[PDF] TEMPERATURE DEPENDENCE OF THE SUBMILLIMETER

View - Download TEMPERATURE DEPENDENCE OF THE SUBMILLIMETER

Previous PDF

Next PDF

TEMPERATURE DEPENDENCE OF THE SUBMILLIMETER ABSORPTION

COEFFICIENT OF AMORPHOUS SILICATE GRAINS

N. Boudet

Centre d"Etudes Spatiales des Rayonnements, 9 Avenue du Colonel Roche, BP 4346,

Toulouse Cedex 4, France; Nathalie.Boudet@cesr.fr

H. Mutschke

Astrophysikalisches Institut und Universita¨ts-Sternwarte, Schillerga¨?chen 3, D-07745 Jena, Germany

C. Nayral

Centre d"Etudes Spatiales des Rayonnements, 9 Avenue du Colonel Roche, BP 4346, Toulouse Cedex 4, France

C. Ja¨ger

Astrophysikalisches Institut und Universita¨ts-Sternwarte, Schillerga¨?chen 3, D-07745 Jena, Germany

J.-P. Bernard

Centre d"Etudes Spatiales des Rayonnements, 9 Avenue du Colonel Roche, BP 4346, Toulouse Cedex 4, FranceT. Henning

Max-Planck-Institut fu¨r Astronomie, Ko¨nigstuhl 17, D-69117 Heidelberg, Germany and

C. Meny

Centre d"Etudes Spatiales des Rayonnements, 9 Avenue du Colonel Roche, BP 4346, Toulouse Cedex 4, France

Receivved 2004 July 19; accepted 2005 June 23

ABSTRACT

We have measured mass absorption coefficients of amorphous silicate materials for wavelengths between 100?m

and 2 mm (5-100 cm ?1 ) and at temperatures between 300 and 10 K. For both interstellar analog MgSiO 3 and simple silica SiO 2

, we find evidence for a strong temperature and frequency dependence. We define two distinct wavelength

regimes, 500?m-1 mm and 100-250?m, for which the absorption coefficient presents different trends with fre-

quency. To evaluate this frequency dependence, we fit our absorption coefficient using two power laws with spectral

index?that varies with temperature. We do not find a significant variation of?with temperature between 100 and

for the observed temperature and frequency dependence and that OH groups could be at the origin of the submil-

silicates,TLScouldalsobeproducedbytheMg +2 ions, which act as network modifiers, similar to how they act with OH groups.

Subject headinggs:dust, extinction — infrared: ISM — ISM: molecules — methods: laboratory —

molecular processes

1. INTRODUCTION

The interstellar medium (ISM) represents 4% of the mass of the Galaxy, and its dust component (composed of submicron- sized solid grains) accounts for roughly 1% of the ISM. Despite with which the dust scatters, absorbs, and reradiates starlight en- sures that it plays a key role in the Galactic energy balance. Dust controls the star formation process and is an important player in the dynamics of protostellar disks (Yorke & Henning 1994). The a dust component composed of grains large enough to be at ther- mal equilibriumwiththeincomingradiationfield, at temperatures ofafewtensof K(Mathisetal.1977;Draine&Lee1984;Desert et al. 1990). An important component of these grains is silicates. They explain in particular the absorption bands at 9.7 and 18?m and are predominantly in the amorphous state. Assuming an op-

ticallythininterstellarmedium(whichisvalidinthesubmillimeterwavelength range) and a single grain temperatureTalong the line

of radiusain the Rayleigh approximation is given by I?

¼?a

2 Q abs N grain B (T);ð1Þ whereI is the spectral intensity (in MJy sr ?1 ),B (T)isthe

Planckfunction,Q

abs istheabsorptionefficiency,andN grain isthe column density of grains. Absorption efficiency at long wave- lengths depends on various parameters: size, shape, chemical see Yorke & Henning 1994). The knowledge of the absorption efficiency and its variation with temperature is of great impor- tance to infer key information from astronomical observations, such as dust mass and temperature, in various sites of the ISM and to estimate reliably the contribution of the dust emission to the observed cosmic microwave background. Such information

272The Astrophysical Journal, 633:272-281, 2005 November 1

#2005. The American Astronomical Society. All rights reserved. Printed in U.S.A. is crucial for the scientific exploitation of data from the Atacama sions. For spherical grains, the absorption efficiencyQ abs can itself be related to the mass absorption coefficient?(?,a)(incm 2 g ?1 through ?(?;a)¼ 3 4?Q abs a:ð2Þ Thus, laboratory measurements of the mass absorption coef- ficient can be used to characterize the absorption efficiency. However, such data for astrophysically relevant materials at far- infrared and millimeter wavelengths and at low temperatures are hardly available. Moreover, direct spectroscopic astronomical observations of such extended thermal dust emission have al- ways been limited by the strong atmospheric absorption, requir- ing balloon-borne or satellite experiments, or measurements of the brightest emission through a few atmospheric windows. In this context of a relative lack of reliable and accurate laboratory and astrophysical data, our view of submillimeter dust emission is still mainly based on theoretical assumptions. It is known that at low temperatures the submillimeter ab- sorption in crystalline dielectric materials results from the long- wavelengthwingofsomefundamental vibrational bands,leading Such a temperature-independent quadratic dependence is also expected in this wavelength range for the free-carrier absorption of light in metallic materials, and for the Debye phonon absorp- tion in three-dimensional amorphous solids. It is thus widely ac- cepted that the thermal dust submillimeter emission spectrum law dependence of the wavelength, assuming a constant and temperature-independent exponent?oftheorder of2(alsocalled the emissivity spectral index), following the expression I 0 k k 0 B k;TðÞ:ð3Þ Recently the French balloon-borne experiment PRONAOS

´triques)mea-

sured the low extended surface brightness of the ISM in four broadband filters covering the wavelength range from 200 to

1100?m (Lamarre et al. 1994; Serra et al. 2001). This enabled

simultaneous measurements of both the temperature and the emissivity spectral index. Over various sites of the ISM, a syn- thesis of PRONAOS observationsrevealed an anticorrelation between the dust grain mean temperatureTand the emissivity spectral index?, with?-values down to 1 at 80 K and up to 2-

2.5 at 10 K (Dupac et al. 2003). It has been argued that if the

origin of such anticorrelation remains unclear, it should not be attributed to distribution of grains with different temperatures along the line of sight; it is more likely due to a temperature dependence of the intrinsic optical properties of the materials that constitute the grains. In addition, the all-sky data of the FIRAS (Far Infrared Absolute Spectrophotometer) instrument on board theCOBE(Cosmic Background Explorer) satellite have emission with respect to a single graybody law. This has been interpreted by several authors as due to the presence of very cold inadditiontothe warmercomponent responsible for the emission maximum near 100?m. However, the millimeter excess appears to be extremely well correlated with the FIR dust emission, even

inviewofmorerecentobservationswithangularresolutionbetterthan FIRAS. This is indicative of a situation in which the excess

could also be produced by unidentified processes intrinsic to the grain component radiating at thermal equilibrium, without re- quiring an additional colder component. The optical properties could strongly change between room temperature and a few tens of K, especially at far-infrared and submillimeter wavelengths (for a review, see Henning & Mutschke 1997). The experimental temperature dependence of the mass absorption coefficient is an important diagnostic of the very few submillimeter spectroscopic studies at variable temper- ature in the range 10-100 K are available on materials of astro- physical interest. Agladze et al. (1996), for example, measured the temperature dependence of the absorption but in a restricted temperature and wavelength range (1.2-30 K and 0.7-2.9 mm) for crystalline enstatite (MgSiO 3 ), forsterite (Mg 2 SiO 4 ), and their amorphous submicron-sized grain precursors. They could observe, in some samples, an anticorrelation between spectral index and temperature in this millimeter and very low temper- ature range. They identified thephysics underlying this behavior as resonant light absorption in a distribution of two-level sys- tems. Mennella et al. (1998) measured the temperature depen- grains (amorphous carbon grains, crystalline silicates, and amor- phous fayalite FeSiO 4 ). They deduced over the whole 100?m-

2 mm range a wavelength-independent spectral index for the

absorption coefficient and found a significant temperature depen- dence of the spectral index, more or less pronounced, depending onmaterialcomposition.For example,amorphous fayaliteshows an increase of its spectral index from 1.35 up to 2.04 as the tem- perature goes down. They attribute this behavior to two-phonon difference processes. Previously, in highly absorbing silica-based glasses, Bo¨sch (1978) observed a strong temperature dependence (above 500?m), characterized by a spectral index of the order of 1.6 at 300 K, which drops down to around 3 at 10 K. Bo¨sch interpreted such temperature-dependent behavior in terms of the TLS ‘‘tunnelling model,"" first formulated by Phillips (1972) and, three different processes: at low temperatures, a resonant absorp- tionbetweenthemillimeterwaveandadistributionof asymmetric two-level systems (as observed by Agladze); and at higher tem- (hopping and phonon-assisted tunneling). To get high spectral index values up to 3, Bo¨sch superposed such thermally activated processes to vibrational absorption in the charge-disordered dis- tribution of the amorphous structure, as modeled by Schlo¨mann (1964). on more amorphous materials than did Mennella et al. (1998), whostudiedonlyanamorphous iron-basedsilicate,andina more extended temperature and wavelength range than did Agladze et al. (1996). This is also of prime interest for the data analysis of the PRONAOS experiment and the futureHerschelandPlanck missions. Furthermore, even if the variations of absorption with mains unclear. More experimental studies are required in order to progress in the understanding of the physical and chemical grain properties responsible for the temperature-dependent optical be- havior observed in the ISM. Our approach is to try to identify properties of the solid state (especially lattice disorder or discrete defects) and the underlying physical mechanisms that could be involved in these phenomena.SUBMILLIMETER ABSORPTION IN SILICATE GRAINS273 to 10 K) of the mass absorption coefficient between 100?mand

2mm(5-100cm

?1 ) for different amorphous silicate materials. We studied three types of amorphous silica materials (SiO 2 )and two kinds of amorphous enstatite (MgSiO 3 ) with different struc- ture, morphology and chemical characteristics. Inx2, the sample choice is explained, and a description of the material structures is given. Section 3 is devoted to a description of sample preparation for spectroscopy and setup used for low-temperature measure- ments. Inx4, we detail our data analysis of transmission spectra, and the experimental results are presented and discussed inx5. Conclusions and perspectives of this work are given inx6.

2. SAMPLE STRATEGY

It is clear that interstellar dust along a line of sight consists of a wide range of complex aggregates, silicates, and carbonaceous grains. Thus, the observed submillimeter behavior should result from common properties of amorphous grains, mostly dominated by silicate species. A better understanding of optical properties does not necessarily require working on realistic interstellar ana- erties.Thatiswhywehavechosentocombine astudyofthemore 2 ,andanamorphous MgSiO 3 , which is a classical astronomical silicate analog. These two families of amorphous silicate samples have been sition and grain morphology. Thus, we compare silica spherical grains (diameters 1.5 and 0.5?m), silica-fumed agglomerates (7 nm spherical particles linked together to form micrometer- sized chains), and MgSiO 3 grains ground in a mill (micrometer- sizedgrainsofundefinedshape).The sizeofall the grainsstudied is much smaller than the wavelengths; no size effect is expected (Rayleighlimit).

Bulk silica consists of SiO

4 tetrahedra sharing their oxygen atoms. Each of the four oxygen atoms is covalently bonded to at least one silicon atom to form either a siloxane ( ?Si?O?Si) or a silanol ( ?Si?O?H) functionality. Amorphous silica is closely related to the cristobalite structure, but the local order is believed to be limited to crystalline domains of up to 2 nm in diameter, which have completely random orientations. Depend- ing on the synthesis processes, silanols are obviously present as such as SiOH groups trapped into the structure (Serp et al. 2002; Iler 1979; Waddell & Evans 1997). We have to make clear the distinction between water content and silanol content; both of them vary quite a lot from sample to sample, depending on the synthesis procedure (dry or wet processes). It is well known (Serp et al. 2002) that several layers of physisorbed water (mo- lecular water weakly bonded to the surface) can cover the hydrophilic monolayer formed by hydroxyl groups (Si ?OH), terminating the bulk at the silica surface. According to the lit- erature (Serp et al. 2002; Ek et al. 2001), heating at 200 C can easily remove physisorbed water (a small difference in the de- sorption temperature values can be found because of different heating rates, gas flows, or pressures, but a general behavior is yet to be described). A thermal treatment above this temperature leads to the condensation of silanol groups (formation of silox- anefunctionality andlossofmolecular water).Differentkindsof surface silanol have to be distinguished: geminal hydroxyl hydroxyl groups (only one OH group bonded to one silicon they can interact weakly (vicinal interaction) by the hydrogen bonding. Dehydroxylation processes occur slowly from 200

Cup to 1200

C; the first step is the condensation of these close OH groups to leave at the surface only geminal and isolated groups.

Above800

(preventing rehydration) and isolated hydroxyl groups, which can befullyremovedattemperaturesofabout1200

C(bycalcination).

Thus, the disorder in amorphous SiO

2 is related to a lack of peri- odicity in the (SiO 2 ) network and, of course, to the nature and content of SiOH defects. It is important to note that in contrast to erties for different amorphous silica samples just because their disorders are not the same (especially the Si ?OH content). ent amorphous magnesium silicates with pyroxene stoichiometry (MgSiO 3 ). The MgSiO 3 glass was produced by melting the ap- propriate precursors at 1913 K and subsequent fast quenching of thehotliquidtoroomtemperature(Ja¨geretal.1994;Dorschner et al. 1995). Thesecondamorphousmagnesiumsilicateof thesamecompo- chemical polymerization of silicates in a liquid phase at low tem- peratures. Metal organic compounds such as tetraethoxysiloxane and magnesium methylate served as precursors. After the evapo- ration of the solvents methanol and water from the produced gels, the remaining magnesium silicate powder was heated to 870 K in order to achieve a densification of the silicate framework and a removal of porosity accompanied by a condensation of OH groups. The complete synthesis is described in a previous paper (Ja¨ger et al. 2003). Crystalline pyroxenesareinosilicates containingchainsofsil- iconoxygen tetrahedra (SOT) that share two oxygens. The (Mg) cations are located between the adjoining chains and possess well-defined coordination spheres. In contrast to the crystalline structure, the amorphous material is composed of a disordered network of SOTs connected by bridging oxygens. The incorpo- ration of metal oxides destroys part of the oxygen bridges and forms nonbridging oxygen accompanied by a widening of the structure. In the case of pyroxene glass it is assumed that, in addition to the disordered network of SOTs, further structure- forming arrangements such as chains, rings, sheets, and isolated tetrahedra occur (Mysen et al. 1982). Consequently, as for the silica, the amorphous structure of the magnesium silicates is not a well-defined state. The relative proportions and the distribution of the isolated SOTs, chains, rings, and sheets can differ in networks of the same average stoichiometry. In the sol-gel-produced magnesium silicates, the Mg +2 ions can also act as a network former (Ja¨ger et al. 2003).In addition, OH groups remaining from the sol-gel synthesis act as network modifiers (Iler 1979) and can strongly influence the silicate"s properties. For instance, it has been found that crystallization of the sili- contains isolated OH groups. This is due to the reduced viscosity, which strongly lowers the activation energy for crystallization, e.g., for the sol-gel silicates (Scholze 1988; Ja¨ger et al. 2003).

The formation of Si

?OH bonds in cosmic magnesium silicate particles has to be taken into account, since H 2

O, as the most abun-

dant oxygen-bearing molecule, plays a very important role in the condensation of circumstellar silicates (Gail & Sedlmayr1998).

3. EXPERIMENTAL METHODS

3.1.Materials

To perform transmission measurement on samples, we worked withtwokinds of matrices: KBr forspectroscopy(Merck)in theBOUDET ET AL.274Vol. 633 mid-infrared wavelength range, and polyethylene for spectros- copy (Merck) in the far-infrared to submillimeter wavelength range. We used commercially available amorphous silica pow- der consisting of monosized spherical particles with 1.5?mdi- ameter (Monospheres 1500, Merck), silica spheres with 0.5?m diameter from Lancaster, and fumed silica from Aldrich. Fumed silica is composed of 0.007?m spheres, which are fused into high surface area. We carried out measurements on two differ- ent enstatite samples of MgSiO 3 : MgSiO 3 sol-gel prepared by a sol-gel process (details of the synthesis are given in Ja¨ger et al.

2003) and MgSiO

3 glass prepared by high-temperature melting (Dorschner et al. 1995). The density and dielectric constants of these materials are given in Table 1.

3.2.Sample Preparation

We carefully mixed our material powder with matrix powder, using ethanol and ground for several minutes in order to obtain a homogenous mixture.Weputthismixtureinanovenat70 Cfor

30 minutes to evaporate the ethanol solvent. After drying, we

pressed our mixture, applying 10 tons of pressure for 3 minutes tomake thepellet.The finalmaterialmass embeddedinthesam- pleis determined by weighting thepellet and takinginto account the dust-to-matrix mass ratio (weighting is performed by means of a microbalance with a sensitivity of 1?g of the starting mix-quotesdbs_dbs44.pdfusesText_44

[PDF] pablo picasso autoportrait 1938

[PDF] chaleur interne de la terre

[PDF] base de frenet acceleration

[PDF] autoportrait dali

[PDF] pablo picasso autoportrait 1972 wikipedia

[PDF] autoportrait au chapeau

[PDF] don quichotte

[PDF] triplement paces temoignage

[PDF] numerus clausus descartes 2017

[PDF] moodle paris descartes 2016-2017

[PDF] paces descartes difficile

[PDF] grafcet niveau 3

[PDF] grafcet verin simple effet

[PDF] grafcet fonctionnel

[PDF] grafcet technologique