Faraday Rotation in the Tail of the Planetary Nebula DeHt 5
aa r X i v : . [ a s t r o - ph . GA ] S e p Accepted for publication in the Astrophysical Journal on 2010September 13.
FARADAY ROTATION IN THE TAIL OF THE PLANETARYNEBULA DeHt 5
R. R. Ransom , , R. Kothes , M. Wolleben and T. L. Landecker [email protected]
ABSTRACT
We present 1420 MHz polarization images of a 5 ◦ × ◦ region around the plan-etary nebula (PN) DeHt 5. The images reveal narrow Faraday-rotation structureson the visible disk of DeHt 5, as well as two wider, tail-like, structures “behind”DeHt 5. Though DeHt 5 is an old PN known to be interacting with the inter-stellar medium (ISM), a tail has not previously been identified for this object.The innermost tail is ∼ > RM ) through the inner tail to be − ± − , and, us-ing a realistic estimate for the line-of-sight component of the ISM magnetic fieldaround DeHt 5, derive an electron density in the inner tail of n e = 3 . ± . − .Assuming the material is fully ionized, we estimate a total mass in the inner tailof 0 . ± .
33 M ⊙ , and predict that 0 . ± .
33 M ⊙ was added during the PN-ISM interaction. The outermost tail consists of a series of three roughly circularcomponents, which have a collective length of ∼ Department of Physics and Astronomy, Okanagan College, 583 Duncan Avenue West, Penticton, B.C.,V2A 8E1, Canada National Research Council of Canada, Herzberg Institute of Astrophysics, Dominion Radio AstrophysicalObservatory, Box 248, Penticton, BC, V2A 6J9, Canada
Subject headings: planetary nebulae: individual (DeHt 5) — ISM: structure —polarization — radio continuum: ISM
1. INTRODUCTION
Interaction with the interstellar medium (ISM) is expected for planetary nebulae (PNe)in a later stage of evolution (e.g., Borkowski, Sarazin, & Soker 1990). In the early stages, theshell of the PN (i.e., the region where the fast wind from the hot central star collides with theslow wind from its progenitor) has high density, and the PN expands essentially unimpededinto the surrounding ISM. As the nebula expands, its density decreases, and at some point thethermal pressure in the nebular shell drops to a value equal to the ram pressure of the ISM.In many cases, the magnetic pressure from the magnetized ISM should also be considered(e.g., Soker & Dgani 1997). Equal internal and external pressures mark the proposed startof the PN-ISM interaction. Since all stars have some peculiar motion, and the ISM pressureis proportional to the square of the speed of the PN system relative to the ambient ISM, thePN-ISM interaction appears first as a bow shock upstream from the central star. Indeed,an asymmetric emission structure is the observational marker for PN-ISM interaction in oldPNe with large proper motions (Tweedy & Kwitter 1996; Xilouris et al. 1996; Kerber et al.2000).Recently, two-dimensional (Villaver, Garc´ıa-Segura, & Manchado 2003) and three di-mensional (Wareing et al. 2006b; Wareing, Zijlstra, & O’Brien 2007a; Wareing et al. 2007c)hydrodynamic simulations have been used to show that the interaction between an evolvedintermediate-mass star ( ∼ ⊙ ) and the ISM likely starts during the asymptotic giantbranch (AGB) stage. The simulations predict that for any system moving with respect tothe ambient (warm neutral or warm ionized) ISM, an upstream bow shock is formed betweenthe slow, dense AGB wind and the ISM. Additionally, the simulations predict that materialis ram-pressure stripped from the upstream interface and deposited downstream to forma comet-like tail behind the moving system. These predictions are confirmed for at leastone star, namely the primary (AGB) star in the Mira AB binary (see Martin et al. 2007;Matthews et al. 2008). Far-ultraviolet observations for this high-proper-motion system showemission from both a bow shock and a ∼ & α observations, while forMira AB, the ∼ α may, if carefully perused, reveal other tails. However, it is unlikelythat they can be used to trace the full mass-loss histories of interacting systems, since thematerial stripped during the AGB phase will be very diffuse and at least partially deionized.On the other hand, the ultraviolet emission in the Mira AB tail, though tracing a significantfraction of the mass-loss period for that system, suggests an unusual, and perhaps unique,emission mechanism. Recent observations of the neutral component in the Mira AB tailsuggest that H I imaging is a useful tool for studying the morphology and dynamics of tailsbehind evolved stars, but high-resolution H I images may be limited to the densest (i.e.,innermost) regions of the tails. A better tool for locating and studying the extended tailsbehind evolved intermediate-mass stars may be radio polarimetric observations. Polarimet-ric observations at frequencies ≤ ≥
10 cm) are very sensitive to Faraday 4 –rotation of the diffuse Galactic synchrotron emission (e.g., Uyanıker et al. 2003), and havebeen used previously to study the interaction region at the upstream interface between PNSh 2-216 and the ISM (Ransom et al. 2008). In propagating through the magnetized ISM,the polarized component of the background emission is rotated at wavelength λ [m] throughan angle ∆ θ = RM λ [rad] , (1)where RM is the rotation measure and depends on the line-of-sight component of the mag-netic field, B k [ µ G], the thermal electron density, n e [cm − ], and the path length, dl [pc],as RM = 0 . Z B k n e dl [rad m − ] . (2)Even in the relatively low electron-density environment expected in the tail behind a PNor AGB star ( n e ≈ n H ∼ − , see Wareing et al. 2007a), a rotation of the backgroundpolarization angle of ∆ θ ∼ ◦ can be expected for a system moving through the ISM near theGalactic plane. Such a rotation will result in a Faraday-rotation signature for the tail whichis readily observable, albeit one which can be corrupted by other, perhaps more turbulent,structures along the line of sight.In this paper, we report the discovery of a Faraday-rotation structure consistent with atail behind the fast-moving PN DeHt 5 (=DHW 5; PK 111.0+11.6). The structure is identi-fied in the 1420 MHz polarization images of the Canadian Galactic Plane Survey (CGPS).DeHt 5 is an old PN known to be interacting with the ISM (Tweedy & Kwitter 1996), butno tail has previously been associated with this object. In § §
3, we describebriefly the preparation of the CGPS polarization images for the region surrounding DeHt 5.In §
4, we characterize the Faraday-rotation structures at the position of DeHt 5 and in thetail behind DeHt 5, and estimate the RM s through the inner and outer portions of the tail.In §
5, we discuss in detail the PN and tail structures, and estimate the electron densities,and thus hydrogen-mass densities, in the inner and outer tails. Finally, in §
6, we give asummary of our results and present our conclusions. We assume for the calculation of ∆ θ an observing frequency of 1420 MHz ( λ = 21 cm), i.e., the principalfrequency of the Galactic Plane Surveys (e.g., Landecker et al. 2010; Haverkorn et al. 2006), and use valuesof 2 µ G and 1 pc for the line-of-sight component of the Galactic magnetic field and the path length throughthe tail, respectively.
2. THE PLANETARY NEBULA DeHt 5
DeHt 5 is an old PN (see Table 1) whose optical morphology clearly demonstrates inter-action with the ISM. Indeed, the narrow-band and emission-line images of Tweedy & Kwitter(1996) show that the intrinsic morphology of the PN may have been largely destroyed bythe PN-ISM interaction. For point of illustration, we show in Figure 1 the Digitized SkySurvey (DSS) R-band ( λ = 657 nm) optical image for the 1 ◦ × ◦ region centered on theDeHt 5 white dwarf central star, WD 2218+706. The brightest and most distinct portion ofthe PN is marked by the filamentary emission to the (Galactic) north of WD 2218+706. Theremainder of the PN is significantly more diffuse and amorphous. The 9 ′ -diameter dashedcircle in Figure 1 (and in subsequent figures featuring the radio polarization images) repre-sents the approximate visible extent of DeHt 5 (see Tweedy & Kwitter 1996), and, for lackof a better reference point, is centered on WD 2218+706. We emphasize, however, that the“visible disk” of DeHt 5 has neither a distinct circular shape nor is it necessarily centered onWD 2218+706. In fact, given the age of DeHt 5, it is very likely that WD 2218+706 is offsetfrom the center of the PN in the direction of motion. We therefore advise the reader to usethe disk outline as it appears in each figure only as a rough guide to the size and position ofDeHt 5. We can use the parallax, proper motion, and radial velocity of WD 2218+706 (seeTable 1) to estimate the UVW motion of DeHt 5 through the ISM, where the U componentis positive toward the Galactic center, V is positive in the direction of Galactic rotation,and W is positive toward the north Galactic pole (e.g., Johnson & Soderblom 1987). Aftercorrecting for a solar motion of (U ⊙ , V ⊙ , W ⊙ ) = (10 . ± . , . ± . , . ± .
38) km s − (Dehnen & Binney 1998), we find velocity components for WD 2218+706 of (U , V , W) =(+54 . ± . , − . ± . , − . ± .
81) km s − . The overall space velocity of DeHt 5through the ambient ISM is then 59 . ± .
70 km s − , with components on the sky and alongthe line of sight of 45 . ± . − and 37 . ± . − , respectively. The sky-projectedspace velocity of DeHt 5, including error cone, is illustrated in Figure 1 (optical image) aswell as Figures 2 and 3 (radio polarization images).The space velocity of DeHt 5 is consistent with the average value ( ∼
60 km s − ) given We do not take into account any motion from an interstellar cloud that may contain DeHt 5. If DeHt 5is indeed part of a cloud, then the quoted space velocity may be off by up to ∼ − (see Bannister et al.2001). ∼ §
3. OBSERVATIONS AND IMAGE PREPARATION
The radio polarization data presented in this paper were obtained at 1420 MHz ( λ =21 cm) as part of the CGPS (Taylor et al. 2003). The particular area of interest actually fallsin the high-latitude extension of the CGPS: a region between Galactic longitudes l = 100 ◦ and l = 117 ◦ for which the nominal coverage in Galactic latitude (i.e., − . ◦ < b < +5 . ◦ ) isincreased to − . ◦ < b < +17 . ◦ (see Landecker et al. 2010). The principal observing instru-ment for the radio component of the CGPS is the synthesis telescope (ST; see Landecker et al.2000) at the Dominion Radio Astrophysical Observatory (DRAO). The ST collects contin-uum data in four 7.5 MHz bands centered on 1406.65, 1414.15, 1426.65 and 1434.15 MHz,respectively. The ST is sensitive at 1420 MHz to emission from structures with angular sizesof ∼ ◦ down to the resolution limit of ∼ ′ . In Stokes-Q ( Q ) and Stokes-U ( U ), data fromtwo single-antenna surveys of the northern sky at ∼ ◦ × ◦ region of the CGPS presented in this paper is ∼ − ( ∼ Note that the frequency corresponding to the midpoint of the four continuum bands is 1420.4 MHz, theneutral hydrogen spin-flip frequency. The 5.0 MHz band about this frequency is allocated to the 256-channelspectrometer (see Taylor et al. 2003).
4. RADIO POLARIZATION IMAGES OF PN DeHt 5
In Figure 2 we show the polarized intensity ( P = p Q + U − (1 . σ ) ; e.g., Simmons & Stewart1985), where the last term gives explicitly the noise bias correction) and polarization angle( θ P = arctan U/Q ) images at 1420 MHz ( λ = 21 cm) for the 1 ◦ × ◦ region about DeHt 5(i.e., the same region presented optically in Figure 1). The images reveal narrow polariza-tion structures on and immediately (to the Galactic) north-east of the visible disk of thePN, as well as an extended polarization structure which runs east-northeast from the edgeof the visible disk to the edge of the displayed region. (Recall, visible disk, or “disk” as itis shortened to hereafter, refers to the approximate visible extent of DeHt 5; see § ◦ × ◦ region, indicates that their appearances are due to the effectsof Faraday rotation; namely localized beam depolarization of the background diffuse syn-chrotron emission and/or background-foreground cancellation (e.g., Gray et al. 1999). Thecoincidence of the narrow structures with the disk of DeHt 5 suggests that their origin is theshell of the PN (see § ◦ × ◦ region about DeHt 5. Theextended Faraday-rotation structure, noted above, is the most distinct feature in the region.It is ∼ ◦ long and ∼ ◦ wide, with a major-axis position angle (p.a.) of ∼ ◦ (northof east). The major axis of the structure is nearly (counter-)aligned with the projectedspace velocity of DeHt 5 (p.a. = 10 . ◦ ± . ◦ , south of west), suggesting that the origin ofthis structure may be a tail (labeled “thick tail” in Figure 3) of ionized material depositeddownstream from the PN. We estimate the RM of the thick tail in § § RM of these structures in § § Figure 2 a and a close-up view in Figure 4 a show a low-polarized-intensity ridge whichruns approximately east-to-west through the projected center of DeHt 5, and a low-polarized-intensity arc which approximately traces the north-east edge of DeHt 5. The mean polarizedintensities for the ridge and arc are 0 . ± .
044 K and 0 . ± .
039 K, respectively, or ∼ ∼ . ± .
047 K (see Table 2). In addition, eachof these structures has small beam-sized (i.e., ∼ ′ -diameter) areas for which the polarizedintensities are virtually zero. The low mean intensities and small-scale spatial variations inboth the ridge and arc point to beam depolarization (rather than background/foregroundcancellation) as the dominant depolarization mechanism in these structures. Moreover, beamdepolarization is consistent with the Faraday rotation for the ridge and arc occurring in theshell of DeHt 5: sharp RM gradients are caused by fluctuations (in projected position onthe sky) in either electron density or line-of-sight magnetic field strength (see Equation 2),each of which is expected in the turbulent shell of a PN (e.g., Ransom et al. 2008).Figure 2 b and a close-up view in Figure 4 b show that the polarization angles for theridge and arc are significantly rotated compared to the off-source value of − ◦ ± ◦ (seeTable 2). Moreover, the polarization angles at the outside edges of the ridge and arc (whiteor near-white pixels) appear to outline a single “bent” structure with clear boundaries onthe north-east side (arc) and south side (ridge). At approximately the midpoint between thearc and ridge (see Figure 4), the polarization angles plateau at − ◦ ± ◦ , a value differentfrom that seen both off-source (i.e., north-west, west, and south of the disk) and in thethick tail immediately north-east of the arc (see Table 2). Compared to the off-source value,the mean polarization angle in the “plateau” is rotated +19 ◦ ± ◦ . We infer positive, i.e.,counter-clockwise, rotation, since the alternative, namely a negative rotation of ∼ − ◦ ,would likely produce a clear (black-to-white) boundary in polarization angle at the northand north-west edges of the disk. No such boundary is seen. Also, the polarization anglesin our four ST bands (see §
3) do not reflect a RM ∼ ◦ “wraps” in polarization angle are excluded by the data.) The mean polarizedintensity over the plateau is 0 . ± .
045 K, a very modest reduction from the off-sourcevalue, which suggests for this small region that background/foreground cancellation is moreimportant than beam depolarization. If we assume that the Faraday rotation takes place inthe shell of DeHt 5, i.e., at a distance of 345 ±
20 pc, and use the values for the foregroundpolarized intensity and polarization angle given in § RM over the plateau region of +18 ±
12 rad m − . We comment on the viability of usingon-source/off-source polarization angles to estimate RM s in § § The ∼ ◦ × ∼ ◦ thick tail (see Figures 2 and 3) stands out noticeably in bothpolarized intensity and polarization angle relative to the (off-source) region near DeHt 5.The 0 . ± .
055 K mean polarized intensity in the thick tail is ∼ . ± .
047 K, and the − ◦ ± ◦ mean polarization angle is rotated ∼− ◦ compared to the off-source value of − ◦ ± ◦ (see Table 2). A negative, i.e., clockwise,rotation makes sense, given the absence of sharp polarized-intensity and polarization-angleboundaries at the edge of the thick tail, and is supported by our Faraday-rotation models(see below). While the reduction in mean intensity is more modest for the thick tail than forthe narrow features on the disk of DeHt 5, small beam-sized “knots” of near-zero polarizedintensity and rapid polarization-angle variation indicate the presence of sharp, localized RM gradients in the thick tail. As with the ridge and arc, beam depolarization probably definesthe knots.If we exclude the knots, the mean polarized intensity in the thick tail increases slightlyto 0 . ± .
039 K, or ∼ − ◦ ± ◦ , a rotation of − ◦ ± ◦ from the off-source value. The ∼
39% reduction in thepolarized intensity compared to that off-source, and relatively small standard deviationsabout the mean values in polarized intensity and polarization angle (compared the ridge andarc), suggest that, outside of the knots, background/foreground cancellation is of comparableimportance to beam depolarization. We constructed Faraday-rotation models using theapproach of Wolleben & Reich (2004). We treated the thick tail (excluding the knots) as aFaraday “screen” that rotates the polarization angle of the background emission. When therotated background is vector-added to the polarized foreground, the net polarized intensitydrops (compared to regions outside of the screen). For each model, we varied the RM anddegree of beam depolarization for the screen as well as the foreground polarized intensityand polarization angle. Our set of acceptable models (judged against the scatter in the on-source and off-source data) gives a RM over the thick tail of − ± − . The models 10 –also give a foreground polarized intensity of 0 . ± .
050 K and foreground polarizationangle of − ◦ ± ◦ , and indicate that beam depolarization accounts for 40% ±
10% of theintensity reduction in the thick tail. The foreground polarized intensity is consistent with the0.12–0.23 K value suggested by Roger et al. (1999) for the 345 ±
20 pc line of sight towardDeHt 5. Note that the polarization angle is very similar to the nominal off-source value.Estimating RM s by comparing on-source and off-source polarization angles is subjectto rather large uncertainties, particularly when the foreground toward the source of interestcontributes a significant fraction of the polarized emission ( & RM using (widely spaced) multi-frequency polarimetric data would be superior. Futurepolarimetric observations at subarcminute resolution and at multiple frequencies in the 1–3 GHz range could provide a detailed RM -map of the region around DeHt 5. The less conspicuous thin tail (see Figure 3) is comprised of three Faraday-rotationstructures, each roughly ∼ ◦ in diameter, and has a total length of ∼ ◦ . The meanpolarized intensity across all three structures is 0 . ± . ∼ . ± .
047 K (see Table 2). The meanpolarization angle across all three structures is − ◦ ± ◦ , a rotation of − ◦ ± ◦ from theoff-source value of − ◦ ± ◦ (see Table 2). We again infer a negative rotation, since thereare no sharp boundaries at the edges of these structures. If we assume that the Faradayrotation takes place at a distance of 345 ±
20 pc, and use the values for the foregroundpolarized intensity and polarization angle given in § Roger et al. (1999) give emissivity measurements at 22 MHz for lines of sight near DeHt 5. Reich & Reich(1988) show that the brightness temperature spectral index between 408 and 1420 MHz for the region aroundDeHt 5 is β = − . T B ∼ ν β ). If we assume that this value applies also to the frequency range 22–408 MHz,then (assuming the emission is ≈
70% polarized) the Roger et al. measurements give a foreground polarizedintensity of 0 . ± .
010 K. However, low-angular-resolution maps between 38 and 408 MHz suggest aslightly flatter index ( β = − .
5; Lawson et al. 1987) at lower frequencies. Applying β = − . β = − . . ± .
015 K. Though roughly consistent with zero, the − ◦ ± ◦ on-source/off-source rotation nevertheless yields thin-tail structures which stand out in contrast relative to the smoother off-source regions. The true standarderror for the rotation is likely smaller than 8 ◦ (computed as the root-sum-square of the standard deviationsof the polarization angles for the thin tail and off-source region), but is difficult to estimate in the absenceof a Faraday-rotation model.
11 –tail, we estimate by comparing off-source and on-source polarization angles a mean RM overthe thin tail of − ±
10 rad m − . We tried to check the consistency of the values for theforeground polarized intensity and polarization angle, but were not able, due to the relativelyweak depolarization signatures in the thin-tail structures, to solve simultaneously (using theapproach of Wolleben & Reich) for the RM and the foreground parameters.
5. DISCUSSION5.1. Evidence in Faraday Rotation of the Shell of DeHt 5
Given the west-southwest projected space velocity of the DeHt 5 central star, WD2218+706, we might expect to see evidence in the optical image (Figure 1) and radio polar-ization images (Figure 2) of a bow shock at the south-west edge of the disk of the PN. Aclose connection between enhanced optical emission and Faraday rotation is observed at theleading edge of at least one other old PN (Sh 2-216; Ransom et al. 2008). Since, however,both enhanced optical emission and Faraday rotation rely on column density of ionized ma-terial, and a large component of the space velocity of DeHt 5 lies along the line of sight (see § − ) show that “hot spots” in the nebular shell migrate overtime downstream from the leading-edge of the interaction, and sit in the more advancedstages of the PN-ISM interaction at approximately the halfway point between the upstreamand downstream sides of the shell (see Wareing et al. 2007a). If this is the case for the & b )may be the outline of a ring which forms a clear boundary with the surrounding regions on thenorth-east (arc) and south (ridge) sides. Indeed, a circular ring, aligned in three-dimensionsperpendicular to the space velocity of DeHt 5, would have a projected shape on the frontside (i.e., the side closest to Earth) similar to that formed by the arc and ridge. (We showin Figure 5 a two-dimensional slice through the ring described here on a plane perpendicularto the plane of the sky.) Furthermore, close inspection of the west and north-west edgesof the disk in both polarized intensity and polarization angle shows that there is a jagged,though very narrow, transition between the plateau and off-source region. There are three 12 –reasons why the ring boundary on the west side may be more subtle than that on the eastside: (1) The polarization angle difference between the plateau and thick tail (immediatelyeast of the disk) is larger than that between the plateau and off-source region immediatelywest of the disk. A smaller mean off-source to on-source rotation results in a lesser degreeof beam depolarization, and consequently a thinner and more subtle boundary. (2) Theremay be spherical (or circular in the case of a ring) asymmetries in the density of material inthe shell (Wareing et al. 2007a,b). (3) Spatial variations in the strength and/or orientationof the magnetic field at the back (west) side of the ring may be smaller in magnitude thanthose on the front (east) side. We elaborate on the magnetic field geometry in the shell ofDeHt 5 below.We estimate in the plateau region between the ridge and arc a mean RM of +19 ±
11 rad m − . Since no estimate is given in the literature for the electron density in the shellof DeHt 5, we cannot derive (see eq. [2]) the magnitude of the line-of-sight magnetic field inthe plateau. However, the sign of the RM indicates that the field in this central region isdirected out of the plane of the sky. If the field is ISM in origin (e.g., Ransom et al. 2008),we can explain simultaneously the out-of-the-sky component near the center of the disk ofDeHt 5 and the into-the-sky component discussed in § § § Are the Faraday-rotation structures behind DeHt 5 really the signatures of a tail of ion-ized material deposited downstream by the interaction between DeHt 5 (and its progenitor)and the ISM? Could the structures in fact trace the history of the interaction first betweenthe AGB wind and the ISM (thin tail) and second between the PN and the previously es-tablished AGB-ISM interaction zone (thick tail)? We look in § § The most compelling pieces of evidence that the ∼ ◦ × ∼ ◦ Faraday-rotation struc-ture is the signature of a PN-ISM tail is its north-east side attachment to DeHt 5 and theapproximate alignment of its major axis with the projected space velocity of WD 2218+706.It is very unlikely that a line-of-sight Faraday-rotation structure, unrelated to DeHt 5, wouldhave one end bounded by DeHt 5, and extend away from DeHt 5 in the direction from whichWD 2218+706 came. But why then is the alignment approximate, and not exact to withinthe error bars of the projected space velocity?There are two reasons why the position angle of the major axis of the thick tail mightdiffer by 21 . ◦ ± . ◦ from that of the projected space velocity of WD 2218+706. First, themeasured proper motion of WD 2218+706 may not represent that of the system. Both theHST (Benedict et al. 2009) and USNO (Harris et al. 2007) observations of WD 2218+706show signs of a binary companion. Using the example provided by Benedict et al. (2009), anM1 V star would be at the limit of detectability if separated from WD 2218+706 by 5.2 AU.The resulting 11 yr orbital period for this system would result in a transverse velocity ashigh as 14 km s − (assuming an orbital eccentricity of zero). If the transverse orbital motionwere approximately perpendicular to the system proper motion, the position angle of theprojected (instantaneous) motion of WD 2218+706 would differ from that of the systemmotion by ∼ ◦ . A more massive K8 V star with a separation of just 3.2 AU would also beat the limit of detectability (see Table 8 in Benedict et al. 2009), and would give a maximumposition-angle difference of ∼ ◦ . Second, the system may be accelerating toward a massive,extended molecular cloud complex. The cloud runs ∼ ◦ from ( l = 100 ◦ , b = 13 ◦ ) to( l = 117 ◦ , b = 22 ◦ ), and contains ∼ × M ⊙ of material (see Grenier et al. 1989). (Theelongated feature which runs diagonally across the top-right portion of Figure 3 may be theFaraday-rotation signature of this extensive cloud.) Distance estimates for the cloud range 14 –between ∼
300 pc (Grenier et al. 1989) and ∼
400 pc (Bally & Reipurth 2001), indicatingthat this mass is in the same region of the local arm as DeHt 5. We elaborate more on thegravitational effect of the molecular cloud on WD 2218+706 (and its progenitor) in § ∼ ◦ major-axis length of the thick tailmay itself provide evidence that this region is the PN-ISM tail of DeHt 5. At a distance of345 ±
20 pc, the angular length of the thick tail corresponds to a projected physical length of ∼ . ± . − projected velocity of WD 2218+706leads to an age for the PN-ISM interaction of ∼ − , but point out that the viscosity inthe tail may be underestimated. Taking the minimum value for the lag in the DeHt 5 tailto be ∼
10 km s − , we derive an upper limit for the interaction age of ∼ . & ∼ ◦ width of the thick tail is also consistent with the simulations of Wareing et al.(2007a), which show, for systems moving at speeds of 50–75 km s − , a PN-ISM tail width ∼ § ∼ If we accept that the thick tail is the signature of the PN-ISM interaction for DeHt 5,then the best evidence that the thin tail is the signature of the earlier AGB-ISM interactionis its relative position behind, and approximate alignment with, the thick tail. Again, theapproximate alignment requires some explanation. The three Faraday-rotation structureswhich make up the ∼ ◦ thin tail fall on a line with position angle ∼ ◦ (north of east). Thisline is ∼ ◦ more northerly than the major axis of the thick tail. The orbit we computedfor WD 2218+706 (see § ∼ § <
100 pc distant, the resulting acceleration would,over ∼ ∼
20 km s − , yieldingthe curved trajectory between the thin and thick tails seen in Figure 3. Indeed, such anacceleration would simultaneously explain the difference in position angles between the thinand thick tails and between the thick tail and projected space velocity of WD 2218+706.At a distance of 345 ±
20 pc, the ∼ ◦ angular length of the thin tail correspondsto a projected physical length of ∼ < ⊙ ) during the later stages of the TP-AGB phase, specificallythe last 2–3 thermal-pulse cycles (Vassiliadis & Wood 1993; Marigo & Girardi 2007). Indeed,the recent models of Marigo & Girardi (2007) show that the mass-loss rates are a factor 2 ormore higher for the last 2–3 pulses, and give a period for the final pulse cycles of 0.1–0.2 Myr.If the thin tail is the signature of the later stages of the TP-AGB phase, then the three-component Faraday-rotation structure suggests a time frame of 0.3–0.6 Myr, consistent withthe low end of the range suggested by the tail’s full length. A direct measurement of thevelocity of the gas in the thin tail would help to put observational constraints on the models. 16 – Based on the evidence and consistencies described in § RM in the thick tail (see § ∼ ◦ or ∼ l av ≈ . µ G. In the local arm, this field points toward l = 84 ◦ ± ◦ (seeBrown & Taylor 2001). At the midpoint of the thick tail ( l = 111 . ◦ ), the azimuthal fieldpoints into the sky, consistent with the sign of the RM , and makes a 27 ◦ ± ◦ angle with theline of sight (see Figure 5). Considering the range of possible combinations of the regularand random components of the Galactic magnetic field (see, e.g., Ransom et al. 2008) alongthis line, we estimate an uncertainty in the azimuthal field strength of 1.3 µ G. With thesevalues, we derive for the line-of-sight component of the ISM field in the region of the thicktail B k = 3 . ± . µ G. Since the azimuthal field lies largely perpendicular to the directionof motion, and the material in the thick tail is ionized (see below), the field is likely pulledforward slightly inside the tail creating a “magnetic wake” (see Figure 5). The result ofthis magnetic wake is a field that has on the near (i.e., Earth-facing) side of the thick taila line-of-sight component smaller than the intrinsic field, and on the far side a line-of-sightcomponent larger than the intrinsic field. The net effect for our purposes, assuming minimalcompression of the field lines within the tail, is a mean line-of-sight component essentially thesame as that quoted above for the intrinsic azimuthal field. Using RM = − ± − ,∆ l av ≈ . B k = 3 . ± . µ G, we derive for the electron density in the thick tail n e = 3 . ± . − .The gas in the thick tail is likely completely ionized (see Wareing et al. 2007a). Underthis assumption, the number density for protons in the thick tail is the same as that forelectrons: n H = 3 . ± . − . Using a cylindrical volume for the thick tail, with length ∼ ∼ . ± . × . This value corresponds to a total mass of0 . ± .
33 M ⊙ . If we assume that the ambient density in the region surrounding DeHt 5 is We used the midpoints of the thick tail in longitude and latitude to project the azimuthal field onto theline of sight. Note that, at a distance of 345 ±
20 pc, the thick tail sits ∼
70 pc above the Galactic plane,well within the predicted ∼
17 – n H ∼ − , then the mass ejected into the thick tail during the PN-ISM interaction stageis 0 . ± .
33 M ⊙ . Using the estimate of the mean RM given in § RM for the thin-tail components is quite large, we use only the nominalvalue ( RM ∼ − − ) for the following derivation, and caution the reader to view thegiven ionized mass value as only a first estimate.To estimate the path length through the thin tail, we approximate each component asa sphere with diameter equal to its projected diameter ( ∼ ◦ or ∼ l av ≈ . § ◦ angle with the line of sight at the midpoint of the thin tail ( l = 112 . ◦ ), yielding B k ≈ . µ G. Using RM ∼ − − , ∆ l av ≈ . B k ≈ . µ G, we derive for theelectron density in each of the thin-tail components n e ∼ . − .In contrast to the thick tail, the material in the thin tail is likely only partially ionized.(The ionization fraction is difficult to estimate, since we have no independent knowledge ofthe ionization age or density of the thin tail material.) Thus, we can consider the electrondensity to be only a lower limit on the mass density in the thin-tail components: n H & . − . Using a spherical volume with diameter ∼ & . × . This valuecorresponds to a total mass of & ⊙ . The mass ejected into the thin tail during theAGB-ISM interaction stage is & ⊙ , based on an ambient density of n H ∼ − . The stripped-mass estimates given above for the thick (0 . ± .
33 M ⊙ ) and thin( & ⊙ ) tails, though subject to some important assumptions and (especially for thethin tail) large uncertainties, indicate that a significant amount of material, originally in theenvelope of the WD 2218+706 progenitor, has been deposited downstream. This result isconsistent with hydrodynamic simulations (Villaver et al. 2003; Wareing et al. 2007a), andhas implications for (1) the PN missing-mass problem, and (2) models which describe the 18 –impact of the deaths of intermediate-mass stars on the ISM. We discuss each of these brieflybelow.Observations show that, on average, only 0.15 M ⊙ of material is present in ionizedform in the shells of PNe (see, e.g., Zhang 1995). For intermediate-mass stars with initialmasses in the 1–3 M ⊙ range (i.e., an initial mass for which a final core mass similar to thatof WD 2218+706 is plausible), this nebular mass represents .
40% of the mass lost duringAGB and post-AGB evolution (e.g., Bloecker 1995a). The “missing” &
60% may be partly inneutral form within the PN (see Aaquist & Kwok 1991), but the majority is likely in the tailsof material ejected during AGB-ISM and PN-ISM interactions. Future radio polarimetricobservations of the tail regions of DeHt 5 and other PNe may yield mass estimates accurateenough to account for all of the progenitor mass loss, and eliminate altogether the missing-mass problem for PNe.PNe are much smaller than SNRs: compare ∼ ∼
30 pc for the average diameter of a Galactic SNR(Kothes et al. 2006). In a one-to-one comparison, an SNR has ∼ ∼ ∼
21. If the tails behind DeHt 5 are representative of those behind other PNe, then thecontrast between the volume of the ISM influenced by the death of an intermediate-mass starand that of a high-mass star is somewhat reduced. Furthermore, since intermediate-massstars complete several orbits of the Galactic center, and oscillate on their orbits up and downthrough the Galactic plane, many regions of the Galactic ISM may have been touched bytheir dynamically active tails. A detailed model, which takes into account the initial massfunction as well as lifetimes and trajectories of disk stars, is necessary in order to fully assessthe relative importance of the deaths of intermediate and high-mass stars in the GalacticISM.
6. CONCLUSIONS
Here we give a summary of our results and conclusions: 19 –1. We have presented 1420 MHz polarization images for the 5 ◦ × ◦ region around thePN DeHt 5.2. Narrow Faraday-rotation structures trace approximately the north-east (arc) andsouth (ridge) edges of the disk of DeHt 5. These narrow features may represent an enhancedring of material in the shell of DeHt 5, which is aligned in three-dimensions perpendicular tothe space velocity of the white dwarf central star, WD 2218+706.3. A small region of relatively high polarized intensity (plateau) sits at approximatelythe midpoint of the disk of DeHt 5. We have estimated via comparison of on-source andoff-source polarization angles a RM through the plateau of +18 ±
12 rad m − . The positivesense for the RM indicates that the magnetic field in the plateau is directed out of the sky.4. A Faraday-rotation structure, ∼ ∼ > RM throughthe thick tail of − ± − . The negative sense for the RM indicates that the magneticfield is directed into the sky.6. We have presented a qualitative model which interprets the into-the-sky magneticfield in the region of the thick tail as the large-scale azimuthal field of the Galaxy. Theout-of-the-sky magnetic field near the center of the disk of DeHt 5 (see point 3 above) isnaturally explained in this model as the deflected and compressed azimuthal field.7. With the RM given above (point 5), we have estimated the electron density in thethick tail to be n e = 3 . ± . − . Assuming that the ambient density in the regionsurrounding DeHt 5 is n H ∼ − , the mass ejected into the thick tail during the PN-ISMinteraction is 0 . ± .
33 M ⊙ .8. A series of three roughly circular Faraday-rotation structures appear behind the thicktail, and have a collective length of ∼ RM through the thin tail of ∼− − . With this value for the RM , we have estimatedthe total ionized mass in the thick tail to be ∼ ⊙ . Assuming, again, that the ambientdensity is n H ∼ − , the mass ejected into the thin tail during the > & ⊙ . (We give a lower limit since the thin tail is likely only partiallyionized.) Targeted, polarimetric observations at multiple frequencies between 1 and 3 GHzwould yield a more precise estimate of the RM in the thin-tail components, and allow for amore rigorous derivation of the mass content.10. The discovery of the tails representing the PN-ISM and possibly the AGB-ISMinteractions for DeHt 5 (and its progenitor) has important implications for the PN missing-mass problem and models which describe the impact of the deaths of intermediate-mass starson both the composition and dynamics of the ISM. Future radio polarimetric observationsof other old PNe and late-stage AGB stars may provide additional insight.ACKNOWLEDGMENTS. We thank the anonymous referees for constructive reviewsof the paper and for comments helpful in the preparation of the final manuscript. RRRthanks the Dean of Science, Technology, and Health at Okanagan College for arrangingrelease time for research. The Canadian Galactic Plane Survey is a Canadian project withinternational partners, and has been supported by a grant from NSERC. The DominionRadio Astrophysical Observatory is operated as a national facility by the National ResearchCouncil of Canada. This research is based in part on observations with the 100-m telescope ofthe MPIfR at Effelsberg. The Second Palomar Observatory Sky Survey (POSS-II) was madeby the California Institute of Technology with funds from the National Science Foundation,the National Geographic Society, the Sloan Foundation, the Samuel Oschin Foundation, andthe Eastman Kodak Corporation. REFERENCES
Aaquist, O. B. & Kwok, S. 1991, ApJ, 378, 599Bally, J. & Reipurth, B. 2001, ApJ, 552, L159Bannister, N. P., Barstow, M. A., Holberg, J. B., & Bruhweiler, F. C. 2001, in AstronomicalSociety of the Pacific Conference Series, Vol. 226, 12th European Workshop on WhiteDwarfs, ed. J. L. Provencal, H. L. Shipman, J. MacDonald, & S. Goodchild , 105–+ 21 –Barstow, M. A., Bannister, N. P., Holberg, J. B., Hubeny, I., Bruhweiler, F. C., & Napi-wotzki, R. 2001, MNRAS, 325, 1149Benedict, G. F., Mc Arthur, B. E., Napiwotzki, R., Harrison, T. E., Harris, H. C., Nelan,E., Bond, H. E., Patterson, R. J., & Ciardullo, R. 2009, AJ, 138, 1969Bloecker, T. 1995a, A&A, 299, 755—. 1995b, A&A, 297, 727Borkowski, K. J., Sarazin, C. L., & Soker, N. 1990, ApJ, 360, 173Brown, J. C. & Taylor, A. R. 2001, ApJ, 563, L31Dauphole, B. & Colin, J. 1995, A&A, 300, 117Dehnen, W. & Binney, J. J. 1998, MNRAS, 298, 387Good, S. A., Barstow, M. A., Burleigh, M. R., Dobbie, P. D., & Holberg, J. B. 2005, MNRAS,364, 1082Gray, A. D., Landecker, T. L., Dewdney, P. E., Taylor, A. R., Willis, A. G., & Normandeau,M. 1999, ApJ, 514, 221Grenier, I. A., Lebrun, F., Arnaud, M., Dame, T. M., & Thaddeus, P. 1989, ApJ, 347, 231Han, J. L. & Qiao, G. J. 1994, A&A, 288, 759Harris, H. C., Dahn, C. C., Canzian, B., Guetter, H. H., Leggett, S. K., Levine, S. E.,Luginbuhl, C. B., Monet, A. K. B., Monet, D. G., Pier, J. R., Stone, R. C., Tilleman,T., Vrba, F. J., & Walker, R. L. 2007, AJ, 133, 631Haverkorn, M., Gaensler, B. M., McClure-Griffiths, N. M., Dickey, J. M., & Green, A. J.2006, ApJS, 167, 230Heiles, C. 1996, in Astronomical Society of the Pacific Conference Series, Vol. 97, Polarimetryof the Interstellar Medium, ed. W. G. Roberge & D. C. B. Whittet, 457–+Johnson, D. R. H. & Soderblom, D. R. 1987, AJ, 93, 864Kerber, F., Furlan, E., Rauch, T., & Roth, M. 2000, in Astronomical Society of the PacificConference Series, Vol. 199, Asymmetrical Planetary Nebulae II: From Origins toMicrostructures, ed. J. H. Kastner, N. Soker, & S. Rappaport, 313–+Kerber, F., Guglielmetti, F., Mignani, R., & Roth, M. 2002, A&A, 381, L9 22 –Kerber, F., Mignani, R. P., Guglielmetti, F., & Wicenec, A. 2003, A&A, 408, 1029Kothes, R., Fedotov, K., Foster, T. J., & Uyanıker, B. 2006, A&A, 457, 1081Landecker, T. L., Dewdney, P. E., Burgess, T. A., Gray, A. D., Higgs, L. A., Hoffmann,A. P., Hovey, G. J., Karpa, D. R., Lacey, J. D., Prowse, N., Purton, C. R., Roger,R. S., Willis, A. G., Wyslouzil, W., Routledge, D., & Vaneldik, J. F. 2000, A&AS,145, 509Landecker, T. L., Reich, W., Reid, R. I., Reich, P., Wolleben, M., Kothes, R., Uyanıker, B.,Gray, A. D., Del Rizzo, D., Furst, E., Taylor, A. R., & Wielebinski, R. 2010, ArXive-printsLawson, K. D., Mayer, C. J., Osborne, J. L., & Parkinson, M. L. 1987, MNRAS, 225, 307Marigo, P. & Girardi, L. 2007, A&A, 469, 239Martin, D. C., Seibert, M., Neill, J. D., Schiminovich, D., Forster, K., Rich, R. M., Welsh,B. Y., Madore, B. F., Wheatley, J. M., Morrissey, P., & Barlow, T. A. 2007, Nature,448, 780Matthews, L. D., Libert, Y., G´erard, E., Le Bertre, T., & Reid, M. J. 2008, ApJ, 684, 603Napiwotzki, R. 1999, A&A, 350, 101Napiwotzki, R. & Schoenberner, D. 1995, A&A, 301, 545Parker, Q. A., Acker, A., Frew, D. J., Hartley, M., Peyaud, A. E. J., Ochsenbein, F.,Phillipps, S., Russeil, D., Beaulieu, S. F., Cohen, M., K¨oppen, J., Miszalski, B.,Morgan, D. H., Morris, R. A. H., Pierce, M. J., & Vaughan, A. E. 2006, MNRAS,373, 79Ransom, R. R., Uyanıker, B., Kothes, R., & Landecker, T. L. 2008, ApJ, 684, 1009Reich, P. & Reich, W. 1988, A&A, 196, 211Reich, W., F¨urst, E., Reich, P., Uyanıker, B., Wielebinski, R., & Wolleben, M. 2004, in TheMagnetized Interstellar Medium, ed. B. Uyanıker, W. Reich, & R. Wielebinski, 45–50Roger, R. S., Costain, C. H., Landecker, T. L., & Swerdlyk, C. M. 1999, A&AS, 137, 7Sch¨onberner, D. & Bl¨ocker, T. 1996, Ap&SS, 245, 201Simmons, J. F. L. & Stewart, B. G. 1985, A&A, 142, 100 23 –Soker, N. & Dgani, R. 1997, ApJ, 484, 277Taylor, A. R., Gibson, S. J., Peracaula, M., Martin, P. G., Landecker, T. L., Brunt, C. M.,Dewdney, P. E., Dougherty, S. M., Gray, A. D., Higgs, L. A., Kerton, C. R., Knee,L. B. G., Kothes, R., Purton, C. R., Uyanıker, B., Wallace, B. J., Willis, A. G., &Durand, D. 2003, AJ, 125, 3145Tweedy, R. W. & Kwitter, K. B. 1996, ApJS, 107, 255Uyanıker, B., Landecker, T. L., Gray, A. D., & Kothes, R. 2003, ApJ, 585, 785Vassiliadis, E. & Wood, P. R. 1993, ApJ, 413, 641Villaver, E., Garc´ıa-Segura, G., & Manchado, A. 2003, ApJ, 585, L49Wareing, C. J., O’Brien, T. J., Zijlstra, A. A., Kwitter, K. B., Irwin, J., Wright, N., Greimel,R., & Drew, J. E. 2006a, MNRAS, 366, 387Wareing, C. J., Zijlstra, A. A., & O’Brien, T. J. 2007a, MNRAS, 382, 1233—. 2007b, ApJ, 660, L129Wareing, C. J., Zijlstra, A. A., O’Brien, T. J., & Seibert, M. 2007c, ApJ, 670, L125Wareing, C. J., Zijlstra, A. A., Speck, A. K., O’Brien, T. J., Ueta, T., Elitzur, M., Gehrz,R. D., Herwig, F., Izumiura, H., Matsuura, M., Meixner, M., Stencel, R. E., &Szczerba, R. 2006b, MNRAS, 372, L63Wolleben, M., Landecker, T. L., Reich, W., & Wielebinski, R. 2006, A&A, 448, 411Wolleben, M. & Reich, W. 2004, A&A, 427, 537Xilouris, K. M., Papamastorakis, J., Paleologou, E., & Terzian, Y. 1996, A&A, 310, 603Zhang, C. Y. 1995, ApJS, 98, 659
This preprint was prepared with the AAS L A TEX macros v5.2.
24 –Table 1. Properties of DeHt 5 and its Central Star WD 2218+706
Parameter Value ReferenceNebula PropertiesAngular Diameter (arcmin) 9 1Linear Diameter (pc) 0 . ± . a b . ± . c +19 − − ) 21 . ± . d − ) − . ± . e T eff (K) 76500 ± f ⊙ ) 0 . ± .
02 4Note. — Units of right ascension are hours, minutes, and seconds, and units ofdeclination are degrees, arcminutes, and arcseconds; mas ≡ milliarcseconds. a The linear diameter reflects the angular diameter given by Tweedy & Kwitter (1996)and the trigonometric parallax given by Benedict et al. (2009). b The kinematic age estimate is adjusted from Napiwotzki (1999) to the trigonometricparallax given by Benedict et al. (2009). c Estimate given is the weighted average of estimates from the Hubble Space Telescope(HST) and the United States Naval Observatory (USNO; see Harris et al. 2007). d Equatorial components for the proper motion: µ α = − . ± . µ δ = − . ± .
08 mas yr − . e A DAO classification is used in some references, since trace amounts of helium arepresent in the photosphere (see Barstow et al. 2001). f T eff estimates in the literature range between 57400 K (Barstow et al. 2001) and76500 K (Napiwotzki 1999).References. — 1. Tweedy & Kwitter 1996; 2. Napiwotzki 1999; 3. Kerber et al.2003; 4. Benedict et al. 2009; 5. Good et al. 2005; 6. Napiwotzki & Schoenberner1995.
25 –Table 2. On-Source and Off-Source Polarized Intensities and Polarization Angles
P.A. σ P . A . P.I. σ P . I . n Region ( ◦ ) ( ◦ ) K K(1) (2) (3) (4) (5) (6)——————–Disk Features——————–Ridge −
74 15 0 .
103 0 .
044 410Arc −
86 12 0 .
081 0 .
039 325Plateau (between Ridge and Arc) −
29 9 0 .
252 0 .
045 140——————–Tails——————–Thick −
72 8 0 .
162 0 .
055 2800Thick (excl. knots) −
66 6 0 .
182 0 .
039 1650Thin −
55 6 0 .
260 0 .
048 2600——————–Off-Source——————–Disk/Thick −
48 6 0 .
297 0 .
047 21500Thin −
48 6 0 .
309 0 .
047 14000Note. — Col. (1) See text for description of ridge, arc, plateau, thick tail,and thin tail. Off-source (Disk/Thick) refers to all pixels of the 1 ◦ × ◦ regioncentered on DeHt 5 (see Figure 2 or Figure 3) that are outside the disk ofDeHt 5 and not in the thick tail. Off-source (Thin) refers to all pixels within0 . ◦ of the thin tail (see Figure 3), but not in the thin tail. Col. (2) Meanpolarization angle in the respective region. (Note that polarization anglesare modulo 180 ◦ ; i.e., angles of − ◦ and +90 ◦ are equivalent.) Col. (3)Standard deviation in polarization angle. Col. (4) Mean polarized intensityin the respective region. Col. (5) Standard deviation in polarized intensity.Col. (6) Number of pixels used to estimate the mean polarization angle andpolarized intensity.
26 –Fig. 1.— R-band DSS optical image of a 1 ◦ × ◦ region centered on the position ofWD 2218+706, the central star of PN DeHt 5 (see Table 1). Here and hereafter, imagesare presented in Galactic coordinates, with Galactic north up and Galactic east to the left.The intensity scale is in photon counts with lighter shades indicating higher counts. Therange of intensities has been adjusted to highlight extended emission. The angular resolutionis ∼ ′′ . The dashed circle drawn on the image, and on each subsequent image, shows theapproximate extent of the “visible disk” (see §
2) of DeHt 5. The solid arrow and adjacentdotted lines on the image, and on the images presented in Figs. 2 and 3, indicate the pro-jected space velocity and space-velocity error cone of WD 2218+706. The length of the solidarrow represents the change in the sky position of WD 2218+706 over the next ≈ m V = 5 . a ) polarized intensity, P = p Q + U − (1 . σ ) , and ( b )polarization angle, θ P = arctan U/Q of the same 1 ◦ × ◦ region presented in Fig. 1. Theintensity scale is in brightness temperature in ( a ) and runs from 0 to 0 .
63 K, with lightershades indicating higher temperatures. The intensity scale in ( b ) extends from − ◦ (black)to +90 ◦ (white). Note that abrupt black-to-white transitions in ( b ) do not represent largechanges in angle, since polarization angles of − ◦ and +90 ◦ are equivalent. The resolvingbeam at the center of each image is 1 . ′ × . ′ (full-width at half-maximum; FWHM)oriented at a position angle (east of north) of − ◦ . The resolving beam varies in size overeach image by ∼ a ) polarized intensity and ( b ) polarization angle for a 5 ◦ × ◦ region around DeHt 5. The intensity scales are as described in Fig. 2. The approximateextents of the thick and thin tails described in the text are indicated. The three regionscomprising the thin tail are labeled 1–3, with the numbers increasing with distance fromDeHt 5. The length of the solid arrow represents (at this scale) the change in the skyposition of WD 2218+706 over the next ≈ Plateau PlateauRidgeRidge ArcArc
Fig. 4.— ( a ) Polarized intensity and ( b ) polarization angle images zoomed-in to a 0 . ◦ × . ◦ region around DeHt 5. The intensity scales are as described for Fig. 2. The disk featuresdiscussed in the text are labeled. 30 – F i e l d I n t r i n s i c To Earth Sp a ce V e l o c i t y A rc ComponentLine−of−Sight
Thick Tail P l a t e a u [From Above] KnotShell Enhanced Ring
Plane of the Sky
Sky Component L i n e o f S i g h t o Magnetic Wake Fig. 5.— Qualitative model showing the interaction between DeHt 5 and the magnetizedISM. The perspective is that of an observer sitting well above the Galactic plane and look-ing down on the center of DeHt 5 (with Galactic east to the left). The space velocity ofWD 2218+706, projected onto a plane parallel to the Galactic plane, is indicated, with line-of-sight and sky components. The intrinsic ISM magnetic field is inclined ≈ ◦ relativeto the line of sight. The field lines immediately surrounding DeHt 5 are compressed anddeflected around the leading edge of the shell of the PN. The deflection leads to a fieldwhich has a rapidly varying line-of-sight component on the east side of the disk (Arc) and amore uniform orientation, with small out-of-the-sky component, near the center of the disk(Plateau). Note that the arc is coincident with the front-side of the Enhanced Ring, whichis oriented roughly perpendicular to the space velocity (see § ∼∼