as shown below.

    tetrahedral ball stack w- n=3.jpg
    Figure 1. Tetrahedral ball stack.
    pyramidal ball stack.jpg
    Figure 2. Square pyramidal ball stack.

    Here are the Excel-friendly algorithms (n = number of spheres along bottom row):

    Density of Tetrahedral Stack (Dt):

    Dt = (4*(2^0.5)*n*(n+1)*(n+2)*Pi())/(3*(2*n+2*3^0.5-2)^3 or

    Dt equation.jpg (1)

    Density of Square Pyramidal Stack (Dp):

    Dp = ( n*(1+n)*(1+2*n)*Pi())/(3*((1+(2^0.5*n)))^3) or

    Dp equation.jpg (2)

    I found that, as the number of spheres approaches infinity for both arrangements, that Dt = Dp = ≈ 0.74048... or π/√18.

    I resorted to Wolfram Alpha with the following query:

    Dt limit.jpg = Dp limit.jpg (3)

    and got the following result:

    Wolfram Proof RTC = SPC.jpg (4)


    I then forwarded my calculations to Dr. Thomas C. Hales, who proved the Kepler Conjecture, asking him if Dt = Dp was correct, and he responded, saying,

    "The reason for the equal densities in a tetrahedron and square pyramid is that they can both be viewed as part of the face-centered-cubic packing, each with a different set of exposed facets."

    My question to you is this: can you provide a detailed proof that (1) = (2) as n→∞, i.e., that Dt equals Dp.jpg is true?

    Thanks for any input!

  2. jcsd
  3. Jun 12, 2016 #2


    User Avatar
    2017 Award

    Staff: Mentor

    You get the equality of the limits by calculating the coefficients of the highest power ##3## of ##n##. It is sufficient to see that
    $$\frac{2 \pi n^3}{3 \cdot \sqrt{2}^3 n^3} = \frac{4 \pi \sqrt{2} n^3}{3 \cdot 2^3 \cdot n^3}$$
    because all other terms don't grow as fast (and division of the complete fractions by ##n^3## (nominator and denominator) gives ##\frac{1}{n}## terms which become zero at infinity).
Share this great discussion with others via Reddit, Google+, Twitter, or Facebook

Have something to add?
Draft saved Draft deleted