The conventional total internal reflection (TIR) microscopy or confocal microscopy renders a detection volume of 20-200 attoliters, limiting single molecule experiments to nanomolar concentrations of fluorescently labeled reagents. However, many biological processes require higher concentrations of proteins or substrates. Here, we present a novel combination of existing techniques with confocal microscope to reduce the detection volume to below 300 zeptoliter, which represents a three orders of magnitude reduction compared to the confocal case. The sub-diffraction focal spot of stimulated emission depletion microscopy accounts for > 20-fold reduction of volume in lateral direction and a simple convex lens provides > 40-fold confinement in axial direction. This method should allow single molecule studies of complex processes that require transient interactions between multiple components.