Scientific Prospects for a Terahertz TelescopeContents: YSOs PDRs Chemistry (H2D+, NH+, LiH)
Nearby galaxies (CO, N+) High-redshift C+ Continuum studies
The terahertz (THz) frequency range (1-3 THz = 100-300 microns) provides an important window on the far-infrared Universe. Recent studies of the transmission spectrum of Earth's atmosphere reveal that several windows in the 1-2 THz range become significantly transparent from high altitude (> 5000m) sites in northern Chile. Shown below in Figure 1 is a zenith transmission spectrum acquired with a Fourier Transform Spectrometer (FTS) built and deployed on Chajnantor by the receiver lab staff. This level of transmission occurs during good weather conditions, but is by no means the best transmission we have recorded. Recently, the FTS has been moved to a higher altitude site (5500m) where the first results have revealed even better transmission. As a result of this work, the prospect for a unique, ground-based terahertz telescope can now be considered for both Galactic and extragalactic astronomy.
[Figure 1: FTS spectrum measured at Chajnantor, Chile]
Spectroscopy -- Galactic Studies
Young stellar objectsAs can be seen in the transmission plot above, several interesting spectral lines will be accessible to a ground-based terahertz telescope. A local catalog of line frequencies is currently being compiled. Strong, high-J rotational transitions of the abundant species CO, HCN and HCO+ will provide unique information on hot cores as well as on young low-mass stars. For example, a new hot component of CO gas (T=1000K) has been discovered with the ISO LWS spectrometer in the low-mass protostellar source L1448-mm. A portion of the spectrum containing five CO transitions is reproduced below (Figure 2). The total on-source integration time was 1.2 hours and the spectral resolution was 300. The diameter of the telescope on ISO was 0.6-meter.
For a molecular region at such high temperatures, the rotation diagram in Figure 3 (shown above right) indicates that the CO rotational transitions appearing in the terahertz windows accessible from the ground (CO 13-12, 12-11, 11-10 and 9-8) also contain the maximum line flux. Moderate resolution (10 km/s) observations of these lines provide a simple way to measure the column density of hot gas in a variety of objects (shocks, hot cores, PDRs) with no interference from cool foreground gas as is often the case in lower transitions. Observations of other molecular species in this hot gas will reveal the chemistry in a temperature regime previously accessible only to ISO.
In the case of L1448-mm, the predicted
line flux in each of these CO lines is 1.0E-19 W cm-2,
which is over 1000 times the flux found in the 2-1 line. The size of
the emitting region is inferred from models to be <12 arcseconds,
which will be completely contained within the beam of a 3 to 6-meter
terahertz telescope. With an intrinsic linewidth of 50 km/s (173 MHz
at 1.036 THz), this flux translates to a flux density of 5.8E-28 W
cm-2 Hz-1 = 5.8E-24 W m-2
Hz-1 = 580 Jansky (Jy). At a resolution of 10 km/s (35
MHz), the peak flux density will be at least 116 Jy/channel. As seen
in the following table, sensitivity calculations for a 3-meter
ground-based terahertz telescope show that lines of this strength can
be readily detected. It is the large gain in collecting area over ISO
which compensates for the added atmospheric losses, while also
providing new information via the superior angular resolution. For
example, in the case of L1448-mm, the angular resolution of a 3-meter
telescope (24'') will reveal whether the hot CO originates from the
accretion disk or from shocks in the outflow lobes.
As a demonstration of these sensitivity calculations, in January 2000 a Hot Electron Bolometer receiver constructed by the receiver lab staff was used to detect the CO 9-8 line at the Heinrich Hertz Telescope (HHT) on Mt. Graham. The spectrum toward Orion IRc2 (shown below in Figure 4) was obtained in only seven minutes of integration time with a zenith transmission of only about 3%. This spectrum is the first ground-based heterodyne detection in the terahertz band. This ease of detection under such poor conditions illustrates that much fainter sources like L1448-mm will be detectable from the more transparent skies in Chile.
[Figure 4: First Terahertz CO spectrum obtained at Mt. Graham with the HEB receiver]
An even better chance for a detection of para-H2D+ comes from observations of this transition in absorption towards bright galactic continuum sources (Orion IRc2, Sgr B2, W49, W51 are examples). We use the expressions given by Spitzer (1978) for interstellar absorption lines and the de-excitation rate of this transition of 3.46E-03 s-1 -- and assume that 90% of the molecules reside in the ground state. We find that the column density would be N(para-H2D+) ~ 2.2E+13 W(km/s) cm-2, where W(km/s) is the total integrated absorption in K*km/s divided by the continuum strength. In the table below we provide detection estimates for two sources Orion IRc2 and W49 - using an estimate of the line width from Boreiko and Betz (1993) and SWAS H2O observations:
A survey of bright continuum sources can be expected to obtain several
detections or set sensitive limits that will challenge chemical theory
which predicts an H2D+ abundance of 3E-11 at 10K.
Water linesA 3m terahertz telescope would also be able to search for a number of high lying transitions of H218O and even some transitions of H2O that are not excited in the Earth's atmosphere. Water is an important astrochemical and biological molecule and these transitions will be able to sample some of the most prolific sites of water production in the galaxy: shocks (Harwit et al. 1998) and hot cores (Jacq et al. 1988). In these regions through high temperature chemistry and grain surface evaporation the water abundance becomes enhanced by several orders of magnitude and water becomes one of the dominant coolants. At present only SWAS (operated at SAO), and perhaps the ODIN satellite, can easily observe these species. A terahertz telescope would offer the opportunity to follow up on discoveries made by SWAS with observations of higher excitation transitions and with a smaller beamsize. This could prove to be quite enlightening as the 4 arcminute SWAS beam encompasses many different physical regimes and the higher resolution terahertz telescope could locate and map the water emission region. Similarly there are several high excitation lines of deuterated water (HDO) that lie in the 1 THz window. Also the 2(21)-1(10) transition of D2O at 1.5287 THz with an upper state energy of 100 K lies within the window. D2O has yet to be detected in an astrophysical context and its detection would yield important clues to the formation pathways of water in the interstellar medium.
Cosmological and Extragalactic Studies
The Primordial Abundance of LithiumPrimordial molecules are believed to play an important role in the evolution of the early universe, as after decoupling of matter and radiation simple molecules such as H2, HD, and LiH are important contributors to the thermal evolution. Despite its importance the primordial abundance of Li is unknown and sensitive searches of low lying transitions of LiH, the simplest Lithium-Hydride, have failed to convincingly detect the molecule (Combes and Wikland 1998). One significant advantage of the 1.23 - 1.4 THz band is access to the J=3(0)-2(0) transition of LiH at 1.329 THz (Bellini et al. 1995). Currently the abundance of LiH in the ISM is unknown and a detection of this transition would be an important step in understanding the chemistry resulting in LiH formation. In the future observations of LiH as a function of redshift will make it possible to derive the true lithium abundance (Combes and Wikland 1998) and a detection at z=0 is the logical starting point. In addition, some models of Big Bang nucleosynthesis require Lithium to be enriched by nearly a factor of 10 by stellar nucleosynthesis (see Puy and Signore 1997, Puy and Signore 1998). If this is the case the terahertz telescope could take advantage of the known galactic gradient of atomic abundances (e.g. Smart and Rolleston 1999). Detections of LiH as a function of Galactic radius, combined with a chemical model, could set important limits on this possibility.
Detection estimates for Lithium HydrideFor our detection estimates of Lithium Hydride, we draw interesting parallels to the excitation of H2O. Both molecules have a high dipole moment (5.6 Debye for LiH). This property gives rise to fast de-excitation coefficients (4 s-1 for the 3(0)-2(0) transitions), high critical densities > 1.0E+11 cm-3, and high expected optical depths, even for a small LiH abundance. Under these conditions, where a molecule is expected to be sub-thermally excited (due to densities below the critical density) and optically thick, every photon that is created will eventually escape the cloud (i.e. there are no collisional de-excitations that result in photon destruction). The emission can therefore be considered effectively thin. This is quite similar to H2O and we use the methods describing water excitation given by Snell et al (2000). Here the emission can be described by the density, temperature and total column density. Thus for Orion BN/KL with a density of 1.0E+06 cm-3, T = 100 K, and a column density of 1.0E+23 cm-2, the LiH abundance limit achieved in 1 hour of integration on a 3-meter telescope would be an incredibly low estimate of 1.0E-14 relative to H2. With a 0.8-meter telescope, the line antenna temperature will probably be lower due to beam dilution (depending on the extent of the emission), thus the abundance limit is more like 1.4E-13. In contrast, the Li abundance relative to H is thought to be about 1.0E-10 (Lemoine et al. 1993). Thus, if all of the Li resides in LiH, a detection would be quite easy. However, we anticipate that only a small amount resides in LiH and long integrations will be required. In any case, we we will either set very sensitive limits or obtain a detection. While Orion BN/KL is a high excitation source the special radiative transfer of LiH will allow for excellent abundance limits < 1.0E-12 relative to H2 to be set towards just about any galatic molecular core.
Ionized gas in nearby galaxies: NIIAs a global tracer of the interstellar medium, NII and CII occupy a similar place with respect to the ionized medium as HI does for the neutral medium and CO does for molecular medium. In view of extragalactic studies, there are two advantages of observing the 205 micron NII emission line over the optical transitions of NII. First, this line is unaffected by extinction so that a view of the entire target galaxy can be gathered. Second, the kinematic information carried by this line can be resolved and studied by heterodyne receivers. A 3-meter terahertz telescope will be able to detect the 205 micron NII line in nearby galaxies, depending on the fraction of the emission that arises in a compact nuclear region. This line was first measured from the 0.915-meter KAO telescope in M82 to be 7.1E-19 W cm-2 in a 54 arc second beam (Petuchowski et al. 1994). Recent ISO measurements found a larger flux of 1.8E-18 W cm-2 in a 70 arc second beam (Colbert et al. 1999). We can estimate the line strength in other galaxies by using the ISO measurements of the 122 micron NII line flux (since both NII lines have similar strength in the Milky Way). For example, the ISO-measured flux of the 122 micron line in Centaurus-A is 1.5E-19 W cm-2 (Unger et al. 2000 ), or about 1/12 of the strength in M82. For the following calculations, we assume that much of this emission originates from a compact nuclear region that is contained by the telescope beam. With a linewidth of 300 km/s (1.5 GHz), this NII line flux corresponds to a flux density of 1.0E-28 W cm-2 Hz-1 = 1.0E-24 W m-2 Hz-1 = 100 Jy. When observed with 100 km/s resolution, the peak flux density will be at least 33 Jy/channel. Similar numbers follow for the normal galaxy NGC4414, which has a 122 micron NII line flux of 1.3E-19 W cm-2 (Braine & Hughes 1999), and should exhibit about 30 Jy/channel in the 205 micron NII line. If the emission is more extended than the beam, then more time will be required to detect the NII emission from these galaxies. Clearly, in order to generate a significant scientific impact, a 6-meter class telescope will be needed for such extragalactic studies.
Molecular gas in nearby galaxies: COObservations of CO at millimeter wavelengths have provided an important measure of the molecular component of the ISM in all types of galaxies. Information such as the mass, dynamics and star formation rate have been derived in surveys (e.g. Young et al. 1995). To date, most observations have been performed in the low J transitions: primarily in 1-0 and 2-1 ( Sage 1993; Braine et al. 1992), some in 3-2 (Mauersberger et al. 1999), and a few in 4-3 (Guesten et al. 1993). In contrast, the higher J transitions trace warmer gas.
Ionized gas in redshifted galaxies: CIIFinally, for many years, people have considered the prospect of observing the forbidden fine-structure line of singly-ionized carbon (CII) at 1.909 THz in high-redshift galaxies. The relative brightness of this interstellar cooling line in the COBE FIRAS spectrum of the Milky Way and in airborne observations of other galaxies (Stacey et al. 1991) suggested that it might be a beacon for detecting distant galaxies. As an example, ISO spectra indicate that the CII line carries 0.4% of the total luminosity of the giant elliptical galaxy Centaurus-A. In M82, the CII line is 7-8 times brighter than the 122 micron NII line, and is surpassed only by the neutral OI line at 63 microns. But ISO surveys of nearby galaxies show a much lower level of CII emission in FIR-bright galaxies (Malhotra et al. 1999) and ultraluminous galaxies (Luhman et al. 1998). As yet, no convincing detection of this line has been made beyond the local Universe. However, most searches have been performed at wavelengths of 1 mm and longer, corresponding to galaxies at redshifts greater than 5. In constrast, the terahertz windows offer a chance to search for this line in relatively lower redshift galaxies in three ranges of z: 0.25-0.29, 0.38-0.43 and 0.80-0.90. A simple search of the IRAS catalog using NED reveals 28 objects with known redshifts within these ranges. 17 of these objects were detected in at least two IRAS bands. All of these objects are ultraluminous galaxies and make interesting targets for a deep search.
Continuum -- Galactic StudiesIn addition to spectral lines, the terahertz band contains the bulk of the dust continuum emission from cold (T=10-20K), prestellar cores. Due to their low temperatures and high column densities, these objects will appear as dark, obscuring clouds from optical through mid-infrared wavelengths. A good example is L1689B (Bacmann et al. 2001), which appears as a dark cloud in the 6.75 micron ISOCAM image (Figure 6), while exhibiting emission at longer wavelengths, as shown by the 1.3mm contours. The spectral energy distribution in Figure 7 (below right) indicates that the flux density from this source peaks exactly in the terahertz band while falling steeply at higher frequencies (due to the Wien limit) and at lower frequencies (due to frequency-dependent dust emissivity).
From these results, it appears likely that these prestellar clouds exist in the early stages of fragmentation and collapse leading to star formation. Terahertz photometry of these objects (about 50 Janskys), combined with millimeter flux densities (about 50 milliJanskys), will tightly constrain the dust temperature and total mass in these clouds. At the same time, terahertz imaging will provide important information on their volume density profile and degree of fragmentation. All of this information is essential for determining the mass distribution of prestellar cores, which should shed new light on the origin of the stellar initial mass function (IMF). However, a 6-meter-class terahertz telescope equipped with a bolometer array camera (12'' resolution) would be required before we could effectively compete with existing cameras such as MAMBO on the IRAM 30-meter at 1.3 mm, SCUBA on the JCMT at 800 microns, and SHARC2 on the CSO at 350 microns. A similar argument holds for extragalactic studies,
Airborne projectsVery little competition currently exists for observations in the terahertz band. On two days per year, the TIFR Balloon Facility at Hyderabad, India launches a 1-meter balloon-borne telescope with a small bolometer array which performs 100-200 micron continuum observations. In a much larger project, the NASA/DLR 2.5-meter airborne observatory SOFIA plans to field a menagerie of ten instruments, three of which can access the terahertz band. These include two non-facility, PI-class spectrometers: a heterodyne receiver (CASIMIR) and a lower resolution Fabry-Perot spectrometer (SAFIRE); along with a facility continuum camera (HAWC). Two additional low-resolution spectrometers, FIFI-LS and AIRES both have a long-wavelength cutoff of 210 microns. Although this suite of instruments holds great promise, the first flight of SOFIA was recently delayed by two years until October 2004, with general observations not beginning until January 2005. In plain terms, that means no competition from them for at least four more years. Even after SOFIA achieves routine operation, only about 100 hours per year will be available for heterodyne observations, in which CfA scientists hold great expertise. Also, a majority of the SOFIA flights will be made from its headquarters in the northern hemisphere (California), whereas a ground-based telescope in Chile will have unique access to the wealth of relatively less explored star forming sites in the fourth Galactic quadrant.
Spaceborne projectsIn terms of space-based competition, the 0.85-meter telescope SIRTF was recently launched and carries MIPS which includes a 2x20 pixel continuum detector for the 170 micron band. Although it will be quite useful for extragalactic studies, this detector is expected to be saturated for all targets within the Milky Way due to high background. In 2003, the Japanese ISAS plan to launch the 0.7-meter telescope ASTRO-F with the Infrared Imaging Surveyor (IRIS) to perform an all-sky survey with better sensitivity than IRAS. In addition, before the 500-day lifetime elapses, there will be some time for pointed observations of continuum and very-low resolution spectroscopy with a long-wavelength limit of 200 microns. More substantial competition in the terahertz band will begin to arrive with the HIFI heterodyne receiver system on the next generation 3.5-meter orbiting infrared observatory Herschel (FIRST). However, this complex satellite remains in the planning stages and will not be launched before 2007 or 2008. Also, only single pixel receivers will be present on this satellite, while a ground-based observatory could benefit from a future array receiver.
In summary, there is a worldwide effort underway to develop terahertz astronomy over the next decade. We feel that a unique opportunity exists for the CfA to lead by example. In fact, the CfA team has already demonstrated the feasibility of ground-based terahertz astronomy by obtaining the first such detection at Mt. Graham.