Berkeley Center for Green Chemistry

Home » Environmental Chemistry » How Robots Can Help Us Understand the Environmental Fate of Nanoparticles

How Robots Can Help Us Understand the Environmental Fate of Nanoparticles

“Assessment of the physico-chemical behavior of titanium dioxide nanoparticles in aquatic environments using multi-dimensional parameter testing” von der Kammer, F.; Ottofuelling, S.; Hofmann, T. Environ. Pollut. 2010, 158, 3472-3481. DOI: 10.1016/j.envpol.2010.05.007

In order to rationally design nanoparticles that are environmentally benign, we need to be able to accurately predict their environmental fate (i.e. will they travel long distances through waterways, get stuck in soils or sediments, etc?).  Though relatively robust modeling tools are available for predicting the environmental fate of organic chemicals, analogous tools for nanoparticles are in their infancy.  This is largely due to the insane variety of nanoparticle properties (e.g., composition, size, shape, surface chemistry, etc) that can be varied, resulting in an equally insane variety of nanoparticles to study.  In addition, we know very little about any of these nanoparticles.  One important property that controls the environmental fate of nanoparticles is their propensity to aggregate together and fall out of suspension, potentially limiting their environmental mobility.

One issue with studying properties like aggregation is that in order to accurately mimic natural aquatic environments in a controlled manner you need to independently vary numerous properties of water (e.g., pH, salt content), but that makes for a lot of experiments.  That is why researchers at the University of Vienna enlisted the help of  a robot that automatically samples and analyzes the nanoparticle suspensions.  They investigated the effect of differing concentrations of NaCl, CaCl2, Na2SO4, Na4P2O7, and SuWanee River Natural Organic Matter (NOM) on suspensions of TiO2 nanoparticles, also varying the pH of the solution at each salt/NOM concentration.  They measured the concentration (mg/L TiO2), zeta potential, and size of the nanoparticles remaining in suspension after a set amount of time.  This data was then visualized in a series of 2-D contour plots like the one above.  This experimental design is conveyed conceptually below:

I have to admit that while I found the collected results of this study interesting, none of the individual results were that mind-shattering.  However, to my knowledge this approach blows all other approaches out of the water in terms of the amount of useful data generated.  I won’t walk you through all of their results, but I will provide a brief overview of a few illustrative findings.

Before we dive in, there are two bits about their rather complex experimental design that bear mentioning.  There are many other important aspects to their experimental design that I won’t go through here, but I feel these two issues are most relevant.

  • Firstly, in order to determine when they should tell the robot to take its measurements, the authors briefly studied the kinetics of the nanoparticle aggregation.  They determined that under various conditions the particles underwent an initial very rapid period of aggregation that lasted anywhere from 1 to 10 hours.  This was followed by a relatively stable period of much slower aggregation, during which the robot did its sampling (at 15 hours).
  • Secondly, in order to judge the extent of aggregation, they determined the concentration of the semi-aggregated particles remaining in suspension using a method called nephelometric turbidity, which measures the amount of light scattered by the colloidally suspended particles.  To convert turbidity into mg/L TiO2 they performed a calibration curve using stabilized (“non-aggregated”) nanoparticles of identical size, as particles of different sizes scatter light differently.  However during their experiments they measured the turbidity of suspensions of nanoparticles with sizes ranging from around 200 to 1000 nm in size.  The authors state that “turbidity is a valid approximation of particle concentration under the conditions of these experiments, where the size distributions are broad and the average particle diameter is in the range of the incident light wavelength,” which was 870 nm.  Though this claim is supported by the fact that many of their findings agree with those that have been reported in the literature, one should keep this in mind when interpreting the results of the study.

Now on to the data!  Starting out with the NOM experiments.  First, here is what you need to know about NOM: at near neutral pH it is anionic (think lots of COOH and OH groups from partially decomposed organic matter).  In this study, the anionic NOM adsorbed to the TiO2 surface across all pH values studied, resulting in highly negative zeta potentials (right graph, below).  You can think of the zeta potential as correlating directly with the surface charge.  Lots of surface charge generally means the nanoparticles will repel one another and not aggregate.  This lack of aggregation (measured as high TiO2 concentration remaining in suspension) is exactly what the researchers found (left graph, below).  This finding is consistent with what has been reported in the literature.  Note that the absolute value of the zeta potential is shown simply because the sign of the surface charge matters less than its magnitude when considering electrostatic repulsion.

This same NOM-stabilization phenomenon has been observed for E. coli bacteria in aquifers, which is pretty cool.  You can think of bacteria in the same way you can think of nanoparticles!!

Onto some cooler graphs.  The results of their tests using NaCl are shown below.  In the zeta potential graph (right) you will notice the dark blobby line running from top to bottom, the center of which corresponds to the point of zero charge of the nanoparticles, or the conditions under which they have no net surface charge (pH≈4.6).  NaCl appears to have little effect on the zeta potential of these particles, indicating NaCl does not interact with specific surface sites on the nanoparticles.  In the TiO2 concentration graph (left) you can see that the particles aggregate at the point of zero charge, regardless of NaCl concentration, which is as expected due to low electrostatic repulsion between the particles at that pH.  However, at low and high pH values, the nanoparticle suspensions are relatively stable, but their propensity to aggregate increases with increasing NaCl concentration.  This is a well-documented phenomenon that is caused by to the sodium ions concentrating near the surface of the nanoparticles, which compresses the electrical double layer around the particle and increases the frequency with which collisions between particles lead to them sticking together (for more see here and here).

Finally, an example using Na2SO4 to show you just how complex this field is.  The concentration graph below (left) looks semi-familiar, with low TiO2 concentrations in the middle of the graph (due to low electrostatic repulsion around the point of zero charge) and top of the graph (due to compression of the electrical double layer at high ionic strength).  However, the zeta potential graph (right) looks a bit odd.  It is evident that the sulfate anion is interacting with the surface of the particles, because at high pH the zeta potential is relatively negative (-20 mV), which explains the stability observed in the concentration graph at high pH values.  At low pH values the sulfate interaction is again evident, as the zeta potential never reaches a value higher than +5 mV (lots of sufate anions counteracting the otherwise positively-charged surface).  At this low of a zeta potential, you would expect the particles to aggregate.  However, the concentration graph reveals they do NOT aggregate under these conditions.  The authors simply state that “especially with multivalent ions, the correlation between low zeta-potential magnitude and low dispersion stability does not always exist.”  We are left to ponder why this is.

Developing a complete predictive model for nanoparticle environmental fate by putting this data together with other data such as the attachment of nanoparticles to environmentally relevant surfaces will likely take a while, but this report certainly represents a step in the right direction.  I hope to see these researchers use this technique to examine nanoparticles of varying composition or surface chemistry, but we’ll have to wait and see!

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: