1: \documentclass[12pt,preprint]{aastex}
2:
3: \begin{document}
4:
5: \title{Neutral Hydrogen Absorption Toward XTE~J1810--197: the
6: Distance to a Radio-Emitting Magnetar}
7:
8: \author{
9: Anthony H. Minter,\altaffilmark{1}
10: Fernando Camilo,\altaffilmark{2}
11: Scott M. Ransom,\altaffilmark{3}
12: Jules P. Halpern,\altaffilmark{2}
13: and Neil Zimmerman\altaffilmark{2} }
14:
15: %\email{tminter@nrao.edu, fernando@astro.columbia.edu, sransom@nrao.edu,
16: %jules@astro.columbia.edu, and neil@astro.columbia.edu}
17:
18: \altaffiltext{1}{National Radio Astronomy Observatory, Green Bank,
19: WV 24944.}
20: \altaffiltext{2}{Columbia Astrophysics Laboratory, Columbia University,
21: New York, NY 10027.}
22: \altaffiltext{3}{National Radio Astronomy Observatory, Charlottesville,
23: VA 22903.}
24:
25: \def\magnetar{XTE~J1810--197}
26: \newcommand{\sgr}{\object{SGR~1806--20}}
27: \newcommand{\snr}{\object{G11.2--0.3}}
28: \newcommand{\darkcloud}{\object{G10.74--0.13}}
29: \newcommand{\wthirtyone}{\object{W~31}}
30: \newcommand{\gtentwo}{\object{G10.2--0.3}}
31: \newcommand{\gtenthree}{\object{G10.3--0.1}}
32: \newcommand{\gtensix}{\object{G10.6--0.4}}
33: \newcommand{\gtenzero}{\object{G10.0--0.3}}
34: \newcommand{\lbv}{\object{LBV~1806--20}}
35: \newcommand{\ourdistance}{3.1--4.3\,kpc}
36: \newcommand{\flatdist}{$3.4_{-0.7}^{+0.5}$\,kpc}
37: \newcommand{\weinerdist}{$4.0_{-0.8}^{+0.3}$\,kpc}
38: \newcommand{\englemaierdist}{$3.7\pm0.6$\,kpc}
39: \newcommand{\englemaierdistboth}{$3.3_{-0.2}^{+0.2}$\,kpc or $4.0_{-0.4}^{+0.3}$\,kpc}
40:
41: \newcommand{\nraoblurb}{The National Radio Astronomy Observatory is a
42: facility of the National Science Foundation operated under cooperative
43: agreement by Associated Universities, Inc.}
44:
45: \begin{abstract}
46: We have used the Green Bank Telescope to measure \ion{H}{1} absorption
47: against the anomalous X-ray pulsar \magnetar. Assuming a flat rotation
48: curve, we find that \magnetar\ is located at a distance of \flatdist.
49: For a rotation curve that incorporates a model of the Galactic bar,
50: we obtain a distance of \weinerdist. Using a rotation
51: curve that incorporates a
52: model of the Galactic bar and the spiral arms of the Galaxy, the distance
53: is \englemaierdist. These values are consistent with the distance
54: to \magnetar\ of about 3.3\,kpc derived from its dispersion measure,
55: and estimates of 2--5\,kpc obtained from fits to its X-ray spectra.
56: Overall, we determine that \magnetar\ is located at a distance of
57: $3.5\pm0.5$\,kpc, possibly not far in front of the infrared dark cloud
58: \darkcloud. We also used the GBT in an attempt to measure absorption
59: in the OH ${\rm ^2\pi_{3/2}(J=3/2)}$ lines against \magnetar. We were
60: unsuccessful in this, mainly because of its declining radio flux density.
61: Analysis of \ion{H}{1} 21\,cm, OH ${\rm ^2\pi_{3/2}(J=3/2)}$, and
62: $^{12}$CO$(2\rightarrow 1)$ emission toward \magnetar\ allows us to
63: place a lower limit of $N_{\rm H} \ga 4.6 \times 10^{21}$ cm$^{-2}$ on
64: the non-ionized hydrogen column density to \magnetar, consistent with
65: estimates obtained from fits to its X-ray spectra.
66:
67: \end{abstract}
68:
69: \keywords{ISM: clouds --- ISM: individual (G10.74--0.13) --- pulsars:
70: individual (XTE~J1810--197) --- radio lines: ISM }
71:
72: \section{Introduction}
73:
74: Prior to the detection of pulsed radio emission from the anomalous
75: X-ray pulsar (AXP) \magnetar\ by \citet{camilo}, pulsed emission
76: from the dozen known magnetars had been detected in X-rays in all
77: instances, and in one case at optical wavelengths. Estimating the
78: distance to a magnetar has relied on associating it with a supernova
79: remnant (SNR) of known distance, or by fitting its X-ray spectrum and
80: parameterizing the energy-dependent absorption by interstellar gas
81: along the line of sight with a non-ionized hydrogen column density
82: $N_{\rm H}$ \citep[][]{mccammon}. Using standard relations, $N_{\rm
83: H}$ is related to visual extinction $A_V$, which is turned into a
84: distance estimate using mean values in the Galactic plane \citep[see,
85: e.g., \S6 of][]{firstdist}, or is directly calibrated as a function of
86: distance using stars of known luminosity in the field \citep{redclump}.
87: Sometimes, the probable location of the X-ray source in a well-studied
88: star cluster can be used to infer its distance \citep{muno}.
89:
90: The unique detection of pulsed radio emission from \magnetar\ allows
91: one to determine its distance using methods that are not applicable to
92: other magnetars. The dispersion measure (DM, the total column density
93: of free electrons along the line of sight) was obtained upon discovery
94: of the radio pulsations \citep{camilo}. A model for the Galactic free
95: electron distribution then yields a distance estimate, in this case $d
96: \approx 3.3$\,kpc using the most recent model \citep{ne2001}. Also,
97: bright, pulsed radio emission allows kinematic distance limits to be
98: obtained by observing spectral lines that are seen in absorption against
99: the magnetar, as we report here using \ion{H}{1}.
100:
101: Obtaining a reliable distance to \magnetar\ allows for a precise
102: determination of the luminosity of the star based on its measured flux
103: in a variety of wavebands. The distance also allows a proper motion
104: \citep[see][]{helfand} to be converted into a tangential velocity.
105: Magnetars are thought to be
106: very young neutron stars, and are expected to be found near star
107: forming regions and/or spiral arms. Knowing the distance to \magnetar\
108: allows this prediction to be tested in this case. Besides providing a
109: kinematic distance, the \ion{H}{1} absorption spectrum can also give an
110: independent estimate of $N_{\rm H}$, which may be compared to results
111: from X-ray spectral fitting.
112:
113: In \S~2 we present the \ion{H}{1} and OH observations and data analysis.
114: This is followed by a determination of the hydrogen absorption spectra
115: toward \magnetar, in \S~3, and of its kinematic distance in \S~4.
116: In \S~5 we comment briefly on some features of the neutral hydrogen
117: toward \magnetar, and in \S~6 on the OH absorption limits. We obtain a
118: limit on the hydrogen column density in \S~7, and comment on models of
119: the line of sight toward \magnetar\ in \S~8. We conclude in \S~9 with
120: a discussion of our main results.
121:
122: \section{Observations and Data Analysis}
123:
124: \subsection{\ion{H}{1} 21\,cm Absorption Observations}
125:
126: \magnetar\ was observed with the National Radio Astronomy Observatory
127: (NRAO\footnote{\nraoblurb}) Robert C. Byrd Green Bank Telescope (GBT)
128: for approximately 3 hr on 2006 June 6 in order to measure the \ion{H}{1}
129: 21\,cm line absorption. The GBT has an unblocked aperture, a spatial
130: resolution of $9\farcm2$ at 21\,cm, and a system temperature on cold
131: sky of 18\,K. The NRAO spectral processor (SP), an FFT spectrometer,
132: was used as the detector. The SP provides good dynamic range for the
133: observations via its 32-bit sampling, and was used in its pulsar mode
134: whereby it accumulates spectra that are folded synchronously with the
135: pulsar period. The SP was configured to have 1024 channels for each
136: linear polarization across a bandwidth of 2.5\,MHz, producing a spectral
137: resolution of 0.52\,km\,s$^{-1}$ per channel. We recorded 128 spectra
138: (phase bins) evenly spaced in time across each individual period of
139: \magnetar. The duty cycle of this 5.54\,s pulsar was such that 4--5 of
140: these bins contained pulsed flux.
141:
142: \subsection{OH ${\rm\bf ^2\pi_{3/2}(J=3/2)}$ Absorption Observations}
143:
144: \magnetar\ was observed with the GBT for approximately 38\,hr in order
145: to measure absorption in the 1612, 1665, 1667 and 1720\,MHz ${\rm
146: ^2\pi_{3/2}(J=3/2)}$ ground state transitions of OH. Observing details
147: are given in Table~\ref{table:absobs}. The GBT's angular resolution
148: is approximately $8'$ for each of the OH transitions. The spectral
149: processor was also used for these measurements and was configured to
150: provide 256 spectral channels for each linear polarization across each
151: of the four OH transitions. A bandwidth of 0.625\,MHz was used, giving
152: a spectral resolution of 0.44\,km\,s$^{-1}$ per channel. We chose this
153: narrower bandwidth (with higher frequency resolution) since we already
154: knew which velocities displayed \ion{H}{1} absorption. As with the
155: \ion{H}{1} measurements, the pulsed flux of \magnetar\ was detected in
156: 4--5 phase bins.
157:
158: The GBT auto-correlation spectrometer (ACS) was used to obtain pulsar
159: ``off'' spectra for the four OH lines. The data were obtained in six
160: position-switching observations, each consisting of two minutes on source
161: followed by two minutes off source, resulting in an effective integration
162: time of 12 minutes. An off position two minutes of time in R.A. offset
163: from the position of \magnetar\ was used so that approximately the same
164: hour-angle coverage was obtained for the on and off positions. The ACS
165: was configured to observe all four of the OH ${\rm ^2\pi_{3/2}(J=3/2)}$
166: lines simultaneously with 8192 channels within a bandwidth of 12.5\,MHz,
167: resulting in a spectral resolution of 0.27\,km\,s$^{-1}$.
168: These observations allow for a better calibration of the pulsar ``off''
169: OH spectra.
170:
171: \subsection{Data Reduction}
172:
173: The data were analyzed using a method similar to that described
174: in \cite{weisberg} and \cite{minter}. For each individual pulse
175: of \magnetar\ and for each spectral line we measured the flux
176: $T(\nu,\phi,\mathcal{P})$ in terms of the equivalent brightness
177: temperature $T$, versus frequency $\nu$, phase $\phi$ across the
178: pulse (equivalent to time), and polarization $\mathcal{P}$. For each
179: polarization and pulse, the data were divided into two separate parts,
180: a pulsar ``on'' spectrum and a pulsar ``off'' spectrum:
181: \begin{eqnarray}
182: {T_{\rm on}(\nu,\mathcal{P})} & = & {{\sum_{\phi_{\rm on}} \left( T(\nu,\phi_{\rm on},\mathcal{P})
183: \left[ { \left< T(\nu, \phi_{\rm on}, \mathcal{P})\right>_{\nu({\rm noabs})} \over T_{\rm sys} } \right]^2 \right)
184: \over
185: \sum_{\phi_{\rm on}} \left[ { \left< T(\nu, \phi_{\rm on}, \mathcal{P}) \right>_{\nu(\rm noabs)}
186: \over T_{\rm sys}}
187: \right]^2} } \\
188: & & \nonumber \\
189: & & \nonumber \\
190: {T_{\rm off}(\nu,\mathcal{P})} & = & { { \sum_{\phi_{\rm off}}
191: T(\nu,\phi_{\rm off},\mathcal{P}) \over N_{\phi_{\rm off}}} }
192: \end{eqnarray}
193: where ${\phi_{\rm on}}$ are the phase bins when the pulsar is on,
194: ${\phi_{\rm off}}$ are the phase bins when the pulsar is off,
195: ${N_{\phi_{\rm off}}}$ is the total number of phase bins when the
196: pulsar is off and ${\rm \left< \right>_{\nu({\rm noabs})}}$ denotes
197: averaging over the frequencies that do not show absorption in the final
198: spectrum\footnote{An iterative approach in the data reduction is necessary
199: in order to determine ${\rm \nu(noabs)}$.}. A pseudo-absorption spectrum
200: for the $i$th pulse is then formed by taking the difference between the
201: pulsar on and pulsar off spectra,
202: \begin{displaymath}
203: T_{\rm on}^i({\nu}, \mathcal{P}) - T_{\rm off}^i({\nu}, \mathcal{P})
204: =
205: \left( T_{\rm p}^i(\nu, \mathcal{P})e^{-\tau(\nu)} + T_{\rm H~I}({\nu}) + T_{\rm sys}
206: \right) - \left( T_{\rm H~I}({\nu}) + T_{\rm sys} \right)
207: \end{displaymath}
208: \begin{equation}
209: =
210: T_{\rm p}^i(\nu, \mathcal{P}) e^{-\tau(\nu)}
211: \end{equation}
212: where ${T_{\rm p}}$ is the brightness temperature of the pulsar. We then
213: take a weighted average of the pseudo-absorption spectrum
214: \begin{equation}
215: {\left< T_{\rm p}(\nu, \mathcal{P}) \right> e^{-\tau(\nu)} =
216: { \sum_i \left( T_{\rm p}^i(\nu, \mathcal{P}) e^{-\tau(\nu)}
217: \left[ { \left< T_{\rm p}^i(\nu, \mathcal{P}) \right>_\nu({\rm noabs}) \over T_{\rm sys} }
218: \right]^2 \right)
219: \over \sum_i \left[ { \left< T_{\rm p}^i(\nu, \mathcal{P}) \right>_\nu({\rm noabs}) \over
220: T_{\rm sys} } \right]^2 }} \label{eq:wght}
221: \end{equation}
222: and store the weights for later use when the polarizations are averaged
223: together. This weighting is proportional to the signal-to-noise
224: ratio obtained for each pulse of \magnetar. A third-order orthogonal
225: polynomial was fitted to the $\left<T_{\rm p}(\nu, \mathcal{P})\right>
226: e^{-\tau(\nu)}$ spectrum in order to determine the intrinsic pulsar
227: brightness, $T_{\rm p}^{\rm fit}(\nu, \mathcal{P})$, at all frequencies,
228: i.e., by extrapolating $\left<T_{\rm p}(\nu, \mathcal{P})\right>$ across
229: the absorption features. This yielded the absorption spectrum
230: \begin{equation}
231: e^{-\tau({\nu})} =
232: { \left< T_{\rm p}(\nu, \mathcal{P}) \right> e^{-\tau(\nu)} \over
233: T_{\rm p}^{\rm fit}({\nu, \mathcal{P}}) }
234: \end{equation}
235: for each polarization. The absorption spectra for the
236: different polarizations were then averaged using the weights from
237: equation~(\ref{eq:wght}) to create the final, measured absorption spectrum
238: for each spectral line.
239:
240: For the \ion{H}{1} data, the pulsar off spectrum (i.e., the normal
241: \ion{H}{1} emission spectrum) was converted from detector counts to a
242: Kelvin scale using a calibrated noise diode that was injected during
243: two-minute calibration scans. Observations of the IAU \ion{H}{1}
244: standard source S6 were also made to put the pulsar off spectrum on the
245: IAU standard brightness temperature scale.
246:
247: \section{The Absorption Spectra Toward XTE~J1810--197}
248:
249: The \ion{H}{1} absorption spectrum toward \magnetar\ is shown in
250: Figure~\ref{fig:hiabs}. \magnetar\ is highly linearly polarized
251: \citep{xtepol} so that the YY polarization signal had a much better
252: signal-to-noise ratio (by a factor of $\approx 4$) than the XX
253: polarization. The two polarizations were averaged together with weights
254: given by their signal-to-noise ratios to produce the spectrum shown.
255: The spectrum has not been smoothed in any way and shows the native
256: resolution of the observations.
257:
258: Five Gaussians were fitted to the opacities determined from the \ion{H}{1}
259: absorption toward \magnetar. The results of the fit are shown in
260: Table~\ref{table:higaussians} and in Figure~\ref{fig:gaussians}.
261: The columns in Table~\ref{table:higaussians} give for each line,
262: respectively, the opacity, Local Standard of Rest velocity ($V_{\rm
263: LSR}$), and full-width at half-maximum (FWHM). The errors listed in
264: Table~\ref{table:higaussians} are $1\,\sigma$ errors from the Gaussian
265: fits.
266:
267: The $V_{\rm LSR} = 7.7$\,km\,s$^{-1}$ absorption line is associated
268: with the Heeschen Cloud \citep{riegel}, which has also been called the
269: Riegel--Crutcher Cloud by \citet{rccloud}. The Heeschen Cloud is a nearby,
270: $d \approx 125\pm25$\,pc, cold cloud ($T_{\rm spin} \approx 40$\,K)
271: that covers Galactic longitudes $345^\circ$ to $25^\circ$ and latitudes
272: $\pm6^\circ$. This cloud is also seen as a self-absorption feature in
273: the \ion{H}{1} emission spectrum (top panel of Fig.~\ref{fig:hiabs}).
274: The weak absorption line at $V_{\rm LSR} = 14.1$\,km\,s$^{-1}$ can be
275: kinematically associated with gas in the Carina--Sagittarius spiral arm
276: assuming that the line arises on the near side of the tangent point and
277: using the Galactic rotation model of \cite{englemaier}. Likewise, the
278: $V_{\rm LSR} = 22.8$ and 25.7\,km\,s$^{-1}$ lines can be kinematically
279: associated with the Crux--Scutum spiral arm.
280:
281: The OH absorption spectra toward \magnetar\ are shown in
282: Figures~\ref{fig:oh1612}--\ref{fig:oh1720}. With the strength of the
283: pulsar emission during the OH measurement epochs being much weaker than
284: was the case for the \ion{H}{1} measurement epoch \citep{fluxvarying},
285: the strong linear polarization of \magnetar\ means that only one
286: polarization effectively contributed to the measurement of the OH
287: absorption against \magnetar. Thus, the OH absorption data shown
288: in Figures~\ref{fig:oh1612}--\ref{fig:oh1720} and discussed in this
289: paper are only from a single linear polarization (YY). The $1\sigma$
290: opacity limits for any OH absorption toward \magnetar\ are given in
291: Table~\ref{table:ohlimits}.
292:
293: \section{Determining the Kinematic Distance to XTE~J1810--197}
294:
295: \subsection{Near or Far Side of the Velocity--Distance Relationship?}
296:
297: \magnetar\ is located at Galactic coordinates $(l,b) = 10\fdg726,
298: -0\fdg158$. At this Galactic longitude the kinematic velocity--distance
299: relationship is double-valued (see Fig.~\ref{fig:flat}). Our first
300: concern is whether it is possible to determine on which side of the
301: tangent point \magnetar\ lies. At this Galactic longitude, the tangent
302: point is about 8.3\,kpc from the Sun and has a $V_{\rm LSR} \approx
303: 168$\,km\,s$^{-1}$ using the flat rotation curve of \cite{butler}.
304: Distance estimates for \magnetar\ based on its X-ray-fitted $N_{\rm H}$
305: are 2.5--5\,kpc \citep{firstdist, gotthelf, redclump}, and the DM-based
306: distance is 3.3\,kpc \citep{camilo}. These strongly suggest that
307: \magnetar\ lies on the near side of the tangent point.
308:
309: Another magnetar, \sgr, which lies $41'$ from \magnetar, has \ion{H}{1}
310: absorption at velocities greater than $V_{\rm LSR} = 50$\,km\,s$^{-1}$
311: and is thought to lie on the far side of the tangent point \citep{hiabs}.
312: Both \sgr\ and the Galactic SNR \snr\ \citep{becker} also have \ion{H}{1}
313: absorption features at negative velocities that must arise from the
314: far side of the Galactic disk if the rotation curve is flat. However,
315: as the Galactic rotation model of \cite{sellwood} shows, these negative
316: velocities can also arise in the Galactic bar at a distance of 7\,kpc
317: toward these sources. The Galactic bar induces large radial motions
318: that deviate from the normally assumed circular Galactic rotation.
319: Along the line of sight to \magnetar, the Galactic bar is responsible for
320: motions $> 90$\,km\,s$^{-1}$ more negative than the prediction of a flat
321: rotation curve (see Fig.~\ref{fig:weiner}). Negative-velocity \ion{H}{1}
322: absorption is not found toward \magnetar, which according to this model
323: establishes an upper-limit to its distance of 7\,kpc (i.e., it is closer
324: than the Galactic bar). This strengthens the conclusion that \magnetar\
325: lies on the near side of the tangent point.
326:
327: \subsection{Is the Last Absorption Feature at the Actual Distance or a
328: Lower Limit?}
329:
330: The number of \ion{H}{1} absorption features per kpc in the inner Galaxy
331: is about 1--3 \citep[][see their Table~3]{garwood}. Assuming a flat
332: rotation curve and a distance of 5\,kpc, corresponding to a maximum
333: $V_{\rm LSR} \approx 40$\,km\,s$^{-1}$, along with a line width
334: of 3\,km\,s$^{-1}$ for the average \ion{H}{1} absorption feature,
335: we expect that 38\%--100\% of the velocities within $V_{\rm LSR} =
336: 0$--40\,km\,s$^{-1}$ should contain \ion{H}{1} absorption along the line
337: of sight toward \magnetar. If $\sim 38$\% of the velocity space were
338: occupied by absorption features, then the highest velocity feature in
339: the absorption spectrum toward \magnetar\ should be taken as indicating
340: a lower limit for the distance, since there is likely a significant
341: distance between \magnetar\ and the \ion{H}{1}-absorbing cloud nearest it.
342: If on the other hand the whole velocity range should contain \ion{H}{1}
343: absorption, it is likely that the last absorbing feature gives the actual
344: distance to \magnetar.
345:
346: It is useful to look at absorption measurements toward objects near
347: \magnetar\ in order to help us determine the \ion{H}{1} absorption
348: velocity coverage. The line of sight to \magnetar\ lies near that to
349: the giant Galactic \ion{H}{2} complex \wthirtyone\ \citep[see Fig.~1
350: of][]{corbel}. \wthirtyone\ is comprised of several individual \ion{H}{2}
351: regions including \gtentwo, \gtenthree, and \gtensix. Roughly on
352: the opposite side of \wthirtyone\ from \magnetar\ lies \gtenzero, the
353: wind-blown bubble of \lbv, and \sgr. At higher Galactic longitude than
354: \magnetar\ lies the SNR \snr. All of these objects have \ion{H}{1}
355: absorption measurements, except for \gtenzero\ and \lbv, which have
356: ${\rm NH_3}$ absorption measurements \citep[][]{corbel}.
357:
358: \citet[][]{becker} claimed that SNR~\snr\ lies well on the far side of the
359: tangent point. However, \citet[][]{green} noted the lack of absorption
360: between $V_{\rm LSR} = 45$\,km\,s$^{-1}$ and the tangent point velocity,
361: which led them to assign a distance of $\approx 5$\,kpc to \snr, while
362: attributing weak absorption at negative velocities to peculiar motions
363: in local gas. A distance of $\approx 5$\,kpc is also in reasonable
364: agreement with the expected expansion of the SNR \citep{green}. If we
365: take the Galactic bar into account using the model of \citet{sellwood},
366: we can reconcile the \ion{H}{1} absorption with the expansion/age-derived
367: distance estimate of \citet{green}. The \ion{H}{1} absorption to \snr\
368: at $V_{\rm LSR} \approx -20$\,km\,s$^{-1}$ arises from the Galactic bar,
369: which is at a distance of $\approx 5.5$--7.0\,kpc along this line of sight
370: (see Fig.~\ref{fig:weiner}). With both the Galactic bar and \snr\ being
371: on the near side of the tangent point toward \snr, \ion{H}{1} absorption
372: is not expected at $V_{\rm LSR} > 45$\,km\,s$^{-1}$. So we judge that
373: \snr\ likely lies within the Galactic bar at $\approx 5.5$--7\,kpc.
374:
375: \citet[][]{corbel} determined that \gtenthree, \gtenzero, \lbv, and
376: \sgr\ all lie on the far side of the tangent point, while \gtentwo\ and
377: \gtensix\ lie on the near side. The \ion{H}{1} absorption of \gtenthree\
378: \citep[Fig.~2 of][]{kalberla} and \sgr\ \citep[Fig.~2 of][]{cameron}
379: fill the range $V_{\rm LSR} = 0$--40\,km\,s$^{-1}$. \ion{H}{1} toward
380: \gtensix\ shows absorption at all velocities within $V_{\rm LSR} =
381: 0$--45\,km\,s$^{-1}$ \citep[Fig.~32 of][]{caswell}. Likewise, the
382: \ion{H}{1} absorption toward \gtentwo\ \citep[Fig.~4 of][]{greisen} and
383: SNR \snr\ \citep[Fig.~3 of][]{becker} completely cover the velocity range
384: $V_{\rm LSR} = 0$--45\,km\,s$^{-1}$. Since \gtensix, \gtentwo\ and \snr\
385: all lie on the near side of the tangent point and are only $16\arcmin$,
386: $33\arcmin$ and $30\arcmin$, respectively, in projection from \magnetar,
387: we argue that all velocities between the Sun and \magnetar\ should show
388: \ion{H}{1} absorption. Therefore, the highest-velocity absorption feature
389: seen along the line of sight toward \magnetar\ gives us its true distance,
390: rather than a lower limit.
391:
392: \subsection{The Kinematic Distance to XTE~J1810--197 }
393:
394: For the kinematic distance models that we discuss we will assume that
395: the Sun is located 8.5\,kpc from the Galactic center. We will also use
396: a velocity of $V_\sun=220$\,km\,s$^{-1}$ as the azimuthal velocity of
397: the LSR about the Galactic center. These are the current IAU standards.
398: However, there is evidence that these values may not be correct \citep[see
399: the discussions in][]{gcdist,englemaier}, which would require a scaling
400: of the kinematic distance estimates presented below.
401:
402: In Figure~\ref{fig:flat} we show the determination of the distance
403: to \magnetar\ using the flat rotation curve of \citet[][]{butler}.
404: Cold \ion{H}{1} clouds have random motions superposed on the uniform
405: Galactic rotation as indicated by their measured velocity dispersion
406: of 7\,km\,s$^{-1}$ \citep{jay}. The dotted lines on either side of
407: the rotation model in Figure~\ref{fig:flat} indicate the deviation from
408: Galactic rotation of $\pm 7$\,km\,s$^{-1}$. (The random motions of the
409: clouds are relative to the Galactic rotation and not to the measured
410: velocity of the cloud. Thus, the 7\,km\,s$^{-1}$ represents an error
411: in converting the model's distance into a velocity and not in using
412: the measured velocity of the cloud to obtain a distance as is usually
413: presumed. This distinction can be quite important in determining the
414: distance and its errors when the observed velocity is at or near a
415: local maximum or minimum in the model. However, in most cases it makes
416: only a minor difference.) The arrow at a constant $V_{\rm LSR}$ in
417: Figure~\ref{fig:flat} indicates the velocity of the highest velocity
418: \ion{H}{1} absorption. The shaded regions in Figure~\ref{fig:flat}
419: indicate the allowed kinematic distances from the flat rotation curve.
420: Figures~\ref{fig:weiner} and \ref{fig:englemaier} are similar to
421: Figure~\ref{fig:flat}, using instead the Galactic rotation models
422: of \citet[][]{sellwood} and \citet[][]{englemaier}\footnote{We have
423: scaled the \citet[][]{englemaier} model for $R_\sun=8.5$\,kpc.},
424: respectively.
425:
426: For the flat rotation curve used in Figure~\ref{fig:flat}, the last
427: \ion{H}{1} absorption feature toward \magnetar\ at $V_{\rm LSR} =
428: 25.7$\,km\,s$^{-1}$ (see Table~\ref{table:higaussians}) gives a kinematic
429: distance of \flatdist\ or $13.3_{-0.5}^{+0.7}$\,kpc. We can rule out
430: the larger distance since this is on the far side of the tangent point.
431: Although a flat rotation model may provide a reasonable distance estimate
432: for \magnetar, there are Galactic structures along the line of sight that
433: also need to be taken into account. Spiral arms can have a substantial
434: effect on the expected Galactic rotational velocities \citep[][]{roberts}.
435: The Galactic bar may also have an influence by adding non-circular
436: motions to the general Galactic rotation.
437:
438: The rotation model of \cite{sellwood} used in Figure~\ref{fig:weiner}
439: takes into account the effects of a strong potential associated with the
440: Galactic bar. This model uses a global Galactic gravitational potential
441: to determine the motions of the gas in the ISM. It uses both the
442: minimum and maximum observed \ion{H}{1} velocities along lines of sight
443: in the inner Galaxy in fitting the properties of the Galactic potential.
444: We can again immediately rule out the distance on the far side of the
445: tangent point. Since no \ion{H}{1} absorption is seen at velocities
446: between 30 and 40\,km\,s$^{-1}$ (see Table~\ref{table:higaussians}) we
447: can eliminate all distances except \weinerdist. The \citet[][]{sellwood}
448: model, however, does not take into account the effects of spiral arms.
449: Since magnetars are expected to lie in or near spiral arms this could have
450: a significant impact on the velocity-determined distance to \magnetar.
451:
452: The rotation model of \cite{englemaier} used in
453: Figure~\ref{fig:englemaier} takes into account the effects of a
454: potential associated with the Galactic bar as well as those associated
455: with spiral arms. The \citet[][]{englemaier} model was determined
456: in a similar fashion as the \citet[][]{sellwood} model. However, in
457: the \citet[][]{englemaier} model only the maximum (minimum) observed
458: velocities for CO were used for positive (negative) Galactic longitudes.
459: It should be noted that CO observations do not trace column densities as
460: low as \ion{H}{1} does. Combined with fitting to only one side of the
461: velocity--longitude data, this means that the \citet[][]{englemaier}
462: model uses a weaker potential for the Galactic bar compared with the
463: \citet[][]{sellwood} model. In fact, the \citet[][]{englemaier} model
464: predicts that there should be no negative velocity gas along the line
465: of sight toward \magnetar, while the \citet[][]{sellwood} model does.
466: As can be seen in the Southern Galactic Plane Survey \ion{H}{1} data
467: \citep{sgpshi}, there is plenty of gas at negative velocities along this
468: line of sight. We suggest that the \citet[][]{englemaier} Galactic bar
469: potential is too weak and that the \citet[][]{sellwood} model should be
470: used for velocities and distances associated with the Galactic bar.
471:
472: We can, however, still use the \citet[][]{englemaier} model to determine
473: a distance for \magnetar\ since it is likely located in front of the
474: Galactic bar as is evidenced by the \citet[][]{sellwood} model results
475: from above. In fact, comparing the \citet[][]{englemaier} result with
476: the \citet[][]{sellwood} result provides an indication of how spiral arms
477: may be affecting the velocity--distance relationship toward \magnetar.
478: We can again rule out distances that lie on the far side of the tangent
479: point and distances that include 30--40\,km\,s$^{-1}$ gas along the line
480: of sight. As can be seen from Figure~\ref{fig:englemaier}, we obtain
481: kinematic distances of \englemaierdistboth. Due to uncertainties in the
482: modeling of the Galactic rotation curve, the $0.1$\,kpc gap in distance
483: between these two values is in effect negligible. We thus combine these
484: two possible distance ranges into a single value, \englemaierdist.
485:
486: All of the above Galactic rotation models are derived from empirical
487: models fit to observed data. There is one model that is fully derived
488: from observational data, that of \citet[][]{brand}. In this model,
489: observations of \ion{H}{2} regions are used. The velocities of
490: the \ion{H}{2} regions are determined from recombination lines and
491: \ion{H}{1} absorption. Independent measurements provide distances to the
492: \ion{H}{2} regions. Using the rotation model of \citet{brand}, shown
493: in Figure~\ref{fig:brand}, we obtain $d = 2.4\pm0.5$\,kpc. However,
494: this model does not have enough \ion{H}{2} regions at $d \ga 2$\,kpc
495: in the general direction of \magnetar\ to be able to provide a good
496: velocity--distance relation \citep[see][]{brand}, so we give it little
497: weight.
498:
499: We conclude that the best estimate of a kinematically determined
500: distance to \magnetar\ is provided by the \citet[][]{sellwood} and
501: \citet[][]{englemaier} models, giving $d=$ \ourdistance.
502:
503: \section{The Neutral Hydrogen Toward XTE~J1810--197: GBT Versus SGPS}
504:
505:
506: The GBT has a resolution of $9\farcm2$ at 21\,cm while the Southern
507: Galactic Plane Survey \citep[SGPS,][]{sgpshi} has a resolution of about
508: $3\farcm3$. In Figure~\ref{fig:gbtsgps} we compare the GBT \ion{H}{1}
509: spectrum with the SGPS spectrum. For the \ion{H}{1} self-absorption
510: associated with the Heeschen Cloud we find the remarkable result that the
511: line width of the absorption increases with increasing spatial resolution!
512: To our knowledge this effect has never been observed before for \ion{H}{1}
513: on arc-minute resolutions.
514:
515: Inspection of the SGPS \ion{H}{1} data cube shows that there is structure
516: within the \ion{H}{1} emission inside of the GBT beam. We convolved
517: the SGPS data with a beam the size of GBT's, so that the two data
518: sets would have the same spatial resolution. As can be seen from
519: Figure~\ref{fig:gbtsgps} the GBT spectrum and $9\farcm2$ resolution SGPS
520: spectrum are in very good agreement. If we compare the line width of the
521: \ion{H}{1} absorption due to the Heeschen Cloud at $\sim 8$\,km\,s$^{-1}$
522: with the SGPS data having \ion{H}{1} absorption seen against \magnetar,
523: we find that the line widths are identical.
524:
525: These properties suggest that there is definite spatial structure in
526: the \ion{H}{1} emission on size scales of $3\farcm3$ to $9\farcm2$
527: (0.12--0.33\,pc for a Heeschan Cloud distance of $d=125$\,pc).
528: That the tiny beam of absorption
529: against the magnetar (limited by the size of the pulsar emission region,
530: which is smaller than the pulsar's light-cylinder radius, or about
531: 1\,light-second) has the same line width as the $0.12$\,pc-wide beam
532: of the SGPS at the Heeschen Cloud, suggests that there are few if any
533: spatial structures left unresolved by the SGPS in the Heeschen Cloud.
534: Since the \ion{H}{1} self-absorption line width does change between the GBT
535: and SGPS resolutions we can infer that either the absorption feature
536: comprises multiple narrow line features or that the non-thermal broadening
537: of the absorption feature changes between different structures in the
538: Heeschen cloud. Attempts to fit the different line-width absorption-line
539: structures in the Heeschen cloud with multiple Gaussian components
540: does not improve the fitting, which suggests that a single Gaussian
541: is sufficient. This implies that the non-thermal contribution to the
542: line width varies within the cloud.
543:
544: The standard assumption is that non-thermal line broadening arises
545: from turbulence. If the turbulence is intermittent, we can expect that
546: the turbulence has decayed more in some places than in other locations
547: \citep{frisch}. This can then easily explain the observed line widths of
548: the Heeschen Cloud \ion{H}{1} absorption. The areas that have undergone
549: more damping of the turbulence will have less turbulent energy and thus
550: have narrower non-thermal line widths.
551:
552: \section{OH Absorption Limits Toward XTE~J1810--197}
553:
554: Although we did not detect any OH absorption against \magnetar, the limits
555: are still interesting. All previous OH absorption detections against
556: pulsars \citep[][]{snezana,joel,toney} have found that the absorption is
557: deeper (larger opacity) and has a narrower line width than the pulsar
558: ``off'' spectra: in the three known cases, the OH absorption against
559: the pulsar is 2--3 times deeper than seen in the pulsar off spectra.
560:
561: For our OH absorption limits against \magnetar\ of $\tau < 0.1$
562: ($1\,\sigma$), we would expect our pulsar off spectra to have OH opacities
563: $\la 0.033$. In Table~\ref{table:ohofftau} we list the opacities of the
564: OH lines in the pulsar off spectra. We subtracted the system temperature
565: from the raw OH spectra, and then fitted a third order polynomial to
566: determine the continuum emission levels. Dividing the spectra by the
567: continuum emission results in $e^{-\tau}$ spectra for the pulsar off
568: spectra.
569:
570: If we assume that the OH pulsar off spectra should have a factor of
571: 2--3 weaker opacity than the OH absorption against the pulsar, then
572: we were likely within a factor of about 2 of detecting OH absorption
573: against \magnetar\ at velocities within the range observed for \ion{H}{1}
574: absorption (Table~\ref{table:higaussians}), e.g., at 9.9\,km\,s$^{-1}$
575: at 1612\,MHz, and 9.8 and 16.9\,km\,s$^{-1}$ at 1720\,MHz (see
576: Table~\ref{table:ohofftau}). Unfortunately, due the decay of the flux
577: of \magnetar\ \citep{fluxvarying}, we were not able to detect any OH
578: absorption.
579:
580: Our OH absorption limits toward \magnetar\ are also meaningful for the
581: OH seen at $V_{\rm LSR} \ga 28$\,km\,s$^{-1}$. The 1665\,MHz OH feature
582: at 30.7\,km\,s$^{-1}$ should have been detected at 2--$3\,\sigma$ if
583: it were between us and \magnetar. However, we would not have expected
584: to detect OH absorption at these velocities based on the \ion{H}{1}
585: absorption results (Table~\ref{table:higaussians}).
586:
587: \section{The Column Density to XTE~J1810--197}
588:
589: From the \ion{H}{1} and OH observations that we have performed, it
590: is possible to estimate the column density of hydrogen $N_{\rm H}$,
591: both atomic and molecular, to \magnetar. This can then be compared
592: with the range of values $N_{\rm H} = (6.5$--14$)\times 10^{21}$\,cm$^{-2}$
593: determined from fitting X-ray spectra \citep[][]{gotthelf,redclump}.
594:
595: \subsection{Estimate of the Atomic Column Density to XTE~J1810--197}
596:
597: We cannot directly measure the \ion{H}{1} column density toward \magnetar,
598: since we do not have enough constraints to perform radiative transfer
599: modeling in this direction. We can however still make an estimate of the
600: column density to \magnetar. To do this, we just integrate the \ion{H}{1}
601: emission spectrum between 0 and 25.7\,km\,s$^{-1}$. Assuming that half
602: of the emission at a given velocity
603: comes from \ion{H}{1} on the near side of the tangent
604: point and the other half comes from gas with similar properties on
605: the far side of the tangent point,
606: we obtain an estimate of the column density to \magnetar,
607: \begin{equation}
608: N_{\rm HI} = {1 \over 2} \int_{0}^{25.7} 1.83 \times 10^{18}~T_{\rm B}(v)~dv.
609: \end{equation}
610: Since there are clouds on the near side of the tangent point that
611: significantly absorb the \ion{H}{1} emission from the far side of the
612: tangent point, as is evidenced by the absorption seen against \magnetar,
613: this method provides a lower limit to the column density of \ion{H}{1}
614: to \magnetar. Upon performing the integration we find that $N_{\rm HI}
615: \ga 1.8 \times 10^{21}$\,cm$^{-2}$.
616:
617: \subsection{Estimate of the Molecular Column Density to XTE~J1810--197}
618:
619: We can use the pulsar ``off'' OH spectra to make an estimate of the
620: molecular column density to \magnetar. From Figures~\ref{fig:oh1612}
621: and~\ref{fig:oh1720} we see that the 1612\,MHz OH lines are in
622: absorption while the 1720\,MHz OH lines are in emission, with both having
623: approximately the same amplitude. Such conjugate emission arises in
624: regions where the OH column density is in the range $10^{14} < N_{\rm OH}/
625: \Delta v < 10^{15}$ cm$^{-2}$\,km$^{-1}$\,s with $\Delta v$ the velocity
626: resolution of the observations \citep[][]{joel,ohbook}. The total column
627: density of OH is found by integrating over the whole line.
628:
629: Integrating over the OH spectrum between 0 and 25.7\,km\,s$^{-1}$ and
630: assuming that half of the OH emission is from beyond the tangent point,
631: we find
632: \begin{equation}
633: (2.4\pm 0.3) \times 10^{14} < N_{\rm OH} < (2.4\pm 0.3)
634: \times 10^{15}~{\rm cm}^{-2}. \nonumber
635: \end{equation}
636: Using standard abundances, the ratio of the number of OH molecules to
637: the number of hydrogen atoms is $N_{\rm OH}/N_{\rm H} = 6 \times 10^{-8}$
638: \citep[][]{ohbook}, which gives
639: \begin{equation}
640: (4\pm 0.7) \times 10^{21} < N_{\rm H} < (40\pm 7) \times 10^{21}~{\rm cm}^{-2} \nonumber
641: \label{eq:nhoh}
642: \end{equation}
643: between us and the magnetar.
644:
645: Measurements of the $^{12}$CO$(2\rightarrow 1)$ spectrum toward
646: \magnetar\ are available from \citet[][]{cosurvey}. The total column
647: density of hydrogen in molecular form can be determined from the
648: $^{12}$CO$(2\rightarrow 1)$ spectrum using
649: \begin{equation}
650: {N_{\rm H} = 2 N_{\rm H_2} = X_{\rm CO} \int_0^{25.7} T_{\rm B}^{\rm CO}(v) dv}
651: \end{equation}
652: where $X_{{\rm CO}}$ is a conversion factor. The commonly used
653: ``standard'' value for this is $X_{\rm CO} = 2 \times 10^{20}$
654: cm$^{-2}$\,K$^{-1}$\,km$^{-1}$\,s. However, $X_{\rm CO}$ is
655: known to vary depending on the line of sight \citep[][]{xfactor}.
656: The $^{12}$CO$(2\rightarrow 1)$ spectrum nearest to the line of sight
657: toward \magnetar\ gives $\int_0^{25.7} T_{\rm B}^{\rm CO}(v) dv = 28.75$
658: K\,km\,s$^{-1}$, of which we will assume that half arises from beyond the
659: tangent point. From Figure~4 of \citet[][]{xfactor} we see that $X_{\rm
660: CO} = 10^{20}$ cm$^{-2}$\,K$^{-1}$\,km$^{-1}$\,s is reasonable for the
661: amount of $^{12}$CO$(2\rightarrow 1)$ toward \magnetar. This then yields
662: $N_{\rm H} \ga 2.8 \times 10^{21}$\,cm$^{-2}$.
663:
664: This limit for $N_{\rm H}$ obtained from the CO spectrum is consistent
665: with the range of values determined using the OH spectra
666: (eq.~\ref{eq:nhoh}). This also tells us that the OH column
667: densities are near the lower limit of $N_{\rm OH} / \Delta v =
668: 10^{14}$ cm$^{-2}$\,km$^{-1}$\,s. Since the CO spectra have a higher
669: signal-to-noise ratio, we will use the CO value for the molecular gas
670: contribution to $N_{\rm H}$.
671:
672: Adding the molecular and atomic components of $N_{\rm H}$, we obtain
673: $N_{\rm H} \ga 4.6 \times 10^{21}$\,cm$^{-2}$. This limit is in
674: agreement with the values determined from the X-ray spectrum of \magnetar.
675: Unfortunately we only have a lower limit for the total column density
676: that is below the values determined from the X-ray spectrum.
677:
678:
679: \section{Models of the Line of Sight Toward XTE~J1810--197}
680:
681: A detailed molecular model of the ISM toward \sgr\ has been developed
682: by \citet[][]{corbelone} and \citet[][]{corbel}, depicted in Figure~8 of
683: the latter. With increasing distance toward \sgr, their model encounters
684: gas at $V_{\rm LSR}=4$, 24, 30, 38, 44, and then 13\,km\,s$^{-1}$ in
685: reaching the Scutum-Crux spiral arm \citep[labeled as the 30\,km\,s$^{-1}$
686: spiral arm in Fig.~8 of][]{corbel}. Since the line of sight toward \sgr\
687: is very close to that of \magnetar\ we might expect to encounter clouds
688: at roughly the same velocities on the line of sight toward \magnetar.
689:
690: The gas at 4\,km\,s$^{-1}$ toward \sgr\ is likely one of the two velocity
691: components of the Heeschen Cloud \citep[][]{riegel} and can be associated
692: with the \ion{H}{1} absorption feature at 7.7\,km\,s$^{-1}$ toward
693: \magnetar, which is from the other velocity component of the Heeschen
694: Cloud. The gas at 24\,km\,s$^{-1}$ toward \sgr\ can be associated with
695: the \ion{H}{1} absorption features at 22.8 or 25.7\,km\,s$^{-1}$ toward
696: \magnetar. We can associate the 13\,km\,s$^{-1}$ gas toward \sgr\ with
697: the \ion{H}{1} absorption seen at 14.1\,km\,s$^{-1}$ toward \magnetar.
698:
699: If the model of \citet[][]{corbelone} and \citet[][]{corbel} is correct,
700: then we should also expect see \ion{H}{1} absorption at velocities of
701: roughly 30, 38, and 44\,km\,s$^{-1}$ toward \magnetar, since we observe
702: absorption that can be associated with the 13\,km\,s$^{-1}$ cloud in
703: their model. In fact, we do not observe any \ion{H}{1} absorption
704: against \magnetar\ at these velocities. This suggests that either
705: (1) the model of \citet[][]{corbelone} and \citet[][]{corbel} is not
706: entirely correct, or (2) it cannot be applied to the line of sight toward
707: \magnetar, which is only $41'$ away, or (3) there is molecular material
708: without \ion{H}{1} along these lines of sight. The last possibility
709: is not likely since molecular clouds are expected to have cosmic-ray
710: ionization and photo-dissociation regions within and at their outer edges,
711: which would produce atomic hydrogen \citep[see][\S~9.2, and references
712: therein]{atomiccloud}.
713:
714: \section{Discussion}
715:
716: Using the $\mbox{DM} = 178\pm5$\,cm$^{-3}$\,pc measured for \magnetar\
717: \citep{camilo}, its distance according to the \citet{ne2001} electron
718: density model is 3.3\,kpc. This model has a claimed average uncertainty
719: of about 20\% which, however, can be much larger for individual objects.
720: For the sake of discussion, we assume an uncertainty of 1\,kpc.
721: The electron density model was derived using distances to pulsars that
722: in many cases were determined via \ion{H}{1} absorption spectra assuming
723: a flat rotation curve, so that it seems most appropriate to compare the
724: DM--derived distance of $3.3\pm1$\,kpc with our flat rotation curve's
725: kinematic distance of \flatdist. These two values agree remarkably well
726: and imply that along the line of sight to \magnetar, the \citet{ne2001}
727: model gives a good representation of the average free electron density
728: out to about 4\,kpc.
729:
730: The distance to \magnetar\ has been estimated from X-ray observations
731: to range over 2.5--5\,kpc \citep{gotthelf,firstdist}. These distances
732: are determined by converting $N_{\rm H}$ values obtained from fits to
733: the X-ray spectra into visual extinction, $A_V$, and an estimate of
734: the $A_V$ per kpc in the Galaxy. This method is limited by a number of
735: complications: (1) the spectral model used to fit the X-ray data, e.g.,
736: two blackbodies versus a blackbody and a power law; (2) the $N_{\rm H}$
737: versus $A_V$ relationship determined locally (within $\sim 1$\,kpc)
738: but used for large distances ($> 1$\,kpc);
739: (3) the large deviations from the fitted $N_{\rm H}$
740: versus $A_V$ relationship for any particular line of sight;
741: and (4) the $d$--$A_V$ relationship also determined locally but used
742: for large distances.
743:
744: \cite{redclump} determined $d = 3.1\pm0.5$\,kpc toward \magnetar\ assuming
745: an X-ray $N_{\rm H} = 14 \times 10^{21}$\,cm$^{-2}$, which is higher
746: than any of the values fitted by \cite{firstdist} or \cite{gotthelf}.
747: We consider the $d$--$A_V$ relation of \citet{redclump} using red clump
748: stars in the line of sight to \magnetar\ to be %a significant
749: an improvement,
750: although it is still necessary to apply an X-ray--fitted value of $N_{\rm
751: H}$ to this relation and then convert it into a visual extinction
752: $A_V$. Using $N_{\rm H} = 6.5 \times 10^{21}$\,cm$^{-2}$ determined
753: by \citet{gotthelf}, which is arguably a lower limit, the extinction
754: toward \magnetar\ becomes $A_V \approx$\,3--4.5\,mag using
755: Figure~3 of \citet{predehl}. This then yields
756: a distance estimate of 2--3.5\,kpc based on Figure~7 of \cite{redclump}.
757:
758: The kinematic distances determined from the \ion{H}{1} absorption
759: measurements presented in this paper rely only on the model of Galactic
760: rotation used to convert the measured velocity into a distance.
761: The \citet{sellwood} and \citet{englemaier} models both give consistent
762: results (see Table~\ref{table:distance}). We prefer the kinematic
763: distance of \ourdistance\ determined in this paper over the DM- and
764: X-ray-derived distances, because our conversion of measured velocity
765: to distance is more direct and better constrained than through these
766: other methods.
767:
768: In Figure~7 of \citet{redclump} we see that the extinction vs.\ distance
769: rises steeply between 3.0 and 3.5\,kpc and that for $A_V \le 13$ the
770: distance can be limited to be less than 4\,kpc. The line of sight
771: toward \magnetar\ contains the Infrared Dark Cloud (IRDC) \darkcloud\
772: \citep[][]{irdc,carey}. IRDCs are thought to be places where Giant
773: Molecular Clouds are either just starting to form massive stars or on the
774: verge of beginning to form stars. The sudden increase in opacity vs.\
775: distance seen by \citet{redclump} is likely associated with this IRDC.
776: We note that the highest opacity regions of the IRDC cover less than
777: half the area on the sky that \citet{redclump} used to determine the
778: visual extinction vs.\ distance toward \magnetar. Also, IRDCs can have
779: extremely high opacities such that only the edge of the cloud is seen
780: even in the infrared \citep{irdc,atomiccloud}. Since the opacities can be
781: very large, it is plausible that \cite{redclump} may have underestimated
782: the actual extinction vs.\ distance toward \magnetar.
783:
784: \cite{co} associated IRDC \darkcloud\ with $^{12}{\rm
785: CO}\left(1\rightarrow 2\right)$ emission centered at 32\,km\,s$^{-1}$,
786: implying $d<3.8_{-0.5}^{+0.5}$\,kpc for a flat rotation curve and
787: $d<4.4_{-0.3}^{+0.6}$\,kpc for the \cite{sellwood} and \cite{englemaier}
788: models. Our OH spectra show a large spectral feature at the same
789: velocities that we also associate with the IRDC. All observed \ion{H}{1}
790: absorption features lie at velocities lower than those associated with the
791: IRDC. This puts \magnetar\ no further than the front edge of the IRDC.
792:
793: We consider the 4\,kpc upper limit from Figure~7 of \citet{redclump}
794: to be a hard upper limit on the distance to \magnetar. This further
795: constrains slightly the distance estimate of 3.1--4.3\,kpc obtained
796: from \ion{H}{1} absorption measurements (\S~4.3). Overall, we can
797: thus summarize that the distance to \magnetar\ is $3.5\pm0.5$\,kpc.
798: Together with the measured proper motion of the AXP, this results in
799: a transverse velocity corrected to the LSR of $212\pm35$\,km\,s$^{-1}$
800: \citep{helfand}, a perfectly ordinary velocity among pulsars.
801:
802: \acknowledgments
803:
804: We would like to thank Jay Lockman for many useful discussions and
805: T. Dame who graciously provided us with the $^{12}$CO$(2\rightarrow 1)$
806: spectrum toward \magnetar. The GBT is operated by the National Radio
807: Astronomy Observatory, a facility of the National Science Foundation
808: operated under cooperative agreement by Associated Universities, Inc.
809: F.C. acknowledges the NSF for support through grant AST-05-07376.
810:
811: \clearpage
812:
813: \begin{thebibliography}
814:
815: \bibitem[Becker et al.(1985)]{becker}
816: Becker, R. H., Markert, T., \& Donahue, M. 1985, \apj, 296, 461
817:
818: \bibitem[Brand \& Blitz(1993)]{brand}
819: Brand, J., \& Blitz, L. 1993, A\&A, 275, 67
820:
821: \bibitem[Burton(1988)]{butler}
822: Burton, W. B. 1988, ``The structure of our Galaxy derived from
823: observations of \ion{H}{1}'', in Galactic and Extragalactic Radio
824: Astronomy (2nd edition), Berlin and New York, Springer-Verlag, 295
825:
826: \bibitem[Cameron et al.(2005)]{cameron}
827: Cameron, P. B., et al. 2005, Nature, 434, 1112
828:
829: \bibitem[Camilo et al.(2006)]{camilo}
830: Camilo, F., Ransom, S. R., Halpern, J. P., Reynolds, J., Helfand, D. J.,
831: Zimmerman, N., \& Sarkissian, J. 2006, Nature, 442, 892
832:
833: \bibitem[Camilo et al.(2007{\natexlab{a}})]{fluxvarying}
834: {Camilo}, F., et al. 2007{\natexlab{a}}, ApJ, 663, in press (astro-ph/0610685)
835:
836: \bibitem[Camilo et al.(2007{\natexlab{b}})]{xtepol}
837: {Camilo}, F., Reynolds, J., Johnston, S., {Halpern}, J.~P., Ransom, S.~M., \&
838: van Straten, W. 2007{\natexlab{b}}, ApJ, 659, L37
839:
840: \bibitem[Carey et al.(2000)]{carey}
841: Carey, S. J., Feldman, P. A., Redman, R. O., Egan, M. P., MacLeod,
842: J. M., \& Price, S. D. 2000, \apj, 543, L157
843:
844: \bibitem[Caswell et al.(1975)]{caswell}
845: Caswell, J. L., Murray, J. D., Roger, R. S., Cole, D. J., \& Cooke,
846: D. J. 1975, A\&A, 45, 239
847:
848: \bibitem[Corbel et al.(1997)]{corbelone} Corbel, S., Wallyn, P.,
849: Dame, T.~M., Durouchoux, P., Mahoney, W.~A., Vilhu, O., \& Grindlay, J.~E.\
850: 1997, \apj, 478, 624
851:
852: \bibitem[Corbel \& Eikenberry(2004)]{corbel}
853: Corbel, S., \& Eikenberry, S. S. 2004, A\&A, 419, 191
854:
855: \bibitem[Cordes \& Lazio(2002)]{ne2001}
856: Cordes, J. M., \& Lazio, T. J. W. 2002, preprint (astro-ph/0207156)
857:
858: \bibitem[Dame et al.(2001)]{cosurvey} Dame, T.~M., Hartmann, D.,
859: \& Thaddeus, P.\ 2001, \apj, 547, 792
860:
861: \bibitem[Durant \& van Kerkwijk(2006)]{redclump}
862: Durant, M., \& van Kerkwijk, M. H. 2006, \apj, 650, 1070
863:
864: \bibitem[Elitzur(1992)]{ohbook} Elitzur, M.\ 1992,
865: Astronomical Masers, Kluwer Academic Publishers (Astrophysics and Space
866: Science Library.~Vol.~170), 365
867:
868: \bibitem[Englmaier \& Gerhard(2006)]{englemaier} Englmaier, P., \&
869: Gerhard, O.\ 2006, Celestial Mechanics and Dynamical Astronomy, 94, 369
870:
871: \bibitem[Frisch(1995)]{frisch} Frisch, U.\ 1995, Turbulence. The legacy
872: of A.~N. Kolmogorov (Cambridge: Cambridge Univ. Press)
873:
874: \bibitem[Garwood \& Dickey(1989)]{garwood}
875: Garwood, R. W., \& Dickey, J. M. 1989, \apj, 338, 841
876:
877: \bibitem[Gotthelf \& Halpern(2005)]{gotthelf} Gotthelf, E. V., \&
878: Halpern, J. P. 2005, \apj, 632, 1075
879:
880: \bibitem[Gotthelf et al.(2004)]{firstdist}
881: Gotthelf, E. V., Halpern, J. P., Buxton, M., \& Bailyn, C. 2004, \apj,
882: 605, 368
883:
884: \bibitem[Green et al.(1988)]{green}
885: Green, D. A., Gull, S. F., Tan, S. M., \& Simon, A. J. B. 1988, \mnras,
886: 231, 735
887:
888: \bibitem[Greisen \& Lockman(1979)]{greisen} Greisen, E. W., \& Lockman,
889: F. J. 1979, \apj, 228, 740
890:
891: \bibitem[{Helfand {et~al.}(2007)Helfand, Chatterjee, Brisken, Camilo, Reynolds,
892: van Kerkwijk, Halpern, \& Ransom}]{helfand}
893: Helfand, D.~J., Chatterjee, S., Brisken, W., Camilo, F., Reynolds, J., van
894: Kerkwijk, M.~H., Halpern, J.~P., \& Ransom, S.~M. 2007, ApJ, 662, 1198
895:
896: \bibitem[Jaffe et al.(1982)]{co}
897: Jaffe, D. T., Stier, M. T., \& Fazio, G. G. 1982, \apj, 252, 601
898:
899: \bibitem[Kalberla et al.(1982)]{kalberla}
900: Kalberla, P. M. W., Goss, W. M., \& Wilson, T. L. 1982, A\&A, 106, 167
901:
902: \bibitem[Lockman \& Dickey(1990)]{jay}
903: Lockman, F. J., \& Dickey, J. M. 1990, ARAA, 28, 215
904:
905: \bibitem[Magnani et al.(1998)]{xfactor} Magnani, L., Onello,
906: J.~S., Adams, N.~G., Hartmann, D., \& Thaddeus, P.\ 1998, \apj, 504, 290
907:
908: \bibitem[McClure-Griffiths et al.(2005)]{sgpshi}
909: McClure-Griffiths, N.~M., Dickey, J.~M., Gaensler, B.~M., Green, A.~J.,
910: Haverkorn, M., \& Strasser, S.\ 2005, \apjs, 158, 178
911:
912: \bibitem[McClure-Griffiths et al.(2006)]{rccloud}
913: McClure-Griffiths, N. M., Dickey, J. M., Gaensler, B. M., Green, A. J.,
914: \& Haverkorn, M. 2006, \apj, 652, 1339
915:
916: \bibitem[McClure-Griffiths \& Gaensler(2006)]{hiabs}
917: McClure-Griffiths, N. M., \& Gaensler, B. M. 2006, \apj, 630, L163
918:
919: \bibitem[Minter et al.(2001)]{atomiccloud} Minter, A.~H., Lockman,
920: F.~J., Langston, G.~I., \& Lockman, J.~A.\ 2001, \apj, 555, 868
921:
922: \bibitem[Minter et al.(2005)] {minter}
923: Minter, A. H., Balser, D. S., \& Kartaltepe, J. S. 2005, \apj, 631, 376
924:
925: \bibitem[Minter(2005)]{toney} Minter, A.~H.\ 2005, Bulletin
926: of the American Astronomical Society, 37, 1301 (in preparation for \apj)
927:
928: \bibitem[Morrison \& McCammon(1983)]{mccammon} Morrison, R., \&
929: McCammon, D.\ 1983, \apj, 270, 119
930:
931: \bibitem[Muno et al.(2006)]{muno}
932: Muno, M. P., et al. 2006, ApJ, 636, L41
933:
934: \bibitem[Predehl \& Schmitt(1995)]{predehl} Predehl, P., \&
935: Schmitt, J.~H.~M.~M.\ 1995, \aap, 293, 889
936:
937: \bibitem[Reid(1993)]{gcdist} Reid, M.~J.\ 1993, \araa, 31, 345
938:
939: \bibitem[Riegel \& Crutcher(1972)]{riegel}
940: Riegel, K. W., \& Crutcher, R. M. 1972, A\&A, 18, 55
941:
942: \bibitem[Roberts(1972)]{roberts} Roberts, W.~W., Jr.\ 1972,
943: \apj, 173, 259
944:
945: \bibitem[Simon et al.(2006)]{irdc}
946: Simon, R., Jackson, J. M., Rathborne, J. M., \& Chambers, E. T. 2006,
947: \apj, 639, 227
948:
949: \bibitem[Stanimirovi{\'c} et al.(2003)]{snezana}
950: Stanimirovi{\'c}, S., Weisberg, J.~M., Dickey, J.~M., de la Fuente, A.,
951: Devine, K., Hedden, A., \& Anderson, S.~B.\ 2003, \apj, 592, 953
952:
953: \bibitem[Weiner \& Sellwood(1999)]{sellwood}
954: Weiner, B. J., \& Sellwood, J. A. 1999, \apj, 524, 112
955:
956: \bibitem[Weisberg(1978)]{weisberg} Weisberg, J.~M.\ 1978,
957: Ph.D.~Thesis, University of Iowa
958:
959: \bibitem[Weisberg et al.(2005)]{joel} Weisberg, J.~M.,
960: Johnston, S., Koribalski, B., \& Stanimirovi{\'c}, S.\ 2005, Science, 309,
961: 106
962: \end{thebibliography}
963:
964: \clearpage
965:
966: \begin{deluxetable}{lcc}
967: \tablewidth{0pt}
968: \tablecolumns{3}
969: \tablecaption{\label{table:absobs} Observations of \magnetar\ at GBT }
970: \tablehead{
971: \colhead{Date} &
972: \colhead{Time (UTC)} &
973: \colhead{Species} }
974: \startdata
975: 2006 Jun 6 & 07:30 -- 11:00 & \ion{H}{1} \\
976: 2006 Jul 22/23 & 23:00 -- 07:30 & OH \\
977: 2006 Aug 31/Sep 1 & 23:30 -- 05:00 & OH \\
978: 2006 Sep 2 & 00:00 -- 05:00 & OH \\
979: 2006 Sep 4 & 00:00 -- 05:00 & OH \\
980: 2006 Sep 24/25 & 19:45 -- 03:40 & OH \\
981: 2006 Oct 21 & 17:00 -- 19:09 & OH \\
982: 2006 Oct 21/22 & 20:55 -- 01:45 & OH
983: \enddata
984: \end{deluxetable}
985:
986: \clearpage
987:
988: \begin{deluxetable}{ccc}
989: \tablewidth{0pt}
990: \tablecolumns{3}
991: \tablecaption{\label{table:higaussians} Gaussian fits to the \ion{H}{1}
992: absorption lines toward XTE~J1810--197 }
993: \tablehead{
994: \colhead{$\tau$} &
995: \colhead{$V_{\rm LSR}$} &
996: \colhead{FWHM} \\
997: \colhead{} &
998: \colhead{(km\,s$^{-1}$)} &
999: \colhead{(km\,s$^{-1}$)} }
1000: \startdata
1001: $2.15 \pm 0.09 $ & $ 7.73 \pm 0.06 $ & $ 3.3 \pm 0.2 $ \\
1002: $0.5 \pm 0.1 $ & $ 14.1 \pm 0.2 $ & $ 1.2 \pm 0.4 $ \\
1003: $1.53 \pm 0.08 $ & $ 19.1 \pm 0.1 $ & $ 3.8 \pm 0.4 $ \\
1004: $1.2 \pm 0.1 $ & $ 22.8 \pm 0.2 $ & $ 2.1 \pm 0.4 $ \\
1005: $0.94 \pm 0.1 $ & $ 25.7 \pm 0.2 $ & $ 2.4 \pm 0.5 $
1006: \enddata
1007: \tablecomments{ See Figs.~\ref{fig:hiabs} and \ref{fig:gaussians} for
1008: data upon which these fits are based. }
1009: \end{deluxetable}
1010:
1011: \clearpage
1012:
1013: \begin{deluxetable}{cc}
1014: \tablewidth{0pt}
1015: %\tablewidth{0.50\linewidth}
1016: \tablecolumns{2}
1017: \tablecaption{\label{table:ohlimits} OH absorption limits toward
1018: XTE~J1810--197 }
1019: \tablehead{
1020: \colhead{$\nu$} &
1021: \colhead{$\sigma_\tau$\tablenotemark{a}} \\
1022: \colhead{(MHz)} &
1023: \colhead{} }
1024: \startdata
1025: 1612 & 0.09 \\
1026: 1665 & 0.10 \\
1027: 1667 & 0.10 \\
1028: 1720 & 0.10
1029: \enddata
1030: \tablenotetext{a}{$1\,\sigma$ limits.}
1031: \end{deluxetable}
1032:
1033: \clearpage
1034:
1035: \begin{deluxetable}{cccc}
1036: \tablewidth{0pt}
1037: \tablecolumns{3}
1038: \tablecaption{Gaussian fits to the OH opacity
1039: in the pulsar ``off'' spectra toward XTE~J1810--197 \label{table:ohofftau}}
1040: \tablehead{
1041: \colhead{$\nu$} &
1042: \colhead{$\tau$} &
1043: \colhead{$V_{\rm LSR}$} &
1044: \colhead{FWHM} \\
1045: \colhead{(MHz)} &
1046: \colhead{} &
1047: \colhead{(km\,s$^{-1}$)} &
1048: \colhead{(km\,s$^{-1}$)} }
1049: \startdata
1050: 1612 & $0.026\pm 0.002$ & $9.9\pm 0.2$ & $4.4\pm 0.6$ \\
1051: 1612 & $0.097\pm 0.002$ & $29.54\pm 0.05$ & $5.4\pm 0.1$ \\
1052: \hline
1053: 1665 & $0.059\pm 0.009$ & $28.7\pm 0.5$ & $7.9\pm 0.4$ \\
1054: 1665 & $0.10\pm 0.01$ & $30.69\pm 0.08$ & $3.9 \pm 0.3$ \\
1055: 1665 & $0.020\pm 0.002$ & $39.9\pm 0.3$ & $5.5\pm 0.8$ \\
1056: \hline
1057: 1667 & $0.045\pm 0.005$ & $28.1\pm 0.6$ & $9.7\pm 0.6 $ \\
1058: 1667 & $0.086\pm 0.007$ & $31.11\pm 0.08 $ & $4.3\pm 0.3$ \\
1059: 1667 & $0.018\pm 0.002$ & $42.0\pm 0.4 $ & $6.3\pm 0.9$ \\
1060: \hline
1061: 1720 & $-0.023\pm 0.002$ & $9.8\pm 0.1$ & $3.1\pm 0.3$ \\
1062: 1720 & $-0.020\pm 0.002$ & $16.9\pm 0.2$ & $4.8\pm 0.5$ \\
1063: 1720 & $-0.021\pm 0.004$ & $26.2\pm 0.2$ & $2.3\pm 0.5$ \\
1064: 1720 & $-0.063\pm 0.002$ & $29.5\pm 0.1$ & $3.8\pm 0.3$ \\
1065: 1720 & $0.014\pm 0.002$ & $44.4\pm 0.3$ & $3.0\pm 0.6$
1066: \enddata
1067: \end{deluxetable}
1068:
1069: \clearpage
1070:
1071: \begin{deluxetable}{llrc}
1072: \tabletypesize{\footnotesize}
1073: \tablewidth{0pt}
1074: \tablecolumns{4}
1075: \tablecaption{\label{table:distance} Distance estimates for XTE~J1810--197 }
1076: \tablehead{
1077: \colhead{Measurement} &
1078: \colhead{Method} &
1079: \colhead{$d$} &
1080: \colhead{Refs.} \\
1081: \colhead{} &
1082: \colhead{} &
1083: \colhead{(kpc)} &
1084: \colhead{} }
1085: \startdata
1086: DM & $n_e$ model \citep{ne2001} & $3.3 \pm 1$ & 1 \\
1087: X-ray: blackbody + power law & ${N_{\rm H} \rightarrow A_V \rightarrow 1.5-2.0~{\rm mag\,kpc}^{-1} \rightarrow } d$ & $\sim 5.0$ & 2 \\
1088: X-ray: two blackbodies & ${N_{\rm H} \rightarrow A_V \rightarrow 1.5-2.0{\rm ~mag\,kpc}^{-1} \rightarrow } d$ & $\sim 2.5$ & 3 \\
1089: X-ray: blackbody + power law & ${N_{\rm H} \rightarrow A_V \rightarrow {\rm red~clump~stars} \rightarrow } d$ & $3.1\pm0.5$ & 4 \\
1090: X-ray: two blackbodies& ${N_{\rm H} \rightarrow A_V \rightarrow {\rm red~clump~stars} \rightarrow }d$ & 2--3.5 & 5 \\
1091: \ion{H}{1} absorption & Flat rotation curve \citep{butler} & \flatdist & 5 \\
1092: \ion{H}{1} absorption & \citet[][]{sellwood} model & \weinerdist &5\\
1093: \ion{H}{1} absorption & \citet[][]{englemaier} model & \englemaierdist &5\\
1094: \ion{H}{1} absorption & \citet[][]{brand} model & $2.4\pm0.5$&5
1095: \enddata
1096: \tablerefs{(1) \citet{camilo}; (2) \citet{firstdist}; (3)
1097: \citet{gotthelf}; (4) \citet{redclump}; (5) this work. }
1098: \tablecomments{See \S~4 for a discussion of the various rotation curve
1099: models that we fit to the \ion{H}{1} data. The X-ray fits done by other
1100: authors are discussed in \S~9. Overall, our best distance determination
1101: comes largely from the direct \ion{H}{1} measurements on \magnetar,
1102: and is $3.5\pm0.5$\,kpc (see \S~9). }
1103: \end{deluxetable}
1104:
1105: \clearpage
1106:
1107: \begin{figure}
1108: \plotone{fig01}
1109: \caption{\textit{Top}: The \ion{H}{1} emission spectrum (pulsar
1110: ``off'' spectrum) toward XTE~J1810--197. \textit{Bottom}: The
1111: \ion{H}{1} absorption spectrum against XTE~J1810--197. \label{fig:hiabs}
1112: }
1113: \end{figure}
1114:
1115: \clearpage
1116:
1117: \begin{figure}
1118: \includegraphics[angle=0, width=6in]{fig02}
1119: \caption{Gaussian fits (\textit{thick line}) to the \ion{H}{1} opacity
1120: (\textit{thin line}) observed toward XTE~J1810--197. Residuals to the
1121: fits are shown as the dotted line. Results of the fits are listed in
1122: Table~\ref{table:higaussians}. \label{fig:gaussians} }
1123: \end{figure}
1124:
1125: \clearpage
1126:
1127: \begin{figure}
1128: \plotone{fig03}
1129: \caption{\textit{Top}: The OH 1612\,MHz pulsar ``off'' spectrum
1130: toward XTE~J1810--197. \textit{Bottom}: The OH 1612\,MHz
1131: absorption spectrum against XTE~J1810--197. \label{fig:oh1612} }
1132: \end{figure}
1133:
1134: \clearpage
1135:
1136: \begin{figure}
1137: \plotone{fig04}
1138: \caption{\textit{Top}: The OH 1665\,MHz pulsar ``off'' spectrum
1139: toward XTE~J1810--197. \textit{Bottom}: The OH 1665\,MHz
1140: absorption spectrum against XTE~J1810--197. \label{fig:oh1665} }
1141: \end{figure}
1142:
1143: \clearpage
1144:
1145: \begin{figure}
1146: \plotone{fig05}
1147: \caption{\textit{Top}: The OH 1667\,MHz pulsar ``off'' spectrum
1148: toward XTE~J1810--197. \textit{Bottom}: The OH 1667\,MHz
1149: absorption spectrum against XTE~J1810--197. \label{fig:oh1667} }
1150: \end{figure}
1151:
1152: \clearpage
1153:
1154: \begin{figure}
1155: \plotone{fig06}
1156: \caption{\textit{Top}: The OH 1720\,MHz pulsar ``off'' spectrum
1157: toward XTE~J1810--197. \textit{Bottom}: The OH 1720\,MHz
1158: absorption spectrum against XTE~J1810--197. \label{fig:oh1720} }
1159: \end{figure}
1160:
1161: \clearpage
1162:
1163: \begin{figure}
1164: \includegraphics[angle=90, width=6in]{fig07}
1165: \caption{The flat Galactic rotation model (\textit{solid curve}) of
1166: \citet{butler}. The $x$-axis is the distance from the Sun along the line
1167: of sight of \magnetar. The $y$-axis is the radial velocity of the gas.
1168: The dotted lines indicate the deviation from Galactic rotation of $\pm
1169: 7$\,km\,s$^{-1}$. The arrow at $V_{\rm LSR}=25.7$\,km\,s$^{-1}$ indicates
1170: the highest velocity \ion{H}{1} absorption seen toward \magnetar\ (see
1171: Table~\ref{table:higaussians}). The shaded regions indicate the allowed
1172: kinematic distances from the flat rotation curve (see discussion in
1173: \S\S~4.2 and 4.3 for interpretation of these figures). \label{fig:flat} }
1174: \end{figure}
1175:
1176: \clearpage
1177:
1178: \begin{figure}
1179: \includegraphics[angle=90, width=6in]{fig08}
1180: \caption{The same as Fig.~\ref{fig:flat}, except that it uses the Galactic
1181: rotation model of \citet[][]{sellwood}, considering potential due to a
1182: bar. \label{fig:weiner} }
1183: \end{figure}
1184:
1185: \clearpage
1186:
1187: \begin{figure}
1188: \includegraphics[angle=90, width=6in]{fig09}
1189: \caption{The same as Fig.~\ref{fig:flat}, except that it uses the Galactic
1190: rotation model of \citet[][]{englemaier}, considering potential due to
1191: a bar and spiral arms. \label{fig:englemaier} }
1192: \end{figure}
1193:
1194: \clearpage
1195:
1196: \begin{figure}
1197: \includegraphics[angle=90, width=6in]{fig10}
1198: \caption{The same as Fig.~\ref{fig:flat}, except that it uses the
1199: Galactic rotation model of \citet{brand}, derived from observations
1200: of \ion{H}{2} regions. The points and error bars are derived from
1201: Fig.~2 of \citet{brand}. The solid line is a quadratic fit to the
1202: points. \label{fig:brand} }
1203: \end{figure}
1204:
1205: \clearpage
1206:
1207: \begin{figure}
1208: \includegraphics[angle=0, width=6in]{fig11}
1209: \caption{Comparison of the GBT \ion{H}{1} spectrum (\textit{black line})
1210: with the full-resolution SGPS \ion{H}{1} spectrum (\textit{red line})
1211: toward \magnetar. The SGPS spectrum convolved to the same resolution
1212: as the GBT is shown as the green line. The top panel shows the full
1213: range of emission, while the lower panel is zoomed in to show the
1214: difference in line widths for the Heeschen Cloud absorption around
1215: 8\,km\,s$^{-1}$. \label{fig:gbtsgps} }
1216: \end{figure}
1217:
1218:
1219: \end{document}
1220: