We present the numerical simulations of coherent inverse Compton scattering (CICS) in a highly magnetized plasma process by means of a full three dimensional particle in cell code (PIC), which is mass and energy conservative. We used the parameters of a pulsar magnetosphere where CICS is one of the most promising models for the generation of the observed highly coherent radio emission. First we show details of the onset and time evolution of strong Langmuir turbulence driven by a relativistic electron beam penetrating a strongly magnetized background plasma. The Langmuir turbulence acts as self-generated wiggler fields which bunch the beam electrons thereby inducing strong coherent emission of the bunches at frequency gamma^2 times the plasma frequency. The emitted power is about 10 GW in a few nanoseconds. This radiation is interpreted in terms of inverse Compton scattering on nonlinear density fluctuations. CICS is the longitudinal version of a laboratory free electron laser and is applicable in strongly magnetized plasmas like pulsars.

We present the outline of a new model for the coherent radio emission of pulsars that succeeds in reproducing the energetics and brightness temperatures of the observed radio emission from the observationally deduced distances of 50-100 pulsar radii above the neutron star in a narrow region. The restrictions imposed by energy conservation, plasma dynamics of the coherent radiation process and propagation effects are used to apply the action of a plasma process like coherent inverse Compton scattering (CICS) (see Benford, 1992). In accordance with our findings (Kunzl et al. 1998a) this process requires Lorentz factors of about 10 which are lower than in most other radio emission models. This implies that no significant pair production can take place near the surface and we expect charge densities close to the Goldreich-Julian value (Goldreich & Julian (1969)). To fulfill the energetic and electrodynamic constraints the model requires constant re-acceleration in dissipation regions which can be interpreted as a voltage drop similar to that in a resistive current circuit built of a battery, connecting copper wires and a resistive load. Using the emission heights for PSR 0329+54 published by Mitra and Rankin (2002) and the spectral data from Malofeev et al. (1994) we find that a constant depth of the dissipation region of about 2 r_NS can account for the observed luminosities and spectral behavior. The extremely high peak fluxes in the substructures of single pulses can be explained by beaming effects as discussed in Kunzl et al. (1998b).

Rough estimates and upper limits on current and particle densities form the basis of most of the canonical pulsar models. Whereas the surface of the rotating neutron star is capable of supplying sufficient charges to provide a current that, given the polar cap potential, could easily fuel the observed energy loss processes, observational and theoretical constraints provide strict upper limits to the charge densities. The space charge of a current consisting solely of particles having only one sign creates a compensating potential that will make the maximum current dependent on potential and distance. In the non-relativistic case this fact is expressed in the familiar Child-Langmuir law. Its relativistic generalization and subsequent application to the inner pulsar magnetosphere provides clear limits on the strength and radial extension of charged currents originating on the polar cap. Violent Pierce-type oscillations set in, if one attempts to inject more current than the space charge limit into a given volume. These considerations apply wherever there is a significant amount of charged current flow, in particular in the gap regions. There they can be used to derive limits on the size of such gaps and their stability.

The ability of the neutron star surface to supply all or only part of the charges filling the pulsar magnetosphere is crucial for the physics prevailing within it, with direct consequences for the possible formation of pair creation regions. We evaluate the conditions for $e^-$ emission from pulsar surfaces for a simple Goldreich-Julian geometry taking both thermal and field emission processes into account. Using recently published estimates for the equation of state at the neutron star's surface, we show, that for a large range of $T_{\rm surf}, B$ and $P$, the liberated charges will fully screen the accelerating B-parallel electric field ${\rm E_\|}$. For surface temperatures $T_{\rm surf}<2\cdot 10^5 $K a balance between field emission of electrons and shielding of the field will occur. Even in the overidealised case of $T_{\rm surf}=0$ one can expect a prodigious supply of electrons which will weaken the accelerating ${\rm E_\|}$. We calculated the motion of electrons along selected polar field lines numerically for the low temperature, field emission scenario yielding their Lorentz factors as well as the produced radiation densities. Inverse Compton and later curvature losses are seen to balance the acceleration by the residual electric fields. We found that the conditions for magnetic pair production are not met anywhere along the field lines up to a height of 1500 pulsar radii. We did not {\it a priori} assume an "inner gap", and our calculations did not indicate the formation of one under realistic physical conditions without the introduction of further assumptions.

We develop a model for the infrared, optical and soft X-ray emission of the Crab pulsar in terms of anisotropic synchrotron emission by relativistic particles in an outer gap scenario with a single energy distribution. It is shown that such a distribution is naturally produced in an efficient pair cascade and that the energy of the primary particles are limited by synchrotron radiation. It is further shown that this synchrotron model is able to reproduce the spectral shape between the infrared and soft X-rays and also the corresponding luminosities. In particular, the long standing problem of the rapid spectral decline towards infrared frequencies is understandable as emission at very small pitch angles from low energy particles. Finally, we show that the scaling of our synchrotron model explains the observed correlation between the X-ray luminosity and the spin-down luminosity of the neutron star.

We calculate the beaming of coherent emission by means of geometrical optics. When spherical elementary waves are phase-coupled in that way that constructive interference occurs along one axis, the intensity is lower for off-axis viewing angles. The spatial angle of constructive interference shrinks to a fraction of $4\pi$ when the length scale of the emission region is larger than the wavelength of the radiation. This leads to an amplification of the received radiation in addition to the relativistic beaming caused by the relativistic radial outflow which is thought to be important for the radio emission of pulsars. The effect will be calculated by numerical methods and can be approximated by a simple analytic expression.

We consider the widespread assumption that coherent pulsar radio emission is based on extended pair production leading to plasma densities highly exceeding the Goldreich-Julian density. We show as an example that the observed low frequency (160 MHz) emission of the Crab pulsar is incompatible to the model of extended pair production. Our results rule out significant pair production if a plasma process is responsible for coherence and the radio emission originates from inside the light cylinder.

We present a comprehensive investigation of the radio polarization properties within the theory of natural electromagnetic wave modes in pulsar magnetospheres. Taking into account the curvature of the field lines, aberration effects and magnetic sweep-back we use the relativistic dielectric tensor in the low-density cold plasma approximation and derive the following polarization characteristics, which are in full agreement with the observational findings. Specifically, we demonstrate that 1. The degree of linear polarization decreases with increasing frequency. 2. The degree of circular polarization increases with increasing frequency. 3. At high frequencies (> a few GHz) the degree of linear polarization is correlated to the spin down luminosity Edot. 4. At high frequencies long-period pulsars exhibit weaker linear polarization than their short-period counterparts. 5. The difference between the refractive indices of the two natural wave modes decreases with increasing frequency which possibly results into a depolarization via superposition.

We consider the widespread hypothesis that coherent curvature radiation is responsible for the radio emission of pulsars. The comparison of energy conservation and the published data and luminosities explicitely proves that coherent curvature radiation cannot be the source for the radio emission of pulsars for frequencies below a few GHz. At higher frequencies coherent curvature radiation can be ruled out because neither the observationally deduced emission heights nor the observed radius to frequency mapping can be reproduced by this mechanism. Our argumentation is in accordance with the more general critics (e.g. Melrose 1992) that no adequate bunching mechanism has been identified for coherent curvature radiation. We present 5 examples (0329+29, 0355+54, 0540+23, 1133+16, 1916+10) of pulsars whose high frequency (larger than 1.4. GHz, up to 32 GHz) luminosities are well known, and as a low frequency example the faintest radio pulsar 0655+64 of the Taylor et al (1993) sample.