Accelerator experiments contradicting general relativity

Vahagn Gharibyan∗
Deutsches Elektronen-Synchrotron DESY - D-22603 Hamburg

The deflection of gamma-rays in Earth’s gravitational field is tested in laser Compton scattering
at high energy accelerators. Within a formalism connecting the bending angle to the photon’s
momentum it follows that detected gamma-ray spectra are inconsistent with a deflection magnitude
of 2.78 nrad, predicted by Einstein’s gravity theory. Moreover, preliminary results for 13–28 GeV
photons from two different laboratories show opposite – away from the Earth – deflection, amounting
to 33.8–0.8 prad. I conclude that general relativity, which describes gravity at low energies precisely,
arXiv:1401.3720v2 [physics.gen-ph] 12 Jul 2014

break down at high energies.

PACS numbers: 04.80.Cc, 04.20.-q, 29.27.-a

Introduction.— Einstein’s general relativity (GR) [1] to substitute c in Eq.(2) by v and increase the bending
is the currently established theory of gravitation and has magnitude.
been confirmed in all observations and experiments to In this Letter, I describe a laboratory method that
date [2]. An essential validity check of GR is based on probes gravitational bending of high-energy photon
gravitational light bending. These deflection measure- beams.
ments, which were started by a spectacular observation of Gravity as a bending medium.— Within the method
starlight deflection during a solar eclipse about a century I use an idea that gravitation for light is equivalent to
ago [3], have been expanded to radio-waves and have be- an optical medium. This idea was suggested by Einstein
come very precise [4]. The most accurate measurements and was employed by many authors; see ref. [11] and ref-
are performed using the gravitational field of the Sun [5], erences therein. Felice [11] has proved that a GR curved
as the bending increases with mass. The results prove space described by Riemannian geometry is identical to
that electromagnetic radiation, from radio to visible light the language of classical optics in a flat space medium.
frequencies, is bent according to GR and follows the cur- The author, however, has warned that the optical de-
vature of space [6]. On a scale of less massive objects, scription may be mathematically more complicated, al-
the light bending strength of the planet Jupiter has also though it could be beneficial for solving certain problems.
been tested and quantified [7]. For the Earth, however, a The deflection of high-energy photons is one such prob-
check of gravitational bending remains infeasible because lem. In a recent paper [12], the authors further developed
of the smallness of the expected deflection, about 3 nrad an optical approach and suggested the following refrac-
(compared with more than 8 µrad for the Sun). The tive index for a spherically symmetric gravitational field:
quoted numbers follow from the well-known expression  2GM 
4GM/c2 R for the deflection angle [8], when light travels n = exp , (3)
c2 R
near a mass M with an impact parameter R (higher-
which, for the Earth’s weak field, reduce to
order terms of deflection are neglected throughout the
Letter). For a light ray grazing the Earth’s surface, the 2GM⊕
n⊕ = 1 + 2 . (4)
total deflection angle is c R⊕
The latter expression has also been derived by other au-
4GM⊕ thors [13–15] and is equivalent to Eqs.(1) and (2) when
≈ 2.78 × 10−9 , (1)
c2 R⊕ one applies optical tracking with such a refractive index.
The main difference between the gravitational ”medium”
where G is the gravitational constant and c is the speed
presented by Eq.(4) and any material medium is the
of light. The bending magnitude for light generated and
bending independence on frequency of light or photon
measured in a laboratory is much smaller and is equal to
energy, which is a consequence of the gravity geometrical
2GM⊕ L interpretation or the curved space-time concept. It will
, (2) help us to test the gravitational bending since scatter-
c2 R⊕
q
L + R2
2
⊕ ing or interaction angles decrease toward high energies;
at some energy, the angle will approach the magnitude
where L is the length of light travel [9]. Thus, for a dis- of refractivity given in the Eq.(4), interfering with the
tance of 1m, this angle is only 2 × 10−16 rad and the light gravity.
shifts by 0.2 femtometer, which is undetectably small, The Compton process in a gravitational field.— A
at least for a direct measurement. A way to overcome proper process to explore is high-energy Compton scat-
this problem is described in ref. [10], which is based on tering, which is sensitive to tiny deviations of the refrac-
the idea of slowing light down to v ≈ 100m/s, in order tive index from unity, as described in ref. [16]. Using
 2

energy-momentum conservation, when at n ≈ 1 a pho- gravity-induced refractivity are presented in Fig. 1. The
ton scatters off an electron with energy E, the Compton plot shows considerable sensitivity, which grows toward
scattering kinematics is given by high energies in a range available to accelerating labora-
 ω 2 tories.
Ex − ω(1 + x + γ 2 θ2 ) + 2ω 1 − γ (n − 1) = 0, (5) Experimental results.— The high-energy accelerators
E
where laser Compton facilities have been operated for
where x = 4γω0 sin2 (θ0 /2)/m, with γ and m being the years, are listed on the upper energy scale of Fig. 1. As
Lorentz factor and mass of the initial electron, respec- can be seen from the plot, 6 GeV storage rings (ESRF
tively. The initial photon’s energy and angle are denoted and VEPP) have low sensitivity while the higher energy
by ω0 and θ0 , while the refractive index n is in effect colliders (HERA, SLC, LEP) have a great potential for
for the scattered photon with energy ω and angle θ; the detecting the gravitational bending effect; see also Table I
angles are defined relative to the initial electron. This . Although all three machines are not operational any-
kinematic expression is identical to Eq.(8) from ref. [16]
and is derived for small refractivity and high energies,
TABLE I. Sensitivity of different accelerators’ Compton facil-
i.e., the O((n − 1)2 ), O(θ3 ), and O(γ −3 ) terms are ne-
ities to the Earth’s gravitational field.
glected. To determine the outgoing photon’s energy, I
solve Eq.(5) for ω and, to leading order of γ 2 (n − 1), I Accelerator Electron Kinematic ωmax ωmax Shift by
obtain: energy factor n = 1 n = n⊕ gravity
Ex  2γ 2 (n − 1)(1 + γ 2 θ2 )  GeV x GeV GeV GeV
ω= 1+ . (6)
1+x+γ θ 2 2 (1 + x + γ 2 θ2 )2 ESRF, VEPP 6.0 0.21 1.05 1.39 0.34
HERA 26.5 0.98 13.1 23.4 10.3
Writing this formula for the maximal energy of the scat-
SLC, LEP 45.6 1.62 28.2 43.7 15.5
tered photons (Compton edge, at θ = 0) in the Earth’s
gravitational field, I obtain:
more, one can analyze available data recorded by these
Ex  2γ 2 (n⊕ − 1) 
ωmax = 1+ , (7) accelerators where laser Compton setups were employed
1+x (1 + x)2 for polarimetry. Expected shifts of the maximal Comp-
where the Earth’s light bending refractivity n⊕ − 1 is ton energies are large and so prominent that they would
amplified by γ 2 , allowing one measure it by detecting not have been missed if this magnitude gravitational in-
the extremal energy of the scattered photons ωmax , or fluence was present there. This is true for the HERA and
electrons E − ωmax . In order to estimate the method’s SLC but not for the LEP Compton polarimeter, which
sensitivity, I calculate the Compton edge for an incident has generated and registered many photons per machine
photon energy 2.32 eV (the widely popular green laser) at pulse [17]. In this multi-photon regime, any shift of the
different energies of the accelerator electrons. The result- Compton edge is convoluted with the laser-electron lu-
ing dependencies for a free space (n = 1) and the Earth’s minosity and can-not be disentangled and measured sep-
arately.
Unlike the LEP, the SLC polarimeter has operated in
ESRF, VEPP HERA SLC, LEP multi-electron mode and has analyzed the energies of in-
teracted electrons using a magnetic spectrometer [18].
Compton edge [GeV]

60
n⊕ The spectrometer converted energies to positions, which
50 then were detected by an array of Cherenkov counters.
40 n=1 The position-energy correspondence has been derived
from the spectrometer magnetic field strength according
30
to the following expression:
20
296.45 GeV · cm
10 Sx = − 9.61 cm , (8)
E′
0
10 20 30 40 50 60 where Sx is the position of the scattered electron with
Electron energy [GeV] energy E ′ = E − ω. The scaling factor is quoted from
ref. [18] and the offset, which depends on the electron
FIG. 1. Compton scattered photons’ maximal energy (Comp- beam position at the laser-electron interaction point, cor-
ton edge) dependence on the initial electron energy for a head- responds to a calibration from ref. [19]. According to this
on collision with 532nm laser light. Solid and dotted lines relation, the SLC polarimeter’s Compton edge electrons
correspond to the refractive index of the gravitational field at with 17.4 GeV energy will enter the detector at a position
the Earth’s surface (n⊕ = 1 + 1.39 × 10−9 ), and free space of 7.43cm. This is what has been measured with 200µm
(n=1) respectively. statistical accuracy by a kinematic endpoint scan and is
 3

presented in Fig. 3-9 of ref. [18]. Could it happen that that the HERA Compton experiment rules out the GR
these authors have measured the GR-supported value of prediction about gamma-ray bending.
1.9 GeV (at n = n⊕ from table I) instead of 17.4 GeV? From the SLC and HERA measurements and the de-
Eq. (8) tells us that the 1.9 GeV electrons will have a rived or quoted numbers, it follows that both Comp-
position of 146.4 cm at the detector location, inconsis- ton facilities have a much higher sensitivity to the
tent with what has been measured (7.43cm). Possible Earth’s gravitational refractivity than that of the GR
instrumental influence is limited to the initial electron value in Eq. (4), n⊕ − 1 = 1.39 × 10−9 . Indeed, as re-
beam position shift, less than 1 cm (to be contained in ported in ref. [19], the anomalous refractivity equals
the accelerator’s magnetic lattice [20]) and an estimated −(4.07 ± 0.05) × 10−13 for the SLC 16.3–28.3 GeV pho-
accuracy of the magnetic spectrometer, better than 2%. tons and −(1.69 ± 0.47) × 10−11 for the HERA 12.7 GeV
These factors add up to a maximum energy uncertainty gamma-rays. At the time of the publication of ref. [19]
or a possible offset of 1.4 GeV for the measured value of and up until now, the source of this refractivity has re-
17.4 GeV, reducing it to 16 GeV, which is still too high mained unknown since possible contributions by a non-
compared with the predicted value of 1.9 GeV. I therefor perfect machine vacuum, electromagnetic stray fields, or
conclude that the SLC polarimeter data do not support hypothetical vacuum polarizations [22–24] are negligibly
GR gravitational bending. small (< 10−20 ). Now, in light of real gravitational field
At the HERA transverse polarimeter, Compton pho- interpretations, the observed bending ability of the labo-
tons are registered by a calorimeter in single particle ratory vacuum could be attributed to Earth’s gravity as
counting mode. A recorded Compton spectrum from the most influental and likely source. Thus, combining
ref.[21] is shown in Fig. 2 superimposed on a background the SLC and HERA results and multiplying by a factor
Bremsstrahlung distribution. In contrast to the Compton of 2 to obtain the integral bending, one can state that
12.7–28.3 GeV gamma-rays are deflecting away from the
CE(1) CE(n⊕) Earth by 33.8–0.81×10−12rad.
Conclusions.— In order to test the gravitational de-
15000
Photons / 0.32GeV

CE BE flection of photons at the Earth, I first described GR

light bending in equivalent, optical refractivity terms.
Next, for the solution, I applied high-energy laser Comp-
10000
ton scattering, which is extremely sensitive to any small
refractivity due to its well-defined initial and final energy
states and fixed momentum transfers. Finally, I explored
5000
available experimental records from the SLC and HERA
Compton polarimeters, finding with high confidence that
gamma-rays are not bending according to GR. The ob-
0 served energy dependence of gravitational bending is in-
5 10 15 20 25
Photon energy [GeV] compatible with a curved space approach and invalidates
GR or other alternative, purely geometrical, metric grav-
ity theories described, for instance, in ref. [25].
FIG. 2. HERA polarimeter Compton and Bremsstrahlung
(hatched area) spectra. Vertical solid lines show measured po- The SLC and HERA data also revealed a much smaller,
sitions of the Compton (CE) and Bremsstrahlung (BE) maxi- negative deflection or repulsion of the high-energy pho-
mal energies. The dotted lines correspond to predicted Comp- tons from the Earth. Such a change of refractivity
ton edge for free (CE(1)) and Earth’s gravitational (CE(n⊕ )) sign from positive(attractive) at low energies to nega-
space. tive(repelling) at high energies is a known feature of pho-
ton interactions with ordinary matter [26]. This anal-
scattering, in the Bremsstrahlung process the momentum ogy could open new perspectives on quantum gravity,
transfer is not fixed, and any small n 6= 1 effect is smeared for which quantization of the GR geometrical gravita-
out and becomes negligible [19]. Hence, following the tion is a major problem [27]. With the detected en-
analysis in ref. [19], I calibrate the energy scale accord- ergy dependent photon scattering, gravity is exposing an
ing to the maximal Bremsstrahlung energy and show in attribute belonging to momentum transfer interactions,
Fig. 2 the free space- and GR-predicted Compton edge which the quantum approach can more conventionally
energies (from table I), relative to the Bremsstrahlung handle. A possible connection of the observed effects to
edge. Comparing a measured maximal Compton energy the Planck scale, where quantum gravity should materi-
of 12.7 ± 0.1 GeV from ref. [19] with the GR expectation alize, has been discussed previously in ref. [16]. Nonethe-
of 23.4 GeV reveals a huge difference that can not be ex- less, despite its possible relation to the quantum regime,
plained by any one of the instrumental mis-measurement repelling gravity is something quite unusual and dedi-
sources discussed in refs. [21] and [19] (or a total of these cated accelerator experiments are needed to check the
systematic uncertainties). Therefore, I have to conclude preliminary SLC and HERA results on the negative re-
 4

fractivity. [14] P. Boonserm, C. Cattoen, T. Faber, M. Visser and

S. Weinfurtner, Class. Quant. Grav. 22, 1905 (2005)
[gr-qc/0411034].
[15] A. K. Sen, Astrophysics 53, 560 (2010).
[16] V. Gharibyan, Phys. Rev. Lett. 109, 141103 (2012)
∗
[email protected] [arXiv:1207.7297 [hep-ph]].
[1] A. Einstein, Annalen Phys. 49, 769 (1916) [Annalen [17] L. Knudsen, J. P. Koutchouk, M. Placidi, R. Schmidt,
Phys. 14, 517 (2005)]. M. Crozon, J. Badier, A. Blondel and B. Dehning, Phys.
[2] J. Beringer et al. [Particle Data Group Collaboration], Lett. B 270, 97 (1991).
Phys. Rev. D 86, 010001 (2012). [18] R. C. King [SLD Collaboration], Nucl. Phys. Proc. Suppl.
[3] F. W. Dyson, A. S. Eddington, C. R. Davidson, Phil. 37B, 23 (1994). SLAC-0452.
Trans. R. Soc. A, 220, 291 (1920) [19] Gharibyan,V. Phys. Lett. B 611 (2005) 231.
[4] D. E. Lebach, B. E. Corey, I. I. Shapiro, M. I. Ratner, [20] A. W. Chao, K. H. Mess, M. Tigner and F. Zimmer-
J. C. Webber, A. E. E. Rogers, J. L. Davis and T. A. Her- mann, Handbook of accelerator physics and engineering
ring, Phys. Rev. Lett. 75, 1439 (1995). (Hackensack, USA: World Scientific, 2013)
[5] S. S. Shapiro, J. L. Davis, D. E. Lebach and J. S. Gregory, [21] D. P. Barber et al., Nucl. Instrum. Meth. A 329, 79
Phys. Rev. Lett. 92, 121101 (2004). (1993).
[6] C. M. Will, Living Rev. Rel. 9 (2006) 3 [gr-qc/0510072]. [22] J. I. Latorre, P. Pascual and R. Tarrach, Nucl. Phys. B
[7] R. N. Treuhaft & S. T. Lowe, Astron. J., 102, 1879-1888, 437, 60 (1995) [hep-th/9408016].
(1991). [23] W. Dittrich and H. Gies, Phys. Rev. D 58, 025004 (1998)
[8] L. D. Landau and E. M. Lifshitz, The Classical Theory [hep-ph/9804375].
of Fields (Pergamon Press, New York, 1975). [24] K. Scharnhorst, Annalen Phys. 7, 700 (1998)
[9] P. Amore and S. Arceo Diaz, Phys. Rev. D 73, 083004 [hep-th/9810221].
(2006) [gr-qc/0602106]. [25] C. M. Will, Theory and experiment in gravitational
[10] N. Kumar, Europhys. Lett. 82, 60002 (2008) physics (Cambridge University Press, New York, 1993).
[11] F. de Felice, Gen. Rel. Grav. 2, 347 (1971). [26] B. L. Henke, E. M. Gullikson and J. C. Davis, Atom.
[12] X. -H. Ye and Q. Lin, J. Opt. 10, 1 (2008) Data Nucl. Data Tabl. 54, no. 2, 181 (1993).
[arXiv:0711.0633 [gr-qc]]. [27] R. P. Woodard, Rept. Prog. Phys. 72, 126002 (2009)
[13] J. C. Evans, K. K. Nandi and A. Islam, Gen. Rel. Grav. [arXiv:0907.4238 [gr-qc]].
28, 413 (1996)