Dynamic HPLC on chiral stationary phases: A powerful tool for the investigation of stereomutation...

Ilaria D’Acquarica1

Francesco Gasparrini1

Marco Pierini1

Claudio Villani1

Giovanni Zappia2

1Dipartimento di Studi diChimica e Tecnologia delleSostanze Biologicamente Attive,Universit� degli Studi di Roma“La Sapienza”, Rome, Italy

2Institute of Chemical Sciencesand Institute of PharmaceuticalChemistry, University of Urbino,Urbino, Italy

Review

Dynamic HPLC on chiral stationary phases:A powerful tool for the investigation ofstereomutation processes

Dynamic HPLC on enantioselective stationary phases has become a well-establishedtechnique to investigate chiral molecules with internal motions that result in ste-reoinversion and occur on the time scale of the separation process. Kinetic param-eters for the on-column interconversion phenomena can be extracted from experi-mental peak profiles by computer simulation or by direct calculation methods. Thetechnique has been used in a wide range of temperatures and is complementary inscope to dynamic NMR spectroscopy.

Keywords: Dynamic HPLC / Chiral stationary phases / Stereomutation, enantiomerization / Com-puter simulation /

Received: March 22, 2006; revised: April 21, 2006; accepted: April 23, 2006

DOI 10.1002/jssc.200600129

1 Introduction

The direct HPLC separation of enantiomers on chiral sta-tionary phases is nowadays one of the preferred modes toquantitatively characterize chiral compounds in termsof enantiomeric composition, and to obtain discreteamounts of pure, single enantiomers on large scales.Enantiomer separation is realized through non-covalent,reversible formation of diastereomers as the two enantio-mers of the analyte interact with the chiral, enantiomer-enriched selector present in the system. Reversible inter-actions of the individual enantiomers with the station-ary phase are usually referred to as primary equilibria[1–5].

The enantioselectivity has a thermodynamic origin andis simply related to a difference in the stability of the dia-stereomeric species formed by the two enantiomers andthe chiral selector. A simple equation (Eq. 1) relates chro-matographically derived enantioselectivity a with thisenergy difference. This value may approach the truevalue of the thermodynamic enantioselectivity of theselector-analyte-solvent system when the associationwith the chiral selector governs the retention of the

enantiomers in the chromatographic system and other,non-selective types of interactions are negligible.

DDG82,1 = DG82 – DG81 = –RTln (k92/k91) = –RTlna2,1 (1)

where dDG82,1 is the difference in the interaction Gibbsenergies between the two enantiomers and the systemselector, R is the gas constant, T is the absolute tempera-ture and a2,1 is the ratio of the retention factors k92 and k91of the two enantiomers.

Both enantioselective and unspecific association of theenantiomers with the stationary phase may have aretarding or activating effect on the internal molecularmotions of stereolabile chiral species during their pas-sage through the chromatographic column.

2 Theory of dynamic HPLC of chiralstereolabile compounds

In recent years enantioselective HPLC has been extendedto the investigation of stereolabile compounds. DynamicHPLC (DHPLC) in the form of variable temperature orvariable flow chromatography can be employed to deter-mine the enantiomerization (the reversible isomeriza-tion of one enantiomer into the other) barrier of chiralstereolabile species when the interconversion takes placeat the time scale of the separation process [6–12].

This technique has been recently used for the study ofinternal molecular dynamics of a range of chiral, stereo-labile compounds, and for the determination of the per-tinent kinetic parameters. Attractive features of this

Correspondence: Professor Claudio Villani, Dipartimento diStudi di Chimica e Tecnologia delle Sostanze Biologicamente At-tive, Universit� degli Studi di Roma ”La Sapienza”, P.le A. Moro5, 00185 Roma, ItalyE-mail : [email protected]: +39-06-4991-2780

Abbreviations: CSP, chiral stationary phase ; DHPLC, dynamicHPLC

i 2006 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.jss-journal.com

1508 I. D’Acquarica et al. J. Sep. Sci. 2006, 29, 1508 –1516

J. Sep. Sci. 2006, 29, 1508 –1516 Dynamic HPLC on chiral stationary phases 1509

technique are: (i) only small amounts of racemic sampleare required, (ii) a wide range of working temperaturesare available, and (iii) the interconverting species arephysically separated and, under certain circumstances,they can be further characterized, e. g., by chiropticalmethods.

A dynamic chromatographic experiment entails tworesolvable enantiomers that interconvert on the separa-tion time scale. The HPLC time scale extends from sec-onds in the cases of fast HPLC on short columns to hourswhen standard columns and low linear velocities of theeluent are used. For a given column format (columndimensions and particle size of the packing material),typical peak shape deformations are observed inresponse to changes in the column temperature and/orin the flow rate of the eluent. Peak shape characteristicof an on-column interconversion are usually observed asa profile between the two resolved peaks that does notreach the baseline (plateau).

The resulting peak shapes contain the necessary informa-tion to extract pertinent kinetic parameters related tothe exchange process between the two species.

Several methods (the theoretical plate model [8], the sto-chastic model [7–10], the continuous flow model [13],peak deconvolution methods [10], approximation func-tions [10], and unified equation [12]) have been used toextract kinetic data from experimental elution profiles,and each method exploits a different theoretical frame-work to describe the dynamic system in which primaryand secondary equilibria take place inside the column(Fig. 1).

The theoretical plate model describes the chromato-graphic separation as a discontinuous process, assuming

that all steps proceed repeatedly in separate uniform sec-tions of a column consisting of N plates. Every theoreticalplate is considered as a distinct chemical reactor inwhich three events take place: distribution of the speciesbetween the mobile and stationary phases, interconver-sion in both phases, and shifting of the mobile phase tothe next plate [8]. The stochastic model describes thechromatographic separation using time-dependent dis-tribution functions [7, 10]. The elution profile of twointerconverting species is given by the sum of the distri-bution functions of the non-interconverted species andthe probability of density functions of the interconvertedspecies. In the continuous flow model, general analyticalexpressions based on the Damk�hler number describethe chromatographic mass distribution at the columnoutlet for a solute subject to secondary equilibria [13].

The effects of changes in some critical parameters (km,kst, retention) on the appearance of the chromatographicprofiles are illustrated in Figs. 2–4, where the theoreticalplate model is used to generate typical plots for a chiralstereolabile analyte resolved on a chiral stationary phase(CSP).

Figure 5 shows the individual traces of two interconvert-ing enantiomer that are concomitantly separated on aCSP: Tailing of the first and fronting of the second elutedenantiomers are due to analyte molecules that haveundergone at least one stereochemistry inversion duringelution.

Recent applications of DHPLC to the study of chiralstereolabile compounds (Fig. 6) will be illustrated here,with a special emphasis on two aspects of the technique:(i) the effect of the stationary phase on the interconver-


Figure 1. Primary (vertical arrows) andsecondary (horizontal arrows) equilibriataking place during chromatography oftwo interconvertible enantiomers R andS on a CSP. Left: actual interconversionsoccurring in the mobile and stationaryphases, with R enantiomer eluting first;right: the apparent equilibria.


sion process, and (ii) cryo-HPLC study of stereolabile com-pounds featuring low barrier to interconversion.

3 Stationary phase effectsThe apparent rate constants, ka

e and kaE, determined by

simulation methods are related to those involved in theactual equilibria occurring in the mobile phase, km

e = kmE,

and in the stationary phase, kste and kst

E. The apparentrate constants are not in principle equal to the rate con-stants of the interconverting processes occurring in themobile and stationary phases. The potential perturbingeffect of the stationary phase on the interconversion canbe studied by combining the DHPLC experiment withindependent measurements that yields the rate con-stants in free solution. Off-column thermal racemizationexperiments are conveniently performed on single enan-

tiomers that are first collected by enantioselective HPLCand then converted back to the racemate, while monitor-ing enantiomer composition as a function of time bychromatography or chiroptical methods (polarimetry orcircular dichroism spectroscopy). Alternatively, free solu-tion rate constants are extracted from dynamic NMR(DNMR) experiments. However, the NMR time scale isdifferent and dynamic spectra are observed across ahigher temperature range compared to DHPLC.

For a series of axially chiral tertiary amides of 1-naphthoic acid, the whole set of rate constants wasobtained by a combination of DHPLC and off-columnthermal racemization experiments. Tertiary amides of 1-naphthoic amides have a ground state conformation in


Figure 2. Transition from slow to fast exchange at constantretention and efficiency. Simulation parameters from a to h:km = kst = 0.1; 0.2; 0.3; 0.4; 0.5; 1.0; 5.0 min – 1; N1 = N2 =1500; t0, t1, t2 = 2.2; 10.0; 20.0 min.

Figure 3. Stationary phase effects. Transition from inter-mediate to fast exchange at constant retention and effi-ciency. Simulation parameters: km fixed at 0.2 min – 1; from ato e: kst = 0.0; 0.1; 0.2; 0.3; 0.5 min-1; N1 = N2 = 1500; t0, t1,t2 = 2.2; 10.0; 20.0 min.

Figure 4. Retention and selectivity effects. Transition fromintermediate to fast exchange at constant efficiency andkinetic parameters. Simulation parameters: km = kst =0.1 min – 1; N1 = N2 = 1500; t0 = 2.2 min; from a to e: t1,t2 = 8.0, 22.0; 10.0, 20.0; 12.0, 18.0; 13.0, 17.0; 14.0, 16.0.

Figure 5. Simulated plots showing separate peak profilesfor the first and second eluted peaks and their convolution togive the total profile featuring the interconversion plateaubetween the resolved peaks.


which the aromatic and carboxamide planes are twistedat about right angles, thus rendering the moleculechiral. Rotation around the Ar-CO bond interconvertsthe two conformational enantiomers, and the substitu-tion pattern at the 2 and 8 positions of the naphthyl ringregulates the enantiomerization rate. Tertiary amides of1-naphthoic acid are not isolable at room temperature(enantiomerization barriers DG‡ lower than 20 kcal/mol),whereas tertiary amides of 2-methyl, 2-ethoxy and 8-methyl substituted 1-naphthoic acid are atropisomeric,i. e., physically separable at room temperature (DG‡

between 21 and 24.8 kcal/mol).

The individual enantiomers of atropisomeric amides 1–5 were isolated at room temperature by HPLC on theWhelk-O1 CSP and subjected to off-column thermal race-

mization at temperatures between 318 and 348 K [14,15]. The enantiomerization rate constants obtained fromthese experiments were used as fixed input parametersfor the computer simulations of exchange broadenedchromatograms of 1–5, collected on the Whelk-O1 CSP,using different combinations of eluent flow rate andtemperature that yielded a visible plateau between theresolved peaks. Comparison between off-column andchromatographically derived rate constants showed thatthe enantiomerization process is always faster in freesolution (km) compared to the adsorbed state (kst), with dif-ferences between the energy barriers in the two phasesin the 0.3–1.3 kcal/mol range. The inhibitory effects ofthe stationary phase were found to be small for the firsteluted enantiomer and more pronounced for the secondeluted one. This finding is consistent with the assump-


Figure 6. Structures of stereola-bile compounds investigated byDHPLC.


tion that, in the solute-selector complex, the transitionstate for the enantiomerization (1808 rotation of the car-boxamide fragment around the CAr-CO bond) cannot bereached unless the hydrogen bond between the carbonyloxygen of the analyte and the selector amidic NH is bro-ken. This interaction, detected both in solution and inthe solid state in similar systems, is more pronouncedfor the second eluted enantiomer and leads to a higherenergy barrier for the conversion of the second into thefirst eluted enantiomer (Fig. 7). A similar study was per-formed on the chiral N-benzyl-1,3,2-benzodithiazole 1-oxide 6, whose stereolabile enantiomers were resolvedby HPLC on the Whelk-O1 CSPs [16, 17]. Comparison ofthe activation barriers for the R/S inversion measured bythermal racemization of the isolated enantiomers in freesolution and by DHPLC (i. e., in the presence of the station-ary phase) revealed only a small, deactivating effect ofthe solid sorbent on the enantiomerization process(DDG‡ = 0.5 kcal/mol at 308.9 K, DDG‡ = 0.8 kcal/mol at320.7 K). In these studies, on-column and off-columnexperiments were performed using the same solventsand exploring the same temperature ranges. Relatedstudies on axially chiral compounds, whose R/S intercon-versions were observed by concomitant DHPLC and off-column thermal racemization or DNMR experiments,revealed that free solution processes are only marginallyaffected by the HPLC stationary phases. In addition, sol-vent and temperature effects were unimportant, asexpected for intermolecular motions occurring throughsingle bond rotation processes. The biphenyl derivatives7 and 8 were resolved by HPLC on Chiralpak AD columnsand studied by variable temperature HPLC. The enantio-mers of 7 gave two well-resolved peaks, with an inter-mediate plateau-like region clearly visible when the col-

umn temperature was in the 15 –258C range [18]. Thedynamic chromatographic profiles changed in a flowand temperature dependent manner and such dynamicdeformations of the experimental chromatograms wereexploited to extract the pertinent kinetic data. Computersimulations of exchange-deformed elution profilesobtained at 258C and using eluent flow rates of 1.0 and0.5 mL/min gave the apparent rate constants for the on-column R/S interconversion an the associated barriersDG‡

12 = 21.8 kcal/mol and DG‡21 = 21.9 kcal/mol at 248C,

using hexane/2-propanol (95 :5) + 0.1% CF3COOH as elu-ent. Off-column racemization of the individual enantio-mers of 7 in methanol/water (5 :1) + 0.1% CF3COOH wasstudied by circular dichroism spectroscopy. Followingthe decay of the circular dichroism signal as a functionof time, the energy barrier to enantiomerization atT = 98C was found DG‡ = 21.8 kcal/mol, indicating thattemperature, solvent and the HPLC stationary phasehave little effect on this enantiomerization. Similarresults were obtained for the biphenyl derivative 8, forwhich the DHPLC experiment [eluent hexane/isopropa-nol (80 :20) at T = –58C] gave an averaged energy barrierDG‡ = 19.7 kcal/mol, close to the NMR (micellar aqueousphase, T = 908C) value of DG‡ = 18.3 kcal/mol [19].

The enantiomers of the O-alkyl aryl oxime 9 wereresolved by HPLC on Chiralcel OD using hexane/isopropa-nol (98 :2) as eluent [20]. The DHPLC experiment carriedout in the temperature range between –40 and –708Cgave the averaged energy barrier of 15.0 kcal/mol, thatcorrelates well with the value of 15.9 kcal/mol deter-mined by DNMR in toluene-d8. The non-symmetrical salo-phen ligand 10 enantiomerizes through a two-step pro-cess in which the CAr-N bonds rotate sequentially, withthe rate determining step represented by the slower rota-tion around the more hindered bond (indicated by theasterisk in Fig. 6): in this case DHPLC carried out on Chir-alcel OD using hexane/isopropanol/methanol as eluentand DNMR in DMSO-d6 gave the energy barriersDG‡ = 21.8 kcal/mol and 21.4 kcal/mol, respectively [21].

Oxazepam 11 is a well-studied chiral compound, whoseenantiomers easily interconvert near room temperaturein a range of solvents. Free solution enantiomerizationbarriers are in the 21.6–22.6 kcal/mol range, while bar-riers extracted by DHPLC are 21.6–21.7 kcal/mol (onNucleodex-b-PM, permethylated b-CD chemically bondedto silica, using a 6 :4 mixture of phosphate buffer andmethanol as eluent, T range 20–608C) [22] and between20.7 and 22.1 kcal/mol (on Chirobiotic R, Chirobiotic T,Whelk-O1 CSPs under HPLC or SFC conditions at 458C)[10]. It has to be pointed out that enantiomerization ofoxazepam 11 (and of related 3-hydroxy-1,4-benzodiaze-pin-2-ones) does not occur through simple bond rota-tions, but presumably proceeds via protonation of the 3-hydroxy-group and elimination of water with generation


Figure 7. Stationary phase effects observed in the DHPLCexperiments on amides 1–5. For each compound verticalbars are (white to black): free energy of activation in themobile phase, forward (conversion from first to secondeluted) and backward (conversion from second to firsteluted) apparent free energies of activation, forward andbackward free energies of activation in the stationary phase.


of an intermediate carbocation that undergoes keto-enoltutomerization. Accordingly, DHPLC is able to correctlymodel also on-column interconversions that requirereversible bond breaking and formation, as well as com-plex processes involving charged intermediate species.

The interconversion between syn and anti rotamers of 12(diastereoisomers, the anti rotamer is chiral) is fast atambient temperature and the energy barrier to rotationabout the naphthylpyridyl axis was determined byDNMR (DG‡ = 17.5 kcal/mol at 40.38C in CDCl3). The syn/anti isomers were separated by HPLC on an achiral cyanocolumn [hexanes/EtOH (3 :2) as eluent] in the tempera-ture range between –65.0 and –438C and the corre-sponding dynamic experiments gave the averagedenergy barrier in the 16.2–16.4 kcal/mol range. A strictlylinear Eyring plot was obtained by combining the DHPLCand DNMR experiments that spanned a temperaturerange of about 1008C, demonstrating the two methodsare complementary and can be used in concert to extractstructural and dynamic information on enantiomer anddiastereomer interconversions [23].

4 DHPLC on CSPs at low temperatureDHPLC has found wide applications in the study of slowmolecular motions with energy barriers up to 25 kcal/mol, barriers that are beyond the usual range of DNMR.Fast interconversion processes can be studied by DHPLCas well, but the typical characteristic time of the tech-nique (minutes) poses a lower limit for the energy barrierof interconversion at around 14 kcal/mol. Such energybarrier translates into half-life times for the individualenantiomers in the millisecond range at temperaturesclose to 258C. To bring these processes on the same timescale of the fastest HPLC separation (seconds), the columnhas to be cooled down to temperatures around –608C,whereas slower HPLC separations that are completewithin minutes require temperatures as low as –808C.Cryo-chromatography at temperatures ranging from–508 down to –808C presents some practical and instru-mental difficulties, mainly related to the increased visc-osity of the mobile phases. Higher viscosities in turnresult in a large pressure drop across the column, withlimitations in the range of available linear velocities ofthe eluent. The back pressure of a packed column can beestimated using Eq. 2, that combines the Darcy law andthe Kozeny-Carman equation for the specific column per-meability for the velocity of the eluent in a laminar flowsystem (Eq. 2) [24], where the pressure drop (DP) is propor-tional to the length of the column (L), the eluent viscosity(g), the column linear velocity (u) and the flow resistancefactor (U), and inversely proportional to the square of theparticle diameter (dp)

DP = L g u (U/dp2) (2)

For typical normal-phase HPLC solvents, a temperaturejump from 258C to –808C results in a three- to fourfoldincrease in both viscosity and pressure drop for a givencombination of column format and linear velocity of theeluent.

However, decreasing the column temperature has addi-tional ramifications that include changes in retention,selectivity, and efficiency. As a general rule, retentionand enantioselectivity increase as the column tempera-ture is lowered. Enthalpic and entropic contribution toretention and selectivity are described using expressions(Eq. 3) and (Eq. 4), respectively, where R is the gas con-stant, T the absolute temperature of the column, b is thecolumn phase ratio, a the enantioselectivity, DH and DSare the enthalpy and entropy changes for the solute-selector interaction and dDH and dDS are the differencesin enthalpy and entropy changes between the two enan-tiomers.

lnk9 = –DH/RT + DS/R + lnb (3)

lna = –dDH/RT + dDS/R (4)

The interaction of enantiomeric solutes with the station-ary phase usually features negative DH and DS, i. e., theprocess is exothermic and accompanied by entropy loss,as expected for a simple bimolecular association process;under these conditions, retention decreases with increas-ing temperature. The linear inverse relationship betweenlna and temperature (Eq. 4) is explained by entropyeffects: The diastereomeric solute-selector combinationhaving the greatest degree of simultaneous bondinggives the more stable (negative dDH term) complex but isalso expected to lose the most degrees of freedom (posi-tive dDS). Hence, temperature reduction decreases theimportance of the adverse entropy term and increasesenantioselectivity [25].

At sub-ambient temperatures column efficiencydegrades rapidly, mainly because adsorption/desorptionevents, mass transport between phases and diffusioninside the stationary phase are slowed down. Althoughmass transfer problems erode column efficiency at lowtemperatures, the increase in the magnitude of a morethan offsets the loss in column performance.

When sufficient retention is obtained at a given low tem-perature (i. e., temperature and eluent composition arefixed parameters), analysis time can be reduced by con-trolling the flow rate and/or the column format. Avail-able column formats range from standard 25-cm lengthpacked with 5-lm particles (25,5 format) to short 5-cmcolumns packed with 3-lm particles (5,3 format).

Several literature examples show that in practice cryo-HPLC with analysis times in the range of few minutes is



conveniently carried out using either classical 25,5, 10,5or 5,3 column formats.

One example of enantioselective HPLC at extremely lowtemperature is represented by the separation of therapidly interconverting enantiomers of naphthyl sulfone13. The experimental setup consisted of a 25064.6 col-umn packed with 5-lm Whelk-O1 CSP, a mobile phasebased on methylene chloride/methanol (98 :2) deliveredat 2.0 mL/min, and a 1 m long connecting tube used asmobile phase precooler [26]. Room temperature HPLC of13 gave a single sharp peak that remained sharp down to–608C, indicating fast interconversion on the time scaleof the separation (about 2 min). Further cooling of thecolumn at 208 K resulted in peak broadening and split-ting. At –708C, the two peaks were separated but on-col-umn enantiomerization was still occurring at this tem-perature. On the other hand, the dynamic process wasslow on the HPLC time scale at –808C, where the twopeaks are baseline resolved. A similar dynamic behaviorwas observed for the corresponding 1,5-disulfone, wherethe two stereogenic Ar-SO2 axes generate a meso and twoenantiomerically related chiral conformers. Interconver-sion between the three species was frozen on the HPLCtime scale at –808C, giving two equally intense peaks cor-responding to the chiral species, along with a major peakcorresponding to the meso form [27].

An additional example of cryogenic DHPLC of fast inter-converting species is given by the secondary phosphineoxide 14 [28]. The two stereogenic elements present in 14(the configurationally stable P atom and the stereolabileP-Ar axis) generate two configurational enantiomers,each of which can generate two distinct diastereomersthat rapidly interconvert at room temperature by rota-tion around the P-Ar axis.

The residual enantiomers of 14 at room temperaturewere resolved by HPLC on 1060.4-cm column packedwith 5-lm (R,R)-DACH DNB CSP. While room temperatureHPLC separation of the stereostable enantiomers of 14was straightforward, the separation of the rotamers ofeach enantiomer was expected to be quite difficult, giventhe low interconversion energy barrier DG‡ = 14.8 kcal/mol measured by DNMR. The use of a short column hadin this case a positive effect as it shortened the analysistime and reduced the operating backpressure due toincreased eluent viscosity at the required low tempera-ture. Indeed, as the column temperature was lowered toabout –508C the two peaks showed considerable broad-ening due to on-column exchange. An interconvertingregion preceding the peaks appeared between –608Cand –708C, consistent with a reduced rotamer intercon-version rate on the separation time scale. Further coolingof the column to –838C prevented on-column exchange,and allowed a complete separation of the four species,

which no longer interconverted during their passagethrough the column, as evident from the absence of anyplateau between the resharpened peaks (Fig. 8). The samelow-temperature experiment performed on the indivi-dual enantiomers (isolated by HPLC at ambient tempera-ture) yielded two peaks for each enantiomer, correspond-ing to the diastereomeric rotamers.

Computer simulations of the chromatographic profilescarried out on the individual enantiomers of 14 yieldedthe isomerization barriers DG‡ = 14.8 kcal/mol for theenantiomer eluted first at room temperature andDG‡ = 15.1 kcal/mol for the remaining species. In the firstcase the value was identical to that obtained by DNMR,whereas in the second case the value was higher due to asmall deactivating effect of the stationary phase. Therelative amounts of the two diastereomeric conformersand the energy barriers separating them, as measured byDNMR and DHPLC methods, are in good agreement, inspite of the different temperature ranges at which theexperiments have been performed.

A short column packed with 3.5 lm silica-bonded chiralphase was used for the resolution of the stereolabileenantiomers of tertiary amide 15, separation thatrequired a column temperature as low as –708C [29]. Asthe column temperature was raised in 58C increments,the two chromatographic peaks gradually broadenedand a plateau appeared between the resolved peaks, indi-cating accelerated on-column interconversion of the twoenantiomers. At –408C a complete peak coalescence dueto fast exchange was observed. Computer simulation ofthe exchange-broadened elution profiles at –708C (Fig. 9,bottom traces), gave the averaged apparent rate constantfor the enantiomer interconversion (C-CO bond rotation)and the corresponding energy barrier DG‡ = 14.3 kcal/


Figure 8. HPLC of secondary phosphine oxide 14 on a10064-mm column packed with 5-lm DACH-DNB CSP.Eluent: CH2Cl2/MeOH (99 :1), flow rate 1.5 mL/min. Tem-peratures 258C (top) and –838C (bottom).


mol, in close agreement with the DNMR value of13.5 kcal/mol extrapolated at the same temperature.

5 ConclusionsEnantioselective DHPLC is now a well-established tech-nique for the investigation of the stereochemical stabi-lity of chiral compounds. The DHPLC experimentcoupled with computer simulation of the experimentalprofiles can yield valuable information on interconvert-ing species (rate constants for the dynamic process, con-former distribution for non-enantiomeric interconvert-ing species, chiroptical data with circular dichroism orpolarimetric detectors) and requires only small amountsof racemic material. Prerequisite for a successful experi-ment is the separability of the species on proper station-ary phases, with the separation rate matching the rate ofthe on-column interconversion. The technique is suitedto study on-column interconversions that exchange ana-lyte stereochemistry and that occurs either through sin-gle bond rotations or by bond breaking-formation pro-cesses [30–36].

The experimental parameters that can be varied to reachthis situation are the column temperature, the eluentflow rate and the column format. Column temperaturehas the greatest effect on the relative rates of the two pro-cesses (separation and interconversion) as it can be setfrom 808C down to –808C: the molecular internalmotions that can be studied in this range of tempera-tures have energy barriers spanning from 14 to 25 kcal/mol. For low temperature runs, a solvent of low viscosityshould be chosen to avoid excessive deterioration of thecolumn kinetic performances and to reduce operational

pressure. The eluent flow rate can also be adjusted tofine-tune the separation and interconversion rates. Thelinear velocity of the eluent can be varied by approxi-mately an order of magnitude, so the impact on the sep-aration rate is much smaller than the effect of tempera-ture on the interconversion rate. Enantioseparations ofstereolabile compounds can be carried out on the timescale of few minutes at cryogenic temperatures usingeither standard columns (25,5 format) or short columnspacked with small particle CSPs (5,3 format) used in con-junction with high eluent flow rates.

The ease with which key experimental parameters arechanged and controlled (temperature, eluents, eluentflow rate) makes DHPLC the ideal approach to extractkinetic data for solute isomerizations that occur withenergy barriers between 14 and 25 kcal/mol. Consideringthe time scale of the experiments, DHPLC is complemen-tary to DNMR in that it can reveal the existence ofexchange phenomena, and yield the correspondingexchange rates for molecular systems with slow internalmotions. In addition, the physical separation of the inter-converting species can be fruitfully exploited to extendtheir characterization by chiroptical methods. Yet, whenthe DHPLC and DNMR techniques can be used in combi-nation, the quality and quantity of information achiev-able on the dynamic system are greatly extended com-pared to each of the two techniques alone.

Support of this research by a grant from MIUR-PRIN (contract no.2005037725_001) is gratefully acknowledged.

6 References[1] Subramanian, G. (Ed.), A Practical Approach To Chiral Sep-

arations By Liquid Chromatography, Wiley-VCH, Weinheim1994.

[2] Ahuja, S. (Ed.), Chiral Separations – Applications and Tech-nology, American Chemical Society, Washington, DC1997.

[3] Allenmark, S., Schurig, V., J. Mater. Chem. 1997, 7, 1955 –1963.

[4] Armstrong, D. W., LC-GC Int. 1998, 4, 22–31.

[5] Gasparrini, F., Misiti, D., Villani, C., J. Chromatogr A 2001,906, 35–50.

[6] Mannschreck, A., Zinner, H., Pustet, N., Chimia 1989, 43,165 –166.

[7] Veciana, J., Crespo, M. I., Angew. Chem. Int. Ed. 1991, 30,74–77.

[8] Jung, M., Schurig, V., J. Am. Chem. Soc. 1992, 114, 529 –534.

[9] Trapp, O., Schoetz, G., Schurig, V., Chirality 2001, 13,403 –414.

[10] Krupcik, J., Oswald, P., Majek, P., Sandra, P., Armstrong,D. W., J. Chromatogr. A 2003, 1000, 779 –800.

[11] Wolf, C., Chem. Soc. Rev. 2005, 34, 595 –608.


Figure 9. DHPLC traces of tertiary amide 15 on a 5064mm column packed with 3.5-lm DACH-DNB CSP. Eluent:CH2Cl2/MeOH (98 :2), flow rate 2.0 mL/min. Simulated ploton the right at –708C.

https://www.researchgate.net/publication/285119458_The_significance_of_the_HPLC_time_scale_An_example_of_interconvertible_enantiomers?el=1_x_8&enrichId=rgreq-9db1d044-dc72-4fa9-81af-b45b404e6e4a&enrichSource=Y292ZXJQYWdlOzY4OTE5MDY7QVM6MTAyMjkxODYwMzYxMjI2QDE0MDEzOTk2ODg5MDI=

https://www.researchgate.net/publication/285119458_The_significance_of_the_HPLC_time_scale_An_example_of_interconvertible_enantiomers?el=1_x_8&enrichId=rgreq-9db1d044-dc72-4fa9-81af-b45b404e6e4a&enrichSource=Y292ZXJQYWdlOzY4OTE5MDY7QVM6MTAyMjkxODYwMzYxMjI2QDE0MDEzOTk2ODg5MDI=


[12] Trapp, O., Anal. Chem. 2006, 78, 189 –198.

[13] Jacobson, J., Melander, W., Vaisnys, G., Horvath, C., J.Phys. Chem. 1984, 88, 4536–4542.

[14] Gasparrini, F., Misiti, D., Pierini, M., Villani, C., Tetrahe-dron: Asymmetry 1997, 8, 2069.

[15] Gasparrini, F., D’Acquarica, I., Pierini, M., Villani, C., J.Sep. Sci. 2001, 24, 941 –946.

[16] Oxelbark, J., Allenmark, S., J. Org. Chem. 1999, 64, 1483 –1486.

[17] Oxelbark, J., Allenmark, S., J. Chem. Soc. Perkin Trans.1999, 2, 1587–1589.

[18] Ceccacci, F., Mancini, G., Mencarelli, P., Villani, C., Tetra-hedron: Asymmetry 2003, 14, 3117–3122.

[19] Andreani, R., Bombelli, C., Borocci, S., Lah, J., et al., Tetra-hedron: Asymmetry 2004, 15, 987 – 994.

[20] Gasparrini, F., Grilli, S., Leardini, R., Lunazzi, L., et al., J.Org. Chem. 2002, 67, 3089 –3095.

[21] Dalla Cort, A., Gasparrini, F., Lunazzi, L., Mandolini, L.,et al., J. Org. Chem. 2005, 70, 8877 –8883.

[22] Trapp, O., Trapp, G., Schurig, V., J. Biochem. Biophys. Meth-ods 2002, 54, 301 –313.

[23] Wolf, C., Tumambac, G. E., J. Phys. Chem. A 2003, 107,815 –817.

[24] Thomson, J. D., Carr, P. W., Anal. Chem. 2002, 74, 4150 –4159.

[25] Pirkle, W. H., Schreiner, J. L., J. Org. Chem. 1981, 46,4988–4991.

[26] Villani, C., Pirkle, W. H., Tetrahedron: Asymmetry 1995, 6,27–30.

[27] Casarini, D., Lunazzi, L., Alcaro, S., Gasparrini, F., Vil-lani, C., J. Org. Chem. 1995, 60, 5515–5519.

[28] Gasparrini, F., Lunazzi, L., Mazzanti, A., Pierini, M., etal., J. Am. Chem. Soc. 2000, 122, 4776–47800.

[29] Borsato, G., Della Negra, F., Gasparrini, F., Misiti, D., etal., J. Org. Chem. 2004, 69, 5785 –5788.

[30] Stephan, B., Zinner, H., Kastner, F., Mannschreck, A., Chi-mia 1990, 10, 336 –338.

[31] Cabrera, K., Lubda, D., J. Chromatogr. A 1994, 666, 433 –438.

[32] Cabrera, K., Jung, M., Fluck, M., Schurig, V., J. Chroma-togr. A 1996, 731, 315 –321.

[33] Friary, R. J., Spangler, M., Osterman, R., Schulman, L.,Schwerdt, J. H., Chirality 1996, 8, 364 –371.

[34] Lorenz, K., Yashima, E., Okamoto, Y., Angew. Chem. Int. Ed.1998, 37, 1922–1925.

[35] Tobler, E., L�mmerhofer, M., Mancini, G., Lindner, W.,Chirality 2001, 13, 641 –647.

[36] Trapp, O., Caccamese, S., Schmidt, C., B�hmer, V.,Schurig, V., Tetrahedron: Asymmetry 2001, 12, 1395 –1398.


Dynamic HPLC on chiral stationary phases: A powerful tool for the investigation of stereomutation...

Documents

Transcript of Dynamic HPLC on chiral stationary phases: A powerful tool for the investigation of stereomutation...