Limitations of naneos devices

Accuracy

The Partector 2 and the Partector 2 pro measure multiple different metrics; Lung-deposited surface area (LDSA), particle number concentration, average particle diameter. In addition the Partector 2 pro measures a rough particle size distribution. At naneos we typically specify a measurement uncertainty of ± 30% of the measured values (see the instrument manual for details). However, if you want to understand the measurement uncertainty a bit better, here are some further details:

The LDSA measurement is by far the most robust metric of the Partector 2. It is derived directly from the measured total charge transfer to the aerosol. Multiple Partector 2 units measuring the same aerosol will usually agree best among each other in the LDSA measurement, often to within ± 5%. In metrology terminology, the precision of the LDSA measurement is best among the different metrics the Partector 2 measures. However, even though the agreement between the instruments is excellent, the accuracy of the LDSA value is less good, as the total charge transfer is only an approximation of the true LDSA, which works well (to within the ± 30% quoted) in the size range of about 20 - 350 nm, but is less accurate for particles below 20 and above 350nm.
Furthermore, LDSA is different for different physical activities, different persons, different breathing modes (nose/mouth), and different particle morphologies and densities.
This means that the LDSA reported by the Partector 2 should always be treated as a surrogate for true LDSA. One can think of it as similar to optical mass measurements e.g. for PM2.5 which work reasonably well in most cases compared to a true gravimetric mass measurement, but may be wrong for special cases.
The Partector 2 standard uses an assumption on the particle size distribution, namely that it is lognormally distributed, and has a geometric standard deviation of 1.9. Based on this assumption and the two measured electrometer currents, it infers the average particle diameter, and from the total charge and the average particle diameter, it calculates the particle number concentration. For most aerosols this works well, however if you measure a special aerosol for which the assumption is not at all fulfilled, errors become larger. One example would be a monodisperse aerosol such as after a size classifier, another example would be a bimodal size distribution with two peaks at very different particle diameters.
The Partector 2 pro uses multiple measurements at different conditions to infer the particle size distribution. For the measurement of the average particle diameter and the particle number concentration, this will work about equally well as the Partector 2 standard for "normal" aerosols (for which the assumption of the lognormal particle size distribution is correct), but much better for extreme cases such as monodisperse and bimodal aerosols.
While the Partector 2 pro can deal better with unusual particle size distributions, it needs to make multiple sequential measurements which take time, which leads to some more limitations. The Partector 2 pro has a lower time resolution than the Partector 2 standard, and if the aerosol changes during a scan, the reported size distribution will be inaccurate, just as in other scanning mode devices. As the particle size distribution measurement relies on the measurement of small differences in signals, it is also generally noisier than for the Partector 2 standard.
In general, measurements towards the lower (10nm) or upper (300nm) end of the particle diameter range will be more uncertain than those at diameters between, due to the inherent limitations of the measurement principle (if it were possible to measure accurately below 10 or above 300nm we would not limit the size range to this interval!), but also because of sub-10nm or above-300nm particles; for more on this see the section on size range.
At low particle concentrations the electrometer signals become more affected by the ubiquitous electrometer noise and the signals become noisier and more uncertain. Since the electrometer signals depend on both particle number concentration and average particle diameter, it is impossible to give hard numbers, but you should expect performance to deteriorate at particle number concentrations below about 1000 pt/cm³. Note that again the Partector 2 pro relies on measuring small signal differences which are even noisier than for the Partector 2 standard. For a detailed discussion on how to improve measurements at low concentrations, please read more on the page on lowlevel measurements.

Environmental

The Partector 2 firmware includes compensations for changes in temperature and air pressure. However, these will only work in a limited range, so be aware that:

The calibration is done in our laboratory in Brugg at about 300 meters above sea level. Since air pressure changes with elevation, if you go to high altitudes (low absolute pressures), the Partector 2 will behave differently than during calibration. Once again, this issue is larger for the Partector 2 pro which relies on small signal differences for an accurate measurement. From comparison measurements at the high alpine site at Jungfraujoch (~3500 masl) with reference devices, we know that the Partector 2 standard still worked well at this altitude / low pressure, whereas the Partector 2 pro had to be recalibrated at the high altitude to give accurate results. For the Partector 2 pro we therefore recommend to not use it (or be aware of its limitations) above 2000m altitude; the Partector 2 standard should work to at least an altitude of 4000m, beyond that we do not know.
Temperature changes are also compensated for, but the Partector 2 will protect its internal high voltage module from high stress at high temperatures by shutting it down above 50°C, so the Partector 2 will stop measuring correctly (and report close to 0 LDSA / particles/cm³).

Size range

When there are particles present which are outside the measurement range of 10-300nm, the measurements will become distorted. Here's how:

Very small particles (below 7nm) cannot be charged in the Partector 2, and therefore are not detected at all. Particles from 7 to 10nm are all detected as 10nm particles, but the smaller they are, the more their number concentration will be underestimated. Therefore, if you have a relevant number of particles below 10nm, the measured particle number concentration will be too low. Note that such "cutoff" effects also happen for other aerosol instruments.
Large particles (above 300nm) will acquire a high charge, and will produce measurable signals in the electrometers. Under normal circumstances, there are two or three orders of magnitude fewer particles larger than 300nm than below; so even though these individual large particles carry a high charge, they hardly affect the measurement. However, in special conditions there may be more larger particles, e.g. in dusty environments, during Sahara dust events, for hazy conditions due to forest fires etc. Generally, such conditions are visible by eye, as the large particles produce visible light scattering to the observer. If this is the case, then the larger particles will affect the measurements: for the Partector 2 standard, they will appear as 300nm particles, but in a larger number as they carry more charge than 300nm particles would. For the Partector 2 pro, they will also appear as 300nm particles in a higher number, but in addition, they will "cannibalize" the measurements in the lower size channels. The figure below shows an experiment where increasingly large particles were generated in an evaporation-condensation process in our lab. The top panel shows the size distribution measured by the SMPS, the bottom panel shows the size distribution measured by the Partector 2 pro. As more particles grow above 300nm in diameter, the data inversion stops working properly, needing to remove the smaller particles to achieve a better fit to the observed currents. As a result of the large particles, the Partector 2 pro reports only 300nm particles in its size distribution with no smaller particles, although the SMPS shows that these particles are present. If you know what to look out for, you can easily spot such events, and interpret them correctly.

Charged particles

The Partector 2 is calibrated with neutralized particles. Neutralized in this context means, they are in a charge equilibrium, where the overall charge of all particles is approximately neutral, but there are individual particles with small numbers of positive or negative charges. The charge distribution depends on the particle diameter, the larger a particle is, the more likely it will be charged, and have a higher absolute charge. This charge equilibrium is the most common state of an aerosol, but there are exceptions: highly charged aerosols of both polarities can be produced in different ways, for example:

Atomization of liquids (e.g. nebulizers, ultrasonic humidifiers)
Electrosprays
Remaining particles after electrostatic filtration

Since the Partector 2 measures charge transfer to the aerosol by charging particles positively, such charged aerosols cannot be measured properly. There will be a larger charge transfer to negatively precharged aerosol than expected, and an inhibited or for very highly positively charged particles, no charge transfer at all. Therefore concentrations of positively precharged aerosols will be underestimated, concentrations of negatively precharged aerosol will be overestimated. To measure precharged aerosols accurately, a neutralizer needs to be used in front of the Partector 2. Such neutralizers come in different forms (radioactive sources of different kinds, or X-Ray sources), but they are usually expensive and a hassle to work with due to regulations. The atmosphere itself also works as a neutralizer, but it is a slow process, with neutralization time in the order of one to a few hours.