Properties of the Diffuse X-ray Background toward MBM20 with Suzaku
A. Gupta, M. Galeazzi, D. Koutroumpa, R. Smith, R. Lallement
PProperties of the Diffuse X-ray Background toward MBM20 with
Suzaku
A. Gupta and M. Galeazzi
Physics Department, University of Miami, Coral Gables, FL 33124 [email protected]
D. Koutroumpa
NASA/GSFC, Code 662, Greenbelt, MD 20771
R. Smith
Harvard Smithsonian Center for Astrophysics, Cambridge, MA 02138
R. Lallement
CNRS Service d’Aronomie, FR 91371 Verrieres Buisson, France
ABSTRACT
We used
Suzaku observations of the molecular cloud MBM20 and a low neu-tral hydrogen column density region nearby to separate and characterize theforeground and background diffuse X-ray emission. A comparison with a pre-vious observation of the same regions with
XMM-Newton indicates a significantchange in the foreground flux which is attributed to Solar Wind Charge eXchange(SWCX). The data have also been compared with previous results from similar“shadow” experiments and with a SWCX model to characterize its O
VII andO
VIII emission.
Subject headings:
X-rays: diffuse background
1. Introduction
Our current interpretation of the diffuse X-ray emission below 1 keV includes a combina-tion of 5 components: Solar Wind Charge eXchange (SWCX), Local Bubble (LB), GalacticHalo (GH), Warm-Hot Intergalactic Medium (WHIM), and unresolved point sources (e.g.,(Gupta & Galeazzi 2009), (Galeazzi et al. 2009)). Resolving the different components is a r X i v : . [ a s t r o - ph . GA ] O c t Suzaku
X-ray observatory. The XIS is an excellenttool for studying the SXRB, due to its low and stable non-X-ray background and goodspectral resolution. The targets are identical to those observed in Galeazzi et al. (2007)using
XMM-Newton .To build a consistent picture of the diffuse X-ray background we compared the
Suzaku observation with the previous
XMM-Newton result and with similar shadow observations inthe direction of the high latitude molecular cloud MBM12 (Smith et al. 2005, 2006) and afilament in the southern Galactic hemisphere (Henley et al. 2007; Henley & Shelton 2008).We also used the model recently developed by Koutrompa et al. (2007) to estimate theemission from SWCX. The model is time dependent and includes factors such as solar cyclephase, the observation position, and the line of sight. 3 –The data reduction is discussed in § §
3. Section4 compares our result with the previous
XMM-Newton observation generally with otherrecent shadow experiments, and § Suzaku and
XMM-Newton observations.
2. Observations of MBM20 and Eridanus Hole
MBM20 and the Eridanus Hole were observed with
Suzaku in February 2008 and July2007 respectively. Notice that the temporal gap between the observations is large comparedto the typical time variation of the SWCX and will be discussed in section 5.The details of theobservations are reported in Table 1. MBM20 is a high-density, high-latitude star-formingcloud located at or within the edge of the Local Bubble (Galeazzi et al. 2007). Its mass is84M (cid:12) and it is located at coordinates l = ◦ (cid:48) (cid:48)(cid:48) . , b = − ◦ (cid:48) (cid:48)(cid:48) . , southwest ofthe Orion star forming complex. Based on interstellar NaI D absorption lines the distance ofMBM20 is evaluated between 112 ±
15 pc and 161 ±
21 pc (Hearty et al. 2000). The Eridanushole, at coordinates l = ◦ (cid:48) (cid:48)(cid:48) . , b = − ◦ (cid:48) (cid:48)(cid:48) . , is a region of low neutral hydrogencolumn density located about 2 degrees from the highest-density part of MBM20 (Fig. 1). We used the
Suzaku data reprocessed to version 2.0 and the analysis was performed withHEAsoft version 6.4 and XSPEC 12.4.0. We started the event screening from the cleanedevent file, in which selection of the event grade and bad CCD column, and removal of hotand flickering pixels by the “cleansis” ftool, were already conducted ( Suzaku
Data ReductionGuide ).In our analysis, we use only data from the XIS1 detector, as this has the greatestsensitivity at low energies. We combined the data taken in the 33 and 55 observation mode.For that, first we convert the 5 × × × × Suzaku passing through the South AtlanticAnomaly (SAA), and when
Suzaku ’s line of sight is elevated above the Earths limb by lessthan 5 ◦ , or is less than 20 ◦ from the bright-Earth terminator. We decided to expand this to See http://heasarc.nasa.gov/lheasoft/ See http://suzaku.gsfc.nasa.gov/docs/suzaku/analysis/abc/abc.html/ ◦ , as there are some excess eventsin the 0.5-0.6 keV band in the 5 ◦ − ◦ range (Smith et al. 2006).Due to Suzaku ’s broad point spread function (half-power diameter ∼ XMM-Newton observations (see Fig. 2). We extracted spectra from the full XIS1 field of view, after re-moving above-mentioned point sources and the corners of the detectors which contained theonboard Fe-55 calibration sources.
Suzaku is in a low-Earth orbit, so it is significantly shielded from the particle back-ground that strongly affects
XMM-Newton and
Chandra . The effectiveness of this shieldingis dependent upon the “cut-off rigidity” (COR) of the Earth’s magnetic field, which varies as
Suzaku traverses its orbit. During times with larger COR values, fewer particles are able topenetrate to the satellite and to the XIS detectors. We excluded times when the COR wasless than 8 GV, which is higher than the default value (COR 4 GV) for both observations,as the lowest background was desired.Although it is reduced by the Earth’s magnetic field,
Suzaku still has a noticeable particlebackground. We can estimate the appropriate particle background from a database of thenight Earth data (NXB). NXB was collected when the telescope was pointed at the nightEarth (elevation less than -5 degree, and pointed at night side rather than day). The eventfiles in the database have been carefully screened for telemetry saturation and other artifacts.We constructed the spectra of the night earth data using Ftool “xisnxbgen” (Tawa et al.2008), which sorts the NXB data by COR values, generates an NXB spectrum and image foreach COR range, and combines them weighted by exposure time ratio of each COR rangeduring GTIs in our spectral data file. The background spectra were then subtracted fromthe corresponding source spectra.
We calculated the XIS detector effective area using the tool “xissimarfgen” (Ishisaki etal. 2007). This tool takes into account the spatially varying contamination on the opticalblocking filters of the XIS sensors which reduces the detector efficiency at low energies(Koyama et al. 2007). For the ancillary response file (ARF) calculations we assumed a 5 –uniform source of radius 20’ and used a detector mask which removed the bad pixel regions.To generate the redistribution matrix file (RMF), we used the ftool “xisrmfgen”.
3. Analysis
We first fit a model to our spectra consisting of 3 components: a Local Bubble compo-nent, modeled as an unabsorbed plasma with thermal emission in collisional ionization equi-librium (CIE); a hotter Galactic halo emission, modeled as an equilibrium thermal plasmacomponent absorbed by the gas in the Galactic disk; and an unresolved extragalactic sourcecomponent, modeled with an absorbed power law. This is the same model used in Galeazziet al. (2007). As extensively discussed in §§ wabs model, which uses cross-sections from Wisconsin (Morrison & McCammon, 1983) and usesthe Anders & Ebihara (1982) relative abundances. We fit the above mentioned model to the Suzaku spectra of MBM20 and the Eridanus Hole. As in Galeazzi et al. (2007), we used the
IRAS µ m maps to evaluate the neutral hydrogen density in the two regions. The IRAS average brightness is 13.34 MJy sr − , and 0.73 MJy sr − for MBM20, and the Eridanus holerespectively. Using the “typical” high-latitude 100 µ m/NH ratio of 0 . × − cm MJy sr − (Boulanger & Perault 1988) the estimated neutral hydrogen densities are 1 . × cm − ,and 0.86 × cm − respectively. The fits are shown in Fig. 3, along with the best-fittingmulticomponent spectral model. The model parameters are reported in Table 2.We also tried to fit the above mention model simultaneously to our MBM20 and EridanusHole Suzaku spectra with a single set of parameters, except for the neutral hydrogen columndensity. The fits are shown in Fig. 4, and the model parameters are presented in Table 2.To extend the analysis further, we also included data from the
ROSAT
ALL-Sky Survey(RASS) in the same directions. We extracted RASS data in the
ROSAT bands R1-R7(Snowden et al. 1998) and scaled them to the same field of view as our
Suzaku data setsfor both MBM20 and the Eridanus Hole. We then performed a global fit of the four datasets simultaneously with a single set of parameters. The fit results are reported in Table 2,and the data are shown in Fig. 5. Overall, the model gives a good fit to the data (reduced 6 – χ = 0 .
87 for 352 degrees of freedom), however, the fit to some of the
ROSAT bands is ratherpoor.We used our fit results to obtain O
VII and O
VIII intensities, since at temperatures ofmillion Kelvins, O
VII and O
VIII lines are the dominant features. In our
Suzaku spectra ofMBM20 and Eridanus Hole, the blended O
VII triplet at 561, 569 and 574 eV is clearly visiblein both observations, while the O
VIII line at 654 eV is barely visible in the MBM20 dataset and lies within the statistical uncertainty in the Eridanus Hole data set. The O
VII andO
VIII line intensities are 2 . ± . − cm − sr − (line units, LU, from now on)and 0 . ± .
48 LU for MBM20, and 5 . ± .
04 LU and 1 . ± .
79 LU for the Eridanushole respectively. Following the same recipe used in Galeazzi et al. (2007), we can evaluateO
VII and O
VIII emission of the foreground (LB+SWCX) and background (GH) components.Using the expression for cross section per hydrogen atom for a cosmic abundance plasmaderived by Morrison & McCammon (1983), we find that MBM20 absorbs about 75% ofthe background O
VII emission and about 61% of the background O
VIII emission, while theEridanus Hole absorbs about 8% of the background O
VII emission and about about 5% ofthe background O
VIII emission. Combining these data with the result of our observations weobtain, for O
VII and O
VIII respectively, 0 . ± .
91 LU and 0 . ± .
01 LU for the foregroundand 5 . ± .
79 LU and 1 . ± .
74 LU for the background.We evaluated the electron density and thermal pressure of the GH and the LB, using thesame procedure discussed in Galeazzi et al. (2007). Assuming the foreground componentis due solely to LB emission, we obtain lower and upper limits for the plasma density of0.015 and 0.018 cm − K and limits of 23,500 and 28,800 cm − K for the plasma pressure.Similarly, assuming that the absorbed plasma component is due solely to GH emission, weobtain a plasma density ranging from 0.0005 to 0.0014 cm − and a pressure between 3.3 × and 5.8 × cm − K .We also used the non-equilibrium plasma model GNEI (Borkowski et al. 2001), a non-equilibrium model characterized by a constant postsock electron temperature and by itsionization age, to fit our data. While we obtained a good fit, similar to that shown in Fig. 5,and an electron density in the range 0.013-0.158 cm − , we derived a value for the age of theLB of ≤
4. Comparing
Suzaku and
XMM-Newton
Observations of the Soft X-rayBackground.
The temperature and emission measures we obtained from the
Suzaku data are signifi-cantly different from those determined from the
XMM-Newton analysis in the same pointingdirections. For a visual estimate of the difference, we folded our
Suzaku model through the
XMM-Newton response and compared it with the
XMM-Newton spectra (see Fig. 6). Thedifference in these spectra would be consistent with a time dependent component of theforeground emission, attributable to SWCX, which we will discuss in detail in the next fewsections. The excess is clearly significant in both data sets.So far only a few targets with the proper characteristics for shadow experiments havebeen observed with any of the three major X-ray satellites (
Chandra , XMM-Newton , and
Suzaku ). In addition to the MBM20 observations discussed, we point out the observations ofthe neutral hydrogen cloud MBM12 performed with
Chandra (Smith et al. 2005) and
Suzaku (Smith et al. 2006) and that of a relatively dense neutral hydrogen filament in the southerngalactic hemisphere (Henley & Shelton 2008).Table 3 summarizes the O
VII and O
VIII flux for all the available observations. Datafrom McCammon et al. 2002 are also reported for comparison. In McCammon et al. a highresolution measurement over a 1 sr field of view near the north Galactic pole was performedusing cryogenic microcalorimeters mounted on a sounding rocket. Tables 4 and 5 give asummary dividing the results in foreground and background emission. Where a fit with aplasma model has been performed, the best fit parameters for temperature and emissionmeasure are also reported.While the amount of available data is limited, we identified a few general trends thatwe want to point out: • Each target has been observed at least twice in the past 8 years, but the results frommultiple observations of the same target do not agree, at various levels, with eachother. This is evidence of a significant contribution from SWCX, the only componentof the diffuse X-ray background that should change with time on such a short timescale. Moreover, when we separate foreground and background oxygen line emission,the component that changes with time seems to be the foreground one, while thebackground does not change, within the errors, between different observations of thesame target, strengthening the notion that the variation is due to SWCX. We want topoint out, however, that multiple observations of the same target have been performedwith different satellites, i.e., different data reduction analysis, background subtractionschemes, etc., which have different systematic uncertainties. 8 – • The change in oxygen line emission between different observations of the same targetcan be used to estimate the typical flux variation of the SWCX emission. The O
VII emission varies between 1 . ± .
61 LU and 4 . ± .
90 LU, while the O
VIII emissionaries between 0 . ± .
72 LU and 2 . ± .
37 LU. The detailed results are reported inTable 6. • High resolution investigations of the diffuse X-ray emission have shown that a simpleone-temperature plasma in equilibrium cannot explain the observed spectra (McCam-mon et al. 2002; Sanders et al. 2001). However, while CCD detectors are a significantstep forward from proportional counters, their resolution is still quite limited and insuf-ficient to investigate the issue. Equilibrium plasma models seem to be still sufficient tofit the spectra and, while there have been attempts at using more sophisticated mod-els, it is impossible to distinguish between them. At this point the available data areadequately fit with a plasma thermal emission from the LB, with temperature around1 million degrees, and either one or two temperature thermal plasma components forthe Galactic halo, with temperatures between 2 and 3 million degrees. • Except for the
Chandra observation of MBM12, with its very unusual O
VIII emission,all other observations seem to indicate that the foreground O
VIII emission is eithervery small or compatible with 0. Typical LB models do not predict significant O
VIII emission and this seems to indicate that the SWCX component does not normally haveany significant emission in O
VIII either. • When the assumption is made that all the foreground emission is due to LB emission,the derived values for the plasma temperature, density, and pressure seem to be ingood agreement with the predictions from the most commonly accepted models of theorigin and structure of the LB (e.g., Smith and Cox 2001).
5. SWCX Model to Data comparison
The heliospheric SWCX model we use for our simulations is extensively described inKoutroumpa et al. (2006, 2007). This model is a self-consistent calculation of the solar windcharge-exchange X-ray line emission for any line of sight (LOS) through the heliosphere andfor any observation date, based on 3-dimensional grids of the inter-stellar (IS) neutral species(H and He) distributions in the heliosphere modulated by solar activity conditions (grav-ity, radiation pressure, and ionization processes which are anisotropic due to the latitudinalanisotropy of the solar wind mass flux and solar radiation). Highly charged heavy solarwind (SW) ions are propagated radially through these grids and the charge-transfer collision 9 –rates are calculated for each of the ion species, including the evolution of their density dueto charge-transfer with the IS atoms. With this process, we establish 3-dimensional emis-sivity grids for each SW ion species, using photon emission yields computed by Kharchenko& Dalgarno (2000) for each spectral line following charge exchange with the correspondingneutral species (H and He individually). Finally, the X-ray line emission is integrated alongany LOS and observation geometry (for each observation date) in order to build the com-plete spectrum of SWCX emission in the given direction. For comparison to present X-rayobservations we use the O
VII triplet at 0.57 keV and the O
VIII line at 0.65 keV, as they arethe strongest spectral features and provide the best signal-to-noise ratio for the observations.We have conducted simulations for each of the MBM20 and Eridanus Hole observationsaccounting as close as possible for average solar activity conditions corresponding to theobservation period. Solar activity is reflected both in the IS neutral distributions (by meansof ionization rates that are increased and less anisotropic in solar maximum), and in the solarwind ionic composition (abundances and charge state distributions) and spatial distribution.Details on the solar activity effect on the neutral H and He distributions, along with thelatitude-dependant ionization rates used in the model for maximum (e.g. 2001), intermediate(e.g. 2003-2004) and minimum (e.g. 2007-2008) solar conditions are given in Koutroumpaet al. (2009).The latitude dependence of the solar wind also affects the highly charged heavy iondistribution, where abundances depend on the solar wind type. During minimum solaractivity, the solar wind is considered to be highly anisotropic, with a narrow equatorial zone(within ± ◦ of the solar equatorial plane) of slow solar wind with an average speed of ∼ km s − and the fast solar wind emitted from the polar coronal holes at a speed of ∼ km s − . The slow solar wind has a proton density of ∼ cm at 1 AU, while the fast flowis less dense at ∼ cm at 1 AU. At solar maximum, the solar wind spatial distribution isconsidered to be a complex mix of slow and fast wind states that is in general approximatedwith an average slow wind flux. The ionic composition of the two flows can be very differentwith the average oxygen content varying from [O/H] = 1/1780 in the slow wind and [O/H]= 1/1550 in the fast flow. The charge-state distributions change as well, with the highercharge-states strongly depleted (or even completely absent, as for example O +8 in the fastsolar wind. For our model we adopt the oxygen relative abundances published in Schwadron& Cravens (2000): ( O +7 , O +8 )=(0.2, 0.07) for the slow wind and ( O +7 , O +8 )=(0.07, 0.0) forthe fast wind, based on data from the Ulysses SWICS instrument.The XMM-Newton observations of MBM20 and the Eridanus Hole were performed dur-ing 2004, which corresponds to intermediate solar conditions, while for the
Suzaku observa-tions, performed in 2007-2008, solar minimum conditions are most appropriate. The main 10 –difference in the SW heavy ion distribution between the two periods (two sets of coupledobservations) is the spatial (latitudinal) distribution of the slow and fast solar wind flows.For the
Suzaku simulations (solar minimum) the slow SW is expanding in interplanetaryspace through a ± ◦ equatorial zone on the solar surface, while the fast SW flow occupiesthe rest of the space. For the intermediate 2003-2004 period ( XMM-Newton observations)we assume that there is no fast wind flow in interplanetary space (same approach as for solarmaximum), in order to estimate the quiescent (outside potential coronal mass ejection orsolar flare) upper limit for the resulting SWCX X-ray emission. Indeed, as demonstratedin Koutroumpa et al. (2006, 2007), for high ecliptic-latitude LOS, as is the case for theMBM20 and Eridanus Hole observations (Declination ∼ − ◦ ), the oxygen line intensity de-creases from solar maximum to solar minimum conditions as the LOS crosses larger fastwind regions where the parent ions are strongly depleted.In Table 7 we summarize the SWCX model results for the oxygen line intensities forthe four observations. As expected, model A, which assumes average solar wind conditionsas described above, predicts a significant decrease in the SWCX oxygen line intensities aswe progress from near solar maximum ( XMM-Newton ) to solar minimum (
Suzaku ), sincewe are observing at high southern ecliptic latitudes. Also, the model predicts a decrease inthe heliospheric SWCX emission when shifting the view direction from the Eridanus Hole(off-cloud) to the MBM20 (on-cloud) direction. This decrease is of the order of 8% for the
XMM-Newton observations and of the order of 30% up to 45% (for O
VII ) for the
Suzaku observations. Such a large difference can be explained by the large interval separating the two
Suzaku observations combined with the inclination of the equatorial SW zone with respectto the ecliptic plane (due to the 7 . ◦ inclination of the solar axis with respect to the eclipticaxis). Indeed, the Suzaku observations of MBM20 and Eridanus Hole were performed at anobserved ecliptic longitude of 142 ◦ and 306 ◦ , respectively, separated by six months, while theLOS was pointing at ∼ ◦ south. Since the solar equator ascending node is Ω = 73 . ◦ , theEridanus Hole LOS was looking through a larger region of the oxygen-rich slow solar windequatorial zone than the MBM20 LOS.One step to further improve the accuracy of our prediction is to apply reasonable as-sumptions to the SWCX simulations. First, evidence from the Ulysses/SWICS O +7 /O +6 ratio data (which is a proxy for the flow speed/type) during the 2007 (minimum) crossingof the equatorial slow wind zone (Fig. 7) shows that this later was in fact more extended inlatitude (more than ± ◦ than during the previous solar minimum (1996) that served as areference for the minimum SW conditions applied in the SWCX model. In order to investi-gate the effect of such a possibility we performed a second simulation of the Suzaku
EridanusHole/MBM20 observations introducing a ± ◦ slow SW zone as input. The results are alsonoted in Table 7 (Model B). However, in situ measurements with ACE at the L1 point show 11 –unusually low O +7 abundances for the Suzaku observations period, almost an order of mag-nitude lower than the average slow wind conditions (11.5% during the MBM20 observationand 20% during the Eridanus Hole one). For solar maximum (
XMM-Newton observations),the O +7 abundance in the ACE data does not show significant deviation from average slowwind values. O +8 measurements are too sparse to allow a significant quantitative analysisof the data, and therefore we will make no assumption for these data. If we apply 11.5%and 20% correction factors to the Suzaku ’s MBM20 and Eridanus Hole O
VII line intensitiespredicted from Model B, we obtain the values noted as Model B1 in Table 7.To compare the model results to the observations of MBM20 and the Eridanus Hole wemust consider that MBM20 absorbs about 75% of the background O
VII emission and about61% of the background O
VIII emission and therefore we expect a significant contaminationfrom the background emission. Table 8 summarizes the final values predicted in the SWCXsimulations (model B1 from Table 7), along with the measured O
VII and O
VII fluxes, theestimated foreground (local) flux from Table 4, and the predicted residual cosmic background(data minus model).As the results in Table 8 show, the SWCX O
VII prediction is comparable, within onesigma, with the measured local emission. Also, the residual O
VII cosmic background has aconstant value, within error bars, for all on-cloud and off-cloud observations that is consistentwith the extrapolated background emission reported in Table 5. Both results seem to indicatethat the O
VII foreground emission is dominated by SWCX. This conclusion is also supportedby a previous application of the model to the MBM12 observations (Koutroumpa et al. 2007).Due to the significantly higher absorption of MBM12, the model results were compareddirectly to the total measured flux and the agreement was within 30%. We point out thatthis conclusion does not preclude the existence or a Local Hot Bubble which is expectedto emit X-rays primarily at lower energy, in the 1 / VII surface brightness of about 0.25 LU.The O
VIII results are not as clear, as the measured data are consistent with a zero localemission, while the model predicts a small, but non-zero emission. However, as mentionedbefore, the ACE O data are too sparse and could not be used as input for our model.The negligible O VIII flux could therefore simply be caused by a smaller than expected O density in the solar wind. 12 –
6. Conclusion
We used
Suzaku observations of the molecular cloud MBM20 and a low neutral hydrogencolumn density region nearby to separate and characterize the foreground and backgrounddiffuse X-ray emission. We measured a foreground flux of 0 . ± .
91 LU and 0 . ± . VII and O
VIII respectively and a background flux of 5 . ± .
79 LU and 1 . ± . VII and O
VIII respectively.The comparison with a previous observation of the same regions with
XMM-Newton indicates a significant change in the foreground flux which we attribute to Solar Wind ChargeeXchange. By combining our results with similar multiple shadow investigation of the sametarget we find that the O
VII emission varies between 1 . ± .
61 LU and 4 . ± .
90 LUbetween multiple observations of the same target. The O
VIII emission, except for a singlecase with a change of 2 . ± .
37 LU, is generally compatible with zero, possibly indicatinga very low density of O in the solar wind.We also compared our results with a SWCX model to constrain its O VII and O
VIII emission. The model is in good agreement with the measured O
VII flux and seems toindicate that most of the O
VII foreground emission is due to SWCX. This is not necessarilyinconsistent with the existence of a local hot bubble which is expected to emit predominantlyat lower energy, in the 1 / ROSAT
All Sky Survey Data for the same targets isconsistent with a foreground plasma emission with T = 0 . × K and EM = 0 .
096 cm − pcand a background plasma emission with T = 2 . × K and EM = 0 . − pc . Wealso obtained a good fit by using a non-equilibrium plasma model for the foreground emission,however the inferred age of the plasma is ≤ . REFERENCES
Anders, E., & Ebihara, 1982, Geochimica et Cosmochimica Acta, 46, 2363Arnaud, K., & Dormer, B. 2002, XSPEC: An X-Ray Spectral Fit-ting package, User’s Guide for version 11.2.x, HEASARC, Explo- 13 –ration of the Universe Division, NASA/GSFC, Greenbelt, MD 20771,http://heasarc.gsfc.nasa.gov/docs/xanadu/xspec/manual/manual.htmlBorkowski, K. J., Lyerly, W. J., & Reynolds, S. P. 2001, ApJ, 548, 820Boulanger, F., & Perault, M. 1988, ApJ, 330, 964Edgar, R. J., & Cox, D. P. 1993, ApJ413, 190Galeazzi, M., Gupta, A., Covey, K., & Ursino, E. 2007, ApJ, 658, 1081Galeazzi, M., Gupta, A., & Ursino, E. 2009, ApJ, 695, 1127Gupta, A., & Galeazzi, M., 2009, ApJ, 702, 270Hearty, T., et al. 2000, A&A, 357, 681Henley, D. B., Shelton, R.L. & Kuntz, K. D. 2007, ApJ, 661, 304Henley, D. B. & Shelton, R.L. 2008, ApJ, 675, 335Ishisaki, Y., et al. 2007, PASJ, 59,113Kharchenko, V. & Dalgarno, A. 2000, J. Geophys. Res., 1051835Koutroumpa, D., et al. 2006, A&A, 460, 289Koutroumpa, e al. 2007, A&A, 475, 901Koutroumpa, e al. 2009, ApJ697, 1214Koyama, K., et al. 2007, PASJ, 59, s23McCammon, D., et al. 2002, ApJ, 576, 188Mitsuda, K., et al. 2007, PASJ, 59s, 1Morrison, R. & McCammon, D. 1983, ApJ, 270, 119Sanders, W. T., et al. 2001, ApJ, 554, 694Schwadron, N. A. & Cravens, T. E. 2000, ApJ, 544, 558Smith, R. K., et al. 2005, ApJ, 623, 225Smith, R. K., et al. 2006, PASJ, 59, 141 14 –Smith, R. K., & Cox, D. P. 2001, ApJS, 134, 283Snowden, S. L., et al. 1998, ApJ, 493, 715Tawa, N., et al. 2008, PASJ, 60s, 11
This preprint was prepared with the AAS L A TEX macros v5.2.
15 –
Fig. 1.— IRAS 100 µ m map of MBM20 and surroundings showing the two pointing used inthis investigation. 16 – Fig. 2.— XIS1 image of MBM20 ( left and Eridanus Hole ( right in the energy range 0.5-2.0keV. Point sources (black circle) and corners of the detector (grey circles) have been removedfor the analysis. 17 –
Fig. 3.— Eridanus Hole(
Top ) and MBM20(
Bottom ) Suzaku spectra, with the best fittingthree component model. 18 – flux (counts s-1 keV-1)
M B M 2 0
E r i d a n u s H o l e e n e r g y ( k e V )
Fig. 4.— Simultaneous fit for MBM20 and Eridanus Hole
Suzaku data. 19 – flux (counts s-1 keV-1)
S u z a k u - M B M 2 0
S u z a k u - E r i d a n u s H o l e
R O S A T - M B M 2 0
R O S A T - E r i d a n u s H o l e e n e r g y ( k e v )
Fig. 5.— Global fits for MBM20 (dark grey) and Eridanus Hole (grey) using data from our
Suzaku observations (circles) and RASS (squares). 20 –
Fig. 6.—
Top : Comparison between
XMM-Newton
MBM20 spectra and
Suzaku model foldedthrough
XMM-Newton response.
Bottom : Same as top for Eridanus Hole. 21 –
Fig. 7.—
Left : Ulysses/SWICS O +7 /O +6 ratio data (plain) during the 1995 crossing of thesolar equatorial plane. The dotted curve marks the spacecraft heliographic latitude on theright axis. The vertical and horizontal plain lines denote the limits of the slow solar windequatorial zone in terms of crossing time and heliographic latitude respectively. Right : Sameas left, except for the 2007 crossing of the solar equatorial plane. 22 –Table 1. Details of our
Suzaku observations.Target Observation ID Start Time(UT) End Time(UT) Exposure(ks)MBM20 502075010 2008-02-11 14:41:19 2008-02-14 16:45:11 69.2Eridanus Hole 502076010 2007-07-30 00:51:47 2007-08-01 05:11:19 84.6 23 –Table 2. Model parameters of the spectral fitsDataset(s) Local component Galactic Halo Power Law χ /dof T E.M. a T E.M. Γ b Norm c (10 K ) /keV cm − pc (10 K ) /keV cm − pc MBM20 EH (MBM20+EH) (MBM20+EH) (MBM20+EH) a Emission Measure b Index of absorbed power law fit c Normalization of power law fit at 1 keV in units of photons keV − s − cm − sr − Suzaku data only Suzaku and RASS data XMM-Newton result from Galeazzi et al. 2007 24 –Table 3. Summary of the oxygen line emission for MBM12, MBM20, and the filament inthe southern galactic hemisphere (SGF). The data from McCammon et al. 2002 are alsoreported.Experiment NH(10 cm − ) O VII O VIII
MBM20 . ± .
56 0 . ± . . ± .
60 0 . ± . Eridanus Hole . ± .
34 1 . ± . . ± .
04 1 . ± . MBM12-on cloud . ± .
55 2 . ± . . ± .
26 0 . ± . MBM12-off cloud . ± .
42 0 . ± . Henley et al. on filament . ± .
80 3 . ± . . ± .
41 2 . ± . Henley et al. off filament . ± .
44 2 . ± . . ± .
61 3 . ± . McCammon et al. 2002 . ± . . ± . VII O VIII K cm − pc LU LU
MBM20
XMM 1.12 0.0088 2 . ± .
78 0 . ± . . ± .
91 0 . ± . MBM12
Suzaku ∼ . . ± .
26 0 . ± . . ± .
55 2 . ± . SGF
Suzaku 0.95 0.0064 1 . ± . . ± . . ± . ≤ VII O VIII K cm − pc LU LU
MBM20
XMM 2.23 0.0034 5 . ± .
98 1 . ± . . ± .
79 1 . ± . MBM12
Suzaku 2 . ± .
33 0 . ± . SGF
Suzaku 1.29/3.16 0.034/0.0065 8 . ± . . ± . . ± VII and O
VIII variations between multiple observations of the same object.Target ∆[O
VII ] (LU) ∆[O
VIII ] (LU)MBM20 1 . ± .
82 0 . ± . . ± .
09 0 . ± . . ± .
61 2 . ± . . ± .
90 1 . ± . . ± .
56 0 . ± .
67 28 –Table 7. Model SWCX oxygen line intensites in LU.ObsId Target Model A Model B a Model B1 b O VII O VIII O VII O VIII O VII O VIII a A larger latitudinal extent ( ± ◦ ) of slow wind heavy ion abundancesis assumed for solar minimum (Suzaku). XMM-Newton simulation as-sumptions remain unchanged. b A real-time O +7 measured density is applied to model B simulationfor the Suzaku observations. O +7 data taken in situ at the L1 point areextrapolated to the whole LOS. XMM-Newton O VII simulations and O
VIII simulations remain unchanged. 29 –Table 8. Data and SWCX model oxygen line intensities in LU. The foreground values arefrom Table 4 for the data and are the average of the Eridanus Hole and MBM20 values forthe models.Data Model B1 ResidualMission Target O
VII O VIII O VII O VIII O VII O VIII
EH 7 . ± .
34 1 . ± .
17 2.04 0.80 5 . ± .
34 0 . ± . . ± .
56 0 . ± .
24 1.88 0.74 2 . ± . ∼ . ± .
78 0 . ± .
43 1.96 0.77 0 . ± . ∼ . ± .
04 1 . ± .
79 0.29 0.42 5 . ± .
04 0 . ± . . ± .
60 0 . ± .
48 0.15 0.38 2 . ± . . ± . . ± .
91 0 . ± .
01 0.22 0.40 0 . ± . ∼∼