Winter 1992 Vol

-- -. CONTENTS

4000 00010

WINTER 1992 VOL. 22, NUMBER 4 FEATURES Editors RENE DONALDSON, BILL KIRK 1COSMOLOGY AFTER COBE What ripples in the cosmic background Contributing Editor radiation imply about the origin, MICHAEL RIORDAN composition, and evolution of the Universe. Editorial Advisory Board Joel R. Primack JAMES BJORKEN, ROBERT N. CAHN, DAVID HITLIN STEWART C. LOKEN, RONALD RUTH 8 PARTICLE PHYSICS-WHERE DO WE GO MARVIN WEINSTEIN, HERMAN WINICK FROM HERE? Talk given at the 75th Anniversary Celebration of the Max-Planck Photographic Services Institute for Physics, Munich. TOM NAKASHIMA BETTE REED James D. Bjorken 15 THE 17-keV NEUTRINO-A EULOGY Illustrations A summary of the odd life story of this TERRY ANDERSON, KEVIN JOHNSTON elusive particle. SYLVIA MACBRIDE, JIM WAHL Hamish Robertson Distribution CRYSTAL TILGHMAN DEPARTMENTS

21 TOWARD THE NEXT LINEAR COLLIDER: The Beam Line is published quarterly by the A REVIEW OF THE ECFA WORKSHOP Stanford Linear Accelerator Center, ON e+e - LINEAR COLLIDERS P.O. Box 4349, Stanford, CA 94309. Telephone: (415) 926-2585 Gregory A. Loew INTERNET: [email protected] BITNET: BEAMLINE@SLACVM 26 THE UNIVERSE AT LARGE FAX: (415) 926-4500 BANG! BANG! YOU'RE DISTANT. SLAC is operated by Stanford University under SUPERNOVAE AND THE VALUE contract with the U.S. Department of Energy. The opinions of the authors do not necessarily OF HUBBLE'S CONSTANT reflect the policy of the Stanford Linear Virginia Trimble Accelerator Center. 32 GUEST EDITORIAL Cover: Map of the variations in the sky temperature, as measured at microwave frequencies by the NASA Wolfgang K. H. Panofsky Cosmic Background Explorer satellite. The broad yellow band across the center is the Milky Way galaxy. Blue and green correspond to regions that are slightly 33 FROM THE EDITORS' DESK cooler than average (2.73 K), while red areas are slightly warmer. (Courtesy of Lawrence Berkeley CONTRIBUTORS Laboratory) 34

Printed on Recycled Paperl- . 36 DATES TO REMEMBER COSMOLOGY AFTER ( by JOEL R. PRIMACK

What ripples in the cosmic background radiationimply about the origin, composition, and evolution of the Universe.

...... ? ?J . ....

== ...... ?,MOLO.YNC.¥ TO - IKFJVIS PHYSICS before the Newtonian tion, or geology before plate Lfundamental theory, an all- igin and evolution of the d pieces of the story. Almost ,ral relativity provides of the h piece. Big Bang Nucleo- It describes the period starting Bang and ending about three ore than 90 percent of all the d, producing most of the

BEAM LINE 1 Far Infrared Diffuse Infrared Absolute Spectrophotometer Background Experiment

'he radiation emitted then rave had a characteristic ther- "blackbody") spectrum with rature of about 3000 K. The 3e has expanded by a factor of thousand since then, and the fting of this radiation has the temperature to drop by re factor, to 2.73 K. But its m should still be that of ther- iation, which COBE has con- to better than one percent y. isions between light nuclei eV energies they had during of Big Bang Nucleosynthesis en studied in the laboratory. ulting nuclear physics data owed computer calculations .ct the abundances of all the s of the lightest elements (hy- helium, and lithium) that have been produced in the Jniverse. These predictions asonably well with estimates r actual primordial abun- made from observations of ements in various astronomi- ironments. success of these Big Bang synthesis predictions is one hree traditional pillars of the d Hot Big Bang model, the vo being the observed expan- the Universe and the thermal m of the CBR. Together, they wincing evidence that the early se was far hotter, denser, and niform than the present one. vCOBE has at last found struc- ;he CBR, which previous mea- nts had shown to be extreme- )th. The typical temperature ons detected by COBE in )ns ten degrees apart are about ,nr-, ,,i'v Inr n nnnnen T\-. inrpe aUULtt OUUUUU yCiL cILC LI1 D18Jl Ll rms -= bju-m nlsv tuI U.UUUUU 1x1; OUI,I-'

2 WINTER 1992 the average temperature T is 2.73 K, toward the origins of the ripples- this corresponds toAT/T 10- 5. possibly during the first 10- 35 Astronomers have been search- second-and forward toward the ing for these ripples ever since the formation of galaxies and clusters. CBR itself was discovered, but the The relative magnitudes of the ripples only non-uniformity found previ- observed at smaller angles versus ously was the dipole component of larger angles is consistent with the the CBR anisotropy, discovered in spectrum suggested twenty years ago 1983. This presumably arises from by Yakov Zel'dovich and others, our (that is, the earth's, sun's, and which has equal power on all distance galaxy's) motion with respect to the scales. This "Zel'dovich spectrum" absolute rest frame in which this of primordial fluctuations is also dipole anisotropy vanishes. Our what occurs in the simplest models motion causes a Doppler blueshift or of cosmic inflation. And now that redshift of the CBR photons, which we know the rough magnitude of the makes the observed temperature a fluctuations that gave rise to struc- COBE maps of the sky temperature little higher in the direction toward ture when the Universe was only made at three microwave frequencies which we are moving, and corre- about 0.002 per cent of its present (31, 53, and 90 GHz). Light regions spondingly lower in the opposite di- age, we have a solid starting point for correspond to slightly hotter areas of the rection. The ripples detected by COBE calculations of their evolution into sky. The maps are plotted in galactic the galactic center at structures observed today. The coordinates, with are two orders of magnitude smaller the the center of each map and the galactic than the dipole anisotropy. observed ripples in temperature north pole at the top. Microwave Now that instruments have at presumably arose from correspond- emission from the Milky Way is respon- last reached the required sensitivity, ing ripples in the density of the matter sible for the bands across the center of an entirely new field of astronomical in the Universe at that early time. each map. Although noise is compa- observation has begun. With its seven But gravity amplifies such ripples: rable to signal in each ten degree pixel, degree resolution, COBE has rela- where there is a little higher density, the larger features visible above and the Universe expands a little more below the galactic plane in the top two tively blurry vision. Measurements maps represent actual large-scale of the ripples at smaller angular sepa- slowly, until eventually the expan- structures in the cosmic background rations and with higher precision sion stops within this clump, its radiation. The dipole anisotropy in the will be crucial in determining which matter falls together, and a galaxy, sky temperature (which peaks in the cosmological theories can survive. cluster, or supercluster forms. The direction of the constellation Hydra, to At a meeting in Berkeley in Decem- COBE results are consistent with such the right of the maps, just above the ber 1992, strong confirmation of the a gravitational origin for structure, plane of the Milky Way) has been removed from these maps. COBE results was reported by an MIT but only if most of the mass in the group that made a map of the ripples Universe is invisible. covering about one-quarter of the The recent COBE discovery finally sky using data from a balloon experi- ends the game of chicken between ment with a resolution of about four theorists and observers. Theorists degrees. Preliminary detections of originally assumed that the Universe ripples at separations of about one was entirely made of ordinary degree were reported there by sev- matter-protons, neutrons, and eral groups using both balloon-borne electrons. Since most of the mass is and ground-based detectors. in the baryons, we call this "baryonic The measurements of ripples in matter" for short. But if the fluctua- the CBR point backward in time tions were adiabatic (that is, matter

BEAM LINE 3 12

and radiation fluctuations coincide, of light years). However, almost all as predicted by most theories, and as detailed computer simulations of I will assume from now on, a purely the evolution of a CDM universe baryonic universe would have needed agree that this model predicts galaxy much greater fluctuations to have velocities on small scales that are formed structure by the present day. Large-scale structure in a galaxy too high, whereas setting b =2.5 does COBE would have seen AT/T- 10-4, an distribution measured by the Harvard- much better in this regard. Thus the order of magnitude higher than it Smithsonian Center for Astrophysics. problem with cold dark matter is of the observed. The Milky Way is at the center only about a factor of two or three. plot, and each dot represents a large By about 1980, observations of But the COBE and large scale galaxy on a slice of the heavens galaxy data are now so good that this the high velocities of the stars and extending out to a distance of almost a of fudge is unacceptable! How- gas in galaxies, and of galaxies within billion light years. The large blank sort clusters of galaxies, had convinced regions are obscured by dust and stars ever, this near-agreement certainly most cosmologists that the bulk of in the Milky Way. The small blank inner does suggest that some-perhaps the matter in galaxies and clusters regions are true voids. [M.J. Geller and most-of the basic assumptions of had to be invisible or "dark." The J.P. Huchra, Science 246, 901 (1989).] CDM may be right. In particular, it early failures to observe fluctuations suggests that cosmological structure in the CBR could possibly also be grew by gravitational amplification explained by supposing that a large millions of years later. Observations of the sorts of primordial fluctuations fraction of all the matter in the Uni- strongly suggest that galaxies are whose thermal signatures COBE verse is dark matter (DM). So theorists much older than superclusters, observed in the CBR, rather than made models of the formation of however, so it seems unlikely that because of giant explosions. Matter galaxies and clusters with various HDM and its top-down structure fell, it wasn't pushed! assumptions about its composition. formation scenario can be correct. Cold dark matter was an edu- Zel'dovich and his colleagues, and cated guess regarding such a funda- others, worked it out using adiabatic mental theory. But, like the original fluctuations and "hot" dark matter COLD DARK MATTER SU(5) grand unified theory in particle (HDM)-for example, neutrinos with physics, it was apparently too simple a mass of about 25 eV. My colleagues and I, and other groups, to be true. In the summer of 1983, I Because they are so light, the HDM subsequently worked out a theory of was working with Martin Rees on a particles would still be relativistic- structure formation using adiabatic draft of the paper* that first put the traveling at nearly the speed of light- fluctuations and "cold" dark matter whole CDM theory together. Once, when the Universe was about a year (CDM) particles, such as Weakly when I was feeling particularly en- old, the temperature was about 107 K Interacting Massive Particles (WIMPs) thusiastic about this theory, I asked (1 keV), and the amount of matter in or axions, that were moving non- Martin's opinion of it. He said he causal contact was about that of a relativistically one year after the Big would give it about a 10 percent large galaxy such as the Milky Way. Bang. The CBR fluctuation amplitude chance of being right. I must have The relativistic motions of the HDM required with standard CDM is AT/T looked crestfallen, since he immedi- particles would have smoothed out = 10-5/b, where the "biasing factor" ately added that this was a much primordial fluctuations on galaxy b is a fudge factor that determines better chance than he had ever given scales. The smallest structures that the height of the fluctuations. With any previous cosmological theory. It can survive in a HDM universe are b = 1, CDM agrees with COBE, and wasn't a bad estimate. superclusters, with masses a thou- also incidentally with essentially all available data on the dis- sand times larger. These would be of the *G. R. Blumenthal, S. M. Faber, the first structures to coalesce, with tribution and velocities of galaxies J. R. Primack, and M. J. Rees, Nature 311, galaxies forming by fragmentation on large scales (hundreds of millions 517-525 (1984).

4 WINTER 1992 HYBRID STRUCTURE MODELS with the lower estimates giving Q2 0.2 and the higher estimates con- After COBE, it appears that the data sistent with Q = 1 as predicted by require a hybrid cosmological theory cosmic inflation. However, once the containing elements of at least two Hubble Space Telescope's optics are simpler theories, such as a mixture fixed, it should be possible to use of cold and hot dark matter. The observations of individual variable resulting theory may thus be a little stars in distant galaxies to measure like the Standard Model of particle the distance to these galaxies, and physics, which is also a hybrid of thereby measure Ho reliably. And QCD and electroweak theories. This rapidly accumulating data on galaxy may be more than just an analogy, velocities may finally allow us to since a true "theory of everything" determine the value of Q accurately. should explain both the connection One variant of CDM that remains between quarks and leptons, and also viable is "low-density CDM." This that between cold and hot dark mat- model has Q1 0.2, with Ho near 100, ter particles. at the upper end of the allowed range. Cosmological data is now flood- In this model it is necessary to include ing in from many workers using new a cosmological constant-which instruments and techniques. The fact amounts to assuming that every part that the standard CDM model nor- of empty space repels all other parts malized to the COBE data (with b = 1) of space. (Einstein first introduced also roughly fits all the data on the the cosmological constant to coun- distribution and velocities of galax- teract the gravitational attraction of ies on large scales is very encourag- matter in order to get a static uni- ing. It suggests that the time may verse; when Hubble discovered that now be ripe to construct a detailed the Universe is actually not static, theory that specifies the nature of but expanding, Einstein said that the the primordial fluctuations and of cosmological constant was his worst the dark matter. From them, we mistake.) In low-density CDM, the might then explain the origin and present epoch is special, with the large-scale distribution of galaxies, Universe now starting into an era of including the "voids" and "walls" late cosmological inflation caused recently observed. by the cosmological constant. At In order to build such a theory, we best, low-density CDM is a hybrid will need to pin down the two key model. cosmological parameters whose val- There are, however, several ues can be determined from obser- indications that Q2 1. For example, vation but still remain uncertain: the distribution of the all-sky map of the Hubble parameter Ho, which brightest infrared galaxies cataloged measures how fast the Universe is by the IRAS satellite is peaked in the expanding, and the cosmological same direction as our motion with density parameter Q. (See box at the respect to the CBR rest frame, which right.) The value of Ho is uncertain produces the dipole anisotropy in by about a factor of two,while Q is this radiation. It is tempting to sup- uncertain by about a factor of ten, pose that these galaxies trace the

BEAM LINE 5 nearby dark matter distribution, The main objection to 0XDM in which caused the gravitational principle is the apparent unlikeli- acceleration that produced our hood of having two different dark velocity with respect to the CBR rest matter components each making frame. If so, the required cos- comparable contributions to the Mpc mological density of this unseen mass density. However, the entire matter corresponds to 12 - 1. particle physics Standard Model begs Similarly, a comparison of the IRAS for further explanation, so it should galaxy density with the data on galaxy not disturb us too much to contem- velocities compiled by the "Seven plate one more feature that, if valid, Samurai" team and others-as well would call for a more fundamental as a new project reconstructing initial justification. conditions from the galaxy The fact that J)DM is a promising velocities-also work only for 2Q 1. model for large-scale structure was While this evidence that 2 = 1 is still established by several simplified cal- not compelling, and arguments for a culations. My colleagues and I have large Hubble parameter and an old just done the first detailed computer Mpc Universe do seem to point toward N-body simulations for JXDM, with smaller Q2, I personally feel it is likely proper initial conditions, sufficiently that 2 = 1. many hot particles to sample veloc- Perhaps the simplest hybrid model ity space adequately, and careful com- that might work is an 12 = 1 cold- parisons to the available data. We plus-hot dark matter mixture, with find that 0XDM is consistent with all U L q u Mpc about twice as much cold mass as the available data. On large scales of hot. These proportions of hot and about 100 million light years and Distribution of particles representing cold dark matter are required to fit more, the 0XDM streaming veloci- cold dark matter (top) and hot dark the data on large-scale structure. I ties are considerably larger than those matter (bottom) in computer simulations predicted by CDM, and in excellent of a XCDM universe by Anatoly Klypin, will use the phrase cold-plus-hot dark Jon Holtzman, Enikd Regds, and the matter (or CXDM, with Xrepresent- agreement with the data. We also author. Note that the hot particles are ing CH) to refer to this model. (XDM find that on galactic scales the high more smoothly distributed and that no has only one additional parameter initial neutrino velocities make the density enhancements in hot particles beyond those of standard CDM: the neutrinos cluster around galaxies less are seen near the small clumps of cold neutrino mass m( vZ), or equivalently than the cold dark matter (see illus- to particles (which would correspond the mass density in neutrinos Q2v tration at left); this results in galaxy small galaxies) at the center right. = [m(v-)/23 eV]. The required value velocities that are smaller than in of m(vr), about 7 eV for Ho = 50 km/ CDM, again in good agreement with sec/Mpc, is consistent with the value the data. Thus in the cXDM model, implied by the currently available the neutrinos that make up about a solar neutrino data plus the most third of the dark matter fix the prob- popular models of neutrino masses lems of standard CDM on both large (see box at right). The neutrinos and small scales. That, plus the fact provide an unclustered dark matter that the neutrino mass needed seems component on small scales, which reasonable from the viewpoint of par- could help explain why measure- ticle physics, is why I think this is ments typically give Q < 1 on these the most attractive hybrid model scales. proposed so far.

6 WINTER 1992 T HE NEAR SUCCESS of cold dark matter in predicting the magnitude of the ripples measured by COBE favors the hypothesis that structure formed in the Universe through gravitational collapse from the sort of initial fluctuations provided by cosmic inflation. In the temperature ripples discovered by COBE, we see on very large scales direct evidence of the density in- homogeneities whose smaller-scale counterparts grew through gravitational amplification into the galaxies, clusters, superclusters, and voids we see today. But difficulties in simul- taneously explaining the small- and large-scale data in the CDM model have led to the demise of CDM and the rise of hybrid schemes such as cold-plus-hot dark matter (JXDM). It is technologically feasible to detect the dark matter candidates whose existence is predicted by current particle physics theories, and sen- sitive experimental searches for all of them are now well underway. The possibility that the neutrino oscil- lation experiments now starting at CERN may confirm the prediction of J)DM that m(vT) - 7 eV is a good illustration of the growing inter- dependence of cosmology and particle physics. With COBE's detection of the ripples in the cosmic background radiation, we at last have a solid foundation for cosmology. It is perhaps not too much to hope that in the present decade we may succeed in constructing a grand cosmological theory that explains the origin, evolution, and matter content of the Universe. 0

BEAM LINE 7 mal;f00ize d,':"'did0 nE t~da, h .ese ...... W;;.... . T IS AN- HONOR AND PRIVILEGE to be here as part-of 1A Eslendid occasion. It is a very German one, and I have a schi2 both an outsider and an insider. The felin insider~isnotonlybecausee of my present attachment to -:-.-Institute as vig-committee (Fachbeirat) member, but,, simply-as a mee of-t he international communt of art ' p-~-~hysics,hysics, of whichch thisinstition, isofu, St : tant component. I orBBeat st o field.t weares minternationalized, and in these as o g nat al;-isit i--s-essential to reaffirm that commitmen.igh-en physics, especially here in Europe, servesaspla fields of endeavor much more important than our own i the splgit of international cooperation and of inter e is much more difficult to attain and maintain thanit i ourselves. Before going on to the main part which he had struggled for some time, of this talk, I should like to pay the second derivative of entropy with homage to Max Planck. Figuratively respect to energy went inversely with speaking, I met Planck when I was the energy. Then when he learned an assistant professor at Stanford, (from experiment!) of the infrared assigned to teach graduate quantum Rayleigh-Jeans limit, he quickly de- mechanics. I wanted to know how termined that in that limit the sec- Planck did it. The usual textbook ond derivative went as the inverse descriptions left the impression of square. This invited an obvious in- black magic, not science. So I went terpolation. The result fitted the data to the source of all wisdom on such so well that he went on to look for matters at that time at Stanford, Mr. the physical explanation. His intro- Chung-Kuei Chang. Mr. Chang was duction of quantum oscillators, along a physics groupie, a dropout from the with a very unconventional use of electrical engineering department counting statistics, was on the one who for many years slept in the attic hand an "act of desperation" (his of the Physics Department, lived on words) that went beyond the logic of one egg a day, and earned a small classical physics, and on the other Chung-Kuei Chang stipend developing film badges in hand a set of instinctive hypotheses the linac laboratory. Mr. Chang lived rooted in his deep understanding of in the physics library and knew his- thermodynamics and electromagne- :: : 777 -- tory, with strong opinions about it (I tism. To me Planck's accomplish- regret to say that he preferred ment stands as a marriage of electro- Notes to Einstein). dynamics with thermodynamics by Poincare and Hilbert Jammer, The Conceptual When I asked about Planck, he led a method that is a marriage of sci- Development of QuantumF~~I :A-~l-:?-_:-: .ii ; me without pause into the stacks ence with art. Abraham Pais, in his (McGraw-Hill Book 2 Mechanics and pulled out a volume of what I marvelous book on Einstein, de- Company, New York1966). : : 2 recall to be the Obituary Notices of scribes this with special eloquence: Abraham Pais, 'Subtle is the Lord the Royal Society of Edinburgh from "His reasoning was mad, but his .. ',The Science and the Life of the 1920s, and opened it to an article madness had that divine quality that Albert Einstein (Oxford University only the greatest transitional figures Press, Oxford, 1982). written by Planck, although I haven't 3H. Georgi and S. L. Glashow, Unity bring to science. It cast Planck, been able to locate it again. In any can ofAll Elementary Particles, Phys. case, I acquired thereby a vivid im- conservative by inclination, into the Rev. Lett. 32, 438 (1974): pression of Planck and learned how role of a reluctant revolutionary. .. " 4 H: Georgi, H. R. Quinn, and he got the formula. Maybe even here How Planck did it is a great example S. Weinberg, Hierarchy of Interac- in this Institute there are those who of the importance of finding the right tions in Unified Gauge Theories, 1 way of looking at a problem, and Phys. Rev. Lett. 33, 451 (1974).i don't remember. It was not simple 5 curve fitting, but rather an exercise must always stand as one of the great Paul Langacker, ProtonDeca, at the Benjamin Franklin sense intellectual adventures in the his- presented carried out with an exquisite Symposium in Celebration of the of taste and deep insight. He looked tory of physics. Discovery of the Neutrino Philadel- at the problem in terms of the en- phia April 29, 992 (UPR-0539 tropy of electromagnetic field modes T IS TIME TO SKIP from then to May 1992, and HEPPH:9210238)., as a function of energy, because that now, and to get on with the sub- 6James Gleick, Genius: fRichard was fundamental thermodynamic ject of this talk. The answer to the Feynman and Modern Physics thinking as practiced by Boltzmann. question posed in the title, "Where (Pantheon, New York, 1992) In the ultraviolet Wien limit, with do we go from here?" seems in some .

BEAM LINE 9 sense trivial. We live in a quiet time social structure evolves toward larger "organization men" (and women) for the field. It is not that nothing is and larger experimental collabora- necessary for the smooth operation happening. The Standard Model, that tions, with fewer and fewer experi- of the experiment. And in a well- incredibly successful product of the ments. That must in large part be our managed experimental group one efforts of a previous generation of future, like it or not. should expect that the individualists experiment and theory, serves us all These boundary conditions have will be heard and respected, just as too well. The recent rounds of beau- for some time been effecting a real they traditionally have been within tiful, precision experiments attest to social change for the field. And it the physics laboratories in the past. the great accuracy to which the Stan- becomes an interesting issue as to And one can also expect that there is dard Model works. And the success whether this social phenomenon a basic sense of satisfaction in being of the Standard Model has created a changes things at the actual scien- a part of an enterprise on the cutting standard answer to the question of tific level: are the actual physics re- edge of the most important develop- the future: for the electroweak sec- sults conditioned by the new sociol- ments in the field. tor the great proton-proton colliders ogy, or is physics physics, no matter Not so long ago I addressed this LHC and SSC (the Large Hadron what the style in which it is done? Of subject and expressed the opinion, Collider at CERN in Switzerland and course styles in physics have always not well researched, that I felt that the Superconducting Super Collider been changing, and it is traditional the social situation in the big col- in Texas), and someday an electron- for the older, outgoing generation- laborations was all right, similar to positron linear collider. And for spe- like me-to bemoan the fact that what existed in laboratories in the cific particles of interest, "factories" things aren't the way they used to be. past. Afterward I faced a firestorm of (kaon, phi, tau-charm, B, Z, eventu- There is probably one feature that is criticism from some of the young ally top, and maybe even Higgs) are handed down more-or-less intact: the people present that it was not as of importance. It should not of course basic way individuals are taught to good as I said. Since then I have heard be taken for granted that they will all think physics is arguably not too both sides, each expressed by young get built. We are most privileged and different now from olden days, al- and old. I don't know what the real fortunate that any of these machines, though I will later mention a pos- situation is, because it is a little hard past and present, will get built. But I sible difference. And it is certain that to see in from the outside. The mod- presume here that sooner or later the native talent in the present gen- ern collaboration is not as completely someone will build them. eration is there in abundance, again open a society as a laboratory. Most I choose here not to utter addi- arguably as much as in any past gen- of the communications are collabo- tional words on why these initia- eration. Nowadays the single big ration confidential, and perhaps a tives are the right way to go. And I do collider experiment is a community model of the social structure is closer not challenge the basic correctness comparable in size to a complete lab- to an industrial lab such as Bell Labs of the argumentation. The most im- oratory such as Stanford Linear Ac- or IBM. I am not sure how much of a portant physics problems do appear celerator Center (SLAC) orDeutsches change this represents. to be sharply defined, despite the Elektronen-Synchrotron Laboratory I do think that big-collaboration present difficulty in doing anything (DESY) in their early days. At that sociology is a problem that in prin- about solving them. The new facili- time those labs seemed monstrous, ciple is solvable given good, sensi- ties are vital in making progress; the Big Science at its biggest. Retrospec- tive management. I know that bright best efforts of theoreticians appear tively they look not at all that way, young physicists go into this dif- to be insufficient. and the science accomplished there ficult sociology with eyes wide open, Because the push to higher ener- was of highest quality. If one looks because that is where the physics is, gies is unavoidable, we will be stuck within the workings of the big col- and they must be part of it. And in with bigger, more costly accelera- laborations, it is certain that there large part I do think they get satisfac- tors and colliders. That means every will be found highly creative rugged tion from it, despite the headaches. thing takes longer to do. And the individualists, as well as the modern But of all these things I am not sure.

10 WINTER 1992 T HERE REMAINS the question by the Monte Carlo generators carry eliminate-and I think that it is even of whether there is a change in the authority of data itself. They harder to do so the larger the the physics itself because of this so- look like data and feel like data, and collaboration. ciological change. One concern is in if one is not careful they are accepted Another possible sociological the area of instrumentation. In a big as if they were data. All Monte Carlo change, one conditioned not only by integrated collider detector, with a codes come with a GIGO (garbage in, the bigness of the enterprise, but by construction time of a good fraction garbage out) warning label. But the the success of the Standard Model, is of a decade, it is unacceptable for a GIGO warning label is just as easy for the predominance of what I call the subsystem to fail. Innovative tech- a physicist to ignore as that little engineered discovery. At present we nologies are disfavored. Design message on a pack of cigarettes is for witness the unfolding of the discov- choices may be frozen many years a chain smoker to ignore. I see nowa- ery of the top quark. Its properties before the detector is commissioned, days experimental papers that claim are believed to be known well enough leading to the danger of technologi- agreement with QCD (translation: that a well-defined search strategy cal obsolescence ab initio. I hear of someone's simulation labeled QCD) can be laid out in advance. And if by examples of this in the space pro- and/or disagreement with an alter- some chance the properties differ, gram, where the instrumentation for native piece of physics (translation: the top may not be discovered at all a mission yet to be flown is already an unrealistic simulation), without in the present round of experiments decades obsolete because of rigid much evidence of any critical ex- even if produced. Not all discoveries design freezes made long ago. No amination of the inputs into those are that way; take for example the matter how honorable the intentions simulations. I am prepared to be- strange particles, which appeared in of a large collider-experiment's man- lieve that the computer-literate gen- cosmic-ray experiments totally un- agement, it becomes more difficult eration (of which I am a little too old invited. But with the gluon, then the to be adventurous. And adventurous to be a member) is in principle no W and Z, and now the top quark, we instrumentation and experimental less competent and in fact benefits are getting used to the engineered design, as well as adventurous ex- relative to us in the older generation approach. Even such a speculative perimental goals, seem to me as im- by having these marvelous new tools. venture as supersymmetry searches portant to nurture as adventurous They do allow one to look at, indeed gets presented in the most stylized theory. It is not just theorists who visualize, problems in new ways. But way nowadays. When I hear a talk must look at the problem the right I also fear a kind of "terminal ill- these days on that subject, I know in way. Experimentalists must look at ness," perhaps traceable to the influ- advance what to expect. There ap- nature in the right way. ence of television at an early age. pears an emotionless young experi- Another change that I find dis- There the way one learns is simply mentalist riffling through the pre- turbing is the rising tyranny of Carlo. to passively stare into a screen and approved, collaboration-rehearsed No, I don't mean that fellow who wait for the truth to be delivered. A stack of transparencies. (It seems to runs CERN, but the other one, with number of physicists nowadays seem be unprofessional these days for a first name Monte. The simultaneous to do just this. real experimentalist to exhibit emo- increase in detector complexity and A further concern of mine is a tion). First comes the description of in computation power has made tendency for bias in physics output the apparatus, then the analysis simulation techniques an essential in favor of the expected result. An method, and then a numbing set of feature of contemporary experimen- analysis that agrees with expecta- exclusion plots (this parameter ver- tation. The Monte Carlo simulation tions is sure to get somewhat less sus that parameter with shaded ar- has become the major means of visu- scrutiny in a big collaboration than eas showing the disallowed values of alization of not only detector perfor- one which disagrees. This bias is as the this and the that, the parameters mance but also physics phenomena. old as physics, well known to all having been taken from the Minimal So far so good. But it often happens experimentalists. But even with the Standard Supersymmetric Model). that the physics simulations provided best precautions it is hard to Then comes the finale, when the

BEAM LINE 11 speaker recites the future plans, for strong force. It indeed eventually was overwhelming emphasis on only which they hope to extend the limits found as a very broad pi-pi s-wave three physics goals: find the top, find on the this and that to such and so. resonance. But in the 1980s it was the Higgs, and explore CP violation And then I ask my own standard removed from the Particle Data in K and B decays. Yes, these are the question: "You mean you don't hope Group compilations because no one Big Three, but what about the rest? to discover the particles?" Not only was interested in it. The key to the Program diversity can be maintained the search for supersymmetry, but strong interactions, quarks and glu- at various levels-by many different also its nondiscovery, is becoming ons, was just not accessible. Experi- kinds of accelerators and colliders, institutionalized. ments beyond everyone's imagina- by a variety of different kinds of I bet that supersymmetry, if true, tion were the real key to the strong detectors within a given collider or manifests itself in utterly surprising interactions. accelerator program, or by experi- ways. Simple changes in the under- What is best understood about ments within which there is as much lying assumptions may lead to wild the Higgs business is that we don't diversity as possible built in from and rich changes in the phenom- understand it. There is the necessity the start. And within a given experi- enology, just as rich as changing neu- of something like it being there by ment it is individuals who will create tron and/or proton mass by a few the time the TeV mass scale is extra diversity by extending the percent, or letting the muon weigh reached. But the first manifestation original functions of the apparatus more than the pion instead of less. In of its existence may be in some to the outer limits of possibility. the phenomenology business, details lucky, non-engineered discovery. In electron-positron colliders are everything. That could even occur at low mass there is relatively little diversity pos- But the most hazardous area of scales, for example if some axion- sible in the detector design. Because discovery may be in the search for like particle or particles exist within of the low collision rate, each event the Higgs particles and related phe- the Higgs sector. It seems to me that can and must be examined with as nomena. The Higgs boson has been the slow timescale of progress tends much acceptance and information described as the Holy Grail of con- to institutionalize the generic op- as possible. Thus all such detectors temporary particle physics. Nowa- tion, despite lack of strong evidence nowadays, such as those at LEP, are days it is often viewed as an object of that it is all that credible. 4 pi and look not too dissimilar. search as well-defined as the top The important case of proton- quark or Z. Beware! This thing need HAVE DWELT at length on diffi- proton colliders is very different. not be that way at all. There is a culties. These are difficulties Only a small fraction of all events strong warning from history that the immersed within a field of great that occur can be recorded, and there Higgs problem is not being viewed in importance and opportunity. My is consequently the possibility of the right way. The Higgs physics has concern is not that the enterprise we selecting in one experiment an its analogue in strong interactions. It face is ill-conceived or not worth entirely different event class than in is the physics of pions and chiral doing, but that it deserves to be done another one, with detectors opti- symmetry breaking, described by in as optimal away as possible. All of mized accordingly and therefore what is known as the sigma model. these concerns argue for maintaining quite different from each other. So it From the perspective of the late at the highest institutional levels of is a little disheartening to me to see 1950s, when all strong-interaction the field as much diversity as the present worldwide lack of physics (other than those uninvited possible. There is convergence in the diversity in the hadron-collider strange particles) focussed on the major scientific interests, as well as programs, present and future. At interaction of pi mesons with nucle- the convergence in the number of big Fermilab, the two major detectors, ons, the analogue of the contempo- experiments in existence. In the most CDF and DO, are not all that different rary battle cry would have been "Find recent U.S. high-energy physics in configuration and physics goals. the sigma! " The sigma particle would Subpanel convened to examine the Similarly the major detectors, SDC be the key to understanding the physics of the 1990s, there was and GEM, being designed for the

12 WINTER 1992 Superconducting Super Collider are collisions and, not to be overlooked, not so different, and it is likely that proton-ion collisions. We should the same situation will prevail at the remember that quantum electro- LHC. Why? It is felt that the top- dynamics experimentation goes well priority physics goals deserve at least beyond precision measurement of two generic detectors to provide the fine structure constant and checks and balances. But building computable elementary processes. two detectors fully commits (or over- All of condensed matter physics is commits) the existing resources, ultimately quantum electrodynam- leaving precious little opportunity ics and not at all uninteresting. It is for anything else. It is quite possible a comparable challenge to explore that we will end up with a worldwide fully the consequences of quantum version of LEP, with four detectors NE MAY ASK what physics chromodynamics, especially beyond optimized for the Higgs search. One opportunities exist beyond the limits of calculability, and cer- must hope that the Higgs is indeed what is covered in the generic tainly beyond the precision measure- what it is supposed to be! And program. Everything to do with ments of the running coupling meanwhile the optimal development neutrinos is clearly important to constant. of B-physics in hadron-collider mode, pursue. And in addition to that and Thus far I have not said much just as one example of an alternative, what I have already mentioned about theorists. Can we hope for will be greatly inhibited if not com- elsewhere, it seems to me that there theorists to lead the breakout from pletely stalled while all this goes on. is a wealth of strong-interaction the Standard Model, as Planck led Another way of establishing phenomena waiting to be explored. the breakaway from classical phys- diversity is via different machines. At hadron-hadron collider energies ics? Of course it is possible, provided The array of "factories" already there are whole classes of processes, not everyone is working at and be- mentioned is one way of doing this, especially involving leading particles, yond the Planck mass scale in 26 one that may be especially valuable that are totally unexplored. This is dimensions, and provided the right in technology innovation as well as an area of great interest to QCD data exists. Maybe the necessary data for solid physics results. In addition, practitioners, who anticipate a new already exists, maybe not. There is there might be "search factories" kind of nonperturbative strong no shortage of information: we have mainly devoted to looking for interaction associated with the the famous 20 parameters of the stan- phenomena beyond the standard blackness of partons at extreme dard model and a simple periodic model. The Brookhaven AGS program energies, even at short distances. And table of the constituents of matter to of the 1980s has been a splendid cosmic-ray data also hint at novel work with. So far that hasn't been example of this. A major investment phenomena. Beyond that, much sufficient, although already Planck- went into looking for rare K decays remains to understand in spectros- style thinking did almost work. and neutrino oscillations. It has been copy, heavy-flavor as well as non- By this I mean the grand- marvelously successful in finding flavor glueball physics. Spin physics unification idea. The first paper of 3 nothing to great accuracy. That is, is an area where theorists tra- Georgi and Glashow was full of both seriously, a true success. Extending ditionally stumble in their pre- insight and madness,with its require- this kind of program in the future is dictions. At present polarization ment of proton decay. The work of 4 going to be a lot harder, but well experiments in deep inelastic Georgi, Quinn and Weinberg that worth serious effort. And the non- scattering provide QCD with one of followed shortly thereafter on accelerator programs deserve to be its most important challenges. In running coupling-constant unifi- advanced, although they will never addition, there is the high- cation turned that madness into be a substitute for the high-energy temperature and high-multiplicity divine madness, because it set a machines. frontier addressed by ion-ion definite scale for the unification and

BEAM LINE 13 predicted a measurable rate of proton along with the long time scales, going that new physics can happen any- decay. The profound difficulties with to mean the end of the line? I doubt time and change our outlook in a big the idea, in particular the awkward- it, at least for the next century. I hurry. Now is indeed a quiet time. If ness in the Higgs sector, had to be suspect that a still bigger proton- I search for a comparable historical ignored. That would have been for- proton collider will eventually be period, the best I can find is the late giveable, just as Planck was forgiven, contemplated, perhaps a true VBA nineteenth century, just prior to had the experiments on proton decay (very big accelerator), to be a world Planck's great discovery. Maybe there cooperated. Unfortunately they did project as the original 20 TeV VBA, is something like that awaiting us; not, and we are now left with an the precursor of SSC and LHC, was maybe not. But no matter what, I inconclusive impasse. It can be contemplated in the 1970s to be. The think it important to be prepared for summarized by noting that there energy will probably be limited by surprises, maintaining at the insti- used to be an annual workshop on the demand that synchrotron tutional level as much diversity as grand-unification theories. It has radiation emitted by the circulating possible. The opportunities for strik- recently been discontinued. Never- proton beams peacefully coexist with ing out in new directions must be theless, this whole development has the superconducting magnets of the maximized. We must provide the for better or worse changed pro- machine. That criterion will probably means for those future physicists foundly the present theoretical terminate the proton-ring sequence who follow the advice of Feynman 6 : climate, so much so that recently at hundreds of TeV in parallel to one well-known reviewer5 stated that what has already happened for It will not do you any harm there is no compelling reason to electron-positron rings at hundreds whatever to think in an original assume that the proton is absolutely of GeV. fashion.The odds that [you] will be in fact right and the general stable. Given the limits of 1032 years Beyond that is a real dilemma. thing that everybody is working on its lifetime, this represents to me Radical linear acceleration tech- on is wrong will be low. But the considerable hubris. Nevertheless, niques, perhaps using laser technol- odds that you, Little Boy most everyone-including myself- ogy such as the plasma beat-wave Schmidt, will be the guy who feels there is something right about idea and others like it, seem a neces- figures things out is not smaller. the grand-unification idea. sary condition, although far from It's very important that we do sufficient. The specifications for a not all follow the same fashion. are Because although it is 90% sure collider that produces physics that the answer lies over there ITH THE LONGER and daunting, and it will take a lot of where Gell-Mann is working, longer time scales for new difficult and expensive R&D to get what happens if it doesn't? facilities and experiments, it is anywhere. Nowadays it is not a popu- probably sufficient here to limit this lar activity, and maybe it is prema- These future physicists may, like talk about the future to the above. ture to struggle very hard with it. Planck, have the combination of deep What we glimpse now should form There is a long and difficult linear- insight and daring to become the the basis of a program lasting several collider program extending into the next reluctant revolutionaries. It will decades. I think the longevity of, say, future as it is, which seems not to take conscious effort at the institu- the ssC, if fully and properly exploited need that radicalism. But I hope that tional level to do a good job of serving to maximize the diverse scientific there are some people out there grap- this future need. The Max Planck opportunities it presents, is at least pling with this long-range challenge. Institute as I have known it has 50 years. And the array of other So that is where I think we are been-and I am sure will be-an facilities could easily provide a heading. I have spent much of my important element in this effort. I healthy supplement over that time time talking about the social issues, wish it all the best for the future. scale as well. But what about the not the physics. Were there news of longer run? Is the increasing difficulty physics beyond the Standard Model I of attaining still higher energies, might have done differently. And

14 WINTER 1992 keVN AK K

1 Z, 0 Electron Energy (keV) mass of 17 keV. Within the last few One of Simpson's spectra from tritium months, however, the strongest implanted in a silicon detector. This plot threads in the fabric unraveled, leav- and others show the residuals between ing physicists bemused that so many data points and a linear Kurie plot. The interlocking strands of evidence curved line is calculated for a 3.3% could have been synthetic. admixture of a 17-ke V neutrino.

coupling would be needed, stronger N 1985, John Simpson at theUni- than the weak interaction. Alto- versity of Guelph in Ontario, gether, too many impossible things Canada, was studying the beta spec- before breakfast. After long deliber- trum of tritium that had been ation, Simpson published. implanted in a silicon detector by Not surprisingly, the paper accelerating the tritium nuclei to aroused intense interest. Five experi- several MeV in the McMaster mental groups in four countries car- University Van de Graaff. It was an ried out experiments on a different ingenious idea that made possible a isotope, sulphur 35 (35S), reasoning precise measurement of the very low- that the admixed neutrino would energy electrons emitted by tritium have to show up in all beta decays. (endpoint energy E0 = 18.6 keV). The results were uniformly nega- Simpson found that it was impossible tive. Simpson was completely unim- to fit the beta spectrum to standard pressed by this, and politely drew Fermi theory with a single, massless attention to serious flaws in each neutrino. Instead, there was a distinct experiment. He made a general com- break in the slope at 1.6 keV (see the ment that we all can heed, "Con- illustration at the top of the next trary to intuition, a null result is not column), which he realized imme- inherently more reliable than a posi- diately could be explained by a second tive one," and he called for better type of neutrino admixed with the experiments. electron neutrino at 3 percent inten- Arguably the finest magnetic beta sity. This new object had to have a spectrometer in the world is the iron- mass of 17 keV. free "Pi-Root-Two" magnetic instru- Such a heavy neutrino flew in the ment at Chalk River Laboratories in face of what was "known" at the Ontario (the polar angle between ob- time. The minimal Standard Model ject and image is 7T 2 radians). It had of particle physics does not accom- been mothballed for some years, but modate neutrino mass (or mixing) at was recommissioned when interest all. A mass of 17 keV fell in the in neutrino mass intensified in the middle of a cosmologically excluded 1980s. A very careful measurement region between about 100 eV and (prompted by Simpson) of the beta 1 MeV for stable neutrinos, so this spectrum of nickel-63 (63 Ni; Eo = 67 neutrino had to decay if it were not keV) revealed no "kink" at 50 keV to shorten the lifetime of the universe (see left illustration on page 17). This to a preposterous 500 million years. was, and remains today, one of the In fact, it would have to decay so most convincing negative results, al- quickly (<30 years) that a new though we will revisit this topic later.

16 WINTER 1992 1.02 m 1.00

cm 0.98

30 40 50 60 Electron Energy (keV) the detector, to make the average Data on 63Ni from the Chalk River mag- angle of incidence more nearly nor- netic spectrometer, showing no mal, and to prevent electrons from III neutrino kink 0 4 8 evidence for a 17-keV scattering into the detector from the cm of the magnitude proposed by Simpson. The solid line shows the expected result walls of the apparatus (right top figure). A much thinner source backing Oxford apparatus used to study 35S and if a 17-keV neutrino were admixed with 63 the electron neutrino at 3% intensity. (2.8 instead of 10 #m of Mylar) was Ni. a-Silicon detector; b-Thin source also used. These changes greatly re- on Mylar backing; c, d-Collimators; e-electron antiscattering baffle. The HEORISTS NOTED TWO effects duced the backscatter tail, and solid line is a normal electron trajectory. in tritium: an improper choice changed its shape from a flat, fea- The dotted line shows a neglected of electron screening correction, and tureless continuum to abroad bump. scattering process discussed in the text. exchange between the outgoing beta The kinks 17 keV below the end- and atomic electrons, that, when points were now very dramatic, es- properly allowed for, reduced the pecially in 3S, where high statistics admixture of heavy neutrino to 1 displayed the precise shape expected neutrino (middle illustra- percent (but not zero). That Simpson for a heavy 0.01 acknowledged to be valid, and tion on the right). the null experiments It was becoming quite difficult to remarked that ( were much less conclusive in ruling explain the positive observations in out such a small admixture even had any terms other than a heavy neu- -0.01 they contained no errors. He and a trino. Nevertheless, it was not until student, Andrew Hime, then carried a new result from the Lawrence Ber- 120 140 160 out an experiment in which tritium keley Laboratory was published that Electron Energy (keV) was implanted in germanium (Ge) the 17-keV neutrino was taken seri- instead of silicon (Si) to address the ously by the physics community at Hime and Jelley's data on 35S, showing odd coincidence that the Si K edge large. It happened that a Ge detector a pronounced kink 17 keV below the curve is a lies at 1.6 keV. The effect was seen in crystal had been grown with a small 167-keV endpoint. The solid fit to a 0.85% admixture of a 17-keV concentration of carbon 14 (14C) as a Ge, just as it had been in Si. But even neutrino. more remarkable was an experiment means for measuring the (very tiny) Hime and Simpson carried out on solubility of carbon in crystalline 35S. The arrangement was quite germanium. The detector lay unused different from that of the tritium for years until rediscovered by Eric 1.0z experiment: instead of an implanted Norman and collaborators. The spec- source, a thin-film source sat a few trum (bottom figure) showed persis- 1.0' centimeters from a cooled Si detector. tent evidence for a kink 17 keV be- Clear evidence was seen for a low the 156-keV endpoint. 4~ 1.0( discontinuity in the 35S spectrum at 150 keV, 17 keV below the endpoint. 0.9O Hime went to Oxford, joined H OW COULD ONE MAKE SENSE 0.9< forces with Nick Jelley, and made of this? On the one hand 100 120 140 160 new Si-detector measurements of the experiments (including many we Electron Energy (keV) spectrum of 35S and 63Ni. The appa- have not mentioned) with magnetic ratus was basically the same as at spectrometers were unanimous in Spectrum from 14C-doped Ge detector Guelph, but baffles and collimators finding no evidence for a 17-keV at the Lawrence Berkeley Laboratory. were incorporated to prevent elec- neutrino, while on the other all the The fitted curve corresponds to a 1.26% trons from clipping the corners of positive reports came from admixture of a 17-keV neutrino.

BEAM LINE 17 Inner Outer experiments with solid-state detec- T HE FIRST BREAKTHROUGH Solenoid Coil Solenoid Coil tors. A few solid-state experiments came from a group at the Insti- Cold , / Si(Li) reported negative results, but at a tute for Nuclear Studies in Tokyo. Surface N N m N Detector confidence level that was not violent- The INS has a three-quarter scale Source 3 ly in conflict with the positive ones. version of the Chalk River X 2 mag- IFq bm- Simpson had long criticized the netic spectrometer, instrumented 1 magnetic-spectrometer experiments with 30 proportional counters on the for their use of arbitrary "shape focal plane. The spectrum of 63Ni correction" parameters. An a priori was measured intensively between 'd 0 calculation or measurement of the electron energies of 39 and 59 keV, W-4 energy-dependence of efficiency of surrounding the expected kink at magnetic spectrometers seems to be 50 keV. A prodigious number of -0.2 0 0.2 0.4 0.6 in Position (m) impossible at the level of accuracy events, 2.4xl09, was accumulated needed, and so it is customary to 30 spectra. As with all other mag- parameterize the efficiency by a poly- netic spectrometer measurements, superconducting solenoid The Argonne correction parameters and silicon detector. A typical axial nomial in energy whose coefficients arbitrary shape magnetic field profile is shown below. are determined from the data itself. (30 of them, a linear term for each That seems innocuous enough-how spectrum) were used to fit the data. could a smoothly varying function However, the very high statistics, hide something as sharp as a heavy- the small magnitude of the param- neutrino kink-but it is in fact quite eters, and the excellent fit to the data dangerous. Giovanni Bonvicini, in probably render the result safe against an important and influential paper, Bonvicini's caution. An upper limit showed that the combination of an of only 0.077% at 95% confidence unknown efficiency function mod- level on the admixture of a 17-keV eled by a first-order polynomial, and neutrino was obtained. Although the modest statistics, could be lethal. A remaining systematic uncertainty systematic error arises that always associatedwith the spectrometer effi- causes an underestimate of the size ciency is not known, the result is of any kink actually present in the assuredly in contradiction to all posi- spectrum. The underestimate can be tive results. a factor of five. So it appeared At the same time, a novel experi- conceivable that all magnetic- ment on 35S was being carried out by spectrometer experiments might be Stuart Freedman and his collabora- suspect, and perhaps the 17-keV tors at Argonne National Laboratory. neutrino really did exist. Freedman made use of a super- In physics the requirements for a conducting solenoid and Si detector really decisive experiment are not system originally built by Zbigniew obvious at first. To resolve the ques- Grabowski for the study of tion, a magnetic spectrometer exper- conversion-electron spectra. A source iment with "infinite" statistics or is placed on the axis of the guide field with no arbitrary shape-correction near an axial maximum (see figure parameters was needed. Alternative- top left). Electrons are constrained to ly a watertight, neutrino-free explan- spiral along the field lines toward the ation for all the positive results was detector. As the electrons move into needed. lower fields, their momentum vector

18 WINTER 1992 swings forward, and they strike the analysis of the Oxford experiment in 0.010 Si detector at near-normal incidence. search of a neglected effect, and Most electrons that backscatter from discovered that scattering from the 4- the Si are reflected by the field for intermediate ("antiscattering"!) another attempt-the result is a baffle had been overlooked by Hime t 0 much reduced backscatter fraction. and Jelley and produced a significant No collimators are needed to define distortion in the spectrum (at the at the detector. It is set several-percent level). (See the the spot size -0.010 purely by the source size and the illustration on page 17.) Yet their ratio of fields at source and detector. calculation did not seem to produce 0.005 Extremely low-activity (thin) sources a spectrum that agreed with the are usable because the efficiency for known beta spectrum shape in the 0 electrons is essentially 50 percent. absence of a heavy neutrino. if any are Hime, by then at Los Alamos (For photons, however, -0.005 emitted by the source, the efficiency National Laboratory, took up this is less than 10- 4). Finally, andperhaps tack and found that scattering from 120 140 160 most significant, no energy- the intermediate baffle did indeed Energy (keV) dependent shape-correction param- produce a distortion, but that and Abashian had over- eters are needed. The Argonne group Piilonen Data from the Argonne device. Top: 35S obtained an upper limit of 0.25 % at estimated its size. For a mono- spectrum residuals, with no evidence of 95% confidence level on the 17-keV chromatic electron line, the dis- a kink at 150 keV. Bottom: spectrum branch. The clincher was a secondary tortion took the form of a broad, residuals from a mixed source, in which experiment to demonstrate that they asymmetric peak at an energy 93- the distortion caused by the 1.3% of 14C 35 had the sensitivity to see a spectral 95% of the full energy. When the admixed with S can be seen. distortion had one been present. They cadmium 109 (109Cd) calibration prepared a source in which 1.3% of lines were re-examined, the structure the activity was 14C, and the balance was clearly present (see illustration 35S. The data (figure on right) easily on the next page). Including it in the revealed the presence of the lower analysis of the 35S data gave X2 values endpoint activity (because of the without the heavy neutrino as good spectrum shape, a harder job than as had been obtained before only by finding a heavy neutrino). including the 17-keV neutrino. Hime also re-examined the Guelph data, in which no baffles were used. ECAUSE THERE WERE SO many Ordinarily, one would be entitled to unrelated but concordant posi- hope that such radically different tive experiments, physicists had not geometries would offer some pro- seriously examined the possibility tection against overlooked scattering that each such result might be due to effects. As it happened, however, the a different artifact. But with the silicon detector was surrounded by Chalk River, Tokyo Institute for an aluminum ring that produced Nuclear Studies, and Argonne quite similar effects. Hime submitted demonstrations that kinks were not a paper describing his new obser- intrinsic to beta spectra, there were vations, effectively demolishing the few alternatives. Leo Piilonen and strongest experimental evidence for Alexander Abashian began a detailed the 17-keV neutrino.

BEAM LINE 19 HAT, THEN, OF ALL THE may appear, the disparate evidence o ( x other experiments that report for a 17-keV neutrino at 1% mixing W 4 a heavy neutrino? The tritium puzzle may all be due to different, unrelated 0 remains unsolved, but it is generally artifacts. U , u felt that if an effect can be reduced a What has been learned from all of factor of 3 by theoretical corrections, this? Is it a black eye for physics? Far x it is too model dependent to be reli- from it; an enormous amount has 0 L~ 2 able. Wolfgang Stoeffl and Dan been gleaned from this experience. 1C Decman at Lawrence Livermore In 1985, no theorist would have given 30 50 /U Laboratory constructed an appara- much credence to the possibility of a Energy (keV) tus to allow observation of the entire 17-keV neutrino. Now, thanks to beta spectrum of gaseous tritium hundreds of theoretical studies, we Hime's reanalysis of 109Cd conversion- with excellent resolution down to have a much clearer idea of what we line data. Part (b) of the figure is a 100- fold expansion of the vertical scale zero energy. They do, in fact, see an know for sure and what we were shown in part (a); the extra structure excess of events below 1.6 keV, but merely assuming. A 17-keV unstable caused by scattering from the intermedi- it is possible that these are all shakeoff neutrino can be accommodated in a ate baffle is shown. Previously known electrons from autoionizing states of number of schemes that do no effects (source and detector backscat- HeT+ . The excess does not have the violence to any experimental fact. tering, energy loss in the detector spectral shape of a heavy neutrino, Indeed, in some scenarios it helps contact, and collimator penetration) but might appear to at low resolu- greatly (for example, in the formation make up the bulk of the tail structure. When those effects are taken out, the tion as in Simpson's experiment. of intermediate-scale structure in the remainder (bottom) shows a good fit to Still, shakeoff processes would not universe). Experimentally, the art of the calculated scattering from the affect calorimetric measurements continuum beta spectroscopy has intermediate baffle. like Simpson's at all. been advanced two orders of magni- The 14C data from Berkeley re- tude, and the possible pitfalls iden- main unexplained, too, but there are tified, although pragmatically it will curious aspects of the data that raise be some time before the reliability of doubts. There is no way to introduce such experiments is again widely a calibration source of electrons accepted. Most important, the entire directly into the detector where the process has been carried out in the 14C is-instead, external photon proper scientific tradition, with free sources must be used. For unknown exchange of ideas, publication in reasons, photon sources and the 14C refereed journals, and none of the behave differently with respect to press conferences or "secret the division of signals between methods" that so tarnished the cold- central and outer ("guard") regions fusion debate. of the detector, which is a warning The search for neutrino mass goes that the true electron response func- on, motivated by the conviction that tion may not have been determined. the Standard Model's prohibition of The report by a Yugoslav group of a it is arbitrary and wrong, and spiced positive effect seen in internal brem- by this vivid reminder that nature sstrahlung emission from 7 1Ge guards her more important secrets appears to contain an invalid statis- with a cunning that borders on tical analysis, and may not have even "mean." the 2-standard-deviation accuracy claimed. Thus, incredible though it o

20 WINTER 1992 I i /::: 'I

oy I

A review of Workshop on E C( r\ / r i i\ I\ I I Ii I I I\ I\ ,\ I i------B-I..\..------d---...... Q...\-··········-··e·····\... I i i i iI\ I ---- 1 i -···--d I n i\\ ii I i I ..1..1...... !. ii ----X---·------ -·--·--·r)······l···.···-··-·····B··------ ---\------i-- .\...... 0... 1 -··-···-····, -···!·······-·-······a·· h i.....·-- ··-d-:\ ---3-" -- 8-·\··-.-.--B- i...... ·) 5· ------··4··--·-"·-··a-----l··- --··---(1 a·-····\-····-· ""D""'r----·-l ; i a i ' L .i ::-!:L: i .:I --\-----9---i---- -o-·-i----·---a O' 'P i a i I- -· \··· ·-r · - I--· ---· - -- 8 )::: 1- I-- -- *-!------n--\----a-- ,--·-·--····-\1)1 i -3---1io·i·-··B - -a 1 ---8---\-- ---a-·--·--------··--· ---t---.- a n a o- " "I---a--\- i B o ·i B B a BQ i f .\ O i O a I a i1 na i , I ,a 1a " "· i1 a· · i -----w-- ------·n----lr_·i ra \:·:· gQ a i: at, i I " a r) a_1 1 a a 3 a 1 a 8 aB aa f a iti f I I I fY aB B o P3 1 9B P i a a I 1 I i r I I I i I , i I I i I I I i I i iI ii I , I I i I , I I iii I r I i i ii i i i i I' ii i i' i ii 1I i I i , I I i i 1 i i ' i iI i i i i I i i i Ii i i I t j r i Ijl i I'i i :ri i I II i 1, \: \' ' \ I V i L/ v V \i i:i/ ii i/ v \i i/ ji';I i/jj V i\I ;I

The first plenary session opened with a particle physics overview by Peter Zerwas from DESY. The rest of the day was spent on presentations for all the linear collider programs around the world. On the morning of the second day, we broke into our working groups and from then on the pattern became quite regular, with plenary sessions for one hour every morning followed by meetings of the working groups during the rest of the day. The tasks outlined for the various groups were self-explanatory. Group 6 was new in the LC series, and its purpose was to pull together the parameters from all the other groups, examine the similarities and differ- ences between the various techniques, and attempt to get an idea of the costs for the different machines. The major projects reviewed during the workshop were TESLA (DESY et al.); the DLC (DESY/Darmstadt); the JLC (KEK); the NLC (SLAC); VLEPP (INP, Protvino); and CLIC (CERN). The tentative parameters of these machines at 500 GeV center-of-mass energy are listed in order of increasing rf frequency and approximately decreasing beam power in the table on the facing page. All the luminosities are in the range between 1033 and 1034 cm - 2 sec- 1. As can be seen, all the machines except for VLEPP and CLIC now have large numbers of bunches per rf pulse. This choice leads to relatively lower charge per bunch (except for TESLA and DLC) and makes these bunches more manageable along the linac and at the final focus (FF). The rf pulses are somewhat longer and thereby make the klystron modulator pulses more efficient because the energy lost in the rise and fall time is less significant. For TESLA and DLC the vertical beam size oy at the FF is relatively large and more forgiving than for the other machines. On the other hand, the beam power is more difficult to handle, and the de- mands on the positron sources and damping rings are more severe. Working Group 6 provided a stimulating experience because it gave us a chance to interview all the other groups to find out where the main design obstacles were being encountered. It also forced us to look at both

22 WINTER 1992 ...... 6...... ~ a·· ·· ....i...... ,--a -\--·-···--u- s·····-~------a ...... -* Y:Ij::I:...... ---~~--)-·--····-·U O..~~~~~...... '...1 a ··-t--·-·-a-l-··· ... ·- ·-·a -c···-l----·s a~~~~~~~~~~~~~la.....

Overall Parameters for Linear Colliders Discussed at LC-92.a

TESLA DLC JLC-I(S) JLC-I(C) JLC-I(X) NLC VLEPP CLIC

Initial energy (c.m.) (GeV) 500 500 500 500 500 500 500 500 RF frequency of main linac (GHz) 1.3 3 2.8 5.7 11.4 11.4 14 30 Nominal luminosity (10: -33cm-2 s 1) 2.64 4.0 7.3 5.4 6 12 0.7-2.7 Luminosity w/pinch (1033cm-2- 1) 11.1 6.5 4.4 6.5 6.3 8.2 15 2.2-8.9 Linac repetitionrate (Hz) 0 50 50 100 150 180 300 1700 No. of particles/bunch at IP (1010) 5.15 2.1 1.30 1.0 0.63 0.65 20 0.6 No. of bunches/pulse 800 172 55 72 90 90 1 1-4 Bunch separation (nsec) 1000 10.66 5.6 2.8 1.4 1.4 0.33 Unloaded gradient (MV/m) 25 2 1 22 40 40 50 108 80 Beam loaded gradient (MV/m)b 25 17 18.4 32.5 28 37.6 96 78-73 Active two-linac rf length (kim) 20 30 28 16.7 17 14 6.4 6.6 Section length (m) 1.04 6 3.6 2 1.3 1.8 1.01 0.273 Two-linac number of sections 19232 4900 7776 8360 13600 7778 5200 24000 Two-linac number of klystrons 1202 2450 1944 4180 3400 1945 1300 2 Number of sections/klystron 16 2 4 4 4 "12000" Klystron peak power (MW) 3.25 150 85 45 70 94 150 700 Klystron pulse length (Msec) 1300 2.8 4.5 3.6 0.840 1.5 0.7 0.011 Pulse length to section (usec) 1300 2.8 1.2 0.6 0.210 0.250 0.110 0.011 Pulse compression ratio - - 3.7 6 4 6 6.3- Pulse compression gain 2.4 4.2 3.2 4 4.22 a/x range (input/output cavit) 0.15 0.154/.08 0.13 0.160/.120 0.236/.138 0.210/.147 0.140 0.2 Beam power/beam (MW) 16.5 7.5 1.6 3.6 3.8 4.2 2.4 0.4-1.6 Total two-linac power(MW)c 137 114 106 193 86 152 91 175 Damping ring energy (GeV) 3 or 14 3.13 1.98 1.98 1.98 1.8 3.0 3

Final Focus: YEx/YE (m-radx10-8) 2000/100 500/50 330/4.5 330/4.5 330/4.5 500/5 2000/7.5 180/20 xj,*/fl(mmy^^^^):0 -t :- - _I I _I 10/5 16/1 10/0.1 10/0.1 10/0.1 10/0.1 100/0.1 2.2/0.16 */(a P(nm) 640/100 400/32 300/3 260/3 260/3 300/3 2000/4 90/8 u*(im)-_ ^^ ^^ 1000t80 6 500 80 67 100 750 170 B(%)N3 4.8 10 8.1 4.5 3 9 6 Upsilon00 i00 ; ; 0 00 0.035 0i 0Q M0.058 0.20 0.20 0.15 0.1 0.06 0.15 Disruptions Dx/Dy 1.2/7.9 0.69/8.6 0.13/13 0.13/11.5 0.07/6 0.08/8.3 0.4/200 1.3/15 Crossing A IngleatIP(rd) 12 2 7.3 8 7.2 3 N.A. 1

aMost parameters are subject to change. bBefore applying further gradient reductions for off-crest running, BNS, etc. (VLEPP excepted). CDESY bases its number on a combined klystron-modulator efficiency of 45%. KEK and SLAC have assumed this number to be closer to 35%. In addition, SLED-I (used for JLC-I(S)) and SLED-II (used for JLC-I(C), JLC-I(X), NLC and VLEPP) are assumed to be about 65% efficient. Power for klystron focusing is not included.

BEAM LINE 23 n~~~~~~~~~~~~~~~~~~~~~~~~~~p i~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~...... ~ ~ ~ ~ ~ ~...... ~ ~ ~ ~ ~ ~ ~ ~ ...... i ...... ~ ~ ~~~ ~ ~ ~ ~ ~ ~ ~ ~ ...... I :i k '~ ...... -...... ~ ~ ~ ~~......

advantages and disadvantages in order to draw mode" cavities proposed by Tsumoru Shintake. These comparisons. Given that the atmosphere at the cavities, of which he had brought some aluminum Workshop was very congenial, the various proponents models, created a considerable amount of interest and felt unthreatened and showed a refreshing candor and discussion. The S-band design is similar to the DLC sense of humor. The TESLA proponents, because of except that it uses SLED-I (SLC-type) pulse compres- their low rf frequency and large beam apertures, strongly sion. The X-band design resembles the NLC (with advertised their advantages in wakefields and alignment slightly lower luminosity) but is somewhat longer tolerances, large bunch spacing and small energy spread, (17 km instead of 14 km), uses more klystrons, lower but were not embarrassed to admit that in order to pulse compression ratio, a lower repetition rate, and become competitive they would have to increase their thereby a smaller amount of ac power. accelerating gradients by a factor of at least four above In turn, the NLC is a shorter machine with a higher existing technology, eliminate any "dark current, " and gradient and a somewhat more difficult klystron (94 MW reduce their linac costs by a factor of 4-5. Also, given peak with 1.5 usec pulse instead of 70MW with 0.84 usec the large beam power, they realized that their damping pulse for the JLC). With a lower beam current per rings and positron source would have to be rather bunch, Gaussian higher order mode detuned structures robust. Similar remarks were made about the DLC by and short betatron wavelengths, it can stand low fre- our DESY/Darmstadt colleagues. Indeed, their machine quency ground motions but still has to meet stringent is also fairly long (30 km), has heavy beam loading, but (ay = 3 nm) focusing conditions at the final focus. looks all the way like the S-band pre-accelerator of the VLEPP is the highest gradient and shortest machine. NLC or JLC, thus not needing a second compressor. Gus With a very high current single bunch, it does not need Voss from DESY, a strong proponent of S-band, flattered detuned or damped structures and its "autophasing" his SLAC colleagues by claiming that one of S-band's scheme is supposed to take care of longitudinal and advantages was indeed that SLAC had already built transverse wakefields. The linac is to be realigned 10% of his machine, to which the author gave him a adaptively (with the beam) at a 10 Hz frequency. hand and told him that it was really 20% (!). The DESY/ Vladimir Balakin, when asked if he saw any disadvan- Darmstadt proponents, together with their INP tages to his machine, unflinchingly answered: "Yes, colleagues, also pointed out that the positron source too many new ideas!" problems could be solved by a bootstrap operation, i.e., CLIC, finally, is the highest rf frequency machine by sending the spent e- beam emerging from the with the lowest beam power. It has a high repetition interaction point through a wiggler, making gammas rate and only uses two (BIG!) klystrons. Admittedly, which in turn would produce e± pairs from which the they are not trivial devices, using superconducting positrons could be sent directly into the e+ damping accelerating cavities and numerous room temperature ring, at the appropriate energy. transfer cavities to generate the 30 GHz rf power. The Concerning the JLC, KEK presented three so-called wakefields and fabricational tolerances are severe. On JLC-I designs, respectively at S-, C- and X-band. The S- the other hand, the two klystrons are semi-passive band and C-band designs, in particular, were invented devices, and they can be located in the same tunnel as in response to a Japanese constituency of physicists the accelerator proper. who would like to know if it is possible to build a 300- Speaking of tunnels and civil construction costs, GeV center-of-mass linear collider well before the end there was a lively discussion about whether conven- of the century. The C-band design is based on a SLED-II tional linacs could coexist in one tunnel with a large pulse compression and cylindrically symmetric "choke- array of klystrons and modulators requiring frequent

24 WINTER 1992 Ilk A I-\ '\

..,... \ i d \ D d 1 d~~~~~~~~~~~~~~~~~~...... i~~~~~~~~~~~~~~~~~~r\...... iI ...... ". ' ...... a---...---r------~~~~~~~~~~~~~~~~~~~~~·--·------····o.s::--l·:·::·:-:--s--; 1 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~-...... ····-g ·--···~---e···· !-----o----··- r-·· ·· ·-·-- ...... ~....8-:...... --- :-L.._...._=...:...... ·- ·- ~l ...·...... s.: "...'.*i I Il - ---ra ·-

maintenance. The pros and cons were presented by the various protagonists, but the feeling was that more work and ideas are still needed to resolve this question. As far as overall machine costs were concerned, it was pointed out that the DLC S-band linacs, in terms of length and number of klystrons and modulators, were equivalent to the present SLC linac multiplied by a factor of 10. Using original SLAC costs escalated to 1992, a machine of modular construction such as the SLC linac, about 30 kilometers long and with about 2450 klystrons, would cost about $4 billion. The working group, however, reached the conclusion that this type of scaling could only give a rough calibration of the overall cost, and that it is premature to try to get much better estimates until technology R&D and demonstra- tion projects at the various frequencies were more advanced. There also was considerable discussion on how the various machines could be expanded from 1/2 to 1 TeV energy in the center-of-mass. The NLC and DLC could be upgraded by quadrupling the peak rf power by doubling the number of klystrons and doubling the peak power, respectively, of the klystrons or via SLED-I pulse compression. To keep the ac power from increasing excessively, one could reduce the repetition rate (NLC) or the pulse length (DLC). JLC-I C-band and TESLA might use a combination of peak power and length increases. VLEPP and CLIC could be doubled in length. The Workshop ended August 1 with nine summary talks that will appear in the proceedings. Despite a healthy sense of competition among the protagonists, the dominant feeling was that the Workshop had strengthened the bonds among all co-workers to the point where they could, at least for now, collaborate towards the best future linear collider. The box on the right lists the major systems construction and tests to which the community can look forward in the next three or four years to evaluate the various technologies and converge on some choice. 0

BEAM LINE 25 O: for

26 WINTER 1992 Meanwhile, not far from Hubble's office on Santa Barbara Street,* Walter Baade and Fritz Zwicky were busy discovering and defining the class of stellar explo- 1000 -6Ua1) sions they called super-novae. In 1934, Baade and 5 Zwicky pointed out that about 1053 ergs per event were E)C IrI.A.lirA] fA) on_ irYU-JLaU LU a0- 50C count for the 12.- 0 phenomena (in- C) cluding acceler- C ation of cosmic

rays) and "with 6 all reserve" ad- 0 106 2 x 10 vocated the col- Distance (parsecs) lapse of a normal star to a The graph of redshift (assumed due to velocity of recession) neutron star as vs. apparent brightness of galaxies (an indicator of distance) the source of as it was known to Edwin Hubble in 1929. [Adapted from Proc. Nat. Acad. of Sci. A 15, 168 (1929).] energy. The hy- phen disappear- ed in 1938, and, in 1939, Olin lines in their spectra. He used apparent brightness of Wilson and, in- galaxies as an indicator of their distances (faint = far), dependently, for which the astronomer's standard unit is the mega- Zwickyhimself parsec (1 Mpc = 3x1024 cm), and converted measured suggested that redshift, z = A/;1L, into velocity, v = cz, which we supernovae normally quote in kilometers per second. H is the slope 1 . .d mignt eventu- of the linear cz vs. distance relation, and so acquires the ally prove to be good distance indicators and thus wondrous units of km/sec/Mpc, thereby effectively calibrators of Hubble's constant. obscuring the fact that H is really a reciprocal time. What is Hubble's constant; and why should we care Even before Hubble, Carl Wirtz in Germany, Knut whether, like the national debt, it is 526 quintillion Lundmark in Sweden, and H. P. Robertson in the U.S. jillion or only 50? Hubble's eponymous discovery was had perceived velocity-distance relations in the data, the linear relationship between the distance to galaxies using several different distance indicators and not outside our Local Group and the redshifts exhibited by always seeing the relationship as linear. There has never been any enthusiasm for relabeling the propor- *Once the office and administrative space for "Mt. Wilson tionality constant W or L or R, and I do not propose it Observatory," the building at 813 Santa Barbara Street, Pasa- here, despite some feeling that fellow Scandinavian dena, later served the same function for "Mt. Wilson and Palomar Observatories," then for "The Hale Observatories," Lundmark, at least, got a raw deal. The velocity v = cz followed by "Mt. Wilson and Las Campanas Observatories" and, is not a Doppler shift, no matter what you may have most recently, for "The Observatoriesof the Carnegie Institute of read in elementary text books. But that is a different Washington." Contraryto appearances, the building has not moved and remains a brisk half-hour walk from Caltech. story (well told in several places by Edward R. Harrison).

BEAM LINE 27 The time, 1/H, is precisely what it would take the All astronomical distance measurements fall into universe to expand from infinite density to the present three classes: (a) roundtrip travel time of light or radar conditions if no acceleration or deceleration occurred signals (useful only in the solar system); (b) comparing along the way. The real expansion time ("age of the the apparent and real brightnesses of objects; and (c) universe") is obviously shorter than 1/H if there has comparing the angular diameters and physical sizes of been deceleration, longer with acceleration. Either of things. By way of clarification, if you know that a star these is possible according to the best available theory (or whatever) emits L erg sec- 1 and you receive a flux of gravity (general relativity), but the probable range is of F erg sec-lcm-2 from it, then d = (L/4tIF)1/2 . And if really only from 2/3 to about 4/3 of 1/H. In Hubble's you know that something is A miles across and sub- day, the oldest earth rocks clocked in at a bit less than tends an angle of a radians in the sky, then d = A/a 2 billion years, and his 1/H was 1.8 Gyr, excellent miles. The former gets called a standard candle, the agreement in his view. Today, the oldest objects we latter a standard meter stick. The units actually used know are the globular clusters (see the Fall 1992 Beam are seconds of arc for angles, parsecs for distances, and Line, Vol. 22, No. 3) at 13-17 Gyr. Thus, whether we astronomical units (1 AU = 93,000,000 miles) for sizes. have reasonable agreement or a glaring discrepancy Once again, one part is relatively straightforward- depends very much on whether 1/H = 20 or 12 Gyr measuring the flux you receive or the angle subtended; (H = 50 or 85 km/sec/Mpc). The numerical value of H the other part (knowing the real values) is much also enters to first or second power in calculations of harder. And again, there are only a couple of possibili- the mass density needed to produce flat space-time ties. You may think you know real luminosity and size ("close the universe"), of the sizes and brightnesses of either because you have understood the physics re- galaxies and quasars, and of all sorts of other things sponsible for energy production and structure or be- that astronomers earn their precarious livings by giv- cause you recognize the object as belonging to a famil- ing or going to seminars about. iar type, whose properties you know from less distant examples. Being able to measure large distances is particularly important because we need to get outside the Local Group of galaxies to see Hubble PROBLEM THE CORE OF THE expansion at all, and to much larger distances of 30-100 Mpc before it dominates over random If we care so much, why can't we get our act together, motions of galaxies milling around each other in establish the correct value of H, and get on with trying clusters. to understand the structure and evolution of objects Hubble's approach, followed by a majority of within a universe of age (1±1/3)/H? Measuring red- later H-evaluators, was to start with common shifts is, if not precisely easy, at least not a topic of types of stars, recognizable by their patterns of dispute for galaxies or quasars displaying more than a variability, and to work outward calibrating couple of spectral lines. The hard part is measuring brighter candles and bigger sticks along the way, distances. All of the historical changes in best-buy until whole galaxies are the distance indicators. values of H and most of the present unpleasantness And from Hubble's time to the present, a few arose from disagreements about the extragalactic renegades have been trying to figure out some way (meaning outside the Milky Way) distance scale. Large to jump over most of the steps in the distance distances mean small values of H and, therefore, long ladder without breaking either their necks or their time scales; smaller distances the opposite. reputations.

28 WINTER 1992 THE DISTANCE LADDER AND BUNGEE of the gas cloud, then only one choice of distance will CORDS yield consistent results, because the two processes depend on different powers of gas temperature, density, Modern step-by-step calibrators (either candles or meter and distance. Joseph Silk and Simon White pointed this sticks or both) include Cepheid and RR Lyrae variable out in 1978. It is possible to doubt whether the Sunyaev- stars, novae, planetary nebulae, supergiant stars, globu- Zeldovich effect has ever really been seen, but lar clusters, ionized regions around young stars, arms examination of the cluster Abell 665 yields H = 45, of spiral galaxies, whole galaxies, and even collective again with sizable and uncertain error bars. properties of clusters of galaxies. The reviews by Jacoby Third comes John Tonry's method of studying et al., van den Bergh, and de Vaucouleurs mentioned in surface brightness fluctuations across the faces of gal- the box on page 31 will tell you at least as much as you axies. The further away a galaxy is, the more stars will want to know about these. Being sure that the distant fall into a given 1 arc second x 1 arc second. Thus the objects are really physically like the nearby ones is N 1/2 /N statistical variations in brightness from pixel to much, but not all, of the problem. The effect of galaxies pixel will scale as (distance)-1. Radio astronomers may milling around each other is also somewhat in dispute. recognize this as the optical analog of the method for In general, most applications of the step-by-step method counting faint, confused* radio sources, called P(D) lead to values of Hin the 85+10 km/sec/Mpc range; and and developed in the early 50s by Peter Scheuer in it is possible to make the climb up this distance ladder Martin Ryle's group at Cambridge. Galaxies studied sound rather dull. this way so far are not distant enough to give H directly, In contrast, the four attempts to leap across cosmic but they are coming in at the short end of possible distances in a single bound each include at least one distance scales, implying a large H eventually. clever idea that it is easy to get excited about, though not necessarily trivial to explain. And the majority produce Hubble constants of 50 ±10 km/sec/Mpc. SUPERNOVA DISTANCE SCALES First, if a distant quasar has been gravitationally lensed by a galaxy between it and us (Beam Line, Spring When Zwicky and Wilson proposed supernovae as 1992, Vol. 22, No. 1), then the image on the sky tells us distance indicators back in 1939, they had in mind just an angular size, while the time between brightness another step on the ladder, though indeed a very big changes of the parts of the image tells us physical size one. At maximum brightness, SNe can briefly outshine (because light travels at the speed of light). Sjur Refsdal their host galaxies, and so be seen at distances of (another Scandinavian, but now working in Hamburg) gigaparsecs. In principle, then, supernova luminosities suggested this possibility in 1964. The first lensed could be calibrated on events in or near the Local Group quasar turned up in 1978, and its assorted variations and take us out to large redshifts in a single step. Olin have led to an estimate of H = 40, but with very large Wilson is a gentleman of the old school and would error bars. These can be reduced only by examining never say "I told you so," but we are pleased that he has other lensed quasars with simpler geometry. the opportunity to do so. A couple of possible bungee Second, the hot, intragalactic gas in rich clusters will both emit X rays and inverse Compton scatter from the 3K microwave background radiation *Confused in this context means so crowded on the sky that you photons cannot resolve the sources individually. Source confusion ("Sunyaev-Zeldovich effect"). If you observe both the frequently leads to astronomer confusion (but then you wouldn't emission and the scattering and try to infer the properties really want to see individual astronomers, would you?).

BEAM LINE 29 cord modes for supernovae have turned up more re- Hubble Space cently. Telescope images of a small region main flavors, Type II, Supernovae come in two within the faint which we continue to blame on neutron star formation spiral galaxy IC a la Baade and Zwicky, and Type Ia (no, I am not going 4182. The arrow in to explain the a just now), which seem to derive their the upper image energy from nuclear sources -the burning of about points toward one nf th P/ :Pnhpni[ 1033 grams of carbon and oxygen mostly to iron and ul i, .C7 --,/ % , l , nearby elements. Each type could, in principle, be a variable stars that have been ob- yard stick of either the "we understand the cosmic served by the HST physics" variety or of the "we recognize the class" in IC 4182. The variety. For instance, if SN Ia's really are the advertized arrow in the lower nuclear explosions, then the same reactions must yield image points to the the energy both for the light output and for blow-off of same Cepheid the ejecta we see. These scale differently with distance, variable as it and a consistency argument leads to H = 61+10. When appeared a few days days later, after it had brightened in its Using the known relation critics say that this is "model dependent," they don't periodic cycle of pulsation. between a Cepheid's intrinsic brightness and its period of mean that it is wrong. But the number will be precisely pulsation, Allan Sandage and his colleagues were able to right only if that is how Type Ia supernovae really determine the distance to IC 4182 (about 5 megaparsec). work. But see text for a caveat. Type II supernovae are known not to be standard candles. SN 1987A, for instance, was a factor of a then we know the distance to the galaxy and the hundred fainter than some others. But their expanding supernova and, therefore, the real brightness of Ia's and eject may constitute a meter stick of the "understand- the distances to other, larger-redshift galaxies containing the physics" sort. The linear size of the emitting ing them. Sandage et al. find that SN Ia's at peak material is just the measured expansion (doppler-shift) brightness outshine our sun by a factor of more than velocity integrated over time since the explosion. Ap- 1010. And H = 45±9. parent brightness plus color temperature tell you the As usual, there is a caveat. The same SN Ia and angular size, and, voila, a distance not dependent on parent galaxy, if put at the distance implied by the any of the traditional calibrators. Early work along brightest (non-exploding!) stars, are much closer. In 9 these lines led to largish values of H, but the most that case, the supernova peaked at "only" about 3x10 recent published one was H = 60±10 km/sec/Mpc. solar luminosities, and H = 86±14. One implication is Type Ia supernova mavens believe they can identify that either the Cepheids or the brightest stars in IC a subset of events that are standard candles, presum- 4182 are not the same brightness as the corresponding ably because they all burn about the same amount of objects near the sun. Another implication is that super- carbon and oxygen. SN 1937C belonged to that subset. novae have still not quite told us the value of H It has recently flared up again, in press releases if not in unambiguously. the sky. Allan Sandage and his colleagues have used the To summarize, most of the traditional, step-by-step much-despised Hubble Space Telescope to locate and approaches to Hubble's constant yield relatively large study Cepheid variables in the parent galaxy IC 4182. If values, and so apparently do Tonry's method and one those Cepheids are like the ones in the Milky Way, way of calibrating Type Ia supernovae. But gravitational

30 WINTER 1992 lensing, the Sunyaev-Zeldovich effect, and three of the year, when they will be reporting to you themselves in four ways of using supernovae, including a calibration these pages.** on Cepheid variables, produce H's between 45 and The same data will have the potential to resolve a 60 km/sec/Mpc. still more ancient issue: is the universe really expanding? It is possible to imagine (and Hubble did) a sort of tired light effect, in which photons lose energy merely THE NEXT MILLENIUM by traveling long distances in a static universe. You would still see a linear redshift-distance relation, but Thirty-some years ago, Sandage began a review of not other effects of expansion. In particular, if SN Ia cosmological parameters by saying that the value of fade with a half-life of 77 days near us, distant ones will the Hubble constant is not presently well established. show the same time constant in a tired light universe. To this we can only say, amen. Some things will get In an expanding universe, on the other hand, the 77 better. Another SN Ia host galaxy is within range of the days is also redshifted or time dilated by (1 + z), and the present HST. The refurbished telescope should see distant SN 1992bi, if followed for a long enough period, Cepheids in the Virgo cluster of galaxies, pinning down would decay with a half-life of 112 days. I'm prepared an important intermediate rung on the ladder (at 14 or to bet (as Hubble eventually did) fairly strongly in favor 22 Mpc, or somewhere in between). More lensed qua- of expansion, but not at all on the value of q0 they will sars have been found and will be studied, traditional find! calibrators be further explored, and so forth. If, as is frequently suggested, we end up with convergence (or exhaustion) somewhere near H = 75 km/sec/Mpc, then the age problem is a real one. Continued disagreement **Editors' Note: Carl Pennypacker of the Berkeley group tells us among the methods would be equally interesting. that within the next three years they expect to find about 30 far-sighted colleagues are looking be- distant supernova events, which can result in a measurement of Meanwhile, the geometry/mass of the universe to an accuracy of about 30%. yond mere measurement of H. The Berkeley distant supernova search has found one event, SN 1992bi, in an anonymous galaxy at z = 0.457, far enough away that the distance associated with a particular redshift de- 0 pends on the geometry of the universe. One swallow has not yet made a deceleration parameter,* but they expect to find several more distant SN Ia's by later this

*A homogeneous, isotropic expanding universe can be fully described by a single function, a(t). The instantaneousvalue of Hubble's constant is simply H = (da/dt)/a. The second derivative can be put into dimensionless form as (d2 a/dt2 )a qo- (da/dt)2 called the decelerationparameter. Its numerical value is 0.5 for a universe that slows monotonically to zero expansion speed in infinite time and 0.0 for a matter-free, coasting universe. The truth probably falls somewhere in between.

BEAM LINE 31 GUEST EDITORIAL

T HIS ISSUE OF THE BEAM LINE marks a door, to adequately communicate with everyone as transition for its founder and present co-editor, to what we are doing and what we are about. The Rill Ki1rr wAihn .UL. . . vv LLJ Beam Line went a long way towards filling that gap. will retire from At the same time, because of its high standards, it SLAC and the co- also became a highly reputable and recognized editorship of the publication for the community at large. Beam Line on Bill Kirk joined SLAC in 1956, that is one year February 28. before the laboratory was even proposed. His initial This guest edi- job was in response to an advertisement by the then torial is to en- Director of the Microwave Laboratory, Professor lighten Beam Edward L. Ginzton, for an assistant, and Bill immedi- Line readers on ately became Ed's right-hand man. When the group Bill's 37 years of planning what was then Project M started meeting contributions to during off hours in order to plot the creation of the particle physics world's leading linear electron accelerator laboratory, and to this jour- Bill coordinated the operation of that group and nal in particular. documented its many conclusions. That early phase In fact, for much of his work culminated in the "Stanford Proposal for of the time since a Two-Mile Linear Accelerator," dated April 1957, the birth of the whose prose is largely Bill's work. Characteristi- BeamD' . LineT in*n: cally, his name does not appear in the proposal, 1974 to its B l Kirk although acknowledgements to other contributors reincarnation in 1988 Bill Kirk was the Beam Line. are profusely strewn about in it. The rest is history- He wrote almost all its text. I watched him doing the the proposal was submitted to three diverse govern- layouts, composing the prose, and doing the typogra- ment agencies, received the support of the then phy in a variety of fonts with the meticulousness Atomic Energy Commission, and from there the which, as far as I can tell, he learned from building path continued through Congressional authorization the "perfect" model railroad at home. and construction of SLAC. The original Beam Line as conceived and ex- Bill has served the Laboratory in a variety of ways, ecuted by Bill was a true innovation. It became a always unobtrusive and always a perfectionist. He popular science journal of high tutorial value. For was the editor of innumerable conference proceed- instance, the April 1978 "High Energy Physics-An ings and his services in that respect were very much Introduction" was widely used to expose lay persons in demand by laboratories other than SLAC through- to the subject. In that spirit there were editions out the world. He mediated innumerable "people describing the workings of PEP, the general operation controversies," usually successfully. He wrote of storage rings, the history of the discovery of the planning documents for the Laboratory and had the psi particles, etc. But then there were also more knack, unusual for such documents, of having "newsy" editions which kept people at SLAC in- content take precedence over form, a practice only formed about the important goings on. very rarely followed today! It is always difficult in a place like SLAC, where Bill's joining what became SLAC and his contin- no final products other than paper come out the back ued service there was to a large extent unplanned.

32 WINTER 1992 FROM1 THE EDITORS' DESK

He had many other ambitions: he IT'S THE PROCESS and not the product. wished to be a novelist and an During the past 30-odd years the people who study the bits and pieces athlete, among other things. The of nature have put together, experimental bit by theoretical piece, a love of those avocations affected wonderfully successful picture of the underlying structure and dynamics much of his work at SLAC. He of the world. This Standard Model of the elementary particles and the insisted on a high level of literacy forces that act upon them seems to work (predict, explain) like and respect for the English lan- Gangbusters. But not everyone is happy. The Model has too many guage throughout the Laboratory, arbitrary parameters. It is unaesthetic. The search is on to find something and particularly in reference to the that doesn't Fit. (See Hamish Robertson's article starting on page 15 for Beam Line. He organized, presided the story of something that almost didn't Fit.) So this is a time to think over, and participated in a multi- about "Where do we go from here?" (James Bjorken, page 8). tude of athletic events, including It's the journey and not the destination. the annual SLAC slugfest between Meanwhile, over in another part of the asylum, the astrophysicists of the theorists and experimentalists our acquaintance are all smiles. What does it mean that the Cosmic in baseball. He himself continued Background Explorer (COBE) has found the eagerly sought freckles on the other athletic activities, both face of the Universe? (Joel Primack, page 1). And the much-maligned personally and with his children. Hubble Space Telescope works (if not yet as well as Gangbusters) at least Burt Richter has placed Bill Kirk well enough to stretch one of the important cosmic distance-measuring in charge of all the Laboratory's scales out to a galaxy beyond the Local Group (Virginia Trimble, information activities. After his page 26). retirement his shoes will be hard So the particle physicists have a master Model but don't seem happy to fill, both in that position and at with it, while the astrophysicists "do not yet have a fundamental theory" the Beam Line. SLAC, the Stanford (Primack) but seem delighted with their bits and pieces of it. community, and scientific people It's the Search and not the Grail. throughout the world owe Bill thanks and admiration for a job well done, for his work on publica- tions and for keeping the SLAC Director out of unintended troubles. ^^*~·~"" Wolfgang K. H. Panofsky

BEAM LINE 33 CONTRIB UTORS

C- n i c-c jT Ec "| I

JOEL PRIMACK is a professor of JAMES BJORKEN (BJ) returned to HAMISH ROBERTSON has carried physics at the University of Califor- SLAC in 1989 as a member of the out experimental research on nia, Santa Cruz. He is a theoretical Theory Group after ten years at Fermi neutrinos for ten years at Los Alamos particle physicist and cosmologist. National Accelerator Laboratory National Laboratory, where he is a His recent research has explored the where he was Associate Director of staff member and Fellow. His back- nature of dark matter, the invisible Physics. While at SLAC from 1963 to ground is in nuclear physics, and material that appears to account for 1979, he played a central role in the between 1971 and 1981 he was on most of the mass in the Universe, development of the quark-parton the faculty at Michigan State Uni- and its role in the evolution of the model. He is currently combining versity. Presently he is a Principal cosmos and the formation of galax- his theoretical and experimental in- Investigator on the Sudbury Neutrino ies. Primack has also worked on pub- terests by proposing a new type of Observatory project, and he has been lic policy issues, most recently space detector called FAD (Full-Acceptance engaged in an experiment on the nuclear power, and he helped create Detector) at the Superconducting beta decay of tritium to measure several influential institutional Super Collider, along with a similar neutrino mass. niches for public interest scientists, initiative (MAX) at the Fermilab such as the Congressional Science Tevatron collider. Fellowship Program.

34 WINTER 1992 WOLFGANG K. H. PANOFSKY (PIEF) is Director Emeritus of Stanford Linear Accelerator Center. He has served in many advisory ca- pacities to various governmental GREGORY LOEW is Deputy VIRGINIA TRIMBLE oscillates at a agencies. He is currently chairman Director of SLAC's Technical frequency of 63 nHz between the of the Committee on International Division and a member of the SLAC University of California, Irvine and Security and Arms Control of the Faculty. In 1958 he joined Project M, the University of Maryland. A National Academy of Sciences and which was later to become SLAC, to facultative feminist, she traces her chairman of the Board of Overseers design the constant-gradient accel- ancestry back to a maternal great- of Universities Research Association erator structure for the three- grandmother who was a Rasmussen for the SSC Laboratory in Dallas, kilometer linac. Since then he has from HillerAd, Denmark. She was Texas. had many diverse assignments at the the 1986 recipient of the U.S. National laboratory and in the international Academy of Sciences Award for accelerator community. He greatly scientific reviewing and currently enjoyed being one of the conveners serves as editor of Comments on for LC92 at Garmisch. The nearby Astrophysics and associate editor of Neuschwanstein Castle, seen in the the Astrophysical Journal. background on the first page of his article, should be a reminder that others in the past have not shied away from designing and building very exotic structures.

BEAM LINE 35 DATES TO REM\ EM BER

Mar 22-26 American Physical Society meeting, Seattle, WA (American Physical Society, 335 E. 45th Street, New York, NY 10017-3483). Mar 29-31 SSC Physics Symposium, Madison, WI (Linda Dolan, University of Wisconsin, Phenomenology Institute, 1150 University Avenue, Madison, WI 53706 or SSCSYMP@WISCPHEN). Mar 29-Apr 1 International Workshop on Supersymmetry and Unification of Fundamental Interactions (SUSY 93) (Linda Harris, SUSY-93, Department of Physics, Northeastern University, Boston, MA 02115 or SUSY93@NUHUB). Apr 26-30 Second International Workshop on Physics and Experiments with Linear e+e- Colliders, Waikoloa, HI (F. Harris, University of Hawaii, Department of Physics, 2505 Correa Road, Honolulu, HI 96822 or LCSW@UHHEPG or FAH@UHHHEPG). May 3-5 1st European Workshop on Beam Instrumentation and Diagnostics for Particle Accelerators (Ch. Parthe, CERN, SL Division, CH-1211 Geneva 23, Switzerland, or DIPAC@CERNVM). May 3_7 4th ICFA Seminar on Future Perspectives in High Energy Physics, Hamburg, Germany (DESY, Notkestrasse 85, 2 Hamburg 52, Germany) May 6-8 5th Annual International Industrial Symposium on the Super Collider (IISSC), San Francisco, CA (Pamela E. Patterson, IISSC, Box 171551, San Diego, CA 92197) May 17-20 Particle Accelerator Conference, Washington, DC (C. W. Leemann, CEBAF, 12000 Jefferson Avenue, Newport News, VA 23606 or LEEMAN@CEBAFVAX). May 31-Jun 4 Physics Computing '93, Albuquerque, NM (American Physical Society, 335 E. 45th Street, New York, NY 10017). Jun 21-Jul 2 Beauty Physics at Proton Accelerators, Snowmass, CO (Vera Luth, SSCL, 2550 Beckleymeade Aveue, MS 2080, Dallas, TX 75237 or BPHYSICS@ SSCVX1). Jul 26-Aug 6 21st Annual SLAC Summer Institute on Particle Physics: Spin Structure in High Energy Processes, Menlo Park, CA (Conference Coordinator, MS 62, SLAC, Box 4349, Stanford, CA or SSI@SLACVM). Aug 10-15 16th International Symposium on Lepton and Photon Interactions, Ithaca, NY (by invitation) (Lepton-Photon '93, Newman Laboratory, Cornell University, Ithaca, NY 14853 or LEPPHO@CRNLNS). Sep 19-23 Third International Workshop on Theory and Phenomenology in Astroparticle and Underground Physics (TAUP 93), Assergi, Italy (by invitation) [Laboratori Nazionali del Gran Sasso, 1-67010 Assergi, Italy, phone (862) 437231 or FAX (862) 410795]

36 WINTER 1992