Cosmic Background Explorer

From Scholarpedia
John C. Mather and Gary F Hinshaw (2008), Scholarpedia, 3(3):4732. doi:10.4249/scholarpedia.4732 revision #135529 [link to/cite this article]
(Redirected from Cosmic background explorer (COBE))
Jump to: navigation, search
Post-publication activity

Curator: Gary F Hinshaw

Figure 1: COBE concept. COBE was launched in 1989 to observe the cosmic microwave background radiation from the Big Bang.

The COsmic Background Explorer (COBE) was a NASA space mission designed to test the Big Bang theory of the origin of the universe by measuring the spectrum the cosmic microwave background radiation (CMB) and mapping its distribution across the sky, and to search for the infrared and submillimeter background light, the possible faint diffuse emission from the first generations of stars and galaxies. The official NASA web site is [1] which gives images, explanations, and data archives.


Discovery and importance of the CMB

The CMB was discovered by Penzias and Wilson in 1965 [2] , and the discovery was recognized by the Nobel Prize in Physics in 1978 [3] . It is the radiant energy of the Big Bang itself, cooled and diluted by the expansion of the universe, but it still fills the universe and comes to us almost uniformly from every direction. The measurements made by the COBE satellite confirmed this picture in every detail: the radiation has the spectrum of a blackbody (a perfect emitter as well as a perfect absorber) within 50 parts per million (rms), it has a temperature of 2.725 +/- 0.001 K, and it has the same brightness in every direction within 10 parts per million (rms), after accounting for the velocity of the Earth through the cosmos. The small spatial variations (anisotropies) have no preferred angular size (as measured with a 7 degree angular resolution), as predicted on general grounds by Peebles and Yu (1970) ; Harrison (1970); Zeldovich (1972) , and as predicted by the inflation theory of the early universe (Guth 1981).

COBE launch and mission design

The COBE was launched from Vandenberg Air Force Base in Lompoc, California at dawn on November 18, 1989, 15 years after it was first proposed (1974) and 7 years after its construction was approved by NASA (1982). It was described by Boggess et al. 1982. The mission was originally proposed for launch on a Delta rocket, but NASA was then directed to use the Space Shuttle for launch. Following the loss of the Space Shuttle Challenger in 1986, the COBE was redesigned again for launch on the Delta rocket. The mission science definition team was appointed in 1976 by NASA. The COBE is illustrated in Figure 1 and its orbit in Figure 2.

Figure 2: COBE orbit altitude was 900 km, and orbit plane was roughly perpendicular to the Sun.

The COBE carried three instruments, the Far Infrared Absolute Spectrophotometer (FIRAS), the Differential Microwave Radiometers (DMR), and the Diffuse Infrared Background Experiment (DIRBE). The instrument package was protected by a conical shield to exclude the intense light and heat from the Sun and Earth, as well as radiation in the bands to which the COBE was sensitive. The COBE was launched by a Delta 2 rocket directly into a circular polar sun-synchronous orbit, 900 km above the Earth. The orbit plane was inclined 99 degrees to the Earth’s equator, and the Earth’s equatorial bulge caused the orbit plane to precess at the rate of one revolution per year. The time of launch was chosen so that the orbit plane was nearly perpendicular to the line to the Sun. In this orbit, the spacecraft can be oriented to always point away from the Earth and approximately perpendicular to the Sun, so that the Sun never illuminates the instruments, and the Earth limb was limited to shining on the instrument package for a maximum of only 20 minutes per orbit for 3 months per year, and not at all otherwise. A higher orbit would have provided more protection from the Earth’s heat but greater exposure to cosmic rays from the Van Allen belts. The spacecraft pointing control system of gyros, magnetometers, reaction wheels, and magnetic torquer bars kept the spacecraft spinning at 1 revolution per 72 seconds.

The entire mission design was aimed at controlling or preventing systematic errors, which had plagued earlier measurements. As the CMB and IR backgrounds are very faint compared to the local environment, rejecting environmental effects was critical. The orbit choice and the conical shield were essential to protect the instruments from the Sun and the Earth. Both the DMR and DIRBE instruments had lines of sight inclined 30 degrees to the spin axis so that they could map large portions of the sky very rapidly, and so that any dependence of the sky brightness on the angle from the Sun could explicitly determined.


The primary objective of the FIRAS instrument was to determine whether the CMB has the predicted blackbody spectrum, which is the result of rapid equilibrium processes in the universe, but could not be readily produced by any other known process. In particular, it would be difficult to explain a precise blackbody spectrum in the steady state theory of the universe, or if the Big Bang were cold instead of hot. The FIRAS provided a null, differential measurement to provide the extremely high precision and accuracy desired. In addition, the FIRAS was designed with its own Big Bang spectrum simulator, a very precise (parts per million) blackbody radiator that could be inserted into the aperture of the instrument for direct comparison with the sky. In addition, the FIRAS was designed as a polarizing Michelson interferometric spectrometer, an intrinsically differential instrument with a place for a second blackbody input to reduce the dynamic range of the measurement. Albert Michelson won the Nobel Prize in Physics in 1907 for his original form of the interferometer. [4]

Figure 3: FIRAS was a symmetrical polarizing Michelson interferometer used as a spectrophotometer.

The FIRAS concept is illustrated in Figure 3. The entire instrument was kept cold in a vacuum space surrounded by liquid helium, at a temperature of 1.5 K, so that emission from the instrument was much less intense than the CMB. Light was collected from the sky by a compound parabolic concentrator, a non-imaging optical device that accepts light from angles up to 3.5 degrees off its axis, which was parallel to the spacecraft spin axis. The light was then expanded and collimated again to pass through the polarizing spectrometer. After the light was initially polarized, it was passed to a polarizing beamsplitter, and divided into two equal parts. The two parts were recombined with a relative delay, their polarization analyzed, and the output directed to the detectors. The delay between the two beams was varied by sweeping two mirrors back and forth on a parallelogram linkage, and the intensity reaching the detectors was measured, digitized, and sent to the ground for analysis. The intensity as a function of delay is called an interferograms. Millions of interferograms were collected and analyzed in a least-squares fitting program to determine the calibration of the instrument and to make maps of the sky at all the wavelengths measured by the FIRAS (from 100 micrometers to 1 cm).

Figure 4: FIRAS spectrum shown to the Astronomical Society with 1% error bars. Final result had error bars of 50 parts per million.

The first scientific result was a spectrum based on 9 minutes of good sky data and a single calibration. The graph shown in Figure 4 was presented to the American Astronomical Society in January 1990 and received a standing ovation (Mather et al. 1990). The blackbody curve is the smooth line and the measurements are the boxes, all of which lie on the predicted curve. When the full data analysis was complete, the errors were reduced to 50 parts per million of the peak intensity. The spectrum of the Big Bang is therefore indistinguishable from that of a blackbody, implying that less than 1 part in 10,000 of the energy in the CMB could have been released more than 1 year after the Big Bang, as summarized by Wright et al. (1994). The FIRAS also observed atomic and molecular emissions and dust from our Galaxy. The final measured temperature of the CMB is 2.725 +/- 0.001 K (Fixsen and Mather 2002)


The purpose of the DMR instrument was to make maps of the CMB. If the radiation is indeed cosmic, it must come almost equally from all directions. There should be one significant effect due to the motion of the Earth relative to the source of the radiation, a Doppler shift that raises the temperature slightly in the direction towards which the Earth is moving, and reduces it slightly in the opposite direction. And although this was not known when the COBE was proposed, there should also be small random variations in the CMB temperature from one direction to another, called anisotropy, to explain the existence of the clusters of galaxies. Theoretically, if the early universe were exactly uniform, there would be no reason why matter should have been able to clump into objects, locally reversing the expansion. So, some initial seeds are apparently required.

Figure 5: Differential Microwave Radiometer concept. A Dicke switch connects the receiver alternately to the two antennas. The output measures the difference in temperature of two directions on the sky.

The DMR concept is illustrated in Figure 5. Each channel used a Dicke-switched microwave radiometer, with a single microwave receiver connected to two antennas (each with a 7 degree beamwidth, with lines of sight 30 degrees from the COBE spin axis) through a polarizing ferrite switch. As the switch was connected alternately to the two antennas at 100 Hz, the output of the receiver changed slightly at the same rate, and the small change was a measure of the difference in brightness seen by the two antennas. The DMR was designed to measure at three different frequencies (31.4, 53, and 90 GHz), with two receivers at each frequency. The 31.4 GHz receivers were warmed to near room temperature, while the other two were cooled down to about 140 K for improved sensitivity. The reason for measuring at three frequencies is to be able to recognize the patterns of emission from the electrons in our own galaxy, which collide with protons to produce brehmsstrahlung radiation, and orbit in the galactic magnetic field to produce synchrotron radiation. Both types are concentrated in the plane of the Milky Way galaxy, and the intensity diminishes rapidly as the frequency increases. There is also radiation from dust grains in the interstellar medium of the Milky Way, which increases rapidly with frequency.

Figure 6: Map of temperature fluctuations obtained from 2 years of DMR observations. Fluctuations are about 1 part in 100,000 and represent density variations in the early universe.

Maps were produced from hundreds of millions of measurements of differences of brightness, as the COBE spin and orbit swept the antenna beams rapidly around the sky. A very complex least-squares fitting program was used to model the systematic errors, determine calibration constants, and produce sky maps. A map of the radiation measured by the COBE is shown in Figure 6, after careful subtraction of the dipole anisotropy due to the motion of the Earth, and the dust and electron emission of our own Galaxy. When the first DMR maps were released, Steven Hawking wrote that this was the most important scientific result of the century, if not of all time. The instrument, calibration, correction for foreground emissions, and the interpretation are described in a set of papers summarized by Bennett et al. (1996) [5]. Since that time, thousands of other papers have been written interpreting these maps.

The significance of these maps was that for the first time, it was possible to understand how we come to exist, and how galaxies could form after the Big Bang. Moreover, the inflation theory predicted that the patterns of hot and cold should have the random nature seen in these maps.


The third instrument on the COBE was the DIRBE. Its purpose was to search for the diffuse infrared and submillimeter background light from the first generations of stars and galaxies. This diffuse background might be released in the early universe by objects that would never be visible with telescopes because they are individually too faint. There were serious predictions before the COBE was proposed that such light might exist (Partridge and Peebles 1967)) but there was no observational evidence for it.

Figure 7: The DIRBE was a small telescope to measure the cosmic infrared background radiation. It observed a single pixel of the sky at a time, at 10 wavelengths from 1.25 to 240 micrometers.

The DIRBE instrument concept is illustrated in Figure 7. Unlike the other two instruments, it was a small telescope specially designed for strong rejection of stray light, with a 20 cm aperture, but because the predicted light is diffuse, the DIRBE had a very modest angular resolution of 0.7 degrees. It observed simultaneously at 10 wavelengths (1.25, 2.25, 3.5, 4.9, 12, 25, 60, 100, 140, and 240 micrometers, and at the three shortest wavelengths it also measured polarization. The DIRBE line of sight was inclined 30 degrees to the spin axis, in order to observe half the sky every day and monitor the apparent seasonal variation of the zodiacal light, the diffuse light emitted by interplanetary dust. This variation was an essential signature of the zodiacal light and was used to facilitate removing that signal from the DIRBE maps.

The main difficulty faced by the DIRBE team was to measure and account for the effects of the diffuse foreground light from dust in our own solar system, and from stars and dust in the interstellar medium of our Milky Way galaxy. The DIRBE results showed that indeed the universe is much brighter than had been expected. As illustrated in Figure 8, there are two parts of the diffuse light of the universe, one at near infrared wavelengths (a few micrometers) and one at far infrared wavelengths (50-300 micrometers).

Figure 8: Cosmic IR background spectrum. Cosmic background intensity Iν times frequency ν as a function of wavelength λ. The circles with error bars are the detections based on DIRBE data after removal of foreground emission at 140 and 240 μm, and those with arrows are 2 σ upper limits with the arrows extending to the measured residuals at 1.25–100 μm. The hatched thick lines are dark sky limits (95% CL) from the DIRBE data at 1.25–240 μm, and the hatched thin lines are dark sky “broad band” limits (95% CL) from FIRAS data at 120–650 μm (Shafer et al. 1998). The crosses are upper limits derived from rocket experiments at 134–186 μm (Kawada et al. 1994) and 2.5–4.0 μm (Matsuura et al. 1994). The dashed line from 1.4–2.6 μm is residual radiation after foreground removal from the rocket data of Noda et al. (1992). The diamonds with arrows are lower limits derived from IRAS counts at 25–100 μm (Hacking & Soifer 1991; 60 μm limit from Gregorich et al. 1995). The dotted curve from 170–1260 μm shows the tentative infrared background determined from FIRAS data by Puget et al. (1996), and the solid curve is the average of the two DIRBE-independent methods of FIRAS analysis used by Fixsen et al. (1998). The triangles are lower limits derived from the Hubble Deep Field at 3600–8100 Å (Pozzetti et al. 1998) and K-band galaxy counts at 2.2 μm (Cowie et al. 1994). The square is an upper limit derived from sky photometry at 4400 Å (Mattila 1990).

From measurements of distant galaxies subsequent to the COBE mission, we now know that the brightness of the NIR and FIR backgrounds is about equal to that of all the galaxies that have ever been cataloged. In particular, the DIRBE discovered a stronger than expected far-infrared cosmic background (140-300 micrometers), implying a surprisingly dusty universe at early times. Furthermore, FIRAS data extended that discovery, providing the first evidence for a strong submillimeter background radiation shortward of the CMB. The task of understanding the sources of this diffuse radiation is still continuing. In particular, a new population of galaxies that are very bright at far IR wavelengths has been discovered, both with space observatories (IRAS, the Infrared Astronomical Satellite, ISO, the Infrared Space Observatory, and the Spitzer Space Telescope), and with ground-based observatories. The detailed results are presented in a series of four papers in 1998: [6], [7] [8] [9].

COBE Science Team

The COBE Mission Definition Science Team, appointed in 1976, included Samuel Gulkis, Michael G. Hauser, John C. Mather, George F. Smoot, Rainer Weiss, and David T. Wilkinson. Hauser was Principal Investigator (PI) for the DIRBE, Mather the PI for the FIRAS, and Smoot the PI for DMR. Mather was the NASA Project Scientist and Weiss was the Chairman of the Science Team. The Science Working Group also included Charles L. Bennett (Deputy PI for DMR), Nancy W. Boggess, Edward S. Cheng, Eli Dwek, Michael Janssen, Thomas Kelsall (Deputy PI for DIRBE), Stephan S. Meyer, S. Harvey Moseley, Tom Murdock, Richard A. Shafer (Deputy PI for FIRAS), Robert F. Silverberg, and Edward L. Wright (Data Team Leader). Wright was the first person to find the CMB anisotropy in the data and bring it to the attention of the science team. The NASA project manager was Roger Mattson and the Deputy was Dennis McCarthy.

In 2006, John Mather and the COBE team received the Gruber Cosmology Prize, and the results of the FIRAS and DMR instruments were recognized by the Nobel Prize in Physics, given to John C. Mather and George F. Smoot.

COBE Data Archives

The COBE data and related images are available on-line at the Legacy Archive for Microwave Background Data Analysis (LAMBDA), [10].


  • Hauser, M. and Dwek, E. 2001. The Cosmic Infrared Background: Measurements and Implications, ARAA (Annual Reviews of Astronomy and Astrophysics), 39, 249.

Internal references

Recommended Reading

  • Chown, M. (1996) Afterglow of Creation: From the Fireball to the Discovery of Cosmic Ripples. University Science Books
  • Guth, A. (1998) The Inflationary Universe: The Quest for a New Theory of Cosmic Origins. Perseus Books Group
  • Lemonick, M. (2003), Echo of the Big Bang. Princeton University Press
  • Mather, J. and Boslough, J. (1996 and 2008) The Very First Light. Basic Books

External Links

See Also

Wilkinson microwave anisotropy probe, Cosmic microwave background radiation, Galactic magnetic field

Personal tools

Focal areas