Exotic PeVatrons: Galactic High-Energy Sources

• Exotic PeVatrons are Galactic particle accelerators that produce photons above 100 TeV and include sources outside traditional supernova remnants.
• Observations identify candidates in diverse environments such as pulsar vicinities, massive star clusters, superbubbles, and unidentified regions with hard, uncurved spectra.
• Spectral analyses, neutrino constraints, and morphology studies are critical in distinguishing between leptonic, hadronic, and CR knee PeVatron scenarios.

Exotic PeVatrons are Galactic particle accelerators discussed in connection with the PeV regime outside the narrow identification of the problem with resolved supernova-remnant shells. The shift in emphasis is explicit in review literature: the first detections of PeVatrons were not associated to supernova remnants, and “all clearly detected SNRs have yet revealed to not be pevatrons,” while several sources detected above 100 TeV are linked instead to pulsar environments, star-forming regions, massive star clusters, superbubbles, H II regions, or remain unidentified (Cristofari, 2021). Ground-based gamma-ray astronomy has now established more than a dozen Galactic sources emitting above 100 TeV, including PeV photons from the Crab Nebula and the Cygnus region, but the central interpretive problem remains whether such objects are leptonic PeVatrons, hadronic PeVatrons, or true “CR knee PeVatrons” capable of supplying the Galactic cosmic-ray knee population (Angüner, 2023).

1. Conceptual scope and classification

The recent literature uses several operational definitions of a PeVatron. A leptonic PeVatron accelerates electrons or positrons to at least $10^{15}\,\mathrm{eV}$, a hadronic PeVatron accelerates protons or nuclei to the same scale, and a CR knee PeVatron accelerates bulk hadrons to the rigidity range relevant to the knee of the Galactic cosmic-ray spectrum (Angüner, 2023). This distinction is not terminological detail: it is the reason why sources with photons above 100 TeV are compelling but not automatically decisive evidence for the origin of the bulk Galactic cosmic-ray population.

The label “exotic” is used in more than one sense. In observational work it often denotes sources with no secure multi-wavelength counterpart, sources not cleanly identified with an SNR shell, or source classes not traditionally placed at the center of Galactic cosmic-ray origin models. In theoretical work it extends further, to compact-object or dark-sector scenarios explicitly proposed as new classes of Galactic PeVatrons (Addazi et al., 30 Sep 2025).

A persistent misconception is that the PeVatron question is equivalent to the SNR question. Review work argues against that identification directly: the scientific goal is solving the origin of Galactic cosmic rays, not merely confirming that SNRs are PeVatrons, and the current source census has broadened accordingly (Cristofari, 2021).

2. Unidentified and morphologically complex candidates

A defining observational example is HESS J1702-420A, revealed by deep H.E.S.S. observations of the previously unidentified VHE source HESS J1702-420. Using a 44.9 hr acceptance-corrected livetime and improved high-energy analysis, H.E.S.S. detected gamma-ray photons up to at least 100 TeV for the first time in the history of the experiment. The emission was decomposed into at least two overlapping components: a hard, compact source, HESS J1702-420A, and a steeper, extended component, HESS J1702-420B. For HESS J1702-420A the spectrum is a power law with $$\Gamma = 1.53 \pm 0.19_{\mathrm{stat}} \pm 0.20_{\mathrm{sys}}$$, the flux above 2 TeV is $$F(>2\,\mathrm{TeV}) = (2.08 \pm 0.49_{\mathrm{stat}} \pm 0.62_{\mathrm{sys}})\times 10^{-13}\,\mathrm{cm}^{-2}\,\mathrm{s}^{-1}$$, the extension radius is $$(0.06 \pm 0.02_{\mathrm{stat}} \pm 0.03_{\mathrm{sys}})^\circ$$, the source is detected at $5.4\sigma$, and the $64\!-\!113$ TeV band is measured at $4\sigma$. The spectral fits are well described by simple power laws, with no evidence for curvature or exponential cutoff up to 100 TeV, and the 95% CL lower limit on the proton cutoff energy is $0.55{-}1.16\,\mathrm{PeV}$, depending on the assumed prior on the proton spectral index (Giunti et al., 2021).

Another prominent case is LHAASO J0341+5258, described as an unidentified PeVatron with no clear astrophysical counterpart. It is extended in nature, notably bright, and has a flux $>20\%$ of the Crab Nebula’s flux above 25 TeV. Multiwavelength follow-up sharpened the anomaly rather than resolving it: VERITAS observed the region for $\sim 50$ hours between 2021 and 2023 and found no significant TeV emission below 30 TeV, whereas HAWC reported an $$\Gamma = 1.53 \pm 0.19_{\mathrm{stat}} \pm 0.20_{\mathrm{sys}}$$0 detection, preferred an extended source template over a point source with $$\Gamma = 1.53 \pm 0.19_{\mathrm{stat}} \pm 0.20_{\mathrm{sys}}$$1, and found a best-fit extension of $$\Gamma = 1.53 \pm 0.19_{\mathrm{stat}} \pm 0.20_{\mathrm{sys}}$$2. A faint XMM-Newton PWN candidate and a partially overlapping dense molecular “half-shell” are discussed, but no supernova remnant, bright PWN with a powerful pulsar, or stellar cluster convincingly matches the UHE emission (Bangale et al., 24 Sep 2025).

A third example is 1LHAASO J1857+0203u, in the area of H II region G35.6$$\Gamma = 1.53 \pm 0.19_{\mathrm{stat}} \pm 0.20_{\mathrm{sys}}$$30.5. LHAASO detected the source at $$\Gamma = 1.53 \pm 0.19_{\mathrm{stat}} \pm 0.20_{\mathrm{sys}}$$4 above 100 TeV. In the WCDA band, $$\Gamma = 1.53 \pm 0.19_{\mathrm{stat}} \pm 0.20_{\mathrm{sys}}$$5 TeV emission is extended with $$\Gamma = 1.53 \pm 0.19_{\mathrm{stat}} \pm 0.20_{\mathrm{sys}}$$6 and a power-law index $$\Gamma = 1.53 \pm 0.19_{\mathrm{stat}} \pm 0.20_{\mathrm{sys}}$$7; in the KM2A band above 25 TeV the source becomes point-like, with a power-law index $$\Gamma = 1.53 \pm 0.19_{\mathrm{stat}} \pm 0.20_{\mathrm{sys}}$$8. Modeling disfavors the scenario in which nearby clouds are illuminated by protons escaped from SNR G35.6$$\Gamma = 1.53 \pm 0.19_{\mathrm{stat}} \pm 0.20_{\mathrm{sys}}$$90.4. The GeV–TeV spectrum can be fit in a hadronic H II region scenario with $$F(>2\,\mathrm{TeV}) = (2.08 \pm 0.49_{\mathrm{stat}} \pm 0.62_{\mathrm{sys}})\times 10^{-13}\,\mathrm{cm}^{-2}\,\mathrm{s}^{-1}$$0 and $$F(>2\,\mathrm{TeV}) = (2.08 \pm 0.49_{\mathrm{stat}} \pm 0.62_{\mathrm{sys}})\times 10^{-13}\,\mathrm{cm}^{-2}\,\mathrm{s}^{-1}$$1, but an evolved PWN origin cannot be ruled out (Collaboration, 2024).

3. Massive star clusters, superbubbles, and H II regions

Massive star clusters and superbubbles have moved from peripheral to central status in the PeVatron discussion. Analytical work on second-order Fermi acceleration in superbubbles shows that it can be as efficient as diffusive shock acceleration when comparable relevant velocities are used, with the maximum energy determined by diffusive escape and therefore by the diffusion coefficient $$F(>2\,\mathrm{TeV}) = (2.08 \pm 0.49_{\mathrm{stat}} \pm 0.62_{\mathrm{sys}})\times 10^{-13}\,\mathrm{cm}^{-2}\,\mathrm{s}^{-1}$$2, the Alfvén velocity $$F(>2\,\mathrm{TeV}) = (2.08 \pm 0.49_{\mathrm{stat}} \pm 0.62_{\mathrm{sys}})\times 10^{-13}\,\mathrm{cm}^{-2}\,\mathrm{s}^{-1}$$3, and the superbubble size. For superbubble radii of $$F(>2\,\mathrm{TeV}) = (2.08 \pm 0.49_{\mathrm{stat}} \pm 0.62_{\mathrm{sys}})\times 10^{-13}\,\mathrm{cm}^{-2}\,\mathrm{s}^{-1}$$4 pc, cosmic rays can be accelerated to $$F(>2\,\mathrm{TeV}) = (2.08 \pm 0.49_{\mathrm{stat}} \pm 0.62_{\mathrm{sys}})\times 10^{-13}\,\mathrm{cm}^{-2}\,\mathrm{s}^{-1}$$5 within $$F(>2\,\mathrm{TeV}) = (2.08 \pm 0.49_{\mathrm{stat}} \pm 0.62_{\mathrm{sys}})\times 10^{-13}\,\mathrm{cm}^{-2}\,\mathrm{s}^{-1}$$6 Myr, provided that $$F(>2\,\mathrm{TeV}) = (2.08 \pm 0.49_{\mathrm{stat}} \pm 0.62_{\mathrm{sys}})\times 10^{-13}\,\mathrm{cm}^{-2}\,\mathrm{s}^{-1}$$7 and $$F(>2\,\mathrm{TeV}) = (2.08 \pm 0.49_{\mathrm{stat}} \pm 0.62_{\mathrm{sys}})\times 10^{-13}\,\mathrm{cm}^{-2}\,\mathrm{s}^{-1}$$8 at 100 TeV. The paper applies these conditions to HESS J1646-458 associated with Westerlund 1 and the Cygnus Cocoon associated with Cyg OB2, and concludes that superbubbles should be considered a possible source of Galactic cosmic rays up to, or beyond, a PeV (Vink, 2024).

A complementary theoretical line argues that collective winds of massive star clusters can reconcile recent LHAASO composition data in a way that individual stellar winds cannot. In that model, stellar winds of individual massive stars fail because they yield distinct rigidity cutoffs for protons and helium, whereas collective winds naturally reproduce the common rigidity-dependent spectral break at $$F(>2\,\mathrm{TeV}) = (2.08 \pm 0.49_{\mathrm{stat}} \pm 0.62_{\mathrm{sys}})\times 10^{-13}\,\mathrm{cm}^{-2}\,\mathrm{s}^{-1}$$9 through the mixing effect of stars at different evolutionary stages. The proposed stellar-dominated picture assigns dominance to SNRs in the GeV–TeV range, to individual stellar winds in the TeV range, and to collective cluster winds in the PeV knee region. It also predicts hardening of carbon and oxygen around $$(0.06 \pm 0.02_{\mathrm{stat}} \pm 0.03_{\mathrm{sys}})^\circ$$0, and no analogous hardening of magnesium in the multi-TV range (Qiu et al., 29 May 2026).

H II regions enter this landscape because they provide dense gas, possible embedded stellar populations, and environments in which hadronic gamma-ray production can be efficient. The G35.6$$(0.06 \pm 0.02_{\mathrm{stat}} \pm 0.03_{\mathrm{sys}})^\circ$$10.5 study explicitly raises the possibility that, if confirmed, this would be the first H II region with PeV-class gamma-ray emission (Collaboration, 2024). This suggests that the “exotic” label increasingly marks environment and source evolution rather than merely a lack of identification.

4. ULX winds, microquasars, and hidden remnants

A distinct compact-object branch of the literature proposes Ultra-Luminous X-ray sources as a new class of Galactic PeVatrons. In this picture, super-Eddington accretion onto stellar-mass compact objects powers winds with $$(0.06 \pm 0.02_{\mathrm{stat}} \pm 0.03_{\mathrm{sys}})^\circ$$2 and kinetic luminosities $$(0.06 \pm 0.02_{\mathrm{stat}} \pm 0.03_{\mathrm{sys}})^\circ$$3, inflating wind bubbles with strong wind termination shocks. The transport equation is solved in spherical symmetry, and the approximate maximum proton energy in the optimistic Bohm-diffusion limit is written as

$$(0.06 \pm 0.02_{\mathrm{stat}} \pm 0.03_{\mathrm{sys}})^\circ$$4

Applied to SS 433, the model finds $$(0.06 \pm 0.02_{\mathrm{stat}} \pm 0.03_{\mathrm{sys}})^\circ$$5 PeV, argues that high-energy protons in the bubble might explain the $$(0.06 \pm 0.02_{\mathrm{stat}} \pm 0.03_{\mathrm{sys}})^\circ$$6 TeV photons and morphology observed by LHAASO, and predicts a neutrino flux close to the 10-year discovery threshold for KM3NeT for extended sources (Peretti et al., 2024).

A related fossil-source scenario treats microquasar remnants as hidden or dark PeVatrons. In this framework, a former super-Eddington microquasar shuts off after mass transfer ceases, but a $$(0.06 \pm 0.02_{\mathrm{stat}} \pm 0.03_{\mathrm{sys}})^\circ$$7 pc cocoon inflated during the active phase remains filled with cosmic rays accelerated at the terminal reverse shocks. Diffusion is suppressed in the cocoon, with $$(0.06 \pm 0.02_{\mathrm{stat}} \pm 0.03_{\mathrm{sys}})^\circ$$8, so the particles remain confined for $$(0.06 \pm 0.02_{\mathrm{stat}} \pm 0.03_{\mathrm{sys}})^\circ$$9 years and then illuminate nearby molecular clouds via $5.4\sigma$0 collisions. The proposal is that remnants of former super-Eddington systems can release particles up to $5.4\sigma$1 PeV and produce gamma rays reaching hundreds of TeV while remaining faint in radio and X-rays, thereby accounting for part of the unidentified LHAASO population (Abaroa et al., 8 Dec 2025).

These compact-object and remnant scenarios differ from cluster and superbubble models in timescale and morphology. Cluster models emphasize ongoing collective acceleration in extended star-forming environments, whereas ULX and microquasar-remnant models emphasize confined bubbles, fossil cosmic-ray reservoirs, and gamma-ray illumination of nearby gas (Peretti et al., 2024, Abaroa et al., 8 Dec 2025).

5. Spectral criteria, hadronic diagnostics, and systematic effects

The observational grammar of PeVatron identification is dominated by the spectrum. In H.E.S.S. work the differential photon spectrum is parameterized as

$5.4\sigma$2

and a hard power law with $5.4\sigma$3, extending without curvature or cutoff to $5.4\sigma$4 TeV, is treated as a strong indicator of PeV-scale parent particles (Giunti et al., 2021). This criterion underlies the status of HESS J1702-420A, whose lack of curvature is one reason it is considered one of the most compelling Galactic PeVatron candidates in H.E.S.S. data (Giunti et al., 2021).

At the same time, the literature explicitly warns against a one-step inference from $5.4\sigma$5 TeV photons to hadronic cosmic-ray acceleration. The Crab Nebula is described as an unambiguously established electron PeVatron, and the review literature distinguishes leptonic, hadronic, and CR-knee PeVatrons precisely because the same gamma-ray data can admit different parent populations (Cristofari, 2021, Angüner, 2023). In HESS J1702-420A, the leptonic interpretation requires very hard electron spectra with $5.4\sigma$6 and electron energies $5.4\sigma$7 TeV, which challenges standard cooling and energetics arguments for typical PWNe (Giunti et al., 2021).

Morphology provides an independent discriminator. For G106.3+2.7, Tibet AS+MD detected gamma rays above 10 TeV up to and above 100 TeV, with a centroid offset by about $5.4\sigma$8 from PSR J2229+6114 and coincident with a molecular cloud. The gamma-ray emission above 10 TeV is therefore well correlated with the cloud rather than the pulsar, favoring a hadronic origin via $5.4\sigma$9 decay over a leptonic inverse-Compton interpretation. The paper emphasizes that an X-ray flux upper limit on the synchrotron spectrum would be decisive for firmly establishing the hadronic scenario (Amenomori et al., 2021).

Systematic propagation effects further complicate interpretation. Very-high-energy gamma rays are attenuated by pair production on the ISRF and CMB, so the observed spectrum is related to the intrinsic spectrum by

$64\!-\!113$0

Using three-dimensional ISRF models, one study finds that spectral cutoffs can be underestimated by factors of $64\!-\!113$1 to $64\!-\!113$2 in the energy range so far sampled by TeV gamma-ray telescopes. For the H.E.S.S. diffuse emission near Sgr A$64\!-\!113$3, the 95% confidence lower limit on the parent proton cutoff rises from $64\!-\!113$4 TeV without ISRF correction to $64\!-\!113$5 TeV with the F98 model, making the emission consistent with a CR accelerator with a spectral cutoff of at least 1 PeV (Porter et al., 2018).

These interpretive issues motivate formal ranking metrics. CTA simulation work introduces a PeVatron figure of merit based on the 95% C.L. lower limit on the gamma-ray cutoff energy, $64\!-\!113$6, when a cutoff is not significantly detected; sources with $64\!-\!113$7 are flagged as robust candidates for follow-up (Angüner et al., 2019).

6. Neutrino tests and multimessenger constraints

Neutrinos remain the decisive hadronic messenger, and their absence is now a major part of the exotic-PeVatron discussion. The IceCube ICEMAN analysis combines a 12.3-year, full-sky, all-flavor dataset and uses both template-based and point-source likelihood methods to search for Galactic PeVatrons. The result is negative at the source level: no significant excess of neutrinos correlated with known Galactic PeVatron candidates or with Galactic gamma-ray sources detected above 100 TeV has been confirmed. At the same time, the analysis reports sensitivities improved by over 20% relative to previous work and an expected median local significance of $64\!-\!113$8 for diffuse Galactic plane neutrinos in the Fermi $64\!-\!113$9 model, underscoring the difference between detecting a Galactic neutrino component and resolving individual PeVatrons (Thiesmeyer et al., 11 Jul 2025).

A targeted IceCube search for neutrinos from the 12 LHAASO ultra-high-energy gamma-ray sources likewise found no significant emissions. The abstract frames the implication sharply: evidence for neutrino emission would unequivocally confirm hadronic acceleration, and in its absence the analysis places constraints on the fraction of gamma-ray flux originating from hadronic processes in the Crab Nebula and LHAASOJ2226+6057 (Abbasi et al., 2022).

Joint likelihood analyses tighten the argument further. The HAWC–IceCube multimessenger study of 22 sources from the 3HWC catalog found no significant neutrino emission from the directions of the HAWC sources. For five sources, the 90% CL neutrino limits imply that the gamma-ray emission observed by HAWC cannot be produced purely from hadronic interactions, and the analysis reports limits on the hadronic fraction of the gamma-ray flux (Alfaro et al., 2024).

These null results do not eliminate exotic PeVatrons as a class. They do, however, constrain the simplest one-zone hadronic interpretations and shift attention toward mixed leptonic-hadronic models, extended-source effects, hidden accelerators, or source classes whose neutrino output remains below current sensitivity (Thiesmeyer et al., 11 Jul 2025, Alfaro et al., 2024).

7. Observational frontier and speculative extensions

The near-term observational frontier is set by three linked goals: extending gamma-ray measurements beyond the PeV range, resolving the X-ray synchrotron counterparts of TeV–PeV electrons, and increasing neutrino sensitivity. Extrapolation studies based on LHAASO and HAWC spectra conclude that searches for photons in the $4\sigma$0 PeV regime are presently challenging because the energy threshold is too high or the detection area too small for existing giant air-shower observatories; dedicated detector concepts are needed to explore the UHE frontier (Niechciol et al., 2024). A related line of work notes that PeV gamma-ray sources provide, in effect, a beam of pure high-energy primary photons, opening the possibility of measuring the photo-production cross section at energies beyond collider reach; future southern arrays such as SWGO are proposed as a route to extend this study to “Super-PeVatrons” (Sciascio, 2024).

At lower photon energies but with higher spatial precision, the HEX-P mission concept is explicitly designed for this problem. Its stated capabilities are $4\sigma$1 FWHM imaging, $4\sigma$2 keV spectral coverage, and effective area far superior to current facilities including XMM-Newton and NuSTAR. The science case emphasizes Galactic PeVatrons, star clusters, superbubbles, microquasar jets, gamma-ray binaries, and the use of hard X-ray synchrotron emission to distinguish leptonic from hadronic scenarios through spatially resolved non-thermal diagnostics (Mori et al., 2023).

A more speculative theoretical literature extends the phrase “exotic PeVatron” to nonstandard compact objects. One proposal considers ultra-spinning black hole vortex-string systems and exotic compact objects such as boson stars, axion stars, and Q-balls. In that framework, millicharged dark matter generates quantized magnetic flux,

$4\sigma$3

and the spin-down power scales as

$4\sigma$4

The paper argues that such objects could exceed the energy reach of conventional sources such as PWNe and SNRs and might be detectable by LHAASO, HAWC, and CTA (Addazi et al., 30 Sep 2025). This is not an observationally established source class; it is a theoretical extension of the PeVatron concept motivated by quantum gravity scenarios and dark matter phenomenology.

Across these observational and theoretical developments, the term “exotic PeVatron” now marks a broad research program rather than a single model. Its empirical core is the discovery of Galactic sources with hard gamma-ray emission above 100 TeV outside the clean SNR-shell paradigm; its unresolved question is whether those sources trace the hadronic accelerators that produce the Galactic knee, a heterogeneous population of leptonic and hadronic engines, or a still wider zoo of hidden and genuinely novel accelerators.