Introduction to Modern Liquid Chromatography, Third Edition part 38 ppt

Ion-exchange retention decreases as mobile-phase ionic strength increases Section 7.5.1, with the result that an increase in buffer concentration will tend to decrease the retention of p

Inertsil ODS-3

Symmetry C18

Discovery C18

Time (min)

Time (min)

1

23

6 + 7 11

5 bases 1–5

1 2

3 + 6

7

11

8

4 5+

bases 1–5

Time (min)

7 6

11

8 2

1 + 9

5 + 10 bases 1–5

(b)

Figure7.9 (Continued)

7.3.2.5 Other Conditions That Can Affect Selectivity

Conditions that are less used today for the control of RPC selectivity include:

• buffer type (e.g., phosphate, acetate, ammonium)

• buffer concentration

• amine modifiers

Buffer type is not commonly considered as a means of controlling selectivity for

the separation of ionic samples As discussed in Section 7.2.3, however, buffer type can affect the ‘‘effective’’ pKaof a solute, which is equivalent to a change in pH The largest changes in relative retention will occur when a basic buffer (e.g ammonium, triethylamine, etc.) replaces an acidic buffer such as phosphate or acetate, and vice versa Another way in which buffer type can contribute to selectivity is by ion pairing (Section 7.2.1.2) More hydrophobic buffers such as trifluoroacetate TFA

or (especially) heptafluorobutyrate HFBA can ion-pair with protonated bases BH+ and selectively increase their retention [30] The increase in retention for protonated bases increases with the positive charge on the solute molecule, as in the case

of peptides which contain multiple, basic amino-acid residues (Fig 13.8; Section 13.4.1.2) The use of TFA and HFBA as buffers is not subject to problems that are common for other ion-pair separations (Section 7.4.3)

Buffer concentration usually has only a minor effect on relative retention for

separations at low pH on modern (type-B) alkylsilica columns [26] However, for separations at pH> 6 and/or older, type-A columns (Section 5.2.2.2), protonated

bases can be retained by ion exchange as a result of interaction with ionized silanols

of silica-based column packings Ion-exchange retention decreases as mobile-phase ionic strength increases (Section 7.5.1), with the result that an increase in buffer concentration will tend to decrease the retention of protonated bases [31]

Amine modifiers, such as triethylamine or tetrabutylammonium salts, have

been added to the mobile phase in the past, primarily as a means of suppressing unwanted silanol interactions (Section 7.3.4.2) By interacting with stationary-phase silanol groups, amine modifiers can suppress ion exchange by the sample, thereby resulting in decreased retention for protonated bases These modifiers are little used today because (1) modern RPC columns (type-B) are largely free of unwanted silanol interactions and (2) the use of amine modifiers can be inconvenient, requiring long column-equilibration times in some cases

7.3.3 Method Development

Method development is similar for the RPC separation of either ionic or neutral

samples, as summarized in Figure 6.21a Seven method-development steps are

defined there, of which only one (step 3: choosing separation conditions) differs significantly for the separation of ionic samples The choice of separation conditions

for either ionic or neutral samples is summarized in Figure 6.21b and includes the

following steps:

1 choose starting conditions

2 select %B for 1≤ k ≤ 10

3 adjust conditions for improved selectivity and resolution

4 vary column conditions for a best compromise between resolution and run

time

Method development for ionic samples differs from that for neutral samples mainly

with respect to steps 1 and 3 above Method development should always start with a

new (unused) column, as exposure of a column to previous samples and conditions

can change its selectivity so as to make it impossible to replicate the column at a later time

7.3.3.1 Starting Conditions (Step 1)

Table 7.3 suggests conditions for the initial separation of a mixture of acids and/or bases, conditions that are similar to those recommended for the initial separation of neutral samples (Table 6.1) The main difference for ionizable samples is the need for a buffered mobile phase Because ionic samples are usually less strongly retained

in RPC, the value of %B for the initial mobile phase is likely to be a bit lower than for neutral samples However, it is best to start development at 80% B, so as

to reduce the risk of missing a late-eluted solute with a mobile phase that is too weak Alternatively (and preferably), an initial gradient-elution run can be used to determine the best value of %B for isocratic separation (Section 9.3.1)

Table 7.3

Representative Conditions for the Separation of Ionic Samples by Means of Reversed-Phase Chromatography

Columna Type: C8 or C18 (type-B)

Dimensions: 100 × 4.6-mm

Particle size: 3 μm

Pore diameter: 8–12 nm

Mobile phase 80 % acetonitrile-buffer; buffer is 10 mM potassium phosphate, adjusted to

pH−2.5b

Flow rate 2.0 mL/minb

Temperature 30 or 35◦Cb

%B Determined by trial and errorc

weight ≤ 10 μg

aAlternatively, use a 150× 4.6-mm column of 5-μm particles; note that a new (unused) column should

always be selected at the start of method development.

bInitial values will be varied during method development (Section 2.5); the starting pH of the mobile phase can also be varied.

cStart with 80%B and adjust further as described in Section 2.5.1.

The choice of starting mobile-phase pH and buffer depends on (1) the sep-aration goals and (2) what the chromatographer knows about the sample We recommend carrying out initial separations with a mobile-phase pH of 2.5 to 3.0, using phosphate for UV detection, or ammonium formate for LC-MS Problems aris-ing from peak tailaris-ing (Section 7.3.4.2), column instability (Section 5.3.1), or a lack

of method robustness (Section 12.2.6) are somewhat less likely for a mobile-phase

pH of 2.5 to 3.0 For samples that contain strong bases, there is increasing use

of high-pH mobile phases (pH > 8), in order to minimize the ionization of basic

solutes during separation Decreased sample ionization results in stronger retention, and can favor symmetrical peaks and more robust RPC methods However, when the mobile phase pH is> 8, special columns are required to avoid the dissolution

of the silica particles with resulting failure of the column (Sections 5.2.5, 5.3) The main advantage of a mobile-phase pH > 8 for strongly basic samples is that a

larger sample weight can be injected (Section 15.3.2.1), with a resulting increase

in detection sensitivity for an assay procedure, or increase in yield for preparative separations Whatever mobile-phase pH is used, care should be taken to ensure adequate buffering capacity (Section 7.2.1.1)

7.3.3.2 Optimizing Selectivity (Step 3)

Any of the separation conditions described in Section 7.3.2 can be varied in order

to improve relative retention and maximize resolution The simultaneous variation

of two different separation conditions will generally prove more effective; the same

two-variable procedures described for neutral samples in Section 6.4.1 can be applied for ionic samples We recommend that temperature and %B should be varied first

over a range that results in solute k-values within the range 0 5 ≤ k ≤ 20; see the

example of Figure 7.7 and the discussions of Section 6.4.1.3 and [32] If further changes in selectivity are needed, simultaneous changes in pH and solvent strength can be used for ionizable samples [33] When varying two conditions simultaneously, simulation software (Section 10.2) is especially helpful for determining conditions that correspond to maximum resolution

More than two conditions can be simultaneously optimized for the control

of selectivity; for example, varying %B and temperature for different columns (Section 6.4.1.4) is a popular and effective strategy A few other examples have been reported of the simultaneous optimization of three different variables [34, 35], each of which can be varied continuously (i.e., excluding column type as a variable) However, this approach can require a formidable number of experiments, for example, 32 experiments for the simultaneous optimization of %B, temperature, and pH in one example [34]

7.3.4 Special Problems

RPC separations of ionic samples are subject to two problems that do not occur for the separation of neutral samples

7.3.4.1 pH Sensitivity

As noted in Section 7.2 for the RPC separation of ionizable samples, relative retention can be quite sensitive to small (unintended) variations in mobile-phase pH The ability of most laboratories to replicate the pH of the buffer by means of a pH meter is typically no better than±0.05 to 0.10 units; variations in mobile-phase pH

of this magnitude may be unacceptable for some separations For this reason the

robustness of the final method in terms of pH should be a major concern during method development for ionic samples.

There are several ways in which the problem of pH sensitivity can be minimized First, determine the pH sensitivity of the method If the mobile-phase pH must be held within narrow limits (±0.1 unit or less), precise pH control can be achieved

by accurately measuring the buffer ingredients (either by weight or volume), rather than by using a pH meter to adjust the buffer to a desired pH (see Appendix II for some examples) Second, as an alternative to the precise adjustment of pH in this way, carry out separations with mobile phases that are, respectively, 0.2 pH units higher and lower than the required pH The inclusion of these chromatograms

in the method procedure can be used by an operator to guide the correction of mobile-phase pH when needed (Section 12.8)

Finally, the best approach for a method that proves to be too pH sensitive is

to re-optimize conditions so as to obtain a method that is more robust This will sometimes require a change in pH to a value that differs by more than±1 pH unit from the pKavalues of critical solutes (those whose resolution can be compromised

by small changes in pH) Minor changes in other conditions can also result in a more robust separation See the further discussion of Section 12.2.2.6 and [36]

7.3.4.2 Silanol Effects

Protonated basic solutes BH+ can interact with stationary-phase silanols by ion exchange (a buffer in the potassium form is assumed):

BH++ SiO−K+⇔ BH+SiO−+ K+ (7.7) This interaction can lead to increased retention, peak tailing, and column-to-column irreproducibility Problems of this kind are most pronounced when older, type-A columns (Section 5.2.2.2) are used, because type-A silica is contaminated by Al3+,

Fe2+, and other heavy metals Metal contamination increases silanol acidity, results

in a higher concentration of SiO− groups for all mobile-phase pH values, and likely contributes to poor column reproducibility Newer columns made from purer, type-B silica are largely free of metal contamination, and fewer associated problems are encountered in their use

Even when newer, type-B columns are used, the separation of basic compounds can lead to peak tailing [37] The origin of peak tailing with type-B columns appears to differ for separations with mobile phases of high or low pH For a mobile-phase pH< 5, tailing peaks usually resemble rounded right triangles, as in

the example of Figure 2.15e For a pH≥ 6, exponential peak tailing as in Figure

2.15a is more often seen Low-pH tailing is now believed due to charge repulsion

between retained ionized molecules (Section 15.3.2.1; [38]) As a result the column overloads more quickly for basic samples than for neutral samples, and peak tailing can become noticeable for injections of more than 0.5 μg of a basic compound

(assumes a column diameter of 4–5 mm) Low-pH peak tailing can be reduced somewhat by an increase in mobile-phase ionic strength For example, the use of buffers at a mobile-phase pH that favors buffer ionization results in a higher ionic strength, even when buffer molarity is unchanged; this approach for reduced peak tailing of bases has been recommended when volatile formate buffers are used for LC-MS [39]

The tailing of basic samples on type-B columns at pH-7 and above is less well understood, but may be the result of slow sorption-desorption of molecules of BH+ [37] The extent of tailing is affected by the nature of the B-solvent [40, 41], with

acetonitrile (worst)> methanol ≈ tetrahydrofuran (best)

Peak tailing is generally decreased by the use of higher column temperatures or higher

%B, conditions that also favor lower values of k The use of columns with smaller

values of C (cation-exchange capacity values; Section 5.4.1) is likely to minimize

peak tailing for a mobile-phase pH< 5 ‘‘Hybrid’’ particles (Section 5.3.2.2) do not

exhibit silanol ionization below pH-8, and the peak shape of protonated bases is good for pH< 8 [16] For a good review of peak tailing for basic solutes, see [16].

Small weights of undissociated carboxylic acids can exhibit tailing peaks for

some columns (Section 5.4.4.1), but peak shape improves for larger samples The

origin of such peak tailing is as yet unknown It is possible nevertheless to identify columns that are less likely to exhibit this problem (Section 5.4.4.1; [42])

In the case of type-A columns, various means have been employed in the past

to reduce ion exchange (Eq 7.7) and associated deleterious effects [43, 44]:

• suppress silanol ionization (use low-pH mobile phases)

• suppress the ionization of basic solutes B (use high-pH mobile phases)

• suppress ion exchange (use high–ionic-strength mobile phases)

• block ionized silanols (add amine modifiers to the mobile phase)

• use end-capped columns

Silica-based RPC columns can degrade more rapidly when the mobile phase

pH is<2.5 or >8.0, which limits the use of extreme pH to control silanol or solute

ionization Some type-B columns are now available for operation outside these pH limits (see the discussion of Section 5.3) Ion-exchange and related adverse silanol effects can also be minimized by the use of higher buffer concentrations; the buffer cation competes with the solute in the equilibrium of Equation (7.7) Buffer cations such as K+, Li+ and NH+4 are more effective than Na+ in suppressing silanols and minimizing peak tailing The addition to the mobile phase of amine modifiers such

as triethylamine and dimethyloctylamine was popular at one time for improving the separation of basic samples, but today the predominant use of type-B columns has rendered these (inconvenient) additives unnecessary End-capping the column (Section 5.3.1) tends to shield silanols from the solute and typically improves peak shape

7.3.4.3 Poor Retention of the Sample

Very polar samples are poorly retained in RPC, as noted for neutral samples in Section 6.6.1 The same problem is even more common for ionic samples—because

of the greater polarity of ionized molecules However, there are additional means for increasing sample retention in this case Poor retention of an ionic solute is usually due to its ionization, which can result in more than a 10-fold decrease in

values of k The simplest approach for solutes that are acidic or basic is a change in

mobile-phase pH that results in decreased solute ionization Alternatively, ion-pair chromatography (IPC, Section 7.4) can be used to similar effect, especially for per-manently ionized solutes such as quaternary-ammonium compounds Hydrophilic interaction chromatography (HILIC) is also effective for very polar samples (Section 8.6) and is usually a better choice than IPC for this purpose

7.3.4.4 Temperature Sensitivity

The relative retention of ionized solutes tends to be more dependent on temperature than is the case for neutral samples Therefore the need for accurate column thermostatting (Section 3.7) can be more important for ionic samples The use of (unthermostatted) separations at ambient temperature is not recommended for any sample, and it is especially problematic for the separation of ionic samples

Ion-pair chromatography (IPC) can be regarded as a modification of RPC for the separation of ionic samples The only difference in conditions for IPC is the addition

of an ion-pairing reagent R+ or R− to the mobile phase, which can then interact with ionized acids A−or bases BH+ in an equilibrium process:

ionized solute ion pair

hydrophilic solute hydrophobic ion-pair (less retained in RPC) (more retained in RPC)

The use of IPC can thus create similar changes in sample retention as by a change in mobile-phase pH (Section 7.2), but with greater control over the retention of either acidic or basic solutes, and without the need for extreme values of mobile-phase pH (e.g., pH< 2.5 or > 8) Typical ion-pairing reagents include alkylsulfonates R–SO3−

(R−) and tetraalkylammonium salts R4N+(R+), as well as strong (normally ionized) carboxylic acids (trifluoroacetic acid, TFA; heptafluorobutyric acid, HFBA [R−]), and so-called chaotropes (BF4−, ClO4−, PF6−)

When first introduced in the 1970s, high-performance IPC was found to reduce peak tailing for basic solutes This and its ability to increase the retention of weakly

retained ionized acids and bases for acceptable values of k were primary reasons

for its use at that time Additionally IPC provides further options for the control of selectivity in the separation of ionic samples Today the predominant use of type-B columns has reduced the importance of peak tailing, and we now have a better understanding of how best to control RPC selectivity The poor RPC retention of very hydrophilic acids and bases (especially strong bases that remain ionized for

pH< 8) can also be addressed in other ways, for example, (1) by the use of high- or

low-pH mobile phases in order to minimize solute ionization and increase retention (combined with the use of columns that are stable at pH extremes) and (2) by the use

of hydrophilic interaction chromatography (HILIC, Sect 8.6) Consequently there is much less need for IPC today because of its greater complexity and other problems (Section 7.4.3)

When developing an HPLC separation, we recommend starting with RPC,

followed by the addition of an ion-pairing reagent only when necessary When, or

for what applications, might IPC be recommended? IPC separation involves two additional variables (type and concentration of the IPC reagent) that can be used for further control of selectivity As will be seen below, the effects of an added IPC reagent on solute retention are reasonably predictable, when we know whether a particular peak corresponds to an acid, base, or neutral Consequently the retention

of both acidic and basic solutes can be varied continuously so as to optimize their separation, when other changes in RPC conditions fail to achieve acceptable resolution

IPC can also be used to narrow the retention range of a sample, so samples that might otherwise require gradient elution can be separated isocratically An example

is shown in Figure 7.10, for a proprietary sample that includes a drug-product X

plus several preservatives and degradants In Figure 7.10a, RPC separation is shown

with a mobile phase of 30% methanol-buffer (pH-3.5) The neutral preservative, propylparaben PP, is strongly retained, while the basic drug X and its degradants

0 20 40

Time (min)

no ion-pairing (0< k < 30)

ion-pairing (0.6< k < 9)

(a)

(b)

X

X1

X3 + HB B

Time (min)

HB

B MP X

PP

X2

3

X1

Figure7.10 RPC separation of a proprietary mixture of acids, strong bases and neutrals Sam-ple: X, strongly basic (proprietary) drug substance; X1, X2, X3, strongly basic degradants of X; MP and PP, methyl and propyl paraben preservatives; B, benzoic acid; HB,

hydroxyben-zoic acid (degradant of MP and PP) Conditions: (a) 150 × 4.6-mm column (5-μm particles);

30% methanol-buffer mobile phases (buffer is pH-3.5 acetate); 30◦C; 2.0 mL/min (b) Same as (a), except mobile phase is 45% methanol-buffer plus 65-mM octane sulfonate; 1.5 mL/min.

Adapted from [46]

X1–X3are weakly retained (because they are in the protonated form as BH+) Two other sample compounds, benzoic acid B (another preservative) and hydroxybenzoic acid HB (a paraben degradant), are acidic, while methyl paraben MP is also a

neutral preservative The separation of Figure 7.10a exhibits an excessive retention

range (0≤ k ≤ 30), which would normally suggest gradient elution as an alternative

(Section 9.1) Because compounds X–X3 are strongly basic, an increase in their isocratic retention (relative to the rest of the sample) by an increase in pH was deemed impractical Thus a mobile-phase pH> 8 would be necessary (requiring a

column that is stable at high pH), but this would lead to k

B and HB, while having no effect on the retention of neutral compounds MP and

PP Thus no practical change in pH is able to narrow the retention range of this

sample so as to provide k-values for all compounds in an acceptable range of values

(e.g., 0.5 ≤ k ≤ 20).

Isocratic elution was preferred for the sample of Figure 7.10, so the use of IPC was investigated as an alternative to gradient elution The addition of a sulfonate IPC reagent would be predicted to lead to strongly increased retention for the sample cations (X–X3), accompanied by a modest decrease in the retention of both sample acids (B and HB) and neutral compounds (MP and PP); see Section 7.4.1.2 below

The separation of Figure 7.10b was therefore carried out with octane sulfonate as

IPC reagent (for a preferential increase in the retention of X–X3), plus a stronger

mobile phase (45% B vs 30% B in Fig 7.10a) for a reduction in k for the neutral

solute PP The resulting decrease in retention range (0.6 ≤ k ≤ 9) now allows the

baseline separation of this sample within a reasonable time (11 min)

The remainder of this section provides a short description of the basis of ion-pair separation, followed by a discussion of how separation depends on various conditions For further details, see [45] and Chapter 7 of [46]

7.4.1 Basis of Retention

Two possible retention processes or ‘‘mechanisms’’ exist for separation by IPC

As an example, we will use the ion-pairing of an ionized acidic solute A− by a tetraalkylammonium IPC reagent R+ The ion-pairing of a protonated basic solute

B+by an alkylsulfonate IPC reagent R− can be described similarly

One hypothesis for IPC retention assumes that an ion-pair forms in solution,

as described by Equation (7.8a) The resulting ion-pair A−R+ is retained by the column; that is, the solute retention equilibrium as described by Equation (2.2) in Section 2.2 is replaced by

A−R+(mobile phase)⇔ A−R+(stationary phase) (7.9) According to this hypothesis, retention is governed by (1) the fraction of solute

molecules A in the mobile phase that are ionized (determined by mobile-phase pH

and the solute pKavalue), (2) the concentration of the IPC reagent and its tendency

to form an ion pair (the equilibrium constant for Eqs 7.8 or 7.8a), and (3) the value

of k for the ion-pair complex A−R+ (which will be greater for more hydrophobic IPC reagents)

An alternative picture of IPC retention assumes that the IPC reagent is retained

by the stationary phase, with retention then occurring by an ion-exchange process

(Section 7.5.1), for example, for an ionized acid A−and IPC reagent R+X−:

A−(mobile phase)+ R+X−(stationary phase)⇔

A−R+(stationary phase)+ X−(mobile phase) (7.9a)

That is, the ion-pair reagent R+X− first attaches to the stationary phase, and then

the sample ion A− replaces the counter-ion X− in the stationary phase Either of these two IPC retention processes (Eqs 7.9 or 7.9a) might predominate for a given separation, but which mechanism plays the more important role is neither easy to determine nor important in practice It has been shown that these two retention mechanisms are virtually equivalent [47], and both provide similar predictions of retention as a function of experimental conditions Consequently either process can

be assumed in practice We will use the ion-exchange process of Equation (7.9a) in the following (simplified) discussion, because this retention mechanism appears to

us to be easier to understand and to apply in practice

7.4.1.1 pH and Ion Pairing

Further insight into IPC retention as a function of mobile-phase pH is provided by Figure 7.11 for the case of an acidic sample (a carboxylic acid RCOOH) and a positively charged IPC reagent (tetrabutylammonium, TBA+) In Figure 7.11a, no

IPC reagent is added to the mobile phase (i.e., RPC separation), so the non-ionized acid RCOOH is preferentially retained by the C8 stationary phase (shown as a

C8C8C8C8C8C8

C8C8C8C8C8C8

C8C8C8C8C8C8

C8C8C8C8C8C8

C8C8C8C8C8C8

C8C8C8C8C8C8

k

pK a

RCOO − + H +

RCOO − + H +

(b)

TBA + TBA + TBA + TBA + TBA + TBA + TBA + TBA + TBA + TBA + TBA + TBA + TBA + TBA + TBA + TBA + TBA + TBA +

pH

pH

Figure7.11 Representation of retention for an acidic solute RCOOH as a function of

mobile-phase pH; (a) RPC and (b) IPC with tetrabutylammonium ion (TBA+) as IPC reagent

in high concentration

surface with attached C8 groups) Also shown on the right side of Figure 7.11a is

a plot of retention k as a function of mobile-phase pH; retention decreases with

increased pH, due to the greater ionization of RCOOH (i.e., characteristic RPC retention for an acidic solute as pH is varied)

In Figure 7.11b, the retention of the same compound RCOOH is shown,

except that the IPC reagent TBA+ has been added to the mobile phase in sufficient concentration to cover the entire stationary phase surface—hence blocking RPC interaction of the sample (non-ionized RCOOH) with the column C8 groups Now the ionized acid RCOO−is preferentially retained by ion exchange with TBA+in the

stationary phase (Eq 7.9a) The dependence of k on mobile-phase pH is seen to be the reverse of that in Figure 7.11a for RPC (no ion-pairing); retention in Figure 7.11b

increases with increasing mobile-phase pH and the consequent increasing ionization

of the solute, due to ion-pairing of the ionized solute If fully protonated bases

BH+ are present in the sample (e.g., strong bases), their retention will decrease for increasing concentrations of the IPC reagent (TBA+), due to the repulsion of the positively charged BH+ ions by the positively charged TBA+ ions in the stationary phase, as well as by the decreased availability of C8groups—which are covered by