1. INTRODUCTION: A BRIEF HISTORY FOR THE STUDIES OF INTERSTELLAR DUST

In 1930 - exactly 75 years ago, the existence of solid dust particles in interstellar space was by the first time firmly established, based on the discovery of color excesses (Trumpler 1930). But the history of the interstellar dust-related studies is a much longer and complex subject, and can be dated back to the late 18th century when Herschel (1785) described the dark markings and patches in the sky as "holes in the heavens". Below is a summary of the highlights of this history. For a more detailed record of the historical development of dust astronomy, I refer the interested readers to Aiello & Cecchi-Pestellini (2000), Dorschner (2003), Li & Greenberg (2003), and Verschuur (2003).

Early Chronology: From "Holes in the Heavens" to the Firm Establishment of Interstellar Dust

• As early as 1785, Sir William Herschel noticed that the sky looks patchy with stars unevenly distributed and some regions are particularly "devoid" of stars. He described these dark regions ("star voids") as "holes in the heavens".

• At the beginning of the 20th century, astronomers started to recognize that the "starless holes" were real physical structures in front of the stars, containing dark obscuring masses of matter able to absorb starlight (Clerke 1903; Barnard 1919), largely thanks to the new technology of photography which made the photographic survey of the dark markings possible. Sir Harold Spencer Jones (1914) also attributed the dark lanes seen in photographs of edge-on spiral galaxies to obscuring matter. Whether the dark lanes in the Milky Way is caused by obscuring material was one of the points of contention in the Curtis-Shapley debate (Shapley & Curtis 1921).

• Wilhelm Struve (1847) noticed that the apparent number of stars per unit volume of space declines in all directions receding from the Sun. He attributed this effect to interstellar absorption. ¹ From his analysis of star counts he deduced an visual extinction of ~ 1 mag kpc^-1. Many years later, Jacobus Kapteyn (1904) estimated the interstellar absorption to be ~ 1.6 mag kpc^-1, in order for the observed distribution of stars in space to be consistent with his assumption of a constant stellar density. This value was amazingly close to the current estimates of ~ 1.8 mag kpc^-1. Max Wolf (1904) demonstrated the existence of discrete clouds of interstellar matter by comparing the star counts for regions containing obscuring matter with those for neighbouring undimmed regions.

• In 1912, Vesto Slipher discovered reflection nebulae from an analysis of the spectrum of the nebulosity in the Pleiades cluster which he found was identical to that of the illuminating stars. It was later recognized that the nebulosity was created by the scattering of light from an intrinsically luminous star by the dust particles in the surrounding interstellar medium (ISM).

• Henry N. Russell (1922) argued that dark clouds accounted for the obscuration and this obscuring matter had to be in the form of millimeter-sized fine dust. Anton Pannekoek (1920) recognized that the obscuration can not be caused by the Rayleigh scattering of gas, otherwise one would require unrealistically high masses for the dark nebulae. He also noticed that, as suggested by Willem de Sitter, the cloud mass problem can be vanished if the extinction is due to dust grains with a size comparable to the wavelength of visible light.

• In 1922, Mary L. Heger observed two broad absorption features at 5780Å and 5797Å, conspicuously broader than atomic interstellar absorption lines. The interstellar nature of these absorption features was established 12 years later by Paul W. Merrill (1934). These mysterious lines - known as the diffuse interstellar bands (DIBs), still remain unidentified.

• In 1930, a real breakthrough was made by Robert J. Trumpler who provided the first unambiguous evidence for interstellar absorption and reddening which led to the general establishment of the existence of interstellar dust. Trumpler (1930) based this on a comparison between the photometric distances and geometrical distances of 100 open clusters. ² If there was no interstellar absorption, the two distances should be in agreement. However, Trumpler (1930) found that the photometric distances are systematically larger than the geometrical distances, indicating that the premise of a transparent ISM was incorrect. ³ Using this direct and compelling method he was able to find both absorption and selective absorption or color excess with increasing distance. ⁴ Trumpler (1930) also concluded that the observed color excess could only be accounted for by "fine cosmic dust".

• In 1932, Jan H. Oort demonstrated that the space between the stars must contain a considerable amount of matter. He derived an upper limit ("Oort limit") on the total mass of the matter (including both stars and interstellar matter) in the solar neighbourhood from an analysis of the motions of K giants perpendicular to the plane of the Galaxy (the z-direction). An upper limit of ~ 1.0 × 10^-23 g cm^-3 on the total mass density was obtained from measuring the gravitational acceleration in the z-direction. The Oort limit has important implications: (1) there has to be more material in the galactic plane than could be seen in stars since the mass density of known stars is only ~ 4.0 × 10^-24 g cm^-3; and (2) the upper limit of ~ 6.0 × 10^-24 g cm^-3 on the mass density of the interstellar matter in the solar neighbourhood places severe restrictions on the source of the obscuration: what kind of material distributed with this density with what mass absorption coefficient could give rise to the observed visual extinction of about 1 mag kpc^-1? apparently, only with small dust grains could so much extinction by so little mass (and the ^-1 wavelength dependence; see below) be explained.

Interstellar Absorption and Scattering

• In 1936, Rudnick by the first time measured the wavelength dependence of extinction in the wavelength range 4000-6300Å based on differential spectrophotometric observations of reddened and unreddened stars of the same spectral type. Rudnick (1936) found that the measured extinction curve was inconsistent with Rayleigh scattering (which has a ^-4 wavelength dependence). This so-called "pair-match" method remains the most common way of deriving an interstellar extinction curve.

• By the end of the 1930s, a ^-1 extinction law in the wavelength range 1-3 µm^-1 had been well established (Hall 1937; Greenstein 1938; Stebbins, Huffer, & Whitford 1939), thanks to the advent of the photoelectric photometry, excluding free electrons, atoms, molecules, and solid grains much larger or much smaller than the wavelength of visible light, leaving solids with a size comparable to the wavelength as the sole possibility.

• In 1936, Struve & Elvey demonstrated the scattering of general starlight by interstellar clouds based on a series of observations of the dark cloud Barnard 15, the core of which is appreciably darker than the rim, although the latter is about opaque as the former. They attributed the increased brightness of the outer region to interstellar scattering.

• In 1941, Henyey & Greenstein confirmed the existence of diffuse interstellar radiation (which was originally detected by van Rhijn [1921]) in the photographic wavelength region. They interpreted the observed intensity of diffuse light as scattered stellar radiation by interstellar grains which are strongly forward scattering and have a high albedo (higher than ~ 0.3).

• In 1943, with the advent of the six-colour photometry (at 3530Å < < 10300Å) Stebbins & Whitford found that the extinction curve exhibits curvature at the near infrared (IR; 1.03 µm) and ultraviolet (UV; 0.35 µm) regions, deviating from the simple ^-1 law.

• In 1953, Morgan, Harris, & Johnson estimated the ratio of total visual extinction to color excess to be A_V / E(B - V) 3.0 ± 0.2. This was supported by a more detailed study carried out by Whitford (1958), who argued that there appeared to be a "very close approach to uniformity of the reddening law" in most directions. A uniform extinction curve with a constant A_V / E(B - V) was welcomed by the astronomical community - at that early stage, interstellar dust was mainly regarded as an annoying mere extinguisher of starlight which prevented an accurate measurement of distances to stars. The proposal of "a uniform extinction curve with a constant A_V / E(B - V)" made it easier to correct photometric distances for the effects of absorption (also because the determination of the color excess E(B - V) for early-type stars was relatively straightforward).

• In 1955, based on the UBV photometry of early O stars in a region in Cygnus, Johnson & Morgan noted that there may exist regional variations in the interstellar extinction curve. The nonuniformity nature of the interstellar extinction curve was later confirmed in Cygnus, Orion, Perseus, Cepheus and NGC2244 by Johnson & Borgman (1963), Nandy (1964), and Johnson (1965). Those authors also found a wide variety of A_V / E(B - V) values (ranging from ~ 3.0 to ~ 7.4) in different regions. Wampler (1961) found a systematic variation with galactic longitude of E(U - B) / E(B - V), the ratio of slopes in the blue to those in the visible region.

• In the 1960s and early 1970s, the extension of the extinction curve toward the middle and far UV (^-1 > 3µm^-1) was made possible by rocket and satellite observations, including the rocket-based photoelectric photometry at = 2600Å and 2200Å (Boggess & Borgman 1964); the Aerobee rocket spectrophotometry at 1200Å < < 3000Å (Stecher 1965); the Orbiting Astronomical Satellite (OAO-2) spectrophotometry at 1100Å < < 600Å (Bless & Savage 1972); and the Copernicus Satellite spectrophotometry at 1000Å < < 1200Å (York et al. 1973). By 1973, the interstellar extinction curve had been determined over the whole wavelength range from 0.2 µm^-1 to 10 µm^-1.

• In 1965, the 2175Å extinction bump was detected by Stecher (1965). Shortly after its detection, it was attributed to graphite (Stecher & Donn 1965). ⁵ It was later found that the strength and width of this bump vary with environment while its peak position is quite invariant.

• Cardelli, Clayton, & Mathis (1989) found that the optical/UV extinction curve in the wavelength range of 0.125 3.5µm which shows considerable regional variations can be approximated by an analytical formula involving only one free parameter: the total-to-selective extinction ratio R_V A_V / E(B - V), whereas the near-IR extinction curve (0.9 µm 3.5µm) can be fitted reasonably well by a power law A() ~ ^-1.7, showing little environmental variations. ⁶

From Metallic Grains to Dirty Ices:
Meteoritic Origin or Interstellar Condensation?

• In the 1930s, small metallic particles were proposed to be responsible for the interstellar extinction, partly because meteoritic particles (predominantly metallic) and interstellar grains were then thought to have the same origin. Reasonably good fits to the ^-1 extinction law were obtained in terms of small metallic grains with a dominant size of ~ 0.05µm (Schalén 1936) or a power-law size distribution dn(a) / da ~ a^-3.6 in the size range 80Å < a < 1 cm (Greenstein 1938).

• In 1935, based on the correlation between gas concentration and extinction, Bertil Lindblad suggested that interstellar grains were formed by condensation from the interstellar gas through random accretion of gas atoms, as speculated by Sir Arthur Eddington (1926) that it was so cold in space that virtually all gaseous atoms and ions which hit a solid particle would freeze down upon it. ⁷ However, it was found later that in typical interstellar conditions, the Lindblad condensation theory would result in a complete disappearance of all condensable gases and the grains would grow to sizes (~ 10 µm) well beyond those which could account for the interstellar extinction.

• In 1946, by introducing a grain destruction process caused by grain-grain collisions as a consequence of interstellar cloud encounters, Jan H. Oort and Hendrik C. van de Hulst further developed the interstellar condensation theory and led to the "dirty ice" model consisting of saturated molecules such as H₂O, CH₄, and NH₃ with an equilibrium size distribution which could be roughly approximated by a functional form dn(a) / da ~ exp[ -5(a / 0.5 µm)³] and an average size of ~ 0.15 µm. What might be the condensation nulcei was unclear at that time.

• In 1946, van de Hulst by the first time made realistic estimations of 10-20 K for grain temperatures. Before that, it was long thought that they had a black-body temperature of ~ 3.2 K (Eddington 1926). Van de Hulst (1946) noted that interstellar grains are much warmer than a 3.2 K black-body because they do not radiate effectively at long wavelengths.

Interstellar Polarization

• In 1949, Hall and Hiltner independently discovered the general interstellar linear polarization by incident - their original objective was to look for intrinsic stellar polarization from eclipsing binaries. The interstellar origin of this polarization was indicated by the correlation of the degree of polarization with reddening and the fact that the direction of polarization is generally parallel to the galactic plane. The interstellar polarization was attributed to the differential extinction of starlight by nonspherical grains aligned to a small degree with respect to the galactic plane.

• In 1951, Davis & Greenstein suggested that interstellar grains could be aligned with respect to the interstellar magnetic field by the paramagnetic relaxation mechanism.

• The variation of interstellar polarization with wavelength was first revealed by Behr (1959) and Gehrels (1960). It was later shown that the wavelength dependence of polarization is well approximated by an empirical formula, often known as the Serkowski law (Serkowski 1973; Coyne, Gehrels, & Serkowski 1974; Wilking et al. 1980). ⁸ But the near-IR (1.64 µm < < 5µm) polarization is better approximated by a power law P() ^-, with 1.8 ± 0.2, independent of _max (Martin & Whittet 1990, Martin et al. 1992).

• In 1972, the interstellar circular polarization which arises from the interstellar birefringence (Martin 1972) as originally predicted by van de Hulst (1957), was first detected along the lines of sight to the Crab Nebula by Martin, Illing, & Angel (1972) and to six early-type stars by Kemp & Wolstencroft (1972).

From Dirty Ices to Graphite: Interstellar Condensation or Stellar Origin?

• In early 1950s - soon after the discovery of interstellar polarization, the validity of the ice model seemed doubtful since ice grains are an inefficient polarizer, and therefore it would be difficult for them to explain the observed rather high degree of polarization relative to extinction (van de Hulst 1950; Spitzer & Tukey 1950; Cayrel & Schatzman 1954).

• In 1954, Cayrel & Schatzman suggested that graphite grains, comprising a small component of the total mass of interstellar dust, could account for the observed polarization-to-extinction ratio because of their strong optical anisotropy.

• In 1962, Hoyle & Wickramasinghe proposed that graphite grains of sizes a few times 0.01µm could condense in the atmospheres of cool N-type carbon stars, and these grains will subsequently be driven out of the stellar atmospheres and injected into interstellar space by the stellar radiation pressure. Hoyle & Wickramasinghe (1962) argued that ~ 10⁴ N-type stars in the Galaxy may be sufficient to produce the required grain density to account for the observed interstellar extinction. They also showed that the extinction predicted from small graphite grains is in remarkable agreement with the observed reddening law (which was then limited to ^-1 < 3 µm^-1). It is interesting to note that the condensation of graphite grains in cool carbon stars was suggested many years earlier by O'Keefe in 1939, while as early as 1933 Wildt had already found that solid grains of carbon, Al₂O₃, CaO, carbides (SiC, TiC, ZrC), and nitrides (TiN, ZrN) might form in N-type stars.

• In 1966, in view of the fact that the albedo of pure graphite grains appear to be too low to be consistent with the observations, Wickramasinghe, Dharmawardhana, & Wyld proposed that interstellar dust consists of graphite cores and ice mantles. Wickramasinghe (1965) argued that graphite grains ejected from stars tend to grow ice mantles in interstellar clouds. Wickramasinghe et al. (1966) showed that graphite grains of radii ~ 0.05-0.07 µm coated by an ice mantle up to twice their radii could satisfy the observed interstellar extinction and albedo.

• In 1968, Wickramasinghe & Nandy argued that solid molecular hydrogen mantles may be accreted by interstellar grains in dense interstellar clouds. Wickramasinghe & Krishna Swamy (1969) showed that graphite core-solid H₂ mantle grains with core radii ~ 0.04-0.06 µm and mantle radii ~ 0.15-0.25 µm are consistent with the observed interstellar extinction in the wavelength range of 0.11 µm < < 2 µm and the albedo and phase function derived from the diffuse Galactic light.

Interstellar Silicate Dust of Stellar Origin

• In 1963, Kamijo suggested that nanometer-sized SiO₂ grains could condense in the atmospheres of cool M-type stars. After blown out of the stellar atmospheres and injected into interstellar space, they could serve as condensation nuclei for the formation of "dirty ices".

• In 1968, Wickramasinghe & Krishna Swamy considered quartz grains covered with dirty ice mantles and found that their match to the observed extinction curve was unsatisfactory.

• In 1969, Gilman found that grains around oxygen-rich cool giants are mainly silicates such as Al₂SiO₅ and Mg₂SiO₄. Silicates were first detected in emission in M stars (Woolf & Ney 1969; Knacke et al. 1969a).

Interstellar Iron, SiC, and Diamond Grains of Stellar Origin?

• In 1965, Cernuschi, Marsicano, & Kimel argued that iron grains could condense out of the expanding supernova explosion ejecta. Schalén (1965) explicitly modeled the interstellar extinction curve in the wavelength range of 0.5 µm^-1 < ^-1 < 4.5 µm^-1 using iron grains of radii ~ 0.01 µm. Hoyle & Wickramasinghe (1970) also argued that a significant fraction of the mass of the heavy elements produced in supernova explosion could condense into solid particles during the expansion phase following explosion. They further suggested that supernovae may constitute a major source of silicate, iron, and graphite grains in the ISM. ⁹

• In 1969, Friedemann showed that silicon carbide grains could condense in the atmospheres of carbon stars and then leave the star and become an interstellar dust component, although they comprise only a minor fraction of the total interstellar dust mass. ¹⁰

• In 1969, Saslaw & Gaustad suggested that carbon may condense in cool stellar atmospheres in the form of diamond grains and are subsequently injected into interstellar space. ¹¹ Presolar nanodiamonds were first detected in primitive carbonaceous meteorites based on their isotopic anomalies (Lewis et al. 1987; see Section 5.4 of Li & Draine 2004a and Jones & d'Hendecourt 2004 for more information regarding interstellar nanodiamonds).

Grain Mixtures with Multi-modal Size Distributions

• The extension of the wavelength base for the interstellar extinction observations into the far-UV and IR provide a strong stimulus for the development of dust models. The fact that the extinction continues to increase in the far UV (e.g., see York et al. 1973) implies that no single grain type with either a single size or a continuous size distribution could account for the observed optical to far-UV interstellar extinction (Greenberg 1973). This led to the abandonment of any one-component grain models and stimulated the emergence of various kinds of models consisting of multiple dust constituents, including silicate, SiC, iron, iron oxide, graphite, dirty ice, solid H₂, etc. ¹² By early 1970s, the two highly refractory components - silicates and graphite have been considered in most dust models, supported by the detection of the conspicuous bump in the interstellar extinction curve at 2175Å and the prominent emission feature at 10 µm of oxygen-rich stars and by the belief that graphite and silicate grains can be produced in stellar atmospheres and expelled into the ISM.

  • In 1969, Hoyle & Wickramasinghe modeled the interstellar extinction in terms of a mixture of silicate grains of radii ~ 0.07 µm and graphite grains of radii ~ 0.065 µm.

  • Wickramasinghe (1970a) found that the interstellar extinction curve in the wavelength range 0.3 < ^-1 < 9 µm^-1 could be reproduced by a mixture of graphite grains with a size distribution of dn(a) / da ~ exp[-0.5{(a - 0.06) / 0.02}²] for 0.03 µm < a < 0.13 µm and silicate grains of radii ~ 0.07 µm. He also found that silicate grains of radii ~ 0.03 µm with an ice mantle of radii ~ 0.14 µm together with the same graphite population could fit the observed extinction curve equally well. ¹³ By modeling the albedoes and phase functions derived from the diffuse Galactic light, Wickramasinghe (1970b) concluded that the graphite-silicate mixture was preferred over the graphite-(ice-coated) silicate mixture.

  • Wickramasinghe & Nandy (1970) found that a mixture of silicate, graphite, and iron grains also achieved a rough fair fit to the interstellar extinction curve at ^-1 < 8 µm^-1.

  • Huffman & Stapp (1971) found that enstatite grains plus 12% small (~ 100Å) iron oxide grains also provided a fairly good fit to the extinction curve up to ^-1 < 8 µm^-1.

  • Gilra (1971) performed extinction calculations for a mixture of graphite, silicate, and SiC and provided close fits to the observed extinction curves. But his model heavily relied on SiC: the required mass of the SiC component was ~ 4 times of that of graphite.

  • Greenberg & Stoeckly (1970) found that ice-coated cylindrical silicate grains together with a population of small bare silicate grains could reproduce the extinction curve from the IR to the UV and the wavelength dependence of polarization.

  • In 1974, Greenberg & Hong suggested that interstellar grains consist of submicron-sized silicate cores surrounded by mantles of heterogeneous molecular and free-radical mixture of O, C, N and H ("modified dirty ices"), and a minor component of very small bare grains of sizes < 100Å whose precise composition was uncertain.

• In a study of the scattering properties of interstellar dust (albedo and phase function) determined from the OAO-2 observations at 1500Å < < 4250Å of the diffuse Galactic light (Witt & Lillie 1973), Witt (1973) first explicitly suggested a bi-modal size distribution for interstellar grains: large grains with radii > 2500Å would provide extinction in the visible region including scattering which is strongly forward directed, and small particles with radii < 250Å would dominate the UV region and contribute nearly isotopic scattering.

The Infrared Era: Ices, Silicates, PAHs and Aliphatic Hydrocarbons

• In the 1960s, the first attempt to search for the 3.1 µm feature of H₂O ice in the diffuse ISM was unsuccessful (Danielson, Woolf, & Gaustad 1965; Knacke, Cudaback, & Gaustad 1969b), although it had long been considered to be a possible constituent of interstellar grains. This was the strongest objection against the dirty-ice model of Oort & van de Hulst (1946).

• In 1973, the 3.1µm H₂O ice feature was finally detected (Gillett & Forrest 1973). But it was recognized that water ice is present only in dense regions (usually with A_V > 3 mag).

• By early 1970s, silicates had been detected in the ISM, first in emission in the Trapezium region of the Orion Nebula (Stein & Gillett 1969), then in absorption toward the Galactic Center (Hackwell, Gehrz, & Woolf 1970), and toward the Becklin-Neugebauer object and Kleinmann-Low Nebula (Gillett & Forrest 1973).

• In 1973, Gillett, Forrest, & Merrill (1973) detected prominent emission features at 8.6, 11.3, and 12.7µm in the planetary nebulae NGC 7027 and BD+30°3639. These features together with the 3.3, 6.2, and 7.7 µm features were collectively known as the "unidentified infrared" (UIR) bands, which are now often attributed to polycylic aromatic hydrocarbon (PAH) molecules (Duley & Williams 1981; Léger & Puget 1984; Allamandola, Tielens, & Barker 1985; Allamandola, Hudgins, & Sandford 1999). ¹⁴

• Willner et al. (1979) detected a strong absorption band at 3.4 µm in the Galactic Center toward Sgr AW. Wickramasinghe & Allen (1980) detected this feature in the Galactic Center source IRS7. Although it is generally accepted that this feature is due to the C-H stretching mode in saturated aliphatic hydrocarbons, the exact nature of this hydrocarbon material remains uncertain (see Pendleton & Allamandola 2002, Pendleton 2004 for recent reviews). This feature has also been detected in a carbon-rich protoplanetary nebula CRL618 (Lequeux & Jourdain de Muizon 1990; Chiar et al. 1998) with close resemblance to the interstellar feature.

Interstellar Depletion: Where Have All Those Atoms Gone?

• In 1973, Morton et al. found that the gas-phase abundances of some heavy elements (relative to hydrogen) measured by the Copernicus UV satellite for interstellar clouds are significantly lower than in the Sun.

• In 1974, Field noted that the depletions of certain elements observed by Morton et al. (1973) correlate with the temperatures for dust condensation in stellar atmospheres or nebulae. He suggested that these elements have condensed into dust grains near stars and other elements have accreted onto such grains in interstellar space after they enter the ISM, forming a mantle composed of H, C, N and O compounds.

• In 1974, Greenberg found that the observed depletion of C, N, and O is significantly greater than could be accommodated by the dust under any reasonable models, using the gas-phase abundances measured by Copernicus for the Ophiuchi sightline (Morton et al. 1973) and the solar abundances as the reference abundances.

• Twenty years later, Sofia, Cardelli, & Savage (1994) found that the interstellar depletions are lowered for C, N, and O if B stars are used as the reference standard. They argued that the solar system may have enhanced abundances of many elements, and therefore the solar abundances are not representative of the interstellar abundances.

• Snow & Witt (1996) analyzed the surface abundances of B stars and field F and G stars and found that not only C, N, and O but also Si, Mg, and Fe and many other elements are underabundant in these stars. This led them to suggest that the interstellar abundances are appreciably subsolar (~ 60%-70% of the solar values). ¹⁵

Dust Luminescence: The "Extended Red Emission"

• In 1980, Schmidt, Cohen, & Margon (1980), detected in the Red Rectangle a far-red continuum emission in excess of what would be expected from simple scattering of starlight by interstellar dust. This continuum emission, known as the "extended red emission" (ERE), consists of a broad, featureless emission band between ~ 5400Å and 9500Å, peaking at 6100 < _p < 8200Å, and with a width 600Å < FWHM < 1000Å. ¹⁶ The ERE has been seen in a wide variety of dusty environments: the diffuse ISM of our Galaxy, reflection nebulae, planetary nebulae, HII regions, and other galaxies (see Witt & Vijh 2004 for a recent review).

• The ERE is generally attributed to photoluminescence (PL) by some component of interstellar dust, powered by UV/visible photons. The photon conversion efficiency of the diffuse ISM has been determined to be near unity (Gordon, Witt, & Friedmann 1998). The ERE carriers are very likely in the nanometer size range because nanoparticles are expected to luminesce efficiently through the recombination of the electron-hole pair created upon absorption of an energetic photon, since in such small systems the excited electron is spatially confined and the radiationless transitions that are facilitated by Auger and defect related recombination are reduced (see Li 2004a).

• The ERE carrier remains unidentified. Various candidate materials have been proposed, but most of them appear unable to match the observed ERE spectra and satisfy the high-PL efficiency requirement (Li & Draine 2002a; Li 2004a; Witt & Vijh 2004). Promising candidates include PAHs (d'Hendecourt et al. 1986) and silicon nanoparticles (Ledoux et al. 1998, Witt, Gordon, & Furton 1998, Smith & Witt 2002), but both have their own problems (see Li & Draine 2002a).

Stochastically Heated Ultrasmall Grains or "Platt" Particles

• In 1956, John R. Platt first suggested that very small grains or large molecules of less than 10Å in radius grown by random accretion from the interstellar gas could be responsible for the observed interstellar extinction and polarization. Platt (1956) postulated these "Platt" particles as quantum-mechanical particles containing many ions and free radicals with unfilled electronic energy bands.

• In 1968, Donn further proposed that PAH-like "Platt particles" may be responsible for the UV interstellar extinction.

• In 1968, Greenberg first pointed out that very small grains with a heat content smaller than or comparable to the energy of a single stellar photon, cannot be characterized by a steady-state temperature but rather are subject to substantial temporal fluctuations in temperature.

• Andriesse (1978) by the first time presented observational evidence for the existence of "Platt" particles in a dust cloud near M17, as indicated by its near-invariant 8-20 µm spectral shape over a distance of ~ 2' through the source and by its broad spectral energy distribution characterized by a combination of widely different color temperatures. He found that the the observed IR spectrum of M17 could be explained by a population of large grains and a population of "Platt" particles of ~ 10Å in size which exhibit temperature fluctuations.

• Sellgren, Werner, & Dinerstein (1983) found that the color temperatures of the 2-5 µm near-IR continuum (~ 1000 K) and the spectral shapes of the 3.3 µm emission features of three visual reflection nebulae NGC 7023, 2023, and 2068 show very little variation from source to source and within a given source with distance from the central star. They attributed the near-IR continuum emission to ultrasmall grains of radii ~ 10Å undergoing large excursions in temperature due to stochastic heating by single stellar photons.

• The presence of a population of ultrasmall grains in the diffuse ISM was explicitly indicated by the 12 µm and 25 µm "cirrus" emission detected by the Infrared Astronomical Satellite (IRAS) (Boulanger & Pérault 1988), which is far in excess (by several orders of magnitude) of what would be expected from large grains of 15-25 K in thermal equilibrium with the general interstellar radiation field. Subsequent measurements by the Diffuse Infrared Background Experiment (DIRBE) instrument on the Cosmic Background Explorer (COBE) satellite confirmed this and detected additional broadband emission at 3.5 µm and 4.9 µm (Arendt et al. 1998). More recently, spectrometers aboard the Infrared Telescope in Space (IRTS) (Onaka et al. 1996; Tanaka et al. 1996) and the Infrared Space Observatory (ISO) (Mattila et al. 1996) have shown that the diffuse ISM radiates strongly in emission features at 3.3, 6.2, 7.7, 8.6, and 11.3 µm.

Interstellar Grain Models: Modern Era

• The modern era of interstellar grain models probably began in 1977 with the paper by Mathis, Rumpl, & Nordsieck (1977). By fitting the interstellar extinction over the wavelength range of 0.11 µm < < 1 µm, Mathis et al. derived a power-law size distribution of dn / da ~ a^-3.5 for a mixture of bare silicate and graphite grains. ¹⁷ With the substantial improvements made by Draine & Lee (1984), this model became one of the standard interstellar grain models with well-characterized chemical composition, size distribution, optical and thermal properties. Modifications to this model were later made by Draine & Anderson (1985), Weiland et al. (1986), Sorrell (1990), Siebenmorgen & Krügel (1992), Rowan-Robinson (1992), Kim, Martin, & Hendry (1994), Dwek et al. (1997), Clayton et al. (2003), and Zubko, Dwek, & Arendt (2004) by including new dust components (e.g., amorphous carbon, carbonaceous organic refractory, and PAHs) and adjusting dust sizes (e.g., deriving dust size distributions using the "Maximum Entropy Method" or the "Method of Regularization" rather than presuming a certain functional form).
  Recent developments were made by Draine and his coworkers (Li & Draine 2001b, 2002b, c; Weingartner & Draine 2001a) who have extended the silicate-graphite grain model to explicitly include a PAH component as the small-size end of the carbonaceous grain population. It has been shown that the IR emission spectrum calculated from this model closely matches that observed for the Milky Way (Li & Draine 2001b), the Small Magellanic Cloud (SMC; Li & Draine 2002c), and more recently the ringed Sb galaxy NGC 7331 (Regan et al. 2004; Smith et al. 2004), including the "UIR" emission bands at 3.3, 6.2, 7.7, 8.6, and 11.3 µm.

• In contrast to the bare silicate-graphite model, Greenberg (1978) proposed that interstellar grains could be coated by a layer of organic refractory material derived from the photoprocessing of ice mantles acquired in molecular clouds and repeatedly cycled into and out of diffuse and molecular clouds. The organic refractory mantles would provide a shield against destruction of the silicate cores. Since the rate of production of silicate dust in stars is about 10 times slower than the rate of destruction in the ISM (mostly caused by sputtering and grain-grain collisions in interstellar shock waves; Draine & Salpeter 1979a, b, Jones et al. 1994), the silicates would be underabundant if they were not protected and thus it would be hard to explain the observed large depletions of Si, Fe and Mg and the strength of the observed 9.7 µm silicate absorption feature, unless most of the silicate mass was condensed in the ISM as suggested by Draine (1990). The most recent development of this model was that of Li & Greenberg (1997), who modeled the core-mantle grains as finite cylinders (to account for the interstellar polarization). In addition, a PAH component and a population of small graphitic grains are added respectively to account for the far-UV extinction rise plus the "UIR" emission bands and the 2175Å extinction bump.

  Modifications to this model were also made by considering different coating materials (e.g., amorphous carbon, hydrogenated amorphous carbon [HAC]), including new dust type (e.g., iron, small bare silicates), and varying dust size distributions (Chlewicki & Laureijs 1988; Duley, Jones, & Williams 1989; Désert, Boulanger, & Puget 1990; Li & Greenberg 1998; Zubko 1999). In particular, Duley et al. (1989) speculated that the silicate cores are coated with a mantle of HAC material arising from direct accretion of gas-phase elemental carbon on the silicate cores in the diffuse ISM.

• Recognizing that grain shattering due to grain-grain collisions and subsequent reassembly through agglomeration of grain fragments may be important in the ISM, Mathis & Whiffen (1989) proposed that interstellar grains may consist of a loosely coagulated structure built up from small individual particles of silicates and carbon of various kinds (amorphous carbon, HAC, and organic refractories). Further developments of this composite model were made by Mathis (1996), Iatì et al. (2001, 2004), Saija et al. (2001, 2003), and Zubko, Dwek, & Arendt (2004) (see Section 2 of Li 2004a for more details).

¹ To be precise, this should be called "extinction" which is a combined effect of absorption and scattering: a grain in the line of sight between a distant star and the observer reduces the starlight by a combination of scattering and absorption. Back.

² The photometric distances were obtained by comparing apparent and absolute magnitudes, with the latter determined from the spectral types of the stars in the clusters. The geometrical distances were determined from the angular diameters of the clusters, assuming that all their diameters were the same. Back.

³ As mentioned earlier in this review, general star counts did suggest the existence of interstellar extinction which increases with distance. However, this evidence is not decisive because interpretation of the star-count data rests on assumptions (generally unproved at the time) as to the true spatial distribution of the stars. Back.

⁴ Trumpler (1930) derived a color-excess of ~ 0.3 mag kpc^-1 between the photographic (with an effective wavelength _B 4300Å) and visual (_V 5500Å) bands, and a general (visual) absorption of ~ 1.0 mag kpc^-1. Back.

⁵ The exact nature of the carrier of this bump remains unknown. It is generally believed to be caused by aromatic carbonaceous (graphitic) materials, very likely a cosmic mixture of polycyclic aromatic hydrocarbon (PAH) molecules (Joblin, Léger & Martin 1992; Li & Draine 2001b). Back.

⁶ Very recently, on the basis of the FUSE observations of 9 Galactic sightlines at 1050Å < < 1200Å, Sofia et al. (2005) found that the CCM prediction for short-wavelengths (^-1 > 8 µm^-1) is not valid for all sightlines. Back.

⁷ Van de Hulst (1949) pointed out that this is not the case for H, He and Ne since they will evaporate rapidly at grain temperatures exceeding ~ 5 K. Back.

⁸ The "Serkowski law" P() / P_max = exp [- K ln²( / _max)] is determined by only one parameter: _max - the wavelength where the maximum polarization P_max occurs; the width parameter K is related to _max through K 1.66 _max + 0.01. The peak wavelength _max is indicative of grain size and correlated with R_V: R_V (5.6 ± 0.3) _max (_max is in micron; see Whittet 2003). Back.

⁹ Many years later, the idea of metallic iron grains as an interstellar dust component was reconsidered by Chlewicki & Laureijs (1988) who attributed the 60 µm emission measured by IRAS for the Galactic diffuse ISM to small iron particles with a typical size of a 70Å (which would obtain an equilibrium temperature of ~ 53 K in the diffuse ISM). But their model required almost all cosmic iron to be contained in metallic grains: ~ 34.5ppm (parts per million) relative to H. Exceedingly elongated metallic needles with a length (l) over radius (a) ratio l /a 10⁵, presumably present in the intergalactic medium, have been suggested by Wright (1982), Hoyle & Wickramasinghe (1988), and Aguirre (2000) as a source of starlight opacity to thermalize starlight to generate the microwave background. Very recently, elongated needle-like metallic grains were suggested by Dwek (2004) as an explanation for the flat 3-8 µm extinction observed by Lutz et al. (1996) toward the Galactic Center and by Indebetouw et al. (2005) toward the l = 42° and 284° lines of sight in the Galactic plane. But these results heavily rely on the optical properties of iron needles (see Li 2003a, 2005b). Back.

¹⁰ Whittet, Duley, & Martin (1990) estimated from the 7.7-13.5 µm spectra (with a spectral resolution of ~ 0.23 µm) of 10 sightlines toward the Galactic Center the abundance of Si in SiC dust to be no more than ~ 5% of that in silicates. Since about half of the dust in the ISM is injected by carbon stars in which an appreciable fraction of the stardust is SiC, it is unclear how SiC is converted to gas-phase and recondense to form silicates in the ISM. Back.

¹¹ Nanodiamonds were identified in the dust disks or envelopes surrounding two Herbig Ae/Be stars HD 97048 and Elias 1 and one post-asymptotic giant branch (AGB) star HR 4049, based on the 3.43 µm and 3.53 µm C-H stretching emission features expected for surface-hydrogenated nanodiamonds (Guillois, Ledoux, & Reynaud 1999; van Kerckhoven, Tielens, & Waelkens 2002). Back.

¹² The reason why so many different materials with such a wide range of optical properties could be used to explain the observed interstellar extinction was that the number of free parameters defining the size distribution was sufficiently large. Back.

¹³ The reason why Wickramasinghe (1970a) considered ice-coated silicate grains was that he thought that graphite grains of a typical size ~ 0.06 µm would attain an equilibrium temperature of ~ 40 K in the ISM and would be too warm to possess an ice mantle, while silicate grains would tend to take up lower temperatures because of their lower optical and UV absorptivity and therefore the condensation of ice mantles could occur on their surfaces. Back.

¹⁴ Since the "UIR" emission bands were initially found to be associated with UV-rich objects, it had been thought that they were pumped primarily by UV photons. Li & Draine (2002b) demonstrated that the excitation of PAHs does not require UV photons - since the PAH electronic absorption edge shifts to longer wavelengths upon ionization and/or as the PAH size increases (see Mattioda, Allamandola, & Hudgins 2005 for their recent measurements of the near-IR absorption spectra of PAH ions), therefore long wavelength (red and far-red) photons are also able to heat PAHs to high temperatures so that they emit efficiently at the "UIR" bands (also see Smith, Clayton, & Valencic 2004). Li & Draine (2002b) have modeled the excitation of PAH molecules in UV-poor regions. It was shown that the astronomical PAH model provides a satisfactory fit to the UIR spectrum of vdB133, a reflection nebulae with the lowest ratio of UV to total radiation among reflection nebulae with detected UIR band emission (Uchida, Sellgren, & Werner 1998). Back.

¹⁵ The most recent estimates of the solar C ([C/H] 245 ppm; Allende Prieto, Lambert, & Asplund 2002) and O abundances ([O/H] 457 ppm; Asplund et al. 2004) are also "subsolar", just ~ 50%-70% of the commonly-adopted solar values (e.g. those of Anders & Grevesse 1989) and close to the "subsolar" interstellar abundances originally recommended by Snow & Witt (1996). If the interstellar abundances are indeed "subsolar", there might be a lack of raw material to form the dust to account for the interstellar extinction. Mathis (1996) argued that this problem could be solved if interstellar grains have a fluffy, porous structure since fluffy grains are more effective in absorbing and scattering optical and UV starlight than compact grains (on a per unit mass basis). However, using the Kramers-Kronig relation, Li (2005a) demonstrated that fluffy dust is not able to overcome the abundance shortage problem. The abundances of refractory elements in stellar photospheres may under-represent the composition of the interstellar material from which stars are formed, resulting either from the possible underestimation of the degree of heavy-element settling in stellar atmospheres, or from the incomplete incorporation of heavy elements in stars during the star formation process. Back.

¹⁶ Very recently, Vijh, Witt, & Gordon (2004) reported the discovery of blue luminescence at < 5000Å in the Red Rectangle and identified it as fluorescence by small three- to four-ringed PAH molecules. Nayfeh, Habbal, & Rao (2005) argued that this blue luminescence could be due to hydrogen-terminated crystalline silicon nanoparticles. Back.

¹⁷ Such a power-law size distribution is a natural product of shattering following grain-grain collisions (e.g. see Hellyer 1970, Biermann & Harwit 1980, Dorschner 1982, Henning, Dorschner, & Gürtler 1989). Back.