| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
a Department of Civil and Environmental Engineering, Vanderbilt University, Nashville, TN 37325
b State Agricultural Laboratory, Arizona Department of Agriculture, 2422 West Holly, Phoenix, AZ 85009
c Environmental Sciences Division, Oak Ridge National Laboratory, P.O. Box 2008, Oak Ridge, TN 37831-6036
* Corresponding author (eugene.j.leboeuf{at}vanderbilt.edu)
Received for publication June 23, 2004.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONREFERENCES |
|---|
), constant-pressure specific heat capacities (Cp), and thermal transition temperatures (Tt) of several aquatic- and terrestrial-derived NOM. For the first time, glass transition behavior is reported for a series of NOM fractions derived from the same whole aquatic or terrestrial source, including humic acid–, fulvic acid–, and carbohydrate-based NOM, and a terrestrial humin. Thermal mechanical analysis (TMA), standard differential scanning calorimetry (DSC), and temperature-modulated differential scanning calorimetry (TMDSC) measurements revealed Tt ranging from –87°C for a terrestrial carbohydrate fraction to 62°C for the humin fraction. The NOM generally followed a trend of increasing Tt from carbohydrate to fulvic acid to humic acid to humin, and greater Tt associated with terrestrial fractions relative to aquatic fractions, similar to that expected for macromolecules possessing greater rigidity and larger molecular weight. Many of the NOM samples also possessed evidence of multiple transitions, similar to ß and transitions of synthetic macromolecules. The presence of multiple transitions in fractionated NOM, however, is not necessarily reflected in whole NOM, suggesting other potential influences in the thermal behavior of the whole NOM relative to fractionated NOM. Temperature-scanning X-ray diffraction studies of each NOM fraction confirmed the amorphous character of each sample through Tt.
Abbreviations: DSC, differential scanning calorimetry • GCH, Georgetown carbohydrate • GFA, Georgetown fulvic acid • GHA, Georgetown humic acid • HCH, Harpeth carbohydrate • HFA, Harpeth fulvic acid • HHA, Harpeth humic acid • HHU, Harpeth humin • NMR, nuclear magnetic resonance • NOM, natural organic matter • PMMA, poly(methyl methacrylate) • TMA, thermal mechanical analysis • TMDSC, temperature-modulated differential scanning calorimetry • XRD, X-ray diffraction
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONREFERENCES |
|---|
Insights into the compositional, structural, and conformational aspects of NOM have increased considerably as a result of advancements in spectroscopic techniques (Chien and Bleam, 1998; Chen et al., 2002), as well as finding insights from polymer science through thermal gravimetric analysis and calorimetry (e.g., Turi, 1997), thermal and dynamic mechanical analysis (Hatakeyama and Quinn, 1994), dilatometry (Hatakeyama and Quinn, 1994), and dielectric analysis (Campbell et al., 2000) techniques. While spectroscopic techniques provide knowledge of NOM structural and functional group compositions, thermal analysis measurements, while macroscopic in nature, are capable of quantifying NOM thermodynamic properties, including glass transition temperature (Tg), crystalline melt temperature (Tm), thermal expansion coefficient (), and constant-pressure specific heat capacity (Cp). Thermal analysis techniques are especially useful in elucidating the influences of elemental and chemical functional group composition, molecular weight, and cross-linking on NOM mechanical behavior, which is essential to evaluating the influence of NOM macromolecular mobility on sorption and desorption behavior of organic compounds (LeBoeuf and Weber, 1997, 2000b, 2001).
Ongoing work on well-characterized synthetic macromolecules has been especially useful in revealing the effects of structure and composition on Tg, Tm, , and Cp (Qu et al., 2001; Zhao et al., 2002). Of note are observations of multiple thermal transitions (Buijs and Damman, 1994; Bershtein et al., 2002) or a broadening of the Tg (Alves et al., 2002; Bhat et al., 2002) within some synthetic macromolecules. Bershtein et al. (2002) attributed multiple thermal transition behaviors to either (i) the heterogeneous nature of the macromolecule in which "soft" polymeric regions are confined in a rigid geometry, or (ii) covalent anchoring of segments or strong interaction with a rigid structural constituent that increases the Tg. Increased energy input in the form of heat resulted in a series of sequential segmental motions represented first by mobility of side-chain functional groups (noncooperative ß relaxation), followed by cooperative, main-chain motions ( relaxation or glass transition) at higher temperatures. For more heterogeneous samples, the distinction between ß and relaxations can be obscured by a "smearing" of the overall thermal signal, resulting in observations of lower-temperature broadenings of the Tg.
Because synthetic macromolecules are primarily synthesized for strength, durability, or thermal stability, advanced thermal characterization is often used to determine the effects of cross-link density (Montserrat, 1995; Jenkins et al., 2000), branching (Hatakeyama and Quinn, 1994; Mohamed and Al-Dossary, 2003), and crystalline content (Alves et al., 2002; Lam et al., 2002) on the intended functions of the macromolecule. An increase in the glass transition temperature and the decrease in the thermal expansion coefficient (Montserrat, 1995) or decrease in the change in specific heat capacity (Canadas et al., 1998) are evident in macromolecules possessing greater cross-link density or crystalline content. Branching of new segments to closely packed, rod-like structures that possess symmetry and much interchain hydrogen bonding also results in greater thermal stability of the macromolecule.
Proteins and biopolymers display similar thermal behaviors to synthetic macromolecules. For example, multiple transitions exhibiting ß and relaxations have been observed in glucose (Gangasharan and Murthy, 1993) and cotton cellulose (Flaque et al., 2000). Moreover, increasing the density of cross-links in tissue cells (Ramos-Sanchez et al., 1999) or adding new crystallites in wood flour–starch cellulose compounds (Liu et al., 2001) have led to decreases in macromolecular mobility as reflected by higher Tg values. Likewise, a broadened and increased Tg, due to hydrophobic and ionic interactions, resulted from the addition of salt to flaxseed proteins (Li-Chan and Ma, 2002).
Recent studies (LeBoeuf and Weber, 1997, 2000a, 2001; Young and LeBoeuf, 2000; DeLapp et al., 2004) involving application of advanced thermal characterization techniques, including differential scanning calorimetry (DSC), temperature-modulated differential scanning calorimetry (TMDSC), and thermal mechanical analysis (TMA), to NOM have focused on homogenized fractions of the source materials (LeBoeuf and Weber, 1997), including fulvic acid and humic acid materials. Glass transition behavior was observed via DSC and TMDSC measurements of several humic acids, including Aldrich humic acid, with a Tg between 57 and 59°C (LeBoeuf and Weber, 1997); Leonardite humic acid, 72 to 76°C (LeBoeuf and Weber, 2001); and a peat humic acid, 43 to 62°C (Young and LeBoeuf, 2000). Young and LeBoeuf (2000) also employed TMA to evaluate the thermal behavior of a peat humic acid and an aquatic fulvic acid, where the measured glass transitions of TMA compared well with DSC and TMDSC measurements. A more comprehensive study of humic and fulvic acids obtained from the International Humic Substances Society (IHSS) using TMA, DSC, and TMDSC (DeLapp et al., 2004) reported glass transition behavior for all samples, and observations of general trends of higher Tg values for humic acids compared with fulvic acids. In addition, several multiple thermal transitions were observed.
The Tg of a German whole soil obtained from a spruce forest was measured through DSC. The measured Tg values were 79°C for air-dried samples and 77°C for pre-moistened samples (Schaumann and Antelmann, 2000). Additional transition behaviors were observed for IHSS bulk soils as well as a mature peat (DeLapp and LeBoeuf, 2004). The reported Tg values ranged between 48 and 71°C, with higher glass transition temperatures exhibited in the more diagenetically altered Leonardite sediment.
Although glass transition behavior has been observed in a number of NOMs, additional questions remain, including:
This work seeks to partially answer the first question by employing variable temperature X-ray diffraction (XRD) to evaluate the relative amorphous contents of each NOM fraction above and below the suspected Tg. The latter questions are addressed in part through examination of additional NOM fractions derived from two bulk sources: a terrestrial soil from Franklin, Tennessee, USA, and a coastal sediment derived from a wetland in Georgetown, South Carolina, USA. These fractions consist of carbohydrate, fulvic acid, humic acid, and, in the case of the terrestrial material, humin materials. Evaluation of their behavior through TMA, DSC, and TMDSC analyses provided quantification of Tg, , and Cp.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONREFERENCES |
|---|
The HHU–HHA–HFA fraction was then neutralized with approximately 4 mL of 1 M NaOH to a pH of 7 under a nitrogen atmosphere, with 146 mL of 0.1 M NaOH added to produce a final liquid to sample ratio of 10 to 1. The HHU solids were separated from the HHA–HFA supernatant following centrifuging at 1700 x g for 5 min. The supernatant was then reacidified with 12 mL of 6 M HCl to a pH of 1, and centrifuged to obtain the HHA precipitate and HFA supernatant. The HFA and HFA–HCH samples were purified further by passing the solution through a glass column packed with a synthetic polymer resin (XAD-8, Lot no. 63117; Rohm & Haas, Philadelphia, PA). The resulting eluent contained the HCH fraction, while the HFA fraction was adsorbed in the columns. The HFA was eluted from the column through use of a two-step backwash; (i) 125 mL of 0.1 M NaOH, followed by (ii) 250 mL of distilled water at a rate of 5 mL min–1. The HFA and HCH fractions were further purified in a cation-exchange resin (AG MP 50, Lot no. 143-0841; BioRad Laboratories, Hercules, CA) column at a rate of 5 mL min–1. This resin, consisting of sulfonic acid functional groups attached to a styrene divinylbenzene copolymer lattice, preferentially adsorbs metal cations and hydrophilic bases, while low molecular weight organic compounds, represented by carbohydrates, proteins, amino acids, and uronic acids, pass through the column.
The HHU and HHA fractions were demineralized through use of repetitive HCl and hydrofluoric acid (HF) washing (Certified ACS Grade; Fisher Scientific). Approximately 200 mL of each fraction was subjected to repeated changes of 500 mL of 6 M HCl until no additional evolution of CO2 gas bubbles was observed. Repeated changes of 500 mL of a 0.1 M HCl–0.3 M HF solution were then used until the ash content was under 3 and 0.5% for the HHU and HHA fractions, respectively. Following demineralization, 250-mL samples of the HHA were also passed through dialysis tubing (10000 Daltons, Lot no. 132118; Spectrum Laboratories, Rancho Dominguez, CA) until a negative chloride test was reached. All fractions (humin, humic acid, fulvic acid, and carbohydrate) were freeze-dried and then stored in amber glass bottles and placed in desiccators.
The aquatic NOM fractions were obtained from a source NOM derived from a wetland pond in Georgetown, SC. Sampling and fractionation procedures, conducted in a separate laboratory, for this aquatic source material have been discussed in previous works (Chen et al., 2002; Gu et al., 1994, 1995). Briefly, this aquatic NOM was collected and concentrated using a reverse-osmosis system as described by Serkiz and Perdue (1990). After removing inorganic impurities via an ion-exchange process, this bulk NOM was freeze-dried for storage. Approximately 1 g of the freeze-dried bulk NOM was dissolved in 500 mL of deionized water at pH approximately 7.0 and further fractionated as follows. The NOM solution was first acidified to pH 1.5 with 6 M HCl to separate the humic acid (precipitates) from the fulvic acid (FA) fraction (supernatant). The precipitated humic acid (designated as GHA, with "G" representing Georgetown) was collected by centrifuging at approximately 23000 x g for 30 min, followed by freeze-drying. The supernatant solution was collected and passed through a prewashed, cross-linked poly(vinyl pyrrolidone) (PVP) column. Most of the colored materials (designated as the polyphenolic-rich NOM or Georgetown fulvic acid [GFA] fraction) were retained by the PVP, whereas the eluent contained the carbohydrate-rich NOM fraction (designated as the GCH fraction) and soluble salts. The column was then backwashed with approximately 300 mL of 0.1 M NaOH to desorb the GFA fraction from the PVP column. Each of these NOM fractions were further purified by passing through a column containing AG 50W-X4 H+ cation exchange resin, followed by a rinse with 100 mL of deionized water. Finally, NOM fractions were freeze-dried, placed into airtight glass bottles, and stored in a desiccator before use.
Thermal Analysis Methods
This study employed TMA (Model TMA 2940 thermomechanical analyzer; TA Instruments, New Castle, DE) to identify thermal transitions through evaluation of sample expansion (collapse) as a function of temperature, including, where possible, quantification of samples' thermal expansion coefficient, . Additional thermal analysis through DSC and TMDSC (TA Instruments Model DSC 2920 differential scanning calorimeter, in standard and temperature modulation modes, respectively) was applied to each sample to further evaluate thermal transition behavior, and quantify Cp. Instrument descriptions, capabilities, and calibration protocols are provided in Young and LeBoeuf (2000). Additional experimental protocols specific to this study are provided below.
Thermal gravimetric analysis (TGA) was performed to provide a means to evaluate sample decomposition temperatures and evolution of physi-sorbed water to ensure that TMA, DSC, and TMDSC protocols (i) do not exceed sample thermal decomposition temperature; and (ii) properly evolve physi-sorbed water to ensure minimal additional losses during subsequent heating and cooling cycles (which could obscure thermal transitions). Water contents are reported in Table 1. Sample thermal decomposition temperatures were determined through use of approximately 5 mg of sample on a Model 7e TGA (PerkinElmer, Wellesley, MA), scanning from ambient (approximately 30°C) to 500°C at a rate of 20°C min–1 under a nitrogen purge gas (99.0% purity; A-L Compressed Gases, Nashville, TN) flowing at 100 mL min–1. Fresh samples were then analyzed by heating from 30 to 120°C at 10°C min–1; holding at 120°C for 30 min, cooling to 30°C, and then heating again to 120°C at 10°C min–1. This analysis revealed minimal volatile losses following the initial heating and cooling cycle (ranging from additional moisture loss of 0.00% for H-NOM to a high of 0.56% for HHA).
|
Before analysis, the TMA underwent a probe, force, dimension, and temperature calibration according to standard operating procedures. A 50-g weight was provided for the force calibration. Aluminum and indium standards were employed for dimension and temperature calibration, respectively. DeLapp et al. (2004) further describe TMA scans of a synthetic polymer, poly(methyl methacrylate) (PMMA), as a means to illustrate thermal transitions and as an additional means to verify instrument performance.
To evolve physi-sorbed water, each sample was heated under a nitrogen purge gas (99.0% purity; A-L Compressed Gases) at 2.5°C min–1, from room temperature to 110°C. The sample was then held at 110°C for 30 min under a constant force of 2.5 N using a quartz macroexpansion probe (contact diameter = 6.25 mm). Following evolution of physi-sorbed water, the sample was cooled to –50°C. Three individual scans consisting of heating from –50 to 120°C at a heating rate of 2.5°C min–1 and a force of 2.5 N, were employed to ensure repeatability of the probe response. As noted above, instrument performance was verified using standard calibration procedures and examination of a PMMA standard. Analysis of each scan involved use of TA5100 Instruments Thermal Advantage software to quantify Tg and .
Differential scanning calorimetry (DSC) and TMDSC were also performed. Sample preparation involved the massing of 10 mg to 20 mg of powder sample into a standard aluminum pan, secured by the crimping of a standard aluminum cover (Lot no. 900786.901 for pan, 900779.901 for cover; TA Instruments). For improved accuracy of specific heat capacity measurements, the mass of the empty sample pan and cover was weighed to ensure that its mass was within 0.1 mg of the mass of the reference pan and cover.
Standard operating procedures for instrument calibration were also followed for the DSC instrument, including baseline slope, cell constant, and temperature. Baseline calibration is used to zero the heat flow signal. An indium standard was used for both the cell constant and temperature calibrations. The calibration procedure for TMDSC mode is the same as DSC, but it also includes an additional heat capacity calibration using a sapphire standard. DeLapp et al. (2004) further described DSC and TMDSC scans on PMMA to illustrate thermal transitions and as an additional means to verify instrument performance.
Similar to the TMA protocol, each sample was preconditioned to evolve physi-sorbed water. For DSC, the reference and sample pans were heated under a nitrogen (99.0% purity; A-L Compressed Gases) purge flowing at 80 mL min–1. The heating protocol involved sample heating at 10°C min–1 from room temperature to 110°C, holding for 30 min to evolve physi-sorbed water, and then cooling at 10°C min–1 to –135°C. The TMDSC protocol followed similar procedures, albeit with a 2.5°C min–1 heating–cooling rate.
Following conditioning, samples were cooled at a rate of 10°C min–1 to –135°C and then heated to 120°C at a heating rate of 10°C min–1 under DSC and 2.5°C min–1 under TMDSC with temperature modulation employing a ±2°C temperature amplitude, and a 90-s period of modulation. Cooling cycles followed the same protocol. Three individual scans were used in DSC, while two individual scans were used in TMDSC. Similar to TMA, instrument performance was verified using standard calibration procedures and examination of a PMMA standard. Analysis of each scan employed TA5100 Instruments Thermal Advantage software to quantify Tg and 
Cp from DSC scans, and Tg, Cp, and Cp (at 0 and 25°C) from TMDSC scans.
Powder X-ray diffraction (XRD) scans were conducted on the Harpeth fractions to investigate the amorphous and crystalline regions of these materials. Approximately 30 mg of sample was analyzed using a Scintag X1 
/ automated powder diffractometer with a Cu target, a Peltier-cooled solid state detector, and a zero-background, Si(510) sample support. Scans were recorded from 2 to 50° with 0.05° 2 steps at a speed of 0.01° min–1 with the DMSNT software (Scintag, 1997) (Hu et al., 2000; Boxall et al., 2002). The first scan was run at ambient temperature (approximately 23°C). A second run above the sample's measured Tg (approximately 100°C) was conducted to investigate the potential for a crystalline melt through the Tg.
|
RESULTS AND DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONREFERENCES |
|---|
|
before and after T2 of 1.4 and 16 µm m–1 °C–1, respectively. Harpeth fulvic acid possesses its first thermal transition (T1) at –7°C, with before and after T1 of 22 and 42 µm m–1 °C–1, respectively. A second thermal transition is observed near 30°C, while a third transition (T3) is observed at 54°C, with before and after T3 of 12 and 17 µm m–1 °C–1, respectively. The observed increases in values following the glass transition temperatures follow similar trends predicted from free-volume theory for glass transitions of macromolecules (Cowie, 1991). At the Tg, the mobility of the macromolecule is expected to markedly increase as the sample transcends into a rubber-like state. Values of following the glass transition are thus expected to be greater than those representing thermal expansion within glass-like matrices. The samples HHA and HFA were the only ones demonstrating a typical expansion curve similar to those discussed in DeLapp et al. (2004) for poly(methyl methacrylate) (PMMA) and other NOM fractions examined in that study.
|
|
Figure 1d illustrates TMA results for HCH. Thermal transitions were observed at –23 and 22°C. Figure 1d differs from results reported in Fig. 1a through 1c in that HCH experiences continual matrix contraction, with increases in the rate of contraction at suspected thermal transitions; similar behavior was observed for GCH. In this case, the relatively soft and pliable matrices of HCH and GCH could not fully support the pressures exerted by the macroexpansion probe. As temperatures increased, the matrices continued to soften as predicted by the Doolittle equation (Eisenberg, 1993), leading to additional matrix contractions. At thermal transitions, abrupt increases in free volume occur, which lead to consequent changes in viscosity, and additional probe penetration. The XRD scans of the HCH reveal a possible crystalline melt at temperatures close to 80°C. Scans below ambient temperature, however, were not possible with the XRD instrumentation used in this study. Therefore, the transitions observed below 23°C cannot be confirmed as glass transition behaviors. Insights from DSC and TMDSC experiments, however, suggest that glass transitions, or similar increases in macromolecular mobility, may be occurring at subambient temperatures. Thermal mechanical analysis for the Georgetown and Harpeth River whole NOMs was not performed due to the large matrix collapse over broad temperature ranges observed in previous experiments on whole soils (DeLapp and LeBoeuf, 2004).
Differential Scanning Calorimetry and Temperature-Modulated Differential Scanning Calorimetry
Selected DSC and TMDSC scans for HHU are illustrated in Fig. 2a and 2b , respectively. The DSC and TMDSC figures with dashed lines represent collected data, while solid lines represent points of analysis before and after an observed transition, where a "jump" in the measured heat flow provides the primary indicator for the transition. The change in the specific heat capacity (Cp) is measured along with the glass transition temperature. Multiple thermal transitions are observed at 9 and 48°C for HHU by DSC with corresponding change in specific heat capacity (Cp) of 0.014 and 0.016 J g–1 °C–1, respectively. Multiple thermal transitions are also indicated in TMDSC analysis illustrated in Fig. 2b with transition behavior at 5 and 66°C, and respective Cp values of 0.013 and 0.001 J g–1 °C–1.
|
Cp, and Cp (at 0 and 25°C) for all NOMs are reported in Table 2. Observed Tt values for HHU and HFA using both methods yield comparable results with those obtained in TMA. Results for HHA appear inconclusive considering the lack of an observed thermal transition using TMDSC. TMDSC typically possesses greater sensitivity to weak thermal transitions since kinetic effects may be separated from the overall transition as a result of the modulated heating profile derived from the heating rate, temperature amplitude, and period of modulation (Schawe and Bergmann, 1997). The coincidence of DSC and TMA measurements of thermal transitions near 45°C for this sample suggests, however, that glass transition behavior may also be occurring in this sample. Figures 3a and 3b provide illustrations of the thermal behavior of GCH derived from DSC and TMDSC experiments, respectively. Thermal transitions are observed at –82°C, between –24 and –19°C, and again between 36 and 43°C. Similar thermal transition behavior is observed for HCH. Numerous characterization studies on carbohydrates have been documented, including the investigation of glass transition behavior through DSC. Roos (1993) reported Tg values for anhydrous glucose at 31°C and for anhydrous sucrose at 62°C. Noel et al. (2000) reported Tg values for several sugars, including 39°C for anhydrous glucose and 11°C for anhydrous fructose. These values compare well with the DSC data for the Harpeth and Georgetown CH thermal transitions occurring at or above ambient temperature.
|
transitions) as opposed to a melt. Additional XRD studies of humic materials (Liu et al., 2001; Naidja et al., 2002) have provided similar evidence of amorphous structures.
Multiple Transitions
As noted above, and summarized in Table 3, multiple thermal transitions were observed for HHU, HFA, and HCH, and GCH. As described by DeLapp et al. (2004), multiple transition behavior may be similar to that exhibited by block copolymers, which Ma et al. (1995) attributed to coexisting macromolecules. Given the heterogeneous nature of NOM, more than one Tg or a range of Tg values may result (e.g., see whole NOMs in Table 3). Similar to that observed for synthetic macromolecules, multiple thermal transitions may also indicate the presence of 
, ß, or transitions. The lower temperature transitions may be associated with either noncooperative, local mobility in the side chain ( transition), or cooperative side chain mobility associated with ß transitions, both of which occur below the main chain or transitions. Synthetic macromolecules, such as poly(oxymethylene) (Bershtein et al., 2002), poly(phenylene-terephthalate) (Buijs and Damman, 1994), and PMMA (Ma et al., 1995; DeLapp and LeBoeuf, 2004), exhibit , ß, or transitions in which the and ß transitions occur at much lower temperatures than the glass transition. Multiple transitions measured with DSC at –160, –35, –50, and –120°C have previously been attributed to , ß, ß, and transitions, respectively (Buijs and Damman, 1994).
Fractionated Natural Organic Matter versus Whole Natural Organic Matter Transitions
The presence of multiple transitions in fractionated NOM samples is not necessarily reflected at the same temperature in whole NOM samples, signifying other potential influences in the thermal transition behavior of the whole NOM. In other words, the individual mobility of a HA, FA, or CH is not represented well in the whole NOM, suggesting a lack of sufficient quantities of these fractions to bring about block copolymer behavior within these two NOM samples, or, perhaps, the macromolecular behavior of whole NOM is truly distinct from the sum of its parts (or fractions). One potential cause for this difference may reside in differing thermal history and water contents; however, effects of sample preparation were minimized wherever possible. For example, samples were stored for extended periods over several weeks in sealed desiccators (with desiccant) before use (thus, each being exposed to the same ambient moisture conditions). While the NOM fractions were freeze-dried, and the whole NOMs were not, the study by Schaumann and LeBoeuf (2005) suggests that the effect of freeze-drying on NOM thermodynamic properties is negligible compared with non-freeze-dried samples. Lastly, in preparation for thermal analysis, each of the NOM samples was subjected to the same thermal history and removal of physi-sorbed water through cycling the NOM with the same heating and cooling protocol that took the sample from ambient temperature. This initial heating and cooling cycle assists in bringing each of the NOM samples "to the table" with the same thermal history.
A second explanation for differing behavior between whole NOM and its fractions may be the presence of greater ash content in the whole NOM relative to the fractions. Because the whole NOM samples are not demineralized, the presence of minerals may serve as additional binding sites for the NOM matrix. This may partially explain the lack of extremely low temperature thermal transitions within the whole NOM relative to the CH fractions in both Harpeth and Georgetown samples. Likely, the most plausible explanation is a combination of the influence of the presence of a mineral matrix, that each NOM fraction is closely intertwined with one another (e.g., linked by van der Waals and hydrogen bonds), and that the fractionation protocol greatly disrupts this relationship, leading to differing macromolecular mobility between the whole NOM and its fractions.
Source of Thermal Transitions
Attempts to identify the nature of multiple transitions in biopolymers (maple and spruce wood) have resulted in several interpretations. Kelley et al. (1987) attributed a subambient transition of –22°C and a glass transition at 50°C to amorphous components of wood. Irvine (Irvine, 1984) mentioned water as a plasticizer that produced transitions between 50 and 120°C, as well as below room temperature. Glasser et al. (1995) discussed the difficulty of pinpointing distinct transitions due to the abundance of hydrogen bonds in cellulose. ß relaxations, discussed as cooperative motions in loosely packed regions, have been designated as a universal property of amorphous materials (Noel et al., 1996). Identification of functional groups responsible for side chain mobility may be performed using 13C NMR. For example, 13C NMR studies (Girlich and Ludemann, 1994) of sugar–water mixtures have revealed that ß relaxations originate from both hydroxymethyl groups and water. Although NOMs investigated in this study are relatively "dry," the presence of residual, bound water within the NOM matrix could similarly explain multiple transition behavior by attributing one or more of the transitions to ß relaxations arising from polar functional group–water interactions, that is, potential hydrogen bond breakage as noted by Schaumann and LeBoeuf (2005). Additional research that investigates the activation energies associated with observed multiple thermal transitions are currently underway.
Challenges in interpreting the source of multiple thermal transitions also exist in block copolymer research. For example, Lin et al. (2001) endeavored to determine the source of multiple endotherms (soft segment and hard segment glass transition temperatures) observed on polyurethane. Previous attempts have been unable to explain the nature of the soft-segment Tg. The study attributed this transition to phase mixing of the soft segments and hard segments; however, the exact nature of the transition continues to be studied in terms of intermolecular interactions due to hydrogen bonding (Liu et al., 2001) or the influence of decreasing molecular weight of the soft segments (O'Sickey et al., 2002).
Specific heat capacity measurements at 0 and 25°C are within comparable ranges for all NOMs, and are similar in magnitude to Cp values reported for International Humic Substances Society (IHSS) NOMs (DeLapp et al., 2004). This suggests that Cp values do not vary much between samples, even among samples of widely differing origins and chemical functional group composition. Evaluations of the magnitude of Cp at thermal transitions suggest a relatively narrow range of values from 0.001 to 0.04 J g–1 °C–1, with the exception of Cp values for both CH fractions, which possess significantly larger values. These measurements for carbohydrates, however, are comparable in magnitude to Cp values expected for glass transitions in organic macromolecules (Borde et al., 2002). The HU, HA, and FA fractions possess Cp values similar to those reported for other NOM (LeBoeuf and Weber, 2001; DeLapp et al., 2004). Measured Cp values for HFA are within the same order of magnitude as those reported for Nordic aquatic humic and fulvic acids (DeLapp et al., 2004), which are one magnitude lower than other samples. On the other hand, the Cp values of the CH fractions are one order of magnitude higher. The Cp of 0.24 J g–1 °C–1 for GCH compares well with the Cp of polysaccharides, typically reported as 0.2 J g–1 °C–1 (Borde et al., 2002).
The magnitude of Cp is a function of the increased mobility observed below and above the transition. For example, ß transitions should inherently possess smaller Cp values relative to transitions since only side-functional groups are being mobilized compared with whole-chain mobility. Further, it is possible that the larger Cp values may also be attributed to more homogeneous materials (LeBoeuf and Weber, 2000a), since Cp values are normalized to the entire mass of the sample. Hence, observation of multiple transitions would indicate only a portion of the total mass of the macromolecule is becoming mobilized at a particular temperature.
Chen et al. (2002) 13C NMR analysis of GCH and GFA indicates a greater abundance of aromatic and methoxyl functional groups, as well as phenolic and carboxyl functional groups, in GFA as compared with GCH. Given the higher aromaticity of GFA, its glass transition temperature may be expected to be higher due to the additional stiffness of aromatic structures. However, this additional stiffness may be offset by the presence of stronger intermacromolecular bonds in GCH due to its increased polar nature. Such increased interactions may be evidenced by the large Cp noted at the higher temperature transition for both GCH and HCH.
Factors influencing the mobility of a macromolecule include (i) stiffness of the macromolecule (more aromatic regions generally are more rigid than aliphatic regions); (ii) cohesive energy density (larger solubility parameters generally coincide with more polar macromolecules providing for increased interaction energies between individual macromolecule segments); (iii) cross-link density (increased number of cross-links reduce segment mobility); and (iv) molecular weight (larger molecular weights tend to possess decreased numbers of mobile endgroups) (Salmen, 1984). Properties that provide for decreased mobilities generally are expected to result in increased energies required for mobilization (or increased glass transition temperature).
The range of temperatures for thermal transitions observed in this study vary widely from –87°C for HCH to 62°C for HHU. General observations provided by this study include: (i) variable-temperature XRD studies suggest that NOM samples are primarily amorphous in nature and that no crystalline melts occur between ambient temperature and temperatures at the higher thermal transitions; (ii) Tt for HHU is similar to HHA at the higher temperature; (iii) lower temperature Tt values are similar for HHA, GHA, GFA, and HFA (upper); (iv) presence of very low temperature Tt in both HCH and GCH; (v) large Cp in carbohydrate fractions suggest the potential influence of strong polar–polar interactions on the higher temperature Tt (likely may be classified as a Tg); and (vi) presence of multiple transitions in fractionated NOM samples is not necessarily reflected in whole NOM samples, suggesting other potential influences in the thermal transition behavior of the whole NOM relative to that observed in the fractionated NOM.
Because a number of factors influence the mobility of a macromolecule, it is difficult to clearly delineate trends in transition temperatures based solely on elemental or functional group composition. Additional work that probes the disruption of particular functional groups and/or sources of intra- and interchain bonds (e.g., van der Waals or hydrogen bonds) through use of water and other probe sorbates may prove useful in further elucidating the components of NOM most responsible for influencing the mobility of the macromolecule.
Use of Thermodynamic Data
Thermodynamic data that quantifies thermal transitions such as , ß, and transitions provide additional insight to the potential role of increased macromolecular mobility on sorption and desorption behavior of organic compounds in aquatic- and terrestrial-based NOM matrices. Sorption within glassy macromolecules, with their relatively rigid character, is characterized by adsorption to surfaces and within microvoids, while sorption in rubbery systems, where increased fluidity of the matrix results in the loss of "frozen-in" microvoids, is similar to uptake within a viscous fluid or gel (Weber et al., 2001). Such differences in mechanisms can manifest themselves in nonlinear sorption, competitive sorption for glassy NOMs, and linear, noncompetitive sorption in rubbery systems. For example, LeBoeuf and Weber (1997)( 2000b) demonstrated the transition of isotherm linearity from nonlinear behavior below the glass transition temperature to linear behavior above the glass transition temperature in several NOM samples, while Xing et al. (1996) demonstrated the presence of competitive sorption in glassy NOM systems. Nonequilibrium behavior may be categorized based on relative rate of solute diffusion and relaxation or reconfiguration of a macromolecular structure, where four different types of diffusion phenomena may occur in organic macromolecules: Case I (Fickian), Case II, Super Case II, and anomalous or non-Fickian diffusion (Hopfenberg and Stannett, 1973), where the rate of macromolecular relaxation is directly related to macromolecular mobility (i.e., diffusion in rubbery matrices is normally not hindered by the relaxation rate of the macromolecule). One may expect glassy macromolecules to possess orders of magnitude lower diffusion rates relative to the same macromolecule in the rubbery state (Vieth and Sladek, 1965). With respect to multiple thermal transitions representing a sequential increase in macromolecular mobility, one may expect particular segments or regions in NOM to become mobilized with increasing temperature (or solute-induced swelling), leading to changes in the sorption mechanism within the affected region.
When used in combination with advanced spectroscopic and computational chemistry-based techniques, advanced thermal characterization can prove to be a useful tool in further understanding the structural conformation of NOMs. For example, one- and two-dimensional 1H and 13C NMR can assist in identifying functional group connectivity and mobility (Chien and Bleam, 1998; Mao et al., 2000; Diallo et al., 2003). Correlating observed NMR-based functional group mobilities with thermal transitions provides a means to quantify energies associated with the movement of functional groups or whole chains from conditions of low mobility (i.e., glass-like) to conditions of high mobility (i.e., rubber-like), or vice versa. Similarly, the effects of sorbate chemical potential on mobilization of NOM matrices may be quantified in terms of reduced energies (or temperatures) for realization of ß and transitions.
Until recently, structural models of humic substances derived from molecular modeling have not extended their application beyond inference from spectroscopic elemental–functional group composition (Schulten and Schnitzer, 1997; Leenheer et al., 1998; Mao et al., 2002). Poerschmann and Kopinke (2000) and Diallo et al. (2001), however, provide insights to the use of molecular modeling to derive basic thermodynamic properties of humic materials such as cohesive energy density (solubility parameter), including comparisons with available experimental data. Additional thermodynamic information, specifically , Tg, and Cp, may also be derived from such models (DeLapp and LeBoeuf, 2004). For example, molecular dynamics simulation-derived and Tg compared favorably with experimental data for a number of materials, for example, polyethylene (Yu and Christie, 2001), PMMA and polymethacrylic acid (Soldera, 2002), amorphous CaAl2Si2O8 (Morgan and Spera, 2001), and deoxyribonucleic acid (DNA) (Norberg and Nilsson, 1996). In the case of humic substances, atomistic simulations and computer-assisted structure elucidation (e.g., Diallo et al., 2003) may benefit greatly by using advanced thermal characterization as additional constraints within an overarching framework for developing three-dimensional structural models of NOMs.
|
ACKNOWLEDGMENTS |
|---|
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONREFERENCES |
|---|
Related articles in JEQ:
