Thermodynamic Study of Enantiomeric Separations Using Normal Phase Chiral HPLC Perkins, Gregory University of Redlands: Senior Thesis Spring 2009 Abstract: Normal Phase Chiral HPLC was used to determine if 3- hydroxypropranoic acid-1-phenyl-ethylester (3HP-1EE) and flavanone enantiomeric separations are enthalpy or entropy driven. A Chiralpak-IA polysaccharide column was used to separate the two compounds and temperature was controlled by a water bath. Both compounds were tested at various temperatures with multiple runs at each temperature. A Van’t Hoff plot graphing ln vs 1/T (1/K) was used to obtain Sand Hvalues for both compounds. 3HP-1EE had a H= -1430 J and a S= -3.73 J/K, showing that is primarily enthalpy driven. Flavanone had a H= -430 J and a S= -0.570 J/K, showing that it is more entropy driven. This means that larger compounds will be entropy 1
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Thermodynamic Study of Enantiomeric Separations Using Normal Phase Chiral HPLC
Perkins, Gregory
University of Redlands: Senior ThesisSpring 2009
Abstract:
Normal Phase Chiral HPLC was used to determine if 3-hydroxypropranoic acid-
1-phenyl-ethylester (3HP-1EE) and flavanone enantiomeric separations are enthalpy or
entropy driven. A Chiralpak-IA polysaccharide column was used to separate the two
compounds and temperature was controlled by a water bath. Both compounds were
tested at various temperatures with multiple runs at each temperature. A Van’t Hoff plot
graphing ln vs 1/T (1/K) was used to obtain S and H values for both
compounds. 3HP-1EE had a H = -1430 J and a S = -3.73 J/K, showing that is
primarily enthalpy driven. Flavanone had a H = -430 J and a S = -0.570 J/K,
showing that it is more entropy driven. This means that larger compounds will be
entropy driven and easily separated at higher temperatures while smaller compounds will
be enthalpy driven and easily separated at lower temperatures.
Introduction:
Separation of pharmaceutical enantiomers is of great importance to the
pharmaceutical industry. Individual enantiomers can exhibit different pharmaceutical,
toxicological, metabolic, and pharmacokinetic properties within chiral environments or
they may react at different rates. The U.S. Food and Drug Administration (FDA) prefer
single enantiomers whenever it is practical and economically feasible. This is because
different enantiomers of the same molecule can have different effects in the body.
1
Today, more than half of the top 500 drugs are single enantiomers. An example of a
single enantiomeric drug is Lipitor, which has generated profits in the multi-billion dollar
range since its release.
Figure 1. A molecular diagram of Lipitor with two R* chiral centers where the OH groups are shown to have stereochemistry. < www.drugs.com/ pro/lipitor.html>
There are several methods for evaluating enantiomeric drug purity. Non-
separating techniques include polarimetry, circular dichroism, Nuclear Magnetic
Resonance (NMR) using shift reagents, enzyme reactions and the use of enantioselective
sensors [1]. Polarimetry and circular dichroism both use polarized light to detect
different enantiomers [1]. NMR with shift reagents is similar to NMR with reagents
added to create chemical shifts that simplify the data [2]. This is usually used to separate
resonance structures of complex molecules that generally overlap. Enzyme reactions use
stereoselective enzymes that are more attracted to a certain enantiomer [1]. The last non-
separation technique uses enantioselective sensors, which are selective to different
enantiomers of a compound.
There are also separation techniques that can be used on samples containing
racemic mixtures. The first separation method is to react each chiral compound with a
2
chiral enantiomer to create two diastereoisomers so they can be separated on a C-18
chiral column using high performance liquid chromatography (HPLC). The second
method is capillary electrochromatography, or CEC, which adds an enantiomer to the
electrophoresis buffer so that one enantiomer in the sample will interact more with
enantiomers in the running buffer to create a differences in electrophoretic mobilities for
each enantiomer.
Another separation technique that will be the main focus of this paper is direct
separation of enantiomers by passing a mixture through a chiral stationary phase HPLC
column. Samples are injected and passed through the column where one of the chiral
centers, either R* or S*, will be more attracted to E* in the column, were E* is the
bonded enantiomeric center on the stationary phase. This difference in interaction will
create a separation between the enantiomers. Two of the most commonly used chiral
stationary phases for HPLC are the polysaccharide type and the Brush, or Pirkle, type.
Polysaccharides are either cellulose or amylose based chains (see Figure 2) which
are held together by hydrogen bonding to form sheets [3]. These sheets are held together
by van der Waals forces and the three-dimensional formation can sometimes create
inclusion complexes. These inclusion complexes are in cavities created during the
bonding of the chains that will interact with enantiomers. Each polysaccharide sub-unit
will contain several chiral centers with chiral activities, but the interactions with
enantiomers are weak, so derivitization by a reaction at the hydroxyl group to add
additional chiral centers can improve enantioselective properties [3]. Recognition of
chiral compounds by the polysaccharide column are based on 3-D shape, chirality, H-
bonding, dipole interactions, pi interactions, and van der Waals forces.
3
Figure 2. An example of amylose chain with likages. (John Innes Centre, UK) <www.jic.ac.uk/STAFF/ cliff-hedley/Starch.html>
Brush-type columns, or Pirkle columns, consist of low molecular weight selectors
bonded to a silica substrate. A man named William Pirkle first introduced these columns
in the 70’s [3]. The design allows three bonded separate chiral selector molecules to act
independently with each solute enantiomer and is compatible with many mobile phase
solvents. The selector molecules are made up of amino acid derivatives that consist of
aromatic rings, creating pi-pi interactions with the solute [3]. Other attractive interactions
that will stabilize the complex are H-bonding, dipole interactions, van der Waals forces,
steric repulsion, and Lewis acid or Lewis base interactions.
4
Figure 3. A chiral selector from a Nucleosil Chiral-2 brush type column. This column will experience Lewis acid/base interactions, H-bonding, dipole-dipole interactions and steric effects. (Macherey-Nagel) < http://www.mn-net.com/HPLCStart/SpecialHPLCphases/HPLCenantiomer/NUCLEOSILCHIRAL/tabid/6150/language/en-US/Default.aspx>
To separate these enantiomers, high performance liquid chromatography,
or HPLC, is used and it can be performed in two ways. The most common type of
mobile phase used for HPLC is reverse phase, where the solvent is more polar than the
stationary phase. Often to examine chiral stationary phases use the normal phase
technique where the mobile phase is less polar than the stationary phase.
The quality of a separation using HPLC can be determined using several different
equations. The number of theoretical plates tells the separation efficiency for the two
enantiomers. The number of theoretical plates can be determined using equation 1.
N = 16 (tr / w)2 (1)
For this equation, N is the number of theoretical plates, tr is the retention time and
w the baseline width of the peak. The capacity factor (equation 2) is a measure of how
long the enantiomer is retained by the column. A larger capacity factor means the
enantiomer is retained longer.
k’ = (tr - tm) / tm (2)
5
For this equation, k’ is the capacity factor, which is the retention time, tr, minus the time
for mobile phase, tm, divided by the time for the mobile phase. Resolution is another term
that can determine the quality of a separation, which shows how well the peaks are
separated between peaks that can be solved using equation 3.
R = tr / w (3)
R is the resolution, which is the difference in the retention time for the two peaks over the
average baseline width of the peaks. The selectivity factor, , or relative retention,
shows the quality of separation between two enantiomers. The greater the selectivity
factor is, the better the separation will be.
= k 'B / k 'A (4)
Alpha the selectivity factor is k’ of species B over that of A, where B is always the
second peak eluded from the column.
Thermodynamic properties can also be determined from HPLC retention data by
determining the capacity factor, k’, and the phase ratio, . The capacity factor, as
described above, can be expressed in the following ways:
k’ = (tr – t0) / t0 (5)
k’ = time of solute in stationary phase/time of solute in mobile phase (6)
Vs is the volume of the stationary phase and Vm is the volume of the mobile phase. The
phase ratio, , is equal to Vs / Vm, which is in the volume of the stationary phase, Vs, over
the volume of the mobile phase, Vm. In protein and polysaccharide columns the phase
ratio is not constant due to changes in the shape and volume of the stationary phase with
the changes in the mobile phase composition.
k’ = KD (Vs / Vm) = KD (7)
6
KD = Cs / Cm (8)
k' is the time in the stationary phase divided by the time in the mobile phase. The
distribution coefficient, KD, for a chromatographic system is also affected by the
thermodynamics of a solution [4]. Cs and Cm are concentrations for the stationary and
mobile phases, respectively.
The Selectivity Coefficient, , reflects k’ even though chromatographic system
may not reach equilibrium so it may not be equal in value. From thermodynamics, at
equilibrium, the free energy, G, equals –RT ln k so
G = - RT ln KD (9)
where R is the gas constant, and T is the temperature in Kelvin, K. Substituting equation
7; we get:
G = - RT ln k’ (Vm/Vs) = -RT ln k’ (1/) (10)
Rearranging equation 10 we get
G = - RT [ln k’ + ln (1/)] (11)
G = - RT [ln k’ - ln ] (12)
Using equation 13
G = H -TS (13)
and substituting it into equation 12:
H -TS = - RT [ln k’ - ln ] (14)
Rearranging equation 14
ln k’ - ln = -H / RT + S / R (15)
7
When this equation is plotted, in what is called a Van’t Hoff plot, with ln k’ vs. 1/T it will
have an intercept of S / R and a slope of -H / R [4]. G, H, S are define as
the differences of G, H and S for a given pair of enantiomers, R and S [5].
Then
G = -RT ln (kr’/ks’) = -RT ln (16)
then
ln = -H/RT + S/R (17)
Since there is no change in in a single isocratic chromatogram, that is, there is no
change in solvent composition, then
(ln R / ln S) = ln 1 = 0 (18)
The H and S values were determined by using a Van ’t Hoff plot of the natural log
of the selectivity factor versus the inverse of Temperature, in Kelvin. The slope of the
line is solved using equation 15 [5], where R is the gas constant.
Slope = -H / R (19)
The intercept is then used to determine the S value [5], with R as the gas constant
once again.
Intercept = S / R (20)
Several thermodynamic studies using these equations have been done and are presented
below
Weng et al. researched how to avoid the trial and error in the selection of
enantioseparation method [5]. The experiment was done on a Kromasil CHI-DMB
Brush-type column, consisting of an immobilized network polymer, with four amino acid
derivatives and two binaphthyl compounds that were enantioseparated successfully under
8
normal conditions. This experiment showed the significance that hydrogen bonding that
takes place in the mobile phase with the amino acids. The mobile phase for this
experiment was a mixture of n-hexane with differing amounts of 2-propanol. The
retention times of tested chiral compounds decreased and the separation of the
enantiomers increased tremendously when the amount of 2-propanol was below 3%. The
experiment also showed that the separated enantiomers of amino acid derivatives have
higher values and better separation than the enantiomers of binaphthyl compounds [5].
Enantiomers of amino acids with a phenyl group also had a larger enantioselectivity than
amino acids derivatives with a methyl group.
F
COOMe
NHCOPh
OH
OH
Figure 4. Example of one Amino Acid Derivitive and Binapthyl compound used in the Weng et al. experiment.
From van’t Hoff plots of ln vs (1/T), Weng et al concluded that
enantioseparations for the amino acids are enthalpy driven [5]. This is because both
H and S values of all four amino acids derivatives were negative, showing that
H is favorable and S is unfavorable
9
Figure 5. Weng et al van’t Hoff plot analyzing the effects of different temperatures and alcohol concentrations. From the graph on the far right, H and S values can be obtained [5].
Strong hydrogen bonding interactions or - interactions occurs with the
enantiomer and the Chiral Stationary Phase, CSP, which are enthalpy driven. The phenyl
group will enhance interactions more than a CH3 group because of additional -
interactions between the aromatic ring of CSP and the benzene ring of the solute. In the
larger compounds the OH groups are vital for the chirality of the compound but do not
enhance chiral recognition.
% 2-Propanol 25 C 30 C 35 C 40 C
H (kJ /m ol)
S (JK/ m ol)
12 1. 6 9 1. 6 3 1. 5 6 1. 5 0 -6. 4 1 -17 . 11
9 1. 7 1 1. 6 4 1. 5 7 1. 5 0 -6. 5 3 -17 . 42
6 1. 7 2 1. 6 5 1. 5 8 1. 5 1 -6. 4 9 -17 . 23
3 1. 7 6 1. 6 8 1. 6 1 1. 5 4 -6. 7 5 -17 . 95
1.5 1. 8 6 1. 7 8 1. 7 2 1. 6 7 -5. 5 1 -13 . 37
1 1. 9 8 1. 9 0 1. 8 6 1. 8 2 -4. 4 1 -9. 1 2
Figure 6. Data for the Amino Acid derivative (pictured above) indicating it is largely enthalpy driven.
Figure 7. Data for the Binaphthyl compound (pictured above) indicating that it is more entropy driven compared to the amino acid derivative.
The binaphthyl compounds have a less negative H value than the amino acids
because of the size of the compound [5]. Looking closer at the binaphthyl compounds,
the small change in H and S values compared to the amino acid derivatives shows
that steric hindrance causes enantioselectivity, not covalent bonding. But with an
increase in temperature the binaphthyl compounds could be more easily separated
because they are more entropy driven. This may reflect a conformational change in the
CSP polysaccharide at higher temperature. The large binaphthyl compounds have
hydrogen bonding because of the OH groups but lack other strong interactions due to
steric hindrance [5]. The results of this experiment indicate that the size of the compound
may determine whether or not it is enthalpy or entropy driven.
Yueqi et al. tried to take a different approach in the study of Biphenyl dimethyl
dicarboxylate derivatives (DDB) with a chiral HPLC column consisting of a CSP of
cellulose tris-(3,5-dimethylphenylcarbamate). DDB is used to treat hepatitis and shows
the potential to act against HIV-1. For this experiment, the cellulose CSP column was
specially made by the researchers but prepared according to standard procedure. In this
experiment, various mixtures of hexane and several different alcohols were tested as the
11
mobile phase. Enantiomers of nineteen different compounds were separated tested under
these conditions including five carboxylic acids, eight -biphenyl dimethyl dicarboxylate
derivatives, and eight -biphenyl dimethyl dicarboxylate and derivatives. For most
compounds, the best mobile phase modifier, MPM, was 2-propanol and for four -DDB
and one -DDB the best MPM was 1-butanol. Little changes in values were observed,
but the k’ value did decrease as the concentration of alcohol increased. Along with
different alcohols, a change in amount of alcohol in the mobile phase was tested. The
results indicated that the alcohol that worked best with all compounds was 2-propanol
over other alcohols [6]. A change in temperature was also tested and it seemed that the
compounds were separated more efficiently (sharper peaks) at higher temperatures even
though enantioselectivity is decreased (less enantiomeric separation). These results are
shown in Figure 5. At higher temperatures, the separations will occur at a faster rate and
the peaks will be much higher and thinner than lower temperature separations. So from
this research, the researchers concluded that enantioselectivity is affected by temperature
and is entropy driven [6].
Figure 8. Data that shows the effect of Temperature on the selectivity factor, Resolution and capacity factor [6].
In Jacobson et al. the chiral stationary phase was immobilized bovine serum
albamin (BSA) on a solid support and this column was used to separate L- and D-isomers
of benzoyl derivatives of amino acids. The CSP uses long protein chains that allow BSA
12
to covalently bond to wide pore silica [7] which is similar to the polysaccharide column
used in the experiments mentioned . In this experiment, there were two different types of
interactions taking place in the stationary phase in the column; chiral interactions and
non-chiral interactions. There were two different chiral interactions that were taking
place between the amino acids and the column. The D-amino acid experienced a three-
point interaction with the column [8]. This three-point interaction is the strongest
possible interaction taking place because it sterically fits. This interaction was a high-
energy interaction that is more sensitive to temperature.
The second chiral interaction in the column, from Jacobson et al, was a two-point
interaction with the L-amino acid [8]. The chiral interactions on the column were easily
saturated because it had a limited number of sites, thus this interaction worked better at
lower concentrations of analyte. The other interactions taking place in the column were
non-chiral interactions. These were one-point interactions, van der Waals forces, polar
interactions and hydrogen bonding which were all equal for both D and L–amino acids.
The relative energies for the two separation modes were found using the enthalpies of
adsorption. At lower temperatures, non-chiral sites were most important and the system
was more ordered. At higher temperature, the system underwent a conformational
change, was less absorbed and became more chirally selective.
For this research plan, a set a similar shaped chiral compounds and test their
separations in a few different ways. Conditions were varied in the mobile phase and
tested with hexane and various alcohols at various concentrations. To further test
separations of the compounds temperatures were varied from 20-40 C. The affect of
entropy and enthalpy on chiral separations were accumulated. Van ‘t Hoff plots were
13
acquired to determine the H and S values for each separation. These methods
were tested on the Chiralpak IA, a polysaccharide column with Amylose tris(3,5-
dimethylphenylcarbamate) as its chiral selector.
Experimental:
The HPLC instrument was two Shimadzu LC-20ACT solvent delivery systems
with dual reciprocating plunger design. The Detector was a Shimadzu SPD-20AV HPLC
UV-Vis Detector. It has a large sensitivity level having a 0.5x10-5 AU max noise level
reading. Samples were injected into the instrument in 10 mL aliquots. The data was
observed on a Windows 98 based computer with an EZstart program.
The Chiralpak IA (Chiral Technologies Inc. and Daicel Chemical Industries,
LTD, West Chester, PA) is a carbohydrate type column that is used with normal mobile
phase. It’s chiral selector is Amylose tris(3,5-dimethylphenylcarbamate with a solid
support diameter of 5 mm. With this column the normal-phase mobile phase consisted of
hexane (HEX) and isopropanol (IPA) in ratios of 95/5, 90/10 and 85/15, respectively. All
mobile phase solvents and reagents were purchased from Aldrich, Milwaukee, WI unless
otherwise stated. Two chiral sample solutions were used during May 2008 which were 3-
hydroxypropranoic acid-1-phenyl-ethylester (3HP-1EE) and flavanone. 3HP-1EE was
400 ppm in IPA and synthesized by Dr. David Soulsby. Flavanone was 300 ppm in 90:
10 HEX: ethanol.
During Fall 2008 and Spring 2009 the compounds were tested to find if the
separations were more enthalpy or entropy driven. To do this a carbohydrate-type
column was placed in a water bath to control the temperature of the column. Each
compound was tested 3 times at a temperature ranging from 20-35C and varying alcohol
14
from 5-15 %. The samples tested were flavanone and 3HP-1EE, the same concentrations
as before. The separation was performed in the normal phase so hexane with isopropanol
was used as the mobile phase. Each Separation run time was about 15 minutes long and
the peaks came off between 5 and 10 minutes depending on conditions
O OH
O
O
O
Figure 9. The compound on the left is 3HP-1EE and Flavanone is pictured on the right.
Figure 10. Both possible chains, Amylose and Cellulose, with possible chiral selectors. The Chiralpak IA is amylose base using the selector on the right.
15
Data/Results:
Figure 11. Separations for 3HP-1EE in 95/5 conditions with temperatures at 23, 25, 28, 31 and 35 C from top left down and right.
16
Figure 12. Shows original data obtained for 3HP-1EE. The separations were done in replicates of 3 at 4 temperatures. 90/10 and 85/15 indicate the percent hexane to isopropanol ratio.
Data values for the figure above are shown in the appendix. The data collected
showed good correlation but it was than tested at 6 different temperatures with duplicates
of 2 runs. This new data was collected in Table 1.
3.278 10.01 11.34 0.51 0.55 1.114 1.93 2.26 2.13 6210 35.0Table 1. Data for 3HP-1EE collectd at 6 temperatures at 95% Hexane, 5 % Isopropanol.
17
3HP-1EE y = 171.61x - 0.4487R2 = 0.9899
0.100
0.105
0.110
0.115
0.120
0.125
0.130
0.135
0.140
0.0032 0.00325 0.0033 0.00335 0.0034 0.003451/T
ln a
Figure 13. Van’t Hoff plot for 3HP-1EE to determine H and S values by using the ln vs. 1/T (1/K) This was using 95%hexane/5%isopropanol.
The slope of the line and y intercept is than used to fins the H and S
values. Using equation 19, the slope of 171.61 is used to determine the H, which is -
1430 J. With equation 20, the y-intercept is used to determine the S value, which is -
3.73 J/K.
The same method was than used on Flavanone.
18
Figure 14. Separations of falvanone under 95/5 conditions varying the temperatures from top left and down. Temperature readings were 20, 25, 27, 30 and 35 C.
19
Figure 15. Van’t Hoff plot separations of Flavanone enantiomers at 4 temperatures in 3 replicates.
The data collected had much less correlation than 3HP-1EE, which could be do to
changes in the column because it had just been hooked up to the pumps and detector.
Additional problems cold be due to the fact so of the readings were done on different
days. To test this readings would be taken at 7 different temperatures in duplicates of 2.