1: %\documentclass[manuscript]{aastex}
2: \documentclass[preprint]{aastex}
3: %\documentclass[preprint2]{aastex}
4:
5: \shorttitle{Fe Abundance in HH Shocks}
6: \shortauthors{B\"ohm \& Matt}
7:
8: \slugcomment{Accepted by \pasp.}
9:
10: \begin{document}
11:
12: \title{An Approximate Determination of the Gas-Phase Metal Abundance
13: in Herbig-Haro Outflows and their Shocks}
14:
15: \author{Karl-Heinz B\"ohm and Sean Matt}
16: \affil{Astronomy Department, University of Washington, Seattle WA 98195\\
17: bohm@astro.washington.edu, matt@astro.washington.edu}
18:
19:
20: \begin{abstract}
21:
22: It is important to determine whether the observed bow shocks in the
23: working surfaces of Herbig-Haro outflows has lead to a destruction of
24: dust grains and consequently to a change in the gas phase metal
25: abundances (say of Fe) in the cooling regions of HH bow shocks.
26: Detailed studies are currently available for only 5 HH outflows
27: (Beck-Winchatz et al.\ 1996). This small number is due to the large
28: observational and theoretical effort required to determine metal
29: abundances in HH objects.
30:
31: Information about metal abundances in more HH objects is badly needed.
32: We therefore use a very approximate method. We introduce a
33: ``characteristic number,'' $A_{\rm me}$, whose definition is based
34: only on the often observed line fluxes of [Fe II] 5159, [Fe II] 7155,
35: [Ca II] 7291, as well as H$\alpha$ and H$\beta$. These fluxes can
36: easily be determined from existing observations. We find a good
37: correlation between $A_{\rm me}$ and the Fe abundance for the 5
38: well-studied HH objects. We use this correlation to determine
39: approximate values of the gas phase Fe abundance in 13 additional high
40: excitation and in 3 additional low excitation HH objects.
41:
42: The results are the following: Of the 16 high excitation HH objects
43: studied, there are 6 which approximately agree with normal population
44: I abundance (i.e., no depletion due to dust formation). The remaining
45: 10 show some (very modest) Fe gas phase depletion which, however,
46: never gets larger than by a factor of 2.5. This result is in
47: agreement with our qualitative expectations that fast shocks
48: efficiently destroy dust grains. Of the 5 low excitation HH objects
49: studied, there are 4 which show a normal population I abundance
50: (strictly speaking, an even slightly higher abundance than this). This
51: is completely unexpected. In low excitation objects, one might expect
52: strong gas phase Fe depletion (showing the unchanged molecular cloud
53: composition), unless the matter has previously gone through shocks of
54: much higher shock velocities. We discuss this possible explanation and
55: the question of whether low excitation HH objects have a different
56: ``history'' than usually assumed.
57:
58:
59: \end{abstract}
60:
61: \keywords{Herbig-Haro objects --- ISM: abundances --- shockwaves}
62:
63: \section{Introduction}
64:
65: It has been generally accepted that the Herbig-Haro outflows (jets)
66: from young stellar objects become optically visible through the shocks
67: (bow shocks and jet shocks, Mach disks) at their working surfaces
68: (Mundt \& Fried 1983; Hartmann \& Raymond 1984; B\"ohm \& Solf 1985;
69: Schwartz 1985; Mundt 1985; Hartigan, Raymond \& Hartmann 1987;
70: Reipurth 1989; Hartigan 1989; Solf \& B\"ohm 1991; Reipurth \&
71: Heathcote 1993; Eisloeffel, Mundt \& B\"ohm 1994; Raga 1995). In
72: other words, in HH objects we see matter (if it is observed in a bow
73: shock) that originally comes from the molecular cloud and has recently
74: gone through the shock and through an essential part of the cooling
75: region of the shock. Recently, it has also become clear that, in a
76: number of Herbig-Haro (HH) objects, we can identify ``internal working
77: surfaces'' and their bow shocks (Raga 1995; Reipurth \& Heathcote
78: 1997; Reipurth et al.\ 1997; Raga \& Noriega-Crespo 1998). That is,
79: in a number of HH objects (e.g., in HH 111 or in HH 34), the cloud
80: matter has passed through multiple shocks.
81:
82: In molecular clouds (including the boundary regions of the clouds,
83: which are accessible to optical HH observations), we expect that the
84: gas phase abundances of elements like Fe or Ca are considerably
85: depleted (by a factor of on the order of 100) because of the formation
86: of dust grains (van Dieshoeck et al.\ 1993). If one takes into account
87: the fact that refractory elements like Fe and Ca can return to the gas
88: phase by the destruction of the dust grains in sufficiently strong
89: shocks ($v = 150$ km s$^{-1}$; e.g., McKee et al.\ 1987; Seab 1987;
90: Draine 1995), we might expect that HH objects will show different Fe/H
91: or Ca/H ratios that depend on the velocity of the shock through which
92: the visible matter has passed before it is observed.
93:
94: The (admittedly very naive) expectation would be that ``low excitation
95: objects'' (slow shocks, $v \la 50$ km s$^{-1}$), would have rather
96: small gas phase Fe and Ca abundances because these elements remain in
97: the dust grains. In ``high excitation objects'' (fast shocks, $v >
98: 85$ km s$^{-1}$), we might expect the dust grains to be destroyed, at
99: least partially, by the shock. The gas phase abundance of elements
100: like Fe or Ca would be high (solar?). This hypothesis inspired
101: Beck-Winchatz, B\"ohm, \& Noriega-Crespo (1994, hereafter BBN94; 1996,
102: hereafter BBN96) to determine the Fe gas phase abundance for a few
103: objects.
104:
105: In both papers the authors decided that it would be most convenient
106: (for the study of dust grain survival) to determine the Fe abundance
107: (they derived the abundance ratio Fe/S and Fe/O). The effort required
108: for carrying out this program for a single HH object is sizable. It
109: requires reliable line spectrophotometry with fairly high spectral
110: resolution in a fairly large wavelength interval and the detailed
111: study of the statistical equlibrium of Fe (and a comparison element,
112: either O or S). For this reason the program was carried out for only
113: 5 HH objects, namely for HH 1, HH 7, HH 11, HH 43A, and HH 255
114: (Burnham's Nebula; BBN94, BBN96).
115:
116: Many of their results are surprising and do not (in general) agree
117: with our above mentioned expectations. First, the Fe abundance for
118: both HH1 and HH7 are essentially solar (with only 18\% and 3\%
119: deviation, respectively). This is surprising because HH 1 is the
120: standard high excitation object, while HH 7 is one of the lowest
121: excitation objects known (so we expect a low Fe abundance). Second,
122: the Fe abundance of the low excitation object HH 11 is (for unknown
123: reasons) somewhat higher (by a factor of 1.68) than the solar
124: abundance. Finally, the authors detected some depletion of Fe in HH
125: 43A (Fe/O about 38\% and Fe/S about 48\% of the solar value) and HH
126: 255 (Fe/O about 71\% and Fe/S about 36\% of the solar value), though
127: these are high excitation objects. In judging these results we have to
128: be aware that HH 255 is a somewhat complicated HH object (see B\"ohm
129: \& Solf 1997; Solf \& B\"ohm 1999), to which some of the standard
130: assumptions (e.g., that the line emission is exclusively caused by a
131: bow shock in the working surface of the outflow) may be only partly
132: applicable. Even if we exclude for the moment HH 255 from our
133: consideration, it is clear that the earlier results do not agree with
134: the simple expectations expressed above. One way to explain why the
135: low excitation objects (HH 7 and HH 11) show no Fe depletion is to
136: assume that the shocked gas we observe has passed through a much
137: faster shock at some earlier time.
138:
139: In order to clarify this problem, it is important to determine the gas
140: phase Fe abundance (and correspondingly the destruction of dust
141: grains) in a larger number of HH objects. However, as discussed
142: above, this is very difficult to do in detail. We therefore consider
143: it presently justified to try to determine preliminary, (very)
144: approximate values for the Fe abundance in a larger number of HH
145: objects, even if they are considerably less reliable than the earlier
146: determinations of BBN96.
147:
148: First, we want to mention related studies by other authors. Hartigan
149: et al.\ (1999, 2000) derived gas phase Fe abundance in HH objects
150: which approximately agree with BBN94 and BBN96. In their 1999 paper,
151: they studied the Fe II and [Fe II] spectrum for HH 47A in unusual
152: detail, covering the bow shock and the Mach disk separately. Their
153: conclusion was that the gas phase Fe abundance is approximately solar.
154:
155: It is also worth noting that there has been another recent study of
156: the gas phase Fe abundance which gave a somewhat lower abundance (on
157: the average by a factor of 3) for HH objects by Mouri \& Taniguchi
158: (2000). They use a plane shock wave model to calculate the
159: theoretical line ratio [Fe II] 8617/[O I] 6300 in considerable detail
160: and compare this to observations of HH objects (which are not
161: individually listed). We agree (of course) that HH objects are caused
162: by shock waves, but an attempt to explain in quantitative detail
163: individual spectra by a (plane or bow shock) model leads to
164: difficulties \citep{ragaea96,bg97}. For this reason, the approach of
165: BBN94 and BBN96 was completely different.
166:
167: BBN94 and BBN96 studied the HH objects individually and tried to
168: introduce (where possible) observationally determined parameters
169: (empirically determined densities, sizes, and electron temperatures of
170: the line emitting regions in HH objects). They used different methods
171: to get information about the ionization. In one method, for example,
172: BBN96 determined the abundance ratio
173: \begin{equation}
174: {{\rm N(Fe)}\over{\rm N(O)}} =
175: {{\rm N[Fe\: II] + N[Fe\: III]}\over
176: {\rm N[O\: I] + N[O\: II] + N[O\: III]}} \nonumber
177: \end{equation}
178: (where N[X] is the number of [X] ions). (This does not require
179: detailed theoretical information about the ionization.) They also
180: based the derived abundance information on many lines (e.g., 42 [Fe
181: II] lines in HH 1).
182:
183: Since the approaches of \citet{mt00} and BBN94/96 are completely
184: different (one based on a detailed, but basically very simple model,
185: the other on determining the empirical parameters for individual HH
186: objects), we presently cannot say which is more correct. This will
187: depend on obtaining more insight into the detailed structure of
188: individual HH objects. However, our following approach has been to
189: employ knowledge of individual objects, and consequently, we follow
190: the work of BBN96.
191:
192: We proceed as follows: In \S \ref{method}, we introduce a
193: characteristic number which uses the line fluxes of only a very few
194: lines which are observed in reasonably many HH spectra (which are not
195: taken with the special purpose of determining the Fe abundance, see
196: Raga, B\"ohm, \& Canto 1996 for a reference to appropriate spectra).
197: We show that the characteristic number is well correlated with the
198: actual Fe abundance for the 5 HH objects for which the Fe abundance
199: has been determined in detail (BBN96). In \S \ref{application}, we
200: use this correlation to determine approximately the gas phase Fe
201: abundance in the other HH objects. Finally, in \S \ref{conclusions},
202: we discuss the enigmatic results and suggest possible explanations.
203:
204:
205:
206:
207:
208: \section{An Approximate Method for Determining Gas-phase Fe Abundance
209: \label{method}}
210:
211: We want to develop a diagnostic method that uses only a few frequently
212: observed lines. We will use the Fe lines [Fe II] 5159 and [Fe II]
213: 7155, which have been observed in many objects. It would be nice to
214: add a third Fe line to the criterion. Since there are no other Fe
215: lines as frequently observed in the optical spectrum as these, we have
216: also used the [Ca II] 7291 line. This may appear somewhat
217: inconsistent, since we wish to determine the Fe abundance. However,
218: one has to remember that the gas phase abundance of Ca behaves similar
219: to that of Fe (with regard to the dust grain destruction). (One also
220: has to keep in mind that we are constructing a characteristic
221: number which will be approximately correlated with the Fe abundance
222: but is not the Fe abundance itself.) Using these lines, we define
223: \begin{equation}
224: \label{ame}
225: A_{\rm me} \propto
226: {{\rm F}_{\rm [Fe\: II]\; 5159}\over{\rm F}_{\rm H\beta}} +
227: {{\rm F}_{\rm [Fe\: II]\; 7155}\over{\rm F}_{\rm H\alpha}} +
228: {{\rm F}_{\rm [Ca\: II]\; 7291}\over{\rm F}_{\rm H\alpha}}
229: \end{equation}
230: where F$_{\rm [Fe II]\; 5159}$ is the measured flux of the [Fe II]
231: 5159 line, and similarly for the [Fe II] 7155 and [Ca II] 7291 lines.
232: The [Fe II] 5159 line is measured relative to the H$\beta$ flux, while
233: [Fe II] 7155 and [Ca II] 7291 are relative to H$\alpha$. This is done
234: to keep the error which is introduced by the different reddening of
235: the HH objects small. In order to get a feeling for the possible
236: error of this assumption, we use the reddening for HH 1
237: \citep{solfea88}. In this case, E$_{\rm B - V} \sim 0.43$ (A$_{\rm V}
238: \sim 1.35$). This leads to an error in F$_{\rm [Fe\: II]\; 5159}/{\rm
239: F}_{H\beta}$ of 15.6\% if we use the directly measured ratio instead
240: of the reddening corrected ratio. For the case of F$_{\rm [Ca\: II]\;
241: 7291}/{\rm F}_{\rm H\alpha}$, the corresponding error is 13.4\%. We
242: choose a proportionality constant so that all $A_{\rm me}$ values are
243: given relative to the value for HH 1. It would also be possible to
244: use the near IR [Fe II] lines (see, e.g., Gredel 1994), but the
245: influence of the reddening on the ratio of, e.g., [Fe II] 12570, 16439
246: to H$\alpha$ would cause additional problems.
247:
248: The question of whether this (somewhat arbitrary) definition of
249: $A_{\rm me}$ can be justified is best addressed by looking at the
250: correlation of $A_{\rm me}$ with the Fe gas phase abundances of the 5
251: HH objects whose abundances have been studied in detail (BBN94,
252: BBN96). In Figure \ref{fig1}, this correlation between $A_{\rm me}$
253: and the determined gas phase abundance of Fe is evident. Here, the
254: iron abundance is normalized to 1 for the solar Fe abundance. (This
255: corresponds to the abundance ratio N(Fe)/N(S) = 2.95 or N(Fe)/N(O) =
256: 0.55, see Grevesse \& Anders 1991.) Is it surprising that there is a
257: reasonable correlation between such a simple interpolation formula as
258: given in Equation \ref{ame} and the carefully determined Fe abundance?
259: One might have expected that the characteristics of an individual HH
260: object in question might have a sizable influence on the relation
261: between $A_{\rm me}$ and the Fe abundance. However, one has to keep
262: in mind that the high excitation objects all have relatively similar
263: shock structure (shock velocities between 85 and 115 km s$^{-1}$, see
264: B\"ohm \& Goodson 1997). It is therefore apparent that, for these
265: objects, the change in $A_{\rm me}$ is mostly caused by a change in
266: gas phase Fe abundance, while changes of the model are much less
267: important.
268:
269: It seems more surprising that there appears to be a continuous
270: extension of the (approximate) correlation from the high excitation
271: objects to the low excitation objects (which would not be the case if
272: the correlation depended only on the shock velocity). However, in
273: this respect, it is important that (for reasons which are not fully
274: understood) there is a typical observed difference between low and
275: high excitation objects. High excitation objects typically have a
276: pre-shock density of $\sim 10^2$ cm$^{-3}$, while low excitation
277: objects have a typical pre-shock density of $\sim 10^3$ cm$^{-3}$
278: \citep{bg97,ragaea96}. This change may contribute to an almost
279: continuous extension of the relation between $A_{\rm me}$ and Fe
280: abundance (Fig. \ref{fig1}).
281:
282: For the purpose of determining the approximate Fe abundance of
283: additional HH objects, we assume that all five objects studied by
284: BBN96 follow a one-dimensional relationship between $A_{\rm me}$ and
285: Fe abundance. The data is well represented by a least squares fit to
286: a second order polynomial. We assume, then, that the Fe abundance
287: (relative to the sun) of an HH object with an $A_{\rm me}$ value
288: (relative to HH 1) in the range of the abscissa of Figure \ref{fig1}
289: is approximately
290: \begin{equation}
291: \label{fit}
292: {\rm Fe \;\; abund.} \approx 0.36 + 0.39 A_{\rm me} + 0.36 A_{\rm me}^2
293: \end{equation}
294: This function overplots the data in Figure \ref{fig1} and is applied
295: to new $A_{\rm me}$ values in Figure \ref{fig2} (see \S
296: \ref{application}). We want to point out that we have no theoretical
297: justification for the mathematical form of Equation \ref{fit}.
298: However, the qualitative results of this work (discussed below) are
299: not at all affected by the function chosen to fit the data.
300:
301:
302:
303:
304:
305: \section{Fe Abundance Determination \label{application}}
306:
307: We determined the $A_{\rm me}$ values for a number of additional HH
308: objects using the photometric data from the reference list put
309: together by Raga et al.\ (1996) in the footnote of their table 1. The
310: additional objects comprise 13 high and 3 low excitation HH objects.
311: We would like to apply the correlation shown in Figure \ref{fig1} and
312: given in equation \ref{fit} to these objects. To do so, we assume
313: that the earlier studied HH objects are somehow ``typical,'' or that
314: $A_{\rm me}$ is insensitive to the differences in the detailed
315: structure of each object. Also, the new objects whose Fe abundances
316: we want to determine must have $A_{\rm me}$ values in the range of the
317: 5 HH objects whose Fe abundances have been determined in detail. The
318: objects studied in full detail cover the range from $A_{\rm me}$ =
319: 0.143 (HH 43A) to 1.452 (HH 11), while the newly studied objects cover
320: the range from $A_{\rm me}$ = 0.111 (HH 123) to 1.601 (HH 10). In
321: other words, the newly added objects cover only a slightly larger
322: range of $A_{\rm me}$. If we take into account that our selection of
323: objects was sort of arbitrary (based on the availability of spectra),
324: the fact that the range in $A_{\rm me}$ is relatively narrow may
325: already tell us that, even among our extended list of HH objects,
326: there is none whose gas phase Fe abundance deviates strongly from the
327: range of the 5 objects which were earlier studied in detail.
328: Consequently, we see no difficulty in this respect with applying our
329: approximate method.
330:
331:
332:
333:
334:
335:
336:
337: \subsection{Gas Phase Fe Abundances for High Excitation HH Objects
338: \label{heos}}
339:
340: At first we restrict ourselves to the study of high excitation objects
341: because, in the present context, they are less problematic. As pointed
342: out above, high excitation objects are physically similar to each
343: other, so the correlation of $A_{\rm me}$ with the Fe abundance
344: (Fig. \ref{fig1}) is rather convincing. We use Equation \ref{fit} and
345: the measured $A_{\rm me}$ values for our list of additional objects to
346: determine their Fe abundance. The position of these objects in the
347: $A_{\rm me}$-Fe abundance diagram is shown in Figure \ref{fig2}.
348:
349: In Figure \ref{fig3} we have plotted the Fe abundance as a function of
350: the excitation of the HH object (i.e., as a function of the shock
351: velocity), which is most easily measured by the line flux ratio [O
352: III] 5007/H$\beta$ in high excitation objects (Fig. 3a) and by the
353: flux ratio [N II] 5200/H$\beta$ in low excitation objects (Fig. 3b).
354: The purpose of doing this is to determine whether the gas phase Fe
355: abundance depends on shock velocity. If there occurs a drastic grain
356: destruction at some ``critical'' velocity (leading to a drastic
357: increase in the Fe abundance), there would be a discontinuity (or a
358: steep slope) in the Fe abundance in Figure 3. A discontinuity could
359: also indicate that, at certain shock velocity, our approximate method
360: for the Fe abundance determination breaks down.
361:
362: Figure \ref{fig3}, however, indicates no dependence of the Fe
363: abundance on the excitation. The results for the high excitation
364: objects (Fig. 3a) show an abundance close to 1 but slightly higher
365: than 1 for 4 objects (HH 54B: 1.05, HH 1: 1.18, HH 40: 1.32, HH 54C:
366: 1.42). The other (12) objects show a (relatively modest) Fe gas phase
367: abundance depletion which seems to reach a limit of approximately a
368: factor 2.5 (HH 123 and HH 43A). Our tentative conclusion is that, in
369: our sample, there are 4 objects which have basically unchanged
370: population I abundances for Fe (the reason being that dust grains have
371: been destroyed by shock waves and the Fe has gone back into the gas
372: phase). We leave it open, whether the very slight overabundance (5\% -
373: 40\%) in these objects has to be considered as real or whether it may
374: be a consequence of our approximate method. In addition to these
375: objects, there are also high excitation objects (including HH 123, HH
376: 43A or HH 2A) where a certain fraction has gone back into the gas
377: phase, but a part may still be contained in dust. Thus, for high
378: excitation HH objects, it seems reasonable that there are no objects
379: for which a reasonably large fraction is still bound in dust.
380: However, if we take into account that the dust destruction is a very
381: sensitive function of the shock velocity \citep{mckeeea87}, it is
382: somewhat surprising that where there is depletion, it is never larger
383: than a factor of 2.5.
384:
385: It is also very interesting to note that different parts of the same
386: HH objects (which may be due to different shock waves, like HH2A and
387: HH 2G; see Eisloeffel, Mundt \& B\"ohm 1994) tend to show the same gas
388: phase Fe abundances. This is obvious for HH 2A and HH 2G as well as
389: for HH 54B and HH 54C. Surprisingly, this is also true for the objects
390: HH 43A and HH 43B,C, where HH 43A is a high excitation and HH 43B,C
391: are low excitation objects (see Fig. 3a and 3b). It may however be
392: true that HH 43A and HH 43B,C form the apex and the ``wing'' of the
393: same bow shock (see the isophotic contour diagram of HH 43 presented
394: in B\"ohm \& Solf 1990, fig. 1), so that in this case, the result is
395: not too surprising.
396:
397: Though there are some aspects of our Fe abundance results for high
398: excitation HH objects which we do not yet completely understand, we
399: can argue that the results in Figure 3a correspond approximately to
400: our expectations. As we shall now see the situation is not as
401: satisfactory when we consider the low excitation HH objects.
402:
403:
404:
405:
406:
407:
408: \subsection{Approximate Fe Abundance for Low Excitation HH Objects
409: \label{leos}}
410:
411: We proceed in an analogous way to the one used for high excitation
412: objects. However, in this case we have available a total of only 5
413: objects, and the somewhat unexpected results have to be interpreted
414: with considerable caution. Of these 5 objects, 2 have been analyzed in
415: full detail for their gas phase Fe abundances (BBN96), namely HH 7 and
416: HH 11. For the other 3 objects, we determined the Fe abundance using
417: the same procedure as for the high excitation objects (\S
418: \ref{heos}). The results are shown in Figures 2 and 3b. In 3b, we
419: have plotted the results again as a function of excitation (shock wave
420: velocity), which is now (for the low excitation objects) measured by
421: the line ratio [N I] 5200/H$\beta$ (Raga et al.\ 1996) because the [O
422: III] line used in the high excitation objects is not emitted in the
423: low excitation objects. Our results show that (surprisingly) 4 out of
424: 5 low excitation outflows show a gas phase Fe abundance which lies
425: somewhat above the solar Fe abundance (by a factor of 1.02 to 1.86).
426: There is only one low excitation HH object, namely HH 43B,C, which
427: shows some Fe depletion. But even in this case, the Fe depletion is
428: almost identical to the Fe depletion of the high excitation object HH
429: 43A, with which it seems to be connected (see above).
430:
431: The fact that 4 of the low excitation objects show an Fe abundance
432: somewhat larger than 1 may not be significant because our approximate
433: method may not be very accurate, especially for low excitation
434: objects. It is possible that the Fe abundance in these objects
435: corresponds reasonably well to the solar abundance. We are, however,
436: surprised that 4 out of 5 low excitation HH objects show no depletion
437: of the gas phase Fe abundances. We would have (naively) expected that
438: these objects show strong Fe depletion (by a factor of roughly 100)
439: because no dust grains would have been destroyed in the weak shock
440: waves of the low excitation objects. We would expect to see the
441: original gas phase composition of the molecular cloud. We must
442: conclude that the matter which is visible now in a low excitation HH
443: object went through a shock wave of a much higher shock velocity at an
444: earlier time.
445:
446:
447:
448:
449:
450:
451:
452:
453: \section{Conclusions \label{conclusions}}
454:
455: Earlier detailed studies of the gas phase Fe abundance for a very few
456: HH objects had led to the conclusion that there is no simple
457: correlation between the Fe abundance and the HH shock velocity (BBN94,
458: BBN96). A correlation would have been expected if the molecular cloud
459: matter enters the HH working surface and dust grains are destroyed in
460: the HH bow shock by the usual processes (mostly by non-thermal
461: sputtering, see e.g., McKee et al.\ 1987; Draine 1995). Since the
462: study of BBN96 is based on only 5 objects (HH 1, HH 7, HH 11, HH 43A,
463: HH 255), and adding more objects to this list using a detailed method
464: requires great effort, we have derived Fe abundances of 16 additional
465: HH objects using a very approximate method.
466:
467: We introduced a characteristic number, $A_{\rm me}$ (Eqn. \ref{fit}),
468: which depends on the (frequently measured) line fluxes of [Fe II] 5159
469: and [Fe II] 7155, as well as on the [Ca II] 7291 line flux and the
470: H$\alpha$ and H$\beta$ fluxes. For the five HH objects (mentioned
471: above) for which we have a detailed gas phase analysis, $A_{\rm me}$
472: is correlated reasonably well with the gas phase Fe abundance. If we
473: assume that this correlation is not only approximately correct for the
474: 5 objects studied in detail, but for all HH objects for which we can
475: determine $A_{\rm me}$, then we can use this correlation to determine
476: their gas phase Fe abundance.
477:
478: For the high excitation HH objects, our results are the following. In
479: addition to the 3 high excitation objects for which the Fe gas phase
480: abundances have been measured in detail (BBN96), we have now
481: determined the Fe gas phase abundance for 13 more objects. Of these 16
482: objects, 4 show a gas phase Fe abundance between 1.0 and 1.44 times
483: the solar abundance. Taking into account the possible errors of the
484: abundance determination, this probably means that 4 of these objects
485: have unchanged population I Fe gas phase abundances (no Fe bound in
486: dust grains). The other 12 high excitation objects show some (but
487: relatively small) Fe depletion which never gets larger than a factor
488: of 2.5.
489:
490: In general, the result looks reasonable. The high excitation HH
491: objects show that, in some cases, Fe has gone back completely to the
492: gas phase, and in other cases, some (but not too much) Fe may still be
493: bound in dust grains, leading to moderate depletion of the gas phase
494: abundance. We also find that similar depletion exists in different
495: parts of the same object (e.g., in HH 2A and HH 2G or in HH 54B and HH
496: 54C). The gas phase Fe abundance for high excitation objects agrees
497: qualitatively with the predictions given, e.g., by McKee et al.\
498: (1987) that dust grains are destroyed by strong shocks. The limit of
499: the depletion factor of 2.5 is interesting but has not been explained
500: in detail.
501:
502: Understanding the results for the low excitation HH objects seems to
503: be more difficult, and some questions remain open. The first problem
504: is that we could study only 5 objects. For these objects we find quite
505: unexpected results. Four of them have an Fe gas phase abundance
506: between 1.02 and 1.86 times the solar Fe abundances. One may argue
507: that our uncertainty (which is supposedly larger in the low excitation
508: than in the high excitation objects) might permit these to be
509: consistent with 1 solar Fe abundance. However, we still have the
510: problem that low excitation objects (for which we do not expect
511: destruction of dust grains) apparently have all or most of their Fe in
512: the gas phase.
513:
514: It is also important that we find no correlation of the Fe abundance
515: with shock velocity. While this is not surprising for the high
516: excitation objects, we would have (naively) expected a large
517: difference between the low and high excitation objects. Specifically,
518: we would have expected a strong Fe depletion (perhaps by a factor of
519: 100) in the low excitation objects, since their shocks are not strong
520: enough to destroy grains \citep{mckeeea87,draine95}. The (perhaps)
521: simplest explanation is that the material we have studied in low
522: excitation objects has passed through a faster shock at earlier times.
523: This may be a surprising result, but we cannot see any way to avoid
524: it.
525:
526: It is difficult to speculate about this process. What would be
527: required is that the matter that enters the low excitation bow shock
528: (which we assume to be molecular cloud matter) has previously gone
529: through a high excitation bow shock (in which the dust grains were
530: destroyed). The simplest assumption would be that the previous high
531: excitation bow shock was generated by the same source star that
532: generates the low excitation HH outflow. This suggests that none of
533: the five low excitation objects studied in this work correspond to the
534: leading edge of the stellar outflow (i.e., all of them are produced
535: some time after the onset of the jet).
536:
537: It is also interesting to note that we have assumed the molecular
538: clouds in which these HH objects reside have approximately solar Fe
539: abundance (in gas and dust phases). Though this is a reasonable
540: assumption, it may be possible that the abundances in present-day
541: molecular clouds could differ from solar by a factor of a few. The
542: entire range of Fe abundances of the objects we have studied is from
543: $\sim 0.4$ to 1.9 times solar. In light of this, one may argue that
544: the dust grains are completely destroyed (by passing through fast
545: shocks) in all 21 objects, and that the observed spread in Fe
546: abundance is due to different abundances in their parent molecular
547: clouds. At present, our data cannot resolve this issue.
548:
549:
550:
551:
552: \acknowledgments
553:
554: We thank the referee, Bo Reipurth, for carefully reading the
555: manuscript and making very useful suggestions for improvement. This
556: research has been supported by NSF grant AST-9729096.
557:
558: \begin{thebibliography}{}
559:
560: \bibitem[Beck-Winchatz et al.(1994)]{beck-winchatzea94}Beck-Winchatz,
561: B., B\"ohm, K.-H., \& Noriega-Crespo, A. 1994, \pasp, 106, 1271
562: (BBN94)
563:
564: \bibitem[Beck-Winchatz et al.(1996)]{beck-winchatzea96}Beck-Winchatz, B.,
565: B\"ohm, K.-H., \& Noriega-Crespo, A. 1996, \aj, 111, 346 (BBN96)
566:
567: %\bibitem[B\"ohm(1995)]{Bohm95}B\"ohm, K.-H. 1995, ApSS, 233, 11
568:
569: \bibitem[B\"ohm \& Goodson(1997)]{bg97}B\"ohm, K.-H., \& Goodson, A. P.
570: 1997, Herbig-Haro Flows and the Birth of Low Mass Stars, IAU
571: Symp. No. 182, ed. B. Reipurth \& Cl. Bertout (Dordrecht: Kluwer
572: Acad. Publ.), p. 47
573:
574: \bibitem[B\"ohm \& Solf(1985)]{bs85}B\"ohm, K.-H. \& Solf, J. 1985,
575: \apj, 294, 533
576:
577: \bibitem[B\"ohm \& Solf(1990)]{bs90}B\"ohm, K.-H. \& Solf, J. 1990,
578: \apj, 348, 297
579:
580: \bibitem[B\"ohm \& Solf(1997)]{bs97}B\"ohm, K.-H. \& Solf, J. 1997,
581: \aap, 318, 565
582:
583: \bibitem[Draine(1995)]{draine95}Draine, B. T. 1995, ApSS, 233, 111
584:
585: \bibitem[Eisloeffel et al.(1994)]{eisloeffelea94}Eisloeffel, J., Mundt,
586: R., \& B\"ohm, K.-H. 1994, AJ, 108, 1042
587:
588: \bibitem[Gredel(1994)]{gredel94}Gredel, R. 1994, \aap, 292, 580
589:
590: \bibitem[Grevesse \& Anders(1991)]{ga91}Grevesse, N. \& Anders, E. 1991,
591: The Solar Interior and Atmosphere, ed. A. N. Cox, W. C. Livingstone,
592: \& M. S. Matthews (Tucson: The University of Arizona Press), p. 1227
593:
594: \bibitem[Hartigan(1989)]{hartigan89}Hartigan, P. 1989, \apj, 339, 987
595:
596: \bibitem[Hartigan et al.(2000)]{hartiganea00}Hartigan, P., Bally, J.,
597: Reipurth, B., \& Morse, J. A. 2000, in Protostars \& Planets IV,
598: ed. Y. Mannings, A. P. Boss, \& S. S. Russell (Tucson, Univ. of
599: Arizona Press), p. 841
600:
601: \bibitem[Hartigan et al.(1999)]{hartiganea99}Hartigan, P., Morse, J. A.,
602: Tumlinson, J., Raymond, J., \& Heathcote, S. 1999, \apj, 512, 901
603:
604: \bibitem[Hartigan et al.(1987)]{hartiganea87}Hartigan, P., Raymond, J.,
605: \& Hartmann, L. 1987, \apj, 316, 323
606:
607: \bibitem[Hartmann \& Raymond(1984)]{hr84}Hartmann, L. \& Raymond, J.
608: 1984, \apj, 276, 560
609:
610: \bibitem[McKee et al.(1987)]{mckeeea87}McKee, C. F., Hollenbach, D. I.,
611: Seab, C. G., \& Tielens, A. G. G. M. 1987, \apj, 318, 674
612:
613: \bibitem[Mouri \& Taniguchi(2000)]{mt00}Mouri, H., \& Taniguchi, Y.
614: 2000, \apj, 534, L63
615:
616: \bibitem[Mundt(1985)]{mundt85}Mundt, R. 1985, in Protostars and Planets II,
617: ed. D. C. Black \& M. S. Matthews (Tucson: University of Arizona
618: Press), p. 414
619:
620: \bibitem[Mundt \& Fried(1983)]{mf83}Mundt, R. \& Fried, J. W. 1983, \apj,
621: 274, L83
622:
623: \bibitem[Raga(1995)]{raga95}Raga, A. C. 1995, RevMex AASC, 1, 103
624:
625: %\bibitem[Raga \& B\"ohm(1986)]{rb86}Raga, A. C. \& B\"ohm, K.-H. 1986,
626: %\apj, 308, 829
627:
628: \bibitem[Raga et al.(1996)]{ragaea96}Raga, A. C., B\"ohm, K.-H., \& Canto,
629: J. 1996, RevMex AA, 32, 161
630:
631: \bibitem[Raga \& Noriega-Crespo(1998)]{rn98}Raga, A. C., \& Noriega-Crespo,
632: A. 1998, \aj, 116, 2943
633:
634: \bibitem[Reipurth(1989)]{reipurth89}Reipurth, B. 1989, in Low Mass Star
635: Formation and Pre-Main Sequence Objects, ESO, ed. B. Reipurth, p. 247
636:
637: \bibitem[Reipurth(1999)]{reipurth99}Reipurth, B. 1999, A General
638: Catalogue of Herbig-Haro Objects, 2. edition,
639: http://casa.colorado.edu/hhcat
640:
641: \bibitem[Reipurth \& Heathcote(1993)]{rh93}Reipurth, B., \& Heathcote,
642: S. 1993, in Astrophysical Jets, ed. D. Burgarella, M. Livio,
643: C. P. O'Dea, Cambridge Univ. Press, p. 35
644:
645: \bibitem[Reipurth \& Heathcote(1997)]{rh97}Reipurth, B. \& Heathcote, S.
646: 1997, in Herbig-Haro Flows and the Birth of Low Mass Stars, IAU
647: Symp. No. 182, ed. B. Reipurth \& Cl. Bertout (Dordrecht: Kluwer
648: Acad. Publ.), p. 3
649:
650: \bibitem[Reipurth et al.(1997)]{reipurthea97}Reipurth, B., Hartigan, P.,
651: Heathcote, S., \& Morse, J. A. 1997, \aj, 114, 757
652:
653: \bibitem[Schwartz(1985)]{schwartz85}Schwartz, R. D. 1985, in Protostars
654: and Planets II, ed. D. C. Black \& M. S. Matthews (Tucson: University
655: of Arizona Press), p. 405
656:
657: \bibitem[Seab(1987)]{seab87}Seab, C. G. 1987, in Interstellar Processes,
658: ed. D. I. Hollenbach \& H. A. Thronson (Dordrecht: Reidel
659: Publ. Corp.), p. 491
660:
661: \bibitem[Solf \& B\"ohm(1991)]{sb91}Solf, J., \& B\"ohm, K.-H. 1991,
662: \apj, 375, 618
663:
664: \bibitem[Solf \& B\"ohm(1999)]{sb99}Solf, J., \& B\"ohm, K.-H. 1999,
665: \apj, 523, 709
666:
667: \bibitem[Solf et al.(1988)]{solfea88}Solf, J., B\"ohm, K.-H., \& Raga,
668: A. 1988, \apj, 334, 229
669:
670: \bibitem[van Dieshoek et al.(1993)]{vandieshoek93}van Dieshoek, E. F.,
671: Blake, G. A., Draine, B. T., \& Lunine, J. I. 1993, in Protostars and
672: Planets III, ed. E. H. Levy \& J. I. Lunine (Tucson: Univ. of Arizona
673: Pr.), p. 163
674:
675: \end{thebibliography}
676:
677: %Comment the next line out if you want figures.
678: %\end{document}
679:
680: \begin{figure}[t]
681: \plotone{fg1.eps}
682:
683: \caption{Five HH objects (whose Fe abundance was determined in detail
684: by BBN96) indicate a possible correlation between the Fe abundance and
685: the characteristic number, $A_{\rm me}$ (see text). High excitation
686: objects are designated by ``plus'' symbols, while ``x'' symbols
687: correspond to low excitation objects. The HH designation is indicated
688: next to each symbol (the number and letter designation identifies the
689: HH object by the name assigned to it in Reipurth's 1999 catalogue).
690: The solid line is the least squares fit of a second order polynomial
691: to the data. \label{fig1}}
692:
693: \end{figure}
694:
695:
696:
697: \begin{figure}[t]
698: \plotone{fg2.eps}
699:
700: \caption{The $A_{\rm me}$ values of 16 additional HH objects (derived
701: from data in the literature; see, e.g., the summary by Raga et al.\
702: 1996) and the second order polynomial fit to the data in
703: Fig. \ref{fig1} (solid line) determine the approximate Fe abundance
704: for each object. The symbols have the same meaning as in
705: Fig. \ref{fig1}, and the HH designation is indicated next to each
706: point. \label{fig2}}
707:
708: \end{figure}
709:
710:
711:
712: \begin{figure}[t]
713: \plotone{fg3.eps}
714:
715: \caption{Gas phase Fe abundance as a function of ``excitation'' (see
716: text, a measure of shock wave velocity) for high excitation (a) and
717: low excitation (b) HH outflows. In (a), excitation increases to the
718: right, in (b) to the left. No obvious correlation exists between the
719: Fe abundance of the 21 HH objects and their excitation. The HH
720: designation is indicated to the right of each point. \label{fig3}}
721:
722: \end{figure}
723:
724: \end{document}
725: