1: \documentclass[12pt,preprint]{emulateapj}
2:
3: %\usepackage{graphicx}
4:
5: \begin{document}
6:
7: \title{Molecular Hydrogen Formation on Amorphous Silicates \\
8: Under Interstellar Conditions }
9:
10: \author{H.B. Perets\altaffilmark{1},
11: A. Lederhendler\altaffilmark{2},
12: O. Biham\altaffilmark{2},
13: G. Vidali\altaffilmark{3},
14: L. Li\altaffilmark{3},
15: S. Swords\altaffilmark{3},
16: E. Congiu\altaffilmark{3,4,5},
17: J. Roser\altaffilmark{6},
18: G. Manic\'o\altaffilmark{5},
19: J.R. Brucato\altaffilmark{7}
20: and
21: V. Pirronello\altaffilmark{5}
22: }
23:
24: \altaffiltext{1}{Faculty of Physics, Weizmann Institue of Science, Rehovot 76100, Israel}
25: \altaffiltext{2}{Racah Institute of Physics, The Hebrew University, Jerusalem 91904, Israel}
26: \altaffiltext{3}{Physics Department, Syracuse University, Syracuse, NY 13244, USA}
27: \altaffiltext{4}{Universit\'a di Cagliari, Dipartimento di Fisica, Cagliari Italy}
28: \altaffiltext{5}{Universit\'a di Catania, DMFCI, 95125 Catania, Sicily, Italy}
29: \altaffiltext{6}{NASA Ames, Mail Stop 245-6, Moffett Field, CA, 94035, USA}
30: \altaffiltext{7}{INAF-Osservatorio Astronomico di Capodimonte, Napoli, Italy}
31:
32: \begin{abstract}
33:
34: Experimental results on the formation of molecular hydrogen on amorphous
35: silicate surfaces are
36: presented for the first time and
37: analyzed using a rate equation model.
38: The energy barriers for the relevant diffusion and
39: desorption processes are obtained.
40: They turn out to be significantly higher than those obtained
41: earlier for polycrystalline silicates,
42: demonstrating the importance of grain morphology.
43: Using these barriers we evaluate the
44: efficiency of molecular hydrogen formation on amorphous silicate
45: grains under interstellar conditions.
46: It is found that unlike polycrystalline silicates, amorphous silicate
47: grains are efficient catalysts of H$_{2}$ formation within a temperature
48: range which is relevant to diffuse interstellar clouds.
49: The results also indicate that the hydrogen molecules
50: are thermalized with the surface
51: and desorb with low kinetic energy.
52: Thus, they are unlikely to occupy highly excited states.
53:
54: \end{abstract}
55:
56: \keywords{dust --- ISM: molecules --- molecular processes}
57:
58: \section{Introduction}
59: \label{intro.}
60:
61: H$_{2}$ is the most abundant molecule in the interstellar medium (ISM).
62: It plays a crucial role in the initial cooling of clouds during
63: gravitational collapse and is involved in most reaction schemes
64: that produce other molecules
65: \citep{Duley1984}.
66: It is widely accepted that
67: H$_{2}$ formation in the ISM takes place on dust grains
68: \citep{Gould1963}.
69: In this process,
70: H atoms that collide with a grain and adsorb on its surface
71: quickly equilibrate and diffuse
72: either by thermal activation or tunneling.
73: They may encounter each other and form
74: H$_{2}$ molecules
75: \citep{Williams1968,Hollenbach1970},
76: or desorb thermally
77: in atomic form.
78:
79: In recent years, we have conducted a series of
80: experiments on the formation of molecular hydrogen on
81: dust grain analogues
82: such as polycrystalline silicates
83: \citep{Pirronello1997a,Pirronello1997b},
84: amorphous carbon
85: \citep{Pirronello1999}
86: and amorphous water ice
87: \citep{Manico2001,Roser2002,Roser2003,Perets2005},
88: under astrophysically relevant conditions.
89: In these experiments, the surface was irradiated by beams of H and D atoms.
90: The production of HD molecules
91: was measured during the irradiation and during
92: a subsequent temperature programmed desorption (TPD) experiment.
93: To disentangle the process of diffusion from the one of desorption,
94: separate experiments were carried out in which molecular species were
95: irradiated on the sample and were later induced to desorb.
96: Related studies were done on
97: amorphous ice surfaces
98: \citep{Hornekaer2003,Hornekaer2005,Dulieu2005,Perets2005,Amiaud2006,Williams2007}.
99:
100: The results were analyzed using rate equation models
101: and the energy barriers for the relevant diffusion and desorption
102: processes were obtained
103: \citep{Katz1999,Cazaux2004,Perets2005}.
104: Using these parameters,
105: the conditions for efficient H$_{2}$
106: formation on different astrophysically relevant surfaces
107: were found.
108: In particular, the
109: formation of H$_{2}$ on polycrystalline silicates
110: was found to be efficient only in
111: a narrow temperature window below 10K.
112: Since the typical dust grain temperature in diffuse interstellar
113: clouds is higher than 10K,
114: these results
115: indicated that polycrystalline silicate grains cannot be efficient
116: catalysts for H$_{2}$ formation in most diffuse clouds.
117:
118: In this \emph{Letter} we present,
119: for the first time,
120: experiments on molecular hydrogen formation
121: on \emph{amorphous} silicates
122: and analyze the results using a suitable rate equation model
123: \cite{Perets2005}.
124: Using the parameters that best fit the experimental
125: results, the efficiency of hydrogen recombination on grains
126: is obtained for
127: a range of conditions pertinent to diffuse interstellar clouds.
128: It is found that
129: unlike the polycrystalline silicate grains,
130: amorphous silicate grains, which are the main silicate
131: component in interstellar clouds
132: \cite{Tielens2005},
133: are efficient catalysts for H$_{2}$ formation within a broad temperature
134: window that extends at least up to about 14K.
135:
136: \section{Experimental Methods and Results}
137:
138: \label{sec:Experimental}
139:
140: The apparatus consists of an ultra-high vacuum
141: chamber housing the sample holder and a detector.
142: The sample can be cooled by liquid helium
143: to $\sim$5K, as measured by a calibrated
144: silicon diode and thermocouple placed
145: in the back of the sample.
146: A heater in the back of the sample is used to maintain
147: a set temperature between 5 and 30K
148: during the irradiation phase of the experiment.
149: The sample and detector can rotate around the vertical axis.
150: Prior to a series of measurements, the sample is heated to 380-400K.
151: During a series of measurements, the sample is taken periodically to
152: 200-250K to desorb condensables.
153: Hydrogen and deuterium gases are dissociated
154: in two radio-frequency dissociation sources,
155: with dissociation efficiency of 80-90\%,
156: and are sent into
157: the sample chamber via two triple
158: differentially pumped beam lines
159: \citep{Vidali2004}.
160:
161: In the experiments reported here, we
162: used beams of low
163: fluxes and short dosing times.
164: Using the standard Langmuir-Hinshelwood analysis,
165: plotting the total yield of HD vs. the exposure time
166: \citep{Biham2001},
167: we estimated the coverage to be a small fraction
168: (a few percent) of a monolayer (ML).
169: This is still far from interstellar values but is within the
170: regime in which results can be safely extrapolated to diffuse
171: cloud conditions
172: \citep{Katz1999,Perets2005}.
173: The interstellar dust analogues we used are amorphous silicate
174: samples,
175: (Fe$_{x}$, Mg$_{1-x}$)$_{2}$SiO$_{4}$,
176: prepared by one of us (J.R.B.) by laser ablation
177: (wavelength 266 nm) of a mixed MgO, FeO and SiO$_{2}$ target in an
178: oxygen atmosphere (10 mbar).
179: The optical and stochiometric characterization
180: of samples produced with this technique is given elsewhere
181: \citep{Brucato2002}.
182: The results reported here are
183: for a sample with $x=0.5$.
184:
185: The experiment consists of adsorbing hydrogen atoms onto the surface
186: while monitoring the amount of hydrogen molecules that are
187: formed.
188: To increase the signal to noise ratio, hydrogen and deuterium
189: atoms are used and the formation of HD is monitored.
190: The measurement of HD formation is done in two steps.
191: First, we record the amount of HD that forms and comes off
192: the surface while the sample is being dosed with H and D atoms
193: (the \emph{irradiation phase}).
194: Next, after dosing is completed, in a
195: TPD experiment, the surface temperature
196: is raised rapidly and the rate of HD desorption
197: is measured
198: (the \emph{TPD phase}).
199: By far, the main contribution comes from the TPD phase.
200:
201: Irradiations with beams of H and D
202: (\char`\"{}H+D\char`\"{} thereafter)
203: were done on an amorphous silicate surface
204: in order to explore the formation processes
205: of HD molecules.
206: The H+D irradiation runs were performed
207: with different irradiation times (15, 30, 60, 120 and 240 s),
208: at a surface temperature of $T_{0} \simeq 5.6$K.
209: In a separate set of experiments, beams
210: of HD molecules were irradiated on the same surface.
211: During the TPD runs, the sample temperature was monitored as a function of
212: time. The temperature ramps $T(t)$
213: deviate from linearity but are highly reproducible
214: (see inset in Fig. 2).
215:
216: \begin{figure}
217: \includegraphics[scale=0.2,angle=270, clip,width=1\columnwidth,keepaspectratio]{f1.ps}
218: %\includegraphics[scale=0.55,angle=270]{f1.ps}
219: \caption{
220: TPD curves of HD desorption
221: after irradiation of
222: HD molecules ($\times$)
223: and
224: H+D atoms ($\square$)
225: on amorphous silicate.
226: Also shown, for comparison,
227: is HD desorption after irradiation
228: with H+D atoms on polycrystalline silicate ($\circ$).
229: }
230: \label{fig:1}
231: \end{figure}
232:
233: \begin{figure}
234: \includegraphics[scale=0.2,angle=270, clip,width=1\columnwidth,keepaspectratio]{f2.ps}
235: %\includegraphics[scale=0.55,angle=270]{f2.ps}
236: \caption{
237: TPD curves of HD desorption
238: after irradiation with H+D atoms on amorphous silicate with
239: irradiation times of 15 ({*}), 30 ($\times$), 60 ($\Box$)
240: 120 (+) and 240 ($\circ$) s.
241: The solid lines are fits
242: obtained using the rate equation model.
243: The temperature ramps (shown in the inset)
244: are identical for all the runs.
245: }
246: \label{fig:2}
247: \end{figure}
248:
249: The desorption rates of HD molecules vs. surface temperature during the
250: TPD runs are shown in Fig. 1,
251: for H+D irradiation on polycrystalline silicate ({\it circles})
252: and amorphous silicate ({\it squares})
253: surfaces, with irradiation times of 120 s.
254: The TPD curve following
255: irradiation of HD molecules on an amorphous silicate
256: surface
257: is also shown
258: ({\it crosses}).
259: The results of current experiments of H+D irradiation on amorphous silicates,
260: clearly differ from those of earlier experiments
261: on polycrystalline silicates.
262: The desorption curves from amorphous silicates contain two wide peaks,
263: located at a significantly higher temperatures than the single narrow
264: peak obtained for the polycrystalline silicate.
265: The higher peak temperatures
266: indicate that the relevant energy
267: barriers are larger, while their large width
268: reflects a broader distribution
269: of the energy barriers of the HD desorption
270: sites.
271: The TPD curve of HD desorption from amorphous silicates, after irradiation
272: with HD molecules (crosses in Fig. 1),
273: is qualitatively similar to the curve obtained for H+D irradiation.
274: In particular, the peak temperatures are the same.
275: The relative weights of the high temperature peaks vs. the
276: low temperature peaks are somewhat different.
277: Also, in similar experiments with higher values of $T_0$,
278: a third peak was observed at higher temperatures.
279: We attribute this behavior to diffusion of HD molecules,
280: which gradually migrate from shallow into
281: deep adsorption sites
282: \cite{Perets2005,Dulieu2005,Amiaud2006}.
283:
284: In Fig. 2 we present a series of
285: TPD curves for HD desorption
286: after irradiation with H+D
287: atoms on an amorphous silicate surface with different
288: irradiation times.
289: Each curve exhibits a large peak at a lower temperature and
290: a broader peak (or a shoulder) at a higher temperature.
291: The position of the high temperature peak is found to be
292: independent of the irradiation time,
293: indicating that this peak exhibits first order kinetics.
294: The low temperature peak shifts to the right
295: as the irradiation time is reduced.
296: Since in the low-coverage regime studied here,
297: the activation energies and the
298: pre-exponential factor are not expected to
299: depend on the coverage,
300: these results indicate that the low temperature
301: peak exhibits second order kinetics.
302: This is unlike the case of irradiation with HD molecules,
303: where the lower-temperature peak does not shift,
304: showing first order kinetics.
305:
306: \section{Analysis of the Experimental Results}
307:
308: \label{sec:analysis}
309:
310: In the model
311: used here, there is no distinction between the H and D atoms, namely
312: the same diffusion and desorption barriers are used for both isotopes.
313: Hydrogen atoms that stick to the surface hop as random walkers
314: and may either encounter each other and form molecules,
315: or desorb from the surface.
316: As the sample temperature is raised, both
317: the diffusion and desorption rates quickly increase.
318: If a large fraction of the energy released when two H
319: atoms recombine
320: is transformed into kinetic energy of the formed molecule, it would
321: immediately desorb from the grain surface in a high
322: ro-vibrational state, and with large translational energy.
323: However, both our experiments on ice
324: \citep{Perets2005}
325: and the current experiments
326: indicate that such prompt desorption does not occur on amorphous surfaces
327: [but see Perry and Price (2003), Tine et al. (2003) and Creighan et al. (2006)
328: for other, more ordered, surfaces].
329: Instead,
330: the newly formed molecules dissipate their energy,
331: probably through multiple collisions with the rough surface or internal
332: pores.
333: These molecules thermalize with the surface and become trapped in adsorption
334: sites before they thermally desorb.
335: Consequently, the desorbed
336: molecules are not highly excited, and desorb only with a thermal energy
337: comparable with the grain surface temperature.
338: Therefore, in our model
339: it is assumed that the newly formed molecules do not promptly
340: desorb, but are trapped in adsorption sites with a range of potential
341: barriers.
342:
343: The experimental results were fitted using
344: the rate equation model described in
345: Perets et al. (2005).
346: The parameters for the diffusion
347: and desorption of hydrogen atoms and molecules
348: on the amorphous silicate surface were obtained.
349: These include the energy barrier
350: $E_{\rm H}^{\rm diff}$
351: for the diffusion of H atoms
352: and the barrier
353: $E_{\rm H}^{\rm des}$
354: for their desorption.
355: The value obtained for the desorption barrier
356: should be considered only as a lower bound,
357: because the TPD results are insensitive
358: to variations in
359: $E_{\rm H}^{\rm des}$,
360: as long as it is higher than the reported value.
361: The desorption barriers of HD molecules adsorbed
362: in shallow (lower temperature peak)
363: and deep (higher temperature peak)
364: sites, are given by
365: $E_{\rm H_{2}}^{\rm des}(j),$
366: where $j=1$, 2, respectively.
367:
368: The rate equation model is integrated using a Runge-Kutta stepper.
369: For any given choice of the parameters, one obtains a set of TPD
370: curves for the different irradiation times used in the experiments.
371: The actual temperature curve of the sample, recorded during
372: the experiment, is used in the analysis
373: (see inset in Fig. 2).
374: In the first step, the barriers
375: $E_{{\rm H_{2}}}^{{\rm des}}(j)$, $j=1$, 2,
376: for the desorption of molecules
377: are obtained using the results of the experiments in which
378: HD molecules are irradiated on the surface.
379: To obtain better fits, we incorporate suitable
380: Gaussian distributions of energy barriers
381: around these two values.
382: In the second step, the barriers for diffusion and desorption
383: of H atoms are obtained,
384: using the model to fit the results of
385: H+D irradiation experiments
386: (Table 1).
387:
388: \begin{table}
389: \caption{
390: Energy barriers for diffusion and desorption.
391: }
392: \vspace{0.1in}
393: \begin{centering}
394: \begin{tabular}{lccc}
395: \hline
396: Material &
397: $E_{\rm H}^{\rm diff}$(meV) &
398: $E_{\rm H}^{\rm des}$(meV) &
399: $E_{\rm HD}^{\rm des}$(meV)
400: \\
401: \hline
402: Polycrystalline Silicate&
403: 25&
404: 32 &
405: 27 \\
406: Amorphous Silicate&
407: 35&
408: 44&
409: 35, 53 \\
410: \hline
411: \end{tabular}
412: \par
413: \end{centering}
414: \label{t:E_barriers}
415: \end{table}
416:
417:
418: The second order behavior of the low temperature
419: peak in the H+D irradiation
420: experiments can be explained as follows.
421: Most HD molecules are formed only when the surface temperature
422: is sufficiently high to enable significant mobility of H and D
423: atoms.
424: At this temperature, the shallow adsorption sites
425: cannot retain the newly formed molecules adsorbed in
426: these sites, which quickly desorb from the surface
427: by thermal activation.
428: However, those newly formed
429: molecules which are trapped in deeper sites
430: do not desorb yet, and remain on the surface
431: until its temperature increases further.
432: Thus, the high temperature peak exhibits
433: first order kinetics and is located
434: at the same temperatures in both HD and H+D irradiation experiments.
435: Note that we cannot exclude the possibility that some fraction
436: of the molecules desorbed in the high temperature peak
437: are formed during irradiation and remain trapped
438: in deep adsorption sites until the temperature becomes sufficiently
439: high for them to desorb.
440:
441: At longer H+D irradiation times, more atoms are adsorbed on the
442: surface, and they find each other more easily.
443: The low temperature peak
444: (corresponding to HD desorption from shallower sites) shifts
445: to lower temperatures with longer irradation times,
446: thus showing second order behavior.
447: This behavior must saturate with even longer irradiation times,
448: when the temperature needed for HD formation becomes
449: lower than the temperature needed for thermal desorption from the
450: shallower HD adsorption sites.
451: The HD molecules then form at
452: low temperatures, but are trapped in the shallow sites.
453: They then await until the temperature is further increased,
454: much like the adsorbed HD molecules in the HD irradiation experiments.
455: Therfore, at longer irradiation times the
456: low temperature peak should saturate
457: into first-order like behavior and the TPD curves of the H+D and HD
458: irradiation experiments should become more similar. This behavior
459: is confirmed by the TPD curves
460: for H+D and HD irradiation shown in
461: Fig. 1.
462:
463: Using the parameters obtained from the experiments we now calculate
464: the recombination efficiency of hydrogen on
465: amorphous silicate surfaces
466: under interstellar conditions.
467: The recombination efficiency is defined
468: as the fraction of hydrogen atoms adsorbed on the surface which come
469: out as molecules.
470: In Fig. 3 we present the recombination efficiency
471: vs. surface temperature for
472: the amorphous silicate sample under flux of
473: 5.2 $\times$ $10^{-10}$ (ML s$^{-1}$).
474: This flux is within the typical
475: range for diffuse interstellar clouds,
476: where bare amorphous silicate grains are expected to play a crucial role
477: in H$_2$ formation.
478: This flux
479: corresponds to gas density of 10
480: (atoms cm$^{-3}$),
481: gas temperature of 100K and
482: density of
483: $7 \times 10^{14}$ adsorption
484: sites per cm$^{2}$ on the surface,
485: obtained using the procedure
486: described in Biham et al. (2001).
487: \begin{figure}
488: \includegraphics[scale=0.3, clip]{f3.eps}
489: %\includegraphics[scale=0.55]{f3}
490: \caption{
491: Calculated recombination efficiency of Hydrogen at steady
492: state on amorphous silicate (solid line) and polycrystalline silicate (dashed line) vs.
493: temperature, using the parameters obtained from
494: the TPD experiments.
495: }
496: \label{fig:3}
497: \end{figure}
498:
499: A window of high recombination efficiency is found between 8-13K,
500: compared to 6-10K for polycrystalline silicate under similar conditions.
501: For gas density of 100
502: (atoms cm$^{-2}$),
503: the high efficiency window for the amorphous silicates
504: surface shifts to 9-14K.
505: At higher temperatures
506: atoms desorb from the surface before they have sufficient time to
507: encounter each other.
508: At lower temperatures diffusion is suppressed and
509: the Langmuir-Hinshelwood mechanism is no longer efficient.
510: Saturation of the surface with immobile
511: H atoms
512: might render the Eley-Rideal mechanism more efficient
513: in producing some
514: recombination
515: \citep{Katz1999,Perets2005}.
516: Our results
517: thus indicate that recombination efficiency of hydrogen on
518: amorphous silicates is high
519: in this temperature range, which is relevant to interstellar clouds.
520: Therefore, amorphous silicates
521: seem to be good candidates for interstellar
522: grain components on which hydrogen recombines
523: with high efficiency.
524:
525: \section{Discussion and Summary}
526:
527: \label{sec:discussion}
528:
529: The analysis of the TPD curves from amorphous silicate surfaces
530: shows that the relevant energy barriers on these
531: surfaces are significantly higher
532: than on polycrystalline silicates
533: \cite{Katz1999,Cazaux2004}.
534: A similar trend was observed in amorphous and porous
535: ice surfaces
536: \cite{Williams2007}.
537: These results confirm the effect of surface
538: morphology on the distribution of energy barriers.
539: This effect can be parameterized using a model
540: that provides a quantitative connection between
541: the roughness and the energy barriers
542: \citep{Cuppen2005}.
543: Our results are consistent with this model
544: and can be used to quantitatively constrain
545: its parameters.
546: More specifically,
547: we find a 1.4-1.5 times increase in
548: the energy barriers of the amorphous
549: silicate vs. the polycrystalline silicate surface.
550: This gives rise to shifting and
551: broadening of the temperature window in which
552: H$_2$ formation is efficient, by a similar factor.
553: Scanning electron microscope images show that the morphology of the
554: amorphous silicate samples is very rough
555: (H. Cuppen, private communication).
556: These samples are made of a broad (log-normal like) distribution
557: of spheres and agglomerates, from
558: several nanometers up to a micron size.
559: It is safe to conclude that
560: our present results, together with those on amorphous carbon
561: \citep{Pirronello1999,Katz1999},
562: show that amorphous silicate and carbon grains
563: are efficient catalysts
564: for the formation of molecular hydrogen in diffuse clouds.
565: However, the results indicate that
566: surface roughness
567: is unlikely to extend the window of efficient recombination
568: to temperatures of the order of
569: 30-50K observed in photo-dissociation regions (PDRs).
570: We thus conclude that surface roughness, by itself,
571: does not explain the high abundance of H$_{2}$ in PDRs.
572:
573: In the model used here, it is assumed
574: that H$_{2}$ molecules do not desorb immediately upon formation.
575: Instead, they stay trapped in the adsorption sites or hop
576: between them until thermal desorption takes place.
577: Consequently, one needs to consider
578: mechanisms for the dissipation of the excess energy acquired from
579: the recombination process in order to prevent prompt desorption.
580: For amorphous, porous ice surfaces it was shown,
581: using time of flight (TOF) measurements,
582: that the kinetic energy of the desorbed molecules is small
583: ($\simeq$3 meV), namely
584: the excess energy is absorbed by the surface
585: \cite{Roser2003,Hornekaer2003,Hornekaer2005}.
586: Although we do not have direct measurements of the TOF of the HD molecules
587: desorped from the amorphous silicate surface,
588: the similarity between the TPD curves obtained after
589: HD and H+D irradiations
590: indicates that newly formed HD molecules
591: reside and then desorb from the same adsorption sites as
592: HD molecules irradiated on the surface.
593: In light of the
594: results on both ice and amorphous silicates,
595: it is likely than H$_{2}$ molecules formed on realistic
596: interstellar dust would have low kinetic energy and would probably
597: not occupy excited vibrational or rotational states.
598:
599: In summary, we have analyzed a set of TPD experiments on molecular
600: hydrogen formation and desorption from
601: amorphous silicate surfaces under
602: conditions relevant to interstellar clouds.
603: Fitting the TPD curves by
604: rate equation models, the essential parameters of
605: H$_{2}$ formation on amorphous silicate surfaces were obtained.
606: These parameters include the energy barrier for
607: diffusion of H atoms
608: as well as their barrier for desorption
609: (considered as a lower bound).
610: The distribution of barriers for desorption of H$_2$
611: molecules is also obtained.
612: Interestingly,
613: a single type of adsorption site for hydrogen atoms is identified,
614: vs. two types of sites for molecules.
615: The fraction of the adsorption sites, which belong to each of the
616: two types is also evaluated.
617: The rate equation model provides a unified description
618: of several first and second order processes.
619: It enables us to extrapolate the production rate
620: of H$_{2}$ molecules from laboratory conditions to astrophysical
621: conditions.
622: It thus provides a quantitative evaluation of the efficiency
623: of amorphous silicate surfaces as
624: catalysts in the formation of H$_{2}$ molecules
625: in interstellar clouds.
626: We find that the recombination efficiency
627: strongly depends on the surface temperature.
628: In particular,
629: the amorphous silicate sample studied
630: here exhibits high efficiency within a range of surface temperatures
631: which is relevant to diffuse interstellar clouds.
632: The comparison of the current results with earlier ones
633: on polycrystalline silicate surfaces shows the
634: importance of surface morphology
635: in molecular hydrogen formation.
636: The results are in agreement with theoretical
637: predictions
638: on the effects of surface roughness
639: \citep{Cuppen2005}.
640: The results also indicate
641: that on amorphous surfaces,
642: newly formed H$_{2}$ molecules are thermalized on the surface
643: and do not promptly desorb.
644: Consequently, H$_{2}$ molecules formed
645: on and desorbed from realistic
646: \emph{amorphous} interstellar dust
647: are expected to
648: have very low kinetic energy and
649: would probably not occupy excited
650: vibrational or rotational states.
651:
652: This work was supported by the Adler
653: Foundation for Space Research and the Israel Science Foundation (O.B),
654: by NASA grants NAG5-11438 and NAG5-9093 (G.V), and by the
655: Italian Ministry for University and Scientific Research grant
656: 21043088 (V.P).
657:
658: \begin{thebibliography}
659:
660: \bibitem[Amiaud et al. 2006]{Amiaud2006} Amiaud, L. et al., J. Chem.
661: Phys., 2006, 124, 094702.
662:
663: \bibitem[Biham et al. 2001]{Biham2001} Biham, O. et al., 2001, ApJ,
664: 553, 595.
665:
666: \bibitem[Brucato et al. 2002]{Brucato2002} Brucato, J.R. et al.,
667: 2002, Plan. Space Sci. 50, 829.
668:
669: \bibitem[Cazaux \& Tielens 2004]{Cazaux2004} Cazaux, S., \& Tielens,
670: A. G. G. M. 2004, ApJ, 604, 222.
671:
672: \bibitem[Creighan et al. 2006]{Creighan2006}
673: Creighan, S.C. et al.
674: 2006, J. Chem. Phys., 124, 114701
675:
676: \bibitem[Cuppen \& Herbst 2005]{Cuppen2005}Cuppen, H. M. \& Herbst,
677: E., 2005, MNRAS, 361, 565
678:
679: \bibitem[Duley \& Williams 1984]{Duley1984} Duley, W.W., \& Williams,
680: D.A. 1984, Interstellar Chemistry (Academic Press, London).
681:
682: \bibitem[Dulieu et al 2005]{Dulieu2005}
683: Dulieu, F. et al. 2005, Chem. Phys. Lett., 404, 187
684:
685: \bibitem[Gould \& Salpeter 1963]{Gould1963} Gould, R.J., \& Salpeter,
686: E.E. 1963, ApJ, 138, 393.
687:
688: \bibitem[Hollenbach \& Salpeter 1970]{Hollenbach1970} Hollenbach,
689: D., \& Salpeter, E.E. 1970, J. Chem. Phys., 53, 79.
690:
691: \bibitem[Hornekaer et al. 2003]{Hornekaer2003} Hornekaer, L. et al.,
692: 2003, Science, 302, 1943.
693:
694: \bibitem[Hornekaer et al. 2005]{Hornekaer2005} Hornekaer, L. et al.,
695: 2005, J. Chem. Phys., 122, 124701.
696:
697: \bibitem[Katz et al. 1999]{Katz1999} Katz, N. et al., 1999, ApJ,
698: 522, 305.
699:
700: \bibitem[Manico et al. 2001]{Manico2001} Manico, G. et al., 2001,
701: ApJ, 548, L253.
702:
703: \bibitem[Perets et al. 2005]{Perets2005}
704: Perets, H. B., et al., 2005, ApJ, 627, 850
705:
706: \bibitem[Perry \& Price 2003]{Perry2003}
707: Perry, J.S.A., \& Price, S.D., 2003,
708: Ap\&SS, 285, 769.
709:
710: \bibitem[Pirronello et al. 1997a]{Pirronello1997a} Pirronello, V.
711: et al., 1997a, ApJ, 475, L69.
712:
713: \bibitem[Pirronello et al. 1997b]{Pirronello1997b} Pirronello, V.
714: et al., 1997b, ApJ, 483, L131.
715:
716: \bibitem[Pirronello et al. 1999]{Pirronello1999} Pirronello, V. et
717: al., 1999, A\&A, 344, 681.
718:
719: \bibitem[Roser et al. 2002]{Roser2002} Roser, J.E. et al., 2002,
720: ApJ, 581, 276.
721:
722: \bibitem[Roser et al. 2003]{Roser2003}
723: Roser, J.E. et al. 2003, ApJ, 596, L55.
724:
725: \bibitem[Tielens 2005]{Tielens2005} Tielens, A.G.G.M., The Physics
726: and Chemistry of the Interstellar Medium, 2005, Cambridge University
727: Press.
728:
729: \bibitem[Tine et al. 2003]{Tine2003} Tine et al., 2003,
730: Ap\&SS, 288, 377.
731:
732: \bibitem[Vidali et al. 2004]{Vidali2004} Vidali, G. et al., 2004,
733: J. Geophys. Res.-Planets, 109, E07S14.
734:
735: \bibitem[Williams 1968]{Williams1968} Williams D.A. 1968, ApJ, 151,
736: 935.
737:
738: \bibitem[Williams et al. 2007]{Williams2007} Williams D.A. et al.,
739: 2007, A\&G, 48, 25.
740: \end{thebibliography}
741:
742: \end{document}
743: