sign in sign up
<<
Home , 39 Tauri, 47 Ursae Majoris, Corona Borealis, Groombridge 1830, HD 130322, HD 187123

Superflares and Giant

From time to time, a few sunlike produce gargantuan outbursts. Large planets in tight might account for these eruptions

Eric P. Rubenstein

nvision a pale blue , not un- bushes to burst into flames. Nor will the lar flares, which typically last a fraction Elike the , orbiting a yellow surface of the planet feel the blast of ul- of an hour and release their energy in a in some distant corner of the . traviolet light and x rays, which will be combination of charged particles, ul- This exercise need not challenge the absorbed high in the atmosphere. But traviolet light and x rays. Thankfully, imagination. After all, astronomers the more energetic component of these this radiation does not reach danger- have now uncovered some 50 “extra- x rays and the charged particles that fol- ous levels at the surface of the Earth: solar” planets (albeit giant ones). Now low them are going to create havoc The terrestrial easily de- suppose for a moment something less when they strike air molecules and trig- flects the charged particles, the upper likely: that this planet teems with life ger the production of nitrogen oxides, atmosphere screens out the x rays, and and is, perhaps, populated by intelli- which rapidly destroy ozone. the stratospheric ozone layer absorbs gent beings, ones who enjoy looking So in the space of a few days the pro- most of the light. So solar up at the sky from time to time. tective blanket of ozone around this flares, even the largest ones, normally During the , these creatures planet will largely disintegrate, allow- pass uneventfully. People living at high would see their shining brightly, ing ultraviolet light to rain down latitudes may see unusually intense providing them with an almost constant unimpeded. Until this shield is re- auroral displays. But unless an espe- bath of warming rays. At night they stored, a process that might take cially strong flare knocks out a com- might gaze at a moon or two against a decades, daylight will remain deadly munication or a power grid, backdrop of seemingly unchanging to the organisms exposed to it. few people take much notice. stars. Over the billions of since life Fortunately, our Sun does not ap- first evolved on this hypothetical planet, pear prone to such disastrous explo- Inconstant this world has been spared any cosmic sions. Indeed, it produces a remarkably Not all stars shine as steadily as the catastrophes. No nearby supernovae steady output of energy—the value of Sun. Some change in brightness over a have exploded. The of the planet which is duly named the solar con- period of hours, days or months. As- has stayed circular. And no great aster- stant. Sensitive instruments placed on tronomers are fascinated by such vari- oids have come crashing down. In show that it varies by only a able stars, and groups of observers have short, none of the many celestial events few parts in a thousand. Astrophysi- organized specifically to examine that would threaten life here have hap- cists do, however, surmise that the so- them. Fortunately for those of us inter- pened. But today is a bad day. lar constant is slowly increasing, by ested in such curious stars, there is an A stellar flare, enormous by normal about one percent each 300 million uncountable number of them. standards, erupts from the surface of years. Over the 4.6 since Astronomers classify these variable the parent star, giving off a sudden the formed, the Sun has stars according to their properties. So- burst of radiation and spewing into brightened by about 40 percent. Al- called geometric variables do not actually space a fusillade of charged particles. though this variation is substantial, it change the amount of energy they give Although it will startle the inhabitants, has taken place so gradually that off. Rather, the geometry of viewing the sudden heat wave will not cause changes in the amount of heat-trap- these multiple star systems causes us ping gases in the atmosphere have to see different amounts of light at dif- compensated, allowing the tempera- ferent times. In most cases the varia- Eric P. Rubenstein received a Ph.D. in astronomy ture of the Earth to remain relatively tion arises because one star in a binary from Yale University in 1997. He remained there steady—and in a range that keeps wa- system passes in front of the other. for two years before becoming a National Science ter liquid—for billions of years. Such eclipses are most likely to be seen Foundation International Research Fellow and Life on Earth clearly benefits from if the two stars revolve around each starting work at the Cerro Tololo Inter-American Observatory in Chile. Last summer he returned to this long-term stability and from the other in a tight orbit. Interestingly, such Yale as a lecturer in the Department of Astronomy. Sun having only moderate levels of binaries, ones with small separations Address: Yale University, Department of Astrono- what astronomers refer to as “activity,” between the stars, commonly exhibit my, P.O. Box 208101, New Haven, CT 06520- meaning violent outbursts. In large additional kinds of variability. Tidal in- 8101. Internet: [email protected] part, this activity manifests itself as so- teractions cause these closely spaced

38 American Scientist, Volume 89 Copyright © 2001 American Scientist Figure 1. Artistic rendition of a distant, sunlike star undergoing a “” hints at the enormity of such an event. These huge eruptions dwarf the largest flares on the Sun, releasing from 100 to 10 million times as much energy. The author proposes that giant planets orbiting close to such stars trigger these occasional outpourings. (Image courtesy of the author.)

Copyright © 2001 American Scientist 2001 January–February 39 stars to spin rapidly, which in turn galaxy. Other examples of this class in- range from several seconds to hun- greatly enhances their magnetic fields, clude systems in which a dreds of days. Some vary in brightness causing strong stellar flares. As- compact pulls in by more than an order of . tronomers call such flaring chromos- from its nearby partner. From time to Astrophysicists have determined that pheric activity, because the energy time the rain of gas gives rise to huge such changes result from instabilities comes from the portion of the stellar thermonuclear explosions on the sur- in either the rate of energy generation atmosphere that also gives the star its face of the white dwarf. Such systems, or in the relationship between temper- characteristic color. dubbed cataclysmic variables, typically ature and opacity in the stellar atmos- Even more spectacular are the erup- experience repeated explosions, so the phere. Pulsators are usually periodic, tive variables, which include super- term “cataclysmic” is not really apt. although some have multiple period- novae—stars that end their lives by ex- Skywatchers also recognize pulsating icities superposed. ploding so violently that they can variables, which alternately increase So variable stars are themselves a outshine all the other stars in their and decrease in size over intervals that varied lot—and a well-studied one. In- deed, astronomers have prided them- 1 Algol selves in their understanding of the mechanisms that bring about these changes. Conversely, we generally be- lieved that we could predict which stars would be as stable as the Sun. We have traditionally assumed, for instance, that if a star has roughly the same surface and as the Sun, relative brightness relative is a single star and rotates at a speed similar to that of the Sun, it will likewise 0 023451 have only modest levels of chromos- days pheric activity. Such stars are commonly called solar analogues. The unspoken as- 1 SS Cygni sumption that all solar analogues are, in essence, interchangeable underlies much of the thinking about habitable worlds, and perhaps life, existing else- where in the cosmos. Although it seems reasonable that solar analogues should be as stable as the Sun, it is clear that this assumption relative brightness relative was premature—and wrong. Astron- omers now know of nine instances of 0 0 24 6810 12 these sunlike stars having eruptions months that were from 100 to 10 million times 1.5 more energetic than the most powerful

1 Sun (kilowatts per square meter) solar flares. In these cases the stars tem- brightness porarily brightened by amounts that range from several percent up to a fac- solar flux 1.0 solar flux tor of 10 or more. Startling as they would be for some- one living nearby, giant flares around 0.5 distant stars can easily elude us earth- lings. For the most part, the sightings relative brightness relative of these events were simply fortuitous. One set of observations was, for ex- 0 0 ample, made on March 6, 1899, when 1980 1985 1990 1995 2000 a newly discovered comet passed years through the and entered the part of the sky that contains Figure 2. Many variable stars exhibit characteristic patterns of changing brightness, which a star called S Fornacis. Four experi- can reveal much about their nature. The star Algol (Beta Persei), for example, shows a repeti- enced and highly respected observers tive pattern of short-lived diminutions in brightness (top). Astronomers regard Algol as a went out early that evening to view the “geometric variable” because the changes in brightness result from the viewing geometry, which allows each star in this binary system to periodically eclipse the other. The star SS comet. Three of them (one in Austria, Cygni, an “eruptive variable,” shows repeated flashes, but the interval between bursts is not one in Italy and one in Germany) noted steady (middle). Another in the same constellation is Chi Cygni, which is consid- that S Fornacis was strangely bright— ered a “pulsator.” It changes dramatically in brightness with a regular period (bottom). Solar estimating by eye that it was some 15 output for the same interval shows so little variation that it appears essentially constant. (Mea- times as luminous as normal. These surements courtesy of the American Association of Variable Star Observers.) three sets of observations were made

40 American Scientist, Volume 89 Copyright © 2001 American Scientist during one 17-minute interval. Less name rotation rate duration energy released than a half-hour earlier, another (kilometers per second) (ergs) trained astronomer in Italy had ob- 1.3 18 minutes 1 × 1035 served S Fornacis and the comet, and Kappa Ceti 8 40 minutes 2 × 1034 he did not comment on any unusual × 35 brightening. A photograph taken at MT Tauri ? 10 minutes 1 10 1 × 33 Harvard Observatory a few hours lat- Pi Ursae Majoris 9.7 > 35 minutes 2 10 er fails to show S Fornacis as any- S Fornacis 7 17Ð367 minutes 2 × 1038 thing other than normal. BD + 10°2783 4 49 minutes 3 × 1034 Although these century-old observa- Omicron Aquilae 3.9 5Ð15 days 9 × 1037 tions of an anomalous brightening are 2 3Ð25 days 7 × 1037 subject to the limits of human percep- UU 6 > 57 minutes 7 × 1035 tion, the combination of three widely separated observers all seeing a dra- Figure 3. Nine stars are known to have experienced at least one superflare. The values given matic change suggests that the star here for the amount of energy released are minimum estimates, because radiation from these was, in fact, unusually luminous for a events was registered for only a limited range of wavelengths. Rotation rates shown are also short time. The reports become even minimum values, because of uncertainties in the viewing geometry. more credible when compared with some later observations of similar and often sparks large-scale flaring and spin axis of the star is tilted toward the episodes. A series of photographs tak- the emission of x rays. line of sight, the estimated rotation ve- en at the Allegheny Observatory in Just how fast are they spinning? As- locity will be smaller than the actual 1939, for example, shows a star called tronomers often estimate how quickly rate. So the number determined using Groombridge 1830, another solar ana- objects move using Doppler shifts in this strategy represents a lower bound logue, becoming almost twice as lumi- the frequency of spectral lines. Just as and need not reflect the true rotation nous as normal for a little longer than a whistle of a passing locomotive rate. In an extreme case, where the spin 18 minutes. More recently, satellites changes in tone depending on the axis of the star is in the line of sight, equipped with x-ray telescopes have speed of the train toward or away from the measured rotation velocity will captured two sunlike stars in the act of the observer, changes in spectral lines seemingly be zero, even with the star giving off such massive flares that they, can reveal motion on the surface of a spinning rapidly. too, brighten substantially for a few distant star. But it is difficult for as- Such unfavorable geometry might minutes to hours. And astronomers tronomers to characterize stellar rota- account for the low rotation rates found know of five other sunlike stars that tion with a single number, because for one or two of these superflaring became strangely luminous for a while. stars are fluid enough that different stars—but it is unlikely to explain all Although any one particular observa- parts rotate at different rates. So we nine. And some indirect evidence also tion may be somehow defective or mis- usually just report the average speed indicates that these stars are not spin- leading, collectively these nine sepa- of motion on the face of the star. ning very quickly. The argument goes rate reports indicate that not all sunlike The fastest measured rotation speed like this: Rapidly rotating stars usually stars are truly sunlike. of the nine known “superflaring” stars contain a lot of lithium, a rather fragile Brad Schaefer, my former colleague is a bit under 10 kilometers per second. element that is destroyed when it gets at Yale (who is now at the University of Young, active stars rotate much faster. mixed into a hot stellar interior. Rapid Texas at Austin), has been investigat- One possibility is that these indications rotation is thought to prevent such ing these objects for more then 10 of slow rotation are misleading. If the mixing. So by estimating the abun- years. His 1989 paper discussing these events introduced the idea that these Groombridge 1830 stars had all experienced what he called “.” But at that time he had not yet collected all the observa- tional evidence needed to demonstrate that these sudden blow-ups did not 12 3 456 come from some strange type of binary star or from young single stars display- comparison star ing expected amounts of chromospher- ic activity. Recently Schaefer, with Jere- my King of the Space Telescope Science Institute and Constantine Deliyannis of Indiana University, revisited these ob- jects and found that they are indeed normal, middle-aged stars: They are Figure 4. Groombridge 1830, a star about 29 light-years away, experienced a superflare that made it noticeably brighter for a brief time. By happenstance, this transient event was captured not closely spaced binaries, and they by workers at the Allegheny Observatory in 1939. The six images of the star (top) were taken are not spinning as fast as young stars consecutively with each exposed for two minutes. Ten minutes separate image 3 from image 4, do. This later point is important be- which shows an obvious increase in brightness. A comparison star captured on the same pho- cause rapid rotation, even of single tographic plate (bottom) displays no such variation, confirming that the temporary brightening stars, boosts chromospheric activity seen in image 4 is not an artifact. (After Beardsley et al., 1974.)

Copyright © 2001 American Scientist 2001 January–February 41 This outing is largely social, but sci- ence usually dominates the conversa- tion. So one November afternoon, as 2 we strolled slowly from our offices to the restaurant, Schaefer began system- atically at the back of the line of as- tronomers and by ones and twos asked for ideas that might help solve the rid- 1 dle of the superflares. Walking near the front, I heard him slowly overtaking my little group as he dismissed one counts per second proposal after another for various rea- sons. “No, they aren’t supernovae, be- 0 cause the time scale and energy is all wrong,” he might have said. “No, they aren’t cataclysmic variables because they aren’t close binaries,” and so forth. Finally, by the time he was just a 14:00 15:00 16:00 17:00 few steps behind me, I had heard the refrain enough times that I could re- time peat to myself the four most important Figure 5. Curious flaring of the star Pi1 Ursae Majoris can be found in a series of x-ray mea- constraints: The energy of a superflare 33 38 surements taken on January 31, 1984, using the orbiting observatory EXOSAT. Counts regis- is between 10 and 10 ergs; the dura- tered by this detector remain near zero for most of the record, then at 16:00 they rise abruptly tion varies from a fraction of an hour to and slowly fall to background levels. (After Landini et al., 1986.) days; the radiation comes out at wave- lengths that span from at least the near dance of lithium, astronomers can immense blast of energy. What could infrared to x rays; and the Sun has not gauge the rotation rate of a star with- possibly trigger these sporadic super- had any superflares recently, nor has a out knowing how the spin axis is tilted. flares? This question drew me into a big one scorched the solar system with- The nine superflaring stars all have collaboration with Schaefer in 1998. At in the last three billion years or so. low lithium values, which confirms that time, Schaefer had just completed a I thought to myself that these prop- that they are indeed spinning compar- series of careful spectroscopic measure- erties sounded quite a bit like some atively slowly. ments of the nine superflaring stars, but highly active binaries I used to study, still he was baffled by what was caus- which are named for the prime exam- The Superflare Riddle ing the outbursts. So he decided to take ple: RS Canus Venaticorum, or RS CVn Clearly these observations call for a advantage of a longstanding tradition: for short. These systems contain two new class of variables—sunlike stars Every Thursday at noon the astronomy stars that orbit in such close proximity that look very normal most of the time faculty and postdocs at Yale go off to- that they have a period of between one and then without warning give off an gether and eat pizza. and 15 days. The small separation means that the stars raise large tides on each other. This deformation causes the rotation of both stars to become syn- chronized to the of the pair. The resulting rapid spin and con- vection within the star generate pow- erful magnetic fields (a process called dynamo action), which in turn produce tremendous amounts of stellar activity in their . The RS CVn stars often have dark spots (similar to sunspots) covering a few tenths of their surfaces—a sign of strong magnetic fields—and they con- tinually give off small-scale flares. In addition to this enhanced but essential- Figure 6. Rotation of a distant star can be detected by changes in spectral lines. Light given off ly normal flaring, RS CVn stars also ex- by a small patch on the surface of the spinning star will be shifted toward the blue end of the hibit sporadic flares of immense pro- spectrum when the rotation moves that zone toward the observer. Similarly, light from that portions. Astrophysicists have proposed patch will be Doppler-shifted toward the red when rotation moves the source away. Because telescopic observations integrate light from all points on the face of the spinning star, as- several explanations for these giant tronomers detect only an overall broadening of spectral lines (left). If, however, the rotation flares. What the different models have axis of the star is aligned closely with the line of sight, there is little motion toward or away in common is that the energy of the from the observer and little widening in the spectral lines (right). This method of measuring flare is released suddenly by a phe- rotation thus gives only a minimum estimate of the rate. nomenon called magnetic reconnection.

42 American Scientist, Volume 89 Copyright © 2001 American Scientist Magnetic reconnection takes place after magnetic-field lines become tan- gled, pulled, stretched and twisted, and then suddenly reorganize into a sim- pler geometry, which stores far less en- ergy in the magnetic field itself. This change is analogous to a bundle of elongated rubber bands suddenly snapping and releasing all of their pent- up energy at once. But magnetic-field lines, no matter how stretched, cannot break outright. The best they can do is to rearrange themselves pair-wise, splitting and simultaneously reconnect- ing so that they are no longer crossed. And whereas snapped rubber bands can fly off and hit things, reconnecting field lines do not separate from the ob- ject to which they are anchored. Rather, they pump energy into the ionized gas,

the plasma, in which they reside. NASA The physics of magnetic reconnection is well studied but poorly understood. Investigators can observe it taking place in the laboratory, where it sometimes happens unexpectedly and releases enough energy to destroy expensive equipment. Magnetic reconnection is also clearly recognized to contribute to the production of flares and promi- nences on the surface of the Sun. Astrophysicists believe that the in- dustrial-strength flares on RS CVn bi- a naries are most likely caused by the magnetic reconnection of field lines threading from one star to the other. In this situation, the entire volume en- closed by the great looping fields of both stars, a region perhaps 20 or more stellar radii across, is suddenly heated to millions of degrees when the lines reconnect. This event violently acceler- ates the ions and electrons in the stellar wind, producing broadband radiation b and sending a burst of charged particles outward at nearly the speed of light. With these thoughts displacing my craving for pizza, I turned around and mentioned to Schaefer that superflares seemed to have the same properties as the large events sometimes seen on RS CVn binaries. He immediately dis- missed my idea—“These aren’t bina- ries.” But as I walked on, my mind kept churning away on the many similari- c ties, and I became more and more con- vinced: If it looks like a duck, walks like Figure 7. Filaments of hot coronal gas reveal looping solar magnetic-field lines in an image obtained from the Transition Region and Coronal Explorer (TRACE) spacecraft (top). The a duck, sounds like a duck and blows magnetic fields around stars with giant planets in close orbit are likely to be even more com- up like a magnetic reconnection event, plicated, with some lines connecting the two bodies (a). At times motion of the planet or of it probably is a reconnection event. the ionized gases in space (white arrows) can force two fields lines together, allowing them to I was, however, troubled by one break from one configuration and reconnect in another (b). This process injects energy into thing. With no stellar companion, the the surrounding plasma, accelerating charged particles and giving off a burst of high-fre- space in which a reconnection event quency radiation (c).

Copyright © 2001 American Scientist 2001 January–February 43 could happen would be little more than Mercury Earth the size of the star, as is the case with solar flares. Because the energy released Sun HD 83443 0.35 is proportional to the volume of the HD 46375 0.25 flaring region, a star-sized reconnection HD 187123 0.54 could never generate enough oomph to Tau Boötis 4.14 explain superflares. A second object BD 103166 0.48 with a magnetic field was clearly need- HD 75289 0.46 ed to make my idea work. Yet a second HD 209458 0.63 star in orbit, even one that was so small 0.46 and dim that it could not be seen, b 0.68 would induce enough of a wobble in HD 168746 0.24 the bigger star to generate telltale HD 217107 1.29 Doppler shifts in its spectral lines. And HD 162020 13.73 despite having made careful spectro- HD 130322 1.15 scopic observations of these stars, HD 108147 0.35 Schaefer had not found any such evi- HD 38529 0.77 dence for an unseen companion. Then 55 Cancri 0.93 in a heartbeat I thought of one such ob- GJ 86 4.23 ject that he would have missed: Jupiter. HD 195019 3.55 Jupiter is only about a thousandth as HD 6434 0.48 massive as the Sun. Yet this gaseous 0.81 HD 192263 planet has a strong magnetic field, HD 83443 C 0.16 which interacts with the solar wind. In- GJ 876 2.07 Rho Corona Borealis 0.99 deed, astrophysicists believe that HD 168443 7.18 episodes of magnetic reconnection can HD 121504 0.89 explain some of the bursts of radio HD 16141 0.22 waves and higher-energy radiation HD 114762 10.96 that Jupiter occasionally emits. 7.42 So as we continued toward our des- HD 52265 1.14 tination, I imagined what would hap- HD 1237 3.45 pen if Jupiter orbited the Sun closely HD 37124 1.13 enough that the magnetic fields of these HD 202206 14.68 two bodies could become fully en- HD 12661 2.83 twined. A few steps later it struck me HD 134987 1.58 that Jupiter-like planets have been de- HD 169830 2.95 tected orbiting very close to their host 2.05 stars. Astronomers now know of more HD 89744 7.17 than a few such planets, sometimes Iota Horologii 2.98 dubbed “hot .” Like the real HD 92788 3.86 Jupiter, they could well have large mag- HD 177830 1.24 netic fields and thus could play the HD 210277 1.29 supporting role of unseen magnetic 2.3 HD 82943 companion. The reconnection of field HD 222582 5.18 lines between such a planet and its star B 1.68 2.60 would then be expected to do just what HD 10697 6.08 it does on an RS CVn binary—generate HD 190228 5.0 immense flares now and again. 4.29 With his “These are not binaries” Epsilon Eridani 0.8 still lingering in the air, I strolled over 5.55 to Schaefer and offered him my solu- tion to the conundrum. We spent the 012345rest of the walk (and more than a few orbital semimajor axis (astronomical units) slices of pizza) with him poking at the idea and trying, unsuccessfully, to find a problem with it. By the end of the Figure 8. Partial roster of suspected extrasolar planets demonstrates that massive, Jupiter-like meal, he agreed that the notion was bodies can circle in tight orbits. These discoveries have been made during the past five years quite compelling: It required no new from spectroscopic measurements that reveal wobbles in the parent star toward and away from Earth. Minimum estimates for the of each object are given here as the number of Jupiter physical process; it did not invoke any . The diameter of the dot shows the relative size of these bodies, assuming that they have exotic type of astrophysical object; and masses close to the values indicated and are of similar density to Jupiter. Such bodies probably it satisfied all of the observations. What support large magnetic fields, which may become entangled with the magnetic field of the par- is more, this explanation predicts that ent star and spark a superflare. the stars involved should have rela-

44 American Scientist, Volume 89 Copyright © 2001 American Scientist tively strong magnetic fields. Although looks like up close. There must be at tice a seemingly normal, sunlike star estimates of stellar magnetic fields are least one orbiting searing- that for a few minutes appears sub- difficult to make, two of the nine su- ly near to the central yellow star—per- stantially brighter than it ever has be- perflaring stars have been measured in haps more closely than Mercury orbits fore. To my mind, this is a good day. this way, and both show high field the Sun. The planet rotates rapidly, a strengths (in excess of a kilogauss) over prerequisite for generating a large Bibliography substantial portions of their surfaces. magnetic field, and this fast spin sets Ashbrook, J. 1959. The S Fornacis puzzle. Sky What is needed now is some direct up a conspicuous pattern of banding and Telescope 18:427. evidence for giant planets in close orbit in its atmosphere. The large striped Beardsley, W. R., G. Gatewood and K. W. Kam- around these stars. Such determina- planet, and perhaps its several moons, per. 1974. A study of an early flare, radial velocities, and residuals for possi- tions are, however, extremely challeng- are probably bathed in x rays and ul- ble orbital motion of HD 103095 (Groom- ing, which explains why only a few traviolet light produced by the stellar bridge 1830). The Astrophysical Journal teams of astronomers have succeeded corona and by the interplay of strong 194:637–643. in discovering extrasolar planets. Geoff magnetic fields. Some of the field lines Klimchuk, J. A. 1997. Magnetic energy release Marcy, who leads such an effort at the of this connect fully with on the Sun. Nature 386:760–761. University of California, Berkeley and those emerging from the star, permit- Landini, M., C. Monsignori Fossi, R. Pallavicini and L. Piro, 1986. EXOSAT detection of an San Francisco State University, has sur- ting charged particles from the stellar X-ray flare from the solar type star π1 UMa. veyed one of the nine superflaring stars corona to rain down on the planet and Astronomy and Astrophysics 157:217–222. (Kappa Ceti) for planets and has found its companions, lighting up the night- Rubenstein, E. P., and B. E. Schaefer. 2000. Are nothing. But he reports that the noise time sky with bright . superflares on solar analogues caused by in these measurements is so large that Over the years, the orbit of this plan- extrasolar planets? The Astrophysical Journal he cannot rule out a planet the size of et has slowly tangled its magnetic field 529:1,031–1,033. orbiting within 0.1 astronomical with that of the star. Like the spiral Schaefer, B. E. 1989. Flashes from normal stars. The Astrophysical Journal 337:927–933. unit (or A.U., about 150 million kilome- spring of a clock, the linked field lines Schaefer, B. E., J. R. King and C. P. Deliyannis. ters) of the star—or even a Jupiter-like are wound around and around. Now is 2000. Superflares on ordinary solar-type stars. planet circling somewhat farther out. the moment they break loose. In just The Astrophysical Journal 529:1,026–1,030. So Schaefer and I will have to wait for minutes, all this potential energy more definite word from Marcy and his courses into the surrounding stellar fellow planet-hunters before we know plasma, rapidly accelerating many of for sure whether we are right. these charged particles. The affected Links to Internet resources for electrons and ions immediately give off further exploration of Spectacular Fireworks a blast of x rays and ultraviolet radia- “Superflares and Giant Planets” Even without conclusive proof of the tion. Some fly outward at nearly the are available on the American Scientist Web site: idea, it is tempting to think that a speed of light. http://www.americanscientist. org/ mechanism for generating superflares Is this the day that some extraterres- articles/01articles/rubenstein.html is now known, at least in outline. It is trial civilization meets its end? Proba- also tempting to try to envision what bly not. But in time someone on Earth one of these strange planetary systems may look up at the night sky and no-

Copyright © 2001 American Scientist 2001 January–February 45