Post on 14-May-2020
transcript
www.elsevier.com/locate/scitotenv
Science of the Total Environm
Theoretical study on fulvic acid structure,
conformation and aggregation
A molecular modelling approach
R.A. Alvarez-Pueblaa, C. Valenzuela-Calahorrob, J.J. Garridoa,*
aDepartment of Applied Chemistry, Public University of Navarra, Campus Arrosadı́a, E-31006 Pamplona, SpainbDepartment of Inorganic Chemistry, Faculty of Pharmacy, University of Granada, E-18071 Granada, Spain
Received 17 February 2003; received in revised form 29 October 2004; accepted 12 November 2004
Available online 26 May 2005
Abstract
The ubiquitous presence of humic substances (HS), combined with their ability to provide multiple sites for chemical
reaction, makes them relevant to numerous biogeochemical processes such as mineral weathering, nutrient bioavailability,
and contaminant transport. The reactivity of HS depends on their functional group chemistry and microstructure, which
are in turn influenced by the composition of the surrounding media. In order to help towards an understanding of
structure conformations and aggregation process of HS in soils and waters and to get a better knowledge of these kinds of
materials, a fulvic acid (FA) has been modelled as a function of its ionic state under different conditions. Our proposed
theoretical model based on the Temple-Northeastern-Birmingham (TNB) monomer fits well with experimental observations
on the solubility (dipolar moment) and electronic and vibrational spectra of FAs. The presence of water molecules has a
great stabilization effect on the electrostatic energy; this effect is greater as ionized rate increases. In vacuum, the non-
ionized aggregated species are more stable than monomers because of the increase in their interaction due to H-bonding
and non-bonding forces. When the molecules are ionized, no aggregation process takes place. In solution, the FA
concentration is a critical factor for the aggregation. The system containing two FA molecules probably did not form
aggregates because its equivalent concentration was too low. When the concentration was increased, the system gave rise
to the formation of aggregates. The ionic state is another critical factor in the aggregation process. The ionized FA has a
0048-9697/$ - s
doi:10.1016/j.sc
Abbreviation
optimized poten
annealing; SE, s
UV–vis, ultravio
* Correspondi
E-mail addre
ent 358 (2006) 243–254
ee front matter D 2005 Elsevier B.V. All rights reserved.
itotenv.2004.11.026
s: FA, fulvic acid; FTIR, Fourier transform infrared spectroscopy; HS, humic substances; MM, molecular mechanics; OPLS,
tials for liquid simulations; PM3/tm, parameterization model 3/transition metals; QD, Quenched Dynamic; SA, simulated
emi-empirical; TIP3P, transferable intermolecular potentials with three point charges; TNB, Temple-Northeastern-Birmingham;
let–visible spectroscopy; ZINDO/S, Zerner’s intermediate neglect differential overlap; MD, molecular dynamics.
ng author. Tel.: +34 948 169601; fax: +34 948 169606.
ss: j.garrido@unavarra.es (J.J. Garrido).
R.A. Alvarez-Puebla et al. / Science of the Total Environment 358 (2006) 243–254244
higher electric negative charge, which increases the energetic barriers and inhibits the approximation of FA caused by the
Brownian movement.
D 2005 Elsevier B.V. All rights reserved.
Keywords: Fulvic acid; Structure; Conformation; Aggregation; Molecular modelling; Molecular dynamics; Electronic spectrum; Vibrational
spectrum
Table 1
Experimental and theoretical selected chemical properties for FA
Experimental data FA (I)
C (%) 40.1 42.26
H (%) 3.57 3.16
N (%) 0.67 1.33
S (%) 0.65 3.05
O (%) 55.0 50.2
Formula C35H38N1S0,3O36 C37H33N1S1O33
Carboxylic groups 6 6
Phenolic groups 3 4
hMwia (g mol�1) 1,058 1,051
a Average molecular weight estimated from Mw=3.99q280+490
according to Chin et al. (1994).
1. Introduction
Humic substances (HS) can affect soil fertility,
mineral weathering, and water acidity; they are
involved in the transport, sequestration, and mitiga-
tion of contaminants; and may even have an impact on
atmospheric chemistry through the carbon cycle, in
which carbon is constantly recycled among plants,
animals, soil, air, and water (Stevenson, 1982).
A special characteristic of HS is its capacity to
show spontaneous changes in their conformation and
aggregation state as a function of solution conditions
like pH and ionic strength (Senesi, 1999). The eluci-
dation of the aggregation mechanism of HS is impor-
tant because these processes have a great influence on
their interaction with nutrients and contaminants in
soils and waters (Wershaw, 1999; Alvarez-Puebla et
al., 2004a). The aggregation process has been studied
using various techniques such as ultraviolet–visible
spectroscopy (Senesi, 1999), cross-polarization
magic-angle-spinning nuclear magnetic resonance
(Tombacz, 1999), light and X-ray scattering (Chin
et al., 1998; Manning et al., 2000) microscopy
(Senesi et al., 1997; Myneni et al., 1999; Alvarez-
Puebla et al., 2004b), size exclusion chromatography
(Swift, 1999) and dialysis and ultrafiltration (Jones
and Bryan, 1998). However, little direct testing of
this phenomenon has been carried out.
Systematic studies of the structures of these sub-
stances are the first step towards understanding how
they interact with other elements and compounds.
Such knowledge will be necessary to predict and
control the impact of chemical and biological
changes in the environment. Developments in soft-
ware and hardware permit an advance of theoretical
approximations in the study of FA aggregation pro-
cesses using molecular modelling (Bruccoleri et al.,
2001). In recent years the number of papers in which
molecular modelling is used to study HS structures
has increased (Schulten, 1995a,b; Sein et al., 1999).
This technique permits the study of physical, chemi-
cal (Schulten and Leinweber, 2000) and electronic
properties (Bruccoleri et al., 2001) at the molecular
level and some interaction mechanisms between HS
and other molecules present in the environment:
mineral phases (Schulten and Schnitzer, 1997; Shev-
chenko et al., 1998), pesticides (Kubicki and Apitz,
1999; Schulten, 1999) and transition metals (Davies
et al., 1997; Kubicki et al., 1999).
Most of the published models have been carried
out under vacuum conditions. This approximation has
been widespread (Schulten, 1995a,b, 1998; Schulten
and Schnitzer, 1997; Shevchenko et al., 1998; Sein et
al., 1999). However, some authors have suggested that
this approximation is unreliable because the natural
media of HS is aqueous solution (Kubicki, 2000;
Schulten and Leinweber, 2000; Bruccoleri et al.,
2001). Therefore, studies have been performed in
which the HS has been solvated with a small number
of water molecules in order to simulate the hydration
effect on their geometry and stability (Davies et al.,
1997; Schulten, 1999; Kubicki, 2000).
The aims of this research are: (i) to propose a
model for FA based on experimental results (ele-
R.A. Alvarez-Puebla et al. / Science of the Total Environment 358 (2006) 243–254 245
mental composition, number of acidic groups and
hMwi); (ii) to validate the proposed model against
the experimental UV–Vis and IR spectra; and, (iii) to
use computational techniques to study the structure,
conformations and aggregation of the FA model as a
Fig. 1. Molecular modelling methodology for the TNB modified FA (I): (i
conformational space exploration (graph); and, (iv) geometry optimization
function of the ionic state in various modelling
conditions. In order to reach these aims the FA
fraction of a commercial HS was extracted, purified
and characterized. Computational techniques were
used to propose the FA model based on the TNB
) 3D structure (II); (ii) geometry optimization with OPLS (III); (iii)
with OPLS of the structure number 7 (IV).
R.A. Alvarez-Puebla et al. / Science of the Total Environment 358 (2006) 243–254246
model (Sein et al., 1999) and additional character-
ization data. This proposed FA model was used to
study the structure, conformations and aggregation
process.
2. Experimental
2.1. Extraction, purification and characterization of
the FA
The FAwas fractionated from a commercial HS by
Acros Organics (Geel, Belgium) by adjusting the pH
of a 40 g L�1 HS solution to 1.0. The obtained FA
was purified using a XAD-8 resin column, converted
to the protonated form by passing it through a proton-
saturated resin and freeze-dried, in accordance with
Fig. 2. Structures obtained by geometry optimization with PM3/TM, for th
with the carboxylic and phenolic groups ionized (VII).
the procedure proposed by the International Humic
Substance Society (IHSS) (Swift, 1996). The C, H, N
and S contents were determined by an elemental
analyser CHNS EA1108 (Carlo Erba Milan, Italy).
The UV–Vis spectrum was recorded by a Lambda 3B
spectrophotometer (Perkin Elmer, Norwalk, USA) in
the range from 900 to 200 nm in accord with Chen et
al. (1977). The IR spectrum was recorded by a FTIR
Avatar 360 spectrometer (Nicolet, Madison, USA)
co-adding up to 200 scans with 4 cm�1 of resolution
(Niemeyer et al., 1992). The COOH groups and the
total acidity of FA were determined by calcium acet-
ate and barium hydroxide methods (Stevenson,
1982), respectively. The phenolic acidic groups
were calculated as the difference between the total
acidity and that of the carboxylic acidic groups. The
acid–base constants were estimated from the end
e non-ionized FA (V), with the carboxylic groups ionized (VI) and
R.A. Alvarez-Puebla et al. / Science of the Total Environment 358 (2006) 243–254 247
points obtained from a potentiometric acid–base titra-
tion in a Metrohm Titrino 702SM autoburette.
2.2. Molecular modelling of the FA in vacuum
conditions
Molecular modelling was carried out with
HyperChem 7.01 Software (Hypercube, 2002a). The
model design was based upon the TNB model (Sein et
al., 1999), in accordance with the elemental composi-
tion, number of acidic groups and hMwi (Table 1). Themodified TNB model (I) (Fig. 1) was optimized by
using the OPLS force field (Jorgensen and Tirado-
Rives, 1988) with the Polak-Ribiere algorithm. In the
present study, the convergence limit was set by a max-
imum acceptable gradient of 0.042 kJ mol�1 nm�1.
Quenched Dynamic (QD) cycles at 700 K with a step
size of 1 fs, were performed on the optimized structure
(III) in order to explore the conformational space
(Balbuena and Seminario, 1999). The most stable
local minima were re-optimized in the same conditions
and the most stable minimum (IV) was optimized by
using the SE method PM3/TM (Hypercube, 2002b)
with a convergence limit of 0.418 kJ mol�1 nm�1
(Young, 2001) (Table 2). For the modelling of the FA
as a function of the ionic state, the carboxylic groups
from structure (I) (Fig. 1) were deprotonated first fol-
lowed by the phenolic groups. The procedure followed
for the molecular modelling was equal to that which
was followed for the protonated FA.
The electronic and vibrational spectra were calcu-
lated on structure (V) (Fig. 2). The electronic spec-
trum was calculated by using the microstates method
(Merchan et al., 1998) carrying out a single point
calculation with ZINDO/S SE method (Zerner,
1991) using the configuration interaction. The vibra-
tional spectrum was calculated by solving the Hessian
matrix formed by the second derivatives of the energy
with respect to atomic Cartesian coordinates by using
the PM3/TM method (Seeger et al., 1991).
Every proposed molecular model for FA should be
taken as the representation of a small portion of the real
mix (Bruccoleri et al., 2001). Since this mix exhibits a
near continuum of exemplars of the typical organic
functional groups it is possible to perform a Gaussian
distribution around every theoretical absorption line,
convoluting the obtained peaks to get a theoretical
spectrum with a continuous absorption in a similar
way as is obtained in an experimental spectrum. The
theoretical electronic and vibrational spectra bands
were built by entering the position and intensity of
the most relevant absorption lines in the Gaussian
distribution equation
y ¼ a0e�1
2
�x�a1a2
�2h i
where a0, a1 and a2 are the intensity, position and band
broadness, respectively; a0 and a1 were entered while
a2 was allowed to vary until it formed a continuous
spectrum due to band overlapping. This convolution
procedure was carried out using a Fourier algorithm
with the PeakFit software (SSPS, 1995) for analysis
and deconvolution.
2.3. Molecular modelling of the FA in aqueous
solution and its aggregation process
For molecular modelling in an aqueous media, a
91,125-nm3 cubic box was built with approximately
2950 TIP3P water molecules (Jorgensen et al., 1983),
enough to avoid the severe edge effects. The geometry
was optimized by using the OPLS force field with
convergence limit of 0.42 kJ mol�1 nm�1. Simulated
annealing (SA) cycles from 700 K to 298 K with a step
size of 1 fs, and a new geometry optimization were
performed on the optimized structure. Aggregation was
studied through the docking of two, four and eight
molecules with the same ionic state, applying SA and
constant temperature molecular dynamics simulations
at 298 K.
3. Results and discussion
3.1. FA structure as a function of ionic state
Fig. 2 shows the results obtained for the FA struc-
ture, simulating its behaviour at a pH ranging from
3.41 to 941 in accordance with the end points
obtained from the titration curve, by geometry opti-
mization with the PM3/TM method (Kubicki, 2000).
The conformation for the protonated FA (V) folds
itself over, maximising Van der Waals, electrostatic
and H-bonding energetic terms. The H-bonds show an
Table 3
Calculated properties for FA model in vacuum by the SE method
PM3/TM as a function of the ionic state
V VI VII
Charge 0 6 10
DHf (kJ mol�1) �5.05d 103 �4.04d 103 �7.43d 102Gradient
(kJ mol�1 nm�1)
1.55 1.94 2.15
l (D) 8.12 16.8 19.9
R.A. Alvarez-Puebla et al. / Science of the Total Environment 358 (2006) 243–254248
extraordinary structural flexibility and variability, act-
ing in a similar way as in Schulten and Leinweber
(2000). As the carboxylic (VI) and phenolic groups
(VII) become ionized, the molecule tends to expand
because of the electrostatic repulsion generated by the
charge increment. These results agree with the con-
clusions obtained from the application of the fractal
theory to the aggregation study (Senesi, 1999), since
the charge rises as the ionization increases and, in
consequence, the inter- and intra-molecular repulsion
increases.
Some of the calculated properties for the FA
model as a function of ionic state are shown in
Table 3. The formation enthalpy, DHf, gives informa-
tion about the conformational stability in the model-
ling conditions. An increase in this parameter with
the ionic state means that the deprotonation global
process is endothermic. The dipolar moment, l, andmolecule polarity are related. As the ionic state
increases, the charge increases, and so does the dipo-
lar moment. This increase means that the conforma-
tion has a greater tendency to solvate when it is
dispersed in a polar solvent. In the case of FA,
which is in continuous contact with aqueous solu-
tions in the environment, the dipolar moment increase
involves a solubility increase or a colloidal stabiliza-
tion in solution.
3.2. Computational model validation
The electronic spectrum simulation of the structure
(V) yielded 26 theoretical absorption lines, where
only 8 of them had relevant absorption intensity.
The theoretical peaks, obtained from Gaussian dis-
Table 2
Energy and gradient values obtained for the molecular modelling of
the FA
Structure I II III IV
Ebonda 6.53d 104 4.04d 102 7.24 8.99
Eanglea 7.66d 103 79.1 33.1 35.4
Edihedrala 66.1 28.4 195 178
EVan der Waalsa 4.18d 1021 3.41d 107 �29.3 �63.6
Eelectrostatica 0 0 33.9 �14.8
ETotala 4.18d 1021 3.41d 107 240 144
Gradientb 8.91d 1023 3.26d 108 0.042 0.042
a kJ mol�1.b kJ mol�1 nm�1.
tribution over each of the theoretical lines and the
spectrum obtained by peak convolution show close
agreement with the experimental spectrum (Fig. 3a).
The experimental spectrum has a maximum at 231
nm, which is located at 249 nm in the theoretical
spectrum. In the region between 360 and 900 nm,
spectra differ slightly, probably due to the presence of
other atomic groups that have not been considered in
this model.
The vibrational spectrum simulation of structure
(V) resulted in 152 theoretical absorption lines,
where 47 of them had relevant absorption intensity.
Absorption at negative wavenumbers was not found,
confirming the energetic minimum of (V) (Kubicki et
al., 1999). Fig. 3b shows the experimental and theo-
retical spectra and the peaks found by the Gaussian
distribution noted over each absorption line. The
values calculated by the PM3/TM method show a
close correspondence to the experimental values.
The error percentage in the wavenumber of every
band does not exceed 10% in any case. The method
predicts O–H and N–H stretches in the interval from
3157 to 3954 cm�1. The C–H stretches of aromatic
carbons and symmetric and asymmetric aliphatic car-
bons have higher wavenumbers: 3096, 3035 and 2975
cm�1, respectively. The two bands at 2719 and 2521
cm�1 are due to the O–H stretches for H-bonds,
which agrees with Davies et al. (1997). The band at
2418 cm�1 is due to the N–H deformations of the
amine group. The carbonyl bands, from acidic groups,
esters, and ketones, and those of aromatic and alkene
C–H deformation, are shifted about 150–200 cm�1,
which is in line with the usual error of the method.
The band at 1408 cm�1 is caused by the C–O defor-
mations, which are experimentally located at approxi-
mately 1200 cm�1. From this point onward, there is a
multitude of low intensity bands caused by bending
and rotational modes of the molecule.
Fig. 3. Experimental and theoretical spectra and theoretical bands calculated using a Gaussian algorithm around the theoretical absorption lines,
(a) electronic and (b) vibrational spectra.
R.A. Alvarez-Puebla et al. / Science of the Total Environment 358 (2006) 243–254 249
Table 4
Energy and gradient values obtained for the FA model in vacuum and solution as a function of the ionic state
COOH/OH COO�/OH COO�/O�
Solution Vacuum Solution Vacuum Solution Vacuum
Ebonda 7.11 8.996 13.43 28.58 20.59 52.26
Eanglea 36.32 35.44 72.51 79.20 86.44 105.9
Edihedrala 220.0 177.8 197.7 184.5 205.5 201.9
EVan der Waalsa 93.01 �63.60 49.62 52.84 162.8 34.60
Eelectrostatica �837.5 �127.6 �1,961 1,758 �2,466 3848
ETotala �481.1 31.06 �1,628 2,103 �1,990 4243
Gradientb 0.042 0.042 0.042 0.042 0.042 0.042
H2Oc 14 – 32 – 48 –
a kJ mol�1.b kJ mol�1 nm�1.c Number of H2O molecules in the 1st hydration sphere.
R.A. Alvarez-Puebla et al. / Science of the Total Environment 358 (2006) 243–254250
3.3. Modelling in solution
The properties shown in Table 4 are determined by
a single point calculation with OPLS on the models
Fig. 4. (a) Aggregation process in vacuum for non-ionized FA. Variation
molecules for: (b) non-ionized FA and (c) with the carboxylic groups ion
obtained by simulation in vacuum and in solution.
The water retention in the first hydration sphere
increases with the ionic state in the FA. This increase
in electrostatic retention of water involves a change in
of the potential energy as a function of the distance between the
ized and with the carboxylic and phenolic groups ionized.
Fig. 5. Aggregation in aqueous solution for a system containing (a) two; (b) four; (c) eight molecules after 12.34 ps of constant temperature MD
simulation; and, (d) eight molecules after 83.5 ps of constant temperature MD simulation.
R.A. Alvarez-Puebla et al. / Science of the Total Environment 358 (2006) 243–254 251
R.A. Alvarez-Puebla et al. / Science of the Total Environment 358 (2006) 243–254252
the molecule conformation and more negative poten-
tial energy values. The variation of the total energy
values shows that the presence of water molecules has
a great stabilization effect on the conformations due to
the great decrease in electrostatic energy, which
becomes more important as the ionic state rises. The
conformation stability in aqueous solution increases
as the ionic state rises, which is in line with the
determined values for the dipolar moment (Table 3).
3.4. Aggregation of FA in vacuum and solution
Fig. 4 shows the initial and final systems for two
FA molecules placed in vacuum at a distance of 1.600
nm without ionization. The non-ionized FA
approached to form an aggregate (Fig. 4a). Stabiliza-
tion is due to the interaction of both molecules by
weak-bonding forces and the formation of four inter-
molecular H-bonds, which makes the system much
more stable when it is aggregated than when it is
separated. The variation of the potential energy, V,
with respect to distance was tracked during the SA
cycles. For the non-ionized FA (Fig. 4b), the decrease
in potential energy, V, can be divided into 5 seg-
ments: (i) from 1.6 to 1.5 nm, the curve shows a
steep slope as a result of the increase in electrostatic
attraction as the distance decreases; (ii) between 1.5
and 1.2 nm, the decrease of the slope shows an
endothermic conformational change in both mole-
cules as a result of their approach; (iii) from 1.2 to
0.95 nm, the curve increases its slope again due to
the electrostatic attraction; (iv) from 0.95 to 0.52 nm,
the slope decreases as a result of a new conforma-
tional change; and, (v) from 0.52 to 0.22 nm, the
slope increases due to the formation of intermolecu-
lar H-bonds. The ionized FA shows a wide separa-
tion because of the great charge they support. The
separation increases with the charge values and the
potential energy decreases as the distance between
the molecules increases due to the decrease of repul-
sive electrostatic interaction (Fig. 4c).
The simulation of the aggregation process in solu-
tion is shown in Fig. 5. The system containing two
FA molecules (an equivalent concentration in FA of
0.035 M) did not lead to any aggregation process
(Fig. 5a). Both molecules showed conformational
changes in order to increase their electrostatic inter-
action with water molecules. When the FA concen-
tration was increased to 0.070 M (4 FA Molecules)
the aggregation process began (Fig. 5b). After 19.21
ps of constant temperature molecular dynamics simu-
lation, three of the four molecules of FA came closer
so as to form a single particle where the different FA
molecules were linked by means of intermolecular
H-bonds (Fig. 5c). Although the simulation contin-
ued until 40 ps, one of the four molecules of FA
remained stable in solution. The behaviour of the
system containing eight molecules of FA (an equiva-
lent concentration of 0.140 M) was similar. At a
simulation time of 12.34 ps, two particles composed
by two and three FA molecules were formed. This
system progressed and gave rise to a compact parti-
cle formed by five FA at a time of 83.49 ps of
simulation (Fig. 5d). Nevertheless, although the
simulation continued until 150 ps, three molecules
of FA remained stable in solution. Systems contain-
ing the same equivalent concentration of ionized FA
did not lead to any aggregation processes.
The behaviour of these systems shows that in
solution the FA concentration is a critical factor for
the aggregation. The system containing two FA mole-
cules probably did not form aggregates because its
equivalent concentration was too low (Wershaw,
1999). When the concentration was increased, the
system gave rise to the formation of aggregates.
Nevertheless, in all the systems there were stable FA
molecules in solution. This can be explained by the
aggregation of some of the system molecules decreas-
ing the effective concentration of FA in the solution to
levels lower than the minimum concentration neces-
sary to give rise to an aggregation process. The ionic
state is another critical factor in the aggregation pro-
cess. The ionized FA has higher electric negative
charge, which increases the energetic barriers and
therefore inhibits the approximation of FA caused by
the Brownian movement (Buffle and Leppard, 1995).
4. Conclusions
The proposed theoretical model based on TNB
monomer fits in well with some properties of FA:
solubility (dipolar moment) and electronic and vibra-
tional spectra. The presence of water molecules has a
great stabilization effect on the electrostatic energy.
This effect is greater as ionic state increases. The ionic
R.A. Alvarez-Puebla et al. / Science of the Total Environment 358 (2006) 243–254 253
state is one of the most important factors in the
aggregation process in vacuum. The non-ionized
aggregated species are more stable than single ones
because of the increment in their interaction due to H-
bonding and non-bonding forces. In solution, the FA
concentration is a critical factor for the aggregation.
The system containing two FA molecules probably
did not form aggregates because its equivalent con-
centration was too low (0.035 M). When the concen-
tration was increased (0.070 M), the system gave rise
to the formation of aggregates. Nevertheless, in all the
systems there were stable FA molecules in solution
because the aggregation of some of the system mole-
cules decreases the effective concentration of FA to
levels lower than the minimum concentration needed
to give rise to an aggregation process. The ionic state
(i.e. pH) is another critical factor in the aggregation
process. The ionized FA has higher electric negative
charge, which increases the energetic barriers and
inhibits the approximation of FA caused by the Brow-
nian movement.
In summary, the power of molecular simulations to
link experimental data with detailed chemical inter-
pretation could provide a useful insight into the beha-
viour of HS and their interactions with nutrients and
contaminants in the environment.
References
Alvarez-Puebla RA, Valenzuela-Calahorro C, Garrido JJ. Retention
of Co(II), Ni(II) and Cu(II) on a purified brown humic acid
Modeling and characterization of the sorption process. Lang-
muir 2004a;20:3657–64.
Alvarez-Puebla RA, Garrido JJ, Aroca RF. Surface-enhanced vibra-
tional microspectroscopy of fulvic acid micelles. Anal Chem
2004b;76:7118–25.
Balbuena PB, Seminario JM, 1999. Molecular dynamics.
Amsterdam7 Elsevier; 1999.
Buffle J, Leppard GG. Characterization of aquatic colloids and
macromolecules: 1. Structure and behaviour of colloidal mate-
rial. Environ Sci Technol 1995;29:2169–75.
Bruccoleri AG, Sorenson BT, Langford CH. Molecular modeling of
humic substances. In: Gabbour EA, Davies G, editors. Humic
substances: structures, models and functions. Cambridge7 The
Royal Society of Chemistry; 2001. p. 193–208.
Chen Y, Senesi N, Schnitzer M. Information provided on humic
substances by E4/E6 ratios. Soil Sci Soc Am J 1977;41:258–352.
Chin Y-P, Aiken G, O’Loughlin E. Molecular weight, polydisper-
sity, and spectroscopic properties of aquatic humic substances.
Environ Sci Technol 1994;28:1853–8.
Chin WC, Orellana MV, Verdugo P. Spontaneous assembly of main
dissolved organic matter in polymer gels. Nature 1998;391:
568–71.
Davies G, Fataftah A, Cherkasskiy A, Ghabbour EA, Radwan A,
Jansen SA, et al. Tight metal binding by humic acids and its
role in biomineralization. J Chem Soc Dalton Trans 1997;21:
4047–60.
Hypercube. HyperChem 7.01. Hypercube Inc., Gainsville; 2002a.
Hypercube. HyperChem Computational Chemistry. Hypercube Inc.,
Waterloo; 2002b.
Jones ML, Bryan ND. Colloidal properties of humic substances.
Adv Colloid Interface Sci 1998;78:1–48.
Jorgensen WL, Tirado-Rives J. The OPLS potential functions for
proteins Energy minimizations for crystals of cyclic peptides
and crambin. J Am Chem Soc 1988;110:1657–66.
Jorgensen WL, Chandrasekhar JD, Madura JD, Impey RW, Klein
ML. Comparison of simple potential functions for simulated
liquid waters. J Chem Phys 1983;79:729–37.
Kubicki JD. Molecular modeling of humic and fulvic acid. Prepr
Ext Abstr ACS Natl Meet, Am Chem Soc Div Environ Chem;
2000. p. 617–9.
Kubicki JD, Apitz SE. Models of natural organic matter and
interactions with organic contaminants. Org Geochem 1999;
30:911–27.
Kubicki JD, Sykes D, Apitz SE. Ab initio calculation of aqueous
aluminium and aluminium–carboxilate complex energetics
and 27Al NMR chemical shifts. J Phys Chem A 1999;103:
903–15.
Manning TJ, Bennett T, Milton D. Aggregation studies of humic
acid using multiangle laser light scattering. Sci Total Environ
2000;257:171–6.
Merchan M, Serrano-Andres L, Slater LS, Roos BO, McDiarmid R,
Xing X. Theoretical spectroscopy of organic compounds. J Mol
Struct Theochem 1998;463:201–10.
Myneni SCB, Brown JT, Martinez GA, Meyer-Ilse W. Imaging of
humic substance macromolecular structures in water and soils.
Science 1999;286:1335–7.
Niemeyer J, Chen Y, Bollag JM. Characterization of humic acids,
composts and peat by DRIFT. Soil Sci Soc Am J 1992;56:
135–40.
Schulten HR. The three-dimensional structure of humic substances
and soil organic matter studied by computational analytical
chemistry. Fresenius J Anal Chem 1995a;351:62–73.
Schulten HR. The three-dimensional structure of soil organo-
mineral complexes studied by analytical pyrolisis. J Anal
Appl Pyrolysis 1995b;32:111–26.
Schulten HR. Molecular modeling of humic substances, soil organic
matter and soil particles: potential and limits. Mineral Mag
1998;62A:1358–9.
Schulten H-R. Interactions of dissolved organic matter with xeno-
biotic compounds: molecular modeling in water. Environ Tox-
icol Chem 1999;18:1643–55.
Schulten HR, Leinweber P. New insights into organic–mineral
particles: composition, properties, and models of molecular
structure. Biol Fertil Soils 2000;30:399–432.
Schulten HR, Schnitzer M. Chemical model structures for soil
organic matter and soils. Soil Sci 1997;162:115–30.
R.A. Alvarez-Puebla et al. / Science of the Total Environment 358 (2006) 243–254254
Seeger DM, Korzeniewski C, Kowalchyk W. J Phys Chem 1991;
95:68–71.
Sein LN, Varnum JM, Jansen SA. Conformational modeling of new
building block of humic acid: approaches to the lowest energy
conformer. Environ Sci Technol 1999;33:546–52.
Senesi N. Aggregation patterns and macromolecular morphology
of humic substances: a fractal approach. Soil Sci 1999;164:
841–56.
Senesi N, Rizzi FR, Dellino P, Acquafredda P. Fractal humic acids
in aqueous suspensions at various concentrations, ionic
strengths and pH values. Colloids Surf A 1997;127:57–68.
Shevchenko SM, Bailey GW, Akim LG. The conformational
dynamics of humic polyanions in model organic and organo-
mineral aggregates. J Mol Struct 1998;460:179–90.
SSPS. PeakFit 4.05. SSPS Inc., Chicago; 1995.
Humus chemistry: genesis, composition and reactions. In: Stevenson
FJ, editor. New York7 John Wiley & Sons; 1982.
Swift RS. Organic matter characterization. In: Sparks D.L., editor.
Method of soil science analysis: chemical methods Part 3.
Madison7 Soil Science Society of America; 1996. p. 1011–69.
Swift RS. Macromolecular properties of soil humic substances: fact,
fiction and opinion. Soil Sci 1999;164:790–802.
Tombacz E. Colloidal properties of humic acids and spontaneous
changes of their colloidal state under variable solution condi-
tions. Soil Sci 1999;164:814–24.
Wershaw RL. Molecular aggregation of humic substances. Soil Sci
1999;164:803–11.
Young D. Computational chemistry. New York7 Wiley-Interscience;
2001.
Zerner M. Semi-empirical molecular orbital methods. In: Lipkowitz
D.B., Boyd D.B., editors. Reviews of computational chemistry.
New York7 VCH Publishing; 1991. p. 313–66.