In his answer Dario Miric has explained the basis of azeotropes clearly, this is a small addition to his answer. Raoult determined experimentally that the partial vapour pressure over some liquid mixtures is directly proportional to the mole fraction $x$ of each component, and thus the total pressure is
$$P=x_1p_1^\text{o}+x_2p_2^\text{o}=x_1p_1^\text{o}+(1-x_1)p_2^\text{o}$$
where $p^\text{o}$ is the vapour pressure of the pure liquid. An ideal solution can be defined as one that obeys Raoult's law at all temperatures. Solutions obeying Raoult's law are, for example, dichloroethane - benzene or ethylene bromide - propylene bromide. As the pressure is the sum of the partial pressures the two species cannot interact to any great extent that would cause their vapour pressures to differ from that expected from the pure solution. Of course this is only true in a few cases.
If the two types of molecule interact with one another then the potential energy of the mixture has an additional term describing this interaction. The result of this interaction is that the total pressure becomes
$$P=x_1p_1^\text{o}e^{\alpha}+(1-x_1)p_2^\text{o}e^{\beta}$$
where $\alpha =(1-x_1)^2)\Delta U/RT$ and $\beta =x_1^2\Delta U/RT$ and $\Delta U$ is a measure of the average difference in interaction energy. When $\Delta U$ is small vs $RT$ only a slight deviation from Raoult's law (pressure $p_{1,2}/p_{1,2}^\text{o}$,vs. $x$) is evident and instead of being a straight line is slightly bent, convex of concave, for example methanol in water. If the plot of total pressure vs mole fraction passes through a maximum or minimum then this is the Azeotrope and the vapour has the same composition as the mixture. Additionally as $dP/dx=0$ at the azeotrope and as the mole fraction is known an estimate of $\Delta U$ can be obtained. Example of azeotropes are dichloromethane - acetone, carbon disulphide - dimethoxymethane and CS$_2$-CHCl$_3$