A few percent of observed galaxies have an unusual activity which is concentrated to the central area i.e nucleus of the galaxy. This activity means a tremendous luminosity, variability in the energy output and possible side effects like jets, lobes and signs of an explosive activity. Also broad spectral lines are observed indicating rapid rotation near the nucleus. The luminosity of an AGN can be ten or even ten thousand times the luminosity of normal galaxy. Sometimes the amount of energy released is so large that it outshines the rest of the galaxy. It's assumed that this kind of activity is only possible for a short period of time, not for the entire lifetime of a galaxy but for one period of it. The cause of the activity is a supermassive black hole (BH) that lies in the center of AGNs. The mass of the black hole is usually in the order of 10⁷ to 10⁹ Solar masses. It is surrounded by gas that has formed an accretion disk around it and the matter in the disk is accreting into the black hole. The larger the mass of the black hole, the faster the accreting material rotates in the disk. This causes heating of the material in the disk and excessive radiation. The accretion releases the gravitational potential energy of particles and this is the source of energy for all observed activity in AGNs. Over 10% of the rest mass of the accreting material is converted into radiation..

Figure 1. AGN with a torus and jets. (Artist impression, Chandra X-ray Observatory web site.)

The supermassive black holes were born when galaxies were young and their stars were very closely packed in the galaxies' core areas. The stars merged with each other as a result of the short distances, coalescing into a massive star, which eventually evolved into a massive black hole. The interstellar media and the stars that were getting too close to the black hole were eaten by it. The stars were broken apart by tidal forces and the matter spiraled to the black hole, forming an accretion disk around it. As long as there is continuous accretion, the nuclei is active. Galaxies with massive black holes in their centre, that are not accreting mass like our own Milky Way, are non-active and quiet.

The AGNs are classified in many different ways but the two 'main' classes are Seyfert galaxies and radiogalaxies. The former are thought to represent the active phase of spiral galaxies and the latter of elliptical galaxies. The radiogalaxies contain quasars and blazars which have supermassive BH in their centers, while the Seyfert galaxies have only moderate-mass BH. Furthermore, there is also a unification scheme that unites the different radio galaxies, stating that the observed type of the galaxy depends only the orientation of the galaxy to the observer. 

Seyfert galaxies

Seyfert galaxies are named after Carl Seyfert who discovered them in 1943 when studying spiral galaxies. About 2% of spiral galaxies are Seyfert galaxies. The typical features of Seyfert galaxies are a morphology of spiral galaxy (if it has been possible to determine), and a bright point-like nucleus where the massive black hole resides. The luminosity of the nucleus is of the order of 10³⁶ to 10³⁹ W (as bright as all the other parts of the galaxy together) and the value can vary with time. Seyfert galaxies are (as mentioned before) radio quiet and also gamma quiet, which is due to the fact that these galaxies have only weak jets.

The Seyfert galaxies are divided into two different subclasses 1 and 2, depending on the angle from which the galaxy is observed. This means that the Seyfert galaxies are observed either central torus viewed face-on, as in the case of Seyfert 1 galaxies or edge.on, as Seyfert 2 galxies. This can be seen in figure 2 below.

Figure 2. Different viewing angles of a torus.(Provided by Chandra X-ray Observatory.)

In the case of Seyfert 1 galaxies,  the optical permitted emission lines are very broad, corresponding to the velocities of  about 10⁴ km/s. They are formed rather close to the central black hole. One sees also narrow forbidden lines which are  formed in clouds probably along the axis of the symmetry far from the center. The X-ray source, which is close to the black hole, can be viewed directly, since the absorption in X-rays is weak. The X-rays are produced either in the hot accretion flow or in the rarefied corona above the cold accretion disk. This intrinsic continuum is reflected from the disk giving rise to so called Compton reflection bump in the spectrum at 30 keV and a fluorescent iron 6.4 keV. 

The X-ray radiation illuminates the accretion disk so that only up to 10 % of the radiation is reflected, and the rest of the energy is thermalized in the disk. This thermalized energy is radiated from the disk. This radiation constitutes the Comptonized part of the spectrum from soft X-rays to UV region. The iron K-alpha line (absorption line) is produced as well. The spectra of Seyfert 1 galaxy is seen in figure 3 below.

Figure 3. The Spectrum of Seyfert I galaxy IC 4329A.(Zdziarski et al.1996)

In the case of Seyfert 2 galaxies,  the optical permitted lines are now rather narrow  with velocities of  10³ km/s. The broad emission line
region is now blocked by the torus from our view. The radiation from the central part can be scattered in the region above the torus and still be observed in  polarized light. The X-ray  radiation from the central source is also blocked. Only the hard X-rays can penetrate through the torus and some radiation can again be scattered to our line of sight. The spectrum of type 2 galaxy is shown in figure 4 below.

Figure 4. The Seyfert 2 galaxy NGC 4945. (From Madejski et al. 2000.)

Black hole mass in NGC 4258 from water masers

Serious evidence for the existence of supermassive black holes in the centers of active galaxies was the discovery of a disk in the center of  NGC4258, a weakly active Seyfert 2 galaxy. Studying the water maser emission (mechanism is similar to the laser, but the radiation is produced in the microwave/radio band instead of the optical)  at 22 GHz Miyoshi et al. (1995)  showed that the disk is nearly Keplerian with velocity of ~1000 km/s  and estimated the mass of the object within ~0.1 pc from the center of about 3.6*10⁷ solar masses. This AGN has a weak jet extending to kpc scales from the center of the galaxy and it is well aligned with the rotational axis of the  disk.  The measurements of masers proper motions and accelerations give an independent way to determine the distance to the galaxy.

Figure 5. NGC 4258. View in optical. Image and a model for  water maser emission is shown in the inset. Credit: Nature.

Radiogalaxies

Radio galaxies are usually giant elliptical galaxies that radiate strongly in the radio band with luminosity of 10³³ to 10³⁸ W (i.e. 10⁴⁰ to 10⁴⁵ erg/s, in radio range). The radio spectrum is non-thermal and originates from the synchrotron radiation of electrons. The source of this radiation is in jets and radio lobes. The power-source of radiogalaxy is, like in all of the cases of AGNs, the super-massive black hole. The brightest radio galaxies observed are Centaurus A, Cygnus A and Virgo A.

Figure 6: Cen A radio galaxy at about 2.5 Mpc. Combined HST image of the galaxy with the dust lanes and the VLA (6cm) image if the jet with  radio lobes.

Observations in different wavelengths give the more total view of a radiogalaxy. Such features like radio-lobes and jets could not be observable in the optical range, but in radio wavelengths these are strong radiators. The jets are two opposite narrow plasma outflows from the nuclear region of the galaxy, and they coincide with the rotational axis of the galaxy. It is thought that they are brought forth as consequence of the winding up of the magnetic field in the accretion disk around the black hole. This winding up of magnetic field provides a way to convert gravitational and rotational energy into these high speed (up to 99.9% of speed of light) flows. Depending on how powerful the galaxy's central engine is, the jets can extend far outside the galaxy or they can be stopped much earlier by the surrounding matter it is pushing through. These plasma outflows consist of high energy electrons, protons (with possible mixture of positrons) and magnetic fields. The radiating electron gas is highly relativistic and radiate synchrotron radiation, when electrons spiral  around magnetic field lines.  The orientation of the jet can vary as a result of the precession of the accretion disk or as a result of the gas density gradients in the surrounding medium.  

Figure 7. A schematic picture of a radio-loud galaxy. From Urry & Padovani (PASP, 1995).

On their way penetrating through the surrounding materia, the jets drag the materia along with them and lose their kinetic energy. The jets have to go through many different zones of materia: the matter near the active nucleus, the interstellar medium of the host galaxy, the outer halo of the galaxy, the intra-cluster medium of group or cluster of galaxies and intergalactic medium. Even the most powerful jets are eventually stopped in the low-density intergalactic medium forming shocks. These shocks are observed as intensively emitting "Hot Spots". These hot spots and the materia dragged along by the jet form a radio lobe. Because of the symmetry of jets, there are more or less symmetric lobes on each side of the galaxy. These lobes can be stretched as far as a million light-years from nucleus of the galaxy and the sizes are usually much bigger than the galaxy itself. Many of them are of the order of 60 kpc in their diameter. As mentioned earlier, the lobes radiate strongly in the radio band and the radiative process is the same as in the case of jets, namely the synchrotron radiation. However, the process of acceleration the electrons in the lobes is not quite clear. The total energy of one lobe can be as high as 10⁵³ Joules, compared with an ordinary supernovae explosion where total energy output is around 10⁴⁴ Joules. These radio galaxies which have lobes are also called DRAGNs, double radio source associated with a galactic nucleus. The plasma in the lobes are thought to be supplied by the jets over millions of years and they keep growing as the central engine keeps functioning and jets are being produced. The smallest DRAGN known is a few tens of parsecs across while the largest is about 6 Mpc. The estimated lifetimes for powerful DRAGNs is 20 million years but galaxies with weak lobes are very common meaning that the jets aren't switched off  at all.

Figure 8. Parts of a DRAGN of Cygnus A.

Quasars = quasistellar radio sources

When quasars were discovered in the 1960s, they first appeared to be stars (i.e point like) that were strong radio sources. That is the origin for the name quasar, quasi-stellar radio source. The peculiarity of quasars can be seen from their spectrum, the lines are tremendously redshifted. This tells that the source is very far away and we are observing antecessors of galaxies when they were young, their age being about one tenth of the universe age.

Actually the quasar that is just the central area of a galaxy, but is many times brighter than the host galaxy itself so only the center is observed. This central area i.e quasar is only a couple of light days its diameter but very luminous, in the order of 10³⁹ to 10⁴¹ W. The brightness of quasars varies very rapidly.

The optical spectra of quasars are similar to the optical spectra of the nucleus of a Seyfert galaxy. This indicates that the mechanisms of producing the spectra are the same in both cases. A quasar is a huge black hole (hundreds of millions of solar masses) surrounded by an accretion disk from which material accretes into the hole. There is also a host galaxy that is out-shined by the luminous nucleus. The accreting material consists of ISM, and because of the huge gravitational field the tidal forces can break apart stars near the center which contribute to the mass of accreting material. The most powerful quasars have accretion rates in the order of several solar masses per year. The quasars produce jets as well. Quasars tend to have rapid variability in the X-ray emission like Seyferts. The timescales for the variations can be as short as days, hours or even minutes.

Figure 9. Quasar 3C273 with a jet (by Chandra X-ray Observatory).

The spectrum of a quasar consists of two different parts, powerlaw from radio to infrared region and a separate emitting region called big blue bump that ranges from visual to UV region and peaks around 3000 Å. The powerlaw part is caused by synchrotron radiation by electrons in the magnetic field that is produced in the central object, and the big blue bumb is caused by thermal radiation from the accretion disk which is the sum of cool outer, warm middle and hot inner part of the disk. The contribution of these different components in the spectra can be seen in figure 10. The spectrum of a quasar contains also very broad emission lines.

Figure 10. A typical spectrum of a quasar.

There are also very densely distributed narrow absorption lines in the spectra, but the origin is different from the emission lines. These Lyman-series of absorption lines are thought to been produced in dust clouds between the quasar and the observer. These dust clouds are either young galaxies or progenitors of them. So the information in absorption lines gives information about the formation of galaxies.

Blazars

The first blazar discovered was BL Lacertae in 1941. It was classified as a variable star by mistake till it was confirmed to be a strong radio-source in 1968. Nowadays BL Lac type sources are classified as a subgroup of blazars which are the most energetic AGNs with certain features. To be classified as a blazar, the AGN has to have at least one of the following properties: the source has a high radio-brightness combined with flatness of radio-spectrum, the optical spectrum is polarized or it shows strong optical variability in very short periods which are less than few days (flare-like behavior). Other features typical for blazars are that it appears to be a point-like object in the sky and that the blazars can emit radiation in a broad range, from radio to gamma-rays. They are the most, energetic radiators of AGNs emitting a significant fraction of their radiation in higher energies than 100 MeV. The blazar is said to be a red-blazar if its flux peaks around 10 MeV-1 GeV, a blue-blazar if at 1 GeV-100GeV and TeV-blazar if the flux peaks around 200 GeV-1TeV. Blazars that emit gamma-rays are thought to have more tightly bound jets than usual radio-galaxies.

The emitted radiation seems to be coming from the jet area of the blazar. The smooth and featurless spectrum indicates that the blazar is observed as one of its jets pointing toward the observer. By smoothness of spectrum is meant that no spectral lines are observed and so the radiation hasn't gone through the torus around the nucleus of the galaxy. So with blazars there is an opportunity to study what is physically going on in jets. The spectral energy distribution is usually double-humped where the first component peaks anywhere between IR to X-rays, the second peaks in gamma-rays. The first component is polarized, rapidly variable and originates from synchrotron emission by electrons in the jet. The second component is produced by mechanism called inverse Compton (IC) scattering, where photons (seed photons) gain energy by scattering from electrons. The origin of the seed photon is not quite clear and it can differ in the cases of red and blue blazars. The seed photons are thought to be external to the jet in the case of the red blazars, perhaps they come from thermal radiation of the disk, broad line regions or from the torus. In the case of the blue blazar, the seed photons are thought to be internal to the jet (synchrotron radiation). In this case the second component is created by a process called synchrotron-self Compton (SSC). See figure 11 show examples of blazar spectra.

Since radio jets of blazars are observed in close alignment to the line-of-sight, the radiation flux observed is enhanced by geometrical effect, relativistic Doppler boosting. The jet contains discrete blobs of materia and these blobs seem to be moving with apparent superluminal speeds, exceeding the speed of light, which is just an optical illusion and the matter  in the jets actually is moving with velocities of about 99% of the speed of light.  Furthermore the brightness temperature of the blobs is very high which also can be explained by Doppler effect. The apparent velocity depends on the angle between the line of sight and the jet axis.

Unification Scheme

The unification scheme explains the connection between different types of radio-loud AGNs. It is based on the orientation of the AGNs disk to the observer. When the radio-loud AGN is observed disks edge-on, i.e the central area is blocked from the direct view, a radiogalaxy is seen. If the radiation from central torus is observed without it being blocked by the central torus, the quasar is seen. The blazar is observed if the central area is observed directly and the jet is facing the observer. The picture is shown below.

Figure 12. The unification scheme for radio galaxies and quasars.

Micro-Quasars

Micro-quasars are X-ray sources in our own galaxy, Milky Way. They are actually X-ray binaries, black holes paired with normal stars. The black hole is produced in the end of evolution of very massive stars. There is a mass transfer from the normal companion star to the black hole via the accretion disk, or the gas can be captured from the wind in case of massive companions. See figure 13 below. The process is so violent that matter spiraling toward the black hole is heated to so high temperatures that X-rays are produced and mass eruptions, i.e jets, occur from the central area of the accretion disk. The jets are the reason for the name micro-quasar, although the jets sizes are smaller than that of quasars. The jets here seem also to be having super-luminal motion. The precession of the disk causes bending of jets with increasing distance from the core. Famous examples of micro-quasars are SS 433 (where jets are moving with 0.26 speed of light) and GRS 1915+105 (with velocities of ~0.92-0.98 c).

Figure 13. Artist impression of micro-quasar SS 433. (Provided by Chandra X-ray Observatory.)

In the case of SS 433, the jets are precessing around a common axis with a long period of 164 days. This is actually the precession period of the disk. The SS433 is a high mass binary system where the mass losing star is an early-type B star and the mass is accreting onto a black hole via the accretion disk. The SS433 is a highly variable radio source. It has jets that move with time in opposite directions at a rate of 3 arcsec per year. The radio maps show two lobes that are aligned with the axis of the jets. X-rays are also produced when the expanding jet material collides with the material that is already slowing down. This produces shocks and temperature of the gas arises to such values that enable the production of X-rays.

Neutron stars and the black holes are born as consequence of supernovae explosions. These are the last phases of the stellar evolution when a massive star cannot gain energy by fusion reactions anymore and blows its outer parts as an expanding shell into the interstellar space, which is observed as a supernova. See figure below as an example of supernova remnant Cassiopeia A.

Figure 14. Cassiopeia A.(Provided by Chandra X-ray Observatory.)

The inner parts of the star start to collapse at the same time and if the mass of collapsing interior is high enough, much bigger than 1,4 solar masses, it continues to collapse forming a black hole. Otherwise, a neutron star is born with a mass of around 1,4 solar masses and a radius of about 10 km. Neutron star consists of materia that is in a degenerated form. As the density of the materia increases during the collapse, the electrons cannot orbit around atoms nuclei anymore, and merge with protons, which leads to the forming of neutrons. One cm³ of this materia would weigh 10¹⁴-10¹⁵ g ie. 100 or 1000 million tons. In the core, the materia is in a superfluid form, while the surface is a two km thick rigid crust. See figure below for a schematic picture of the structure. These objects are very compressed and the black holes are even more, so usually they are referred to as compact objects. In addition, white dwarfs belong to this group, they are the final evolution stages of a low mass stars, white dwarf having a radius of 10000 km.

Figure 15. Structure of a neutron star. Credit: NASA.

Neutron stars can be observed in binary systems but they can be single stars as well. Single neutron stars (or in a binary system without gas transfer)  are detected as radio pulsars. The emission is produced when particles are accelerated in the magnetosphere of the pulsar. How exactly this happens is still a matter of active research.  The radiation is mostly non-thermal synchrotron. The radiation is beamed in small cones. As the rotational axis and the magnetic axis do not usually coincide, the cones are rotating and when it crosses the line of sight, the pulse of radiation is observed. The energy source is the rotational energy of the star, and as a consequence of the energy loss due to radiation, the rotation period of neutron star becomes longer. The neutron star is spinning down slowly and as time goes by, the radiation is no longer produced as the rotation is really slow. A sudden change of the rotational period can happen if there is a "starquake" when the stellar size suddenly changes by ~1mm. The angular momentum must be conserved and so the neutron star spin changes slightly. 

When detected in binary systems the neutron stars are usually paired either with a low mass star that has a mass less than one solar mass or with a high mass star, mass bigger than 15 solar masses. These binary systems are usually very tight meaning the orbital periods are short, distances are short between stars, and the tidal forces affect greatly in these systems. In these cases the companions are much more luminous in the optical range than neutron stars which could easily remain undetected if there wouldn't be a mass flow from the companion star to the neutron star's surface. The accretion mechanism is different in the cases of low mass companions (LMXB=low mass X-ray binary) and high mass companions (HMXB=high mass X-ray binary).

In the LMXB case, the accretion becomes possible when the low mass companion evolves and starts to expand. These are old systems, their age being in the order of billion years. The neutron stars magnetic field is the order of 10⁸-10¹⁰ Gauss. When matter reaches the gravitational equilibrium surface of the system (so called Roche Lobe), the neutron stars gravitational force can overcome that of the companion star and the plasma from the companion star can begin to move closer to the neutron star via the Lagrangian point. This gas has a non-zero angular momentum (because of the orbital motion) and it cannot go directly on the neutron stars surface. So it forms an accretion disk around it. See figure 16 below as an example of an accretion via disk in the section concerning millisecond pulsars.

Figure 16. Accretion onto SAX J1808.4-3658 via accretion disk. Credit: NASA

When HMXB is concerned, the companion star is a massive O or B (early spectral type) star or a Be-star. HMXB with a Be-companion is a more common combination from these. OB and Be stars are young (age less than 10 million years), very luminous and have strong stellar winds. Since they are young, the binary systems are also young. The magnetic fields of neutron stars are very powerful, being in the order of 10¹² Gauss. It is thought that the strength of the magnetic field decays with time.

In the case of OB companion stars (these are so called persistent sources, meaning the accretion is continuous, see below), the matter is transferred on to the neutron star by the stellar wind and no accretion disk is formed (or perhaps a very small one) because of  almost zero angular momentum of the stellar wind.

The accretion in the case of HMXB with Be-companion is quite different. The companion star is un-evolved (deep inside its Roche Lobe) and spins rapidly expelling matter to its equatorial plane. The orbital period varies between 15 days to over a year and the neutron star revolves in a very eccentric orbit penetrating to the expelled matter ring near the perihelion. The accretion occurs only when the neutron star is near the Be star and so it's called a transient source. At this point, an accretion disk is formed and the neutron star is spun up. When receding from Be star, the accretion is terminated and the neutron star begins to spin down.

If the neutron star has a magnetic field in the order of 10⁸ -10¹² G, the accreting matter  will follow the magnetic field lines to the neutron stars magnetic poles. As the material is falling towards the surface of the star, a shock front is formed close to the surface. The potential energy of plasma particles is transformed to kinetic energy and then into the heat and into radiation. Because of the rotation of the neutron star and the fact that the magnetic axis and rotational axis do not necessarily coincide, the magnetic polar caps change their positions relative to the observer. So the area that is emitting X-rays seems to be changing and  the observer sees pulsation. The pulse is strongest when the polar cap is closest to the observer, weakest when furthest. In some cases the both poles are seen giving origin to a double peaked lightcurve. However, in a number of cases, just one pole is seen producing a single peaked lightcurve. Factors here are the angle between magnetic- and rotational axes and also the angle between rotational axis and the observer. In addition, gravitational light-bending plays a role here which is an effect of general relativity. The spacetime is curved near compact objects and the photons seem to be following bend trajectories, which makes them visible even when produced on the other side of the neutron star. In these cases, the both magnetic poles can be seen at the same time.

Figure 17. The origin of the X-ray pulsation. Credit: Chandra X-ray Observatory.

Figure 18. Pulse profiles from GX 1+4. a) low (1.5-3 keV) energy band and b) the high (30-40 keV) band (Doty et al. 1981)

The accretion in LMXB happen in a different way as the magnetic field of a neutron star is weak because the system is old.  The matter is not  gathered in polar areas but spreads all over the neutron star surface. The hydrogen  burns steadily (i.e there is a fusion reaction going on) on the surface of the neutron star, producing helium. As the temperature and density grows the helium begins its fusion which is observed as a flash. At the beginning of the  burst the temperature is as high as 30 million degrees and produces hard X-rays but after 10 seconds it has decreased to 15 million degrees.

Figure 19. X-ray burst  (yellow curve) from the LMXB source 4U 1728-34 and the corresponding dynamic power density spectrum (in color). For the duration of the burst, we see oscillations around 364 Hz, which current theories attribute to the X-ray emission from a hot spot on the surface of the spinning neutron star. This provides evidence for a millisecond spin period for a neutron star which is not a radio pulsar. Credit: Tod Strohmayer.

The amount of angular momentum transfer to the neutron star is much higher in the case of LMXB, where the accretion happens via accretion disk. The duration of the accretion in these systems is 100 million years to 1000 million years and is much longer than in HMXB systems, where the corresponding duration is a few ten thousands of years. In LMXB systems, the neutron star is spun up to millisecond rotational periods. As  accretion finally stops (because of the evolution of companion star) and the neutron star is left alone it can begin to radiate as a millisecond radio pulsar. These neutron stars transferred from an X-ray millisecond pulsar to a radio millisecond pulsar are said to be recycled. To spin up a neutron star to millisecond periods requires a rather low magnetic field. A proof of the recycling theory was gained in 1998 when persistent pulsation were discovered in a transient source SAX J1808.4-3658.  The spin period can be seen from the power spectrum of the source, where one sees a peak at 401Hz corresponding to a spin period of 2,5 milliseconds. By September 2005 there are seven such objects known, with the  fastest pulsar having a period of just 1,67 ms.

Figure 20. Power Spectrum of RXTE light curve of SAX J1808.4-3658 from 1998 April 11. (Wijnands & van der Klis 1998, Nature, 394, 344)

Figure 21. The fastest X-ray millisecond pulsar with the spin period of 1,67 ms (599 Hz) was discovered in December 2004 by INTEGRAL and RXTE. From Falanga, Kuiper, Poutanen et al. (2005).

Magnetars

Magnetars are extremely magnetical neutron stars. The strength of the magnetic fields is in the order of 10¹⁴ -10¹⁵ Gauss. Actually magnetars are the most strongly magnetized objects known. These stars radiate soft gamma-rays or hard x-rays and are called soft gamma-ray repeaters (SGR) because of historical reasons. The first SGR was detected in 1979 and was distinguished from gamma-ray bursts to their own class in 1987. Differences between the gamma-ray bursters and SGR:s are that the gamma-ray bursts origin from one-shot events and produce gamma-rays instead of SGR's less energetic X-rays. But SGRs are the most energetic sources of radiation that repeat themselves.

The structure of a magnetar is similar to that of neutron stars.  In ordinary pulsars, the crust is quite stable, not like in the case of magnetar (see later in the text). The magnetic field must be 10¹⁴ Gauss or more to break the crust. The shape of the atoms is greatly affected by the strong magnetic field and they are compressed into the shape of a thin needle. The width of the atoms is 1% of their length. This kind of atoms can form carpets of polymer-like molecular chains or fibers on the surface of neutron stars.

Magnetic fields of neutron stars are generated by circulating electric currents inside the star. As the neutron star is born, the hot, ultra-dense neutron star fluid that consists of also protons and electrons (among neutrons) conducts electricity very well. The energy transport from inside outwards is done by convection of the fluid and magnetic field lines caught in the fluid initially are transported outwards by convection. The star's overall magnetic field is created by dynamo action, which is a combination of the star's rotation and convection. In order to create a magnetar kind of magnetic field, the spin  of the star must be high enough when it is born. For instance, the famous Crab nebula pulsar had a spin period of 33 milliseconds which isn't enough. So the large scale dynamo action has failed in the regular radio pulsars. High enough spin period and dynamo that works with ideal efficiency can generate a magnetic field with a strength up to 10¹⁶ Gauss. The cooling of the star happens very quickly in a time between 10 and 20 seconds and the convection and the dynamo action are terminated. However, ten seconds is enough for the creation of a 10¹⁶ Gauss field and it becomes trapped inside the neutron star by the stratified liquid of neurons and protons.

Figure 22. Convection in a neutron star.(Provided by NASA.)

The reason why magnetars aren't observed as powerful radio sources like radio pulsars is due to the fact that the rotation is spun-down very quickly after the magnetic field is created. The loss of rotational energy is efficient because of the strong magnetic field and so the rotational energy as a source of radiation becomes negligible very quickly. The magnetic field lines are anchored to the neutron star surface and as the star rotates, the field has to rotate as well. This produces magnetic waves outwards with diffuse winds of charged particles. As in the case of radio pulsars, the spin rate of neutron star must decrease. (The strength of the magnetic field can be estimated from the measured spin-down rate.)

The magnetic field itself operates as a source for observable emissions. It pushes matter inside the star and moves the crust as well, so the magnetic energy is dissipated during the first ten thousand years. The emission consists of weak steady X-ray emission and  repeating bursts.

A young magnetar (age less than few times 10 000 years) is kept hot by moving material in its interior and the star's surface is glowing brightly in X-rays. Furthermore, the electrical currents along the magnetic field lines give their contribution to the emission. The charged particles scatter the X-ray photons making the photons even more energetic. The particles also slam against the surface of the star  following the field lines and heat the surface. These heating patches glow in X-rays too.

The changing and twisting magnetic field drives dissipative currents above the star and the particles trapped in the exterior magnetic field are energized. At the same time, the magnetic field is rearranging itself to a lower energy state. This leads to bright outbursts that release tremendous amount of magnetic energy in a very short time. These bursts seem to come in bunches. If the magnetic field becomes unstable in larger scales (as they do occasionally) a much bigger shift in the crust happens and a giant flare is created.

Figure 23. Light curve of August 27, 1998 flare by Ulysses and BeppoSAX.
(Figure by: M. Feroci, K. Hurley, R. C. Duncan & C. Thompson 2001, Astrophysical Journal, 549, p.1021.)

Magnetic flares can power also fireballs. These are almost mass-free explosions where only electron-positron pairs are present which is typical to hot gasses. The little amount and lightness of particles enables the emission of gamma-rays. After a fireball has disappeared, the positron-electron gas is still trapped in the magnetic field of the star and the particles are slamming against the surface of the star. Now the X-rays and gamma-rays are emitted into this magnetic bottle (trapped gas in magnetic field) but because of scattering between the gas particles and photons, the photons cannot escape from the bottle. Furthermore, the annihilation that occurs in the gas is compensated by pair production, where the gamma-rays generate positron-electron pairs even more. The cooling of the gas, ie. escaping of gamma-rays, can only happen at the surface of the bottle. Inevitably, the size of the bottle will shrink in time and it gets dimmer. In the case of March 5th in 1998 SGR the trapped fireball evaporated in 3 minutes.

Figure 24. Evaporation of fireball, final stages of August 27 1998 flare.
(From M. Feroci, K. Hurley, R. C. Duncan & C. Thompson 2001, Astrophysical Journal, 549, p.1021.)

Gamma-ray bursts (GRBs) are the most energetic phenomenae studied in astronomy. They were discovered by the Vela satellite, which was launched by the U.S Department of Defence in the  1967. Astronomers were not informed about this new discovery until 1973. The distribution of GRBs at the sky is random as seen in figure below, and they occur at the rate of roughly one per day. GRBs show no signs of repeating themselves, i.e these are one-shot events. They can be classified according to the duration of the burst, 2 seconds being the dividing line.  Approximately one third of observed GRBs are short duration bursts, while two thirds are long bursts.

Figure 25. Uniform distribution of gamma-ray bursts detected by BATSE.
(From Stern B.E., Tikhomirova Y., Kompaneets D.,  Svensson R., Poutanen J. 2001, ApJ, 563, 80.)

Every GRB is different as seen from the light curves below.

Figure 26. Every gamma-ray burst is different.Credit: J.T. Bonnell, NASA/GSFC.

Long duration GRBs originate from very great distances, billions of light years away. Some of them were created when the universe was just a few billion years old. The positions can be determined if afterglows are detected in X-rays, as well as optical and radio-waves. Usually the source seems to be in a very distant galaxy. The most popular model for the production of long duration GRBs, is a collapse of a massive star with a supernova explosion and a following formation of a black hole. The black hole is spinning rapidly and accreting at a very high rate dense matter. The jet is then produced and aligned with the rotational axis. If the observer happens to be on the line of sight towards the jet, the GRB may be observed as the jet punches its way out through the surrounding gas. In another model, a  rapidly spinning newly born magnetar can be the central engine, instead of the accretion disk  around a black hole.

The following spectrum of GRB afterglow gives a support for this GRB to originate from massive supernova explotion. In the spectrum, the presence of chemical elements like magnesium, silicon and sulfur can be seen. These are elements produced by massive star just before supernova explotion.

Figure 27. Spectrum of GRB 011211. Credit: Reeves et al., Nature 416, 512-515, 4 April 2002.

Short duration GRB:s are distributed all over the observable universe as the long duration GRB:s. However, there exist only a few indication for the afterglows in these sources.  Some of short GRBs could probably be due to magnetar flares. Most probably short GRBs are the result of  merging of two neutron stars (or neutron star and a black hole). The Swift Gamma-Ray Burst Mission satellite is launched in mid 2005 by NASA. This satellite is equipped with X-ray and UV/optical telescopes that will make a follow-up of the possible afterglow whenever the gamma-ray burst is detected by Swift gamma-ray detector. This should bring some insight to the short duration bursts.

One of the big puzzles of GRB research for long time was very hard gamma-ray spectra that peak at a few hundred keV.  In addition to that, a very high energy component was discovered at least in one GRB:   GRB941017. The Oulu  high-energy astrophysics group has developed a model which explains most of the puzzling properties.

Figure 28. Spectral evolution predicted in the synchrotron self-Compton model of the GRB emission.
  From Stern B.E. & Poutanen J., 2004, MNRAS, 352, L35-L39.