Particle accelerator accidents cause extinction—black holes won’t evaporate, high energy densities create vacuums, and strange matter is more stable than regular matter

Particle physics is the study of the elementary constituents of matter and radiation, and the interactions between them. A major experimental method in particle physics involves the use of particle accelerators such as the RHIC and LHC to bring beams of particles to near the speed of light and then collide them together. This focuses a large amount of energy in a very small region and breaks the particles down into their components, which are then detected. As particle accelerators have become larger, the energy densities achieved have become more extreme, prompting some concern about their safety. These safety concerns have focused on three possibilities: the formation of ‘true vacuum’, the transformation of the earth into ‘strange matter’, and the destruction of the earth through the creation of a black hole. 4.1 True vacuum and strange matter formation The type of vacuum that exists in our universe might not be the lowest possible vacuum energy state. In this case, the vacuum could decay to the lowest energy state, either spontaneously, or if triggered by a sufficient disturbance. This would produce a bubble of ‘true vacuum’ expanding outwards at the speed of light, converting the universe into different state apparently inhospitable for any kind of life (Turner and Wilczek 1982). Our ordinary matter is composed of electrons and two types of quarks: up quarks and down quarks. Strange matter also contains a third type of quark: the ‘strange’ quark. It has been hypothesized that strange matter might be more stable than 11 normal matter, and able to convert atomic nuclei into more strange matter (Witten 1984). It has also been hypothesized that particle accelerators could produce small negatively charged clumps of strange matter, known as strangelets. If both these hypotheses were correct and the strangelet also had a high enough chance of interacting with normal matter, it would grow inside the Earth, attracting nuclei at an ever higher rate until the entire planet was converted to strange matter — destroying all life in the process. Unfortunately strange matter is complex and little understood, giving models with widely divergent predictions about its stability, charge and other properties (Jaffe, Busza et al. 2000). One way of bounding the risk from these sources is the cosmic ray argument: the same kind of high-energy particle collisions occur all the time in Earth’s atmosphere, on the surface on the Moon and elsewhere in the universe. The fact that the Moon or observable stars have not been destroyed despite a vast number of past collisions (many at much higher energies than can be achieved in human experiments) suggest that the threat is negligible. This argument was first used against the possibility of vacuum decay (Hut and Rees 1983) but is quite general. An influential analysis of the risk from strange matter was carried out in (Dar, De Rujula et al. 1999) and formed a key part of the safety report for the RHIC. This analysis took into account the issue that any dangerous remnants from cosmic rays striking matter at rest would be moving at high relative velocity (and hence much less likely to interact) while head-on collisions in accelerators could produce remnants moving much at much slower speeds. They used the rate of collisions of cosmic rays in free space to estimate strangelet production. These strangelets would then be slowed by galactic magnetic fields and eventually be absorbed during star formation. When combined with estimates of the supernova rate, this can be used to bound the probability of producing a dangerous strangelet in a particle accelerator. The resulting probability estimate was < 2 ! 10-9 per year of RHIC operation.8 While using empirical bounds and experimentally tested physics reduces the probability of a theory error, the paper needs around 30 steps to reach its conclusion. For example, even if there was just a 10-4 chance of a calculation or modelling error per step this would give a total P(¬A)!"!0.3%. This would easily overshadow the risk estimate. Indeed, even if just one step had a 10-4 chance of error, this would overshadow the estimate. A subtle complication in the cosmic ray argument was noted in (Tegmark and Bostrom 2005). The Earth’s survival so far is not sufficient as evidence for safety, since we do not know if we live in a universe with ‘safe’ natural laws or a universe where planetary implosions or vacuum decay do occur but we have just been exceedingly lucky so far. While this latter possibility might sound very unlikely, all observers in such a universe would find themselves to be in the rare cases where 8 (Kent 2004) points out some mistakes in stating the risk probabilities in different versions of the paper, as well as for the Brookhaven report. Even if these are purely typesetting mistakes, it shows that the probability of erroneous risk estimates is nonzero. 12 their planets and stars had survived, and would thus have much the same evidence as we do. Tegmark and Bostrom had thus found that in ignoring these anthropic effects, the previous model had been overly narrow. They corrected for this anthropic bias and, using analysis from (Jaffe, Busza et al. 2000), concluded that the risk from accelerators was less than 10-12 per year. This is an example of a demonstrated flaw in an important physics risk argument (one that was pivotal in the safety assessment of the RHIC). Moreover, it is significant that the RHIC had been running for five years on the strength of a flawed safety report, before Tegmark and Bostrom noticed and fixed this gap in the argument. Although this flaw was corrected immediately after being found, we should also note that the correction is dependent on both anthropic reasoning and on a complex model of the planetary formation rate (Lineweaver, Fenner et al. 2004). If either of these, or the basic Brookhaven analysis is flawed, the risk estimate is flawed. 4.2 Black hole formation The Large Hadron Collider experiment at CERN was designed to explore the validity and limitations of the Standard Model of particle physics by colliding beams of high energy protons. This will be the most energetic particle collision experiment ever done, which has made it the focus of a recent flurry of concerns. Due to the perceived strength of the previous arguments on vacuum decay and strangelet production, most of the concern about the LHC has focused on black hole production. None of the theory papers we have found appears to have considered the black holes to be a safety hazard, mainly because they all presuppose that any black holes would immediately evaporate due to Hawking radiation. However, it was suggested by (Dimopoulos and Landsberg 2001) that if black holes form, particle accelerators could be used to test the theory of Hawking radiation. Thus critics also began questioning whether we could simply assume that black holes would evaporate harmlessly. A new risk analysis of LHC black-hole production (Giddings and Mangano 2008) provides a good example of how risks can be more effectively bounded through multiple sub-arguments. While never attempting to give a probability of disaster (rather concluding "there is no risk of any significance whatsoever from such black holes") it uses a multiple bounds argument. It first shows that rapid black hole decay is a robust consequence of several different physical theories (A1). Second it discusses the likely incompatibility between non-evaporating black holes and mechanisms for neutralising black holes: in order for cosmic ray–produced stable black holes to be innocuous but accelerator-produced black holes to be dangerous, they have to be able to shed excess charge rapidly (A2). Our current understanding of physics suggests both that black holes decay and that even if they didn’t, they would be unable to discharge themselves. Only if this understanding is flawed will the next section come into play. 13 The third part, which is the bulk of the paper, models how multidimensional and ordinary black holes would interact with matter. This leads to the conclusion that if the size scale of multidimensional gravity is smaller than about 20 nm, then the time required for the black hole to consume the Earth would be larger than the natural lifetime of the planet. For scenarios where rapid Earth accretion is possible, the accretion time inside white dwarves and neutron stars would also be very short, yet production and capture of black holes from impinging cosmic rays would be so high that the lifespan of the stars would be far shorter than the observed lifespan (and would contradict white dwarf cooling rates) (A3). While each of these arguments have weaknesses the force of the total argument (A1,A2,A3) is significantly stronger by the combination of them. Essentially the paper acts as three sequential arguments, each partly filling in the grey area (see figure 1) left by the previous. If the theories surrounding black hole decay fail, the argument about discharge comes into play, and if against all expectation black holes are stable and neutral the third argument shows that astrophysics constrains them to a low accretion rate.