Limitations of Nanoparticle Tracking Analysis (NTA)
Nanoparticle tracking analysis relies on following a particle over time and determining its size from its motion (see: difference between detection and tracking of nanoparticles and Sources of error in measuring Brownian Motion ).
The first limit to the analysis is that the accuracy is related to the number of frames during which the same particle can be localized, and the localization accuracy of each particle.
Assuming one has no control on the diffusion properties of the particles, the number of frames can be increased by increasing the frame-rate. On the other hand, if one has control on the geometry of the diffusion (see: nanoparticle tracking in hollow optical fibers ), one can increase the time particles stay in the field of view. The localization accuracy of each measurement is related to: the amount of recorded photons and the exposure time. Localization microscopy has already shown the dependence of accuracy and photon budget. But, long exposure times will have smeared out images of particles which are free to diffuse in that time span.
note : In water, it takes few microseconds for a 20nm particle to diffuse over a distance equal to its diameter.
Therefore, high frame rates with low exposure times increase the total number of frames acquired per particle and lower the uncertainty in the localization due to the movement itself. On the other hand, lower photon budgets increase the uncertainty in localization. Increasing the illumination power can yield direct benefits, unless the sample is too polydisperse. In such case, the dynamic range of the camera will play have a limiting effect on the accuracy of the measurement.
Another obvious limit to the method is that it can only work with particles that are visible. Even though it seems trivial, it is not. The fact that a particle is visible or not depends on multiple factors. For instance, a particle surrounded by a medium of the same refractive index will be invisible. However, with fluorescent labeling of particles even in the case just described, we would be able to see the particles.
Therefore, we must separate the limitations of the technique depending on whether one is working in Scattering mode and Fluorescent mode .
Limitations of Working in Scattering
Scattering mode requires a minimal sample preparation. Optically speaking is also the easiest to construct, since there are no filters involved and almost any light source would work fine. The only requirement is that particles scatter enough light and that the background is sufficiently low.
The limits of NTA in scattering mode are related therefore to the mismatch between the refractive index of the particle and the medium, the size of nanoparticles , and the background signal. Dim particles (i.e. small particles or those with a refractive index similar to the medium) can still be detected by improving on the camera specifications or the illumination strategy. This is the approach of Dispertech and the hollow optical fiber .
If, however, there is a high background signal given by, for example, rests of proteins or anything else that is not a particle of interest, little can be done from an optical perspective. In this case, proper sample preparation is of utmost importance. Size exclusion chromatography , or High Performance Liquid Chromatography are two common techniques to purify samples.
Limitations of Working in Fluorescence
One of the biggest technical limitations of NTA is that it needs to acquire enough data points per particle in order to calculate its size. If particles go too fast out of the field of view or if the dye used photobleaches too quickly, then the number of detected locations will be too small to calculate the diffusion coefficient ( @fo2020 Tracking and Analyzing the Brownian Motion of Nano-objects Inside Hollow Core Fibers ). The trivial way of overcoming bleaching is by using [quantum dots]
The biggest disadvantage is that QDs are relatively big (20nm) and thus alter the size of the particles under study. The way most NTA software works is by assuming spherical particles and therefore the result would be an effective hydrodynamic radius, which will differ from the sample's radius by a non-trivial amount.
Another problem is related to the fluorescent labeling mechanism itself, but is not exclusive to the NTA method. If labeling is not specific or it generates by products such as dye molecules not bound to any target, it will increase the background signal, or it won't help distinguishing the object of study from random elements in the medium ( @libregts2018 Flow cytometric analysis of extracellular vesicle subsets in plasma: impact of swarm by particles of non-interest ).
This is one of the strengths of flow cytometer that do not need to overcome photobleaching since they look at just one point and for a short period of time. The non-specificity is, however, a challenge, that can be seen as the Effects of swarm of particles in flow cytometers . Again, the hollow optical fiber approach of Dispertech may be a good solution to overcome both limitations.
I am not sure about what Particle Metrix software looks like, but the software provided by NanoSight is extremely limiting. Papers relate to thresholds and camera setpoints which are irreproducible outside of the devices themselves. The papers I've seen don't report exposure times, nor specify which cameras they are using (this last point may be understandable). Then, the step to actually analyze the data adds another layer of custom defined parameters that make it virtually impossible to translate to other software.
I understand that many researchers don't want to deal with specific values, and they just want to click a button saying "high sensitivity" or "low sensitivity", but it really looks like they are doing a disservice to science by hiding away the true values. Without even counting the nanosight proprietary software update that included the Walker algorithm and some "calibration".
References: @faez2015 Fast, Label-Free Tracking of Single Viruses and Weakly Scattering Nanoparticles in a Nanofluidic Optical Fiber @thane2019 Improved methods for fluorescent labeling and detection of single extracellular vesicles using nanoparticle tracking analysis
These are the other notes that link to this one.