Electron and Force Microscopy Characterization of ...

Microsc. Microanal., page 1 of 10 doi:10.1017/S1431927614000038

© MICROSCOPY SOCIETY OF AMERICA 2014

Electron and Force Microscopy Characterization of Particle Size Effects and Surface Phenomena Associated with Individual Natural Organic Matter Fractions Lee W. Hoffman,1,* Gabriela Chilom,1 Swaminathan Venkatesan,2 and James A. Rice1 1

Department of Chemistry & Biochemistry, South Dakota State University, Box 2202, Brookings, SD 57007-0896, USA Department of Electrical Engineering and Computer Science, South Dakota State University, Box 2202, Brookings, SD 57007-0896, USA 2

Abstract: Natural organic matter (NOM) generically refers to organic substances found in soils, waters, and sediments. It is the brown-to-black, heterogeneous organic material produced through the diagenetic alteration of plant tissue and microbial biomass via a myriad of biotic and abiotic reactions. Since NOM is the primary source of organic carbon in the earth’s surficial environment, understanding the processes by which NOM is produced is integral to understanding carbon sequestration, contaminant fate and transport, and other earth surface processes. NOM samples (HA0) consist of nonamphiphilic (HA1), lipid-like (L0 and L1), and strongly amphiphilic (HA2) components. Here we present the structure and morphology of self-assembled NOM components based on scanning electron microscopy (SEM), atomic force microscopy (AFM), and electrostatic force microscopy (EFM) characterizations. Effects of surface charge and hydrophobicity/hydrophilicity of the amphiphile on the interaction and resulting structures were investigated using SEM, AFM, and EFM. Data shows that the component’s amphiphilic nature plays a key role in the formation of NOM. SEM data show that aggregates form while AFM/EFM analysis verifies the existence of hydrophobic/hydrophilic moieties in different fractions of HA0. Subsequently, the amphiphilic nature of HA2 will have a substantial effect on interfacial interactions and subsequent self-assembly of HA0’s components. Key words: natural organic matter, atomic force microscopy, scanning electron microscopy, electrostatic force microscopy, self-assembly, humic acid

I NTRODUCTION The sequestration of organic carbon in soils and sediments is being investigated as a means to mitigate global climate change because the natural organic matter (NOM) they contain is a significant carbon sink. NOM in these environments has residence times measured in hundreds to thousands of years (Six & Jastrow, 2006). It is defined as the brown to black, chemically heterogeneous and polydisperse mixture of by-products of the biotic and abiotic degradation of plant tissue and microbial biomass (Kononova, 1966; Orlov, 1985; Thurman, 1985; Stevenson, 1994). Part of the enigma of the persistence of NOM is that this material is not chemically inert but possesses an abundance of reactive functional groups. The presence of these functional groups in NOM is essential for the performance of its various natural functions in the soil; in the case of NOM, persistence and chemical reactivity are not antagonistic properties. MacCarthy and Rice (MacCarthy & Rice, 1991) proposed that not only is the heterogeneity of its chemical component the defining characteristic of NOM, but in fact it represents a necessary characteristic if NOM is to persist Received May 20, 2013; accepted January 3, 2014 *Corresponding author. [email protected]

and to perform its other ecological functions. A plausible explanation for its persistence may be that its molecular irregularity presents a confusing microscopic “buffet” to micro-organisms; such disorder could preempt NOM from serving as a fixed template for guiding evolution of future generations of organisms capable of rapidly utilizing it as an energy or nutrient source. In this perspective, the apparent lack of specific control on NOM formation to produce such a complex mixture is a thermodynamically efficient way to achieve persistence and stability against rapid mineralization because the rapid degradation of NOM would require either an extremely versatile enzyme or a very diverse enzymatic consortium to achieve rapid NOM degradation and mineralization. But if NOM is a mixture, is there an overarching order to how its components are organized? The heterogeneity of each NOM fraction limits the application of a conventional structural determination approach to identify the constituent elements and their connectivity; one only needs to review the ESI FTICR and LDI-TOF mass spectra published (Swift, 1985; Rice & MacCarthy, 1991, 1992) to visualize the challenge that the level of complexity that we are referring poses. Significant progress has been made in the structural study of many complex systems (Balazs et al., 2000; Nielsen et al., 2005; Alber et al., 2007; Brown, 2008). While many of

2

Lee W. Hoffman et al.

these systems are comprised of complex molecules or materials (e.g., proteins or carbon nanotubes), they generally consist of a relatively small number of different chemical components, in many cases only two different materials, each with a relatively well-defined chemical structure. If the defining chemical characteristic of NOM is, in fact, that it is an extremely complex mixture, then a primary structure of NOM in the classical structural sense does not exist (though certainly the individual components that comprise them have primary structures). Even the determination and modeling of secondary structures becomes problematic from this perspective. Left to explore are structures that can be assembled from molecule–molecule, aggregate–molecule, and aggregate–aggregate interactions to produce supramolecular or particulate assemblies. An expanding literature acknowledging this realization has begun to appear. Rice and MacCarthy viewed the insoluble components of NOM as an organo-mineral nanocomposite consisting of clay, lipid, and amphiphilic humic acid-like components. Because some NOM fractions are surface active in an aqueous, alkaline solution, Wershaw (Wershaw, 1993) proposed a model that describes the organization of NOM’s components as a micelle. Piccolo (2001, 2002) and MacCarthy (2001a) described NOM organization as “supra-” and “super-” molecular assemblies, respectively. In a review of conceptual models of OM, Wershaw (2004) concluded that the “preponderance of evidence” favors molecular aggregation models. Two recent reviews have also addressed NOM’s organization and the interactions that may impact it, concluding that NOM’s molecular structure is formed from a diverse assemblage of relatively low molecular weight components that self-assemble under favorable conditions which support a model of NOM as a dynamic system (Langford & Melton, 2005) consisting of an extremely heterogeneous assembly of organic materials (MacCarthy, 2001b). Chilom and Rice (2009) have unambiguously demonstrated that NOM components self-organize using differential scanning calorimetry and heat capacity measurements. The NOM self-organization process is a result of explicit interactions generating materials that self-assemble via a hierarchical process, thus yielding a more complex composite soft material at each level. The next logical step, then, is to gain a better understanding of the individual NOM components and the properties responsible for component interaction when they come into contact. In this paper, microscopy is used to describe a sample’s topography and composition as well as probe a surface’s electrical conductivity, allowing the application of these techniques to extract information about a material’s structure and interfacial properties as components assemble. In this study we describe the nature of the interfaces produced by the self-assembly of NOM’s component using scanning electron microscopy (SEM) to probe the size and structure of the individual NOM components, tapping mode atomic force microscopy (AFM) phase imaging for the detection of polar and nonpolar groups, and electrostatic force microscopy (EFM) to examine surface charge densities.

MATERIALS

AND

METHODS

Materials Two NOM source materials were used in this study, a leonardite (LEO) and a peat. The LEO is a low-grade coal from Gascoyne Mine in Bowman County, North Dakota, and was purchased from the International Humic Substances Society (http://www.ihss.gatech.edu). The peat sample, referred to as Guanella peat (GP), is a Cryohemist soil sample (Moore, 1986) collected from Guanella Pass, Clear Creek County, Colorado (Kohl, 1999). Methanol and benzene were obtained from Fisher Scientific (HPLC grade, ≥99.9%) and used without further purification. Mica disks (15 mm, V1 grade; Ted Pella Inc., Redding, CA, USA) were coated with 500 nm gold using a CrC-150 sputtering system (Plasma Science Inc., acquired by Torr Technologies, Inc., Auburn, WA). After coating the mica with gold, the substrate was annealed to produce a flat surface (Nogues & Wanunu, 2004).

NOM Extraction and Fractionation An alkali extraction method (Chilom et al., 2009) was used to isolate a fraction referred to as HA0 from each parent material. This material is traditionally referred to as humic acid in the soil chemistry literature (Stevenson, 1994). A lipid fraction, L0, was isolated via Soxhlet extraction for 72 h using a benzene–methanol azeotrope (3:1, v:v) and cellulose thimbles. The residue left after extraction is referred to as HA1. The L0 fraction was further separated into an amphiphilic (HA2) and a lipid-like component (L1) using dilute NaOH. This scheme is summarized in Figure 1 mass balance and an organic carbon balance were prepared for each fraction, and the proportion of each fraction in the

Figure 1. Extraction of natural organic matter (NOM) components. HA0 can then be further separated into nonamphiphilic (HA1) and lipid (L0) fractions, and L0 can be fractionated into a strongly amphiphilic fraction (HA2) and a second lipid (L1) fraction.

Microscopy Characterization of Natural Organic Matter Fractions

starting material was calculated. Total organic carbon (TOC) contents were determined with a Shimadzu TOC–VSCN total organic carbon analyzer fitted with the SSM-5000 Solid Sampling Module. The instrument operates by catalytically oxidizing organic matter under a flow of CO2-free air and measuring the amount of CO2 produced via infrared absorption.

Solid-State 13C NMR Analysis The chemical characteristics of each material were determined from quantitative solid-state 13C NMR spectra obtained with a Bruker Avance 300 spectrometer (The Woodlands, TX, USA). The 13C NMR spectra were recorded using direct polarization magic-angle spinning (DPMAS) at a frequency of 75 MHz and high rotation speeds (13 kHz), combined with a TC1 correction obtained from a CP/T1—TOSS experiment (Mao et al., 2000). The recycle delays used for DPMAS were between 3 and 6 s. The samples were packed in a 4-mmdiameter zirconia rotor with a Kel-F cap, and the number of scans recorded varied between 5,000 and 100,000.

Surface Analysis Each material (HA0, HA1, L0, HA2, or L1) was dissolved in a benzene:methanol azeotrope at a concentration of 0.1 mg/mL. A small aliquot of the dissolved material was deposited onto a flat, gold-coated substrate spinning at ∼100 rpm, covered, and allowed to air-dry overnight before surface analysis.

AFM and EFM An Agilent 5500 AFM operating in noncontact mode and equipped with a MAC III controller and environmental chamber was used for all AFM and EFM measurements. Microscope tips (Budget Sensors MULTI75E-G, obtained through Ted Pella, Redding, CA, USA) had a conductive coating of Pt/Ir. Two lock-in amplifiers were used, one to track topography and the other for electrostatic force imaging allowing both measurements to be done simultaneously via a single pass method. For topographic images, the first resonant frequency of the tip (66 kHz) was used. For EFM images, the second resonant frequency of the tip (395 kHz) was used. Depending on the sample analyzed, varying the tip bias between −2 and +2 V optimized the EFM signal. For each sample, the bias was chosen to provide the greatest possible change in voltage without affecting the topography signal response. AFM/EFM analyses were performed in a controlled environment chamber with a relative humidity of 65 ± 5%. All data collected was processed using Gwyddion version 2.31 (Nečas & Klapetek, 2012), an open source software downloaded from http://gwyddion.net/.

Scanning Electron Microscopy A Hitachi VP-SEM S-2400N SEM (Clarksburg, MD, USA) equipped with a tungsten filament and operating at an

3

accelerating voltage of 5 kV was used. It provided an ∼10 nm resolution limit. Using a CrC-150 sputtering system, samples were coated with 5 nm of gold before analysis to minimize charging of the nonconductive organic material. Images were collected at a magnification of between 1,000 and 50,000, with particle size calculations obtained from images at a magnification of 10,000. Particle size distributions were calculated using ImageJ (Schneider et al., 2012) on TIFF image files. Particle sizes for each fraction were calculated from a data set of at least 300 individual particles. National Institute of Standards and Technology traceable silica microspheres with a diameter of 0.99 ± 0.02 μm (Thermo Scientific, Waltham, MA, USA) were used as standards to verify size calculations made using the ImageJ software analysis. They were prepared and characterized as described above.

RESULTS Characterization of Fractions The mass balance of the fractions in each of the two materials studied is presented in Table 1. While L0 represents ~30% of the total carbon in the HA0 of both materials the component distributions differ by source material. In LEO L0 the HA2 fraction is 72% of the material TOC and the HA2:L1 ratio is 3.7 while in GP L0 HA2 represents 58% of the TOC and the HA2:L1 ratio is 1.3. Aromatic carbon (110–180 ppm) dominates the 13C NMR spectrum of the LEO HA0 while aliphatic carbon (0–50 ppm) dominates the spectrum of GP HA0. The spectra of both HA1 fractions display contributions from all four major carbon-types typically observed in NOM; aliphatic, carbohydrate (50–100 ppm), aromatic and carboxyl/carbonyl (180–210 ppm). The L1 fractions display a NMR fingerprint typically associated with lipids isolated from soils (Wu et al., 1995; Chilom & Rice, 2005; Nambu et al., 2005). These spectra display distinct resonances attributed to amorphous and crystalline polymethylene components (30 and 32 ppm, respectively). The spectrum of LEO HA2 is strongly aromatic, conversely the spectrum of the GP HA2 is strongly aliphatic mirroring the trends in their parent L0 fractions.

Scanning Electron Microscopy The particle size distribution for each material calculated from SEM data is presented in Figure 2. Earlier investigations into the morphology of the analogous soil material (humic

Table 1. Fraction Mass Balance in Each HA0 Samples (g OC/ g TOC). Sample

HA1 (%)

HA2 (%)

L1 (%)

LEO GP

71.1 ± 0.6 72.6 ± 1.0

20.7 ± 0.6 15.6 ± 1.0

8.3 ± 0.6 11.8 ± 1.0

LEO, leonardite; GP, Guanella peat.

4


Figure 2. Particle size distributions of leonardite (LEO) (left) and Guanella peat (GP) (right) for natural organic matter fractions obtained from ImageJ analysis of scanning electron microscopy images.

acid placed on a surface under acidic conditions) used a “static drop” technique (Chen & Schnitzer, 1976; Tan, 1985; Senesi et al., 1997). While these studies typically describe these materials as forming “fibers” or “bundles of fibers,” work in our lab using a “drop” method to prepare samples of HA0 and its subsequent components show that fibers are, in fact, a potential artifact of sample preparation/drying. More recent studies have appeared utilizing thin films obtained by a spinning technique to investigate NOM properties (Mertig et al., 2002; Mayes et al., 2013). These films were obtained from aqueous solutions. Due to the amphiphilic properties of HA2 (Guetzloff & Rice, 1994; Guetzloff & Rice James, 1996), we chose to create thin, sub-monolayer films by spin-coating from nonpolar organic solutions thereby minimizing the molecular attraction of a polar aqueous solution. Particle sizes for both LEO and GP HA0 components increase as the particles assemble according to the hierarchy depicted in Figure 1 (i.e. HA2 & L1 < L0 & HA1 < HA0). The LEO L0 distribution suggests a single size distribution centered at ~200 nm, while GP L0 shows multiple distributions (centered at ~120, ~240, and ~360 nm), indicating multiple particle populations are formed from a roughly 120 nm particle. The size distribution for individual components HA2 and L1 is shifted toward smaller values compared with that of L0. The L1 fractions have similar monomodal particle size distributions that are truncated at the lower limit indicating that particle sizes approach the measurement resolution limit (~10 nm). The HA2 fractions differ from each other; the HA2 LEO is primarily monomodal, while the GP sample shows multiple distributions.

AFM/EFM AFM has become a widely used technique for studying organic systems including lipids and proteins and their molecular interactions (e.g. Ikai, 1997; Nikiforov & Bonnell, 2007; Paananen, 2007; Gross, 2011; Qi, 2011; Krafft, 2012). In EFM, electrostatic forces acting on surface structures observed by AFM are probed by applying a voltage to the probe. Variations in surface components such as surface charge, surface component dielectric constants, etc. result in changes to the electrostatic forces between the tip and sample, and is usually presented as spatial variation in the surface electrical properties. A detailed explanation of EFM can be obtained elsewhere (Sarid, 1991; Dekanski, 1997; Valdre, 1999). Figure 3 contains AFM/EFM signal images from all components from the LEO sample (HA0, HA1, HA2, L0, and L1). A bias to the tip was applied during the entire analysis while the bias direction was changed during data collection, as indicated by the horizontal black lines in Figure 3. Analyses of AFM topography data for the different LEO fractions are presented in Table 2. While there is little statistical difference between these values (based on the large standard deviations), there is a slight size increase as NOM is fractionated to its more basic components (i.e. from left to right in Table 2). GP-derived materials were slightly larger to start with (HA0) and the size increase from left to right is more pronounced. We interpret the large size and standard deviation observed in GP HA2 to its stronger surfactant character than LEO HA2. By definition, self-assembly via a hierarchal pathway will lead from less-ordered starting materials to more ordered products. We interpret the AFM


5

Figure 3. Atomic force microscopy/electrostatic force microscopy (AFM/EFM) analysis for LEO components HA0, HA1, HA2, L0, and L1. The topography, phase, and EFM images are presented in the left, middle, and right columns, respectively. Each image is 1 μm2. The black lines through each set of images represents the point where the tip bias was reversed. All images were obtained by scanning from top to bottom. Taller features in the topography image show a positive potential as compared to the bulk material when a positive bias is applied, while these same features show a negative potential when a negative bias is applied.

data presented in Table 2 as support that NOM components go from a less ordered, more disperse L1 and HA2 to a more ordered, hierarchal structure in forming L0 and HA0. For HA0, AFM data shows aggregates have formed from two and three particles stacked on one another. This “stacking” to

Table 2.

Component Heights (in nm) for LEO and GP Calculated by Gwyddion Software.

Material component LEO GP

form aggregates can be detected by AFM while going unnoticed by SEM because AFM (a) is measured normal to the surface while SEM measures within the surface dimensions, and (b) has roughly an order of magnitude better resolution than the SEM.

HA0

HA1

L0

L1

HA2

3.9 ± 5.1 4.4 ± 5.1

2.7 ± 2.4 4.0 ± 3.9

5.1 ± 3.3 6.5 ± 9.0

4.4 ± 4.2 8.2 ± 7.5

5.3 ± 4.5 16.7 ± 23.7

LEO, leonardite; GP, Guanella peat. At least twenty individual particle heights were measured from three different samples of each component, and the average value and standard deviation (based on the population measured) are reported.

6


Figure 4. Atomic force microscopy/electrostatic force microscopy (AFM/EFM) analysis for GP components HA0, HA1, HA2, L0, and L1. The topography, phase, and EFM images are presented in the left, middle, and right columns, respectively. Each image is 1 μm2. The black lines through each set of images represents the point where the tip bias was reversed. All images were obtained by scanning from top to bottom. Taller features in the topography image show a positive potential as compared to the bulk material when a positive bias is applied, while these same features show a negative potential when a negative bias is applied.


No change in the AFM phase signal with change in the tip bias was observed for HA0, HA2, L0, or L1. A change in the phase response for HA1 with a change in the applied tip bias is observed. The EFM signal response suggests that differences exist in the local charge density across the surface of all LEO fractions. In addition, the observed responses for EFM of all LEO fractions are similar. The AFM/EFM images for GP HA0, HA1, HA2, L0, and L1 are presented in Figure 4. The observed trend in topography measurements from sample to sample is similar for GP samples when compared with the LEO samples. The GP fractions varied in particle sizes (Table 2). Similar to LEO, HA1 was the only component showing the effects of the applied bias on the phase signal. EFM signal response was also similar for GP HA0, HA2, and L1, as it was for LEO, where a positive change in sample topography resulted in a negative EFM response relative to background when a negative bias was applied, while the opposite occurred when a positive bias was applied. In contrast to HA0, HA2, and L1, the EFM response for L0 is opposite (i.e. positive EFM response when a negative bias is applied to the cantilever tip, and negative EFM response when a positive bias is applied). The behavior of the GP HA1 showed both positive and negative EFM response to features observed in AFM topography regardless of whether a positive or negative bias was applied.

D ISCUSSION Scanning Electron Microscopy The two HA0 samples have significantly different carbon type distributions (Fig. 5). LEO is primarily aromatic, with significantly less carbon contribution from aliphatic carbon, while the aliphatic carbon dominates for all components of GP. The various particle sizes and size distributions (Fig. 2) are a function of the fractionation (HA2 < HA1/L0 < HA2/L1) which is consistent with a hierarchal self-assembly process in which the smaller NOM components combine to form larger structures. Most LEO and GP fractions (HA0, HA1, L0, and L1) do not show much difference in size and size distribution characteristics, while HA2 for LEO and GP are different. LEO HA2 displays a monomodal size distribution while multiple size distributions are observed in the GP data, suggesting aggregates are formed when GP HA2 is placed on the gold surface. Prior work by Chilom et al. (2009) has shown that HA2 extracted from different soil types is strongly amphiphilic (Guetzloff & Rice, 1994, 1996). The increased particle size of the GP material would be a consequence of GP’s more pronounced amphiphilic character and its associated tendency to self-assemble (Chilom et al., 2009). LEO HA2 is primarily aromatic carbon while GP HA2 is a primarily aliphatic material with significant carboxyl carbon (Fig. 5). Hydrogen bonding can occur between the carboxyl groups, which will lead to increased molecular interaction between individual GP HA2 components.

7

AFM/EFM Yu et al. (2006b) have shown that by applying a potential to an AFM tip the phase response of AFM can be used to differentiate between hydrophobic and hydrophilic groups on a surface. Subsequently, it has been employed to investigate changes in local surface charge densities on a number of different materials (Leng & Williams, 1993; Takano et al., 2000; Kikunaga et al., 2011). Self-assembly involving amphiphilic materials is influenced by the hydrophobic/hydrophilic nature of the amphiphile (Israelachvili et al., 1976; Wang et al., 2009), and recent work discusses thermodynamic properties contributing to this process (Dougan et al., 2010). Surface topography data are consistent with the SEM particle size results (Table 2). The lateral dimension is measured for SEM whereas with AFM the vertical dimension is more accurately measured. Since these are soft materials, the structure of the materials will collapse when placed on a hard surface. In these experiments, the sizes measured along the z-axis (AFM) will be appreciably less than size measured in the x, y-dimension by SEM. Even so, comparison of the AFM topography measurements are consistent with those obtained by SEM; LEO and GP HA0 and HA1 are similar, HA2, L0, and L1 for GP show significantly larger particle sizes and ranges than the LEO materials. This supports our interpretation that the GP materials have a stronger tendency to form aggregates because it is a stronger amphiphile than LEO. Much work on the self-assembly of amphiphilic materials suggests that the hydrophilic/hydrophobic character of the molecular components significantly contributes to the resulting structures (e.g. Alexandridis & Lindman, 2000; Park et al., 2011; Wang et al., 2012). Because the gold substrate possesses a neutral surface, no preferential binding should occur when individual NOM fractions are placed onto it. Based on the work of Yu et al. (2006a), we anticipated some change in phase response as a result of a change in tip bias for materials containing both polar and nonpolar components. In the context of this study, we expect that phase images for L0 and L1 (nonpolar) will not respond to a change in tip bias while HA1 and HA2 (polar) will respond to changes in tip bias. For all materials except HA1, no change was observed in phase images after switching the tip bias (Figs. 3, 4). While HA1 is not amphiphilic, the presence of hydrophilic moieties is confirmed by the respective phase images. Since HA2 contains both hydrophobic and hydrophilic moieties, we expected a change in the corresponding phase signal. The lack of change in phase response for HA2 indicates no hydrophilic groups are detected on the exterior of HA2 when placed on the gold surface. We interpret these results to indicate the hydrophilic groups of HA2 are attracted to each other and subsequently have been directed inward. Therefore, the hydrophilic groups have a greater attractive force than the hydrophobic groups when placed on a chemically inert surface, indicating that interactions of these hydrophilic groups will play a significant role in the selfassembly process.

8


Figure 5. 13C DPMAS NMR spectra of leonardite (LEO) (left) and Guanella peat (GP) (right) natural organic matter components. The definitions of HA0, HA1, HA2, L0, and L1 are given in the text.

The EFM signal response suggests a difference in the local charge density on the surface of L1 when compared with HA2 and L0. In particular, we were interested in the response differences as a result of the changes in tip bias. The observed EFM response of LEO HA2 and L0 (Fig. 3) are similar, the taller features in the topography image show a positive potential as compared with the bulk material when a

positive bias is applied, while these same features show a negative potential when a negative bias is applied. The opposite behavior is observed for L1. These results (L0 responds similar to HA2 for LEO) suggest that during the self-assembly of LEO L0 from HA2 and L1, HA2 encapsulates L1. We are pursuing Langmuir Blodgett studies of this system. Both GP HA2 and L1 provide similar EFM responses


that are different from L0. It is possible the stronger amphiphilic nature of GP HA2 is a key to these differences, and we are currently exploring potential surfactant properties that could support this conclusion.

CONCLUSIONS SEM data shows an aggregated L0 parent material. An AFM analysis confirms that different surfaces (i.e. hydrophobic/ hydrophilic) are present in different fractions of HA0 and the hydrophobic/hydrophilic balance of the amphiphilic material (HA2) has a substantial effect on interfacial interactions and subsequent self-assembly. EFM data reveals different surface charge densities for different fractions of HA0 that, as with the surfactant properties of HA2, will likely influence component interactions during the assembly process. This work has provided further foundation to support the idea that NOM self-assembles, and does so via a hierarchical process. Further studies are underway to explore the factors that trigger and affect it. Because HA2 is a stronger amphiphile, we anticipate that it will play a major role in the self-assembly process of NOM.

ACKNOWLEDGEMENTS This work was supported by the National Science Foundation through grant number 1012648. SEM measurements were performed at the Materials Characterization Lab in the Department of Electrical Engineering and Computer Science at South Dakota State University.

REFERENCES ALBER, F., DOKUDOVSKAYA, S., VEENHOFF, L.M., WENZHU, Z., KIPPER, J., DEVOS, D., SUPRAPTO, A., KARNI-SCHMIDT, O., WILLIAMS, R., CHAIT, B.T., ROUT, M.P. & SALI, A. (2007). Determining the architectures of macromolecular assemblies. Nature 450(7170), 683–694. ALEXANDRIDIS, P. & LINDMAN, B. Eds. (2000). Amphiphilic Block Copolymers: Self-Assembly and Applications. Amsterdam, The Netherlands: Elsevier Science B.V. BALAZS, A.C., GINZBURG, V.V., QIU, F., PENG, G. & JASNOW, D. (2000). Multi-scale model for binary mixtures containing nanoscopic particles. J Phys Chem B 104(15), 3411–3422. BROWN, F.L.H. (2008). Elastic modeling of biomembranes and lipid bilayers. Annu Rev Phys Chem 59(1), 685–712. CHEN, Y. & SCHNITZER, M. (1976). Scanning electron microscopy of a humic acid and a fulvic acid and its metal and clay complexes. Soil Sci Soc Am J 40, 682–686. CHILOM, G., BRUNS, A.S. & RICE, J.A. (2009). Aggregation of humic acid in solution: Contributions of different fractions. Organic Geochem 40(4), 455–460. CHILOM, G. & RICE, J.A. (2005). Glass transition and crystallite melting in natural organic matter. Organic Geochem 36(10), 1339–1346. CHILOM, G. & RICE, J.A. (2009). Structural organization of humic acid in the solid state. Langmuir 25(16), 9012–9015. DEKANSKI, A.B. (1997). Material surface characterization by electronic techniques. II. Electron microscopy. Hem Ind 51, 403–410.

9

DOUGAN, L., CRAIN, J., FINNEY, J.L. & SOPER, A.K. (2010). Molecular self-assembly in a model amphiphile system. Phys Chem Chem Phys 12(35), 10221–10229. GROSS, L. (2011). Recent advances in submolecular resolution with scanning probe microscopy. Nat Chem 3, 273–278. GUETZLOFF, T.F. & RICE, J.A. (1994). Does humic acid form a micelle? Sci Total Environ 152(1), 31–35. GUETZLOFF, T.F. & RICE, J.A. (1996). Micellar Nature of Humic Colloids. In Humic and Fulvic Acids, Gaffney, J.S., Marley, N.A. & Clark, S.B. (Eds.), pp. 18–25. Washington, DC: American Chemical Society. IKAI, A. (1997). STM and AFM of bio/organic molecules and structures. Surf Sci Rep 26, 261–332. ISRAELACHVILI, J.N., MITCHELL, D.J. & NINHAM, B.W. (1976). Theory of self-assembly of hydrocarbon amphiphiles into micelles and bilayers. J Chem Soc, Faraday Trans. 2(72), 1525. KIKUNAGA, K., TOOSAKA, K., KAMOHARA, T., SAKAI, K. & NONAKA, K. (2011). A study of electrostatic charge on insulating film by electrostatic force microscopy. J Phys 301(1), 012043. KOHL, S. (1999). Role of Lipids in the Sorption and Binding of Nonpolar Anthropogenic Organic Compounds to Soil Organic Matter. (PhD Dissertation), South Dakota State University. KONONOVA, M.M. (1966). Soil Organic Matter. Oxford: Pergamon. KRAFFT, M.P. (2012). Large organized surface domains selfassembled from nonpolar amphiphiles. Acc Chem Res 45, 514–524. LANGFORD, C.H. & MELTON, J.R. (2005). When should humic substances be treated as dynamic combinatorial systems? In Humic substances, Ghabbour, E.A. & Davies, G.R. (Eds.), pp. 65–78. New York, NY: Taylor & Francis. LENG, Y. & WILLIAMS, C.C. (1993). Molecular charge mapping with the electrostatic force microscope. Proc SPIE-Int Soc Opt Eng 1855, 35–39. MACCARTHY, P. (2001a). The principles of humic substances. Soil Sci 166(11), 738–751. MACCARTHY, P. (2001b). The principles of humic substances: An introduction to the first principle. Spec Publ – R Soc Chem 273, 19–30. MACCARTHY, P. & RICE, J.A. (1991). An ecological rationale for the heterogeneous nature of humic substances. In Scientists on Gaia Schneider, S.H. & Boston, P.J. (Eds.), pp. 339–345. Cambridge, MA: MIT Press. MAO, J.D., HU, W.G., SCHMIDT-ROHR, K., DAVIES, G., GHABBOUR, E.A. & XING, B. (2000). Quantitative characterization of humic substances by solid-state carbon-13 nuclear magnetic resonance. Soil Science Society of America Journal 64, 873–884. MAYES, M., JAGADAMMA, S., AMBAYE, H., PETRIDIS, L. & LAUTER, V. (2013). Neutron reflectometry reveals the internal structure of organic compounds deposited on aluminum oxide. Geoderma 192(0), 182–188. MERTIG, M., KLEMM, D., ZÄNKER, H. & POMPE, W. (2002). Scanning force microscopy of two-dimensional structure formation in thin humic acid films. Surf Interface Anal 33(2), 113–117. MOORE, R. (1986). Soil survey of the Pike and San Isabel National Forest, Northern Part, Soil Survey No. 8657. Pueblo, Co.: US Forest Service. NAMBU, K., VAN HEES, P.A.W., ESSÉN, S.A. & LUNDSTRÖM, U.S. (2005). Assessing centrifugation technique for obtaining soil solution with respect to leaching of low molecular mass organic acids from pine roots. Geoderma 127(3–4), 263–269. NEČAS, D. & KLAPETEK, P. (2012). Gwyddion: An open-source software for SPM data analysis. Centr Eur J Phys 10(1), 181–188.

10


NIELSEN, S.O., SRINIVAS, G., LOPEZ, C.F. & KLEIN, M.L. (2005). Modeling surfactant adsorption on hydrophobic surfaces. Phys Rev Lett 94(22), 228301. NIKIFOROV, M.P. & BONNELL, D.A. (2007). Scanning Probe Microscopy in Materials Science, pp. 929–968. New York: Springer. NOGUES, C. & WANUNU, M. (2004). A rapid approach to reproducible, atomically flat gold films on mica. Surface Sci 573(3), L383–L389. ORLOV, D.S. (1985). Humus Acids of Soils. New Delhi: Amerind Publishing. PAANANEN, A. (2007). On the Interactions and Interfacial Behaviour of Biopolymers. An AFM Study, vol. 637. VTT Publ, A101–A102. PARK, C., LEE, J. & KIM, C. (2011). Functional supramolecular assemblies derived from dendritic building blocks. Chemical Communications 47(44), 12042–12056. PICCOLO, A. (2001). The supramolecular structure of humic substances. Soil Sci 166(11), 810–832. PICCOLO, A. (2002). The supramolecular structure of humic substances: A novel understanding of humus chemistry and implications in soil science. In Advances in Agronomy, Sparks, D. (Ed.), pp. 57–134. San Diego, CA: Academic Press. QI, Y. (2011). Investigation of organic films by atomic force microscopy: Structural, nanotribological and electrical properties. Surf Sci Rep 66, 379–393. RICE, J.A. & MACCARTHY, P. (1991). Composition of humin in stream sediments and peat. In Organic Substances and Sediments in Water, Vol. 1: Humics and Soils, Baker, R.A. (Ed.), pp. 35–46. Chelsea, MI: Lewis Publications. RICE, J.A. & MACCARTHY, P. (1992). Disaggregation and characterization of humin. Sci Total Environ 117–118(0), 83–88. SARID, D. (1991). Oxford Series on Optical Sciences, Vol. 2: Scanning Force Microscopy with Applications to Electric, Magnetic, and Atomic Forces. New York, NY: Oxford University Press. SCHNEIDER, C.A., RASBAND, W.S. & ELICEIRI, K.W. (2012). NIH Image to ImageJ: 25 years of image analysis. Nat Meth 9(7), 671–675. SENESI, N., RIZZI, F.R., DELLINO, P. & ACQUAFREDDA, P. (1997). Fractal humic acids in aqueous suspensions at various concentrations, ionic strengths, and pH values. Colloids Surf A 127(1–3), 57–68. SIX, J. & JASTROW, J.D. (2006). Organic matter turnover. In Encyclopedia of Soil Science Lal, R. (Ed.), pp. 936–942. Boca Raton, FL: Taylor & Francis. STEVENSON, F.J. (1994). Humus Chemistry: Genesis, Composition, Reactions. Hoboken, NJ: Wiley.

SWIFT, D.L. (1985). Fractionation of soil humic substances. In Humic Substances in Soil, Sediment, and Water: Geochemistry, Isolation, and Characterization, Aiken, G.R., McKnight, D.M., Wershaw, R.L. & MacCarthy, P. (Eds.), pp. 387–408. New York: John Wiley and Sons. TAKANO, H., WONG, S.-S., HARNISCH, J.A. & PORTER, M.D. (2000). Mapping the subsurface composition of organic films by electric force microscopy. Langmuir 16(12), 5231–5233. TAN, K.H. (1985). Scanning electron microscopy of humic matter as influenced by methods of preparation. Soil Sci Soc Am J 49, 1185–1191. THURMAN, E.M. (1985). Organic Geochemistry of Natural Waters. Dordrecht; Boston; Hingham, MA, USA: M. Nijhoff; Distributors for the US and Canada, Kluwer Academic. VALDRE, G. (1999). Correlative microscopy and probing in materials science. NATO Sci Ser, Ser E 364, 455–472. WANG, C., WANG, Z. & ZHANG, X. (2012). Amphiphilic building blocks for self-assembly: From amphiphiles to supraamphiphiles. Acc Chem Res 45(4), 608–618. WANG, Y., XU, H. & ZHANG, X. (2009). Tuning the amphiphilicity of building blocks: Controlled self-assembly and disassembly for functional supramolecular materials. Adv Mater 21(28), 2849–2864. WERSHAW, R. (1993). Model for humus in soils and sediments. Environ Sci Technol 27(5), 814–816. WERSHAW, R. (2004). Evaluation of conceptual models of natural organic matter (humus) from a consideration of the chemical and biochemical processes of humification. Scientific Investigations Report 2004–5121, pp. 49. Reston, VA: USGS. WU, Q., SCHLEUSS, U. & BLUME, H.-P. (1995). Investigation on soil lipid extraction with different organic solvents. Zeitschrift für Pflanzenernährung und Bodenkunde 158(4), 347–350. YU, A.A., STONE, P.R., NORVILLE, J.E., VAUGHN, M., PACSIAL, E.J., BRUCE, B.D., BALDO, M., RAYMO, F.M. & STELLACCI, F. (2006a). A simple atomic force microscopy method for the visualization of polar and non-polar parts in thin organic films. J Exp Nanosc 1(1), 63–73. YU, A.A., STONE, P.R., NORVILLE, J.E., VAUGHN, M., PASCAL, E.J., BRUCE, B.D., BALDO, M., RAYMO, F.M. & STELLACCI, F. (2006b). A simple atomic force microscopy method for the visualization of polar and non-polar parts in thin organic films. J Exp Nan 1(1), 63–73.