J. van Eerden

Molecular-dynamics simulation of crystalline 18-crown-6: thermal shortening of covalent bonds

Molecular-dynamics simulations of crystalline 18-crown-6 have been performed in a study of the apparent thermal shortening of covalent bonds observed in crystal structures. At 100 K, a shortening of 0.006 _+ 0.001 A for C----C and C----O bonds was obtained. This result was found to be independent of details of the force field and the simulation. There was agreement between computational and experimental values for the thermal parameters, as well as for the molecular geometry (bond and dihedral angles) of 18-crown-6. Some differences are attributed to the inability of the force field to reproduce hydrogen-bonding geometries. Simulation at 295 K resulted in an estimated shortening of 0.019_+ 0.005 A. Thus at room temperature for C--C bonds (apparent) thermal shortening and (real) chemical shortening, resulting from the electronegative oxygen substituents, are of the same order of magnitude. In the simulation at 295 K occasional dihedral transitions were observed, which may reflect the proximity of the melting point (312 K).

Modelling of the diffusion of carbon dioxide in polyimide matrices by computer simulation

Computer aided molecular modelling is used to visualize the motion of CO2 gas molecules inside a polyimide polymer matrix. The polymers simulated are two 6FDA-bases polyimides, 6FDA-4PDA and 6FDA-44ODA. These polymers have also been synthesized in our laboratory, and thus the simulated properties could directly be compared with “real-world” data. The simulation experiments have been performed using the GROMOS1 package. The polymer boxes were created using the soft-core method, with short (11 segments) chains. This results in highly relaxed and totally amorphous polyimide matrices. The motion of randomly placed CO2 molecules in the boxes during molecular dynamics runs was followed, revealing three types of motion: jumping, continuous- and trapped motion. The calculated diffusivities are unrealistic, but possible shortcomings in our model are given.

Potential of mean force by thermodynamic integration: molecular-dynamics simulation of decomplexation

“Umbrella sampling” has been incorporated in the thermodynamic integration method to obtain a potential of mean force by slow growth molecular-dynamics simulations. The method was tested for liquid argon, for which good agreement was obtained with a standard potential of mean force, as derived from the radial pair-correlation function. For a sodium chloride ion-pair in aqueous solution the calculations showed resonable agreement with a literature result. The method was also applied to the decomplexation of 18-crown-6 and a potassium cation in aqueous solution.

Molecular-dynamics simulations of interfaces between water and crystalline urea

Molecular-dynamics simulations of several water-crystalline urea interfaces have been performed. The structure and dynamics of water close to the urea crystal surface are discussed in terms of density profiles, positional and orientational distribution functions, and diffusion coefficients. The water structure close to the interface is strongly determined by the structure of the crystal surface: the (001) and (111) interfaces reveal strong adsorption of water while the (110) and () interfaces do so to a lesser extent. Assuming that the growth rate of a specific crystal face decreases with increasing solvent adsorption, the appearance of only (111) on the urea growth form is predicted. We argue that on the other hand the dominance of (110) over (001) cannot be explained using a simple layer growth model.

Structure of 2,6-pyrido-18-crown-6–guanidinium perchlorate–deuterochloroform

C 15H23NOs.C H6N+.CIO4.CDC13, M r = 577.27, orthorhombic, Pna21, a= 10.799(1), b-22.671 (5), c = 10.561 (2)A, V = 2586 (1)A3, Z = 4 , D x = 1.48 g cm -3, ;L(Mo K~t) = 0.71069 A, /z(Mo Kt~) = 5 . 1 c m -~, F(000)=1200, T = I 6 8 K , final R = 4.7% for 1762 observed reflections. Each guanidinium cation is hydrogen-bonded to two 2,6-pyrido-18-crown6 molecules, as the macrocyclic cavity of one crown molecule cannot encapsulate the cation completely. Similarly, each crown molecule is hydrogen-bonded with two cations. As a result, the structure consists of chains with an alternating sequence of crown molecules and cations. The perchlorate anion is involved in short contacts with two 2,6-pyrido-18-crown-6 molecules and one solvent molecule of deuterochloroform. Experimental. The title compound was obtained in an extraction experiment. A solution of 1 mmol of 2,6pyrido-18-crown-6 in 2 ml CDCI 3 was equilibrated with a solution of 2 mmol of guanidinium sulfate and 2 mmol LiC10 4 in 2 ml H20. The organic layer was separated off and the amount of guanidinium perchlorate that was extracted into the organic phase was determined from the intensities in the ~HNMR spectrum. Only 0.32 mmol of guanidinium perchlorate proved to be transferred. Upon addition of 0.5 ml of * IUPAC name: 3,6,9,12,15-pentaoxa-21-azabicyclo[15.3.1]henicosa1 (21), 17,19-triene. 0108-2701/87/122453-03501.50 accepted 30 June 1987) diethyl ether the complex crystallized and was filtered off; m.p. 353-356 K (Uiterwijk, van Staveren, Reinhoudt, den Hertog, Kruise & Harkema, 1986). Intensities were measured at 168 K on a Philips PW l l00 diffractometer (MoK~t radiation, graphite monochromator)..Lattice parameters determined by least squares from 25 centered reflections (4.5 < 0 < 9.5°). A total of 2404 independent reflections up to 0 = 2 5 ° ( 0 < h < 1 2 , 0 _ < k < 2 6 , 0 < l < 1 2 ) were measured in the 0/20 scan mode (scan speed 0.05 ° s -l, scan width 1.4°); 1762 reflections considered observed [Fo2> 3tr(Fo2)]. The intensity variation of three standard reflections, measured every hour, was less t han 3%. No absorption correction. The structure was solved with MUL TAN (Germain, Main & Woolfson, 1971) and refined by full-matrix least squares. Weights for each reflection in the refinement (on F) were calculated from w=4Fo2/ tr2(Fo2), a2(Fo2)=tr2(1)+ (PFo2)2; the value of the instability factor p was determined as 0.06. All H atoms were located on difference Fourier maps; they were placed in calculated positions and treated as riding on their parent atoms [bond distance 0.96 A, B~so(H ) = 1.2 Beq(parent)]. The number of parameters refined was 308: scale factor, isotropic extinction parameter ]final value 1.1 (6) x 10-7], positional and anisotropic thermal parameters for the non-H atoms. Refinement converged at R =4 .7%, wR = 6.0%, (A/a)max=O.11. Largest peak on final difference