2
TAHERI ET AL.
and hence considerable care may cause too much time for
preparation. Although it is completely acceptable for indus-
trial applications, in practical processes of research where
time and sample size are limited (enantiomers at the range
of grams), scientists prefer to use rapid batch chromatogra-
phy in semipreparative scales.
and propranolol had to be inspected. Chiral separation on a
polar stationary phase with a nonpolar MP containing an
organic modifier was ordered as NP chromatography. The
effect of the MP composition on the retention in NP chroma-
tography, operating under conditions of a linear isotherm, has
been described using theoretical models of adsorption, devel-
oped in previous studies.3-6 In these models, the adsorption
process on a polar adsorbent surface was defined as a compe-
tition between the molecules of the solute and those of a
modifier for adsorption sites. Despite some differences, all
the models after simplifications led to the same simple
equation describing the retention of the solute as a function
of concentration of a strong solvent (the modifier):
An attractive practical method for preparation of pure
components from binary mixtures by the use of a single col-
umn is to utilize the stacking mode for sample injection, mak-
ing the total column length work, which eventuates in
maximum throughput and consequently low production
costs. To maximize sample loading, touching bands of the
two enantiomers takes place automatically. Since in concen-
tration overloading, throughput is determined by a selectivity
factor (α), a high value of α is always recommended.2 How-
ever, in practice, the solubility limitation in the mobile phase
(MP) causes band touching of the peaks with volume
overloading and this does not allow all the stationary phase
to work, so that the column diameter determines throughput.
Also, big band broadening causes a long delay between injec-
tions. Stacked injection does not require a big α, but needs
baseline separation; therefore, optimizing the resolution (Rs)
is more important than α for stacked injection. Since the
MP composition in stacked injection has to be the same as
that of the feed solvent to avoid baseline perturbation, another
difficult task is optimization of the MP to keep the baseline
separation alongside the dissolving power. This is the main
aim of this work: to show that stacked injection could over-
come low throughput due to the solubility limitation. While
an MP providing sufficient Rs as well as high solubility of
the enantiomers would be excellent for stacked injection on
a single‐batch column, usually low solubility of the enantio-
mers in MP convinces experts to change the MP or sample
solvent, or to use another resolving method like crystalliza-
tion. To solve this problem for stacked injection of poorly sol-
uble enantiomers, a new method development is introduced
according to the simultaneous optimization of MP and dis-
solving power. A well‐known drug, tramadol, with low solu-
bility in normal phase (NP) was used as a model enantiomer
of this kind.
ki ¼ k0;iðCmodÞ−m
(1)
i
where k0 is the retention factor of the solute in pure modifier,
Cmod is the percentage of modifier, and mis the empirical
constant, which is determined by fitting to the set of experi-
mental retention data acquired at different modifier contents
in the MP. α (related to the two components) easily gains
from the above equation:
k2 k0;2ðCmodÞ−m
k0;2
k0;1
2
∝
¼
¼
¼
ðCmodÞ−m þm
(2)
2
1
1;2
k1 k0;1ðCmodÞ−m
1
Once 1 and 2 are enantiomers, in a pure modifier it is
acceptable to assume k0 , 1 ≅k0 , 2 and m1 > m2, which
results in
∝
¼ ðCmodÞþm
(3)
1;2
This model shows a very straightforward linear positive
relation between α and Cmod. It is a useful estimation but does
not provide further information about the combination effects
with other factors, such as flow rate and/or concentration of
the additive. In some cases, this model is not complete at
all to fulfill the entire requirements of stacked injection.
Therefore, a central composite design (CCD) on surface
response methodology including three factors of modifier
and additive concentrations as well as the flow rate, with fur-
ther analysis of variance (ANOVA) analysis (which easily
allows modeling of the combined effects) was used for accu-
rate investigations.7
Careful culling of the responses is probably the most
important concern of this kind of modeling. Fortunately,
there are well‐known responses in chromatography to satisfy
all the requirements such as k, α, Rs, run time (RT), peak area
(A), etc.8 Meanwhile, α is the most important response in pre-
parative chromatography, but since stacked injection needs
touching‐band resolution, Rs might be more useful. The use-
fulness of α and/or Rs in chiral separation with stacked injec-
tion will be discussed further. To take the highest mass
loading in the lowest band broadening, A and RT were
For highly soluble enantiomers in MP, usually
chromatographists prefer to use nonlinear overloading
instead of baseline resolution, but this is labor‐intensive to
optimize the recovery and purity at the same time, either
experimentally or theoretically. Undoubtedly, stacked injec-
tion by touching‐band overloading, which provides 100%
recovery and purity, is simpler and more rapid for lab‐scale
production. Therefore, throughput of stacked injection for
propranolol was investigated as a model of highly soluble
enantiomers in MP.
To optimize a separation, we need to discover how differ-
ent parameters affect the separation. Thus, the influence of
the most effective factors on chiral separation of tramadol