Remarkable effects of substitution on stability of complexes and origin of the C–H⋯O(N) hydrogen bonds formed between acetone's derivative and CO₂, XCN (X = F, Cl, Br)

Abstract

The interactions of the host molecules CH₃COCHR₂ (R = CH₃, H, F, Cl, Br) with the guest molecules CO₂ and FCN (X = F, Cl, Br) induce significantly stable complexes with stabilization energies, obtained at the CCSD(T)/6-311++G(3df,2pd)//MP2/6-311++G(2d,2p) level, in the range of 9.2–14.5 kJ mol⁻¹ by considering both ZPE and BSSE corrections. The CH₃COCHR₂⋯XCN complexes are found to be more stable than the corresponding CH₃COCHR₂⋯CO₂ ones. The overall stabilization energy has contributions from both the >C=O⋯C Lewis acid–base and C–H⋯O(N) hydrogen bonded interactions, in which the crucial role of the former is suggested. Remarkably, we propose a general rule to understand the origin of the C–H⋯O(N) hydrogen bonds on the basis of the polarization of a C–H bond of a proton donor and the gas phase basicity of a proton acceptor. In addition, the present work suggests that the >C=O group can be a valuable candidate in the design of CO₂-philic and adsorbent materials, and in the extraction of cyanide derivatives from the environment.

---

1. Introduction

The miscibility and dissolution of materials in liquids and supercritical CO₂ (scCO₂) have attracted much attention due to the advantage of CO₂ in industrial and chemical processes over more conventional organic solvents, and in many potential applications of “green” chemistry.¹ Accordingly, during the last three decades, studies of the interaction between organic and/or inorganic compounds and CO₂ have been carried out on a large scale not only theoretically but also experimentally to rationalize the origin of the interactions in order to be able to control the solubility between macromolecules or colloidal particles and CO₂.^2–4 Recently, direct sol–gel reactions in scCO₂ have been used in the synthesis of oxide nanomaterials, oligomers and polymers.^5,6 Nevertheless, due to a lack of polarity and a dipole moment, scCO₂ is a poor solvent for most polar solutes and solvents. In this context, much effort has been dedicated to enhancing the applicability of CO₂ as a solvent through the use of “CO₂-philes” that can be incorporated into the structure of insoluble and poorly soluble materials, making them soluble in CO₂ at soft temperatures and pressures.⁷ Most of the available studies have concentrated on the complexes of hydrocarbons and their fluorinated derivatives with CO₂, such as CH_4−nF_n⋯CO₂, C₂H₆⋯(CO₂)_n and C₂F₆⋯(CO₂)_n (n = 1–4)^8–13 and suggest that the fluoro-substitution increases the solubility of hydrocarbons in scCO₂. However, these fluorine-based CO₂-philes are less favorable both economically and environmentally. Therefore, it is necessary to develop novel CO₂-philic materials which are cheaper and more benign towards human beings. There is also a great interest in understanding the origin of the interactions between molecules and CO₂ at the molecular level in order to effectively use CO₂ in different purposes.

In recent years, a large number of studies concerning the interaction of simple functionalized organic molecules, such as CH₃OH, CH₃CH₂OH,^14–16 CH₃OCH₃, CH₃OCH₂CH₂OCH₃,^17–19 HCHO, CH₃CHO, CH₃COOCH₃, CH₃COOH^20–23 and XCHZ (X = CH₃, H, F, Cl, Br; Z = O, S),²⁴ with CO₂ have been performed using quantum chemical methods. The strength of these complexes has been assigned to the main contribution of the Lewis acid–base interaction and/or an additional contribution from the C–H⋯O hydrogen bonded interaction. However, the role of the C–H⋯O hydrogen bond in increasing solubility remains questionable. Additionally, for a clearer understanding of chemical origins, it can be expected that other model molecules possessing electron-deficient carbon atoms and electron-rich N atoms, such as FCN, ClCN and BrCN, would be potential candidates to act as Lewis acids and Lewis bases in the presence of carbonyl compounds. Despite the fact that cyanides are not safe in solute–solvent processes, some of them are used in studies of intermolecular interactions.²⁵ Furthermore, the selection of these three cyanides interacting with carbonyl compounds is in order to understand the origin of interactions that may guide the use of substituted carbonyl polymer surfaces to adsorb and extract cyanide derivatives from the environment.

The A–H⋯B hydrogen bond is a weak non-covalent interaction whose significant importance is shown not only in chemistry and biochemistry but also in physics and medicine.²⁶ More noticeably, the existence of C–H⋯O(N) hydrogen bonds has been revealed in proteins, DNA, RNA, etc. Consequently, special attention has been paid to the C–H bond donors involved in this hydrogen bond in the last decade.^27–29 Up to now, several hypotheses and models have been proposed to unravel the reasons for the differences between contraction and elongation, which are respectively accompanied by a blue shift and a red shift in the stretching frequency of the A–H bond length upon complexation.^30–34 However, no general explanation has been formulated for the origin of the blue shifting hydrogen bond. Most hypotheses have been focused on explaining the origin of a specific blue shifting hydrogen bond when the hydrogen bonded complexes are already formed. It might be more appropriate if one considers the origin of a blue shifting hydrogen bond on the basis of the inherent properties of isolated isomers that are proton donors and proton acceptors, as reported in the literature.^24,32,34,35

In this study, we focus on the interactions between carbonyl compounds, including acetone (CH₃COCH₃) and its doubly methylated and halogenated derivatives (CH₃COCHR₂, with R = CH₃, F, Cl, Br), with CO₂ and XCN (X = F, Cl, Br) in order to probe the existence and the role of the >C=O⋯C Lewis acid–base interaction along with the C–H⋯O(N) hydrogen bonded interaction on the stabilization of the complexes examined. To the best of our knowledge, an investigation into these systems has not been reported in the literature. Another important purpose is how the durability of the complexes formed by the interactions of these compounds with CO₂ and XCN will be changed upon substitution. Remarkably, this work also aims to obtain the origin of the C–H⋯O(N) blue-shifting hydrogen bond on the basis of the polarizability of the C–H covalent bond and the gas phase basicity of the O and N atoms.

2. Computational methodology

Geometry optimizations for the monomers and complexes formed in the interactions of CH₃COCHR₂ (R = CH₃, H, F, Cl, Br) with CO₂ and XCN (X = F, Cl, Br) were carried out using the MP2/6-311++G(2d,2p) level of theory. Computations of the harmonic vibrational frequencies at the same level of theory followed to ensure that the optimized structures were all energy minima on potential energy surfaces, and to estimate the zero-point energy (ZPE). In order to avoid vibrational couplings between the CH₃ stretching modes of CH₃COCH₃ and CH₃COCH(CH₃)₂, the harmonic frequencies in both the monomers and relevant complexes were calculated by means of the deuterium isotope effect. Single point energy calculations were done in all cases at the CCSD(T)/6-311++G(3df,2pd) level based on the MP2/6-311++G(2d,2p) optimized geometries. Basis set superposition errors (BSSE) resulting from the CCSD(T)/6-311++G(3df,2pd)//MP2/6-311++G(2d,2p) level were obtained using the counterpoint procedure.³⁶ The interaction energies were derived as the difference in total energy between each complex and the sum of the relevant monomers, corrected for ZPE only (ΔE) or for both ZPE and BSSE (ΔE*). All of the calculations mentioned above were carried out using the Gaussian 09 program.³⁷ Topological parameters of the complexes were defined by AIM2000 software³⁸ based on Bader's Atoms in Molecules theory.^39,40 Finally, the electronic properties of the monomers and complexes were examined through natural bond orbital (NBO) analysis using the GenNBO 5.G program⁴¹ at the MP2/6-311++G(2d,2p) level.

3. Results and discussion

3.1 Interactions of CO₂ with CH₃COCHR₂ (R = CH₃, H, F, Cl, Br)

Four stable shapes of the complexes, which are denoted as H1, H2, H3 and H4, and their interaction energies at the CCSD(T)/6-311++G(3df,2pd)//MP2/6-311++G(2d,2p) level are shown in Fig. 1 and Table 1. Further evidence for the existence of intermolecular contacts in the complexes by means of AIM analysis are given in Fig. S1 and Table S1 in the ESI.† Indeed, as listed in Table S1 in the ESI,† the electron density and Laplacian values of bond critical points (BCP) of intermolecular contacts, including O6⋯C11 and O12⋯H8 in H1, O6⋯C11 in H2, O6⋯C11 and O12⋯C5 in H3, and O12⋯H3(9) and O13⋯H4(10) in H4, fall within the limitation criteria for the formation of weak interactions.⁴⁰ Accordingly, they are Lewis acid–base and hydrogen bonded interactions, both contributing to the strength of the complexes examined.

Fig. 1 The stable complexes from the interactions between CH₃COCH₃ and CO₂ (distances in Å).

Table 1 Interaction energies (given in kJ mol⁻¹) corrected for only ZPE, for both ZPE and BSSE, and BSSE of the complexes displayed in Fig. 1

	H1	H2	H3	H4
ΔE	−12.7	−11.3	−12.7	−4.7
BSSE	2.4	1.9	3.5	2.3
ΔE*	−10.3	−9.4	−9.2	−2.4

---

As shown in Table 1, the interaction energies obtained are quite negative, and increase in the order H1 < H2 ≈ H3 < H4. This means that the stability of the complexes reduces in the same order. The interaction energy of −10.3 kJ mol⁻¹ with both ZPE and BSSE corrections for H1 is between the values of −11.1 kJ mol⁻¹ reported in ref. 42 at CCSD(T)/aug-cc-pVTZ and −8.8 kJ mol⁻¹ reported in ref. 43 at MP2/aug-cc-pVDZ. Notably, in this work, the interaction of CH₃COCH₃ with CO₂ induces H3 to be less stable than H1, which is different from the results reported by Ruiz-Lopez et al.⁴⁴ The authors carried out the calculations at MP2/aug-cc-pVDZ and CCSD(T)/aug-cc-pVDZ, and suggested that H3 is more stable than H1 by an average value of 1.0 kJ mol⁻¹. Their predictions were obtained from the interaction energies without taking the BSSE correction into account since they reported a close BSSE value of 2.3 kJ mol⁻¹ for both H1 and H3. Our calculated BSSE values for these two structures at CCSD(T)/6-311++G(3df,2pd)//MP2/6-311++G(2d,2p) are 2.4 and 3.5 kJ mol⁻¹. It is clear that the contribution from BSSE to the overall stabilization energy for H3 is significantly larger than for H1. This leads to a larger magnitude in the strength of H1 compared to H3, as estimated in Table 1. In addition, to gain a more reliable evaluation, a higher level of theory (CCSD(T)/aug-cc-pVTZ//MP2/aug-cc-pVTZ) was used to obtain the interaction energies, which are −13.4 and −12.3 kJ mol⁻¹ for only the ZPE correction, and −11.7 and −9.5 kJ mol⁻¹ for both ZPE and BSSE corrections in the cases of H1 and H3, respectively. The results reliably suggest that H1 is more stable than H3, although their strengths are comparable when considering only the ZPE correction (cf. Table 1). The present work also locates two new stable geometries, denoted as H2 and H4, of the interaction between CH₃COCH₃ and CO₂, in which H2 (−9.4 kJ mol⁻¹) is negligibly more stable than H3 (−9.2 kJ mol⁻¹) when both ZPE and BSSE corrections are included.

Apart from the most stable H1 structure in all the CH₃COCH₃⋯CO₂ shapes, the presence of both Lewis acid–base and hydrogen bond interactions in this structure, the demand to evaluate the solubility of carbonyl compounds in scCO₂ and to reveal the role of the interactions in contributing to the strength of the formed complexes, we replaced two H atoms in a CH₃ group of CH₃COCH₃ by two CH₃, F, Cl and Br alike groups (denoted by CH₃COCHR₂, and considered as host molecules), and set out to investigate their interactions with the CO₂ guest molecule at the molecular level. The most stable geometries of the F, Cl and Br derivatives are similar to H1 and there is only a slight difference in the shape of the complex in the case of R = CH₃ (Fig. 2). The selected parameters of the complexes are collected in Table 2. In general, all the O⋯C and O⋯H contact distances are shorter or close to the sum of the van der Waals radii of the two relevant atoms (3.22 Å for the former and 2.72 Å for the latter). They are indeed in the range of 2.85–2.94 Å for O⋯C contacts and 2.38–2.79 Å for O⋯H contacts. Consequently, it can be roughly intimated that these interactions are >C=O⋯C (CO₂) Lewis acid–base type and C–H⋯O hydrogen bonds. An AIM analysis to lend further support to their existence and contribution to the complex strength is given in Table S2 of the ESI.†

Fig. 2 The stable shapes of the complexes between CH₃COCHR₂ and CO₂ (with R = H, CH₃, F, Cl, Br).

Table 2 Interaction energies (kJ mol⁻¹), BSSE (kJ mol⁻¹), changes in the bond length (Δr, Å), stretching frequency (Δν, cm⁻¹) and infrared intensity (ΔI, km mol⁻¹) of the C7–H8 bond in the complexes relative to the relevant monomers

	CH3COCHR2⋯CO2
	R1	R2(3)	ΔE	BSSE	ΔE*	Δr(C7H8)	Δν(C7H8)	ΔI(C7H8)
a For the C10–H17 bond in CH3COCH(CH3)2⋯CO2.b For the value of R3.
R = H	2.87	2.61	−12.7	2.4	−10.4	−0.00025	10.9	−3.1
R = CH3	2.85	2.77	−13.6	2.8	−10.7	−0.00054	14.1	−2.1
		2.79^b				−0.00054^a	6.0^a	−3.4^a
R = F	2.94	2.51	−11.9	2.8	−9.2	−0.00084	16.3	−10.1
R = Cl	2.92	2.40	−13.8	3.8	−10.1	−0.00068	15.0	10.6
R = Br	2.91	2.38	−13.8	2.4	−10.4	−0.00065	14.8	14.8

---

All the interaction energies are significantly negative, and range from −11.9 to −13.8 kJ mol⁻¹ when considering only ZPE, and from −9.2 to −10.7 kJ mol⁻¹ when considering both ZPE and BSSE (cf. Table 2). These obtained results are consistent with the suggestion of a larger magnitude in the strength of carbonyl relative to fluorocarbons and other functionalized compounds on interacting with CO₂. Thus, at the MP2/aug-cc-pVDZ level, the interaction energies are in the range of −3.7 to −4.9 kJ mol⁻¹ for the complexes of CO₂ with hydrocarbons such as CH₄, C₂H₆, CF₄, C₂F₆, and from −2.4 to −7.8 kJ mol⁻¹ for the complexes of CO₂ with CH_4−nF_n (n = 0–4).^9,11 In our previous work, the complexes of CO₂ with carbonyl and thiocarbonyl compounds such as XCHZ (X = CH₃, H, F, Cl, Br; Z = O, S) possess the interaction energies (ΔE*) from −5.6 to −10.5 kJ mol⁻¹ at CCSD(T)//aug-cc-pVTZ//MP2/aug-cc-pVTZ.²⁴ The fact that all the interaction energies of these complexes are considerably more negative than that of the dimer of CO₂ (ref. 22 and 24) (ΔE* ≈ −5.5 kJ mol⁻¹) suggests the CH₃COCHR₂⋯CO₂ complexes more stable than the dimer. In other words, the compounds functionalized with the >C=O counterpart could be an effective approach for the design of CO₂-philic materials.

We now discuss in more detail the substitution effects on the contribution of the interactions to the overall interaction energy in CH₃COCHR₂⋯CO₂. Generally, the association of CH₃COCHR₂ with CO₂ leads to a slight increase in the interaction energy (by including both ZPE and BSSE corrections, cf. Table 2) in the order CH₃ < H ≈ Br < Cl < F. This is in accordance with a report on the effect of substitution on the strength of complexes formed by halogenation of formaldehyde and acetaldehyde, and CO₂.^11,24 To evaluate strength of the complexes investigated, we calculated the proton affinity (PA, using CCSD(T)/aug-cc-pVTZ//MP2/aug-cc-pVTZ) at the O site of the >C=O group and the deprotonation enthalpy (DPE, using CCSD(T)/aug-cc-pVTZ//MP2/6-311++G(2d,2p)) of the C–H bond of the –CHR₂ group in isolated CH₃COCHR₂ monomers. The obtained values are listed in Table 3. The polarization of the C–H bond increases in the order CH₃ < H < F < Cl < Br, and the gas phase basicity at the O site increases in the order F < Cl < Br < H < CH₃. This is evidence for withdrawing electron density from the O atoms in the halogenated compounds, causing a larger decrement in the electron density at the O site on going from the Br- via Cl- and F-substituted derivative. In contrast, a CH₃ substitution results in an enhancement of the electron density at the O site in CH₃COCH(CH₃)₂ compared to CH₃COCH₃. Accordingly, along with the strengthening order of the interaction energy mentioned above, the total stabilization energy of the complexes contains contributions from both the >C=O⋯C (CO₂) Lewis acid–base interaction and the C–H⋯O hydrogen bond, in which the former dominates the latter. This is in agreement with previous results on the additional contribution of the hydrogen bond in stabilizing complexes and enhancing solubility in scCO₂.^22–24 In conclusion, the strength of CH₃COCHR₂⋯CO₂ complexes is gently increased when substituting two H atoms in a CH₃ group by two CH₃ groups for CH₃COCH₃, while it is slightly decreased by replacement with two halogen atoms (2F, 2Cl and 2Br). This is understood by the electron-donating effect of the CH₃ group and electron-withdrawing effect of the halogen groups, which makes electron density at the O atom in the methylated monomer larger than in the halogenated monomers and acetone.

Table 3 Deprotonation enthalpy of the C–H bond of the –CHR₂ group and the proton affinity at the O site of the >C=O group in the relevant monomers (all in kJ mol⁻¹)

	CH3COCH3	CH3COCH(CH3)2	CH3COCHF2	CH3COCHCl2	CH3COCHBr2
a Single point energy of the CH3COCR2 anions calculated at the respective geometry of the isolated monomer without optimization.
DPE^a	1704.6	1707.8	1669.9	1579.4	1558.4
PA	812.7	832.9	738.6	762.1	776.1

---

We continued the investigation into the character of the C–H⋯O hydrogen bond in these complexes. Its formation results in a shortened C–H bond length of 0.00025–0.00084 Å, and a blue-shifted stretching frequency of 6.0–16.3 cm⁻¹, when compared to those in the relevant monomers (cf. Table 2). It is, however, remarkable that the C–H infrared intensity is reduced in the range of 2.1–10.1 km mol⁻¹ for CH₃COCHR₂⋯CO₂, with R = CH₃, H, F, while it is enhanced by 10.6 and 14.8 km mol⁻¹ for CH₃COCHCl₂⋯CO₂ and CH₃COCHBr₂⋯CO₂, respectively, in spite of a contraction of the C–H bond length and a blue shift of its stretching frequency. Nevertheless, this observation is consistent with our previously reported results.^24,34 With the all obtained results, we would suggest that the C–H⋯O blue shifting hydrogen bond, which partly contributes to the complex strength, is also present in the complexes examined. This finding is different from Besnard's results^43,45 where they reported only the presence of a Lewis acid–base interaction between the electron donor O atom of CH₃COCH₃ and the electron acceptor C atom of CO₂ for CH₃COCH₃⋯CO₂.

It should be noted here that the general trend in the magnitude of the C–H bond length contraction is in accordance with the magnitude order of the polarity of the C–H bond in the isolated monomers. Thus, on going from F via Cl to Br, the polarization magnitude of the C–H bond in the isolated monomers increases, and this is accompanied by a decrease in the magnitude of the C–H bond length contraction and its stretching frequency enhancement when the complexes are formed (cf. Tables 2 and 3). This is not observed in the case of the CH₃ substitution group in the present work since the shape of the CH₃COCH(CH₃)₂⋯CO₂ complex differs from the remaining ones.

As reported by Joseph, Jemmis³³ and Szostak,⁴⁶ there is a good correlation between the NBO charge on the H atom of the proton donor involved in a hydrogen bond and the change in the bond length and stretching frequency upon complexation. They suggested that the blue shifting hydrogen bond was more likely to occur for donors bearing smaller positive charges on the H atom, and on the contrary, a red shifting hydrogen bond occurred for molecules with larger positive charges on the H atom. Our results further confirm this remark. Thus, the NBO charges at the MP2/6-311++G(2d,2p) level on the H atoms of the –CHR₂ group in the CH₃COCHR₂ monomers are calculated to be 0.216, 0.206, 0.139, 0.217 and 0.223 e for the H, CH₃, F, Cl and Br substituted derivatives, respectively.

An NBO analysis at the MP2/6-311++G(2d,2p) level was performed to evaluate the electron density transfer (EDT) between the host and the guest molecules, the electron density in the σ*(C7–H8) antibonding orbitals, the percentage of s-character at the C7(H8) hybrid orbitals and the intermolecular hyperconjugation energies. Selected NBO results are given in Table S3 of the ESI.† A positive EDT value implies electron transfer from the host to the guest molecules, and the inverse for a negative value. Following complexation, there are electron density transfers from CH₃COCH₃ and CH₃COCH(CH₃)₂ to CO₂, while reverse transfers are observed for CH₃COCHR₂⋯CO₂, with R = F, Cl and Br (cf. Table S3, ESI†). This implies that the C7–H8⋯O12 hydrogen bonded interactions become stronger on going from the CH₃ via H- to F- to Cl- and finally to the Br-derivative. A slight increase of 0.12–0.66% in the s-character percentage of the C7(H8) hybrid orbitals is obtained for all the examined complexes. Such an enhancement of the s-character contributes to the contraction of the C7–H8 bond. Remarkably, there is a different variation in the σ*(C7–H8) electron densities in the complexes compared to that in the relevant monomers. They are indeed reduced by 0.0002–0.0003 e for CH₃COCHR₂⋯CO₂, with R = F, Cl, Br, and are enhanced by 0.0004 e and 0.0009 e for CH₃COCH(CH₃)₂⋯CO₂ and CH₃COCH₃⋯CO₂, respectively. Therefore, a contraction of the C7–H8 bond along with a blue shift of its stretching frequency in the former complexes arises from both a decrease in the occupation of the σ*(C7–H8) orbital and an increase in the s-character percentage of the C7(H8) hybrid orbital, while in the latter complexes it is due to an overriding enhancement of the C7(H8) s-character relative to an increase in the σ*(C7–H8) electron density following complexation.

In a word, the bond contraction and the blue shift of the frequency of a C–H bond involved in hydrogen bonded complexes depend on its polarization in the isolated monomer. In particular, the weaker the polarization of a C–H covalent bond acting as a proton donor, the stronger its distance contraction and frequency blue shift as a result of complex formation, and vice versa.

3.2 Interactions of the guest molecules XCN (X = F, Cl, Br) with the host molecules CH₃COCHR₂ (R = H, CH₃, F, Cl, Br)

The interactions of CH₃COCHR₂ with XCN induce stable shapes for the complexes, similar to that of CH₃COCHR₂⋯CO₂ shown in Fig. 2. There is only a slight difference in the structures by replacing the O12 and O13 atoms of CO₂ by the N12 and X13 atoms of XCN, respectively, and their geometric shapes are presented in Fig. S2 of the ESI.† Some of the typical data are tabulated in Table 4. Most of the O6⋯C11 and N12⋯H8(H17) contact distances are in turn in the range of 2.82–3.15 Å and 2.27–2.76 Å, shorter than or comparable to the sum of the van der Waals radii of the two relevant atoms (3.22 Å and 2.75 Å for the O⋯C and N⋯H respective contacts). Consequently, >C=O⋯C Lewis acid–base and C–H⋯N hydrogen bond interactions exist in CH₃COCHR₂⋯XCN, in which the latter is quite weak. Further evidence for the existence and the stability of the mentioned interactions is provided by the results of the AIM analysis given in Table S4 of the ESI.†

Table 4 Intermolecular contact distances (in Å), interaction energies (in kJ mol⁻¹), and changes in the bond length (Δr, in Å), stretching frequency (Δν, in cm⁻¹) and infrared intensity (ΔI, in km mol⁻¹) of the C7–H8 bond in the complexes relative to the respective monomers

	CH3COCHR2⋯XCN
	R1	R2(3)	ΔE	ΔE*	Δr(C7H8)	Δν(C7H8)	ΔI(C7H8)
a For the C10–H17 bond.
R = H, X = F	2.84	2.57	−16.7	−13.9	−0.00013	10.0	−2.8
R = H, X = Cl	3.09	2.55	−15.1	−11.8	−0.00010	8.9	−1.7
R = H, X = Br	3.13	2.56	−13.5	−11.1	−0.00006	8.0	−2.0
R = CH3, X = F	2.82	2.76	−17.9	−14.4	−0.00084	12.8	−1.6
		2.74			−0.00084^a	9.1^a	−5.7^a
R = CH3, X = Cl	3.06	2.74	−18.3	−12.4	−0.00082	12.2	−1.9
		2.76			−0.00080^a	8.1^a	−3.8^a
R = CH3, X = Br	3.11	2.71	−15.1	−11.9	−0.00081	12.0	−2.0
		2.76			−0.00080^a	8.1^a	−3.2^a
R = F, X = F	2.89	2.46	−16.4	−13.0	−0.00090	17.5	−13.4
R = F, X = Cl	3.11	2.43	−16.0	−12.1	−0.00078	15.8	−13.5
R = F, X = Br	3.15	2.41	−14.8	−11.7	−0.00076	15.4	−13.4
R = Cl, X = F	2.87	2.33	−18.7	−14.1	0.00009	−0.2	24.1
R = Cl, X = Cl	3.08	2.29	−18.5	−13.3	0.00035	−0.4	40.1
R = Cl, X = Br	3.13	2.28	−17.4	−13.0	0.00038	−0.8	44.1
R = Br, X = F	2.87	2.31	−18.6	−14.5	0.00014	−1.2	42.9
R = Br, X = Cl	3.08	2.28	−18.4	−13.7	0.00038	−1.8	48.8
R = Br, X = Br	3.13	2.27	−17.3	−13.4	0.00040	−2.2	51.2

---

All the interaction energies of the complexes examined are significantly negative, more negative than those of CH₃COCHR₂⋯CO₂. In particular, they are in the range of −11.1 to −14.5 kJ mol⁻¹ for both ZPE and BSSE corrections, and from −13.5 to −18.7 kJ mol⁻¹ for only the ZPE correction (cf. Table 4). The obtained results suggest a larger magnitude in the strength of CH₃COCHR₂⋯XCN relative to CH₃COCHR₂⋯CO₂. In other words, replacement of the CO₂ by FCN or ClCN or BrCN guest molecule leads to an increase in the strength of the formed complexes. Nevertheless, the variations in the magnitude of their stabilization energies is not considerable, only about 1.0–1.5 kJ mol⁻¹.

As shown in Table 4, the strength of the complexes of acetone and its substituted derivatives with FCN increases in the order F < H < Cl < CH₃ ≈ Br, and H < F < CH₃ < Cl < Br for ClCN and BrCN. The obtained results show that the stability of the complexes contains contributions from both the >C=O⋯C Lewis acid–base interaction and the C–H⋯N hydrogen bond, since there are increases in both the C–H (–CHR₂) polarity and O-gas basicity on going from the F- via Cl- to Br-substituted derivative of CH₃COCHR₂ (cf. Table 3). Nevertheless, an enhanced contribution from the C–H⋯N hydrogen bond energy to the total stabilization energy should be suggested for the examined complexes, since CH₃COCH₃⋯XCN is, in general, less stable than CH₃COCHR₂⋯XCN (R = F, Cl, Br), in spite of the larger O-gas basicity of CH₃COCH₃. The considerable stability of CH₃COCH(CH₃)₂⋯FCN, which is close to the largest stability of CH₃COCHBr₂⋯FCN, might be mainly assigned to the >C=O⋯C Lewis acid–base interaction (due to the largest gas phase basicity at the O site and the largest electron-accepting capacity of FCN) and an additional cooperation between the two C–H⋯N hydrogen bonds. From the discussion of the comparison of the complex strength, it indicates that the C–H⋯N hydrogen bond is more stable than the C–H⋯O hydrogen bond.

For the same host molecules, the stability of all the CH₃COCHR₂⋯XCN complexes decreases in the order of the guest molecules from FCN via ClCN and to BrCN. This tendency is opposite to the increasing order of the PA at the N sites of the three guest molecules. Thus, the PAs at the N sites in the guest molecules calculated at the CCSD(T)/6-311++G(3df,2pd)//MP2/6-311++G(2d,2p) level are 690.1, 733.9 and 747.5 kJ mol⁻¹ for FCN, ClCN and BrCN, respectively. Remarkably, at the N site of FCN, our estimated PA of 690.1 kJ mol⁻¹ is very close to that of 690.3 kJ mol⁻¹ at the G2 level reported by Rossi et al. in ref. 47. In order to explain this observation, an NBO analysis for the guest molecules was performed using the MP2/6-311++G(2d,2p) level. The NBO charge values at the C atoms are estimated to be in turn 0.662, 0.163 and 0.072 e for FCN, ClCN and BrCN. This implies a decrease in the >C=O⋯C Lewis acid–base interaction in CH₃COCHR₂⋯XCN going from FCN to BrCN. The NBO analyses for the monomers and their complexes (given in Table S5 of the ESI†) indeed indicate an electron density transfer in decreasing order from the n(O) lone pairs of CH₃COCHR₂ to the π*(C≡N) orbital of XCN for each of the CH₃COCHR₂⋯XCN series going from FCN to BrCN. Remarkably, an additional transfer of electron density from the n(O) lone pairs of CH₃COCHR₂ to the σ*(C–F) orbital of FCN is observed following complexation. On the contrary, there is a slight increase in the stability of the C–H⋯N hydrogen bond from FCN to BrCN for each host molecule (cf. Table S5, ESI†). In summary, the crucial contribution to the overall stabilization energy in CH₃COCHR₂⋯XCN is dominated by the >C=O⋯C Lewis acid–base interaction, which overwhelms the C–H⋯N hydrogen bonded interaction. However, an enhancement in the role of the C–H⋯N hydrogen bond should be suggested for CH₃COCHR₂⋯XCN on going from FCN to BrCN.

As indicated from Table 4, there is an enhancement in the stabilization energy for each CH₃COCHR₂⋯XCN relative to the corresponding CH₃COCHR₂⋯CO₂ series. This is due to the fact that the PA at all the N sites in XCN is larger than that at the O site in CO₂, and more noticeably, the PA value is enhanced in the order of FCN to BrCN. Indeed, the PA at the O atom of CO₂ is 541.6 kJ mol⁻¹ at the CCSD(T)/6-311++G(3df,2pd)//MP2/6-311++G(3d,2p) level, which is significantly smaller than the PAs at the N atoms of XCN. These results firmly indicate a larger magnitude in the strength of the C–H⋯N interaction relative to the C–H⋯O interaction in stabilizing the complexes. In brief, substitution of the two H atoms in a CH₃ group of CH₃COCH₃ by two alike R groups (R = CH₃, F, Cl, Br) results in an increase in the strength of CH₃COCHR₂⋯XCN compared to CH₃COCH₃⋯XCN, while it negligibly affects the strength of CH₃COCHR₂⋯CO₂ relative to CH₃COCH₃⋯CO₂.

Following complexation, there are different changes in the C7–H8 bond length, its stretching frequency and infrared intensity in the examined complexes with respect to the relevant monomers. The C7–H8 bond lengths in CH₃COCHR₂⋯XCN (with R = H, CH₃, F) are slightly shortened by ca. 0.0001 Å, accompanied by increases of 8.0–17.5 cm⁻¹ in the stretching frequency and decreases of 1.6–13.5 km mol⁻¹ in the infrared intensity. In contrast, the interactions of CH₃COCHR₂ (with R = Cl, Br) with XCN lead to slight elongations (0.0001–0.0004 Å) of the C7–H8 bond length and tiny decreases (0.2–2.2 cm⁻¹) in its stretching frequency, along with enhancements (24.1–51.2 km mol⁻¹) to the corresponding infrared intensity compared to those in the relevant host derivatives. These characteristics point out that the C7–H8⋯N12 intermolecular interaction in the CH₃COCHR₂⋯XCN complexes belongs to the blue shifting hydrogen bond in the case of the CH₃-, H- and F-substituted R host derivatives and the red shifting hydrogen bond in the case of the Cl- and Br-substituted complexes.

In the case of the alike substituted derivatives (R = CH₃, H or F) interacting with XCN, there is a tiny decrease in the magnitude of the shortening of the C7–H8 bond length and the blue shift of its stretching frequency on going from the F- to Br-substituted guest molecule. Going in the same order of the guest molecules, an increase in the magnitude of the C7–H8 bond length elongation and its stretching frequency red shift is observed in each pair of CH₃COCHR₂⋯XCN (R = Cl, Br) (cf. Table 4). These results are due to both an increase in the gas phase basicity at the N atoms from FCN to BrCN, and a stronger polarization of the C7–H8 bonds in the CH₃COCHR₂ (R = Cl, Br) relative to the CH₃COCHR₂ (R = H, CH₃, F) host molecules (cf. Table 3). Accordingly, a proton acceptor with a stronger basicity should lead to a weaker contraction of the C–H bond acting as the proton donor and a weaker frequency blue shift, and vice versa. Thus, a red shift of the C7–H8 stretching frequency is predicted in the case of CH₃COCHR₂⋯XCN, with R = Cl, Br. In addition, as shown in Table 4, for each XCN, there is a shortened-to-lengthened change in the C7–H8 bond length and a blue-to-red shift of its stretching frequency in the examined complexes relative to the respective monomers. The obtained results should be firmly assigned to an increase in the polarity of the C7–H8 covalent bond on going from the CH₃ via H- to F- to Cl- and finally to the Br-substituted derivative.

Consequently, we would suggest that for the same proton acceptor, the weaker the polarization of a C–H bond involved in the hydrogen bond, the larger its bond contraction and frequency blue shift upon complexation, and also for the same C–H proton donor, the weaker the gas phase basicity of the proton acceptor, the larger its bond contraction and frequency blue shift, and vice versa. Thus, a similar trend in the change in the C7–H8 bond length and its stretching frequency is also obtained for the CH₃COCHR₂⋯CO₂ complexes. The contraction of the C7–H8 bond length and the blue shift of its stretching frequency are larger for each of the CH₃COCHR₂⋯CO₂ series than for each of the CH₃COCHR₂⋯XCN series, respectively (cf. Tables 2 and 4). Generally, an electron density transfer from the XCN guest molecules to the CH₃COCHR₂ host molecules is predicted in the complexes examined, except for the two CH₃COCH₃⋯FCN and CH₃COCH(CH₃)₂⋯FCN complexes (cf. Table S5 of the ESI†). This observation is similar to that obtained in the case of CH₃COCHR₂⋯CO₂, in which electron density is transferred from CO₂ to CH₃COCHR₂ for CH₃COCHR₂⋯CO₂ (R = F, Cl, Br), and a reverse tendency is seen for CH₃COCH₃⋯CO₂ and CH₃COCH(CH₃)₂⋯CO₂. Upon complexation, there are electron density increases of 0.0001–0.0022 e in the σ*(C7–H8) orbitals and C7(H8) s-character percentage enhancements of 0.26–0.97% in CH₃COCHR₂⋯XCN (R = H, CH₃, F) with respect to the relevant monomers. As a result, the enhancement of the C7(H8) s-character overcoming the increase in the occupation of the σ*(C7–H8) orbital plays a decisive role, giving rise to the contraction and the blue shift of the C7–H8 stretching frequency. However, the elongation and the red shift of the C7–H8 stretching frequency in CH₃COCHR₂⋯XCN (R = Cl, Br) are determined by the significant increases of 0.0007–0.0019 e in the population of the σ*(C7–H8) orbital dominating the increases of 1.23–1.53% in the C7(H8) s-character percentage as a result of complexation. A large increase in the electron density in the σ*(C7–H8) orbitals is due to the stronger interaction transferring electron density from the n(N) and π(C≡N) orbitals of XCN to the σ*(C7–H8) orbital of the host molecules on going from F via Cl and Br guest molecules (cf. Table S5, ESI†). This observation differs from the case of CH₃COCHR₂⋯CO₂, as discussed above.

4. Concluding remarks

The significantly stable structures from the interactions between the CH₃COCHR₂ (R = H, CH₃, F, Cl, Br) host molecules with the CO₂ and XCN (X = F, Cl, Br) guest molecules were located on the potential energy surface at MP2/6-311++G(2d,2p). The stability of the CH₃COCHR₂⋯CO₂ and CH₃COCHR₂⋯XCN complexes is due to the crucial role of the >C=O⋯C Lewis acid–base interaction and an additional cooperation from the C–H⋯O(N) hydrogen bond interaction. The CH₃COCHR₂⋯XCN complexes are found to be more stable than the CH₃COCHR₂⋯CO₂ ones, which is due to a stronger contribution from the C–H⋯N interaction relative to the C–H⋯O interaction to the overall stabilizing energy. Generally, the substitution of the two H atoms in a CH₃ group of CH₃COCH₃ by two alike R groups leads to an increase in the strength of CH₃COCHR₂⋯XCN relative to CH₃COCH₃⋯XCN, while it negligibly affects the strength of CH₃COCHR₂⋯CO₂ relative to CH₃COCH₃⋯CO₂. It is noteworthy that FCN is the strongest Lewis acid among the four guest molecules. This revelation is assigned to an additional transfer of electron density from the n(O) lone pairs of CH₃COCHR₂ to the σ*(C–F) orbital of FCN, which is not observed in the other cases, following complexation. The obtained results suggests that, for the same proton acceptor, the weaker the polarity of a C–H bond involved in the hydrogen bond, the larger its bond contraction and frequency blue shift as a result of complexation. Similarly, for the same C–H proton donor, the weaker the gas phase basicity of the proton acceptor, the larger its bond contraction and frequency blue shift, and vice versa.

Acknowledgements

This research is funded by the Vietnam National Foundation for Science and Technology Development (NAFOSTED) under grant number 104.03-2012.12. NTT thanks Prof. M. T. Nguyen for valuable discussions and Katholieke Universiteit Leuven for extending their computational facilities.

References

1. C. A. Eckert, B. L. Knutson and P. G. Debenedetii, Nature, 1996, 383, 313–318.
2. T. W. Zerda, B. Wiegand and J. Jonas, J. Chem. Eng. Data, 1986, 31, 274–277.
3. J. Sadlej, J. Makarewicz and G. Chalasinski, J. Chem. Phys., 1998, 109, 3919–3928.
4. A. J. Cox, T. A. Ford and L. Glasser, THEOCHEM, 1994, 312, 101–108.
5. R. Sui, J. M. H. Lo and P. A. Charpentier, J. Phys. Chem. C, 2009, 113, 21022–21028.
6. Y. Wang, L. Hong, D. Tapriyal, I. C. Kim, I.-H. Paik, J. M. Crosthwaite, A. D. Hamilton, M. C. Thies, E. J. Beckman, R. M. Enick and J. K. Johnson, J. Phys. Chem. B, 2009, 113, 14971–14980.
7. E. J. Beckman, J. Supercrit. Fluids, 2004, 28, 121–191.
8. A. Cece, S. H. Jureller, J. L. Kerschner and K. F. Moschner, J. Phys. Chem., 1996, 100, 7435–7439.
9. P. Diep, K. D. Jordan, J. K. Johnson and E. J. Beckman, J. Phys. Chem. A, 1998, 102, 2231–2236.
10. Y. K. Han and H. Y. Jeong, J. Phys. Chem. A, 1997, 101, 5604.
11. P. Raveendran and S. L. Wallen, J. Phys. Chem. B, 2003, 107, 1473–1477.
12. C. R. Yonker and B. J. Palmer, J. Phys. Chem. A, 2001, 105, 308–314.
13. M. Tafazzoli and A. Khanlarkhani, J. Supercrit. Fluids, 2007, 40, 40–49.
14. J. M. Stubbs and J. I. Siepmanna, J. Chem. Phys., 2004, 121, 1525–1534.
15. E. Dartois, K. Demyk, L. d'Hendecourt and P. Ehrenfreund, Astron. Astrophys., 1999, 351, 1066–1074.
16. P. Lalanne, T. Tassaing, Y. Danten, F. Cansell and S. C. Tucker, J. Phys. Chem. A, 2004, 108, 2617–2624.
17. J. J. Newby, R. A. Peebles and S. A. Peebles, J. Phys. Chem. A, 2004, 108, 11234–11240.
18. K. H. Kim and Y. Kim, J. Phys. Chem. A, 2008, 112(7), 1596–1603.
19. P. V. Ginderen, W. A. Herrebout and B. J. van der Veken, J. Phys. Chem. A, 2003, 107, 5391–5396.
20. R. Rivelino, J. Phys. Chem. A, 2008, 112(2), 161–165.
21. M. A. Blatchford, P. Raveendran and S. L. Wallen, J. Phys. Chem. A, 2003, 107, 10311–10323.
22. P. Raveendran and S. L. Wallen, J. Am. Chem. Soc., 2002, 124(42), 12590–12599.
23. P. Raveendran and S. L. Wallen, J. Am. Chem. Soc., 2002, 124(25), 7274–7275.
24. N. T. Trung, N. P. Hung, T. T. Hue and M. T. Nguyen, Phys. Chem. Chem. Phys., 2011, 13, 14033–14042.
25. E. Muchova, V. Spiro, P. Hobza and D. Nachtigallova, Phys. Chem. Chem. Phys., 2006, 8, 4866–4873.
26. G. R. Desiraju and T. Steiner, The weak hydrogen bond in structural chemistry and biology, Oxford University Press, New York, 1999.
27. X. Li, L. Liu and H. B. Schlegel, J. Am. Chem. Soc., 2002, 124, 9639–9647.
28. K. S. Rutkowski, A. Karpfen, S. M. Melikova, W. A. Herrebout, A. Koll, P. Wolschann and B. van der Veken, Phys. Chem. Chem. Phys., 2009, 11, 1551–1563.
29. E. S. Kryachko, in Hydrogen Bonding – New Insights, ed. S. J. Grabowski, Springer, Dordrecht, 2006, p. 293.
30. O. Donoso-Tauda, P. Jacque and J. C. Santos, Phys. Chem. Chem. Phys., 2011, 13, 1552–1559.
31. P. Hobza and Z. Havlas, Chem. Rev., 2000, 100, 4253–4264 , and references herein.
32. I. V. Alabugin, M. Manoharan, S. Peabody and F. Weinhold, J. Am. Chem. Soc., 2003, 125, 5973–5987.
33. J. Joseph and E. D. Jemmis, J. Am. Chem. Soc., 2007, 129, 4620–4632.
34. N. T. Trung, T. T. Hue and M. T. Nguyen, J. Phys. Chem. A, 2009, 113, 3245–3253.
35. A. Yong Li, J. Chem. Phys., 2007, 126, 154102–154113.
36. S. F. Boys and F. Bernadi, Mol. Phys., 1970, 19, 553–566.
37. M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria, M. A. Robb, J. R. Cheeseman and G. Scalmani et al., Gaussian 09 (version A.02), Gaussian Inc., Wallingford CT, 2009.
38. AIM 2000, designed by Friedrich Biegler-König, University of Applied Sciences, Bielefeld, Germany, 2000.
39. R. F. W. Bader, Chem. Rev., 1991, 91, 893–928.
40. P. Popelier, Atoms in Molecules, Pearson Education Ltd., Essex, U.K., 2000.
41. E. D. Glendening, J. K. Baderhoop, A. E. Read, J. E. Carpenter, J. A. Bohmann and F. Weinhold, GenNBO 5.G, Theoretical Chemistry Institute, University of Wisconsin Madison, WI, 1996–2001.
42. N. T. Trung and M. T. Nguyen, Chem. Phys. Lett., 2013, 581, 10–15.
43. Y. Danten, T. Tassaing and M. Besnard, J. Phys. Chem. A, 2002, 106, 11831–11840.
44. M. Altarsha, F. Ingrosso and M. F. Ruiz-Lopez, ChemPhysChem, 2012, 13, 3397–3403.
45. M. Besnard, M. I. Cabaco, S. Longelin, T. Tassaing and Y. Danten, J. Phys. Chem. A, 2007, 111, 13371–13379.
46. R. Szostak, Chem. Phys. Lett., 2011, 516, 166–170.
47. F. Bernardi, F. Cacace, G. Occhiucci, A. Ricci and I. Rossi, J. Phys. Chem. A, 2000, 104, 5545–5550.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c3ra47321j

---