John R Shakeshaft Early CMBR observations at the Mullard Radio Astronomy Observatory

John Shakeshaft is an Emeritus Fellow at St Catharine's College, Cambridge. He served for many years as Editor of Monthly Notices of the Royal Astronomical Society.

At the time of publication of the Penzias and Wilson (1965a) paper, I was a member of staff in the Radio Astronomy Group of the Cavendish Laboratory, the physics department of the University of Cambridge, having been an undergraduate and graduate student at Cambridge, the latter under the inspiring supervision of Martin Ryle. I had had an interest in cosmology and measurements of cosmic radio radiation for over ten years. Indeed my first published scientific paper, in 1954, had the title The Isotropic Component of Cosmic Radio-Frequency Radiation, although I advise readers not to bother to search it out. At that date, low-noise receivers for the microwave range had not yet been developed, so interest was concentrated at lower frequencies. Westerhout and Oort (1951) had shown that the survey of galactic radiation at 100 MHz (or Mc/sec as we called it) by Bolton and Westfold (1951) could be explained by assuming that most of the radiation came from "radio stars" distributed through the Galaxy in the same way as the common Population II stars of types G and K, although it was necessary to add in an isotropic component besides. They suggested three possible explanations for this extra component but found none to be satisfactory. Subsequent to their paper, extragalactic sources much more intense than normal galaxies had been identified, such as the so-called "colliding galaxies" Cygnus A, and I attempted an estimate of the integrated contribution due to these. Interestingly perhaps, in view of later controversies about the number counts of radio sources and their cosmological significance, I concluded that the isotropic component could be accounted for by standard relativistic cosmo-logical models but not by the steady state theory. Shortly after publication, however, the general realization that galactic radio emission is largely due to synchrotron radiation from cosmic ray electrons in the interstellar magnetic field vitiated both the Westerhout and Oort model and my conclusion from it.

Toward the end of that decade I began work, with graduate student Ivan Pauliny-Toth, on a survey of the background radiation at 404 MHz (A = 74 cm). This was intended as part of a study over a range of frequencies to determine the spectrum of the galactic radiation as a function of direction, which could provide information on the dependence of cosmic ray electron density and interstellar magnetic fields on position within the Galaxy. It was therefore important, if brightness temperatures at different frequencies were to be compared, for these temperatures to be absolute values rather than merely relative values in different directions. We used an 8-m diameter dish on an alt-azimuth mount (it was in fact a German radar dish "liberated" after World War II by Martin Ryle, and one of the two dishes that Graham Smith used as an interferometer to determine accurate positions of the sources Cygnus A and Cas A, enabling their optical identifications by Walter Baade and Rudolph Minkowski), an electron beam parametric amplifier and a Dicke-type radiometer with a liquid nitrogen reference source. The published survey (Pauliny-Toth and Shakeshaft 1962) was used later both at Cambridge and elsewhere to correct CMBR measurements for the contribution due to galactic radiation. The experience gained on determining losses in the antenna and connections, and ground radiation in the far-out side lobes, was also helpful in measurements of the CMBR a few years later.

The news in 1965 of the serendipitous result from Holmdel of 3.5-K CMBR at 4080 MHz (the second such major astronomical discovery from there, the first being Karl Jansky's accidental finding of the galactic radio emission in 1931) was received at the Mullard Observatory with great interest but no real surprise, since our work on radio source surveys over the previous 12 years had left us convinced that the universe was evolving and not in a steady state; the discovery of radiation from a big bang therefore fitted naturally with these ideas. Although the data from the 2C survey had been overinterpreted in terms of actual sources, the ingenious P(D) probability analysis by Peter Scheuer (1957) of the deflections D of the interferometer records themselves, without the identification of individual sources, showed conclusively (to us, at least) that the slope of the radio source counts N(>S) was proportional to S_L8, significantly steeper than the N(>S) proportional to S"L5 expected for a uniform Euclidean model, and even more so than the values expected for Friedmann and steady state models. By 1965, the increase in the numbers of actual identifications of distant radio galaxies and quasars had confirmed the excess of sources at large redshifts, and subsequent studies have shown that Scheuer's result for the slope was indeed correct. One of the merits of the steady state theory was said to have been that it gave specific predictions, unlike the Friedmann models, but its proponents seemed very reluctant to accept that these predictions were in conflict with the observations.

I realized that we were in a position fairly easily to check the Holmdel temperature value at a different frequency, namely 1407 MHz (A = 21.3 cm), which would help to determine whether this component of radiation had a thermal spectrum as predicted. With the aid of graduate student Tim Howell, a copper horn of beamsize 13° x 15° was set up inside the 8-m dish mentioned above, itself surrounded by a wire mesh screen, 30 m square, lifted at the edges, so the horn was doubly screened from ground radiation. The horn was connected to a Dicke radiometer, with a termination in liquid helium as the reference source. This consisted of a metal film resistor at the end of a 75-cm length of low-loss coaxial line. The temperature distribution along the line was measured and the effective noise temperature at the upper end calculated to be TL = 5.9 ± 0.2 K. When the leads from the horn and cold load were interchanged in the circuit, the alteration of receiver output gave a direct measure of the temperature difference with no contribution from any asymmetry of the switch and leads. The receiver output was calibrated by measuring the differences between terminations immersed in water at various temperatures.

Observations were made at night with the horn directed toward the zenith (declination 5 = 52°) at a right ascension such that the galactic radio emission was at a minimum. A temperature difference Th — Tl = 0.9 ± 0.1 K was found, implying that Th = 6.8 ± 0.3 K, representing the combined contributions from (a) galactic radiation and the CMBR, (b) atmospheric emission, (c) ground radiation, and (d) losses in the horn and waveguide-coaxial connection. For (b) we took the value of 2.2 K, derived by Dave Hogg for a wavelength of 20.7cm (see below), and assumed an error of ±0.2K. For (c), we measured the polar diagram of the horn and estimated a value less than 0.1 K, and for (d) we calculated a contribution of 1.3 ± 0.2K. The sum of (b), (c), and (d) was 3.5 ± 0.3 K, leading to a value of 3.3 ± 0.5K for the minimum background brightness temperature. The galactic contribution to this was found by convolving the reception pattern of the horn with the brightness temperatures measured in the survey at 74 cm mentioned earlier and scaling the result to 20.7 cm by assuming T to be proportional to A2'7.16 The result was 0.5 ± 0.2 K, leaving a CMBR value of 2.8 ± 0.6 K (Howell and Shakeshaft 1966), which turns out to be gratifyingly - if fortuitously - close to the currently accepted value of 2.725 K.

The atmospheric absorption for frequencies up to 8 GHz is due predominantly to nonresonant absorption by molecular oxygen, and Hogg (1959) had calculated values from 400 MHz up, on the assumption of a line-broadening constant of 0.75 GHz per atmosphere. Our search of the literature for experimental measurements of the absorption by observations of the extinction of extraterrestrial sources as a function of zenith angle had revealed a relatively wide scatter of values, with some in very poor agreement with Hogg's predictions. To throw more light on this problem, we carried out measurements of our own at 408 and 1407 MHz. The interpretation of these involved consideration of the change of apparent angular size of the source in question due to differential refraction in the atmosphere. After applying the necessary corrections, our results fitted well with Hogg's curve, but we then realized that some of the earlier workers had either not applied the refraction correction or applied it with the wrong sign. Judicious reworking of the earlier results, where necessary, then produced a satisfactory agreement between theory and experiment (Howell and Shakeshaft 1967a).

On completion of the initial measurement of the CMBR at 1407 MHz, we tried to check whether this component of the cosmic radiation could be detected at the lower frequencies of 408 and 610 MHz, although the dominance of galactic radiation in this range would cause increased uncertainties. Studies by Peter Scheuer (1975)17 and by Ray Weymann (1966) had suggested that deviations from a blackbody spectrum might be present at low frequencies. We used optimal scaled horns with beamwidths of

16 To my embarrassment, Jim Peebles, in reviewing this piece, has noticed that the actual wavelength corresponding to a frequency of 1407 MHz is of course 21.3 cm rather than 20.7 cm, as appeared in the original paper and as I unthinkingly copied above. He is the first person in the last 40 years to have pointed out this blunder to me. At this late date I do not have the original working material available and so cannot determine whether the quoted temperature of 2'8 ±0'6K might require modification. Any such change would only be in the second decimal place, already omitted due to the size of the error.

17 Presented in an article written in 1965 for Galaxies and the Universe, and eventually published in revised form in 1975.

15°, screened from ground radiation by wire mesh, and Dicke radiometers with liquid helium reference sources as before. After applying corrections for other contributions as at 1407 MHz, the effective brightness temperatures from the region of the celestial North Pole were 24.3 ± 0.9K at 408 MHz and 10.4 ± 0.7K at 610 MHz, the ratio of these, 2.3 ± 0.2, being significantly less than the ratio of 3.1 ± 0.1 expected for the galactic contribution with a temperature spectral index of —2.8 (T rc A2'8). This implied that there was indeed an extra component of radiation characterized by a temperature close to independent of wavelength, that is, a blackbody spectrum. Further analysis indicated that, if the spectrum of this component were blackbody, the excess temperature would be 3.7 ± 1.2 K (Howell and Shakeshaft 1967b). Unfortunately, the error in this value was such that no new upper limit could be put on the epoch of ionization of the intergalactic gas.

This work concluded for over 20 years observational studies at the Mullard Radio Astronomy Observatory of the CMBR, since other groups much better equipped for work at high frequencies had vigorously entered the field, but they were subsequently taken up again with the building by Paul Scott and others of the Cosmic Anisotropy Telescope (CAT), a three-element interferometer which, in 1996, was the first telescope to detect structure in the CMBR on angular scales smaller than the main peak in the angular spectrum (Scott et al. 1996). This was followed by the Very Small Array (VSA), now observing from Tenerife the anisotropies on angular scales between 15arcmin and 2°, and the Arcminute MicroKelvin Imager (AMI) to study the Sunyaev-Zel'dovich effect in high-redshift clusters and proto-clusters of galaxies. In addition to this observational work, there has been theoretical modeling of background fluctuations, and the Cambridge Planck Analysis Centre has been set up in preparation for the launch in 2009 of the European Planck Surveyor satellite.

Other authors in this volume have noted that, if the attention of obser-vationalists had been drawn to the matter, the CMBR could perhaps have been detected (or recognized as such) years earlier than in fact it was. It is, for example, unfortunate that in neither of the two editions (1952 and 1960a) of his influential textbook Cosmology did Hermann Bondi refer to the possibility, nor did Fred Hoyle (1959b) in his paper The Relation of Radio Astronomy to Cosmology at the Paris Symposium on Radio Astronomy. We must hope that sufficient of the astronomical literature is now available on the World Wide Web for rapid searches which could prevent oversights of this kind in the future.

Telescopes Mastery

Telescopes Mastery

Through this ebook, you are going to learn what you will need to know all about the telescopes that can provide a fun and rewarding hobby for you and your family!

Get My Free Ebook


Post a comment