Artificial ground freezing (AGF) has been applied over 150 years for mining, and was adopted for civil engineering purposes more than 30 years ago. AGF is used as a construction method to provide structural support and to exclude groundwater. Due to 9 percent volumetric expansion of pore water during freezing, soil experiences primary heave. In fine-grained soils heave due to formation of distinct ice lenses may occur. This secondary heave may contribute significantly to the total heave. After thawing, favourable effects of the freezing process may turn into unfavourable soil conditions due to reduction in strength properties and thaw settlements. The primary thesis objective is investigating one-dimensional ice lens growth and its influence on physical and mechanical properties and deformation of a silty soil. To focus on important aspects only, three secondary objectives were formulated, being establishing an overview of literature, designing test equipment and performing lab tests on frozen soil. Freeze-thaw cycles (FTCs) consist of three phases: initiation phase, preservation phase and degradation phase. In the initiation phase a temperature gradient is introduced, initiating heat extraction. The temperature within the sample starts to drop now. At the end of this phase thermal equilibrium is established, which is maintained during the preservation phase. Between the frozen and unfrozen zone lies the frozen fringe, which is partially frozen soil with reduced permeability. Water from the unfrozen zone can flow through this zone, until it freezes. Here ice crystals may form distinct layers, known as ice lenses. The lens thickness can increase as long as the frozen fringe supplies groundwater. Two theories exist regarding moment of ice lens initiation: a pressure- and a temperature-based theory. During thawing phase, pore ice melts and water is released by the thawing soil. In fine-grained soils, drainage of excess pore water is limited by the permeability, resulting in excess pore pressure. Therefore, effective stress and shear strength may be reduced after thawing. Soil tends to reach a residual void ratio that is approached after a number of FTCs. The result is densification of loose soils and loosening of dense soils. Remarkably, the particle rearrangement always causes an increase in permeability. With regard to changes in mechanical properties after thawing, different conclusions were found, sometimes even contradictory. This may be because of different researchers studied different soils and applied different test conditions. A lack in study is recognised on the difference between natural and reconstituted soils, between the same soil at different densities and between soil types. During FT tests on natural clay the inter-particle bonding was damaged and disappeared after five FTCs. As laboratory-prepared samples lack particle cementation, they are not as affected by FT damage. Triaxial test data for this project was not accurate enough for determination of undrained shear strength and permeability. By frost heave observations useful data was obtained. For the frost tests Illite and K122 was used, given their high potential for ice lens formation. Frost heave progresses non-linearly, starting rapidly, but slowing down after a while. This is probably because of temperature boundary effects of the setup. With decreasing cold-side temperature, the frost heave increases. This is partially the result of a larger frozen zone, as seen on ?CT scans. With this non-destructive method the internal rearrangement of the soil particles was monitored, revealing vertical cracking and expansion of existing cracks. Numerical test results show a linear relationship between sample height and frost penetration height and thus primary heave. The secondary heave is not height-related, as lens growth is a distinct process dependent on time. Other test results show a non-linear relationship between sample height and time to establish thermal equilibrium.