Primordial Deuterium and the Big Bang

Nuclei of this hydrogen isotope formed in the ﬁrst moments of the big bang. Their abundance offers clues to the early evolution of the universe and the nature of cosmic dark matter

by Craig J. Hogan

he big bang model of the early ticles such as neutrons and protons con- markably well with observations. Be- universe is extraordinarily sim- stantly changed back and forth into one cause hydrogen is the principal fuel of Tple: it has no structure of any another, “cooked” by interactions with the stars of the universe, its predomi- kind on scales larger than individual el- plentiful and energetic electrons, posi- nance is the basic reason for starlight ementary particles. Even though the be- trons and neutrinos. Neutrons are slight- and sunlight. havior it predicts is governed only by ly heavier than protons, however; as During the formation of helium nu- general relativity, the Standard Model things cooled, most of the matter settled clei, perhaps only one in 10,000 deuter- of elementary particle physics and the into the more stable form of protons. As ons remained unpaired. An even smaller energy distribution rules of basic ther- a result, when the temperature fell below fraction fused into nuclei heavier than modynamics, it appears to describe the 10 billion degrees and the intertransmu- primordial fireball almost perfectly. tation stopped, there were about seven Atomic nuclei that formed during the times as many protons as neutrons. first seconds and minutes of the universe provide additional clues to events Out of the Primordial Furnace in the early universe and to its composition and structure today. The big bang hen the universe was a few min- produced a universe made almost en- Wutes old (at a temperature of tirely of hydrogen and helium. Deuteri- about one billion degrees), the protons um, the heavy isotope of hydrogen, was and neutrons cooled down enough to made only at the beginning of the uni- stick together into nuclei. Each neutron verse; thus, it serves as a particularly im- found a proton partner, creating a pair portant marker. The ratio of deuterium called a deuteron, and almost all the ) to ordinary hydrogen atoms depends deuterons in turn stuck together into strongly on both the uniformity of mat- helium nuclei, which contain two pro- ter and the total amount of matter tons and two neutrons. By the time pri- formed in the big bang. During the past mordial helium had formed, the density enhancement color few years, astronomers have for the first of the universe was too low to permit time begun to make reliable, direct mea- further fusion to form heavier elements surements of deuterium in ancient gas in the time available; consequently, al- clouds. Their results promise to provide most all the neutrons were incorporat- a precise test of the big bang cosmogony. ed into helium. The expansion of the universe appears Without neutrons to hold them to- to have started between 10 and 20 bil- gether, protons cannot bind into nuclei lion years ago. Everything was much because of their electrical repulsion. Be- closer together and much denser and cause of the limited neutron supply in ( GRACE OF CRAIG J. HOGAN; LAURIE COURTESY hotter than it is now. When the universe the primordial fireball, six of every sev- KECK TELESCOPE (right) on Mauna Kea, Ha- was only one second old, its temperature en protons must therefore remain as waii, gathered light from a distant quasar and was more than 10 billion degrees, 1,000 isolated hydrogen nuclei. Consequently, concentrated it on the photodetector of a high- resolution spectroscope. The resulting bands of times hotter than the center of the sun. the big bang model predicts that about color (above) are marked by dark lines where in- At that temperature, the distinctions be- one quarter of the mass of the normal tervening gases have absorbed light of specific tween different kinds of matter and en- matter of the universe is made of helium wavelengths. Analysis of the characteristic line ergy were not as definite as they are un- and the other three quarters of hydro- patterns for hydrogen gas can reveal the presence der current conditions: subatomic par- gen. This simple prediction accords re- of the heavy isotope of the element deuterium.

68 Scientific American December 1996 Copyright 1996 Scientific American, Inc. Primordial Deuterium and the Big Bang ROGER RESSMEYER Corbis Copyright 1996Scientific American,Inc. helium, such as lithium. (All the other familiar elements, such as carbon and oxygen, were produced much later inside stars.) The exact percentages of helium, deuterium and lithium depend on only one parameter: the ratio of protons and neutrons—particles jointly cat- egorized as baryons—to photons. The value of this ratio, known as η (the Greek letter eta), remains essentially constant as the universe expands; because we can measure the number of photons, knowing η tells us how much matter there is. This number is impor- tant for understanding the later evolution of the universe, because it can be compared with the actual amount of matter seen in stars and gas in galaxies, as well as the larger amount of unseen dark matter. For the big bang to make the observed mix of light elements, η must be very small. The universe contains fewer than one baryon per billion photons. The temperature of the cosmic background radiation tells us directly the number of photons left over from the big bang; at present, there are about 411 photons per cubic centimeter of space. Hence, baryons should occur at a density of somewhat less than 0.4 per cubic meter. Although cosmologists know that η is

) small, estimates of its exact value cur- rently vary by a factor of almost 10. The painting most precise and reliable indicators of η are the concentrations of primordial AMAJIAN (

. K light elements, in particular deuterium. A ﬁvefold increase in η, for example,

); ALFRED T would lead to a telltale 13-fold decrease

inset in the amount of deuterium created. The mere presence of deuterium sets an upper limit on η because the big bang

AIG J. HOGAN ( is probably the primary source of deuterium in the universe, and later

TESY OF CR processing in stars gradually OUR

C destroys it. One can think

of deuterium as a 9 BILLION kind of partially spent fuel YEARS AGO like charcoal, left over because (APPROXIMATE) there was originally not time for all of it to burn completely to ash before the ﬁre cooled. Nucleosynthesis in the big bang lasted only a few minutes, but the nuclear burning in stars lasts for millions or billions of years; as a result, 9.5 BILLION any deuterium there is converted to he- YEARS AGO (APPROXIMATE) lium or heavier elements. All the deuterium that we ﬁnd must therefore be a

Copyright 1996 Scientific American, Inc. Primordial Deuterium and the Big Bang MEASURING THE EARLY MAKEUP of the universe is complicated because so much matter has been transmuted inside stars. Nevertheless, radiation from quasars several billions of light-years distant, at the edge of the observable universe, offers one method. Long ago this light passed through clouds of fairly pure primordial gas, possibly in the outskirts of a forming galaxy (a computer model of one primordial gas cloud is shown in the inset at the left). Hydrogen and PRESENT deuterium in such clouds remove characteristic wavelengths of this light; these changes can be detected and measured on the earth.

called because of the dark line it leaves in a spectrum) corre- sponds to the wavelength of a photon exactly energetic enough to excite the electron in a hydrogen atom to a particular energy level. These lines have colors that lie deep in the ultraviolet and cannot usually be seen from the ground because of atmospheric absorption; even times. As a result, looking the reddest (and most prominent) line, at nearby gas clouds can sug- Lyman alpha, appears at a wavelength gest only a lower limit to primor- of 1,215 angstroms. Luckily, the expan- dial deuterium abundance. sion of the universe causes a “cosmo- It would be much better if one could logical redshift” that lengthens the wave- get hold of some truly pristine primor- lengths of photons that reach the earth dial material that had never undergone to the point where hydrogen absorption chemical evolution. Although we can- lines from sufﬁciently distant gas clouds not bring such matter into the laborato- reside comfortably within the visible remnant of the big ry, we can look at its composition by its range. bang—even the one mole- effect on the spectrum of light from dis- Lyman alpha appears in light from a cule in 10,000 of seawater that tant sources. Bright quasars, the most typical quasar hundreds of times, each contains a deuterium atom in place luminous objects in the universe, are so time from a different cloud along the line of a hydrogen atom. far away that the light we see now left of sight at a different redshift and there- them when the universe was only one fore at a different wavelength. The re- Quasars and Gas Clouds sixth to one quarter of its present size sulting spectrum is a slice of cosmic his- and perhaps only a tenth of its present tory, like a tree-ring sample or a Green- etermining the primordial ratio of age. On its way to us, the light from land ice core: these quasar absorption Ddeuterium to ordinary hydrogen these quasars passes through clouds of spectra record the history of the con- should be highly informative, but it is gas that have not yet condensed into version of uniform gas from the early not easy, because the universe is not as mature galaxies, and the light absorbed big bang into the discrete galaxies we simple as it used to be. Astronomers can by these clouds gives clues to their com- see today over an enormous volume of measure deuterium in clouds of atomic position. Some of the clouds that have space. This multiplicity of spectra offers hydrogen gas between the stars of our been detected contain less than one thou- another way to test the primordial char- galaxy, but the element’s fragility ren- sandth the proportion of carbon and acter of the absorbing material: the big ders the results suspect. We live in a pol- silicon (both stellar fusion products) seen bang model predicts that all gas clouds luted, dissipated, middle-aged galaxy in nearby space, a good sign that they from the early universe should have whose gases have undergone a great retain very nearly the composition they more or less the same composition. Mea- deal of chemical processing over its 10- had immediately after the big bang. suring the abundances of different clouds billion-year history. Deuterium is very There is another advantage to looking at vast distances from us and from one readily destroyed in stars, even in their so far away. The main component of another in both time and space will di- outer layers and their early prestellar these clouds, atomic hydrogen, absorbs rectly test cosmic uniformity. evolution. Stars eject their envelopes light at a sharply deﬁned set of ultravio- In some of these clouds, we can de- when they die, and the gas in our gal- let wavelengths known as the Lyman se- termine from the quasar spectra both axy has been in and out of stars many ries. Each of these absorption lines (so how much ordinary hydrogen there is

Primordial Deuterium and the Big Bang Copyright 1996 Scientific American, Inc. Scientific American December 1996 71 a b c d ORPOLE IAN W NUCLEOSYNTHESIS, the formation of atomic nuclei, started form deuterons, but because the former outnumber the latter, instants after the big bang, as the universe cooled, when the fun- most of the protons remained alone and became hydrogen nu- damental particles called free quarks (a) condensed into protons clei (c). Almost all the deuterons in turn combined to form heli- and neutrons (b). Protons (red) and neutrons (blue) paired off to um nuclei (d), leaving a tiny remnant to be detected today. and how much deuterium. We can sep- 20 years, these observations proved too smaller telescopes, my colleagues An- arate the signal from deuterium because difficult. Many of us have spent long toinette Songaila and Lennox L. Cowie the added mass in the deuterium nucleus nights waiting for photons to drip one of the University of Hawaii were allocat- increases the energy required for atomic by one onto the detectors of the world’s ed their university’s first science night on transitions by about one part in 4,000 largest telescopes, only to find that the the Keck Telescope for this project in (twice the ratio of a proton’s mass to an weather, instrument problems and, ulti- November 1993. They trained the tele- electron’s mass). As a result, the absorp- mately, just lack of time had prevented scope on a quasar known as 0014+813, tion spectrum of deuterium is similar to the accumulation of enough light for a famous among astronomers for its that of single-nucleon hydrogen, but all convincing result. The technique is now brightness—indeed, it was for some years the lines show a shift toward the blue practical only because of improved, more the brightest single object known in the end of the spectrum equivalent to that efficient detectors, the 10-meter Keck universe. From earlier studies by Ray J. arising from a motion of 82 kilometers telescope in Hawaii and advanced high- Weymann of the Observatories of the per second toward the observer. In resolution, high-throughput spectro- Carnegie Institution of Washington and spectrographic measurements of a hy- graphs such as the Keck HIRES. Frederic Chaffee, Craig B. Foltz and Jill drogen cloud, deuterium registers as a After many unsuccessful attempts on Bechtold of the University of Arizona faint blue-shifted “echo” of the hydrogen. These spectra also record the INCOMING LIGHT velocity and temperature distribution of the atoms. Atoms PRIMARY traveling at different velocities GRATING absorb light at slightly different FOCUSING wavelengths because of the MIRROR Doppler effect, which alters the ap- parent wavelength of light according to the relative motion of transmitter and SILICON receiver. Random thermal motions im- PHOTODETECTOR pel the hydrogen atoms at speeds of CORRECTIVE LENSES about 10 kilometers per second, causing a wavelength shift of one part in 30,000; COLLIMATOR because they are twice as heavy, deuterium atoms at the same temperature move at only about seven kilometers per second and therefore have a slightly different velocity distribution. A modern spec- trograph can resolve these thermal velocity differences, as well as larger-scale collective flows. SECONDARY GRATING Waiting for the Light

lthough spectrographs can easily re- SPECTROSCOPE attached to the 10-meter Keck Tele- scope can distinguish 30,000 different colors. Two op- A solve the wavelength differences tical gratings, acting as prisms, spreads out the light. between ordinary hydrogen and deuter- Additional components focus the beam on a silicon ium, splitting the light of a distant qua- wafer a few centimeters square to produce an image sar into 30,000 colors leaves very little like that on page 68. The wafer contains four million ORPOLE intensity in each color. For more than tiny photodetectors, each only 20 microns on a side. IAN W

72 Scientific American December 1996 Copyright 1996 Scientific American, Inc. Primordial Deuterium and the Big Bang and their collaborators, we knew that a San Diego and Xiao-Ming Fan of Co- ) fairly pristine gas cloud lay in front of lumbia University, who have found a ra- ement this quasar. tio that is apparently almost a factor of The ﬁrst Keck spectrum, obtained in 10 lower than our estimate. It remains olor enhanc only a few hours, was already of sufﬁ- to be seen whether their result represents c ciently high quality to show plausible the true primordial value. The lower signs of cosmic deuterium. That spec- abundance might be a result of deuteri- ORPOLE ( ; IAN W trum showed the absorption pattern for um burning in early stars or a sign that Y OR T hydrogen gas moving at various veloci- the production of deuterium was per- A V ties, and it showed an almost perfect haps not as uniform as the big bang

echo of the Lyman alpha line with the model predicts. AL OBSER V characteristic blueshift of deuterium. The amount of absorption in this sec- Clues to Dark Matter ond signal would be caused by about

two atoms of deuterium per 10,000 f our higher value is correct, the TESY OF THE U.S. NA

atoms of hydrogen. The result has since amount of primordial deuterium OUR I C been independently confirmed by Rob- would fit very well with the standard ert F. Carswell of the University of Cam- predictions of the big bang model for a QUASAR, or quasistellar object, 0014+ bridge and his colleagues, using data value of η around two baryons per 10 813 is one of the brightest objects known to exist in the cosmos. It appears here in a from the four-meter Mayall Telescope billion photons. With this value of η, radiotelescope image. The light from this at the Kitt Peak National Observatory the big bang predictions are also consis- supermassive black hole at the center of a in Arizona. Subsequent analysis has re- tent with the amounts of lithium in the very young galaxy near the edge of the vealed that the deuterium absorption oldest stars and estimates of primordial observable universe provided the first indeed displays an unusually narrow helium seen in nearby metal-poor gal- measurements of primordial deuterium. thermal spread of velocities, as expected. axies. Confirmation of this result would It is possible that some of the absorp- be fabulous news. It would verify that tion we saw was caused by a chance in- cosmologists understand what happened 10 times the mean density of the visible terposition of a small hydrogen cloud only one second after the beginning of baryons. Thus, our high deuterium abun- that just happens to be receding from us the expansion of the universe. In addi- dance indicates that this mass is not at 82 fewer kilometers per second than tion it would indicate that the history made of ordinary atomic matter. the main cloud we observed. In that of matter at great distances is like that of Cosmologists have proposed many case, the deuterium abundance would nearby matter, as assumed in the sim- candidates for such nonbaryonic forms be less than we think. Although the a plest possible model of the universe. of dark matter. For example, the big priori chance of such a coincidence on This estimate of η fits reasonably well bang predicts that the universe has al- the first try is small, we ought to regard with the number of baryons we actually most as many neutrinos left over as this estimate as only preliminary. Nev- see in the universe today. The observed photons. If each one had even a few bil- ertheless, the effectiveness of the tech- density of photons calls for about one lionths as much mass as a proton (equiv- nique is clear. Absorbing clouds in front atom for every 10 cubic meters of space. alent to a few electron volts), neutrinos of many other quasars can be studied This is about the same as the number of would contribute to the universe rough- with the new technology; we will soon atoms counted directly by adding up all ly as much mass as all the baryons put have a statistical sampling of deuterium the matter in the known gas, stars, together. It is also possible that the early in primordial material. In fact, our planets and dust, including the quasar universe created some kind of leftover group and others have now published absorbers themselves; there is not a huge particle that we have not been able to measurements and limits for eight dif- reservoir of unseen baryons. At the same produce in the laboratory. Either way, ferent clouds. time, observations suggest that an enor- the big bang model, anchored by obser- One of the most intriguing results is a mous quantity of dark matter is neces- vation, provides a framework for pre- measurement by David Tytler and Scott sary to explain the gravitational behav- dicting the astrophysical consequences Burles of the University of California at ior of galaxies and their halos—at least of such new physical ideas. SA

The Author Further Reading

CRAIG J. HOGAN studies the edge of the visible The First Three Minutes. Steven Weinberg. Basic Books, 1977. universe. He is chair of the astronomy department The Physical Universe: An Introduction to Astronomy. F. H. Shu. Universi- and professor in the departments of physics and as- ty Science Books, Mill Valley, Calif., 1982. tronomy at the University of Washington. Hogan A Short History of the Universe. J. Silk. W. H. Freeman and Company, 1994. grew up in Los Angeles and received his A.B. from Deuterium Abundance and Background Radiation Temperature in High- Harvard College in 1976 and his Ph.D. from the Redshift Primordial Clouds. A. Songaila, L. L. Cowie, C. J. Hogan and M. University of Cambridge in 1980. After postdoctoral Rugers in Nature, Vol. 368, pages 599–604; April 14, 1994. fellowships at the University of Chicago and the Cal- Cosmic Abundances. ASP Conference Series, Vol. 99. Edited by S. S. Holt and G. ifornia Institute of Technology, he joined the faculty Sonneborn. Astronomical Society of the Paciﬁc, 1996. of Steward Observatory at the University of Arizona The History of the Galaxies. M. Fukugita, C. J. Hogan and P.J.E. Peebles in for ﬁve years. He moved to Seattle in 1990. Nature, Vol. 381, pages 489–495; June 6, 1996.