Hydration of Carboxylate Anions - American Chemical Society

is based on the empirical relationship between the ROO and νOD values.48 The ...... (55) Yilmaz M.; Memon S.; Tabakci, M.; Bartsch R. A. In New Front...
0 downloads 0 Views 1MB Size
Download PDF
8128

J. Phys. Chem. B 2009, 113, 8128–8136

Hydration of Carboxylate Anions: Infrared Spectroscopy of Aqueous Solutions Emilia Gojło, Maciej S´miechowski, Aneta Panuszko, and Janusz Stangret* ReceiVed: December 23, 2008; ReVised Manuscript ReceiVed: April 1, 2009

Hydration of carboxylate ions was studied in aqueous solutions of sodium salts by means of FTIR spectroscopy using the HDO molecule as a probe. The quantitative version of the difference spectra method has been applied to determine the solute-affected water spectra. They display two-component bands of affected HDO at ca. 2550 and 2420 cm-1. These bands are attributed to the -COO- group of the R-COO- ion (R ) H, CH3, C2H5), because water molecules surrounding the substituent R behave roughly as molecules in the bulk phase. For the studied carboxylates the net water structure making effect is observed, which increases with electron-donor ability of R, by means of changing the relative intensity of solute-affected HDO component bands. The observed splitting of the carboxylate-ion-affected HDO band is unique for these anions. The experimental results were confronted with DFT-calculated structures of small gas-phase and polarizable continuum model (PCM) solvated aqueous clusters to establish the structural and energetic states of carboxylate ions hydrates. This was achieved by comparison of the calculated optimal geometries with the interatomic distances derived from HDO band positions. Different possibilities have been considered to explain the peculiar spectral results. The plausible explanation assumes symmetry breaking of the carboxylate ion induced by interaction with water solvent: C-O bond lengths of RCOO- and electric charge localization become unequal. It is demonstrated by nonequivalent interaction of oxygen atoms of the RCOO- anion with water molecules. Taking into account only the energetic effect, the phenomenon is explained by the anticooperative H-bond formation of the carboxylate group with water molecules, which increases with the electron-donor ability of the substituent R. In this interaction two water molecules play an important part, as appears from the calculated clusters. They interact with oxygen atoms of the RCOO- ion, forming a cooperative system, within which solvent molecules are nonequivalent with respect to H-bond formation with both proton-accepting sites of the solute. This additionally enhances solvent-induced symmetry breaking of carboxylate anion. Strongly hydrogen-bonded solvent is more effective in inducing symmetry breaking; thus, increasing the temperature decreases the splitting of the carboxylate-ion-affected water, as experimentally observed. Introduction The important constituents of biomolecules often experience mixed types of hydration, i.e., hydrophobic hydration of the alkyl chains combined with electrophilic/hydrophilic hydration of the charged groups or the electron pair donor/acceptor atoms.1-3 The cooperative effects of both types of hydration determine the overall interaction of the molecule with the surrounding aqueous medium.2 The study of entire biomolecules or biopolymers in an aqueous solution, especially using vibrational spectroscopy, is often inconvenient, because of the difficulties with separating different effects originating from many types of interactions.4 Therefore, it is often more suitable to study simple model compounds, having an isolated type of functional group encountered in biological systems. The carboxylic group is ubiquitous in biological molecules. It is found, among others, in amino acids, fatty acids, lipid bilayers, and surface-active agents. The active sites of enzymes are often equipped with carboxylic functional groups for ligand binding.5 At typical physiological pH around 7 they are found primarily in the deprotonated state, in the form of carboxylate anions.6 However, in certain circumstances (such as in the strongly acidic gastric juice) they can also appear in the protonated state. Vibrational spectroscopy of aqueous solutions showed that both of these forms interact with water in a unique and different manner.6-8 * To whom correspondence should be addressed.

Anions of carboxylic acids in aqueous solutions have been hitherto studied with a variety of theoretical and experimental methods. The computational approaches include static ab initio calculations,5,9,10 as well as classical11 and quantum molecular dynamics simulations.12 The more commonly encountered experimental techniques comprise different variations of vibrational spectroscopy,6-8,13-19 dielectric spectroscopy,20 diffraction studies,21-23 and volumetric measurements.24 An intriguing structural aspect of carboxylate anions in aqueous solutions is the equivalence of both oxygen atoms of the -COO- group. The picture changes from total charge delocalization in the free gas-phase anion9,10 to the apparent nonequivalence of the oxygen atoms connected with asymmetry of the functional group in larger aqueous clusters.10 The former model was usually used in classical MD simulations to define the anion-water interaction potential,11 but this arbitrary assumption remains disputable. From a practical point of view, vibrational spectroscopy allows registering short-lived species formed by water molecules influenced by a solute due to its favorable time scale of response (ca. 10-14 s in the range of water fundamentals). Consequently, even instantaneous configurations of water molecules around the carboxylate anion should be registered in the infrared spectrum. Sodium salts of the three simplest homologous carboxylic acids, i.e., formic, acetic, and propionic acids, have been selected for the present study. The formate anion is the only one that lacks a hydrophobic group, whereas acetate and propionate anions serve as models for an increasingly hydrophobic character

10.1021/jp811346x CCC: $40.75  2009 American Chemical Society Published on Web 05/13/2009

Hydration of Carboxylate Anions of the alkyl chain. To investigate the hydration of the just mentioned carboxylates, we used vibrational (FTIR) spectroscopy of HDO isotopically diluted in aqueous solutions of the studied solutes in ordinary water, which is widely considered an ideally suited method in studies of solute hydration.25-27 The quantitative method of spectral data analysis formulated by Lindgren and co-workers, and independently in our laboratory,28-30 allows separation of the solute-affected water spectrum based on series of spectra measured at increasing concentration of the solute. Further interpretation of affected water spectra is made possible by application of the Badger-Bauer rule,31 linking the OD band position with the intermolecular interaction energy of water,13 for a description of the energetic state of HDO molecules, as well as by transformation of the HDO band contour to water oxygen-oxygen distance probability distribution function, based on the published procedure,30,32 in order to reveal the structural state of hydration spheres. To further link the structural and spectral data, we also recently applied static ab initio quantum mechanical calculations of optimal structures of small aqueous clusters of the solutes in the gas phase or in the external solvent simulated by the polarizable continuum model (PCM).3,33 Here we extend this method to study hydration of carboxylate anions. Experimental Section Sodium formate (99+%, ACS reagent, Sigma-Aldrich), sodium acetate (99+%, ACS reagent, Sigma-Aldrich), nickel(II) acetate tetrahydrate (p.a., cobalt free, POCh Poland), and sodium propionate (minimum 99%, Sigma-Aldrich) were used as supplied. D2O came from two sources: the Institute of Nuclear Investigation, Poland (99.84% isotopic purity), and Aldrich (99.96% isotopic purity). Stock solutions were prepared by dissolving weighted amounts of respective solutes in double-distilled water. The series of solutions were prepared by dissolving weighted amounts of respective stock solution in double-distilled water. Sample solutions were made by adding 4% (by weight) of D2O relative to H2O and reference solutions by adding the same molar amounts of H2O. Densities of the salt solutions were measured with an Anton Paar DMA 5000 density meter at 25.000 ( 0.001 °C. The solution series of sodium carboxylates included samples of the following approximate molalities: 0.1, 0.2, 0.4, 0.6, 0.8, and 1.0 mol · kg-1. For nickel acetate samples molalities were ca. 0.1, 0.2, 0.3, 0.4, and 0.5 mol · kg-1. FTIR spectra were recorded on IFS 66 Bruker and also on Nicolet 8700 spectrometers. Two hundred fifty six scans were made with 4 cm-1 resolution. A cell with CaF2 windows was employed. The path length was 0.0299 mm, as determined interferometrically. The spectrometer was purged with dry air free of carbon dioxide (Bruker spectrometer) or dry nitrogen (Nicolet spectrometer). The temperature, monitored by a thermocouple inside the cell, was kept at 25.0 ( 0.1 °C by circulating thermostatted water through mounting plates of the cell. Spectral dependence with temperature of sodium acetate aqueous solution (m = 1 mol · kg-1) was measured in the range of 10-60 °C in 10 °C intervals using a Refrigerated and Heating Circulator F12-ED “JULABO”. Spectral Data Analysis The spectra have been handled and analyzed by commercial programs GRAMS/32 4.01 (Galactic Industries Corp., Salem, MA) and RAZOR (Spectrum Square Associates, Inc., Ithaca, NY) run under GRAMS/32.

J. Phys. Chem. B, Vol. 113, No. 23, 2009 8129 Spectral data have been analyzed following the published procedures, leading to separation of the spectrum of soluteaffected water only, based on the spectra of the entire solution series and the bulk HDO spectrum.29,30 The algorithm is based on the assumption that water in solution may be divided into additive contributions of bulk (b) and solute-affected (a) water. The vibrational spectrum of the latter, εa, may be calculated for each wavenumber from eq 1, using the pure HDO spectrum, εb, the “affected number”, N, and the derivative of absorption vs molality in the infinite dilution limit. M denotes the mean molar mass of water (H2O + 4% D2O).

εa )

1 ∂ε NM ∂m

( )

m)0

+ εb

(1)

An approximation of the experimental spectra ε(νi) vs molality m (including m ) 0) at each wavenumber νi by the least-squares method allows the calculation of the respective derivative. Normally, a low-order polynomial provides a satisfactory fit for this purpose.30 Equation 1 contains two unknowns: εa and N. The latter parameter is equal to the number of moles of water affected by one mole of solute. The proper value of it is found based on the published algorithm.29,30 In brief, the trial solute-affected water spectrum for a given N value is fitted using the baseline, analytical bands, and the bulk water spectrum. The product of the Gaussian and Lorentzian peak functions is used as the starting analytical band shape, but it might be replaced later by a pure function for a given band if the contribution of the other component is found to be negligible. The maximum value of N, for which the solute-affected water spectrum still contains a negligible amount of the bulk water spectrum (the practical threshold value is set at 0.5% of the total integrated intensity of the εa spectrum), is considered as the “true” value of N and the corresponding εa spectrum as the “true” affected water spectrum. Thus, both unknowns are obtained simultaneously. Computational Details Quantum mechanical calculations were performed in the framework of density functional theory, utilizing the hybrid B3LYP combination functional.34,35 Calculations were performed with the use of a standard basis set 6-311++G(d,p),36-38 which is of the triple-zeta type and is augmented with polarization and diffuse functions on all atoms. Numerical integrations were done with an UltraFine grid (99 radial shells per atom and 590 points per shell). The Berny algorithm with Tight convergence criteria was used for optimizations.45 Vibrational analysis was performed on optimized clusters to check for the absence of imaginary frequencies and thus confirm the existence of local energetic minima. Zero-point vibrational energies (ZPVE) and thermal corrections to energy, enthalpy, and Gibbs free energy were simultaneously obtained during the vibrational analysis. The ZPVE values were empirically scaled by the recommended scale factor equal to 0.9877.39 Gas-phase cluster structures were subsequently used as starting points for geometry optimization in the self-consistent reaction field (SCRF) approach. The polarizable continuum solvation model (PCM) was used,41-43 with the united atom topological model applied on the atomic radii of the UFF force field (the UA0 radii set). The GDIIS method44 was used in the PCM calculations to speed up the optimization process. All calculations were performed with the Gaussian 03 system.45 GaussView 3.0 (Gaussian, Inc., Wallingford, CT) and

8130

J. Phys. Chem. B, Vol. 113, No. 23, 2009

Gojło et al.

Figure 1. Derivatives (∂ε/∂m)m)0 for HDO spectra of aqueous solutions of sodium formate (blue), sodium acetate (red), and sodium propionate (green). The bulk HDO spectrum (black, dashed) has been shown for comparison purposes (in molar absorptivity scale divided by 25).

HyperChem 6.0 (Hypercube, Inc., Gainesville, FL) served as front-end interfaces and visualization tools. Results and Discussion The absorptivity derivatives shown in Figure 1 (together with the bulk HDO band for better comparison) have been used to determine the solute-affected HDO spectra according to eq 1 and the band shape analysis performed following the previously published procedure.29,30 Derivatives shown correspond to the linear regression of ε(ν) values vs solution molalities. Spectra of HDO affected by sodium formate (N ) 5.0), acetate (N ) 3.0), propionate (N ) 1.8), and nickel acetate (N ) 12.8) have been determined and are shown in Figure 2. A component band at 2735 cm-1 and above 2800 cm-1 in Figure 2a arises from absorption of the formate ion. We could expect that the total hydration numbers for sodium carboxylates are much higher than the affected numbers (N) found. It is well known, however,30,46 that the correspondence between affected number and hydration number is perfect only under certain circumstances, and usually N is lower than the experimentally or theoretically obtained hydration number, so that the affected spectra describe water status in a “concentrated” form. As it can be seen, HDO influenced by sodium carboxylates reveals two component bands at 2550 ( 5 and 2420 ( 6 cm-1. They changed their relative intensity, increasing the 2420 cm-1 band in the order from formate, acetate, to propionate ion (Figure 2a, 2b and 2c, respectively). Thereby, the center of gravity of this composite-“affected” HDO band shifts to lower wavenumbers in the same order, that is, of electron-donor ability of the carboxylic anion substituent. In fact, the lower wavenumbers component is slightly asymmetric, being the superposition of two symmetric analytical components. The higher wavenumbers band position is close to the well-known position for the Na+affected HDO.13,46 This fact does not explain, however, the above-mentioned changes of the component bands relative intensity. To test this supposition, the affected spectrum for nickel acetate has been determined and is shown in Figure 2d. It is known that HDO affected by the Ni2+ ion in the first hydration sphere displays a band at ca. 2420 cm-1 (ref 12 and references cited therein). The weak component band at 2625 cm-1 belongs to the Ni2+-affected HDO in the third hydration sphere, which is influenced by the anion.13 It can be added that HDO affected by Ni2+ in the second hydration sphere shows a band, which resembles the bulk water and therefore could be subtracted in the procedure of determining the affected spectrum.13

Figure 2. Decomposition of HDO spectra, affected by (a) sodium formate, (b) sodium acetate, (c) sodium propionate, and (d) nickel acetate, into analytical component bands: (solid line) original affected spectrum; (dashed lines) component bands; (dotted line) sum of the component bands.

At that point it seems clear that in the latter case the lower wavenumber band is a superposition of the Ni2+-affected and the acetate-anion-affected HDO and that the higher wavenumbers band at ca. 2550 cm-1 also belongs to the anion. On the other hand, in the case of sodium carboxylates the band at ca. 2550 cm-1 is a superposition of the Na+-affected and respective carboxylate-anion-affected HDO. To summarize the results for carboxylate ions, they display two component bands of affected HDO at ca. 2550 and 2420 cm-1. As it has been already extensively proved, water molecules surrounding nonpolar groups of a solute behave spectrally roughly as water in the bulk phase.1-3,30,47 Accordingly, we do not expect the substituent (R)-affected HDO band component. The relative intensity of observed bands in the carboxylate-affected water spectra changes with the electrondonor ability of the substituent: for propionate the band at 2420 cm-1 distinctly dominates, while for tifluoroacetate, as shown

Hydration of Carboxylate Anions

J. Phys. Chem. B, Vol. 113, No. 23, 2009 8131

Figure 3. Interatomic oxygen-oxygen distance distribution functions derived from the HDO spectra affected by sodium formate (blue), sodium acetate (red), and sodium propionate (green), along with the bulk HDO (black, dashed) distance distribution curve.

previously in our laboratory,13 it is virtually invisible. The single CF3COO--affected HDO band at 2552 ( 10 cm-1 is present in aqueous solution of lithium salt. The formate and acetate ions occupy intermediate positions in this order (Figure 2). The center of mass of these complex HDO solute-affected bands equals 2419, 2442, and 2490 cm-1 for sodium propionate, acetate, and formate, respectively. This unusual observation required detailed explanation. Further analysis may be enabled by transforming the molar absorptivity vibrational spectrum εa(νOD) to the probability distribution of the interatomic oxygen-oxygen distance P(ROO). The details of this transformation were given elsewhere.30,32 It is based on the empirical relationship between the ROO and νOD values.48 The correlation gives results in good agreement with that from X-ray scattering studies for liquid water30 as well as molecular dynamics simulations for aqueous solutions.3 Several correlations linking these values have been already published.1,3,30,33,46,60 The obtained probability distributions are shown in Figure 3 and compared with bulk water. This graph illustrates the series of water “structure-making” properties of the studied carboxylates: HCOONa < CH3COONa < C2H5COONa. It is also apparent that the probability distribution becomes more asymmetric in the reverse order, and the curve for sodium formate is distinctly structured. The experimental data on many solutes reveal a systematic dependence of the OD band position on the pKb of the solute, as shown in Figure 4. The two approximate curves illustrate the trend for monovalent ions and nonelectrolytes. The course of the line for nonelectrolytes resembles the previously obtained correlation between the most probable intermolecular oxygenoxygen distance of hydrated water (as a function of the band position) and the electron-donating power of the aprotic solute.1 This distance approaches a limit for strong electron donors because of restricted water compressibility. It appears now from Figure 4 that the electric charge of a solute delivers the energy necessary for additional collapse of the water structure: the observed shift of the water band to lower wavenumbers of ca. 75 cm-1 corresponds to a decrease of about 0.08 Å of the intermolecular oxygen-oxygen distance of hydrating water. Preliminary data obtained for some divalent anions, as well as for the trivalent ones (not shown here), seem to build up two further separate corresponding curves. These hydration patterns observed for anions resemble “band”-type hydration proposed previously for cations.13 Correlations shown in Figure 4 are based on a single point (single solute-affected HDO band) for each solute with the exception of carboxylate anions. Closer inspection of the figure

Figure 4. Correlation of the solute-affected HDO band position in maximum and pKb value of this solute. pKb values of triethylamine (TEA) and C2H5COO- taken from ref 49, F-, Cl-, Br-, I-, and ClO4from ref 50, dimethyl sulfoxide (DMSO) and tetrahydrofuran (THF) from ref 51, HDO, acetone (AC), and OH- from ref 52, CF3COO-, CH3COO-, HCOO-, and CF3SO3- from ref 53, sec-butylamine from ref 54, acetonitrile (AN) from ref 55, and acetamide (A), N-methylacetamide (NMA), N,N-dimethylacetamide (DMA), dimethylformamide (DMF), and urea from ref 56. Values of the OD stretching band position for F-, Cl-, Br-, I-, ClO4-, CF3SO3-, and CF3COO- were taken from ref 13, DMSO from ref 57, THF from ref 58, TEA, AC, AN, and HDO from ref 1, sec-butylamine from ref 2, A, NMA, DMA, and DMF from ref 3, urea from ref 59, and OH- from ref 60. Curve of trend for aprotic nonelectrolytes (black solid line), curve of trend for monovalent anions (red solid line), band position of the higher wavenumber component band of carboxylate-affected HDO (dashed blue line), which is at the level of the band position of acetone-affected HDO.

can lead to the conclusion that the proton-accepting ability of the first oxygen atom of the carboxylate group corresponds to that of the OH- ion, whereas for the second one it pairs with the oxygen atom of ketones. This should correspond to the wellknown mesomeric structure of the carboxylate anion, where the whole electric charge is localized on one oxygen atom, forming a single covalent bond with the carbon atom, whereas the second oxygen atom devoid of electric charge forms a double bond. The physical basis of this possibility in aqueous solution will be further discussed. Irrespectively, one ought to consider the explanation shown previously for hydration of HCOO- ion by Lindgren, Hermansson, and Wo´jcik.61 From the calculated densities of states for the first hydration shell surrounding the formate ion, the authors claimed that HDO molecules interacting with the COO- end of the anion give rise to two bands, one on either side of the bulk water band. The lower wavenumber band arises from OD oscillators pointing toward the ion, whereas the higher wavenumber one arises from OD oscillators pointing away from the anion. According to the authors, this is a pure electrostatic effect and should be observed for anions, which polarize the OD/OH bond of the HDO molecule more strongly than takes place in the bulk phase. They experimentally observed two bands also for F- and SO42- hydration water.62 Taking into account the band position of the high-wavenumber component bands reported for the mentioned ions, we proposed ascribing them to water molecules within the last hydration sphere of accompanying cations, second or third, which are simultaneously influenced by the anion.13,46 We did not observe the highwavenumber band corresponding to the carboxylates case for either the F- 30 or the OH- 60 ion in our former spectral results. To utilize the above-cited idea for explanation of results obtained in this paper would require observation that separation

8132

J. Phys. Chem. B, Vol. 113, No. 23, 2009

Gojło et al. TABLE 1: Optimized Intermolecular Oxygen-Oxygen Distances, ROO, in Calculated Propionate Water Clusters, C2H5COO- · n(H2O), in the Gas Phase ROOb n

Figure 5. HDO spectra affected by sodium acetate at 10 (blue) and 60 °C (red). The intensities of the spectra have been scaled to the same value at maximum for better comparison.

between both component bands of anion-affected HDO decreases with decreasing energy of the carboxylate interaction with water. Our spectral results do not show band shifts between the anions studied, despite the fact that the change of relative band intensities is observed. Thereby, the obtained results cannot be explained according to this concept. To clear up opinion on carboxylate hydration, the influence of temperature on carboxylate-affected HDO spectra has been

a

1 1a 2 3 4 5 6 7 8

1

2

3

4

2.863 2.844 2.849 2.806 2.820 2.816 2.823 2.784 2.737

2.669 2.682 2.702 2.703 2.716 2.612 2.602

2.924 2.963 2.882 2.900 2.876 2.857

2.746 2.752 2.674 2.804 2.811

a Number of water molecules. b All distances are in Angstroms; the numbers refer to respective labels in Figure 6.

determined in the range of 10-60 °C. The result is shown in Figure 5 for sodium acetate solution at extreme temperatures; affected spectra have been determined for N ) 3.0. It is evident that the intensity of the high-wavenumber band decreases at higher temperature and that the expected effect of the blue shift of both bands takes place with increasing temperature.

Figure 6. Hydrated propionate clusters, C2H5COO- · n(H2O), in their B3LYP/6-311++G(d,p) gas-phase-optimized geometries. Red spheres denote oxygen atoms; gray spheres denote hydrogen atoms. The numbers over the hydrogen bonds refer to columns in Table 1.

Hydration of Carboxylate Anions

J. Phys. Chem. B, Vol. 113, No. 23, 2009 8133

TABLE 2: Average Intermolecular Oxygen-Oxygen Distances in Respective Structural Positions of the Propionate Water Clusters, C2H5COO- · n(H2O), in the Gas Phase and the Corresponding Vibrational Frequencies na

ROO (R)b

ROO (β)c

1 2 3 4 5 6 7 8 expf

2.863 2.849 2.865 2.892 2.849 2.861 2.830 2.797 2.905 ( 0.009

2.844 2.669 2.682 2.724 2.728 2.695 2.708 2.707 2.757 ( 0.005

νODo (R)d νODo (β)d ∆νOD(R-β)e 2520 2509 2522 2541 2509 2519 2493 2463 2550 ( 6

2505 2301 2322 2381 2385 2340 2359 2357 2420 ( 6

0 ( 15 208 ( 104 200 ( 100 161 ( 80 125 ( 63 179 ( 90 134 ( 67 106 ( 53 130 ( 12

a

Number of water molecules. Average oxygen-oxygen distances (Å) refer to averaged values. b 1,3 from Table 1. c 2,4 from Table 1. d OD stretching band position (cm-1) calculated from the average distance for oxygen R and oxygen β; numbers and letters refer to Figure 6. e Differences between the OD stretching band position (cm-1). f Experimental data obtained in this work from HDO spectra.

Further arguments in the discussion are provided by quantum mechanical calculations. Figure 6 shows DFT-optimized gasphase geometries of propionate ion water clusters for 1-8 water molecules. The numbers over hydrogen bonds in the figure refer to columns in Table 1, where interatomic oxygen-oxygen distances, ROO, are shown. The comparison of computational and experimental results might be achieved by comparing interatomic distances with band positions of carboxylate-affected HDO spectra. We used the already mentioned ROO versus νOD correlation curve of Berglund et al.48 Band positions νoOD calculated this way and the corresponding ROO are shown in Table 2. We did not apply analytically calculated vibrational frequencies for the comparison with affected HDO spectra, as we verified that they poorly correlated with experimental results for liquid water, even when scaled or otherwise corrected for anharmonicity. Stretching vibrations of water are strongly anharmonic, and this anharmonicity is significantly dependent on water hydrogen-bond strength (stretching frequencies). Interaction of one water molecule with a carboxylate group is practically symmetric (Figure 6). This structure would give rise to a single band of hydrated HDO at a rather highwavenumber position, Table 2. It is interesting to observe that for clusters containing two and more water molecules interaction with oxygen atoms of carboxylate group is unsymmetrical in the sense of oxygen-oxygen distances and corresponding band positions of affected HDO, in agreement with experimental results. The geometries shown in Figure 6 seem to suggest the importance of repeating a specific structure of two water molecules interacting with the carboxylate group. It is characteristic that the water molecule interacting with other ones via the oxygen atom simultaneously forms the shortest H bond with the carboxylate group. It is a demonstration of the effect of water H-bonds cooperativity. This last effect is enhanced by the anticooperativity of H-bonds formation through both oxygen atoms of the carboxylate anion simultaneously. Respective calculation has been performed for RCOO- · nH2O clusters, where R ) CF3, H, CH3, C2H5 and n ) 1 or 2, Figure 7. Results are included in Table 3, where the effect of energy cooperativity/anticooperativity has been also defined (in the footnote). Negative values mean a cooperativity effect, while positive values mean anticooperativity. The effect ranges from 9.08 to 12.36 kJ · mol-1, depending on the substituent R. With increasing electron-donor abilities

Figure 7. Monohydrated and dihydrated clusters of carboxylate anion, RCOO-, in the gas phase. Violet balls denote substituent R, where R ) CF3, H, CH3, C2H5; the sketch refers to Table 3.

TABLE 3: Anticooperative Effects in RCOO-(H2O)2 Clusters in the Gas Phase R

∆EACa

CF3 H CH3 C2H5

9.08 11.25 11.97 12.36

a Hydrogen-bond energy differences between dihydrated clusters and sum of hydrogen-bond energy of monohydrated clusters (kJ · mol-1): ∆EAC ) ∆EHB[RCOO- · (H2O)2] - 2 · ∆EHB[RCOO- · (H2O)]; ∆EHB[RCOO- · (H2O)n] ) E298[RCOO- · (H2O)n] E298[RCOO-] - nE298[H2O]; see Figure 7.

of R, anticooperativity of forming H bonds with water molecules increases. This effect promotes an asymmetric-type interaction with the carboxylic group and appears to be quite strong for the alkyl substituent, of energy comparable with water H bonds in the bulk. From the calculated structures, it is possible to verify the hypothesis that has been formulated on the basis of the correlation in Figure 4. Accordingly, interaction with water could stabilize one of the mesomeric forms of the carboxylate anion: the one for which the whole electric charge is localized on one oxygen atom, forming a single covalent bond with the carbon atom. From the optimized structure for propionic acid (not shown) it appears that the length of (H)O-C(C2H5) is 1.34 Å, whereas the length of OdC(C2H5) is 1.22 Å. For the structures shown in Figure 6 for n g 2, the length of the O-C(C2H5) bond for the oxygen atom interacting stronger with water molecules is equal to 1.27 Å, whereas for the second one the corresponding length equals 1.25 Å. The difference equals 0.11 and 0.02 Å, respectively; thus, water stabilization of the discussed carboxylate mesomeric form seems to be an oversimplification. The tautomeric form of OdC(R)-O- does not exist in an aqueous solution, at least as a static limiting individuum. Mulliken atomic charges on oxygen atoms of the carboxylate group within RCOO-(H2O)4 hydrates reveal that significant nonequivalence still exists. Charges have been calculated for the gas phase and the PCM model for distinguished clusters structures (shown in Figure 8) and are included in Table 4. Possibly excluding one considered structure of the cluster (IIPCM), important charge asymmetry takes place on the oxygen atoms of carboxylate anions. On the other hand, the electronic charge is equally distributed on the oxygen atoms of anions interacting symmetrically with two water molecules, as shown in Figure 7. Results of respective calculations for the gas phase and the PCM model have not been shown.

8134

J. Phys. Chem. B, Vol. 113, No. 23, 2009

Gojło et al.

Figure 8. RCOO-(H2O)4 clusters in the gas phase and PCM model; violet balls denote group R, where R ) CF3, H, CH3, C2H5.

TABLE 4: Mulliken Atomic Charges on Oxygen Atom r and β of RCOO-(H2O)4 Clusters in the Gas Phase and PCM Model Calculated for Structures I and tII as in Figure 8 R

structure

R

β

R - βa

CF3

I-gas II-gas I-PCM II-PCM I-gas II-gas I-PCM II-PCM I-gas II-gas I-PCM II-PCM I-gas II-gas I-PCM II-PCM

-0.513 -0.543 -0.601 -0.535 -0.641 -0.602 -0.669 -0.612 -0.659 -0.608 -0.702 -0.626 -0.649 -0.589 -0.694 -0.607

-0.455 -0.449 -0.518 -0.506 -0.528 -0.494 -0.579 -0.573 -0.533 -0.493 -0.594 -0.580 -0.509 -0.485 -0.556 -0.576

-0.058 -0.094 -0.083 -0.029 -0.113 -0.108 -0.090 -0.039 -0.126 -0.115 -0.108 -0.046 -0.140 -0.104 -0.138 -0.031

H

CH3

C 2H 5

a

Difference of charges of oxygen R and oxygen β.

Unsymmetrical interaction of water molecules with the carboxylic group takes place because of nonequivalence of the proton-donor ability of water molecules H bonded to both oxygen atoms of the anions. This is a result of the already mentioned specific arrangement of water molecules solvating the COO- group and the cooperativity of water H bonds. The effect of nonequivalence in the interaction is reinforced according to the anticooperativity of H-bonds formation with the RCOO- anion (Table 3). In this light only a low number of water molecules interacting directly with the carboxylate oxygen (one or two) could justify the low-wavenumber band position (ca. 2420 cm-1) observed for anion-affected HDO. Low values of affected numbers N for the spectra in Figure 2 might confirm this statement. It is also likely that the number of water molecules directly hydrogen bonded to the two carboxylate oxygen atoms fluctuates with time. Taking into account the time scale resolution of infrared spectroscopy, this could account for instantaneous fine structure splitting of the HDO spectra. This hypothesis does not require any specific arrangements of water molecules around the carboxylate group of the anion; however, it does not seem to explain directly the dependence of the R type of the RCOOanion on HDO band splitting observed. The established phenomenon resembles solvent-induced symmetry breaking of a solute, already noticed in the literature. This was theoretically predicted for triiodide63 and nitrate anion63 in hydrogen-bonded solvents, like ethyl alcohol and especially water. In the latter case, symmetry breaking was also experimentally proved by resonance Raman spectroscopy. We previously observed similar behavior when studying metal ions interaction with diethyl phosphate:65 some cations (kat+) at

TABLE 5: Average Intermolecular Oxygen-Oxygen Distances in Respective Structural Positions for Calculated RCOO-(H2O)4 Clusters and the Corresponding Vibrational Frequencies in the Gas Phase R

structure

ROO(R)a

ROO(β)b

ROOc

νOD(R)d

νOD(β)d

CF3

I II I II I II I II

2.829 2.826 2.736 2.752 2.725 2.741 2.725 2.755

2.981 2.760 2.907 2.728 2.886 2.713 2.892 2.704

2.905 2.793 2.821 2.740 2.805 2.727 2.808 2.730

2492 2490 2395 2414 2381 2401 2381 2418

2594 2424 2552 2385 2537 2365 2541 2353

H CH3 C2 H 5

a Average distance for oxygen R (Å); numbers and letters refer to Figure 8. b Average distance for oxygen β (Å); numbers and letters refer to Figure 8. c Average value of ROO(R) and ROO(β) (Å). d OD stretching band position (cm-1) calculated from the average distance for oxygen R and oxygen β, numbers and letters refer to Figure 8.

specific concentrations in aqueous solutions demonstrated an asymmetric-type interaction with the phosphate group, forming kat+-O-P(dO)(O-Et)2 species, analogous to protonated diethyl phosphate. Inspection of νoOD values from Table 2 indicates that clusters with four or five water molecules give the results most resembling the experimental ones, taking into account both the bands positions as well as their difference. To examine this structure more closely, water clusters of the type 4H2O · R-COO- have been calculated, where R ) C2H5, CH3, H, and CF3. It appears that two different structures of comparable energy correspond to each of the carboxylate clusters. They are visualized in Figure 8. The first structure consists of two water molecules interacting with each oxygen atom of the carboxylate group, structure I, and the second structure with one water molecule interacting with one oxygen atom and two molecules interacting directly with the second oxygen of the carboxylate group (the forth molecule interacting with the latter water molecules), structure II. They are calculated by two methods, in the gas phase and within the polarizable continuum model (PCM) for an aqueous environment. Geometries for different carboxylic anions vary only quantitatively, but structure I calculated by the PCM method has been changed (as shown in Figure 8) resembling more structure II than the original structure I. Average intermolecular oxygen-oxygen distances in respective structural positions and the corresponding vibrational frequencies in the gas phase have been shown in Table 5 whereas for the PCM model in Table 6. Several conclusions arise from Tables 5 and 6. Only structure I in the gas phase indicates two component bands of affected water molecules, which is in accordance with the experiment. This seems to suggest that the specific molecular arrangement closest to the carboxylic group determines the hydration of the anion. It is not noticeable, when influenced by the surrounding

Hydration of Carboxylate Anions

J. Phys. Chem. B, Vol. 113, No. 23, 2009 8135

TABLE 6: Average Intermolecular Oxygen-Oxygen Distances in Respective Structural Positions for Calculated RCOO-(H2O)4 Clusters and the Corresponding Vibrational Frequencies in the PCM Model R CF3 H CH3 C 2H 5

structure ROO(R)a ROO(β)b I II I II I II I II

2.824 2.826 2.757 2.771 2.747 2.752 2.744 2.760

2.828 2.805 2.779 2.761 2.757 2.744 2.754 2.732

ROOc 2.826 2.816 2.768 2.766 2.752 2.748 2.749 2.746

νOD(R)d νOD(β)d νODe 2488 2490 2420 2436 2408 2414 2404 2424

2492 2470 2444 2425 2420 2405 2417 2390

2490 2480 2432 2430 2414 2409 2410 2407

a Average distance for oxygen R (Å); numbers and letters refer to Figure 8. b Average distance for oxygen β (Å); numbers and letters refer to Figure 8. c Average value of ROO(R) and ROO(β) (Å). d OD stretching band position (cm-1) calculated from the average distance for oxygen R and oxygen β; numbers and letters refer to Figure 8. e Average value of the OD stretching band position (cm-1).

TABLE 7: Total Energy, Differences of Energy, Enthalpy, Gibbs Free Energy, and Entropy Contribution at 298 K between Structure I and Structure II Calculated in the Gas Phase (Figure 8) R

structure

CF3

I II I II I II I II

H CH3 C2 H 5

E0a

∆E0b

-832.237467 6.08 -832.239783 -495.085347 2.03 -495.086119 -534.389488 2.02 -534.390257 -573.685678 -0.07 -573.685652

∆H298c ∆G298d T∆S298e 6.55

3.51

3.04

3.36

-1.92

5.15

3.23

-1.18

4.41

1.33

-4.00

5.33

a

Total energy, including scaled zero-point vibrational energy (Hartree). b Energy differences at 0 K (kJ · mol-1). c Enthalpy differences at 298 K (kJ · mol-1). d Gibbs free energy differences at 298 K (kJ · mol-1). e Entropy differences contribution at 298 K, calculated as ∆H298 - ∆G 298 (kJ · mol-1).

continuum solvent environment, but persists for large water clusters (n > 4) in the gas phase, as can be seen in Figure 6 and Table 2. Structure II gives practically single water bands, taking into account the full width at half-height of the corresponding affected bands. The position of this band corresponds to the experimental low-wavenumber component at ca. 2420 cm-1 in the case of propionate, acetate, and also formate anion. It is worth noticing that in the case of the trifluoroacetate anion, from structure I in the gas phase, the lower wavenumber component band appeared at a position characteristic of bulk water, which could have been removed in the procedure of determination of the solute-affected water spectrum. The structures considered above seem to be the most stable ones but of course are not unique in solution. Structure I in the gas phase could explain two spectral bands observed for HDO hydrating carboxylate group, but it does not explain the relative change of bands intensities in the studied order of anions. The differences of thermodynamic functions between structure I and structure II calculated in the gas phase are shown in Table 7. As can be seen, structure II appears to be more stable energetically. The enthalpy difference is roughly comparable to the thermal energy at 298 K, but it distinctly increases with the increasing electron-accepting ability of substituent R. This order has been generally reversed in the Gibbs function differences, because of the entropy contribution. Structure I, as more symmetric, corresponds to higher entropy than structure II. The entropic effect is very important in this respect,

determining the relative stability of aqueous clusters. It reminds once again how illusory could be predictions based solely on energetic effects. The order of the ∆G298 values (Table 7), obtained for calculated structures I and II in the gas phase (Figure 8), qualitatively explains the spectral results concerning the relative intensity of the component bands for carboxylate-affected HDO. Accordingly, hydration of RCOO- anion can be considered as some kind of equilibrium between solvent structure which induce symmetry breaking of the anion (structure I) and other less specific structures (like structure II), which do not induce such effect. The ∆G298 value for R ) CF3 correctly predicts the lack of band splitting for CF3COO- anion-affected HDO.13 For the rest of carboxylates, ∆G298 values predict the band splitting, as experimentally observed. The only discrepancy concerns the value for R ) C2H5, which is too negative with respect to the experimental order of the relative intensity of band components: H > CH3 > C2H5 . CF3. In this order the contribution of both component bands in the affected spectra decreases (Figure 2). Calculated structures have one essential disadvantage, as they do not provide the existence of a hydrophobic cage of water H-bonded network surrounding a solute-like carboxylate anion. Water molecules interacting directly with the oxygen atoms seem to depend more on this water cage than on the effect of bulk water molecules (approximated by the PCM model); they should be embedded into the hydrophobic hydration shell of a carboxylate anion. Conclusions Interaction of the solvating water with the carboxylic group of the carboxylate anion forces an asymmetric electron distribution between oxygen atoms of this group. This asymmetry perhaps fluctuates according to rotation of the solvent molecules directly involved in this interaction. Two water molecules appear to play the most important role in this phenomenon. They form a cooperative system, which plays a role as a bifunctional and nonequivalent proton donor with respect to both oxygen atoms of the carboxylate group. The entropic effect involved in this specific interaction appears to be important for explanation of experimental results obtained for the studied carboxylate aqueous solutions. The symmetry-breaking effect depends on the electron-donating ability of the substituent R in the RCOOanion and decreases with increasing temperature. The last effect may correspond to weaker proton-donor abilities of solvating water molecules at higher temperatures, because of the cooperativity effect of water H bonds. The water-induced symmetry breaking of the carboxylate anion, evidenced in this paper, seems to have important implications for hydration of numerous molecules of biological significance. Acknowledgment. This work was supported by the Republic of Poland scientific funds as a research project, within grant no. N N204 3799 33. Calculations were carried out at the Academic Computer Center in Gdan´sk (TASK). References and Notes (1) Gojło, E.; Gampe, T.; Krakowiak, J.; Stangret, J. J. Phys. Chem. A 2007, 111, 1827. (2) Gojło, E.; S´miechowski, M.; Stangret, J. J. Mol. Struct. 2005, 744747, 809. (3) Panuszko, A.; Gojło, E.; Zielkiewicz, J.; S´miechowski, M.; Krakowiak, J.; Stangret, J. J. Phys. Chem. B 2008, 112, 2483. (4) Mare´chal, Y. The Hydrogen Bond and the Water Molecule: The Physics and Chemistry of Water, Aqueous and Bio Media; Elsevier: Amsterdam, 2007; Chapter 10.

8136

J. Phys. Chem. B, Vol. 113, No. 23, 2009

(5) Beveridge, A. J.; Heywood, G. C. Biochemistry 1993, 32, 3325. (6) Max, J.-J.; Chapados, C. J. Phys. Chem. A 2004, 108, 3324. (7) Max, J.-J.; Chapados, C. J. Phys. Chem. A 2002, 106, 6452. (8) Max, J.-J.; Be´rube´, G.; Trudel, M.; Groleau, S.; Chapados, C. Langmuir 1998, 14, 5051. (9) Masamura, M. J. Phys. Chem. 1993, 97, 3157. (10) Markham, G. D.; Bock, C. L.; Bock, C. W. Struct. Chem. 1997, 8, 293. (11) Jorgensen, W. L.; Gao, J. J. Phys. Chem. 1986, 90, 2174. (12) Leung, K.; Rempe, S. B. J. Am. Chem. Soc. 2004, 126, 344. (13) Stangret, J.; Gampe, T. J. Phys. Chem. A 2002, 106, 5393. (14) Bergstro¨m, P.-Å.; Lindgren, J. J. Mol. Struct. 1991, 245, 221. (15) Kameda, Y.; Fukuhara, K.; Mochiduki, K.; Naganuma, H.; Usuki, T.; Uemura, O. J. Non-Cryst. Solids 2002, 312, 433. (16) Hartman, K. A. J. Phys. Chem. 1966, 70, 270. (17) Zundel, G. In The Hydrogen Bond: Recent deVelopment in theory and experiments; Schuster, P., Zundel, G., Sandorfy, C., Eds.; NorthHolland: Amsterdam, 1976; pp 683-766. (18) Molt, K.; Niemo¨ller, A.; Cho, Y. J. J. Mol. Struct. 1997, 410, 565. (19) Paquette, J.; Joliceur, C. J. Solution Chem. 1977, 6, 403. (20) Loginova, D.; Lileev, A.; Lyashchenko, A.; Aladko, L. J. NonCryst. Solids 2005, 351, 2882. (21) Kameda, Y.; Sasaki, M.; Yaegashi, M.; Tsuji, K.; Oomori, S.; Hino, S.; Usuki, T. J. Solution Chem. 2004, 33, 733. (22) Marcus, Y. Chem. ReV. 1988, 88, 1475. (23) Caminiti, R.; Cucca, P.; Monduzzi, M.; Saba, G.; Crisponi, G. J. Chem. Phys. 1984, 81, 543. (24) Banipal, T. S.; Singh, K.; Banipal, P. K. J. Solution Chem. 2007, 36, 1635. (25) Waldron, R. D. J. Chem. Phys. 1957, 26, 809. (26) Hornig, D. F. J. Chem. Phys. 1964, 40, 3119. (27) Falk, M.; Ford, T. A. Can. J. Chem. 1966, 44, 1699. (28) Kristiansson, O.; Lindgren, J.; de Villepin, J. J. Phys. Chem. 1988, 92, 2680. (29) Stangret, J. Spectrosc. Lett. 1988, 21, 369. (30) Stangret, J.; Gampe, T. J. Phys. Chem. B 1999, 103, 3778. (31) Badger, R. M.; Bauer, S. H. J. Chem. Phys. 1937, 5, 839. (32) Kristiansson, O.; Eriksson, A.; Lindgren, J. Acta Chem. Scand. A 1984, 38, 613. (33) S´miechowski, M.; Stangret, J. J. Chem. Phys. 2006, 125, 204508. (34) Becke, A. D. J. Chem. Phys. 1993, 98, 5648. (35) Stephens, P. J.; Devlin, J. F.; Chabalowski, C. F.; Frisch, M. J. J. Phys. Chem. 1994, 98, 11623. (36) Krishnan, R.; Binkley, J. S.; Seeger, R.; Pople, J. A. J. Chem. Phys. 1980, 72, 650. (37) Frisch, M. J.; Pople, J. A.; Binkley, J. S. J. Chem. Phys. 1984, 80, 3265. (38) Clark, T.; Chandrasekhar, J.; Spitznagel, G. W.; Schleyer, P. V. R. J. Comput. Chem. 1983, 4, 294. (39) Andersson, M. P.; Uvdal, P. J. Phys. Chem. A 2005, 109, 2937. (40) Peng, C.; Ayala, P. Y.; Schlegel, H. B.; Frisch, M. J. J. Comput. Chem. 1996, 17, 49. (41) Cance`s, M. T.; Mennucci, B.; Tomasi, J. J. Chem. Phys. 1997, 107, 3032. (42) Mennucci, B.; Tomasi, J. J. Chem. Phys. 1997, 106, 5151. (43) Cossi, M.; Scalmani, G.; Rega, N.; Barone, V. J. Chem. Phys. 2002, 117, 43. (44) Csaszar, P.; Pulay, P. J. Mol. Struct. (Theochem) 1984, 114, 31.

Gojło et al. (45) Frisch, M. J. ; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; Montgomery, J. A., Jr.; Vreven, T.; Kudin, K. N.; Burant, J. C.; Millam, J. M.; Iyengar, S. S.; Tomasi, J.; Barone, V.; Mennucci, B.; Cossi, M.; Scalmani, G.; Rega, N.; Petersson, G. A.; Nakatsuji, H.; Hada, M.; Ehara, M.; Toyota, K.; Fukuda, R.; Hasegawa, J.; Ishida, M.; Nakajima, T.; Honda, Y.; Kitao, O.; Nakai, H.; Klene, M.; Li, X.; Knox, J. E.; Hratchian, H. P.; Cross, J. B.; Bakken, V.; Adamo, C.; Jaramillo, J.; Gomperts, R.; Stratmann, R. E.; Yazyev, O.; Austin, A. J.; Cammi, R.; Pomelli, C.; Ochterski, J. W.; Ayala, P. Y.; Morokuma, K.; Voth, G. A.; Salvador, P.; Dannenberg, J. J.; Zakrzewski, V. G.; Dapprich, S.; Daniels, A. D.; Strain, M. C.; Farkas, O.; Malick, D. K.; Rabuck, A. D.; Raghavachari, K.; Foresman, J. B.; Ortiz, J. V.; Cui, Q.; Baboul, A. G.; Clifford, S.; Cioslowski, J.; Stefanov, B. B.; Liu, G.; Liashenko, A.; Piskorz, P.; Komaromi, I.; Martin, R. L.; Fox, D. J.; Keith, T.; Al-Laham, M. A.; Peng, C. Y.; Nanayakkara, A.; Challacombe, M.; Gill, P. M. W.; Johnson, B.; Chen, W.; Wong, M. W.; Gonzalez, C.; Pople, J. A. Gaussian 03, Revision C.02; Gaussian, Inc.: Wallingford, CT, 2004. (46) S´miechowski, M.; Gojło, E.; Stangret, J. J. Phys. Chem. B 2004, 108, 15938. (47) Pienia¸z˙ek, P. A.; Stangret, J. Vibr. Spectrosc. 2005, 39, 81. (48) Berglund, B.; Lindgren, J.; Tegenfeldt, J. J. Mol. Struct. 1978, 43, 179. (49) Reusch, W. An Introduction to Organic Chemistry: Molecular Structure and Acidity; http:/www.cem.msu.edu/∼reusch/VirtualText/ suppmnt2.htm (accessed Dec 22, 2008). (50) Scudder, P. H. Electron Flow in Organic Chemistry; Wiley: Hoboken, NJ, 1982; pp 261-263. (51) Reusch, W. An Introduction to Organic Chemistry: Basicity. http:/ www.cem.msu.edu/∼reusch/VirtualText/basicity.htm (accessed Dec 22, 2008). (52) pKa Values for Organic and Inorganic Bronsted Acids at 25 °C. http:/www.uaf.edu/chem/322Fa08/pkas.html (accessed Dec 22, 2008). (53) Reusch, W. An Introduction to Organic Chemistry: Ionization Constants of Heteroatom Organic Acids. http:/www.cem.msu.edu/∼reusch/ VirtualText/acidity2.htm (accessed Dec 22, 2008). (54) Gokel, G. W., Ed. Dean’s Handbook of Organic Chemistry, 2nd ed.; McGraw-Hill: New York, 2004. (55) Yilmaz M.; Memon S.; Tabakci, M.; Bartsch R. A. In New Frontiers in Polymer Research; Bregg, R. K., Ed.; Nova Science Publishers: Hauppauge, NY, 2006; Chapter 6. (56) Grant, H. M.; McTigue, P.; Ward, D. G. Aust. J. Chem., 1983, 36, 2211. (57) Stangret, J.; Kamien´ska-Piotrowicz, E. Proceedings of Polish Chemical Society Conference, Gdan´sk, Sept 22-26, 1997, S5-P35. (58) Stangret, J.; Gampe, T. J. Mol. Struct. 2005, 734, 183. (59) Panuszko, A.; Gojło, E.; Stangret, J. Proceedings of Polish Chemical Society Conference, Gdan´sk, Sept 18-22, 2006, S5-P32. (60) S´miechowski, M.; Stangret, J. J. Phys. Chem. A 2007, 111, 2889. (61) Lindgren, J.; Hermansson, K.; Wo´jcik, M. J. J. Phys. Chem. 1993, 97, 5254. (62) Bergstro¨m, P.-Å.; Lindgren, J.; Kristiansson, O. J. Phys. Chem. 1993, 95, 8575. (63) Zhang, F. S.; Lynden-Bell, R. M. Phys. ReV. Lett. 2003, 90, 185505. (64) Waterland, M. R.; Stockwell, D.; Kelly, A. M. J. Chem. Phys. 2001, 114, 6249. (65) Stangret, J.; Savoie, R. Can. J. Chem. 1992, 70, 2875.

JP811346X