Selectively excavated carious lesions remain radiographically detectable. Radiopaque tagging could resolve the resulting diagnostic uncertainty. We aimed to evaluate if tagging depends on lesions depths, is antibacterial, or affects dentin bond-strengths. Artificial lesions (depth-range: 152-682 μm, n = 34/group) were induced in human dentin samples, evaluated using wavelength-independent microradiography, treated with one of two tagging materials (70% SnCl2, 30% SnF2) and re-evaluated. To evaluate antimicrobial effects, 40 dentin samples were submitted to a Lactobacillus rhamnosus invasion-model. Infected samples were treated with placebo, 0.2% chlorhexidine, SnCl2, SnF2 (n = 10/group). Dentin was sampled and colony-forming units/mg determined. Micro-tensile bond-strengths of adhesive restorations (OptiBond FL, Filtek Z250) to tagged or untagged, sound and carious dentin were assessed (n = 12/group). Tagged surfaces were evaluated microscopically and via energy-dispersive X-ray-spectroscopy (EDS). Tagging effects of both materials decreased with increasing lesion depths (p < 0.001). Un-/chlorhexidine-treated dentin contained significantly more viable bacteria (median 7.3/3.7 × 10(5) CFU/mg) than tagged dentin (no CFU detectable, p < 0.001). Tagging decreased bond strengths (p < 0.001) on sound (-22%/-33% for SnCl2/SnF2) and carious dentin (-50%/-54%). This might be due to widespread tin chloride or fluoride precipitation, as detected via microscopy and EDS. While radiopaque tagging seems beneficial, an optimized application protocol needs to be developed prior clinical use.