In the simplest model, the diffuse 1/4 keV background at high galactic latitude is a combination of emission from three components: the local interstellar medium (LISM), the galactic halo, and the extragalactic background. Separation of the LISM contribution from the distant ones can be accomplished via the detection of intensity variations in the 1/4 keV background caused by absorption by intervening neutral material. ROSAT PSPC observations in the direction of high galactic latitude enhancements in IRAS 100 mm emission, identified by Désert, Bazell, & Boulanger (1988, ApJ, 334, 815) as likely molecular cirrus clouds, provide good candidates for seeing these shadows. For several such fields at both high northern and southern galactic latitudes, I have generated maps of the intensity in the 1/4 keV band following the procedure suggested by Snowden et al. (1994, ApJ, 424, 714) and fit for an anticorrelation with the 100 mm intensity. The derived average brightness of the galactic halo in both the north and the south is ~ 1.0×10^-3 counts s^-1 arcmin ^-2. Assuming coronal equilibrium at 10⁶ K and solar abundances, this intensity would be produced by an emission measure of ~ 5.3×10^-3 cm^-6 pc. Viewed face-on this emission measure results in a 1/4 keV band luminosity per unit area of ~ 1.8×10³⁷ ergs s^-1 kpc ^-2.

1  Introduction

The existence of a hot galactic halo was suggested as a mechanism to confine clouds at high galactic scale height[16] long before the capability to directly observe it existed. With the first observation of the 1/4 keV background[2] questions arose as to how much of the observed intensity might be due to a galactic halo; various combinations of local, distant, and distributed emission have been proposed. Observations of edge-on galaxies[3,10,19] have demonstrated that some have extended hot halos; while observations of face-on spiral galaxies[5] provide upper limits to emission from their halos.

Prior to the launch of ROSAT the simplest explanation[11] that fit the available data for our galaxy was that the 1/4 keV background was predominantly local in origin (i.e. that it arose in a local, hot, low-density cavity, in front of most of the neutral material - the Local Bubble). Observations with the PSPC on ROSAT have provided the first definitive evidence for emission from outside the local interstellar medium[4,12,14,18]. The detection of shadows in the 1/4 keV background allows us to divide this background into its individual components.

2  The Clouds

Ideal targets for shadowing observations with the ROSAT PSPC should have low total column density ( £ 2×10²⁰ cm^-2) but have significant variation in column density on angular scales of ~ 1^° so that the PSPC can image any shadow. Désert, Bazell, & Boulanger (1988 hereafter DBB) compiled a list of clouds based on excess IRAS 100 micron emission relative to HI column density, which provides an ideal starting point to search for appropriate shadowing candidates. Additional cloud candidates can be found in other papers that have used the IRAS 100 micron data to find candidates for molecular gas at high latitude[7]. I have searched the ROSAT public archive for PSPC observations in the direction of clouds from the DBB survey with |b|  > 60^°.

Tables 1 and 2 list clouds for which suitable PSPC observations were found; the fields are labeled with their number from DBB where applicable. Also given are the galactic longitude and latitude of the center of the observed field, the HI column density in this direction, the average IRAS 100 micron intensity and its range, and the name of the ROSAT observation. Not all of the clouds meet the ideal candidate criteria listed above; some have fairly large total HI column density. Figure 1 shows the ROSAT all-sky survey maps of the 1/4 keV background toward the galactic poles[15]; the locations of the observed fields are indicated.

^a HI in units of 10²⁰ cm^-2 from Stark et al. 1992
^b Average IRAS 100 mm intensity in units of MJ  sr^-1
c 1/4 keV band ate in units of 10^-6 counts s^-1 arcmin^-2

Figure 1: ROSAT all-sky survey images of the 1/4 keV band for the north and south galactic poles. The directions of the fields toward the clouds considered in this poster are indicated.

3  ROSAT Data Reduction

Observations of extended emission with the ROSAT PSPC require special procedures to remove background contamination and to properly determine the exposure across the detector. I follow the procedures described in Snowden et al. (1994a). The 1/4 keV band I used consists of the R1L and R2 pulse height bands[13]. Large variations in the background level for some of the observations required the elimination of large fractions of the total observed time from analysis. Maps of the counts, exposure, and modeled background were made with ~ 15 arcsec pixels for each of the fields. Point sources were searched for by sliding a circular source region and concentric annular background region over the maps; the diameters of the circle and annulus were adjusted to account for the variation in the telescope point spread function with off-axis angle. Areas where counts within the source region were inconsistent with the expected background at the 3-s level were masked out during subsequent analysis as sources. Regions in which the exposure was less than 10% of the average exposure over the central 40 arcmin diameter of the PSPC were also excluded from analysis.

Tables 1 and 2 include the average 1/4 keV intensity and an estimate of the depth of any shadow that may be present for each of the observed fields. Several fields do not show any evidence for a shadow. Figure 2 shows one of the better examples of shadowed 1/4 keV emission (cloud 306 or G211+63). Both the 1/4 keV and the 100 micron maps have been smoothed with a gaussian with a ~ 1 arcmin sigma for display purposes. The color image is the 1/4 keV intensity; the contours are 100 micron intensity.

Figure 2: Map of the 1/4 keV emission toward one of the better examples of shadowing, cloud 306 (G211+63); color indicates brightness (blue - faint / red - bright). Contours show IRAS 100 micron emission and are spaced every 0.5 MJy sr-1.

4  Comparison to IRAS 100 mm emission

Photoelectric absorption by neutral gas associated with the dust responsible for the IRAS 100  mm emission causes an anticorrelation between the 1/4 keV and 100 mm intensities. Photoelectric absorption in the 1/4 keV band is predominantly due to H atoms (from both HI and H₂) and HeI in the neutral gas. A simple model for the expected 1/4 keV band counting rates is shown schematically in figure 3. The model has two x-ray emitting components: a local uniform component across the target field and partially absorbed emission from a uniform distant component located beyond all of the absorbing gas.

Figure 3: Cartoon of the absorption geometry.

In order to obtain reasonable statistical uncertainties on the 1/4 keV band rates, the counts, exposure, and background maps were binned to ~ 8 arcmin cells and rates for each cell were calculated. The average 100 mm intensity for each cell was also calculated. The observed rates for each field were fit with a model of the form:

Figure 4: Example of the fit to the absorption of the 1/4 keV emission by neutral material traced by 100 micron emission. The field shown in this fit is the same one displayed in figure 2.

5  Results & Discussion

Tables 3 and 4 list the results of the fits described in the previous section. In many of the fields the absence of a strong shadow result in an upper limit on the brightness of the component beyond absorbing material or strongly coupled derived brightnesses of the two x-ray emitting components. A comparison of the average rates with the fit local rates indicates that most of the observed intensity has an origin in front of the absorbing material (the Local Bubble). To truly say that the fit values R_D reflect a halo component we must: know that the absorbing material causing the shadow is within the ``disk'' of the galaxy and assess the contribution from any extragalactic component.

The distances to these clouds are not known. Interstellar absorption line measurements toward several stars at a variety of distances in the direction of each of the clouds could be used limit the distance to the cloud; without such distance limits we must resort to other determinations. Most of the clouds in this data set for which velocities have been determined are at low velocity. Statistical arguments have been made[8] that assign these clouds to the edges of the Local Bubble.

The extragalactic intensity at 1/4 keV is unknown but observations searching for shadows caused by external galaxies[1,6,9] have provided both upper ( 62 keV cm^-2 s^-1 sr^-1 keV^-1) and lower ( 32 keV cm^-2 s^-1 sr^-1 keV^-1) limits on the extragalactic 1/4 keV surface brightness. If we use the average of these two values as an estimate of the surface brightness of the extragalactic component and convert to rates, the extragalactic contribution is ~ 580×10^-6 counts s^-1 arcmin^-2. This value includes a correction for absorption by neutral material within the extended warm, partially ionized gas traced by Ha emission; this correction was not included in the fits that I performed and so for comparison the above rate should be reduced to ~ 390×10^-6 counts s^-1 arcmin^-2.

The variation in the values of R_D suggest that the region causing the halo emission does not extend far from the disk of the galaxy. It may be reasonable to assume that much of the hot halo gas lies closer than the gas responsible for the Ha emission. If this is the case the rates for the halo emission do not require a correction for absorption by that extended, warm material.

In the north the average value of all the points after removing a contribution from extragalactic emission is ~ 1000×10^-6 counts s^-1 arcmin^-2. A plasma in collisional equilibrium at 10⁶ K will produce this rate at an emission measure of ~ 5.3×10^-3 cm^-6 pc. While it is unlikely that the emission is in collisional equilibrium, this model for the emission provides a convenient reference for comparison with other results. A 10⁶ K collisional plasma has a power per unit emission integral of 1.13×10^-22 erg cm³ s^-1 so an emission measure of 5.3×10^-3 cm^-6 pc implies an average surface brightness of the local galactic halo of ~ 1.8×10³⁷ ergs s^-1 kpc^-2. Two of the lines of sight (clouds 8 and 12) are in the direction of the North Polar Spur (NPS), a region of enhanced emission associated with the Sco-Cen OB association. This nearby ``bubble'' is large enough that it may have broken out into the lower portion of the galactic halo and could be a source of hot gas into the halo. If the two points toward the NPS are ignored the average drops to ~ 760×10^-6 counts s^-1 arcmin^-2.

In the south the average value of the points that are not just upper limits is ~ 925×10^-6 counts s^-1 arcmin^-2, after removing a contribution from the extragalactic emission. This value is about the same as in the north but since the upper limits are not included and most of the upper limits are less than this value even before accounting for an extragalactic contribution it is clear the local halo in the south is not as bright as in the north. Calculating an average in the south using all of the points and using half the value of the upper limits as a value for R_D results in a halo rate of ~ 500×10^-6 counts s^-1 arcmin^-2.

References

[1]

Barber, C. R., & Warwick, R. S. 1994, MNRAS, 267, 270.

[2]

Bowyer, C. S., Field, G. B., & Mack, J. E. 1968, Nature, 217, 32.

[3]

Bregman, J. N. & Pildis, R. A. 1994, ApJ, 420, 570.

[4]

Burrows, D. N. & Mendenhall, J. A. 1991, Nature, 351, 629.

[5]

Cui, W., Sanders, W. T., McCammon, D., Snowden, S. L., & Womble, D. S. 1996, ApJ, 468, 102.

[6]

Cui, W., Sanders, W. T., McCammon, D., Snowden, S. L., & Womble, D. S. 1996, ApJ, 468, 117.

[7]

Heiles, C., Reach, W. T., & Koo, B.-C. 1988, ApJ, 332, 313.

[8]

Magnani, L., Hartmann, D., & Speck, B. G. 1996, ApJS, 106, 447.

[9]

McCammon, D. & Sanders, W. T. 1990, ARA&A, 28, 657.

[10]

Pietsch, W., Vogler, A., Kahabka, P., Jain, A. & Klein, U. 1994, A&A, 284, 386.

[11]

Snowden, S. L., Cox, D. P., McCammon, D., & Sanders, W. T. 1990, ApJ, 354, 211.

[12]

Snowden, S. L., Mebold, U., Hirth, W., Herbstmeier, U., & Schmitt, J. H. M. M. 1991, Science, 252, 1529.

[13]

Snowden, S. L., McCammon, D. Burrows, D. N., & Mendenhall, J. A. 1994a, ApJ, 424, 714.

[14]

Snowden, S. L., Hasinger, G., Jahoda, K., Lockman, F. J., McCammon, D., & Sanders, W. T. 1994b, ApJ, 430, 601.

[15]

Snowden, S. L., Egger, R., Freyberg, M. J., McCammon, D., Plucinsky, P. P., Sanders, W. T., Schmitt, J. H. M. M., Truemper, J., Voges, W. 1997, ApJ 485, 125.

[16]

Spitzer, L. 1956, ApJ, 124, 20.

[17]

Stark, A. A., Gammie, C. F., Wilson, R. W., Bally, J., Linke, R. A., Heiles, C., & Hurwitz, M. 1992, ApJS, 78, 77.

[18]

Wang, Q. D., & Yu, K. C. 1995, AJ, 109, 698.

[19]

Wang, Q. D., Waterbos, R. A. M., Streakley, M. F., Norman, C. A. & Braun, R. 1995, ApJ, 439, 176.