I was able to find or derive an expression for most other Platonic or Archimedean solids, but for the snub cube I was not able to find a value nor find an expression anywhere.
2 Answers
There is a general expression of the solid angle subtended by the snub cube at any of its $24$ vertices is given by the general expression $$\boxed{\Omega=8\sin^{-1}\left(\frac{(2x-1)\sqrt{3x^2-1}}{\sqrt{3}(4x^2-1)}\right)+2\sin^{-1}\left(\frac{(2x-1)\sqrt{2x^2-1}}{4x^2-1}\right) \space sr}$$
where, $x$ is a parametric variable of an 8th degree polynomial obtained from HCR's Theory of Polygon, given as follows
$$128x^8-352x^6+256x^4-72x^2+7=0\ \ \quad \forall \ x>1$$
After $5$ successive iterations using Newton-Raphson's method, the value of $x$ can be fairly approximated as follows
$x\approx 1.343713374$
Now, substituting the above value of $x$, one should get the approximate value of solid angle subtended by the snub cude at any of its 24 vertices ( all lying on the spherical surface) as follows
$$\Omega=\left.8\sin^{-1}\left(\frac{(2x-1)\sqrt{3x^2-1}}{\sqrt{3}(4x^2-1)}\right)+2\sin^{-1}\left(\frac{(2x-1)\sqrt{2x^2-1}}{4x^2-1}\right)\right|_{x\approx 1.343713374} \\ \approx 3.589629551 \space sr$$
For solid angles subtended by all the Archimedean solids, kindly go through Table of solid angles subtended by Archimedean solids at their vertices by HCR
- 38,565
-
It seems pointless to give $\Omega$ exactly in terms of $K$, when $K$ is not given exactly. Is there some polynomial defining $K$? – mr_e_man Aug 28 '23 at 11:21
-
1You are right. There is a polynomial equation to compute the value of K. I lost that pretty old document. But it can be derived easily although a bit lengthy. I'll have to recompute the value of K. – Harish Chandra Rajpoot Aug 28 '23 at 12:23
-
@mr_e_man Thank you very much for pointing this out. I have recomputed and added sufficient details to my answer. – Harish Chandra Rajpoot Aug 28 '23 at 13:55
I'll derive the angles for the snub dodecahedron with edge length $1$. For the snub cube, just replace $\varphi=(1+\sqrt5)/2$ with $\varphi=\sqrt2$. (Replace "regular pentagon" with "square", etc.)
Since the polyhedron is vertex-transitive, all vertices are on a sphere, whose centre is the average of the vertices. Since the edge lengths are all equal, all vertices adjacent to one vertex are on another sphere, whose radius is the edge length. The intersection of two spheres is a circle. And a circle is contained in a plane. Thus, the vertices adjacent to one vertex form a planar polygon inscribed in a circle. In fact it's an irregular pentagon; four edges are $1$, and the fifth edge is $\varphi$ (the diagonal of a regular pentagon).
Ignoring that fifth edge for the moment, we have a chain of four equal line segments with three equal angles $\theta$ between them. Let $z^2=e^{i(\pi-\theta)}=-\cos\theta+i\sin\theta$ be the complex number that rotates one segment to the next segment. If the complex plane is oriented such that one end of the chain is $0+0i$ and the next vertex is $1+0i$, then the vertices of the chain are
$$0,\quad1,\quad1+z^2,\quad1+z^2+z^4,\quad1+z^2+z^4+z^6.$$
Let $s=|1+z^2|=z^{-1}+z$ be the short diagonal of the chain. It follows that the other two diagonals are
$$|1+z^2+z^4|=z^{-2}+1+z^2=(z^{-1}+z)^2-1=s^2-1$$
and
$$|1+z^2+z^4+z^6|=z^{-3}+z^{-1}+z+z^3=(z^{-1}+z)^3-2(z^{-1}+z)=s^3-2s.$$
(We didn't need to use complex numbers; you could find a different way to derive these expressions for the diagonal lengths.)
Incorporating the fifth edge, we get an equation that determines $s$:
$$s^3-2s=\varphi.$$
Now let $\alpha$ be the dihedral angle between two triangular faces. Since the altitude of a regular triangle is $\sqrt3/2$, the law of cosines gives
$$\cos\alpha=\frac{(\sqrt3/2)^2+(\sqrt3/2)^2-s^2}{2(\sqrt3/2)(\sqrt3/2)}=1-\tfrac23s^2.$$
And let $\beta$ be the dihedral angle between a triangular and a pentagonal face. Consider the irregular tetrahedron with edges $s^3-2s,\;s^2-1,\;1,\;1,\;1,\;1$. The angle $\phi$ in the triangle with edges $s^2-1,\;1,\;1$ is also given by the law of cosines: $\cos\phi=(1^2+1^2-(s^2-1)^2)/(2(1)(1))=\tfrac12+s^2-\tfrac12s^4$. The angle in a regular triangle is of course $60^\circ$, with $\cos60^\circ=\tfrac12$. The angle in a regular pentagon is $108^\circ$, with $\cos108^\circ=1-\tfrac12\varphi^2$. Then the spherical law of cosines can be applied to a small sphere centred on the tetrahedron's vertex with edges $1,\;1,\;1$:
$$\cos\beta=\frac{\cos\phi-\cos60^\circ\cos108^\circ}{\sin60^\circ\sin108^\circ}=\frac{\tfrac12+s^2-\tfrac12s^4-\tfrac12(1-\tfrac12\varphi^2)}{\tfrac{\sqrt3}2\sqrt{1-(1-\tfrac12\varphi^2)^2}}$$ $$=\frac{4s^2-2s^4+\varphi^2}{\sqrt3\sqrt{4\varphi^2-\varphi^4}}=\frac{-2\varphi s+\varphi^2}{\sqrt3\,\varphi\sqrt{4-\varphi^2}}=\frac{\varphi-2s}{\sqrt{12-3\varphi^2}}.$$
Finally, the solid angle $\Omega$ is the area of the spherical pentagon gotten by intersecting the polyhedron with a small sphere centred on the vertex:
$$\Omega=3\alpha+2\beta-3\pi.$$
Here are the numerical values. I take radians and steradians to be the default units, but degrees can be used secondarily: $1^\circ=\tfrac{\pi}{180}=\tfrac{\pi}{180}\text{rad}=\tfrac{\pi}{180}\text{sr}$. So a whole sphere is $4\pi=720^\circ$. The values on the left are for the snub dodecahedron, and the values on the right are for the snub cube.
$$\varphi=\quad1.6180339887,\quad1.4142135624$$ $$s=\quad1.71556149969736783,\quad1.68501832488972093$$ $$\alpha=\quad2.8654006883\text{ rad},\quad2.6744480835\text{ rad} \\ \quad=\quad164.17536606^\circ,\quad153.23458771^\circ$$ $$\beta=\quad2.6691306336\text{ rad},\quad2.4955316306\text{ rad} \\ \quad=\quad152.92992028^\circ,\quad142.98343007^\circ$$ $$\Omega=\quad4.5096853715\text{ sr},\quad3.5896295510\text{ sr} \\ \quad=\quad258.38593872^\circ,\quad205.67062329^\circ$$
- 5,986
-
1You got two different values of solid angle subtended at a vertex by snub cube but there must be unique value as all 24 vertices are identical. How do you know which one is correct? – Harish Chandra Rajpoot Aug 30 '23 at 17:44
-
@HarishChandraRajpoot - I didn't get two different values. My answer was edited after your comment; is it clear now? – mr_e_man Oct 28 '24 at 20:29
-
I would not have converted stersdians to degrees. Degrees are for two-dimensional angles. – Oscar Lanzi Feb 05 '25 at 15:23
-
@OscarLanzi - That's reasonable. But angles are often considered dimensionless, so $1^\circ\approx0.017$ is just a number, and doesn't say whether it's a planar angle or a solid angle. – mr_e_man Feb 06 '25 at 20:51
-
The formula for spherical area in terms of angles depends on this non-dimensionality. Otherwise we'd have to write something like $$\frac{\Omega}{1\text{ sr}}=\frac{3\alpha+2\beta}{1\text{ rad}}-3\pi$$ or $$\frac{2\Omega}{4\pi\text{ sr}}=\frac{3\alpha+2\beta}{2\pi\text{ rad}}-\frac32.$$ – mr_e_man Feb 06 '25 at 20:59
-
I have seen square degrees used. A square degree is $(\pi/180)^2$ of a steradian, thus a sphere with $4\pi$ steradians covers $129600/\pi\approx41253$ square degrees. – Oscar Lanzi Feb 09 '25 at 11:04
