Far-Ultraviolet Observations of the Spica Nebula and the Interaction Zone
Yeon-Ju Choi, Kyoung-Wook Min, Kwang-Il Seon, Tae-Ho Lim, Young-Soo Jo, Jae-Woo Park
aa r X i v : . [ a s t r o - ph . GA ] J u l D RAFT VERSION O CTOBER
16, 2018
Preprint typeset using L A TEX style emulateapj v. 5/2/11
FAR-ULTRAVIOLET OBSERVATIONS OF THE SPICA NEBULA AND THE INTERACTION ZONE Y EON -J U C HOI , K YOUNG -W OOK M IN , K WANG -I L S EON , T AE -H O L IM ,Y OUNG -S OO J O , J AE -W OO P ARK Draft version October 16, 2018
ABSTRACTWe report the analysis results of far ultraviolet (FUV) observations, made for a broad region around α Vir(Spica) including the interaction zone of Loop I and the Local Bubble. Whole region was optically thin and ageneral correlation was seen between the FUV continuum intensity and the dust extinction, except in the neigh-borhood of the bright central star, indicating the dust scattering nature of the FUV continuum. We performedMonte-Carlo radiative transfer simulations to obtain the optical parameters related to the dust scattering as wellas the geometrical structure of the region. The albedo and asymmetry factor were found to be 0.38 ± ± + - pc and 40 + - pc, respectively. The diffuse FUVcontinuum in the northern region above Spica was mostly the result of scattering of the starlight from Spica,while that in the southern region was mainly due to the background stars. The C IV λλ II * λ II region. This indicates that the C IV line arises mostly at the shell boundaries of the bubbles, with a largerportion likely from the Loop I than from the Local Bubble side, whereas the Si II * line is from the photoionizedSpica nebula. Subject headings:
ISM: individual (Spica Nebula) — H II region — ultraviolet: ISM Interaction Zone INTRODUCTIONObservations in the far ultraviolet (FUV) wavelengths (900-1750 Å) provide a wealth of information regarding the physi-cal and chemical processes in the interstellar medium (ISM).For example, many important ion emission lines associatedwith the cooling hot gas of T ∼ . -10 . K, such as C
III λ VI λλ IV λλ II re-gions around bright stars. One example is the Spica Nebula.Spica ( α Vir), located at ( l , b ) = (316 ◦ .11, 50 ◦ .84), is a binarysystem with a primary star of the B1 III-IV type and a sec-ondary star of the B2 V type, and is at the distance of ∼
80 pcfrom the Sun in the direction toward the Scorpius-Centaurus(Sco-Cen) association. The Spica Nebula is relatively iso-lated, as revealed in the Wisconsin H α Mapper (WHAM) im- email: [email protected] Korea Advanced Institute of Science and Technology (KAIST), 373-1Guseong-dong, Yuseong-gu, Daejeon, Korea 305-701, Republic of Korea Korea Astronomy and Space Science Institute (KASI), 61-1 Hwaam-dong, Yuseong-gu, Daejeon, Korea 305-348, Republic of Korea Korea Intellectual Property Office (KIPO), Government ComplexDaejeon Building 4, 189 Cheongsa-ro, Seo-gu,Daejeon, Korea 305-348,Republic of Korea age (Haffner et al. 2003), and is characterized by a low gasdensity of 0.2 – 0.6 cm - (Reynolds 1985; Park et al. 2010).Park et al. (2010) analyzed its spectral images made for theSi II * 1533 and Al II II * intensity in the southern region, a feature also seenin the H α image. They attributed the feature to the densityincrease in the southern region. In contrast, the image of theAl II II region from the analysis of the InfraredAstronomical Satellite (IRAS) 60 µ m and 100 µ m observa-tions (Zagury et al. 1998). Murthy & Henry (2011) noted theFUV halo around Spica and attributed it to dust scattering ofthe central star Spica. They tried to constrain the distance tothe dust layers with a single scattering model. We will furtherdiscuss their results later for comparisons with our model.The Spica Nebula is located close to the so-called the "in-teraction zone" of the Local Bubble (hereafter, LB) and theLoop I superbubble (Egger & Aschenbach 1995). The ex-istence of the LB was inferred from the soft X-ray back-ground observations (Cox & Reynolds 1987). It is a low-density ( ∼ . × - cm - ) region with a radius of ∼
100 pc, inside which the solar system is embedded. Anumber of direct and indirect observations have revealeda cool and dense neutral hydrogen gas layer with a col-umn density N (H I ) of ∼ . × cm - (Cox & Reynolds1987; Lallement et al. 2003) as an expanding boundary ofLB (Knude 1978; Ferlet et al. 1985; Fruscione et al. 1994;Welsh et al. 1994). Nevertheless, the shape, origin, and theionization structure of LB are still under debate (Cox 1998;Welsh et al. 2010; de Avillez & Breitschwerdt 2012). TheLoop I is a large radio loop centered on ( l , b ) = (329 ◦ ± ◦ .5,17 ◦ .5 ± ◦ ) with an angular radius of ∼ ◦ . It is believed tohave been formed by multiple supernova explosions and/orstrong stellar winds of the Sco-Cen OB association located at ∼
170 pc from the Sun. The boundary of Loop I was observedto be expanding with a velocity of ∼
20 km s - into the neu-tral ambient medium of density n ∼ - , to form a denseneutral shell with column density N (H I ) of ∼ . × cm - at the terminal shock (Sofue et al. 1974; Egger & Aschenbach1995). There is a wealth of evidence that Loop I consistsof a low density ( ∼ . × - cm - ), hot ( ∼ . K) andhighly ionized X-ray emitting gas, generated by the interac-tion between the shock waves from the recent supernova ex-plosions within the Sco-Cen association and the ambient neu-tral medium (Egger & Aschenbach 1995; Breitschwerdt et al.2000; Willingale et al. 2003; Welsh & Lallement 2005). TheX-rays, together with the young OB stars of the Sco-Cen as-sociation, are believed to provide intense radiations that areresponsible for the ionization of neutral gas at the Loop Iboundary, probably on the inner side of the boundary wall(Welsh & Lallement 2005).The interaction zone is a huge ring-like feature of a denseneutral matter within the apparent boundary of the Loop Ishell, as identified by Egger & Aschenbach (1995) from theanalysis of diffuse X-ray and 21 cm neutral hydrogen maps.The feature has been suggested to be a colliding structureof Loop I with LB. Yoshioka & Ikeuchi (1990) predicted thering-like feature using hydrodynamical simulations; when ashock at the boundary of a bubble in a radiative stage makescontact with a shock of another bubble, a dense neutral wallof a ring-like shape with a density is 20 – 30 times higherthan those of the ambient medium, would be formed at theintersecting points of the two bubbles. Egger & Aschenbach(1995) noted that the N (H I ) of this region suddenly jumpsfrom ∼ cm - to ∼ . × cm - at a distance of ∼
70 pcfrom the Sun using the absorption measurements of metal ionsby Fruscione et al. (1994), and they suggested the distance tothe interaction zone to be ∼
70 pc. Assuming a toroidal shapeof thickness of ∼ ◦ , corresponding to ∼
15 pc at 70 pc, itsdensity was estimated to be ∼ - . Breitschwerdt et al.(2000) set the upper limit of the distance of the interactionzone to be 80 – 100 pc since Loop I, being still active withongoing star formations, may have a higher pressure than theadjacent LB and push the interaction shell towards the LB.Corradi et al. (2004), using the Strömgren photometry andNa I column density measurements, estimated the distance tothe interaction zone to be 120 – 150 pc. On the other hand,Reis & Corradi (2008) suggested that the ring-shaped inter-action zone may be folded and warped as both the distance tothe ring and the color excess values vary along the ring; theyestimated the distance to the interaction zone to be 110 ± ±
50 pc on the eastern side.In this paper, we present the analysis results of the FUVobservations of an extended region around the Spica Nebulathat includes the interaction zone. The main goal of this studyis to understand the morphological relationship between theSpica Nebula and the neighboring interaction zone as well asthe scattering properties of the associated dust clouds usingthe FUV observations and radiative transfer models. DATAWe employed two datasets for this study: one is from thearchival FUV data of Galaxy Evolution Explorer (GALEX;Morrissey et al. (2007)), obtained as part of the All-SkyImaging Survey (AIS), and the other is the dataset used byPark et al. (2007, 2010), obtained from the Far-ultravioletImaging Spectrograph (FIMS; Edelstein et al. (2006a,b)) aboard the Korean microsatellite STSAT-1. Both observationscover very similar FUV wavelength bands of 1350–1780 Åand 1330–1720 Å for GALEX and FIMS, respectively. Theyare complementary to each other in that the GALEX data havegood statistics but no spectral information, while the FIMSdata provides spectral information but with rather large sta-tistical fluctuations due to relatively short exposure time andlower sensitivity. Hence, we use the GALEX data for pho-tometric analyses as it gives more reliable estimations of thediffuse FUV intensities than the FIMS data, and the FIMSdata for spectral analyses as it can discriminate the FUV lineemissions from the stellar continuum emissions. The GALEXsky background data (with an extension of skybg), providedwith a resolution of 16 ′ (Morrissey et al. 2007) from the web-site http://galex.stsci.edu/GR6/, were smoothed to obtain thefinal image with a resolution of 0.5 ◦ . The data reduction pro-cedures of the FIMS data are described in detail by Park et al.(2007). More information on the instrument FIMS can befound in Edelstein et al. (2006a,b). RESULTSFigure 1 shows the GALEX FUV image, together with themaps obtained from the observations made in other wave-lengths. The figures are shown for an extended region of 24 ◦ × ◦ around the Spica Nebula, in which the central star α Vir is marked with an asterisk at ( l , b ) = (316 ◦ .11, 50 ◦ .84).The black circle of the 8 ◦ radius, defined as the boundary ofthe Spica Nebula, is also drawn in each of the figures. InFigure 1(a), the missing data regions shown in black are theregions of bright stars that were intentionally avoided duringthe observations. The H α map of Figure 1(b), scaled in unitsof Rayleighs (1R = (10 /4 π ) photons cm - sr - s - ), was ex-tracted from the H α sky survey map of Finkbeiner (2003) andHaffner et al. (2003). The H I column density map of Figure1(c), shown in units of 10 cm - , was taken from the LeidenArgentine Bonn Galactic H I Survey (Kalberla et al. 2005).The dust extinction map of Figure 1(d) was obtained from theGalactic reddening map of the Schlegel et al. (1998), whichwas derived from the Cosmic Background Explorer (COBE)and the Infrared Astronomical Satellite (IRAS) observations.First, we note that the H α intensity is mostly confinedwithin a finite volume of the ionized nebula and dropsabruptly at the southern boundary near the interaction zone ofLB and Loop I, while the FUV intensity continues southwardbeyond the boundary of the nebula. Park et al. (2010) arguedthat the H II region is "ionization bounded" at the boundaryin the southern region with N (H I ) ∼ . × cm - , belowwhich a good correlation is seen between the H I column den-sity and the H α intensity. This conclusion accords well withthe fact that the estimated distance to the interaction zone is ∼
70 pc, which is comparable to that of the Spica Nebula.There are two noticeable features in the FUV map: a haloaround the central star α Vir and an extended feature thatgenerally traces dust in the interaction zone. The FUV haloaround α Vir is best seen in the northern region above the starand the halo is a dust-scattered feature of the α Vir starlight.The extended FUV feature in the southern region is predom-inant, as will be discussed in Section 3, due to scattering ofstarlight originating from other background stars by dust inthe interaction zone.Figure 2(a) shows a pixel-to-pixel correlation plot betweenthe FUV intensity and the dust extinction in the whole re-gion of 24 ◦ × ◦ except the bright central part of an angular F IG . 1.— The extended Spica Nebula region observed in various wavelengths: (a) the GALEX FUV intensity in CU ( photons sr - s - cm - Å - ), (b) the H α intensity in Rayleigh (R = (10 /4 π ) photons cm - sr - s - ), (c) the H I column density in units of 10 cm - , and (d) the dust extinctioin given by E ( B-V ). Theimages are shown in galactic coordinates. The black circles of radius 8 ◦ denote the boundary of the Spica Nebula. The central star α Vir is marked with anasterisk at at ( l , b ) = (316 ◦ .11, 50 ◦ .84).F IG . 2.— Correlation plots between the FUV intensity and dust color excessin the region outside the bright circular region with an angular radius of 3.5 ◦ .(a) is for the entire region and (b) is an enlarged plot for E ( B-V ) < 0.03. Thedata points marked with blue crosses represent the northern region above thecentral star α Vir and those with black circles represent the southern regionbelow the star. size of 3.5 ◦ around Spica. In the figure, the data points with E ( B-V ) < 0.06, denoted in blue, were mostly obtained in the northern region above the central star, while those with E ( B-V ) > 0.06, denoted in black, were obtained in the southernregion. It should be noted that the color excess E ( B-V ) inthe whole region is less than 0.14, which corresponds to theoptical depth of τ ∼ E ( B-V ) increases, except that the data points are severelyscattered for the values of E ( B-V ) from 0.04 to 0.08 whiletheir intensities are limited to ∼ ∼ E ( B-V ). Seon et al.(2011a) & Seon (2013) noted this property in the analysesof the FUV continuum background, and attributed it to thelognormality of the FUV intensity, caused by the turbulenceproperty of ISM. It is well known that the probability distri-bution functions of the ISM density and column density arevery close to lognormal (Vazquez-Semadeni 1994; Klessen2000; Burhart & Lazarian 2012; Seon 2012a). Since the dust-scattered intensity is roughly proportional to the dust columndensity, the FUV intensity should exhibit a signature of thelognormal density structure. They also found that the disper-sion of the FUV intensity is consistent with that observed in F IG . 3.— Pixel-to-pixel comparison of N (H I ) with E ( B-V ), as obtainedfrom Figure 1(d) and Figure 1(c). The blue crosses and the black circles arethe data points obtained from the northern and the southern regions, respec-tively. The red line represents the best-fit ratio of N (H I ) to E ( B-V ). the turbulent molecular clouds.It is also interesting to see that, in Figure 2(b), which isan enlarged plot for the low extinction part of Figure 2(a),both the FUV intensity and the dust extinction E ( B-V ) haveminimum values of ∼
200 CU and ∼ ∼
200 CU is comparable to theso-called "isotropic component" of ∼
300 CU (Bowyer 1991;Henry 2002; Seon et al. 2011a), although some part of it mayalso come from airglow (Sujatha et al. 2010). These mini-mum values could be regarded as background values that arenot relevant to the region considered in this study.Figure 3 shows a good correlation between the neutralatomic hydrogen column density N (H I ) and E ( B-V ). The fig-ure reveals that the ratio of N (H I ) to E ( B-V ) was ∼ . × atoms cm - mag - . There have been several studies to esti-mate the ratio in the general diffuse interstellar medium. Us-ing 21 cm data, Knapp & Kerr (1974) and Burstein & Heiles(1978) obtained an average ratio of ∼ . × atoms cm - mag - . Savage & Mathis (1979) used the Ly α absorptiondata obtained toward 100 stars by the Copernicus space-craft and found a ratio of 4 . × atoms cm - mag - .Diplas & Savage (1994a) employed 393 sample stars ob-served with the International Ultraviolet Explorer (IUE) satel-lite and obtained 4 . × atoms cm - mag - in the regimeof E ( B-V ) ≤ . × atoms cm - mag - in thelower color excess values of E ( B-V ) ≤ N (H I ) to E ( B-V ) ratio in the lower color regime was at-tributed to the fact that a significant amount of hydrogen isin a molecular form for E ( B-V ) > 0.3. The present result ishigher than these previous values, implying that dust is prob-ably less abundant in the shell boundaries of the LB and LoopI. The evaporation of dust by the hot gas of LB and Loop Iand/or the expelling of dust by the strong radiations pressurefrom the young OB stars of the Sco-Cen association might beresponsible for this deficiency of dust.The image of Figure 4(a), made by utilizing the HEALPixscheme (Górski et al. 2005) with a pixel resolution of 0.5 ◦ ,was constructed from the FIMS data of the 1360 – 1660 Åwavelength band, excluding the airglow line of λ II line of λ ∼
50 bright pixelswere removed based on the TD-1 stellar catalog. The brightstreak inside the narrow rectangular box is an artifact due tothe instrumental scattering of the bright central star along theslit direction. The color levels are the same as those adopted for the GALEX image in Figure 1(a). The two images ofGALEX and FIMS are very similar, although the FIMS inten-sity varied less than the GALEX data because of smoothingof the FIMS image. In Figure 4(b), the top panel shows thespectrum of the core region inside an angular radius of 0.5 ◦ ,and the middle and bottom panels show the spectra of the dif-fuse emission in the two outer regions A and B, respectively.The spectrum of the top panel is dominated by the centralstar ( α Vir), and its shape is consistent with the stellar spec-trum of Spica observed by IUE. The Si II * and C IV ion linesare found in the outer region spectra at wavelengths ∼ ∼ II * line was studied byPark et al. (2010) using a photo-ionization model appropriateto the Spica Nebula. The Si II * emission was mostly observedwithin the Nebular and its intensity was slighter higher in thesouthern part than in the northern part. We fitted the C IV dou-blet lines observed in the outer region spectra using the IDL-based MPFIT (Markwardt 2009) by assuming two Gaussiansfunctions, as done in Seon et al. (2006); the best-fit intensitiesare ∼ ± - cm - sr - (line unit; LU) and ∼ ± IV intensity fluctuated significantly and that the ex-posure time was rather short. We also note that the molecularhydrogen fluorescence lines are found in the spectrum of thesouthern region corresponding to the interaction zone. MODELING AND DISCUSSIONSpica is likely located close to the interaction zone of theLB and Loop I, as mentioned in the Introduction. Since LoopI is observed to be expanding into the ambient medium of ∼ - , the density of the ambient medium would be sim-ilar to that the Spica Nebula region. Reynolds (1985) andPark et al. (2010) estimated the density of ∼ - and ∼ - respectively, implying that Spica is lo-cated in the ambient medium outside the two bubbles. Itis also interesting to note that the H I column density toSpica was estimated to be ∼ cm - (Shull & van Steenberg1985; Fruscione et al. 1994), which is similar to the columndensity of the shell boundary of the LB (Cox & Reynolds1987; Lallement et al. 2003). The shell boundary of LoopI had a column density of ∼ cm - (Sofue et al. 1974;Egger & Aschenbach 1995). These observations suggest thatSpica is located behind the LB shell but in front of the Loop Ishell. The density ( ∼
15 cm - ; Egger & Aschenbach (1995))of the interaction zone was much higher than that of the SpicaNebula. This indicates that Spica is located slightly awayfrom the interaction zone. Taking all these considerations intoaccount, the relative geometry between LB , Loop I and Spicacould be represented by a schematic diagram shown in Figure5. As we discussed previously, the interaction zone is rich indust and its column density N (H I ) is proportional to the levelof dust extinction, implying that dust responsible for the scat-tering of starlight should be located close to the central starand the interaction zone. This will be confirmed by modelingthe scattered stellar photons by assuming dust distributions.A similar study was conducted by Murthy & Henry (2011)for the halo around Spica observed by GALEX. They focusedon the region close to the star and ignored the dust scatteringeffect in the interaction zone. They employed a single dust-scattering model and assumed only a single star (i.e, α Vir). F IG . 4.— (a) The FUV intensity map observed by FIMS is given in CU ( photons sr - s - cm - Å - ). The bright streak in the narrow rectangular box is due tothe instrumental scattering of the bright central star along the slit direction. The black circle of radius of 8 ◦ indicates the boundary defining the Spica Nebula. Thecentral star, α Vir, is marked with an asterisk at at ( l , b ) = (316 ◦ .11, 50 ◦ .84). (b) Spectra observed by FIMS: the top panel shows the spectrum for the central coreregion within an angular radius of 0.5 ◦ ; the middle and bottom panels present the spectra for the regions A and B, as defined in (a), respectively. The verticaldashed lines indicate the Si II * and C IV lines, and the brightest H fluorescence lines.F IG . 5.— Schematic representation of the geometrical relation between theLocal Bubble, Loop I , the interaction zone and the central star Spica. In their first model, they placed a thin dustsheet at a distanceof 3 pc from the star, inferred from the 60 µ m and 100 µ minfrared observations, and obtained a very low albedo valueof 0.10 ± ∼ φ ( θ ) = a π (1 - g )[1 + g - g cos( θ )] . (1)where a and g are the albedo and the phase function asym-metry factor of dust scattering, respectively. The simulationcode employed the peeling-off method (Yusef-Zadeh et al.1984) for simulation efficiency; while each photon from radi-ation sources propagates through the dust medium and expe-riences multiple scatterings into random directions, the codestores the scattered fraction of the photon in the direction to-ward the Sun from each scattering, and these fractions areadded up to give the final intensity. The parameters to be de-termined by the simulation are the albedo a, the asymmetryfactor g, and the spatial distribution of dust. More detail de-scriptions on the simulation technique can be found in Jo etal. (2012) and Lim et al. (2013).The diffuse interstellar FUV emissions are composed ofthree different components in addition to the dust-scatteredstarlight: ionic emission lines from hot gases of 10 . –10 . K, H fluorescent emission lines originating from pho-todissociation regions irradiated by interstellar UV photons,and two-photon continuum emission from the ionized gasof about 10 K. Therefore, in order to obtain only the dust-scattering component from the total observed FUV intensity,it is necessary to estimate various contributions to the ob-served FUV intensity. In addition, the isotropic backgroundderived from the FUV intensity extrapolated to N (H I ) = 0should be removed from the observed FUV intensity (Bowyer1991; Henry 2002). In this study, we subtracted the mini-mum intensity of 200 CU as the contribution of the isotropicbackground. The two-photon emission was estimated usinga simple formula of 60 CU per 1 Rayleigh of H α intensity(Reynolds 1990). However, the contribution is insignificanteven in the Spica Nebula with ∼
6% contribution of the to-tal FUV intensity. The contribution of two-photon contin-uum emission in the diffuse ionized gas or warm ionizedmedium outside of the Spica Nebula is also unimportant (seeSeon et al. (2011a)). The contribution of H fluorescent emis-sion is also negligible as it was estimated to be smaller than10% even for the interaction zone.The simulation domain consisted of 400 × ×
400 rect-angular grids cells each with cell size of 1 pc. The central axiswas chosen along the line of sight toward Spica, and the Sunwas placed at the center of the front face of the simulation box.A total of 21,225 stars from the TD-1 and Hipparcos catalogswere disposed in the simulation box as radiation sources. Atotal of 10 photons were generated and assigned to the starsin proportion to their intrinsic luminosities. The spatial distri-bution of dust was obtained based on the Schlegel, Finkbeinerand Davis (SFD) dust map (Schlegel et al. 1998). As the mapis two-dimensional without the distance information along theline of sight, a dust slab with a thickness (t) was assumed to belocated at a distance (d). The total dust column density alongeach sightline was uniformly distributed among the dust cellsin the sightline. The optical depth at 1565 Å was derived fromthe E ( B-V ) value of the SFD map with R V = 3.1, the averagevalue for the Milky Way. We assumed that the minimum E ( B-V ) of 0.015 found in Figure 2 is not related to the dust aroundSpica and subtracted the value from the dust extinction map.Comparison between the simulation and the observation wasmade only for the region with an angular distance larger than3.5 ◦ from the central star as the GALEX data may not be re-liable for the bright region close to the central star. The sim-ulation parameters were varied as follows: the distance to thefront face of the dust slab was varied from 50 to 90 pc in stepsof 2 pc and the thickness from 10 pc to 50 pc in steps of 2pc. We also varied the albedo (a) from 0.30 to 0.50 in steps of0.02 and the asymmetry factor (g) from 0.40 to 0.60 in stepsof 0.02. Each simulation was compared with the GALEX ob-servation to obtain a set of best-fit parameters.A Monte Carlo simulation of dust scattering does not con-strain the scattering parameters effectively when only a sin-gle star is present, as discussed in Murthy & Henry (2011).In fact, such a simulation may produce many parameter setsthat yield similar FUV intensities. Hence, we first fitted thesouthern region to obtain the optical parameters of dust, ex-cluding the northern region where Spica dominates the FUVintensity to reduce the effect caused by the bright centralstar. The dust parameters determined from the southern re-gion will be adopted for the simulation of the northern re-gion. The resulting best-fit parameters were d = 70 + - pc, t= 40 + - pc, a = 0.38 ± g = 0.46 ± F IG . 6.— FUV continuum map of the southern region obtained from (a) thedust scattering simulation, and (b) the GALEX observation. A pixel-to-pixelcorrelation plot of the FUV intensity obtained from (a) and (b) is shown in(c). ulation and the observation. It should be also noted thatEgger & Aschenbach (1995) suggested that the location of theinteraction zone was ∼
70 pc, which agrees with the presentbest-fit distance. Furthermore, the obtained optical param-eters a = 0.38 and g = 0.46 accord well with the studiesconducted for other objects (Jo et al. 2012; Lim et al. 2013),as well as the theoretical predictions ( a = 0.4 – 0.6 and g = 0.55 – 0.65) from the carbonaceous-silicate grains model(Weingartner & Draine 2001). Hence, we use a =0.38 and g =0.46 for further discussions.The contributions of the central star and background starsto the total FUV intensity are shown in Figures 7(a) and 7(b),respectively. Figure 7(a) shows that the contribution from thecentral star is most important for E ( B-V ) ≤ E ( B-V ). On the other hand, the backgroundstars become more important than the central star for E ( B-V )> 0.06. In fact, the structures of the interaction zone in theobserved FUV map appear to match the simulation map ob-tained only with the background stars, which implies that theobserved FUV intensity in the interaction zone is dominatedby scattering of the photons of the background stars. It is alsointeresting to note that the simulated intensity is also limitedby a maximum intensity of 1500 CU, due to the finiteness ofthe local stellar radiation field in the region.We also performed radiative transfer simulations for thenorthern region, adopting the values of a = 0.38 and g = 0.46and a single slab model. The distance to the front face andthe thickness of the dust slab were 70 pc and 30 pc, respec-tively. That result implies that ∼
33% of the dust is locatedin front of the central star and the approximate remaining 67% behind the star. Figure 8 presents the resulting simulationmap together with the GALEX observation and compares thesimulation intensity with the GALEX intensity. As seen in the F IG . 7.— Simulated FUV intensity plotted against E ( B-V ) for the southernregion: (a) simulation with the central star only and (b) simulation with thebackground stars only. figure, the simulation and observation are in good agreement.With ∼
33% of the dust located in front of the central star,the H I column density of the northern region, obtained us-ing the net mean E ( B-V ) value of ∼ N (H I ) and E ( B-V ) found in Figure 3, is estimated to be ∼ × cm - ,which is similar to that ( ∼ cm - ) obtained for the shellboundary of the LB (Cox & Reynolds 1987; Lallement et al.2003). Adopting a thickness of 10 pc, this gives a densityof ∼ - , which is consistent with those ( ∼ - or0.2 – 0.6 cm - ) estimated for the H II region (Reynolds 1985;Park et al. 2010). While we used a single slab with a constantdensity and thickness, it is remarkable that the results accordwell with those of other independent observations. We alsotried two-slab models, representing the LB and Loop I bound-aries, but the two slabs could not be separated effectively, per-haps because the density of the LB is similar to that of theambient medium and thus the present models are not able todiscriminate the shell of the LB from the ambient medium.Sallmen et al. (2008) analyzed the results of the FUV ob-servations made by the Far Ultraviolet Spectroscopic Explorer(FUSE) satellite towards a single direction ( l , b ) = (277 ◦ , 9 ◦ )at the boundary of the interaction zone. By comparing theO VI and C III intensities of two other neighboring directionsacross the boundary of the interaction zone, one just outsidethe interaction zone and the other inside the interaction zone,they were able to discriminate the contributions of Loop Ifrom those of the LB, by considering the fact that the inter-action zone with a neutral hydrogen column density of N (H I ) ∼ × cm - makes a shadowing effect on the two FUVlines. The result shows that the O VI emission comes mostlyfrom the interface on the Loop I side of the interaction zonewhile the C III emission is produced on the LB side of theinteraction zone.As the FIMS wavelength band includes the C IV doubletlines at ∼ F IG . 8.— FUV continuum map of the northern region obtained from (a)the best-fit dust scattering simulation, and (b) the GALEX observation. Apixel-to-pixel correlation plot of the FUV intensity obtained from (a) and (b)is shown in (c). with different extinctions: region B, being part of the interac-tion zone, suffers more extinction for the emission originatingfrom Loop I than region A does. The average E ( B-V ) for re-gion A is ∼ E ( B-V ) for region B is0.075 after the background subtraction with a correspondingoptical depth of 0.54 at 1565 Å. With the C IV intensities I A = ∼ ± B = ∼ ± LB = ∼ ± LI = ∼ ± A and I B are used toderived I LB and I LI . The result seems to imply that a largerportion of the C IV emission comes from Loop I than fromLB, although this is not conclusive because of the large errorrange. CONCLUSIONSWe have analyzed the datasets of the FUV observations per-formed by GALEX and FIMS, together with the maps con-structed in other wavelengths, for the Spica Nebula and itsneighborhood, which includes the interaction zone of Loop Iand the LB. The following are the main findings of this study.1. The diffuse FUV radiation observed in the northern re-gion above Spica was mainly due to the dust-scatteringof starlight originating from Spica, while the diffuseFUV radiation in the southern region was attributed toscattering of photons originating from background starsby dust in the interaction zone.2. The FUV intensity showed a general correlation withthe dust color excess E ( B-V ). 3. The ratio of neutral hydrogen to dust was about 7 . × atoms cm - , which is slightly higher than previousstudies estimated for diffuse ISM.4. Diffuse C IV emission was observed throughout thewhole region, including the interaction zone. By con-sidering the shadowing effect in the interaction zone,we found that a larger portion of the C IV emissionarises in the inner side of the shell boundary Loop Ithan in the LB.5. Molecular hydrogen fluorescence lines were also ob-served in the interaction zone.6. Based on the dust scattering simulation using a MonteCarlo radiative transfer code, we estimated the opticalparameters of dust scattering as follows: a = 0.38 ± g = 0.46 ± ∼
70 pcaway from the Sun, with a thickness of ∼
40 pc, cov-ering the southern neighborhood of the Spica Nebula.The central star Spica is likely to be located betweenthe two shells of LB and Loop I.FIMS/SPEAR is a joint project of KAIST and KASI (Ko-rea) and UC Berkeley (USA), funded by the Korea MOSTand NASA grant NAG5-5355. This research was supportedby Basic Science Research Program (2010-0023909) and Na-tional Space Laboratory Program (2008-2003226) through theNational Research Foundation of Korea (NRF) funded by theMinistry of Education, Science and Technology. The Wis-consin H-Alpha Mapper is funded by the National ScienceFoundataion.40 pc, cov-ering the southern neighborhood of the Spica Nebula.The central star Spica is likely to be located betweenthe two shells of LB and Loop I.FIMS/SPEAR is a joint project of KAIST and KASI (Ko-rea) and UC Berkeley (USA), funded by the Korea MOSTand NASA grant NAG5-5355. This research was supportedby Basic Science Research Program (2010-0023909) and Na-tional Space Laboratory Program (2008-2003226) through theNational Research Foundation of Korea (NRF) funded by theMinistry of Education, Science and Technology. The Wis-consin H-Alpha Mapper is funded by the National ScienceFoundataion.