Observation of VHE Gamma Radiation from HESS J1834-087/W41 with the MAGIC Telescope

Recently, the HESS array has reported the detection of gamma-ray emission above a few hundred GeV from eight new sources located close to the Galactic Plane. The source HESS J1834-087 is spatially coincident with SNR G23.3-0.3 (W41). Here we present MAGIC observations of this source, resulting in the detection of a differential gamma-ray flux consistent with a power law, described as dN/(dA dt dE) = (3.7 +/- 0.6)*10^(-12) (E/TeV)^(-2.5 +/- 0.2) \ cm^(-2)s^(-1)TeV^(-1). We confirm the extended character of this flux. We briefly discuss the observational technique used, the procedure implemented for the data analysis, and put this detection in the perspective of the molecular environment found in the region of W41. We present 13CO and 12CO emission maps showing the existence of a massive molecular cloud in spatial superposition with the MAGIC detection.


ABSTRACT
Recently, the HESS array has reported the detection of γ-ray emission above a few hundred GeV from eight new sources located close to the Galactic Plane. The source HESS J1834-087 is spatially coincident with SNR G23.3-0.3 (W41). Here we present MAGIC observations of this source, resulting in the detection of a differential γ-ray flux consistent with a power law, described as dN γ /(dAdtdE) = (3.7 ± 0.6) × 10 −12 (E/TeV) −2.5±0.2 cm −2 s −1 TeV −1 . We confirm the extended character of this flux. We briefly discuss the observational technique used, the procedure implemented for the data analysis, and put this detection in the perspective of the molecular environment found in the region of W41. We present 13 CO and 12 CO emission maps showing the existence of a massive molecular cloud in spatial superposition with the MAGIC detection.

Introduction
In the Galactic Plane scan performed by the HESS Cherenkov array in 2004, with a flux sensitivity of 3% Crab units for γ-rays above 200 GeV, eight sources were discovered (Aharonian et al. 2005(Aharonian et al. , 2006. One of the newly detected γ-ray sources is HESS J1834-087 which is found to be, in projection, spatially coincident with SNR G23.3-0.3 (W41). The possibility of a random correlation between the VHE source and SNR G23.3-0.3 was estimated to be 6% for the central region of the Galaxy (Aharonian et al. 2005). The high energy source could also be connected to the old pulsar PSRJ1833-0827 (Gaensler & Johnston 1995), which would be energetic enough as to power HESS J1834-087. However, its location at 24 arc minutes away from the center of HESS J1834-087, renders an association unlikely (Aharonian et al. 2005(Aharonian et al. , 2006. In addition, there is also no extended PWN detected so far, whereas HESS J1834-087 has been found to be extended: A brightness distribution ρ ∼ exp(−r 2 /2σ 2 ) with a size σ = (0.09 ± 0.02) • has been reported by HESS (Aharonian et al. 2006).
Here, we present observations of HESS J1834-087 with the Major Atmospheric Gamma Imaging Cherenkov telescope (MAGIC). We briefly discuss the observational technique used and the procedure implemented for the data analysis, derive a very high energy γ-ray spectrum of the source, and analyze it in comparison with other observations, including the molecular environment found in the region of W41.

Observations
MAGIC (see e.g., Baixeras et al. (2004); Cortina et al. (2005) for a detailed description) is the largest single dish Imaging Air Cherenkov Telescope (IACT) in operation. Located on the Canary Island La Palma (28.8 • N, 17.8 • W, 2200 m a.s.l.), the telescope has a 17-m diameter tessellated parabolic mirror, supported by a light weight space frame of carbon fiber reinforced plastic tubes. It is equipped with a 576-pixel 3.5 • field-of-view enhanced quantum efficiency photomultiplier (PMT) camera. The analogue PMT signals are transported via optical fibers to the trigger electronics and are read out by a 300 MSamples/s FADC system.
At La Palma, HESS J1834-087 culminates at about 37 • zenith angle (ZA). This ZA increases the energy threshold for MAGIC observations, but also, it provides a larger effective collection area. The sky region around the location of HESS J1834-087 has a relatively high and non-uniform level of background light. Within a distance of 1 • from HESS J1834-087, there are 3 stars brighter than 8 th magnitude, with the star field being brighter in the region NW of the source. MAGIC observations were carried out in the false-source tracking (wobble) mode (Fomin et al. 1994). The sky directions (W1, W2) to be tracked are such that in the camera the sky brightness distribution relative to W1 is similar to the one relative to W2. The source direction is in both cases 0.4 • offset from the camera center. These two tracking positions are shown by white stars in figure 1. For each tracking position two background control regions are used, which are located symmetrically to the source region (denoted by the central white circle) with respect to the camera center. During wobble mode data taking, 50% of the data is taken at W1 and 50% at W2, switching (wobbling) between the 2 directions every 30 minutes. This observation mode allowed a reliable background estimation least affected by the medium-scale ZA and the inhomogeneous star field. HESS J1834-087 was observed for a total of 20 hours in the period August-September 2005 (ZA ≤ 45 • ). In total, about 12 million triggers have been recorded.

Data Analysis
The data analysis was carried out using the standard MAGIC analysis and reconstruction software (Bretz & Wagner 2003), the first step of which involves the calibration of the raw data (Gaug et al. 2005). It follows the general steps presented in (Albert et al. 2006a,b): After calibration, image cleaning tail cuts of 10 photoelectrons (ph. el.) for image core pixels and 5 ph. el. (boundary pixels) have been applied (see e.g. (Fegan 1997)). These tail cuts are accordingly scaled for the larger size of the outer pixels of the MAGIC camera. The camera images are parameterized by image parameters (Hillas 1985). In this analysis, the Random Forest method (see Bock et al. (2004);Breiman (2001) for a detailed description) was applied for the γ/hadron separation (for a review see e.g. Fegan (1997)) and the energy estimation. For the training of the Random Forest a sample of Monte Carlo (MC) generated γ-showers (Majumdar et al. 2005) was used together with about 1% randomly selected events from the measured wobble data. The MC γ-showers were generated between 35 • and 45 • ZA with energies between 10 GeV and 30 TeV with a SIZE distribution equal to the one of the selected data events for the training. The source-position independent image parameters SIZE, WIDTH, LENGTH, CONC (Hillas 1985) and the third moment of the ph. el. distribution along the major image axis were selected to parameterize the shower images. After the training, the Random Forest method allows to calculate for every event a parameter, so-called hadronness, which is a measure of the probability that the event is not γ-like. The γ-sample is defined by selecting showers with a hadronness below a specified value. An independent sample of MC γ-showers was used to determine the cut efficiency.
The analysis at similar ZA angles was developed and verified using Crab nebula data taken in September 2005, see also (Albert et al. 2006c). The Crab energy spectrum, as determined by our studies, was consistent with measurements from other experiments (see Fig. 4, dot-dashed line).
For each event the arrival direction of the primary γ-ray candidate in sky coordinates is estimated using the DISP-method (Fomin et al. 1994;Lessard et al. 2001;Domingo-Santamaria et al. 2005). A conservative lower SIZE cut of 200 ph. el. is applied to select a subset of events with superior angular resolution. The corresponding analysis energy threshold is about 250 GeV. Figure 1 shows the sky map of γ-ray candidates (background subtracted, see e.g. (Rowell 2003)) from the direction of HESS J1834-087. It is folded with a two-dimensional Gaussian with a standard deviation of 0.072 • and a maximum of one. The MAGIC γ-ray PSF (standard deviation of a two dimensional Gaussian fit to the non-folded brightness profile of a point source) is 0.1 ± 0.01 • . The folding of the sky map serves to increase the signal-to-noise ratio by smoothing out statistical fluctuations. However, it somewhat degrades the spatial resolution. The sky map is overlayed with contours of 90 cm VLA radio data (green) from White et al. (2005) (20 cm radio data from the same reference are overlayed in the following figures) and 12 CO emission contours from Dame et al. (2001) (black), integrated in the velocity range 70 to 85 km/s, the range that best defines the molecular cloud associated with W41. The MAGIC excess is centered at (RA, DEC)=(18 h 34 m 27 s , -8 • 42'40"). The statistical error is 0.5', the systematic pointing uncertainty is estimated to be 2' (see Bretz et al. (2003)). A fit of a two dimensional Gaussian brightness profile to the non-folded sky map yields after subtraction the MAGIC γ-ray PSF inquadrature an intrinsic source extension of σ = (0.14 ± 0.04) • (the extension reported by HESS is 0.09 ± 0.02) • (Aharonian et al. 2006)). Both, position and extension, coincide well with the shell-type SNR G23.3-0.3 (W41). Figure 2 shows the distribution of the squared angular distance, θ 2 , between the reconstructed shower direction and the excess center. The observed excess in the direction of HESS J1834-087 has a significance of 8.6σ for θ 2 ≤ 0.1deg 2 . Figure 3 shows images of HESS J1834-087 with three different lower cuts on SIZE (200, 300, 600 ph.el), corresponding to energy thresholds of about 250, 360 and 590 GeV. As in figure 1 the background subtracted sky maps are folded with a two-dimensional Gaussian, but here the color scale shows directly the excess significance. The total observed excess significance for θ 2 ≤ 0.1deg 2 (corresponding to the sky region inside the central white circle of figure 1) are 8.6σ, 7.8σ and 7.3σ for the three lower cuts on SIZE. Overlayed are contours of 20 cm VLA radio data from White et al. (2005) (green) and 13 CO emission contours (black) from Jackson et al. (2006). The contours of the radio emission are at 0.0035 Jy/beam. The 13 CO contours are integrated from 70 to 85 km/s in velocity, as was the 12 CO data in figure 1. For all three SIZE cuts the MAGIC PSF is about 0.1 • , and the source position, extension and morphology stay roughly constant. The characteristics of the MAGIC observation are compatible within errors with the measurement of HESS (Aharonian et al. 2006).
For the spectral analysis a sky region of maximum angular distance of θ 2 = 0.1deg 2 around the excess center (indicated by the white circle in Figure 1) has been integrated. Figure 4 shows the reconstructed very high energy γ-ray spectrum (dN γ /(dE γ dAdt) vs. true E γ ) of HESS J1834-087 after correcting (unfolding) for the instrumental energy resolution (Anykeev et al. 1991). The horizontal bars indicate the bin size in energy, the marker is placed in the bin center on a logarithmic scale. The full line shows the result of a forward unfolding procedure: A simple power law spectrum is fitted to the measured spectrum (dN γ /(dE γ dAdt) vs. estimated E γ ) taking the full instrumental energy migration (true E γ vs. estimated E γ ) into account as described in Mizobuchi et al. (2005). The result is given by (χ 2 /n.d.f = 7.4/7): The quoted errors are statistical. The systematic error is estimated to be 35% in the flux level determination and 0.2 in the spectral index, see also (Albert et al. 2006b). Within the observation time (weeks) no flux variations exceeding the measurement errors have been observed. Also, the flux is compatible within errors with the measurement of HESS made one year earlier.

Discussion and concluding remarks
SNRs as gamma-ray sources have been extensively discussed in the past (e.g., see Torres et al. 2003 for a review). Due to the spatial coincidence between the VHE γ-ray source and the SNR G23.3-0.3 (W41), this SNR appears to be the natural candidate for generating the observed γ-ray emission. W41 is an asymmetric shell-type SNR, with a diameter of 27 arc minutes. It is included in Green's catalog (Green 2004), and has a spectral index of 0.5, and a flux density of 70 Jy at 1 GHz. It was mapped in radio with the VLA array at 330 MHz (Kassim 1992) and at 20 and 90 cm (see White, Becker & Helfand 2005), following earlier studies (see, e.g., Ariskin &Berulis 1970, Shaver &Goss 1970, andreferences therein). It is partially overlapping with SNR G22.7-0.2 (see, e.g., Fig. 9 of Kassim 1992), although the latter is not in coincidence with the peak of the very high energy source (see Figure 1 above). No Chandra and XMM observations of W41 are publicly available yet.
W41 was associated with a very large molecular complex called "[23,78]" in Dame et al. (1986). There, it was concluded that there are probably two large clouds blended at that position in l − b − v space, one in the near side of the 4 kpc arm and another in the far side of the Scutum Arm. The giant molecular cloud associated with W41 is best defined by integrating the CO emission from 70 to 85 km/s in velocity. The CO emission peaks near l=23.3 • , b=−0.3 • , v=78 km/s; the near kinematic distance of this peak is 4.9 kpc. The peak is marked by the central black contour in Figure 1, which lies very close to the VHE source. The total H 2 mass of the cloud, computed over the range l=22 • to 24.25 • , b=−0.75 • to 0.5 • , and v=70 to 85 km/s, and assuming a distance of 4.9 kpc, is 2.1 × 10 6 M ⊙ . This mass is necessarily an upper limit since, as mentioned above, there is certainly an emission contribution from unrelated gas at the far kinematic distance. Still, the CO peak is so strong and well defined that it most likely arises from gas primarily at one location, near the VHE source, rather than being a random blend of emissions from the near and far distances. The total H 2 mass of the CO emission peak in Figure 1 (computed over the region l=23.2 • to 23.4 • , b=−0.35 • to −0.15 • , v=70 to 85 km/s) is 8.8 × 10 4 M ⊙ . The higher-resolution 13 CO map in figure 3, which was derived from the recently completed Galactic Ring survey (Jackson et al. 2006), confirms that the VHE source lies toward the a local enhancement of molecular material, a giant molecular cloud.
At 5 kpc, the luminosity of HESS J1834-087 between 250 GeV and a few TeV is about 5 × 10 34 erg s −1 , similar to the luminosity of HESS J1813-178 (Albert et al. 2006a) if that source is considered associated with SNR G12.8-0.0 at a distance of ∼ 4 kpc. We note, though, that J1813-178 has been found to be nearly point like whereas in the present case, a significant extension is observed. The γ-ray spectrum of HESS J1834-087 is steeper than the one of J1813-178. From the observed γ-ray luminosity, and assuming an acceleration efficiency of hadrons in the order of 3% and a supernova power of 10 51 erg, the required density of matter in the γ-ray production region for hadrons to be mainly responsible of the observed radiation can be estimated from the formula L γ ∼ (velocity of light) × (density) × (efficiency of acceleration) × (supernova power) × (p − p cross section) × (γ p − p inelasticity), and it is of about ∼ 11 cm −3 (see Torres et al. 2003 and references therein). With the extension of HESS J1834-087, and the gas mass found to be in the innermost contour of the CO map, i.e., in close superposition with the very high energy source, there is enough mass to generate the high energy radiation hadronically, even if only part of the gas is interacting with the SNR shock.
All in all, the observation of HESS J1834-087 using the MAGIC Telescope confirms a new very high-energy extended γ-ray source in the Galactic Plane. A reasonably large data set was collected from observations at medium-scale zenith angles to infer the spectrum of this source up to energies of a few TeV. Above 200 GeV, the differential energy spectrum can be fitted with a power law of slope Γ = −2.5 ± 0.2. The results of the independent observations of the HESS and MAGIC telescopes are in agreement within errors concerning the level of flux, the spectral shape, the morphology, and the extension of the source. The coincidence of the VHE γ-ray source with SNR G23.3-0.3 (W41) poses this SNR as a natural counterpart, and although the mechanism responsible for the high energy radiation remains yet to be clarified, a massive molecular cloud has been identified in the region.  GeV. The color scale shows the excess significance. Overlayed are contours of 20 cm VLA radio data from White et al. (2005) (green) and 13 CO emission contours (black) from Jackson et al. (2006). The contours of the radio emission are at 0.0035 Jy/beam. The 13 CO contours are at 10/20/30 K km/s, integrated from 70 to 85 km/s in velocity, as was the 12 CO data in figure 1. The white circle indicates the MAGIC PSF which is about 0.1 deg for all three lower SIZE cuts.  (Aharonian et al. 2006). The dashed line shows the spectrum of the Crab nebula as measured by MAGIC (Wagner et al. 2005).