Click Here to Return to Risk Evaluation Forum
Home Page
Discussion of the Problem
Experiments at the Large Hadron Collider (LHC) at the CERN research facility in Europe just might destroy Earth. There has been controversy about this for years. Some aspects of both the controversy and of the physics on which it is based have been published in the popular press and in peer-reviewed journals. Some of these publications are cited in our reference section, accessible from our home page.
The LHC uses a ring of super-conducing magnets, over five miles in diameter, to accelerate two beams of particles moving in opposite directions. The beams intersect at four points, points surrounded by enormous particle detectors. At those points, particles collide with tremendous energy. When they collide their energy is converted into new particles, including, it is hoped, particles never before seen by scientists. Some theories suggest that the LHC might create particles that could be dangerous: micro black holes that might swallow Earth, and strangelets that could catalyze conversion of Earth into a small ball of strange matter.
The risk has never been high because it is based on speculative, but respectable, physics, and because there have been several safety considerations. However, successive safety considerations have evaporated, and have had to be reinforced by successive safety studies. See "Evaporating Safety Considerations" below. The risk is a bit higher right now because 1) the LHC has been rebuild to function at a higher energy level. Uneventful operation at its previous energy level does not demonstrate that this new level is safe, and 2) A new physics paper predicts creation of black holes at the new energy level, if certain theories and parameters are true. (See [Alia et al] in our reference section.)
Evaporating Safety Considerations
At first, particle colliders seemed safe because there were no known failure modes. An environmental impact statement for the LHC mentioned radiation, but this would be controlled because the LHC was to be located underground.
Micro black holes are impossible
In 1999 a letter published in Scientific American asked whether upcoming colliders might create micro black holes. This was answered by calculations that showed that black hole creation would require energy beyond the reach of any collider. At about the same time, several physics papers based on new string theory predicted black hole production at colliders, eliminating the safety consideration that such production was impossible. [Giddings & Thomas], [Dimopoulos & Landsberg], [Landsberg], [Arkani-Hamed et al], & [Kanti] (See these citations in our reference section.)
Hawking radiation will make black holes go away
The black hole papers contained the safety consideration that micro black holes would evaporate via Hawking radiation, a prediction echoed by a safety study commissioned by CERN. [Blaizot et al] Shortly afterward, several papers were published that questioned whether Hawking radiation, which has never been observed, would work as predicted. [Helfer], [Unruh & Schützholdand], and [Belinski] These papers were independent of the collider controversy, but they reduced the validity of Hawking evaporation as a safety consideration.
Strangelets are positive
Another potential collider product was a strangelet, a particle of strange matter which might catalyze conversion of normal matter into more strange matter, turning Earth into a small ball of strange matter. A safety study for an earlier collider said that this was unlikely because strangelets would have positive charge and not attract normal matter. [Madsen], and [Busza et al] A subsequent physics paper predicted that strangelets would have negative charge. [Peng et al]
An analogy with cosmic rays shows colliders to be safe
Another safety consideration was that cosmic rays with more energy than the LHC have apparently been hitting
Earth since its formation, and we are still here. By analogy, if they didn't harm us, the LHC will not harm us
either. However, micro black holes made by cosmic rays would retain the momentum of the cosmic ray. If they behave
like neutrinos, most would zip right through Earth without hitting anything. Even if they do hit a particle, they
would absorb that particle and continue. A black hole with the momentum of a cosmic ray would have to absorb thousands
of particles to slow below escape velocity from Earth. Given neutrino-level collisions, the bionomial probability of
any cosmic-ray-created black hole ever accreting enough particles to slow below escape velocity, even given creation of billions of black holes over Earth's history, is very small. However, the LHC collides two particles moving in opposite directions. Their momentum would cancel. In most cases it would not cancel precisely, leaving enough momentum so that most micro black holes made by this process would still be moving faster that escape velocity. However, calculations indicate that hundreds per year would be moving slower than escape velocity. [Based on discussions between Blodgett and Landsberg, formerly posted here.] These would be captured by Earth. Therefore this safety consideration did not survive scrutiny in this form, and required the attention of Giddings amd Mangano, discussed below.
After loss of these safety considerations, CERN, to their credit and the credit of a few critics who badgered them, commissioned a new safety study. [Ellis et al] As part of this new study, Giddings and Mangano published a paper that developed a new safety consideration involving white dwarf stars and neutron stars, bodies whose high gravity should reliably capture micro black holes, breathing new life into the cosmic ray consideration. [Giddings & Mangano] Subsequently two scientists posted papers positing separate models in which colliders do cause trouble despite this safety consideration. [Plaga] and [Rössler]. Also see [Penrose] and [Rahman (LHC Black Hole Risks)] in our forum section.
- - - -
[Johnson] reports a history of much of this.
Conclusion
We contend that the risk was never high because science that enables risk was always speculative. We content that this risk was reduced by Giddings and Mangano because circumventing their new safety consideration requires a new level of speculative science. However, speculative science that enables trouble exists. The risk is not zero because normal science is sometimes wrong [Ord at al] and speculative science is sometimes true. Risk assessment for this issue has been far from best practices. [Leggett, 2008] New theory predicting black hole production at the energy level of the rebuilt LHC increases risk. [Ali et al]
Policy Implications
Policy implications are complex. Some experiments should never be conducted. An experiment with even a small probability of destroying Earth seems in this class. The negative expected value (value times probability) is still humongous even if the probability is low because the value to be lost is seven billion human lives, plus all future humans, plus sentient animals and our beautiful biosphere. However, risks should be compared with benefits. CERN is not making this comparison, because they claim that there are no risks. Basic research sometimes produces unexpected benefits, sometimes large benefits. The results of dangerous science just might save us from other risks, or just might enable expansion of humanity into the universe, enabling a very large number of human lives. But is dangerous science the path and the only path to these benefits? They do not seem a likely result of LHC research, and CERN, to their credit, promises none of them. Our estimate of the effectuality of unknown science will at best be highly subjective and speculative. The existential conundrum in the Sartre sense is that we are forced to choose some path, but we can only see a short way along that path. Another criteria for choice is deontological ethics. Thou shall not kill. In this context, that implies: thou shall not risk human lives, especially all of them, unless that risk reduces other risks. Unfortunately, as imperfect humans we are unlikely to handle this optimally.