Environ. Sci. Technol. 2002, 36, 3960-3967
Black Carbon and Kerogen in Soils and Sediments. 1. Quantification and Characterization JIANZHONG SONG,† P I N G ’ A N P E N G , † A N D W E I L I N H U A N G * ,‡ State Key Laboratory of Organic Geochemistry, Guangzhou Institute of Geochemistry, Chinese Academy of Sciences, Wushan, Guangzhou 510640, P.R. China, and School of Environmental Science, Engineering and Policy, Drexel University, Philadelphia, Pennsylvania 19104
A comprehensive wet chemical procedure was developed by combining acid demineralization, base extraction, and dichromate oxidation for fractionation and quantitative isolation of soil/sediment organic matter (SOM) into four fractions: (1) humic acids + kerogen + BC (HKB); (2) kerogen + BC (KB); (3) humic acid (HA); and (4) BC. The soil/ sediment samples tested were collected from the suburban areas of Guangzhou, a rapidly developing city of China. The results show that BC and kerogen constitute 57.880.6% of the total organic carbon (TOC) and that the relative content of BC ranges from 18.3% to 41.0% of the TOC, indicating that both BC and kerogen are major organic components in soils and sediments from this industrialized region. Systematic characterization of the isolated SOMs shows that both BC and kerogen have sizes ranging from a few microns to above 100 µm, relatively low O/C and H/C atomic ratios, and low contents of oxygen-containing functional groups. The isolated BC has unique fusinite and semifusinite macerals, highly porous nature, and structures indicative of its possible origins. The study indicates that SOM is highly heterogeneous and that humin, the nonextractable humus fraction, consists mainly of kerogen and BC materials in the tested soil/sediment samples. The presence of these materials in soils and sediments may have significant impacts on pollutant mass transfer and transformation processes such as desorption and bioavailability of less polar organic chemicals in surface aquatic and groundwater environments.
Introduction A concept of “soft carbon” (or amorphous) vs “hard carbon” (or condensed) soil organic matter (SOM) has recently been invoked to operationally delineate chemical heterogeneity of SOM and to mechanistically interpret sorption of hydrophobic organic chemicals (HOCs) by soils and sediments (1-5). While the conceptual model can explain a spectrum of the observed nonpartitioning phenomena such as nonlinear isotherms, sorption-desorption hysteresis, and slow rates of sorption and desorption, the nature of the hard carbon SOM remains ambiguous. It has been hypothesized in the literature that humic acids, kerogen, and black carbon * Corresponding author phone: (215)895-4911; fax: (215)895-2267; e-mail:
[email protected]. † Chinese Academy of Sciences. ‡Drexel University. 3960
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 36, NO. 18, 2002
are all possibly responsible for the nonpartitioning phenomena (1-15). Humic acid, an SOM fraction soluble in bases but insoluble in acids (16, 17), was found to be a relatively condensed or glassy phase at temperatures lower than its glass transition temperature (5, 9, 18), and to exhibit nonlinear sorption isotherms for less polar HOCs (e.g., ref 5). It may not be the only SOM fraction responsible for the overall nonpartitioning phenomena by soils and sediments, however (e.g., 19). Humin, the residual SOM fraction after base extraction (20, 21), often exhibited more nonlinear sorption isotherm than the humic acid extracted from the same soil or sediment for a given HOC (19, 22). Kerogen, a natural organic matter fraction that is insoluble in organic solvents and acids/bases and has undergone various diagenetic alterations in geological histories (23), was shown to have very nonlinear sorption isotherms with much greater capacities for HOCs (e.g., 2, 7, 9, 11). Black carbon (BC) formed during incomplete oxidation of biomass and fossil fuels was believed to be a super-sorbent because of its high specific surface areas and relatively reduced chemical nature (6, 8, 12-15). Indeed, modern BC was shown to exhibit sorption capacities several times greater than those of humic acids (14), and soot in harbor sediments was believed to be responsible for the nonlinear isotherms observed for pyrene (15). As increasing data suggest the important roles of both kerogen and BC in the sorption and sequestration of HOCs in soils and sediments, identifying and quantifying these materials, characterizing their physical and chemical properties, and establishing quantitative correlation between their properties and the measured HOC sorption parameters are much needed. This study was initiated to isolate, quantify, and characterize BC and kerogen associated with soils and sediments and to examine their roles in the sorption of HOCs in aquatic environments. It is expected that the operationally defined humin fraction of soils and sediments in heavily industrialized areas may consist of particulate hard carbon SOM; i.e., BC and kerogen. Limited studies showed that humin may be comprised of humic acid-like materials (20), plant residues (e.g., lignin), and kerogen-like materials (21); but the major physicochemical properties of humin may vary dramatically (21). As suggested by a research team who recently claimed that BC could be a “super-sorbent” for HOCs (24), isolation of BC may be very difficult due to the complexity of soils. In fact, existing methods for quantifying BC contents in soils, sediments, and sedimentary rocks involve thermal and/or chemical oxidation to destructively remove non-BC fraction and/or mineral matrix (12, 13, 15, 25-34). The remaining inert organic carbon is operationally defined as BC, char, soot, or elemental carbon and can be quantified from mass balance or elemental composition or using spectroscopic means. Due to the heterogeneity of BC and varied capability for removing non-BC fraction and altering BC of different origins, BC contents measured using different methods for a given sample may vary dramatically (30, 31). In this study we utilized a comprehensive procedure consisting of major steps of demineralization, base extraction, and chemical oxidation, each of which has been used in prior studies for isolation of specific SOM fraction(s) such as humic acids (16, 17), kerogen (23), and BC (25, 26). Unlike prior studies that focused on quantification and/or isolation of a specific fraction but discarded others, we combined these procedures to fractionate bulk SOM to a spectrum of fractions including BC, BC + kerogen mixture, and humic acids and to quantify them on an organic carbon content basis. Because quantification, rather than characterization, of soil/sediment10.1021/es025502m CCC: $22.00
2002 American Chemical Society Published on Web 08/17/2002
FIGURE 1. A flowchart of the experimental procedure for fractionation and quantification of the four different SOM fractions. associated BC was the focus of prior studies (25-34), this study was designed to characterize major physical and chemical properties of BC as well as other SOM fractions using different techniques. The heterogeneity of SOM was thus delineated with relative distributions of different types of SOM and their differences in morphology, sizes, and elemental compositions. The impact of SOM heterogeneity on sorption and desorption equilibria of HOCs was examined by conducting isotherm experiments using the SOM fractions and the original soil/sediment samples as the sorbents and using phenanthrene and naphthalene as the HOC probes. This paper, the first in a two-part series, presents the SOM isolation procedure and the contents and major physical and chemical properties of the isolated SOM fractions. The measured sorption equilibria and the interrelationship among sorptive phenomena, contents, and properties of BC and kerogen will be presented in the companion paper.
Experimental Section Soil/Sediments. Four soil/sediment samples were collected from the suburban areas of Guangzhou City, P. R. China, which has a population of over 7 million and a total area of over 7000 km2. These samples were selected from a set of ∼50 soils and sediments collected in 2000 from this area and are representative of different surface aquatic environments polluted variously during the rapid development and industrialization over the past two decades. They likely contain large quantities of coal and BC particles due to extensive utilization of fossil fuels and the resulting high atmospheric particulate matter concentrations. They include the following: (1) a brownish/yellowish sandy soil (SS) from a paddy field; (2) a marine sediment (MS) collected at depth of 0-15 cm from the Wanshan Archipelagos (113°49′20′′E, 22°22′05′′N) of Zhuhai; (3) a river sediment (RS) collected at depth of 0-15 cm from the Pearl River; and (4) a pond sediment (PS) collected at depth of 0-10 cm from Baishigang. After collection, the soil/sediment samples were air-dried and passed through a 1.0-mm sieve, and plant debris was hand picked. They were stored in glass bottles under an N2 atmosphere until use for isolation of SOM. A modern BC sample was used in this study for comparison of the BC particles isolated from the four samples. It is a char-like material produced from arbor branches in a cooking kiln with limited air entry under uncontrolled temperature condition.
Fractionation and Quantification of SOM. A comprehensive procedure was developed in this study to quantify contents of BC, kerogen, and humic acids and to isolate BC and kerogen in large quantities from the soil/sediment samples for characterization and sorption studies. The isolated four SOM fractions include the following: (1) humic acids + kerogen + BC (HKB); (2) kerogen + BC (KB); (3) humic acid (HA); and (4) BC. A total of 15 SOM fractions were obtained for the four samples; the mass of the HA fraction isolated from MS was not sufficient for further study. A flowchart of the procedure is outlined in Figure 1, and the details are described below. The first step of the procedure was to obtain the HKB fraction by removal of the dominant mineral matrix of a sample using a hot HF + HCl acid mixture. In this step, a predetermined amount (70 g) of a sample was treated at 60 °C for 20 h in 300 mL of 6 M HCl acid contained in plastic centrifuge bottles. After digestion, the contents were centrifuged (1870g) for 30 min, the supernatants were decanted, and the residue was rinsed three times with 2 M HCl. The solid residue was then transferred to a Teflon beaker and demineralized at 60 °C for 20 h in concentrated HCl (6 M) + HF (22 M) mixture (300 mL) at a volumetric ratio of 1:2. After digestion, the content of the beaker was transferred to a plastic centrifuge bottle and centrifuged, and the supernatant was decanted. The residue was rinsed with Milli Q water and then soaked at ∼60 °C for 10 h in 6 M HCl (200 mL) to remove such minerals as fluorite formed during the demineralization procedure. The residue was rinsed again with Milli Q water until the aqueous phase became neutral (pH ∼ 7). After being dried at 60 °C, the residual solid was Soxhlet extracted for 72 h using a mixture of methanol, acetone, and benzene at a volumetric ratio of 2:3:5 to remove extractable organic matter such as hydrocarbons, lipids, and asphaltenes (23, 35) as well as organic pollutants bound tightly to SOM matrices. The Soxhlet extraction was performed after, rather than before, the demineralization because the solid to be extracted and hence the volume of the solvent mixture used were reduced dramatically. After extraction, the solid (HKB) was vacuum-dried at ∼60 °C to remove solvent molecules. To obtain sufficient quantity for each of the four SOM fractions for both characterization and sorption studies, replicates with a total mass of kilograms of a soil or sediment sample were used in the first step. The HKB fractions obtained VOL. 36, NO. 18, 2002 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
3961
from several replicates were combined, mixed, and homogeneously split to several portions for use in subsequent fractionation and characterization of the SOMs and sorption experiments. The second step was to fractionate the HKB to HA and KB following a standard base extraction procedure (16). The KB fraction should be the major component of the operationally defined humin fraction of SOM. The base extraction was performed after, rather than before, the demineralization procedure to completely remove both HA that may not be fully base-extractable when bound to mineral surfaces and products of acid hydrolysis of such organic matter as amorphous kerogen and biopolymers (35). A predetermined amount of the HKB solid was placed in a centrifuge bottle containing N2 gas-purged NaOH solution (0.1 M). After extracted for 12 h, the content of the bottle was centrifuged at 1870g for 30 min, and the supernatant was transferred to another bottle. The extraction procedure was repeated 5 times until the supernatant became colorless. The base-extracted residue containing KB was freeze-dried and stored. The extracts of different batches were combined and acidified to pH ∼ 1 using a concentrated HCl acid (6 M), and the HA was allowed to precipitate. After centrifugation, the HA was freezedried and stored. The last step was to isolate BC using the KB fraction as the starting material and a dichromate/sulfuric acid solution (K2Cr2O7 (0.1 M) + H2SO4 (∼2 M)) to oxidize kerogen, while BC remains fairly unchanged. Several different oxidants such as persulfate (36), nitric acid (29, 31), and dichromate (25, 26, 34) were used in prior studies for chemically isolating BC from sediments. We chose dichromate because it has been shown to be most effective for oxidizing kerogen with minimal effect on BC including charred materials (25). In brief, a predetermined amount of the KB solid was placed in a bottle containing the aqueous dichromate solution, and the KB material was disaggregated in an ultrasonic bath for 10 min for better solid-solution contacts. The bottle was then placed in a water bath at 55 ( 1 °C for 60 h. This reaction time was chosen based on prior studies (25, 26), and our preliminary test that showed a maximal digestion of kerogen and coal and minimal depletion of BC. During oxidation, the solution phase was replaced at least twice with the fresh dichromate solution to maintain relatively constant pH and overdosed oxidant, and Milli-Q water was added every 2 h to maintain a constant volume of solution phase. At the end of the reaction, the bottle was placed in a cold-water bath for ∼5 min to lower the temperature of the solution and then centrifuged for 30 min. After the supernatant was decanted, the residual was rinsed ∼5 times with Milli Q water and oven dried at 60 °C for 48 h. The above procedure was modified slightly to determine the contents of humic acids, kerogen, and BC by quantitative isolation of the four SOM fractions. Three different batches of experiments were set out, each batch having six replicates initiated simultaneously with 3 g of a given original sample that had been Soxhlet extracted following the procedure described above. The three batches were terminated at the end of the first, second, and third step, respectively, and a total of 24 different SOM samples were obtained for a given soil or sediment. The mass of each SOM fraction was determined on a microbalance ((10 µg). TOC and Elemental Analyses. A high-temperature combustion-based Heraeus elemental analyzer (CHN-O-RAPID) was utilized for determinations of the total organic carbon (TOC) contents of the original samples and the quantitatively isolated SOM fractions and for measurements of the C, H, N, and O contents of each large-quantity SOM fraction and the modern BC sample. The original samples were decarbonated using a 0.5 M HCl before TOC analysis. The ash contents of large-quantity SOM fractions were determined 3962
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 36, NO. 18, 2002
independently by complete oxidation of each SOM fraction (∼0.1 g) under 950 °C in a furnace. The final data are reported based on the analyses of triplicates for each sample, and the calculated relative standard deviation is less than 6%. Particle Sizes and Specific Surface Areas. Particle sizes were measured using a SALD-3001 laser scattering particle analyzer having precision below 5% for particles between 2 µm and 125 µm. Before measurement, SOM samples were ultrasonically treated and dispersed in a solution containing 20 wt % of Na3PO4. Specific surface areas (SSAs) were calculated from the N2 gas adsorption isotherm data collected for each sample at liquid nitrogen temperature (NOVA-1000, Quantachrome Corporation). X-ray Diffraction (XRD). A standard powder X-ray diffraction method was used to determine the mineral species that had not been removed during the demineralization process and to estimate the interplanar spacing of aromatic sheets of organic matrices. The X-ray diffractometer used was set up with Cu target, Ni filtering, and a radiation wavelength of 1.5405 Å (Rigaku D/MAX-1200). Organic Petrographic Examination. The organic facies and the shapes, sizes, and degree of maturation of the SOM fractions were examined under an optical microscopy (Leitz MPV-3) in three different modes: transmitted, reflected, and fluorescent; the latter was induced by blue excitation (546 nm). The SOM thin sections used for examinations in transmitted modes were prepared by evenly spreading SOM powders on glass slides, which were then cemented with glycerol and covered with cover glass (37). The polished sections used for reflected and fluorescent modes were prepared by spreading SOM powers on porcelain slides and cementing the powders with low-fluorescence 502 mucilage glue. The slides were then thin-sectioned and polished. Scanning Electron Microscopy (SEM). The shapes, sizes, and morphology of SOM particles of different fractions were examined using an SEM (Hitachi S-3500N SEM with Oxford Link ISIS 300 EDS (energy-dispersive X-ray spectrometer)). 13C-Nuclear Magnetic Resonance (NMR) Spectroscopy. The functionalities of the isolated SOM fractions were determined from the solid-state 13C NMR spectra obtained using Bruker MSL-300 on a 7.5-mm probe at 75.47 MHz carbon frequency with a cross-polarization/magic angle spinning (CP/MAS). The signals were recorded at 1.5-ms cross-polarization contact time and approximately 1000 data points were obtained for each sample. The time periods for signal acquisition for a given sample varied from ∼1 to 10 h.
Results and Discussion Contents and Elemental Compositions of the Isolated SOM Fractions. The relative contents of HKB, KB, HA, and BC were calculated on an organic carbon content basis from the following: (i) weight of a given SOM fraction isolated (MSOMi); (ii) TOC content of the SOM fraction (TOCSOMi); (iii) weight of the original sample from which the SOM fraction was isolated (Mo); and (iv) TOC content of the original sample (TOCo). The following equation was used for such calculation
SOMi(wt% ) )
TOCSOMi × MSOMi TOCO × MO
× 100
The average value of SOMi % was calculated from the determinations of six replicates, and the relative standard deviations were less than 6%. The final results of the relative contents of different SOM fractions are summarized in Table 1 for the four samples studied. The elemental compositions and ash contents of the 15 extracted SOM fractions and the modern BC sample, along with the H/C and O/C atomic ratios, are listed in Table 2.
TABLE 1. TOC (wt %) and the Relative Contents (% TOC) of the SOM Fractions
sample
TOC
sandy soil (SS) river sediment (RS) pond sediment (PS) marine sediment (MS)
HA
1.55 15.0 3.17 5.05 2.10 6.02 0.87 0.29
K
BC
24.0 47.6 30.8 39.4
35.9 33.0 41.0 18.3
hard carbon HKB (KB) 74.8 85.6 77.8 58.0
59.7 80.6 71.8 57.8
TABLE 2. Elemental (C, H, O, N) Compositions (wt %), Ash Contents (wt %), and Specific Surface Areas of the Isolated SOM Fractions sample
C
H
O
atomic atomic O/C H/C
N
ash
SSAa (m2/g)
SS RS PS MS
52.3 49.1 54.4 17.1
4.51 17.0 3.39 16.2 3.40 13.0 2.26 8.81
HKB 2.36 1.26 1.90 0.81
0.24 0.25 0.18 0.39
1.03 0.83 0.75 1.59
20.57 1.93 28.5 7.29 23.6 8.59 67.5 11.0
SS RS PS MS
48.6 50.0 57.5 17.5
3.66 12.7 3.89 11.1 3.16 13.4 2.27 8.54
KB 1.26 1.27 1.66 0.82
0.20 0.17 0.17 0.37
0.90 0.93 0.66 1.55
32.1 32.1 21.1 69.2
1.66 0.94 2.01 16.8
SS RS PS
43.3 38.4 39.7
4.03 23.4 4.30 19.5 3.61 22.2
HA 2.74 2.03 2.47
0.41 0.38 0.42
1.12 1.34 1.09
25.4 NAb 27.5
0.31 NAb 0.35
SS RS PS MS modern
42.6 2.58 8.40 61. 3.18 9.01 57.4 2.44 10.1 29.7 1.77 8.47 67.8 12.9 2.09
BC 0.77 1.02 1.15 0.68 1.56
0.15 0.11 0.13 0.24 0.14
0.73 0.63 0.51 0.72 0.37
43.0 22.1 27.0 56.2 15.7
0.59 2.97 1.09 2.99 26.2
a N -BET special surface area. 2 mass.
b
Not analyzed due to insufficient
The results listed in Tables 1 and 2 indicate that (1) about 58-85.6% of the TOC are retained for the four samples tested; (2) the particulate kerogen and BC constitute 57.8-80.6% of the TOC; (3) BC is 18.3% of the TOC for MS and 33-41% of the TOC for the other three samples; and (4) the elemental compositions of the four SOM fractions differ dramatically. The estimated BC contents for the four samples are comparable to those reported in the literature. For instance, charcoal was reported to account for 5-45% of TOC for soils (38, 39), and BC contents are within ranges of 3-13% TOC for the sediments of New England harbors (40), 5-39% TOC for lake sediments of France (26), and 15-30% TOC for marine sediments (29). The major sources of error in the measured BC using this method include the loss of fine particles to decantation of supernatant after base extraction, oxidation of BC particles, and incomplete digestion of kerogen by dichromate. Following the base extraction procedures, dispersed dark-colored fine particles that are presumably BC or kerogen materials were sometimes observed unsettled after centrifugation for hours likely due to their surface hydrophobicity and low bulk density. The amount of such unsettleable particles is estimated to be
