The effect of quasispherical and chainlike solutes on the nematic to isotropic phase transition in liquid crystals

Abstract

The effects of solute molecular structure (size, shape and flexibility) and solvent molecular structure (length‐to‐breadth ratio and end‐chain flexibility) on nematic phase stability in dilute binary mixtures of nonmesomorphic solutes and nematogenic solvents are examined through experiment and theory. Addition of the perturbing solute to the liquid‐crystalline solvent leads to depression of the nematic–isotropic (NI) transition temperature and formation of a two‐phase region. Directly determined moduli of the slopes, β_n and β_i, of the nematic and isotropic phase boundary lines in the reduced NI transition temperature (T*) vs solute mole fraction (x₂) diagrams are reported for quasispherical and chainlike solutes with two nematogenic solvents. The systems studied are the quasispheres Et₄C (tetraethylmethane) and R₄Sn (R = CH₃, C₂H₅, n‐C₃H₇ and n‐C₄H₉) and the chains n‐C₈H₁₈ through n‐C₁₄H₃₀, mixed with p‐methoxybenzylidine‐p′‐n‐butylaniline (MBBA) and p‐n‐pentyl‐p′‐cyanobiphenyl (5CB). Also reported are indirectly determined β^∞_n and β^∞_i values (limit as x₂→0), using a novel approach combining differential scanning calorimetry (for the pure solvent contribution) and gas–liquid chromatography (for the solution contribution), for Et₄C and n‐C₅H₁₂ through n‐C₁₁H₂₄, with MBBA, 5CB, p‐azoxyanisole (PAA), and p,p′‐di‐n‐hexyloxyazoxybenzene (DHAB). For the systems in common, the average difference between the directly and indirectly determined β values is approximately 10% and, qualitatively consistent with lattice model predictions, the comparison suggests slight curvature of the phase boundary lines. The experimental β values, as a function of increasing solute size, are found to double (roughly) for the quasispheres and increase only slightly for the chains, reflecting the concurrent behavior of the solution contribution to β. The thermodynamic results for the quasispherical solutes are compared in some detail with predicted values from statistical–mechanical theories based on rigid‐rod solvent molecules: (a) lattice model, (b) virial expansion treatment, (c) molecular‐field model (after Maier and Saupe), and (d) van der Waals model. All four models correctly predict the observed trend of increasing β_n and β_i with increasing solute size and yield predicted slopes which are within a factor of 2 of experiment. All are found to be deficient to a minor or major extent in their predictions of the solvent and solution contributions to the β values. The more tractable lattice model is used to examine the chainlike solutes and the effect of solvent end‐chain flexibility. It correctly predicts the qualitative features of the observed dependence of β on solute size for the different solute structures (including rigid‐rod solutes) and indicates that dissolved n‐alkane solutes have appreciable (effective) chain flexibility in nematic solvents. Incorporation of some solvent end‐chain flexibility in the lattice model markedly improves agreement with experiment, primarily through better quantitative prediction of the solution contribution.