If you look carefully, you will see that almost every chromatographic peak tails toa certain degree. Although many of today’s liquid chromatography (LC) columns are less problematic from this standpoint than
were their ancestors, the problem has not been eliminated. This month’s “LC Troubleshooting” discusses the causes of peak tailing, how to measure it, some preventive steps you can take and why newer column types are less prone to tailing.
What Is Peak Tailing?
A peak is labelled as tailing or asymmetrical when it deviates from the ideal, symmetrical shape of a Gaussian peak. Figure 1 shows an example of a tailing peak. The later eluted half of the peak is wider than the front half, and the broadening appears to be emphasized near the baseline. Although fronting peaks can occur in chromatography, they are rare compared with tailing peaks. Because peak tailing can influence the quality of a separation, it is a good idea to quantify the amount of tailing a peak has. Two widely used measurements of peak tailing exist.
Workers in the pharmaceutical industry use the US Pharmacopeia (USP) tailing factor (Tf): where a and b are the front and back half-widths at 5% of the peak height, as shown in Figure 1. Most other workers use the asymmetry factor (As) where a and b are measured at 10% of the peak height. If a b, both measurements yield a value of 1.0 — the value for a non-tailing peak. For values of less than approximately 2, As and Tf values are roughly the same, as is shown in Table 1. It really doesn’t matter which measurement of tailing is used — unless it is dictated by an industry regulation — as long as the same technique is used consistently so that changes in peak tailing can be tracked.
What Is Acceptable?
Some degree of peak tailing is unavoidable, but how much can be tolerated? Figure 2 shows some examples of peaks with various asymmetry factors. When a column is new, most manufacturers will allow it to pass the
acceptance criteria if As is between 0.9 (slight fronting) and 1.2 (slight tailing). Most practical workers using real samples strive for asymmetry factors of no more than 1.5, but they will settle for less than 2.0 if necessary. Generally, when asymmetry factors are greater than 2.0, something is wrong with the separation and the problem should be addressed.
Potential Problems
A look at the peaks in Figure 2 will provide some indication of the potential problems created by tailing peaks. Compare the peak with an asymmetry factor of 1.0 with the peak that has an asymmetry factor of 4.0. The differences are dramatic. For example, although the peak areas are the same, there is a threefold difference in peak height. When using a method for trace analysis, the peak height becomes a critical factor in determining the lower limit of quantification, and the tailing peak is clearly inferior. Another potential problem is the performance of the data system. Peakdetection algorithms determine the start and end of a peak by detecting a change in the slope of the baseline. For all the peaks shown, the start of the peak is the same, and it is easy for the data system to determine when the peak starts. The more the peak tails, the more gradual the return to baseline. In addition, the end of the peak is difficult to determine, especially if the baseline is not smooth. If the data system uses tick marks to indicate the start and end of the peak, it is likely that the end of the peak would be shown in different places for multiple injections of the same tailing peak. This results in an error in the area measurement.
Tailing peaks are also problematic if minor peaks occur in the same sample as major ones; for example, during a stabilityindicating assay or a measurement of metabolites or impurities. With stabilityindicating assays, International Committee on Harmonization of Technical Requirements for Registration of Pharmaceuticals for Human Use (ICH) guidelines require reporting of peaks of 0.1% of the main peak or larger. It is easy to see that a minor peak could hide under the peak tail and that ICH requirements for minor peaks might not be met if severe peak tailing occurred. LC methods used for routine analysis must usually provide baseline resolution of all the peaks and run times that are as short as possible. Figure 2 shows that baseline resolution of tailing peaks will require longer run times than that for peaks that do not tail or that have only minor tailing. So, tailing peaks not only reduce the quality of the chromatogram, they also reduce the ability to quantify components in the sample. Furthermore, because of questions about peak purity and proper quantification, analysts’ confidence in their results is weakened when separations contain severely tailing peaks
What Causes Peak Tailing? Peaks tail because more than one retention mechanism is present in a separation, and one of those mechanisms is overloaded. Chemists tend to think of a reversed-phase separation—for example, with a C18 column—as being a pure hydrophobic retention mechanism. It might be true if the stationary phase was only C18 groups. However, approximately one-half of the silica surface is unbonded, which leaves exposed silanol (Si–OH) groups that can interact with the sample. The silica particles that form the solid support are made of a polymer of silicon and oxygen. Like carbon, silicon is tetravalent, so silica’s internal structure comprises four oxygen atoms attached to each silicon atom in a three-dimensional polymer. At the surface, the polymer terminates in silanol groups. Figure 3 shows the different possible silanol configurations. Chromatographers most commonly draw the surface as free silanols, but silicon atoms with two hydroxyl groups in a geminal configuration also are present. If the silanol groups are positioned next to each other, an associated silanol group can share a proton with an adjacent group. The free silanol groups are more acidic than the geminal and associated groups, and they interact more strongly with basic solutes. The result is peak tailing commonly associated with the separation of bases. To further complicate matters, trace metals (e.g., iron and aluminium) can be present in the silica matrix. These trace metals can act as ion-exchange sites or, when adjacent to free silanols, can withdraw electrons, which makes the silanols even more acidic. In the past, column packings designated as Type A silica tended to have all theforms of silanols shown in Figure 3. These Type A materials usually yielded tailing peaks for basic compounds.
In the past 10
years, column manufacturers have learned
how to make high-purity silica or Type B
silica. These products are made carefully in
metal-free environments so that trace
metal levels are extremely low. Proprietary
technologies have enabled manufacturers
to simultaneously reduce the free-silanol
content and increase the number and spatial
distribution of geminal and associated silanols.
The result is a much less acidic silica surface.
In addition, Type B silica columns have less
of a tendency to generate severe peak
tailing with basic compounds. What to Do about Tailing?
With Type A columns, reducing peak tailing
usually meant modifying the mobile phase
to include a tail-suppressing compound.
Triethylamine was a popular mobile-phase
additive for reducing peak tailing. Used at
concentrations of 20 mM or more in the
mobile phase, triethylamine could reduce
peak tailing significantly with Type A
columns. In my laboratory, my co-workers
and I perform contract separation services
for the pharmaceutical industry. When we
depended on Type A columns, triethylamine
was present in most mobile phases.
Today, with the availability of Type B
silicas, peak tailing for basic compounds is
much less of a problem. In my laboratory
we now use Type B columns almost
exclusively; very seldom do we add
triethylamine to the mobile phase. As a
general rule, low-pH mobile phases (pH 3)
also act to suppress silanol ionization,
which further reduces tailing problems.
Today, the added reliability of the columns
and the simplicity of the mobile phases
make for more dependable and robust LC
methods. I see no reason to develop a new
separation on a Type A column—you are
just asking for problems that easily could
be avoided by using a Type B packing material.
Peak tailing is primarily a problem with
silica-based packing materials. Reversedphase
stationary phases attached to organic
polymers, zirconia or other non-silica supports are available as alternatives to silica-based
columns. These speciality columns could
solve your separation problem and produce
peaks with little or no peak tailing.
What about Gradients?
Generally, analysts are most concerned
about peak tailing with peaks generated by
isocratic separations because gradients
have two characteristics that minimize
tailing. The first characteristic is more of an
illusion than a real difference. All the peaks
in a gradient separation are approximately
the same width, whereas an isocratic
separation generates peaks that increase in
width with retention. Tailing often is more
visually obvious with broader peaks such as
the later eluted peaks in an isocratic
separation. Because gradient peaks tend to
be narrower, tailing is less obvious.
The second characteristic that minimizes
the problem of peak tailing with gradients
is peak compression, a process that acts on
all gradient peaks. Think of a gradient on a
microscopic scale—the concentration of
organic solvent gradually increases over
time in a reversed-phase gradient. At any
point in a column, the solute molecules
passing through it experience a weaker
mobile phase than those that pass through
it later. For a given peak, this phenomenon
means that the molecules at the front of
the peak are traveling in a weaker solvent
than those at the back edge of the peak.
A stronger solvent (more organic) elutes
molecules through a column faster than a
weaker solvent does. Thus, under gradient
conditions, the molecules at the trailing
edge of a peak are always travelling faster
than those at the front edge of the peak.
In other words, the tail is constantly being
pushed back into the main peak, and peak
tailing is reduced. For this reason, peak
tailing in gradient separations is less of a
problem than the tailing for the same
compound under isocratic conditions and
with the same column and mobile-phase
components.
Conclusions
Peak tailing is a real and unavoidable
problem with most LC separations.
Manufacturers have made great strides in
the development of Type B silica support
materials, and although these materials
significantly reduce peak tailing, they
seldom eliminate it. Peak tailing can cause
qualitative and quantitative problems with
LC methods, so it is important to monitor
peak tailing so that it does not compromise
your analytical results. Measure peak tailing
using either the USP tailing factor or the asymmetry factor. With today’s Type B
packing materials, adding reagents to
suppress peak tailing is generally
unnecessary. However, strongly basic
solutes might still require the use of
triethylamine or another reagent to reduce
peak tailing. Another alternative is to
explore the use of a non-silica–based
packing material.
The use of silica-based packing materials
can be a love–hate relationship. I
remember when organic polymer–based
supports were introduced in the mid-1970s.
Many workers thought they would be the
death knell for silica because they did not
have the tailing problems associated with
silica. But then they realized that the silanol
effects of the stationary phase played a
very important role in selectivity of the
chromatographic separation. Although
polymer-based columns could generate
symmetrical peaks, they were not as useful
as silica-based columns for separating
complex mixtures. The perfect LC column
has yet to be invented.
Reference
1. L.R. Snyder, J.J. Kirkland, and J.L. Glajch,
Practical HPLC Method Development (John
Wiley & Sons, New York, 2nd ed.), p. 180
(1997).
“LC Troubleshooting” editor John W.
Dolan is vice-president of BASi Northwest
Laboratory of McMinnville, Oregon, USA; a
training consultant for Rheodyne LLC, the
LC Resources Training Group, of Walnut
Creek, California, USA; and a member of
the Editorial Advisory Board of LC•GC
Europe. Direct correspondence about this
column to “LC Troubleshooting,” LC•GC
Europe, Advanstar House, Park West,
Sealand Road, Chester CH1 4RN, UK.
Readers can also direct questions to the
on-line Chromatography Forum at
http://www.chromforum.com
WITH BIG THANKS TO John W. Dolan, BASi Northwest Laboratory, McMinnville, Oregon, USA. LC•GC Europe September 2003
YOURS CHROMATOGRAPHICALLY
KAUSHIK ZALA
