Preamble

Vacuum: Candidate for Ether MinYu Li 1Yunnan Observatories, Chinese Academy of Sciences, Kunming, 650216, People’s Republic of China (Dated: November 6, 2025) Vacuum is also a type of matter. It is the only medium for light propagation. In regions far from galaxies, the vacuum is flat. When an atomic nucleus exists, the surrounding vacuum becomes curved, similar to a sponge. When photons pass near atomic nuclei, they also travel along curved paths. This microscopic effect cannot be directly detected, but it does exhibit a macroscopic representation. That is, the “speed” of light slows down in the “media,” such as air, water, and glass. Because the actual microscopic path length increases while the speed of light remains unchanged. Meanwhile, light with longer wavelengths is more likely to bypass obstacles, so it travels fewer detours and exhibits “higher speed.” Additionally, the vacuum is anisotropic in a microscopic region at the boundary between two “media.” As photons pass through this region, the spatial structure determines their deflection. Furthermore, when matter is in motion, its internal vacuum is entirely dragged along.

This is the actual reason for the null result of the Michelson-Morley experiment conducted on Earth’s surface. However, the principle of the constancy of the speed of light cannot serve as an explanation.

Therefore, it is predicted that repeating this experiment on the lunar surface would yield observable shifts in the interference fringes. If the experiment is successful, then vacuum is the long-sought ether.

INTRODUCTION

Why are photons so “smart”? For example, when they enter water from air at a 45◦ angle, how can they instantly “know” the refractive index (n=1.333) of the approaching medium and refract at 32◦ upon contacting the water surface?

Another question is why the speed of light is decreased in a medium? It is commonly believed that the speed of light decreases due to its interaction with electrons (or molecules) in a medium. However, if this were the case, the reduction in speed would occur continuously within the medium. In other words, as the length of the medium increases, the speed of light should decrease progressively.

Yet, this phenomenon has not been observed experimentally. An alternative view suggests that light interacts only with the first layer of electrons (or molecules) when it enters the medium. But how could photons distinguish which electrons or molecules belong to the first layer?

This explanation would also seem to endow photons with some kind of intelligence.

The statement “matter tells space how to curve, space tells matter how to move” may provide a clue to these questions. In the model presented here, the first “matter” refers to the atomic nuclei in the so called “medium”, and the second “matter” refers to the photons.

2. RESULTS

Photons, of course, have no intelligence, and they simply move along geodesics in vacuum, which serves as the only medium for light propagation. When vacuum is flat, light travels along straight paths; when vacuum is curved, light follows curved paths; when the matter moves, the internal vacuum within it is entirely dragged along.

When photons move through a “medium”, their speed does not change, but they travel a longer path at the microscopic level. Since the vacuum inside the “medium” is isotropic, the photons still travel in a straight line at the macroscopic level. Therefore, since the actual path traveled is longer, the “speed” of photons decreases at the macroscopic level. Meanwhile, light with longer wavelengths is more likely to bypass obstacles, so it travels fewer detours and exhibits higher “speed”. In a microscopic region at the boundary between two “media”, the vacuum is anisotropic. When photons pass through this region, the spatial structure determines their deflection.

This structure also allows photons of different wavelengths to be deflected at different angles. Further research is needed to characterize the precise form of this spatial structure in vacuum.

Additionally, when reviewing the Michelson-Morley experiment, note that the principle of the constancy of the speed of light cannot explain the null result. Because this principle states that the speed of light (magnitude) remains constant in different inertial frames while its direction changes. A plausible explanation for the null result is that the vacuum within the air on Earth’s surface is entirely dragged, and light is isotropic in the same inertial frame. Therefore, this work predicts that a shift in the interference fringes would be observed if the Michelson-Morley experiment were conducted on the lunar surface.

If the experimental results are confirmed, then vacuum is the ether that had been sought after for so long.

3. ANALYSIS

A compact celestial body can curve the space around it. This is proved by the deflection of starlight passing near the Sun. However, one key point was neglected: this curve was detected using only two lines of starlight, one from the ordinary night sky and one from a total solar eclipse. Note that to determine whether a two-dimensional plane is curved, three lines must be used to form a triangle, and the sum of its interior angles must be measured to see if it equals 180◦. It can be inferred that detecting curvature in three-dimensional space requires a greater number of lines. Therefore, the curvature around the Sun is implausible to be a curvature of three-dimensional space; rather, it is a curvature of vacuum.

This paper proposes the following physical scenario: vacuum is also a type of matter, and is the only medium for light propagation. Similar to a sponge, vacuum can be curved by a compact celestial body composed of densely packed atomic nuclei. Furthermore, it can be inferred that such curvature originates from the contribution of each individual atomic nucleus. Therefore, it can be concluded that a single atomic nucleus can also curve its surrounding vacuum, although the curvature is extremely weak. It can be further inferred that photons can also go through a slightly curved path when passing near the atomic nuclei. This curved path is microscopic and currently undetectable by direct observation. However, this physical mechanism exhibits macroscopic representation.

3.1 Stationary “medium”

The model presented in this work suggests that the currently recognized optical transmission media (e.g., air, water, and glass) are not the actual medium. Instead, the vacuum within these matters constitutes the actual medium through which light propagates. Within materials, the vacuum is isotropic. Consequently, light propagates along macroscopically straight paths. Microscopically, however, the photons move through curved paths, actually increasing their path length, while light speed (c) remains unchanged. Since the actual path length is longer, the “speed” of light in the “medium” appears reduced.

Moreover, qualitatively speaking, longer-wavelength light more easily bypasses obstacles—that is, it takes fewer detours—and thus traverses shorter microscopic paths, exhibiting “higher speed” in the “medium”.

The propagation time through such a “medium” can be expressed by Equation (1):
$$t = \frac{l'}{c} = \frac{l}{v}$$
where $l'$ is the microscopical path length, $c$ denotes the speed of light in vacuum, $l$ stands for the macroscopic distance of light transfer through the “medium”, and $v$ is the experimentally measured velocity of light in the “medium”. Therefore, $l'$ is proportional to the refractive index $n$:
$$l' = n \cdot l$$

Additionally, let $l' = l + \Delta l$, which leads to:
$$n = 1 + \frac{\Delta l}{l}$$
where $\Delta l/l$ is solely related to the physical properties of the “medium”.

For further investigation, Figure 1 [FIGURE:1] represents the relation of refractive indices $n$ and $m \cdot R_e^{-1} \cdot V_M^{-1}$ of various common “media”, where $m$ is the mass of the atomic nuclei, $R_e$ is the equivalent radius of the atomic nuclei, and $V_M$ is the molar volume of the matter. For a single-atom molecule, $R_e$ is the atomic nucleus radius; for a multi-atom molecule, $R_e$ is derived using the surface area equivalent radius formula:
$$R_e = \sqrt{\sum_{i=1}^{N} r_i^2}$$
where $r_i$ is the radius of each atomic nucleus, and $N$ is the number of atomic nuclei in the molecule.

In General Relativity, the deflection angle of distant starlight passing near the Sun is proportional to $M_\odot \cdot R_\odot^{-1}$. Therefore, the $-1$ power of $R_e$ is more reasonable than other powers. For ideal gases, the molar volumes are the same, so the inclusion of the $V_M$ term in the left panel is negligible. However, for liquids and solids, the molar volumes are much smaller, resulting in a higher number of atomic nuclei per unit volume. This results in stronger spatial curvature within the material and, consequently, a larger refractive index. Therefore, $V_M$ is inversely proportional to the refractive index $n$.

The figure shows that water vapor (left panel), liquid water (right panel), and ice (right panel) all fit well with the blue line. In the left panel, the blue points indicating the single-atom molecular gases helium (He), neon (Ne), argon (Ar), krypton (Kr), and xenon (Xe) are fitted to a straight line: $y = ax + b$. The blue line shows the results: $a = 354508(13)$ and $b = 0.99993706(2)$. Since $b$ is very close to 1, this straight line equation is consistent with Equation (3).

Therefore, an empirical equation relating the material properties of the “medium” to refractive index $n$ for single-atom molecule gas is obtained:
$$n = a \cdot m \cdot R_e^{-1} \cdot V_M^{-1} + 1$$
where $a = 354508(13)$ kg$^{-1}$·m$^4$·mol$^{-1}$.

However, the upper boundary of the light-blue region in Figure 1 corresponds to $1.5a$, implying that the actual uncertainty should be larger for multi-atom molecules. It should be noted that Equation (4) is only an approximation, as demonstrated more clearly by CS$_2$ and CaF$_2$ in the right panel, warranting further in-depth investigation.

3.2 “Medium” in motion

The aforementioned analysis is reasonable when dealing with stationary “media”. Furthermore, when the “medium” is in motion, is the vacuum within it stationary, partially dragged, or entirely dragged?

The null result of the Michelson-Morley experiment sufficiently demonstrates that the vacuum in air is entirely dragged by moving air. It should be emphasized that the principle of the constancy of the speed of light cannot explain this null result. The Lorentz velocity transformation equations are as follows:
$$u'\parallel = \frac{u\parallel - v}{1 - u_\parallel v/c^2},$$
$$u'\perp = \frac{u\perp}{\gamma(1 - u_\parallel v/c^2)},$$
where $\gamma = 1/\sqrt{1 - v^2/c^2}$. In the parallel direction, using Equation (6), when $u_\parallel$ is equal to $c$, $u'\parallel$ is equal to $c$ too. In the perpendicular direction, when $u\perp = c$, and in this case $u_\parallel = 0$, then $u'\parallel = -v$ according to Equation (6) and $u'\perp = c/\gamma$ according to Equation (7). Consequently, $u' = \sqrt{(u'\parallel)^2 + (u'\perp)^2} = c$, but its direction is no longer perpendicular. Note that the principle of the constant speed of light refers specifically to the invariance of light’s speed (magnitude), while its direction of propagation changes in non-parallel directions.

However, in the Michelson-Morley experiment, the two light paths are orthogonal. A plausible explanation for the null result is that the air on Earth’s surface moves synchronously with the Earth, and the vacuum within the air is entirely dragged. Consequently, the experiment involves only one inertial frame, where light is isotropic, thus yielding the null result.

In contrast, the Fizeau experiment actually involves two inertial frames: one being the vacuum in the air, and the other is the entirely dragged vacuum in the moving water. The observed “partial drag” effect is essentially a result of the Lorentz transformation between these two inertial frames.

Based on the above analysis, a feasible approach would be to conduct the Michelson-Morley experiment on the lunar surface. In Figure 2 [FIGURE:2], let $v$ be the velocity of the Moon relative to vacuum. The lengths of $AC$ and $AB'$ are both equal to $l$. In the parallel direction, the time of light is:
$$T_{ACA} = \frac{2l}{c - v}$$

In the perpendicular direction, note that if the mirror is placed at point $B'$, the returning light will not be received. Therefore, the mirror must be positioned at point $B$, which is slightly offset from $B'$. Additionally, the mirror angle must be adjusted to ensure the reflected light returns to point $A$. The time it takes light to travel segment $AB$ at speed $c$ is equal to the time it takes to travel segment $AB'$ at speed $u'$:
$$T_{ABA} = \frac{2l}{\sqrt{c^2 - v^2}}$$

Therefore, the time difference between the two lights is:
$$\Delta T = T_{ACA} - T_{ABA} = \frac{2l}{c}\left(\frac{1}{1-\beta} - \frac{1}{\sqrt{1-\beta^2}}\right)$$
where $\beta = v/c$.

4. DISCUSSION

Equation (5) is incomplete, as it does not account for light of different wavelengths. Meanwhile, Equation (4) appears somewhat rough and unsatisfactory. The geometric structure of vacuum within material media remains a topic requiring deeper investigation. Here, only a qualitative analysis is provided. Inside such “media”, vacuum is isotropic. As light propagates through it, its path curves at the microscopic scale but appears straight at the macroscopic scale. The speed (magnitude) of light remains unchanged; however, the actual path length increases, which leads to the macroscopic measurement that the speed of light is “slowed down” in the “medium”. In a microscopic region near the boundary between two such “media”, vacuum becomes anisotropic. The geometric configuration of this region determines how light is deflected when it passes through. This configuration also allows photons of different wavelengths to be deflected at different angles. Additionally, some anisotropic media, such as calcite (CaCO$_3$) and quartz (SiO$_2$), likely have an asymmetrical spatial arrangement of atomic nuclei.

Vacuum is also a type of matter, which may not be continuous. It can be inferred that curved vacuum generates a repulsive force, which may serve as the origin of gravity and could also potentially act as the medium for its propagation.

If experiments on the lunar surface yield positive results, then vacuum may indeed be the ether. However, it is not absolutely rest, as was previously assumed. At the very least, it can be curved or dragged. Whether the ether in cosmic space is absolutely rest or exhibits relative motion with respect to logical space requires further experimentation.

The existence of an absolute inertial frame does not contradict Special Relativity. It can simply be treated as an ordinary inertial frame without any “privileged” status. If clocks were distributed throughout the observable universe, the total number of clocks would be finite. Therefore, at least one clock would tick the fastest. The inertial frame corresponding to this clock would constitute the absolute inertial frame.

In the 19th century, Fizeau, Michelson and Morley designed ingenious experiments to search for the ether and obtained crucial results. In 1905, Einstein discovered Special Relativity. In 1911, Rutherford revealed that matter contains vast vacuum within. In 1915, General Relativity demonstrated that space (indeed the vacuum) could be curved. In 1930, Dirac proposed that the vacuum is not truly empty. These groundbreaking achievements served as vital sources of inspiration for the present discovery. These brilliant scientists deserve the highest tribute.

5. CONCLUSION

The conclusions of this work are summarized as follows:

1. Vacuum is a type of matter. It is the only medium through which light can propagate. Air, water, glass, etc., are not actual media.

2. The curvature around a compact celestial body is implausible to be a curvature of three-dimensional space; rather, it is a curvature of vacuum.

3. The atomic nuclei can also curve the surrounding vacuum, causing light to follow a curved path at the microscopic level during propagation in the “media”. This physical mechanism results in the macroscopic measurement of a “slowed” light speed within the “media”. Equation (5) is an empirical equation relating the material properties to refractive index $n$ for single-atom molecules.

4. In a microscopic region at the boundary between two “media”, the vacuum is anisotropic. When photons pass through this region, the spatial structure determines their deflection.

5. This paper argues that the null result of the Michelson-Morley experiment is due to the entire dragging of the internal vacuum by the Earth’s surface air, rather than to the principle of the constancy of the speed of light.

6. It is predicted that repeating the Michelson-Morley experiment on the lunar surface would yield observable shifts in the interference fringes.

7. Curved vacuum generates a repulsive force, which may serve as the origin of gravity and could also potentially act as the medium for its propagation.

8. Given that vacuum can be curved and entirely dragged, it may not be continuous.

9. The existence of an absolute inertial frame does not contradict Special Relativity. It can simply be treated as an ordinary inertial frame without any “privileged” status.

If the experiment is successful, then vacuum is ether. But it is not absolutely rest.

FIG. 2. The schematic diagram of the Michelson-Morley experiment on the lunar surface.

As a result, it can be predicted that repeating the Michelson-Morley experiment on the lunar surface will yield observable shifts in the interference fringes. If so, then the vacuum is the long-sought ether.

In addition, this work cautiously suggests that a null result would be obtained even when conducting this experiment in an Earth-based vacuum environment because the vacuum inside the glass cover would remain entirely dragged.

ACKNOWLEDGMENTS

This paper does not require any reference.

Appendix A: Physical parameters of the “media”

TABLE I. Refractive indices and molar volumes of the “media”.
| molecular formula | molar volume (m³·mol⁻¹) | refractive index |
|-------------------|------------------------|------------------|
| H₂O(gas) | 2.241e-2 | 1.000261 |
| He | 2.241e-2 | 1.000035 |
| Ne | 2.241e-2 | 1.000067 |
| Ar | 2.241e-2 | 1.000281 |
| Kr | 2.241e-2 | 1.000427 |
| Xe | 2.241e-2 | 1.000702 |
| H₂O(liquid) | 1.807e-5 | 1.333 |
| CH₃OH | 4.047e-5 | 1.329 |
| C₂H₅OH | 5.836e-5 | 1.361 |
| C₃H₈O₃ | 7.300e-5 | 1.473 |
| C₃H₆O | 7.360e-5 | 1.359 |
| C₁₀H₇Br | 1.388e-4 | 1.658 |
| H₂O(solid) | 1.965e-05 | 1.310 |
| CaCO₃ | 3.693e-05 | 1.680 |
| CaF₂ | 2.455e-05 | 1.434 |
| CS₂ | 2.269e-05 | 1.628 |

These values were obtained from: https://refractiveindex.info and the CRC Handbook of Chemistry and Physics, under standard temperature and pressure (STP) conditions and a wavelength of 589.3 nm.

TABLE II. The masses and radii of atomic nuclei.
| element | atomic mass (u) | radius (fm) |
|---------|-----------------|-------------|
| H | 1.007 | 0.8783 |
| He | 4.0026 | 1.6755 |
| C | 12.011 | 2.4702 |
| N | 14.006 | 2.5582 |
| O | 15.999 | 2.6990 |
| Na | 22.989 | 3.0055 |
| Cl | 35.446 | 3.3654 |
| Ar | 39.948 | 3.3960 |
| K | 39.098 | 3.3876 |
| Br | 79.901 | 4.1629 |
| Kr | 83.798 | 4.2100 |

u: 1.6605e-27 kg. The masses were obtained from: https://physics.nist.gov/cgi-bin/Compositions/stand_alone.pl.

fm: 1e-15 m. The radii were obtained from: https://www-nds.iaea.org/radii.