Evgeni,

That sensor is a simple resistor that has a value which changes in the presence of some types of gases.

It cannot discriminate between two gases with the same 'reducing' capacity.

A given concentration CO will have the same effect as some other concentration of ammonia: it just measures their chemical activity of a gas - and many gases will create a change in the sensor.

May I ask what you are trying to achieve?

Isn't the sensor MOS transistor based? But yeah, I get it.

Does that means that if I measure Ammonia and there's a change in Hydrogen concentration, it will mess up my readings?

I'm trying to make a mobile air quality station (nothing new).

Evgeni,

A1: Yes. But the fabrication method is neither here nor there.

A2: Yes. Unless you can, somehow, infer the content of hydrogen.

Your problem is not new to me.

All sensors (apart from mass-spectrometers) are going to infer the presence of a compound from some bulk property. Speed of sound, thermal conductivity, specific hear, chemical activity, etc.

None of these are unique to any one material - and even spectroscopic methods are plagued by the 'overlap' between species (I work with CO2 and N2O and use a variety of methods, galvanic, spectroscopic, etc.)

I suggest that you focus on the *probable* contaminants and go from there.

VOCs, CO, and PM2.5s would be my first choices.

VOCs can be inferred from electrically heated wheatstone sensors coated with catalysts.

CO, well, that's a chemical modality and can be 'spoofed'.

PM2.5s detectors are simple backscatter sensors.

Maybe throw an O2 sensor in there too. The LuminOx range use an O2-quenched dye fluorescence method that is quite specific to oxygen. Not perfect, but in the average environment, it's pretty specific.

Thanks James, that's a really helpful answer.

So, since there's so many variables to take into account, is there even a point in measuring such gasses? I mean luftdaten has a network of simple gas sensors (mainly PM2.5 though), and if I built one but for NH3, NO2, CO is there any point since the data won't be reliable? Or is the interference of other gasses not that big and the data is still somewhat useful?

I also know humidity and sometimes temperature changes the readings of these sensors, but at least I can take those variables into account.

Evgeni,

There's certainly utility in building this - and I'd have an array of electrochemical sensors. Sure, they're consumible (and so have to be replaced every so often) but they do allow for more specific detection.

eg.

https://www.sainsmart.com/products/mq-137-ammonia-detection-sensor-nh3-gas-sensor

(but even here, it'll sense other airborne amines (!))