The SIMP Miracle
10/28/2014
691 citations (654 excluding self-citations). Established a new paradigm for thermal relic dark matter, recognized by the Particle Data Group alongside WIMPs and FIMPs as a standard dark matter mechanism.
The Problem
The standard picture of thermal relic dark matter, the WIMP miracle, assumes dark matter particles annihilate in pairs (2-to-2) into Standard Model particles, freezing out when the expansion rate of the universe exceeds the annihilation rate. This mechanism predicts dark matter masses in the GeV-TeV range with weak-scale interaction cross-sections. By 2014, direct detection experiments (LUX, XENON) and the LHC had excluded large portions of the WIMP parameter space without a discovery. The WIMP miracle was not dead, but the unconstrained territory was shrinking.
Meanwhile, astrophysical observations were hinting at a separate problem. N-body simulations of cold dark matter predicted cuspy density profiles in galaxy centers, but observations showed flat cores. Simulations predicted too many massive satellite galaxies and too much substructure. These “small-scale structure problems” could be resolved if dark matter had sizable self-interactions, but WIMPs at the weak scale generically have negligible self-interaction cross-sections.
The Key Idea
The SIMP mechanism changes the freeze-out arithmetic by replacing the standard 2-to-2 annihilation with 3-to-2 number-changing processes within the dark sector. Three dark matter particles interact, producing two dark matter particles plus kinetic energy. The dark sector stays in thermal contact with the Standard Model through elastic scattering, which maintains kinetic equilibrium, but the relic abundance is set by the 3-to-2 rate rather than by annihilation into visible particles.
The dimensional analysis shifts dramatically. The 3-to-2 cross-section scales differently with mass than 2-to-2, and the thermal relic calculation points to dark matter masses in the MeV-GeV range rather than the GeV-TeV range of WIMPs. At these masses, the same strong self-coupling that drives the 3-to-2 process also produces sizable 2-to-2 self-interactions, exactly what the small-scale structure observations require. The mechanism connects two seemingly independent problems: the relic abundance and the self-interaction scale.
Impact
The SIMP mechanism entered the Particle Data Group’s Review of Particle Physics as a standard dark matter paradigm, appearing in all four editions from 2016 through 2024. Tulin and Yu’s review of self-interacting dark matter (1,260 citations) treats the SIMP mechanism as a primary theoretical motivation for dark matter self-interactions. The comprehensive dark matter reviews by Arcadi et al. (“The Waning of the WIMP?”, 1,080 citations) and Roszkowski et al. (842 citations) discuss SIMPs as a leading alternative to the WIMP paradigm.
The sub-GeV mass prediction directly motivated new experimental programs. The SENSEI collaboration (477 citations) developed skipper-CCD detectors with sensitivity to sub-GeV dark matter through electron scattering, a technique designed for exactly the mass range SIMPs predict. Essig, Mardon, and Volansky’s proposal for semiconductor-target direct detection (528 citations) was motivated in part by the SIMP mass window. The US dark matter community planning documents, Cosmic Visions 2017 (914 citations) and the Dark Sectors 2016 workshop report (709 citations), cite the SIMP mechanism as a key driver for expanding the experimental program beyond the traditional WIMP search window.
The companion paper, “Model for Thermal Relic Dark Matter of Strongly Interacting Massive Particles” (441 citations), provided the first concrete realization: a QCD-like hidden sector where pseudo-Nambu-Goldstone bosons are the dark matter and the Wess-Zumino-Witten term mediates the 3-to-2 process.
Recollections
The SIMP Miracle is the longest project from conception to completion I ever worked on: six years from first idea to publication.
It started in late 2007. I felt I didn’t truly understand the WIMP miracle, so I worked out a clean five-step derivation that gave mDM ~ α(mp η MPl)1/2 ~ 1 TeV. Once I had that, I did what any scientist does: tried it under different circumstances. If dark matter had no 2-to-2 annihilation channel to the Standard Model, the leading process would be 3-to-2 within the dark sector. Working the math gave mDM ~ α(mp2 η2 MPl)1/3 ~ α × 1 GeV. Strong interactions at the strong interaction scale give the right dark matter abundance. I tongue-in-cheek dubbed it the SIMP miracle.
The problem was thermodynamics. When SIMPs annihilate via 3-to-2, they heat up the dark sector relative to the Standard Model, breaking the assumptions of the calculation. I needed a “steam release valve”: elastic scattering (DM SM → DM SM) to maintain kinetic equilibrium. But any interaction strong enough to thermalize the sectors also enables DM DM → SM SM, which is just the WIMP miracle and overrides the SIMP mechanism. I couldn’t make a model work.
My graduate student Tomas Rube challenged the dynamical assumptions and did the first numerical calculations over the winter of 2008-2009. We found a related 1980s paper by Lawrence Hall, who had taught me quantum field theory and was the advisor to my advisor. Hall had proposed a similar 3-to-2 idea but without the steam release valve. Tomas eventually started working on a no-go theorem. In June 2009, he transitioned to another project. I kept the idea in the mental filing cabinet.
For three years I showed the mechanism to senior colleagues. Savas Dimopoulos didn’t love the aesthetics. Nima Arkani-Hamed was enthusiastic but didn’t see a way out. Most people were skeptical or thought I was pulling a fast one with non-standard derivations. I stopped talking about it.
In spring 2013, I visited Tomer Volansky at Tel Aviv University, where Eric Kuflik and Yonit Hochberg were postdocs. After a day of talking trash about various trends, Tomer said something like “what the fuck makes us think we’re so great, we’re just sitting around with no new ideas.” So I pulled the SIMP derivation out of the filing cabinet. Instead of glossing over the steam release valve problem, Tomer said: “I think we can solve this, this doesn’t seem so hard.” We filled blackboards for several days and left with messy notes that we thought contained a solution, though we couldn’t cleanly rederive it. The key was Tomer’s interest in light dark matter scattering off electrons.
Back at Stanford, I found the simple reason it worked. The ratio between the rates of (DM SM → DM SM) and (DM DM → SM SM) differs by a factor of nDM/nSM. I had always tried coupling dark matter to baryons, where this ratio is roughly 1. But if dark matter couples to electrons, photons, or neutrinos, the ratio is suppressed by the baryon-to-photon ratio η ~ 10-10. The steam release valve works because the thermalization process is 10 billion times faster than the annihilation process. Five years of delay for that one realization.
Eric and Yonit did the majority of the hard work on the paper. Eric validated our back-of-the-envelope calculations with real numerics. Yonit explored experimental limits and did much of the writing. My daughter was born in July 2013 and we were actively drafting around that. We published in February 2014. By the end of the year it had 13 citations — not bad, but nothing extraordinary.