Intrinsic Optical Activity and Conformational Flexibility: The Role of

Oct 24, 2013 - *Electronic Mail Address: [email protected]. Telephone/Fax Number: (203) 432-3975/(203) 432-6144. ... The intrinsic behavior of ...
0 downloads 5 Views 2MB Size
Article pubs.acs.org/JPCA

Intrinsic Optical Activity and Conformational Flexibility: The Role of Size-Dependent Ring Morphology in Model Cycloketones Priyanka Lahiri, Kenneth B. Wiberg, and Patrick H. Vaccaro* Department of Chemistry, Yale University P.O. Box 208107, New Haven, Connecticut 06520-8107, United States S Supporting Information *

ABSTRACT: The optical rotatory dispersion of two monocyclic ketones, (R)-3-methylcyclopentanone [R-3MCP] and (R)-3-methylcyclohexanone [R-3MCH], has been investigated under isolated and solvated conditions to explore the role of ring size/morphology and to elucidate the impact of environmental perturbations. Vapor-phase measurements of specific rotation, [α]Tλ , were performed at 355/633 nm by means of cavity ring-down polarimetry while complementary solution-phase work employed a canonical discrete-wavelength polarimeter to probe five distinct solvents. The magnitude of [α]Tλ was found to increase upon solvation, albeit to different extents for the two species of interest, with the attendant sign switching between the solution and vapor phases for λ ≥ 510.7 nm in the case of R-3MCH. Quantumchemical analyses suggest two low-lying conformers to exist for each ketone, distinguished by an equatorial or axial arrangement of the methyl substituent. Linear-response calculations built upon density-functional [DFT(B3LYP)/aug-cc-pVTZ] and coupled-cluster [CCSD/aug-cc-pVDZ] frameworks gave antagonistic chiroptical parameters for these isomers, which were combined with various energy metrics in a conformer-averaging ansatz to simulate the response for a thermally equilibrated ensemble. The intrinsic behavior of R-3MCP was reproduced best by averaging DFT optical-activity predictions according to relative populations deduced from free-energy differences; however, less satisfactory agreement was realized for isolated R-3MCH molecules. The sizable circular birefringence of R-3MCP can be attributed to inherent chirality of its twisted carbon ring whereas the more modest response of R-3MCH stems mainly from the lone stereogenic center. The implicit polarizable continuum model treated solvation effects in R-3MCP with moderate success, but failed to replicate solvent-dependent trends in R-3MCH. The relationship of dispersive optical activity to bulk characteristics of the surrounding medium, including dielectric constant, refractive index, and polarizability, is discussed with the goal of bridging the gap between isolated and solvated chiroptical properties.

I. INTRODUCTION The intrinsic chiroptical response of a molecule,1,2 which reflects the manifestation of natural optical activity in the absence of environmental perturbations, affords an important metric for elaborating chiral structure−property relationships and assessing attendant quantum-chemical predictions. Recent years have witnessed the rapid development of ultrasensitive spectroscopic techniques designed to probe such circulardifferential polarization characteristics under rarefied (vaporphase) conditions that allow for the direct observation of isolated-molecule behavior.2 In particular, cavity ring-down polarimetry (CRDP)3,4 has been exploited to interrogate the dispersive (nonresonant) phenomenon of circular birefringence (CB) for a variety of chiral gases,5−9 leading to optical rotatory dispersion (ORD) profiles that reveal the structural/electronic provenance of innate properties and accentuate the “solvation effects” sustained from a surrounding medium. The present work represents an extension of these efforts to a new family of molecules epitomized by two monocyclic ketones, (R)-3methylcyclopentanone [R-3MCP] and (R)-3-methylcyclohex© 2013 American Chemical Society

anone [R-3MCH], which differ primarily in the number of atoms forming the carbon ring. Ambient (25 °C) solutionphase and vapor-phase measurements of specific rotation, [α]Tλ [in deg dm−1(g/mL)−1], performed at discrete excitation wavelengths (λ) have been combined with computational analyses built upon the linear-response frameworks of densityfunctional (DFT) and coupled-cluster (CC) theory to elucidate the distinct chiroptical signatures engendered from these otherwise similar species. Owing to the presence of low-lying electronic manifolds that facilitated systematic investigations of the Cotton effect at readily accessible wavelengths, 3MCP and 3MCH were the targets of choice for various chiroptical studies performed during the last century. French, et al. reported the rotatory dispersion of 3MCH within its near-ultraviolet absorption band as early as 1936,10 while later studies of 3MCP and 3MCH in Received: September 5, 2013 Revised: October 19, 2013 Published: October 24, 2013 12382

dx.doi.org/10.1021/jp4089194 | J. Phys. Chem. A 2013, 117, 12382−12400

The Journal of Physical Chemistry A

Article

The low-lying conformations of 3MCH embody the canonical chair arrangement of the six-membered carbon ring, with the equilibrium between axial and equatorial orientations of the methyl substituent being explored extensively in both condensed and vapor phases. Variable-temperature ECD studies by Lightner and Crist30 determined the difference in free energy between these structural isomers, ΔG0, to be 2.1 and 3.1 kJ/mol in polar and nonpolar solvents, respectively. These results are in keeping with those obtained by analogous NMR investigations based on the lanthanide-induced shift methodology,31 which found ΔG = 4.6 ± 0.4 kJ/mol in CDCl3. The enthalpy change between gaseous 3MCH(A) and 3MCH(E), ΔH0, has been measured to be 6.49 ± 0.50 kJ/ mol by REMPI spectroscopy performed in a low-temperature molecular beam environment.32 All of these endeavors concur in asserting the (E) form to reside at lower energies than its (A) counterpart, although the relative extent of stabilization appeared to be affected strongly by the nature of the surrounding medium. Ongoing advances in quantum chemistry have spurred a renewed interest in dispersive optical activity,33−38 including that of the targeted cyclic ketones. He and co-workers16 reported the moderately large specific rotation of R-3MCP in carbon tetrachloride, acetonitrile, and methanol solutions for sodium D-line (589.3 nm) excitation, with concentrationdependent measurements being extrapolated to the limit of infinite dilution so as to give 150.3 ± 0.2, 144.9 ± 0.4, and 138.9 ± 0.3 deg dm−1(g/mL)−1, respectively. Relying on an independent conformer-averaging ansatz whereby chiroptical properties computed for the (E) and (A) forms contribute to the overall response in proportion to their relative populations, DFT (B3LYP) predictions of [α]D with augmented basis sets of double−ζ and triple−ζ quality were found to overestimate observed rotatory powers. Attempts to account for implicit solvation through a polarizable continuum model (PCM) further increased the discrepancies between experiment and theory. The dependence of [α]Tλ on wavelength was explored in a cyclohexane medium by Al-Basheer, et al.,29 finding the general shape of the ORD profile to be in good agreement with that estimated by introducing the PCM treatment into the B3LYP/aug-cc-pVDZ linear-response framework. In the case of R-3MCH, an early series of HF/DZP calculations were performed by Polavarapu and Zhao39 on chair and boat conformations optimized at the HF/6-31G level of theory. Focusing on the anomalous rotatory dispersion proximate to the π* ← n electronic transition, these authors compared their isolated-molecule predictions with the cyclohexane results of French and Naps10 to conclude that the chair form dominates under ambient conditions. The work described below builds on such pioneering efforts to execute synergistic experimental and computational analyses of optical rotatory dispersion in R3MCP and R-3MCH, with complementary solution-/vaporphase polarimetric data elucidating conformational degrees of freedom and their alteration by environmental perturbations (viz., solvation).

tandem revealed the difference in ring size/morphology to imbue the former with rotatory powers of substantially larger magnitude.11 Djerassi and co-workers examined the electronic circular dichroism (ECD) of these species as a function of temperature,12 with the pronounced changes in absorptive circular-differential signatures noted upon cooling of solutionphase samples furnishing evidence for active conformational degrees of freedom. More recent spectroscopic developments have motivated interest in the vibrational optical activity of both cyclic ketones, leading to synergistic experimental and computational analyses based upon the burgeoning techniques of vibrational circular dichroism13−16 (VCD) and Raman optical activity14,17−20 (ROA). The archetypical stature of 3MCP among chiral ketones led to its use as a model system for the first quantitative oneelectron analyses of rotatory strength in π* ← n transitions21 and for early theoretical descriptions of optical activity in carbonyl compounds;22−24 however, all of these pioneering efforts assumed a rigid molecular structure built upon a planar five-membered ring. Motivated by experimental evidence for conformational equilibria taking place under ambient conditions, Richardson, et al.25 performed semiempirical molecular orbital calculations that predicted the most stable geometries to embody a twisted carbon framework and either an axial (A) or an equatorial (E) disposition of the methyl substituent, where the former was estimated to be more stable than the latter by ∼5.9 kJ/mol. These authors also proposed distortion of the carbocylic framework to be intrinsically chiral and a more potent contributor to the composite chiroptical response than the lone asymmetric-carbon center. Although the bulk-gas microwave spectrum of 3MCP by Li26 did not provide conclusive structural information, comparison of extracted rotational constants to those deduced for possible conformers suggested observed features could be assigned primarily to those expected for a buckled cyclopentanone ring supporting an equatorial −CH3 group. Both 3MCP(A) and 3MCP(E) were identified in resonance-enhanced multiphoton ionization (REMPI) spectra acquired under cold supersonic expansion conditions,27 with relative intensities indicating the equatorial form to dominate − an assertion in keeping with molecularmechanics (MM3) and rudimentary Hartree−Fock (HF/631G*) computations. By applying REMPI detection in a variable-temperature molecular beam source, Kim and Baer28 determined the enthalpy difference, ΔH0, between these gaseous species to be 4.98 ± 0.59 kJ/mol in favor of the E isomer. He, et al.16 have reported a comprehensive investigation of R-3MCP conformational dynamics in the condensed phase that combined specific-rotation measurements with vibrational absorption (VA) and circular dichroism (VCD) probes. Analysis of temperature-dependent infrared profiles for the neat liquid indicated that ΔH0 = 4.84 ± 0.08 kJ/mol, with the assumption of a negligible entropy term (TΔS0) yielding an equilibrium (298 K) (A)/(E) population ratio of 0.13/0.87. This thermodynamic estimate for ambient conformer fractions was found to be in reasonable accord with those deduced from quantum-chemical modeling of VA and VCD intensity patterns. Al-Basheer and co-workers29 have used analogous van’t Hoff treatments to describe the T-dependence of the π* ← n ECD spectrum for R − 3MCP dissolved in 34 solvents possessing a broad range of physicochemical properties, giving ΔH0 and ΔS0 values of 3.63 ± 0.05 kJ/mol and 1.96 ± 0.10 J/K-mol, respectively, in the case of an acetonitrile medium.

II. EXPERIMENTAL AND COMPUTATIONAL METHODS Requisite samples of (R)-3-methylcyclopentanone and (R)-3methylcyclohexanone were obtained commercially (SigmaAldrich) with stated purities of 99% and 98%, respectively, and used without additional treatment. Quantitative assessment of percentage enantiomeric excess (%ee) did not prove feasible with available chiral gas-chromatographic instrumentation, but 12383

dx.doi.org/10.1021/jp4089194 | J. Phys. Chem. A 2013, 117, 12382−12400

The Journal of Physical Chemistry A

Article

Table 1. Observed Dispersive Optical Activity of R-3MCPa specific optical rotation [deg dm−1 (g/mL)−1] wavelength (nm) 355.00 365.02 435.83 546.07 578.39 589.30 633.00

isolated vapor 1033.8(50) 830.6c 326.4c 151.2c 128.1c 121.6c 100.23(56)

CH3OH solution

CH3CN solution

b

1048 860.88 364.41 170.20 144.77 137.71 114.2b

b

1137 924.50 381.19 175.74 148.52 142.33 117.8b

(C4H9)2O solution b

1310 1048.8 417.50 191.25 160.63 151.88 126.6b

C6H12 solution b

1349 1077.8 426.31 193.53 162.53 157.02 128.9b

CHCl3 solution 1241b 1006.0 412.86 189.30 162.26 152.64 127.0b

a

The intrinsic specific-rotation values determined for isolated (vapor-phase) R-3MCP molecules at 355 and 633 nm are compared with those measured at discrete visible and near-ultraviolet excitation wavelengths for five dilute solutions of this species. bExtrapolated from discrete solutionphase measurements using an analytical ORD profile. cInterpolated from discrete vapor-phase measurements using an analytical ORD profile.

was specified by the manufacturer to be 99%ee in the case of R3MCH, with similarly high levels being assumed for R-3MCP. As such, reported chiroptical parameters have not been corrected for chemical/stereochemical purity. Samples destined for vapor-phase studies were sealed in glass vessels that incorporated greaseless vacuum stopcocks and subjected to at least three freeze−pump−thaw cycles prior to filling of the CRDP apparatus so as to eliminate potential interference from entrained atmospheric gases. Solution-phase measurements of dispersive optical activity exploited an ambient (25 °C) quartz sample cell (10.000 ± 0.005 cm length) and a commercial polarimeter (Perkin-Elmer 341; ± 0.002° angular accuracy), the latter operating at discrete wavelengths filtered from Na I and Hg I atomic-emission lamps. All solvents were of spectrophotometric grade, with solutions maintained at sufficient dilution to discount the influence of solute−solute interactions (as gauged by limited studies of concentration dependence). For the specific-rotation parameters tabulated here, typical mass (molar) concentrations for R3MCP and R-3MCH were ∼0.0014 g/mL (0.014 mol/L) and ∼0.0043 g/mL (0.038 mol/L), respectively. Measurements of isolated-molecule rotatory powers were performed at two discrete excitation wavelengths, 355 and 633 nm, by exploiting instrumentation and procedures described elsewhere for the dispersive implementation of CRDP.5,6 The core of the vapor-phase polarimeter entailed a high-finesse cavity of L = 170.656(64) cm total length (one standard deviation uncertainty on final two significant digits in parentheses) formed from two concave mirrors (Los Gatos Research; 6 m radius of curvature) having specified reflectivities of ≥99.95% for 355 nm or ≥99.99% for 633 nm. An intracavity quarter-wave retardation plate of composite zero-order design with superior antireflection coatings (Alpine Research Optics; optically contacted with