Exploring the Structural Properties of Simple Aldehydes: A Monte

Jun 22, 2009 - The structure of simple linear alkanals from propanal to nonanal was studied utilizing configurational bias Monte Carlo (MC) simulation...
0 downloads 0 Views 397KB Size
Download PDF
J. Phys. Chem. B 2009, 113, 9429–9435

9429

Exploring the Structural Properties of Simple Aldehydes: A Monte Carlo and Small-Angle X-ray Scattering Study Andrej Lajovic,† Matija Tomsˇicˇ,† Gerhard Fritz-Popovski,‡ Luka´sˇ Vlcˇek,§ and Andrej Jamnik*,† Faculty of Chemistry and Chemical Technology, UniVersity of Ljubljana, AsˇkercˇeVa 5, SI-1000 Ljubljana, SloVenia, Institute of Chemistry, UniVersity of Graz, Heinrichstrasse 28, A-8010 Graz, Austria, and Institute of Chemical Process Fundamentals, Academy of Sciences of the Czech Republic, RozVojoVa 135, CZ-165 02 Prague 6, Suchdol, Czech Republic ReceiVed: March 30, 2009; ReVised Manuscript ReceiVed: May 22, 2009

The structure of simple linear alkanals from propanal to nonanal was studied utilizing configurational bias Monte Carlo (MC) simulations of the aldehydes modeled according to the transferable potential for phase equilibriasunited atom force field (TraPPE-UA) and was compared to experimental small-angle X-ray scattering (SAXS) results. This was done by exploiting a recently developed approach for calculating the scattering intensities from theoretically obtained MC data by utilizing the Debye equation (Tomsˇicˇ et al. J. Phys. Chem. B 2007, 111, 1738). Similar calculations were also performed utilizing a well-established approach based on the reciprocal lattice. Comparison of the calculated scattering data with the experimental SAXS results in the first instance revealed information on the molecular organization in simple aldehydes and in addition also served as a good structural test of the TraPPE-UA force field used to model the aldehydes studied. However, it turned out that such a structural test is a rather strict test for the model which otherwise showed good agreement with the experimental data from the thermodynamic point of view. 1. Introduction As an extensive group of organic compounds, aldehydes are in general important substances that are in widespread use in various industrial processes and are also not to be neglected from the health-risk and ecological points of view. They are used for the synthesis of diverse chemicals and pharmaceuticals, as solvents, disinfectants, repellents, detergents, perfumes, and some even as food additives.1-5 Due to their reactivity (they are able to interact with electron-rich biological macromolecules) they also raise adverse health issues such as general toxicity, allergenic reactions, mutagenicity, and carcinogenicity.5-7 Our present interest lies in the simple aldehydes, i.e. linear alkanals, from propanal, butanal, pentanal, hexanal, heptanal, octanal, to nonanal. The motivation for structural study of simple aldehydes originates in the studies of the enzymatic activity of horse liver alcohol dehydrogenase (HLADH) in microemulsion systems of the nonionic surfactant Brij 35, water and one of the simple alcohols from ethanol to 1-decanol. In such biocatalytic reactions a simple aldehyde appears as the oxidation product of the simple alcohol, which on the other hand constitutes the microemulsion reaction medium itself and is also involved also in its selfassembly structure. This structure and also its effect on the activity of HLADH have already been thoroughly investigated.8-12 Nevertheless, the structural effect of the aldehydes that emerge in course media during the time of the biocatalytic reaction is still to be investigated. Furthermore, the simple alcohols, like the simple aldehydes, themselves show some organized molecular structure. Recently we have conducted a study on the * Corresponding author. E-mail: [email protected]. Tel: +386 1 24 19 434. Fax: +386 1 24 19 425. † University of Ljubljana. ‡ University of Graz. § Academy of Sciences of the Czech Republic.

structure of the pure simple alcohols13,14 which are commonly used as cosurfactants in various surfactant systems. In the present study we aimed to obtain similar information for the simple aldehydes of interest. Namely, such information is rather important, because it sheds some light on the effects that such compounds could have in more complex microemulsion systems if present in higher concentrations. The simple alcohols, like the aldehydes, can be considered as representatives of the socalled structured organic solvents, which usually show some pronounced organization on the molecular level. It has been shown that some interesting solvation effects can be observed in such solvents.15,16 However, they drew our attention to the problems that they usually cause in the evaluation of structural small-angle X-ray scattering (SAXS) data of microemulsion systems rich in such solvents, where the scattering signal arising from their molecular organization can mask the detailed structure of the surfactants, yielding the latter nonresolvable in such data.12 In this sense the topic of study also has a more general relevance. In this paper we present configurational bias Monte Carlo (CBMC) simulation results17,18 for the simple aldehydes based on the transferable potential for phase equilibriasunited atom (TraPPE-UA) force field,19 and compare them to the experimental SAXS results in the regime 0.3 < q < 25 nm-1, where q represents the length of the scattering vector q ) 4π/λ · sin(ϑ/ 2), with λ being the wavelength of the X-rays and ϑ the scattering angle. In this respect a recently developed procedure for calculating the scattering intensities from the MC simulation data based on the Debye equation13,14,20 and also a procedure based on the reciprocal lattice approach14,21 were utilized. The calculated scattering results obtained turned out to be a rather strict structural test for the TraPPE-UA results on simple aldehydes; therefore we also checked the performance of this model from the thermodynamic point of view, utilizing the calculation of some vapor-liquid coexistence curves with the

10.1021/jp9028485 CCC: $40.75  2009 American Chemical Society Published on Web 06/22/2009

9430

J. Phys. Chem. B, Vol. 113, No. 28, 2009

corresponding Clausius-Clapeyron plots and comparison with experimental data from the literature. 2. Experimental Details and Methods 2.1. Materials. The aldehydes from propanal to nonanal were purchased from Fluka (propanal, purity >96%; butanal, purity >97%; pentanal, purity >97%; hexanal, purity >97%; heptanal, purity >95%; octanal, purity >98%; nonanal, purity >95%) and were used without further purification. The samples were investigated at 25 °C. Since the aldehydes used (with the exception of octanal) were denoted to contain up to 2% of individual carboxylic acids, the effect of a small amount of carboxylic acid on the aldehyde structure was tested experimentally for the example of octanal and octanoic acid, and the result is available in Figure S1 of the Supporting Information. These results demonstrate that the effect of such a small amount of acid on the aldehyde structure is experimentally practically undetectable. 2.2. Small-Angle X-ray Scattering Measurements. Smallangle X-ray scattering spectra were measured with a “SAXSess” (Anton Paar KG, Graz, Austria) evacuated high performance SAXS instrument, which was attached to a conventional X-ray generator (Philips, The Netherlands) equipped with a sealed X-ray tube (λCuKR ) 0.154 nm) operating at 40 kV and 50 mA. The samples were measured in a standard quartz capillary for the SAXSess camera (with an outer diameter of 1 mm and wall thickness of 10 µm). A Fuji BAS 1800 2D-imaging plate detection system with a spatial resolution of 50 × 50 µm per pixel at a “sample to detector distance” of 265 mm was used. After a 30 min measurement the scattering data were read off from the imaging plate, then corrected for the absorption of X-rays in the sample and further transformed to the q scale (program SAXSQuant; Anton Paar KG, Graz, Austria). The scattering intensities were further corrected for the empty capillary scattering and subsequently transformed to an absolute scale using water as a secondary standard22 (program PDH; PCG software, Institute of Chemistry, Graz, Austria). However, the scattering intensities obtained in this way were experimentally smeared because of the finite dimensions of the primary beam23 and as such did not represent the real “absolute scattering intensities” which are free of smearing effects. In order to be able to compare the experimental SAXS curves directly to the simulated theoretical scattering patterns, the latter had to be theoretically smeared.23 2.3. Model and Monte Carlo Simulation Details. The TraPPE-UA (transferable potential for phase equilibriasunited atom) force field19 was used to model the aldehydes from propanal to nonanal. In this model, CHx groups (CH3, CH2, and CH) are treated as single sites: the model molecules therefore consisted of a linear chain of united CHx atoms with an oxygen atom bonded to the terminal one. The configurational bias Monte Carlo (CBMC)17 simulation was used to generate configurations of these aldehydes in the NVT ensemble. The numbers of molecules N and the side lengths a of the simulation boxes were as follows: propanal (400, 36.25 Å), butanal (350, 37.42 Å), pentanal (300, 37.7025 Å), hexanal (300, 39.5283 Å), heptanal (200, 36.05 Å), octanal (200, 37.375 Å), and nonanal (200, 38.6186 Å). A spherical truncation at 14 Å was used for the Lennard-Jones part of the potential, and analytical corrections for the tail were employed. The Ewald summation was used for the long-range Coulombic interactions with the summation range in reciprocal space κmax ) 5 and the Ewald convergence parameter κR ) 5.6. Three types of Monte Carlo steps were used (relative probabilities of their usage are given in brackets): configurational bias partial molecule regrowth (40%), center-

Lajovic et al. of-mass molecule translation (30%), and rotation about centerof-mass (30%). Maximum displacements for the latter two were automatically adjusted during equilibration to yield 50% acceptance ratios. Each system was equilibrated for at least 10,000 cycles (one cycle consisted of N Monte Carlo steps), followed by a production run of 40,000 cycles. During this run, a snapshot of particle coordinates was saved every 400 cycles, and the resulting 100 configurations were later used to calculate the scattering intensities. The vapor-liquid coexistence curves and vapor pressures of propanal, butanal and pentanal were determined using the NVT Gibbs ensemble Monte Carlo (GEMC) method.24 Six types of Monte Carlo steps were employed (relative probabilities of their usage are given in parentheses): isotropic volume change (20%), rotational-bias two box molecule transfer (10%), configurational bias two box molecule transfer (10%), configurational bias partial molecule regrowth (20%), center-of-mass molecule translation (30%), and rotation about center-of-mass (10%). The equilibration and production periods were 15,000 and 25,000 cycles, respectively. The Clausius-Clapeyron equation was used to calculate the boiling points at standard pressure, and the critical temperatures and densities were determined using the saturated density scaling law and the law of rectilinear diameters with a scaling exponent β* ) 0.325.25 All simulations were performed with the MCCCS Towhee Monte Carlo simulation code, version 5.2.11.26 2.4. Calculation of the Scattering Intensities. The scattering intensities were calculated for the CBMC results utilizing two different approaches. The first one was based on the Debye equation20 combined with a special procedure to eliminate the background scattering caused by the finite size of the simulation box, while the other used reciprocal lattice theory. A detailed description of the methods can be found in refs 13 and 14, respectively. Only a short overview is given here. In the so-called “brute force approach”, a stored configuration was used to construct a larger system by stacking 27 simulation boxes into a 3 × 3 × 3 grid. Boxes of predetermined size were then cut out of such a system, and their scattering intensities were further calculated via the Debye equation:20 N

I(q) ) k

N

∑ ∑ φi(q) φj(q) i)1 j)1

sin qrij qrij

(1)

where rij represents the distance between a pair of pseudoatoms i and j, functions φ(q) represent the scattering amplitudes of the pseudoatoms, and k represents an appropriate constant to bring the scattering curves to an absolute scale.13 The oscillations in the scattering intensities calculated directly from the original CBMC boxes, which arise due to the finite size of the boxes, were smoothed by weighted summation of normalized scattering intensities over the whole set of boxes. Such smoothing resulted in a scattering curve superimposed on a q-4 background, which was then easily subtracted. Finally, the result was normalized by a factor to establish the correct absolute scattering intensity. Such a procedure was used for all 100 configurations, and the results were averaged to yield the overall scattering curve of an individual aldehyde on an absolute scale. A second, conceptually different method based on the reciprocal lattice approach14,21 was also used to calculate the scattering intensities. In this process, the system is constructed by stacking the simulation cells into an infinite three-dimensional lattice. This essentially makes the system crystalline-like and allows one to use the theories developed for crystallography.

Structural Properties of Aldehydes

Figure 1. Calculated scattering curves of modeled aldehydes from propanal (bottom) to nonanal (top). The results of the “brute force method” (solid lines) and of the “reciprocal lattice approach” (symbols with the dashed lines) are shown. Symbols indicate the discrete nature of the q-scale in the latter case, and the dashed line is added to guide the eye. The curves from butanal on are shifted upward by multiples of 0.1 for the sake of clarity.

Because of its periodicity, the scattering of such a system is not continuous but rather displays discrete scattered beams in particular directions which correspond to nodes in the reciprocal lattice. In our calculation we used the nodes (1,0,0), (1,1,0), (1,1,1), (2,1,0), (2,1,1) and their equivalents, totaling 43 nodes. The substantial advantage of this method is speed, as it is several orders of magnitude faster than the “brute force approach” described above. However, it only allows the calculation of scattering intensities at discrete values of q. 3. Results and Discussion 3.1. Theoretical and Experimental X-ray Scattering Intensities. The calculated scattering curves obtained from the CBMC simulation results by the “brute force method” and the “reciprocal lattice approach” are shown in Figure 1. Generally they show rather similar results. Nevertheless, one can still observe some slight oscillations superimposed on the curve produced by the former method. These oscillations are most probably the remains of the background scattering caused by the MC simulation boxes. The ripples mainly affect q values below 5 nm-1 and are less pronounced elsewhere. On the contrary, the results of the reciprocal lattice approach readily extend toward low q values without noticeable distortions, since there is no background contribution involved per se. As this method yields the results at discrete q values, the curves connecting the intensities at individual q are less smooth. Comparison of the two sets of curves also shows that the method based on the reciprocal lattice consistently gives somewhat lower intensities, but the position and appearance of the scattering peaks are very similar. In order to be able to compare the theoretical results to the experimental SAXS curves, the former need to be numerically

J. Phys. Chem. B, Vol. 113, No. 28, 2009 9431

Figure 2. Experimental SAXS curves of aldehydes (symbols) from propanal (bottom) to nonanal (top) together with the numerically smeared calculated scattering curves obtained by the “brute force method” (solid lines) and by the “reciprocal lattice approach” (dashed lines). The curves from butanal on are shifted upward by multiples of 0.2 for the sake of clarity.

smeared to establish the same frame of reference. This necessity stems from the fact that the experimental results depend on the size and shape of the primary beam. While numerical smearing is relatively easy to perform, the reverse processsdesmearingsis a much more tedious task and is seldom attempted in practice. The experimental SAXS curves together with the numerically smeared calculated scattering results of CBMC simulations obtained by both methods used for the scattering intensity calculations are displayed in Figure 2. Their most prominent feature is a pronounced peak at q ≈ 14 nm-1. Its position shows very weak dependence on the type of the aldehyde and remains almost constant. Keeping in mind that the q-scale is a measure of the reciprocal space and as such bears a reciprocal relationship with the real (actual) r-space, we may calculate the characteristic real distance corresponding to the value of q. Using the relation r ) 2π/q, the reciprocal distance 14 nm-1 thus translates into a real distance of approximately 4.5 Å, which obviously corresponds to some average effective distance between the aldehyde molecules. As will be seen in the following, we interpret this scattering peak as mainly the result of the average correlations mostly between the hydrocarbon parts of the aldehyde molecules, similarly to what was found for the primary liquid alcohols.13 The scattering curves in Figure 2 also show a smaller peak at lower q values (inner peak) that shifts in position with change in the size of the aldehyde molecule. This peak can barely be discerned in the case of propanal, but increases in height with increasing aldehyde chain length and shows itself clearly as a distinctive scattering peak in aldehydes longer than pentanal. In parallel its position shifts substantially as well, from the vague scattering peak at the rough estimate of q ≈ 7 nm-1 for propanal to the relatively well-defined peak at q ≈ 3.5 nm-1 for nonanal. These scales translate into approximately 9 Å and 18 Å in real space, respectively. Judging from its strong

9432

J. Phys. Chem. B, Vol. 113, No. 28, 2009

Lajovic et al.

Figure 4. A schematic representation of a united-atom heptanal molecule.

Figure 3. (a) Results of the simulated “contrast matching experiment” for the modeled heptanal. The numerically smeared total calculated scattering curve (solid line) and partial scattering contributions to the scattering of heptanal: ICHx(q) (dashed line) and IO(q) (dotted line) are shown, with the cross-term contribution HOCHx(q) (dash-dotted line). (b) Calculated partial structure factors of heptanal: HCHxCHx(q) (solid line), HOCHx(q) (dashed line), and HOO(q) (dotted line).

dependence on the chain length of the aldehyde molecule and from some results presented in the following, we can conclude that the inner peak corresponds to the correlations of the oxygen and CH pseudoatoms of adjacent aldehyde molecules. As was also the case in our previous study of alcohols,13,14 the results quantitatively show a systematic underestimation of the inner scattering peak height, although this might be partly due to the fact that the range of the correlations involved is near one-half the MC simulation box side length, at which the reliability of the simulation is already limited. Nevertheless, the main features of the scattering curves such as their general appearance and the scattering peak positions are reproduced rather well. In this sense such a structural test was again recognized as a rather strict test for the model used. In spite of the fact that by modeling individual groups of atoms as pseudoatoms this model in this regard is rather simple, it still provides satisfactory results. In any case, as already recognized in previous studies19 and as will also be illustrated on some additional tests presented in the following text, it performs rather well from the thermodynamic point of view. In order to unambiguously address the origin of the two scattering peaks in Figure 2, we made a simulated contrast matching experiment,13 similar to the well-known experimental method used in small-angle neutron scattering (SANS) measurements. By simply considering only a subset of atoms when calculating the scattering of CBMC configurations, we could assess the effective contributions of individual atom types to the total scattering intensity. The results of such a simulated contrast matching experiment are shown for the case of heptanal in Figure 3a. The scattering contributions IR(q) represent the scattering intensities according to eq 1 with only atoms of type R taken into account. Similarly, the so-called partial structure factors HRβ represent the cross-terms of R and β atoms in the MC box (see eq 1). For the example of heptanal depicted in Figure 3a it turns out that the CHx pseudoatoms mainly

contribute to scattering in the regime 10 nm-1 < q < 20 nm-1 with a scattering maximum around 14 nm-1. This corresponds to the position of the outer peak on the scattering curve. Similarly, the scattering contribution of oxygen atoms IO(q) together with the cross-terms HOCHx(q) appear to introduce the inner peak at around 4 nm-1 and suggest that its emergence is related to the long-distance correlation of oxygen atoms (see also the discussion on pair distribution functions in the following text). Interestingly, very similar features have been observed in a previous study on alcohols, although somewhat more clearly. In that case rather strong sequential hydrogen bonding of the alcohol molecules was observed. It was also shown that the inner scattering peak unambiguously corresponds to the oxygen-oxygen correlations between neighboring -OH skeletons.13 The pronounced hydrogen bonding of alcohol molecules is actually the main reason for their interesting molecular organization and consequently also for the clearly expressed inner scattering peaks. Due to the different nature of the alcohol and aldehyde molecules no such pronounced hydrogen bonding can be expected for aldehydes, which in a sense explains the weaker inner scattering peaks in Figure 2. Nevertheless, to some extent a very similar organization of aldehyde molecules into larger aggregates was in any case observed in the CBMC configurations and is discussed in more detail in the next subsection. Similarly to the individual atom/pseudoatom type scattering contributions IR(q), the cross-terms in the form of partial structure factors HRβ(q) were also calculated. In the latter case, only the atom/pseudoatom type pair Rβ were considered in the CBMC configuration and the condition i * j was applied in the eq 1. Since in similar theoretical investigations the partial structure factors HRβ(q) and HRR(q) are usually reported, we also depict all of the partial structure factors of modeled heptanal in Figure 3b. 3.2. Pair Distribution Functions g(r). The collection of equilibrium configurations was also used to calculate the pair distribution functions g(r) for various pairs of atom types (a typical united-atom aldehyde molecule is shown schematically in Figure 4). The distances between atoms of the same molecule, especially between nearest neighbors, are virtually constant and tend to produce sharp spikes in the pair distribution functions. To avoid this, only atoms situated on different molecules were taken into consideration. An example of pair distribution functions involving the oxygen atom in heptanal is shown in Figure 5. Note that each CH2 site was treated separately, since they lie in different parts of the molecule and therefore experience different surroundings. The gOO(r) function immediately draws attention as it reveals a rather strong intermolecular correlation at the oxygen site. Interestingly, the pair distribution function of the intermolecular CH and O sites gOCH(r) shows a comparable degree of association, which surprisingly occurs at an even smaller distance. This can be explained by recognizing that in contrast to other groups in the

Structural Properties of Aldehydes

J. Phys. Chem. B, Vol. 113, No. 28, 2009 9433

Figure 5. Selected pair correlation functions between the intermolecular pseudoatoms of the modeled aldehydes: gOO(r) (solid), gOCH(r) (long-dashed), gOCH2/1(r) (short-dashed), gOCH2/2(r) (dotted), gOCH2/3(r) (long dash-dot), gOCH2/4(r) (short dash-dot), gOCH2/5(r) (dash-dashspace), and gOCH3(r) (dash-dot-dot:). Sites designated CH2/X are Xth CH2 pseudoatom counting from oxygen toward CH3.

Figure 6. Histograms of the fraction of heptanal molecules versus the aggregate size (striped) and the fraction of heptanal molecules in a special type of aggregates containing at least one closed cycle (crisscross).

Figure 7. Vapor-liquid coexistence curves calculated for the TraPPE-UA force field for propanal (squares), butanal (circles) and pentanal (triangles). Experimentally determined critical points are marked by stars. The dotted lines are the fits of theoretical data according to the saturated density scaling law and the law of rectilinear diameters.25,33

Figure 8. Clausius-Clapeyron plots of the saturated vapor pressure versus inverse temperature for propanal (squares), butanal (circles) and pentanal (triangles) as predicted by the TraPPE-UA force field. Solid lines represent the experimental saturated vapor pressure data.34

TABLE 1: Fraction of Heptanal Molecules with a Specific Number of Associated Neighboring Molecules number of neighbors fraction of molecules [%]

0 40.5

1 41.2

2 15.8

3 2.4

4 0.1

molecule, the CH and O sites are partially charged, which leads to pronounced Coulombic attraction between them. Inspecting further the pair distribution functions of CH2 groups along the aldehyde molecule, the primary peaks of the pair distribution functions broaden, become less intense and shift to longer distances (with the exception of the last two CHx groups in the molecule which are again closer to oxygen). Considering the whole picture, it becomes obvious that the molecules tend to organize in such a way that their CH-O ends come close together in a manner where the CH of one molecule is close to O of the other one. Such an orientation is in agreement with the very short intermolecular CH-O distances and at the same time the slightly longer intermolecular O-O distances. Also, the relatively high second peak on gOCH reveals that there may frequently be more than two aldehyde molecules grouped in such a fashion. Indeed, if one looks at a 3D model of the configuration, such aggregates can be easily seen (see the PDB formatted material available in the Supporting Information). At this point we have to note that these aggregates are not unlike the structures found in the related alcohols, where oxygen and hydrogen atoms of adjacent molecules are correlated via hydrogen bonds.13 A rather simple criterion for an intermolecular association can be defined: two molecules are considered associated if at least one of their intermolecular CH-O distances is equal to or smaller than the distance of the first peak on the pair distribution function gOCH(r). In the case of heptanal, this

Figure 9. Enthalpy of vaporization versus temperature for the TraPPEUA force field modeled propanal (squares), butanal (circles) and pentanal (triangles).

distance is r ≈ 3.3 Å. A statistical evaluation of the configurations obtained from CBMC simulations results in a diagram of the fraction of aldehyde molecules vs the number of aldehyde molecules in an aggregate, which is presented in Figure 6. Interestingly, the general distribution of heptanal molecules within the aggregates shows an exponential dependence on the aggregate size. A similar histogram considering only the aggregates which contain at least one closed cycle is also shown. One can see that the distribution of molecules within the latter no longer follows an exponential course, but instead reaches a maximum at an aggregate size of five molecules. Statistical evaluation of aggregation phenomena yielded also the fractions of heptanal molecules with a specific number of associated neighboring molecules, which are given in Table 1. Despite the fact that the nature of alcohol and aldehyde molecules is different, one can surprisingly notice some similarities in this type of structural result.

9434

J. Phys. Chem. B, Vol. 113, No. 28, 2009

Lajovic et al.

TABLE 2: Comparison of Calculated and Experimental Thermodynamic Quantities of Propanal, Butanal and Pentanal propanal (sim) exptl butanal (sim) exptl pentanal (sim) exptl

Tboil [K]

Tcrit [K]

Fcrit [g/cm3]

Pcrit [Bar]

∆Hvap [kJ/mol]

317 ( 3 322 ( 227 344 ( 5 348 ( 227 372 ( 5 376 ( 227

502 ( 8 504.4 ( 0.728 535 ( 7 537.2 ( 0.830 566 ( 9 566.1 ( 0.830

0.28 ( 0.01 0.285 ( 0.00529 0.277 ( 0.008 0.280 ( 0.00529 0.272 ( 0.009 0.275 ( 0.00529

55 ( 8 52.6 ( 0.628 50 ( 7 43 ( 130 46 ( 8 40 ( 130

28.7 ( 0.2 29.7 ( 0.227 31.3 ( 0.3 33.7 ( 0.431 34.3 ( 0.4 38.1 ( 0.132

In the discussion of the results of the simulated contrast matching experiment in the previous subsection, we ascribed the emergence of the inner peak on the scattering curve to longdistance oxygen-oxygen and oxygen-CHx correlations. The inner peak on the scattering curve of heptanal is situated at q ≈ 4 nm-1, which corresponds to a real-space distance of around 16 Å. Looking at the oxygen-oxygen pair distribution curve of heptanal, there is no obvious correlation visible at that distance. Inspecting it closely, it can be recognized that the density slightly exceeds the average, forming a very broad and unexpressed maximum. Considering the possibility that this indicates a correlation with a very long spatial period, we suggest that a larger simulation box be used in subsequent studies to allow it to be faithfully established. The inadvertent neglect of such long-range correlations is in fact a possible explanation of the systematically underestimated heights of the inner peaks found in this study. 3.3. Vapor-Liquid Phase Equilibria. Since we found that the structural tests are a severe test of the model used in this investigation of aldehydes, we additionally decided to check its performance further on some thermodynamic results. Although the model has generally already been tested from this point of view,19 we nevertheless present some additional thermodynamic tests in the following. We performed a number of Gibbs ensemble Monte Carlo simulations to calculate the saturated vapor-liquid coexistence curves and related thermodynamic quantities of propanal, butanal and pentanal. The corresponding saturated vapor-liquid coexistence curves are shown in Figure 7, the Clausius-Clapeyron (vapor pressure) plots in Figure 8 and the enthalpies of vaporization vs temperature in Figure 9. The theoretical and experimental thermodynamic quantities are gathered in Table 2. As expected from a force field developed specifically for prediction of phase equilibria and related quantities, the simulation data is in good agreement with experiment. The margins of error of normal boiling points and critical temperatures overlap with the experimental values, although the theoretical values appear to be slightly (but consistently) too low. A similar conclusion can be reached about the vapor pressure plots. This is a known bias of the TraPPE-UA, attributed by the authors of the force field to parametrization of CH2 and CH3 sites.19 4. Conclusions In this study, we used the TraPPE-UA force field in combination with the configurational bias Monte Carlo method to simulate various primary aldehydes. The resulting aldehyde configurations were then used to predict the small-angle X-ray scattering curves via two different methods: the Debye equation coupled with a background removal procedure, and the reciprocal lattice approach. The results compare favorably with experimental data, but show a consistently underestimated inner peak: it is believed that this could be in part corrected by using a larger simulation box. Also, the intensity in general is lower than expected. It must be stressed that comparison of X-ray scattering from simulation data to experimental scattering curves

poses an extremely strict test for a force field, as it effectively probes the structure of the liquid. From that point of view, the TraPPE-UA force field actually performs rather well considering the fact that it is based on a very simplified model of molecules. To conclude, the results of our study show that this model does a better job in predicting thermodynamic than structural properties. While the results for vapor-liquid phase equilibria agree practically quantitatively with experimental data, the agreement between the measured X-ray scattering intensities and the calculated scattering functions of the primary aldehydes is somewhat worse. However, also in the latter case a qualitative agreement between the theoretical predictions and experimental data is observed. Acknowledgment. We are thankful to Gu¨nther Scherf for the assistance with the SWAXS experiments. A.L. is grateful for his warm welcome in the group of Prof. Otto Glatter. We acknowledge financial support from the Slovenian Research Agency through the Physical Chemistry Research Programme 0103-0201 and from the Federal Ministry for Education, Science, and Culture of Austria (BI-AT/09-10-022). Supporting Information Available: Figure S1 depicts experimental SAXS curves of octanal, 2 wt % octanoic acid in octanal and 5 wt % octanoic acid in octanal. Also, five saved configurations of heptanal from CBMC simulations are available as PDB files. One set of files contains unmodified configurations with whole molecules, while the other set contains only the CH and O groups to allow a less obstructed view of intermolecular association. This material is available free of charge via the Internet at http://pubs.acs.org. References and Notes (1) Bruze, M.; Johansen, J. D.; Andersen, K. E.; Frosch, P.; Lepoittevin, J. P.; Rastogi, S.; Wakelin, S.; White, I.; Menne, T. J. Am. Acad. Dermatol. 2003, 48, 194. (2) Cheng, S. S.; Liu, J. Y.; Tsai, K. H.; Chen, W. J.; Chang, S. T. J. Agric. Food Chem. 2004, 52, 4395. (3) Sato, K.; Krist, S.; Buchsauer, G. Biol. Pharm. Bull. 2006, 29, 2292. (4) Placzek, M.; Fromel, W.; Eberlein, B.; Gilbertz, K. P.; Przybilla, B. Acta Derm.-Venereol. 2007, 87, 312. (5) Watanabe, K.; Matsuda, M.; Furuhashi, S.; Kimura, T.; Matsunaga, T.; Yamamoto, I. J. Health Sci. 2001, 47, 327. (6) Benigni, R.; Passerini, L.; Rodomonte, A. EnViron. Mol. Mutagen. 2003, 42, 136. (7) Benigni, R.; Conti, L.; Crebelli, R.; Rodomonte, A.; Vari, M. R. EnViron. Mol. Mutagen. 2005, 46, 260. (8) Meziani, A.; Touraud, D.; Zradba, A.; Pulvin, S.; Pezron, I.; Clausse, M.; Kunz, W. J. Phys. Chem. B 1997, 101, 3620. (9) Preu, H.; Zradba, A.; Rast, S.; Kunz, W.; Hardy, E. H.; Zeidler, M. D. Phys. Chem. Chem. Phys. 1999, 1, 3321. (10) Schirmer, C.; Liu, Y.; Touraud, D.; Meziani, A.; Pulvin, S.; Kunz, W. J. Phys. Chem. B 2002, 106, 7414. (11) Tomsˇicˇ, M.; Besˇter-Rogacˇ, M.; Jamnik, A.; Kunz, W.; Touraud, D.; Bergmann, A.; Glatter, O. J. Phys. Chem. B 2004, 108, 7021. (12) Tomsˇicˇ, M.; Besˇter-Rogacˇ, M.; Jamnik, A.; Kunz, W.; Touraud, D.; Bergmann, A.; Glatter, O. J. Colloid Interface Sci. 2006, 294, 194. (13) Tomsˇicˇ, M.; Jamnik, A.; Fritz-Popovski, G.; Glatter, O.; Vlcˇek, L. J. Phys. Chem. B 2007, 111, 1738. (14) Tomsˇicˇ, M.; Fritz-Popovski, G.; Vlcˇek, L.; Jamnik, A. Acta. Chim. SloV. 2007, 54, 484–491.

Structural Properties of Aldehydes (15) Reeves, R.; Kaiser, R. S. J. Org. Chem. 1970, 35, 3670. (16) Auvray, X.; Petipas, C.; Lattes, A.; Ricco-Lattes, I. Colloids Surf., A 1997, 123, 247. (17) Siepmann, J. I.; Frenkel, D. Mol. Phys. 1992, 75, 59. (18) Vlugt, T. J.; Martin, M. G.; Smit, B.; Siepmann, J. I.; Krishna, R. Mol. Phys. 1998, 94, 727. (19) Stubbs, J. M.; Potoff, J. J.; Siepmann, J. I. J. Phys. Chem. B 2004, 108, 17596. (20) Debye, P. Phys. Z. 1930, 31, 419. (21) Pedersen, J. S. Monte Carlo Simulation Techniques Applied in the Analysis of Small-Angle Scattering Data from Colloids and Polymer Systems. In Neutrons, X-Rays and Light Scattering Methods Applied to Soft Condensed Matter; Lindner, P., Zemb, T., Eds.; Elsevier Science B.V.: Amsterdam, 2002; p 381. (22) Orthaber, D.; Bergmann, A.; Glatter, O. J. Appl. Crystallogr. 2000, 33, 218. (23) Glatter, O. Data treatment. In Small angle x-ray scattering; Glatter, O., Kratky, O., Eds.; Academic Press Inc. London Ltd.: London, 1983; p 119. (24) Frenkel, D.; Smit, B. Understanding Molecular Simulation: From Algorithms to Applications; Elsevier Science & Technology Books: 2002. (25) Rowlinson, J. S.; Widom, B. Molecular theory of capillarity; Clarendon Press: Oxford, 1984.

J. Phys. Chem. B, Vol. 113, No. 28, 2009 9435 (26) MCCCS Towhee, http://towhee.sourceforge.net. (27) NIST Chemistry WebBook, NIST Standard Reference Database Number 69 (http://webbook.nist.gov); Linstrom, P. J., Mallard, W. G., Eds.; National Institute of Standards and Technology: Gaithersburg, MD 20899, retrieved January 29, 2009. (28) Gude, M. T.; Teja, A. S. The Critical Properties of Several n-Alkanals, Tetralin, and NMP. In DIPPR Data Series; 1994; Vol. 2; p 174. (29) Teja, A. S.; Anselme, M. J. AIChE Symp. Ser. 1990, 86, 128. (30) Teja, A. S.; Rosenthal, D. J. AIChE Symp. Ser. 1990, 86, 133. (31) Buckley, E.; Cox, J. D. Trans. Faraday Soc. 1967, 63, 895. (32) D’yakova, G. N.; Korichev, G. L.; Korkhov, A. D.; Vasil’eva, T. F.; Vasil’ev, I. A. Zh. Prikl. Khim. (J. Appl. Chem. USSR) 1981, 54, 1644– 1646. (33) Rowlinson, J. S.; Swinton, F. L. Liquids and liquid mixtures, 3rd ed.; Butterworth: London, 1982. (34) Dykyj, J.; Svoboda, J.; Wilhoit, R. C.; Frenkel, M.; Hall, K. R. Vapor Pressure and Antoine Constants for Oxygen Containing Organic Compounds; Springer-Verlag, 2000; Vol. IV/20B.

JP9028485