Historical research guide · 0557 · 2020

The PVLAS experiment: A 25 year effort to measure vacuum magnetic birefringence

A. Ejlli · F. Della Valle · U. Gastaldi · G. Messineo · R. Pengo · G. Ruoso · G. Zavattini

In one minute

Quantum electrodynamics says the vacuum is not empty but polarisable, and that a strong magnetic field should make it behave like a crystal: light polarised along the field travels very slightly slower than light polarised across it. The predicted difference is minute — about 2.5 parts in ten thousand million million million at 2.5 tesla — and no laboratory has yet resolved it directly. Aldo Ejlli, Guido Zavattini, Giuseppe Ruoso and their colleagues spent twenty-five years, funded by Italy’s INFN, building the instrument that came closest. This report is the whole story: rotating superconducting magnets first, then two rotating permanent magnets; a three-metre optical cavity of finesse near seven hundred thousand, which made the light cross the field hundreds of thousands of times; five million seconds of running. Their final measurement, the best ever obtained by optical means, brackets the vacuum’s birefringence to within a factor of about seven of the predicted value — and the paper names the magnet that would close the gap.

Connection

Chapter 2 says the vacuum is a real, structured medium, and vacuum magnetic birefringence is the cleanest optical signature it has: bend light with empty space and a magnet, and the medium is no longer a figure of speech. PVLAS is the experiment that pinned down how large that effect can still be, and its careful accounting of every noise source is the kind of work chapter 1’s evidence ladder is built out of. It also leaves a specific, funded next step on the table — a spare LHC dipole magnet, and a Letter of Intent already at CERN.

From the source

Quantum electrodynamics predicts that an external magnetic field makes the vacuum birefringent: light polarised parallel to the field acquires an index of refraction of one plus seven times the parameter A-sub-e times the square of the field, and light polarised perpendicular to it one plus four times that parameter times the square of the field, so the difference is three times A-sub-e times the square of the field. With A-sub-e equal to 1.32 times ten to the minus twenty-four per tesla squared, the Euler-Kockel Lagrangian gives a birefringence of 2.5 times ten to the minus twenty-three at 2.5 tesla.Sect. 1, Eqs. (2) and (4); Sect. 2.2.1, Eq. (37)
The four-field photon interaction that produces the effect has already been seen in two other regimes. Light-by-light scattering at high energies has been observed by the ATLAS experiment at the Large Hadron Collider in peripheral lead-lead collisions, and optical polarimetry of an isolated neutron star led Mignani and colleagues to publish evidence of vacuum magnetic birefringence. What remains is a direct laboratory verification at low photon energy, at the macroscopic level.Sect. 1, Introduction
The final PVLAS-FE apparatus was a polarimeter on a single vibration-isolated bench in a clean room at Ferrara: a 1064 nanometre laser locked to a Fabry-Perot cavity 3.303 metres long with a measured finesse of 770,000, the beam passing through the bores of two rotating permanent dipole magnets of 2.5 tesla and 0.82 metres field length each, with the signal read at twice the magnet rotation frequency. It reached an optical path difference sensitivity of 3.5 times ten to the minus nineteen metres per root hertz at 16 hertz.Sect. 5.1, Summary, and Sect. 5.2
After a total run time of about five million seconds the experiment’s final limits are a vacuum magnetic birefringence of twelve plus or minus seventeen times ten to the minus twenty-three, and a vacuum magnetic dichroism whose absolute value is ten plus or minus twenty-eight times ten to the minus twenty-three, both at 2.5 tesla. These are the current best limits on these quantities obtained by optical means, and the integrated noise level is a factor of about seven larger than the predicted effect.Sect. 8.1, Eqs. (214) and (215)
The same polarimetric measurements bound physics beyond the Standard Model. They set laboratory exclusion limits at 95 percent confidence on axion-like particles in the coupling-versus-mass plane, and on millicharged particles in both fermion and scalar form — the fermion plot applying to all types of neutrinos and limiting their charge to less than about three times ten to the minus eight of the electron charge for masses below ten millielectronvolts. Chameleon fields, the screened scalars proposed as a source of dark energy, behave as axion-like particles with respect to photons in an experiment of PVLAS type.Sect. 8.2, and Sect. 2.5, Chameleons and Dark energy
The limit was set by wide-band noise intrinsic to the Fabry-Perot cavity, which the authors believe to be of thermal origin, and which does not improve with higher finesse — so the way forward is to increase the field-length integral rather than the optics. Of the three routes the paper costs out, the most attractive uses a spare Large Hadron Collider dipole at 9 tesla over 14.3 metres of field length, giving about 1158 tesla squared metres, which would in principle reach a signal-to-noise ratio of one in less than a day. A new collaboration is coalescing, a Letter of Intent has been submitted to CERN, and initial feasibility testing is underway.Sect. 9, Conclusions, Eq. (220) and the following list

The way in

https://doi.org/10.1016/j.physrep.2020.06.001

Historical source-access record. This guide links to the original; it does not reproduce the paper text.

LICENCE. Two independent records establish the Creative Commons licence. The arXiv posting of this paper, arXiv:2005.12913v1, submitted 26 May 2020, carries CC BY 4.0 on its abstract page. The Crossref licence record for the version of record in Physics Reports, volume 871 (2020), pages 1 to 74, also names creativecommons.org/licenses/by/4.0, and the Unpaywall record for this DOI reports it as hybrid open access under cc-by. The two Elsevier text-and-data-mining URLs that accompany the Crossref record are not licences and are disregarded. TEXT. The publisher’s PDF is served behind a block, so the text below is taken from the author version on arXiv, which is the copy carrying the explicit CC BY 4.0 grant. It is a seventy-four-page review of roughly fifty-one thousand words, far beyond this library’s length limit, so the abstract, the complete introduction and the complete conclusions are reproduced in full, together with the load-bearing passages on the predicted effect, the measurement method, the final apparatus and the results, and the omissions are marked where they fall. Running heads, page numbers, table and figure furniture and reference-number markers are dropped; the 175-item bibliography is omitted and the complete list is at the source. The review’s sixty-four figures are optical layouts, noise spectra and exclusion plots that cannot be reproduced as text. Displayed equations reached the library with Greek letters, vector arrows and several subscripts lost in extraction, so they are given as named results in plain words with their original numbers, and inequalities are written in words.

How to cite it

A. Ejlli, F. Della Valle, U. Gastaldi, G. Messineo, R. Pengo, G. Ruoso, G. Zavattini (2020) The PVLAS experiment: A 25 year effort to measure vacuum magnetic birefringence doi:10.1016/j.physrep.2020.06.001

Where it sits in the curriculum

What the vacuum is