In UV-curable printing technology, an ink or a varnish is exposed to the UV-light source within a fraction of a second. Subsequent polymerization therefore proceeds in the dark, referred to as dark curing. Because dark curing can be critical and contributes to a large part of the overall conversion and to the final properties of the printed ink, the aim of this work was to get a better understanding of the dark curing process. Because an ink is a complex mixture of components, simpler systems representative of free radical and cationic chemistry were studied. Special attention was given to free radical polymerization with a detailed kinetic modeling. For cationic systems, the study was more application oriented. Attenuated-Total-Reflection Fourier-Transform Infrared (ATR-FTIR) spectroscopy was found to be the method of choice to follow in real-time the progress of the reaction under UV irradiation and in the dark. This technique allowed the study of curing conditions close to those encountered in industry. Dark curing of free radical was examined in systems consisting of a monomer (multifunctional acrylate) and a photoinitiator. The reactivities of different monomers were investigated and it was found that the chemistry of the spacer between the acrylate functional groups had an influence on the extent of dark curing. For example, TPGDA (tripropylene glycol diacrylate) led to higher overall conversion than HDDA (hexanediol diacrylate) although both are diacrylate monomers. Several explanations were put forward, such as the effect of dissolved oxygen, primary cyclization and initiation efficiency. The efficiency of photoinitiators (PIs) was examined and different behaviours were highlighted. For example, it was observed that DMPA (dimethoxyphenyl acetophenone) has a better quantum yield for α-cleavage compared to HCPK (1-hydroxycyclohexylphenyl ketone), whereas HCPK is more efficient in initiating the reaction, resulting in a better overall conversion in the dark. It is suggested that the two primary radicals resulting from the cleavage of the PI behave differently for DMPA and HCPK. One radical effectively initiates the reaction and the second is mainly involved in the termination process. These two different behaviours were observed under specific experimental conditions: (i) by covering the sample with a quartz plate and taking into account dissolved molecular oxygen and (ii) by curing under inert conditions with nitrogen. Photo-crosslinking polymerization reactions are strongly influenced by diffusion effects due to the formation of an infinite network. Therefore, it was proposed that a redistribution of reactivity within the system occurs, such as the diffusion of unreacted photoinitiators towards concentrations of unreacted monomers. This diffusion process has two positive effects: (i) for the same UV dose, but different intensities, a higher extent of conversion was reached with longer irradiation times under the condition investigated; (ii) multiple exposures led to a higher overall conversion than for a single exposure, for a fixed overall UV dose. Other parameters were explored, such as temperature and diffusing oxygen from air. With temperature, different trends were observed depending on the photoinitiator and the start of dark curing, either starting before the maximum rate of polymerization (Rp,max), or after Rp,max. The development of a kinetic model to describe dark curing of free radical polymerization was proposed. Existing models based on physical parameters, such as those incorporating diffusion-controlled phenomena and free-volume are relatively complex and require extensive information on the system studied. Therefore, a semi empirical approach was chosen in order to study the variation of fitting parameters when changing curing parameters, such as the photoinitiator concentration and type, the irradiation time, the UV light intensity and the temperature. The model was derived from kinetic data and it was assumed that rate coefficients are time-dependent and that termination, which was considered to be a first-order process, is propagation dependent. The kinetic equation has the form of a stretched exponential function and incorporates four fitting parameters. The model was evaluated for a large number of experimental data and very good fits were obtained with a random distribution of residuals. The exponent β was found to be sensitive to the termination process and reflected different reactivities in the dark as in the case of DMPA and HCPK. Dark curing of cationic systems continues until complete monomer conversion but at a slower rate. Thus systems with high reaction rates, both during UV irradiation and during the first seconds of the dark period, are of most importance for UV-curable printing technology. The most widely used diepoxide monomer in cationic ink formulation is 3,4-epoxycyclohexylmethyl-3,4-epoxycyclohexane carboxylate (DECC). In order to improve its reactivity during the first seconds of dark curing, the influence of several parameters was examined, such as the addition of an oxetane co-monomer (TMPO), the influence of relative humidity, UV dose, light intensity, temperature, and film thickness. The enhanced reactivity of DECC with TMPO was confirmed. During the cationic curing process, two reaction mechanisms compete; an activated chain end (ACE) mechanism and an activated monomer (AM) mechanism. It was observed that these two processes depend on the presence of hydroxyl containing compounds such as TMPO monomer, polyol compounds, and, most important, water. When the concentration of these components is increased the AM mechanism is enhanced, leading to higher conversion but at a slower curing rate, and a degradation of the mechanical properties. The presence of water is unavoidable and its concentration in the ink varies in presence of hydrophilic compounds, such as polyol. Furthermore, the water content increases with increasing relative humidity in air. It was shown that the reactivity of the system decreased as the percent of relative humidity increased. Moreover, although a higher overall conversion is achieved, the crosslink density of the film was severely reduced. This lower crosslink density was the result of the predominance of chain transfer reactions, leading to more pendant and short chains which were hydroxyl terminated. Moreover, the extent of conversion was thickness dependent since less water diffuses into the deeper layers occurred in thicker films.