Preparation and evaluation of chiral stationary phases based on N,N-2,4-(or 4,8)-disubstituted 4,5-(or 2,5)-dichloro-1,3-dicyanobenzene

Regioselective functionalization of 2,4,5,6-tetrachloro-1,3-dicyanobenzene (TCDCB) by nucleophilic substitution of the chlorine at C(4) with L-Ala, L- Phe or L-Pro, followed by amide-bond formation to lipophilic amines contain ing strong pi-donor group, and by final introduction of the spacer 3-aminop ropyltriethoxysilyl (APTES), provided a number of new brush-type chiral sel ectors in the form of 1-2:1 mixture of 2,4 and 4,6-di(alkyl)amino regioisom ers (8/9, 10/11, 12/13, 14/15, 20/21, 23/24). Linking these to silica gel ( Nucleosil 100-5) gave new chiral stationary phases for HPLC columns (CSP I- CSP VI). Being strong pi-basic selectors, most of these columns exhibited g ood resolution properties for pi-acid test racemates (TR 1-TR 9), specifica lly rac 3,5-dinitrobenzoyl-alpha-amino acid isopropyl-esters (DNB-AA). CSP V [1,3-dicyano-2,5(5,6) dichloro-6(2)-(gamma'-silica bound propylamino)-4-N -{[N-butyl]-N'-[(1R)-cyclohexylethyl]-N'[naphthylmethyl]acetamido}-aminoben zene] and particularly the dipeptide-containing CSP VI [2,5 (5,6)-dichloro- 6 (2)-(gamma'-silica bound propylamino)-4-N-(3',5'-dimethylanilido) L-alany l-L-prolyl-aminobenzene] proved to have the highest efficiency, comparable with the best commercial brush-type columns with pi-donor properties. Furth er evidence revealed that multiple hydrogen bonding via the amide group in the chiral environment and pi-pi interaction play a major role in chiral re cognition, whereas steric perturbations via nonbonding VDW interactions con tribute substantially only to the resolution of CSP III [2,5(5,6) -dichloro -6 (2)- (gamma'-silica bound propylamino)-4-N- (cyclohexylamido) -L-alanyl- aminobenzene]. This contribution is minor for the other CSPs. Chirality 12: 63-70, 2000. (C) 2000 Wiley-Liss, Inc.