Hi Dshedoo,

Our experiments show that SCiO can detect alcohol levels in drinks with good accuracy. We conducted these tests as part of our feasibility study for future applications we will develop. We are continuing to build these data collections and models to make them available for future use.

Evaluating for other ingredients in the beverage may also be possible, although we have not tried it ourselves. It should be noted that any component in the beverage which is below a concentration of 1% will not be detectable by SCiO. In addition, as SCiO is currently designed for the scanning of solid samples, a liquid sampling accessory (available soon) will be required in order to sample.

Unfortunately your suggestion to scan blood for insights about components found in concentrations of few ppm is not likely to be feasible with SCiO. While we have not tested these thoroughly, we believe that SCiO’s NIR spectroscopy technology is most likely NOT suitable for trace detection of such components due to their low concentrations, which are below SCiO’s range of at least 0.1%.

However, scanning in concentrations above 0.1% with SCiO is generally feasible.

Ayelet,

The Consumer Physics Team

Could you tell me when the liquid sample will be available?

Tnx

The most successful detection threshold is: 0.1%.

Pffttt! AlcoSense TruTouch is just a gimmick. Why would anybody pay 300 pounds/week for it is mind-boggling. Notice that there aren’t any real specs in the website. Ethanol is one of the simplest organic molecules and the NIR spectrum is composed of just overtones and combination bands for only 2 types of bonds, C-H and O-H, which are present in many, many molecules in the human body. I can guarantee that anybody can win a lawsuit against any entity using that device against them.

I started yesterday to play with my scio to measure alcohol in wines

first I am using medical alcohol at 90% and made dilutions

using a small polycarnonate cavity sticked on alumnium reflective foil  and micropipette to get  reproducible thickness I reach for the 0-90%  R2 =0.94

which is not too bad ( deviation of 6 % in volume )

for 8% to 33% R2 = 0.89 with about 2 % deviation

each time with only 6 samples repeated 2or 3 times

however we can’t hope a precision better than 1%

now I have to do this with water + sugar ( and then water alcohol and sugar )

problem I see is the shape of the meniscus with different chemical stuff in liquids

so best solution would be liquid between 2 plates but the upper one would absorbe part of the IR but to which extend ?

I don’t hope to measure anything else than sugar and alcool in beverage : the IR range is not large enough to get acids

( malic lactic acetic )

• This reply was modified 8 years, 1 month ago by bernard.

It works very well.

We’ve now built an ethanol model with R2 = 0.997 | RMSE = 0.642

We’ve tested it extensively and had the results verified by the Public Analyst.