Chemical characterization of the early evolutionary phases of high-mass star-forming regions

The formation of high-mass stars is a very complex process and up to date no comprehensive theory about it exists. This thesis studies the early stages of high-mass starforming regions and employs astrochemistry as a tool to probe their different physical conditions. We split the evolutionary sequence into four observationally motivated stages that are based on a classification proposed in the literature. The sequence is characterized by an increase of the temperatures and densities that strongly influences the chemistry in the different stages. We observed a sample of 59 high-mass star-forming regions that cover the whole sequence and statistically characterized the chemical compositions of the different stages. We determined average column densities of 18 different molecular species and found generally increasing abundances with stage. We fitted them for each stage with a 1D model, such that the result of the best fit to the previous stage was used as new input for the following. This is a unique approach and allowed us to infer physical properties like the temperature and density structure and yielded a typical chemical lifetime for the high-mass star-formation process of 105 years. The 18 analyzed molecular species also included four deuterated molecules whose chemistry is particularly sensitive to thermal history and thus is a promising tool to infer chemical ages. We found decreasing trends of the D/H ratios with evolutionary stage for 3 of the 4 molecular species and that the D/H ratio depends more on the fraction of warm and cold gas than on the total amount of gas. That indicates different chemical pathways for the different molecules and confirms the potential use of deuterated species as chemical age indicators. In addition, we mapped a low-mass star forming region in order to study the cosmic ray ionization rate, which is an important parameter in chemical models. While in chemical models it is commonly fixed, we found that it strongly varies with environment