Mecanismo Adsorção Odor Pelo Ricenoleato de Zinco
Mecanismo Adsorção Odor Pelo Ricenoleato de Zinco
Mecanismo Adsorção Odor Pelo Ricenoleato de Zinco
ABSTRACT: Zinc ricinoleate [Zn(Ri)2] is widely used in surfac- industry. Most of the odor-active materials include chemi-
tant and detergent mixtures for the adsorption of odor-active cally reactive functional groups. Organic oxygen, sulfur,
compounds. The mechanism of this process is not known. In this and nitrogen compounds are the predominant species
initial study, we discuss the results of molecular dynamics com- leading to odor in gaseous materials (1).
puter simulations that were performed to get more information In industrial applications, silicate and charcoal filtration
and detailed insights into the interaction mechanism between
are often used to prevent unpleasant odors. The effect is
Zn(Ri)2 and odor-active substances. The calculations, based on
that odor-active materials are bound by adsorption to the
simple molecular mechanics approximations, simulated the dy-
namic features of the molecular structures of Zn(Ri)2 in vacuum, inner surface of the charcoal silicates.
in the oil phase, and in aqueous solution. We determined actual Application of zinc ricinoleate [9-octadeceonic acid, 12-
molecular conformations and simulated an adduct of ammonia hydroxy-, zinc salt (2:1), (9Z,12R)] as an odor-removing
with Zn(Ri)2. On close inspection, in the vacuum and oil phase agent is well-known (2–19). The chemical structure is given
structures, the Zn2+ ion is almost completely shielded by the oxy- in Figure 1, and for simplification we denote zinc rici-
gen ligands. Calculated structural transitions caused by the inter- noleate as Zn(Ri)2 in the subsequent discussion.
action of Zn(Ri)2 with water-solvent molecules resulted in a This substance, commercially available as TEGO®Sorb
weakening of the electrostatic shield. Nucleophilic attack of (Goldschmidt AG, Essen, Germany), is widely used as a
odor-active compounds to the relatively unprotected Zn2+ atom deodorizing agent in cosmetic and hygienics applications
is easy to achieve in aqueous solution. Simulation of the addi-
(2–14). The compound is produced by alkaline saponifica-
tion product of Zn(Ri)2 with ammonia revealed an elementary
tion of technical ricinus oil and subsequent reaction with
structural change, resulting in an increase of the solubility and
adsorption activity of Zn(Ri)2. Molecular dynamics simulations aqueous zinc sulfate solution. To some extent Zn(Ri)2 be-
showed that the results coincide with experimental observations. haves like a surfactant molecule and is compatible with
Paper no. S1130 in JSD 3, 335–343 (July 2000). other surfactants and detergents. Dissolution in water
without addition of solubilizers is difficult.
KEY WORDS: Molecular dynamics simulation, molecular Zn(Ri)2 offers a crucial advantage because of excellent
structure, odor-adsorption, zinc ricinoleate. selectivity with sulfur and nitrogen compounds (18). Fur-
thermore, Zn(Ri)2 is able to bind these molecules chemi-
cally with a consequence that unpleasant odor-active sub-
With comprehensive industrialization and production, stances are removed from the ambient air. Due to its im-
foul-smelling wastes are increasing steadily in exhaust air portance in the technology of specific odor-adsorption, the
and sewage. In the cosmetic area, decomposition products toxicological, pharmacological, and dermatological prop-
appear and often can be removed with odor-deleting de- erties were examined (14–17). Recently, the adsorption of
odorizing agents. Generally, the active substances that odor-active compounds with Zn(Ri)2 was studied (18,19).
cause intensive odor are compounds with low odor thresh- Zn(Ri)2 is available as a viscous oil, and up to now ef-
old values. Inevitably, at a certain concentration, the com- forts made toward crystallization have not been successful
pounds become nasal nuisances. (19). Consequently, X-ray structural analyses have failed.
Through aerobic or anaerobic degradation processes, The molecular structure of Zn(Ri)2 is so far unknown, lead-
odor-active substances are released to the environment. ing to the fact that the mechanism of odor depletion is not
Products of bacterial decomposition processes are present
in human perspiration, in sour gasoline, in natural gas, in
paper facilities, in the businesses of waste and sewage
treatment, and in the agriculture and the food-processing
*To whom correspondence should be addressed at University of
Essen, Department of Physical Chemistry, Universitaetsstrasse 3-5,
D-45141 Essen, Germany. E-mail: hubert.kuhn@uni-essen.de FIG. 1. Chemical formula of zinc ricinoleate [Zn(Ri]2.
Copyright © 2000 by AOCS Press Journal of Surfactants and Detergents, Vol. 3, No. 3 (July 2000) 335
336 H. KUHN ET AL.
unequivocally proven. The detailed mechanism of its reac- is a component of the molecular dynamics program
tion with nitrogen or sulfur compounds is discussed with DISCOVER (23). This forcefield was especially developed
differing conclusions (12,13). Furthermore, the distinctive for molecular modeling of organic, inorganic, and organo-
effectiveness of odor-adsorption, especially in aqueous so- metallic systems as well as transition metal complexes. It
lution, cannot be explained. Consequently, the reaction of covers almost all of the elements in the periodic table. ESFF
odor-active substances with Zn(Ri)2 is related to the is primarily used for structural prediction of isolated mol-
molecular structure in solution. Generally, the dynamic ecules as well as condensed phases. Recently, molecular
structures of molecules surrounded by solvent molecules mechanics and molecular dynamics studies have proven
are barely understood, and until recently there have been the reliability of the ESFF forcefield for conformation pre-
no experimental techniques for probing dynamic con- dictions in organometallic complex systems and solubi-
formation adjustments resulting from the kinetic energy lized carbohydrates (24–29). Atomic parameters in the
of the system and the specific interactions with solvent ESFF forcefield have been taken from experiments or are
molecules. generated by specific rules.
Due to recent progress in computer techniques, it is now The total potential energy of the molecular system is cal-
possible to use molecular dynamics computer simulations culated by the sum of intramolecular interaction energies
for a detailed insight into the structural behavior of mole- and the interatomic potential energy contributions. The
cules. Interatomic forces such as hydrogen bonding, van procedure for the calculation of the specific potential en-
der Waals and electrostatic interactions can be calculated ergy terms used in our simulation is given in Equations 1
by appropriate interatomic energy functions. The molecu- and 2.
lar dynamics technique allows sampling of the molecular
∑ ( ) ∑ ( cos θ )
2
0 2
Db 1 − exp 1 + α rb − rb0 +
intra Ka
configuration space and therefore a study of the dynamic Epot = a − cos θ a
sin 2 θ0a
motion of molecules. In this way information is obtained b a
1 2
regarding the time response of molecular systems, which
sin 2 θ ⋅ sin 2 θ
∑
n n
2 + sign sin θ1 ⋅ sin θ2 ⋅ cos nφ
is of great interest to chemists and biochemists. An + Dφ 1
( )
sin θ1 ⋅ sin θ2
2 0 2 0 n 0 n 0
sin θ1 ⋅ sin θ2
overview of currently used computational methods for the φ
calculation of dynamic characteristics on an atomic or mo- 3
lecular scale can be found from several sources (20–22).
In this paper we focus attention on the molecular dy-
+
∑
o
Do χ2 [1]
system, and the interactions between Zn2+ and all other well as potential energy, is necessary but in fact not suffi-
atoms (C, H, O) are described by nonbonded electrostatic cient to verify the structural relaxation of Zn(Ri)2 to the
and van der Waals interactions. Fixed partial atomic thermodynamic equilibrium. It is obvious that only a de-
charges were used to calculate the electrostatic interac- tailed analysis of correlation functions can prove this state-
tions. This is a simplification because polarization effects ment. Additionally, to verify equilibration, in all simula-
are not taken into account. Nevertheless, in the classical tions we ensured that the initial dihedral distributions con-
mechanics model, this approximation is justified. We as- verged to constant distributions in the equilibrium state.
signed an energy cutoff to a distance of 2.0 nm in connec- In the thermodynamic equilibrium, the simulation box
tion with an explicit image model to truncate the range of relaxed to the dimension of Lx = Ly = Lz = 3.83 nm. Equilib-
intermolecular van der Waals and electrostatic interac- rium water density was slightly increased to 1.179 g/cm3,
tions. In practice, these restraints were necessary to reduce somewhat larger than the water density at standard condi-
the amount of computer time. tions. Subsequently, the simulation was continued under
At this point we describe the procedure of generating the conditions of constant temperature, constant volume
molecular starting configurations of the computer simula- of the simulation box, and a time period of 80 ps. Average
tions II. One Zn(Ri)2 molecule was surrounded by 2150 values and structural results were obtained from the
water molecules in a cubic box with the dimension Lx = Ly 135–175 ps range of the final production run.
= Lz = 4.0 nm. In Figure 2 the simulation box is shown con- To simulate adducts of ammonia and Zn(Ri)2, we per-
taining Zn(Ri)2 and water molecules. Periodic boundary formed molecular dynamics simulations of Zn(Ri)2-NH3
conditions were used to include the bulk solvent influence. complexes in which the bond distance between zinc and
Since the molecular arrangement of the water phase was nitrogen was varied from 350 to 180 pm (100 pm ≡ 1 Å).
based on a previous isothermic-isobaric constant NPT Starting from the fully equilibrated aqueous solution of
equilibrium molecular dynamics simulation of a periodic Zn(Ri)2 obtained from simulation II, a water molecule was
boundary box consisting of 1000 water molecules, the ini- exchanged by one NH3 particle representing an odor-ac-
tial water density amounted to 1 g/cm3. tive nitrogen compound. The NH3 molecule was placed in
Total potential energy was minimized to generate a the simulation box with the condition that the N-Zn2+ dis-
well-defined starting point. The minimization process was tance was kept constant during the molecular dynamics
performed with the Steepest Descent method which selects simulation. Different N-Zn2+ distances and simulation
and moves down the steepest slope of the potential energy times are shown in Table 1.
surface of the molecular structure (30). After minimization, Simulation of the constant N-Zn2+ distance is accom-
random velocities, according to a Gaussian distribution, plished by a steep harmonic function that forces the bond
and a temperature of 298 K, were assigned to each atom. to the specified target value given in Table 1. A plot of the
Pressure was set at 105 Pa, temperature was fixed at 298 K total energy vs. the N-Zn2+ distances is shown in Figure 3.
and was held constant by direct velocity scaling. Ander- In Figure 3 it is seen that the minimum of the total energy
sen’s method of pressure control (31) was applied to main- occurs at a N-Zn2+ distance of 260 pm, indicating the most
tain the pressure at a constant value. To calculate the mo- stable complex at this bond distance.
lecular dynamics of the water molecules and of Zn(Ri)2, we
integrated the Newtonian equations of motion of all atoms RESULTS AND DISCUSSION
with the Verlet Velocity Integrator algorithm (32), setting
the integration time step to 1 femtosecond. All calculations First we will discuss the results of simulation I. The poten-
were done on an IBM-SP2 parallel computer (8 CPU units) tial energy as a function of simulation time is plotted in
connected to a Silicon Graphics R4400 workstation. The Figure 4. The minimum represents molecular conforma-
trajectory data were obtained using the molecular dynam- tions with the lowest potential energies (S1–S6). S1 corre-
ics program DISCOVER (23). sponds to the energetically favorable conformation that
Because the starting configuration is far from the ther- was generated at a simulation time of 16 ps. This means
modynamic equilibrium of the dissolved Zn(Ri)2, the sys- that in the absence of water molecules the molecular struc-
tem needed a definite simulation time period for equilibra- ture of S1 may be a representative complex structure.
tion. The main factors for this process are the reorganiza- In Figure 5 the molecular structure of S1 is plotted with
tion of water molecules near the Zn(Ri)2 molecule and spheres obtained from the van der Waals radius for each
generation of an energetically favorable trans-gauche-con- atom. Figure 5 shows that Zn2+ forms a complex with the
formation distribution in the alkyl chains of Zn(Ri)2. Opti- carboxylate and hydroxyl group oxygen atoms. A detailed
mization of the intramolecular interactions like bond analysis of the S1 conformation reveals that the 6 Zn-O
lengths and bond angles was also an important factor. bond distances are in the range of 215–218 pm. Further-
During a 130 ps molecular dynamics simulation with more, the structure is symmetrical with respect to the C1
constant pressure, we initially allow the system to equili- point group. If reaction with odor-active substances like
brate to an energy fluctuation around a constant average NH3 is considered, it is obvious that the Zn2+ ion is almost
value of total energy. Investigation of the total energy, as completely shielded by oxygen ligands. An additional
FIG. 3. Plot of the total energy vs. the N-Zn2+ distance. The bars show
the constant fluctuations of the total energy around their average values.
TABLE 1
Simulated Zn2+-NH3 Bond Distances for the Adduct
of One NH3 Molecule to Zn(Ri)2a
R(N-Zn2+) (pm) t (ps)
350 140
300 140
260 200 FIG. 4. Simulation of the molecule Zn(Ri)2 in the vacuum phase at 298
220 100 K. The potential energy as a function of simulation time is plotted. Con-
180 100 formations S1–S6 possess very low potential energy. See Figure 1 for ab-
a
Zinc ricinoleate = Zn(Ri)2. breviation.
FIG. 7. Simulation of Zn(Ri)2 in aqueous solution. Plot of the bond distances between Zn2+
and the oxygen ligand atoms as a function of simulation time (135–175 ps). Average values
and their standard deviations are also given. See Figure 1 for abbreviation.
ment with the ESFF forcefield results is acceptable. From with two larger O−-Zn2+ bonds with a difference of about
analysis of the vacuum simulation, it is concluded that the 25 pm. Conversely, the ESFF-forcefield calculation sug-
carboxylate groups are perpendicular to each other, which gests a complex with approximately equal O−-Zn2+ bond
also corresponds with the ab initio calculations. Only the distances. Nevertheless, the agreement between ab initio
O−-Zn2+ bond distances are somewhat larger in the Zn(Ri)2 and ESFF is qualitatively good. Additionally, the quantum
molecule under vacuum. The perpendicular arrangement mechanical calculation suggests a disturbance of the com-
of both carboxylate groups is distorted in Zn(Ri)2 in the plex symmetry if NH3 is added to Zn2+ which is sur-
viscous oil phase. Obviously, this comes from bulk interac- rounded by two carboxylate ligands. Actually, this con-
tions between Zn(Ri)2 molecules. firms our simulation results for the Zn2+-NH3 adduct.
Table 3 shows ab initio DZP and ESFF results for the Simulation in vacuum and oil phases means that inter-
model system (HCOO)2Zn(NH3). The ab initio results from actions of Zn(Ri)2 with solvent molecules are not consid-
the model compound (HCOO)2Zn(NH3) show a complex ered. The result is a symmetrical complex under vaccum
FIG. 9. Simulation of the adduct between Zn(Ri)2 and one NH3 molecule in aqueous solution. Plot of the bond distances between Zn2+ and the
oxygen ligand atoms as a function of simulation time (0–140 ps). The distance between Zn2+ and N was set to a constant value of 260 pm during
the simulation. Calculated average values and standard deviations are inserted into the diagrams. See Figure 1 for abbreviation.
Hubert Kuhn (1961) Dr. obtained his degree from the Univer-
sity of Munster (1992) in theoretical organic chemistry. After
working a period at the chamber of commerce in Wuppertal (Ger-
many) in the scope of environmental chemistry, he attended the
University of Essen. In 1995 he established a molecular model-
ing group at the university. His research topics are focused on
the application of computer simulation in colloid chemistry. His
main scientific interest is the molecular dynamics computer sim-
ulation of complex molecular systems in aqueous solution and
interface properties of surfactant films.
Felix Muller (1964) obtained a Ph.D. in organic chemistry
from the University of Munster in 1992. After working as a
FIG. 10. Simulation of the NH3 adduct with Zn(Ri)2 in aqueous solu-
tion. Representative molecular structure corresponding to thermody- management consultant, he joined Th. Goldschmidt AG, Essen,
namic equilibrium and a simulation time of 170 ps. During the simula- in 1993. There he started in the applied research department of
tion, the distance between Zn2+ and N was set to a constant value of the surfactants business unit and currently is technical director
260 pm. The N atom is green-colored, and for better clarity the water of the Home-Care business unit of Th. Goldschmidt AG. His sci-
molecules and hydrogen atoms are not drawn. See Figure 1 for abbrevi-
entific interest is in novel surfactant and emulsion systems, sili-
ation.
cone surfactants, and organic performance additives for home-
care applications.
27. Barlow, S., A.L. Rohl, S. Shi, C.M. Freeman, and D. O’Hare, Jorg Peggau (1962) joined Th. Goldschmidt AG, Essen, in
Molecular Mechanics Study of Oligomeric Models for 1978. From 1981 to 1994 he worked in the research and techni-
Poly(ferrocenysilanes) Using the Extensible Systematic Force-
field (ESFF), J. Am. Chem. Soc. 118:7578 (1996). cal service unit for Goldschmidt’s disinfectant business. In 1987
28. Asensio, J.L., M. Martín-Pastor, and J. Jiménez-Barbero, The he graduated as chemical engineer/technician. Since 1995 he has
Use of the MM3 and ESFF Forcefields in Conformational been working in the applied research lab of the Home-Care busi-
Analysis of Carbohydrate Molecules in Solution: The Methyl ness unit.
α-Lactoside Case, J. Mol. Struct. 395–396:245 (1997). Reinhard Zekorn (1956) started at Henkel KGaA, Dusseldorf,
29. Martins, J.C., R. Willem, and M. Biesemans, A Comparative
Investigation of the Consistent Valence and Extensible Sys- in the technical service skin-care lab in 1977. In 1988 he gradu-
tematic Forcefields. A Case Study on the Conformation of ated as chemical engineer/technician. In 1989 he joined
Erythromycin A in Benzene, J. Chem. Soc., Perkin Trans. 2:1513 GRILLO-WERKE AG, Duisburg, as lab manager technical ser-
(1999). vice and had his first contact with zinc ricinoleate. Since 1995 he
30. Press, W.H., B.P. Flannery, S.A. Teukolsky, and W.T. Vetter- has been working at Th. Goldschmidt AG, Essen. He started in
ling, Numerical Recipes, Cambridge University Press, London,
1986. the technical service skin-care products, and since 1999 he has
31. Andersen, H.C., Molecular Dynamics Simulations at Constant been working as sales manager eastern Europe in the personal-
Pressure and/or Temperature, J. Chem. Phys. 72:2384 (1980). care business unit.