Einstein was wrong

Disentangling Data – Disproving Einstein’s Theory of Locality

Making the intangible tangible is no mean feat. In October 2015, Professor Ronald Hanson and his team, including featured PhD student Bas Hensen,  proved Einstein’s theory on locality to be wrong. They demonstrated that the so-called ‘entanglement’ of particles at a large distance exists. In other words, objects separated by great distance can instantaneously affect each other’s behaviour, without being physically connected by another medium. With diamonds set up in identical labs 1.3 km apart on opposite sides of the TU Delft campus, the team proved the phenomenon.

Scientists had been using the methodology (the so-called Bell Test) to try and disprove Einstein’s theory of locality since 1964. However, on every occasion the performed experiments had contained loopholes that allowed critics to say that there may be other physical connections (‘hidden variables’ in Einstein’s terminology) between the particles. Thus the predictions of quantum mechanics could never quite be verified.

It was a significant discovery

That was until the work done by the Delft-led team demonstrated there could be no physical connection between the entangled particles. It was a significant discovery that gained much attention both in scholarly and general audiences.

Have data open for proof and check for consistency

Bas Hensen

Bas Hensen

Bas Hensen: “We knew we were working on a high impact paper and realised there would be requests for the raw data, so that the experiment could be validated and the data checked for consistency. Given that scientists had been using this experimental method since the 1960s, and results had always been contested, there was a tradition for of sharing data related to this experiment. So we knew from the start we would open up the data. As a comment with the preprint in arXiv we already stated that “raw data will be made available after publication”. And there is no mystery around the data. It is a small dataset of an experiment that takes time to execute. There is not too much in there that we did not get out of it already. In a way the experiment is quite standard but it is difficult to execute, so there was no chance to cheat. But it’s good to have the proof online for all to check.”

Open up in 4TU.Centre for Research Data archive

Since they already knew they wanted to open up the raw data, an e-mail by 4TU.ResearchData offering their archiving services was very welcome.

Their initial plan was to deposit the data at Nature’s Scientific Data. Since this is a data journal they would need to submit an extensive description in a prescribed format. They could have added the data as supplementary material to their publication in Nature, but the article was already in print.

So they decided to go for the more efficient way of publishing in the data archive of 4TU.ResearchData. 4TU.Research Data is the data repository for TU Delft; any Delft researcher can deposit, free of charge, up to a 100GB for long-term archiving of data.

Hosted at the TU Delft Library, and run on behalf the two other founding members (the universities of Eindhoven and Twente) of the 4TU Federation, 4TU.ResearchData has been in existence since 2008. It now has over 7,000 datasets, archived and openly available for others to re-use.

Making use of the library’s connection with the standards body DataCite, each deposited dataset receives a stable DOI (Digital Object Identifier). This means that the researcher can cite the data’s existence in any article, and be sure that the link will work on a long-term basis.

The dataset is gaining interest

Bas Hensen and Professor Hanson in their lab

Bas Hensen and Professor Hanson in their lab

Data Usage Following Publication

Professor Hansen and his team are happy with the possibility to link to their publication in Nature for further description of the data and having a DOI . Now they only have to send the DOI to people who are interested in their raw data.

A couple of months since its publication and the dataset is already gaining interest. In the first six months since its deposit, the first dataset has been viewed 650 times. The second dataset has been viewed 56 times in the first three weeks. Bas: “This figure meets our expectations. It shows that nearly all of the world’s other research groups involved in experimental quantum mechanics have accessed the dataset.”






More information: library@tudelft.nl

Photography: Frank Auperlé

TU Delft Library