316 IONIC SAMPLES: REVERSED-PHASE, ION-PAIR, AND ION-EXCHANGE CHROMATOGRAPHY
A
−
or BH
+
by ion-pairing:
ionized solute ion pair
(acids) A
−
+ X
+
⇔ A
−
X
+
(7.5)
(bases) BH
+
+ X
−
⇔ BH
+
X
−
(7.5a)
hydrophilic (less retained in RPC) hydrophobic (more retained in RPC)
Buffers that ion-pair are usually organic ions, and they tend to be more hydrophobic
than inorganic ions. For example, trifluoroacetate is a significant ion-pairing agent
(Section 13.4.1.2), whereas phosphate is not. Ion-pairing by inorganic buffers is
usually not significant, although some studies suggest that even phosphate may
undergo weak ion-pairing with protonated bases [8]. For further information on

There is increasing use of higher pH mobile phases (pH
>
8) as a result of the
development of RPC columns that are stable at high pH (Section 5.2.5). Borate and
ammonia have been used to some extent as buffers for high-pH operation, but note
the further discussion of Section 5.8 concerning column stability at high pH. See
7.2 ACID –BASE EQUILIBRIA AND REVERSED-PHASE RETENTION 317
also Appendix II for details on the more convenient preparation of some common
buffers of required pH.
7.2.2 pK
a
asaFunctionofCompoundStructure
When selecting a range of mobile-phase pH values within which to carry out method
development (i.e., optimization of pH), it can be useful to know the approximate
pK
a
values of the various sample components. This information allows the mobile
phase to be restricted within an appropriate range of pH values. For example, a
mobile-phase pH range that varies from pK
a
− 1topK
a
+ 1 (for the sample) is useful
for controlling retention and selectivity by changes in mobile-phase pH. Similarly
a mobile-phase pH outside this range will result in a more robust method that
is less sensitive to small (unintended) changes in pH. Values of pK
a
vary widely
for different organic compounds, but a large number of pK
a

a
values in water for some common acid- or
base-substituent groups present in typical sample molecules. It is also possible to
infer values of pK
a
from experimental plots of retention against pH, as for peaks
2and4inFigure7.3a.
7.2.3 Effects of Organic Solvents and Temperature on Mobile-Phase pH
and Sample pK
a
Values
This detailed section is less essential in everyday operation, so the reader may wish
to skip to Section 7.3. Nevertheless, the conclusions of this section are potentially
useful for an accurate interpretation of the relationship between sample retention
and mobile-phase pH.
Literature values of pK
a
for different compounds (as in Table 7.2) are usually
reported for aqueous solutions at near-ambient temperatures. Quite often, however,
RPC separations of ionic samples are carried out at higher temperatures, with mobile
phases that contain varying amounts of organic solvent. Both mobile-phase pH and
values of pK
a
for the sample can be affected by added organic (specifically, the
value of %B) and by temperature. However, a knowledge of the true (mobile-phase)
pH and solute pK
a
values as a function of %B and temperature has little practical
importance, so far as routine RPC assays are concerned; it is only important that %B
and temperature are maintained constant for all runs so that solute ionization and

Sulfoxide, –SO– 2
Thiazole 2
Amine, –NH
2
,–NR
2
10 5
Pyridine 5
Imidazole 5
Piperazine 10
Note: Values can vary by 1 to 2 pK
a
units or more as a result of adjacent groups in the molecule.
a
See Figure 13.1 for values of individual amino acids.
and temperature on the pK
a
values of both solute and buffer (Tables 7.1 and 7.2).
Effective pK
a
values for the solute can be used with the pH of the buffer (not
the buffer-organic mobile phase) to estimate solute ionization and retention as a
function of pH. An effective pK
a
value is equivalent to the value that can be inferred
from an experimental plot of retention against mobile-phase pH (i.e., buffer pH, as
in Fig. 7.2).
7.2.3.1 Effect of %B on Values of Effective pK
a
for the Solute

7.3 SEPARATION OF IONIC SAMPLES BY REVERSED-PHASE CHROMATOGRAPHY (RPC) 319
each 1% increase in %B (experimental data of [17, 18]). For example, consider the
separation of several substituted anilines, using 25% methanol-phosphate buffer,
as reported in [17]. Effective pK
a
values for these solutes would be expected to be
25 × 0.03 = 0.75 units lower than literature values. For eight different solutes with
literature pK
a
values of 2.7 to 5.3, experimental pK
a
values (as in Fig. 7.2) for this
sample were lower by an average 0.7 ± 0.1 units (1 SD). Similarly, for six substituted
benzoic acids separated with 40% methanol and acetate buffer [17], effective pK
a
values were the same as literature values (±0.1 unit, 1 SD). The effect of added
acetonitrile on effective pK
a
values is similar to that for the addition of methanol.
The effect on relative retention of changes in effective pK
a
values with %B is
equivalent to a change in mobile-phase pH (the ‘‘effective’’ pH of the mobile phase),
which suggests that a change in %B can have a larger effect on relative retention
and selectivity for ionic samples than for neutral samples. This has been observed
for gradient elution as a function of gradient steepness [19, 20], which is equivalent
to a change in %B for isocratic elution (Section 9.1.3).
In the discussion above, the mobile-phase pH is equated to that of the measured
pH of the aqueous buffer, a procedure that we recommend for reasons given in
Section 7.2.1. The pH of the final, water-organic mobile phase could be measured

C is typical), the effect of temperature on solute
pK
a
values will usually be small and can therefore be ignored when using estimated
values of pK
a
for method development.
7.3 SEPARATION OF IONIC SAMPLES BY REVERSED-PHASE
CHROMATOGRAPHY (RPC)
Reversed-phase chromatography (RPC) should be a first choice for the separation
of mixtures of ionizable organic compounds. Method development for the RPC
separation of ionic samples (Section 7.3.3) proceeds in similar fashion as for neutral
samples, with some important differences that are developed in the remainder of this
320 IONIC SAMPLES: REVERSED-PHASE, ION-PAIR, AND ION-EXCHANGE CHROMATOGRAPHY
section. The following information on RPC separation (Sections 7.3.1, 7.3.2) should
be useful for both method development and in troubleshooting routine separations.
7.3.1 Controlling Retention
Following an initial experiment, mobile-phase strength (%B) can be varied to obtain
a desirable retention range (e.g., 1 ≤ k ≤ 10), the same way as for neutral samples
(Sections 2.5.1, 6.2.1). Alternatively (and generally preferable), initial experiments
can be carried out using gradient elution, as discussed in Section 9.3.1. Once a value
of %B has been selected for the reasonable retention of the sample, the next step is
the adjustment of separation selectivity for optimal relative retention and maximum
resolution.
7.3.2 Controlling Selectivity
For the separation of neutral samples, selectivity can be varied by changing solvent
strength (%B), temperature, the B-solvent, or the column. These variables also
affect the separation of ionic samples, usually to a greater extent than for neutral
samples. In addition separation selectivity for ionic samples is strongly affected
by mobile-phase pH, and to a lesser extent by the kind and concentration of the

a
≈ 5–10) will be fully ionized,
and acids (pK
a
≈ 5) will be in the neutral form. This is only approximately true,
since it overlooks the effects of %B and temperature on values of pK
a
,aswellas
changes in pK
a
that can result from the presence of different substituents in the
solute molecule. Consequently, while pH selectivity is usually reduced at low pH, it
can still be significant—depending on the sample and the value of %B.
7.3 SEPARATION OF IONIC SAMPLES BY REVERSED-PHASE CHROMATOGRAPHY (RPC) 321
Another example of a change in relative retention with pH is shown in
Figure 7.6 for a mixture of substituted benzoic acids (peaks 1–4) and anilines
(peaks 5–7). As mobile-phase pH is increased from 3.2 to 4.3 (Fig. 7.6a–c), the
retention of acids 1 to 4 decreases, while the retention of bases 5 to 7 (shaded peaks)
increases. For a mobile-phase pH of 4.3 or higher, the acidic compounds 1 to 4 are
mainly in the ionized form and therefore retained weakly; similarly, at higher pH
the basic compounds 5 to 7 are largely non-ionized and more strongly retained. As
a result for a mobile-phase pH
>
4 there is a separation of these acids and bases
into two groups of peaks. An optimum mobile-phase pH = 3.4(Fig.7.6d)provides
acceptable resolution of the sample. However, even at this relatively low pH, the
separation of Figure 7.6d is seen to be somewhat sensitive to small changes in pH;
0 2 4 6 8 10 12 14
5
1

Time (min)
Time (min)
Time (min)
Time (min)
5
1
2
6
3
4
7
02468101214
2
1
3
5
4
6
7
02468101214
2
1
3
4
5
6
+
7
1-4 acids
5-7 bases

sample bases have decreased as a result of an increase in either %B or temperature,
which is equivalent to an increase in mobile-phase pH for these basic compounds
Time (min)
13% B, 35
°
C
5 ≤ k ≤ 16
R
s
= 0.6
28% B, 35
°
C
1 ≤ k ≤ 5
R
s
= 0.3
19% B, 49
°
C
3 ≤ k ≤ 9
R
s
= 3.3
28% B, 60
°
C
1 ≤ k ≤ 4
R
s

024 024
5
1
2
3
4
+
6
7
5
2
3
4
6
7
02468
Time
(
min
)
5
1
2
3
6
4
7
(e)
Figure 7.7 Effect of mobile-phase strength (%B) and temperature on the separation of a
mixture of acids and bases. Sample: same as in Figure 7.6; conditions also the same, except

±
, this relationship can be simplified to
k ≈ k
0
(1 − F
±
), (7.6)
where k
0
is the value of k for the neutral (non-ionized) molecule, and F
±
is the
fractional ionization of the solute for a given mobile-phase pH. An increase in either
temperature or %B will lead to a decrease in values of k
0
for the solute, regardless of
whether it is ionic or neutral. Additionally a change in conditions that also changes
the ‘‘effective’’ mobile-phase pH (and therefore values of F
±
) can have a further
effect on the separation of an ionic sample. Thus in Figure 7.7 an increase in either
%B or temperature appears to increase mobile-phase pH slightly (equivalent to a
decrease in pK
a
values for these solutes)—with a preferential retention of basic
solutes 5 to 7.
7.3.2.3 Solvent Type
A change in solvent type (e.g., methanol replacing acetonitrile) is expected to have
a comparable effect on the relative retention of both ionic and neutral samples. In
addition any change in ‘‘effective’’ pK

= 35
Luna phenyl-hexyl
2 ≤ k ≤ 8, R
s
= 0.4
F
s
= 33
Time (min)
1
5
2
3
4
+
6
7
1
2
+
5
3
4
7
6
1
2
+
3
4

(Section 5.4.2):
Symmetry/Altima, F
s
= 35; Symmetry/Luna, F
s
= 33. In each case these F
s
values
suggest significant differences in column selectivity, although much larger differences
can be achieved with other pairs of columns.
Another example of the effect of the column on selectivity is shown in
Figure 7.9 for the separation of a mixture of five, fully protonated strong
bases (1–5), five partly ionized weak acids (6–10), and a neutral reference
7.3 SEPARATION OF IONIC SAMPLES BY REVERSED-PHASE CHROMATOGRAPHY (RPC) 325
compound (11; shaded peak). In Figure 7.9a,mixturesofeither the strong bases
or weak acids plus neutral compound 11 are separated on each of these three
columns. In Figure 7.9b corresponding separations of samples containing all 11
compounds are shown. The relative retention of the fully protonated strong bases
(1–5) of Figure 7.9 is most affected by values of the ion-exchange capacity C for the
column (Section 5.4.1); larger values of C mean a greater retention of protonated
bases. Values of C at pH-2.8 for these columns are, respectively, −0.47 (Inertsil),
−0.30 (Symmetry), and 0.18 (Discovery). As expected, the relative retention of
basic solutes 1 to 5 increases in proceeding from the Inertsil to the Symmetry to the
Discovery column (note the retention ranges for peaks 1–5 in Fig. 7.9b, indicated at
the top of each chromatogram by arrows). The relative retention of the weak acids
6 to 10 and neutral compound 11 are quite similar on the three columns because
their retention is not affected by values of C.
02
Time (min)
2460

11
8
9
10
6
7
11
8
9
10
6
7
11
8
9
10
Strong bases Acids
Inertsil ODS-3
C = −0.47
Inertsil ODS-3
Symmetry C18
C = −0.30
Symmetry C1
8
Discovery C18
C = 0.18
Discovery C18
3
(a)
Figure 7.9 RPC separation of an ionic sample as a function of column type. Sample: (bases)