Enantioseparation of α-Hydroxyallylphosphonates and Phosphonoallylic Carbonate Derivatives on Chiral Stationary Phases Using Sequential UV, Polarimetric, and Refractive Index Detection

Article abstract of DOI:10.1002/chir.22627

Chromatographic separation of the enantiomers of parent compounds dimethyl α-hydroxyallyl phosphonate 1a and 1-(dimethoxyphosphoryl) allyl methyl carbonate 1b was demonstrated by high-performance liquid chromatography (HPLC) using Chiralpak AS-H and ad-H chiral stationary phases (CSP), respectively, using a combination of UV, polarimetric, and refractive index detectors. A comparison was made of the separation efficiency and elution order of enantiomeric α-hydroxyallyl phosphonates and their carbonate derivatives on commercially available polysaccharide AS, ad, OD, IC-3, and Whelk-O 1 CSPs. In general, the α-hydroxyallyl phosphonates were resolved on the AS-H CSP, whereas the carbonate derivatives 1b and 2b were preferentially resolved on the ad-H CSP. The impact of aryl substitution on the resolution of analytes 1d, 1e, 1f and 2, 3, 4, 5, 6, 7, 8 was evaluated. Thermodynamic parameters determined for enantioselective adsorption hydroxyphosphonates 1a and 4 on the AS-H CSP and carbonate 1b on the ad-H CSP demonstrated enthalpic control for separation of the enantiomers. Chirality 28:656–662, 2016.

Full text of DOI:10.1002/chir.22627

CHIRALITY (2016)
Enantioseparation of α-Hydroxyallylphosphonates and
Phosphonoallylic Carbonate Derivatives on Chiral Stationary Phases
Using Sequential UV, Polarimetric, and Refractive Index Detection
*
BRUCE C. HAMPER, MICHAEL P. MANNINO, MELISSA E. MUELLER, LIAM T. HARRISON AND CHRISTOPHER D. SPILLING
Department of Chemistry and Biochemistry, University of Missouri-St. Louis, St. Louis, Missouri, USA
ABSTRACT
Chromatographic separation of the enantiomers of parent compounds dimethyl
α-hydroxyallyl phosphonate 1a and 1-(dimethoxyphosphoryl) allyl methyl carbonate 1b was
demonstrated by high-performance liquid chromatography (HPLC) using Chiralpak AS-H and
AD-H chiral stationary phases (CSP), respectively, using a combination of UV, polarimetric,
and refractive index detectors. A comparison was made of the separation efficiency and elution
order of enantiomeric α-hydroxyallyl phosphonates and their carbonate derivatives on
commercially available polysaccharide AS, AD, OD, IC-3, and Whelk-O 1 CSPs. In general, the
α-hydroxyallyl phosphonates were resolved on the AS-H CSP, whereas the carbonate derivatives
1b and 2b were preferentially resolved on the AD-H CSP. The impact of aryl substitution on the
resolution of analytes 1d–f and 2–8 was evaluated. Thermodynamic parameters determined for
enantioselective adsorption hydroxyphosphonates 1a and 4 on the AS-H CSP and carbonate 1b
on the AD-H CSP demonstrated enthalpic control for separation of the enantiomers. Chirality
00:000–000, 2016. © 2016 Wiley Periodicals, Inc.
KEY WORDS: α-hydroxy phosphonates; chiral stationary phases; enantiomer separation; HPLC;
multiple detectors; phosphonoallylic carbonates; polarimetric detection
Direct chromatographic analysis of the synthetically
valuable enantiomers of dimethyl α-hydroxyallyl phosphonate
1a and 1-(dimethoxyphosphoryl) allyl methyl carbonate 1b
has been a significant challenge (Fig. 1). These particular
compounds do not contain strong UV-absorbing
chromophores and coupled with limited retention on a num-
ber of chiral stationary phases (CSPs) under both reverse
phase and normal phase conditions pose difficulties for chro-
matographic analysis. Recently, we reported in collaboration
with the Welch-Regalado group at Merck the moderate reso-
lution of phosphono allylic carbonate and acetate derivatives
1b and 1c, respectively, using technologies developed in
the pharmaceutical industry.1 Supercritical fluid chromatog-
raphy (SFC) screening using various Chiralcel and Chiralpak
CSPs coupled with mass spectrometry (MS) detection
afforded detection and resolution of the enantiomers of
carbonate 1b and acetate 1c. Optimized separation used a
Chiralpak IC column with isocratic elution at 4% methanol in
CO2 and detection of the 2 M + 1 extracted ion. The resolu-
tion of the parent compound 1a was not achieved using this
approach. While this represents the first observed analytical
separation of the enantiomers of 1b and 1c, it does require
specialized equipment that is not readily available in
academic organic synthesis laboratories.
We are interested in evaluating the enantiomeric purity of
α-hydroxyallylphosphonates 1a and 1b due to their utility
as valuable chiral synthetic intermediates.2 They have been
employed for the preparation of a wide variety of biologically
active compounds including both organophosphorus and, by
chirality transfer, nonphosphorus-containing target mole-
cules.3 The ability to selectively prepare either enantiomer
of 1b has led to the synthesis of nematocidal oxylipids4 and
antimicrobial cyclophostins.5 Palladium coupled cross-
metathesis reactions have been employed for transfer of
chirality of α-hydroxyallylphosphonates to the gamma
© 2016 Wiley Periodicals, Inc.
position of the carbon chain,6,7 resulting in enantioselective
synthesis of cyclopentanone substituted phosphonate,8 cyclic
ethers,9 and (S)-tumerone.10 Enantiomerically pure
dimethoxyphosphorylallyl methyl carbonates 1b can be pre-
pared on a multigram scale from readily available precursors
by coupling an asymmetric Pudovik reaction with enzyme cat-
alyzed kinetic resolution.2 This three-step process initially
provides phosphonate 1a in 60–70% enantiomeric excess
(ee), which after conversion to a carbonate derivative can be
kinetically resolved to yield carbonate 1b in high ee. The de-
gree of enantiomeric purity is dependent on the specific con-
ditions for the kinetic resolution and can be facilitated by
direct monitoring of the enantiomeric purity of the product
1b in the reaction mixture.
Determination of the enantiomeric purity of aryl
substituted α-hydroxyphosphonates such as 2a and 3 by
high-performance liquid chromatography (HPLC) using ei-
ther Chiralpak AS or Whelk-O 1 CSPs was first reported in
1995 by Spilling and colleagues.11,12 In these initial studies,
it was shown that the aromatic (R = aryl) phosphonates 2a
and 3 were baseline resolved with the R enantiomer eluted
first on an AS CSP; however, resolution of the parent com-
pounds 1a–c were not addressed. Recently, the analytical res-
olution of reaction mixtures by HPLC using CSPs has proven
to be critical for determination of the effectiveness of catalysts
used for asymmetric hydrophosphonylation. Chiral aluminum
complexes have been reported as catalysts for highly
enantioselective hydrophosphonylation of aldehydes.13,14
*Correspondence to: Bruce C. Hamper, Department of Chemistry and Bio-
chemistry, University of Missouri-St. Louis, One University Blvd., St. Louis,
MO 63121. E-mail: hamperb@umsl.edu
Received for publication 18 June 2016; Accepted 11 July 2016
DOI: 10.1002/chir.22627
Published online in Wiley Online Library
(wileyonlinelibrary.com).

HAMPER ET AL.
Analytical normal phase chiral HPLC was performed using
a
ThermoSeparations P4000 pump or Rainin HPX, Rheodyne 7125 injector
equipped with a 20-uL sample loop and a set of sequential Waters (Mil-
ford, MA) UV 484 variable wavelength, Advanced Laser Polarimeter
(ALP, PDR-Chiral, LLC), and ERC-7512 (ERC, Huntsville, AL) refractive
index detectors. Data collection of the multiple detectors was achieved
using PeakSimple SRI Model 302 data collection system (SRI Instru-
ments, Menlo Park, CA) for simultaneous monitoring of all three data out-
puts. HPLC columns employed were 250 x 4.6 mm I.D., containing 5 μm
particle size chiral stationary phases Chiralpak AD-H, AS-H, OD, OC, OJ,
IC (Chiral Technologies, West Chester, PA) or (S,S)-Whelk-O 1 (Regis
Chemical, Morton Grove, IL). The IC-3 HPLC column was 100 x
4.6 mm I.D., containing 3 μm particle size CSP. Chromatographic runs
were performed at a flow rate of 1.0 mL/min and at a temperature of
25 °C, unless otherwise stated. Temperature studies were carried out
using a 100 x 4.6 mm I.D. column containing 3 μm particle size AS-H or
AD-H CSP equipped with a Cera Column Cooler 250 digital Peltier temper-
ature controller (Cobert Associates, St. Louis, MO). Resolution (Rs) and
theoretical plate count (N) for the enantiomers was determined by the
method of measurement of mid-height of the peaks.21
Fig. 1. Chemical structure of the chiral analytes 1–8.
Self-assembled bifunctional catalysts have also been shown to
afford exceptionally efficient chiral catalysts for synthesis of
2a and 3.15 Ooi and colleagues used an alternative approach
for chiral hydrophosphonylation using a phosphonium dialkyl
phosphite.16 Typically, the analytical details are relegated to
the supplementary material and the relative merits of the
chromatographic separations are not addressed.
(Dimethyloxyphosphoryl)allyl 3,5-dinitrobenzoate (1d)
To allylic hydroxyphosphonate 1a (0.17 g, 1.0 mmol) in 4 mL of CH2Cl2
was added trimethylamine (0.22 g, 2.1 mmol), DMAP (12 mg, 0.1 mmol),
and 3,5-dinitrobenzoyl chloride (0.23 g, 1.0 mmol). After stirring at room
temperature (rt) overnight, the mixture was quenched with 1 M KHSO4
and extracted three times with ethyl acetate. The combined extracts were
concentrated and purified by chromatography (10 g SiO2, 50% ethyl
acetate/hexanes) to afford 90 mg (25%) of a clear, colorless oil: HPLC (re-
verse phase) 2.63 min (>98% by UV 254 and ELS); TLC (50%
EtOAc/hexanes) Rf = 0.22; IR (ATR) 3096, 2957, 1736 (C = O), 1627,
1539, 1021; 1H NMR (CDCl3, 300 MHz) δ 3.85 (d, 3H, 3.2 Hz), 3.87 (d,
3H, 3.2 Hz), 5.50 (d, 1H, 9.5 Hz), 5.58 (dd, 1H, 19.8 Hz, 4.2 Hz), 5.99
(dd, 1H, 13.3 Hz, 8.0 Hz), 6.07 (m, 1H), 9.17 (s, 2H), 9.23 (s, 1H); 31P
NMR (CDCl3) δ 18.8; 13C NMR (CDCl3) δ 53.1, 71.4 (d, 170 Hz), 121.2
(d, 11.6 Hz), 123.0, 128.4 (d, 4 Hz), 129.7, 133.1, 149.0, 161.3 (d,
7.1 Hz); MS(FAB+) 361 (MH+, 66), 149 (100); HRMS (FAB, NaI, M+)
calcd for C12H13N2O9PNa: 383.0247; found: 383.0256.
In this report we describe the resolution of substituted α-
hydroxyallylphosphonates 1–8 and their derivatives on com-
mercially available chiral stationary phases using sequential
UV, polarimetric, and refractive index detectors. The analysis
was carried out using an HPLC with normal phase solvents
and allows direct determination of the enantiomeric purity
of synthetically valuable carbonate 1b from gram-scale
reaction mixtures.
A
comparison is made of the
enantioselectivity of various CSPs for separation of the target
analytes. A temperature study of the remarkably selective
separation of furan derivative 4 allowed determination of
the thermodynamic parameters leading to enantioselectivity
on the AS-H CSP.
(E)-dimethyl(3-(furan-2-yl)-1-hydroxyallyl)phosphonate 4
and enantiomerically enriched 4(S)
To a solution of trans-3-(2-furyl)acrolein (1.22 g, 10 mmol) in 35 mL of
anhydrous toluene cooled in a wet ice-acetone bath was added a catalytic
amount of quinine (42 mg, 0.13 mmol) and dimethyl phosphite (2.28 mL,
2.74 g, 25 mmol). After stirring for 72 h at rt., the mixture was concen-
trated, purified by chromatography (50 g SiO2, ethyl acetate), and concen-
trated in vacuo to give 2.0 g of a reddish oil. Enantiomeric composition of
the product was determined by 31P NMR in the presence of 5 equivalents
of quinine as a chiral solvating agent.22 31P NMR (CDCl3 + quinine) δ 23.7
(R enantiomer, 40%), 23.97 (S enantiomer, 60%). Repeated trituration with
CCl4 gave 1.08 g (46.5%) of a tan solid which was nearly racemic 4. Con-
centration of the mother liquors gave an orange solid which was 40%
enriched in the S enantiomer 4(S); mp 55–60 °C (dec). HPLC (reverse
phase) 1.64 min (>98% by UV 254 and ELS); TLC (EtOAc) Rf = 0.23; IR
(ATR) 3270 (br, OH), 2953, 2848, 1457, 1021; 1H NMR (CDCl3,
300 MHz) δ 3.82 (s, 6H), 4.71 (dd, 1H, 5.8 Hz, 1.5 Hz), 6.22–6.37 (m,
3H), 6.63 (d, 1H, 15.8 Hz), 7.34 (s, 1H); 31P NMR (CDCl3) δ 23.5; 13C
NMR (CDCl3) δ 53.7 (d, JCP = 7.5 Hz), 54.0 (d, JCP = 7.5 Hz),68.7 (d,
JCP = 162 Hz),108.8, 111.4, 120.4 (d, JCP = 15.8 Hz),122.2 (d, JCP = 4.5 Hz),
142.2 (d, JCP = 1.5 Hz), 152.2 (d, JCP = 4.5 Hz); HRMS (FAB, NaI, M+)
calcd for C9H13O5PNa: 255.0398; found: 255.0405.
MATERIALS AND METHODS
All reactions were carried out in oven-dried glassware under an atmo-
sphere of argon unless otherwise noted. Nuclear magnetic resonance
(NMR) spectra were recorded in CDCl3 at 300 or 500 (1H), 75 or 125
(
13C), and 121 (31P) MHz, respectively. 1H NMR spectra were referenced
to residual CHCl3 (7.27 ppm) and collected phosphorus-31 decoupled, 13
C
NMR spectra were referenced to the center line of CDCl3 (77.23 ppm),
and 31P NMR spectra were referenced to external 85% H3PO4 (0 ppm).
Reverse phase HPLC analysis was conducted using an Agilent (Palo Alto,
CA) 1100 system equipped with UV diode array and evaporative light
scattering (ELS) detectors. A Zorbax XBD-C18 column was employed
using a gradient of 95% H2O/5% CH3CN to 95% CH3CN over 6 min
followed by a hold of 95% CH3CN for 3 min. Mobile phase solvents were
prepared from HPLC-grade CH3CN and H2O containing 0.1% TFA.
General procedures for the preparation of racemic and enantiomerically
enriched dimethyl (1-hydroxyallyl)phosphonates 1a–c and (E)-dimethyl
(1-hydroxyl-3-phenylallyl)phosphonates 2a-c have been previously
reported.7,17–19 Compounds 3 were prepared by hydrophosphonylation
from the appropriate aldehyde and dimethyl phosphite and were obtained
from our in-house collection.20 Enantiomerically enriched samples of
allylic phosphonates 1b–f and cinnamylic phosphonates 4b–c were pre-
pared from enriched (R)-hydroxyphosphonates 1a and 2a, respectively.
The (R) enantiomers of 1a–f gave rise to a negative rotation, whereas
the (R) enantiomers of 2a–c resulted in a positive rotation at 670 nm
using the polarimetric ALP detector.
(E)-dimethyl(3-(4-chlorophenyl)-1-hydroxyallyl)
phosphonate 5
A stirred solution of 4-chlorocinnamaldehyde (0.5 g, 3.6 mmol) dis-
solved in 1 mL of CH2Cl2 was cooled in a wet ice-acetone bath and
treated sequentially with trimethylamine (0.25 ml, 3.4 mmol) and di-
methyl phosphite (0.35 ml, 3.63 mmol). After stirring for 3 days, the
Chirality DOI 10.1002/chir

α-HYDROXYALLYLPHOSPHONATES AND PHOSPHONOALLYLIC CARBONATE DERIVATIVES
mixture was concentrated and purified by chromatography (SiO2, ethyl
phosphonates 1–3 and derivatives, we found that the sign of
rotation from the ALP was identical to the reported sign of ro-
tation at the sodium D line (598 nm). The (R)-1a and its de-
rivatives (R)-1b-f show negative rotation, while the aromatic
substituted (R)-2a-c and (R)-4 exhibit a positive rotation at
both 598 and 670 nm.
acetate) to afford 0.88 g (88%) of a yellow crystalline solid; mp 94–95.5 °
C. TLC (ethyl acetate) Rf = 0.24; HPLC (reverse phase) 2.37 min (>98%
by UV 254 and ELS); IR(ATR) 3252, 2957, 2846, 1688, 1488, 1024; 1H
NMR (CDCl3, 300 MHz) δ 3.75 (s, 6H), 3.89 (bs, 1H), 4.65 (dd, 1H,
6.0 Hz, 1.3 Hz), 6.22 (dd, 1H, 15.9 Hz, 6.0 Hz), 6.68 (d, 1H, 15.9 Hz),
7.22 (m 4H); 31P NMR (CDCl3) δ 23.7; 13C NMR (CDCl3) δ 53.8 (d,
7.4 Hz), 54.0 (d, 7.1 Hz), 69.1 (d, 161 Hz), 124.3 (d, 4.3 Hz), 127.9 (d,
1.8 Hz), 128.8, 131.2 (d, 13.0 Hz), 133.7, 134.8 (d, 3.0 Hz); HRMS (FAB,
NaI, M+) calcd for C11H14O4PNa 299.0216; found: 299.0221.
Analysis of 1a using a Chiralpak AS-H CSP and sequential
detectors shows nearly baseline resolution and, as expected,
the R-(ꢀ) enantiomer elutes first (Fig. 2A). Determination of
the enantiomeric purity can be made using either the UV at
210 nm or polarimetric detectors. Significantly higher concen-
trations of the analytes were required for RI detection of 1a.
However, the presence of all three detection modes allows
for easy identification of the two enantiomers even in more
complex mixtures. Attempts to resolve carbonate 1b on the
AS-H or WhelkO-1 CSPs were unsuccessful. However, base-
line resolution and determination of the enantiomeric purity
of 1b was achieved using the Chiralpak AD-H CSP with 10%
isopropanol in hexanes as the mobile phase (Fig. 2B). For
carbonate 1b, no UV signal was observed. The (S)-(+)-1b
eluted first, as indicated by the ALP, and enantiomeric purity
can be easily determined using either the ALP or RI detec-
tors. It is interesting to note that the carbonate functionality
of 1b resulted in improved efficiency and column perfor-
mance for the AD-H CSP (first peak N = 1700) compared to
the resolution of hydroxyl bearing 1a on AS-H (first peak
N = 460). Since both columns were of similar dimensions
and particle size, this is likely due to the hydrogen bonding
character of the OH group of 1a resulting in nonideal chro-
matographic performance.
RESULTS AND DISCUSSION
Chromatographic Resolution of Parent Compounds 1a–c
Our initial goal was development of a suitable HPLC
method for evaluation of the enantiomeric purity of the parent
α-hydroxyallylphosphonate 1a and its carbonate derivative
1b for use in monitoring scale-up synthesis of these useful
chiral building blocks. Analysis of 1a and 1b by reverse-
phase HPLC equipped with UV and ELS detection showed
moderate response for 1a at 210 nm; however, the carbonate
derivate was not visible even at lower limits of UV detection.
Neither compound exhibited a response by ELS. Even in
the case of UV detection of 1a, minor impurities can domi-
nate the chromatogram and obscure measurements of purity.
To circumvent these difficulties, we assembled in sequence
multiwavelength UV, laser polarimeter (ALP), and refractive
index (RI) detectors. The ability to detect 1a-c by ALP and
RI was confirmed by analysis of racemic and enantiomerically
enriched samples using an achiral column. The laser polarim-
eter uses 670 nm plane polarized laser light source which can
have a different specific rotation and occasionally sign of rota-
tion compared to the standard sodium D line. For the known
The previously reported SFC-MS study1 showed separation
of the carbonate 1b (α = 1.14) and acetate 1c (α = 1.38) on IC
Fig. 2. Chromatograms showing the retention and separation of the enantiomers of compounds 1a and 1b. (A) Resolution of 1a using polarimetric and UV
(210 nm) detectors. Column: AS-H CSP; mobile phase: 10% isopropyl alcohol in hexanes. (B) Resolution of 1b using polarimetric and refractive index detectors.
Column: AD-H CSP; mobile phase: 10% isopropyl alcohol in hexanes.
Chirality DOI 10.1002/chir

HAMPER ET AL.
amylose derived CSP with baseline resolution, but no separa-
tion of the parent compound 1a (Table 1). Having shown the
utility of the sequential polarimetric and RI detectors for ana-
lyte 1b, we decided to reevaluate use of the IC column using
standard normal phase conditions. Surprisingly, when the IC
column was investigated using hexane-isopropanol as a mo-
bile phase, no resolution of the phosphonate derivatives 1b
or 1c was observed. After evaluation of eight different CSPs
(Chiralpak AS-H, AD-H, OD, OC, OJ, IC-3, IA-3, and Whelk-
O 1) we found that the AS-H CSP was best for resolution of
1a and AD-H for the carbonate 1b. Acetate 1c was not suc-
cessfully resolved on the Whelk-O 1 or AS-H CSP and showed
only moderate separation on AD-H. This was somewhat sur-
prising, since 1c had previously performed well in SFC mode
with an IC-3 stationary phase.
Benzoate
derivatives
of
simple
aliphatic
α-
hydroxyphosphonates have been previously prepared to en-
hance resolution by CSP-HPLC and facilitate detection by
UV.23 Since this approach had not been reported for allylic
phosphonates, we prepared three ester derivatives, DNB
(3,5-dinitrobenzoyl) 1d, benzoate 1e, and naphthoate 1f from
either racemic 1a or enantiomerically enriched 1a (Fig. 3).
The DNB 1d was successfully resolved on the amylose based
CSPs, AS-H, AD-H, and IC-3, with the largest separation
(α = 1.50; Rs = 4.9) observed on the AD-H CSP. Benzoate 1e
was not resolved on AD-H, but did give modest resolution on
the AS-H and IC-3 CSPs. Overall, the esters 1d–f were consis-
tently resolved on the AS-H CSP with the R enantiomer eluted
first. Resolution factors (Rs) were not optimized since we
were typically introducing a slight column overload to en-
hance polarimetric detection and injection concentration
could be reduced to improve measurement of enantiomeric
composition by UV. The IC-3 CSP consistently resolved
analytes 1d–f and, although the separation factors were
smaller, superior performance of the column allowed nearly
baseline resolutions. The elution order of the enantiomers,
however, was not consistent on the IC-3 CSP.
Behavior of the phenyl allyl phosphonates 2a-c was signif-
icantly different than the simple allyl phosphonates 1a–c.
While the Regis Whelk-O 1 CSP did not resolve allyl
phosphonates 1a-1e, all three compounds 2a–c were
baseline-resolved. Interestingly, the S(ꢀ) enantiomer of the
hydroxyphosphonate 2a eluted first, while the R(+) enantio-
mer of 2b and 2c eluted first on Whelk-O 1 CSP. The behav-
ior of the hydroxyl 2a and carbonate 2b on the amylose-
based CSPs reflected what we previously observed for
allylphosphonates 1a and 1b. The free hydroxyl group of
2a afforded the best resolution on AS-H (α = 2.40; Rs = 2.9),
while the carbonate derivative 2b gave the best results on
AD-H (α = 1.65; Rs = 2.7). For the purpose of comparison, we
also investigated archival compounds 3a-c from our in-house
collection. The naphthyl 3b and benzyl 3c afforded signifi-
cantly larger separation factors on the AS-H CSP (α = 3.60
and 4.44, respectively) compared to phenyl substituted hy-
droxyl phosphonate 3a. Enantiomers of compound 3b were
also well separated on the Whelk-O 1 and AD-H CSP.
To gain insight into the structural requirements for
enantioselectivity of substituted hydroxyallylphosphonates
on CSPs, we prepared a set of aryl substituted compounds
4–8 (Fig. 3). Treatment of dimethylphosphonate with
commercially available aryl acrylaldehydes in the presence
of triethylamine afforded the target hydroxyphosphonates
4–8. Enantiomerically enriched samples of the analytes were
Chirality DOI 10.1002/chir

α-HYDROXYALLYLPHOSPHONATES AND PHOSPHONOALLYLIC CARBONATE DERIVATIVES
Fig. 3. Preparation of substituted α-hydroxyallylphosphonates 1d–f and 4–8.
prepared by asymmetric synthesis using either quinine24,25
the parent compounds 1a–b, encouraged us to carry out a
temperature study to determine the thermodynamic parame-
ters associated with the enantioseparation (Table 3).26,27 Sep-
aration factors were obtained for seven individual
chromatographic runs collected in 10 °C increments from
2 °C to 55 °C using a jacketed Peltier temperature controller.
The measured α values of furan substituted 4 showed a
significant impact, with temperature with measurements at
2 °C affording a separation factor of 9.19 and decreasing to
2.90 at 55 °C. Differences in the enthalpy (Δ(ΔHo)) and
entropy (Δ(ΔSo)) of adsorption of the two enantiomers by
the AS-H CSP can be determined by the following equation:
or titanium isopropoxide/dimethyl tartrate19 as a catalyst.
Chromatographic resolution of the racemic compounds was
screened on several CSPs with the best results obtained on
Whelk-O 1, AS-H, and OD phases (Table 2). Separation
factors for all six analytes on the Whelk-O 1 were similar,
ranging from 1.32 for 4 (Ar = 2-furyl) to 1.51 for 6 (Ar = 4-
methylphenyl). In all cases, the (S)-(ꢀ) enantiomer eluted
first. Larger separation factors were observed on the AS-H
CSP with the furyl substituted 4 providing an α value of
6.54. Compound 4 has the most electron-deficient aryl group
of the set and analytes 5, 6, and 7, with electron-donating
groups exhibiting progressively lower separation factors. In
all cases, baseline resolution was observed on AS-H CSP,
with the (R)-(+) enantiomer eluted first. The six aromatic
analytes 3–8 showed less affinity for the OD CSP, requiring
only a 5% isopropanol in hexanes mobile phase to achieve
reasonable chromatographic retention. Resolution (Rs) was
lower for most analytes on OD CSP; however, reasonable α-
values were obtained for o-methyl substituted 7 (α = 1.72)
and diphenyl substituted 8 (α = 1.92).
lnα ¼ ꢀΔðΔHoÞ=RT þ ΔðΔSoÞ=R
The van’t Hoff plot of ln α versus the reciprocal of tempera-
ture in Kelvin gave a linear correlation coefficient greater
than 0.95 in all three cases. Enthalpic and entropic contribu-
tions to the enantioselective adsorption of the enantiomers
of
4 were calculated from the van’t Hoff plot as
The unusual selectivity observed for furanyl substituted
phosphonate 4 and the desire to optimize the separation of
ΔΔHo = ꢀ15.7 kJ/mol and ΔΔSo = ꢀ39.0 J/Kmol, respectively.
Since the contribution from the entropic term (ꢀTΔ(ΔSo)) is
TABLE 2. Chromatographic data for aryl substituted α-hydroxyallylphosphonates 2a and 4–8 on three CSPs.a
Compoundb
WhelkO-1
AS-H
OD
k’
α
Rs
k’
α
Rs
k’
α
Rs
2a
4
1.44(ꢀ)
2.31(ꢀ)
2.03(ꢀ)
4.08(ꢀ)
3.62(ꢀ)
11.59(ꢀ)
1.43
1.32
1.34
1.51
1.36
1.35
1.9
2.0
1.6
3.4
1.8
2.8
2.36(+)
4.79(+)
2.49(+)
5.10(+)
3.11(+)
5.67(+)
2.49
2.5
5.2
3.0
1.6
1.2
1.0
6.91(ꢀ)
5.90
6.10(ꢀ)
9.83(ꢀ)
2.77(+)c
8.39(ꢀ)d
1.30
NS
1.08
1.24
1.72
1.92
1.2
6.54
4.40
1.93
1.53
1.46
5
0.5
1.1
0.6
3.0
6
7
8
aFor key to chromatographic data, see Table 1.
bCompounds 2–8 exhibit (+) sign of rotation for R enantiomer. Mobile phase isopropanol/hexanes (I/H) composition for CSPs: Whelk-O 1, 20% I/H; AS-H, 25% I/H
and OD at 5% I/H, unless otherwise stated.
c10% I/H.
d20% I/H.
Chirality DOI 10.1002/chir

HAMPER ET AL.
TABLE 3. Thermodynamic values for chromatographic sepa-
rations of hydroxyphosphonate 1a, carbonate 1b, and furan 4a
or 2b. Although separation factors were lower, the Whelk-O
1 CSP consistently provided baseline resolution of the aryl
substituted α-hydroxylphosphonates 2–8. A temperature
study of the resolution of furanyl-α-hydroxy allylphosphonate
4 and parent compound 1a on the AS-H CSP allowed deter-
mination of enthalpic control for the resolution and retention
of the enantiomers.
Compound
Δ(ΔHo)
(kJ/mol)
Δ(ΔSo) (J/
mol K)
-TΔ(ΔSo)
(kJ/mol)
Δ(ΔGo)
(kJ/mol)
1a
1b
4
ꢀ3.5
ꢀ0.68
ꢀ15.7
ꢀ8.8
ꢀ0.07
ꢀ39.0
2.6
0.02
11.4
ꢀ0.9
ꢀ0.66
ꢀ4.3
aSee Figure 4 for experimental details.
ACKNOWLEDGMENTS
We thank the University of Missouri-St. Louis for support of
undergraduate research (M. Mannino and M. Mueller) and
support of the Students and Teachers as Researchers
(STARS) program (L. Harrison).
SUPPORTING INFORMATION
Additional supporting information may be found in the
online version of this article at the publisher’s web-site.
LITERATURE CITED
1. Regaldo EL, Kozlowski MC, Curto J, Ritter T, Campbell MG, Mazzotti AR,
Hamper BC, Spilling CD. Support of academic synthetic chemistry using
separation technologies from the pharmaceutical industry. Org Biomol
Chem 2014; 12: 2161–2173.
2. Roy S, Sutivisedsak N, Hamper BC, Lyss AM, Spilling CD. A practical and
scalable synthesis of (S)- and (R)-1-(dimethoxyphosphoryl)allyl methyl
carbonates. Synthesis 2015; 47: 3669–3672.
3. Spilling CD, Malla RK. Synthesis of non-racemic α-hydroxyl phosphonates
via asymmetric phospho-aldol reaction. In: Montchamp J-L, editor. Phos-
phorus chemistry II, synthetic methods. Springer International Publish-
ing: Switzerland; 2015. p 83–136.
4. Roy S, Spilling CD. An expeditious total synthesis of both diastereoiso-
meric lipid dihydroxytetrahydrofurans from Notheia anomala. Org Lett
2012; 14: 2230–2233.
5. Point V, Malla RK, Carriere F, Canaan S, Spilling CD. Enantioselective in-
hibition of microbial lipolytic enzymes by nonracemic monocyclic
enolphosphonate analogues of cyclophostin. J Med Chem 2013; 56:
4393–4401.
6. Cruz AD, He A, Thanavaro A, Yan B, Spilling CD, Rath NP. Allylic hy-
droxy phosphonates: versatile chiral building blocks. J Organomet Chem
2005; 690: 2577–2592.
7. He A, Yan B, Thanavaro A, Spilling CD, Rath NP. Synthesis of nonracemic
allylic hydroxy phosphonates via alkene cross metathesis. J Org Chem
2004; 69: 8643–8651.
Fig. 4. Temperature study van’t Hoff plot of ln α vs. 1/T for resolution of
1a, 1b, and 4 on CSPs. The CSP and mobile phase isopropyl alcohol in hex-
ane (I/H) composition: 1a) CSP: AS-H, 20% I/H; 1b) AD-H, 10% I/H; 4) AS-H,
10% I/H.
smaller in value than the enthalpic term (Δ(ΔHo)), the
enantioselective adsorption in enthalpy controlled and the α
value will decrease with increasing temperature. The reten-
tion times of both enantiomers also decreased with higher
temperatures, indicating enthalpic control for analyte reten-
tion. From a practical standpoint, optimal resolution as mea-
sured by Rs was observed at 20 °C. Lower temperatures
improved the separation factor, but also gave rise to band
broadening, which negatively impacted the peak shape and
resolution. Parent hydroxyphosphonate 1a exhibited a
similar ratio of the contributions of enthalpy and entropy to
the free energy value (Δ(ΔGo)); however, the overall values
were smaller corresponding to smaller separation factor α.
We investigated the resolution of carbonate 1b on the AD-H
CSP and found the relative contributions of Δ(ΔHo) and
Δ(ΔSo) were significantly different compared to resolutions
of 4 or 1a on the AS-H CSP. While the separation of 1b is
enthalpically controlled, the entropic contribution is much
smaller. As a result, the separation factors for 1b are not
greatly influenced by small changes in temperature.
8. Yan B, Spilling CD. Synthesis of cyclopentenones via intramolecular HWE
and the palladium-catalyzed reactions of allylic hydroxy phosphonate de-
rivatives. J Org Chem 2008; 73: 5385–5396.
9. He A, Sutivisedsak N, Spilling CD. Stereoselective synthesis of cyclic
ethers via the palladium-catalyzed intramolecular addition of alcohols to
phosphono allylic carbonates. Org Lett 2009; 11: 3124–3127.
10. Rowe BJ, Spilling CD. Stereospecific Pd(0)-catalyzed arylation of an allylic
hydroxy phosphonate derivative: formal synthesis of (S)-(+)-ar-turmerone.
J Org Chem 2003; 68: 9502–9505.
11. Kozlowski NP, Rath NP, Spilling CD. Determination of the enantiomeric
purity and absolute configuration of α-hydroxy phosphonates. Tetrahe-
dron 1995; 51: 6385–6396.
12. Welch CJ, Szczerba T, Perrin SR. Some recent high-performance liquid
chromatography separations of the enantiomers of pharmaceuticals and
CONCLUSION
We have demonstrated a useful and practical HPLC
method for analysis of enantiomeric purity of dimethyl α-
hydroxylallyl phosphonate 1a and dimethoxyphosphoryl allyl
methyl carbonate 1b using sequential refractive index, polar-
imetric, and UV detectors. These methods can be used for
the direct analysis of reaction mixtures for the preparation
of substituted hydroxyphosphonates as valuable synthetic
building blocks. Comparison of seven chiral stationary
phases demonstrated that the AS-H CSP provided superior
resolution of α-hydroxyphosphonates, while AD-H was pre-
ferred for the resolution of carbonate derivatives such as 1b
Chirality DOI 10.1002/chir
other compounds using the Whelk-O
1 chiral stationary phase. J
Chromatogr A 1997; 758: 93–98.
13. Suyama K, Sakai Y, Matsumnoto K, Saito B, Katsuki T. Highly
enantioselective hydrophosphonylation of aldehydes: base-enhanced
aluminum-salalen catalysis. Angew Chem Int Ed 2010; 49: 797–799.
14. Zhou X, Liu X, Yang X, Shang D, Xin J, Feng X. Highly enantioselective
hydrophosphonylation of aldehydes catalyzed by tridentate Schiff base
aluminum(III) complexes. Angew Chem Int Ed 2008; 47: 392–394.
15. Yang F, Zhao D, Lan J, Xi P, Yang L, Xiang S, You J. Self-assembled bifunc-
tional catalysis induced by metal coordination interactions: an exceptionally

α-HYDROXYALLYLPHOSPHONATES AND PHOSPHONOALLYLIC CARBONATE DERIVATIVES
efficient approach to enantioselective hydrophosphonylation. Angew Chem
Int Ed 2008; 47: 5646–5649.
16. Uraguchi D, Ito T, Ooi T. Generation of chiral phosphonium dialkyl phos-
22. Zymanczyk-Duda D, Skwarxzynski M, Lejczak B, Karfarski P. Accurate
assay of enantiopurity of 1-hydroxy- and 2-hydroxyalkylphosphonate es-
ters. Tetrahedron Asymmetry 1996; 7: 1277–1280.
phite as
a
highly reactive P-nucleophile: application to asymmetric
23. Wang D-Y, Hu X-P, Huang J-D, Deng J, Yu S-B, Duan Z-C, Xu X-F,
Zheng Z. Highly enantioselective synthesis of a-hydroxy phosphonic
acid derivatives by Rh-catalyzed asymmetric hydrogenation with
phosphine–phosphoramidite ligands. Angew Chem Int Ed 2007; 46:
7810–7813.
hydrophosphonylation of aldehydes. J Am Chem Soc 2009; 131: 3836–3837.
17. Roy S, Spilling CD. Synthesis of the C(18)-C(34) fragment of
amphidinolide C and the C(18)-C(29) fragment of amphidinolide F. Org
Lett 2010; 12: 5326–5329.
24. Wynberg H, Smaardijk AA. Asymmetric catalysis in carbon-phosphorus
18. Yan B, Spilling CD. Stereospecific Pd(O)-catalyzed malonate additions to
allylic hydroxy phosphonate derivatives: a formal synthesis of (ꢀ)-
enterolactone. J Org Chem 2004; 69: 2859–2862.
19. Rowe BJ, Spilling CD. The synthesis of 1-hydroxy phosphonates of high
enantiomeric excess using sequential asymmetric reactions: titanium
alkoxide-catalyzed P_C bond formation and kinetic resolution. Tetrahe-
dron Asymmetry 2001; 12: 1701–1708.
bond formation. Tetrahedron Lett 1983; 24: 5899–5900.
25. Smaardijk AA, Noorda S, van Bolhuis F, Wynberg H. The absolute config-
uration of α-hydroxyphosphonates. Tetrahedron Lett 1985; 26: 493–496.
26. Fornstedt T, Sanjonz P, Guiochon G. Thermodynamic study of an unusual
chiral separation. Propranolol enantiomers on an immobilized cellulase. J
Amer Chem Soc 1997; 119: 1254–1264.
20. Baraldi PG, Guarnieri M, Moroder F, Pollini GP, Simoni D. Synthesis of 1-
phthalimidoalkanephosphonates. Synthesis 1982; 653–654.
21. Snyder LJ, Kirkland JJ, Dolan JR. Introduction to modern liquid chroma-
27. ILisz I, Gecse Z, Lajko G, Forro E, Fulop F, Lindner W. High-performance
liquid chromatographic enantioseparation of cyclic β-amino acids on Zwit-
terionic chiral stationary phases based on cinchona alkaloids. Chirality
2015; 27: 563–570.
tography, 3rd ed. John Wiley & Sons: Hoboken, NJ; 2010. p 35–46.
Chirality DOI 10.1002/chir