This thesis deals with a laboratory study of heterogeneous reactions involving bromine and chlorine containing reservoir species on alkali salt substrates which are relevant to the marine boundary layer, and on ice substrates which are relevant to the lower stratosphere. The aim is to obtain reliable data on uptake coefficients γ and reaction mechanisms for the sake of extrapolation of results obtained in the laboratory to atmospheric conditions. The experiments have been performed in a Teflon coated Knudsen flow reactor equipped with a quadrupole mass spectrometer. The reactions of BrONO2 with solid alkali halides have been studied at ambient temperature. The initial uptake coefficients γ0 of BrONO2 on NaCl and KBr substrates are γ0 = 0.31 ± 0.10 and γ0 = 0.32 ± 0.09, respectively. For NaCl substrates BrCl, Br2 and HCl and for KBr both Br2 and HBr are observed as products. HCl and HBr result from the interaction of NaCl and KBr, respectively, with HNO3 generated in the hydrolysis of BrONO2 with H2O condensed on the salt sample. Using NaCl we observed Br2 which is formed from heterogeneous BrONO2 decomposition occurring on solid KBr samples. Halogen exchange reactions are competing with hydrolysis and decomposition which also take place on non-reactive salts (NaNO3, Na2SO4). A high rate of initial uptake (0.2 ≤ γ0 ≤ 0.4) and a Br2 yield on the order of 50 % are observed on non-reactive salts with a closed mass balance for Br2. The ClONO2-salt system has been reinvestigated. Experiments on NaCl, KBr (reactive salts) and on NaNO3 and Na2SO4 (non-reactive salts) have been performed. The hydrolysis of ClONO2 was not observable in contrast to heterogeneous decomposition which occurred on non-reactive salts with a Cl2 yield of (25 ± 5) %. Calculations have been performed using a 0-D box model in order to observe the effects of heterogeneous chemistry on sea salt in the marine boundary layer. Chlorine and bromine chemistry is activated by heterogeneous reactions occurring on NaCl and KBr which are the sole sources of chlorine and bromine in the model. Simulations performed with chlorine heterogeneous chemistry and gas phase chemistry indicate: 1) Denitrification of the troposphere (NOy converted to solid nitrates). 2) Increase of the oxidative capacity of the troposphere by the release of Cl atoms. When bromine chemistry is added to the model, the recombination reaction of NO2 with BrO leads to an increase of tropospheric denitrification through formation of BrONO2. Calculations show that at high [NOx], the most important bromine reservoir is BrONO2, while at low [NOx] HOBr is the most important bromine reservoir. The uptake of HNO3 on ice, solid sulfuric acid solutions and solid ternary solutions (STS) has been investigated at 180 to 211 K. The uptake coefficient γ decreases from 0.30 at 180 K to 0.06 at 211 K. The Arrhenius representation shows two distinct regimes. The first at low temperatures (180-190 K) shows a constant value γ of 0.3, whereas the second regime at T > 195 K corresponds to an activation energy of Ea = -7 ± 1 kcal/mol. At a fixed temperature, γ is independent of [HNO3] thus confirming a rate law first order in [HNO3]. The γ values of HNO3 on H2SO4/H2O solid solutions linearly decreases from 0.20 (0.10) at 10 wt % H2SO4 to 0.05 (0.03) at 98 wt % H2SO4 at 180 K (200K). The γ values of HNO3 with STS of H2SO4/HNO3/H2O is equal to 0.10 ± 0.03 in the temperature range 185-195 K at conditions where the composition of the interface was held constant at the given temperature by adding an external flow of H2O. Br2O hydrolysis on ice surfaces occurs rapidly in the chosen temperature range of 180 to 210 K. At a fixed temperature the uptake kinetics follow an apparent first order rate law in [Br2O]. The observed negative temperature dependence leads to an activation energy Ea for heterogeneous hydrolysis of -2.3 ± 0.6 kcal/mol. These facts point towards a complex reaction mechanism implying that the interaction of Br2O with ice is not an elementary reaction. BrONO2 hydrolysis has been measured on bulk (B), condensed (C) and single crystal (SC) ice in the temperature range 180-210 K. On all types of ice, HOBr and Br2O are observed as products. At a fixed temperature the rate law is first order in [BrONO2] with γ ≈ 0.3 at 180 K. The observed negative temperature dependence leads to an activation energy Ea for the hydrolysis of BrONO2 on pure ice of -2.0 ± 0.2, -2.1 ± 0.2 and -6.6 ± 0.3 kcal/mol on (C), (B) and (SC) ice, respectively, pointing towards different kinetics of BrONO2 on these different types of ice. Doping the ice samples with HBr leads to the formation of Br2, preceded by BrONO2 hydrolysis. The Arrhenius representation corresponds to an activation energy of Ea = -1.2 ± 0.2 kcal/mol. We investigated the structural properties of the near-surface region of various types of HX (X = Cl or Br) doped ice. Basically, we studied the time dependent [HX] in the interface region using the fast reaction titration XONO2 + HX → X2 + HNO3. These experiments reveal that HX is located near the surface of the substrate in a well defined region of thickness I. The HX molecules composing the interface are immediately available for the titration reaction, whatever the flow of XONO2. In the temperature range 190-200 K, we measured a value of IHCl = 43 ± 16, 250 ± 50 and 444 ± 120 nm for (SC), (C) and (B) ices, respectively and IHBr = 110 ± 15 nm on (C) ice at 205K. Finally, we assessed the bulk diffusion coefficient for HX, DHX by modeling our results according to Fick's laws of diffusion. We obtained values of DHCl = 4.5·10-15 - 1.0·10-12 cm2/s at 190 K and DHBr = (6.5 ± 3.0)·10-15 cm2/s at 205 K.