Sample tubing and particle loss is affected by the following mechanisms:
- Turbulent inertial deposition
- Inertial deposition at tube bends
- Sample tube diameter
- Charged particle loss
- Sample size
As you can see, there are a lot of mechanisms! The good news is that they have all been well studied and modeled to better help us understand how they can be a significant factor of variation between sampling or transport schemes. Read more about each mechanism below.
Diffusion is when extremely small particles (≤ 100 nm in diameter) move across bulk mass flowlines with what appears to be random Brownian-type motion. Generally, movement is from an area of high concentration to lower concentration. Tube walls become a sink for these small particles where they will accumulate, leading to low volumes of small particle concentrations near tubing walls. Prevention of tube wall deposition from diffusion involves the use mildly heated tubing.
Thermophoresis: Thermal Gradients in Tubing
Whenever there is a temperature gradient across the tubing there will be a net flux of aerosol particles from the hot areas to cold areas due to difference in air molecule momentum at different temperatures. If tubing walls are colder than the air sample, particles are driven to the walls, and if tubing walls are warmer than the air sample, particles are repelled from the tubing walls. Particle losses due to thermophoresis are negligible for most systems with a temperature differential of < 40 °C between the walls and aerosol.
Sedimentation can be significant for particles ≥ 2 – 3 μm, which are more likely to deposit on horizontal tube surfaces. Gravitational forces generally cause this particle deposition and can be minimized by having all sample tubing oriented vertically whenever possible.
Turbulent Inertial Deposition
Large particles will be more likely to deposit onto tube walls if the fluid (air) flow through the tubing is not laminar. Turbulent flows have more eddy currents causing particles of all sizes to impact tube walls. Larger particles will impact with relatively higher kinetic forces and their deposit leads to under sampling of those sizes. Once deposited on the sample tube walls, turbulent flow also promotes re-entrainment to the bulk airflow of previously deposited larger particles at a later time, possibly leading to oversampling and a misrepresentation of the particle size distribution at the sampling time.
Inertial Deposition in Sample Tube Bends
Inertial deposition will happen whenever a tube directionally bends, therefore all sample tubing bends should be eliminated or minimized both in number and in degrees of turn in each bend. Airflow streamlines change direction in the bends and the inertia of large particles may be too great to follow them, causing a higher probability of inertial deposition on the tube walls in the bend.
Sample Tube Diameter Increases and Decreases
Particle losses occur whenever there is a tube inner diameter increase or decrease. This is due to impaction and deposition. When the tubing diameter increases, the cross-sectional area increases, and eddy currents form as a result of curved airflow streamlines toward the tubing wall. This may cause large particles to deposit due to their inertia. When the tubing diameter contracts there is a change in the direction of the airflow streamlines and large particles may not follow due to their high inertia. They may deposit onto the tubing walls in front of the tubing contraction.
Charged Particle Loss
Losses of charged particles can be significant if tubing is non-conductive (e.g., PTFE). Losses of charged particles is negligible for conductive tubing impregnated with conductive or dissipative fibers (e.g., carbon fibers in Bev-a-line) or if the particle diameter is much smaller than the tubing diameter. Therefore, larger tube diameters and transport tubes designed to be electrostatically dissipative helps reduce losses leading to counting variations when particles carry a charge bias.
Sample Size Considerations
ISO 14644-1 states that a statistically significant sample must capture at least 20 particles in the size channel used for certification purposes. Often this means the larger particle size channels will not have enough particles in a 1-minute sample if the flowrate is relatively low (e.g., 0.1 CFM), and it is recommended to collect a larger sample volume. Lower sample collection times also increase the probability that a transient particle source is captured by the counter, which can skew the particle size distribution towards higher-than-normal counts. Increasing the collection sample times can help normalize these events and bring the total average counts per volume down to be more representative of the normal size distribution.
Learn more… Get the full paper here.