Assembling internationalscience in Japan

8 December 2008 Physics Today 2008 American Institute of Physics, S-0031-9228-0812-210-4

From to precision cosmology: The amazinglegacy of a wrong paper Michael S. Turner

Michael Turner is the Bruce V. and Diana M. Rauner Distinguished Service Profes-sor at the University of Chicago and a founding member of its Kavli Institute forCosmological Physics.

Youd have to be living in a cave inAfghanistan not to know that cosmol-ogy is in the midst of an extraordinaryperiod of discoveryperhaps even agolden age. But you might not knowthat it all started on April Fools Day60 years ago. Ralph Alpher, Hans Bethe,and George Gamow published a Letterto the Editor entitled The Origin of theChemical Elements in the April 1 issueof Physical Review. Gamow asked Betheto add his name to the paper he and hisstudent Alpher were writing to createthe author list alpha, beta, gamma;Bethe agreed. The paper markedthe birth of the hot Big Bang cosmologyand started the march to precision cos-mology. It is also exhibit 1 in my casethat an interestingly wrong paper canbe far more important than a triviallyright paper; recall Wolfgang Paulis fa-mous putdown, It isnt even wrong.

In 1948 cosmology was practiced bya handful of hardy individuals, mostlyastronomers; determinations of theHubble constant were almost 10 timesas large as they are today; the redshiftsof less than a hundred relatively nearbygalaxies had been measured; and the200-inch Hale telescope on Mount Palo-mar was a year away from first light.Cosmology is now center-stage scienceand attracts a thousand researchers,both physicists and astronomers; twoNobel Prizes have been awarded (1978and 2006, and more to come); an ar-mada of telescopes, experiments, andeven accelerators has been brought tobear on the problems of the universe;and precision cosmology is no longeran oxymoron.

Cosmic nuclear reactorIn the late 1930s, buoyed by the success ofsolving the riddle of the energy source ofstars, nuclear physicists were turningtheir attention to the origin of the chemi-cal elements. A decade later it was be-coming clear that equilibrium nuclearprocesses in stars (or elsewhere) wouldntwork, for the simple reason that the meas-ured abundances do not correlate withnuclear binding energies.

Gamow took a bold new tacknon-equilibrium physics in the expandinguniverse. If the universe began in a hot,dense state comprising pure neutrons,the periodic table could be built up bysuccessive neutron captures. Becauseneutron capture cross sections roughlyfollowed the observed abundances, theidea had the right smell. Gamowsyoung collaborators, Alpher and RobertHerman, carried out the calculationsand broke new ground in cosmology.

As it turns out, the basic idea of nu-cleosynthesis by neutron capture waswrong, and most of the calculationswere irrelevant. The lack of stable nu-clei of mass 5 and mass 8 and the rapidincorporation of free neutrons into he-lium-4 prevent the scheme from work-ing. Interestingly enough, did an-ticipate the so-called r-process, todaysparadigm for the production of theheaviest nuclei by rapid neutron cap-ture in stellar explosions.

Sometimes a wrong paper can bevery influential and important (PhysicalReview Letters referees take note!). Thatcertainly was the case with .

Although only the lightest nucleiwere made in the Big Bang and not byneutron capture, Big Bang nucleosyn-thesis (BBN) is a cornerstone of moderncosmology. It led to the prediction of arelic thermal radiationthe cosmic mi-crowave background or CMBwhichhas turned out to be a cosmic Rosettastone. Paradoxically, Gamows Big Bangmodel spurred Fred Hoyle to thinkmore creatively about the stellarnucleo synthesis to keep his steady-statemodel competitive and in 1957, withGeoffrey Burbidge, Margaret Burbidge,and William Fowler, he worked out thecorrect theory of how the bulk of the elements were made in stars.

So what was wrong with ? Although nonequilibrium nuclearprocesses are an essential ingredient,equilibrium processes are just as im-portant. At very early times, when den-sities and temperatures in the universewere high, nuclear reaction rates wererapidso rapid that thermal equilib-

rium abundances among nuclei and nu-cleons (so-called nuclear statisticalequilibrium, or NSE) were establishedat temperatures higher than 1011 K, cor-responding to a time of less than 0.01 safter the bang and thermal energiesgreater than tens of MeVs. However, atthose temperatures, when thermal en-ergies were greater than nuclear bind-ing energies, entropy favored free nu-cleons and the NSE abundances ofnuclei were tiny.

As the universe expanded andcooled, the binding energies of nucleibecame large compared with thermalenergies; that condition favored nucleiover free nucleons, and the NSE abun-dances of nuclei rose. However, nuclearreaction rates also dropped because oflower densities and cross sections thatbecame exponentially suppressed dueto Coulomb barriers between nuclei.Eventually, nuclear reactions becamerare and the epoch of early nucleosyn-thesis ended.

Predicting the CMB temperatureThe yield of our cosmic reactor involvesthe interplay between the slowing ofnuclear reactions and the rising of NSEabundances and is determined by howhot the Big Bang was, which in turn isquantified by the number of photonsper baryon. That number remains con-stant as both the temperature andbaryon density decrease with expan-sion. More photons per baryon (hotterBig Bang) means a higher CMB tem-perature today, more dissociating pho-tons per baryon during the epoch of nu-cleosynthesis, and lower yields ofnuclei; conversely, fewer photons perbaryon lead to higher yields. Cosmolo-gists prefer the inverse of the photon-to-baryon ratio, the baryon-to-photonratio (), and its value is now knownto be 6 1010.

Using the simple physics above, it ispossible to predict from first principlesthe acceptable range for and therebythe CMB temperature. For very small (very hot Big Bang), there is essentiallyno nucleosynthesis, while for very large

www.physicstoday.org December 2008 Physics Today 9

(very cold Big Bang), most of the nu-cleons wind up in the nuclei with thelargest binding energies (iron uni-verse). The Goldilocks range (forthose not familiar with the childrensbedtime story, see http://en.wikipedia.org/wiki/Goldilocks) is from = 1011 to108. Since the number density of baryonsis just the baryon density divided by themass of a baryon (nB = B/mB) and the pho-ton number density is given by a famil-iar thermodynamic formula, n = aT 3(where a is a constant), knowledge of thebaryon density today translates into aprediction for the CMB temperaturetoday, T = (B/amB)1/31/3. For the Goldi -locks range, the prediction is T ~ 1 to 10K, consistent with the value of 2.725 K 0.001 K measured by NASAs CosmicBackground Explorer (COBE) satellite.

The various predictions made byAlpher and Herman were based on theneutron capture model. To produce theobserved pattern of abundances, theyrequired that the density of nucleonstimes the age of the universe ( fn) beabout 1018 s/cm3 when the temperatureof the universe ( Tn) was about 1010 K.That requirement leads to a differentformula, T = (Tn/1010 K)1/3(B/mB)1/3fn1/3,and a wrong prediction, 70 K usingmodern values, reflecting the incorrect-ness of the underlying physics.

Birth of hot Big Bang cosmologyComputer codes with extensive nuclearreaction networks and precise nucleardata allow the accurate prediction ofthe yields of BBN. The discovery of theCMB in 1965 and the uncertain knowl-edge of the baryon density meant that was between 1010 and 109, and for thisrange only deuterium, helium-3, he-lium-4, and lithium-7 are produced insignificant amounts. By far, the yield of4He is the greatest, a mass fraction ofaround 25%. The consistency of thatprediction with the unexplained, largeprimordial abundance measured by as-tronomers was an early home run forthe hot Big Bang cosmology. Together,the CMB and 4He were the last nails inthe coffin of the steady-state cosmology.Strangely, no tribute was paid to ,the paper that started it all.

In the 1970s David Schramm andothers realized that the rapid fall in theproduction of deuterium with thebaryon density and the fact that subse-quent astrophysical processes only de-stroy deuterium make it a good bary-ometer. An upper limit to the baryondensity follows directly from any meas-urement of the present-day deuterium,and a determination of the primordialdeuterium abundance accurately pegsthe baryon density.

In the 1980s measurements of thedeuterium abundance in the local inter-stellar medium led to an upper limit tothe baryon density of about 10% of thecritical density (the energy density thatseparates the high-density universesthat are positively curved from the low-density universes that are negativelycurved). A decade later the primordialabundance of deuterium was measuredin high-redshift clouds of hydrogen,and the baryon density was determinedto be 4.5%. Beginning in the 1980s,measurements of the total matter den-sity indicated a significantly highernumber, around 20% of critical density,and a composition that was predomi-nantly dark matter. That BBN-baseddiscrepancy, which grew in size andsignificance, became the linchpin in theargument that the dark matter is notmade of baryons.

The road to precision cosmologyIn 1992 COBE detected anisotropy inthe CMB temperature at the level ofabout 30 microkelvin (or 1 part in 105).Those variations in the temperature be-tween two points on the sky, separatedby roughly 10 degrees, provided crucialevidence for the underlying variationsin the matter density needed to seed theformation of all the structure in the uni-versefrom galaxies to superclustersof galaxiesand the first evidence forinflation, the best explanation for theorigin of the seed inhomogeneities.

The spectrum of anisotropy dependsnot only on two or three inflationary pa-rameters but also on cosmologicalonescurvature of space, total matterdensity, baryon density, Hubble con-stant, and age of the universe. In par-ticular, the angular power spectrumtakes the form of a series of harmonic oracoustic peaks whose strengths and po-sitions (as a function of angle) encodeinformation about cosmological param-eters: The position of the first peak in-dicates the curvature; the strength ofthe first peak, the matter density; theratio of the strengths of the odd to evenpeaks, the baryon density; and so on(see my article with Charles Bennettand Martin White, PHYSICS TODAY, No-vember 1997, page 32).

The COBE discovery triggered a raceto measure the wiggles in the CMB an-gular power spectrum. And a series ofground-based and balloon-borne CMBexperiments, mostly in Antarctica, andNASAs Wilkinson Microwave AnisotropyProbe have now determined the CMBpower spectrum from about 0.1 to 90degrees. That spectrum, together withmaps of the large-scale structure in theuniverse today, have determined a host

of cosmological parameters to percent-level precision. The Hubble constant isnow known to be 70 1.3 km/s/Mpc; theage of the universe is fixed at 13.73 0.12 Gyr, its curvature is within 0.6% ofthe flat critical density model, and thevalues of the various components ofmass and energy have been determinedwith error bars of less than 2% (seebelow). Finally, measurements ofnearby and distant supernovae have di-rectly pinned down the expansion ratetoday and long ago, revealing that theexpansion rate is speeding up and notslowing down.

Todays wealth of cosmological dataalso permits crosschecks and has pavedthe way for precision cosmology. Theposter child is the baryon density. Frommeasurements of the primordial deu-terium abundance, the baryon densityis fixed at 4.0 0.2 1031g/cm3, whileCMB anisotropy measurements give4.2 0.1 1031g/cm3an agreementand precision of about 5% (see my Ref-erence Frame in PHYSICS TODAY, De-cember 2001, page 10).

For all its success and precision, cos-mology is not yet solved (thank good-ness!). Particle dark matter accounts for23.3% 1.3% of the universe, but whichparticle? The bulk of the universe(about 72% 1.5%) is made of a myste-rious dark energy whose gravity is re-pulsive and is causing the expansion ofthe universe to speed up. The crazycombination of atoms, particle darkmatter, and dark energy that is our uni-verse is without explanation. What hap-pened before the Big Bang and the des-tiny of the universe still elude us. Andlast but not least, the full extent of theuniverse is unknownis it WYSIWYGor a multiverse of disconnected pieces?All of that is why cosmology is so ex-citingbig questions that seem to bewithin reach of our powerful instru-ments and ideas.

The road to precision cosmologystarted on April Fools Day 60 years agowith a game-changing ideathat justafter the Big Bang the universe was a nu-clear reactor. Though Alpher, Bethe, andGamow didnt get the physics right, theywere right about the importance of nu-clear physics (and physics in general) inthe early universe and the existence ofthe CMB (though not its temperature),and they broke new ground in cosmol-ogy by studying the early radiation-dominated phase that is the focus ofmuch of theoretical cosmology today.Although that groundbreaking paper re-ceived little attention when the CMBwas discovered in 1965, with hindsighttoday we can trace the beginning oftodays revolution in cosmology to it.

10 December 2008 Physics Today 2008 American Institute of Physics, S-0031-9228-0812-220-7

Peter Westwicks interesting featurearticle The Strategic Offense Initiative?The Soviets and Star Wars (PHYSICSTODAY, June 2008, page 43) stimulateda few memories that may add to the his-tory he partially documents. I was a USsenator from New Mexico in 197782and thus had some direct involvementin events leading up to PresidentRonald Reagans March 1983 an-nouncement of the Strategic DefenseInitiative.

In 1979 and 1980, I had become in-creasingly interested in the potential ofproviding the US with a defense againstballistic missiles to counter the knownSoviet efforts to construct high-powered ground-based lasers as well asa national infrastructure that could sur-vive in the event of a nuclear exchange.In the course of my reading on the sub-ject, I ran across an article in a Novem-ber 1979 New Yorker by my then col-league, the late Senator Daniel PatrickMoynihan.1 Moynihan cited separatelypublished arguments by AndreiSakharov and Freeman Dyson againstthe existing doctrine of mutually as-sured destruction (MAD) and in favorof mutually assured protection. Moyni-han found the SakharovDyson argu-ments persuasive and added a few fa-vorable ones of his own.

Having discovered our joint interestin strategic defense, Moynihan and Idecided that we would sponsor a floordiscussion during the Senate MorningHour when the Democratic and Repub-lican leadership made time available forpresentations by individual senators.

He agreed, and we sent our colleaguesan invitation to join us at a specific timeand day for that purpose. Unfortu-nately, no one showed up for our dis-cussion except the two of us.

Possibly stimulated by reports of thisattempt and other statements I hadmade on the subject and related tech-nology matters, President-elect Reaganasked me to discuss the subject with himin December 1980. At that meeting, Rea-gan showed both a deep concern and adeep knowledge about the absence ofany means to protect the US from an ac-tual missile attack. He said that the con-tinued production and deployment ofweapons of mass destruction could notpreserve the peace indefinitely and thatwe should search for defensive alterna-tives. He asked what I thought of the fea-sibility of Edward Tellers suggestionthat space-based lasers could ultimatelybe used to destroy missiles or warheads.I said it then appeared to be technicallyfeasible but would require a great deal ofdevelopment work once ongoing re-search indicated which laser candidateswere most attractive. In the exchange, Ihad the impression that Reagan andTeller had discussed the issue long be-fore the 1982 date suggested in West-wicks article. Further archival researchmay confirm this.

On the question of what Reagan be-lieved relative to defensive versus of-fensive use of space-based weapons,note his response to a query from Wal-ter Mondale during a presidential de-bate in 1984. Mondale asked if Reaganwas serious about sharing strategic de-fense technology with the Soviets. Rea-gans answer: Why not? His responsewould seem to imply that his focus waspurely on missile defense. After partic-ipating in the first SDI war game at thePentagon in 1983, I continued to exam-ine the potential of a shared strategicdefense in more detail and concludedthat Reagans intuition on the matterwas correct.2

With todays proliferation of missilesby rogue nations, some having a nuclearpotential, this may be an even better timefor the US, Japan, and Europe to discussshared strategic defense with Russia,China, and other concerned nations.

References1. D. P. Moynihan, New Yorker, 19 November

1979, p. 104.2. H. H. Schmitt, Ann. N. Y. Acad. Sci. 577,

245 (1985).Harrison H. Schmitt(hhschmitt@earthlink.net)

Albuquerque, New Mexico

Westwick replies: I thank HarrisonSchmitt for his firsthand knowledge ofevents. Existing evidence suggests thatby 1980 President Ronald Reagan hadlearned about new concepts for missiledefense, including Edward Tellers,from various sources, but that Tellerhimself was frustrated by his lack ofpersonal access to the president untilSeptember 1982. His July 1982 letterwas an effort to provide his views. Fur-ther research may indeed clarify thischronology.

I agree that Reaganand most oth-ers in the USviewed the Strategic De-fense Initiative as purely defensive, andfurthermore that his personal offer to share SDI technology was sincere.My point is that the Soviets did not be-lieve him.

Peter Westwick(westwick@history.ucsb.edu)

Santa Barbara, California

Scientists protestprofessors dismissal

We, the undersigned plasma physicists,are familiar with magnetic mirror re-search, and we are concerned about therecent actions of the administration ofthe University of Tsukuba in Japan.Teruji Cho, a professor there, was dis-missed from his position as director ofthe universitys plasma research centeron 6 March 2008, allegedly for inten-tionally manipulating experimentaldata that appeared in Physical ReviewLetters.1 That publication, in fact, con-tains results that are extremely inter -esting and far-reaching in their sig -nificance. Chos team definitivelydemonstrated that flow shear stabili -zation can be directly controlled using

Remembering Reagan and SDIletters

Letters and opinions are encouragedand should be sent by e-mail to ptletters@aip.org (using your surnameas Subject), or by standard mail to Let-ters, PHYSICS TODAY, American Center forPhysics, One Physics Ellipse, CollegePark, MD 20740-3842. Please includeyour name, affiliation, mailing address,e-mail address, and daytime phonenumber on your attachment or letter.You can also contact us online athttp://w w w.physicstoday.org/pt/contactus.jsp. We reserve the right toedit submissions.

12 December 2008 Physics Today www.physicstoday.org

off-axis electron cyclotron resonanceheating. In addition to the dismissal, theuniversity requested that the PRL edi-torial staff retract the paper.

The accusation against Cho, andagainst three other senior staff mem-bers, was initiated by graduate studentswho filed a complaint to an adminis-trative oversight committee that washeaded by Hiroshi Mizubayashi. Afteran investigation, the committee de-manded of Cho that the PRL paper beretracted. An additional university in-vestigatory committee headed byKazuhiko Shimizu supported that de-mand. However, Cho and his seniorcollaborators refused to make such a re-traction because they are convinced ofthe integrity of their data. They submit-ted to the university committee a reportaddressing the controversial issues, andthey submitted to the journal Physics ofPlasmas (PoP) a more detailed paper forpublication. The university committeesrejected Chos report without substan-tive scientific comments.

Meanwhile, Chos manuscript wasjudged to be scientifically sound and tomerit publication in PoP2 on the basis offavorable standard refereeing and re-ports from two additional experts whowere consulted when the PoP editorialstaff became aware of the scientific con-troversy associated with Chos work. Webelieve that the PoP editors acted cor-rectly; the second paper convincinglyconfirms the correctness and reliabilityof the results published in the PRL paper.However, the university administrationapparently did not accept the opinion ofthe PoP editorial board. Instead, they ter-minated Chos professorial position on29 August 2008, an action that was an-nounced in the worldwide press.

Many scientists who are familiarwith magnetic mirror research, espe-cially that conducted at Tsukubasplasma research center, are deeply con-cerned about the accusations againstCho and his colleagues. At least four let-ters have been sent to Yoichi Iwasaki,president of the university, to informthe administration of support for thescientific integrity of Chos claims.None of those letters were acknowl-edged. We find it troubling that the uni-versity appears to be uninterested in theopinions of experts in the field.

It is clear to us that neither Cho norhis close colleagues on the GAMMA-10team intentionally misrepresenteddata. We cannot understand why theUniversity of Tsukuba administrationhas taken the extreme action of dis-missing a distinguished investigator.Cho has been open about his experi-

mental and analytical techniques andhas shared his data and methodologywith his research team and with foreigncollaborators from Russia and the US.We are concerned that the universitysactions against Cho constitute a form ofscientific censorship. We believe that anappropriate international scientificpanel should investigate the univer-sitys behavior in this matter.

References1. T. Cho et al., Phys. Rev. Lett. 97, 055001

(2006).2. T. Cho et al., Phys. Plasmas 15, 056120

(2008).Herbert L. Berk

(hberk@mail.utexas.edu)University of Texas at Austin

Nathaniel J. Fisch(fisch@princeton.edu)Princeton University

Princeton, New JerseyAlexander Burdakov

(burdakov@inp.nsk.su)Gennadi I. Dimov(g.i.dimov@inp.nsk.su)

Alexander A. Ivanov(a.a.ivanov@inp.nsk.su)

Eduard P. Kruglyakov (e.p.kruglyakov@inp.nsk.su)

Budker Institute of Nuclear PhysicsAkademgorodok, Russia

Vladimir Moiseenko(moiseenk@ipp.kharkov.ua)

National Science CenterKharkov Institute of Physics and Technology

Kharkov, UkraineKlaus Noack

(k.noack@fz-rossendorf.de)Research Center Dresden-Rossendorf

Rossendorf, GermanyVladimir P. Pastukhov

(past@nfi.kiae.ru)Kurchatov Institute

Moscow, RussiaShigetoshi Tanaka

(tanakashigetoshi@yahoo.co.jp)Kyoto University

Kyoto, JapanOlov gren

(olov.agren@angstrom.uu.se)Uppsala University

Uppsala, Sweden

Stellarator pro and con

The cancellation of the National Com-pact Stellarator Experiment (PHYSICSTODAY, July 2008, page 25) leaves a holein the US and world fusion programsthat are focused on ITER. Two physicspoints define the importance of the holethat NCSX filled. First, the shape of theplasma is the primary design freedom ofmagnetically confined fusion plasmas.The other determinants of plasma equi-libria, which are the pressure and current

profiles, are largely self-determined. Sec-ond, the excellent confinement of toka-maks, such as ITER, does not require axisymmetry. Only quasi-axisymmetryis required, which greatly increases thefreedom of plasma shaping.

In quasi-symmetry the magnetic fieldlines lie on nested toroidal surfaces, andthe magnetic field strength on those sur-faces has a symmetryeven when theshape of the surfaces does not. Particletrajectories are determined by the mag-netic field strength, independent of theshape of the magnetic surfaces, andquasi-symmetry ensures the preserva-tion of the constant of the motion thatgives good confinement in axisymmetry.The deviation from axisymmetry canhave any magnitude as long as it is con-strained by quasi-axisymmetry. Axisym-metric shapingaspect ratio, ellipticity,triangularity, and squarenessis con-sidered essential to achieving the ITERmission, but most of the shaping free-dom of toroidal plasmas requires thebreaking of axisymmetry.

The NCSX stellarator was the onlyexperiment in the world designed tostudy quasi-axisymmetric shapingother than in the axisymmetric limit. Al-though the project is canceled, its costsdo establish a required financial scale.The highest cost estimates for NCSXconstruction and research were about15% of the annual US non-ITER con-struction budget for fusion, or about 1%of the envisioned world ITER budget.Expertise on quasi-axisymmetric shap-ing would give the US unique capabili-ties in exploiting the information fromITER to make fusion a reality, if that ex-pertise were developed by the time theITER information becomes available.

As the primary design freedom,quasi-axisymmetric shaping is clearlyimportant. It is the only type of non-axisymmetric shaping that can be ap-plied to ITER-like plasmas when the fusion program moves to the design ofa demonstration power plant. Non-axisymmetric shaping provides theonly known solutions to a number of issues that must be addressed beforemagnetic fusion energy can be a reality.1

Management problems led to thecancellation of NCSX. Such problemscannot be allowed to undermine thefundamental strategic objectives of USfusion research: to develop the knowl-edge base for fusion energy, to have aworld-leading fusion program, and toensure the success of the ITER mission.

Reference1. For a discussion of issues facing magnetic

fusion, see Priorities, Gaps, and Opportuni-ties: Towards a Long-Range Strategic Plan for

14 December 2008 Physics Today www.physicstoday.org

Magnetic Fusion Energy, Fusion Energy Sci-ences Advisory Committee, US Departmentof Energy, Washington, DC (2007); availableat http://www.ofes.fusion.doe.gov/FESAC/Oct-2007/FESAC_Planning_Report.pdf.

Allen H. Boozer(ahb17@columbia.edu)

Columbia UniversityNew York City

The recent cancellation of the Na-tional Compact Stellarator Experiment(NCSX; PHYSICS TODAY, July 2008, page25) calls to mind the fact that exactly 40years ago the amazing Russian T-3 toka-mak results burst upon the world andblindsided the US stellarator program.The ensuing shutdown of stellaratorwork at the Princeton Plasma PhysicsLaboratory and the rapid adoption oftokamaks at PPPL and other US labora-tories were arguably the most impor-tant episodes in the US magnetic fusionprogram.

Successively more powerful toka-maks with ever more impressive per-formance came on line. Nevertheless,new stellarator projects were eventu-ally funded by the US Department ofEnergy (DOE) at fusion labs in Ten-nessee, Wisconsin, and elsewhere, withat best lackluster results and usually farworse. As suggested by your article,stellarators are more complicated mag-netic confinement devices than toka-maks, and thereby have always ap-pealed to theoreticians who possesscomplicated minds and access to su-percomputers, but nature is indifferentto both.

Having learned nothing fromdecades of tokamak progress and con-tinued stellarator debacles, in the mid-1990s the directorate at PPPL and itscounterpart at DOE reversed the196869 revolution: They decided toshut down the flagship US tokamak fu-sion test reactor and replace it with astellarator of unimaginable complexity,the recently aborted NCSX. Those fool-ish decisions have served only to expe-dite the ongoing demise of the US mag-netic fusion program. Now, with thewell-deserved termination of the NCSXproject, perhaps limited resources canbe refocused on the tokamak family asthe only proven approach to magneticfusion energy.

Daniel JassbyPlainsboro, New Jersey

Peering into peerreview

Given that publications play an impor-tant role in the making or breaking of apersons academic career, I think a re-

examination of the peer-review processis in order. Over the years, as Ive writ-ten and submitted papers, I have comeacross reasonable reviews, horrible re-views, and even personal attacks em-bedded in mediocre reviews. I suspectmany researchers have received similartreatment. And in the end product ofpapers published in journals, we see thegood and sometimes the awful.

I think its time for each of us to takeresponsibility for what we say. I pro-pose that reviews and reviewers namesbe made public after each review iscomplete. The original intention of ananonymous review system, presum-ably, was that it would protect thewriter and the reviewer, but the systemhas been abused.

Reviewers need to be responsiblefor what they say by revealing theiridentity and their comments. If thatwere done, Im sure reviewers wouldbe much more cautious about whatthey write, and we would see both thereviews and the published papers im-prove. Fewer erroneous reviews wouldbe passed on authoritatively to the ed-itors, and personal attacks in the re-views would cease. This revised sys-tem would require reviewers to focuson a papers science content rather than allowing them to air their per-sonal feelings.

We have the resources for this task.With the growth of online journals, itwont take much to post the paper,whether accepted or rejected, onlinewith the reviews alongside it. That way,we can at least have an idea of whetherthe reviewers did their job properly andappropriately. We can also go a step fur-ther with online forums that allowreader feedback on papers and reviews.

Tai-Yin Huang(tuh4@psu.edu)

Penn State Lehigh ValleyFogelsville, Pennsylvania

Early x-ray burstsighting

We were intrigued by the story X-rayOutburst Reveals a Supernova Before ItExplodes (PHYSICS TODAY, August2008, page 21), which describes thelikely discovery of a core-collapse su-pernova by Alicia Soderberg and col-leagues.1 The storys figure 1 resemblesa similar x-ray light curve, reported bycollaborators at Los Alamos NationalLaboratory,2 from an x-ray outburst thatoccurred on 7 July 1969 and precededby two days the x-ray nova CentaurusXR-4.3

The spin periods of the Vela satellites

that recorded the 1969 event wereroughly 1 minute, and any locationwithin the instruments field of viewwould be sampled for 2 or 3 seconds outof that period, followed by subsequentsamplings every 60 seconds or so.When first observed, the precursor tothe Cen XR-4 nova was already at itshighest level, but the subsequent de-cline is almost identical to that of SN2008D.

The outburst was discernable abovebackground for seven minutes;2 thePHYSICS TODAY item indicates a similarduration for the outburst of SN 2008D.The x-ray nova part of the transient CenXR-4 was observed two days later on9 July 1969, the next time the satellitesdetector scanned that part of the sky.

An article about the original discov-ery of Cen XR-4 was published rightaround the time the nova phase wasrapidly declining. By 24 September1969, the source was no longer visibleabove background. In a second articlecovering the known life of the Cen XR-4x-ray nova,3 we stated that there was nodefinite optical identification of CenXR-4; a nova outburst had not been re-ported at the location of the source.

It is not clear whether Cen XR-4 wasa core-collapse supernova as the simi-larities between it and SN 2008D sug-gest. But it is certainly clear that the occurrence of x-ray precursors to ener-getic cosmic processes was docu-mented in the 1972 event and again in2008.

References1. A. M. Soderberg et al., Nature 453, 469

(2008).2. R. D. Belian, J. P. Conner, W. D. Evans,

Astrophys. J. Lett. 171, L87 (1972).3. W. D. Evans, R. D. Belian, J. P. Conner,

Astrophys. J. Lett. 159, L57 (1970).Richard D. Belian

(rdbelian@lanl.gov)Mario R. Perez(mperez@lanl.gov)

Los Alamos National LaboratoryLos Alamos, New Mexico

Coleman tributeRegarding Sheldon Glashows tributeto Sidney Coleman in the May 2008issue of PHYSICS TODAY (page 69), Ishould add another side of Sidney. Hewould come to the physics graduatestudents parties and sit on the floor,back to the wall, and recite all the wordsLord Byron ever wrote. As a physicsgraduate students wife and a humani-ties major, I so enjoyed that Sidney.

Sandy AlyeaBloomington, Indiana

16 December 2008 Physics Today 2008 American Institute of Physics, S-0031-9228-0812-320-1

Physics Nobel Prize to Nambu,Kobayashi, and Maskawa fortheories of symmetry breaking In particle physics, some symmetries are so severely broken that theyre hard to recognize. Others are so slightly broken that the imperfection is hard to find.

For their contributions to theunderstanding of symmetrybreaking in particle physics, threetheorists have been awarded thisyears Nobel Prize in Physics. TheRoyal Swedish Academy of Sci-ences awarded half the prize tothe University of ChicagosYoichiro Nambu for the discov-ery of the mechanism of sponta-neous broken symmetry in sub-atomic physics. The other half isawarded jointly to MakotoKobayashi of KEK, Japans high-energy accelerator research organization in Tsukuba, andToshihide Maskawa of Kyoto Uni-versity for the discovery of theorigin of the broken symmetrywhich predicts the existence of atleast three families of quarks innature.

All three laureates were bornand educated in Japan. ButNambu, born in Tokyo in 1921, isa generation older than Kobayashi andMaskawa, who were born in the early1940s. And unlike them, Nambu hasspent most of his career in the US, hav-ing left war-ravaged Japan in 1952 forwhat he has described as the paradiseof Princeton, New Jersey.

Spontaneous symmetry breaking Introduced into particle physics byNambu in 1960, spontaneous symmetrybreaking was to become a pillar of thefields standard model, which since itscompletion in the mid-1970s has sur-vived every experimental challenge.When a physical state does not exhibitall the symmetries of the dynamicallaws that govern it, the violated sym-metries are said to be spontaneouslybroken.

The idea had been around for a longtime in classical mechanics, fluid dy-namics, and condensed-matter physics.An oft-cited example is ferromagne -tism. Its underlying laws of atomicphysics are absolutely invariant underrotation. Nonetheless, below a criticaltemperature the atomic spins sponta-

neously line up in some arbitrary direc-tion to create a state that is not rota-tionally symmetric. Similarly, the cylin-drical symmetry of a state in which apencil is perfectly poised on its tip isspontaneously broken when the pencilinevitably falls over. But such examplesgive little hint of the subtlety and powerof the notion once Nambu began ex-ploiting it in quantum field theory.

It began with a paper Nambu wrotein 1959 about gauge invariance in su-perconductivity.1 The paper exhibits hisvirtuosity in two disparate specialtiesquantum field theory and condensed-matter theory. He became conversantwith both as a graduate student at theUniversity of Tokyo after he was mus-tered out of the army in 1945. Eventu-ally he began working with the grouparound Sin-itiro Tomonaga, one of thecreators of modern quantum electrody-namics (QED). Tomonaga was actuallybased at another university in Tokyo.But the University of Tokyo was strongin condensed-matter physics. SoNambu started out working on theIsing model of ferromagnetism.

After two years at the Insti-tute for Advanced Study inPrinceton, Nambu came to theUniversity of Chicago in 1954,just before the untimely deathof Enrico Fermi. When JohnBardeen, Leon Cooper, andRobert Schrieffer publishedtheir theory of superconductiv-ity in 1957, Nambu and othersnoted that the BCS supercon-ducting ground state lackedthe gauge invariance of the un-derlying electromagnetic the-ory. In classical electrodynam-ics, gauge invariance refers tothe freedom one has in choos-ing the vector and scalar po-tentials. In QED that freedom islinked to the freedom to changethe phase of the electron wave-function arbitrarily from pointto point in space. Did thegauge-symmetry violationmean that the BCS theory was

simply wrong? Or perhaps supercon-ductivity was a manifestation of someyet unknown force beyond electromag-netism and atomic physics.

Having heard Schrieffer give a talkabout the new theory in 1957 withoutmentioning gauge invariance, Nambuspent the next two years thinking aboutits role in the theory. He recast the BCStheory into the perturbative quantum-field-theoretic formalism with whichRichard Feynman had solvedindependently of Tomonagathe prob-lem of the intractable infinities in QED.From that reformulation, Nambu con-cluded that the superconducting groundstate results from the spontaneous break-ing of the underlying gauge symmetry.He showed that all the characteristicmanifestations of superconductivityincluding the expulsion of magnetic fluxand the energy gap that assures losslesscurrent flowfollow simply from thatspontaneous symmetry breaking.

Exploiting an analogyI began this work with no thought thatit might be relevant to particle physics,

&

Nambu

UNIVERSITYO

FCHICAG

O

www.physicstoday.org December 2008 Physics Today 17

recalls Nambu. But his analysis re-vealed several possible connections:

The spontaneous symmetry break-ing in the BCS theory generated collec-tive excitations of quasiparticle pairswhose frequencies vanished in the limitof long wavelength. In other words, en-ergy vanished at zero momentumthehallmark of a zero-mass particle.Nambu proposed that such massless,spinless particles or excitations are aninevitable consequence of any theorywith spontaneous breaking of a contin-uous (as distinguished from discrete)symmetry. And indeed a year later theorist Jeffrey Goldstone revisitedNambus conjecture and offered a morerigorous and general proof. Then formuch of the decade, the supposed in-evitability of these so-called NambuGoldstone bosons was to pose a frus-trating but extraordinarily fruitfulproblem for theorists seeking quantumfield theories of the strong and weaknuclear forces.

Furthermore, Nambu noted that themechanism by which the spontaneoussymmetry breaking generates the BCSenergy gap was suggestively analogousto what one would need to generate anonvanishing nucleon mass in a theoryof the strong interactions whose under-lying symmetry requires the nucleon tobe massless.

Nambu promptly applied thoseideas to outstanding problems in parti-cle physics, where, he argued, the ana-logue of the BCS ground state that man-ifests the broken symmetry would bethe vacuum itself.2

In Fermis 1934 theory of beta decay,the weak interaction of the hadrons (at

that time only the proton and neutron)is ascribed to a so-called weak hadroniccurrent analogous to the electron cur-rent of electrodynamics. By 1958, a yearafter the discovery that the weak inter-actions violate parity conservation, itwas clear that one had to add an axial-vector current to the vector current ofthe Fermi theory. Vector and axialvector refer to the transformationproperties of the currents under mirrorinversion (parity).

The vector current was known to beconserved in the sense that like the elec-tromagnetic current, it obeys a continu-ity equation. It was appealing to sug-gest that the axial-vector current wasalso conserved. Conserved quantitiesimply symmetries. In particular, a con-served axial current implied a chiralsymmetry that might serve as the basisfor a theory of the strong interactions.Chirality, or handedness, refers towhether a nucleon (or any other spin-12fermion) spins like a left- or right-handed screw. A chirally symmetrictheory of the strong interactions wouldbe invariant under independent globalphase shifts of the left- and right-handed components of the theorys fun-damental fermions.

The problem was that strict chiralsymmetry makes all the fermions in thetheory masslesswhich the nucleoncertainly is not. If, on the other hand, thenucleon acquires its mass through spon-taneous breaking of the chiral symmetry,that breaking should generate a triplet ofmassless, spinless NambuGoldstonehadrons. But they were just as nonexis -tent as the massless nucleons.

To resolve that bind, Nambu

Kobayashi Maskawa

KEK

YUKA

WA

INST

ITUT

E,KY

OTO

UNIV

ERSI

TY

See www.pt.ims.ca/16307-14

goodfellow.com

1-800-821-2870

For custom items and special sizes and quantities

Goodfellow has the flexibility to deliver

just what you need.

info@goodfellow.com

For standard catalog items

To speak to a real, live person

2008 Goodfellow Corporation

Small quantities FAST!

moc.wollefdoogroF motsuc nasmeti

nauqdnasezislaiceps

edotytilibiexlfehteenuoytahwtsuj

roF drrdadnats igolatac

sahwollfedooG

mdn

seitit

revil.de

smet

s

0782-128-008-1

.wollefdoog@ofni

peveil,laerraotkaepsoTTo

2008 Goodfellow Corporatio

Small quantities FA

moc

nosrep

on

FAST!

18 December 2008 Physics Today www.physicstoday.org

proposed that the chiral symmetry is in-deed spontaneously broken. But, he ar-gued, the chiral symmetry of the un-derlying dynamics is slightly imperfecteven before the spontaneous breaking,and therefore the NambuGoldstonebosons dont have to be strictly mass-less, just much lighter than the nucleon.And for that, the three charge states ofthe pion (with about 17 the nucleonsmass) fit the bill.

In two follow-up papers with Gio-vanni Jona-Lasinio, Nambu fleshed outthat idea with a simple illustrative the-ory in which the nucleon, before spon-taneous symmetry breaking, has abare mass about 1% that of the phys-ical nucleon.3 That nonvanishing baremass mars the underlying chiral sym-metry just enough for the spontaneousbreaking to generate the pions as boundstates of nucleonantinucleon pairs.

All this was 3 years before the quarkmodel and 12 years before the formula-tion of quantum chromodynamics, thestandard models theory of the stronginteractions to whose conceptual basisNambu would soon make seminal con-tributions. In QCD, nucleons and pionsare bound states of quarks about aslight as the bare nucleons of Nambus il-lustrative model. And in QCD, its thequark masses that mar the theorys un-derlying chiral symmetry.

So whats the lasting value of such anunrealistic model, which made no pre-tense of being a fundamental theory ofthe strong interactions? Chiral symme-try was to become the basis of a veryuseful effective-field theory. Nambusprescient work in the early 1960s pro-foundly deepened our understandingof mass, says Princeton University the-orist Curtis Callan. It lets us explaintoday not only why the pion is so lightbut also how the proton, a bound stateof three almost massless quarks, can beso much heavier than its constituents.

Electroweak unificationThe crowning triumph of spontaneoussymmetry breaking was the successfulcreation of a unified theory of the weakand electromagnetic interactions in1967 by Steven Weinberg and inde-pendently by Abdus Salam. But achiev-ing electroweak unification first re-quired finding the exception toGoldstones theorem.

In 1963 condensed-matter theoristPhilip Anderson had pointed out thatthe massless excitations Nambu foundin the BCS theory actually acquire masswhen one includes Coulomb interac-tions that the theory explicitly neglects.4He suggested that the mechanism by

which those excitations acquire massalso applies to spontaneous symmetrybreaking in a particular class of fieldtheories in particle physics. They arethe so-called local gauge theories,which, like QED, remain invariantunder independent, arbitrary phaseshifts or more complicated transforma-tions of the wavefunction at every pointin space and time.

A year after Andersons paper, anumber of particle theorists worked outin detail how spontaneous symmetrybreaking in local gauge theories avoidsthe embarrassment of spinless, mass-less NambuGoldstone bosons thatdont exist.5 (The chiral symmetry towhich the pions were attributed is aglobal rather than a local symmetry.) Insome local gauge theories, for exampleQED and QCD, the gauge symmetry re-mains unbroken. Such theories requirethe existence of massless, spin-1 gaugebosons like the photon and the gluons.The theorists concluded that if thegauge symmetry is spontaneously bro-ken, the NambuGoldstone bosons mixwith the gauge bosons in a way thatmakes the former disappear and the lat-ter become massive.

Massive gauge bosons had beenmuch sought after. The massless pho-ton mediates the infinite-range electro-magnetic force. The very-short-rangeweak force, by contrast, appeared to re-quire mediators much heavier than theproton. Weinberg proposed that thephoton and the massive mediators of

the weak interaction are in fact gauge-boson siblings in a unified local gaugetheory.6 Their gross disparities of massand coupling strength, he argued, fol-low from a spontaneous symmetrybreaking that avoids the unwantedNambuGoldstone bosons.

The experimental confirmation ofthe resulting electroweak theory culmi-nated in 1983 with the discovery of thepredicted charged (W) and neutral (Z0)weak gauge bosonsalmost a hundredtimes heavier than the proton. For me,the most wonderful implication ofNambus introduction of spontaneoussymmetry breaking in the early 1960swas that there were heavily disguisedsymmetries in nature that remained tobe discovered, recalls Weinberg.

Kobayashi and MaskawaWhen it was discovered in 1957 that theweak interactions are not invariantunder mirror inversion (denoted by theparity operator P), theorists generally as-sumed that the laws of particle physicsnonetheless remain invariant under CP,the combined transformations of parityand charge conjugation (C). In otherwords, particle interactions should beindistinguishable from those of theirantiparticles as viewed in a mirror. Butthat strategic retreat became untenablein 1964, when it was found that aboutone time in a thousand, a neutral K meson decays in a way that violatesCP invariance.

Parity violation had been explainedby positing that only the left-handedspin components of the fundamentalspin-12 fermions (and the right-handedcomponents of their antiparticles) par-ticipate in the weak interactions. Butsuch schemes appeared to be CP- invariant, thus leaving the violation un-explained. Theorist Lincoln Wolfen-stein promptly suggested that the ob-served CP violation might involve asuperweak interaction outside thepurview of the ordinary weak interac-tions that was, indeed, unique to theneutral-kaon system.

Not until 2001 was CP violation fi-nally observed in another system: thedecay of neutral B mesons. The Bmesons, 10 times heavier than the kaons,carry the very heavy bottom quark,which was discovered in 1977. But its ex-istence had been predicted by Kobayashiand Maskawa in their 1972 quest to ex-plain CP violation as a consequence ofthe ordinary weak interactions.7

Maskawa and Kobayashi were bothborn in Nagoya (in 1940 and 1944) anddid their PhDs in the particle-theorygroup led by Shoichi Sakata at Nagoya

W+

e+q

u, c, or t

qd, s, or b

1/3

+2/3

Feynman diagram of a typical weakinteraction mediated by the chargedW boson. A quark changes flavor anda neutrinopositron pair is created.There are nine possible quark flavorchangesfor example, the commonup (u) to down (d) of beta decay or themore exotic top (t) to strange (s). Therelative coupling strengths and phasesfor those nine processes at theq+2/3q1/3W+ vertex are the elements ofthe 3 3 Cabibbo-Kobayashi-Maskawamatrix from which one can predict theviolation of CP symmetry within thestandard model of particle physics.

www.physicstoday.org December 2008 Physics Today 19 See www.pt.ims.ca/16307-15

University. In the 1950s Sakata had for-mulated an influential precursor to thequark model. Maskawa and I first metwhen he was a teaching assistant in oneof my undergraduate courses, recallsKobayashi. After Kobayashi finishedhis PhD, the two were reunited as post-docs at Kyoto University in 1972 and to-gether undertook the investigation thatwould win them the Nobel Prize.

By 1972, even though the predictedW and Z bosons were still to be discov-ered, the WeinbergSalam model wasalready the presumptive theory of theweak interactions. But to make predic-tions about decays involving hadrons,one has to augment the model with in-formation about the quark content ofthe weak hadronic currents. At the time,only three flavors of quarks wereknown: up (u+2/3), down (d1/3), andstrange (s1/3). But Sheldon Glashow,John Iliopoulos, and Luciano Maianihad already made a strong theoreticalcase (the so-called GIM model8) for theexistence of a fourth quark, which theycalled charmed (c+2/3). The charmedquark would be discovered in 1974.

In a quantum field theory, the CP op-erator transforms fields into their com-plex conjugates. Unless one embellishesthe basic WeinbergSalam model withadditional scalar fields, the only way itcould yield CP violation is if one ormore of the coupling constants thatcouple the W boson to the differentquark flavors is irreducibly complexin the sense that its phase cannot bemade to vanish by clever redefinitionsor phase conventions.

Before Kobayashi and Maskawa setto work seeking the origin of CP viola-tion within the confines of the basicWeinbergSalam model, it was alreadyclear that a theory with only the threeknown quarks could yield no such com-plex coupling constant. Therefore theylooked first at the GIM scheme, inwhich the four quarks come in twopairs(u,d) and (c,s)closely related,respectively, to the lepton pairs (e,e)and (,). Each quartet of two quarksand two leptons is nowadays called afamily of the fundamental fermions.

Mixing flavorsIn the GIM scheme, the quark flavoreigenstates u, d, s, and c are not quitethe weak-interaction quark eigenstatesin the hadron currents that couple to theW. The two sets of basis states are re-lated by a rotation through a small mix-ing angle in two dimensions of the fla-vor space. Called the Cabibbo angle, was introduced by Nicola Cabibbo inpre-quark days (1963) to preserve the

universality of the weak interactions inspite of natures observed preference fordecays that do not change strangeness.

Kobayasahi and Maskawa pointedout that, in this four-quark scenario, therelevant coupling constants amongwhich one must seek an irreduciblephase are just the matrix elements of the2 2 Cabibbo rotation matrix (times auniversal constant g). For example, thecoupling constant for the udW vertex inthe Feynman diagram on page 18would be g cos, while the weaker cou-pling constant for the strangeness-changing vertex usW would be g sin.All the coupling constants would bereal. They would be determined by theone parameter , with no extra degreeof freedom for an irreducible phase.Therefore the basic WeinbergSalammodel with only the four GIM quarksoffered no mechanism for CP violation.

Then Kobayashi and Maskawa wenton to systematically eliminate all theother, less natural four-quark scenar-ios that were consistent with the Wein-bergSalam model. Those that did allowan irreducible phase turned out to be in-compatible with well-established em-pirical characteristics of hadronic de-cays. Thus the authors had shown thatno realistic four-quark scheme woulddo. So they went on to consider an expansion of the GIM scheme byadding a putative new pair of quarks to make six.

Called by their present names, thenew quarks would be the bottom (b1/3)and the top (t+2/3). The Cabibbo matrixthen generalizes to become a 3 3 mix-ing matrix describing a rotation in threedimensions of flavor space. The ex-panded Cabibbo-Kobayashi-Maskawa(CKM) matrix has four degrees of free-dom: three rotation angles and, mostimportant, an irreducible phase thatmight be responsible for the observedCP violation.

In effect, Kobayashi and Maskawawere predicting that unless the CP vio-lation was due to new physics beyondthe basic WeinbergSalam model, therehad to be a third family of quark and lep-ton pairs. Our paper began to attractreal attention when the [heavy] leptonwas found in 1975, says Kobayashi.And then, after the discovery of the bot-tom quark two years later, the paper be-came one of the most cited in the historyof particle theory. The top quark, almost200 times heavier than the proton, wasnot discovered until 1994.

B factoriesThe nine elements of the unitary CKM matrix are the relative coupling

20 December 2008 Physics Today www.physicstoday.org

constants for the nine different q+2/3q1/3W+vertices indicated in the figure. If the ir-reducible phase is not zero, no phaseconvention can make them all real. Onecan measure all the CKM matrix ele-ments in a variety of processes that donot exhibit CP violation, and from thempredict where and how strongly CP vio-lation will occurassuming that anonzero CKM phase is indeed the cause.

Because the B mesons are so heavy,they were expected to offer experi-menters a rich variety of decay modesthat would exhibit CP violation muchbetter than the kaon decays for compar-ison with the CKM predictions. To thatend, the B factories KEKB at KEK andPEPII at SLAC were built in the late1990s. Both facilities first confirmed CPviolation in B decays in 2001 (see PHYSICSTODAY, May 2001, page 17). And sincethen, all their results have been consis-

tent with what the CKM matrix predicts.Thats gratifying for particle theory

but problematic for cosmology. Theoverwhelming preponderance of mat-ter over antimatter in a cosmos that pre-sumably began with neither requiressome source of CP violation. (See the ar-ticle by Helen Quinn in PHYSICS TODAY,February 2003, page 30.) But cosmolo-gists conclude that the CP violation of-fered by the KobayashiMaskawamechanism, though it explains whatsbeen seen in meson decays, is far tooweak to explain the cosmic asymmetry.There must be additional CP-violatingphenomena in nature.

Kobayashi, who retired from the di-rectorship of KEKs Institute of Particleand Nuclear Studies in 2005, is still in-vestigating the symmetry-violating im-plications of various theories that pro-pose new physics beyond the standard

model. Maskawa, having retired fromthe faculty of Kyoto UniversitysYukawa Institute for TheoreticalPhysics in 2003, is now a professor atKyoto Sangyo University.

Bertram Schwarzschild

References1. Y. Nambu, Phys. Rev. 117, 648 (1960).2. Y. Nambu, Phys. Rev. Lett. 4, 380 (1960).3. Y. Nambu, G. Jona-Lasinio, Phys. Rev. 122,

345 (1961); 124, 246 (1961).4. P. W. Anderson, Phys. Rev. 130, 439 (1963).5. F. Englert, R. Brout, Phys. Rev. Lett. 13, 321

(1964); P. W. Higgs, Phys. Rev. Lett. 13, 508(1964); G. S. Guralelnik, C. R. Hagen, T. W. B. Kibble, Phys. Rev. Lett. 13, 585(1964).

6. S. Weinberg, Phys. Rev. Lett. 19, 1264(1967).

7. M. Kobayashi, T. Maskawa, Prog. Theor.Phys. 49, 652 (1973).

8. S. L. Glashow, J. Iliopoulos, L. Maiani,Phys. Rev. D 2, 1285 (1970).

The 2008 Nobel Chemistry Prize honors the development of a fluorescent tag for bioscienceResearchers can now program cells to make their own dyes, which illuminate the activities of proteins within a cell.

A humble, green-glowingjellyfish has unwittinglyrevolutionized how re-searchers study proteinsand their activities in livingcells. Three researcherswhose independent workled to research tools basedon the jellyfishs fluorescentprotein have been awardedthe 2008 Nobel Prize inChemistry.

The three equal winnersare Osamu Shimomura ofthe Marine Biological Labo-ratory (MBL) in WoodsHole, Massachusetts, andBoston Universitys MedicalSchool in Boston; MartinChalfie of Columbia University; andRoger Y. Tsien of the Howard HughesMedical Institute and the University ofCalifornia, San Diego (UCSD). TheNobel Prize cites all three for the dis-covery and development of the greenfluorescent protein, GFP.

A key feature of GFP is that it doesnot require the action of an enzyme orother cofactor to turn on its fluores-cence. It emits green light in response tostimulation by UV or blue light. Thusresearchers can genetically insert GFPto create precisely targeted intracellulargenetic tags. The cells then express GFPin conjunction with (or in place of) aprotein that researchers want to study.The telltale green glow reports when

the gene for a given protein is active.Chalfie likens GFP to a flashlight illu-minating the activities in a living cell.

Medical researchers often study thecells protein machinery because manyillnesses stem from deviations in themachinerys normal operation. Themyriad applications of GFP includestudies of how nerve cells develop inthe brain, how insulin-producing cellsare created in the pancreas of a growingembryo, and how calcium ions flowwithin the chambers of a beating heart.A particularly colorful application offour differently colored GFP-like pro-teins is the brainbow image1 of amouse brainstem seen in the figure onpage 22.

Before the advent of GFP, biologistshad to label a protein by inserting intothe cell an antibody for that protein andtagging that antibody with a dye. Themethod was invasive and could not bedone in vivo. GFP allows researchers tosee proteins in action in a living cell.

Collecting jellyfishThe discovery of GFP in jellyfish beganwhen Shimomura, now 80, was hired byNagoya University in the late 1950s toextract the substance that caused a bio-luminescent mollusk to glow. When Shi-momura accomplished the dauntingtask in only one year, the universitygranted him a PhD in organic chemistry.Frank Johnson of Princeton University

Shimomura Chalfie Tsien

UNIV

ERSI

TYO

FCA

LIFO

RNIA

,SAN

DIE

GO

EILE

ENBA

RRO

SO

TOM

KLEI

ND

INST

www.physicstoday.org December 2008 Physics Today 21

See www.pt.ims.ca/16307-16

attracted Shimomura to his lab in 1960 tohelp him study the luminescent jellyfishAequorea victoria. Soon after his arrival,Shimomura traveled with Johnson andan assistant, Yo Saiga, to Puget Sound inWashington State, where the three la-bored long hours to cut and filter the bio -luminescent edges of nearly 10 000 jelly-fish. They then attempted to extract thebioluminescent protein from the re-maining squeezate.

Shimomura now recalls the frustrat-ing attempts to use conventional tech-niques to isolate the fluorescent sub-stance. He drifted for long hours in arowboat to think deeply about the prob-lem. The key to the extraction proved tobe the dependence of the jellyfish bio-luminescence on the presence of cal-cium ions found in seawater. In 1962 theresearchers reported the discovery of ablue-emitting protein, aequorin, ob-tained from the jellyfish.2 Only inciden-tally did they mention finding a secondprotein, which did not produce its ownlight but fluoresced green. That secondprotein, GFP, turned out to be the im-portant one; it does not require an agentsuch as calcium to fluoresce.

In the late 1960s and early 1970s, Shi-momura and others found that the bluelight emitted by aequorin was close toone of GFPs two absorption lines. Thematch suggested that the two proteinswork as a team in the jellyfish. That is,there is a direct, radiationless transfer ofthe energy of the chemically excitedelectric dipole in aequorin to excite theelectric dipole of GFP, which emitsgreen light as it returns to the groundstate.3 Experimental work in the mid-1970s by Shimomura and colleaguesand, independently, by William Wardand his group at Rutgers Universityconfirmed that such fluorescent reso-nance-energy transfer was indeed oc-curring between the two proteins. Theprocess explains why the jellyfishglows green and not blue.

Shimomura also studied the chemi-cal structure of GFP, trying to elucidatethe nature of the chromophore respon-sible for its optical properties. The chro-mophore structure was not fully re-solved until the early 1990s by Wardand his colleagues.4 In 1982 Shimomurawent to MBL and there he pursuedother bioluminescent systems.

Cloning the geneThe next step in developing GFP into anindispensable fluorescent tag was toproduce the gene for the protein. Thattask was undertaken by DouglasPrasher as a postdoc in the Universityof Georgia laboratory of Milton

Cormier. Cormier had been isolatingand characterizing bioluminescent pro-teins since the 1950s but in the 1980s herefocused his lab on the cloning ofgenes involved in bioluminescence.Prashers first task was to clone the genefor aequorin, GFPs partner in the jelly-fish. He and his colleagues in Cormierslab then turned to the GFP gene.Prasher had moved to the Woods HoleOceanographic Institution by the timehe got the complete gene for GFP in1992.5 One key input to getting the genewas the structure of the chromophoreas determined by Ward, a formerCormier postdoc.

At the time, no one knew whetherGFP required the action of an enzyme orother agent before it could fluoresce. Un-fortunately, before Prasher could put theGFP gene into bacteria to see if the hostcell spontaneously produced the fluo-rescent form, his funding ran out. As henow explains, few funders were inter-ested in bioluminescence at that time.Prasher went on to work at different in-stitutions and in other research areas. In2006 NASA terminated the mission onwhich he had been working, and todayhe drives a shuttle van for a Toyota deal-ership in Huntsville, Alabama.

When Chalfie first heard about GFPat a seminar around 1989, he was im-mediately excited about its prospects.He thought GFP might be just the thingfor looking, literally, into the transpar-ent roundworms (Caenorhabditis ele-gans) that he had been studying for 15years. He had always stressed in his lec-tures that one beauty of the worms wastheir transparency.

Chalfie immediately contactedPrasher to ask for a copy of the gene, butit was not yet completed. He calledagain once the gene paper was pub-lished in 1992. Prasher shared a copy ofthe gene with Chalfie and with Tsien,who had also asked for it by then.

According to Chalfie, the DNA cod-ing that Prasher isolated had some extraDNA on it that, in retrospect, wouldhave prevented it from making GFP.Chalfies group did not include theextra DNA, leaving only the instruc-tions for making the GFP protein. Theyinserted the gene into the DNA ofEscherichia coli, which then glowedgreen. The GFP produced by the bacte-ria spontaneously folded into a fluores-cent form, without the aid of any en-zymes. The chromophore in the proteinis fluorescent only if the protein folds injust the right manner.

Chalfie and his colleagues next usedGFP to study roundworms. Chalfie hadthe idea to insert the GFP gene in the

position on the DNA chain normally occupied by the gene for the protein -tubulin. The worms cells then pro-duced GFP everywhere they wouldnormally produce -tubulin, namely inthe worms six touch-receptor neurons.Chalfies team was able to see where thetouch receptors were located and whenduring development they were turnedon.6 After Chalfies paper appeared, thefield exploded; more than 20 000 papersinvolving GFP have been publishedsince 1992.

Not only can researchers replace agiven protein with GFP to see where agene is active, but they can also attachthe GFP to a protein to study the pro-teins motion and interactions. Suchprotein fusion was demonstrated bya group at Columbia led by Tulle Hazel-rigg,7 Chalfies wife. Hazelriggs teamshowed that the protein still functionednormally even with GFP attached to itand that GFP still glowed under excita-tion. It helps that GFP is a relativelysmall molecule.

Chalfie, who is 61, received a PhD inphysiology from Harvard University in1977. After spending five years at theLaboratory of Molecular Biology inCambridge, UK, he went to Columbiain 1982.

Enlarging the paletteWhen Tsien got his copy of the GFPgene, he and his coworkers at UCSDstarted studying and modifying theproperties of the glowing protein. In1994 they ascertained that GFP requiredoxygen, but no other agents, to fluo-resce.8 More strikingly, they found thatthey could alter the absorption and emis-sion spectra by introducing random mu-tations into the genes. One of the fourfluorescent proteins produced by themutant genes emitted blue rather thangreen light. To the researchers surprise,that mutation had caused an amino acidto be inserted into the center of the light-producing chromophore.

The robustness of the chromophoreto such insertions suggested a way toengineer the optical properties. Thatmethod has been widely exploited bymany researchers. In 1995 Tsiens teamfound a mutation that greatly improvedthe optical properties of GFP by con-verting the double-peaked excitationspectrum of the naturally occurringprotein to one with a brighter, singlepeak, which gave brighter fluores-cence.9 Tsiens group has also intro-duced mutants that fluoresce in a widespectrum of colors, including cyan,blue, and yellow.

Completing the spectrum with the

color red remained a challenge for GFP-based proteins. One answer came fromSergey Lukyanov and his colleagues atthe Shemyakin and Ovchinnikov Insti-tute of Bioorganic Chemistry inMoscow. They found a red fluorescentprotein called DsRed in an organismclosely related to a reef coral.10 In thesame paper, they reported cloning asmany as six proteins, including red, yel-low, and cyan variants. According toMarc Zimmer of Connecticut College,no one before the Russian team hadlooked for GFP in corals because theprotein does not generate its own light.(They thought GFP would be foundonly in bioluminescent organisms.)

Tsien found that DsRed was toolarge to attach easily to proteins be-cause it consisted of four amino acidchains instead of one, as in GFP. He andhis colleagues engineered it into a moreuseful monomeric form that retained itsred fluorescence.11

Another far-reaching technique in-troduced by Tsien and his colleagues12is circular permutation. Its a way toinsert entire proteins into GFP withoutlosing the fluorescence. In some cases,light emission is contingent on the in-teraction of the intruder protein withsome other element in the cells envi-ronment. For example, GFP modifiedby the insertion of the protein calmod-ulin glows only when calmodulin bindsto certain other proteins, which it doesonly when Ca2+ ions are present (seePHYSICS TODAY, May 2006, page 18).

Recently Tsien has concentrated onusing GFP rather than developing it as

a research tool. He is hoping to advancecancer research by directing imagingagents and chemotherapy drugs to tu-mors. Born in New York in 1952, Tsiengot a PhD in physiology in 1977 fromCambridge University. After doing re-search at Cambridge and at the Univer-sity of California, Berkeley, he joinedUCSD and Howard Hughes MedicalInstitute in 1989.

Barbara Goss Levi

References1. J. Livet et al., Nature 450, 56 (2007).2. O. Shimomura, F. H. Johnson, Y. Saiga,

J. Cell. Comp. Physiol. 59, 223 (1962).3. J. G. Morin, J. W. Hastings, J. Cell Physiol.

77, 303 and 313 (1971).4. C. W. W. Cody, D. C. Prasher, W. M.

Westler, F. G. Prendergast, W. W. Ward,Biochemistry 32, 1212 (1993).

5. D. Prasher, V. Eckenrode, W. Ward, F.Prendergast, M. Cormier, Gene 111, 229(1992).

6. M. Chalfie, Y. Tu, G. Euskirchen, W. W.Ward, D. C. Prasher, Science 263, 802(1994).

7. S. Wang, T. Hazelrigg, Nature 369, 400(1994).

8. R. Heim, D. C. Prasher, R. Y. Tsien, Proc.Natl. Acad. Sci. USA 91, 12501 (1994).

9. R. Heim, A. B. Cubitt, R. Y. Tsien, Nature373, 663 (1995).

10. M. V. Matz, A. F. Fradkov, Y. A. Labas,A. P. Savitsky, A. G. Zaraisky, M. L.Markelov, S. A. Lukyanov, Nat. Biotech-nol. 17, 969 (1999).

11. R. E. Campbell et al., Proc. Natl. Acad. Sci.USA 99, 7877 (2002).

12. G. S. Baird, D. A. Zacharias, R. Y. Tsien,Proc. Natl. Acad. Sci. USA 96, 11241(1999).

22 December 2008 Physics Today www.physicstoday.org

A brainbow of colorshelps researchers seeindividual neurons inthe brainstem of amouse. By genetic engi-neering, researchers geteach neuron to activatea random mixture offour color genes to pro-duce a total of 90 hues.(Photo courtesy of JeanLivet and Jeff Lichtman,Harvard University.)

See www.pt.ims.ca/16307-17 See www.pt.ims.ca/16307-18

X-ray light valve emerges as a low-cost, digitalradiographic imagerThe instrument combines the physics of amorphous semiconductors, liquid crystals, and the commondocument scanner.

Nearly 15 years ago, University ofTorontos John Rowlands helped pioneerwhat has become the state of the art indigital x-ray imagingthe active-matrixflat-panel imager. In the device, a layerof amorphous selenium (a-Se) convertsincoming x rays directly to charge carri-ers that migrate, under the influence ofan electric field, into an embedded arrayof thin-film transistors, amplifiers, andsubsequent analog-to-digital converters.The digitized signal can then be dis-played, processed, and stored.

The same kind of flat-panel systemcan also be based on indirect conversion,using both a phosphor layer that emitslight when hit by an x ray and an arrayof photodiodes that convert the light intoan electrical signal. But light scattering in the phosphor makes that a lower-resolution approach. (See the article byRowlands and Safa Kasap in PHYSICSTODAY, November 1997, page 24.)

In both cases, image quality is excel-lent, electronic noise near the quantumlimit, and data acquisition fast enoughto allow fluoroscopyreal-time moni-

toring of a changing scene. But becauseeach pixel of the image is individuallyaddressed by its own tiny transistor, active-matrix systems are expensive; asingle unit can cost up to $200 000. Thatputs them out of reach of small hospi-tals, clinics, and most of the under -developed world. Fortunately, Row-lands and colleagues have nowdeveloped a device that avoids the ex-pensethe x-ray light valve.1,2 Like anactive-matrix system, the XLV relies ona-Se to convert x rays into charge. Butunlike that system, the XLV doesntmeasure the charge directly. Instead, itreads the electro-optical effects of thecharge through a birefringent liquidcrystal. Figure 1 outlines the process.

The cost of the XLV could be anorder of magnitude lower than active-matrix systems. I never saw this as alow-cost system, says Rowlands, butone that simply contained some beauti-ful physics. It took a National Institutesof Health grant for me to realize the costof a flat-panel imager could be cut with-out sacrificing image quality. In No-

vember, Rowlands presented the con-cepts behind his prototype at the Indo-US Workshop on Low-Cost Diagnosticand Therapeutic Technologies held inHyderabad, India.

An imagers ingredientsAmorphous selenium is nearly ideal forradiography. Its exquisitely sensitive tox rays: A single x-ray photon of 50 keVspawns about a thousand electronholepairs. The material has a bandgap ofabout 2 eV, high enough that little darkcurrent flows, and yet low enough to re-main a reliable photoconductor. Andbecause its amorphous, a-Se can beevaporated as a thick film onto largeareas and still retain its optoelectronicproperties. The thickness is often tai-lored to the application: High-energyexposures, such as those used for chestx rays, require a millimeter of a-Se to capture most x rays, whereas a 200-micron layer suffices for the lower-energy exposures of mammography.

Thanks to Chester Carlson, whosework using a-Se made photocopying

24 December 2008 Physics Today www.physicstoday.org

possible in the 1960s, the literature isrich with accounts of the materialsproperties. At room temperature, a-Seis close to its glass transition tempera-ture. So, after the material is evaporatedonto a substrate, its defects end up dif-fusing there and to the free surface inthe course of a few days, leaving thebulk largely defect free. Thus, when thematerial absorbs x rays and a bias volt-age is applied across it, the resultingelectrons freely drift along field linesuntil they become trapped at the myr-iad defect sites at the free surface.

Detecting the presence of thosecharges is where the liquid-crystal cellcomes in. The trapped charge distribu-tion creates a varying electric potentialacross the cell when it is placed adjacentto the a-Se surface. The liquid crystal actsas a valve. Charge variations induce in-tensity variations in outside light pass-ing through the cell. The result is a mod-ulated optical image. Moreover, if thelatent charge image is not intentionallyerased by flooding the a-Se with lightabove its absorption edge, the image re-mains for up to tens of minutes, whichgives a radiologist plenty of time to cap-ture it in digital form using a separateCCD camera or scanner.

One reason that Rowlands and com-pany hit on liquid crystals is that, un-like the Kerr effect, liquid crystals elec-tro-optic effect shows up even whendriven by a low voltage. The liquid-crystal layer can also be made as thin as5 m or less. The thinner the layer, thesmaller the blur introduced during theconversion of a charge image into anoptical image.

Indeed, because the XLV interro-gates the charge distribution optically,it has the potential to exceed the spatialresolution of active-matrix flat-panelimagers. Like a-Se, liquid crystals arepixel-less, discrete only at the molec-ular level. In active-matrix systems, the

pixel size of the detector (typicallyabout 100 m) limits the spatial resolu-tion. Moreover, the larger the area of thematrix, the greater the distributed ca-pacitance along wires that load the sig-nal amplifier and the greater the elec-tronic noise. The XLV has no array ofwires, making its noise properties scale-invariant.

Early on, to extract the optical image,Rowlands coupled a CCD camera to theliquid-crystal screen. But cameras are notefficient light collectors. The inability ofthe lens to capture all the photons fromthe screen confers a statistical penaltyknown as a secondary quantum sink.

The problem prompted him to re-place the CCD with an off-the-shelfpaper scannera brilliant move for-ward, according to J. Anthony Seibert,a medical physicist at the University ofCalifornia, Davis. Apart from avoidingthe secondary quantum sink, the scan-ner simplifies digitization: A linear

array of detectors replaces a two-dimensional matrix of them. The scan-ner can also be configured to viewevery part of the liquid crystal from thesame angle; its thin, which allows it tosit close to the latent image; and it pro-vides remarkably high spatial resolu-tionthe photodiodes resolve up to1200 dots per inch (21.2 m), even finerthan commercial systems designed fordigital mammography. Not least signif-icant, its inexpensive.3

In practice, the research teamachieves a dynamic range of just 8 bitsfrom the scanner, although that can beimproved by averaging repeated scans.By comparison, medical imagers typi-cally achieve a 10- to 12-bit dynamicrange. Theres also plenty of room to op-timize the system, says Seibert. Elec-tronic noise from the scanner can be re-duced, for example, by increasing thebrightness of the LED that illuminatesthe liquid-crystal image and by

+

+

+

+

+

+

V V

- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -

+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +

X rayElectrode

Electrode

Liquidcrystal

a-Se

Optical scannerReadout light

a b c Figure 1. In essence, the x-ray light valve is a layer ofamorphous selenium and thinliquid crystal, sandwichedbetween two transparentelectrodes. (a) An x ray pene-trating the a-Se layer creates a cloud of electronhole pairsvia the photo electric effect.(b) With an electric fieldapplied, some of the pairs

drift toward oppositely charged surfaces of the photoconductor where they are trapped. (c) After the applied electric field isremoved, the charge distribution from the x-ray exposure induces a visible image in a birefringent liquid crystal by virtue of thecrystals dielectric anisotropy, which affects the propagation of outside light through it. The optical image is then digitized usinga scanner that sends light from an LED through the crystal; the light reflects from the a-Se interface and into the scanners arrayof photodiodes. To reset the light valve for another exposure, the system is flooded with light having energy above the absorp-tion edge of a-Se. (Adapted from ref. 1.)

Figure 2. X-ray images of a phantom chest using (a) an active-matrix flat-panelimager and (b) an x-ray light valve, shown as boxed insets against the AMFPIchest x ray. Imperfections from the prototype x-ray light valve appear as artifactsin the image. (Courtesy of John Rowlands.)

a b

www.physicstoday.org December 2008 Physics Today 25

Trampoline model of vertical earthquake ground motion.Seismic sensors at the surface of a borehole near the epicenter ofa magnitude-6.9 earthquake this year in Japan revealed unpre-dicted asymmetry in the vertical wave amplitudes at the soil sur-face: The largest upward acceleration was more than twice thatof the largest downward acceleration. The data also showed thatthe soil surface layer was tossed upward at nearly four times

the gravitational accelerationmore than twice the peak hori-zontal acceleration. These find-ings run contrary to currentstructural engineering models,which presume that seismicwaves from earthquakes shakethe ground horizontally morethan vertically. Shin Aoi and col-leagues at Japans NationalResearch Institute for Earth Sci-ence and Disaster Preventionpropose what they call a tram-

poline model to explain the observed nonlinear bouncingbehavior. In their model, the soil undergoes compression in theupward direction and behaves as a rigid mass with no intrinsiclimit on acceleration, much like an acrobat rebounding from atrampoline (figures 1 and 3). In the downward direction, though,dilatational strains break up the soil and the loose particles fallfreely at or below gravitational acceleration (figures 2 and 4). Theobserved seismographic data was simulated by combining thetheoretical waveform from the trampoline model with selectedborehole data that resembled elastic deformation of adeformable mass. The researchers say that other events need tobe analyzed to learn how material conditions affect verticalground response during high-magnitude earthquakes. (S. Aoi etal., Science 322, 727, 2008.) JNAM

Sensing superbug stress under drug binding. Overuse ofantibiotics has spawned strains of bacteria whose cell walls areimpervious to the crippling blows once delivered by penicillinand its derivatives. One such so-called superbug, methicillin-resistant staphylococcus aureus, although found primarily inprisons and hospitals, has now spread beyond those confines.Despite the controlled use of the drug vancomycin, a last line of

defense against MRSA, the latest threat comes from vancomycin-resistant bacteria, which mutate by deleting a key hydrogenbond that allows the drug to bind and inhibit cell wall growth,thereby mechanically weakening the bacteria. Rachel McKendryat University College London and her collaborators recentlydemonstrated a nanoscale cantilever system that is sensitive

enough to detect the difference between the native drug-sensitive bacteria and the mutated resistant form with the miss-ing hydrogen bond. The researchers coated silicon cantileverswith vancomycin-resistant (DLac in the schematic) and van-comycin-sensitive (DAla) bacterial cell-wall analogues, thenimmersed them in a solution containing free vancomycin mole-cules. As expected, the molecules preferentially bound to thecantilevers coated with the drug-sensitive analogue; those can-tilevers experienced a marked deflectionas measured by an optical detectorthat equated to an 800-fold difference inbinding compared with the cantilevers coated with the drug-resistant analogue. The researchers believe their system will leadto sensitive, nondestructive, and rapid nanomechanical bio -sensors for high-throughput drugtarget interaction studies andwill aid in the design of more effective drugs. (J. W. Ndieyira et al.,Nat. Nanotechnol. 3, 691, 2008.) JNAM

Tracking mercury by its isotopes. Different isotopes of thesame element dont always behave identically in chemical reac-tions. As a result, naturally occurring samples can have measura-bly different ratios of stable isotopes. In most observed isotopefractionation, deviations in reactivity vary with the mass differ-ence between isotopes, due either to kinetic effects or to differ-ences in the zero-point vibrational energy of chemical bonds.Last year Bridget Bergquist and Joel Blum of the University of

Large reboundforce

(>gravity)

Ground at lowest(larger upward

acceleration)

Ground at highest(smaller downward

acceleration)

Gravity

1

3 4

2

These items, with supplementary material, first appeared at http://www.physicstoday.org.

Vancomycinin solution

PEG(reference)

DLac

DAla

improving the scanners optics. The liq-uid crystal itself is also somewhat tun-able. A bias voltage across the crystal isrequired to shift its characteristiccurvethe relation between reflectanceand applied fieldto a region in whichthe nonlinear crystals optical responseaccurately mimics the spatial variationsin the charge image.

Rowlands envisions the XLV beinguseful at first for static imaging (in con-trast to fluoroscopy) and chest x rays, anapplication well matched to the systemsdynamic range. Hes now exploring itsclinical practicality at Canadas ThunderBay Regional Health Sciences Center in

northwestern Ontario, where he is set-ting up a new imaging research institute.Figure 2 compares a standard digital ra-diograph of a phantom chest using anactive-matrix system with some smallerpatches using a prototype XLV; phan-tom here refers to a dummy body partthat replicates absorption properties ofhuman anatomy.

In developing countries, a cryingneed exists for simple devices that canensure bones are set properly and canscreen for diseases such as tuberculosis.But Rowlands speculates that reducedcost may affect the technologys use evenin the developed world. In the US alone,

hundreds of millions of x-ray exams areperformed annually. Its somewhat fan-ciful, but just as PCs and laser printersare now ubiquitous, one can imagineeach hospital bed in the intensive careunit outfitted not just with its own heartmonitor, but its own x-ray imager.

Mark Wilson

References1. R. D. MacDougall, I. Koprinarov, J. A.

Rowlands, Med. Phys. 35, 4216 (2008).2. C. A. Webster et al., Med. Phys. 35, 939

(2008).3. P. Oakham, R. D. MacDougall, J. A. Row-

lands, Med. Phys. (in press).

26 December 2008 Physics Today www.physicstoday.org

Michigan in Ann Arbor found that photochemical reactions ofmercury can result in isotope fractionation that does not fit themass-dependent pattern: Odd-numbered Hg isotopes behavedifferently from even-numbered ones. Such mass-independentfractionation, observed in only a few elements so far, may be dueto spinspin interactions between nuclei and the unpaired elec-trons created in light-initiated reactions. Now, Abir Biswas, work-ing with Blum and other Michigan colleagues, has found that Hg stored in coal deposits shows the effects of both mass-dependent and mass-independent fractionation. Moreover, coalsamples from different regionsthe US, China, and RussiaKazakhstanbear different Hg isotopic signatures. Theresearchers suggest that those signatures could provide someinformation about how Hg pollution (produced when the coal isburned) circulates in the environment, a process that is poorlyunderstood. (A. Biswas et al., Environ. Sci. Technol.,doi:10.1021/es800956q.) JLM

Two-dimensional melting in a dusty plasma. The meltingtransition has long fascinated physicists, both for its ubiquity innature and industry and for the sophisticated physics of thephase transition in general. Two-dimensional systems can mimic

surfaces, which melt differently from bulk matter. One such systemis a 2D dusty plasma: Background gas in a vacuum chamber is ion-ized when RF power is applied to an electrode. With sufficient care,one can levitate a single layer of charged dust microspheresabove the electrode; electrostatic repulsion spreads the particlesapart, usually in a stable 2D crystalline pattern. At Ohio NorthernUniversity, Terrence Sheridan came up with a new way to heat onlythe layer of dust. He modulated the RF power at a resonance fre-quency so as to jiggle the dust up and down; some of that mo -tional energy coupled to an in-plane acoustic instability, increas-ing the dusty plasmas effective temperature. The panels show thedust distributions for different modulation amplitude levels. At1.0%, the entire system oscillates vertically as a crystalline rigidbody. As the hexagonal crystal is heated, the coupling becomesevident in the central region at 1.6%. The crystal begins to melt at2.2% and enters a hexatic liquid-crystal phase; it fully melts at2.8%. For more on dusty plasmas, see PHYSICS TODAY, July 2004,page 32. (T. E. Sheridan, Phys. Plasmas 15, 103702, 2008.) SGB

Shocking start for the solar system. In the 1970s, the hypothe-sis arose that our solar system was formed by a passing shock

wave from a supernova, which triggered the collapse of an inter-stellar cloud into a dense region of gas and dust that further con-tracted to become the Sun and its orbiting planets. The originalevidence came from very old meteorites that contained magnesium-26, a daughter product of the short-lived radioac-tive isotope (SLRI) aluminum-26produced in stellar nucleosyn-thesis. Further evidence came from another SLRI, nickel-60,which can only be produced in a supernovas furnace. In astro-nomical terms, short-lived means a half-life of about a millionyears; any SLRIs would have been transported to, and droppedoff in, the pre-solar cloud faster than that time scale. Computermodelers from the late 1990s, however, could not produce boththe collapse and the injection of supernova material unless theyartificially prevented the shock wave from heating the cloud.That situation has now been remedied by a group from theCarnegie Institution of Washington, who used a modern, adap-tive-grid computer code with an improved treatment of heatingand cooling. Their new models show that a supernovas shockwave moving into an otherwise stable solar-mass cloud can bothtrigger the collapse and leave behind enriched gas and dust,including the SLRIs whose products are found in meteorites. Fur-thermore, the researchers found that a protostar began to formin less than 200 000 years, in the blink of an astronomical eye.(A. P. Boss et al., Astrophys. J. Lett. 686, L119, 2008.) SGB

Ruffling a membrane. Soft biological tissue is often subjectedto forces that affect the tissues geometry, and finite elasticityprovides a robust theoretical framework for analyzing themechanical behavior of those tissues. Although the theory canaccommodate anisotropic, nonlinear, and inhomogeneousprocesses subjected to large stresses and strains, its complexitymakes many problems intractable. For growing tissue, though,the slow addition of cells generates shape- or size-changingstresses that are small enough to model successfully (see PHYSICSTODAY, April 2007, page 20). So, too, are simple geometries for tis-sues in equilibrium, even after those tissues are subjected tolarge stresses. Two recent papers have looked at applying thetheory to those cases in thin elastic disks. In one recent study,Julien Dervaux and Martine Ben Amar (both of cole NormaleSuprieure, Paris) looked at anisotropic growth rates: If growthwas faster in the radial than in the circumferential direction, thedisk became conelike, while a reversal of rates generated saddleshapes. A separate study by Jemal Guven (National AutonomousUniversity of Mexico) along with Martin Mller (ENS) and BenAmar looked at excessively large circumferences for a givenradius. Using the fully nonlinear theory, the researchers found aninfinity of quantized equilibrium states for an ever-increasingperimeter at fixedradius. The ripplesaround the edge grewin size and numbernot unlike the flowerpetals shown hereeventually crowdingtogether enough totouch, like the ruffledcollar in a portrait byRembrandt. For moreon the elasticity of thinsheets, see the article inPHYSICS TODAY, February2007, page 33. (J. Dervaux, M. Ben Amar, Phys. Rev. Lett. 101,068101, 2