XMM OBSERVATIONS OF TWO QUADRUPOLAR OUTFLOWS

The first of the two regions, L1455, contains three low-luminosity IRAS sources, two of them in the central portion of the cloud where the most intense activity is concentrated (Bachiller & Cernicharo 1986). One of the IRAS sources, IRAS 03247+3001, coincides with a faint nebulous T Tau star, RNO 15 (Cohen & Kuhi 1979; Cohen 1980) and the Class II submillimeter peak L1455 IRS 2 (Wu et al. 2007). The other, IRAS 03245+3002, coincides with the Class I FIR and submillimeter source, L1455 IRS 1 (Wu et al. 2007; Shirley et al. 2000). The FIR source of Davidson & Jaffe (1984) lies slightly east of IRS 1, but is very likely the same object. Both IRAS sources have near-infrared counterparts in the c2d Spitzer IRAC images of L1455 (Jørgensen et al. 2006). Two other sources are prominent in the IRAC images: IRS 4, which is a bright Class I or Class 0 submillimeter source (Wu et al. 2007; Shirley et al. 2000; Hatchell et al. 2007), and IRS 5, which is a heavily reddened Class II star that is also known as L1455 FIR 2 (Tapia et al. 1997). The CO map of Goldsmith et al. (1984) traces out the high-velocity gas in this main section of L1455. It shows an elongated structure with red- and blue-shifted peaks that are well separated along an northwest–southeast axis, which defines the direction of the outflow. The source of the outflow has not been firmly established; however, based on the geometry of the flow and the relative locations of the known YSOs, IRS 4 is a strong candidate, as it lies closest to the center line of the flow.

Bally et al. (1997) have identified a number of HH knots and filaments from Hα and [S ii] images of L1455. Some of the same features are also visible in the near-infrared H₂ v = 1–0 S(1) maps of the central part of the cloud made by Davis et al. (1997a). HH 318, among the brightest of the Hα knots, and three other H₂ knots are closely aligned and coincide with a narrow northeast–southwest ridge of CO 3–2 emission (Davis et al. 1997b). IRS 1 lies along the same ridge, centered between the lobes of red-shifted CO to the northeast and blue-shifted gas to the southwest. IRS 1 thus appears to be the origin of a separate, narrowly collimated outflow of its own. This second flow overlaps but is independent of, and nearly orthogonal to, the CO flow that is possibly launched by IRS 4 from virtually the same location in the cloud.

The L723 star-forming region, the second dark cloud in the present study, is rather inconspicuous at optical wavelengths. It contains a single low-luminosity Class 0 source, IRAS 19156+1906 (Reipurth et al. 1993; Shirley et al. 2000). Starting with the original CO map of Goldsmith et al. (1984), L723 has been surveyed in increasing detail by Avery et al. (1990); Hirano et al. (1998); and Lee et al. (2002). The region has also been extensively mapped in CS (Hirano et al. 1998) and NH₃ (Girart et al. 1997). Shirley et al. (2000) mapped the area using the Submillimeter Common-User Bolometer Array (SCUBA) on the James Clerk Maxwell Telescope (JCMT) and found a single 450  μ and 850 μ peak coinciding with the IRAS source. Later work by Young et al. (2006), also using SCUBA, placed the 850 μ peak 75 farther east (but still within the IRAS error ellipse) at the position of an elongated radio jet, VLA 2 (Anglada et al. 1996). Another radio source in the same area, VLA 1, is unresolved and appears to be an extragalactic source that is unrelated to L723 (Anglada et al. 1996). The FIR source of Davidson (1987), L723 FIR, also coincides with VLA 2. Dartois et al. (2005) reported the detection of a very large column of CO₂ ice from spectra taken with the Spitzer Infrared Spectrograph, apparently in the direction of VLA 2. However, the exact coordinates of the observation were not given and cannot be determined from that work.

The interpretation of the CO observations of L723 has been controversial. Some authors have favored a single bipolar outflow from either the IRAS position or the radio position in the cloud core, but others have argued for a pair of independent bipolar flows that emerge from a common center there. The highest resolution CO maps reveal blue-shifted gas east and north of the cloud core and red-shifted gas both west and south of it (Figure 5, Hirano et al. 1998; Figure 1, Lee et al. 2002). Avery et al. (1990) have interpreted the morphology of the CO maps in terms of a single bipolar outflow whose CO emission follows the walls of a cavity that has been hollowed out by a wide-angle wind from the IRAS source. Hirano et al. (1998) have proposed a similar shell model but favor an origin from the VLA radio position. Lee et al. (2002) as well as Girart et al. (1997) prefer two independent flows that originate from the cloud core, one with a wide opening angle in the east–west direction (in position angle P.A.∼100 deg), the other a more highly collimated flow in the north–south direction (in P.A.∼30 deg). Girart et al. identify the wide east–west flow with VLA 2 and attribute the narrow north–south flow to a second, thus far undetected driving source at the location of a prominent hotspot in their NH₃ map (Figure 8, Girart et al. 1997), which lies within 3'' of the IRAS source. Lee et al. (2002) have suggested that both flows originate from the vicinity of the radio position, the east–west outflow from VLA 2 itself, the north–south flow possibly from an unresolved binary companion of VLA 2. Using the VLA in its high-resolution A configuration, Carrasco-González et al. (2008) have resolved the extended emission of VLA 2 and confirmed that the source is, in fact, a close radio binary. It is quite possible, therefore, that the structure of the quadrupolar flow in L723 is the result of separate flows from the two binary components of VLA 2, one outflow emerging in the north–south direction, the other east–west.

Table 1 provides the details of both XMM pointings. L1455 was observed for an exposure time of 44 kiloseconds (ks) during revolution 1399 on 2007 July 31. The observation ID is 0503670101. The aim point of the observation was centered on L1455 FIR, the FIR source of Davidson & Jaffe (1984). With that orientation, the young emission-line star RNO 15 and its FIR counterpart, IRAS 03247+3001, were placed within 2' of the optical axis of the telescope, where the best images are formed. The EPIC pn camera (Strüder et al. 2001) and the two MOS cameras (Turner et al. 2001) were operated in full-window mode, using the medium filter. L723 was observed with XMM on 2007 May 9 during revolution 1358 for a nominal exposure time of 39 ks. The observation ID is 0503670201. The field of view was centered on L723 FIR, the FIR source of Davidson (1987), and on the adjacent radio source VLA 2 (Anglada et al. 1996). The EPIC cameras were run in full-window mode. The medium filter was used.

To analyze both EPIC data sets, I used version 7.1 of the XMM Science Analysis Software (SAS; Watson et al. 2001) and a complete set of updated calibration files, which were downloaded from the XMM Science Operations Center (dated 2007 November 13). The SAS epchain and emchain tasks were used to create X-ray event lists for each camera. The lists were subsequently filtered using the SAS xmmselect task to pick out the photon events with pulse height energies in the broad 0.3–10 keV energy band. Events at higher or lower energy outside that band were discarded. Following the data analysis threads described in the XMM data analysis ABC Guide,¹ I retained only the pixel events that were assigned PATTERN values of 0–12 for the MOS and 0–4 for the pn. Hot and other bad pixels (i.e., those outside the field of view or close to the edges of the CCD chips, etc.) were removed by filtering out poor-quality events that were tagged with FLAG ≠0 during the normal pipeline processing.

Light curves were extracted in the 10–15 keV energy band from source-free images in each camera in order to remove any intervals of flaring particle background or other times marked by high background emission. Both observations experienced a moderately higher than average X-ray background level in the 0.3–10 keV energy band for most of their scheduled exposure time. The event files for both data sets were therefore filtered using slightly higher than normal rejection levels in order to identify the "good time intervals" of low background in the light curves, 0.35 count s⁻¹ for the two MOS cameras and 2.5 count s⁻¹ for the pn. Photon events that were recorded during the rejected segments of each exposure were eliminated and ignored in the rest of the analysis. To identify likely sources in the EPIC images, I ran the SAS edetect_chain science thread for each observation, searching all three cameras simultaneously. Separate source lists were compiled for a soft energy band of 0.5–2.0 keV, a hard energy band of 2.0–7.5 keV, and the total energy band of 0.3–10 keV. The three independent source lists were edited to remove spurious entries (e.g., false detections along the edges of the EPIC detectors), merged, and then purged of duplications. A final source list was determined for each observation after careful inspection and measurement of the individual broadband EPIC images.

The X-ray counts for each source were measured using the funcnts tool in the FUNTOOLS software package from the Smithsonian Astrophysical Observatory. For sources within 7' of the optical axis, I used a circular extraction cell of radius r = 8'' so as to enclose at least 50% of the energy in the local point-spread function (PSF). For sources located farther off-axis, I chose a larger radius, r = 15'', which also ensured an enclosed energy fraction of 50% or more. The measured counts in each case were extrapolated to 100% enclosed energy using the analytical expression given by Simon & Dahm (2005) and corrected for background emission, as estimated from a nearby, offset background region with r ⩾ 40''. To compute the X-ray count rates for individual sources, I derived exposure maps for each camera in the 0.3–10 keV energy band via the SAS eexpmap task. The exposure maps account for vignetting related changes in the effective response of the telescope across the field of view. Optical and near-infrared identifications were established for the X-ray sources by visually comparing the XMM images with images from the Digitized Sky Survey (DSS) of the Space Telescope Science Institute (STScI) and The Two Micron All Sky Survey (2MASS) Point Source Catalog, and also by cross-matching the pn and MOS image coordinates with entries in the SIMBAD database.

Table 2 summarizes the results of the X-ray photometry for L1455. For completeness I present the measurements from the entire XMM field of view, although the discussion below concentrates on the cluster of YSOs in the cloud core at the center of the field. Results for L723, also the full set of measurements, follow in Table 3. An identification number for each source is provided in Column (1) of both tables. Listed next is the name of the source in IAU format, composed in the usual fashion from its right ascension and declination (J2000.0). Column (3) gives the angular offset of the source from the bore sight of the XMM pointing. The name of any optical counterpart appears in Column (4). Names with a 2G prefix denote entries from the second edition of the Hubble Space Telescope Guide Star Catalog (GSC; Lasker et al. 2008). The offset between the XMM centroid position and the position of the nearest optical counterpart is shown in Column (5). The next two columns specify the effective exposure times for the source in the pn and MOS cameras. Columns (8) and (9) list the background-corrected source count rates and their 1σ statistical errors for the pn camera and the average of the two MOS cameras, respectively. The remaining columns in the tables list the successful matches of the XMM sources with the 2MASS catalog and the corresponding JHK_s photometry from the SIMBAD database.

Six of the 21 X-ray sources detected by XMM in L1455 and 10 of the 18 sources in L723 can be identified with near-infrared sources in the 2MASS survey. The corresponding JHK photometry from 2MASS is plotted in a color–color diagram in Figure 1 after conversion to the CIT system via the transformations given in Carpenter (2001). The open circles denote sources from L1455, those from L723 are plotted as filled symbols. Three X-ray sources plotted in the figure have very red colors and occupy the same part of the color–color plot as highly reddened normal stars and embedded protostars (see, for example, Lada & Adams 1992). The reddest of the three, source 8 in L1455, can be identified with the embedded Class II star known as L1455 IRS 5 (Figure 16, Jørgensen et al. 2006; Figure 7, Wu et al. 2007). Tapia et al. (1997) identified the same object as FIR 2 and from its near-infrared colors derived an optical extinction of A_v ≈ 30 mag. The 2MASS photometry plotted in Figure 1 confirms that earlier reddening estimate.

Zoom In Zoom Out Reset image size

The second reddest object in the infrared color–color diagram is L1455 source 10, which is the classical TTS RNO 15. The optical spectrum of RNO 15 shows variable Hα, O i, and the Ca⁺ infrared triplet strongly in emission (Cohen & Kuhi 1979; Cohen 1980; Levreault 1988). Br γ also appears in emission (Carr 1990). Following Equation (2) in Muzerolle et al. (1998), the luminosity of the Brackett line suggests a disk accretion rate of $\dot{M}_{\rm acc}$ ∼few × 10⁻⁷M_☉ yr⁻¹, which is quite typical for an active CTTS. The mid-infrared spectra of RNO 15 obtained by Spitzer display solid state ice features of H₂O and CO₂ in absorption (Boogert et al. 2008; Pontoppidan et al. 2008). Both features appear in much greater strength in the spectra of the adjacent submillimeter sources, IRS 1 and IRS 4, and are presumably formed by ice mantles that are attached to very cold dust grains in the cloud. The presence of these ice bands in the spectral scans of RNO 15 is thus a strong indication that this YSO is deeply buried within the L1455 cloud. The location of RNO 15 in the infrared two-color plot in Figure 1 suggests a minimum extinction of A_v ≈ 7 mag and leads to a similar conclusion.

There are two other stars of note in the color–color plot. The first is source 15 in L1455, which appears to be a moderately reddened (A_v = 2.2 mag) CTTS member of the cloud. Alternatively, if this object is a normal field star that is projected by chance on the line of sight to L1455, its 2MASS colors and magnitudes imply a spectral type of dM5, an extinction of A_v = 3.8 mag, and a distance of just 55 pc. At such a small distance, the star lies well in front of L1455 (d = 320 pc; Hatchell et al. 2007) but so close to the Sun that it is difficult to understand how it could suffer the nearly four magnitudes of extinction required in Figure 1. It is much more plausible to conclude that the star is a modestly reddened CTTS that belongs to L1455. The other notable object in Figure 1 is L723 source number 8. This source has a faint optical counterpart with a GSC red magnitude of R_F = 17.8. Based on its 2MASS colors, it is either a nearly unreddened M giant in the field or else a very modestly reddened TTS within L723. If it is a giant or a bright giant, the observed GSC and 2MASS K magnitudes together suggest a minimum distance of d = 16 kpc, in which case the star would be much too far away to be detected in X-rays since most late-type giants are intrinsically weak X-ray sources. On the other hand, the 2MASS colors of the source are almost identical to those of the marginal TTS KP2-44 in L1251B (Simon 2006, 2009), which is known to be the origin and driving source of the prominent HH 189 jet in that cloud.

The brightest source in the XMM observation of L1455, source 18, is at the very edge of the field of view of the EPIC cameras. It can be identified with a bright field star that is located at the outer boundary of the obscuration that defines the cloud. The star was observed to be slightly polarized in the optical linear polarization survey of the Perseus region by Goodman et al. (1990), where it appears as the 15th entry in their Table 3 (Per 15). The 2MASS and GSC colors of the star are compatible with little or no reddening, a spectral type of late F or early G, and a distance of ∼200 pc. The star is clearly not a pre-main-sequence member of the cloud. Two other sources with moderately large count rates, numbers 4 and 14, also lie at the southern edge of the XMM field of view but have no optical or infrared counterparts.

A near-infrared image obtained from the Spitzer public archives of the area surrounding RNO 15 is presented in Figure 2. The image was acquired at a wavelength of 4.5 μ with the IRAC camera in the course of the c2d Legacy program. Various objects of interest are numbered in the figure and identified by name in Table 4. Three XMM sources in the field are marked by plus signs. The location of RNO 15 is denoted by a cross symbol and labeled as object #2. The positions of RNO 15 and X-ray source 10 are obviously in close agreement, differing by less than 1''. The internal error in the X-ray centroid returned by the SAS edetect tool is ⩽05 at the 95% confidence level in both right ascension and declination. For comparison, the pn camera has a pixel size of 41 (Strüder et al. 2001) and the MOS cameras a pixel size of 11 (Turner et al. 2001). The position of RNO 15 corresponds to the centroid of emission in the IRAC 4.5 μ image. The near-infrared position falls between two 350 μ emission peaks in the map acquired by Wu et al. (2007) using the SHARC-II camera at the Caltech Submillimeter Observatory (CSO) on Mauna Kea, Hawaii, lying somewhat closer to the fainter western extension IRS 2W (i.e., object #3 in Figure 2 and Table 4) than to the main peak IRS 2E (object #4) shown in their map.

Zoom In Zoom Out Reset image size

Two other X-ray sources detected by XMM are plotted in Figure 2. X-ray source 8 (object #16 in Figure 2), a very weak detection, coincides with the heavily obscured Class II YSO IRS 5. IRS 5 is not known to be associated with any molecular outflow, radio jet, or HH structure. X-ray source 9 (object #7 in Figure 2), at best a marginal detection, lies between a pair of nonthermal background radio sources that are unrelated to the L1455 cloud (Schwartz et al. 1985; Meehan et al. 1998). Neither one of the embedded Class I/0 outflow sources in the cloud, IRS 1 or IRS 4, was detected in X-rays. The 3σ upper limit on the count rate signal at the locations of both YSOs in the EPIC pn image is 1.0 count ks⁻¹.

An X-ray light curve and X-ray spectrum were generated for RNO 15 from the filtered EPIC event files using a large circular extraction cell (r = 16'') and offset background region (r = 32''). The broadband count rate of RNO 15 was slightly elevated during the first 4 ks of the exposure (by ∼40%), but was extremely steady during the remainder of the observation, showing no sign of variability on timescales of 600 ks or longer. To analyze the X-ray spectra, I used the XSPEC spectral analysis package from the NASA High Energy Astrophysics Science Archive Research Center (HEASARC) to fit simple multitemperature, optically thin thermal plasma models to the pulse height distribution of the extracted photons. The pn and two MOS spectra were fitted jointly over the 0.9 keV to ∼8.0 keV energy range, varying the hydrogen column density (N_H), the source plasma temperature (kT) and metal abundance (Z), and the volume emission measure (EM). Parameters of the optimum model were determined by a χ² minimization technique. The spectra were binned to a minimum of 15 counts per spectral channel. The latest XMM response files were generated for each camera by means of the SAS arfgen and rmfgen tasks.

The best-fitting model is compared with the observed spectra in Figure 3. The numerical results of the calculation are presented in Table 5. An absorbed one-temperature MEKAL model (Mewe et al. 1995) produced the minimum χ² and hence the best fit. The temperature of kT = 4.2 keV in this isothermal model is much higher than is typical for a quiescent TTS of Class II (e.g., Getman et al. 2002). Moreover, for a normal gas-to-dust ratio, the hydrogen column density in the model corresponds to an extinction at visible wavelengths of A_v ≈ 23 mag (see Bohlin et al. 1978), which is three times the amount suggested by the near-infrared color indices of this star in the color–color plot in Figure 1. A χ² contour diagram, plotting values of kT against N_H (see Figure 4), however, demonstrates that the plasma temperature and column density of the best-fitting model are well constrained by the observation. In addition, no XSPEC model was able to fit the observed spectra with a lower assumed value for the column density; all such models that I tried led to a considerably poorer fit at energies below 2 keV and a much larger value of the χ². Two-temperature models also offered no significant improvement in the fit. As for the EM and intrinsic X-ray luminosity of RNO 15, the values derived here are at the high end of the range for an active, young TTS, but are by no means exceptional (e.g., Imanishi et al. 2003; Gagné et al. 2004).

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

The remaining X-ray sources listed in Table 2 are much fainter than RNO 15, consequently none of them could be modeled in the same fashion as that star. I obtained a crude estimate for the X-ray luminosity of the other embedded YSO of obvious interest, XMM source 8, by matching its observed count rate with a one-component MEKAL model. I assumed a value of 2 keV for the unknown plasma temperature and adopted a N_H value appropriate to the extinction of A_v = 30 mag that was derived earlier from the near-infrared colors of its infrared counterpart IRS 5. Given those assumed values, the observed count rate yields an extremely tentative estimate of L_X ∼ 10^30.4 erg s⁻¹ for the X-ray luminosity.

The center of the XMM field of view is shown in the IRAC 4.5 μ image of L723 in Figure 5. Table 4 provides a key to the symbols plotted there. The cross labeled as object #1 corresponds to X-ray source 8. The uncertainty in the right ascension and declination of the X-ray centroid determined by the SAS edetect task for that source is ±25 at the 95% confidence level. The nominal position of IRAS 19156+1906 lies 2'' east of the X-ray position and has an uncertainty in right ascension of ±12'' (Moshir et al. 1990). Hence, the X-ray centroid falls within the error ellipse of the IRAS source. The two objects are the likely shorter and longer wavelength counterparts, respectively, of the faint Hubble Space Telescope (HST) guide star and its 2MASS and anonymous near-infrared complement that is found at the same location in the IRAC image.

Zoom In Zoom Out Reset image size

Virtually nothing is known about the optical guide star according to the SIMBAD database. As I noted in Section 2.3, the 2MASS colors of the star are compatible with those of an unreddened or lightly reddened normal field giant. However, if that were so, the 2MASS magnitudes of the star would place it at a distance of >16 kpc and consequently too far away to have been detected as a background X-ray source. The more reasonable alternative is that the star lies within the L723 cloud and is a marginal or weak-line TTS with a very small infrared excess and at most a slight amount of reddening. The X-ray source itself is too faint to be reliably modeled using XSPEC. A rough guess at the X-ray luminosity of the star can be deduced from the observed count rate using a very simple one-component MEKAL model. Assuming a nominal extinction of A_v = 2 mag, a plasma temperature kT in the range of 1–4 keV, and a distance of 300 pc for L723 (Goldsmith et al. 1984), I derive a luminosity of L_X ≈ 10^29.3 erg s⁻¹. Although highly uncertain, that L_X is very typical of the X-ray luminosities of weak-line TTSs (e.g., Neuhäuser 1997). Nearly the same value, L_X ≈ 10^29.36 erg s⁻¹, was found for KP2-44, the borderline TTS that powers the HH 189 jet in L1251 (Simon 2006), and which shares the same infrared colors as the source in L723.

Several other noteworthy objects in the XMM field of view toward L723 deserve mention here. VLA 2 is the double radio source whose binary components have been proposed as the separate origin and driving source of the two high-velocity flows in L723 (Carrasco-González et al. 2008). No X-ray emission was detected at their location. The 3σ upper limit on possible emission at the radio position in the pn camera is 1.84 count ks⁻¹. VLA 1, a second radio source located 7'' south of the X-ray source, was discovered at the same time as VLA 2. It was initially proposed as the possible launch site for the narrow north–south flow in L723 (Anglada et al. 1991), but was subsequently dismissed as a likely background source (Anglada et al. 1996). No X-ray emission was observed at the position of VLA 1. Source number 1 (XMMU J191651.1+191520), the brightest X-ray source detected by XMM in the L723 exposure, was outside the MOS field of view and very close to the outer edge of the pn image. The X-ray SED and 2MASS photometry are consistent with a one-component MEKAL model in which A_v = 2.8 mag and kT = 1.4 keV, the distance is d = 125 pc, and the spectral type is mid-G. I infer a source luminosity of L_X = 10^29.92 erg s⁻¹, which suggests a very active field star, possibly of the RS CVn variety.