Five replicate columns were prepared and analyzed for each treatment. Metal concentrations for all three ENMs are reported as ionic, although neither CeO2 nor TiO2 were expected to dissolve to a significant degree under the conditions used in this experiment. TiO2 is known to be highly insoluble in water and CeO2 is similarly insoluble at pHs similar to those found in the soils used here. However, Cu2 has been shown to undergo partial dissolution under acidic to neutral conditions, although at acidic pHs less dissolution occurs in media with high concentrations of dissolved organic matter.Based on this, dissolution of Cu2 is not expected to occur to a significant degree under the conditions and time scales used in this experiment. To measure size distribution of particles throughout the column, air-dried samples of contaminated soils were collected from the top and bottom 3 cm of columns and analyzed using environmental scanning electron microscopy with back scattering electron detection and energy-dispersive X-ray spectroscopy to confirm identification of CeO2, Cu2, or TiO2 ENMs. Beam voltage was set at 12 kV, spot size at 6.0, water vapor pressure was kept at 2.7 Torr,square plastic pot and working distance averaged around 10.5 cm. These settings were chosen in order to minimize X-ray subsurface penetration for EDS analysis. Elemental hypermap data was collected over a period of 6 min per image.
ImageJ image analysis software was used to determine particle or aggregate size. Soil solution extracts of potting, grassland, and farm soils were prepared following Rhoades, although no Na3PO4 was added in order to avoid influencing ENM physicochemical behavior. Soil solution extracts were stored at 4°C until use.Gravity-driven vertical transport of ENMs through unsaturated soil was found in general to follow the hypothesis that the majority of ENMs would be retained in the upper portion of the column, but as predicted was found to be highly dependent on soil type with increased retention occurring in the denser, less porous natural soils . However, ENMs coated with natural organic matter did not have increased vertical transport, and in fact were retained more in potting soil. TiO2 and CeO2 aggregate sizes were seen to decrease with column depth, suggesting physical straining to be the primary impediment to transport. Aggregate hydrodynamic diameters tended to be larger in soil solution extracts than Nanopure H2O and were also generally larger with NOM-coated particles, with several exceptions .All three ENMs largely passed through the entire length of potting soil columns, being present in lower concentrations than the hypothetical homogeneous concentrations at all points, although there was some retention in the upper 0-6 cm that was increased with NOM-coated particles . These trends can likely be explained by the primarily organic composition of the potting soil, which gave it very low density, high porosity, and high reactivity . The low density and high porosity prevented aggregates from being physically strained, which is shown for TiO2 and CeO2 by the similar aggregate sizes in the tops and bottoms of columns for both uncoated and NOM-coated particles .
If physical straining was strongly influencing particle transport in potting soil it is unlikely similar aggregate sizes would be observed throughout the length of the column, but rather would result in smaller aggregates or particles penetrating through the column while larger aggregates would be retained at the surface – as was seen in the two natural soils. All three ENMs had similar hydrodynamic diameters in potting soil solution , although NOM-coated aggregates were significantly smaller than uncoated aggregates . ζ-potentials for all three ENMs in soil solutions from all three soils were also similar , although again the presence of NOM coatings, as well as the ENM and soil types, had significant impacts on ζ-potentials . Coating particles with NOM appears to increase their affinity for the organic components of the potting soil. This resulted in the increased overall retention of NOM-coated CeO2 as well as the decreased vertical transport of NOM-coated TiO2 and Cu2 . Evidence for this can be found in Figures 2.3A-B, which show that both NOM-coated and uncoated aggregates have nearly identical hydrodynamic diameters in potting soil solution extract, so the additional retention of NOM-coated aggregates is unlikely to be due to increased physical straining. This was visually confirmed in micrographs of NOM-coated TiO2 in potting soil including Figure 2.4D, which revealed the formation of TiO2 encrustations occurring primarily on the organic components of potting soil over the Al/Si/Na/K perlite minerals. These encrustations may have been caused in part or whole to interactions between the NOM coating and the organic matter in the potting soil.
This finding is counter to several previous transport studies using TiO2, CeO2, and ZnO25 in quartz sand that found organic coatings decreased ENM retention by increasing electrostatic repulsion between coated aggregates and the sand grains, which further suggests interactions between the organic coating and organic soil components. In the grassland and agricultural soils CeO2 and Cu2 shared similar transport profiles , forming large aggregates in the soil solutions that were retained almost entirely in the upper 0-3 cm of the soil columns. However, the widely variable background concentrations of Ti in these natural soils prevented precise measurement of TiO2 ENM distribution throughout the soil columns by ICP-AES , the majority of TiO2 aggregates were confirmed to be retained immediately at the surface through both visual identification of white buildup on the column surfaces and through BSE/EDS analysis. As shown in Figure 2.4, both uncoated and NOM-coated TiO2 ENMs formed large encrustations on the surfaces of all three soils with the exception of uncoated TiO2 in potting soil. Despite having nearly identical surface charges in soil solution extracts , CeO2 formed large porous sponge-like aggregates instead of the more solid encrustations seen with TiO2. These differences in aggregate morphology may be due to differences between the primary particle shapes of these two ENMs, with TiO2 being nanospheres and CeO2 being nanorods. Afrooz, et al. found that spherical Au ENMs had higher attachment efficiencies and deposition rates than rod-like Au ENMs identical in composition, which they attributed to differences in electrosteric and physical packing characteristics. Similarly, Zhou, et al. found the critical coagulation concentration of TiO2 nanospheres was directly related to particle diameter while the CCC of TiO2 nanrods was better explained by particle surface area, which they postulated was a consequence of differences in exposed crystal faces. It has alsobeen shown that metal oxide nanospheres and nanorods interact differently with NOM,square plant pot which may also be a factor in explaining the differences in aggregate morphology seen here.Little research has been done on the effects of ENM exposure on soil properties. In one of the only available studies available on this subject, Ben-Moshe, et al. 1 observed that CuO and Fe3O4 ENMs did not change the total organic content or macroscopic properties of two types of soil but altered the humic substances in the soils. The authors also observed an effect on the soil microbial community, which has been reported in other studies , but did not attempt to link changes in important soil properties with these effects. VandeVoort, et al. found that silver ENMs could limit denitrification processes in soil, but that the effects were dependent on ENM concentration and coating. While previous studies in this area suggest that the effects of ENMs on soil properties are somewhat limited, there may be additional impacts not considered in these studies. For example, metal oxide surfaces are amphoteric, capable of producing both protons and hydroxide ions , but tend to be predominantly acidic in nature.
Due to this metal oxide ENMs may be able to alter the pH of soil pore water and consequently the overall pH of the soil. pH has been called one of the “master variables” for soil systems7 because it controls a number of critical physical and chemical properties, and if ENMs are able to alter soil pH when present above certain concentrations they may pose a hazard to organisms that rely on the soil for habitat or sustenance. However, soils are typically well-buffered, and may be able to withstand ENM accumulation without changing pH. Additionally, ENMs will likely aggregate as a result of the high ion content of soil solutions, thus decreasing total surface area and, potentially, proton/hydroxide production. Additionally, metal oxide ENMs bear many similarities to naturally occurring nano-scale poorly crystalline metal oxide minerals known as short-range order minerals. SRO minerals have been shown to influence nutrient availability in natural soils via sorptive processes, and metal oxide ENMs may also demonstrate this effect. In particular, metal oxides are well known for their ability to covalently adsorb phosphate ions and, depending on the strength of this interaction, may prevent organisms from accessing this important nutrient. Two hypotheses were addressed in these series of experiments. First, I hypothesized that none of the soils would experience a significant change in pH after spiking with ENMs due to the presence of buffering compounds in the soils. Second, I hypothesized that these ENMs would sorb soil nutrients, including phosphate, and reduce their mobility in the soil.The effect of ENM contamination on soil pH were tested over a range of ENM concentrations by adding potting, grass, or farm soil with 0, 0.1, 1, 10, and 100 µg g-1 TiO2, CeO2, or Cu2 ENMs with and without the addition of 10% NOM. Soil aliquots were then air dried and mixed with Nanopure water to make a 20% soil paste from which the pH was measured. All treatments were performed in triplicate. Changes in soil ion release due to the presence of ENMs was tested by mixing aliquots of potting, grass, or farm soil with 100 µg g-1 TiO2, CeO2, or Cu2 ENMs as suspensions, centrifuging at 8000 x g for 10 min, and analyzing the supernatant for ion concentrations. NO3 – was measured via colorimetric methods and Al, Ca, Fe, K, Mg, Na, P, and S were measured via ICP-AES after acidification to 10% HNO3. The influence of the three ENMs on the bio-availability and mobility of P was investigated further by contaminating agricultural, grassland, or potting soil samples with 100 µg g-1 ENMs and testing P content in three fractions: water extractable P, bio-available P, and immobile P. Soil aliquots were first mixed with water for 1 min, centrifuged at 8000 x g for 10 min, then the supernatant was removed and acidified to 10% HNO3. The same soil aliquots were then mixed with Bray extract11 for 1 min, centrifuged at 8000 x g for 10 min, then the supernatant was removed and acidified to 10% HNO3. The soil aliquots were then acid digested in 1:3 HNO3:HCl at 200°C for 1.5 hours in a microwave digestion system, and all samples were analyzed for P content via ICP-AES. Soil solution extracts of potting, grassland, and farm soils were prepared following Rhoades, although no Na3PO4 was added in order to avoid influencing ENM physicochemical behavior. Soil solution extracts were stored at 4°C until use. Despite varying ENM concentrations over four orders of magnitude, changes in soil pH due to ENM contamination were largely independent of both ENM type and concentration . Contrary to the first hypothesis, changes in soil pH due to ENM contamination did occur, but they were found to be highly dependent on soil type. All three ENMs increased grass soil pH , decreased farm soil pH , and had no effect on potting soil pH . Additionally, the presence of NOM had no effect on the influence of ENMs on soil pH except in the case of farm soil, where a slight buffering effect was seen . As nearly all changes in soil pH were independent of ENM concentration it is unlikely these ENMs directly influenced soil pH through the production of H+ /OH-due to their amphoteric properties. One possible alternate explanation is that the ENMs increased the release of ions that act as buffering or pH-altering agents, such as Al3+ , Ca2+, H+ , K+ , Mg2+, Na+ , and OH- , by replacing them on the mineral surfaces of the soil matrix. Since there is a limited pool of ions available for desorption in a unit of soil, changes in ion release due to ENM sorption would be relatively independent of ENM concentration beyond the point at which total sorption/desorption occurs.