IMPLICATIONS OF ULTRAHIGH ENERGY AIR SHOWERS FOR PHYSICS AND ASTROPHYSICS F.W. STECKER Laboratory for High Energy Astrophysics NASA Go,:ldard Space Flight Center, Greenbelt, MD, USA ABSTRACT The primary ult_ahigh energy particles which produce giant extensive air showers in the Earth's atmosphere present an intriguing mystery from two points of view: (1) How are l.hese particles produced with such astounding energies, eight orders of magnitude !_igher than those produced by the best man-made terrestrial accelerators? (2) Sin,;e they are most likely extragalactic in origin, how do they reach us from extrag._lactic distances without suffering the severe losses expected from interactions witch the 2.7 K thermal cosmic background photons - the so- called GZK effect? The answers to these questions may involve new physics: violations of spe- cial relativity, grand unification theories, and quantum gravity theories involving large extra dimensions. They may involve new astrophysical sources, "zeva- trons". Or some he_-etofore totally unknown physics or astrophysics may hold the answer. I will di:;cuss here the mysteries involving the production and extra- galactic propagation of ultrahigh energy cosmic rays and some suggested possible solutions. Subject headings: ultrahigh energy cosmic rays, active galactic nuclei, gamma-ray bursts, topological &_fects, grand unification 1. Introduction About once per cenl_lry per km 2 of the Earth's surface, a giant shower of charged particles produced by a primary particle with an energy greater than or equal to 16 joules (100 EeV = 1020 eV) pl,)ws through the Earth's atmosphere. The showers which they produce can be detected I)y arrays of scintillators on the ground; they also announce their presence by producing a trail of ultraviolet flourescent light, exciting the nitrogen atoms in the atmosphere. The existence of such showers has been known for ahnost four decades -2- (Linsley 1963). The number of giant air showers detected from primaries of energy greater than 100 EeV has grown into the double digits and may grow into the hundreds as new detectors such as the "Auger" array and the "EUSO" (Extreme Universe Space Observatory) and "OWL" (Orbiting Wide-Angle Light Collectors) satellite detectors come on line. These phenomena present an intriguing mystery from two points of view: (1) How are particles produced with such astounding energies, eight orders of magnitude higher than are produced by the best man-made terrestrial accelerators? (2) Since they are most likely extragalactic in origin, how do they reach us from extragalactic distances without exhibiting the predicted cutoff from interactions with the 2.7K cosmic background radiation? In these lectures, I will consider possible solutions to this double mystery. 2. The Data Figure 1 shows the published data (as of this writing) on the ultrahigh energy cosmic ray spectrum from the Fly's Eye and AGASA detectors. 1 Other data from Havera Park and Yakutsk may be found in the review by Nagano and Watson (2000) are consistent with Figure 1. Additional data are now being obtained by the HiRes detector array and should be available in the near future (T. Abu-Zayyad, el al. , in preparation). For air showers produced by primaries of energies in the 1 to 3 EeV range, Hayashida, et al. (1999) have found a marked directional anisotropy with a 4.5a excess from the galactic center region, a 3.9a excess from the Cygnus region of the galaxy, and a 4.0a deficit from the galactic anticenter region. This is strong evidence that EeV cosmic rays are of galactic origin. A galactic plane enhancement in EeV events was also reported by the Fly's Eye group (Dai, et al. 1999). As shown in Figure 2, at EeV energies, the primary particles appear to have a mixed or heavy composition, trending toward a light composition in the higher energy range around 30 EeV (Bird, et al. 1993; Abu-Zayyad, et al. 2000). This trend, together with evidence of a flattening in the cosmic ray spectrum on the 3 to 10 EeV energy range (Bird, et al. 1994; Takeda et al. 1998) is evidence for a new component of cosmic rays dominating above 10 EeV energy. The apparent isotropy (no galactic-plane enhancement) of cosmic rays above 10 EeV (e.g. Takeda, et al. 1999), together with the difficulty of confining protons in the galaxy at 1The AGASA data have been reanalysed and the number of events determined to be above 100 EeV has been lowered to eight. (Teshima, private communication.) 10 to 30 EeV energies, pwvide significant reasons to believe that the cosmic-ray component above 10 EeV is extragal;_ctic in origin. As can be seen from Figure 1, this extragalactic component appears to ext,.nd to an energy of 300 EeV. Extention of this spectrum to higher energies is conceivable t)e,:ause such cosmic rays, if they exist, would be too rare to have been seen with present delectors. We will see in the next section that the existence of 300 EeV cosmic rays gives us ;_ new mystery to solve. 3. The GZK Effect Thirty seven years ag(,, Penzias and Wilson (1965) reported the discovery of the cosmic 3K thermal blackbody ra, liation which was produced very early on in the history of the universe and which led to the undisputed acceptance of the "big bang" theory of the origin of the universe. Much re.ire recently, the Cosmic Background Explorer (COBE) satellite confirmed this discovery, showing that the cosmic background radiation (CBR) has the spectrum of the most perl)ct thermal blackbody known to man. COBE data also showed that this radiation (on angular scales > 7°) was isotropic to a part in 10s (Mather et al. 1994). The perfect thenm, i character and smoothness of the CBR proved conclusively that this radiation is indeed cosmological and that, at the present time, it fills the entire universe with a 2.7K spectrum of r:tdio to far-infrared photons with a density of _ 400 cm -a. Shortly after the discovery of the CBR, Greisen (1966) and Zatsepin and Kuz'min (1966) predicted that pioll-producing interactions of ultrahigh energy cosmic ray protons with CBR photons of targ,.,t density ,,_400 cm -a should produce a cutoff in their spectrum at energies greater than -,_50 EeV. This predicted effect has since become known as the GZK (Greisen-Zatsepin-Kuz'miJ_) effect. Following the GZK papers, Stecker (1968) utilized data on the energy dependence of the photomeson production cross sections and inelasticities to calculate the mean energy loss time for protons propagating through the CBR in intergalactic space as a function of ene_gy. Based on his results, Stecker (1968) then suggested that the particles of energy above tile GZK cutoff energy (hereafter referred to as trans-GZK particles) must be coining from witt_in the "Local Supercluster" of which we are a. part and which is centered on the Virgo Clu>ter of galaxies. Thus, the "GZK cutoff" is not a true cutoff, but a supression of the ultrahiI,_h energy cosmic ray flux owing to a limitation of the propagation distance to a few tens of l_Ipc. The actual position ot' the GZK cutoff can differ fi'om the 50 Ee\/predicted by Greisen. In fact, there could actually be an er_hancement at or near this energy owing to a "pileup" of cosmic rays starting oul at higher energies and crowding up in energy space at or below the predicted cutoff energy (Puget, Stecker and Bredkamp 1976; Hill and Schramm 1985; -4- Fly's Eye v • A•.•,• a_.•A, • vll Ie+24 % te+"3"e÷17 1o+18 le+t9 1e-¢-20 le+21 Er_q w(eV) Fig. 1.-- The ultrahigh energy cosmic ray spectrum data from Fly's Eye and AGASA. 900 .... , .... , .... , .... _.... , .... , .... 850 8OO 750 E 700 E X 650 • [!_ _' Haverah Park [8] 600 "__lgVLL 2.1 [1"1] --- _--- OG_ET[12] ..... 55O 5OO , , , , l , , • , i .... I .... I .... I .... i .... 16.5 17 17.5 18 18.5 19 19.5 20 IOglo (E(eV) ) Fig. 2.-- Average depth of shower maximum (Xm_) vs. energy compared to the calculated values for protons (upper curves) and Fe primaries (lower curves)(from Gaisser 2000; see references therein). 5 BerezinskyandGrigor'eva 1988; Stecker1989; SteckerandSalamon1999).The existence and intensity of this predicted pileup dependscritially on the flatnessand extent of the sourcespectrum,(i.e., the numberofcosmicraysstarting out at higherenergies),but if its existenceisconfirmedin thefuture by moresensitivedetectors,it wouldbeevidencefor the GZK effect. Scully and Stecker (2002) have determined the GZK energy, defined as the energy for a flux decrease of l/e, as a fimction of redshift. At high redshifts, the target photon density increases by (1 + z) 3 and both the photon and initial cosmic ray energies increase by (1 + z). The results obtained by S,:ully and Stecker are shown in Figure 3. 4. Acceleration and Zevatrons: The "Bottom Up" Scenario The apparent lack of a GZK cutoff has led theorists to go on a hunt for nearby "ze- vatrons", i.e., astrophysic,,1 sources which can accelerate particles to energies O(1 ZeV = 102aeV). In most theoretical a,)rk in cosmic ray astrophysics, it is generally assumed that the diffusive shock acceleration process is the most likely mechanism for accelerating particles to high energy. (See, e.9., Jones (2000) and references therein.) In this case, the maximum obtainable energy is given by Emaz = keZ(u/c)BL, where u _< c is the shock speed, eZ is the charge of the particle T,eing accelerated, B is the magnetic field strength, L is the size of the accelerating region an, t the numerical parameter k = O(1) ( Drury 1994). Taking k = 1 and u = c, one finds = 0.gz(BL) with E in EeV, B in t*G and R in kpc. This assumes that particles can be accelerated efficiently up until the moment when they can no longer be contained by the source, i.e. until their gyroradius bec_,mes larger than the size of the source. Hillas (1984) used this relation to construct a plot of B vs. L for various candidate astrophysical objects. A "Hillas plot" of this kind, recently constructed by Olinto (2000), is shown in Figure 4. Given the relationship between E_ and BL as shown in Figure 4, there are not too many astrophysical candidates for zevatrons. Of these, galactic sources such as white dwarfs, neutron stars, pulsars, an,l magnetars can be ruled out because their galactic distribution would lead to anisotropie_,: above 10 EeV which would be similar to those observed at lower energies by Hayashida el al (1999), and this is not the case. Perhaps the most promising - 6 - 215 g t8 t*z Fig. 3.-- The GZK cutoff energy versus redshift (Scully and Stecker 2002). 15 I • I .... I • " ' ' I .... I Neutron ._._. Stars N 5 -,_,..\o._ J.obe_ ¢ -o_ .. /, c,ua._=j - _-.. //c,_t_.s .... o -5 .... , .... ,, ,--_,-4_oT.... ,"._.';... -10 5 I0 15 20 25 30 Log[Size(c,n) ] Fig. 4.-- A "Hillas Plot" showing potential astrophysical zevatrons (from Olinto 2000). The lines are for B vs. L for Emax = 0.1 ZeV for protons and iron nuclei as indicated. -- 7 -- potential zevatrons are radio lobes of strong radio galaxies (Biermann and Strittmatter (1987) . The trick is tha_ such sources need to be found close enough to avoid the GZK cutoff (e.g., Elbert and Sommers 1995). Biermann has further suggested that the nearby radio galaxy M87 may be _he source of the observed trans-GZK cosmic rays (see also Stecker 1968; Farrar and Piran 2000). Such an explanation would require one to invoke magnetic field configurations capable' of producing a quasi-isotropic distribution of > 1020 eV protons, making this hypothesis questionable. However, if the primary particles are nuclei, it is easier to explain a radio galaxy origin for the two highest energy events (Stecker and Salamon 1999; see section 6). It has also been sugge,sted that since all large galaxies are suspected to harbor super- massive black holes in th_qr centers which may have once been quasars, fed by accretion disks which are now used up, that nearby quasar remnants may be the searched-for zeva- trons (Boldt and Ghosh 1!199; Boldt and Lowenstein 2000) . This scenario also has potential theoretical problems and needs to be explored further. In particular, it has been shown that black holes which arc not accreting plasma cannot possess a large scale magnetic field with which to accelerate p_t.rticles to relativistic energies (Ginzburg and Ozernoi 1964; Krolik 1999; Jones 2000). Observational evidence also indicates that the cores of weakly active galaxies have low magnetic fields (Falcke 2001 and references therein). Another proposed zevatron, the ")'-ray burst, is discussed in the next section. 5. Gamma-Ray Burst Zevatrons and the GZK Problem In 1995, it was sugge,_d;ed that cosmological 7-ray bursts (GRBs) were the source of the highest energy cosmic rays (Waxman 199.5; Vietri 1995). It was suggested that if these objects emitted the same amount of energy in ultrahigh energy (,,_ 1014 MeV) cosmic rays as in ,-, MeV photons, there would be enough energy input of these particles into intergalactic space to account for the _,bserved flux. At that time, it was assumed that the GRBs were distributed uniformly, ind,,pendent of redshift. In recent years, X-ray, optical, and radio afterglows of about a dozen GRBs have been detected leading to the s_lbsequent identification of the host galaxies of these objects and consequently, their redshifts. The host galaxies of GRBs appear to be sites of active star formation. The colors and morphological types of the host galaxies are indicative of ongoing star formation, as is the detection of Lya and [Oil] in several of these galaxies. Further evidence suggests that bursts themselves are directly associated with star forming regions within their host galaxies their positions correspond to regions having significant hydrogen column densities with evi,lence of dust extinction. It now seems more reasonable to assume -8- that a moreappropriate redshift distribution to take for GRBsis that of the averagestar formation rate. Todate,some14GRBsafterglowshavebeendetectedwith a subsequentidentification oftheir hostgalaxies.Asofthis writing, 13ofthe 14areat moderateto highredshiftswith the highestone(GRB000131)lying at aredshift of4.50(Andersen,et al. 2000). A good argument in favor of strong redshift evolution for the frequency of occurrence of the higher luminosity GRBs has been made by Mao and Mo (1998), based on the nature of the host galaxies. Other recent analyses have also favored a GRB redshift distribution which follows the strong redshift evolution of the star formation rate (Schmidt 1999; Penimore and Ramirez-Ruiz 2000). If we thus assume a redshift distribution for the GRBs which follows the star formation rate, being significantly higher at higher redshifts, GRBs fail by at least an order of magnitude to account for the observed cosmic rays above 100 EeV (Stecker 2000). If one wishes to account for the GRBs above 10 EeV, this hypothesis fails by two to three orders of magnitude (Scully and Stecker 2002). Even these numbers are most likely too optimistic, since they are based on the questionable assumption of the same amount of GRB energy being put into ultrahigh energy cosmic rays as in ,-* MeV photons. Figure 5, from Scully and Stecker, (2002) shows the form of the cosmic ray spectrum to be expected from sources with a uniform redshift distribution and sources which follow the star formation rate. The required normalization and spectral index determine the energy requirements of any cosmological sources which are invoked to explain the observations. Pileup effects and GZK cutoffs are evident in the theoretical curves in this figure. As can be seen in Figure 5, the present data appear to be statistically consistent with either the presence or absence of a pileup effect. Future data with much better statistics are required to determine such a spectral structure. An unusual nearby Type Ic supernova, SN 1998bw, has been identified as the nearby source of a low luminosity burst, GRB980425, with an energy release which is orders of magnitude smaller than that for a typical cosmological GRB. Norris (2002) has given an analysis of the luminosities and space densities of such nearby low luminosity long-lag GRB sources which are identified with Type I supernovae. For these sources, he finds a rate per unit volume of 7.8 x 10.7 Mpc-3yr -1 and an average (isotropic) energy release per burst of 1.3 x 1049 erg over the energy range from 10 to 1000 keV. The energy release per unit volume is then ,-_ 104a erg Mpc-ayr -1. This rate is more than an order of magnitude below the rate needed to account for the cosmic rays with energies above 10 EeV. -9- le+26 ........ , ........ _ ........ , ......... . J le+25 ft ...... ., l_-_I i 1e+24 "\J -, i >{ "K 1e+23 ",, % i....... I le+22 ........ ' ........ ' ........ L . L..... le+17 le+18 le+19 le+20 le+21 Energy (eV) Fig. 5.-- Predicted spect] n for cosmic ray protons as compared with the data. The middle curve and lowest curve assume an E -2"7_ source spectrum with a uniform source distribution and one that follows the : distribution of the star formation rate respectively. The upper curve is for an E -2"35 source spectrum which requires an order of magnitude more energy input and exhibits a "pile_lp effect" (Scully and Stecker 2002). - 10 ¸ 6. The Heavy Nuclei Origin Scenario A more conservative hypothesis for explaining the trans-GZK events is that they were produced by heavy nuclei. Stecker and Salamon (1999) have shown that the energy loss time for nuclei starting out as Fe is longer than that for protons for energies up to a total energy of--_300 EeV (see Figure 6). Stanev, et al. (1995) and Biermann (1998)1 have examined the arrival directions of the highest energy events. They point out that the --_200 EeV event is within 10° of the direction of the strong radio galaxy NGC 315. This galaxy lies at a distance of only ,-- 60 Mpc from us. For that distance, the results of Stecker and Salamon (1999) indicate that heavy nuclei would have a cutoff energy of ,-- 130 EeV, which may be within the uncertainty in the energy determination for this event. The ,,_300 EeV event is within 12° of the direction of the strong radio galaxy 3C134. The distance to 3C134 is unfortunately unknown because its location behind a dense molecular cloud in our own galaxy obscures the spectral lines required for a measurement of its redshiff. It may be possible that either cosmic ray protons or heavy nuclei originated in these sources and produced the highest energy air shower events. An interesting new clue that we may indeed be seeing heavier nuclei above the proton- GZK cutoff comes from a very recent analysis of inclined air showers above 10 EeV energy (Ave, et al. 2000). These new results favor proton primaries below the p-GZK cutoff energy but they appear to favor a heavier composition above the p-GZK cutoff energy. It will be interesting to see what future data from much more sensitive detectors will tell us. 7. Top-Down Scenarios: "Fraggers" A way to avoid the problems with finding plausible astrophysical zevatrons is to start at the top, i.e., the energy scale associated with grand unification, supersymmetric grand unification or its string theory equivalent. The modern scenario for the early history of the big bang takes account of the work of particle theorists to unify the forces of nature in the framework of Grand Unified Theories (GUTs) (e.g., Georgi and Glashow 1974). This concept extends the very successful work of Nobel Laureates Glashow, Weinberg, and Salam in unifying the electromagnetic and weak nuclear forces of nature (Glashow 1960; Weinberg 1967; Salam 1968). As a consequence of this theory, the electromagnetic and weak forces would have been unified at a higher temperature phase in the early history of the universe and then would have been broken into separate forces through the mechanism of spontaneous symmetry breaking caused by vacuum fields which are known as Higgs fields.