The LHAASO PeVatron Bright Sky: What We Learned

Abstract

The recent detection of 12 $\gamma$-ray galactic sources well above $E>100$ TeV by the LHAASO observatory has been a breakthrough in the context of the search for the origin of cosmic rays (CR). Although most of these sources remain unidentified, they are often spatially correlated with leptonic accelerators, such as pulsar and pulsar wind nebulae (PWNe). This dramatically affects the paradigm for which a $\gamma$-ray detection at $E>100$ TeV implies the presence of a hadronic accelerator of PeV particles (PeVatron). Moreover, the LHAASO results support the idea that sources other than the standard candidates, supernova remnants, can accelerate galactic CRs. In this context, the good angular resolution of future Cherenkov telescopes, such as the ASTRI Mini-Array and CTA, and the higher sensitivity of future neutrino detectors, such as KM3NeT and IceCube-Gen2, will be of crucial importance. In this brief review, we want to summarize the efforts made up to now, from both theoretical and experimental points of view, to fully understand the LHAASO results in the context of the CR acceleration issue.

1. Introduction

The origin of cosmic rays (CRs) is among the most studied topics in high-energy (HE) astrophysics. CRs are relativistic particles, mainly protons ($92\%$) and ions, that fill our Galaxy. Their very extended spectrum spans from a few MeV to beyond ${10}^{20}$ eV (see Figure 1, left). Historically, this was described by a single power law with an index $\alpha =2.7$ up to the so called “knee” (∼3 × ${10}^{15}$ eV), due to the galactic CR contribution. However, the recent results of CALET/DAMPE show that the region below 1 PeV may not be as featureless as we thought, showing hardening and steepening between 100 GeV and 100 TeV [1,2,3,4,5] (see Figure 1, right). After the “knee”, there is a spectral steepening ($\alpha =3.1$) up to the so-called “ankle” (∼3 × ${10}^{18}$ eV), where the spectrum hardens again, likely due to the emerging contribution of the extra-galactic component. A “second knee” at ∼10${}^{17}$ eV is probably due to heavy nuclei and, according to the latest experimental results, the transition from galactic to the extra-galactic component is likely between the “second knee” and the “ankle” energies [6,7,8].

The complexity of the CR spectrum perfectly fits into the HE astrophysics context of recent years, especially after the latest results from the LHAASO Observatory.

1.1. Status of the Field before the LHAASO Results

One of the main channels for investigating the nature of galactic CRs and their features is the nonthermal HE ($E>100$ MeV) $\gamma$-ray emissions, produced either by electrons, mainly through Bremsstrahlung and inverse Compton (IC) processes, and by protons via pion decay from p-p and p-$\gamma$ interactions. The difference between the two main types of processes, leptonic or hadronic, forms the base of CR acceleration phenomena. The $\gamma$-raydetection of the so-called “pion bump” at about 100 MeV (due to ${\pi }^0$ rest mass) is one of the direct proofs that confirm the presence of energized CRs in a source, and until recently, a detection of ultra-high energy (UHE, $E>100$ TeV, this energy threshold for the UHE definition is valid in the $\gamma$-ray context. If we consider the particle context, the UHE range begins above E $>{10}^{17-18}$ eV. $\gamma$-ray photons was considered unquestionable evidence of freshly accelerated CRs. Indeed, at these energies, electron IC should be limited by the Klein–Nishina regime (In the relativistic regime, $h{\nu }_i>m_e{c}^2$, the Compton scattering is inelastic and characterized by ${\sigma }_{KN}$, derived by the quantum electrodynamics, and not by ${\sigma }_T$) and UHE $\gamma$-ray should only be explained through the decay of neutral pions produced by collisions of PeV protons (CRs) with target protons (for recent reviews see [10,11]). According to this view, the term “PeVatrons” has been coined to indicate galactic CR accelerators (sources emitting $\gamma$-ray at E $>100$ TeV).

In the standard paradigm, supernova remnants (SNRs) are the main contributors to galactic CRs [8]. We know that they energize CRs at the low $\gamma$-ray energies (MeV-GeV) thanks to their detection around the “pion-bump” energies by AGILE and Fermi-LAT [12,13], likely even through re-acceleration [14], but in no SNRs were detected $\gamma$-ray photons with E $>100$ TeV. According to theoretical models, the reason for this is that SNRs can accelerate CRs at these energies only in the first 100 years of their life, and all the known SNRs are older [15,16]. The only way we can detect their UHE emission is in the case CRs are trapped in near giant molecular clouds (GMCs), emitting UHE photons for $t>100$ yrs [11,17,18].

The lack of SNRs at UHE supports the possibility that other sources can accelerate CRs, and this possibility has been widely investigated. In 2011, the Fermi-LAT satellite observed extended $\gamma$-ray emissions associated with the superbubble surrounding the Cygnus OB2 region, then called “Cygnus Cocoon”, for the first time confirming the hypothesis proposed by [19], that star forming regions are able to accelerate particles. Deeper analysis of this region, and further very-high energy (VHE, 100 GeV $<E<$ 100 TeV) detections from other massive star clusters (MSCs), supported this possibility (see Section 2.2 and Section 3.10 for more details).

Later, the detection of several pulsar wind nebulae (PWNe) at $E>56$ TeV with different instruments [20,21,22,23], and in particular of the Crab Nebula at $E>150$ TeV [20,23], opened the possibility that the PWNe, leptonic accelerators “by definition”, could also accelerate hadrons and, consequently, CRs. Moreover, UHE $\gamma$-ray detection from the Crab Nebula made it evident that leptons are able to emit beyond the Klein–Nishina limit. Several works have shown that leptonic emission through IC can predominate at these energies in radiation-dominated environments [24] or in the surroundings of high spin-down powers pulsars [25,26].

In this new context, another fundamental channel comes into play: neutrino detection. This is an unquestionable sign of CR acceleration, since neutrinos can be produced only by the decay of charged pions produced by p-p and p-$\gamma$ interactions [27]. Neutrino detection in correspondence with PeVatron candidates is the main instrument for confirming the nature of $\gamma$-ray emitting sources [28]. These neutrinos can be detected by current (IceCube [29] and Baikal-GVD [30]) and the future (P-ONE [31], IceCube-Gen2 [32], KM3NeT-ARCA [33]) VHE neutrino detectors [34,35,36]. However, up to now, no neutrinos from galactic sources have been detected with high significance [37]. We have tentative evidence of neutrino emissions from only two extra-galactic sources: TXS 0506+056 (a blazar) and NGC 1068 (a Seyfert/starburst), associated with IceCube HE neutrinos at a significance around 3$\sigma$ [37,38]. In addition, most of the recent diffuse neutrinos detected over a period of 7.5 years [39] were found to be of extra-galactic origin [40,41,42].

1.2. Status of the Field after the LHAASO Results

The recent LHAASO discovery of several galactic sources emitting UHE $\gamma$-rays [43,44,45] represents a strong breakthrough in the context of PeVatron research. Most of the sources seem to correlate to pulsars, and/or their nebulae, PWNe (e.g., the Crab Nebula for all), a class of well known leptonic accelerators. These observations indicate the presence at these sites of ultra-relativistic electrons with energies reaching at least a few ${10}^{15}$ eV. As a consequence, the $\gamma$-ray detection at UHEs cannot itself be considered proof of hadronic acceleration. This challenges the definition of “PeVatron” itself. In this work, we will generically call “PeVatron” an object capable of accelerating particles (hadrons or leptons) up to the PeV (=10${}^{15}$ eV) range (and emitting $\gamma$-ray above 100 TeV). Consequently, to indicate a source that specifically accelerates hadrons (CRs), we need to use the term “hadronic PeVatron”.

The LHAASO results also suggest that many classes of sources can accelerate CRs up to PeV energies: SNRs, MSCs, and maybe also the PWNe (see [46], and references therein). In the case of PWNe, there is also another complication: the standard first-order Fermi acceleration mechanism, underpinning diffusive shock acceleration (DSA) theory, may be not the main one (see [47,48], for a review) and magnetic reconnection (MR) may be a valid alternative, even if simulations suggest that it may be a process producing impulsive events (flares) but not a stable VHE emission [49,50,51]. This discussion, however, is beyond the aims of this review.

Actually, some of the LHAASO PeV detection regions could be correlated with SNRs, keeping open the possibility of having PeV emissions from this kind of source, even after their 100 yrs [52,53,54,55]. Several studies are in progress to establish a “look up table” of parameters useful for future instruments to distinguish which SNRs are hadronic [56,57].

The association of VHE/UHE LHAASO emissions with a determined kind of source in the observed region is challenging, due to the very low angular resolution of LHAASO and similar extended air shower (EAS) arrays. For this reason, future imaging atmospheric Cherenkov telescopes (IACTs) such as ASTRI Mini-Array [58,59,60,61,62], and CTA [63] will make a difference (see Section 4 and Figure 17, right).

The LHAASO results triggered several studies on the constraints entailed by potential neutrino emissions on the $\gamma$-ray fluxes from the candidate PeVatron sources. In [64], the authors exploited the IceCube measurements in the surrounding of the LHAASO sources, in order to constrain their possible hadronic contribution; and in their subsequent paper [35], the same authors carried out a Bayesian analysis on the 10-year IceCube data (combined with the ANTARES [65] ones) with the same aim. The results obtained with the two approaches were very similar and can be summarized as showing the impossibility of strong constraints on the LHAASO sources (apart from the Crab) with the current instruments.

These results were also confirmed by the search for neutrinos from the 12 LHAASO sources with 11 yrs of track-like events (neutrino colliding with matter in or near the detector, resulting in a HE muon that traveled a long distance) from the IceCube Collaboration [66]. In this latter work, the authors also performed a stacking analysis combining (1) all LHAASO sources, (2) LHAASO sources with SNRs as potential TeV counterparts, (3) LHAASO sources with PWNe as potential TeV counterparts. Even in this case, with the strong assumption that all of the flux measured by LHAASO is hadronic, there were no significant detection, but a $90\%$ UL on the predicted hadronic flux. This is an indication that there is a low chance that all six sources with a SNR as candidate actually are SNRs and the same is valid for the nine possible PWNes (see Figure 2).

In all the works, however, the great potential represented by future instruments is stressed, such as IceCube-Gen2, to detect neutrino emissions by LHAASO PeVatrons, in the cases where there are hadronic PeVatrons (see Section 4).

Several theoretical works have been developed after the LHAASO results, analyzing which kinds of source are able to satisfy the requirements needed to explain the CR density in our galaxies, in spite of the acceleration mechanism at the origin (DSA is the main source, but there is the magnetic reconnection that can contribute in microquasar jets or to the wind of the PWNe) [5,67,68,69,70]. One of the main hypotheses so far is that, since the CR spectrum in the energy range around the “knee” shows various features, maybe we can have different contributions from different kinds of sources at different energies ([71] and Figure 3). A parallel analysis of $\gamma$-ray and neutrinos from different source types can give us the key to interpreting the recent results from LHAASO and from the other VHE-UHE $\gamma$-ray instruments [42,64,66,72], in order to discern hadronic from leptonic accelerators. This kind of analysis is also important for comparing known and well-studied sources with the same types of candidate sources of PeVatrons, to understand which physical conditions are behind the differences.

We can summarize the century of studies into the CR context in the following way:

• CRs were discovered in 1912 but, despite the enormous efforts made in recent years, the primary three questions about their origin still need to be clarified: what are their sources, how are they accelerated, and how do they propagate?
• Two direct pieces of evidence were identified in $\gamma$-rays as a signature of sources of galactic CRs: the “pion bump” below 100 MeV, due to the ${\pi }^0$ decay, and $\gamma$-ray emissions at E $>100$ TeV.
• In the standard paradigm, the main candidates for galactic CR acceleration are SNRs, mainly for energetic reasons. In spite of the evidence of the “pion bump” in some of them, no SNRs have been detected at E $>100$ TeV, and the hypothesis that other kinds of source could contribute to CR acceleration emerged.
• The first observational evidence of VHE emissions without the sign of a cutoff from MSCs, microquasars, and PWNe confirmed this latter hypothesis and questioned the emissions at E $>100$ TeV as the “smoking gun” of hadronic acceleration.
• The detection of 12 Galactic PeVatrons by LHAASO confirmed that sources other from SNRs can accelerate CRs and that electrons can emit UHE $\gamma$-ray. Consequently, we need further help to understand which sources are really CR accelerators.
• Neutrino detection and better morphological reconstruction are fundamental tools to achieve this aim, and future instruments can be of great help.

Starting from this context, we think that a summary of the work performed on the LHAASO sources after the publication of the 12 PeVatron paper [43] could be very useful to the scientific community, and also for future instruments, to determine the right action plan. We are aware that other instruments before LHAASO, e.g., HAWC and MAGIC, obtained results that hinted at what LHAASO revealed subsequently. However, we think the results from LHAASO were a breakthrough for two main reasons: the extension by an order of magnitude of the energy range, and the relatively large number of sources. Despite the fact that the portion of the galaxy observable from the LHAASO site is smaller than that from the HAWC site (and it does not include the galactic center), LHAASO tripled the number of known sources emitting above 100 TeV, demonstrating that this feature is quite common for galactic sources. The number of sources for which a hadronic scenario is largely excluded similarly increased. Moreover, the detection of photons near 1 PeV challenges the astrophysical explanation of what we are observing (assuming both hadronic or leptonic origins). Let us also note that LHAASO obtained these results with only 10 months of observations, suggesting these 12 sources are only the tip of the iceberg of what will be detected in future observations by both LHAASO and other instruments based on the same technology (i.e., SWGO [73] at the South and TAIGA [74]).

Figure 3. Figure from [75]. Here, the all-particle CR spectrum is explained with the contributions from different kinds of galactic source. In particular, the dominant contribution from the SNRs in MSCs is evident.

Figure 3. Figure from [75]. Here, the all-particle CR spectrum is explained with the contributions from different kinds of galactic source. In particular, the dominant contribution from the SNRs in MSCs is evident.

In Section 2, we briefly summarize what we know about the source types that may contribute to galactic CR acceleration: SNRs (Section 2.1), MSCs (Section 2.2), and PWNe (Section 5). We also consider TeV halos (Section 2.4) because, even if they are expressions of already accelerated CR diffusion, more than one LHAASO source could be associated with a halo. For completeness, we is evident include microquasars in this section (Section 2.5) because, although these kinds of source are not associated with any LHAASO PaVatron, they are considered to be among the possible hadronic PeVatrons. In Section 3, we focus on each of the 12 LHAASO UHE detections and on the information collected about them up to now. Finally, we conclude with a look at the expectations regarding future instruments (Section 4) and with final remarks (Section 5).

2. PeVatron Candidate Source Typologies

Despite some features revealed in the SNRs confirming that these sources accelerate the low energy component of galactic CRs, we do not have proof of their emission of UHE $\gamma$-ray. Even the youngest SNRs, with a high shock velocity, amplified magnetic field, and dense environment, show spectra with a cutoff well below $E=100$ TeV (see Figure 4). Moreover, there are other challenges related to the role of SNRs in CR acceleration; their observed $\gamma$-ray spectra, which are different from the theoretical ones, and some problematic issues correlated with CR abundances (see [8], and reference therein), which motivated the search for alternative sources.

Furthermore, in recent years, the assumption that SNRs may not be the only source of CRs had already been supported by several detections of VHE/UHE $\gamma$-ray emissions from different source types before the LHAASO results. This possibility is also supported by the higher complexity of the CR spectrum shown by CALET (see Figure 1, right) which seems to be an indication that we need several kinds of source to explain all of the spectral features of CRs. Indeed, different sources can contribute to the CR spectrum at different energies [71,75]: in this hypothesis, for example, the CR “knee” could be the transition region between the contribution of standard isolated SNRs and that of the SNRs in MSCs (see Figure 3).

With the large amount of experimental data collected so far and with theoretical models developed in the context of CR acceleration, we have important constraints with which a certain kind of source must comply to be considered a hadronic PeVatron.

From a theoretical point of view, the hadronic PeVatron requirements can be summarized in the following way [67]:

• Energetics—Hadronic PeVatrons must explain the CR power of the Galaxy, estimated to be about $P_{CR}\sim$10${}^{40}$–10${}^{41}$ erg/s [11];
• Power-law injected spectrum—According to DSA, the accelerated spectrum at the source must follow a power-law that, after correction for transport effects and CR anisotropy, is $\propto E^{-2.1÷2.3}$ [10,11];
• Maximum energy—The acceleration efficiency at the source must be able to explain the PeV particle energies;
• Anisotropy—The source distribution must be correlated with the CR anisotropy in the arrival direction predicted by a diffusive model for CRs [11,76];
• Composition—Galactic CR sources must explain why their composition that is slightly different with respect to solar CRs (e.g., large overabundance of light elements, overabundance of metals with respect to H and He) [10,11]. Moreover, the composition varies with energy (e.g., different elements have different spectral indices), and this energy dependence also needs to be reproduced.

From an observational point of view, a candidate hadronic PeVatron must have certain precise features [77]:

• Evidence of UHE $\gamma$-ray emission ($E>100$ TeV)—If we have CR acceleration, we expect protons that are accelerated at least up to 1 PeV and, consequently, $\gamma$-ray emission well above 100 TeV without a cutoff in the spectrum;
• Spectral curvature—The parent particle power-law is modified by the existence of a maximum energy and/or by some breaks due to losses;
• Correlation with target material—The p-p interaction is most favored in the presence of a dense target in the source surroundings [76,78]
• Extended emissions with a likely energy-dependent morphology—CRs diffuse through the media surrounding the source and diffusion (as in cooling) is energy dependent;
• Multi-wavelength counterpart—Secondary non-thermal electrons produced in the p-p interaction are emitted in radio/X-ray bands.

All these observational requirements can also be found for leptonic PeVatrons. Consequently, a cross-check between experimental data and theoretical models is fundamental, in order to disentangle PeVatron and CR sources (hadronic PeVatrons).

In the following, we summarize the classes of source that could be associated with LHAASO candidate PeVatrons and, consequently, that are also candidate CR accelerators (hadronic PeVatrons): Supernova remnants, massive star clusters and pulsar wind nebulae (see [46], for a brief and good review). For completeness, we also dedicate a paragraph to TeV halos (the new class of VHE objects that are due to leptonic emissions but that could be associated with LHAASO UHE emissions) and to microquasars (candidate hadronic PeVatrons but without any associated LHAASO UHE emissions).

2.1. The Standard Candidates: Supernova Remnants

Since the beginning of CR study, SNRs have been considered “the source” of galactic CRs, because they respect most of the theoretical and observational requirements of hadronic PeVatrons (see [8], for a recent review). First of all, their power is able to sustain the CR flux measured in our galaxy: ${\dot{E}}_{SN}\approx {10}^{42}$ erg s${}^{-1}$, consequently, $E_{CR}\approx 10\%{\dot{E}}_{SN}$ [79]. Thus, they are ideal laboratories for particle acceleration using the DSA mechanism [80,81,82]; the first-order Fermi acceleration mechanism works very well in their environment, producing the required pure power-law spectrum. Various experimental results have supported this hypothesis with indirect proofs. First of all, there have been several detections of their extended emission at HE and VHE $\gamma$-ray, as well as their very frequent association with high-density MCs. Then, the presence of non-thermal and fast variable X-ray emissions coincident with TeV $\gamma$-rays indicates the presence of magnetic field amplification, a fundamental condition of reaching VHE energies [83,84,85,86]. The modified Balmer line detection is evidence of the CR back-reaction to the DSA [87], and finally their spatial distribution is compatible with the CR distribution and enough SNRs exist to explain the anisotropy [11]. Furthermore, a first real breakthrough came in 2011, when the AGILE satellite showed, for the first time, direct proof of the presence of CRs in a shock of a SNR, the SNR W44 [12,88], which was then confirmed using Fermi-LAT in the same source W44 and in other similar middle-aged SNRs, such as IC443, W51c, and W49b [13,89,90]. All these sources were detected at E < 100 MeV with a spectrum clearly correlated to the presence of a “pion bump”.

However, these detections confirmed that SNRs can accelerate CRs with low energies ($E<100$ GeV) but a $\gamma$-ray clear detection at UHE is still missing. Theoretical works have shown that this could be due to the intrinsic properties of SNRs, which accelerate PeV particles only in the first 100 years of their evolution [15,16,91] and only very rarely (1 every ${10}^4$ yrs) can high-power ($E_{SN}\sim$ 5–10 × ${10}^{51}$ erg) explosions accelerate CRs to PeV energies at the beginning of their Sedov phase [92]. The only way to detect UHE $\gamma$-rays from SNRs is if the accelerated CRs, after their escape from the sources, are trapped in near GMCs thanks to suppression of the diffusion [11,17,18,93]. The LHAASO results show that about 6 of the its 12 PeVatrons could be associated with (middle-aged) SNRs. Consequently, if a deeper study of these UHE emission regions could confirm their SNR nature, we would know if the emission is actually from near GMCs or, in case the UHE emission comes from the SNR itself, if we need to improve our knowledge of SNR microphysics [5,92].

Anyway, there are also other difficulties related to the link between SNRs and CR acceleration. The predicted spectrum does not match the observed ones, requiring modifications to the original DSA theory (see [8,10], for a recent review). Moreover, if SNRs are the only CR sources, there are several composition issues that are not easy to explain: the most popular example is the ${}^{22}Ne{/}^{20}Ne$ abundance, which in Galactic CRs is about five-times larger than the one in Solar CRs and which cannot be explained in a SNR context. In the face of these many challenges, the search for other galactic CR contributors is well motivated.

Figure 4. Differential luminosity curves for some of the most commonly studied SNRs at TeV energies. It is evident how even the youngest SNRs cannot reach $E>50$ TeV (Figure from [94]).

Figure 4. Differential luminosity curves for some of the most commonly studied SNRs at TeV energies. It is evident how even the youngest SNRs cannot reach $E>50$ TeV (Figure from [94]).

2.2. Superbubbles and Young Massive Clusters

The first interest in the role of superbubbles (SBs, the shell structure originating from multiple stars in OB associations) around Young MSCs (YMSCs, $M\gtrsim {10}^4{M}_{\odot }$) in the CR context was in the early 1980s [95]. Indeed, they have several possible acceleration sites, since they host hundreds of massive stars with their fast winds [96], have a turbulent interior [97], and have a lot of potential SNRs expanding inside them [98]. The entire bubble can reach dimensions greater than 100 pc, creating expectations of extended sources, which was confirmed by several VHE detections of regions surrounding young MSCs [94,99], e.g., Cygnus Cocoon [100,101,102,103,104], Westerlund 2 [105], 30 Dor in the Large Magellanic Cloud [106], the Central Molecular Zone [107,108], and Westerlund 1 [109] ( complete list in [67]). Moreover, the collective effects of shocks from the SN explosions and the wind termination shocks in PWNe can explain the mismatch of the models with the observed ${}^{22}Ne{/}^{20}Ne$ ratio [68,110], solving one of the main issues linked to the role of SNRs in CR acceleration (see [111], for a review).

The power injected by a single stellar wind is between 10–50% of the power generated by a SNR shock [67], with an average energy per massive star of about ${10}^{48}$ erg [4,46]. Consequently, isolated massive stars fail to supply the required CR energy budget in our galaxy. However, in the most compact clusters (YMSCs with a size of a few pc, also called wind-blowing clusters by [75]), powerful winds from the youngest stars and Wolf Rayet stars can merge into a collective outflow that accelerates CRs in the entire region [46,75]. The active mechanism is probably the less efficient second-order Fermi acceleration [112] but this lasts up to a million years, a much longer time with respect to the efficient phase of SNRs.

The long duration of CR injection was also confirmed by the spatial $\gamma$-ray distribution detected in Cygnus OB2, Westerlund 1, and Westerlund 2, which implies a CR distribution $\propto 1/r$ [94], indication of a continuous injection. Moreover, magnetic turbulence with $B\sim$100 $\mathsf{\mu }$G could be reached inside the core [113]. This allows YMSCs to reach E > PeV with a spectrum in agreement with the observations [114] and to explain the CR spectrum at knee energies and above; even if with an expected and strong dependence on the diffusion coefficient [67,68,75,114]. According to these models, the “knee” of the CR particle spectrum can be explained by the transition between the isolated standard SNRs and the SNRs in compact YMSCs (see Figure 3).

In summary, according to observational and theoretical criteria, YMSCs and Superbubbles may be one of the main contributors to galactic CRs. However, the theoretical models are still incomplete and the higher angular resolution of future IACTs, together with multi-wavelength and multi-messengers observations, will be fundamental for resolving the morphology of these regions.

2.3. Pulsar Wind Nebulae: The Crab and the Others

PWNe should be the most numerous of the different kinds of galactic source [48] and their $\gamma$-ray emission is estimated to last a very long time (∼100 kyrs) [70]. This is likely one of the reasons why more than 30% of the candidate PeVatrons detected by LHAASO seem to be associated with PWNe [26]. They are the most efficient accelerators in the galaxy (up to $30\%$ acceleration efficiency), with no signs of thermal particles, and are the only sources in which we have direct proof of PeV electron acceleration, thanks to their synchrotron spectrum. The location of particle acceleration is likely the termination shock of the young PWN, generated by the friction between the pulsar wind and the parent SNR.

With their UHE detection by LHAASO and other EAS experiments, PWNe respect all the observational requirements to be PeVatrons (and possibly hadronic PeVatrons): extended HE $\gamma$-ray emission; multiwavelength counterparts, from radio to X-rays; and a nonthermal spectrum, with spectral curvature [26,70]. The main mechanism producing this UHE emission is the IC scattering of UHE accelerated electrons with cosmic microwave background radiation (CMBR). From a theoretical point of view, however, the context is more complex. The first issue is the energetics. At the LHAASO detected energies, the Klein–Nishina regime is the dominant one and, in order to explain leptonic UHE $\gamma$-ray, a very high acceleration rate is required [69]. Indeed, in spite of the acceleration mechanism of PWNe still being unknown, we know that there are two constraints on the maximum reachable particle energy. The first strong limit is due to the polar cap potential drop: $E_{max}=q{(\dot{E}/c)}^{1/2}$, where q is the particle charge (see [26,115], for a quick review). In order to make a direct link between theory and observation [26], we can write this relation as: $E_{\gamma ,max}\approx 0.9{\eta }_e^{1/3}{\eta }_B^{0.65}{\dot{E}}_{36}^{0.65}$ PeV, where ${\eta }_B$ is the fraction of the PWN energy flux transformed in magnetic energy density and ${\eta }_e$ is the ratio between the electric and magnetic fields. This expression tells us that only a very energetic pulsars ($\dot{E}\gtrsim {10}^{36}$ erg/s) can reach PeV energies. The second constraint is due to the condition that the acceleration rate must be higher than the electron radiative losses: $E_{\gamma ,max}\approx 2.7{\eta }_e^{0.65}{\eta }_B^{-0.33}{R}_{0.1}^{0.65}{\dot{E}}_{36}^{-0.33}$ PeV, where $R_{0.1}$ is the termination shock radius in units of 0.1 pc [26]. This condition is more stringent for young energetic pulsars ($\dot{E}\gtrsim {10}^{37}$ erg/s), such as the Crab Nebula.

In this scenario, the presence of accelerated hadrons cannot be excluded, but some basic calculations show that PWN cannot be the main CR contributors. In order to produce an amount of energy comparable with the energy expected from SNRs, they would have to have too small an initial spin period with respect to that provided by theoretical models [46]. This implies that, even if PWN contribute to galactic CRs, they have no chance of explaining CR composition or anisotropy.

A very good summary of LHAASO PWNe candidate PeVatrons and their characteristics can be found in [26], where all the LHAASO sources with a PWN as potential TeV counterpart are analyzed in the PWN parameter space based on energetic constraints. According to this work, all of these LHAASO $\gamma$-ray emissions near a PSR could have been produced by a PWN, except LHAASO J2032+4127 (see Figure 5 and the paragraph dedicated to this in Section 3.10).

2.4. TeV Halos

The discovery of TeV halos surrounding pulsars can be attributed to the extended TeV emission detected around the Geminga and Monogem (B0654+14) pulsars by MILAGRO [116] and HAWC [117]. The extension of these $\gamma$-ray sources is about $2^{\circ }$, indicating that CR propagation near these pulsars is more constrained with respect to that in the ISM [118]. TeV halos can represent a local environment for studying CR propagation with small-scale diffusion models [119]. Their emission is due to a slow diffusion of the electrons in the ISM around the pulsars (several hundred times smaller than the galactic diffusion coefficient [45]). Although we do not know the nature of the environment that implies a slowing down of the diffusion, we know that the $\gamma$-ray emission is due to the IC scattering of electrons accelerated by the pulsar in the galactic and cosmological radiation fields [119]. The discovery of this new class of objects is also important because they could allow the discovery of a population of “invisible” pulsars that have no beamed radio or HE beamed emissions. Moreover, the spectral study of TeV halos is an alternative method for constraining the injection of PWNe electrons because their spectrum only depends on the electron spectral cutoff.

The large extension (∼1 degree) of many of the LHAASO sources and their correlations with pulsars increased the interest around this new class of sources (see Figure 19). The only firm TeV halo detected by LHAASO is LHAASO J0621+3755 [45], which is not treated in this review because of its low UHE $\gamma$-ray significance (<5 $\sigma$ at 100 TeV). It is one of the currently known TeV halos, together with the Geminga halo, Monogem halo, and HESS J1831-098 [119]. The main problem related to TeV halo detection is the possibility of confusing them with extended SNRs [118]. In order to verify the TeV halo origin of a source, beyond the co-location of the bulk of its emission with a pulsar, this pulsar needs to have a spin-down luminosity sufficient to generate a halo and needs to be enough old to allow electrons to diffuse in the ISM (bow-shock nebula [70]). Furthermore, the extension of the halo has to be larger than the possible PWN X-ray emission, otherwise the emission could be a $\gamma$-ray PWN, and its morphology must be explained by slow-diffusion, otherwise this could indicate a hadronic accelerator [119].

The low angular resolution of the EAS experiments does not allow resolving the TeV halo morphology and, unfortunately, the size of these objects is challenging for current IACTs because of their small field of view. Indeed, Geminga and Monogem are not detected by these instruments, despite their high TeV flux. The future ASTRI Mini-Array, with a FoV of ${10}^{\circ }$ at E $>10$ TeV, and CTA SSTs in the southern hemisphere, with a FoV $>8^{\circ }$, could finally resolve the morphology of some of these objects and detect a high-resolution spectrum. In the ASTRI Mini-Array core science paper [62], a first simulation of the Geminga halo was developed, building a radial profile and showing a high-significance spectrum up to 50 TeV with a single observation. Multiple observations will allow the ASTRI Mini-Array to better track the source morphology.

2.5. Gamma-Ray Binaries and Microquasars

The first interest in $\gamma$-ray binary systems (i.e., microquasars) such as galactic CR accelerators was born from their similarities to AGN, which are likely the main component responsible for extra-galactic CRs [46]. We know that microquasars accelerate electrons to VHE [120,121]: their environment is radiation-dominated, making the Klein–Nishina losses ineffective in preventing acceleration, and $\gamma -\gamma$ absorption is transparent for 100 TeV $\gamma$-rays. However, they are much less abundant with respect to the other kinds of candidate PeVatrons, such as SNRs or PWNe. Consequently, even if they could accelerate protons, they cannot explain either the CR energy density or the anisotropy and composition in our galaxy.

This does not mean that some microquasars cannot be hadronic accelerators. So far, several $\gamma$-ray binaries system have been detected at VHE with no hint of a cutoff above 10 TeV: PSR B1259-63/LS2883 [122], LS 5039 [106,123], V4641 Sagittarii [124] and SS 433 [125,126,127,128]. In particular, SS443 was detected at E$>10$ TeV with a jet that exceeded 10-times the required injection power, $L_{kin}={10}^{39}$ erg/s [5,121]. The microquasar V4641, a black hole binary with ultrarelativistic jets, introduced a new class of $\gamma$-ray sources of great interest for future instruments. In general, the detected emission seems to be dependent on the orbital phase, confirming that it was likely produced at binary system scale and that there is some active acceleration mechanism. One of the open issues concerns if this is the same mechanism seen in the Crab Nebula or if we need to add a special mechanism of action for binary systems, as was suggested by a recent X-ray analysis of the SS 433 [128].

No microquasars are found among the 12 PeVatron candidates of LHAASO, but this class of sources represents an important target for LHAASO and the future EAS and IACTs instruments.

3. Overview of the 12 LHAASO Sources

The Large High altitude Air Shower Observatory (LHAASO), located in Sichuan (province of China) at 4410 m altitude, is a complex of EAS detectors, with the aim of studying both CRs and $\gamma$-rays in the sub-PeV to 1000 PeV energy range [129]. It is composed of three arrays: a kilometer square array (KM2A, 10 TeV–100 PeV, UHE $\gamma$-ray detector with an energy resolution $\le 20\%$, angular resolution of $0.{25}^{\circ }$, and sensitivity (at 50 TeV) of ${10}^{-14}$ erg cm${}^{-2}$ s${}^{-1}$), a water Cherenkov detector array (WCDA, down to 0.1 TeV, angular resolution of $0.2^{\circ }$, and sensitivity ($E>2$ TeV) of ${10}^{-12}$ erg cm${}^{-2}$ s${}^{-1}$), and a wide field-of-view Cherenkov telescope array (WFCTA, 0.1–1000 PeV) [130].

The 12 Galactic PeVatrons were detected by the half-completed LHAASO-KM2A, with an angular resolution of $0.4^{\circ }$ and an energy resolution of $28\%$ at 30 TeV [43] (see Figure 6). The indicated $E_{max}$ in the table represents the maximum energy detected for each source (for the Crab Nebula, there is an updated value, see the dedicated Section 3.1). Among them, the UHE emissions in the regions of LHAASO J1825-1326, LHAASO J1908+0621, and LHAASO J2018+3651 had already been detected by HAWC ([21], eHWC J1825-134, eHWC J1907+063, eHWC J2019+368), and LHAASO J0534+2202 (Crab Nebula) was detected at $E>150$ TeV by HAWC and Tibet AS$\gamma$ [20,23]. LHAASO J0534+2202 is the only candidate PeVatron detected by LHAASO with a firm association. The low angular resolution of the LHAASO instruments implies that for every candidate there are at least two candidate TeV counterparts.

In addition to the 12 well-known PeVatrons published in [43], the LHAASO collaboration published two papers on two other sources detected at VHE: LHAASO J0341+5258 [44] and LHAASO J0621+3755 [45]. LHAASO J0341+5258 is a unidentified source associated with GMCs [131] but without any detected PWN or SNR in the surroundings. LHAASO J0621+3755 is one of the four TeV halos of our galaxy (see Section 2.4) detected and identified to date [119,132]. It is a new source in the TeV domain [131] and is spatially coincident with the Fermi-LAT PSR J0622+3749 that has a spin-down sufficient to explain the halo luminosity. Moreover, its spectrum can be explained using super-diffusion models [133]: all the criteria to be a TeV halo are respected [119].

These two sources are not in the main paper [43] nor in the PeVatron Table in Figure 6 because they have a significance at 100 TeV $<5\sigma$, and for this reason we do not take them into account in this work, where we summarize what has been done from a theoretical and experimental point of view on every LHAASO PeVatron source in [43]. Our aim is to clarify what we know and what we have learned about every LHAASO source and their origin, both from a $\gamma$-ray and neutrino point of views. A theoretical interpretation and/or analysis of the existent models is beyond the scope of this work.

In the Figure 19, we summarize the main conclusions about the origin of $\gamma$-ray emissions from the LHAASO sources, together with their coordinates, detected maximum energies, and detection significance.

3.1. LHAASOJ0534+2202 (Crab Nebula)

LHAASO J0534+2202 is the only LHAASO candidate PeVatron with a specified counterpart, the Crab Nebula. LHAASO detected its emission with the highest significance, $\sigma =17.8$. Its maximum energy was about $E_M\approx 900$ TeV during its first detection by KM2A and WFCTA [43] but, in a second shower detected by KM2A one year later, this reached $E_M=1.12\pm 0.09$ PeV [130].

The Crab Nebula is one of the most studied sources of our galaxy in every electromagnetic band, from radio to $\gamma$-ray (see [69], for a recent and complete review). In the HE band, its flaring nature at MeV energies [134,135,136] was a breakthrough for all theoretical models but even its steady VHE $\gamma$-ray emission is very interesting. Indeed, its synchrotron emission was one of the first pieces of evidence for the presence of PeV electrons in its environment [137] and its Pevatron nature was confirmed by the HAWC [20] and Tibet AS$\gamma$ [23] results, which detected it at $E>150$ TeV; only the sensitivity of water and air shower detectors can reach these extreme energies. The very high sensitivity at UHE of the LHAASO KM2A (∼10${}^{-14}$ erg cm${}^2$s${}^{-1}$ at $E>50$ TeV in 1 yr) showed, for the first time, that the $\gamma$-ray component of its spectrum extended up to 1.1 PeV, transforming it into a super-PeVatron [130]. Following what we said in Section 2.3, the Crab Nebula is one of the youngest and most powerful PWNs and, consequently, the maximum energy of its $\gamma$-ray emission is radiation limited. This means that its detection at 1.1 PeV implies an acceleration rate ${\eta }_e\approx 0.16$, three orders of magnitude greater than that in SNRs [43,130].

After the LHAASO detection, the main issue connected to the super-PeVatron nature of the Crab Nebula is related to its acceleration mechanism but also to the possibility of a contribution from a hadronic component. Even before this discovery, the possibility of a contribution from hadrons had been investigated (see [138], and reference therein) but now, around 1 PeV, the LHAASO spectrum shows a hardening of the spectrum that is really challenging to explain with a leptonic component [130]. The LHAASO error bars are very large and, consequently, this hardening needs to be confirmed, but if it is, this will support the idea of a UHE hadronic component: a component that cannot explain the overall $\gamma$-ray luminosity but that can have a non-negligible contribution at the highest energies, with the condition of proton confinement in the nebula [130]. Recent models estimated that this hardening is compatible with a quasi-monochromatic distribution of protons around 10 PeV [69,139], to be added to the confirmed leptonic distribution. The Lorentz factor and the spin-down luminosity of protons are two of the most important parameters that characterize this component [62].

Several theoretical models for constraining this hypothetical hadronic component are being developed [115,138] (see Figure 7), and future instruments such as the ASTRI Mini-Array and CTA will be fundamental for confirming a lepto-hadronic model of the $\gamma$-ray emission from the Crab Nebula. In the ASTRI Mini-Array core science paper [62], a simulation of a Crab spectrum up to 300 TeV was produced, showing how the ASTRI Mini-Array could be decisive in constraining the likely hadronic component at the highest energies. The Crab Nebula is not the only source for which the neutrino analysis carried on with different methods [35,64,66] can put constraints confirming that the hadronic component is not the dominant one explaining its $\gamma$-ray emission.In [36], it was shown that the expected neutrino flux from the Crab is also lower than the IceCube sensitivity in the case of pure hadronic emissions.

Nevertheless, the Crab Nebula is a very unique source, both for the fact that it is the only PWN where we know that its internal synchrotron radiation is the target of IC and because, due to its young age, its energy losses are dominated by radiation reaction and not by the potential drop, as for most of the other PWNe [26,69]. This is likely the reason why flaring emission was seen only in the the Crab Nebula.

3.2. LHAASO J1825-1326

The candidate PeVatron LHAASO J1825-1326 is one of the sources (the others are the Crab, LHAASO J1908+0621 and LHAASO J2226+6057) detected with the highest significance at E $>100$ TeV ($\sigma =16.4$) and for which there is a UHE SED without any hint of a cutoff (see Figure 8, bottom panel, and paper [43]). Its maximum energy is $E_M\approx 420$ TeV and the gradual steepening at the highest energies may be due to photon–photon absorption [43]. Its potential TeV counterparts are HESS J1825-137/2HWC J1825-134 and HESS J1826-130, both PWNe in a region smaller than the LHAASO angular resolution (see Figure 8, top panel).

HESS J1825-137 is the name of the PWN associated with the PSR B1823-13, revealed by the first Galactic Plane Survey of HESS [140] and is one of the brightest and most extended sources in the galaxy at VHE, with an intrinsic diameter of about 100 pc [22,141]. This was also detected by HAWC as a unique extended region, eHWC J1825-134 [117], then resolved into three different $\gamma$-ray emitting components (HAWC J1825-138, HAWC J1826-128, and HAWC J1825-134). In particular, HAWC J1825-134 is at only 0.03${}^{\circ }$ for the LHAASO source (see Figure 8 top right) and its spectrum extends beyond 100 TeV without a cutoff, whereas HAWC J1825-138 seems to confirm the spectral behavior of HESS J1825-137 [141], with a cutoff below 100 TeV. We cannot exclude that HESS J1825-137 is actually a composition of two different TeV sources. This HESS $\gamma$-ray source shows a strong morphological dependence at GeV and TeV energies [142,143], and the asymmetry of the nebula in the X-ray band [144] could be due to the reverberation phase (interaction of the termination shock with the parent SNR reverse shock) and/or due to the presence of an MC in the surroundings [145]. The extraordinary extension of this source makes it a possible candidate TeV halo [146], despite it not being possible to explain it within the standard dynamics of the PWN [147].

Another one the three HAWC resolved sources, HAWC J1826-128, is detected up to 100 TeV and is spatially coincident with HESS J1826-130 [99,148], the other potential TeV counterpart of the LHAASO PeVatron candidate, and both sources seem to be related with the “Eel” PWN11 (PWN G18.5-0.4) [149]. HAWC J1826-128 was detected in X-ray and $\gamma$-ray but not in the radio band, a strange peculiarity compared to other similar PWNe. In the same field, there is also the very poorly known SNR G18.45-0.42 [150], which cannot be ruled out as the counterpart of HESS J1826-130. According to the developed models, the leptonic nature of this emission is favored [149].

The LHAASO angular resolution cannot resolve the region at the origin of the detected emission, consequently we cannot exclude that this PeVatron emission is due to HESS J1826-130 instead of HESS J1825-137. Actually, the spectrum of LHAASO J1825-1326 is one order of magnitude above the spectrum of HESS J1825-137 [22] and the associated HAWC J1825-138 [141] (see Figure 8, bottom). This could be due to the fact that all three resolved HAWC sources (HAWC J1825-138, HAWC J1826-128, and HAWC J1825-134) can contribute to the LHAASO TeV region (see their VHE spectra shown in [141]). This implies that a deeper analysis of this TeV $\gamma$-ray bright LHAASO region is necessary. It is clear that the better angular resolution of the ASTRI Mini-Array and CTA will be fundamental in order to resolve this interesting region.

In the IceCube neutrino analysis carried out by [64] (see Section 1.2), LHAASO J1825-1326 was constrained to being leptonic dominant up to 200 TeV. The Bayesian approach used in [35] could not constrain this source, stressing how these neutrino-based estimations are strongly biased by statistical methods.

Figure 8. LHAASO J1825-1326. (Top Left): HESS excess count map from [22] adapted with the addition of the LHAASO centroid and a rough estimation of the extension of the LHAASO detection from the image in [43] (white circle). (Top right): HESS 1-10 TeV excess map from [149] adapted by [43,128] and with the addition of the LHAASO centroid and its extension from [43] (orange circle). The red ellipse indicates the Eel PWN. In this way, the reason why the association of the LHAASO emission is so challenging is evident. (Bottom): Spectrum of HESS J1825-137 from [22] with the LHAASO data points from [43] overlain. Reproduced with permission from [43], Springer Nature, 2023.

Figure 8. LHAASO J1825-1326. (Top Left): HESS excess count map from [22] adapted with the addition of the LHAASO centroid and a rough estimation of the extension of the LHAASO detection from the image in [43] (white circle). (Top right): HESS 1-10 TeV excess map from [149] adapted by [43,128] and with the addition of the LHAASO centroid and its extension from [43] (orange circle). The red ellipse indicates the Eel PWN. In this way, the reason why the association of the LHAASO emission is so challenging is evident. (Bottom): Spectrum of HESS J1825-137 from [22] with the LHAASO data points from [43] overlain. Reproduced with permission from [43], Springer Nature, 2023.

3.3. LHAASO J1839-0545

LHAASO J1839-0545 was detected by LHAASO with an $E_M\approx 200$ TeV in a very complex region. The two challenging potential TeV counterparts are HESS J1841-055 and 2HWC J1837-065/HESS J1837-069/MAGIC J1837-073 [43,131].

The first, HESS J1841-055, was revealed by HESS in 2007 as an extended TeV source (∼$0.4^{\circ }$) [151] and then confirmed by HAWC [117] and ARGO [152], but it does not have a counterpart and up to now remains an unidentified source. In 2022, the MAGIC collaboration confirmed a TeV detection coincident with this HESS detection, finding some hot-spots in the extended emission that could indicate the presence of different sources in the same region that had not been resolved so far [153,154]. Their analysis considered both PWN and SNR (the middle-aged G26.6-01) origins of the TeV emission as plausible, due to the presence of a dense MC in the surroundings. The spectrum produced through Fermi-LAT and MAGIC data analysis, however, only extended up to E $<10$ TeV, with an evident cutoff at larger energies. Consequently, this evidence could exclude the association of LHAASO J1839-0545 with this complex source.

On the contrary, HESS J1837-069 is more interesting for the interpretation of the LHAASO emissions. This was first detected by HESS during its Galactic Plane Survey [131] and then by HAWC [117] and MAGIC [155], and it is postulated to be associated with the PWN of the PSR J1838-0655 [131]. A very detailed study of that region by the MAGIC collaboration at GeV-TeV energies [156] found that the TeV emission coincident with the HESS J1837-069 can be interpreted as a hadronic emission, due to runaway CRs from that source with an MC at the location of MAGIC J1837-073. In this model, a maximum energy up to 70 TeV is considered reachable; LHAASO detected this source up to 200 TeV, maybe giving an important boost to that interpretation.

Nevertheless, there is a further interesting hypothesis. In 2017, a deep analysis of Fermi-LAT data detected from the G25 region (the region that also includes the two HESS sources) seemed to indicate that there is an extended OB association, G25.18+0.26, with very similar characteristics and behavior to the Cygnus Cocoon region [157] (see Paragraph dedicated to Section 3.10). The spectrum from this region does not show any bending and could be a candidate for the TeV emission detected in this region. The very low LHAASO angular resolution also does not allow excluding this possible association, which could be another confirmation of the important role of MSCs in CR acceleration.

As shown in [156], the future IceCube-Gen2 will have the necessary sensitivity to detect possible neutrino emissions from this source, confirming or excluding the hadronic origin of this TeV emission.

3.4. LHAASO J1843-0338

LHAASO J1843-0338, detected with an $E_M\approx 260$ TeV, is in a very complex region (see Figure 9, left), and is associated problem includes the two HESS VHE detections, HESS J1844-030 and HESS J1843-033, which are candidates as TeV counterparts.

The first source, HESS J1844-030, is a faint point-like VHE source, spatially coincident with a number of distinct objects. Initially, it was reported as a component of HESS J1843-033 [158] and then was disentangled in the HESS Galactic Plane Survey [99]. In particular, this emission turned out to be spatially coincident with the radio source G29.37+0.1, which has a very complex morphology. Later, a deep study developed in [159] concluded that this radio source is the superposition of a radio galaxy, a composite SNR with a shell, a pulsar powered component, and maybe also a neutron star. At the same time, the authors could not explain if the nature of TeV $\gamma$-ray was produced in the lobe of the radio galaxy or in the SNR/PWN environment. A further analysis explained all the multiwavelength emissions from G29.37+0.1 (radio, X-ray and $\gamma$-ray) as due to PWN–SNR interaction, excluding an extra-galactic counterpart [160].

The centroid of this LHAASO source is more coincident with the second potential TeV counterpart, eHWC J1842-035/HESS J1843-033. This is an unidentified source discovered by HESS [99], with a spectrum extended up to 30 TeV. Its complex morphology could be a hint of the presence of overlapping sources, merged in the Galactic Plane Survey. In TevCat, this is associated with ARGO J1841-0332 [161] and 2HWC J1844-32 [21], and the Tibet As$\gamma$ array recently detected a TeV emission above 25 TeV near the position of HESS J1843-033, TASG J1844-038 [162]. The Tibet-AS$\gamma$ emission has a measured spectrum above 100 TeV, in perfect agreement with the LHAASO results [162] (see Figure 9, right). The authors discussed the origin of this emission, concluding that both the associations with the SNR G28.6-0.1, filled with non-thermal X-ray [163], and PSR J1844-0346, can be considered valid.

3.5. LHAASO J1849-0003

LHAASO J1849-0003 was detected by LHAASO with an $E_{max}\approx 350$ TeV [43] and one of the possible TeV counterparts is HESS J1849-000/2HWC J1849+001 [99,117].

HESS J1849-000 is firmly associated with the PWN G32.6+0.5 generated by the young and energetic PSR J1849-001 [131,164] and is one of the the nine sources detected above 56 TeV by HAWC [21]. According to [43], the extension of the detected emission could also be coincident with a YMSC in that region, W43, one of the closest and most luminous star-forming regions in the galaxy, hosting a GMC and the Wolf-Rayet WR 121a [131,165,166]. The complexity of the region and several observational challenges do not permit confirmation of this association [131] but a study of LHAASO J1849-0003, in the context of other PeV emission detections from similar YMSCs, could indicate if W43 respects all the criteria for explaining a UHE emission.

No neutrino constraints are present so far for this source.

Figure 9. LHAASO J1843-0338. (Left): Tibet AS$\gamma$ significance map of TASG J1844-038 at E $>25$ TeV [Figure from [162]]. (Right): SED of TASG J1844-038, assuming that the LHAASO data point is also correlated with this source [Figure from [162]].

Figure 9. LHAASO J1843-0338. (Left): Tibet AS$\gamma$ significance map of TASG J1844-038 at E $>25$ TeV [Figure from [162]]. (Right): SED of TASG J1844-038, assuming that the LHAASO data point is also correlated with this source [Figure from [162]].

The complexity of this region requires a VHE/UHE morphological reconstruction, in order to disentangled all the existing $\gamma$-ray sources. No neutrino constraints are present so far.

3.6. LHAASO J1908+621

LHAASO J1908+621 is one of the LHAASO sources (the others are the Crab Nebula, LHAASO J1825-1326, and LHAASO J2226+6057) detected with the highest significance at E $>100$ TeV (17.2 $\sigma$) and for which there is a SED at UHE [43]. Its maximum energy is $E_M\approx 450$ TeV and it shows a spectral steepening, possibly due to photon–photon absorption. The potential TeV counterpart is the well-known MGRO J1908+06 [116] also detected by HESS [167], ARGO [168], VERITAS [169], and finally by HAWC at E $>56$ TeV [21,170], with an extension confirmed up to the highest energies. With an off-pulsed analysis, Fermi-LAT detected two GeV excesses corresponding with the LHAASO J1908+621 contours, Fermi J1906+0626, and 4FGL J1906.2+0631 [171]. Possible associations include the middle-aged shell-type SNR G40.5-0.5, which has a MC complex in the surrounding [172,173]; the PWN of the radio PSR J1907+0631; and finally the most powerful PSR J1907+0602 (with aspin-down luminosity sufficient to explain the VHE emission), all in the same region as LHAASO J1908+621 [131] (see Figure 10, left).

In [173], the authors tried to model the whole HE spectrum, concluding that a leptonic or lepto-hadronic model was the most likely with respect to a purely hadronic one. These conclusions were confirmed also by [174], with a dominant hadronic component at the highest energies, and then by [175], with a leptonic component from the PWN J1907+0602 explaining the GeV-TeV component, and a hadronic component from the SNR/MC interaction explaining the UHE spectrum (see Figure 10, right). The one-component leptonic model was also the underdog in the time-dependent model developed in [176]. Only in [177] was a purely leptonic model favored, but this could have depended on the fact that their Fermi-LAT data analysis obtained GeV spectral points different from the ones used in the other papers [172]. A crosscheck in the GeV band is needed to better understand this complex region.

In the IceCube neutrino analysis carried out by [64], LHAASO J1908+621 was constrained to be leptonic dominant up to 200 TeV. A Bayesian approach, however, could not constrain this source, stressing how these estimations are strongly biased by statistical methods [35]. In [36], the authors estimated the neutrino fluxes expected in the case where LHAASO J1908+621 was a SNR with a completely hadronic emission, concluding that the IceCube sensitivity is insufficient to detect possible neutrino emissions. However, a low-significance neutrino emission seemed to come from this region [37], supporting the (hadronic) PeVatron nature of this source.

Figure 10. LHAASO J1908+621. (Left): radio map of the LHAASO J1908+621 region from [173] with the addition of LHAASO centroid and extension, roughly estimated from the map in [43]. The white solid lines are the VERITAS contours. (Right): one of the lepto-hadronic models developed in order to explain the SED of this source, where the contribution of hadrons at the highest energies is evident [Figure form [175]].

Figure 10. LHAASO J1908+621. (Left): radio map of the LHAASO J1908+621 region from [173] with the addition of LHAASO centroid and extension, roughly estimated from the map in [43]. The white solid lines are the VERITAS contours. (Right): one of the lepto-hadronic models developed in order to explain the SED of this source, where the contribution of hadrons at the highest energies is evident [Figure form [175]].

The higher sensitivity of future neutrino experiments, together with the higher angular resolution of future IACTs, will give a fundamental boost to the understanding of this source. This source will be observable by the ASTRI Mini-Array site for a total of about 550 h per year, and simulations have shown that already with 100 h, a high-significance spectrum up to 300 TeV can be obtained, perfectly correlated with the one detected by LHAASO [62]. Moreover, the ASTRI Mini-Array angular resolution will be able to resolve the possible counterparts in the $\gamma$-ray error box [178].

3.7. LHAASO J1929+1745

The maximum energy of LHAASO J1929+1745 is estimated to be $E_M\approx 700$ TeV [43,131] and could be associated with two different TeV counterparts: 2HWC J1928+177/HESS J1928+181 and 2HWC J1930+188/HESS J1930+188/VER J1930+188, separated by about 1 degree (see Figure 11).

2HWC J1928+177/HESS J1928+181 has almost the same centroid as the LHAASO source and is an extended $\gamma$-ray emitter discovered by HAWC [117] and then confirmed by HESS [179] (VERITAS had no detection because of the large extension of the source [180]). The recent study by [181] hypothesized three different origins for the VHE/UHE emission detected, confirmed in the third catalog as 3HWC J1928+178. Owing to the spatial correspondence with the young PSR 1928+1746, located at 0.03${}^{\circ }$, the two most probable (the same as assumed by [179]) are IC emission due to electrons diffusing away from the PWN, and CR protons produced by the pulsar then interacting with a nearby MC. However, the estimated age of the PSR (∼80,000 yrs) and the lack of X-ray detection in its correspondence open the possibility that 3HWC J1928+178 could be another TeV halo. This possibility is also favored by the need to add a further extended component to spatially model the surroundings of this source [181]. A recent study [182] analyzed the affect of the PSR proper motion on the TeV halo morphology, estimating a formula for the maximum possible offset depending on the $\gamma$-ray energies, PSR transversal velocity, and the distance. They concluded that, for E $>10$ TeV, an off-set between a pulsar and an associated extended emission cannot be explained by PSR proper motion. In particular, in order to explain LHAASO J1929+1745 with a TeV halo from the PSR 1928+1746, we need a very high transversal velocity (>2700 km/s).

The second source, 2HWC J1930+188/HESS J1930+188/VER J1930+188, was detected by all the current IACTs as a point-like source [99,117,183], and it is strongly associated with the PWN in the 2000 yr old SNR G54.1+0.3 [131], with no possibility of understanding if the $\gamma$-ray emission is from the shell or from the PWN. This also has a point-like GeV no-pulsed counterpart detected by Fermi-LAT, 3FGL J1928.9+1739 [131,180]. The latest analysis by [181] confirmed the VHE steep spectrum detected by VERITAS, which could be a hint of the presence of a cutoff (too faint to be validated). If this cutoff is confirmed, this source could be ruled out as the origin of the UHE $\gamma$-ray emission detected by LHAASO.

No neutrino constraints are present for this source so far.

Figure 11. LHAASO J1929+1745. [Figure from [181]] The complexity of the region is shown in the VERITAS excess map. White contours are HAWC significance contours, and the centroid of the LHAASO source has been added (we have no clues about its extension).

Figure 11. LHAASO J1929+1745. [Figure from [181]] The complexity of the region is shown in the VERITAS excess map. White contours are HAWC significance contours, and the centroid of the LHAASO source has been added (we have no clues about its extension).

3.8. LHAASO J1956+2845

LHAASO J1956+2845, with an $E_M\approx 420$ TeV [43], has just one potential TeV counterpart, the 2HWC J1955+285/3HWC J1954+286 [131], even if this is $0.{33}^{\circ }$ away (see Figure 12). Two candidate sources exist for the origin of this VHE/UHE emission: the shell-type middle-aged radio SNR G65.1+0.6, and the pulsar PSR J1954+2836 detected by Fermi-LAT at $E>10$ GeV [131,184].

Figure 12. LHAASO J1956+2845. [Figure from [184]] Fermi-LAT off-pulse TS map for the region of PSR J1954+2836 for 0.5–500 GeV, where all the catalog sources have been removed. The cyan contour is the approximate radio shape of the SNR G65.1+0.6.

Figure 12. LHAASO J1956+2845. [Figure from [184]] Fermi-LAT off-pulse TS map for the region of PSR J1954+2836 for 0.5–500 GeV, where all the catalog sources have been removed. The cyan contour is the approximate radio shape of the SNR G65.1+0.6.

The analysis of the GeV $\gamma$-ray data from the Fermi-LAT was the first (and the only so far) attempt to disentangled SNR and PSR emissions, in order to associate this VHE/UHE emission to one of the sources [184]. In this work, the GeV emission was attributed to SNR G65.1+0.6, which is unlikely to be the source of the VHE emission because of the lack of spatial coincidence between the HAWC and Fermi-LAT emissions. On the other hand, no pulsating emission has been detected from the PSR J1954+2836; however, the authors estimated a possible TeV halo generated from this source (based on the Geminga one parameter), suggesting that this is the VHE/UHE emission source.

Once again, the need to resolve the VHE/UHE emission region morphology with a high spatial resolution analysis and to carry out a parallel study with a comparison with the other known TeV halos and/or PWN is evident, in order to analyze the constraints on being galactic CR accelerators.

No neutrino constraints are present for this source so far.

3.9. LHAASO J2018+3651

LHAASO J2018+3651 has a maximum energy of about 270 TeV, and it is one of the three LHAASO UHE detections located in the Cygnus region (the others are LHAASO J2032+4102 and LHAASO J2108+5157), and it has two potential TeV candidates: VER J2016+371 and VER J2019+368 [43,131].

These two VERITAS sources are both associated with MGRO J2019+37, an extended $\gamma$-ray emission detected the first time by the Milagro observatory [185]. This region, one of the brightest of the Cygnus region, overlaps many different sources, but the PWN of the PSR J2021+3651 together with the star-forming region Sharpless 104 (Sh 2-104) were suggested as the main contributors [186,187] (see Figure 13, left). Later, VERITAS resolved this TeV region into the two different source candidates for the LHAASO J2018+3651 counterpart, even if the extension of the LHAASO emission cannot be resolved with the same angular resolution.

The first candidate, VER J2016+371, is an unidentified point-like source and spatially coincident with part of MGRO J2019+37 [188] and also confirmed by HAWC (HAWC J2016+371) [189]. The most likely counterpart is the PWN in the SNR CTB 87, both because of the co-location of VHE and X-ray emissions and its luminosity in the two bands [189].

Figure 13. LHAASO J2018+3651. (Left): HAWC significance map of the region of interest, with the addition of the LHAASO centroid (Figure from [189]). (Right): SED of the PWN G75.2+0.1, where the leptonic model can perfectly explain the multiwavelength spectrum from this source (Figure from [190], reproduced with permission from [190], Elsevier, 2023.).

Figure 13. LHAASO J2018+3651. (Left): HAWC significance map of the region of interest, with the addition of the LHAASO centroid (Figure from [189]). (Right): SED of the PWN G75.2+0.1, where the leptonic model can perfectly explain the multiwavelength spectrum from this source (Figure from [190], reproduced with permission from [190], Elsevier, 2023.).

The other candidate, VER J2019+368, is a bright extended source (1 degree), with an energy dependent morphology, and it represents the bulk of the emission from MGRO J2019+37. Its emission was also confirmed by HAWC (HAWC J2019+368) [189], with a high integrated flux above 56 TeV. The morphological and spectral features in both the TeV and X-ray bands point towards a PWN origin for this VHE/UHE emission [189]; particularly from the Dragonfly PWN (G75.2+0.1) powered by the PSR J2021+3561. This hypothesis was supported by a recent multiwavelength work from radio to X-ray to UHE $\gamma$-ray, also taking into account the LHAASO spectral point [190] (see Figure 13, right).

Even if we cannot say if the LHAASO J2018+3651 UHE detection corresponds with VER J2016+371 or VER J2019+368, according to the models and considerations developed up to now, the contribution from the star-forming region Sh 2-104 seems not to be the dominant one.

No neutrino constraints are present so far for this source.

3.10. LHAASO J2032+4102

LHAASO J2032+4102 has the highest maximum energy among the 12 PeVatron candidates: 1.4 PeV. Similarly to LHAASO J2018+3651 and (the next) LHAASO J2108+5157, this is located in the Cygnus region, and there are several potential TeV counterpart candidates: TeV J2032+4130/2HWC J2031+415/VER J2032+414, ARGO J2031+4157, and MGRO J2031+41 (see Figure 14).

The first source, TeV J2032+4130/2HWC J2031+415/VER J2032+414, was first detected at TeV energies by HEGRA [100] and then by HAWC [117] and VERITAS [191]. It is associated with the PWN of the PSR J2032+412 and overlaps with one of the nine sources of the first HAWC catalog, emitting at E $>56$ TeV [131]. According to [118], it is a TeV halo candidate, because it is associated with a pulsar >100 kyrs and has a TeV halo expected-flux at least as large as 2% of the Geminga flux. However, no other analyses have confirmed (or ruled out) this hypothesis. According the graph produced in that work [26] (see Figure 5), LHAASO J2032+4102 is one of two (the other is LHAASO J2108+5157, which has PSR in its surroundings) LHAASO candidate PeVatrons that cannot be explained by an associated pulsar; it lies above the absolute maximum electron energy producible by a PSR limited by radiative losses or a potential drop (see Section 2.3). This important work indicated that we need to find other sources for the origin of the UHE LHAASO emission detected from this region.

In this context, an important candidate is ARGO J2031+4157, which is the TeV counterpart of the Cygnuns Cocoon [131], which was detected for the first time at GeV energies by Fermi-LAT [101] and then confirmed at E $>100$ TeV by HAWC [103] and Tibet AS$\gamma$ [104]. In [192], the authors tried to explain the entire GeV-TeV spectrum of the Cygnus Cocoon with a model based on [19]: particle acceleration and propagation in a superbubble with multiple shocks of different strengths produced by the powerful winds of massive stars and supernovae. Their model well explains the Fermi-LAT, ARGO, and HAWC data, but it does not fit only one LHAASO spectral point. However, it can also explain the behavior of the $\gamma$-ray spectrum from Westerlund 2, suggesting that this model can work with MSCs, and questioning the association of the LHAASO spectral point with Cygnus Cocoon. Further and deeper analyses are needed to confirm this conclusion.

In [193], with a simpler acceleration model without assumptions about propagation, the authors explained the multiwavelength spectrum of the Cygnus region using a lepto-hadronic scenario (see Figure 14, right), also estimating a neutrino flux in agreement with an IceCube neutrino (see Figure 4 of [193], right). The Cygnus Cocoon, indeed, could be the only galactic source with evidence of neutrino emission (IceCube-201120A, at the IceCube sensitivity limit, not confirmed), coincident with a $\gamma$-ray excess detected by the Carpet-2 experiment [194]. If this neutrino event was correlated with the Cocoon, also taking into account the UHE LHAASO photons, this could be a hint that this source may, not only be a confirmed LHAASO PeVatron, but also a hadronic one.

The last possible TeV counterpart, MGRO J2031+41, was identified with TeV J2032+4130, but Milagro speculated that its emission could be due to more than one source and, consequently, it is listed separately in the TeVCat [131].

Figure 14. LHAASO J2032+4102. (Left): HAWC significance map of the Cocoon Region, after subtraction of the known sources, and with the addition of the LHAASO source centroid (Figure from [103], reproduced with permission from [103], Springer Nature, 2023). (Right): the multiwavelength data from the Cygnus cocoon fitted by the lepto-hadronic scenario in [193] (Figure from in [193]). From this model the authors estimated the neutrino flux expected if the source was hadronic.

Figure 14. LHAASO J2032+4102. (Left): HAWC significance map of the Cocoon Region, after subtraction of the known sources, and with the addition of the LHAASO source centroid (Figure from [103], reproduced with permission from [103], Springer Nature, 2023). (Right): the multiwavelength data from the Cygnus cocoon fitted by the lepto-hadronic scenario in [193] (Figure from in [193]). From this model the authors estimated the neutrino flux expected if the source was hadronic.

3.11. LHAASO J2108+5157

LHAASO J2108+5157 is a unique case because it has no detected TeV counterparts and it was the first source ever detected at UHE [131,195] (see Figure 15, left). Its maximum energy is about 430 TeV, and it was found to be point-like; however, without the possibility of ruling out a possible extension. It is located in the complex Cygnus region (CygnusOB7 MC), and it is the only source without a bright pulsar found in the surroundings (the first is at 3${}^{\circ }$) [26]. The Fermi-LAT extended source 4FGL J2108+5155e was found in its proximity (∼0.13${}^{\circ }$) but a physical connection is not obvious because of spectral differences [195].

Instead, a correlation with the MC [MML2017]4607 seemed to be confirmed [195] and a possible second cloud was proposed in [196], where the authors presented a pioneering study of the COB7 region, with low-resolution ${}^{12,13}$CO(2→1) observations made with the 1.85 m radio-telescope at Osaka Prefecture University, in order to investigate the nature of the LHAASO sources. This second cloud is [FKT-MC]2022, and the authors proposed it as a candidate source of the UHE from LHAASO J2108+5257, with a hadronic origin favored (MCs illuminated by accelerated CRs). Consequently, we have two MCs that suggest a hadronic scenario (see also [197]), even if the hard spectral index requires a nonlinear DSA or stochastic acceleration hypothesis.

The multiwavelength analysis carried on in [198], putting together LST-1, XMM-Newton, and Fermi-LAT results, could not rule out any scenario, leptonic or hadronic (see Figure 15, right). LST-1 and LHAASO data can be explained by the IC scenario but the absence of X-ray emissions implies a weak magnetic field, favoring a TeV Halo scenario over the PWN one. However, there is no evidence of a pulsar. Nevertheless, in their work, a more complete combined analysis of both leptonic and hadronic components in the region is missing, and this could allow obtaining more constraining parameters.

The estimated neutrino flux in a complete hadronic scenario is within the sensitivity range of future instruments, giving to us a chance to understand this challenging LHAASO source [197].

Figure 15. LHAASO J2108+5157. (Left): significance map around LHAASO J2108+5157 as observed by KM2A at E $>100$ TeV [Figure from [195]], © AAS. Reproduced with permission.]. (Right): the leptonic (top) and hadronic (bottom) components that can explain the multiwavelength SED of this source in [198] (Figure from [198]).

Figure 15. LHAASO J2108+5157. (Left): significance map around LHAASO J2108+5157 as observed by KM2A at E $>100$ TeV [Figure from [195]], © AAS. Reproduced with permission.]. (Right): the leptonic (top) and hadronic (bottom) components that can explain the multiwavelength SED of this source in [198] (Figure from [198]).

3.12. LHAASO J2226+6057

LHAASO J2226+6057 is one of the four LHAASO sources (the other are the Crab Nebula, LHAASO J1825-1326 and LHAASO J1908+621) detected with the highest significance at E $>100$ TeV (13.6 $\sigma$) and for which there is a SED at UHE. Its maximum energy is $E_M\approx 570$ TeV and it shows spectral steepening, maybe due to photon–photon absorption [43,131]. Its TeV counterparts are VER J2227+608/HAWC J2227+610 [131], and there are two possible sources explaining this VHE/UHE emission: the SNR G106.3+2.7, with the associated MC, and the Boomerang PWN, associated with the PSR J2229+6141. The SNR is located in the “tail” of the VHE emission, with the PWN on the “head”.

The UHE detection by HAWC [199], Tibet AS$\gamma$ [104], and finally by LHAASO [43] has a low angular resolution and, consequently, we were not able to say if it is from the head or the tail region. Recently, a 12-year Fermi-LAT GeV analysis of the region showed that at the highest energies (10–500 Gev), only the tail is emitting $\gamma$-ray (see Figure 16, top left) and a hadronic model from the SNR/MC interaction can explain the entire HE/VHE/UHE spectrum [54]. Recently, the MAGIC collaboration resolved the VHE/UHE emission and detected E $>10$ TeV only from the tail region [52,200], leaning towards a hadronic explanation of the $\gamma$-ray emission [199,200], and linked to the SNR/MC interaction (or MC illumination form CRs escaping from the SNR [53]). The hadronic model used by the MAGIC collaboration also took into account the GeV Fermi-LAT data points from the tail region from [201] (consistent with the more recent results from [54]) (see Figure 16, top right).

The same detection was explained using hadronic emission, but from the PWN and not from the SNR, by [201], due to the hard spectral index and low energy content of CRs. A PWN origin was also investigated by [55], concluding that, if a part of the UHE $\gamma$-ray emission comes from the PWN, it has to be of hadronic origin.

The real origin of the hadronic emission will only be understood with a deep analysis of the microphysics of the region but, regardless of a PWN or SNR origin, a hadronic origin is also supported by the analysis of nonthermal X-ray radiation [202,203], detected everywhere but with an enhancement of the luminosity in the head region, and well correlated with the radio emissions but not with the $\gamma$-ray emissions.

Nevertheless, there have been some works that analyzed a possible leptonic or lepto-hadronic origin of UHE emissions of LHAASO J2226+6057. In [204], before the LHAASO and Tibet AS$\gamma$ detection, the entire spectrum from VER J2227+608 was explained with both leptonic and hadronic scenarios, and the authors concluded that only UHE data can discern the two models. However, an overlap of the later LHAASO and Tibet AS$\gamma$ data on both their models makes it clear that neither model is confirmed (see Figure 16, bottom), requiring a further analysis with constraints on parameters computed using the LHAASO results. In both [176,205], the authors developed a simplistic one-zone leptonic model (but complete with reverberation effect and PWN age estimation), assuming that the UHE emission comes from the Boomerang PWN. They obtained a good fit but a challenging low value for the magnetic field and a PWN size too large, maybe due to the very simple model used.

Figure 16. LHAASO J2226+6057. (Top left): Fermi-LAT residual significance map of the region for 0.1–500 GeV, where the concentration of the GeV emission in the tail region is evident. White contours are the radio continuum emissions at 1420 MHz (Figure from from [54], reproduced with permission from [54], American Physical Society, 2023.). (Top right): the hadronic model used by the MAGIC collaboration to fit all data from the tail region. Milagro, HAWC, TibetAS$\gamma$, and LHAASO shown here are from extraction regions that partially include the head, consequently they could be contaminated (Figure from [200]). (Bottom): Spectral fit to the spectral energy distribution of Ver J2227+608 from [204], with the leptonic model on the left and the hadronic one on the right. This was developed before the latest Tibet AS$\gamma$ and LHAASO data, which we have overlapped, showing that they rule out both models (Figure from [204], © AAS. Reproduced with permission.).

Figure 16. LHAASO J2226+6057. (Top left): Fermi-LAT residual significance map of the region for 0.1–500 GeV, where the concentration of the GeV emission in the tail region is evident. White contours are the radio continuum emissions at 1420 MHz (Figure from from [54], reproduced with permission from [54], American Physical Society, 2023.). (Top right): the hadronic model used by the MAGIC collaboration to fit all data from the tail region. Milagro, HAWC, TibetAS$\gamma$, and LHAASO shown here are from extraction regions that partially include the head, consequently they could be contaminated (Figure from [200]). (Bottom): Spectral fit to the spectral energy distribution of Ver J2227+608 from [204], with the leptonic model on the left and the hadronic one on the right. This was developed before the latest Tibet AS$\gamma$ and LHAASO data, which we have overlapped, showing that they rule out both models (Figure from [204], © AAS. Reproduced with permission.).

It is clear that this source would be very interesting for future IACTs such as ASTRI Mini-Array and CTA. Both the collaborations have made simulations of this region, in order to evaluate the contribution of these two instruments to understanding it. In [206], it is shown how the CTA-North could reconstruct the spectrum with a very high precision; in [62], it is shown that a detection with the ASTRI Mini-Array could disentangled hadronic and leptonic models. Moreover, the very good angular resolution of this array could definitely resolve the VHE emission location [178].

Assuming a completely hadronic origin of its $\gamma$-ray emission, in [36], the authors estimated the neutrino fluxes expected, concluding that the IceCube sensitivity was insufficient to detect possible neutrino emissions.

4. The Future

Our space and “on-Earth” facilities have allowed us to know a lot about CRs and their sources. However, this is not sufficient to have a definitive answer to the question “What is the origin of CRs?”.

After the review of all the LHAASO candidate PeVatrons, it is quite clear what ingredients are required to have a better understanding of galactic CRs:

• Higher angular resolution of the future VHE/UHE $\gamma$-ray instruments;
• Higher sensitivity of the future neutrino experiments;
• Multiwavelength and multi-messenger analysis of the same source;
• Improvements in the comprehension of the micro-physics of different kinds of source.

According to the observational evidence, a very high angular resolution is the most important feature in an instrument for looking into these sources. This would allow a much better source association of detected $\gamma$-ray emissions, a deeper understanding of a possible energy dependent morphology, and to strengthen multiwavelength correlations. The highest possible energy sensitivity is also fundamental, but LHAASO has proven that this is sufficient to identify the presence of PeVatrons in our Galaxy but not to make a firm association with a parent source and to discern hadronic from leptonic emissions. Morphological information will also be crucial for understanding certain specific parameters of candidate PeVatrons. For example, compact sub-degree regions in PWNe indicate a high magnetic field and fast synchrotron losses, and the detection of extended X-ray nebula should mean that LHAASO $\gamma$-ray detection originates from the up-scattering of the 2.7 K CMB radiation in PWNe [26]. Such a detection is challenging for the current X-ray instruments (XMM-Newton, Chandra), but will be possible for the eROSITA satellite, which will have a larger FoV [26].

Even if the IACTs have some evident disadvantages with respect to the EAS arrays (e.g., limited duty cycle (10–15%) or limited FoV), their low-energy threshold and high angular resolution (<$0.1^{\circ }$ at 1 TeV) make IACTs ideal for detailed morphological studies. For this reason, instruments such as the ASTRI Mini-Array and CTA will make a difference (see Figure 17, right). In addition, the higher sensitivity of the ASTRI Mini-Array, and in particular of CTA South, will allow enhancing the number of VHE sources detected in the TeV band, up to a factor of several hundred [70] (see Figure 17, left).

Figure 17. (Left): Angular resolution of the current and future instruments from [207] with the addition of the ASTRI Mini-Array updated curve with optimized cuts from [208] and the LHAASO curve from [77]. (Right): Differential sensitivity curves of the current and future VHE/UHE instruments from [209], expanded with the addition of the ASTRI Mini-Array curves for 50 h and 500 h and other instrument curves [208].

Figure 17. (Left): Angular resolution of the current and future instruments from [207] with the addition of the ASTRI Mini-Array updated curve with optimized cuts from [208] and the LHAASO curve from [77]. (Right): Differential sensitivity curves of the current and future VHE/UHE instruments from [209], expanded with the addition of the ASTRI Mini-Array curves for 50 h and 500 h and other instrument curves [208].

The ASTRI Mini-Array will have its first three telescopes operative at the beginning of 2024 and the whole array in the first half of 2025. This means that its first results will be parallel or just before the full operation of IceCube-Gen2 (2030s), KM3NeT-ARCA, and P-ONE (after 2030) in the next 5–10 years, preparing the ground for a future $\gamma$-ray–neutrino combined analysis.

In [35], the authors estimated the observational time it takes for a 5$\sigma$ detection using the so-called Planetary Neutrino Monitoring System (PLE$\nu$M) [210], a global repository of HE neutrino observations by current and future neutrino telescopes. According to their computations (and relative assumptions on the threshold energy and models), they found that in the next 20 years the PLE$\nu$M could discover neutrino emissions from the Crab Nebula, LHAASO J1825-1326, LHAASO J1839-0545, LHAASO J1908+0621, LHAASO J2018+3651, and LHAASOJ2226+6057. According to the models developed in [36], IceCube-Gen2 and KM3NeT will be able to detect neutrinos from SNRs (with a parameter space based on the two SNRs analyzed) with the condition that the neutrino fluxes are above ${10}^{-3}$ TeV cm${}^{-2}$ s${}^{-1}$ for $E>10$ TeV, and IceCube-Gen2 will be able to detect all the LHAASO sources in case where they are neutrino emitters (see Figure 18). These estimations are strongly dependent on the models used to estimate the parameter space; however, they show that there will be a “neutrino bright” future for real hadronic PeVatrons.

All the improvements in the technology of future $\gamma$-ray and neutrino experiments will have to be accompanied by increasingly detailed theoretical models able to characterize in a more stringent manner the different source types and exploiting more detailed experimental data.

Figure 18. Detection horizon of LHAASO sources (colored stars) for different neutrino detectors (figure and details in [36]). This is a model-dependent estimation, but we show it in order to stress the important role of future instruments in the PeVatron context. The orange band shows the energy integrated ${\nu }_{\mu }$ flux as a function of distance for the parameters used in Figure 3 of [36].

Figure 18. Detection horizon of LHAASO sources (colored stars) for different neutrino detectors (figure and details in [36]). This is a model-dependent estimation, but we show it in order to stress the important role of future instruments in the PeVatron context. The orange band shows the energy integrated ${\nu }_{\mu }$ flux as a function of distance for the parameters used in Figure 3 of [36].

5. Conclusions

This is a very exciting moment for VHE/UHE astrophysics and in particular for the CR origin issue. After just over a century, we not have only a deeper understanding of the physics behind CR acceleration but also have a lot of data that are allowing us to correct our path and better know our galaxy and their sources. We know that supernova remnants accelerate the lowest energy component of the galactic CRs, but we need other types of sources to explain CR PeV energies, their spectrum, and their composition. Superbubbles and massive star clusters are the main candidates, explaining PeV energies and CR composition, but we need more detailed models, in order to understand the acceleration mechanism in their environment. The standard leptonic-accelerator PWNe contributes to CR acceleration but likely a small percentage with respect to the other source typologies.

Moreover, most of the LHAASO candidate PeVatrons seem to be associated with PWNe, and this can be explained as a statistical bias due to the fact that PWNe are the most numerous sources in our $\gamma$-ray bright galaxy. Only LHAASO J0534+2202 has a firm candidate, the Crab Nebula, and all the other sources could be associated not only with PWNe but also with SNRs, MSCs, and TeV halos. Only a higher sensitivity and spatial resolution multiwavelength and multi-messenger analysis will be able to clarify most of these confusing associations.

LHAASO has demonstrated, once again, how an improvement in technology leads to an advance in knowledge, especially in the field of astrophysics. We can thus define an epoch “after LHAASO” ( A.L.), when the term PeVatrons “exploded” in scientific papers [209], and in this brief review, we have tried to summarize the huge effort made by the scientific community over these last two years, based on a strong theory and experimental background before LHAASO (B.L.), to make sense of the very large amount of information that we have in our hands. A summary of the main information about every LHAASO source is shown in the Table in Figure 19.

Figure 19. A summary table of all LHAASO VHE sources, with their coordinates, significance at 100 TeV, and maximum energy from [43]. The last three columns are based on the review in Section 3: the “origin” column has in bold the most probable type of source, the “neutrino constraints” indicates if we can obtain information on $\gamma$-ray emission from neutrino analysis/estimation, and the last column indicates the most likely origin of the emission according to the analyses performed to date.

Figure 19. A summary table of all LHAASO VHE sources, with their coordinates, significance at 100 TeV, and maximum energy from [43]. The last three columns are based on the review in Section 3: the “origin” column has in bold the most probable type of source, the “neutrino constraints” indicates if we can obtain information on $\gamma$-ray emission from neutrino analysis/estimation, and the last column indicates the most likely origin of the emission according to the analyses performed to date.

The future instruments for $\gamma$-ray and neutrino detection will have a huge responsibility, because our comprehension of galactic PeVatrons will depend on their performance, which we hope will be as good or better than we are estimating during their development. When we will have the first data, there is a great chance to put one more piece into the CR origin puzzle. At the same time, further measurements of the CR spectrum, composition, and anisotropy could help us have a better understanding of the physics behind CR acceleration, escape, and propagation.

We want to conclude with a quote from Felix Aharonian about the origin of CRs: “It is not correct to still speak of mystery because we know a lot of things about them”. A lot of work has been done with beautiful and outstanding results, and a bright future in $\gamma$-ray and neutrino research is waiting for us.