Direct photons have been proposed as a promising signature for the quark-gluon plasma (QGP) formation in relativistic heavy-ion collisions. Recently WA98 presented the first data on direct photons in P b+P b-collisions at SPS. At the same time RHIC started with its experimental program. The discovery of the QGP in these experiments relies on a comparison of data with theoretical predictions for QGP signals. In the case of direct photons new results for the production rates of thermal photons from the QGP and a hot hadron gas as well as for prompt photons from initial hard parton scatterings have been proposed recently. Based on these rates a variety of different hydrodynamic models, describing the space-time evolution of the fireball, have been adopted for calculating the direct photon spectra. The results have been compared to the WA98 data and predictions for RHIC and LHC have been made. So far the conclusions of the various models are controversial.The aim of the present review is to provide a comprehensive and up-to-date survey and status report on the experimental and theoretical aspects of direct photons in relativistic heavy-ion collisions.
References 83Now we will discuss the various attempts for calculating the production rate of energetic photons (E ≫ T ) from an equilibrated QGP.Pre-HTL rate: Before the invention of the Hard-Thermal-Loop (HTL) improved perturbation theory (see below), the QGP photon rates have been calculated using the perturbative matrix elements for the processes of Fig. 1 together with Eq.(2) [15,16,20,22]. In Ref.[20], even bremsstrahlung has been considered in this way. The derivation of the differential production rate of energetic photons (E ≫ T ), 4 The strong coupling constant at finite temperature depends on the temperature (effective, temperature-dependent running coupling constant) [29]. However, for most applications in the following we will use a mean value of α s = 0.2 -0.5, which is typical for temperatures reachable in relativistic heavy-ion collisions.7 In Ref.[49] a numerical error led to an overestimation of the rate by a factor of 4 [51].