Physics Diary

Metastable de Sitter Solutions in 2-dimensions

To obtain stable de Sitter solutions in string theory that also avoid the Swampland is no easy task. It is difficult for many reasons. To give one example, from 10-dimensional Type IIA* and Type IIB* theories, one finds de Sitter related solutions but they come with ghosts that give a negative cosmological constant. Full de Sitter spacetime is a Lorentzian manifold that is the coset space of Lorent groups, dS^d \simeq O(d,1)/O(d-1,1), the cosmology of which, as many readers will likely be familiar, is based on a positive cosmological constant. Hence, in this example, we obtain the wrong sign. Many other examples can be cited in reference to the difficulty we currently face.

To speak generally, it has gotten to the point historically that one possibility being considered is that string theory may simply conspire against de Sitter space – that is, there is some deep incompatibility between our leading theory of quantum gravity and de Sitter vacua. The No-de Sitter Conjecture is an example of an attempt to formalise such a logical possibility, motivating beyond other things the need to understand better the structure of vacua in string theory. This conjecture is by no means rigorous, but it is supported by the fact that historically de Sitter solutions have been elusive. Having said that, one must be cautious about any claims regarding de Sitter constructions, either from the perspective of obtaining de Sitter solutions or from arguments claiming de Sitter belongs to the Swampland. Our present theory is still very much incomplete and there is a lot of work to be done, and I think it is fair to suggest that the fate of de Sitter in string theory is still uncertain. As an expression of opinion, I would currently say that it is more likely de Sitter will come from something much more fundamental than some of the current strategies being proposed. It should also be kept in mind that there may be a larger issue here. Quantum gravity in de Sitter is fundamentally difficult beyond pure string theory reasons. Indeed, it is possible that de Sitter is simply unstable in quantum gravity – that it is a space that simply cannot exist quantum mechanically. For instance, think of complications in QFT in curved space. Indeed, there is a lengthy list of authoritative papers that one may cite when it comes purely to perturbative quantum gravity (see, for example, this page for further references).

Image: de Sitter geometry courtesy of

Given the sociology and history, when a new paper appears claiming to have found de Sitter solutions, one takes notice. That is precisely what happened last week when Miguel Montero, Thomas Van Riet, and Gerben Venken uploaded to the archive their recent work claiming to have found metastable de Sitter solutions in lower dimensions. More precisely, these parametrically controlled solutions appear when compactifying from 4-dimensions to 2-dimensions, particularly as a result of some clever work that invokes abelian p-1-form gauge fields to stabilise the runaway potential giving in general dS_{d-p} \times S^{p} solutions. However, not all solutions are stable with controlled saddle points. Indeed, the authors find D - p > 2 solutions to be generically unstable. On the other hand, the instability in the homogeneous mode disappears when D - p = 2. In this 2-dimensional case, the solutions relate to the near horizon dS_2 \times S^2 geometry of Nariai black holes. It is also worth pointing out that these solutions are not in string theory, but the work highlights some interesting implications and raises some important questions in quantum gravity more generally, including also when it comes to the Swampland.

If I am not mistaken, I think I remember seeing similar kinds of ideas in studies of dS / CFT where people look into cases of rotating Nariai black holes maximally large at dS_2 \times S^2. The geometry is generally interesting because in that some Nariai constructions with quintessence generalise quasi-de Sitter solutions, utilising for example decompositions from 4-dimemensions to things like 2-dimensional dilaton gravity theory (this may make one recall SYK models), one will often find descriptions of the geometric space as ds^2 = \Gamma(\theta) [-d\tau^2 +\cosh^2 \frac{\tau}{l} d\psi^2 + \alpha(\theta) d\theta^{2}] + \gamma(\theta) (d\phi + k \tanh \frac{\tau}{l} d\psi)^2 which we can think of in terms of an S^2 fibered over a dS_2 base space. I kind of see it as a bit of a playground useful to experiment with and probe. In section 2 of the de Sitter paper, one can read a short overview of Nariai black hole solutions of Einstein-Maxwell theory as a precursor to the main study.

I’ve only had time to skim through the Montero et al. paper, so I still need to give it careful attention. But on quick glance I noticed a number of interesting calculations, not least the derived inequality (p -1) \mid \frac{V^{\prime}}{V} \mid \leq \mid \frac{f^{\prime}}{f} \mid that states how, in order for a dS_{d-p} \times S^{p} saddle point to exist, this constraint that relates the gauge kinetic function of the p−1-form field and the potential must be satisfied. This is an example of one of a handful of points to which I would like to go back and think about more deeply. Additionally, it seems an open question whether, if we can obtain controlled lower-dimensional de Sitter solutions from runaway potentials, do such approaches and constructions fully escape the Swampland? It is worth reading their paper with that in mind.

As alluded at the outset, one should always approach any claims about de Sitter with a healthy dose of scepticism. At the same time, van Riet is a physicist who I admire, because he is one of a number championing the need for greater mathematical rigour in string theory and quantum gravity moving forward. So, for me, this paper immediately comes with some weight and authority. If one thing is certain today, at least from my current vantage, it is the need for thoughtful caution and careful mathematical analysis; and it appears that on a few occasions the authors also stress this point in their analysis.