JOURNALOF

CHROMATOGRAPHYA
ELSEVIER

Journal of Chromatography A, 770 (1997) 23-28

Retention behavior of carboxylic acids on highly cross-linked

poly(styrene-divinylbenzene)-based and silica-based cation
exchangers
a*

C h r i s t i a n W. K l a m p f l ' , W o l f g a n g B u c h b e r g e r ,

Guenter Rleder,

G u e n t h e r K. B o n n b

aDepartment of Analytical Chemistry, Johannes-Kepler-University, Altenbergerstrasse 69, A-4040 Linz, Austria

blnstitute of Analytical Chemistry and Radiochemistry, University of lnnsbruck, lnnrain 52a, A-6020 lnnsbruck, Austria

Abstract

A highly cross-linked porous poly(styrene-divinylbenzene) resin was reacted with sulfuric acid to give strong cationexchange resins with capacities from 0.2 to 1.9 mequiv./g. The retention behavior of mono- and dicarboxylic acids on these
resins was compared with results obtained from silica-based cation exchangers functionalized with alkylsulfonic acid or
phenylsulfonic acid groups. The contribution of different retention mechanisms like ion exclusion, hydrophobic interaction
and adsorption to the separation of the analytes on these chromatographic supports was discussed. Additionally the
applicability of silica-based cation exchangers was tested in the field of silage analysis.

Keywords: Stationary phases, LC; Carboxylic acids

1. Introduction

Since ion-exclusion chromatography has been

introduced by Wheaton and Baumann [1] it has
become an important technique for the analysis of
carboxylic acids. Usually a strong cation-exchange
resin is used as stationary phase which separates
these analytes according to their degree of ionization.
Strong acids like mineral acids are completely
excluded from the pores of the packing material by
electrostatic repulsion, whereas neutral compounds
e.g., methanol penetrate into the pores of the resin.
Between these two extreme cases, partially ionized
molecules are retained corresponding to their degree
of ionization. Under the conditions commonly used
in ion-exclusion chromatography, carboxylic acids
are not fully dissociated so that the retention order
depends on their pK a values [2,3].
*Corresponding author.

Most ion-exclusion separations described in the

literature for carboxylic acids have been performed
with highly sulfonated poly(styrene-divinylbenzene)
(PS-DVB) resins of low cross-linkage. It is known
that these resins have in addition to actual ionexclusion effect, additional mechanisms, like size
exclusion and hydrophobic interactions involved in
the retention of analytes [4]. The inherent reversedphase character of the separation system can be a
useful variable in the optimization of the separation
if mobile phases containing organic solvents are
employed [5-7]. Unfortunately, due to an unfavorable swelling behavior the compatibility of low
cross-linked PS-DVB resins with organic solvents is
limited to approximately 20%. Taking full advantage
of the reversed-phase characteristics in addition to
the ion-exclusion properties requires the use of an
highly cross-linked cation-exchange resin. So far
only a few papers have reported separations of
carboxylic acids on fully solvent-compatible ion-

0021-9673/97/$17.00 Copyright 1997 Elsevier Science B.V. All rights reserved

PII S 0 0 2 1 - 9 6 7 3 ( 9 6 ) 0 1 0 7 5 - 8

C.W. Klampfl et al. / J. Chromatogr. A 770 (1997) 2 3 - 2 8

24

exclusion columns [8-10]. The range of applications

seems to be not yet fully exploited so that further
investigations should be justified, especially as the
availability of highly cross-linked monodisperse P S DVB particles has increased during recent years
[11,12]. In this context it was also interesting to
investigate as to how far this packing material would
be useful in narrow-bore columns. This type of
columns offers an enhanced detection limit and an
improved compatibility with mass spectrometric
detection, which is becoming increasingly important
when structure-selective information is required.
An estimation of the extent of reversed-phase
mechanisms in ion-exclusion chromatography looks
possible if the hydrophobic PS-DVB base material
is replaced by a hydrophillic material. For this reason
silica modified with sulfonic acids via aliphatic as
well as aromatic spacers has been investigated in this
work. Silica-based ion-exclusion columns have been
rarely used for the separation of carboxylic acids
[13], although most recently Ohta et al. [14] have
reported the use of unmodified silica for the separation of organic acids. The present work aims at a
direct comparison of solvent-compatible PS-DVB
and sulfonated silica and at the elucidation of the
parameters of the mobile phase that are crucial for
the separation. Finally, the applicability of these new
stationary phases are demonstrated for real samples
in the field of silage analysis.

2. Experimental

ford, MA, USA), an ABI 785A programmable

absorbance detector (Applied Biosystems, San Jose,
CA, USA) and an HP 3359A chromatographic data
acquisition system (Hewlett-Packard, Palo Alto, CA,
USA). For column packing a Knauer pneumatic
HPLC pump (Knauer, Berlin, Germany) was used.
2.2. Columns
PS-DVB beads were prepared by a two-step
microsuspension method similar to the procedure
described elsewhere [11,12]. The resulting particles
had a mean diameter of 4 p~m (---0.3 I~m), a specific
surface area between 92 and 94 m2/g and a crosslinking degree of 50%. For the sulfonation of the
material and the determination of the exchange
capacity a procedure similar to the one described by
Dumont and Fritz was used [15]. General reaction
conditions for the functionalization of 5 g of the
resin are given in Table 1. All polymer resins were
suspended in acetonitrile-water (1:1), sonicated and
packed into 100X2 mm I.D. stainless-steel columns
for 30 min at 35 MPa.
Exsil 100/5 SCX was purchased from Alltech
(Lancs., UK). The material was suspended in methanol, sonicated and packed into 2 5 0 x 4 mm I.D.
stainless-steel columns for 30 min at 40 MPa. A
Nucleosil 100-5 SA 250X4 mm I.D. column was
purchased from Machery-Nagel (Diiren, Germany)
and a 250X4 mm I.D. column filled with Exsil
100/5 from SRD (Vienna, Austria).

2.1. Instrumentation

2.3. Reagents and chemicals

The chromatographic system consisted of a Waters

625 LC-system equipped with a Waters 600E system
controller, a Waters 717 autosampler (Waters, Mil-

All chemicals, except those used for the synthesis

of the polymeric resin, were of analytical-reagent
grade. Gradient grade acetonitrile was purchased

Table 1
Reaction conditions for resin sulfonation
Capacity (mequiv./g)
0.0
0.2
0.6
1.2

1.9

H2SO 4 (ml)

Acetic acid (ml)

Reaction time (min)

Temperature (C)

30
15
15
15

30
50
50
50

4
60
90
120

0
0
75
80

C.W. Klampfl et al. / J. Chromatogr. A 770 (1997) 23-28

from J.T. Baker (Deventer, Netherlands). Doubly

distilled water was used.
Sulfuric acid eluents were prepared by dilution of
0.05 M sulfuric acid with doubly distilled water or
organic solvent-water. Buffer solutions were obtained by dissolving NaH2PO 4 in doubly distilled
water and adjusting the pH value by addition of
phosphoric acid.
Press-liquors of silage samples were obtained by
passing the silage through a hand-driven press. Prior
to the analysis the liquid was diluted with water by a
factor of 500 and passed through a solid-phase
extraction cartridge filled with 100 mg of unmodified
PS-DVB resin.

25
- - x - - oxMc ec~

---m-- f omlc ocld

- - i - - nce~c ~,kJ

1
iIg

-X

,~cher~e capacity [meo/gl

Fig. 1. Influence of the exchange capacity of the resin on the

distribution coefficient of selected acids on sulfonated PS-DVB.
Chromatographic conditions: eluent: 5 mM H2SO4; flow-rate: 0.2
ml/min.

3. Results and discussion

3.1. Chromatographic behavior of highly crosslinked PS-DVB-based cation exchangers

As can be seen in Table 1, a highly cross-linked
P S - D V B resin has been functionalized by introduction of sulfonic acid groups to give a series of
cation-exchange resins with different exchange capacities. Sulfonation of this material led to a maximum exchange capacity of 1.9 mequiv./g, which is
less than 50% of the value specified for low crosslinked PS-DVB resins, commonly used in ion-exclusion chromatography. A further difference was the
dimension of the columns used for separation. Conventional ion-exclusion columns are filled with nearly the forty fold amount of resin than the 100x2 mm
I.D. column used in this work. Fig. 1 shows the
influence of the degree of sulfonation on the retention behavior of a selection of carboxylic acids.
For most of the investigated solutes a maximum of
retention was obtained on a resin with an exchange
capacity between 0.2 and 0.7 mequiv./g. These data
are similar to some results obtained on a sulfonated
PRP-1 resin published earlier by Lee and Lord [8].
As can be seen from this plot the two structural
similar acids, maleic acid and fumaric acid, can not
be separated on unmodified PS-DVB. By introduction of a small amount of sulfonic acid groups on
the surface of the resin both analytes show an
increase in retention. This may be explained by the
fact that the higher polarity of the sulfonated P S -

DVB particles enhances the ability of the eluent to

wet the resin. By a further increase of sulfonation
maleic acid and fumaric acid are separated corresponding to their degree of ionization. The distribution coefficient obtained for maleic acid, a
strong acid (pK a 1.83), decreases if the exchange
capacity of the resin and with it the ionic repulsion is
increased. On the other hand more than 90% of the
fumaric acid molecules (pK a 3.03) are protonated
under the selected conditions and therefore may
penetrate into the pores of the beads, leading to
increased hydrophobic interaction between the solute
and the PS-DVB resin. This different behavior of a
strong and a weak carboxylic acid indicates that an
ion-exclusion mechanism is present despite of the
low exchange capacity of the column.
Regarding the distribution coefficient for all investigated mono- and dicarboxylic it turns out that
only formic acid, oxalic acid and malonic acid show
distribution coefficients in the range of 0 and 1. The
retention of all higher carboxylic acids is predominantly affected by hydrophobic interaction between
the analytes and the resin. Because of the solvent
compatibility of these columns this effect can be
employed for the separation of higher monocarboxylic acids by gradient elution with up to 70%
acetonitrile in the eluent. Under these conditions
acids with a carbon chain length up to twelve could
be separated within less than 20 min; the separation
of acids with even longer chain length might be
successful but has not been tried in this work.

26

C.W. Klampfl et al. I J. Chromatogr. A 770 (1997) 2 3 - 2 8

3.2. Chromatographic behavior of silica-based

cation exchangers
3.2.1. Cation exchangers with phenylsulfonic acid
groups
The chromatographic behavior of mono- and
dicarboxylic acids has been investigated using a
2504 mm I.D. column filled with Nucleosil 100-5
SA, a silica-based cation exchanger functionalized
with phenylsulfonic acid groups and an exchange
capacity of 1.0 mequiv./g. Compared to the P S DVB-based resin mentioned above this cation exchanger is based on a hydrophillic base material,
though to some extent still hydrophobic interactions
between the analytes and the phenyl groups of this
material should be possible.
Fig. 2 shows the distribution coefficients obtained
for the investigated solutes. Similar to the results
obtained on PS-DVB-based columns, this Fig. demonstrates that between pH 5 and pH 3 the distribution
coefficients of the analytes increase in a way that
corresponds to the pKa values of the solutes. Below
pH 3, especially in the case of dicarboxylic acids the
plots for strong acids e.g., oxalic acid and malonic
acid significantly differ from those obtained for
weaker acids like adipic acid and pimelic acid. The
increasing distribution coefficients for the stronger
acids can be explained by their high degree of
ionization at pH 3 which is slightly decreased by
lowering the pH to 2. For weaker acids an effect
running in opposite direction can be observed.
Experiments using NaH2PO 4 buffers (pH 3) with
different molarity showed that increasing ionic
strength of the eluent from 5 mM to 50 mM led to a
decrease of the distribution coefficient of about 0.1
to 0.15 units. By adjusting the pH values of the
eluents with phosphoric acid, the ionic strength of
the buffer solutions increased with decreasing pH. At
pH 3 already less than 10% of the weaker acids are
ionized. Therefore the effect of lower distribution
coefficients caused by increasing ionic strength gains
more influence on the retention behavior of these
solutes. This leads to slightly dropping graphs for
these analytes below pH 3.
Regarding the retention volumes obtained for the
selected analytes only formic acid, oxalic acid,
malonic acid and succinic acid elute within the
confines of the interstitial volume of the eluent and

--x-- forrdcacid

acellcacid

3.5.

3.0.

vab~ alcd

2.5,
_

'" -'- - - - , ~ , ~ . ~

'

i 2.0'
1.5.
1.0"
0.50.0-

?...\

pH
B

3.0'
2.5, ~

--x-- oxalicacid
I ~
acid
succhlc acid
- - l - - glutaricacid
----0--9d~ acU

-~

~zo-

" ~

1.5

--o-- p~lc =u

a
I

1.o
0.5. _

~s

"

;.o

0.0"

zo

;5

3'.o

"

~5

"

5'.o

pH

Fig. 2. Influence of the pH of the eluent on the distribution

coefficient on Nucleosil 100-5 SA. Chromatographicconditions:
eluent: 5 mM NaH2PO4, pH adjusted with H3PO 4. (A) Monocarboxylic acids; (B) dicarboxylicacids.

the retention volume obtained for a neutral compound like methanol. This behavior may be explained partially by hydrophobic interaction between
the phenyl groups of the spacers and the analytes.
However, also adsorption phenomena between the
silica gel and the solutes as reported by Ohta et al.
[14] should be considered.

3.2.2. Cation exchangers with alkylsulfonic acid

groups
Using the Exsil 100/5 SCX column, an ion-exchange column packed with propylsulfonic acid

C.W. Klampfl et al. / J. Chromatogr. A 770 (1997) 2 3 - 2 8

functionalized silica gel and an exchange capacity of

0.45 mequiv./g, a chromatographic behavior similar
to the one on the phenylsulfonate silica could be
observed. It should be mentioned, that compared to
the Nucleosil material much lower distribution coefficients were obtained on this column. Nevertheless
all monocarboxylic acids with more than two carbon
atoms and all dicarboxylic acids with more than five
carbon atoms show longer retention times than
methanol. This effect may be attributed exclusively
to adsorption phenomena between the silica gel and
the analytes, because of the lack of hydrophobic sites
on this chromatographic support. In each of the
series of mono- and dicarboxylic acids with increasing carbon chain length, the number of acids eluting
in front of methanol is increased by one.
In Fig. 3 a separation of a standard mixture of
mono- and dicarboxylic acids on Exsil SCX 100/5
as well as Nucleosil 100-5 SA is depicted. It can be
seen that a change in selectivity occurs concerning
the elution order of butyric acid and pimelic acid.

j,,,,._
4

27

Regarding the influence of the pH on the distribution

coefficients in the case of the Nucleosil column both
carboxylic acids only show a slight decrease in
retention going from pH 2.5 to 3.5. In contrast to this
behavior the distribution coefficient obtained for
butyric acid is nearly constant from pH 2.5 to pH 4,
whereas pimelic acid shows a maximum at pH 2.5
and a steady drop till pH 5 on the Exsil column.
Therefore at pH 3.5 the elution order of butyric acid
and pimelic acid is reversed when switching from the
Nucleosii column to the Exsil column. Additionally
it should be mentioned that the separations achieved
on sulfonated silica-based cation exchangers could
not be performed on a Exsil 100/5 column packed
with unmodified silica.

3.3. Analysis of fermenting acids in silage samples

To demonstrate the suitability of silica-based
cation-exchange columns for the analysis of real
samples the amount of acetic acid, lactic acid,
propionic acid and butyric acid in a press-liquor of a
silage sample has been determined. From the different contents of these fermenting acids conclusions on
the quality of the silage sample can be drawn. Fig.
4A shows the separation of a standard mixture of the
investigated solutes at a concentration of 25 ppm
each in less than 5 min. Linearity of the calibration
plots prepared for these solutes was checked from 5
ppm to 100 ppm. A chromatogram of a diluted
press-liquor sample containing lactic acid and acetic
acid is depicted in Fig. 4B. The investigated sample
contained 2.4% lactic acid and 0.5% acetic acid.

4. Conclusion
2

Fig. 3. Separation of carboxylic acids on Exsil SCX 100/5 (A)

and Nucleosil 100-5 SA (B). Peaks: 1=malonic acid; 2=formic
acid; 3=acetic acid; 4=propionic acid; 5=pimelic acid; 6 =
butyric acid; 7=valeric acid. Eluent: 5 mM NaH2PO4, pH 3.5;
flow-rate: 0.8 ml/min. Detection: UV 210 nm.

The investigations described in this paper on the

chromatographic behavior of carboxylic acids indicate that various retention mechanisms are involved
in the separation of these solutes on cation-exchange
columns. As can be seen from the literature even in
the case of low cross-linked highly sulfonated P S DVB resins hydrophobic interactions between the
analytes and the resin cannot be neglected besides
ion exclusion. Going to solvent compatible highly
cross-linked PS-DVB supports with low exchange
capacities a distinctly increasing influence of an

28

C.W. Klampfl et al. / J. Chromatogr. A 770 (1997) 2 3 - 2 8

A
1
3
4

base material, the degree of sulfonation as well as

the type of spacer used in the case of silica gels leads
to additional selectivity in the separation of carboxylic acids. As an example for the suitability of
the chromatographic supports discussed in this work,
the content of fermenting acids in a silage pressliquor sample was determined within 5 min using a
silica-based cation exchanger functionalized with
phenylsulfonic acid groups.

mlautm

Acknowledgments

The authors would like to thank Mr. A. Ackerlauer

for assistance in the experimental work.

References

Fig. 4. Separation of fermenting acids on Nucleosil 100-5 SA. (A):

Standard mixture containing 25 ppm of each acid; (B): silage
press-liquor (diluted 1:500) containing 2.4% lactic acid and 0.5%
acetic acid. Peaks: l=lactic acid; 2=acetic acid; 3=propionic
acid; 4=butyric acid. Eluent: 5 mM NaH2PO4, pH 4.0; flow-rate:
0.8 ml/min. Detection: UV 210 rim.

hydrophobic interaction mechanism can be observed.

Even by a change in the base material of the cation
exchanger from PS-DVB to the more hydrophillic
silica gel does not lead to an exclusive contribution
of ion exclusion in the retention of carboxylic acids.
Besides adsorption phenomena on the silica gel also
hydrophobic interactions may occur to some extent if
aromatic spacers are used to fix the sulfonic acid
groups on the silica support. The results obtained in
this work demonstrated that a careful selection of the

[1] W. Wheaton and W.C. Bauman, Ind. Eng. Chem., 45 (1953)

228.
[2] P.R. Haddad and P.E. Jackson, Ion Chromatography: Principles and Applications, Elsevier, Amsterdam, 1990.
[3] J.S. Fritz, J. Chromatogr., 546 (1991) 118.
[4] B.K. Glod, Chem. Anal., 39 (1994) 399.
[5] K. Tanaka and J.S. Fritz, J. Chromatogr., 361 (1986) 151.
[6] E. Papp and P. Keresztes, J. Chromatogr., 506 (1990) 157.
[7] R. Widiastuti, P.R. Haddad and P.E. Jackson, J. Chromatogr.,
602 (1992) 43.
[8] D.P. Lee and A.D. Lord, LC.GC, 5 (1987) 261.
[9] P. Walser, J. Chromatogr., 439 (1988) 71.
[10] K. Kihara, S. Rokushika and H. Hatano, J. Chromatogr., 410
(1987) 103.
[11] T. Ellingsen, O. Aune, J. Ugelstad and S. Hagen, J. Chromatogr., 535 (1990) 147.
[12] S. Wongyai, J.M.Varga and G.K. Bonn, J. Chromatogr., 536
(1991) 155.
[13] B.K. Glod and W. Kemula, J. Chromatogr., 366 (1986) 39.
[14] K. Ohta. K. Tanaka and P.R. Haddad, J. Chromatogr. A, 739
(1996) 359.
[15] P.J. Dumont and J.S. Fritz, J. Chromatogr. A, 691 (1995)
123.