The Goal of Separation

Band Broadening & Its

Prevention

The molecules in a sample are exposed to a number of

physical actions on their transport through the
chromatographic system. The molecules are introduced
in the mobile phase and interact with the stationary
phase during the chromatographic process. One of the
Important parameters for the result of the separation is
the flow rate of the mobile phase.

van Deemter and coworkers were already studying the

effect of the mobile phase on the efficiency of the GC
separation and they could express the efficiency in this
formula the van Deemter equation:

This formula expresses the correlation between

the efficiency given by H, the height equivalent to a
theoretical plate, and the band broadening
phenomena A, B and C as a function of the flow rate, u.

A-term Eddy diffusion

The term A describes eddy diffusion of component
molecules and the variable pathways the mobile phase
may follow through the stationary phase packing in the
column. The component molecules may therefore travel
different distances as they pass through a unit length of
column ( Figure 1)

Figure 1. Peak broadening due to eddy diffusion

Figure 1

The particles also cause eddy or turbulence in the

mobile phase resulting in mixing and hence dispersion
of the molecules. The term A is independent of mobile
phase velocity , but it is a function of the size of the
stationary phase particles and the way they are packed
in the column or coated on a TLC plate.

B-term Longitudinal diffusion

Molecules dispersed in a liquid or gaseous phase will be
in random and will diffuse in all directions independently
of the direction of flow (just as a sugar lump dissolves
slowly in water even without being stirred)
(Figure 2&3).
This forward and backward diffusion will therefore
contribute to molecules moving ahead or lagging
behind the main part of the band resulting band
broadening.

Figure 2

Band broadening by longitudinal

diffusion. Left: Sample zone immediately
after injection. It will spread out in all
three axes of space (arrow directions).
Right: Sample zone at a later moment. It
is larger now due to diffusion and it has
also been transported by the flowing
mobile phase
Figure 3

Longitudinal diffusion is time dependent since the

longer a band takes to elute the more there will be for
diffusion to take place ; B is therefore inversely
proportional to the velocity of mobile phase.
The extent of diffusion will be hindered by particles of
column packing and the coefficient diffusion of the
component in the mobile phase.

The longitudinal diffusion has a disadvantageous

effect on plate height only if:
(a) small stationary phase particles,
(b) too low a mobile phase velocity in relation to the
particle diameter, and
(c) a relatively large sample diffusion coefficient
coincide in the chromatographic system.

C terms resistance to mass transfer

Figure 4c: mobile phase mass transfer. This refers to
differing flow rates for different parts of a single flow
stream or path between surrounding particles. In (c),
where the flow stream between particles 1 and 2 is
shown, it is seen that liquid adjacent to a particle
moves slowly or not at all, whereas liquid in the center
of the flow stream moves fastest.

Figure 4

As a result, in any given time, sample molecules near

the particle move a short distance and sample
molecules in the middle of the flow stream move a
greater distance. Again this results in a spreading of
molecules along the column.

Figure 4d shows the contribution of stagnant mobilephase mass transfer to molecular spreading. With
porous column-packing particles, the mobile phase
contained within the pores of the particle is stagnant or
unmoving. In d we show one such pore, for particle 5.
sample molecules move into and out of these pores
by diffusion.

Those molecules that happen to diffuse a short distance

into the pore and then diffuse out, return to the mobile
phase quickly, and move a certain distance down the
column. Molecules that diffuse further into the pore
spend more time in the pore and less time in the
external mobile phase. As a result, these molecules
move a shorter distance down the column. Again there
is an increase in molecular spreading.

In Figure 4e is shown the effect of stationary-phase

mass transfer. After molecules of sample diffuse into a
pore, they penetrate the stationary phase (crosshatched region) or become attached to it in some
fashion. If a molecule penetrates deep into the
stationary phase, it spends a longer time in the particle
and travels a shorter distance down the column - just as
in (d).

Molecules that spend only a little time moving into and

out of the stationary phase return to the mobile phase
sooner, and move further down the column.

The first principle on which a good column is based is

that the packing should be composed of particles with as
narrow a size distribution as possible. The ratio between
the largest and the smallest particle diameters should not
exceed 1.5 -2.0
The broadening due to eddy diffusion and flow
distribution is little affected, if at all, by the mobile phase
flow velocity.

The second principle is that the mobile phase flow

velocity should be selected so that longitudinal
diffusion has no adverse effect. This applies when
u>2Dm/dp, where u is the linear flow velocity of the
mobile phase, Dm the diffusion coefficient of the
sample in the mobile phase and dp the particle
diameter.

The third principle is that small particles or those

with a thin, porous surface layer should be used as
the stationary phase.
The fourth principle is that low-viscosity solvents
should be used.
The fifth principle is that high analysis speed is
achieved at the expense of resolution and vice
versa. However, this effect is much less pronounced
with smaller than with larger particles.

Optimum Mobile Phase Velocity

A chromatography column is judged by its ability to
separate complex mixtures using a given mobile
phase. Highest column efficiencies , N, will be
obtained when equilibrium step or plate height is at
a minimum, since N = L/H.
This will be achieved when the velocity of the mobile
phase is at optimum and the band broadening
processes described in the van Deemter are at a
minimum.

The A term is independent of flow velocity, B is

inversely proportional to flow velocity and C is
directly proportional to flow velocity .

The theoretical plate height, can be expressed as

a function of mobile phase flow velocity, u
(Figure 5). The H/u curve is also called the van
Deemter curve. The optimum flow rate uopt
depends on the properties of the analyte.

Figure 5. The van Deemter plot showing the

optimum efficiency (at minimum H) and the
optimum flow rate

Figure 5

Correcting Peak Tailing Problems in

Reversed Phase HPLC

Peak tailing in reversed phase HPLC continues to

be a common complaint. It is particularly
prevalent when separating basic compounds
and, therefore, a source of constant problems to
those analyzing pharmaceutical compounds by
HPLC.

Peak tailing causes a number of problems:

lower resolution,
reduced sensitivity,
and poorer precision and quantitation.

As peak tailing (T, Tailing factor)

increases from 1.0 to 2.0,
resolution (Rs ) decreases from 1.5
to 1.0. Sensitivity (peak height)
also decreases with peak tailing
since the peak volume increases
and the sample concentration
decreases.

Tailing peaks make it more

difficult for data systems to
identify exactly where a peak
ends.
Because of this, accuracy and
precision can suffer. In this
example, the peak area measured
at point B is 3% less than the peak
area measured at point A.

Causes of Peak Tailing

The major causes of peak tailing include:
injecting sample in a solvent that is significantly
stronger than the mobile phase,
sample mass overload,
stationary phase silanol interactions with amines,
adsorption of acidic compounds onto silica, and
a void in the column packing bed.

TABLE 1
Peak Tailing: Causes and Cures
Causes
Sample solvent stronger
than the mobile phase

Sample mass overload

Cure
Dissolve sample in mobile
phase or at least reduce the
strength of the sample solvent
as much as possible
Reduce the amount (mass)
of sample injected. Refer to
Table 2 for recommended
sample injection size for
different column configurations

 Silanol interaction with

amines
See Figure 6 for a ranking of
C18 phases by silanol activity

Reduce mobile phase pH to

< 3.0
Increase mobile phase ionic
strength. (25mM to 50mM
recommended)
Add a competing amine to
the mobile phase.
10 mM TEA is usually sufficient.
Select a stationary phase
with a lower silanol activity.

 Adsorption of acids on
silica

Increase salt
concentration in the mobile
phase
25 mM to 50 mM is usually
sufficient
Reduce the pH of the
mobile phase to < 3.0
Add a competing organic
acid.
1% acetic acid or 0.1% TFA is
usually sufficient

 Column void

Replace the HPLC column.

Attempts to fill-in the void
are seldom worth the effort.

Peak Tailing Caused by Injecting Sample in a Solvent

that is Significantly Stronger than the Mobile Phase
If the sample is dissolved in a solvent that is stronger
than the mobile phase, broad and even tailing peaks
can occur. There are several clues to help you identify if
this is the cause of peak tailing. The first clue is
that early eluting peaks have more tailing than late
eluting peaks. Another clue is that peak tailing
improves if less sample volume is injected or if the
sample is diluted with the mobile phase.

Peak Tailing Caused by Sample Mass Overload

When the sample mass injected begins to exceed the
capacity of the column packing, the peaks will take
on the look of a right triangle. As more sample mass is
injected, the front of the peak will become sharper
and the back of the peak will tail more. Another clue
that the column is being overloaded is that retention
will decrease as greater sample mass is injected. (See
Figure 3)

As sample overload
occurs, the front of
peaks will become
sharper, the back of the
peak will tail,and the
retention time will
decrease slightly. If the
peak shape looks
somewhat like a right
triangle, sample
overload is likely
occurring.

The cure for peak tailing caused by sample mass overload

is to inject less sample. Table 2 provides a list of column
diameters and the recommended maximum sample mass
that can be injected before sample overload appears as a
problem. Table 2 provides a range of sample size because
the actual amount will depend on the column packing
(packing materials with higher surface area have higher
sample loading), the analyte (larger molecules have lower
loading), and other factors such as sample solubility in
the mobile phase.

TABLE 2
Recommended Maximum Sample
Mass to Inject onto an HPLC Column
Column i.d.(mm)
a)
b)
c)
d)
e)

10.0
4.6
3.0
2.1
1.0

Recommended Maximum Sample

Mass (mg)

35 - 65
4.5 - 9.0
2.0 - 4.0
1.0 - 2.0
0.2 - 0.4

Peak Tailing Caused by Stationary Phase Silanol

Interactions with Amines
A common cause of peak tailing in reversed phase HPLC
is the secondary retention that occurs when an ionexchange interaction takes place between a positively
charged solute (amine) and an acidic silanol on the
surface of silica stationary phase support particles
(Figure 4). It is observed most often when using
HPLC columns packed with stationary phases that have
significant silanol activity. It usually is worse at
neutral pH (6 to 8) than at acidic pH (<3).

Acidic silanols on the surface of silica stationary

phase supports can form ion-exchange sites that
interact with basic compounds. This ion-exchange
interaction will often contribute to peak retention
(secondary retention) and cause peak tailing when
separating amines by reversed phase HPLC.

Acidic or neutral compounds are not affected, and some basic

compounds are more adversely affected than others. If you
think that silanol interaction is the cause of your peak tailing,
there are several steps you can take to correct the problem. First,
make sure that the mobile phase is properly buffered (seeTable
3) and operate at a pH below 3, if possible. Using sufficient
buffer controls the pH and reduces ion-exchange interactions.
Operating at a pH below 3 protonates silanol groups on the silica
stationary phase support (pKa of silanol is 3.5) and thereby
makes the silanols less available for interacting with solutes.

TABLE 3
Commonly Used Buffers for
Reversed Phase HPLC
Buffer

pKa

Buffer Range

UV Cutoff (nm)

Phosphate

2.1
7.2
12.3

1.1 - 3.1
6.2 - 8.2
11.3-13.3

200

Acetate

4.8

3.8 -5.8

210

Citrate

3.1
4.7
5.4

2.1 - 4.1
3.7- 5.7
4.4 - 6.4

230

Tris

8.3

7.3 - 9.3

205

Triethylamine

11.0

10.0 - 12.0

200

Another solution to peak tailing is to add a competing

amine to the mobile phase. Triethylamine (TEA) is
commonly added to mobile phases for this purpose. TEA
interacts strongly with silanols and inhibits them from
interacting with amines in your sample. Figure 5 is an
example of how TEA improves peak shape. About
10mM TEA is sufficient for most applications.

The use of TEA can often be avoided by selecting stationary

phases that have very low silanol activity. Figure 6 ranks some
popular C18 reversed phase columns according to silanol activity.
The ranking in Figure 6 was obtained by measuring the
asymmetry of amitriptyline, an amine that is commonly used to
measure silanol activity of stationary phases. The less tailing
(lower asymmetry value) exhibited by a stationary phase when
running amitriptyline, the less silanol activity that stationary
phase exhibits and the less peak tailing it will have when
separating other basic compounds.

Column: 4.6 x 150 mm, 5 m

Mobile Phase: 80% Methanol+20% 0.05 M Phosphate buffer, pH 7
Flow Rate: 2.0 ml/min
Temperature:24C
Sample:Amitriptyline

Peak Tailing Caused by Adsorption of Acidic

Compounds onto Silica
Although much less common than peak tailing of amines,
acids can sometimes show peak broadening and peak
tailing because of adsorption onto the silica stationary
phase support. To correct peak tailing in these cases
increase the salt concentration of the mobile phase to
suppress secondary interactions, reduce the mobile phase
pH to protonate silanols and solutes and, if necessary,
add a competing acid to the mobile phase (Figure 7).

Peak Tailing Caused by a Void in the Column's

Packing Bed
A void at the head of the HPLC column's packing bed
will cause peaks to be broad and, sometimes, tail.
If, however, you normally get good performance from
your HPLC column and then suddenly you start to see
tailing on all of your peaks, you may have a column
void. The early eluting peaks will be more affected than
late eluting peaks by a column void.

Although some chromatographers will attempt to

repair a column by filling-in the void with a similar
stationary phase material, it has been our
experience that this is seldom worth the effort. The
best cure for a column that is giving tailing peaks
because of a void in the packing bed, is to replace it.
This saves time, money, and frustration.

Optimization of Separation

The most important goal of the chromatographer is

to achieve adequate resolution between all peaks in
the chromatogram in a reasonable amount of time.
A quantitative measure of resolution between two
adjacent chromatographic peaks has been
developed and appears next.

A resolution value of 1.5 or greater between two

adjacent peaks will ensure that the sample
components are well (baseline) separated- to a
degree at which the area or height of each peak
may be accurately measured.

Resolution Values

Figure 11

The previous graphic illustrates the resolution values

for overlapping chromatographic peaks. When two
chromatographic peaks are baseline resolved, the
resolution value is 1.5.
Quantification will not be precise when two adjacent
chromatographic peaks with resolution values of 1.25
or less are involved.
.

During method development, the analyst may wish to

achieve a minimum resolution value of 2 in order to
ensure robust method performance as the column
degrades or in case of minor alterations in experimental
conditions

Resolution can be reduced due to losses in

efficiency (N) or selectivity ().Resolution is
directly related to the capacity factor ( k)

How to control resolution

The fundamental resolution equation , indicates that

resolution degree of resolution between two
chromatographic peaks is dependent upon three
important parameters:

Retention (capacity) factor( k)

Is a means of measuring the retention of an analyte to

chromatographic column

Imad Abu Reid

66

 A high k value indicates that the sample is

highly retained and has spent a significant
amount of time interacting with the stationary
phase.
The retention factor is independent of small
flow rate variations and column dimension

 The most effective and convenient way to alter the

retention factor of a peak is to adjust the solvent
strength of the chromatographic mobile phase.

Imad Abu Reid

68

 Changing the retention factor (k)value

between 1 and 5 has largest impact on
chromatographic resolution.
Above a retention factor of 10 , any increase in
k makes little difference to resolution the
analytical run time will also be prohibitive.
At very low values of k ( 1) it is often difficult
to obtain meaningful separation.

Selectivity (separation ) factor ()

Selectivity (separation ) factor ()

The selectivity factor is the ability of the
chromatographic system to chemically distinguish
between sample components.
High values indicate good separation power and a
good separation between the apex of each peak.
By definition selectivity is always greater than 1. as
when is equal to one , the two peaks are co-eluting.

 As the selectivity of the separation is

dependent upon the chemistry of the analyte ,
mobile and stationary phases, all these factors
may be altered in order to change or optimize
the selectivity of an HPLC separation.

Parameter

Organic solvent

Mobile phase pH

Usage

Changing to a different solvent (e.g.

methanol to acetonitrile)will alter
selectivity in reversed phase
Can affect the degree of ionization of some
analytes affecting their hydrophobicity

Solvent strength
and additives

Can be adjusted to affect selectivity as well

as capacity factor

Stationary phase

One of the most popular ways to alter

separation selectivity

Temperature

Can have an effect with certain analytes in

reversed phase

Organic solvent type

Mobile phase pH

Solvent strength and additives

Stationary phase type

Column temperature

Effects of selectivity on resolution

Efficiency
The efficiency of a chromatographic peak is a
measure of the dispersion of the analyte band
as it travelled through the HPLC system and
column.
In an ideal world , chromatographic peaks
would be pencil thin lines however , due to
dispersion effects the peaks take on their
familiar Guassian shape.

 The plate number (N) is a measure of the peak

dispersion on the HPLC system and column.
Each plate is the distance over which the sample
components achieve one equilibrium between the
stationary and mobile phase. Therefore , the more
(theoretical) plates available within a column ,the more
equilibration possible and the better quality separation.

 The higher the number of plates, the lower

will be the distance between each plate.
Therefore , for high efficiency separation, the
plate number (N) will be high and the plate
height will be (H) low.
H = L/N

How to change efficiency

There are many factors that contribute to the
broadening of the band:
The biggest contributor to band broadening
(and hence low efficiency) is usually the column
Itself:
The quality of the column packing
Particle size
Column dimensions
Presence of void

 As the column length increases the peaks

become narrower ( more efficient).
As the peak efficiency increases the separation
quality increases.
As the column length is increased , the analysis
time increases significantly.
Increasing the column length by an order of
magnitude ( 2.5 to 25 cm) the efficiency of the
peaks also increases by about one order of
magnitude.

Several other factors also need taken into

account:
Injection volume
Dead volume
Flow rate ( reducing flow rate may increase
efficiency)

Doubling column length, doubles efficiency,

double the analysis time but only increase
resolution by a factor of 1.42

The fundamental equation states that resolution

increase for all the bands when N is increased as long
as values of k and do not change. So, if resolution
need to be improved after adjusting k and values , an
increase in N is one option.