A Very High Energy γ-Ray Survey toward the Cygnus Region of the Galaxy

In the vicinity of Cygnus OB2, the γ-ray source TeV J2032+4130 was the first unidentified VHE γ-ray source. It was discovered with the HEGRA IACTs (Aharonian et al. 2005) and has been further studied by the Whipple (Konopelko et al. 2007), MAGIC (Albert et al. 2008), and VERITAS (Aliu et al. 2014a) IACTs. As the position of the source is coincident with the Cygnus OB2 star association and located north of the microquasar Cygnus X-3, it was suggested that these two sources could be the origin of the VHE γ-rays (Aharonian et al. 2002). More recently, it has been argued that the VHE source could be the PWN of the GeV pulsar PSR J2032+4127 (Aliu et al. 2014a; Camilo et al. 2009). Lyne et al. (2015) suggested that PSR J2032+4127 is part of a binary system with a 15 M_⊙ Be star, though the authors argue that this binary might not be able to fully power the VHE emission. Observations of the region during the 2017 November periaston found correlated X-ray and γ-ray emission from the direction of the pulsar, confirming the binary nature of the system (Veritas & MAGIC Collaborations 2017); a paper is in preparation to present these results.

VER J2031+415 is located in the Cygnus Cocoon, a large excess of hard spectrum emission detected by Fermi-LAT that is associated with the star-forming region Cygnus X and has the potential to be detected at VHE energies (Ackermann et al. 2011). The extensive air shower arrays Milagro and ARGO-YBJ have detected extended emission coming from the same region (MGRO J2031+41, Abdo et al. 2007; ARGO J2031+4157, Bartoli et al. 2012) at energies above 20 and 1 TeV, respectively, and these sources have been associated with the Cygnus Cocoon. The Cygnus Cocoon overlaps with, but extends beyond, TeV J2032+4130, and the extent of this emission makes it unlikely to be detectable by VERITAS using conventional techniques. Localized sources of emission (of which TeV J2032+4130 could be an example) within the Cygnus Cocoon may be detectable, and IACT observations may be able to identify the sources that are powering this emission. Identification of the origin of the VHE emission from TeV J2032+4130 could help clarify the origin of the emission from this complex region.

In this analysis, VER J2031+415 was observed by VERITAS at a peak (local) significance of 10.1σ using the Extended integration radius. As in Aliu et al. (2014a), the source was found to be extended and asymmetrical. The extension was fit with an asymmetrical Gaussian function and is given in Table 4.

Table 4.  Source Extension of the VERITAS Sources

Source	Centroid		σsemimajor	σsemiminor	Rotation Angle
	(l, b)	(${\alpha }_{J2000}$, ${\delta }_{J2000}$)	(deg)	(deg)	(deg)
VER J2031+415	(8025 ± 001stat ± 001sys, 120 ± 001stat ± 001sys)	(${20}^{{\rm{h}}}{31}^{{\rm{m}}}{33}^{{\rm{s}}}$, $41^\circ 34^{\prime} 48^{\prime\prime}$)	$0.19\pm {0.02}_{\mathrm{stat}}\pm {0.01}_{\mathrm{sys}}$	$0.08\pm {0.01}_{\mathrm{stat}}\pm {0.03}_{\mathrm{sys}}$	$13\pm {4}_{\mathrm{stat}}\pm {1}_{\mathrm{sys}}$
VER J2019+407	(7830 ± 002stat ± 001sys, 255 ± 001stat ± 001sys)	(20h20m05s $40^\circ 45^{\prime} 36^{\prime\prime}$)	$0.29\pm {0.02}_{\mathrm{stat}}\pm {0.02}_{\mathrm{sys}}$	$0.19\pm {0.01}_{\mathrm{stat}}\pm {0.03}_{\mathrm{sys}}$	$176.7\pm {0.1}_{\mathrm{stat}}\pm {2}_{\mathrm{sys}}$
VER J2019+368	(7497 ± 002stat ± 001sys, 035 ± 001stat ± 001sys)	${20}^{{\rm{h}}}{19}^{{\rm{m}}}{23}^{{\rm{s}}}$, $36^\circ 46^{\prime} 44^{\prime\prime}$)	$0.34\pm {0.02}_{\mathrm{stat}}\pm {0.01}_{\mathrm{sys}}$	$0.14\pm {0.01}_{\mathrm{stat}}\pm {0.02}_{\mathrm{sys}}$	$127.0\pm {2.6}_{\mathrm{stat}}\pm {0.1}_{\mathrm{sys}}$
VER J2018+367*	(7487 ± 001stat ± 001sys, 042 ± 001stat ± 001sys)	(${20}^{{\rm{h}}}{18}^{{\rm{m}}}{48}^{{\rm{s}}}$, $36^\circ 44^{\prime} 24^{\prime\prime}$)	$0.18\pm {0.01}_{\mathrm{stat}}\pm {0.04}_{\mathrm{sys}}$	Symmetric
VER J2020+368*	(7513 ± 001stat ± 001sys, 019 ± 001stat ± 001sys)	(${20}^{{\rm{h}}}{20}^{{\rm{m}}}{31}^{{\rm{s}}}$, $36^\circ 49^{\prime} 12^{\prime\prime}$)	$0.03\pm {0.01}_{\mathrm{stat}}\pm {0.01}_{\mathrm{sys}}$	Symmetric
VER J2016+371	(7494 ± 001stat ± 001sys, 116 ± 001stat ± 001sys)	(${20}^{{\rm{h}}}{15}^{{\rm{m}}}{57}^{{\rm{s}}}$, $37^\circ 12^{\prime} 31^{\prime\prime}$)	Point Source

Note. The source extension was determined by fitting a sky map of the excess events with a 2D Gaussian distribution convolved with the VERITAS PSF. The rotation angle is given east of galactic north.

Download table as:  ASCIITypeset image

The Fermi-LAT emission in the region is dominated by the pulsar PSR J2032+4127 (3FGL J2032.2+4126), which cuts off sharply at a few GeV. To investigate the possibility of an additional source (e.g., a PWN) that is masked by the emission from PSR J2032+4127, a gated analysis was performed to remove the pulsed component using the method outlined in Section 3, with the OFF-pulse selected using phases 0.1 to 0.4 and 0.5 to 0.9. This analysis showed an extended residual, as depicted in Figure 11, with a centroid of (l, b) = (8024 ± 003_stat, 104 ± 003_stat) ((${\alpha }_{J2000}$, ${\delta }_{J2000}$) = (${20}^{{\rm{h}}}{32}^{{\rm{m}}}{13}^{{\rm{s}}}$, $41^\circ 28^{\prime} 39^{\prime\prime}$)). This residual is fit by an extended symmetric Gaussian source of 68% containment radius of $0\buildrel{\circ}\over{.} {15}_{-0\buildrel{\circ}\over{.} 03}^{+0\buildrel{\circ}\over{.} 02}$ centered on this location. Thus, we name the source FGL J2032.2+4128e. This source has a TS of 321.1, and the TS of extension is 28.6. A uniform disk source was also tested; the best fit had a 68% containment radius of $0\buildrel{\circ}\over{.} 15{}_{-0\buildrel{\circ}\over{.} \,02}^{+0\buildrel{\circ}\over{.} \,02}$ and a TS of extension of 24.7. The Gaussian source morphology was preferred and used for the rest of this work.

Zoom In Zoom Out Reset image size

A cross-check using an OFF-pulse analysis identified a similar residual with a comparable morphology, although, with the smaller statistics due to the combined effect of the phase cut and limited time period of the ephemeris (about a factor of 2 reduction in live time), the signal is weaker. The similarity in the morphology between these two methods of removing the pulsar and the offset and extension in the residual is supportive of this residual being due to a new source rather than an artifact due to errors in fitting to PSR J2032+4127.

An additional analysis was conducted above 5 GeV; this has the advantage that the PSF is reduced at these energies (to around 03 at 5 GeV), which will improve the ability to separate the two sources. This also found moderately extended emission potentially associated with VER J2031+415 with a TS of 31.7 as shown in Figure 11. A fit to this emission was conducted, and it was found to peak at (l, b) = (8021 ± 004_stat, 110 ± 004_stat) ((${\alpha }_{J2000}$, ${\delta }_{J2000}$) = (${20}^{{\rm{h}}}{31}^{{\rm{m}}}{54}^{{\rm{s}}}$, $41^\circ 29^{\prime} 24^{\prime\prime}$) with an extension of $0.14\pm {0.05}_{\mathrm{stat}}$ and a TS of extension of 10.6.

It is noted that an analysis of the Fermi-LAT data from the Cygnus Cocoon region is very complex, and in this analysis, a simplistic model of the cocoon emission was used (a single symmetric Gaussian as in the 3FGL). Further work with a more detailed morphology of the Cygnus Cocoon has the potential to provide significantly more insight here but is beyond the scope of this work. In particular, there are significant challenges associated with disentangling the emission from the various objects and identifying whether an observed region of emission is associated with a peak in the emission from the Cygnus Cocoon or another source. At present, VHE emission is detected from either a significantly extended region (Milagro and ARGO-YBJ) associated with the Cygnus Cocoon or a smaller, moderately extended region (VERITAS, MAGIC, HEGRA, HAWC, and Whipple) associated with TeV J2032+4130. Understanding the relationship between these VHE observations will significantly improve our ability to understand this region in other wavelengths. To do this will require further development of tools such as the "3D Maximum Likelihood Method" currently being developed (Weinstein 2014; Cardenzana 2017), which will allow for observations of significantly extended objects by VERITAS. This will then allow the superior PSF of VERITAS to be used to study the morphology of the Cygnus Cocoon, providing information that could be used to guide the HE analysis. Until then it is not possible to definitively claim the detection of an HE counterpart to VER J2031+415, nor to fully understand the relationship between the two objects. However, the presence of extended emission with similar morphology to those observed in VER J2031+415 is highly suggestive that there is HE emission associated with the source.

The spectra of VER J2031+415 and FGL J2032.2+4128e are shown in Figure 12. The VERITAS spectrum is given in Table 5, with the spectral points given in Table 8, and is consistent with the spectrum presented in Aliu et al. (2014a). The spectrum of HE emission can be described by a power law with index $2.52\pm {0.07}_{\mathrm{stat}}$ and a normalization of $(1.44\,\pm {0.09}_{\mathrm{stat}})\times {10}^{-8}\,{\mathrm{GeV}}^{-1}\,{\mathrm{cm}}^{-2}\,{{\rm{s}}}^{-1}$ at 2.27 GeV. Comparing the spectra of FGL J2032.2+4128e and 3FGL J2032.2+4126 shows that, below 1 GeV, the extrapolated flux from FGL J2032.2+4128e would be stronger than that from 3FGL J2032.2+4126. This is clearly not the case, as it has not been detected in lower-energy analyses. It is likely, therefore, that at low energies some of the emission from 3FGL J2032.2+4126 and potentially the Cygnus Cocoon is being included in the measured flux from FGL J2032.2+4128e.

Zoom In Zoom Out Reset image size

Table 5.  Observation Results for the VERITAS Sources

Source	Integration	ON	OFF	α	Sig.	Fit Range	N0	E0	γ	χ2/dof
	Region				(σ)	(GeV)	(GeV−1cm−2 s−1)	(GeV)
VER J2031+415	Extended	871	5181	0.114	10.1	422–42200	(1.24 ± 0.24stat ± 0.24sys)E−16	2300	$2.00\pm {0.19}_{\mathrm{stat}}\pm {0.20}_{\mathrm{sys}}$	3.02/5
VER J2019+407	Extended	598	2767	0.151	7.6	750–7500	(5.01 ± 0.93stat ± 1.00sys)E−16	1500	$2.79\pm {0.39}_{\mathrm{stat}}\pm {0.20}_{\mathrm{sys}}$	3.27/2
VER J2019+368	Extended	951	4230	0.153	10.3	422–42200	(1.02 ± 0.11stat ± 0.20sys)E−16	3110	$1.98\pm {0.09}_{\mathrm{stat}}\pm {0.20}_{\mathrm{sys}}$	8.73/6
VER J2018+367*	Point	196	3692	0.028	7.8	750–23700	(5.12 ± 0.94stat ± 1.48sys)E−17	2710	$2.00\pm {0.21}_{\mathrm{stat}}\pm {0.2}_{\mathrm{sys}}$	2.81/4
VER J2020+368*	Point	191	3456	0.030	7.5	750–23700	(3.00 ± 0.56stat ± 0.60sys)E−17	3270	$1.71\pm {0.26}_{\mathrm{stat}}\pm {0.2}_{\mathrm{sys}}$	4.71/4
VER J2016+371	Point	157	2370	0.038	6.3	680–14700	(2.8 ± 1.2stat ± 1.2sys)E−17	2510	$2.1\pm {0.8}_{\mathrm{stat}}\pm {0.4}_{\mathrm{sys}}$	0.38/2

Note.  ON, OFF, and α are from the RBM analysis. The spectral analysis is conducted using the RR method (Section 2), and all sources were best fit with a power law (Equation (1)).

Download table as:  ASCIITypeset image

A joint fit to the VERITAS and Fermi-LAT data points for VER J2031+415/FGL J2032.2+4128e is well fit with a PL of index $2.39\pm {0.03}_{\mathrm{stat}}$ and normalization $(3.61\pm {0.21}_{\mathrm{stat}})\,\times {10}^{-10}\,{\mathrm{GeV}}^{-1}\,{\mathrm{cm}}^{-2}\,{{\rm{s}}}^{-1}$ at 4.04 GeV, with a χ² of 9.9 for 9 degrees of freedom (dof). However, it is clearly evident that the low-energy Fermi-LAT spectral points are contaminated with emission from 3FGL J2032.2+4126. To reduce the impact of this and better match the regions from which the VERITAS spectrum was extracted, the analysis above 5 GeV was used. This spectrum (index = $2.33\,\pm \,{0.22}_{\mathrm{stat}}$, normalization = $(8.80\,\pm {1.76}_{\mathrm{stat}})\times {10}^{-12}\,{\mathrm{GeV}}^{-1}\,{\mathrm{cm}}^{-2}\,{{\rm{s}}}^{-1}$ at 15.1 GeV, Figure 12) shows good agreement with the VERITAS spectrum for VER J2031+415, and together they are well fit with a PL of index $2.22\pm {0.06}_{\mathrm{stat}}$ and normalization $(2.28\pm {0.33}_{\mathrm{stat}})\,\times {10}^{-15}\,{\mathrm{GeV}}^{-1}\,{\mathrm{cm}}^{-2}\,{{\rm{s}}}^{-1}$ at 221 GeV with a χ² of 4.9 for 9 dof.

VER J2031+415 has been detected by HAWC and is listed in the 2HWC catalog as 2HWC J2031+415. The VERITAS location lies just outside the HAWC 1σ uncertainty at a distance of 011 (though the HAWC position lies within the 1σ extent of the VERITAS emission). HAWC sees evidence of extension with the source fit with both a point source and uniform disk of radius 07. This extension is larger than that measured by VERITAS and closer to the results from Milagro. Using both extensions, HAWC detects a higher flux ($(3.24\pm {0.32}_{\mathrm{stat}})\times {10}^{-17}\,{\mathrm{GeV}}^{-1}\,{\mathrm{cm}}^{-2}\,{{\rm{s}}}^{-1}$ and $(6.16\,\pm {0.44}_{\mathrm{stat}})\times {10}^{-17}\,{\mathrm{GeV}}^{-1}{\mathrm{cm}}^{-2}\,{{\rm{s}}}^{-1}$ at 7000 GeV for the point and extended fit, respectively) and softer spectral index ($-2.57\pm {0.07}_{\mathrm{stat}}$ and $-2.52\pm {0.05}_{\mathrm{stat}}$) than reported in this work (and in previous observations by IACTs); this is likely due to contributions from the extended emission that is not detected by VERITAS.

With similar morphologies and spectral properties, it is likely that both the HE and VHE emission share a common origin. Given the proximity of the emission from both sources to the pulsar PSR J2032+4127, a PWN origin of this emission is a strong possibility. However, Aliu et al. (2014a) comment that, taking into account the hard spectral index (∼−2) and due to the Klein–Nishina effect, a cutoff in the spectrum should be seen close to 10 TeV in order to be consistent with the PWN interpretation. Though it cannot be excluded using these data alone, there is no evidence of a spectral cutoff in these data, which would be expected with a PWN origin for the emission.

Multiwavelength images of TeV J2032+4130 and its vicinity are shown in Figure 13. The infrared images from Spitzer are dominated by bright diffuse emission exhibiting complex structure caused by star-forming activity taking place in the Cygnus X complex. In addition, as noted in Aliu et al. (2014a), nearly all the VHE γ-ray emission happens to be confined within a void. One possible explanation proposed by Butt et al. (2003) is that the large, mechanical power density from the stellar winds of the OB stars could be powering the emission. However, even though the energy required to power the VHE emission is a fraction of the estimated wind energy, almost all of the massive stars are located outside of the VHE emission region. It is also plausible that a supernova exploded within Cygnus OB2 and its remnant is expanding into the surrounding medium, creating the void and powering the γ-ray emission.

Zoom In Zoom Out Reset image size

The γ-ray emission in the vicinity of the SNR G78.2+2.1, located in the region of Gamma Cygni, was first detected by the EGRET instrument on board the Compton Gamma Ray Observatory (CGRO; Thompson et al. 1995). There is a bright Fermi-LAT pulsar, PSR J2021+4026 (3FGL J2021.5+4026), at the center of the remnant (Abdo et al. 2010a, 2010b), as well as extended γ-ray emission above 10 GeV (3FGL J2021.0+4031e) with an extension consistent with that seen in radio observations of G78.2+2.1 (Lande et al. 2012). A VHE source, VER J2019+407, was detected by VERITAS at a level of 7.5σ on the northwestern rim of the radio shell of the SNR G78.2+2.1 (Aliu et al. 2013). HAWC has detected a source (2HWC J2020+403) that lies within G78.2+2.1, 043 away from the location of VER J2019+407 and close to PSR J2021+4026. They do not report on any evidence of extension, and they report a power-law spectrum for flux $(1.85\,\pm {0.26}_{\mathrm{stat}})\times {10}^{-17}\,{\mathrm{GeV}}^{-1}\,{\mathrm{cm}}^{-2}\,{{\rm{s}}}^{-1}$ at 7000 GeV and spectral index ($-2.95\pm {0.10}_{\mathrm{stat}}$).

3FGL J2021.0+4031e is generally modeled as a uniform disk of radius 063 (Acero et al. 2015), though Ackermann et al. (2011) found an improvement in their modeling of the region if, in addition to a disk source for the entire remnant, they included an elliptical source based on the morphology presented in Weinstein et al. (2009). Fraija & Araya (2016) conducted an analysis above 4 GeV that showed a region of enhanced emission in the GeV energy range over and above the uniform disk model that coincides with the VHE emission from VER J2019+407.

The radio emission is brightest at the opposite side of the SNR from VER J2019+407, on the southern edge of the remnant. Leahy et al. (2013) discuss the radio and X-ray emission in detail and note that there are several possible sources of radio emission in the same line of sight as the southern edge of the remnant, which could make the radio emission in that region appear higher than it actually is. In contrast, by examining ROSAT X-ray observations, they show that the diffuse emission is brightest at the northern and western edges. It is noted that the northern emission may, in part, be due to the presence of a large X-ray shell associated with the early-type star V1685 Cyg, which lies in the line of sight of G78.2+2.1 but at between one-half and one-third of the distance (980 pc, versus 1.7–2.6 kpc for G78.2+2.1; Green 2014). G78.2+2.1 lies within the Cygnus Cocoon and could power at least some of the high-energy emission seen from the region.

In this analysis, VER J2019+407 was observed by VERITAS at a peak (local) significance of 7.6σ using the Extended integration region. The source was found to be extended and asymmetric, unlike the morphology reported in Aliu et al. (2013), where a symmetric fit was favored. It was fit with an asymmetric Gaussian with the results given in Table 4. This difference in morphology likely reflects the improved PSF in this analysis in comparison to the discovery paper (Aliu et al. 2013) resulting from the requirement of three telescopes in the reconstruction and a higher image intensity requirement, as well as the larger data set.

The morphology of 3FGL J2021.0+4031e was examined by producing a TS map of the region with 3FGL J2021.0+4031e removed from the model of the region (Figure 14). The Fermi-LAT emission showed a disk-like structure, of similar size to the remnant detected in the CGPS 1420 MHz survey (Taylor et al. 2003). However, the HE γ-ray emission was not uniform across the disk; rather, it was enhanced at the northern rim, where the VERITAS emission is detected.

Zoom In Zoom Out Reset image size

The spectral fit to the VERITAS data is given in Table 5, with the spectral points given in Table 9. The updated analysis has a spectrum that is softer, but within the statistical errors of the earlier spectrum reported in Aliu et al. (2013), with a normalization at 1000 GeV that is in agreement with the value in that work.

Two Fermi-LAT SEDs were produced for 3FGL J2021.0+4031e. The first uses the full disk of radius 063 as in the 3FGL catalog. For the second, 3FGL J2021.0+4031e was broken up into two sources: the region inside the VERITAS Extended integration region (a disk of radius 023 centered on (l, b) = (78.38, 2.56), 3FGL J2021.0+4031eA) and the region from which the VERITAS spectrum presented in Table 5 is extracted, as well as the remainder of the 063 radius disk outside of the VERITAS integration region, 3FGL J2021.0+4031eB. This was done to investigate whether the emission seen by VERITAS has a common origin with that observed with the Fermi-LAT. Also, it was used to test whether the level of enhancement seen in the Fermi-LAT TS map of the remnant is reflected in a change in the spectral properties. In both cases, the favored fit is a PL. When treated as a single disk, 3FGL J2021.0+4031e is detected at a TS of 910, with spectral parameters of γ = $2.02\pm {0.03}_{\mathrm{stat}}$, N₀ = $(2.50\,\pm {0.10}_{\mathrm{stat}})\times {10}^{-13}\,{\mathrm{GeV}}^{-1}\,{\mathrm{cm}}^{-2}\,{{\rm{s}}}^{-1}$ at E₀ = 6.78 GeV. When divided into two sources, they are detected at test statistics of 186 and 474 for 3FGL J2021.0+4031eA and 3FGL J2021.0+4031eB, respectively. 3FGL J2021.0+4031eA is fit with the parameters γ = $2.00\pm {0.07}_{\mathrm{stat}}$, N₀ = $(5.77\pm {0.52}_{\mathrm{stat}})\,\times {10}^{-14}\,{\mathrm{GeV}}^{-1}\,{\mathrm{cm}}^{-2}\,{{\rm{s}}}^{-1}$ at E₀ = 6.78 GeV, whereas 3FGL J2021.0+4031eB is fit by γ = $2.02\pm {0.04}_{\mathrm{stat}}$, N₀ = $(1.84\pm {0.10}_{\mathrm{stat}})\times {10}^{-13}\,{\mathrm{GeV}}^{-1}\,{\mathrm{cm}}^{-2}\,{{\rm{s}}}^{-1}$ at E₀ = 6.78 GeV. Figure 15 shows the SEDs for 3FGL J2021.0+4031e and 3FGL J2021.0+4031eA. 3FGL J2021.0+4031eB is not shown for clarity. All have the same spectral indices within statistical errors, which suggests that they share a common origin of their emission. A joint fit was conducted using the VERITAS and the Fermi-LAT spectral points for 3FGL J2021.0+4031eA with the PL, BPL, and LP spectral models. The best fit is with a BPL with parameters N₀ = $(1.93\pm {0.50}_{\mathrm{stat}})\times {10}^{-14}\,{\mathrm{GeV}}^{-1}\,{\mathrm{cm}}^{-2}\,{{\rm{s}}}^{-1}$, E_b = 405 GeV, γ₁ = $1.97\pm {0.07}_{\mathrm{stat}}$ and γ₂ = $2.79\pm {0.22}_{\mathrm{stat}}$, and p-value = 0.55 versus 0.33 for the LP and 0.11 for the PL.

Zoom In Zoom Out Reset image size

With a morphology that matches G78.2+2.1, the HE emission from 3FGL J2021.0+4031e is associated with the SNR. The agreement in the fluxes from 3FGL J2021.0+4031eA and VER J2019+407 suggests that these two objects are associated and thus that the VERITAS emission is also associated with G78.2+2.1, with the current VERITAS observations only able to detect the brightest part of the remnant. The peak of this emission is located 05 away from the pulsar (PSR J2021+4026), with no evidence of VHE emission from the region between VER J2019+407 and the pulsar. Though the pulsar is often offset from the center of the VHE emission, it generally lies within the emission region and/or is associated with another region of VHE emission as discussed in Aliu et al. (2013) and references therein. The level of separation we have observed here, significantly more than the angular extent of the VHE emission, makes it unlikely that the emission is due to a PWN associated with PSR J2021+4026.

Comparing the joint spectral fit conducted in this work with that presented in Fraija & Araya (2016) shows a softer spectral index above and below the break (though within statistical errors), with a higher break energy. The softer spectral index in this work is in agreement with that predicted by standard test-particle shock acceleration, reducing the requirements for nonlinear effects as discussed in that work.

The source detected by HAWC is offset from the VERITAS emission and has a higher flux and a softer spectral index. The origin of these differences is uncertain; it may be due to the higher energy threshold of the HAWC observations, although there is no evidence of this in the VERITAS data. The other possibility is that the significant diffuse emission in the region is affecting the analyses. Further work is required to understand the emission in this region across all wave bands. Deeper VERITAS observations, in particular with the lower energy threshold following the 2012 camera upgrade, may enable the detection of emission from more of the remnant and the exploration of any energy-dependent morphology that could explain the differences between the three instruments.

MGRO J2019+37, one of the brightest sources observed in VHE γ-rays, is located within the vicinity of the OB star association Cygnus OB1. This source was observed by the Milagro Collaboration to have a flux level of 80% of the Crab Nebula where it was fit with an exponentially cutoff power law (Equation (3)) of parameters N₀ = $({7}_{-2}^{+5})\times {10}^{-17}{\mathrm{GeV}}^{-1}{\mathrm{cm}}^{-2}{{\rm{s}}}^{-1}$ at 10,000 GeV, γ = ${2.0}_{-1.0}^{+0.5}$, and E_c = ${29000}_{-16000}^{+50000}$ GeV (Abdo et al. 2012), with no identified multiwavelength counterpart(s). VERITAS has resolved MGRO J2019+37 into two components: an extended source VER J2019+368, which constitutes the bulk of the emission, and a point source VER J2016+371, which is coincident with the SNR CTB 87 (Aliu et al. 2014b) and discussed in Section 6.4.

VER J2019+368, with an extent of over 1° along the major axis, is spatially close to several sources that could potentially power (at least part of) the VHE emission. These include the H ii region Sh 2-104, Wolf-Rayet stars, the pulsars PSR J2021+3651 and PSR J2017+3625, (which are both detected in HE γ-rays by the Fermi-LAT), the PWN G75.1+0.2 (powered by PSR J2021+3651), and the hard X-ray transient IGR J20188+3657 (Aliu et al. 2014b).

PSR J2021+3651 (3FGL J2021.1+3651) lies to the east of the VER J2019+368 region and is associated with the X-ray PWN G75.2+0.1, which stretches back toward the VHE emission region. The possibility of these objects being linked to MGRO J2019+37 is discussed in detail in Hou et al. (2014).

Gotthelf et al. (2016) report on NuSTAR observations of the western edge of the VER J2019+368 region. They identified a number of hard X-ray sources in the region; three sources of particular interest were identified: 3XMM J201744.7+365045 (a bright point source that is most likely an active galactic nucleus behind the Galactic plane, though it could be a pulsar), NuSTAR J201744.3+364812 (a nebula, most likely a galaxy cluster hidden behind the Galactic plane), and emission associated with a young massive stellar cluster enclosed by a radio spur at the eastern edge of Sh 2-104. Further observations, for example, with Chandra, are required to determine whether these sources are related to the observed emission.

3FGL J2017.9+3627 lies to the southwest of VER J2019+368 and was recently identified as a γ-ray pulsar (PSR J2017+3625) by the Einstein@Home distributed computing project (Pletsch & Clark 2014; Clark et al. 2017). The pulsar has a spin frequency of 6.00 Hz with a first frequency derivative of −4.89 × 10⁻¹⁴ Hz s⁻¹, giving a characteristic age of 1943 kyr and a spin-down power of 1.2 × 10³⁴ erg s⁻¹ (Clark et al. 2017). 3FGL J2017.9+3627 is coincident with a region of low-energy VHE γ-ray emission (E < 1 TeV; Aliu et al. 2014b). Gotthelf et al. (2016) concluded that this pulsar could power the observed emission with an observed efficiency $({L}_{{TeV}}/\dot{E})\sim 0.03$, which is plausible for a >10⁵ yr old pulsar.

In this analysis, VER J2019+368 was observed at a peak (local) significance of 10.3σ using the Extended integration region. Significantly extended and asymmetrical, the source is best fit with an asymmetric Gaussian (Table 4) as in Aliu et al. (2014b), with the fit parameters in agreement with the results presented there.

In addition, a sky map produced using the Point integration region shows that the emission appears to be strongest in two regions, each of which is detected at a level greater than 7σ_local. Between the two sources lies a valley in the significance where the significance drops below 4σ_local, giving a change in the significance that is greater than 3σ_local. This suggests that the emission may be the result of two sources that were previously unresolved (though whether these are two independent sources, two enhancements of a single, extended source, or the result of statistical fluctuations is, as yet, uncertain).

To examine the extension of these two regions, a joint fit was conducted using two symmetric Gaussian sources (Table 4). This did not produce a significant improvement in the fit in comparison to a single, asymmetric Gaussian model (χ²/dof = 6511/5992 for the two sources versus 6600/5994 for the single, asymmetric Gaussian). We also produced a section through the excess map along the major axis of the fit to VER J2019+368 (Figure 16). The data were binned using rectangles of width 005 and height 02 along a length of 10. This was fit with two models: a single Gaussian function and the sum of two Gaussian functions. The single Gaussian function fits the data well, with a χ²/dof of 15.1/17, and the sum of two Gaussian functions mildly overfits the data, with a χ²/dof of 8.2/14. Neither of these tests provides statistically significant evidence that the emission is due to two independent sources; therefore, we name the potential sources VER J2018+367* and VER J2020+368*, where the * indicates that the two sources are still candidates and may be related. Further observations are required to understand this morphology.

Zoom In Zoom Out Reset image size

Aliu et al. (2014b) reported evidence of energy-dependent morphology with a low-significance (3σ) region of lower-energy (E < 1000 GeV) emission to the southwest of the main emission region around (${\alpha }_{J2000}$, ${\delta }_{J2000}$) = (${20}^{{\rm{h}}}{18}^{{\rm{m}}}$, $+36^\circ 26^{\prime}$). This is not visible in this analysis with either the Point or Extended integration regions (Figure 17). The reason for this is due to either the higher energy threshold used in this analysis or the fact that the original apparent energy dependence was a statistical fluctuation that has disappeared with additional data. Further observations with a lower energy threshold following the 2012 camera upgrade may clarify this. Interestingly, both VER J2018+367* and VER J2020+368* are also regions that show low-energy emission in Figure 17. However, at the moment, it is not possible to say whether this is the cause of these excesses (regions of lower-energy emission over a background of higher-energy emission) or whether this is just an effect of having brighter emission from these regions across both energy ranges.

Zoom In Zoom Out Reset image size

We reconstructed spectra for all three sources, with the fit parameters given in Table 5 , the spectral points given in Tables 10–12, and both shown in Figure 18. In the case of VER J2019+368 there is some evidence of curvature. When fit with an LP spectrum, the fit improved in comparison to the PL spectrum that we report in Table 5, but not significantly (F-test = 1.75σ). In comparison with the result presented in Aliu et al. (2014b), the spectrum has a significantly lower flux. This is due to the smaller integration region used in this analysis (023 radius rather than 05).

Zoom In Zoom Out Reset image size

2HWC J2019+367 lies between VER J2018+367* and VER J2020+368*, 007 away from VER J2019+368 (within the 009 statistical uncertainty of the HAWC location). As a point source (it was also detected with an extension of 07), it is fit with a softer spectral index ($-2.29\pm {0.06}_{\mathrm{stat}}$) and a higher normalization ($(3.02\pm {0.31}_{\mathrm{stat}})\times {10}^{-17}\,{\mathrm{GeV}}^{-1}\,{\mathrm{cm}}^{-2}\,{{\rm{s}}}^{-1}$ at 7000 GeV) than observed for VER J2019+368 in this work (γ = $1.98\pm {0.09}_{\mathrm{stat}}\pm {0.20}_{\mathrm{sys}}$, N₀ = $(1.03\pm {0.11}_{\mathrm{stat}}\pm {0.2}_{\mathrm{sys}})\times {10}^{-16}\,{\mathrm{GeV}}^{-1}\,{\mathrm{cm}}^{-2}\,{{\rm{s}}}^{-1}$ at 3110 GeV). It is noted that VER J2016+371 is not resolved in the 2HWC catalog, though there is evidence of excess emission in that direction in the maps presented in the associated paper. Examining a Fermi-LAT TS map, there is no evidence of emission coming from either of the two sources.

3FGL J2017.9+3627 has recently been identified as a γ-ray pulsar (PSR J2017+3625) by the Einstein@Home distributed computing project (Clark et al. 2017) and lies to the southwest of VER J2018+367* at a distance of 034. As pulsations have only recently been detected, no publicly available ephemeris is available. This prevented us from using either of the analysis methods presented in Section 3 for determining whether a PWN is present but hidden by the strength of the pulsed emission. Instead, we performed an analysis above 10 GeV to test for any emission potentially associated with a PWN; none was detected, and we place a 95% integral flux upper limit on the point-source emission of 6.2 × 10⁻¹¹ cm⁻² s⁻¹ between 10 and 500 GeV. We can compare the observed separation of PSR J2017+3625 divided by the extension of VER J2018+367* with the values reported in H.E.S.S. Collaboration et al. (2018b). The observed value of 1.9 is approximately a factor of 2 higher than all of the firm associations reported and above the rating criterion of 1.5 they applied to post-select associations. However, with a characteristic age of 1943 kyr, this is significantly older than all of the firm identification that they report and could explain the larger ratio. Assuming a distance of 450 pc (from Gotthelf et al. 2016 based on the gamma-ray efficiency of the pulsar), the TeV efficiency is 0.003, right at the lower edge of the scatter predicted by their varied model. In contrast, assuming a larger distance (Gotthelf et al. 2016 place a lower limit on the distance using an empirical relationship between pulsars and their spin-down power at 1 kpc) gives a TeV efficiency of greater than 0.01, well within the expected range. Given the large separation and lack of evidence of an extended PWN in multiwavelength observations, we are not able to associate the two objects. However, based on the observed properties of younger systems, it is possible that VER J2018+367* is a PWN associated with PSR J2017+3625.

In addition to PSR J2017+3625, other nonthermal sources are detected in the area that may power the observed VHE emission. Gotthelf et al. (2016) discuss the potential for a number of hard X-ray sources identified using NuSTAR data to power at least the western part of the emission from VER J2019+368, the emission that is now associated with VER J2018+367*. We note that all of the identified sources lie to the west of VER J2018+367* and thus are unlikely to be the sole contributor to the emission.

Observations with Suzaku presented in Mizuno et al. (2017) showed no evidence of X-ray emission in the 0.7–10 keV energy range in the region around VER J2018+367*. The closest source, a point source that they suspect is a background active galactic nucleus, lies in the gap between VER J2018+367* and VER J2020+368* and is unlikely to be the dominant contributor to the observed, extended VHE emission.

VER J2018+367* is spatially coincident with IGR J20188+3647, which remains a possible source for the emission from this region. Based on the fast rise time of the emission from IGR J20188+3647 (∼10 minutes), followed by the slower decay (∼50 minutes) and the spectral parameters, Sguera (2009) consider this to be a candidate supergiant fast X-ray transient. They concluded that though it is spatially coincident with MGRO J2019+37, due to the extended and diffuse nature of the VHE emission region, it is unlikely to be associated with it. However, the resolution of VER J2019+368 into two sources has significantly reduced this constraint.

3FGL J2021.1+3651 (PSR J2021+3651) is a γ-ray pulsar that lies 011 away from VER J2020+368*. The spectrum cuts off sharply at a few tens of GeV (E_c = 5.14 ± 0.36 GeV), and thus it is not expected to contribute to the observed VHE emission, but it could be masking emission associated with a PWN. Using the methods outlined in Section 3, an examination was conducted for a possible counterpart using an OFF-pulse phase of 0.7 to 1. The BRIDGE pulse phase from 0.2 to 0.4 was included in the ON-pulse phase. No evidence of a source was present in either of the analyses. The lack of detection of a Fermi-LAT PWN does not rule out the presence of such an object given the analysis challenges associated with such a bright pulsar, nor does it prevent the known PWN visible in other wavelengths from being responsible at least in part for the VHE emission.

XMM-Newton clearly shows a bright point source spatially coincident with the Fermi-LAT detected pulsar PSR J2021.1+3651, with associated extended emission likely from a PWN (G75.2+0.1) powered by the pulsar. The PWN is also visible in VLA 20 cm data. Aliu et al. (2014b) discuss the potential for G75.2+0.1 to contribute to the VHE emission in detail. The authors note that the bulk of VERITAS emission lies to the west of these objects, though the pulsar is moving in an eastward direction with G75.2+0.1 stretching back toward VER J2020+368*. Mizuno et al. (2017) reported on Suzaku observations that have allowed for a more detailed study of this westerly emission from G75.2+0.1 in the 0.7–2 and 2–10 keV bands. They show significant emission that extends to the west of PSR J2021+3651 back toward VER J2020+368*, though, as in the earlier observations, the X-ray emission they observe does not cover the region of highest VHE emission. Comparing the X-ray and VHE emission, they explore the relationship between the earlier results on VER J2019+368 presented by VERITAS in Aliu et al. (2014b) and the X-ray data they studied. The results presented here show a similar spectral index for VER J2020+368* as for the results reported for VER J2019+368 in Aliu et al. (2014b) ($-1.71\pm {0.26}_{\mathrm{stat}}\pm {0.2}_{\mathrm{sys}}$ versus $-1.75\pm {0.08}_{\mathrm{stat}}\,\pm {0.2}_{\mathrm{sys}}$), though the 1–10 TeV flux is approximately a factor of six lower (7.3 × 10⁻¹⁰ GeV cm⁻² s⁻¹ rather than 4.2 × 10⁻⁹ GeV cm⁻² s⁻¹). This is due to the smaller integration region size used (01 versus 05), but with an area that is 25 times smaller, the flux per unit solid angle is higher. The significant decrease in the TeV flux results in an increase in the ratio U_mag/U_ph from about 1, as presented in Mizuno et al. (2017), to about 6, with the smaller integration region better matching the size of the region from which the X-ray flux was extracted. This change results in a larger average magnetic B field of about 18 μG. This implies a lower energy population of electrons powering the observed X-ray emission (by a factor of about $\sqrt{6}$), bringing them down to about the same energy as those powering the VERITAS emission. Thus, the difference in the spectral indices between the X-ray and γ-ray observations cannot be explained by a break in the electron spectrum. The similarities between the populations also make it difficult to explain the origins of the observed morphology differences between the two wave bands. This higher magnetic field is closer to that detected in HESS J1825-137 (Uchiyama et al. 2009), though VER J2020+368* has a much smaller difference between the X-ray and γ-ray spectral indices. In summary, it is likely that VER J2020+368* is associated with G75.2+0.1 (the PWN of PSR J2021+3651), though more detailed observations are required to fully understand the system, in particular to understand the relationship between the observed properties of the X-ray and γ-ray emission and the high inferred magnetic field.

In addition to the extended emission from VER J2019+368, Aliu et al. (2014b) resolved from MGRO J2019+37 a point source of emission coincident with the SNR CTB 87 (VER J2016+371). There is a bright source detected by Fermi-LAT that lies close to the position of VER J2016+371 in the 1FGL, 2FGL, and 3FGL catalogs (1FGL J2015.7+3708, 2FGL J2015.6+3709, and 3FGL J2015.6+3709, respectively) and also in the higher-energy 1FHL, 2FHL, and 3FHL catalogs (1FHL J2015.8+3710, 2FHL J2016.2+3713, 3FHL J2015.9+3712) and in the 1st Fermi-LAT Supernova Remnant Catalog (1SC; Acero et al. 2016). The location of this source is slightly different in each of the three energy ranges. In the 1FGL/2FGL/3FGL catalogs the source is located close to QSO J2015+371, a BL Lacertae object at z = 0.859 (Shaw et al. 2013), whereas in the 1FHL/2FHL/3FHL catalogs it lies closer to CTB 87. At the lower energies of the 1/2/3 FGL catalog analysis, there is evidence of flux variability, which has, along with the location, led to the source being associated with the blazar QSO J2015+371. The higher energies see reduced evidence of variability and also see a hardening of the spectral index from −2.53 ± 0.04 in the 3FGL to −1.74 ± 0.46 in the 2FHL. Combined with the change in the source location, this has led to it being associated with CTB 87 in those catalogs. This suggests that there could be two HE γ-ray sources that lie close together and are not being resolved in the Fermi-LAT data.

In this analysis, VER J2016+371 was observed using the ring background technique and the Point integration region at a level of 6.3σ_local. Previously reported as a point source (Aliu et al. 2014b), we found no improvement when fitting with a single symmetric Gaussian. Thus, we still consider it to be a point source. To identify its best-fit location, we modeled it as a point source, and the best-fit location is given in Table 4.

A spectral fit was conducted using all of the data taken with four telescopes operational and with a pointing offset of less than two degrees (a looser cut on pointing than used for the other sources, since the majority of the observations that had the source in the FOV were targeted at VER J2019+368 and thus have a larger pointing offset). A SED was produced using three logarithmically spaced bins per decade in energy and the Point integration region (Figure 19, Tables 5 and 13). The updated spectrum is consistent (within statistical errors) with that reported in Aliu et al. (2014b).

Zoom In Zoom Out Reset image size

In the Fermi-LAT analysis, when the source is fit at the location of 3FGL J2015.6+3709 ((l, b) = (7487, 1186)), an additional source is detected at (l, b) = (7499, 116) (see Table 6 for details). The location of the Fermi-LAT source is known to vary depending on the energy threshold, so this additional source is likely due to a change in the source location associated with the higher energy threshold in this analysis in comparison to the 3FGL. The fermipy method localize was run on 3FGL J2015.6+3709 without the additional point source included in the model. This resulted in an updated source position of (l, b) = ($74.893^\circ \pm {0.006}_{\mathrm{stat}}^{^\circ }$, $1.200^\circ \pm {0.004}_{\mathrm{stat}}^{^\circ }$). This position lies 003 from the catalog location of 3FGL J2015.6+3709 and is shown relative to the Fermi-LAT catalog positions and overlaid on an XMM-Newton counts map with VERITAS excess and CGPS 1420 MHz contours in Figure 20. With the updated source location, no second source is detected. For the updated location, the fermipy method extension was run to determine whether there was any evidence of spatial extension; none was detected, and a 95% confidence level upper limit of extension was found of 007 with a 2D symmetrical Gaussian as the source template. Given the location of the source, it is likely dominated by emission from QSO J2015+371, and the SED (Figure 19) is consistent with that from 3FGL J2015.6+3709 as presented in the 3FGL (shown in Figure 7).

Zoom In Zoom Out Reset image size

Table 6.  Results of an Examination of the HE Emission in the Region of VER J2016+371

Model		TS	${\alpha }_{J2000}$	${\delta }_{J2000}$	l	b	N0	E0/ Eb	γ	β
			(deg)	(deg)	(deg)	(deg)	(GeV−1cm−2 s−1)	(GeV)
Single source	3FGL location	1888	303.91	37.16	74.86	1.19	(4.43 ± 0.18)E–12	1520	2.62 ± 0.09	0.12 ± 0.06
Single source	Best-fit location	1964	303.91	37.19	74.893 ± 0.006	1.200 ± 0.004	(4.86 ± 0.17)E–12	1520	2.63 ± 0.07	0.02 ± 0.04
Two sources	3FGL J2015.6+3709	1363	303.91	37.16	74.86	1.19	(4.43 ± 0.18)E–12	1520	2.62 ± 0.09	0.12 ± 0.06
	New source	59.8	304.02	37.25	74.99	1.16	(1.19 ± 0.34)E–12	1000	2.11 ± 0.12	N/A
Two sources	CTB 87	102	304.01	37.21	74.95	1.15	(3.53 ± 0.86)E–12	1000	2.44 ± 0.10	N/A
	QSO J2015+371	1087	303.87	37.18	74.87	1.22	(1.13 ± 0.09)E–11	1000	2.77 ± 0.06	N/A

Note. All errors are statistical only.

Download table as:  ASCIITypeset image

Table 7.  Catalog Values for the 3FGL Sources in the Survey Region ((835 > l > 655) and (55 > b > −25)) that Were Not Significantly Detected in This Analysis (TS < 25) and Removed

Name	Spectrum Type	3FGL Sig. (σ)	N0 (GeV−1cm−2 s−1)	E0/ Eb (GeV)	γ	β
3FGL J1958.6+3844	Power law	4.84	($1.63\pm {0.31}_{\mathrm{stat}}$)E–9	0.8	$2.64\pm {0.13}_{\mathrm{stat}}$	N/A
3FGL J2011.1+4203	Power law	4.31	($1.23\pm {0.25}_{\mathrm{stat}}$)E–9	0.95	$2.54\pm {0.12}_{\mathrm{stat}}$	N/A
3FGL J2014.4+3606	Power law	4.48	($4.49\pm {0.94}_{\mathrm{stat}}$)E–10	1.69	$2.40\pm {0.11}_{\mathrm{stat}}$	N/A
3FGL J2024.6+3747	Log parabola	7.04	($6.30\pm {0.87}_{\mathrm{stat}}$)E–9	0.87	$2.39\pm {0.58}_{\mathrm{stat}}$	$0.49\pm {0.22}_{\mathrm{stat}}$
3FGL J2026.8+4003	Log parabola	9.03	($2.00\pm {0.21}_{\mathrm{stat}}$)E–8	0.7	$2.61\pm {0.19}_{\mathrm{stat}}$	1
3FGL J2033.3+4348*	Power law	4.39	($6.57\pm {1.52}_{\mathrm{stat}}$)E–11	3.84	$2.24\pm {0.14}_{\mathrm{stat}}$	N/A
3FGL J2036.8+4234*	Power law	5.38	($1.52\pm {0.29}_{\mathrm{stat}}$)E–10	3.21	$2.25\pm {0.10}_{\mathrm{stat}}$	N/A
3FGL J2037.4+4132*	Power law	6.45	($9.67\pm {1.66}_{\mathrm{stat}}$)E–11	4.17	$2.18\pm {0.11}_{\mathrm{stat}}$	N/A
3FGL J2043.1+4350	Log parabola	8.41	($3.32\pm {0.38}_{\mathrm{stat}}$)E–8	0.47	$2.46\pm {0.19}_{\mathrm{stat}}$	1

Note. The asterisk denotes that the source lies in the field of the Cygnus Cocoon.

Download table as:  ASCIITypeset image

To test the relative contributions of the two associations, instead of using a single source to model the emission, two power-law sources were used and placed at the radio locations of CTB 87 and QSO J2015+371, respectively (the positions are given in Table 6). When this model was fit to the data, the two sources had test statistics of 102 and 1087 for CTB 87 and QSO J2015+371, respectively. It is not possible to definitively state that the Fermi-LAT data support the presence of two sources that are currently unresolved, but there is evidence to suggest that this is the case.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

The spectra of the two sources are noticeably different, with the source located at the site of CTB 87 being weaker and harder. Plotting these spectra with the spectrum from the updated 3FGL location and with the VERITAS spectrum of VER J2016+371 (Figure 19) shows good agreement between the spectrum of the Fermi-LAT source located at CTB 87 and the spectrum of VER J2016+371, whereas the source located at the location of QSO J2015+371 would require a spectral hardening to fit the VERITAS results. Conducting a joint fit to both the spectral points from the Fermi-LAT emission at the CTB 87 position and VERITAS emission from VER J2016+371 with a PL gives the parameters N₀ = $(6.67\,\pm {0.60}_{\mathrm{stat}})\times {10}^{-11}\,{\mathrm{GeV}}^{-1}\,{\mathrm{cm}}^{-2}\,{{\rm{s}}}^{-1}$ at E₀ = 5.2 GeV and spectral index $-2.39\pm {0.05}_{\mathrm{stat}}$ with a χ² of 4.4 and 9 dof.

Combined, the location of VER J2016+371 and the spectrum of the Fermi-LAT emission, when fit as two sources, suggest that the Fermi emission from the direction of CTB 87 and VER J2016+371 are the same source and that they are associated with the SNR CTB 87. However, a contribution from QSO J2015+371 to the VHE emission cannot be ruled out.

Examining the region in other wavelengths highlights the two sources identified above, along with a number of other, weaker, nonthermal sources. Figure 20 shows an XMM-Newton counts map of the region (Obs. ID 0744640101) overlaid with contours from the CGPS 1420 MHz survey. In both of these observations, CTB 87 and the blazar QSO J2015+371 are clearly visible, with the VERITAS emission centered on a location that lies within the radio shell of CTB 87, supporting the theory that the VHE γ-ray emission is from that source. In contrast, the lower-energy Fermi-LAT emission is located between the two sources and outside of any of the X-ray emission regions. Chandra observations of the region clearly show a bright point source (CXOU J201609.2+371110) that lies within the X-ray emission from CTB 87 at (l, b) = (74.9438, +01.1140); though pulsations have yet to be detected, it has been suggested that this is the location of the pulsar powering the PWN. The implications of this, as well as the location and morphology of the X-ray emission, are discussed in detail in Matheson et al. (2013), concluding that CTB 87 is an evolved (∼5–28 kyr), crushed PWN with the extended radio emission and likely a "relic" PWN, as in Vela-X and G327.1–1.1, both of which are detected in VHE γ-rays. Saha (2016) conducted a multiwavelength study of the object, where they attributed excess emission in the MeV–GeV range to an additional Maxwellian component. In contrast, the single power-law fit presented here does not require this component. However, the emission above 1 TeV still requires the BPL electron distribution as presented in Figure 4 of that work. This BPL electron population can be understood in the context of the multiwavelength data with the electron distribution before the break responsible for the radio emission and the population after the break responsible for the X-ray emission. Due to the relatively poor angular resolution of VERITAS in comparison to the radio and X-ray observations, it is not possible to observe any differences in the VHE spectrum from these two locations.