Question: Does particle composition that the purple air monitors measure greatly skew the data output?

Cbarnes9 is asking a question about air-sensors
Follow this topic

by Cbarnes9 | November 09, 2018 00:29 | #17495


The optical characteristics of aerosols such as organic carbon and silica would likely reflect differently considering the purple air monitor uses laser spectroscopy. I work for a research team for the University of Wisconsin Eau Claire, and we are attempting to discern the validity of the data that is generated from the purple air monitors. Listed following this is a link to the AQI field test of the purple air monitor. http://www.aqmd.gov/docs/default-source/aq-spec/field-evaluations/purple-air-pa-ii---field-evaluation.pdf?sfvrsn=2



4 Comments

Hi. Given that you mention "silica" and you are at UW Eau Claire ( western wisconsin sand mining...) I take your question to really ask whether one can assume the purpleair monitor can accurately count silica particulate, or another way to ask this is whether the concentrations recorded by purpleair data would include silica particulate? I don't see any reason why it wouldn't. But purpleair is an inexpensive particulate monitor designed for general purpose monitoring. Since silica in general is found in very low concentrations except in unique situations I would doubt purpleair designers gave any thought to silica in particular. And if one wanted to use such a monitor in a very unique situation for a unique purpose you would have to do your own work to determine its appropriateness. I believe the work needed to do this would be very complex and expensive. Ask yourself how silica concentration is usually determined and how you might integrate that to answer your question.

Is this a question? Click here to post it to the Questions page.

Reply to this comment...


Following on @jeffalk answer, please expand on your question. In my opinion (and I've worked in aerosol science for a number of years) the main point of concern about ANY optical particle sensor around silica has nothing to do with the sensor itself but rather with the conditioning of the sample. The hygroscopicity of the sample is particularly relevant as if it will absorb or adsorb water it will change its size. Low cost sensors in general don't condition the sample so they don't remove any water and that's why most of the "calibration" exercises done include temperature and humidity in their corrections.

The second issue is to do with accuracy of sizing. The counting of the particles is relatively easy, the sizing of them is where more assumptions are made, particularly about their shape. All "PM" classifications are defined in terms of "aerodynamic diameters" (the diameter of a sphere that has the same settling velocity) as size descriptors while their optical size (the diameter of a sphere that scatters the same amount of light) can be very different depending on the actual shape of the particles. So, I have no doubt that the PAII will see any particle that's larger than ~300nm but whether it will classify it in the correct size bin and then estimate the correct PM concentrations is a different question.

Reply to this comment...


Hi! I had to look up "adsorb" -- (it means accumulating on the surface rather than interior) but that makes sense -- basically @guolivar, you're saying that water in the air sticks to the particles, making them seem bigger. Conditioning -- does this mean like drying the particles out, for this case? (which would seem to be pretty tough/complicated/expensive for a real-time monitor -- like, somehow integrating a heating element and still managing airflow...)

If it's helpful, @guolivar has also posted a fantastic long note on the Purple Air: https://publiclab.org/notes/guolivar/01-08-2018/thoughts-on-low-cost-air-quality-sensors which among other things starts to address some humidity questions.

I would guess the shape and composition of a particle would affect its tendency to adsorb water.

Would a combination of periodic particle sample collection /paired with/ a Purple Air be a good template for examining composition when using Purple Airs? Forgive me but perhaps there's been work like this already done that someone could link to?

Thank you! All good questions!

Is this a question? Click here to post it to the Questions page.

Reply to this comment...


Please look at the data sheet for the purple air sensors. Specifically, the r2 data for the pm1.0, pm2.5, and pm 10.

At least in the medical field, the correlation coefficients of 0.66 for pm10 is unacceptable. On the other hand, the values for the pm1.0 and pm 2.5 were in the very good and good range.

Depending on how you are using the data, this could have an affect. Myself, I'd base data on the pm1.0 and pm2.5. Whether you use the pm10 data is up to you.

Purple labs knew about the problem.

Silica is an interesting material. Often having a very large surface area per volume ratio. So many strange things can happen. Does your silica have a large surface area to volume ratio?

Is this a question? Click here to post it to the Questions page.

Reply to this comment...


Log in to comment