Recent Results of the Milagro Experiment
        To date, our most important published observational results with Milagro have been:

In addition, we have published results on the detection TeV gamma rays from the active galaxies Mrk 501 [5], Mrk 421 [6] plus the Crab Nebula [7], set stringent upper limits on the prompt TeV emission from several gamma ray bursts [8], and performed the most sensitive survey of the northern hemisphere at TeV energies [6]. 
       In this document we will review these results, plus show new results on two more newly discovered unidentified Galactic sources plus Geminaga that have been presented at meetings and are being submitted for publication. In addition, we show three more potential sources that are detected with >4.5s, slightly less than the 5s post trials threshold we set for discovery.
Fig 1        It should be noted that the most exciting discoveries from Milagro have been published since our 2005 review. Figure 1 shows the significant improvement that we have made in the last few years with the addition of the outrigger counters and improved gamma-hadron separation and data analysis.

Survey of the Galactic Plane at 12 TeV with Milagro and the Discovery of MGRO J1909+06 and MGRO J2033+42

This analysis of approximately 6 years of Milagro [9] data shows that an excess of TeV gamma rays also exists along the plane. Most of the activity occurs in the Cygnus region of the Galaxy, as seen in Figures 1 &

2. (The Cygnus region is also where EGRET saw the strongest gamma ray emission outside the galactic center which is not visible from the northern hemisphere).  This is also seen in a previous analysis of the plane using 3 years of data in [10], where the median energy is 3.5 TeV. In the past year, the sensitivity of the Milagro experiment has been substantially increased through improvements in gamma/hadron separation (that utilizes information from the outrigger array) and data analysis. We now use a gamma/hadron separation parameter, A4, which is capable of rejecting background with high efficiency at high energies. We improved sensitivity with the introduction of an analysis technique that weights events based on the likelihood that they are due to gamma rays. This method is equivalent to the Likelihood ratio method in the large N limit. These two improvements result in an increase in sensitivity of ~2.5x, as confirmed by observations of the Crab nebula, and an increase in median energy from to ~12 TeV.  Note that this energy at which Milagro attains its peak sensitivity is well above that of ACTs. 

Fig 2
Fig 3Figure 2.  Significance map of the Galactic plane visible to Milagro.  There are 7 interesting spots of TeV gamma-ray emission with > 4.5 standard deviations within |b| < 5°.  Four of these spots are in the Cygnus Region of the Galaxy from l Î [65°, 85°].
We compute the fluxes reported here by assuming a Crab-like source spectrum with a spectral index of -2.62 taking into account the changing exposure of Milagro with respect to the declination.  This information is applied to the data to give absolute fluxes for gamma-ray sources and diffuse fluxes from different regions of the Galaxy.  These fluxes are scaled to the differential HEGRA measured Crab flux extrapolated to 12 TeV [11].
Table 1 gives the flux from different longitude ranges in the plane. 
The Cygnus region is the most luminous source of TeV gamma rays visible to Milagro with a significance of 10.5 standard deviations above the isotropic background. The flux measurement is 8.7 standard deviations above the GALPROP prediction for diffuse emission, where GALPROP is fit to agree with the EGRET diffuse GeV measurements for the plane [12]. This large discrepancy indicates that the VHE emission from the plane is not dominated by diffuse emission from cosmic ray interactions with matter, but is due to an additional hard component likely associated with localized sources.

Fig 4

TABLE 1.  Gamma-Ray Emission from the Galactic Plane


Position Box (deg)
|b| < 2°


Flux1 (x10-12)
(TeV-1cm-2s-1sr 1)

GALPROP (x10-12)


Galactic Plane

30 < l < 110


3.0 ± 0.3


30 < l < 65


4.1 ± 0.6


65 < l < 85


3.8 ± 0.3


85 < l < 110


< 1.3 (95% CL)


136 < l < 216


< 1.4 (95% CL)


1 errors are statistical; There is a 20% error due to the systematic uncertainty in our overall energy scale.

Gamma-Ray Source Discovery

A search for localized sources of TeV emission from the plane was conducted by identifying spots with more than 4.5 standard deviations (pre-trials) above the background. Seven candidate sources are seen, excluding the Crab Nebula, for Galactic latitudes |b| < 5°.  One candidate is seen at Galactic longitude l ~ 106°, 4 sources are seen in the Cygnus Region, and two more are observed near l ~ 40°.  The source MGRO J2019+37 is the brightest spot in the Cygnus Region with a flux of  ~ 0.5 Crab.  The TeV gamma-ray fluxes from MGRO J1909+06 and MGRO J1854+01 are consistent with that seen from the Crab Nebula for a -2.62 spectrum, as shown in Table 2.

TABLE 2.  Galactic Sources and Source Candidates


Location1 (deg)
(l, b)

(std. deviations)

Flux3 (x10-14)

Crab Nebula

-175.4 ± 0.1, -5.7 ± 0.1


5.1 ± 0.4

MGRO J2019+37

75.1 ± 0.1, 0.3 ± 0.1


2.8 ± 0.4

MGRO J1909+06

40.5 ± 0.1, -1.0 ± 0.1


5.2 ± 0.8

MGRO J2033+42

80.4 ± 0.1, 1.0 ± 0.3


2.0 ± 0.4


195.7 ± 0.3, 4.0 ± 0.3


3.2 ± 0.5

MGRO J2032+37

76.3 ± 0.1, -1.9 ± 0.2


1.2 ± 0.2

MGRO J2043+36

77.2 ± 0.2, -4.0 ± 0.2


1.0 ± 0.2

MGRO J2233+60

105.8 ± 0.5, 2.0 ± 0.8


1.1 ± 0.4

Items in bold have >5s post trials significance.
1 add 0.3° error for pointing systematics and unknown source region morphologies
2 pre-trials significance
3  quoted at 12 TeV – errors are statistical - add 20% systematic uncertainty for our energy scale

The number of trials is based on the large regions of the Galactic plane searched and results in ~ 100,000 trials factor that should be applied to the pre-trials significances of Table 2.  The post-trials significance is greater than 5 standard deviations for MGRO J2019+37 MGRO J1909+06, MGRO J2033+42. Each of these detections is the discovery of a new VHE gamma-ray source. Geminga which is the brighest GeV source in the northern sky is seen at 5 s at a known location.
MGRO J2019+37 is coincident in location with EGRET unidentified sources 3EG J2016+3657 and 3EG J2021+3716. The former is possibly associated with AGN of unknown redshift and the latter with latter with pulsar wind nebula G75.2+0.1. An open star cluster, Berkeley 87, is also nearby and could be a site of particle acceleration.
Fig 5MGRO J2033+42 is in agreement with the location of EGRET source 3EG J2033+4118 as well as HEGRA TeV 2032+4130 (recently also reported by VERITAS).  The HEGRA source has a differential photon flux at 12 TeV of
(0.5±0.2)x10-14TeV-1cm-2s-1. This is 12% of the HEGRA measured Crab flux extrapolated to 12 TeV. This discrepancy between Milagro and HEGRA may be understood as due to both uncertainty in the spectrum or a possible diffuse component underlying the HEGRA identified point source that makes a greater contribution to the Milagro detection due to Milagro’s angular resolution of ~0.5 degrees as compared to HEGRA’s of ~0.1 degrees.
Text Box: Figure 5. The Crab plus the four newly discovered Milagro sources.
Fig 6Energy Spectra with Milagro
Since the 2005 review, we have developed and refined our event energy reconstruction.  Figure 4 shows preliminary result from this effort. Note that the spectrum measured by Milagro is consistent with the ACT measurements at low energies and at the highest energies (~50TeV) the error in the Milagro measurement is the smallest.  This is achieved because of Milagro’s enormous exposure, which can only be achieved by a wide-field continuously operational instrument.
We anticipate that further forthcoming improvement in the high energy reconstruction will further improve the sensitivity for Milagro at energies >50 TeV, an energy regime critical to the understanding of Galactic sources and the origin of Galactic cosmic rays, but difficult to access from ACTs.

The success of Milagro clearly demonstrates the ability of a water Cherenkov detector to perform all-sky surveys in the TeV energy band.  Given this success and the widely recognized importance of surveys at all other wavelengths, it is natural to investigate further improvement in the sensitivity of such an instrument.  The HAWC (High Altitude Water Cherenkov) detector is a concept to re-deploy the Milagro PMTs and electronics at a high altitude site in a somewhat different configuration.  Based upon our Milagro experience we believe that this detector will be over an order of magnitude more sensitive than Milagro and have a substantially lower energy response (700 GeV median energy).  Thus HAWC presents the possibility of opening the extra-galactic sky to continuous TeV observations.  This will lead to the discovery and understanding of transient phenomena from active galaxies and possibly from gamma-ray bursts.  In addition to being an excellent transient source detector HAWC has high sensitivity to extended sources.  As is clear from the HESS data as well as our own Milagro data, many of the galactic sources are diffuse. HAWC will have exceptional capability to perform an all-sky survey for diffuse sources.

The synergy of an all-sky gamma-ray instrument in the northern sky to IceCube is obvious. IceCube is already the world’s largest neutrino telescope. However, even when it is completed and under optimistic assumptions about neutrino emission, the potential neutrino signal in IceCube from AGN and GRBs may be small. The atmospheric neutrino background after cuts will result in 1-2 neutrinos/yr/km3/deg2 above 1 TeV.  For a three year observation of an AGN in the northern sky one could expect ~5 background events. Lacking other information about AGN emission in the TeV nearly 20 neutrino signal events would have to be observed in order to claim an unambiguous detection.  HAWC would be an ideal complement for IceCube, but Milagro has a role to play.
While HAWC will monitor the entire northern sky everyday for transients with fluxes less than 1 Crab in the TeV energy range and will detect stronger outbursts, such as have been observed for Mrk 501 and Mrk 421 within 10 minutes, Milagro is currently the only instrument available to monitor the same sky that IceCube views continuously.



1. Atkins, R. et al. 2005 Phys Rev Lett 95, 251103.
2. Abdo A. et al. 2007, to appear AP J. 658: March 20, 2007.
3. Smith, A. et al. 2005, Review, Rapporteur, and Highlight Papers of 29th ICRC in Pune India, 10, 227.
4. Atkins, R. et al. 2000, ApJ Lett, 533, 119 and Atkins, R. et al. 2003, ApJ, 583, 824.
5. Atkins, R. et al. 1999, ApJ Lett, 525, L25.
6. Atkins, R. et al. 2004, ApJ, 608, 680.
7. Atkins, R. et al. 2003, ApJ, 595, 803.
8. Atkins, R. et al. 2005, ApJ, 630, 996 and Atkins, R. et al., 2004 ApJ Lett, 604, 25.
9. Atkins, R,. et al., Nucl. Instr. Meth., A449, 478 (2000).
10.Atkins, R., et al., Phys. Rev. Lett. 95, 251103 (2005).
11.Aharonian, F., et al., Astrophys. J., 614, 897 (2004).
12.Strong, A.W., Moskalenko, I.V., and Reimer, O., Astrophys. J., 613, 962 (2004).
13. Atkins, et al., Astrophys. J., 630 (2): 996-1002 (2005).
14. Abdo A. et al. 2007, "Milagro Constraints on Very High Energy Emission from Short Duration Gamma-Ray Bursts" submitted to Ap. J.