Free Access
Issue
A&A
Volume 534, October 2011
Article Number A130
Number of page(s) 8
Section Extragalactic astronomy
DOI https://doi.org/10.1051/0004-6361/201117215
Published online 19 October 2011

© ESO, 2011

1. Introduction

The high-frequency-peaked BL Lac object (HBL) 1ES 0229+ 200 is located at αJ2000 = 2h32m48.62s, δJ2000 =  +20°17′17.45″ (Rector et al. 2003) and has a redshift of z = 0.14 (Woo et al. 2005).

HBL are characterized by two peaks in their spectral energy distribution (SED) which are located in the UV-X-ray and the GeV-TeV band, respectively. These are commonly interpreted in terms of leptonic models (e.g. Marscher & Gear 1985) as synchrotron and inverse Compton (IC) emission from a population of relativistic electrons upscattering their self-produced synchrotron photons (synchrotron self-Compton (SSC) models).

1ES 0229+200 was discovered in the Einstein IPC Slew Survey (Elvis et al. 1992), and classified as a HBL based on its X-ray to radio flux ratio (Giommi et al. 1995).

The VLA observations of 1ES 0229+200 reveal a core flux of 51.8mJy and show curved jets with an extension of  ~30″ at 1.4 GHz and jet position angles of PA =  −10° and PA = 180° (Rector et al. 2003).

1ES 0229+200 is not detected in the high energy γ-ray range (100 MeV  <  E  <  100 GeV) by Fermi/LAT in two years of observations and hence it is not mentioned in the second Fermi catalog (Abdo et al. 2011).

In 1996, it was originally predicted to be a potential VHE γ-ray source based on its SED (Stecker et al. 1996), however, Whipple, HEGRA, and Milagro have only reported upper limits (Horan et al. 2004; Aharonian et al. 2004; Williams 2005). Very high-energy (VHE, E > 100 GeV) emission up to 10TeV was first detected with the High Energy Stereoscopic System (H.E.S.S.) in 2006 (Aharonian et al. 2007). In this study, a hard spectrum with a photon index of Γ = 2.5 ± 0.19stat ± 0.1sys was reported. Besides 1ES 1426+428 (Aharonian et al. 2003), 1ES 0229+200 is the only source at redshift z > 0.1 whose spectrum was measured up to this high energy. 1ES 0229+200 has a very hard intrinsic VHE spectrum, hence it is well-suited to EBL studies (e.g. Aharonian et al. 2007; Kneiske & Dole 2010). Its spectral characteristics have been used to probe intergalactic magnetic fields (e.g. Neronov & Vovk 2010; Tavecchio et al. 2010). At the same time, these hard spectra are challenging for blazar models. We have launched dedicated, simultaneous multi-wavelength observations with XMM-Newton (X-ray, UV) and ATOM (optical) to determine the broad-band spectra in more detail.

2. Multi-wavelength observations and data analysis

2.1. X-ray data from XMM-Newton, Swift, and RXTE

XMM-Newton observations of 1ES 0229+200 were carried out on August 21 and 23, 2009 for 23 and 28 ks, respectively. The observations were conducted with MOS1 and PN in full imaging mode and MOS2 in timing mode, all with a thin filter. The two grating spectrometers onboard XMM-Newton RGS 1, 2 were also used to acquire data. The data analysis of the XMM-Newton observations was performed with SAS v.9.0. The data from the X-ray instruments were reprocessed as described in the SAS user-guide1. Both observations are influenced by short soft proton flares at the end of each observation, detected in the high emission  >12keV for the PN (>10keV for MOS) detector. Therefore, the good time intervals were determined using a cut at 0.4 cts/s for PN (0.35 cts/s for MOS) for the high-energy count rate resulting in an effective exposure of 7.3 ks for PN (23.1 ks for MOS) for the first and 12.1 ks for PN (24.4 ks for MOS) for the second pointing. No significant variations were found in either observation using different binning down to a timescale of 100 (10) seconds for the imaging mode (timing mode) during these pointings.

For the imaging mode, spectra were extracted from a source region of 30″ radius for the PN (80″ for MOS) around the position of 1ES 0229+200 and background regions with the same radii, on the same chip of the detectors that contained the source region. No significant pileup was found in the spectra of the source regions using the tool epatplot. The RMF (redistribution matrix file) and ARF (ancillary response file) are calculated for each spectrum using the tools rmfgen and arfgen.

The grating spectrometers RGS 1, 2 onboard XMM-Newton measure in the energy range 0.35 to 2.5 keV. The data are analysed using the tool rgsproc.

From October 19 to November 23, 2009, sixteen Swift observations were targeting on 1ES 0229+200. During these pointings with observations of 0.5 − 4ks each, the XRT detector (Burrows et al. 2005) was operated in windowed-timing (WT) mode in the energy range 0.2 − 10keV. In 2008, 1ES 0229+200 was observed on August 5, 7, and 8 in photon-counting (PC) and WT mode. However, only the PC mode observations are taken into account in the analysis, since only 120 s were observed in WT mode resulting in an insufficient amount of counts. For the Swift spectral analysis, XRT exposure maps were generated with the xrtpipeline to account for some bad CCD columns that are masked out on-board. The masked hot columns appeared when the XRT CCD was hit by a micrometeoroid. Spectra of the Swift data in PC-mode have been extracted with xselect from an annulus region with a radius of 0.8′ at the position of 1ES 0229+200, which contains 90% of the PSF at 1.5 keV. The background was extracted from a circular region with radius of 3′ near the source. For the WT-mode, boxes (~1.6′ × 0.3′) covering the region with source photons and a background region of similar size were used to extract the spectra. The auxiliary response files were created with xrtmkarf and the response matrices were taken from the Swift package of the calibration database caldb 4.1.32.

X-ray observations with the Proportional Counter Array (PCA) detector onboard RXTE (Bradt et al. 1993) were obtained in the energy range 2−60keV from January 1 to October 13 2010 with exposures of 1−2ks per pointing. Only RXTE/PCA data of PCU2 and the top layer 1 were considered to ensure the highest signal-to-noise ratio. The data were filtered to account for the influence of the South Atlantic Anomaly, tracking offsets, and electron contamination using the standard criteria recommended by the RXTE Guest Observer Facility (GOF). For the count rate of  ~1cts/s for this observations, the faint background model, provided by the RXTE GOF was used to generate the background spectrum with pcabackest and the response matrices were created with pcarsp.

The second instrument onboard RXTE, the HEXTE (High Energy X-ray Timing Experiment) onboard RXTE takes data in the energy range 15 to 250 keV. Since 2006, the HEXTE cluster A has been operating only in ON-source mode and since the end of 2009 the cluster B is permanently operating in OFF-source mode. For the spectral analysis, the cluster B data are used as background information for the cluster A data. The sum of all observations from January 1 until October 13, 2010 show no significant signal from 1ES 0229+200 in this energy range.

Table 1

X-ray observations of 1ES 0229+200 showing the results for the annual binning of XMM-Newton, Swift, and RXTE, as well as historical observations from Einstein (Elvis et al. 1992), ROSAT (Brinkmann et al. 1995) and BeppoSAX (Donato et al. 2005).

2.2. Additional X-ray sources in the XMM-Newton field of view

The source detection tool edetect_chain revealed 20 point sources in the field of view of the PN detector and one source that is extended beyond the PSF of the instrument. Two point sources are coincident with IRAS sources, four point sources are coincident with NVSS sources, and 15 with stars from the Guide Star Catalogue (GSC, Lasker et al. 2008, see Table 3). The remaining point sources do not have a counterpart in these catalogues and remain unidentified.

The extended source XMMU 023318.0+201237 is located at αJ2000 = 2h33m18.05s, δJ2000 =  +20°12′37.19″ and is very close (58″ offset) and likely connected to a point-like source XMMU 023315.5+201323 that is positional coincident with an infrared source IRAS 02304+2000 with coordinates αJ2000 = 2h33m14.6s, δJ2000 =  +20°13′30″ and a flux of 0.4 Jy at 60 μm (Moshir 1990). This infrared (IR) source has a radio counterpart NVSS023314+201330 that has an elliptical shape in the NVSS sky map and a flux of 12.4 ± 1.1 Jy at 1.4 GHz (Condon et al. 1998). One of the R-band observations by ATOM cover the region of this source and an object with 16.5 mag is measured. A closer look yields the detection of two very close (4″ separation) point sources of which the brighter one can be identified with USNO-B1.0 1102-0028956.

2.3. UV data from XMM-Newton/OM and Swift/UVOT

The optical monitor (OM) (Mason et al. 2001) onboard XMM-Newton observed 1ES 0229+200 in the filters UVM2 (231 nm), UVW1 (291 nm), and U (344 nm) simultaneously with the X-ray telescope. The analysis of these data was performed with the analysis described in the webpage3.

The UVOT instrument (Roming et al. 2005) onboard Swift measures the UV and optical emission in the bands UVW2 (188 nm), UVM2 (217 nm), UVW1 (251 nm), U (345 nm), B (439 nm), and V (544 nm) simultaneously with the X-ray telescope with an exposure of 0.4 − 3ks each. The instrumental magnitudes and the corresponding flux (see Poole et al. 2008, for the conversion factors) are calculated with uvotmaghist taking into account all photons from a circular region of radius 5″ (standard aperture for all filters). An appropriate background was determined from a circular region of radius 5″ near the source region without contamination of sources.

2.4. Optical data from ATOM

The 75-cm telescope ATOM (Hauser et al. 2004), located at the H.E.S.S. site in Namibia, monitored the flux in the different filters B (440 nm), V (550 nm), R (640 nm), and I (790 nm) according to Bessell (1990). The obtained data were analysed using an aperture of 4″ radius. Photometric calibration was done using the standard fields SA 113 and SA 95 from Landolt (1992).

3. Spectral data in the synchrotron range

All XMM-Newton, Swift, and RXTE spectra are binned with at least 25 counts and xspec v12.5 was used for the spectral analysis. For the XMM-Newton spectra, the energy ranges were restricted to 0.1 − 10keV for MOS and 0.2 − 15keV for PN following the suggestion of the calibration information4. The MOS1 and PN spectra are fitted simultaneously using a constant parameter to account for the different normalizations. A power-law model of the form F(E) = N0(E/E0) − Γ was used to fit the X-ray spectra taking into account the Galactic absorption of NH = 7.9 × 1020cm2 (LAB Survey, Kalberla et al. 2005). As can be seen in Fig. 1a, the residuals of the fit to the X-ray spectra deviate from the expected form, which implies that there is either additional absorption or different spectral characteristics.

Fitting a power-law model with unconstrained absorption results in a much better description of the data with χ2/d.o.f. = 979.7/903 = 1.08. The fit parameter and the goodness of each model is given in Table 2. Residuals are shown in panel b of Fig. 1.

Alternatively, the deviations shown in Fig. 1 could be avoided by taking into account the Galactic absoption and fitting a broken power-law. This fit results in photon indices of Γ1 = 0.8 ± 0.3, Γ2 = 1.78 ± 0.01, and a break at 0.58 ± 0.09keV (χ2/d.o.f. = 935/887). However, the low energy extrapolation of the X-ray spectrum would not fit the UV-optical range, hence this model is disfavoured.

To identify the location of the additional absorption, we test the hypotheses that either the total NH is located at redshift z = 0 or that the additional absorption is located at redshift z = 0.14. The resulting models with two different redshifts are indistinguishable and have similar goodnesses of fit (at z = 0.14: χ2/d.o.f. = 961.3/888; at z = 0: χ2/d.o.f. = 959.3/888). Hence the location of the additional absorption could be either intrinsic to 1ES 0229+200 or in the line of sight to the observer or in the Milky Way. We note that both a local enhancement of Galactic NH by 25% along the line of sight to 1ES 0229+200, as well as an intrinsic column within the source of 2.9 × 1020cm2, are plausible.

thumbnail Fig. 1

XMM-Newton MOS1 (black) and PN (red) spectrum of 1ES 0229+200 from August 21, 2009. The spectra can be well fit with a power-law model taking into account an absorption larger than the Galactic absorption. In panel a), we plot the residuals for a power law considering the Galactic absorption as fixed parameter, and in panel b) residuals for a power law with a free absorption.

Open with DEXTER

Table 2

XMM-Newton fit parameter, goodness of fit, and unabsorbed flux resulting from a simultaneous fit of the MOS1 and PN X-ray spectra.

The XMM-Newton/RGS spectra show no significant line emission and the continuum spectra are well-described by a power law with additional absorption comparable to the PN and MOS spectra.

The spectrum of the extended source detected in the PN and MOS observations close to 1ES 0229+200, was extracted from a region of radius 1.7′ for the PN detector (1′ for the MOS detector due to the CCD gaps) and reveals a faint source with a flux of  ≈ 1 × 10-13ergcm-2s-1. We were able to fit the spectrum with a binning of at least 15 photons per bin with a fixed Galactic absorption and a power law model with Γ = 2.0 ± 0.2 (χ2/d.o.f. = 73/96) and slightly better by a thermal model (mekal) with kT = 4 ± 2keV (χ2/d.o.f. = 87/96).

The annual averages of the Swift spectra in 2008 and 2009 are shown in Table 1. The consecutive pointings of the 2009 data obtained between October and November were binned in three ten-day intervals each to increase the statistics and are shown in Fig. 2. Initially, a simple power law and a free absorption was used to fit these spectra and result in the best description of the spectral shape. No significant change in absorption was detected with values of NH = (1.6 ± 0.4) × 1021cm2 in 2008 and NH = (1.3 ± 0.5) × 1021cm2, NH = (0.9 ± 0.5) × 1021cm2 and NH = (1.3 ± 0.2) × 1021cm2 for the spectra of 2009 binned in ten-day intervals. These values are comparable to the one obtained from the XMM-Newton spectra, which provide the most precise determination of the additional absorption. A total column of NH = 1.1 × 1021cm2 was assumed to obtain the photon indices and the fluxes (see Table 1 and Fig. 2). The Swift data from August 2008 were analysed and are shown in the left panel in Fig. 2 (2 454 685 JD). It should be noted here that these data were already presented in Tavecchio et al. (2009). Our re-analysis reveals that the integrated flux between 2 and 10 keV is lower by a factor of  ~7, but confirms the spectral indices. This is independent of the choice of a higher value of NH.

thumbnail Fig. 2

The X-ray flux of 1ES 0229+200 varied by a factor of  ~2 in October 2009 as shown in the combined lightcurve taken with the instruments XMM-Newton (red cirlces), Swift (black open squares), and RXTE (blue crosses) in the energy range 2–10 keV. For comparison, the Swift observation of 2008 is also shown. The lower panel displays the spectral indices derived by fitting power-law models to the spectra.

Open with DEXTER

The RXTE spectra have been summed covering about 30 days to achieve higher statistics. The energy range from 2–3 keV and 20–30 keV were excluded from the spectral analysis, because of instrumental features (e.g. spikes in the background files, see website5). A simple power-law was used to fit these spectra. The resulting photon indices and fluxes can be seen in Fig. 2.

The historical spectrum of BeppoSAX was fitted with a power law of photon index Γ = 1.99 ± 0.05 and a free absorption of NH = (10 ± 5) × 1020cm-2 in the energy range 0.1 − 50keV without any indication for a cut-off at the high energy end (Donato et al. 2005).

The Swift/BAT spectrum taken from the 58 month catalog6 (Baumgartner 2010) is well fit by a power law with photon index of Γ = 2.1 ± 0.3 () and result in a flux between 14 and 195 keV of F = (28.2 ± 5.5) × 10-12ergcm-2s-1. The BeppoSAX PDS spectrum of 2001 (Donato et al. 2005) is in good agreement with this high energy spectrum of Swift/BAT.

4. Temporal study

Regular monitoring observations with ATOM starting in 2006 with 7 observations in 2006, 47 in 2007, 86 in 2008, 61 in 2009, and 24 in 2010 show no significant variations (<1%) in the R band over five years of observations. The average brightness is mR = 16.39 ± 0.01mag and mB = 18.38 ± 0.02mag.

thumbnail Fig. 3

Ultraviolet fluxes measured by XMM-Newton and Swift for 1ES 0229+200 in 2008 (average over 3 days) and 2009 (U: black circles, UVW1: blue open squares, UVW2: red crosses, UVM2: brown open diamonds). The flux points at JD 2 455 064.6, 2 455 066.6, 2 455 117.8, and 2 455 118.9 are shifted by  ± 0.4 days for better visibility. Marginal variations are detected in the UVW1 and UVM2 bands.

Open with DEXTER

Table 3

X-ray sources detected in the XMM-Newton observation on 1ES 0229+200.

The UV results from XMM-Newton and Swift of 2008 and 2009 are shown in Fig. 3. The U and UVW2 band emission was stable and does not show significant variation (pχ2 = 70% and pχ2 = 98% for a constant, respectively.). The data in the UVW1 and UVM2 bands display a marginally significant drop by  ~30% and  ~20%, respectively, at JD 2 455 120, around the time of an increase in the X-ray flux (see below).

The X-ray flux of 1ES 0229+200 instead shows an increase by a factor of around two starting at JD 2 455 120, as can be seen in Fig. 2. For all of 2010, only a marginal variation could be detected. During 2008–2010, marginally significant spectral changes could be detected as variations in the photon index (pχ2 = 1% for a constant) that were uncorrelated with the flux variation (see Fig. 2). During all epochs, the source can be well-described with a power-law model and an absorption in excess of the Galactic absorption, as described in Sect. 3. At energies above 14 keV, measured by Swift/BAT, marginally significant variation is detected in the monthly lightcurve with a probability for the fit of a constant of pχ2 ≈ 1% over the 58 months of observation7 (Baumgartner 2010). However, since a long integration time is needed, a behaviour similar to that in the 2–10 keV range would be difficult to detect.

5. Spectral energy distribution

thumbnail Fig. 4

Synchrotron emission of 1ES 0229+200 with simultaneously measured optical, UV, and X-ray emission by ATOM and XMM-Newton of August 21, 2009 (black dots). The optical and UV emission is corrected for both the host galaxy and Galactic extinction, and the X-ray emission observed by XMM-Newton/MOS is corrected for the absorption. The bars in the UVW1 and UVM2 bands are explained in Sect. 5.1. The grey open squares represent the non-simultaneous Swift/XRT spectrum (corrected for detected absorption) with the highest flux in 2009. The Swift/BAT 58 month (Dec. 2004–Oct. 2009) spectrum is shown in the energy band  >10 keV (black crosses). The dashed line (butterfly) represents the BeppoSAX spectrum from 2001 (Donato et al. 2005). The grey upper limits represent historical UV observations with GALEX that are extinction corrected and of origin discussed in the text.

Open with DEXTER

The simultaneous observations in the optical by ATOM, the UV, and X-rays by XMM-Newton are considered to study the synchrotron spectrum (see Fig. 4, simultaneous data of August, 21 2009).

5.1. Host galaxy of 1ES 0229+200

The host galaxy of 1ES 0229+200 is an elliptical galaxy with a brightness of mhost,R = 15.85 ± 0.01mag and a half-light radius of re = 3.25 ± 0.07″ (Urry et al. 2000). Other observations in the Bessel R-band with the Nordic Optical Telescope (NOT) (Falomo & Kotilainen 1999) show results with mhost,R = 15.76mag and re,R = 4″. The host galaxy profile of 1ES 0229+200 was also studied in the Bessel U, B, and V-bands with the Nordic Optical Telescope (NOT) by Hyvönen et al. (2007). The results are mhost,B = 18.75mag, mhost,U = 18.83mag, and mhost,V = 17.58mag with half light radii re,B = 5.65″, re,U = 2.75″, and re,V = 4.90″. In order to correct for the host galaxy light, a de Vaucouleurs profile of the galaxy was assumed and the measured brightness was transformed to that of an aperture of 4″ using Eq. (4) of Young (1976). ATOM photometry was performed with a 4″ aperture. The resulting host-galaxy corrected fluxes would result in an unphysical bump in the V band in the SED (in νFν) since the calculated influence of the host galaxy in the V band is only 30%, while in the R and B band it is 90% and 30%, respectively. Hence, the influence of the host galaxy in the R, B, V, and U filters were also calculated using a spectral template for a nearby elliptical galaxy by Fukugita et al. (1995). Here, the R-band magnitudes and the half-light radius detected by Falomo & Kotilainen (1999) were used. The influence of the host galaxy is then  ~90%,  ~80%, and  ~57% for R, V, and B, respectively, which are the values that we used to correct the measured fluxes (shown in Fig. 4). The host-galaxy corrected flux in the R-band is compatible with the detected nucleus magnitude of Falomo & Kotilainen (1999) and Urry et al. (2000) as expected from the absence of variability. The host galaxy influence in the UVW1 and UVM2 bands is unknown. Figure 4 therefore shows two values, connected by a bar. The upper ones correspond to the measured values corrected only for extinction, the lower ones also assume a correction for the host galaxy of 30% (the value derived in the adjacent U band).

In an independent check, the spectral slope was extracted using the Sloan Digital Sky Survey (SDSS) observations (five bands taken simultaneously). The resulting slope is identical to the one measured by ATOM. Since the data do not match the epoch of the ATOM observations, the absolute fluxes were not considered.

5.2. Galactic extinction/absorption correction

Since the host galaxy magnitudes were not extinction corrected, the influence of the host galaxy was first subtracted. The extinction correction was thereafter applied to the AGN light.

The measured UV fluxes were corrected for dust absorption using E(B − V) = 0.135 mag (Schlegel et al. 1998) and the Aλ/E(B − V) ratios given in Seaton (1979), resulting in a correction of 70%,57%,48% for the UVM2, UVW1 and U-band, respectively. For the optical filters, the values for the extinction were derived from the interstellar reddening curve and table given by Zombeck (1990)8. The correction of NH absorption was applied as described in Sect. 3.

thumbnail Fig. 5

Spectral energy distribution of 1ES 0229+200 with simultaneous measured optical, UV, and X-ray fluxes, all corrected for host galaxy emission, Galactic extinction, and Galactic absorption is shown as black data points. The 58 months Swift/BAT spectrum is shown  >10 keV (black crosses). In grey (filled and open circles), historical radio and UV data are shown and their origin is discussed in the text. The grey upper limits in the GeV energy range are taken from Dermer et al. (2011) and represent the upper limit in the energy bins 0.1−1GeV, 1−10GeV, and 10−100GeV by Fermi observations (Aug. 2008 to Sep. 2010). The VHE γ-ray spectrum measured by H.E.S.S. (black circles, taken from Aharonian et al. 2007), as well as the EBL corrected, intrinsic source spectrum (grey open diamonds) with a hard photon index, which implies an inverse Compton emission peaking at very high frequency (>1027Hz), is shown. The solid line represents an SSC model that can describe the intrinsic synchrotron and inverse Compton emission of 1ES 0229+200, and the dashed line represents the absorption by the EBL (details about these models are described in the text).

Open with DEXTER

5.3. Historical multi-wavelength data

GALEX observed 1ES 0229+200 at October 29, 2007 in the far UV (152.8nm) and near UV (227.1nm) leading to a measured flux of 31.68μJy and 50.36μJy, respectively, taken from GalexView9. The measured near and far UV fluxes were corrected for dust absorption using E(B − V) = 0.135 mag (Schlegel et al. 1998) and the Aλ/E(B − V) ratios given in Seaton (1979). A correction of 64% and 68% resulted for the far and near UV bands, respectively. Owing to a lack of information about the host galaxy influence in these wavebands (this influence should be very small compared to the extinction correction), the measured fluxes are shown as upper limits for the intrinsic synchrotron spectrum of 1ES 0229+200 (shown as grey arrows in Fig. 4).

In 2006, 2008, and 2009 INTEGRAL also observed 1ES 0229+200 several times for approximately 66ks with ISGRI (17 − 80keV) and 61ks with JEM-X (3 − 10keV), but the source appears faint and too few photons were detected so that no reasonable light curve or spectrum could be extracted10.

In historical snapshot observations of VLA at 6 cm in 1992 fluxes of 41.5mJy (Schachter et al. 1993) and 49.09mJy (Perlman et al. 1996) were detected (shown as grey open circles in Fig. 5). With 6 cm VLBA observations this flux was resolved into a core of 22.7mJy and a jet of 7.7mJy (Rector et al. 2003). In the same study, the core flux at 1.4 GHz was detected by VLA observations to be 51.8mJy, and 1ES 0229+200 was found to have curved jets to the north and south with extensions of 30″. The core fluxes are shown as grey filled circles in Fig. 5.

5.4. SSC model

Katarzyński et al. (2006) demonstrated that synchrotron spectra with a high minimum Lorentz factor can explain very hard IC spectra. The simultaneously obtained data in the optical, UV, and X-ray bands cover the whole synchrotron emission. This allows an empirical determination of the minimum Lorentz factor in 1ES 0229+200. The high accuracy of the XMM-Newton measurements constrains the synchrotron spectrum very efficiently.

The photon index Γ = α + 1 of the X-ray spectrum gives a direct estimate of the spectral index n of the electron distribution N(E)dE ∝ E − ndE with the relation n =  −1−2α. A broken power law with a canonical cooling break of Δn = 1 is assumed for the electron distribution. The variability of 1ES 0229+200 detected in the X-ray emission places constraints on the maximum size of the emission volume considered in the SSC model which is inferred to be R < (Δt/s) × D × c ≈ D × 2.6 × 1014m.

The peaks in the SED are commonly explained by leptonic models, such as those of Marscher & Gear (1985), as synchrotron and IC emission from a population of relativistic electrons up-scattering their self-produced synchrotron photons (SSC). The code of Krawczynski et al. (2004) for a one-zone SSC model was used to describe the intrinsic emission of 1ES 0229+200. For this model, a spherical emission volume of radius R, moving with a bulk Doppler factor D towards the observer and a magnetic field B was assumed. The electron distribution was described by a broken power-law between minimum and maximum energy.

The X-ray spectrum can be fitted well by a single power-law model over 1.5 orders of energy. The extrapolation fits the absorption and host-galaxy corrected UV data, suggesting that there is a single power law slope for the energy range 0.005 − 100keV. The absorption and host galaxy corrected UV-optical emission below this range is significantly steeper and strongly constrains the minimum energy of the electron distribution function, yielding a very high value. The well-determined spectral index in the X-ray regime has to describe the low-energy spectral index of the electron distribution n1 =  −2.6. The uncooled electron spectral slope hence has n2 =  −3.6 following the canonical break. The radius was kept as the maximum value obtained from the variability timescale (R = 1 × 1016m). This radius is an upper limit and is fixed independently of the SED modelling. Smaller values of radii would also be consistent with the variability constraint. Tavecchio et al. (2009) used a smaller value in their attempts to describe the SED compiled in their study. The SED shown in Fig. 5 cannot be reproduced with a radius (R < 1 × 1015m) (see below). To account for the high energy peak of the IC emission, the underlying electron distribution must be very narrow. Together these constraints are best met with the model parameters Emin = 2 × 1011eV (γmin = 3.9 × 105), Ebreak = 3.2 × 1013eV (γbreak = 6.2 × 107), Emax = 1 × 1014eV (γmax = 1.9 × 108), and a magnetic field of B = 3.2 × 10-5G. The Doppler factor of D = 40 was chosen as the smallest one possible to reproduce the spectra. This model can describe the simultaneously obtained spectra in the synchrotron regime, the long-term hard X-ray spectrum by Swift/BAT, and the non-simultaneous, intrinsic TeV spectrum. The peak frequency of the synchrotron and IC emission based on the described model are νsy,peak = 3.5 × 1019Hz and νIC,peak = 1.5 × 1027Hz. The peak fluxes show a slight Compton dominance with νsy,peakFνsy,peak = 9.7 × 10-12ergcm-2s-1 and νIC,peakFνIC,peak = 1.1 × 10-11ergcm-2s-1. This model is shown in Fig. 5 as a solid line, while the dashed line represents the absorption by the EBL using the model of Franceschini et al. (2008). This model is in good agreement with the EBL absorption model used in Aharonian et al. (2007) to correct the observed TeV spectrum (shown in Fig. 5).

The cooling break is at very high energies such that a single power-law model for the electron distribution with n =  −2.6 fits the data equally well and is an alternative model to describe the measured spectra of the SED.

An earlier attempt to utilize the suggestion of Katarzyński et al. (2006) was reported by Tavecchio et al. (2009). They assumed a narrow single power-law electron distribution with a high minimum Lorentz factor to fit the data. As in the modelling presented here, the slope of the synchrotron emission below γmin is given by Fsy ~ ν1/3, leading to a slope for the IC part that is similar to that is required to fit the TeV spectrum. Attempting to fit the overestimated X-ray flux, Tavecchio et al. (2009) derived a different set of parameters, including a higher (factor 10) magnetic field and a smaller radius. The initial model presented above used an upper limit to the radius, which was not optimized in the SED fitting. Attempts were made to reproduce the SED with models using smaller radii. The precise determination of the low energy cutoff in the UV range and the well-determined spectral shape in the X-ray range can only be reproduced by invoking a very high Doppler factor (D > 100). While providing an adequate fit to the synchrotron range, they do not reproduce the IC emission peak well.

The cut-off of the synchrotron emission above 100 keV represents the maximum energy of the electron distribution. No change in spectral shape could be detected in the Swift/BAT spectrum. Therefore only a lower limit to the maximum cut-off energy could be determined empirically. The precise measurement of the maximum energy would require a soft γ-ray detection. In the IC part of the spectrum, the maximum energy cannot be determined emprically either because any increase in Emax does not result in a change in the IC spectrum because of Klein-Nishina suppression. Assuming that the high energy cut-off of the synchrotron emission is at the high energy end of the Swift/BAT spectrum (~100keV), this results in an electron distribution that extends only over five orders of energy.

Changes in the peak fluxes of the synchrotron emission (such as the difference between the XMM-Newton and the Swift/XRT spectrum, as shown in Fig. 4) were also tested. Any changes in the parameters of the emission volume corresponding to an increase in flux by a factor of  ~2 yield a change in the IC peak flux of a factor of  ~2 as well. This implies that changes in the VHE fluxes in the TeV energy range are also expected to occur in 1ES 0229+200.

Alternative models were also explored. Apart from the attempts to describe the SED by SSC emission from a very compact source, other assumption would also lead to models involving a high Doppler factor. Assuming that the derived curvature in the optical regime would be represented by the break of the electron distribution with n1 =  −1.6, n2 =  −2.6, a separate model can be constructed. This scenario cannot be excluded by our simultaneously measured SED above 1014Hz, but would require a large Doppler factor of about D = 100 in order to account for the high energy peak of the IC emission. This alternative model also underpredicts the radio flux of the core by a large factor, e.g. around 70%. Interestingly, the radio flux of the jet, as measured by Rector et al. (2003), could be represented by such model, indicating that the whole synchrotron emission originates from the jet. Another possible way of explaining the high IC peak using such an electron distribution function, but with a lower Doppler factor, would be to invoke an additional external Compton contribution resulting from the scattering of the CMB as discussed in Böttcher et al. (2008) for 1ES 1101-232, a distant TeV blazar with a hard TeV spectrum. This model would decouple the X-ray and TeV emission because most of the TeV emission would result from the interaction with the CMB further away from the synchrotron emission region of the jet. This model would imply no VHE variability on short timescales. The effect of the CMB scattering depends on the distance to the source, so that it is more likely to be detectable in distant sources such as 1ES 0229+200 and 1ES 1101-232.

6. Conclusions

The X-ray spectral information that have been presented here cover all available X-ray data up to October 2010. Together with the simultaneous UV and optical observations, they yield a very good coverage of the synchrotron emission. Our detailed study of the high quality XMM-Newton spectrum has inferred an absorption higher than the nominal Galactic column density, which could be intrinsic to the source or caused by Galactic excess absorption.

The host-galaxy, extinction-corrected optical and UV fluxes have been shown to provide strong evidence that the cut-off of the low energy part of the synchrotron emission is located between the optical and X-ray regime. Therefore, the minimum energy of the electron distribution has to be rather high to account for this cut-off. As suggested by Katarzyński et al. (2006), an electron distribution with a high minimum Lorentz factor is needed to reproduce a hard TeV spectrum. A narrow electron distribution, as indicated by the high minimum energy, results in a hard intrinsic VHE γ-ray spectrum as deduced for 1ES 0229+200. This hard VHE spectrum also shows that the IC peak is at very high energies (>1027Hz), hence the initial electrons are very energetic.

An earlier attempt to utilize the suggestion by Katarzyński et al. (2006) was reported by Tavecchio et al. (2009), assuming a high minimum Lorentz factor. The main differences between the model presented here and their attempts are caused by the new, precisely determined low-energy cut-off in the optical regime and the simultaneously determined X-ray spectrum.

A broken power-law could be an alternative description to the additional absorption found in the X-ray spectra. However, the low energy extrapolation of the X-ray spectrum would not fit the optical range and is therefore disfavoured.

The hard X-ray spectrum up to 15 keV together with the long-term spectrum by Swift/BAT have been found to show that the synchrotron peak is extended up to  ~100keV without any significant cut-off in the X-ray spectrum.

1ES 0229+200 is defined as a high-frequency-peaked BL Lac object, and the measured synchrotron emission peaks at higher frequencies (>100 keV) than usual for HBL and belongs therefore to the class of extreme blazars. The exact peak frequency cannot be determined, since the hard X-ray spectrum does not show any significant cut-off. Assuming that the high energy cut-off of the synchrotron emission is at the high energy end of the Swift/BAT spectrum (~100keV), the underlying electron distribution extend only over five orders of magnitude in energy, which is a narrow range. For no other extreme blazar has this range been demonstrated to be so narrow.

The low energy part of the synchrotron emission has not been found to show significant variation over four years in the optical R band. Only minor variation has been detected in the ultraviolet flux during 2008 and 2009. The 2–10 keV X-ray flux instead varied by a factor of  ≈ 2 within around 20 days in 2009. In the high energy part of the synchrotron emission, measured by Swift/BAT, variations in this amplitude and timescale could not be detected because of the longer integration time needed.

Several blazars are known as so-called extreme blazars with synchrotron emission extending to high energies (Costamante et al. 2001). It has been found that Mkn 501 revealed a very high energy peak of synchrotron emission around 100 keV in a flare with a detected large shift in the peak frequency compared to previous observations (Pian et al. 1998). The monitoring by BeppoSAX of 1ES 2344+514 has identified huge changes in the synchrotron peak frequency within one year with different spectral shapes (Giommi et al. 2000). INTEGRAL observations of 1ES 1426+428 display a synchrotron peak frequency around 100 keV (Wolter et al. 2008). These extreme blazars have similar synchrotron peak frequencies to 1ES 0229+200, but their very high energy spectra in the TeV range are much softer than for 1ES 0229+200. Hence, 1ES 0229+200 has the highest IC peak frequency among the extreme blazars known up to date.


Acknowledgments

The authors acknowledge the support by the XMM-Newton team in arranging simultaneous observations. The execution and availability of the RXTE and Swift observations and the use of the public HEASARC software packages are acknowledged. S.K. and S.W. acknowledge support from the BMBF through grant DLR 50OR0906.

References

All Tables

Table 1

X-ray observations of 1ES 0229+200 showing the results for the annual binning of XMM-Newton, Swift, and RXTE, as well as historical observations from Einstein (Elvis et al. 1992), ROSAT (Brinkmann et al. 1995) and BeppoSAX (Donato et al. 2005).

Table 2

XMM-Newton fit parameter, goodness of fit, and unabsorbed flux resulting from a simultaneous fit of the MOS1 and PN X-ray spectra.

Table 3

X-ray sources detected in the XMM-Newton observation on 1ES 0229+200.

All Figures

thumbnail Fig. 1

XMM-Newton MOS1 (black) and PN (red) spectrum of 1ES 0229+200 from August 21, 2009. The spectra can be well fit with a power-law model taking into account an absorption larger than the Galactic absorption. In panel a), we plot the residuals for a power law considering the Galactic absorption as fixed parameter, and in panel b) residuals for a power law with a free absorption.

Open with DEXTER
In the text
thumbnail Fig. 2

The X-ray flux of 1ES 0229+200 varied by a factor of  ~2 in October 2009 as shown in the combined lightcurve taken with the instruments XMM-Newton (red cirlces), Swift (black open squares), and RXTE (blue crosses) in the energy range 2–10 keV. For comparison, the Swift observation of 2008 is also shown. The lower panel displays the spectral indices derived by fitting power-law models to the spectra.

Open with DEXTER
In the text
thumbnail Fig. 3

Ultraviolet fluxes measured by XMM-Newton and Swift for 1ES 0229+200 in 2008 (average over 3 days) and 2009 (U: black circles, UVW1: blue open squares, UVW2: red crosses, UVM2: brown open diamonds). The flux points at JD 2 455 064.6, 2 455 066.6, 2 455 117.8, and 2 455 118.9 are shifted by  ± 0.4 days for better visibility. Marginal variations are detected in the UVW1 and UVM2 bands.

Open with DEXTER
In the text
thumbnail Fig. 4

Synchrotron emission of 1ES 0229+200 with simultaneously measured optical, UV, and X-ray emission by ATOM and XMM-Newton of August 21, 2009 (black dots). The optical and UV emission is corrected for both the host galaxy and Galactic extinction, and the X-ray emission observed by XMM-Newton/MOS is corrected for the absorption. The bars in the UVW1 and UVM2 bands are explained in Sect. 5.1. The grey open squares represent the non-simultaneous Swift/XRT spectrum (corrected for detected absorption) with the highest flux in 2009. The Swift/BAT 58 month (Dec. 2004–Oct. 2009) spectrum is shown in the energy band  >10 keV (black crosses). The dashed line (butterfly) represents the BeppoSAX spectrum from 2001 (Donato et al. 2005). The grey upper limits represent historical UV observations with GALEX that are extinction corrected and of origin discussed in the text.

Open with DEXTER
In the text
thumbnail Fig. 5

Spectral energy distribution of 1ES 0229+200 with simultaneous measured optical, UV, and X-ray fluxes, all corrected for host galaxy emission, Galactic extinction, and Galactic absorption is shown as black data points. The 58 months Swift/BAT spectrum is shown  >10 keV (black crosses). In grey (filled and open circles), historical radio and UV data are shown and their origin is discussed in the text. The grey upper limits in the GeV energy range are taken from Dermer et al. (2011) and represent the upper limit in the energy bins 0.1−1GeV, 1−10GeV, and 10−100GeV by Fermi observations (Aug. 2008 to Sep. 2010). The VHE γ-ray spectrum measured by H.E.S.S. (black circles, taken from Aharonian et al. 2007), as well as the EBL corrected, intrinsic source spectrum (grey open diamonds) with a hard photon index, which implies an inverse Compton emission peaking at very high frequency (>1027Hz), is shown. The solid line represents an SSC model that can describe the intrinsic synchrotron and inverse Compton emission of 1ES 0229+200, and the dashed line represents the absorption by the EBL (details about these models are described in the text).

Open with DEXTER
In the text

Current usage metrics show cumulative count of Article Views (full-text article views including HTML views, PDF and ePub downloads, according to the available data) and Abstracts Views on Vision4Press platform.

Data correspond to usage on the plateform after 2015. The current usage metrics is available 48-96 hours after online publication and is updated daily on week days.

Initial download of the metrics may take a while.