Introduction We report development of a prototype transdermal ethanol (EtOH) monitor, in wristwatch format, to provide a modern device for consumer, clinical and law enforcement applications. Our vision is an attractive, comfortable smart watch, which will measure transdermal alcohol concentration (TAC) and predict blood alcohol concentration (BAC) during drinking. Current commercially available alcohol monitors are bulky, ankle-fixed devices, predominantly used in court-ordered monitoring and some research studies [1]. There has been little advancement in wearability, comfort and analytical performance of these devices. Our long-term goal is to realize a new generation of wearable monitor that combines a sensitive ethanol measurement and smartwatch capabilities in a comfortable and attractive package. This device could greatly expand the possible applications employing this technology. Experimental Watch Fabrication. The sensor platform was fabricated based on a scaled down version of our dual gas sensor platform and printed amperometric gas sensor [2, 3]. The simple prototype device contains a single EtOH sensor, analog front end, 24 bit ADC, Bluetooth LE chip and T/P/RH sensor. The board is contained in a commercially available wristwatch enclosure (BodyCase B1606117) as in Fig. 1. Initial tests of linearity and S/N showed that the circuit was linear over at least 0-300 ppm EtOH at 25 C (R2 = 0.9996 – 0.9998 for 4 boards) with sensitivity 23 ± 2 nA/ppm (for 4 boards) and limit of detection (LOD) 0.26 ± 0.06 ppm.Laboratory Testing. A flow cell was designed to deliver 0 – 0.4 w/o EtOH (simulated BAC) through a Strat-M membrane, which served as a skin surrogate [4]. EtOH solutions in phosphate buffered saline pH 7.1 were used. The cell was configured such that EtOH permeated the Strat-M, with vapor entering the headspace above to which the watch was interfaced. Results and Conclusions Flow Cell Tests. EtOH steps between 0 and 0.4 %EtOH (simulated BAC) were applied with 30 min exposure at each step. A 6 hr stability test was included. High stability and S/N were demonstrated. The EtOH concentrations in simulated BAC solutions were standardized and calibrated to headspace vapor phase concentrations, giving a calibration curve for %EtOH in the PBS buffer (simulated %BAC) as a function of measured EtOH (ppm EtOH) in headspace above the membrane. For two devices, we obtained highly linear calibrations (R2 = 0.9983, 0.9992, respectively) with slope of 0.0013% and 0.0011% BAC/ppm EtOH, with near zero intercepts (-0.009 and -0.006 %BAC, respectively). From these two tests and the baseline noise we determined a limit of detection (LOD) of ±0.005% BAC. In later tests (e.g., Fig. 2) the LOD was considerably improved. The 6 hr stability test at 0.12% EtOH showed very stable output of 0.116 ±0.003 % EtOH.Human Tests. An example of real transdermal alcohol measurement using the watch prototype on a human subject is shown in Fig. 2. A male volunteer ingested six 6 oz. aliquots of 5 v/o alcoholic beverage over ca. 1 hr. Two different watches were used, one fixed to each wrist. The measured temperature at the devices was typically in the range 28 – 34 C during tests. The data in Fig. 2 are not temperature compensated. A commercial BACtrack Trace breathalyzer was used to collect contemporaneous breath alcohol data (BrAC).The TAC measurements taken on two wrists correlated quite well with each other. The TAC values tracked BrAC trend qualitatively, but with a shift to longer times, with the peak BrAC occurring ca. 1.5 hours before the peak TAC value. In other tests we have also observed lags on the order 30 min to 2 hr for TAC vs. BrAC measures. This lag is well-known for transdermal ethanol measurements [5] and presents a challenge for reliable instantaneous BAC prediction from TAC data. We note that the lag can depend on many factors including person-to-person variability of skin permeability, presence of sweat (humidity), food consumption and variable rates of alcohol consumption and metabolism, to name a few examples. Mechanical effects such as sudden movements and impacts can also affect the measurement.We demonstrated use of a printed amperometric sensor for TAC measurement. Controlled clinical trials are currently being performed in collaboration with Boston Medical Center. The ultimate goal is to reliably predict BAC from TAC data across a broad range of individuals and use cases.