1
Evaluation of DAX-8 fractionation of soil DOM
2
DAX-8 fractionation of dissolved organic matter (DOM) from soils:
3
calibration with test components and application to contrasting soil
4
solutions
5
F. AMERY, C. VANMOORLEGHEM AND E. SMOLDERS
6
Section Soil and Water Management, Department Earth and Environmental Sciences,
7
K.U.Leuven, Kasteelpark Arenberg 20, 3001 Heverlee, Belgium
8
Correspondence: F. Amery. E-mail: Fien.Amery@ees.kuleuven.be
1
9
Summary
10
Most methods to fractionate natural dissolved organic matter (DOM) rely on sorption of
11
acidified DOM samples onto XAD-8 or DAX-8 resin. Procedural differences among methods
12
are large and their interpretation is limited due to lack of calibration with DOM model
13
molecules. An automated column based DOM fractionation method was set up for 10 mL
14
DOM samples dividing DOM into hydrophilic (HI), non-neutral hydrophobic (nn-HO) and
15
neutral hydrophobic (n-HO) fractions. Fifteen DOM model components were tested in
16
isolation and in combination. Aliphatic low molecular weight acids (LMWA’s) and
17
carbohydrates were classified as HI DOM, but maleic, succinic and propionic acid showed
18
also a partial HO character. Aromatic LMWA’s and polyphenols partitioned in the nn-HO
19
fraction, menadion (quinon) and geraniol (terpenoid) in n-HO DOM. Molecules with log Kow
20
> 0.5 had negligible HI fractions. The HO molecules except geraniol had specific UV
21
absorbance (SUVA, measure for aromaticity) > 3 l g-1 cm-1 while HI molecules had SUVA
22
values < 3 l g-1 cm-1. The DOM of 8 different soil solutions was fractionated with this method.
23
Distributions ranged from 31-72% HI, 25-46% nn-HO and 2-28% n-HO of total dissolved
24
organic carbon with notable effects of DOM sampling method. The SUVA of the HI DOM
25
was lower compared to the HO DOM, but its variation coefficient was only slightly lower for
26
the nn-HO fraction and even larger for the HI fraction compared to the unfractionated DOM.
27
The calibration data allow interpreting results of the fractionation method. Our data show that
28
isolated soil DOM is not structurally more homogeneous than the unfractionated sample,
29
contrasting the general paradigm in fractionation approaches.
30
2
31
Introduction
32
Dissolved organic matter (DOM) has an important role in aquatic and soil systems for
33
contaminant transport, for biochemical and geochemical processes, and serves as a substrate
34
for microorganisms (Leenheer & Croué, 2003; Marschner & Kalbitz, 2003). The influence of
35
DOM on these processes does not only depend on the dissolved organic carbon (DOC)
36
quantity, but also on the DOM quality, i.e. the structural features (Raber et al., 1998; Kalbitz
37
et al., 2003). The composition of DOM in the soil solution varies with soil type, season, water
38
regime, soil depth and vegetation or crop.
39
Due to its complex and heterogeneous composition, soil DOM is commonly analyzed by
40
measuring bulk properties (e.g. spectroscopy) or by fractionating DOM based on chemical or
41
physical properties. The underlying idea of fractionation is that DOM fractions are
42
structurally more homogeneous than the bulk samples (Leenheer, 1981), however this is
43
rarely verified. A frequently used fractionation procedure is based on sorption of acidified
44
DOM onto the weakly hydrophobic resin XAD-8 dividing DOM into hydrophobic (HO) and
45
hydrophilic (HI) DOM fractions (Kaiser, 1998; Van Zomeren & Comans, 2007). The division
46
of DOM into the 2 fractions shows environmental relevancy. Metals transported in soil are
47
particularly associated with hydrophilic DOM (Guggenberger et al., 1994). This fraction has
48
also shown larger bioavailability (Marschner & Kalbitz, 2003), whereas polycyclic aromatic
49
hydrocarbons have larger affinity to hydrophobic DOM (Raber et al., 1998). Reported
50
distributions of soil DOM vary largely: 25 to 90% HI and 10 to 75% HO of total DOC (Table
51
1). The fractionation does not only change with soil depth, DOC concentration, sampling
52
method, land use and solution chemistry, but also depends on the fractionation method.
53
Most of the fractionation procedures are based on the (first step of the) fractionation
54
scheme of Leenheer (1981), that classifies dissolved organic molecules into hydrophobic and
55
hydrophilic fractions based on their pH dependent sorption onto the XAD-8 resin. As the
3
56
production of XAD-8 has been ceased, it is now replaced by SupeliteTM DAX-8, also a
57
polymethylmethacrylate resin. The sorption-desorption properties of the latter are slightly
58
more precise compared to the former (Peuravuori et al., 2002). Many modifications to this
59
method have been proposed, but generally the procedure is as follows. After acidifying the
60
DOM sample to pH 2 to neutralize all weak acid groups, the solution is pumped over a
61
column with DAX/XAD-8 resin. The column is rinsed with 0.01 M HCl. The DOM not
62
sorbed onto the resin is classified as hydrophilic. The hydrophobic DOM except for the
63
neutral molecules can be removed from the resin by rinsing with 0.1 M NaOH. The
64
interpretation of the fractionation in terms of the octanol water partition coefficient (Kow),
65
molecular weigth (MW) or acidity is, however, unknown, and the classification of the solutes
66
into hydrophilic and hydrophobic fractions depend on the procedure used (Peuravuori et al.,
67
1997). Authors modify the procedure according to the volume, concentration and application
68
of the sample (Marhaba et al., 2003). In aquatic studies, the resin is frequently used to
69
concentrate and isolate DOM, thereby neglecting the hydrophilic part (Thurman & Malcolm,
70
1981). In contrast to aquatic samples, fractionation of soil solution DOM has often sample
71
volume limitations, requiring adapted fractionation techniques.
72
Despite the limited interpretation of fractionation, few studies have examined the
73
classification of specific organic molecules into the hydrophilic and hydrophobic fraction.
74
The hydrophobic fraction of natural DOM is assumed to contain lignin-derived, partly
75
aromatic molecules, nucleic acids, quinones, hydrocarbons and fatty acids. Hydrophilic DOM
76
consists of amino acids and sugars, small organic acids and carbohydrates and contains less
77
aromatic components than HO fractions (Cilenti et al., 2005; Simonsson et al., 2005). The list
78
of compounds to be found in each fraction is often based on spectroscopic measurements of
79
the fractions (Leenheer, 1981; Guggenberger & Zech, 1994; Kaiser et al., 2001a; Dilling &
80
Kaiser, 2002; Croué et al., 2003; Cilenti et al., 2005), instead of subjecting model DOM
4
81
components to the fractionation procedure. Among the exceptions are Qualls & Haines (1991)
82
who tested 5 compounds in their modified fractionation procedure. Over 90% of rutin and
83
quercitin (flavonoids) was found in the hydrophobic fraction, whereas over 85% of oxalic
84
acid, albumin and glucose phosphate was classified as hydrophilic DOM. David et al. (1989)
85
recovered over 89% of oxalic, succinic and citric acid as hydrophilic acids; over 95% of
86
benzoic and salicylic acid was found in the HO fraction. Thurman et al. (1978) determined
87
the capacity factors of 20 nonionic organic solutes on XAD-8 and concluded that the resin
88
favors aliphatic over aromatic over alicyclic components. The logarithm of the capacity factor
89
correlated inversely with the logarithm of the aqueous molar solubility, but it remains unclear
90
which components can be expected to appear in the hydrophilic fraction.
91
Numerous modifications to the fractionation procedures have been described varying in
92
resin/volume ratio, flow rate in the columns (often not reported) and in further fractionation of
93
HI or HO fractions. These modifications are known to affect the fractionation and it is striking
94
that no calibration steps are systematically adopted in modified procedures. We developed a
95
fractionation procedure with DAX-8 for only 10 ml of DOM solution. Fifteen soil DOM
96
model components were subjected to this fractionation scheme. Soil solutions from different
97
sources were subsequently fractionated. Specific UV absorbance (SUVA, 254 nm, Weishaar
98
et al., 2003) of the DOM and their fractions were measured to determine relationships with
99
the hydrophobicity and to study characteristics of the fractions. It was investigated if the
100
structural variability (linked to SUVA) of the bulk samples is lower in the corresponding
101
fractions.
102
Materials and methods
103
Fractionation procedure
5
104
The procedure to fractionate DOM solutions is based on preliminary experiments with water
105
extracts of soil. The first step of the fractionation scheme of Leenheer (1981) was adapted to
106
clearly separate hydrophilic and hydrophobic DOM of small volumes of soil solutions.
107
SupeliteTM DAX-8 resin was cleaned according to the method of Thurman & Malcolm
108
(1981). The resin was extracted with 0.1 M NaOH, replacing the solution and decanting off
109
fines daily. After 5 days, the resin was soxhlet-extracted for 24 hours with methanol,
110
acetonitrile, diethyl ether and again methanol. The resin was stored in methanol/water until it
111
was packed in an 8.0 cm x 1.6 cm column, giving a bed volume (BV) of 16 ml. The column
112
was connected to a compact liquid chromatographic system (Äkta Prime Plus, GE Healthcare)
113
with fraction collector and online pH, conductivity and UV (254 nm, path length 0.2 cm)
114
detectors. The void volume (VV) between the injection valve and the detectors was 8 ml. The
115
resin was rinsed with ultrapure water until the DOC concentration of the column effluent was
116
<0.5 mg l-1. Before and after each fractionation procedure, the resin was rinsed with
117
successive cycles of 3 VV’s of 0.01 M HCl and 3 VV’s of 0.1 M NaOH (both 2 VV at 30 VV
118
h-1 and one VV at 7.5 VV h-1). This rinsing was repeated until the DOC concentration was
119
less than 0.5 mg l-1 and the UV absorbance of the effluent stabilized.
120
The DOM solution was acidified to pH 2 with 2 M HCl and filtrated (0.45 µm). The
121
column was rinsed with 0.01 M HCl until pH was stable. The UV absorbance signal was set
122
to zero. Ten ml of the filtered DOM solution was injected onto the column using a superloop
123
(22.5 VV h-1). After the sample, the column was rinsed with 4.5 VV of 0.01 M HCl (22.5 VV
124
h-1). All DOM collected before the end of this rinsing was classified as hydrophilic. The flow
125
direction was subsequently reversed, and the non-neutral hydrophobic (nn-HO) DOM was
126
eluted with 2.25 VV of 0.1 M NaOH, followed by 4 VV of ultrapure water (7.5 VV h-1). After
127
1 VV, the eluate pH increased from 2 to >10 within 0.1 VV. All eluting solution was collected
128
in 30 subfractions with a volume varying from 2 to 5 ml. All DOM subfractions were
6
129
acidified to pH 2 before DOC analysis and inorganic C was removed by sparging samples
130
with O2 for 30 seconds. Concentrations of DOC in the subfractions were measured by
131
catalytic combustion (800°C) and infrared measurement of formed CO2 (Multi N/C 2100S,
132
AnalytikJena). The DOC present in the neutral hydrophobic fraction (n-HO, not eluted by 0.1
133
M NaOH) was calculated by difference between DOC in the original sample and the DOC
134
quantities in fractions recovered.
135
Test components
136
Fifteen components were subjected to the fractionation procedure (Table 2). The components
137
were selected because of their presence in soil solutions or because they are a model for
138
compound classes present in soil solutions. Molecules were also selected to obtain a range in
139
Kow, acidity constant (pKa) and MW. The components were dissolved in or diluted with
140
ultrapure water to obtain DOC concentrations of approximately 20 mg l-1. Beside fractionation
141
of the isolated components, 3 of them were also fractionated in combination: 11.0 mg DOC l-1
142
oxalic acid, 5.0 mg DOC l-1 rutin and 3.3 mg DOC l-1 geraniol, giving a combined
143
concentration of 19.3 DOC mg l-1. Also a solution of Suwannee River Fulvic Acid (SRFA), a
144
fulvic acid isolated by the International Humic Substances Society (IHSS), was fractionated.
145
After acidifying the DOM samples to pH 2 with 2 M HCl and filtration (0.45 µm), DOC
146
concentrations (AnalytikJena) and UV absorbance (254 nm, path length 1 cm, Perkin-Elmer
147
Lambda 20) were measured. Fractionations were performed in duplicate.
148
Fractionation of soil solutions
149
The topsoil of 4 uncontaminated soils (1-4) was sampled, air-dried (25°C), sieved (< 2 mm)
150
and stored dry for various periods (4-12 years). The air-dried soils were wetted with deionized
151
water to field capacity (pF 2), covered in plastic reservoirs and incubated moist during various
152
days at 21°C (Table 3). Topsoil 5 was not dried after sieving (4 mm) but immediately
7
153
incubated moist during 4 days. After incubation soil solutions were obtained by centrifugation
154
(pore water) or by extraction. Pore water was sampled by the ‘double chamber method’. A 30
155
mL syringe without plunger was filled with quartz wool and approximately 60 g of moist soil.
156
The syringe was placed in a 50 ml centrifuge tube and centrifuged during 30 minutes (2500
157
g). The obtained pore water was filtered (0.45 µm). Extractions were performed in a 50 ml
158
centrifuge tube filled with 5 g moist soil and 20 ml 1 mM CaCl2. After shaking end-over-end
159
for 24 hours, the suspensions were centrifuged during 30 minutes (2500 g). The supernatant
160
was filtered (0.45 µm). Soil solutions were sometimes diluted before fractionation to obtain
161
DOC concentrations below 150 mg l-1. Fractionations were performed in duplicate, except for
162
the pore water of soil 5.
163
Specific UV absorbance
164
The specific UV absorbance (SUVA, l g-1 cm-1) of DOM was calculated by
SUVA 
165
A 254
b  DOC
(1)
166
where A254 is the UV absorbance at 254 nm measured by a spectrophotometer (Perkin-Elmer
167
Lambda 20 or online with GE Healthcare Äkta Prime Plus) (dimensionless), b the path length
168
of the quartz cell of the spectrophotometer in cm and [DOC] is in g l-1. The SUVA of DOM
169
with DOC concentration < 1.5 mg l-1 was not calculated because of accuracy problems. The
170
specific UV-absorbance is used as an estimate of the aromaticity of DOM (Weishaar et al.,
171
2003).
172
The SUVA of a collected subfraction was calculated by equation (1) using the average
173
UV absorbance of the eluting subfraction. The SUVA of the HI or nn-HO fraction (SUVAfr, l
174
g-1 cm-1) was calculated by
8
 A
n
175
SUVA fr 
i 1
n
254
i
 Vi

b   DOCi  Vi 
(2)
i 1
176
where n is the number of subfractions in the fraction fr and A i254 , Vi and [DOC]i the average
177
UV absorbance (-), volume (l) and DOC concentration (g l-1) of subfraction i, respectively.
9
178
Results and discussion
179
Fractionation of the test components
180
The fractionation of the 15 test components varied largely (Table 2). Carbohydrates, proteins
181
and aliphatic LMWA’s showed a (partial) HI character, whereas aromatic LMWA’s and
182
polyphenols ended up in the HO fraction. Quinons and terpenoids did not desorb from the
183
resin during base rinsing and were therefore classified as n-HO molecules. The peak of the
184
DOC concentration of the HI fraction of oxalic acid and other pronounced HI components
185
was situated only after 2 VV with DOC concentrations dropping to zero only after 5 VV
186
(Figure 1). This was also the case for the HI fraction of soil DOM (see below). The DAX-8
187
resin showed affinity for the HI molecules as well, resulting in retardation of the components.
188
This emphasizes the need for sufficiently rinsing of the resin and the importance of reporting
189
cumulative resin/solution ratio in the procedure for the distribution of DOM in HI and HO
190
fractions. When no or few rinsing is applied, part of the retarded HI molecules will be
191
classified as HO. The choice for rinsing with 4.5 VV in this method was based on preliminary
192
fractionation experiments with soil solutions. Concentrations of DOC dropped only below 0.5
193
mg l-1 after rinsing for more than 4 VV with 0.01 M HCl. However 3 of the aliphatic
194
LMWA’s (maleic, succinic and propionic acid) showed an even larger retardation on the
195
column at pH 2 (Figure 1), resulting in a partial HI and partial HO classification based on this
196
fractionation procedure. This can be related to the Kow and pKa of the components (see
197
below). Aromatic LMWA’s, e.g. salicylic acid, were almost completely retained by the resin
198
at pH 2 but were eluted with NaOH (Figure 1). This was also the case for polyphenols (e.g.
199
rutin). Geraniol and menadion were classified as n-HO, but also some HI and especially some
200
nn-HO components showed a partial n-HO character (Table 2). This was most pronounced for
201
components with high molecular weight (e.g. ovalbumin and tannic acid). Peuravuori et al.
202
(1997) suggested already that the sorption mechanism of the DAX/XAD-8 resin is partly
10
203
based on molecular size of the solute. Qualls & Haines (1991) also found rutin in the
204
hydrophobic fraction and oxalic acid and albumin in the hydrophilic fraction. David et al.
205
(1989) also classified oxalic acid as hydrophilic and salicylic acid as hydrophobic, but they
206
found over 89% of succinic acid in the hydrophilic fraction. They used the fractionation
207
method of Leenheer (1981) who only rinsed the XAD-8 resin with 1 BV compared to 4.5 VV
208
(=2.25 BV) in our method. This should lead to a smaller HI DOC amount, however they
209
found higher DOC recovery in their HI fraction compared to this study (59%).
210
Fractionation of the combined components (oxalic acid, rutin and geraniol) gave similar
211
results as their isolated fractionations. The HI and nn-HO fractions were slightly lower (49.5
212
± 1.3% and 16.4% ± 0.1% of total DOC respectively, average values ± standard deviation)
213
compared to the calculated values based on the isolated fractionation results (53.8 ± 1.1% and
214
21.1 ± 1.6%). These lower values can be an indication of minor interactions of the
215
components. The neutral hydrophobic geraniol can interact with oxalic acid and rutin,
216
increasing slightly their n-HO fraction. Almost all DOM of the SRFA appeared in the HO
217
fraction: 76.2 ± 0.4% in the nn-HO and 23.5 ± 0.0% in the n-HO fraction. The negligible part
218
of DOM in the HI fraction (0.3 ± 0.4%) can be explained by the SRFA isolation method of
219
the IHSS. All SRFA DOM originates from desorption from the XAD-8 resin with 0.1 M
220
NaOH after percolating Suwannee River water at pH 2 over the resin. Surprisingly, 23.5% of
221
the sorbed SRFA on the DAX-8 resin could not be desorbed in our method. A possible
222
explanation is the use of different resins: XAD-8 by the IHSS method and DAX-8 in our
223
method. The method to desorb DOM from the resin also varies: 3 VV with 0.1 M NaOH at
224
350 ml min-1 in the IHSS method (Malcolm et al., 1995 and E.M. Perdue, personal
225
communication) compared to 2.25 VV with 0.1 M NaOH and 4 VV with MQ water at 1 ml
226
min-1 in our method. Structural conformations of the isolated fulvic acid during further IHSS
11
227
sample preparation (passing through cation exchange resin, hydrogen saturation and freeze
228
drying) are also possible.
229
The affinity of the DAX-8 and XAD-8 resin for components is assumed to be related to
230
the hydrophobicity of the components. A negative linear relationship was found between the
231
logarithm of aqueous solubility and the logarithm of the capacity factor of molecules on
232
XAD-8 (Thurman et al., 1978). It was investigated if the Kow of the test components, a
233
measure for the hydrophobicity, showed a relationship with the classification of the molecule
234
into the HI, nn-HO and the n-HO fraction (Table 2 and Figure 2). Components with log Kow <
235
-1.5 appeared almost completely in the HI fraction. An exception is rutin which has a low log
236
Kow value (-2.02) because of the carbohydrate part of the molecule. The sorption by the DAX-
237
8 resin at pH 2 is probably due to the aromatic part (polyphenol). This was also the case for
238
tannic acid which has a quite low log Kow (-0.19) as well because of its carbohydrate part.
239
Maleic acid, succinic acid and propionic acid have intermediate Kow values (-1 < log Kow <
240
0.5) and have some affinity for the DAX resin, resulting in a partly HI and partly HO
241
character. Of these 3 molecules maleic acid is less retained, probably because it is charged at
242
pH 2 (pKa1 = 1.83), as is oxalic acid. Succinic acid and propionic acid are not ionized at pH 2
243
(pKa1 = 4.16 and pKa = 4.87 respectively). The Kow of a charged molecule is more than 100
244
times smaller compared to that of the non-dissociated molecule (Schwarzenbach et al., 1993).
245
Propionic acid has the largest log Kow (0.33) and showed most retardation by the resin.
246
Components with log Kow > 0.5 were all classified as HO. Although 2,4,6-trihydroxybenzoic
247
acid is ionized at pH 2 (pKa1 = 1.68) it was also found in the HO fraction. The 2 molecules
248
classified as n-HO, geraniol and menadion, showed a very large log Kow (3.47 and 2.20
249
respectively). Salicylic acid is classified as nn-HO despite its larger log Kow (2.26) compared
250
to menadion. This is probably due to the presence of the carboxyl group with pKa = 2.98 in
251
the salicylic acid molecule. All other molecules classified as nn-HO have also pKa < 11,
12
252
resulting in a charged molecule during rinsing with 0.1 M NaOH causing desorption from the
253
DAX-8 resin.
254
The SUVA of the eluted DOM showed similar values as the SUVA of the DOM
255
measured in the unfractionated DOM solution. Exceptions were tannic acid and especially
256
2,4,6-trihydroxybenzoic acid which showed 25 times larger SUVA values in the 0.1 M NaOH
257
eluate. This was probably because of the deprotonation of the carboxylic groups in the
258
alkaline conditions resulting in resonance and therefore larger UV absorption. The SUVA of a
259
molecule can also be an indicator of the hydrophobic character. Test components with SUVA
260
< 3 l g-1 cm-1 were classified as (partly) hydrophilic while hydrophobic components had
261
SUVA > 3 l g-1 cm-1. Exceptions were maleic acid and geraniol. Maleic acid has a SUVA of
262
12.0 l g-1 cm-1 because of the resonance possibilities of the double bond, but the short chain
263
and the presence of carboxylic groups give the molecule a partial hydrophilic character.
264
Geraniol has a very low SUVA (0.6 l g-1 cm-1) because of the lack of resonance, but the long
265
aliphatic chain renders the molecule hydrophobic. The relationship of SUVA and the
266
retention by the resin can be explained by the presence of aromatic structures in the
267
molecules. Aromaticity increases the hydrophobic character, and SUVA is a good measure for
268
aromaticity (Weishaar et al., 2003).
269
From these relationships it is clear that the affinity of the DAX-8 resin for molecules is
270
primarily based on hydrophobic structures in the components. Aromatic and long aliphatic
271
parts sorb onto the resin at pH 2, even if the molecules have hydrophilic parts (e.g. rutin).
272
Hydrophobic components can be desorbed from the resin if they are ionized during base
273
rinsing. Potentiometric titration of the resin between pH 2 and 13 revealed that no charges are
274
formed on the resin, even after extensive use (details not shown).
275
Fractionation of the soil solutions
13
276
The DOC concentration and SUVA of DOM in the eluate of the DAX-8 column changed
277
during the fractionation of the pore water DOM of the 83 days incubated soil 1 (Figure 3).
278
The DOC concentration and SUVA patterns were similar for the other fractionations. As
279
mentioned before, DOC concentrations of the hydrophilic fraction peaked after 2 VV’s and
280
dropped to zero only after 5 VV’s. This demonstrated the retardation of hydrophilic
281
components by the resin. The SUVA of this DOM increased gradually during elution. This
282
confirms that the hydrophilic fraction has a heterogeneous character and that retardation of
283
hydrophilic DOM increases with aromaticity of the hydrophilic components. The
284
hydrophobic DOC peak also tailed for several VV’s. The hydrophobic DOM is also
285
heterogeneous, demonstrated by the changing SUVA during elution. The first eluted
286
hydrophobic DOM had the highest SUVA values. This was probably because DOM with the
287
highest hydrophobicity and aromaticity was sorbed onto the first DAX-8 particles of the
288
column. This DOM comes first out of the column during base rinsing because the flow
289
direction was inversed after the acid rinsing.
290
Characteristics and fractionation results of all soil solutions are presented in Table 3.
291
Distribution of HI and HO DOC are within the ranges found in other studies (Table 1): 31-
292
72% HI, 25-46% nn-HO and 2-28% n-HO of total DOC. The DOC concentration was lowest
293
in the pore water of soil 5 which was not air-dried but sampled freshly from the field. This
294
DOM had also the smallest fraction of HI DOC. Soils that were stored air-dry for several
295
years and rewetted before incubation showed extremely high DOC concentrations in the pore
296
water (> 1000 mg l-1). This was less expressed in soil 1, the soil that was kept air-dry for the
297
shortest time (4 years). The high amount of DOM probably originated from cell lysis of soil
298
biomass due to soil drying (Merckx et al., 2001). Microbial biomass is assumed to consist of
299
simple molecules like sugars, amino acids and nucleotides (Merckx & Martin, 1987). This
300
was reflected in the relatively low SUVA values (< 11 l g-1 cm-1) and high HI fraction (>
14
301
49%) of the DOM in pore water of shortly incubated soils (3-13 days). The fractionation of
302
the test components showed already the hydrophilic character of carbohydrates and proteins.
303
Dissolved organic matter released by drying-wetting cycles is easily degradable (Merckx et
304
al., 2001; Amery et al., 2007). This was also observed in this study as prolonged incubation
305
of soil 1 resulted in lower DOC concentrations and higher SUVA of DOM (Table 3). The
306
fraction of HO DOC increased during incubation as a result of smaller degradation of the HO
307
fraction (-21%) compared to the HI fraction (-54%). Incubation experiments with extracts of
308
soil and plant material showed that resistance of DOM to microbial degradation is mostly
309
found in the HO DOC fraction (Kalbitz et al., 2003).
310
Extracts had lower DOC concentrations compared to corresponding pore water, but
311
extractions sampled more DOC when expressed in mass DOC extracted per kg soil. As this
312
DOM had higher SUVA compared to pore water, it is assumed that the extra DOM sampled
313
by extraction originates from the solid organic matter which is more humified and has higher
314
aromaticity (Amery et al., 2007). The extracts of soil 2 and 3 contained relatively more HO
315
DOC compared to the pore waters, demonstrating the hydrophobic character of the extra
316
DOM that is released by soil extraction. Again, this shows that soil extraction should be
317
discouraged as it dissolves DOM not present in solution as stated before (Amery et al., 2007).
318
The average SUVA of the HI DOM was lower compared to the SUVA of nn-HO DOM
319
(Table 3). This was expected on the basis of the fractionation of the test components as
320
aromatic compounds ended up in the HO fraction. Studies dealing with forest floor and
321
aquatic DOM showed also higher specific UV absorbance for the HO compared to the HI
322
fraction (Martin-Mousset et al., 1997; Imai et al., 2001; Dilling and Kaiser, 2002; Croué et
323
al., 2003; Simonsson et al., 2005). The SUVA of the HI DOM varied largely from 2 to 17 l g-
324
1
325
(Table 3). Not only the distribution of DOM into the HI, nn-HO and n-HO fractions differed
cm-1 and even overlapped the range of SUVA of the nn-HO DOM, from 14 to 35 l g-1 cm-1
15
326
among the soil solutions, but also the characteristics of one fraction depended on the DOM
327
source. This shows that soil solution DOM is not a varying composition of different fractions
328
with constant properties since there is considerable variation in composition and quality
329
within fractions. The statement that the DAX/XAD-8 fractionation procedure isolates similar
330
groups of compounds (Leenheer, 1981) is not proven. The variation coefficient of the SUVA
331
values of the nn-HO DOM (31%) is only slightly smaller than that of the unfractionated DOM
332
(41%). The variation coefficient of the SUVA of HI DOM is even much larger (72%). The
333
standard deviation of the SUVA of the nn-HO DOM is larger (7.5 l g-1 cm-1) and of the HI
334
DOM is smaller (5.0 l g-1 cm-1) compared to the standard deviation of the SUVA of the bulk
335
DOM (6.0 l g-1 cm-1). This is in contrast with Dilling & Kaiser (2002) who found large
336
similarities in the chemical composition of HI and HO fractions of 3 forest floor extracts
337
despite the large range in distribution within the bulk samples (37-71% HI DOC). As a result,
338
the HO DOC concentration could be estimated well using the SUVA of the bulk sample. The
339
attempt of Simonsson et al. (2005) to estimate HO DOC fractions and concentrations of floor
340
leachates based on UV absorbance (210-300 nm) was, however, not very successful due to
341
non-consistent UV absorptivity of the HI and HO DOM. No significant correlation was found
342
between the SUVA of the bulk soil DOM and the fraction HO DOC in this study. These
343
findings question the practical role of DOM fractionation in predicting DOM properties. We
344
did not found smaller SUVA variation in the HI and nn-HO fraction compared to the
345
unfractionated DOM. Fractionation is probably useful to interpret properties of the DOM such
346
as metal and hydrophobic contaminant affinity, biodegradability etc. Predictions of these
347
characteristics based on fractionation can only be successful if these structural properties are
348
more homogenous within the fractions, however our SUVA data show that the variation
349
coefficient and standard deviation can be even larger within a fraction.
350
Conclusions
16
351
A DAX-8 fractionation procedure for small volumes of DOM solution was developed.
352
Subjecting 15 soil DOM model components to this fractionation method showed that Kow,
353
pKa and SUVA of the components can be used to predict the hydrophobic character of the
354
molecule. Components with aromatic or long aliphatic parts sorb onto the DAX-8 resin even
355
if the molecule has negative log Kow values due to hydrophilic parts. Hydrophilic organic
356
components also showed DAX-8 resin affinity, resulting in retardation on the column.
357
Researchers using a different fractionation method are encouraged to subject DOM model
358
components to their procedure as well. In this way, calibration, understanding of the
359
fractionation mechanism and comparison of methods is possible.
360
Fractionation of 8 different soil solutions resulted in a large variation of the distribution:
361
31-72% HI, 25-46% nn-HO and 2-28% n-HO of total DOC. Soil solution DOM of soils
362
stored air-dry for several years before moist incubation had large HI fractions. Soil solutions
363
sampled by extraction had larger HO fractions compared to isolating soil solution by
364
centrifugation of the same soil. Within one soil solution the SUVA of the hydrophilic DOM
365
was always lower than the SUVA of the hydrophobic DOM. However the SUVA values of
366
the fractions of the different soil solutions were not constant. The standard deviation of the
367
SUVA of the nn-HO DOM was larger than that of the unfractionated DOM and the relative
368
variation in SUVA of the hydrophilic DOM was also larger compared to the unfractionated
369
DOM. Fractionation reveals environmentally relevant information, however it is unlikely to
370
predict structural or functional properties of DOM in isolation.
371
Acknowledgements
372
This research was funded by the Onderzoeksfonds K.U.Leuven under the project number
373
GOA/2006/07-TBA, and was supported by the fund for Scientific Research-Flanders
374
(F.W.O.) through a doctoral fellowship awarded to F. Amery.
17
375
References
376
Amato, M. 1983. Determination of carbon 12C and 14C in plant and soil. Soil Biol. Biochem.,
377
378
15, 611-612.
Amery, F., Degryse, F., Degeling, W., Smolders, E. & Merckx, R. 2007. The copper-
379
mobilizing-potential of dissolved organic matter in soils varies 10-fold depending on soil
380
incubation and extraction procedures. Environmental Science & Technology, 41, 2277-
381
2281.
382
Chow, A.T., Tanji, K.K., Gao, S.D. & Dahlgren, R.A. 2006. Temperature, water content and
383
wet-dry cycle effects on DOC production and carbon mineralization in agricultural peat
384
soils. Soil Biology & Biochemistry, 38, 477-488.
385
Cilenti, A., Provenzano, M.R. & Senesi, N. 2005. Characterization of dissolved organic
386
matter from saline soils by fluorescence spectroscopy. Environmental Chemistry Letters,
387
3, 53-56.
388
Croué, J.-P., Benedetti, M.F., Violleau, D. & Leenheer, J.A. 2003. Characterization and
389
copper binding of humic and nonhumic organic matter isolated from the South Platte
390
River: Evidence for the presence of nitrogenous binding site. Environmental Science &
391
Technology, 37, 328-336.
392
David, M.B., Vance, G.F., Rissing, J.M. & Stevenson, F.J. 1989. Organic carbon fractions in
393
extracts of O and B horizons from a New England Spodosol: Effects of acid treatment.
394
Journal of Environmental Quality, 18, 212-217.
395
Dilling, J.& Kaiser, K. 2002. Estimation of the hydrophobic fraction of dissolved organic
396
matter in water samples using UV photometry. Water Research, 36, 5037-5044.
18
397
Guggenberger, G., Glaser, B. & Zech, W. 1994. Heavy metal binding by hydrophobic and
398
hydrophilic dissolved organic carbon fractions in a spodosol A and B horizon. Water Air
399
and Soil Pollution, 72, 111-127.
400
Guggenberger, G.& Zech, W. 1994. Dissolved organic carbon in forest floor leachates :
401
simple degradation products or humic substances? Science of the Total Environment, 152,
402
37-47.
403
404
405
Kaiser, K. 1998. Fractionation of dissolved organic matter affected by polyvalent metal
cations. Organic Geochemistry, 28, 849-854.
Kaiser, K.& Guggenberger, G. 2005. Storm flow flushing in a structured soil changes the
406
composition of dissolved organic matter leached into the subsoil. Geoderma, 127, 177-
407
187.
408
Kaiser, K., Guggenberger, G., Haumaier, L. & Zech, W. 2001a. Seasonal variations in the
409
chemical composition of dissolved organic matter in organic forest floor layer leachates
410
of old-growth Scots pine (Pinus sylvestris L.) and European beech (Fagus sylvatica L.)
411
stands in northeastern Bavaria, Germany. Biogeochemistry, 55, 103-143.
412
Kaiser, K., Guggenberger, G. & Zech, W. 2001b. Organically bound nutrients in dissolved
413
organic matter fractions in seepage and pore water of weakly developed forest soils. Acta
414
Hydrochimica et Hydrobiologica, 28, 411-419.
415
416
417
Kalbitz, K., Schmerwitz, J., Schwesig, D. & Matzner, E. 2003. Biodegradation of soil-derived
dissolved organic matter as related to its properties. Geoderma, 113, 273-291.
Leenheer, J.A. 1981. Comprehensive approach to preparative isolation and fractionation of
418
dissolved organic carbon from natural waters and wastewaters. Environmental Science &
419
Technology, 15, 578-587.
19
420
421
Leenheer, J.A.& Croué, J.P. 2003. Characterizing aquatic dissolved organic matter.
Environmental Science & Technology, 37, 18A-26A.
422
Leenheer, J.A., Nanny, M.A. & McIntyre, C. 2003. Terpenoids as major precursors of
423
dissolved organic matter in landfill leachates, surface water, and groundwater.
424
Environmental Science & Technology, 37, 2323-2331.
425
Malcolm, R.L., Aiken, G.R., Bowles, E.C. & Malcolm, J.D. 1995. Isolation of fulvic and
426
humic acids from the Suwannee River. In: Humic substances in the Suwannee River,
427
Georgia: Interactions, properties, and proposed structures (eds. Averett, R.C., Leenheer,
428
J.A., McKnight, D.M., Thorn, K.A.), pp. 23-25. United States Government Printing
429
Office, Washington, DC.
430
431
432
433
434
Marhaba, T.F., Pu, Y. & Bengraine, K. 2003. Modified dissolved organic matter fractionation
technique for natural water. Journal of Hazardous Materials, 101, 43-53.
Marschner, B.& Kalbitz, K. 2003. Controls of bioavailability and biodegradability of
dissolved organic matter in soils. Geoderma, 113, 211-235.
Martin-Mousset, B., Croué, J.P., Lefebvre, E. & Legube, B. 1997. Distribution and
435
characterization of dissolved organic matter of surface waters. Water Research, 31, 541-
436
553.
437
Merckx, R., Brans, K. & Smolders, E. 2001. Decomposition of dissolved organic carbon after
438
soil drying and rewetting as an indicator of metal toxicity in soils. Soil Biology &
439
Biochemistry, 33, 235-240.
440
441
Merckx, R.& Martin, J.K. 1987. Extraction of microbial biomass components from
rhizosphere soils. Soil Biology & Biochemistry, 19, 371-376.
20
442
Page, D.W., van Leeuwen, J.A., Spark, K.M. & Mulcahy, D.E. 2002. Pyrolysis
443
characterisation of plant, humus and soil extracts from Australian catchments. Journal of
444
Analytical and Applied Pyrolysis, 65, 269-285.
445
Peuravuori, J., Lehtonen, T. & Pihlaja, K. 2002. Sorption of aquatic humic matter by DAX-8
446
and XAD-8 resins - Comparative study using pyrolysis gas chromatography. Analytica
447
Chimica Acta, 471, 219-226.
448
Peuravuori, J., Pihlaja, K. & Välimäki, N. 1997. Isolation and characterization of natural
449
organic matter from lake water: two different adsorption chromatographic methods.
450
Environment International, 23, 453-464.
451
Qualls, R.G.& Haines, B.L. 1991. Geochemistry of dissolved organic nutrients in water
452
percolating through a forest ecosystem. Soil Science Society of America Journal, 55,
453
1112-1123.
454
Raber, B.& Kögel-Knabner, I. 1997. Influence of origin and properties of dissolved organic
455
matter on the partition of polycyclic aromatic hydrocarbons (PAHs). European Journal of
456
Soil Science, 48, 443-455.
457
458
459
Raber, B., Kögel-Knabner, I., Stein, C. & Klem, D. 1998. Partitioning of polycyclic aromatic
hydrocarbons to dissolved organic matter from different soils. Chemosphere, 36, 79-97.
Rasmussen, C.B., Henriksen, A., Abelskov, A.K., Jensen, R.B., Rasmussen, S.K., Hejgaard, J.
460
& Welinder, K.G. 1997. Purification, characterization and stability of barley grain
461
peroxidase BP 1, a new type of plant peroxidase. Physiologia Plantarum, 101, 247.
462
Schulze, W.X. 2005. Protein analysis in dissolved organic matter: what proteins from organic
463
debris, soil leachate and surface water can tell us - a perspective. Biogeosciences, 2, 75-
464
86.
21
465
466
Schwarzenbach, R.P., Gschwend, P.M. & Imboden, D.M. 1993. Environmental organic
chemistry. Wiley-Interscience, United States of America.
467
Simonsson, M., Kaiser, K., Danielsson, R., Andreux, F. & Ranger, J. 2005. Estimating nitrate,
468
dissolved organic carbon and DOC fractions in forest floor leachates using ultraviolet
469
absorbance spectra and multivariate analysis. Geoderma, 124, 157-168.
470
471
472
473
474
475
476
Stevenson, F.J. 1994. Humus chemistry. Genesis, composition, reactions. United States of
America.
Strobel, B.W. 2001. Influence of vegetation on low-molecular-weight carboxylic acids in soil
solution - a review. Geoderma, 99, 169-198.
Thurman, E.M.& Malcolm, R.L. 1981. Preparative isolation of aquatic humic substances.
Environmental Science & Technology, 15, 463-466.
Thurman, E.M., Malcolm, R.L. & Aiken, G.R. 1978. Prediction of capacity factors for
477
aqueous organic solutes adsorbed on a porous acrylic resin. Analytical Chemistry, 50 ,
478
775-779.
479
Tipping, E., Woof, C., Rigg, E., Harrison, A.F., Ineson, P., Taylor, K., Benham, D., Poskitt,
480
J., Rowland, A.P., Bol, R. & Harkness, D.D. 1999. Climatic influences on the leaching
481
of dissolved organic matter from upland UK moorland soils, investigated by a field
482
manipulation experiment. Environment International, 25, 83-95.
483
Van Zomeren, A.& Comans, R.N.J. 2007. Measurement of humic and fulvic acid
484
concentrations and dissolution properties by a rapid batch procedure. Environmental
485
Science & Technology, 41, 6755-6761.
486
Weast, R.C.& Astle, M.J. 1980. CRC Handbook of chemistry and physics. CRC, Florida.
487
Weishaar, J.L., Aiken, G.R., Bergamaschi, B.A., Fram, M.S., Fujii, R. & Mopper, K. 2003.
488
Evaluation of specific ultraviolet absorbance as an indicator of the chemical composition
22
489
and reactivity of dissolved organic carbon. Environmental Science & Technology, 37,
490
4702-4708.
23
491
FIGURE CAPTIONS
492
Figure 1 Average DOC concentration (mg l-1) in DAX-8 column eluates versus elution
493
volume (VV) of fractionation of oxalic acid (■), maleic acid (◊) and salicylic acid (∆).
494
Standard deviations are represented by error bars (2 replicates). Left from the dashed line
495
eluates the hydrophilic DOM, right the hydrophobic DOM.
496
497
Figure 2 Average percentage hydrophilic (HI) DOC of total DOC versus log Kow of the
498
test components (no Kow data for starch and albumin, value of log Kow = -0.48 used for
499
maleic acid). Standard deviations are represented by error bars (2 replicates).
500
501
Figure 3 Average DOC concentration (□) and SUVA (■) of the subfractions of the eluent
502
of the DAX-8 column versus elution volume (VV) of the fractionation of the ½ diluted
503
pore water of soil 1 incubated for 83 days. Standard deviations are represented by error
504
bars (2 replicates). Left from the dashed line eluates the hydrophilic DOM, right the
505
hydrophobic DOM.
24
FIGURE 1
25
Hydrophilic DOM
-1
DOC concentration /mg l
506
Hydrophobic DOM
20
15
oxalic acid
10
maleic acid
salicylic acid
5
0
0
2
4
6
8
Elution volume /VV
10
12
25
FIGURE 2
100
80
%HI
507
60
40
rutin
20
tannic acid
0
-4
-2
0
2
4
log Kow
26
FIGURE 3
25
Hydrophilic DOM
-1
70
Hydrophobic DOM
60
-1
20
15
40
10
30
20
5
-1
50
SUVA /l g cm
DOC-concentration /mg l
508
10
0
0
0
2
4
6
8
Elution volume /VV
10
12
27
509
TABLES
510
Table 1 Reported distributions of soil solution DOM into the HI and HO fraction (% of total
511
DOC) using XAD-8 resin.
Soil solution
Column leachate O+A horizon of
Spodosol
Column leachate O+A+B horizon
of Spodosol
Water extract Ap horizon of Haplic
Luvisol
Water extract O horizon of acidic
forest soil
Water extract non-saline soil
Water extracts 2 saline soils
Pore water A horizon of 7
agricultural soils
Water extract A horizon of
agricultural soil
Leachate A horizon of forest soil
Pore water 90 cm depth of Arenosol
with Scots pine
Pore water 90 cm depth of Leptosol
with European beech
Lysimeter leachate of brown earth
Lysimeter leachate of micropodzol
Lysimeter leachate of peaty gley
Water extract A horizon of 3
agricultural soils
Water extract 3 agricultural topsoils
in a fen area
Water extract forest topsoil in a fen
area
Extract surface oxidized peat soil
Extract subsurface reduced peat soil
[DOC]
/mg l-1
58.6
%HI
/%
41
%HO
/%
68
Reference
4.7
66
36
Guggenberger et al. (1994)
25.1
48
53
Raber & Kögel-Knabner (1997)
100.8
37
63
Kaiser (1998)
38-66
73
75-89
41-67
27
25-11
33-59
Cilenti et al. (2005)
Cilenti et al. (2005)
Raber et al. (1998)
22
69
33
Raber et al. (1998)
12
9.3-37.2
40
47-53
50
47-53
Qualls & Haines (1991)
Kaiser et al. (2001b)
6.4-35.6
63-90
10-37
Kaiser et al. (2001b)
variable
variable
variable
7.2-13.9
56
87
31
45-55
44
13
69
45-55
Tipping et al. (1999)
Tipping et al. (1999)
Tipping et al. (1999)
Kalbitz et al. (2003)
18.4-40.5
33-52
48-67
Kalbitz et al. (2003)
75.7
25
75
Kalbitz et al. (2003)
296
21
43
64
57
36
Chow et al. (2006)
Chow et al. (2006)
Guggenberger et al. (1994)
28
Table 2 Test components of the fractionation procedure with chemical characteristics (octanol water partition coefficient Kow, acid
constants pKa, molecular weight MW, SUVA at pH 2), average percentage of total DOC in the hydrophilic (HI), non-neutral
hydrophobic (nn-HO) and neutral hydrophobic (n-HO) fraction (2 replicates, standard deviation in parentheses) and classification of
the component according to the fractionation in isolation. Each component or class of the component has been identified in natural
DOM according to the corresponding reference.
reference
component
class
log Kow
pKa1a
pKa2a
Strobel (2001)
Strobel (2001)
Strobel (2001)
Strobel (2001)
Strobel (2001)
Strobel (2001)
Strobel (2001)
Oxalic acid
Maleic acid
Succinic acid
Propionic acid
Phthalic acid
Salicylic acid
2,4,6-trihydroxybenzoic acid
Glucose
Aliphatic LMWA
Aliphatic LMWA
Aliphatic LMWA
Aliphatic LMWA
Aromatic LMWA
Aromatic LMWA
Aromatic LMWA
-1.74b
-0.48c;0.46c
-0.59c
0.33c
0.73c
2.26c
1.10b
1.23
1.83
4.16
4.87
2.89
2.98
1.68
Carbohydrates
-3.24c
Starch
Carbohydrates
Ovalbumin
Geraniol
Rutin
Kaiser &
Guggenberger (2005)
Kaiser &
Guggenberger (2005)
Schulze (2005)
Leenheer et al. (2003)
Qualls & Haines
(1991)
Page et al. (2002)
Kaiser &
Guggenberger (2005)
Stevenson (1994)
a
Tannic acid
Coniferyl alcohol
(subunit lignin)
Menadion
SUVApH=2
/L g-1 cm-1
2.1
12.0
0.3
0.4
13.7
6.1
4.1
%HI
/%
93 (1)
63 (0)
58 (2)
25 (2)
1 (1)
0 (0)
0 (0)
%nn-HO
/%
2 (0)
29 (2)
42 (1)
65 (1)
105 (16)
90 (2)
85 (4)
%n-HO
/%
5 (1)
8 (2)
0 (3)
10 (2)
10 (2)
15 (4)
classif.
4.19
6.07
5.61
5.51
13.6
-
MW
/g mol-1
90
116
118
74
166
138
170
-
-
180
0.2
97 (0)
2 (1)
1 (1)
HI
-
-
-
105-108
1.2
86 (0)
5 (1)
9 (1)
HI
Proteins
Terpenoids
Polyphenols
(flavonoids)
Polyphenols
Polyphenols
3.47c
-2.02b
8-10.5d
-
~45,000
154
610
1.1
0.6
55.9
73 (3)
3 (4)
1 (0)
2 (-)
6 (7)
73 (4)
25 (3)
91 (7)
26 (5)
HI
n-HO
nn-HO
-0.19b
1.19b
8-10.5d
9.4e
-
1701
180
51.7
30.2
1 (1)
1 (1)
68 (4)
78 (1)
31 (4)
22 (1)
nn-HO
nn-HO
Quinons
2.20c
HI
HI/nn-HO
HI/nn-HO
HI/nn-HO
nn-HO
nn-HO
nn-HO
172
101.1
0 (-)
7 (-)
93 (-)
n-HO
Weast & Astle (1980); EPISuite: estimate based on chemical structure with KOWWIN v1.67; c Experimental Kow database in EPISuite v3.20; d Not measured,
b
pKa estimations based on pKa of phenols; e Rasmussen et al. (1997).
29
Table 3 Characteristics of the soils used for soil solution sampling, the incubation time (IT), and the average percentage of total DOC in the soil
solution and SUVA (L g-1 cm-1) of DOM in the hydrophilic (HI), non-neutral hydrophobic (nn-HO) and neutral hydrophobic (n-HO) fraction (2
replicates, standard deviation in parentheses).
Soil
Origin
pHa
Org. Cb
1
Boris (Denmark)
5.6
1.3
2
Ochamps (Belgium)
5.8
5.7
3
Rhydtalog (U.K.)
4.2
12.9
4
5
Zegveld (The Netherlands)
Ter Munck (Belgium)
4.7
6.9
23.3
0.9
a
IT
/days
3
83
13
13
13
13
13
4 (fresh)
Soil solution
Pore water
Pore water
Pore water
Extract
Pore water
Extract
Pore water
Pore water
DOC
/mg l-1
77.8
42.8
1192
102.9
1335
111.0
1152
26.4
SUVA
/L g-1 cm-1
10.9
22.9
9.9
19.4
7.0
17.9
7.0
20.1
/%
72 (5)
60 (0)
49 (1)
33 (1)
52 (0)
36 (1)
61 (2)
31 (-)
%HI
SUVA
5 (0)
17 (0)
3 (0)
8 (0)
2 (0)
6 (0)
2 (0)
12 (-)
%nn-HO
/%
/SUVA
25 (5)
23 (1)
38 (1)
31 (1)
38 (1)
18 (0)
39 (1)
35 (0)
36 (1)
14 (1)
39 (2)
33 (0)
27 (0)
16 (0)
46 (-)
26 (-)
%n-HO
/%
3 (7)
2 (1)
12 (1)
28 (2)
13 (0)
25 (3)
12 (2)
23 (-)
Measured in 0.01 M CaCl2; b Organic C content was measured by a wet combustion technique using K 2Cr2O7 and H2SO4 followed by CO2 analysis (Amato, 1983).
30