Background Accurate knowledge of cell-/tissue-delivered dose plays a pivotal role in inhalation toxicology studies, since it is the key parameter for hazard assessment and translation of in vitro to in vivo dose-response. Traditionally, (nano-)particle toxicological studies with in vivo and in vitro models of the lung rely onin siliocomputational or off-line analytical methods for dosimetry. In contrast to traditional in vitro testing under submerged cell culture conditions, the more physiologic air-liquid interface (ALI) conditions offer the possibility for real-time dosimetry using quartz crystal microbalances (QCMs). However, it is unclear, if QCMs are sensitive enough for nanotoxicological studies. We investigated this issue for two commercially available VITROCELL (R) Cloud ALI exposure systems. Results Quantitative fluorescence spectroscopy of fluorescein-spiked saline aerosol was used to determine detection limit, precision and accuracy of the QCMs implemented in a VITROCELL (R) Cloud 6 and Cloud 12 system for dose-controlled ALI aerosol-cell exposure experiments. Both QCMs performed linearly over the entire investigated dose range (200 to 12,000 ng/cm(2)) with an accuracy of 3.4% (Cloud 6) and 3.8% (Cloud 12). Their precision (repeatability) decreased from 2.5% for large doses (> 9500 ng/cm(2)) to values of 10% and even 25% for doses of 1000 ng/cm(2)and 200 ng/cm(2), respectively. Their lower detection limit was 170 ng/cm(2)and 169 ng/cm(2)for the Cloud 6 and Cloud 12, respectively. Dose-response measurements with (NM110) ZnO nanoparticles revealed an onset dose of 3.3 mu g/cm(2)(or 0.39 cm(2)/cm(2)) for both cell viability (WST-1) and cytotoxicity (LDH) of A549 lung epithelial cells. Conclusions The QCMs of the Cloud 6 and Cloud 12 systems show similar performance and are highly sensitive, accurate devices for (quasi-) real-time dosimetry of the cell-delivered particle dose in ALI cell exposure experiments, if operated according to manufacturer specifications. Comparison with in vitro onset doses from this and previously published ALI studies revealed that the detection limit of 170 ng/cm(2)is sufficient for determination of toxicological onset doses for all particle types with low (e.g. polystyrene) or high mass-specific toxicity (e.g. ZnO and Ag) investigated here. Hence, in principle QCMs are suitable for in vitro nanotoxciological studies, but this should be investigated for each QCM and ALI exposure system under the specific exposure conditions as described in the present study.