Article pubs.acs.org/JPCA
All-Atom Force Field for Molecular Dynamics Simulations on Organotransition Metal Solids and Liquids. Application to M(CO)n (M = Cr, Fe, Ni, Mo, Ru, or W) Compounds Carlos E. S. Bernardes,*,†,‡ José. N. Canongia Lopes,‡ and Manuel E. Minas da Piedade† †
Centro de Química e Bioquímica e Departamento de Química e Bioquímica, Faculdade de Ciências, Universidade de Lisboa, 1749-016 Lisboa, Portugal ‡ Centro de Química Estrutural, Instituto Superior Técnico, Universidade de Lisboa, 1049-001 Lisboa, Portugal S Supporting Information *
ABSTRACT: A previously developed OPLS-based all-atom force field for organometallic compounds was extended to a series of first-, second-, and thirdrow transition metals based on the study of M(CO)n (M = Cr, Fe, Ni, Mo, Ru, or W) complexes. For materials that are solid at ambient temperature and pressure (M = Cr, Mo, W) the validation of the force field was based on reported structural data and on the standard molar enthalpies of sublimation at 298.15 K, experimentally determined by Calvet-drop microcalorimetry using samples corresponding to a specific and well-characterized crystalline phase: ΔsubH°m = 72.6 ± 0.3 kJ·mol−1 for Cr(CO)6, 73.4 ± 0.3 kJ·mol−1 for Mo(CO)6, and 77.8 ± 0.3 kJ·mol−1 for W(CO)6. For liquids, where problems of polymorphism or phase mixtures are absent, critically analyzed literature data were used. The force field was able to reproduce the volumetric properties of the test set (density and unit cell volume) with an average deviations smaller than 2% and the experimentally determined enthalpies of sublimation and vaporization with an accuracy better than 2.3 kJ·mol−1. The Lennard-Jones (12-6) potential function parameters used to calculate the repulsive and dispersion contributions of the metals within the framework of the force field were found to be transferable between chromium, iron, and nickel (first row) and between molybdenum and ruthenium (second row).
1. INTRODUCTION
An important point is the fact that the development and validation of the force field has been supported by experimental ° ) or determinations of enthalpies of vaporization (ΔvapHm sublimation (ΔsubHm° ) of the studied compounds, taking particular care in the characterization of the solid samples in terms of phase purity. Indeed, although the structural features of the force field can be assessed from an ample structural database,17 the same is not true for the energetic part because reliable ΔsubH°m and ΔvapH°m data that can be used as measures of the cohesion energy of the condensed phases are very scarce18−20 and few of the available ΔsubH°m values can be safely assigned to a definite crystal structure.18−20 In the present work the force field was expanded to a series carbonyl compounds of first-, second-, and third row transition metals. For materials that are solid at ambient temperature and pressure (T = 298.15 K, p = 1 bar) the validation of the force field was based on ΔsubH°m values obtained by performing Calvet-drop microcalorimetry measurements on samples corresponding to a single crystalline phase. For liquids, where
Molecular dynamics and Monte Carlo simulations based on judiciously developed force fields can be very useful to rationalize and predict the properties of organotransition metal molecular solids and liquids. Yet, most efforts to develop force fields for this class of compounds have essentially been focused on the calculation of molecular structures and intramolecular interactions and not so much on the volumetric or energetic properties of the materials as a whole, such as densities or lattice energies.1−9 This led us to recently initiate the development of an all-atom force field for molecular dynamics (MD) and Monte Carlo simulations of transition metal organometallic compounds to be used within the framework of statistical mechanics.10,11 Because general applicability was sought, the setup of the force-field has been progressing by an incremental approach: first ferrocene was studied;10 then solid and liquid compounds of the ferrocene family (dimethylferrocene, decamethylferrocene, acetylferrocene, and diacetylferrocene) were addressed;11 and later, cyclopentadienyltricarbonylmanganese, Mn(η5C5H5)(CO)3, was considered.12 In all cases the parametrization of most organic residues (methyl, acetyl, etc.) was borrowed from the widely used OPLS-AA/AMBER force field.13−16 © XXXX American Chemical Society
Received: August 2, 2013 Revised: September 24, 2013
A
dx.doi.org/10.1021/jp407739h | J. Phys. Chem. A XXXX, XXX, XXX−XXX
The Journal of Physical Chemistry A
Article
Table 1. Force Field Parameters for M(CO)n, with M = Cr, Fe, Ni, Mo, Ru, and Wa atoms
q/a.c.u.
atoms
q/a.c.u.
atoms
ε/kJ·mol−1
σ/Å
bonds
ro/Å
angles
θ/deg
Cr Fe Ni Mo Ru W CCr OCr CMo OMo CW
−0.690 −0.658 −0.288 −0.186 −0.313 −0.072 0.319 −0.204 0.194 −0.163 0.172
OW CFe * O*Fe CFe OFe C*Ru O*Ru CRu ORu CNi ONi
−0.160 0.459 −0.226 0.246 −0.182 0.393 −0.202 0.117 −0.140 0.229 −0.157
Cr Fe Ni Mo Ru W C O
1.200 1.200 1.200 1.55 1.55 2.60 2.76 0.60
3.11 3.11 3.11 3.40 3.40 3.40 3.30 3.09
CrC FeC FeC* NiC MoC RuC RuC* WC CO
1.889 1.792 1.801 1.805 2.046 1.937 1.940 2.056 1.164b
MCO CFeC C*FeC* CFeC* CRuC C*RuC* CRuC* CNiC CMC
180.0 90.0 180.0 90.0 90.0 180.0 90.0 109.47 90.0
a The asterisk denotes a carbon or oxygen atom in an axial position (see text for details). bAverage CO distance computed from all DFT molecular structural parameters obtained in this work.
11.092(11) Å, c = 6.332(6) Å.23 The onset (Ton) and maximum (Tmax) temperatures and enthalpy (ΔfusH°m) of the fusion peak measured by DSC, at a heating rate β = 5 K·min−1, were Ton = ° = 25.2 ± 426.57 ± 0.04 K, Tmax = 428.05 ± 0.25 K, and ΔfusHm 1.5 kJ·mol−1 (average of two determinations; the quoted uncertainty is twice the mean deviation), respectively. The obtained Ton and ΔfusH°m are higher than those previously reported: Tfus = 423.15 ± 2.0 K24 and Tfus = 422−423 K25 and ° = 23.4 ± 0.4 kJ·mol−1.24 ΔfusHm Mo(CO)6. Molybdenum hexacarbonyl ([CAS Registry No. 13939-06-5], Aldrich, 99.9+%) was used without further purification. Elemental analysis for Mo(CO)6: expected C 27.30%; found C 27.63 ± 0.57% (average of three determinations; the uncertainty quoted is twice the mean deviation). 13C NMR (400 MHz, CDCl3): δ 201.2 ppm (C O). No NMR peaks corresponding to impurities were observed in the spectra. The obtained NMR results are in excellent agreement with previously reported results (δ 204.1 ppm).22 The powder pattern recorded at 293 ± 2 K was indexed as orthorhombic, space group Pnma, with a = 12.009(2) Å, b = 11.405(3) Å, and c = 6.485(1) Å. These values are in good agreement with published information from single crystal X-ray diffraction experiments carried out at ambient temperature (283−303 K): a = 12.019(2) Å, b = 11.415(2) Å, c = 6.488(1) Å.26 The onset and maximum temperatures and enthalpy of the fusion peak measured by DSC, at a heating rate β = 5 K·min−1, were Ton = 421.82 ± 0.10 K, Tmax = 423.29 ± 0.42 K, and ° = 25.1 ± 0.4 kJ·mol−1 (average of two determinations; ΔfusHm the indicated uncertainty is twice the mean deviation), respectively. The obtained Ton and ΔfusH°m are compatible with those previously reported: Tfus = 419.15 ± 2.0 K24 and Tfus = 423.15 K25 and ΔfusHm ° = 26.8 ± 0.4 kJ·mol−1.24 W(CO)6. Tungsten hexacarbonyl ([CAS Registry No. 1404011-0], Aldrich, 99.9+%) was used as received. Elemental analysis for W(CO)6: expected C 20.48%; found C 20.73 ± 0.26% (average of two determinations; the uncertainty corresponds to twice the mean deviation). The powder pattern recorded at 293 ± 2 K was indexed as orthorhombic, space group Pnma, with a = 11.950(2) Å, b = 11.385(3) Å, and c = 6.450(1) Å. These results are in good agreement with those previously obtained from a single crystal X-ray diffraction analysis performed at ambient temperature (283−303K): a = 11.944(1) Å, b = 11.370(1) Å, c = 6.459(1) Å.27 The fusion peak detected by DSC led to Ton = 442.69 ± 0.07 K, Tmax = ° = 26.39 ± 1.54 kJ·mol−1 (average 444.63 ± 0.75 K, and ΔfusHm of two determinations; the uncertainty quoted is twice the
problems of polymorphism or phase mixtures are absent, critically analyzed literature data were used.
2. MATERIALS AND METHODS 2.1. General Information. All samples were handled under a nitrogen atmosphere using standard Schlenk techniques. Elemental analysis was performed on a CE-Instruments EA1110 CHNS-O automatic analyzer. X-ray powder diffraction patterns were obtained on a Philips PW1730 diffractometer, with automatic data acquisition (APD Philips v.35B), operating in the θ−2θ mode. The apparatus had a vertical goniometer (PW1820), a proportional xenon detector (PW1711), and a graphite monocromator (PW1752). A Cu Kα radiation source was used. The tube amperage was 30 mA and the tube voltage 40 kV. The diffractograms were recorded at ∼293 ± 2 K in the range 5° < 2θ < 35°. Data were collected in the continuous mode, with a step size of 0.015° (2θ) and an acquisition time of 1.5 s/step. The samples were mounted on an aluminum sample holder. The indexation of the powder patterns was performed using the program Checkcell.21 Differential scanning calorimetry (DSC) measurements were carried out on a DSC 7 from Perkin-Elmer. The experiments were performed at a heating rate of 5 K·min−1 under a nitrogen (Air Liquide N45) flow rate of 0.5 cm3·s−1. The temperature and heat flow scales of the instrument were calibrated at the same heating rate with indium (Perkin-Elmer; 99.999%; Tfus = 429.75 K, ΔfusH° = 28.45 J· g−1). The samples, with masses in the range 3−13 mg, were sealed in air, inside aluminum crucibles, and weighed with a precision of ±0.1 μg on a Mettler XP2U ultramicrobalance. The 13C NMR spectra were obtained in CDCl3 (Aldrich 99.80%,
