betaFeedback

WorksPhysics and Astronomy

On the Derivation of the Cosmological Gurzadyan's Theorem

Abstract

In cosmology, the Gurzadyan's theorem identifies the most general force law consistent with the finding of Newton's first shell theorem – that a spherical symmetric mass exerts the same gravitational force as a point mass at its center. This theorem has found important applications in cosmological modeling, particularly in the context of MoND (Modified Newtonian Dynamics), which has recently gained renewed attention as a potential alternative to dark matter. The derivation by Gurzadyan is written in an extremely concise and dense style, making it difficult to follow. Recent proofs of the theorem based on power-series methods offer valuable perspectives, though they differ from the original derivation, which is based on perturbation analysis. Our note aims to clarify the underlying logic in a pedagogical way – accessible to advanced high school or undergraduate students – while preserving conceptual clarity and mathematical elegance of his insight.

To find the force law in the Gurzadyan’s theorem, it suffices to consider a spherical shell instead of the full volume. This follows from the additive nature of the interaction, which allows the total force to be built from contributions of individual infinitesimal-thin shells. If the theorem holds for a single uniform shell – producing the correct effect on a test particle outside – it must also hold for any spherically symmetric distribution.

Consider a spherical shell of radius bb, centered a distance a>ba > b from a test particle (see Figure 1A), and having total unit mass – the mass density per unit surface area is thus μ=1/4πb2\mu= 1/4\pi b^2. We use spherical coordinates (ρ,θ,ϕ)(\rho, \theta, \phi) with the origin located at the center of the shell, where ρ\rho is the radial distance, θ=[0,π]\theta= [0, \pi] is the polar angle measured from the axis connecting the shell center to the test particle, and ϕ=[0,2π]\phi= [0, 2\pi] is the azimuthal angle. Let rr be the distance between the test particle and an infinitesimal mass element on the shell, each element exerts a force along r⃗\vec{r}. Due to symmetry, the transverse components cancel upon integration, leaving only radial ρ\rho-components along the axis connecting the shell center and the test particle. The total gravitational acceleration at the position of the test mass, denoted g(a,b)g(a,b), depends on the distance aa and the shell radius bb, and is given by:

Schematic of the spherical shell and test particle configuration

Figure 1. Schematic of the spherical shell and test particle configuration used to calculate the gravitational acceleration g(a,b)g(a,b) in Eq. (1).

g(a,b)=μ∫0πb dθ∫02πbsin⁡θ dϕ F(r)cos⁡δ=12∫0πdθsin⁡θ F(r)cos⁡δ ,(1)\begin{aligned} g(a,b) &= \mu\int_{0}^{\pi} b\,d\theta\int_{0}^{2\pi} b\sin\theta\,d\phi\ F(r)\cos\delta\\ &= \frac{1}{2}\int_{0}^{\pi}d\theta\sin\theta\ F(r)\cos\delta\,, \tag*{(1)} \end{aligned}

where rr and δ\delta are calculated from:

r=(a2+b2−2abcos⁡θ)1/2,cos⁡δ=a−bcos⁡θr.r = \left(a^{2}+b^{2}-2ab\cos\theta\right)^{1/2}, \quad\cos\delta= \frac{a-b\cos\theta}{r}.

Here, the force law is represented by F(r)F(r), such that the interaction force between any pair of masses is given by their product multiplied by this function.

We seek the most general form F(r)F(r) that is consistent with the finding of Newton’s first shell theorem [2]. For our setting, this translates to the condition for the gravitational acceleration calculated in Eq. (1):

g(a,b)∣b<a=g(a,0).(2)\left.g(a,b)\right|_{b<a}=g(a,0). \tag*{(2)}

If this equation holds for all b<ab<a, then it must also hold in the limit b→0b\to0, where bb can be treated as a perturbative parameter. We can then Taylor’s expand g(a,b)g(a,b) as:

g(a,b)≈g(a,0)+bg1(a)+12b2g2(a)+O(b3),(3)g(a,b)\approx g(a,0)+bg_1(a)+\frac{1}{2}b^2g_2(a)+\mathcal{O}(b^3), \tag*{(3)}

and then set all gn(a)=0g_n(a)=0. Note that, Eq. (1) shows that g(a,b)g(a,b) is an even function of bb, i.e. satisfying g(a,b)=g(a,−b)g(a,b)=g(a,-b). Thus, in Eq. (3), the odd-order term should always vanish gn∈2N−1(a)g_{n\in2\mathbb{N}-1}(a). Setting gn∈2N(a)=0g_{n\in2\mathbb{N}}(a)=0 gives us infinitely many differential equations of FF. The leading-order contribution in the perturbative expansion is the 2nd-order term.

We expand the integrand in Eq. (1), i.e. F(r)cos⁡δF(r)\cos\delta, to the 2nd-order in bb:

F(r)cos⁡δ≈F(a)+bcos⁡θF′(a)−b22a2[sin⁡2θF(a)−asin⁡2θF′(a)−a2cos⁡2θF′′(a)]+O(b3),(4)\begin{aligned} F(r)\cos\delta&\approx F(a)+b\cos\theta F'(a)\\ &\quad-\frac{b^2}{2a^2}\left[\sin^2\theta F(a)-a\sin^2\theta F'(a)-a^2\cos^2\theta F''(a)\right]+\mathcal{O}(b^3), \tag*{(4)} \end{aligned}

in which the terms F′F', F′′F'' are 1st- and 2nd-derivatives of FF. A more detailed treatment of this step is provided in the Appendix A. With this, we do the θ\theta-integration in Eq. (1) and then match with the order expansion in Eq. (3) to obtain:

g(a,0)=F(a),g1(a)=0,g2(a)=13a2[2F(a)−2aF′(a)−a2F′′(a)].(5)\begin{aligned} g(a,0)&=F(a),\quad g_1(a)=0,\\ g_2(a)&=\frac{1}{3a^2}\left[2F(a)-2aF'(a)-a^2F''(a)\right]. \tag*{(5)} \end{aligned}

We outlined the derivation for this step in Appendix B. Setting g2(a)=0g_2(a)=0 is equivalent to solving Eq. (4) of the original derivation in [1], i.e.

2F(a)−2aF′(a)−a2F′′(a)=0.(6)2F(a)-2aF'(a)-a^2F''(a)=0. \tag*{(6)}

It is unfortunate that the steps leading from Eq. (1) and Eq. (2) to Eq. (6) are completely omitted in [1]. However, we have now reconstructed them.

Eq. (6) provides a necessary condition on FF, but not a sufficient one, as FF must also satisfy an infinite set of additional differential equations of the form gn∈2N+2(a)=0g_{n\in2\mathbb{N}+2}(a)=0. However, once a “candidate” solution FF is obtained by solving Eq. (6), we can substitute it back into Eq. (1) for general bb and verify whether Eq. (2) holds. If it does, then the function FF satisfies all required conditions non-perturbatively and is the solution we seek.

The general solution for Eq. (6), which is a 2nd-order homogeneous linear differential equation with variable coefficients and can be solved systematically [8], has the form:

F(r)=Ar+Br−2(7)F(r) = Ar + Br^{-2} \tag*{(7)}

which is the combination between Coulomb’s law F(r)∝r−2F(r) \propto r^{-2} and Hook’s law F(r)∝rF(r) \propto r.

Plugging Eq. (7) into Eq. (1) does indeed result in Eq. (2), which means it is the most general force law consistent with the finding of Newton’s first shell theorem [2]. This is the statement of the cosmological Gurzadyan’s theorem [1].

We would like to point out that Gurzadyan’s theorem had already been studied, proven, and even generalized [9, 10, 11, 12] prior to the publication of [1]. It would be interesting to explore whether similar theorems might arise in exotic spaces – such as fractal geometries – where conventional calculus breaks down [13, 14, 15] and new forms of long-range interaction may emerge.

Proving the cosmological Gurzadyan’s theorem was part of the training for the Vietnamese IPhO (International Physics Olympiad) team in 2024. We thank Tran D. Huy for helpful discussions, and the xPhO club for encouraging us to share this with a broader audience.

Derive Eq. (4)

We want to find the Taylor’s expansion in bb of the integrand F(r)cos⁡δF(r)\cos\delta in Eq. (1), so let us begin by finding the series expansions of the central force F(r)F(r) and the trigonometric factor cos⁡δ\cos\delta.

First, we note that the distance rr between a shell element and the test particle can be expanded in powers of bb using a Taylor series as follows:

r≈r∣b=0+b∂br∣b=0+12b2∂b2r∣b=0+O(b3)(8)r \approx\left.r\right|_{b=0} + \left.b\partial_b r\right|_{b=0} + \frac{1}{2}\left.b^2\partial_b^2 r\right|_{b=0} + O(b^3) \tag*{(8)}

Given that r=(a2+b2−2abcos⁡θ)1/2r=(a^2+b^2-2ab\cos\theta)^{1/2} as mentioned in Eq. (1), r∣b=0=a\left.r\right|_{b=0}=a, and we can calculate its derivatives to arrive at:

∂br∣b=0=−acos⁡θ+b(a2+b2−2abcos⁡θ)1/2∣b=0=−cos⁡θ,\left.\partial_b r\right|_{b=0}=\left.\frac{-a\cos\theta+b}{(a^2+b^2-2ab\cos\theta)^{1/2}}\right|_{b=0}=-\cos\theta,
∂b2r∣b=0=a2sin⁡2θ(a2+b2−2abcos⁡θ)3/2∣b=0=sin⁡2θa.(9)\left.\partial_b^2 r\right|_{b=0}=\left.\frac{a^2\sin^2\theta}{(a^2+b^2-2ab\cos\theta)^{3/2}}\right|_{b=0}=\frac{\sin^2\theta}{a}. \tag*{(9)}

Therefore, Eq. (8) can be rewritten as:

r≈a−bcos⁡θ+b22asin⁡2θ+O(b3).(10)r \approx a - b\cos\theta+ \frac{b^2}{2a}\sin^2\theta+ \mathcal{O}(b^3). \tag*{(10)}

Then, we Taylor’s expand the central force F(r)F(r) around r=ar=a:

F(r)≈F(a)+(r−a)F′(a)+12(r−a)2F′′(a)+O[(r−a)3].(11)F(r) \approx F(a) + (r-a)F'(a) + \frac{1}{2}(r-a)^2F''(a) + \mathcal{O}\left[(r-a)^3\right]. \tag*{(11)}

Applying the result of Eq. (10) and organizing by powers of bb, we find:

F(r)≈F(a)−bcos⁡θF′(a)+b22a2[asin⁡2θF′(a)+a2cos⁡2θF′′(a)]+O(b3).(12)F(r) \approx F(a) - b\cos\theta F'(a) + \frac{b^2}{2a^2}\left[a\sin^2\theta F'(a) + a^2\cos^2\theta F''(a)\right] + \mathcal{O}(b^3). \tag*{(12)}

Next, we need to expand cos⁡δ\cos\delta, mentioned in Eq. (1), in powers of bb. From the expansion of rr given in Eq. (10), we can calculate its reciprocal:

1r≈1a+ba2cos⁡θ+b2a3(1−32sin⁡2θ)+O(b3),(13)\frac{1}{r} \approx\frac{1}{a} + \frac{b}{a^2}\cos\theta+ \frac{b^2}{a^3}\left(1-\frac{3}{2}\sin^2\theta\right) + \mathcal{O}(b^3), \tag*{(13)}

and thus obtain:

cos⁡δ=a−bcos⁡θr≈1−b22a2sin⁡2θ+O(b3).(14)\cos\delta= \frac{a-b\cos\theta}{r} \approx1-\frac{b^2}{2a^2}\sin^2\theta+\mathcal{O}(b^3). \tag*{(14)}

Together with Eq. (11), we get:

F(r)cos⁡δ≈F(a)−bcos⁡θF′(a)+b22a2[−sin⁡2θF(a)+asin⁡2θF′(a)+a2cos⁡2θF′′(a)]+O(b3),(15)\begin{aligned} F(r)\cos\delta&\approx F(a)-b\cos\theta F'(a) \\ &\quad+\frac{b^2}{2a^2}\left[-\sin^2\theta F(a)+a\sin^2\theta F'(a)+a^2\cos^2\theta F''(a)\right]+\mathcal{O}(b^3), \tag*{(15)} \end{aligned}

which is identical to Eq. (4).

Derive (5) from (4)

We can do the integration in Eq. (1) order by order in powers of bb, since the integrand is already expressed in Eq. (4) as a power series in bb:

  • For the b0b^0 order:

12∫0πdθsin⁡θ F(a)=F(a).(16)\frac{1}{2}\int_{0}^{\pi}d\theta\sin\theta\,F(a)=F(a). \tag*{(16)}
  • For the b1b^1 order:

b2∫0πdθsin⁡θcos⁡θF′(a)=0.(17)\frac{b}{2}\int_{0}^{\pi}d\theta\sin\theta\cos\theta F'(a)=0. \tag*{(17)}
  • For the b1b^{1} order, there are 3 contributions:

−b24a2∫0πdθsin⁡θsin⁡2θF(a)=−b23a2F(a)b24a∫0πdθsin⁡θsin⁡2θF′(a)=b23aF′(a)b24∫0πdθsin⁡θcos⁡2θF′′(a)=b26F′′(a).(18)\begin{aligned} -\frac{b^{2}}{4a^{2}}\int_{0}^{\pi} \mathrm{d}\theta\sin\theta\sin^{2}\theta F(a) &= -\frac{b^{2}}{3a^{2}}F(a) \\ \frac{b^{2}}{4a}\int_{0}^{\pi} \mathrm{d}\theta\sin\theta\sin^{2}\theta F'(a) &= \frac{b^{2}}{3a}F'(a) \\ \frac{b^{2}}{4}\int_{0}^{\pi} \mathrm{d}\theta\sin\theta\cos^{2}\theta F''(a) &= \frac{b^{2}}{6}F''(a). \tag*{(18)} \end{aligned}

Adding them all up, we get:

g(a,b)=12∫0πdθsin⁡θF(r)cos⁡δ≈F(a)−b26a2[2F(a)−2aF′(a)−a2F′′(a)]+O(b3).(19)g(a,b)=\frac{1}{2}\int_{0}^{\pi}\mathrm{d}\theta\sin\theta F(r)\cos\delta\approx F(a)-\frac{b^{2}}{6a^{2}}\left[2F(a)-2aF'(a)-a^{2}F''(a)\right]+\mathcal{O}(b^{3}). \tag*{(19)}

This result can then be matched with Eq. (3) to give Eq. (5).

References

References

  1. [1]V. G. Gurzadyan, The Observatory 105, 42 (1985).
  2. [2]I. Newton, Philosophiae naturalis principia mathematica, Vol. 1 (G. Brookman, 1833).
  3. [8]G. Strang, Calculus, Vol. 1 (SIAM, 1991).
  4. [9]I. N. Sneddon and C. K. Thornhill, Proceedings of the Cambridge Philosophical Society 45, 318 (1949).
  5. [10]C. W. Kilmister, The Mathematical Gazette 58, 129 (1974).
  6. [11]C. J. Chapman, The Mathematical Gazette 67, 14 (1983).
  7. [12]A. Barnes and C. K. Keogh, The Mathematical Gazette 68, 138 (1984).
  8. [13]V. E. Tarasov, Physics Letters A 336, 167 (2005).
  9. [14]V. E. Tarasov, Physics Letters A 341, 467 (2005).
  10. [15]T. V. Phan, T. H. Cai, and V. H. Do, Physics of Fluids 36 (2024).

Paper details

vlnc:W_myvgjc9h
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 0

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Physics and Astronomy

Not recorded for this paper

View statement
Looking for related papers…
Contents