pubs.acs.org/Langmuir © 2009 American Chemical Society
Structural Organization of Humic Acid in the Solid State Gabriela Chilom* and James A. Rice South Dakota State University, Department of Chemistry and Biochemistry, Box 2202, Brookings, South Dakota 57007-0896 Received March 2, 2009. Revised Manuscript Received April 14, 2009 Studies of the structural organization of humic acid in solution have suggested that it is composed of three fractions, two humic acid-like and one lipid-like, that have distinct roles in the process of its self-organization. The present work extends the study of humic acid’s structural organization to the solid state by directly comparing its organization with a physical mixture composed of its three fractions using differential scanning calorimetry. Comparative measurements of the specific heat capacity as a function of temperature reveal differences between the original humic acid and the mixture with the same chemical composition. These differences provide direct evidence that humic acid’s solid state structure is more than just a mixture of components and is determined by specific interactions between its components. This study indicates that humic acid in the solid state has a hierarchical or “structure within a structure” architecture. The lowerlevel structure is determined by the self-assembly of amphiphilic components of humic acid with lipids into a nanostructured composite material. A higher-level structure is formed by the association of this composite material with the nonamphiphilic components of humic acid.
Introduction Humic materials have been described as refractory, heterogeneous, polyelectrolytic organic substances that color water, sediments, or soils brown or black. They can be operationally divided into three fractions based on their solubility in aqueous solutions as a function of pH. Humic acid is the fraction soluble in an alkaline aqueous solution, fulvic acid is the fraction soluble in an aqueous solution regardless of pH, and humin is the fraction insoluble at any pH value. In a review of conceptual models of humic materials, Wershaw1 concluded that the “preponderance of evidence” favors molecular aggregation models that describe humic substances as large molecular associations formed by the aggregation of smaller molecules. Formation of micelle-like humic acid aggregates in aqueous solution has been observed by a variety of methods, including concomitant surface tension and solubility measurements as well as with spectroscopic techniques.2-12 The formation of humic acid aggregates has been shown to depend on parameters such as pH, temperature, the nature and concentration of cationic species, and the nature of the humic acid itself. Guetzloff and Rice6 suggested that not all humic acid components *Corresponding author. Voice: 605-688-4782. Fax: 605-688-6364. E-mail:
[email protected]. (1) Wershaw, R. L. Evaluation of Conceptual Models of Natural Organic Matter (Humus) from a Consideration of the Chemical and Biochemical Processes of Humification; Scientific Investigations Report, USGS Series, Report 2004-5121, 2004; http://pubs.er.usgs.gov/usgspubs/sir/sir20045121 (accessed Feb 6, 2006). (2) Wershaw, R. L.; Burcar, P.; Goldberg, M. Environ. Sci. Technol. 1969, 3, 271–273. (3) Chen, Y.; Schnitzer, M. Soil Sci. 1978, 125, 7–15. (4) Rochus, W.; Sipos, S. Agrochimica 1978, 22, 446–454. (5) Hayase, K.; Tsubota, H. Geochim. Cosmochim. Acta 1983, 47, 947–952. (6) Guetzloff, T. F.; Rice, J. A. Scie. Total Environ. 1994, 152, 31–35. (7) Chien, Y.-Y.; Kim, E.-G.; Bleam, W. F. Environ. Sci. Technol. 1997, 31, 3204–3208. (8) Engebretson, R. R.; von Wandruszka, R. Org. Geochem. 1997, 26, 759–767. (9) Engebretson, R. R.; von Wandruszka, R. Environ. Sci. Technol. 1998, 32, 488–493. (10) Ferreira, J.; Nascimiento, O. R.; Martin-Neto, L. Environ. Sci. Technol. 2001, 35, 761–765. (11) Martin-Neto, L.; Traghetta, D. G.; Vaz, C. M. P.; Crestana, S.; Sposito, G. J. Environ. Qual. 2001, 30, 520–525. (12) Von Wandruszka, R. Soil Sci. 1998, 163, 921–930.
9012 DOI: 10.1021/la900750z
participate in micelle-like aggregate formation. In a recent study, Chilom and Rice13 provided experimental verification of this supposition by showing that approximately one-third of humic acid’s components behave as surfactants and hence have amphiphilic properties. Their conclusion supports the idea advanced by other authors that humic acid’s components display a continuum of amphiphilic character.14 The structural organization of humic materials in the solid state is less well-documented than that for solutions, and most of our understanding comes from indirect studies such as sorption experiments. A variation in structural arrangement has often been used to account for differences in sorption behavior when the chemical characteristics of the humic sorbent alone could not explain it. Bonin and Simpson15 reported that the sorption coefficients of phenanthrene for soil, humic acid, and humin samples did not show any consistent trend with 13C NMR data. They suggested that, in addition to the chemical characteristics, changes in physical conformation such as an increase in the accessibility of favorable sorption domains were responsible for these differences in sorption properties. Pan et al.16 studied the sorption properties of single humic acid in the solid form (H+ form) and in solution (Na+ form) in an effort to eliminate the influence of chemical characteristics on the observed behavior. They reported differences in phenanthrene sorption coefficients for the two samples, and suggested that these differences were due to “different self-organization patterns of humic molecules”. While all of these studies identify the important role that solidstate structural organization might play in the sorption characteristics of humic materials, and implicitly for contaminant fate in the environment, they do not provide direct evidence for the existence of an organized nature in these materials. The goal of the (13) Chilom, G.; Bruns, A. S.; Rice, J. A. Org. Geochem. 2009, 40, 455-460. (14) Von Wandruszka, R.; Engebretson, R. R.; Yates, L. M. In Understanding Humic Substances: Advanced Methods, Properties and Applications; Ghabbour, E. A., Davies, G., Eds.; Royal Society of Chemistry: Cambridge, U.K., 1999; pp 79-85. (15) Bonin, J. L.; Simpson, M. J. Environ. Sci. Technol. 2007, 41, 153–159. (16) Pan, B.; Xing, B.; Tao, S.; Liu, W.; Lin, W.; Xiao, D.; Dai, H.; Zhang, X.; Zhang, Y.; Yuan, H. Chemosphere 2007, 68, 1262–1269.
Published on Web 05/01/2009
Langmuir 2009, 25(16), 9012–9015
Chilom and Rice
Article Table 1. Mass Balance Showing Fraction Distributions within the Leonardite Humic Acid Sample (g OC/g OC of HA0)
replicate 1 replicate 2
Figure 1. Sample fractionation procedure.
present study is to fill this gap. In order to achieve this goal, we extended our study of the structural organization of humic acid from the solution to the solid state by taking advantage of the opportunities provided by a fractionation method that we have previously developed.13 In that study we were able to separate humic acid into fractions that have distinct roles in the aggregation process in solution. An amphiphilic humic fraction has an initiating role in the aggregation process, while the lipid fraction facilitates the formation of micelle-like aggregates. That study showed that the organization of humic acid in solution can be viewed as an emergent characteristic that is the result of the interaction of these fractions. By physically mixing these fractions in the ratio in which they were present in the original humic acid sample, we prepared mixtures that possessed the same chemical composition as the original humic acid. Since the specific heat capacity of a material is directly related to its structure, comparative measurements of the specific heat capacity as a function of temperature of the original humic acid and a mixture made of these fractions in identical proportions as the original sample should reveal differences if the original humic acid exists in an organized state. Such differences would provide direct evidence that humic acid’s solid state structure is more than just a mixture of components and is determined by specific interactions between the components that comprise it.
Experimental Section Humic Acid Samples. Humic acid was extracted from the IHSS leonardite sample material (Lot No. BS104L) using a traditional alkali extraction method as previously described.13 When used with no further treatment or fractionation, this humic acid sample is referred to as HA0. The lipid extract L0 was obtained from HA0 by Soxhlet extraction with a benzenemethanol azeotrope (3:1, v:v) for 72 h. After removing L0, the extracted humic acid sample is referred to as HA1. The L0 fraction was then separated into two fractions: one that was humic acidlike and soluble in an alkaline aqueous solution that is referred to as HA2, and a second that was lipid-like (soluble in nonpolar solvents) that is referred to as L1. The fractionation procedure is outlined in Figure 1. A mass balance was prepared for each fraction, and the abundance of each fraction in the starting material was calculated. Two different mixtures that mimic HA0 and one mixture that mimics L0 were prepared by mixing individual fractions in the ratios determined from the mass balance. The first mixture, intended to model HA0, was prepared by mixing the three Langmuir 2009, 25(16), 9012–9015
HA1
HA2
L1
recovery (% OC)
61.2% 69.6%
32.2% 25.3%
6.6% 9.5%
104% 102%
fractions HA1, HA2, and L1. It is referred to as HA0/3. The second mixture, also intended to model HA0, was prepared by mixing the fractions HA1 and L0. It is referred to as HA0/2. The third mixture, made to model L0, was prepared by mixing the fractions HA2 and L1. It is referred to as L0/2. Total Organic Carbon Analysis. The total organic carbon in the solid samples was determined with a Shimadzu TOC-VSCN total organic carbon analyzer combined 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. The 13C NMR characterizations were performed on a Bruker Avance 300 spectrometer at a 13C frequency of 75 MHz. Quantitative 13C NMR spectra were recorded using direct polarization magic-angle spinning (DPMAS) at high rotation speeds (13 kHz), combined with a T1C correction obtained from a CP/T1 - TOSS experiment.17 The recycle delays used for DPMAS were individually determined for each sample, and were between 2 and 9 s. The samples were packed in a 4-mm-diameter zirconia rotor with a Kel-F cap, and the number of scans recorded varied between 5000 and 15000. The spectra were integrated using software supplied with the spectrometer operating system over the following chemical shift regions: 0-50 ppm, alkyl carbon; 50-108 ppm, O-alkyl carbon; 108- 163 ppm, aromatic carbon; and 162-220 ppm, carboxyl/carbonyl carbon. Differential Scanning Calorimetry. Differential scanning calorimetry (DSC) experiments were performed using a Shimadzu DSC-50 differential scanning calorimeter equipped with a LTC50 low temperature-cooling unit. Temperature calibration was performed using the melting points of indium and zinc standards. Temperatures were accurate to ( 0.5 K. Heat capacity calibration was performed using synthetic sapphire (NIST SRM 720), and the uncertainty was estimated to be 0.2%. Three successive measurement runs were performed in order to obtain each heat capacity. The measuring cell was empty in the first run, filled with the reference material (synthetic sapphire) in the second run, and the sample was measured in the third run. The reference cell was empty during all runs. The measurements were carried out with a heating rate of 5 °C/min from 10 to 70 °C with isothermal periods of 5 min at the beginning and the end of each experiment. Prior to analysis, the samples were preheated (at 20 °C/min from room temperature to 100 °C and held for 30 min) to remove traces of water or organic solvents. The typical mass of samples used was 10 mg. Standard aluminum cells, unsealed, were utilized in all experiments.
Results and Discussion Chemical Characteristics. Two replicate samples of the humic acid isolated from the IHSS leonardite were used in the fractionation procedure. The distribution of fractions (Table 1) was calculated as the percent of the total organic carbon present in each individual fraction. The variations in fraction distribution between replicates were low considering the heterogeneous nature of humic materials. Moreover, high recovery rates of organic carbon were obtained for both samples, indicating that no material loss occurred during the course of the fractionation. This observation is critical in the context of this study because it (17) Mao, J. D.; Hu, W. G.; Schmidt-Rohr, K.; Davies, G.; Ghabbour, E. A.; Xing, B. Soil Sci. Soc. Am. J. 2000, 64, 873–884.
DOI: 10.1021/la900750z
9013
Article
Chilom and Rice Table 2. 13C NMR Characterizations of Humic Acid Fractions and Their Corresponding Mixturesa 13
C NMR regions (%)
sample
% OC
aliphatic
carbohydrates
aromatic
carboxyl & carbonyl
50.0 13.8 10.5 58.8 16.9 HA0 49.8 13.4 10.5 59.7 16.4 HA0/3 50.0 13.9 11.5 59.4 15.2 HA0/2 54.3 23.9 11.8 47.7 16.3 L0 53.7 21.9 12.0 49.9 16.2 L0/2 a HA0: whole humic acid; HA0/3: mixture of HA1 & HA2 & L1; HA0/2: mixture of HA1 & L0; L0: whole extract; L0/2: mixture of HA2 & L1.
13 C DPMAS spectra of humic acid’s fractions: HA1 (top), HA2 (middle), L1 (bottom).
Figure 2.
represents the first step in the preparation of the mixtures that reconstitute the original samples in the experiments below. For both humic acid replicate samples, the most abundant fraction was HA1 which accounted for almost two-thirds of the total organic carbon. When dissolved in an alkaline aqueous solution, HA1 did not exhibit surfactant properties.13 Fraction HA2, the second most abundant fraction for both samples, accounted for almost one-third of the total organic carbon. This fraction significantly decreased the surface tension of water when dissolved in an alkaline aqueous solution.13 We classified it as a strong surfactant. The 13C NMR spectra of HA1 and HA2 showed that, while they had similar carbon type distributions, there were differences (Figure 2). The HA2 fraction displayed a lower aromatic carbon content than HA1, while the alkyl content of HA2 was slightly higher than that of HA1. Fraction L1 was the least abundant fraction of humic acid and did not dissolve in alkaline aqueous solution. Its 13C NMR spectrum (Figure 2) displayed a large alkyl resonance between 29 and 33 ppm that corresponded to polymethylene chains and weaker resonances in the carboxyl and aromatic carbon portions of the spectrum. This is characteristic of soil and sediment lipids.18-20 Comparisons of the total organic carbon content for the original samples and their corresponding mixtures were made for each step in the procedure as well as for the whole fractionation process. The data in Table 2 show that HA0 had the same carbon content as the mixtures HA0/3 and HA0/2, and fraction L0 had the same carbon content as mixture L0/2. This mass balance confirms that the original humic acid and the humic acid reconstituted as a mixture had identical bulk compositions based on their respective carbon contents. However this carbon balance did not provide any insight into possible chemical differences between the original and reconstituted samples. This aspect was studied further by determining the carbon-type distributions of these samples from their solid-state 13C NMR spectra. The carbon type distributions obtained for the original samples and their corresponding :: (18) Wu, Q.; Schleuss, U.; Blume, H.-P. Z. Pflanzenernahr. Bodenkd. 1995, 158, 347–350. (19) Chilom, G.; Rice, J. A. Org. Geochem. 2005, 36, 1339–1346. (20) Lodygin, E. D.; Beznosikov, V. A. Geoderma 2005, 127, 253–262.
9014 DOI: 10.1021/la900750z
mixtures were similar within experimental error and showed that they had similar chemical compositions (Table 2). Structural Characteristics. Figure 3 shows the heat capacities of the original samples and their corresponding mixtures at temperatures ranging from 20° to 65 °C. The outstanding feature of these thermograms was that the heat capacities of the original samples were different from those of the corresponding mixtures. These differences were as large as 30%, a difference unlikely to result from experimental artifacts because (i) they were greater than the experimental errors in the heat capacity measurement (Figure 3) and (ii) each material, the original sample and its corresponding mixture, had the same thermal history. The heat capacity depends strongly on the nature of the substance and its physical state. Because both the original and the mixture samples had the same chemical composition, it is likely that the differences in the observed heat capacities resulted from the different physical states(s) of the components making up each HA0 and L0. This is direct evidence that humic acid and L0 are not just physical mixtures of their components. In addition, the differences in heat capacities between the mixtures of two and three components (HA0/2 and HA0/3) as well as the differences observed for the fraction L0 and its mixture suggest that humic acid has a hierarchical structure with two successive levels of organization or assembly. A hierarchical model of humic acid was described by Wershaw21 that viewed humic acid as a hierarchical aggregation of structural elements starting with simple phenolic, quinoid, and benzene carboxylic acids that aggregated into chemically distinct small particles, then into larger homogeneous aggregates and finally into mixed aggregates. These aggregates exist in aqueous solutions and are held together by weak interactions between chemically similar particles. While that concept of hierarchy is similar to this work, the components isolated here differ from those described by Wershaw21 because they did not all show aggregation behavior in solution and they were mixtures of many compounds. The first organizational level is represented by L0, and the second level by HA0. While each level was characterized by a significant variation of heat capacities from the values of the corresponding mixtures, the trends were in opposite directions. Fraction L0 had a higher heat capacity than its mixture (L0/2), and HA0 had a lower heat capacity than its mixture (HA0/2). Because calorimetric experiments cannot offer specific details about the nature of these differences, inferences can be made from the literature on similar materials such as inclusion compounds and coal where heat capacity data has been often used to assess structural rearrangements. It is generally recognized that a lower heat capacity describes a more rigid, less flexible structure due to increasing molecular (21) Wershaw, R. L.; Pinckney, D. J.; Booker, S. E. U.S. Geol. Surv. 1977, 5, 565–569.
Langmuir 2009, 25(16), 9012–9015
Chilom and Rice
Article
Figure 4. DSC thermogram for L0 and the mixture L0/2.
Figure 3. Variation of heat capacity with temperature for samples in this study. Top: HA0 and the mixtures HA0/3 and HA0/2 prepared to replicate its composition. Bottom: L0 and the mixture prepared to replicate its composition. 22
interactions or cross-linking. Larsen et al. showed that the heat capacities of coals change when exposed to any swelling solvent and the change depends on the coal’s rank. A decrease in heat capacity for lower rank coals suggested a rearrangement to a more highly associated structure, while an increase in heat capacity for the upper rank coals suggested they rearrange to a looser, more weakly associated structure. The authors speculated that, while in both cases the release of strain was a driving force, the rearrangement of higher rank coals was entropy driven compared to lower rank coals where greater associative interactions are present. Changes in heat capacity of inclusion compounds compared to the bulk components were observed by Zhang et al.23 They showed that the heat capacity of an inclusion compound was smaller than that of a physical mixture of the starting materials by 5% to 20%. They attributed the difference to the quantum confinement effect where small particles were subjected to greater spacial and electronic confinement in the inclusion complex, and therefore their contributions to the overall heat capacity from the various modes of motion or numbers of states are minimized. But quantum confinement is not the only phenomenon that affects heat capacity when a particle’s size decreases to the nanoscale. There are reports of higher heat capacities for nanocrystalline materials than for their extended counterparts that were explained by the larger contribution made by surface energy to the overall energy of the system.24-26 Chilom and Rice13 suggested that self-organization of HA0 in solution is determined by an intimate association between HA2, the fraction responsible for the amphiphilic character of humic acid, and L1, the lipid component. We extrapolate this observation to the solid state and assume that L0 represents an assembly (22) (23) (24) (25) (26)
Larsen, J. L.; Flowers, R. A.; Hall, P. J. Energy Fuels 1997, 11, 998–1002. Zhang, L.; Tan, Z.-C.; Wu, M. Pure Appl. Chem. 1997, 69, 2281–2288. Suriinach, S.; Malagelada, J.; Baro, M. D. Mater. Sci. Eng. 1993, 168, 161. Hellstern, E.; Fecht, H. F.; Zu, Z. J. Appl. Phys. 1989, 65, 305. Rupp, J.; Birringer, R. Phys. Rev. B 1987, 36, 7888–7890.
Langmuir 2009, 25(16), 9012–9015
Figure 5. DSC thermogram for HA0 and its fraction L0.
of clusters or nanoparticles formed by self-organization of humic acid’s amphiphilic components and lipids. As discussed above, the formation of nanoaggregates results in an increase in the contributions of the interfacial component to the overall energy of the system and therefore explains the increases in the heat capacity observed for L0 when compared to its mixture. And if our interpretation of the data is correct, then L0 should have lower thermal stability than L0/2 because nanomaterials are more subject to thermal degradation due to their larger surface area. The thermograms of L0 and L0/2 (Figure 4) clearly shows that thermal decomposition occurred at lower temperatures for L0, indicating that this is in fact the case. We speculate that L0 nanoaggregates further interact with the components of HA1 to form the HA0 composite in which the components that comprise HA1 could be attached to the surface of the L0 nanoaggregates forming a multilayer shell that in effect encapsulates L0. In this way the L0 nanoparticles are confined by a HA1 matrix, explaining the decrease in the heat capacity observed for HA0. In addition, the interaction between L0 nanoparticles and HA1 resulted in a higher melting point for lipids in HA0 compared to L0 and L1 fractions (Figure 5). The increase of 7 °C in the melting temperature also suggests an increase in the molecular interactions in HA0.
Conclusions This study has shown that humic acid in the solid state has a hierarchical or “structure within a structure” architecture. A lower-level structure is the result of the self-assembly of amphiphilic components of humic acid and lipid components into a nanostructured material. A higher level structure is formed by the association of these nanostructures and the nonamphiphilic components of humic acid. The data suggest that the structure of humic acid is determined by the interactions at and between these two levels of structure.
DOI: 10.1021/la900750z
9015
