Molecular modelling - opportunities for soil research

Dedicated to Univ. Prof Dipl.-Ing. Dr. Dr. h.c. mult. Winfried E. H Blum on the occasion o[his 6(Jh birthday

M. H. Gerzabek, A. J. A. Aquino, G. Haberhauer, D. Tunega and H. Lischka

Molekülmodellierung - Perspektiven für die Bodenforschung

Herrn O. Univ. Prof DipL-Ing. Dr. Dr. h.c. multi Winfried E H Blumanläßlich seines 60. Geburtstages gewidmet

1. Introduction

Soil has a large number of functions wirhin the biosphere.

Besides of being the basis for crop production, soil acts as

one ofthe main sinks for atrnospheric CO2 and various pol­

lutants. Especially the latter fact is ofincreasing importance.

The soil cover in many cases is the only effective filter and

buffer against pollutants entering the food chain. The use

of agrochemieals, the application of organic residues con­

taining different organic and inorganic pollutants to agri­

culturalland and immisions ofsubstances originating from

combustion processes and regular or accidental releases

from nuclear facilities led to a large variety of pollutants

already being immobilised in the soil matrix. The main

classes of substances of concern are heavy metals, radionu­

clides, pesticides, hydrocarbons, PAH's, PCB's, dioxins and

- more recently under discussion - endocrine disrupters

(e.g. GÜLDEN et al., 1997). Especially the number of rele­

vant organic compounds is increasing rapidly and data on

the behaviour and fate of these substances are urgently

needed. Thus, the questions we ask in soil science are get­

ting even more complex than a few decades aga, when soil

productivity was the main matter of concern. The largest

efforr in this context is needed to investigate the specific

adsorption and fixation processes for a given compound,which comprises the most important basis for the estima­tion of its environmental behaviour. Laboratory andlysimeter studies (FÜHR et al., 1998; GERZABEK et al., 1998)

are the most important methods in this connection. Any­

how, these studies are quite time consuming and may not

be readily available for a firstquick risk assessment of a new

compound being just in development or recently disco­vered in the environment. Theoretical studies, thus, could

be useful to assist the risk assessment and to help under­

standing the processes of interactions with the soil matrix

on a molecular level.

Computational chemistry (Ce) is a quite yaung seience

and its potential is intimately connected with the capabili­

ties of computer hard- and software (see chapter 2). The

rapid increase in computational power available enables

ZusammenfassungDie Computerchemie (Ce) ist eine relativ junge Wissenschaft, die sich rasch entwickelt. Dies hängt auch mit der

zunehmenden Leistungsfähigkeit von pes und Workstations zusammen. Derzeit wird die Computerchemie selten in

der bodenkundlichen Forschung verwendet, obwohl zahlreiche Anwendungsgebiete - insbesondere auf dem Gebiet

der Erforschung des Verhaltens von Umweltschadstoffen im Boden - vorauszusehen wären. Neue Techniken, wie bei­

spielsweise die ONIOM-Methode ermöglichen in diesem Zusammenhang relativ genaue Berechnungen von Inter­

aktionen auch innerhalb großer Systeme mit bis zu 100 Atomen. Ein weiteres interessantes Forschungsfeld ist die

Untersuchung der hypothetischen Struktur von Huminstoffmodellen und von Interaktionen mit verschiedenen

Schadstoffen. In der vorliegenden Arbeit werden Beispiele für Anwendungen der Computerchemie in der Bodenfor­

schung präsentiert: (i) die Interaktion von organischen Säuren mit Aluminium, (ii) die hypothetische Struktur von

Huminsroffmodellen und deren Wechselwirkung mit Pflanzenschutzmitteln und (iii) die Anwendung der ONIOM

Methode für das Studium der Adsorptionsstellen an Tonmineraloberflächen.

Schlagworte: Aluminium, Bodenforschung, Computerchemie. Humistoffe, Tonminerale.

Die Bodenkultur 133 52 (2) 2001

M. H. Gerzabek, A.]. A.Aquino, G. Haberhauer, D. Tunega and H. Lischka

SummaryCornputational chemistry (Ce) is a young science which develops rapidly. This is intirnately connected with the

increase in computer power and the recent rapid improvement ofPCs and work stations. Although at present seldomused, numerous applications in soil science, especially in pollution research, can be envisaged. New techniques as e.g.rhe ONIOM method make it possible to investigate interactions within quite large systems of up to 100 atoms withhigh accuracy. Another inreresting field for CC are studies of the hyporhetical structure of humic substance models.These can be used for investigations of interactions with various pollutants (e.g, "bound residues"). In this paper we

examplify quire different applications of CC in soil research: (i) the interaction of organic acids with aluminium,(ii) the hypotherical structure of humic substances and their interactions with pesticides and (iii) the use of the

ONIOM technique ro study adsorption sites on day minerals.

Key words: aluminium, day minerals, computational chemistry, soil research, humic substances.

more and more accurate calculations and an application toa wider range of systems. Since even personal computers andsmall workstations nowadays allow to perform modellingcalculations if rhe described systems are not too large, com­

putational chemistry has entered more applied sciences likesoil science. Anyhow, these theoretical methods are seldom

used in soil science up to now and applications are restriet­ed to a few topics. The potential of CC has been alreadyshown e.g. for studies on Al-organic acid complexes(AQUINO et al., 2000; TUNEGA et al. 2000), for investiga­tions of clay mineral interlayers and their surface geochem­istry (CHANG et al., 1998; 1999, SPOSITO et al., 1999) andmodelling smaller systems of humic substances (SCHULTENand SCHNITZER, 1997). Thus, CC is and will be used to

investigate soil processesin the micro-scale in the future.The present paper atrernpts (i) to review some ofthe work

already been done in soil science with the tool CC and (ii)to give examples ofrecent studies highlighting the potentialof CC rnethods for future research.

2. Recent developments in computational ehe­mistry

Computational Chemistry is a rapidly growing section ofTheoretica1 Chemistry wirh emphasis on application-ori­ented molecular problems being solved on the computer.The fundamental equation determining the properties ofatornic and molecular systems is the Schrödinger equation,Its analytic solution is only possible for a few simple modelcases. Based on the rapid progress in computer technology,very powerful numerical methods and respecrive computerprograms have been developed for the quantum chemicalcomputation of atoms and molecules. These methods have

been applied very successfully in practically all fields ofchemistry. Several dasses of cornputational procedures areavailable ranging from ab initio and OFT (density func­tional theory) level via semi-empirical methods to simple

force-fields and interatornic potentials. In these series ofmethods the firsr two ones are computational most expen­sive. However, they give on rhe average the most reliableresults. Serni-empirical methods are computationally lessexpensive, but require well-tuned parameter sets for suc­

cessful application. Finally, force-field and interatomic

potential methods are by far the cheapest ones. They lackany quantum chemical formalism and rely heavily on the

empirical adjustment of parameters. From the large nurn­ber ofrnethods and computational techniques only a fewofthem, which are directly relevant to our work, will be dis­cussed here, For more information and for an overview of

typical applications see e.g. ]ENSEN (1999).Sufficient flexibility and reliability of the computational

methods are of particular importance in soil sciences, sincewe have to deal with various types of interactions and ehe­

mical bonding such as intermolecular forces between soluteand solvent, adsorption processes on surfaces and chemicalbonds occurring in the inorganic and organic constituentsof the soil. Besides this great variability in the interactions,

it is also the size ofthe individual subsystems, which makesthe ca1culations very time-consuming. For example, the

adsorption process on the clay surface requires either thecomputation of a larger cluster cut out from the crystalstructure or the treatment as a periodic system with a largeunit cell including the adsorbed species. The humic sub­stances are polymers with a very irregular structure andtherefore difficult to describe.

In the present contribution we concentrate on OFT

methods. They belong - as has been stated above - to the

Die Bodenkultur 134 52 (2) 2001

Molecular modelling - oppoftunities für soil research

most "expensive" ones, but combine the required flexibility

and reliability with cornputational efficiency. Most of the

calculations described in this work have been set up as

benchmark investigations which shall guide future investi­

gations on even larger and more complex systems.

One of the most crucial practical bottlenecks in an ab ini­

tio or DFT calculation besides computer time is the huge

amount ofdata produced. This problem has been overcome

by the introduction of "direct" methods where storage of

these data is avoided by recomputation every time they are

needed. This teehnique is available in all major quantum

chemical program packages (e.g, GAUSSIAN98, 1998 or

TURBOMOLE (AHLRICHS et al., 1989; ARNIM, 1998)).Several additional techniques have been developed wirhinab initio and OFT approaches, which are especially dedi­

cated to the reduction of the effort required for the compu­

tation of larger molecules. One of those is the RI method

(VATRAS et al., 1993; EICHKORN et al., 1995) in which the

calculation of the most time consuming rwo-electron inte­

grals is substantially reduced bythe introduction ofan auxi­

liary basis set. Using thesemcthods, DFT calculations on

fullerenes C60 up to Cso and inorganie clusters such as[Cd32Se14(SeH)36(PH3)4] have been reported (AHLRICHS

et al.. , 1998). Another alternative for the calculation oflarge

molecules is the ONIOM method (DAPPRlCH et al., 1999).The basic idea of this method is the partitioning of the

entire system in rwo or more layers, where the important

part (inner layer) is treated at a higher level of theory and

the remaining layers at lower levels. In total, calculation of

moleeules with 100 atoms or more are feasible nowadays.

Many of these calculations can be carried out even on stan­

dard pes under Linux, which provides a very cost-effective

alternative to high-end workstations or central computers.

Solvati 0 n effects play a crucial role in soil processes. The

inclusion of solvation poses very challenging problems for

quantum chemical calculations. One eonceptually simple

approach is the embedding of the solute moleeule into a

cluster ofsolvent moleeules and to treat the whole system as

a supermoleeule. However, iE the solute is polar or even

charged, strang and far-reaching interactions berween the

solute and solvent moleeules will occur, In order to accom­

modate for these extended interactions, very large clusters

of solvent moleeules would have to be taken into account,

which increases the cost for the cornputation drastically.

Therefore, continuum solvation models have been deve­

loped, which provide a good and cast-effective alternative.

In this approach the solute (eventually together with a first

solvation sphere) is embedded into a polarisable dielec-

tricum. Early continuum models (BORN, 1920; ONSAGER,

1936; KIRKWOOD, 1935) were oflimited value for quanti­

tative predictions due to the intrinsic sirnplicity of these

rnodels, In the pioneering work of MIERTUS et al. (1981)the polarisable continuum model (peM) was developed

which is now together with the SCRF (self-consistent reac­

tion field) method the basis for most of the recent compu­

ter implementations (TOMASI and PERSICO, 1994). In the

peM model the solute moleeule is placed into a cavity cre­

ated wirhin the continuum. This cavity is constructed bymeans of overlapping spheres centered on the atoms of the

solute, which allows a realistic descri ption of the shape of

the moleeule. The electrostatic potential of the solvent is

represented as point charges at the surface ofthe cavity. The

quanrum chemieal part of the calculation includes thenuclear charges and electrons ofthe solute moleeule and the

point charges describing the continuum. The mutual pola­

rization berween solute and solvent is taken into account bythe aforementioned SCRF method.

3. Examples for applications of computationalchemistry in soil science

3.1. Interaction of aluminium with organie acids­a theoretical study

The aluminium cation plays an important role in soil ehe­

mistry; Free farms (e.g. as hexaaquo complex) present toxi­

cological risks for organisms and especially agricultural

crops and affect the quality of natural sources of water.

Organie acids occurring in the soil have strang influence on

the chemical activity of aluminium. Formation of various

stable complexes can deerease negative factors of the occur­

rence of aluminium in soil. Thus, it is important to under­

stand the related chemical processes in more detail in order

to develop better strategies for the deactivation of free alu­

minium. The formation of complexes between aluminium

and organie acids has been investigated extensively, mainlybased on studies of dissolution processes of aluminium­

containing minerals (see e.g. PALMER and BELL, 1994; FEIN

and HESTRlN, 1994). The analysis of these measurernents

in terms of stabilities of individual compounds is compli­

cared by the fact that a large number of coupled chemical

equilibria oeeur in solution, Their resolution yields good

global information on different groups of complexes but

does not give details about the structures involved. Quan­

tum chemical methods can be used to obtain such detailed

Die Bodenkultur 135 52 (2) 2001

M. H. Gerzabek, A. J.A. Aquino, G. Haberhauer, D. Tunegaand H. Lischka

'"

Figure 1: OprimizedBLYP/SVP+sp geometriesfor waterassolvent-a) monodenrate-[Al(H20)sAc]2+b) bidemate-[Al(H20)4Ac]2+c) bidencare-Alac,d) bidenrate-[AlOx3]3.

Abbildung I: Oprimierre BLYF/SVP+sp Geometrien für Wasser alsLösungsmittel-a) monodenrares [Al(H20)sAc]2+b) bidenrares [Al(H20)4Ac]2+c) bidenratesAlAc3d) bidentates [AlOx3]3-

J''\."",

in absolute values. Asexpected, the solvent effect is propor­tional to the total charge ofthe system. For example, the sol­vent energy of the hexaaquo complex [A1(H20)6P + is-467.5 kcal/mol while for the [AIAc3]O it is only -4.1kcal/mol. In general, the size of the solvation energy of thecomplexes berween aluminium and the organic acids willdepend heavily on the amount of charge cornpensarionwhich takes placewhen the cornplex is formed. This strongvariation in solvation energies ofthe individual species has,ofcourse, also a significant influence on the overall solventeffect for the reactions given by Eq. (1).

Comparing .1Hr(l) values in Table 1 it is found that themost stable complexes in eachset (acetate or oxalate)are thosewhere the water molecules in the hexaaquo complex are fullyreplaced by the organic acid ligands. The oxalate complexesare in general more stable than the acetate complexes.Whilebidenrate acerate complexesare lesssrable than monodenrate

informations. Such investigations have, e.g. been carriedout on several complexes between aluminium and organicacids by KUBICKI at al. (1996, 1999).

The aluminium cation has an amphoteric character. Thismeans that the pH of the solution will have a strong influ­ence on the question which complexes will be formed. Wefocused on complexes where aluminium is sixfold coordi­nated which is characteristic for a pH below 7. At such con­ditions organic acids exist in aqueous solution in dissocia­red form. Thus, aluminium can form complexes in aqueoussolution with fuUy or partially deprotonated organic acids,hydroxyl anions and with water molecules. A typical com­plex structure can be characterised by the formulaA1(H20)p(OH)qLr where L, stands for a deprotonated acidas ligand. In our theoretical investigations (TUNEGA et al.,2000; AQUINO et al., 2000) we restricted ourselves to com­plexes without OH- group in the A13+ coordination sphere.The studied species were complexes with the acetate andoxalate anions (L =Ac' and Ox2-) as ligands. The bondinglinks from these ligands to aluminium are the oxygen atomsfrom the carboxylate groups. Bonding can occur in mo­nodentate and bidenrate form (seeexamples given in Fig. 1aand 1b). Since A13+ possesses a sixfold coordination sphere,complexes with several ligands are possible. Moreover, inthe case ofcomplexes with two acerare or two oxalate anionscis- and trans- isomers can be formed. The calculations havebeen carried out at the OFT levelusing the PCM model forthe inclusion ofsolvent effects.The quantum-chemical cal­cularions were performed with the GAUSSIAN98 package.The BLYP density functional rnethod and a split valencepolarization (SVP) basis set augmented with diffuse sand pfunctions on the oxygen and carbon atoms (SVP+sp) wereused in all calculations. For the relative dielectric constant ethe value of78.54 ofwater was chosen. All thermodynamicdata have been evaluated for a temperatureT = 298.15 K.For more information on technical details seeTUNEGA et al.(2000) and AQUINO et al. (2000).

The following complexation reactions have been studied:

[A1(H20)6]3+ (1) + p V· (1) ~ [A1(H20)6.p*qLP-r*p (1) +p*q H20 (1) (1)

where pis the number ofligands (p =1, 2, 3) and q = 1 or 2for mono- and bidentate ligands, respectively. In Table 1solvation energies Eso1v' reaction enthalpies OHr and Gibbsfree reaction energies OGr are given according to Eq. 1 forthe rnost stable aluminium-acetate and aluminium-oxalatecomplexes. One can see that solvation energies arevery large

Die Bodenkultur 136 52 (2) 2001

Molecular modelling - opportunities for soil research

137

Table 1: Calculared solvent energies, reaction enrhalpies and reactionGibbs free energies (Eq.l) ar the B3LYP/SVP+Sp level for alu­minium-acerate and -oxalate complexes, respecrively (Allenergies are in kcal/mol)

Tabelle1: BerechneteLösungsenergien, Reaktionsenthalpien und Gibbs' ­sehe Reaktionsenergien (GI. 1) berechnet auf dem B3LYP/SVP+Sp Niveau für Aluminium-Azetat bzw, Aluminium­Oxalat Komplexe (alle Energien in kcal/mol)

Complex Eso1v ilH/1) ilGr (I)

m-[A1(H2O)5Ae]2+ -198.2 5.2 3.5b-[A1(HP)4Ac]2+ -204.1 15.9 8.1m-[A1(Hp)4Ac2]+-cis -59.7 -15.3 -11.9b-[A1(HP)0S]+-cis -59.9 6.8 -11.7b-[A1Ac3]° -4.1 -4.9 -28.2m-[A1(H2O)50x]+ -100.9 9.8 8.7b-[A1(HP)40x]+ -107.5 5.6 -6.3b-[A1(H2O)20x2]- -80.6 -9.4 -31.7b-[A10x3]3- -335.0 -20.1 -47.4

ones according to ÖHr' for oxalate complexes the situation isreversed (cf. the pair ofthe first/second and the sixth/seventhrows in Table 1). The reason for this behaviour comes fromthe fact that in the acetäte case only one carboxylate group isinvolved in the bidentare binding whereas it is two different

carboxylate groups in case of oxalate. Characteristic diffe­rences between ÖHr and ÖGr are observed for mono- andbidentate complexes. Reactions with bidentate structures aresignificantlyenhanced at the ÖGr level. In case of acerate li­gands the formation of'bidentare complexes becomes almostequivalent to that of the monodentate species. For oxalate

complexes the difference berween the formation of mono­and bidenrare structures becomes even more favourable for

the latter.The differences berween ÖHr and LlGr are relative­ly small for the formation of monodentate species (e.g. 1.7kcal/mol for monodenrare [Al(HzO)SAc]2+) but much larg­er for bidentate cases (e.g. 7.8 kcal/mol for bidentare[Al(HzO)4Ac]z+). A detailed analysisperformed by AQUINOet al. (2000) showed that the reason for this behaviour is the

increasein entropy because ofthe additional number ofwatermoleculesreleased from the aluminium coordination sphereafter substitution by a bidentate ligand as compared to amonodenrate one. Therefore, itwas concluded thar reactionsfor monodenrare complexes are energy-driven, whereas thosefor the bidentare complexes are entropy-driven.

Experimenrally derermirred ÖG r values were -3.8 kcal/mol and -6.2 kcal/ mol for the AlAc2+ and AlAc2+complexes

(PALMER and BELL, 1994) and -15.5, -18.2 and -19.6 kcal/mol for the AlOx+, AlOxz- and AlOx3

3- complexes (FEIN

and HESTRIN, 1994). In both experimental investigations

Die Bodenkultur

only global chemical reactions of the type Al3+ + mLn- -7

AlLm

3-m*n were taken as basis and no details of the coordi­

nation sphere ofAl3+ such as the number ofwater molecules

or other ligands were specified, Moreover, it is not possibleto distinguish mono- and bidentate structures from the

experimental data, In view ofthis situation we have foundrelatively good agreement with experimental values and ourpredictions of the most srable complexes agrees with expe­rimental observations.

3.2. Hypothetical structures of humic substances andmodelling of interactions of soil organic matterwith pesticides

Humic substances as a very stable and reactive patt of soil

organic matter playa crucial role in rhe fate and behaviour oforganic pollurants and heavy metals in the environment.Adsorption of organic contaminants or heavy metals ontosoil organic matter has the potential to be a major controllingfacror in their bioavailability, Models for estimating theamount and stability of sorbed contaminants based simplyon the fraction of organic carbon in a soil can oversimplifythe process ofsorption in the environment. Itwasshown that

origin, structural characteristic and/or inreraction ofHS withday minerals affect the sorption behaviour. In order to help

to understand sorption processes, several models of soilorganic matter components, which are possible substrates for

sorption, have been developed on a molecular level.Molecular modelling of humic substances is restricted

due to the availabiliry of strucrural information. Experi­mental models, based on breaking down humic substances

through pyrolysis (SCHNITZER and SCHULTEN, 1992) orthermochemolysis (DELRio and HATCHER, 19%) and the

derermination of the formed components were generated.The analysed fragments are used to rebuild the overall struc­

ture. However, both side reactions during the pyrolysisprocess and the lack ofinformation how these fragments arelinked together Iimit the use of this information to createtwo dimensional models of humics. Infrared specrroscopyand NMR spectroscopy supplies information on an appro­ximate distribution of functional groups wirhin humicacids. Togerher with these IR, NMR, UV data, mass spec­rrometry results using soft ionisation techniques, likeMALDI-TOF MS or ESI-MS, demonstrate the complexi­tyofhumics (HABERHAUER et al., 1999; BROWN and RICE,2000; SOLOULKl et al., 1999). Thus, humic substances,which can be assumed to be complex mixtures of corn-

52 (2) 2001

M. H. Gerzabek, A.]. A. Aquino, G. Haberhauer, D. Tunega and H. Lischka

pounds with no prevailing single molecule, present a chal­lenging task for rnodelling,

models have in eommon that few scientific structural proveexists for any of the proposed constitutions.

2-dimensional modelsIn order to visualise interactions ofhumic substances (HS)with other moleeules two dimensional models were deve­loped. Such representations were constructed using theexisting limited structural information. So, it is alwaysimportant to keep in mind that such models are only anapproximation to illustrare selected processes and thar thesemodels can not reflect the whole complex properties ofhumie acids, Such two dimensional models were proposedby many authors and many of them are now used in text­books (e.g. SCHACHTSCHABEL et al., 1989). Two dimen­sional models can be used to visualise and graphicallydescribe interactions berween organie matter and pollu­tants or day minerals. A possible exarnple ofsuch a modelis shown in figure 1. According to average struetural datathis model is composed of aromatic, lipophilic alipharic,polar and carboxylic moieties, Several two dimensionalmodels were published in the last ten years. Some consist ofmaeromolecules (SCHNITZER and SCHULTEN, 1992) orpolymerizable units (DAVIS et al., 1997) while other modelspropose that humic acids are erude mixtures of low mole­cular weight eompounds (STRUMPF, 1998). However, all

3-dimensional burnie substance modelsA two dimensional presentarion ofstructures does not allowto understand and visualise many types of interactionsberween moleeules. Three dimensional representations arenecessary to describe complex reactions in biological sys­tems, Therefore, a demand of suitable three dimensionalhumic substance models exists in order to describe severaltypes ofadsorption processes. In case ofinteractions of daymaterials and organic compounds sorne recently publishedpapers demoristrate the suitability of computational ehe­mistry methods (SPOSITO et al., 1999; PARK et al., 1997;CHANG et al., 1997). Molecular dynamic methods wereemployed to examine sorption behaviour of organic com­pounds in day interfacial structures (TEPPEN et al., 1998).The availability of day mineral structures and moleeularstructures of the investigated organic compounds enablescomparison of the theoretical data to experimental results.The fast development of eomputational chemistry andunderstanding of structural requirements allow the con­struction and modelling ofthree dimensional models out oftwo dimensional concepts (Figure 2). Three dimensionalconstruction ofhumic substance models using several types

'0

c polarandloraromatic sugarmoiety.. Ci

" c

..)

1'1

" .c

C D......" ... ,

o

o-,

arnlnoacld -moiety

,·0

o

Iipophilic"

<,.....

', ...•,

Figure 2: Two (left) and three (righr) dimensional graphical representation ofa hyphotetical humic substance modelAbbildung 2: Zwei (linksl- und dreidimensionale(rechrs) graphische Darstellung eines hypothetischen Huminstoffmodells

Die Bodenkultur 138 52 (2) 2001

Molecular modelling - opportunities for soil research

of available molecular modelling software packages andsubsequent minimisation ofthe structures using force fieldor semi empirical methods will generate stable conforma­tions (seeFigure 2 and SCHULTEN and SCHNITZER, 1995).SCHNITZER and SCHULTEN (1997) used molecular model­ling programs to build a humic substance model using theirrwo dimensional concept. SEIN et al. (1999) investigatedthe conformational flexibility of their humic substancebuilding block. Three dimensional structures of HS wereused to describe (SCHULTEN, 1996; SCHULTEN, 1999;SCHULTEN and LEINWEBER, 2000) and illustrate (VONWANDRUSZKA, 1998) the interaction with xenobiotic com­pounds. Strong H-bondingwas predicted to play an impor­tant role wirhin the formation ofhumics and organic corn­pounds complexes (KUBICKI and APITZ, 1999).

Even if the use ofcornpurational chemistry in that respectseems to be very attractive there are several limitationsregarding this humic substance representation approach.First, one of the most serious limitations for the applicationand relation of such studies ro environmental processes isthe lack of structural support for the chosen rwo dimen­sional models. Second, it is questionable if folding of suchmacromolecules (In many casesminimisation was used for2 D to 3D transformation) will give any realisric results.Even if modern compurational methods are used the mo­delling of folding of weIl characrerised macromoleculessuch as proteins (SOCCI et al., 1998) is resniered by manyparameters. Third, the conformational space of flexiblemacromolecules allows a huge number ofdifferent confor­mations. Identification of the environmental importantconformations requires at least some basic structural data(LEACH, 1996). Additionally, the used force fields are oftennot optimised to study organie - cation or organic mineralbonding. Both boding types are of great importance in soilenvironments, Thus, the adaptation and development offorce fields for such interactions should be considered.

Macromolecular models were used for the description ofsingle interactions (e.g. H-bonding) in a defined environ­ment (= conformation/structure ofhumics), but due ro thelack ofconformational and structural data no direct relarionofthe output to experimental results can be gained. In orderto circumvent these problems other approaches were cho­sen ro describe the interaction ofhumic substance moietiesto organic compounds.

Reductive approaches 0/thedescription 0/humicsubstancesTherefore, one of the first approaches to model humic sub­stances were based on meso-size models. These models con-

sist of hydrophilie and hydrophobie parts and can be usedto describe the amphoteric and surface active behaviour ofhumics. In comparison to molecular models these so calledsecondary structural models ofhumic materials (WERSHAWet al., 1986) are directly developed from experimentalresults and used to explain certain behaviour of humics.Therefore, less arbitrary assumptions are necessary to createthem. Such models can be utilised to describe the built upand the formation of humics in soil and aqueous environ­ments. They have brought an important input in our nowa­days understanding ofhumic substances (CONTE and PIC­COLO, 1999; WERSHAW, 1993). Another approach applyingcomputer modelling to structural problems ofhumics wirh­out describing the complete molecular structure was chosenby BIRKETT et al. (1997). The authors commenced rheirmodelling defining humics bya set ofgroups and not on anatornic level. Humic substances are represented by rhe mo­lecular volume, amount ofcarboxylic, aromatic and severalother groups. The average group information can be direct­ly obtained from experiments. Information on how thesegroups are connected and how humic molecules are consti­ruted (data which are not total available) are not required.Therefore, in comparison to the molecular strucrural mo­dels no additional assumptions have to be made.

However, borh above mentioned conceptional approach­es on a meso-scale do not allow ro model or describe inter­actions between small organic molecules such as pesticidesor various organic pollutants and humic substances on anatornic level. Models, which consist of functional moietiesof humic substances only, can be used to describe sorptionserengeh of different sorptive sites (KUBICKI and APITZ,1999). Such models rely on experimental structural infor­mation. Additionally, the small size of the models - theycontain mainly less then 100 atoms - allows the use of pre­eise quantum chemical methods. Thermodynamic data canbe obtained using such methods. These theoretical resultscould be related to experimental data (TUNEGA et al.,2000). The application of self-consistent field methodsmimics solvation or hydrophobie environments (KUBICKIet al., 1999). Thus both single interactions in hydrared andnon hydrared environments can be examined. However,since sorption oforganic compounds in soil is governed byseveral different binding sites, a number of possible inter­action types must be considered. Quantum chemical calcu­lations were used for rough estimations of possible inrerac­tion energies of a herbieide with mineral or soil organicmatter moieties (HABERHAUER et al., 2000a). The systemsinvestigated consist ofeither a mineral moiery, which can be

Die Bodenkultur 139 52 (2) 2001

M. H. Gerzabek, A.]. A. Aquino, G. Haberhauer, D. Tunega and H. Lischka

140

regarded as a model for a mineral surface edge, or/and anorganic molecule representing a structural rnoiery of soilorganic matter and rhe herbicide. The calcularions pro­nounced the importance of several types of polar interac­

tions and are in agreement with the experimental results.Anorher application of computational chemistry to soil

sorption processes is the relation of rhe experimentallydetermined adsorption directly to the structures of thecompounds investigated. Struetural parameters, such assize,polarity or/and charge distribution can be compared toadsorption and desorption dara of a group of strucrurallysimilar compounds (HABERHAUER et al., 2000b). Theinfor­

mation on chemical structures was used to estimate soil par­tirion coefficients (LOHNINGER, 1994). Modification in

structure can effect the sorption behaviour and could beused for the interpretation of sorption mechanism of a

group ofstrueturally similar organic compounds.

3.3. ONIOM study ofadsorption sites on the 001surface of 1:1 day minerals

Clay minerals represent an important inorganic componentof soils and significantly affect physicochemical proeessestherein. Clays are usually small powder particles « 2 mm)witha high specifie surface area and a high chemical surface

aetivity. Adsorption of mobile chemical species from soilsolution on mineral surfaces is a very important processoceurring in soil. Soil solurions represent chemically verycomplex systems containing inorganic ions and organiemolecular species originated from natural biochemicalprocessesor from human aetivities. Especially the behaviourofpollutants in soil (radionuclides, heavy rnetals, pesticidesetc.) is - due to their potential adverse effects on the eeosys­

tem - the main foeus ofmany investigations (SCHACHTSCH­ABEL et al., 1989). Among organie species, acids and theirderivatives play an important role because of rhe relatively

high ehemical activity of their carboxyl groups. Organicacids affeet the pH ofsoils and can also strongly interactwithother eompounds in solution formingvarious associates andcomplexes, or with the surfaces of the inorganiclorganiesolids forming strongly bound adsorbates. Experimentally, itisvery difficult to distinguish energetieally different adsorp­tion sites on mineral surfaces (e.g, berween "regular" 001surfacesand edgesurfaces (110, 01 0) with broken bonds)orto extract side effeets aecompanying adsorption (like forma­tion ofcomplexes or intercalations),

Computer simulation methods are very useful for the

Die Bodenkultur

detailed description ofinteracrions ofmolecular species withclay minerals. Conventional molecular dynamic or MonteCarlo methods based on classical force fields have been usedin the srudy of interaction of water or srnall organic corn­

pounds on day mineral surfaces or inside of the interlayerspace (DELVILLE, 1995; BRlDGEMAN and SKIPPER, 1997;

TEPPEN et al., 1998; SHROLL and SMITH, 1999; SMIRNOVand BOUGEARD, 1999). A disadvantage of rhose methods is

the difficulty to determine a balanced set of parameters forthe empirical potentials, mainly for the interaetions betweendifferent kinds of atoms in the clay and in rhe adsorbedmolecules. Periodie ab initio total energy pseudopotential

calculations were performed on the ralc-water and pyro­phyllite-watet systems (BRlDGEMAN et al., 1996). A mole­cular cluster approach can also be used in studies ofthe inter­actions ofsmall molecular species with day mineral surfaces.In this approach only a certain fragment of the solid-state

structure is treated. Only a very few papers (CHATTERjEE etal., 1997; ZHANPEISOV et al., 1999; PELMENSCHIKOVandLESZCZYNSKl, 1999) have been published on moleeular clus­ters which were realistically describing a day layer structureusing an ab initio molecular orbital approach. Such clustermodels would conrain about fifty atorns and more, whiehimplies a compurarionally very demanding task wirhin the

ab initioor DFT framework. The aim of the present workwas the study of the interaetion of water, acetic acid and

acerate with different adsorption sites of rhe 001 surface of

minerals of kaolinite-type applying the duster modelapproach togetherwirh the ONIOM layertechnique (DAP­PRICH et al., 1999). These model systems represent the basisfor future investigations of pesticide-clay interactions,

One individuallayer ofkaolinite eonsists ofrwo connected

sheets - a tetrahedral sheet formed from corners sharing SiO4

tetrahedra and an octahedral sheet consisting of edges shar­ing Al06 tetrahedra, Both sheets have a common plane of

apical oxygen atorns,One third ofall possible ocrahedral cen­tral positions are empty. This creates cavities in the ocrahedralsheet and also causes deformation of the other octahedra.Hydroxyl groups participating in hydrogen bonds with basaloxygen atoms from adjacent layers cover the ourer surface ofthe octahedral sheet,The tetrahedral side of the single layerhas characteristie ditrigonal cavities and the surface is formed

from the plane ofbasal oxygen atoms each shared by rwo sil­icon atorns. These two planes (tetrahedral basal oxygen atoms

and octahedral surface hydroxyls) are parallel to rhe crysral­lographic 001 surface and can be considered as electronical­ly saturated, Each of these two planes forms an importantpart of the surface of the macroscopic mineral partiele.

52 (2) 2001

Molecular modelling - opportunities for soil research

The clustermodel ofthe single kaolinite layer was derivedfrom the structure of the mineral dickite (JOSWIG andDRITS, 1986). Ir consists of 78 atoms and contains oneditrigonal tetrahedral ring and one octahedral ring. Dan­gling bonds were saturated with hydrogen atoms. Thestructure of this layer fragment can be seen in Figures 3-6.The adsorbantswere placed above holes on borh sides ofthelayer - water and acetic acid (HAc) above the tetrahedralside and water and acerate anion (Ac') above the octahedralside. The rwo-layered ONIOM rnethod implemented inthe GAUSSIAN98 package was used for all studied sys­tems. In caseofadsorption on the tetrahedral side the innerONIOM layer is the tetrahedral sheet, and in case ofadsorption on the octahedral side it is the ocrahedral sheet.The B3LYP/SVP method was used in all ONIOM geome­try optimisation calculations for the inner part and thesemiempirical PM3 method for the outer layer. Partialgeometry optimisation was performed wich the followingconstraints: the position ofthe adsorbate with respect to theday surface and its geometry were fully optimised. Thepositionsofatoms wirhin the day duster were kept fixed atthe experimental geometry ofdickite in order to reduce anyartefactsdue to finite duster sizes induced by the geometryoptimization. Only the six nearest neighbour hydroxylgroups in case of adsorption on the octahedral side wereoptimised also. Für the optimised geometries single pointcalculations replacing the PM3 method by MNDO in theONIOM approach and extending the SVP basis to SVP+sp(similar to the aluminium complexes) on atorns (excepthydrogen) directly involved in the adsorbent-adsorbatebinding were performed. Additionally, the qualiry of theONIOM approximation was controlled by performingB3LYP calculations for the whole complex. Interactionsenergieswere corrected to the basis set superposition error(BSSE) according to the counterpoise rnethod ofBovs andBERNARDI (1970).

In Table2 calculated results for all studied systems are dis-

played. Figures 3-6 present optimised geometries and alsocontain the most important geometrical parameters.

The D(O)-H20 system (Figure 3). The water is locatedabove the centre of the octahedral cavity and forms hydro­gen bonds with the three surface hydroxyl groups sur­rounding the empty octahedral hole. It is oriented such thatthe lone pairs ofthe oxygen atoms are involved in hydrogenbonds wirh rwo surface hydroxyl groups and one proton ofthe water molecule shows towards one oxygen atorn of asurface hydroxyl group. Calculated interaction energies(Table 2) do not differ much between individual methods,Only a small difference is observed between ONIOM andthe full approach. More important is the extension of theSVP basis set to the SVP+pqualiry. The difference betweenB3LYP/SVP and B3LYP/SVP+sp is 2.2 kcal/mol whichamounts to about 25% of the B3LYP/SVP energy.

The D(O)-Ac system (Figure 4). The acetate anion is alsopositioned above the empty octahedral hole. The C-C bondis almost perpendicular to the hydroxyl surface of the layerwith the methyl group pointed awayfrom the surface. Bothoxygen atorns of the carboxylate group are involved inhydrogen bonding with the four adjacent surface hydroxyl

Figure 3: Dickite (octahedralj-Hjf) system: ONIOM(B3LYF/SVP:PM3) oprirnised geometry. 'Ioptlefr) and sldeüighr) view

Abbildung 3: Das Diekir (oktahedral)- HzO Syscem: ONIOM(B3LYF/SVP:PM3) optimierte Geomerrie. Aufsicht (links) undSeitenansicht (rechts)

Table 2: Calculated interacrion energies of water, acetare and aceric acid with ehe dickire 001 surface (All energies are in kcal/mol)Tabelle2: Berechnete Wechselwirkungsenergien von Wasser, Azetat und Essigsäuremit der Dickit 001 Oberfläche (alle Energien in kcal/mol)

System ONIOM ONIOM(B3LYP/SVP: B3LYP/SVP (B3LYF/SVP+sp: B3LYP/SVP+sp

MNDO) MNDO)

D(O)-Hzoa -10.02 -10.47 -7.17 -8.29D(O)-Acb -73.00 -70.08 ·69.67 -67.13

D(T)-HzÜc -4.51 -3.82 -4.73 ·4.14D(T)-HAcd -2.92 -2.64 -4.06 -2.79

a Dickite(occahedralside)-HzO, b Dickice(occahedral sidel-Ac', C Dickice(cecrahedral side).HzO; d Dickice(cecrahedral side)-HAc

Die Bodenkultur 141 52 (2) 2001

M. H. Gerzabek, A. J.A. Aquino, G. Haberhauer, D. Tunega and H. Lischka

Figure 4: Dickire (ocrahedrall-Ac system. ONIOM(B3LYP/SVP:PM3) optimised geometry. Top(lefi) and sidefright) view

Abbildung 4: Das Dicklt (oktahedral)- Ac' System: ONIOM(B3LYP/SVP:PM3) optimierte Geometrie. Aufsicht (links) undSeitenansicht (rechts)

Figure 5: Dickite (tetrahedral).H20 system. ONIOM(B3LYP/SVP:PM3) optimised geometry. Top(left) and sidetrighr) view

Abbildung 5: Das Diekir (rerrahedral)-Hp System: ONIOM (B3LYP/SVP:PM3) optimierte Geometrie. Aufsicht (links) und Sei­tenansicht (rechts)

From the caleuIated Interaction energies it is dear that therwo 001 surfaees (octahedral and rerrahedral) of the 1:1kaolinite layer differ significantly, The tetrahedral sideforms onIy weak hydrogen bonds (energiesabout -4 keal/molfor water and aeetic acid, respectively). On the other hand,the octahedral side covered by hydroxyl groups is involvedin the relatively strong hydrogen bonds (about -8 kcal/rnolfor water) as one can see also from the shorter respectivebond distances. Figure 3 also shows that one water mole­eule has more possibilities to form hydrogen bonds on theoctahedral side than on the tetrahedral one (see Fig. 5).Theyare even stronger beeause of eooperative effects. TheONIOM approach turned out to be a powerful and effee­tive method for investigations of the interactions of rela­tively Iarge moleeular systems interacting with day rnine­ral surfaees.

groups. Three bond lengths of the hydrogen bridges are~ 1.75 Awhile the fourth one is much larger (~2.00 A). Theinteraction energies are much bigger (in absolute value)than those for the D(O)-H20 system because of the netcharge of the acerare anion. As in rhe previous case, theONIOM approximation has onIy a small influence. Therelative difference berween B3LYP/SVP and B3LYP/SVP+spenergies are even smalier (about 9 %) than before.

The D(T)-H20 system (Figure 5). The water moleculeis not located directly above the centre of the ditrigonaltetrahedral hole but is shifted to the side of the ditrigonalring to form a better contact with two neighbouring oxygenatoms in that ring. The water molecule is oriented with thehydrogen atorns directed towards the basal oxygen atornsand the H-H vector is almost parallel to the plane of theseoxygen aroms, The distances berween the protons and thecontacting oxygen atoms are 2.142 Aand 2.515 A, respec­tively.This large bond distances indicate weaker interactionthan in the D(O)-H20 case. CalcuIated interaction ener­gies (Table 2) vary wirhin one kcal/mol. The average valueis about -4.2 kcal/mo1. As in rhe previous cases ONIOMenergies do not differ substantially from the full benchmarkcalculations.

Tbe D(T)-HAc system (Figure 6). The interaetion of theacetic acid molecule with the tetrahedral surface is very simi­lar to the previous case, The molecuIe is positioned almostdirecdy above the centre ofthe ditrigonal hole and forms oneweak hydrogen bond with one basal oxygen atom with abond length of 1.940 A. The calculated interaction energiesareslightly lower (in the absolute value) than those for the sys­tem D(T)-H20. The variation in inreraction energiesbetween different merhods isalsosirnilar to the previous ease.

Figure 6:

Abbildung 6:

Dickire (retrahedrall-Hac sysrem, ONIOM(B3LYP/ SVP:PM3) optimised geometry.Top(left) and sideiright) viewDas Diekir (retrahedral)- HAc System: ONIOM(B3LYP/SVP:PM3) optimierte Geometrie. Aufsicht (links) undSeitenansicht (rechts)

Die Bodenkultur 142 52 (2) 2001

Molecular modelling - opportunities far soil research

4. Conclusions

Basedon the reviewed literature it may be summarized thatcomputational chemistry is an emerging science wirhin soilresearch. Interesting examples of its application werealready presented in literature, especially in the field ofstructural investigations ofsolid soil constituents,

The studies presented in this paper support that:• inreractions of small molecules (Al-acetate, Al-oxalate)

can be calculated by DFT-methods with relatively high

accuracy• different humic substance models as basis for further

investigations can be derived by molecular modellingmethods based on structural information provided byvarious analytical investigations

• linking molecular structural information of organic pol­lutants to their adsorption behaviour in soil might pro­vide new insights in sorption mechanisms

• the development ofduster models seems to be a prornis­ing way for the investigation ofmolecular interactions inlarger systems like the day mineral/soil water interface.

Acknowledgements

This work was supported by the Austrian Science Fund,project no. PI2969-CHE.

References

AHLRlCHS, R., M. BÄR, M. HÄSER, H. HORN and C KÖL­MEL (1989): Electronic structure calculations on work­starion computers: the program systern TURBOMOLE.Chem. Phys. Lett, 162, 165-169.

AHLRlCHS, R., S. D. ELUOT and U. HUNlAR (1998): Quan­tum Chemistry: large molecules - small computers. Ber.Bunsenges. Phys. Chem. 102,795-804.

AQUINO, A, D. TUNEGA, G. HABERHAUER, M. GERZABEKand H. LISCHKA (2000): A density functional theoreticalstudyon solvated Al3+-oxalate complexes: structures andthermodynamic properties. Phys. Chem. Chem. Phys. 2,2845-2850.

ARNIM, M. v. A (1998): Performance of parallel TURBO­MOLE for density functional calcularions, J. Comp.ehern. 19,1746-1757.

BIRKETT,J. W:,M. N. JONES, N. D. BRYAN and E R. LIVENS(1997): Computer modelling of partial specific volumes

of humic substances. European J. Soil Science 48, 131­137.

BORN, M. (1920): Volumen und Hydrararionswärrne derIonen. Z. Phys. 1, 45-48.

Bovs, S. F. and E BERNARDI (1970): The calculation ofsmall molecular interactions by the differences ofseparatetotal energies. Some procedures with reduced errors. Mol.Phys. 19,553-566.

BRlDGEMAN, C H. and N. T. SKlPPER (1997): A MonteCarlo study ofwater at an uncharged day surface. J. Phys.Condens.-Mat. 9,4081-4087.

BRlDGEMAN, C H.,A D. BUCKINGHAM, N. T. SKlPPERandM. C PAYNE (1996): Ab initio total energy study ofuncharged 2:1 days and their interaction with water.Mol. Phys. 89, 879-888.

BROWN, T.1. and ], A RICE (2000): Effectofexperimentalparameters on the ESI FT-ICR mass spectrurn of fulvicacid. Anal. Chem. 72, 384-390.

CHANG, E-R. C, N. T. SKlPPER and G. SPOSITO (1997):Monte Carlo and Molecular Dynamics Simulation ofInterfacial in Lithium- Montmorillonite Hydrates. Lang­muir 13, 2074-2082.

CHANG, E-R. C, N. T. SKlPPER and G. SPOSITO (1998):Monte Carlo and molecular dynamics simulations ofelectrical double-layer structure in potassium-rnontmo­rillonite hydrates, Langmuir 14, 1201-1207.

CHANG, E-R. C, N. T. SKlPPER, K. REFSON, J. AGREATHOUSE and G. SPOSITO (1999): Interlayer molecu­lar structure and dynamics in Li-, Na-, and Kmonrmoril­lonite-water systems. In: D. 1. SPARKS and T. J. GRUNDL(eds.):Mineral-Water Interfacial Reactions - Kinetics andMechanisms. ACS Symposium Series 715,88-106.

CHATTER]EE, A, T. IWASAKl, T. EBINA and H. HAVASHI(1997): Quantum chemical calculation on clay-waterinterface. Applied Surf. Sdence 1211122, 167-170.

CONTE, P. and A. PICCOLO (1999): Conformationalarrangement of dissolved humic substances. Influence ofsolution composition on association ofhumic molecules.Environ. Sei. Technol. 33, 1682-1690.

DAPPRlCH, S., 1. KOMAROMI, K. S. BVUN, K. MOROKUMAand M. J. FRISCH (1999): A new ONIOM implementa­tion in Gaussian98. Part 1. The calculation of energies,gradients, vibrational frequencies and electric field deri­vatives. J, MoL Structure (Theochem) 461-462, 1-21.

DAVIES, G., A. FATAFTAH, A CHERKASSKlV, E. A GHAB­BOUR, A. RADWAN, S. A.JANSEN, S. KOLLA, M. D. PACI­OLLA, 1. T. SEIN, w: BUERMANN, M, BALASUBRAMANlAN,J. BUDNICK and B. XING (1997): Tight rnetal binding by

Die Bodenkultur 143 52 (2) 2001

M. H. Gerzabek, A.]. A. Aquino, G. Haberhauer, D. Tunega and H. Lischka

144

humic acids and ist role in biomineralization, J. ehern.Soc., Dalton Trans., 4047-4060.

DEL RIo, J. C. and l? G. HATCHER (1996): Structural cha­racterization of humic substances using thermolysis withtetramethylammonium hydroxide. In: GAFFNEY, MAR­

LEY, CLARK (eds.): Humic and Fulvic acids: Isolation,Structure and Environmental Role, ACS Symp. Ser., 651.American Chemical Sociery, Washington, oe, 78-95.

DELVILLE, A. (1995): Monte Carlo simulation of surfacehydration: An application to claywetting. J. Phys. ehern.99, 2033-2037.

EICHKORN, K., O. TREUTLER, H. ÖHM, M. HÄSER and R.AHLRICHS (1995): Auxiliary basis sets to approximateCoulomb potentials. Chern. Phys, Lett, 240, 283-290.

FEIN, J. B. and ]. E. HESTRIN (1994): Experimental studiesof oxalate complexation at 80° C: Gibbsite, amorphoussilica, and quartz solubilities in oxalate-hearing fluids.Geochim. Cosmochim. Acta 58,4817- 4829.

FÜHR, F., R. J. HANCE, J. R. PLIMMER and J. O. NELSON

(eds.) (1998): The lysimeter concept. ACS Symposium

Series, 699.

GAUSSIAN98 (1998): Revision A.7, FRISCH, M. J.; TRUCKS,

G. ~; SCHLEGEL, H. B.; SCUSERlA, G. E.; ROBB, M. A.;CHEESEMAN, ]. R.; ZAKRzEWSKl, V. G.; MONTGOMERY,

JR.,]. A.; STRATMANN, R. E.; BURANT, J. C.; DAPPRICH,

5.; MILLAM, V; DANIELS, A. D.; KUDIN, K. N.; STRAJN,

M. C.; FARKAS, 0.; TOMASI, J.; BARONE, v, COSSI, M.;CAMMl, R.; MENNUCCI, C. POMELLI, C. ADAMO, S. CLIF...

PORD, J. OCHTERSKI, G. A. PETERSSON, l? Y. AYALA, B.;CU!, Q.; MOROKUMA, K., MALrcK, D. K.; RABUCK, A.D.; RAGHAVACHARI, K.; FORESMAN,]. B.; CIOSLOWSKI,].;

ORTIZ,]. Y.; STEFANOV, B. B.;LIU, G.; LIASHENKO, A.;PISKORZ, E; KOMAROMI, L; GOMPERTS, R.; MARTIN, R.L.; Fox, D. J.; KEITH, T.; AL-1AHAM, M. A.; PENG, C. Y.;NANAYAKKARA, A.; GONZALEZ, ·C..; CHALLACOMBE, M.;GILL, P. M. W; ]OHNSON, B.j eHEN, W; WONG, M. W;ANDRES, J. L.; GONZALEZ, C.; HEAD-GORDON, M.;REPLOGLE, E. S.; POPLE,J. A. GAUSSIAN, Inc., Pittsburgh

PA.GERZABEK, M. H., F. STREBEL and B. TEMMEL (1998):

Plant uptake of radionuclides in lysimeter experiments.Environmental Pollution 99,93-103.

GÜLDEN, M., A. TURAN and H. SEIBERT (1997): Sub­stanzen mit endokriner Wirkung in Oberflächengewäs­sern. Texte, 46/97;Umweltbundesamt, Berlin,

HABERHAUER, G., L. PFEIFFER and M. H. GERZABEK

(2000a):Investigation ofmolecular structure related sorp­tion behavior of phenoxyalkanoic acid -herbicides on

Die Bodenkultur

agricultural soil and soil particle size fractions. J. Agr.Food Chern., in print.

HABERHAUER, G., L. PFEIFFER, M. H. GERZABEK, H.KIRCHMANN, A. AQUINO, D. TUNEGA and H. LISCHKA

(2000b):Response ofSorption Processes ofMCPA to theAmount and Origin of SOM in a Long-Term Field

Experiment. Submitted.HABERHAUER, G., W BEDNAR, M. H. GERZABEK and E.

ROSENBERG (1999): MALDI-TOF-MS Analysis ofHumic substances - a new approach to obtain additionalstructural information? In: DAVIS and GHABBOUR (eds.):

Understanding Humic Substanees: Advanced Methods,Properties and Applications. Royal Chemical Society,

Cambridge, 121-128.]ENSEN, F. (1999): Introduction into computational ehe­

mistry. lohn Wiley.JOSWIG, W and V. A. DRITS (1986): The orientation of the

hydroxyl graups in dickite by X-ray diffraction. N. Jb.Miner. Mh.,19-22.

KIRKWOOD, J. (1935): Statistical mechanics of fluid mix­tures. J. ehern. Phys. 3, 300-313.

KUBICKI,]. D. and S. E. APITZ (1999): Models of naturalorganie matter and interactions with organic contami­nants. Organic Geochemistry 30, 911-927.

KUBICKI, J. D., G. A. BLAKE and S. E. APITZ (1996): Mo­leeular orbital models ofaqueous aluminum-acetate com­plexes. Geochim. Cosmoehim. Acta 60,4897-4911.

KUBICKI, J. D., G. A. BLAKE and S. E. APITZ (1999): Mo­lecular models of benzene and selected polycyclic aro­matic hydrocarbons in the aequeous and adsorbed states.Environmental Toxicology and Chemistry 18, 1656­1662.

KUBICKI, J. D .. , D. SYKESand S. E. APITZ (1999): Ab initiocalculation of aqueous aluminum and aluminum-car­boxylate complex, energetics and 27Al NMR chemical

shifts. J. Phys. ehern. A 103, 903-915.LEACH, A. R. (1996): Molecular Modelling Principles and

Applicarions. Addison Wesley Longman Limited, Eding­burgh Gate.

LOHNINGER, H. (1994): Estimation ofsoil partition coeffi­cients of pesticides from their chemical structure.Chemosphere 29, 1611-1626.

MIERTUS, S., E. SCROCCO and]. TOMASI (1981): Electro­static interaction of a solute with a continuum. A directutilization of ab initio molecular potentials for rhe previ­sion of solvent effects. ehern. Phys. 55, 117-129.

ONSAGER, L. (1936): Electric moments of moleeules in li­quids, J.Am. ehern. Soc. 58,1486-1493.

52 (2) 2001

Molecular modelling - opporruniries for soil research

PALMER, D. A. and J. L. S. BELL (1994): Aluminum speci­ation and equlibra in aqueous solution: IV: A potentio­

metrie study ofaluminum acetate complexation in acidic

NaCl brines to 150 0 C. Geochim. Cosmichim. Acta 58,

651-659.PARK, S., A. FITCH and Y. WANG (1997): Computarional

studies compared to electrochemical measurements of

intercalation of cationic compounds in Wyoming mont­

morillonite. J. Physical Chemistry BIO 1, 4889-4896.

PELMENSCHIKOV, A. and J. LESZCZYNSKI (1999): Adsorp­

tion of 1,3,5-trinitrobenzene on the siloxane sites of day

minerals: Ab initio calculations of molecular models. ].

Phys, Chem. B 103, 6886-6890.

SCHACHTSCHABEL, P., H. P. BLUME, G. BRÜMMER, K.-H.HARTGE und U. SCHWERTMANN (1989): Lehrbuch derBodenkunde. Ferdinand Enke Verlag, Sruttgart,

SCHNITZER, M. and H.-R. SCHULTEN (1992): The analysis

of soil organic matter by pyrolysis-field ionization mass

spectrometry, Soil Sei. Soc. Am.J. 56, 1811-1817.

SCHULTEN H. R. and P. LEINWEBER (2000): New insights

inro organic-mineral particles: composition, properties

and models of molecular structure, Biology and Ferrility

of Soils 30, 399-432.

SCHULTEN, H. R. (1996): Three-dimensional, molecular

struetures of humic acids and their interactions with

water and dissolved contaminants. Intern. J. Environ.

Anal. Chem. 64,147-162.

SCHULTEN, H. R. (1999): Inreractions ofdissolved organic

matter wich xenobiotic compounds: molecular modelingin water, Environmental Toxicology and Chemisrry 18,

1643-1655.

SCHULTEN, H. R. and M. SCHNITZER (1993): Astate ofthe

art structural concept for humic substances. Naturwis­

senschaften 80, 29-30.

SCHULTEN, H. R. and M. SCHNITZER (1995): Three­

dimensional models for humic acids and soil organic mat­

ter. Naturwissenschaften 82, 487-498.

SCHULTEN, H. R. and M. SCHNITZER (1997): Chemical

model structures for soil organic matter and soils. Soil Sei.162, 115-130.

SEIN, L. T., J. M. VARNUM and S. A. JANSEN (1999): Con­

formational modeling of a new building block of humic

acid: approaches to the lowest energy confomer. Environ.

Sei. Technol. 33, 546-552.SHROLL, R. M. and D. E. SMITH (1999): Molecular dyna­

mies simulations in the grand canonical ensemble: Appli­

cation ro day mineral swelling. J. Chem. Phys. 111,

9025-9033.

SMIRNOV, K. S. and D. BOUGEARD (1999): A moleculardynamics srudy of structure and short-time dynamics of

water in kaolinite. J. Phys. Chem. B 103, 5266-5273.SOCCI, N. D.,J. N. ONUcHICandP. WOLYNES (1998): Pro­

tein folding mechanism and the multidimensional fold­

ing funnel. Pro teins: Structure, Function and Genetics

32, 136-158.SOLOULKI, T., M.A. FREITAS andA.ALoMARY (1999): Gas­

phase hydrogen/deuterium exchange reactions of fulvic

acids: An electrospray ionization Fourier transform cydo­

tron resonance mass spectral study. Anal. Chem. 71,

4719-4726.SPOSITO, G., N. T. SKIPPER, R. Strrrox, S.-H. PARK, A. K.

SOPER and ]. A. GREATHOUSE (1999): Surface geoche­

mistry of the day minerals. Proc. Natl, Acad. Sei. USA

96, 3358-3364..STRUMPF, T. (1998): Ein Strukturmodell für den Aufbau

von Humus zur Beschreibung von Wechselwirkungen

nicht-extrahierbarer (gebundener) Rückstände von

Pflanzenschutzmitteln in Humus. Agribiol. Res. 51,

357- 368.TEPPEN, B. J., c.u. Yu, D. M. MILLER and L. SCHÄFER

(1998): Molecular dynamies simulations of sorption of

organic compounds at the day mineral/aqueous solution

interface. J. Computational Chemistry 19, 144-153.

TOMASI, J. and M. PERSICO (1994): Molecular interactions

in solution: an overview ofmethods based on continuous

distributions of the solvent. Chem. Rev. 94,2027-2094.

TUNEGA, D., G. HABERHAUER, M. H. GERZABEK and H.

LISCHKA (2000): ): Interaction of acetate anion with

hydrared Al" cation: A theoretical srudy. ]. Phys. Chem.A 104, 6824-6833.

VATRAS, 0., ]. ALMLÖF and M. W. FEYEREISEN (1993):

Integral approximations for LCAO-SCF calculations.

Chem. Phys. Lett. 213, 514-518.

VON WANDRUSZKA, R. (1998): The micellar model of

humic acid: evidence from pyrene fluorescence measure­

ments, Soil Seience 163, 921-930.

WERSHAW, R. L., K. A. THORN and D.]. PICKNEY (I 986):

Application ofa membrane model to the secondary struc­

ture of humic marerials in peat. In: FUCHSMANN (ed.):

Peat and Water. Elsevier Applied Seience Publisher Ltd,

133-157.

WERSHAW. R. L. (1993): Model for humus. Environ. Sei.Technol. 27,814-816

ZHANPEISOV, N. U., J. W. ADAMS, S. 1. URSON, C. A

WEISS, B. Z. ZHANPEISOVA, D. LESZCZYNSKA and J.

LESZCZYNSKI (1999): Cluster quantum chemical study of

Die Bodenkultur 145 52 (2) 2001

M. H. Cerzabek, A. J. A. Aguino, G. Haberhauer, D. Tunega and H. Lischka

triaminotoulene interaction with a model clay surface.Struct, ehern. 10,285-294.

Addresses of authors

Univ. Prof. D.1. Dr. Martin H. Gerzabek, D.1. Dr. GeorgHaberhauer, Department of Environmetal Research, Aus­trian Research Center Seibersdorf A-2444 Seibersdorf;e-mail: martin.gerzabekts'arcs.ac.arUniv. Prof. Dr, Hans Lischka, Institute for TheoreticalChemistry and Structural Biology, University of Vienna,Währingerstrasse 17, A-l 090 Vienna;Dr.. AdeliaAquino und Dr, Daniel Tunega sind an heidenInstitutionen tätig.

Eingelangt am 21. August 2000Angenommen am 12. Januar 2001

Die Bodenkultur 146 52 (2) 2001