In what scientific units are VOC (BME688) data reported?

I have searched the community forums, the website, and the device specs, but cannot get a clear understanding of the units associated with the VOC data produced by the PA-II-FLEX. I believe this unit uses the BME688, the data sheet for which can be found on Bosch’s website here.

The BME688 data sheet indicates this unit can produce raw sensor resistance values, ppm estimates (“breath” VOC equivalents, I think?), gas scan % estimates, and IAQ estimates (0-500), all of which appear to be informed by testing performed under the ISO16000-29 standard (See page 8 on the datasheet). “Breath” VOC (bVOC), ethanol, CO, and H2S are used to assess the sensor.

I understand the VOC reporting is in beta, but the data must correspond to some scientific unit. Is the PA-II-FLEX reporting ppm or IAQ “units” (I understand AQI is technically unitless but I’m not sure how else to Fram the question, and I have seen numbers well beyond 100, so assume it is not gas scan %)? Page 17 of the BME688 data sheet suggests the data maybe in the form of a 0-500 IAQ index?

Further, do the PA-II-FLEX VOC data correspond to bVOC, ethanol, H2S, or some combination of these + CO?

Thanks, and my apologies if I have missed some obvious PurpleAir resource that describes the answer to this (I’d be very appreciative, if it exists, if someone could point me in its direction!).

EDIT: well, not sure how I missed it, but now I see “Bosch Static IAQ” is listed next to VOC on the PA map. Oiy. I gather this is taken directly from the BME688 sensor and produced according to the process laid out in the data sheet, is that right?

12 According to the guidelines issued by the German Federal Environmental Agency, exceeding 25 mg/m3 of total VOC can lead to headaches and further neurotoxic impact on health. The BSEC software auto-calibrates the low and high concentrations applied during testing to IAQ of 50 and 200, respectively.

(from the BME688 data sheet)

ppm: https://atmotube.com/atmotube-support/standards-for-indoor-air-quality-iaq (see Indoor Air Quality Levels by German Federal Environmental Agency)

ug/m3: https://www.catsensors.com/media/pdf/Sensor_Sensirion_IAM.pdf (see 2.2.1)

1 Like

Note that the map uses static-iAQ, which can report higher than 500 (IAQ, non-static, cannot breach 500). I think non-static IAQ exists because 500 is normally never breached, so the non-static IAQ continuously calibrates according to 50/250 as mentioned earlier. However, this self-calibration can be problematic. It seems like PurpleAir is trying to figure out the best approach and interpretation still.

I might be wrong about something here, or maybe someone else can add more info. I only glanced this information briefly.

Some more discussion is available here: Solved: Re: Units and Ranges for IAQ, IAQ accuracy, Static...

1 Like

TVOC units are pretty meaningless anyway, due to the fact that the measurement isn’t selective and sensitivity to different species can be pretty dramatic. Folks try to add context to these measurements by using gas ratios that are typical for an indoor environment, but the fact remains that these sensors only produce meaningful data when you already know what you’re measuring.

1 Like

That’s a very good point. Even a much more expensive PID detector has different sensitivities to each gas and can’t really give you a meaningful total voc for the same reason.

Regarding the BME688, I guess that’s why Bosch does heavily instruct in their development guide to train for specific gases individually; there is no expectation given for being able to do multiple gases.

It does feel like TVOC should be replaced with a more honest total sensitivity metric. I’d imagine TVOC is simply calculated by some predetermined coefficients and models.

I do feel like this somewhat applies to particle sensors like the PurpleAir too. The particle sizes are assumed or standardized from what I could read, or at least there’s very little evidence I’ve seen that suggests otherwise. So I don’t there’s any actual 1, 2.5, 10 differentiation happening. Even a condensation particle counter requires a separate particle sizer/filter device.

note: am just a curious layman, so I might be wrong.

For the most part, I think folks are in agreement that the plantower sensors are simple and cheap integrating nephelometers. That means what they measure is total light scattering of particles >0.3um in diameter. In principle, for an integrating nephelometer to produce accurate mass concentration values, they have to be measuring a known aerosol (the particle size distribution and optical properties of an aerosol mass can have significant impacts on light scattering). So, you’re correct. All low-cost particle sensors of this type are susceptible to large errors when there is a change in the aerosol type (e.g., wildfire smoke, dust storms, etc.). For that reason, the data need to be validated and corrected regularly, usually by collocation with some regulatory monitor (i.e., EPA FEM or FRM monitors).