TechnologyObserving FacilitiesEducation and OutreachAbout Haystack


Digital receivers have become widely employed in communication. These receivers employ digital baseband conversion, FIR filtering. The processing power needed has only recently become available at reasonable cost with the advent of the GrayChip digital receiver and the Analog Devices ADSP21161N DSP. The advantages of digital processing are

  1. High dynamic range » 90 dB
  2. High bandpass stability
  3. Excellent out of band rejection
  4. Software implementation of special algorithms

These advantages offer improvements in spectral line performance, electronic beam steering, simultaneous processing of multiple primary beams and interference rejection for radio astronomical and other applications. We plan to demonstrate these improvements by building an array at the Haystack Observatory for the following science projects:

  1. A more sensitive search for the 327 MHz hyperfine line of deuterium than previous attempts. Detection will place better limits on the baryon density without the potential confusion with hydrogen that exists in the UV/optical Lyman series lines, for which the hydrogen and deuterium spectral features differ by only 0.03 percent in wavelength.
  2. High sensitivity imaging of the hydrogen and carbon recombination lines (H271a , H272a , C271a , C272a ) in the 322 to 328.6 MHz radio astronomy band.
  3. Searches for other spectral lines like the recombination lines of other elements.

In addition the array will provide a test bed for technical developments of large arrays at low frequency, which can benefit the Low Frequency Array (LOFAR) and Square-Kilometer Array (SKA), such as mitigation of radio frequency interference, digital receivers, and wideband communication systems, as well as serve as a training opportunity for students in instrumentation.

Science goals and history of the 327 MHz deuterium line

Starting in the late 1950s and early 1960s the detection of deuterium in the interstellar gas has been considered one of the most important efforts in radio astronomy. Its measurement constrains the photon to baryon ratio, and hence the cosmological baryon density. This measurement, combined with dynamical measurements in clusters and other estimates of the overall mass density, provide a gauge of the amount of non-baryonic dark matter in the universe. Furthermore, the degree to which deuterium is depleted in the interstellar medium of our Galaxy and other galaxies provide a tracer of stellar activity. As discussed below, the small isotope shift in the optical lines make the deuterium measurement extremely difficult and subject to systematic error at optical wavelengths. Recent estimates of the primordial deuterium-to-hydrogen ratio have differed by more than an order of magnitude. In contrast, at radio wavelengths the hyperfine lines of deuterium and hydrogen are separated by more than a factor of three in wavelength. Detection of the deuterium line at radio wavelengths would introduce a new tool, sharper than our existing tools, for studying deuterium abundances.

The primordial abundance of deuterium is determined by the competition between the nuclear reaction rates and the universal expansion rate. Deuterium is thought to be created by cosmological nucleosynthesis, but then is destroyed (astrated) by stellar nucleosynthesis (Epstein, Lattimer & Schramm 1976, Nature, 263, 198.) Measurement of the present D abundance is therefore a critical diagnostic of stellar and chemical evolution in the Galaxy, as well as an essential prelude to determining the primordial cosmic abundance. At present D/H ratios have been derived for cosmologically distant sources (quasars) and for local (< 500 pc) stellar Milky Way sources. The 327 MHz transition of deuterium offers the potential of measuring the ISM D/H abundance ratios for sources anywhere in the Galaxy. A deuterium abundance measurement good to a few tens of percent accuracy leads to very interesting constraints on stellar processing and chemical evolution models. With the instrument proposed here we should be able to achieve this level of accuracy, which will constitute an important step towards a full understanding of primordial deuterium abundances.

Several groups have tried to detect the deuterium hyperfine line including Sander Weinreb whose thesis experiments (Weinreb, 1962) set an upper limit for ND/NH = 8 &times 10 - 5. More recent limits by Anantharamaiah and Radhakrishnan (1979), Chengalur et al (1997), Heiles et al (1993) and Blitz and Heiles (1987) are comparable or slightly better. UV/Optical measurements give ND/NH » 2 &times 10-5 in the spectra of quasars and stars (Dupree et al, 1977), however these measurements are subject to confusion with hydrogen at a different redshift (Lemoine et al, 1999). Figure 3 shows the dependence of the baryon to photon ratio created in the Big Bang from the theory of nucleosynthesis (Walker et al, 1991) on the deuterium abundance. In greatly simplified terms, after the primordial fire has cooled enough to allow nuclear reactions, neutrons and protons combine to form deuterons which are the nuclei of deuterium atoms. The theory predicts a limited supply of neutrons so that 75% of the protons will remain to become hydrogen atoms. Most of the deuterons undergo further reactions to form helium and small amounts of helium 3 and lithium. The fraction of deuterons which remain to acquire an electron and become deuterium decreases with an increase in the density of deuterons. The temperature at which deuteron pairing to form helium occurs determines the photon density at that epoch and hence the fraction of deuterium to hydrogen decreases with an increased density ratio of baryons to photons. While most of the UV/optical measurements are consistent with the radio limit of S. Weinreb, some of the measurements of large deuterium abundance in the absorption spectra of quasars (Rugers and Hogan, 1996) are now thought (Burles et al, 1999) to be the result of the confusion with the strong hydrogen absorption which is very difficult to separate from any weak deuterium absorption. There are also UV/optical measurements of ND/NH as low as 7 &times 10-6 (Jenkins et al, 1999). Modest destruction of deuterium through reprocessing (Tosi et al, 1998) may explain some of the variation. The current status of deuterium abundance is reviewed by Lemoine et al (1999).

Additional data from the radio spectrum could make an important contribution to a better understanding of deuterium abundance variations due to the destruction of deuterium by reprocessing or apparent enhancement in interstellar clouds where much of the hydrogen is molecular. To reliably detect the 327 MHz line for ND/NH = 2 &times 10-5 we need to improve the sensitivity, over that achieved by Weinreb, by a substantial factor. Sensitivity can be improved by the following factors:

  1. 12-bit/sample avoids 1-bit clipping correction to gain 1.6
  2. No need for comparison switching gains a factor of 2
  3. Use of the array of 32 stations each with both polarizations gains a factor of Ö64 »8

The largest sensitivity improvement, which comes from the averaging of the spectra from the array stations, is valid for the detection of interstellar lines if absorption or emission is sufficiently extended so that the signals received at each station are uncorrelated. For the search of the deuterium line we propose to make the station sub-arrays large enough to ensure that the deuterium line should appear in absorption against the Galactic synchrotron emission in the central part of the Galaxy. The Galactic center continuum should be about 500 K with a 25 element sub-array with 23 dB gain, 12 degree beamwidth and 12 m2 effective aperture. If the sub-arrays are separated by about 15 m (4 degrees fringe spacing) the continuum emission which is extended will be largely resolved thereby decorrelating the signals. While the deuterium line is expected in absorption against a background brightness temperature greater than the excitation temperature, the hydrogen and carbon recombination lines, which have been previously observed, (for example see Roshi and Anantharamaiah, 1997) are seen in emission. We will also look for deuterium emission in the region of the Galactic anticenter. In this direction the non thermal background is only about 70 K and if the deuterium excitation temperature is about 130 K the line should appear in emission. Since the emission in this region should be diffuse the array has a large advantage over previous experiments, which used a single dish. While previous observations of the Galactic anticenter by Chengular et al. (1997) and Blitz and Heiles (1987) show a possible detection at 0 km/s of about 2-3 mK, the proposed array should reach 1 sigma noise of about 100 mK in 400 hours at a resolution of 10 km/s.

In addition we can simultaneously process spectra from different primary beams at each station, which would allow simultaneous mapping of any deuterium line with the 12 degree resolution of each station. This multibeaming capability multiplies the effective integration time, and for the purposes of measuring spectra in many different directions, constitutes a major additional sensitivity improvement over existing instruments and studies.

Chengular et al used the 14 separate antennas of the Westerbork array to improve sensitivity. The 32 station array we propose will be about a factor of 4 more sensitive than the Westerbork array for the same amount of observing time. In addition we expect to dedicate most of the observing time to measurement of the deuterium line, with corresponding sensitivity gains over the time constrained Westerbork experiments.

In order to study the effects of the reprocessing of deuterium we propose to observe the deuterium line in at least two directions: towards the Galactic center and towards the anticenter. However, since this will be a dedicated instrument, and in contrast to any existing instrument, sufficient observing time will be available to observe at a variety of Galactic longitudes. This capability is powerfully leveraged by the multibeaming capacity of the dipole array design, mentioned above. Velocity crowding in the Galaxy occurs in several different directions, offering the opportunity to probe D/H over the entire Galactic disk. Such measurements can provide a powerful and detailed constraint on models of stellar processing, which predict a deuterium abundance gradient, with the lowest abundances occurring in the highly processed inner galaxy.

Recently Lubowich et al. (2000) have detected 2 millimeter wave transitions of DCN in emission in a molecular cloud near the Galactic center. Their result suggests that the higher than expected D/H ratio is the result of a recent infall of unprocessed gas. Our proposed absorption measurement towards the Galactic center will be able to map the D/H along the line of sight through the many regions which have different velocities, providing complementary information with which to help trace the processes affecting abundances in the region.

It is realized that the conversion of these measurements to a D/H ratio is difficult due to the lack of the precise knowledge of the optical depths and the excitation temperature of H in the gas clouds along the line of sight. The difficulty of making accurate isotopic abundance measurements has long been recognized (Wilson and Rood (1994); and Rogers and Barrett, (1966). However, systematic errors exist in every method of determining the isotopic abundance ratios so that it is important to make and compare measurements using different techniques.

With a change in the elevation of the stations we can also look for deuterium in absorption in Cassiopeia A. I n this case we would need to combine adjacent stations to achieve sufficient collecting area to reach the point at which the Cas A signal dominates. Since Cas A is relatively compact we would not gain the advantage of many independent spectra, so that the sensitivity would not be significantly better than that obtained by Carl Heiles et al (1993) who accumulated very long integrations during which some narrow features were seen and then not confirmed. However, the use of digital receivers and the ability to simultaneously gather data from multiple primary beams combined with interference rejection should lead to smaller systematic errors, and the prospect of performing very long integrations without being limited by systematics. The success of such measurements would represent a strong incentive to develop 327 MHz capabilities in arrays of the future (as, perhaps, in LOFAR) and these arrays would have sufficient sensitivity and resolution to probe lines of sight toward many extragalactic sources.

Figure 3 Predicted deuterium abundance from Big Bang theory (from Walker et al, 1991)

*The original proposal was to build 64 stations. The number has been reduced to 24 as a result of the limited available funds.



Haystack Radio Telescope

Wideband VLBI

Radio Arrays (MWA, SKA, MAPS, Deuterium)

Small Radio Telescope (SRT)




Westford Radio Telescope

Mark 4 VLBI Correlator

VLBI Technique Development

International VLBI Service (IVS)


Atmospheric Sciences


Millstone Hill Observatory


Array Systems

Open Source Projects

Space Science Resources



Radio Arrays (MWA, SKA, MAPS, Deuterium)

Very Long Baseline Interferometry (VLBI)

Ionospheric Techniques

Open Source Projects

Observing Facilities

Haystack Radio Telescope

Westford Radio Telescope

Millstone Hill Radar

Deuterium Array

Atmospheric Optics Facility

Education and Outreach

Undergraduate Research

Research Experiences for Undergraduates (REU)

Research Experiences for Teachers (RET)

Pre-College Resources

Public Outreach Activities

About Haystack

Northeast Radio Observatory Corp. (NEROC)


News Archive



Positions Available

Art at Haystack





Safety Manual