OBJECTIVE To compare the depth and consistency of penetration of three different root canal sealer cements into dentinal tubules in extracted teeth and to measure the penetration of an epoxy resin-based sealer cement in vivo. METHODS Root canals of 50 extracted human pre-molar teeth were prepared and obturated using three different sealer cements based on epoxy resin (AH26), zinc oxide eugenol (Pulp Canal Sealer EWT) and methacrylate resin (EndoREZ). Five teeth filled without sealer were used as controls. Teeth were sectioned and prepared for observation using scanning electron microscopy. A further 12 teeth with a history of successful root filling and subsequent extraction were collected and sectioned. The depth of sealer penetration into dentinal tubules was measured and the consistency and appearance of the sealer within the tubules observed. RESULTS AH26 demonstrated the deepest penetration (1337 microm), followed by EndoREZ (863 microm) and Pulp Canal Sealer EWT (71 microm). The difference in penetration between all sealer groups was found to be statistically significant (P < 0.05). The resin-based sealers appeared to penetrate tubules more consistently. In the clinical cases, all teeth demonstrated sealer penetration to varying depths (98-1490 microm). CONCLUSIONS The depth and consistency of dentinal tubule penetration of sealer cements appears to be influenced by the chemical and physical characteristics of the materials. Resin-based sealers displayed deeper and more consistent penetration. Penetration depths observed for the epoxy resin-based sealer in vivo were consistent with that found in the experimental model.