R.P.W.Scott Liquid chromatography
-
Upload
indre-gailiute -
Category
Documents
-
view
588 -
download
30
Transcript of R.P.W.Scott Liquid chromatography
This eBook is protected by Copyright law and International Treaties. All rights are reserved. This book is covered by an End User Licensee
Agreement (EULA). The full EULA may be seen at http://www.library4science.com/eula.html.
1
BOOK 3
Chrom-Ed Book Series
Raymond P. W. Scott
LIQUID
CHROMATOGRAPHY
This eBook is protected by Copyright law and International Treaties. All rights are reserved. This book is covered by an End User Licensee
Agreement (EULA). The full EULA may be seen at http://www.library4science.com/eula.html.
2
COPYRIGHT @2003 by LIBRARY4SCIENCE, LLC
ALL RIGHTS RESERVED
Neither this book or any part may be reduced or transmitted in any form
or by any means, electronic or mechanical , including photocopying,
microfilming, and recording or by any information storage and retrieved
system without permission in writing from the publisher except as
permitted by the in-user license agreement.
World Wide Web
http://www.library4science.com/
This eBook is protected by Copyright law and International Treaties. All rights are reserved. This book is covered by an End User Licensee
Agreement (EULA). The full EULA may be seen at http://www.library4science.com/eula.html.
3
Chrom-Ed Book Series
Book 1 Principles and Practice of Chromatography
Book 2 Gas Chromatography
Book 3 Liquid Chromatography
Book 4 Gas Chromatography Detectors
Book 5 Liquid Chromatography Detectors
Book 6 The Plate Theory and Extensions for
Chromatography Columns
Book 7 The Thermodynamics of Chromatography
Book 8 The Mechanism of Retention
Book 9 Dispersion in Chromatography Columns
Book 10 Extra Column Dispersion
Book 11 Capillary Chromatography
Book 12 Preparative Chromatography
Book 13 GC Tandem Systems
Book 14 LC Tandem Systems
Book 15 GC Quantitative Analysis
Book 16 Ion Chromatography
Book 17 Silica Gel and Its Uses in Chromatography
Book 18 Thin Layer Chromatography
Book 19 Chiral Chromatography
Book 20 Bonded Phases
Book 21 Chromatography Applications
COPYRIGHT @2003 by LIBRARYFORSCIENCE, LLC
ALL RIGHTS RESERVED
This eBook is protected by Copyright law and International Treaties. All rights are reserved. This book is covered by an End User Licensee
Agreement (EULA). The full EULA may be seen at http://www.library4science.com/eula.html.
4
Introduction ........................................................................................... 1
The Basic Liquid Chromatograph .......................................................... 2
The Gradient Programmer and the LC Pump ..................................... 4
The Gradient Programmer ............................................................. 4
The LC Pump .................................................................................... 6
The Pneumatic Pump ..................................................................... 6
Non-Return Valves ........................................................................ 8
The Syringe Pump ......................................................................... 9
The Rapid Refill Pump ................................................................ 12
The Twin-Headed Pump. ............................................................. 13
The Diaphragm Pump.................................................................. 14
The Sample Valve ........................................................................... 16
Column Switching ....................................................................... 19
Column Ovens .................................................................................... 20
Detectors ............................................................................................. 21
The UV Detector ............................................................................. 22
The Fixed Wavelength Detector ....................................................... 23
The Multi-Wavelength Detectors ..................................................... 26
The Multi-Wavelength Dispersive Detector ................................. 26
The Diode Array Detector ............................................................... 28
The Electrical Conductivity Detector ................................................... 29
The Fluorescence Detector................................................................... 32
The Refractive Index Detector ............................................................. 34
The Tridet Multi Functional Detector .................................................. 36
Liquid Chromatography Stationary Phases .......................................... 41
Silica Gel ......................................................................................... 42
The Preparation of Spherical Silica Gel ....................................... 43
The Structure of Silica Gel ........................................................... 44
The Thermogravimetric Analysis of Silica Gel ............................. 47
Bonded Phases ................................................................................ 49
The Synthesis of Bonded Phases .................................................. 49
Bonded Phase Synthesis by Reaction in a Solvent........................ 50
The Fluidized Bed Method for Bonded Phase Synthesis .............. 53
Choosing a Bonded Phase ............................................................... 56
Types of Bonded Phase ................................................................... 57
Oligomeric Phases ....................................................................... 58
Bulk Phases ................................................................................. 59
This eBook is protected by Copyright law and International Treaties. All rights are reserved. This book is covered by an End User Licensee
Agreement (EULA). The full EULA may be seen at http://www.library4science.com/eula.html.
5
Interactions Between 'Brush' and 'Bulk' Reverse Phases and
Aqueous Solvents ........................................................................ 60
The Retention Properties of Bulk and Brush Phases. ........................... 63
Macroporous Polymers ........................................................................ 66
LC Mobile Phases ............................................................................... 68
Solvent/Solute Interactions with the Silica Gel Surface .................... 69
Solute Stationary Phase Interactions ............................................ 71
Solvent/Solute Interactions with the Reversed Phase Surface ........... 74
Molecular Interactions in the Mobile Phase ..................................... 76
Aqueous Solvent Mixtures ............................................................... 78
Chiral Stationary Phases ...................................................................... 81
Macrocyclic Glycopeptide Phases .................................................... 83
Cyclodextrin .................................................................................... 89
Liquid Chromatography Applications .................................................. 93
References ......................................................................................... 101
This eBook is protected by Copyright law and International Treaties. All rights are reserved. This book is covered by an End User Licensee
Agreement (EULA). The full EULA may be seen at http://www.library4science.com/eula.html.
1
Introduction
Liquid chromatography (LC) was the first type of chromatography to be
discovered and, in the form of liquid-solid chromatography (LSC), was
originally used in the late 1890s by the Russian botanist, Tswett (1), to
separate and isolate various plant pigments. The colored bands he
produced on the adsorbent bed evoked the term chromatography (color
writing) for this type of separation. Initially the work of Tswett was not
generally accepted, partly due to the original paper being in Russian and
thus, at that time, was not readily available to the majority of western
chemists and partly due to the condemnation of the method by
Willstatter and Stoll (2) in 1913. Willstatter and Stoll repeated Tswett's
experiments without heeding his warning not to use too "aggressive "
adsorbents as these would cause the chlorophylls to decompose. As a
consequence, the experiments of Willstatter et al. failed and their
published results, rejecting the work of Tswett, impeded the recognition
of chromatography as a useful separation technique for nearly 20 years.
In the late 1930s and early 1940s Martin and Synge introduced a form
of liquid-liquid chromatography by supporting the stationary phase, in
this case water, on silica gel in the form of a packed bed and used it to
separate some acetyl amino acids. They published their work in 1941 (3)
and in their paper recommended the replacement of the liquid mobile
phase with a suitable gas, which would accelerate the transfer between
the two phases and provide more efficient separations. Thus, the concept
of gas chromatography was born. In the same paper in 1941, Martin
and Synge suggested the use of small particles and high pressures in LC
to improve the separation that proved to the critical factors that initiated
the development of high performance liquid chromatography.
To quote Martin's original paper,
"Thus, the smallest H.E.T.P. (the highest efficiency) should be
obtainable by using very small particles and a high pressure difference
across the column".
This eBook is protected by Copyright law and International Treaties. All rights are reserved. This book is covered by an End User Licensee
Agreement (EULA). The full EULA may be seen at http://www.library4science.com/eula.html.
2
The statement made by Martin in 1941 contains all the necessary
conditions to realize both the high efficiencies and the high resolution
achieved by modern LC columns. Despite his recommendations,
however, it took nearly fifty years to bring his concepts to fruition.
Activity in the field of liquid chromatography was eclipsed in the 1950s
by the introduction of gas chromatography and serious attempts were
not made to improve LC techniques until the development of GC neared
completion in the mid 1960s. The major impediment to the development
of LC was the lack of a high sensitive detector and it was not until the
refractive index detector was developed by A. Tiselius and D. Claesson
(4) in 1942 could the technique be effectively developed.
Tswett's original LC consisted of a vertical glass tube, a few centimeters
in diameter and about 30 cm high, packed with the adsorbent (calcium
carbonate). The plant extract pigments was placed on the top of the
packing and the mobile phase carefully added to fill the tube. The
solvent percolated through the packing under gravity, developing the
separation, which could be seen as different colored bands at the wall of
the tube. The simple apparatus of Tswett contained all the essentials to
achieve a chromatographic separation.
The contemporary chromatograph, however, is a very complex
instrument operating at pressures up to 10,000 p.s.i providing flow rates
ranging from a few microliters per minute to 10 or 20 ml/minute
depending on the type of LC that is carried out. Modern detectors can
detect solutes at concentration levels of 1x10-9 g/ml and an analysis can
be completed in a few minutes with just a few micrograms of sample.
The Basic Liquid Chromatograph
The basic liquid chromatograph consists of six basic units. The mobile
phase supply system, the pump and programmer, the sample valve, the
column, the detector and finally a means of presenting and processing
the results. A block diagram of the basic liquid chromatograph is shown
in figure 1
This eBook is protected by Copyright law and International Treaties. All rights are reserved. This book is covered by an End User Licensee
Agreement (EULA). The full EULA may be seen at http://www.library4science.com/eula.html.
3
Mobile Phase Supply System
Column and Thermostat
Detector
Data Acquis ition and Processing
Pump and Programmer
Sample Valve
RecorderPrinter
Figure 1. The Basic Liquid Chromatograph
The Mobile Phase Supply System
The mobile phase supply system consists of number of reservoirs (200
ml to 1,000 ml in capacity). At least two reservoirs would be necessary
and are usually constructed of glass or stainless steel and contain an exit
port open to air. Stainless steel, however, is not considered satisfactory
for mobile phases buffered to a low pH and containing certain materials
that can cause corrosion. Each reservoir is usually fitted with a gas
diffuser through which helium can be bubbled. Many solvents and
solvent mixtures (particularly aqueous mixtures) contain significant
amounts of dissolved nitrogen and oxygen from the air. These gasses
This eBook is protected by Copyright law and International Treaties. All rights are reserved. This book is covered by an End User Licensee
Agreement (EULA). The full EULA may be seen at http://www.library4science.com/eula.html.
4
can form bubbles in the chromatographic system that cause both serious
detector noise and loss of column efficiency. As helium is very insoluble
in most solvents, it purges the oxygen and nitrogen from the solvent but
does not produce bubbles in the system itself. Applying a vacuum to the
reservoir is not a permanent solution to dissolved air as, on releasing the
vacuum to allow the solvent to pass to the pump, air again dissolves in
the solvent. The solvent is filtered through a stainless steel or sintered
glass filter to remove any solid contaminants. Depending on the type of
solvent programmer that is employed, the supply from each reservoir
may pass either to a pump or to a valved blending device. Solvent
reservoirs are not usually thermostatted but, when necessary, the solvent
can be brought to the column temperature by the use of an appropriate
heat exchanger. The solvent containers are often situated in an enclosure
that protects the user from toxic solvent vapors such as chloroform or
aromatic hydrocarbons. Such enclosures also isolate the solvents from
atmospheric moisture.
The Gradient Programmer and the LC Pump
The Gradient Programmer
There are two basic types of solvent programmer. In the first, the solvent
mixing occurs at high pressure and in the second the solvents are
premixed at low pressure and then passed to the pump. The high-
pressure programmer is the simplest but most expensive as each solvent
requires its own pump. Theoretically, there can be any number of
solvents involved in a mobile phase program, however, most LC
analyses require only two solvents, nevertheless, up to four solvents can
be accommodated. The layout of a high-pressure gradient system is
shown in figure 2 and includes, as an example, provision for three
solvents to be mixed by appropriate programming.
Solvent passes from each reservoir directly to a pump and then to a
mixing manifold from which it passes to the sample valve and column.
The pumps control the actual program and are usually driven by
stepping motors. The volume delivery of each solvent is controlled by
This eBook is protected by Copyright law and International Treaties. All rights are reserved. This book is covered by an End User Licensee
Agreement (EULA). The full EULA may be seen at http://www.library4science.com/eula.html.
5
the speed of the respective pump, which is precisely determined by the
frequency of its power supply.
Solvent
Reservoir 1
Solvent
Reservoir 2
Solvent
Reservoir 3
Pump 1
Pump 2
Pump 3
To Sample Valve
Mixing Manifold
Programme r
Figure 2. The High Pressure Gradient Programmer
The controlling frequency can be generated either by external oscillators
or, if the chromatograph is computer controlled, directly from the
computer itself.
In a low pressure programmer, the solvent from each reservoir passes to
an oscillating valve, the output from which is connected to a mixing
manifold. A diagram of the layout of a typical low-pressure solvent
programmer is shown in figure 3.
The manifold receives and mixes solvents from each of the programmed
valves. The valves are electrically operated and programmed to open and
close for different periods of time by adjusting the frequency and wave-
form of the supply. Thus, a pre-determined amount of each solvent is
allowed to flow into the manifold. The valves can also be driven either
by oscillators contained in a separate electronic programmer or by the
This eBook is protected by Copyright law and International Treaties. All rights are reserved. This book is covered by an End User Licensee
Agreement (EULA). The full EULA may be seen at http://www.library4science.com/eula.html.
6
chromatograph computer which modifies the wave-form and frequency
to control the flow of each solvent.
Solvent
Reservoir 1
Solvent
Reservoir 2
Solvent
Reservoir 3
Pump
To Sample Valve
Mixing Manifold
Programmer
Valve 3
Valve 2
Valve 1
Figure 3. A Low Pressure Solvent Programmer
The LC Pump
There are a number of different types of pumps that can provide the
necessary pressures and flow-rates required by the modern liquid
chromatograph. In the early years of the LC renaissance, there were two
types of pump in common use; they were the pneumatic pump, where
the necessary high pressures were achieved by pneumatic amplification,
and the syringe pump, which was simply a large, strongly constructed
syringe with a plunger that was driven by a motor. Today the majority of
modern chromatographs are fitted with reciprocating pumps fitted with
either pistons or diaphragms.
The Pneumatic Pump
The pneumatic pump has a much larger flow capacity than the piston
type pumps but, nowadays, is largely used for column packing and not
This eBook is protected by Copyright law and International Treaties. All rights are reserved. This book is covered by an End User Licensee
Agreement (EULA). The full EULA may be seen at http://www.library4science.com/eula.html.
7
for general analysis. The pneumatic pump can provide extremely high
pressures and is relatively inexpensive, but the high-pressure models are
a little cumbersome and, at high flow rates, can consume considerable
quantities of compressed air. A diagram of a pneumatic pump is shown
in figure 4.
Air
Air
Non-Re turn Valve
Non-Re turn Valve
To Sample Valve
From S olvent Reservoir
x
y
(1)
(2)
Figure 4. A Diagram of the Pneumatic Pump
It is seen in figure 4 that the total air pressure on the piston, diameter (y),
is transferred to a piston controlling the liquid pressure, of diameter (x).
Because the radii of the pistons differ, there will be a net pressure
amplification of y2
x2. For example, if the upper piston is 5 cm in
diameter and the lower piston 1 cm in diameter, then the amplification
factor will be 52
12 25. It is seen that the system can provide very high
pressures in a relatively simple manner.
The pump operates in the following manner. Air enters port (1) and
applies a pressure to the upper piston that is directly transmitted to the
This eBook is protected by Copyright law and International Treaties. All rights are reserved. This book is covered by an End User Licensee
Agreement (EULA). The full EULA may be seen at http://www.library4science.com/eula.html.
8
lower piston. If connected to a liquid chromatograph, or the column
packing manifold, liquid will flow out of the left hand side non-return
valve as shown in the diagram which continues until the maximum
movement of the piston is reached. At the extreme of movement a micro
switch is activated and the air pressure transferred from port (1) to port
(2).
The piston now moves in the opposite direction and draws mobile phase
through the right hand non-return valve refilling the cylinder with
solvent. Again on reaching the limit of the piston movement a second
micro switches the air supply from port (2) back to port (1) and the
process repeated.
The refilling process is extremely fast and, if an appropriate pulse
dampener is used, the outlet pressure remains reasonably constant during
the refill cycle.
Non-Return Valves
For efficient function, it is important that while the piston compresses
the solvent to express it from the exit port, the flow is completely
stopped at the inlet port. Conversely, when the pump draws fresh solvent
into the cylinder during refill, the non-return valves must allow solvent
to flow through the inlet valve but, flow-back from the exit valve must
be completely stopped. This is achieved by the use of efficient non-
return valves. Most non-return valves are of similar design and the
construction of a typical valve is shown in figure 5.
The critical part of the valve consists of a synthetic sapphire ball resting
on a seat. The seat may be of stainless steel, captured PTFE or, more
commonly, also of sapphire. When the flow is directed against the ball
the ball moves forward allowing the liquid to flow past it. When the
direction of pressure changes resulting in potential flow-back through
the valve, the ball is driven back onto its seat and stops the flow.
This eBook is protected by Copyright law and International Treaties. All rights are reserved. This book is covered by an End User Licensee
Agreement (EULA). The full EULA may be seen at http://www.library4science.com/eula.html.
9
With careful design and exacting construction these types of valve can
be extremely efficient. In practice, to ensure the most effective
performance, a single non-return valve assembly usually contains two
non-return ball valves connected in series as shown in figure 5.
Ball Valves
Flow Di recti on
Flow Impede d
Valve Se atings
Figure 5. The Design of a Typical Non-Return Valve
The Syringe Pump
The syringe pump is a large, electrically operated simulation of a
hypodermic syringe. Although used in the early days of LC renaissance,
it is rarely used today as, due to its design, it can provide only a limited
pressure and the volume of mobile phase available for use is restricted to
the pump volume. Unless the separation is stopped while the pump is
refilled and the development subsequently continued, the pump can only
elute solutes that have retention volumes equal or less than the pump
capacity. A diagram of a syringe pump is shown in figure 6.
This eBook is protected by Copyright law and International Treaties. All rights are reserved. This book is covered by an End User Licensee
Agreement (EULA). The full EULA may be seen at http://www.library4science.com/eula.html.
10
The pump takes the form of a large metal syringe, the piston being
propelled by an electric motor and driven by a worm gear. The speed of
the motor determines the pump delivery. Another motor actuates the
piston by a different system of gearing to refill the syringe rapidly when
required. The solvent is sucked into the cylinder through a hole in the
center of the piston and between the piston and the outlet there is a coil
that acts as a dampener.
Piston Drive Motor
Fast Re fill Motor
To Col umn
Mobile Phase Reservoir
Drive Control Manual Rewind
Courtesy of the Perkin Elmer Corporation
Figure 6. The Syringe Pump
This type of pump is still occasionally used for the mobile phase supply
to microbore columns that require small volumes of mobile phase to
develop the separation. It is also sometimes used for reagent delivery in
post column derivatization as it can be made to deliver a very constant
reagent supply at very low flow rates.
The Single Piston Reciprocating Pump
The single piston reciprocating pump was the first of its type to be used
with high efficiency LC columns (columns packed with small particles)
This eBook is protected by Copyright law and International Treaties. All rights are reserved. This book is covered by an End User Licensee
Agreement (EULA). The full EULA may be seen at http://www.library4science.com/eula.html.
11
and is still very popular today. It is simple in design and relatively
inexpensive. A diagram of the single piston pump is shown in figure 7.
Most pistons of modern LC pumps are made of synthetic sapphire to
reduce wear and extend the working life of the pump. The cylinder is
usually made of stainless steel and is attached to two non-return valves
in line with the inlet and outlet connections to the pump.
From S olvent Reservoir
Cylinde r
Cam Driving the Piston
Driving Motor
Sapphi re Pi ston
Solvent to Column System
Non-Re turn Valve
Non-Re turn Valve
Re turn Spring
Figure 7. A Single Piston Reciprocating Pump
The piston is driven by a stainless steel cam, which forces the piston into
the cylinder expressing the solvent through the exit non-return valve.
After reaching the maximum movement, the piston follows the cam and
returns as a result of the pressure exerted by the return spring. During
this movement the cylinder is loaded with more solvent through the inlet
non-return valve. The shape of the cam is cut to provide a linear
movement of the piston during expression of the solvent but a sudden
return movement on the refill stroke. In this way the pulse effect that
results from the refill action is reduced.
This eBook is protected by Copyright law and International Treaties. All rights are reserved. This book is covered by an End User Licensee
Agreement (EULA). The full EULA may be seen at http://www.library4science.com/eula.html.
12
The pulses, however, are not completely eliminated and the detector
noise resulting from these pulses is probably the most serious
disadvantage of the single piston pump. Nevertheless, as a result of its
low cost it remains one of the more popular LC pumps.
The Rapid Refill Pump
In order to avoid the refill pulses resulting from a single piston pump, a
number of rapid refill systems have been developed. The designs have
ranged from cleverly designed actuating cams to drive the piston rapidly
in the refill mode to electronically operated piston movements. One
successful approach to this problem is exemplified by the pump design
shown in figure 8.
To Column
To Column
1 ml pe r mi nute
De livery Pi ston Solvent Supply
Pis ton
From S olve nt Supply System
De livery Pi ston
Solvent Supply Pis ton
100 ml per minute
From S olve nt Supply System
Courtesy of Perkin Elmer Inc.
Figure 8. The Rapid Refill Pump
This eBook is protected by Copyright law and International Treaties. All rights are reserved. This book is covered by an End User Licensee
Agreement (EULA). The full EULA may be seen at http://www.library4science.com/eula.html.
13
The pump consists of two cylinders and a single common piston. The
expression of the solvent to the column is shown in the upper part of
figure 8. As the piston progresses to the right, solvent is displaced to the
column system and, simultaneously, fresh solvent is withdrawn from the
solvent reservoir into the right hand chamber. When the piston arrives at
the extent of its travel, a step in the driving cam is reached and the
piston is very rapidly reversed. As a result the contents of the chamber
on the right-hand side are displaced into the left-hand chamber. This
situation is shown in the lower part of figure 8. The transfer rate of the
solvent to the left-hand chamber is 100 times faster than the delivery rate
to the column and consequently reduces the refill-pulse very
significantly. In addition, if a solvent gradient is being used and the
right-hand chamber is being filled with a solvent mixture, excellent
mixing is achieved during the refill of the left-hand chamber. It is clear,
however, that there will not be a smooth transition from one solvent
concentration to the next but will be a step-wise change.
An alternative approach to the elimination or reduction of pump pulses
and one which is probably the more successful (though more expensive)
is the use of twin pump heads. During the operation of a two-headed
pump, one cylinder is filled while the other is delivering solvent to the
column.
The Twin-Headed Pump.
The cylinders and pistons of a two-headed pump are constructed in the
same manner as the single piston pump with sapphire pistons and
stainless steel cylinders fitted with non-return valves to both the inlet
and outlet. The driving cams of both pistons are carefully cut to provide
an increase in flow from one pump while the other pump is being filled.
This compensate for the loss of delivery during the refill process and the
consequent fall in pressure. A diagram of a twin-headed pump is shown
in figure 9. It is seen that there is a common supply of mobile phase
from the solvent reservoir or solvent programmer to both pumps and the
output of each pump joins and the solvent then passes to the sample
valve and then to the column. In the diagram, a single cam drives both
This eBook is protected by Copyright law and International Treaties. All rights are reserved. This book is covered by an End User Licensee
Agreement (EULA). The full EULA may be seen at http://www.library4science.com/eula.html.
14
pistons, but in practice, to minimize pressure pulses, each pump usually
has its own cam drive from the motor.
The cams are carefully designed to produce a virtually pulse-free flow.
The displacement volume of each pump can vary from 20 or 30
microliters to over one milliliter but the usual displacement volume for
the typical pump is about 250 l.
From S olvent Reservoir
Driving Motor
Sapphire Pis ton
Solvent to Column System
Non- Re turn Valve
Re turn Spring
Re turn Spring
Cylinde r
Cam Driving the Piston
Non-Re turn Valve
Sapphire Pis ton
Cylinde r
Non-Re turn Valve
Non-Re turn Valve
Figure 9. The Twin-Headed Pump
A stepping motor drives the pump and thus the delivery depends on the
frequency of the supply fed to the motor. As a consequence, the pump
can have a very wide flow rate range from a few microlitres per minute
to over 10 ml per minute.
The Diaphragm Pump
This eBook is protected by Copyright law and International Treaties. All rights are reserved. This book is covered by an End User Licensee
Agreement (EULA). The full EULA may be seen at http://www.library4science.com/eula.html.
15
The unique property of the reciprocating diaphragm pump is that the
actuating piston does not come into direct contact with the mobile phase
and thus, the demands on the piston-cylinder seal are not so great. The
diaphragm has a relatively high surface area and thus, the movement of
the diaphragm is relatively small and consequently the pump can be
operated at a fairly high frequency. High frequency pumping results in a
very significant reduction in pulse amplitude and, in addition, high
frequency pulses are more readily damped by the column system.
Nevertheless, it must be emphasized that diaphragm pumps are not
pulseless. A diagram of a diaphragm pump, showing its mode of action
is depicted in figure 10
Solvent from Reservoir
Solvent from Reservoir
To Column
To Column
2
34
1
Figure 12. The Action of a Diaphragm Pump
The wheel driving the crank rotates in an anti-clockwise direction and in
position (1) the diaphragm has been withdrawn causing the pumping
cavity behind the diaphragm to be filled with solvent. In position (2), the
piston advances and when it passes the pumping fluid inlet, it starts
This eBook is protected by Copyright law and International Treaties. All rights are reserved. This book is covered by an End User Licensee
Agreement (EULA). The full EULA may be seen at http://www.library4science.com/eula.html.
16
compressing the diaphragm expressing solvent to the column. In
position (3) the diaphragm has been compressed to its limit and the
piston starts to return. In position (4) the piston moves back
withdrawing the diaphragm sucking liquid into the pumping cavity
ready for the next thrust. The inlet from the solvent supply and the outlet
to the column are fitted with non-return valves in the usual manner. Due
to the large volume of the pumping cavity, any gradient profile would be
seriously distorted so this type of pump is not often used for analytical
purposes but is often used in preparative chromatography.
The Sample Valve
In general, LC sample valves must be able to sustain pressures up to
10,000 p.s.i., although they are most likely to operate on a continuous
basis, at pressures of 3,000 p.s.i. or less. The higher the operating
pressure the tighter the valve seating surfaces must be forced together to
eliminate any leak. It follows that any abrasive material, however fine,
that passes into the valve can cause the valve seating to become scored
each time it is rotated which will ultimately lead to leaks. This will cause
the sample size to vary between samples and eventually affect the
accuracy of the analysis. It follows that any solid material must be
carefully removed from any sample before filling the valve. The two
basic types of LC sample valve have been discussed in Book 1. In LC
however, there is a third type of valve which is similar to the external
loop valve but contains an extra loading port and behaves like an
internal loop valve. a diagram of which is shown in figure 13.
This eBook is protected by Copyright law and International Treaties. All rights are reserved. This book is covered by an End User Licensee
Agreement (EULA). The full EULA may be seen at http://www.library4science.com/eula.html.
17
Mobile Phase
Mobile Phase
Sample LoopSample Loop
To ColumnTo Column
Sample In Sample In
To WasteTo Waste
Position A Position B
Ve ntVe nt
Ne edle Port Ne edle Port
Courtesy of Rheodyne Instruments Inc.
Figure 13. The Modified External Loop Sample Valve
The basic difference between this type of valve and the normal external
loop sample valve is the introduction of an extra port at the front of the
valve. This port allows the injection of a sample by a syringe directly
into the front of the sample loop. Position (A) shows the inject position.
Injection in the front port causes the sample to flow into the sample
loop. The tip of the needle passes through the rotor seal and, on
injection, is in direct contact with the ceramic stator face. Note the
needle is chosen so that its diameter is too great to enter the hole.
This eBook is protected by Copyright law and International Treaties. All rights are reserved. This book is covered by an End User Licensee
Agreement (EULA). The full EULA may be seen at http://www.library4science.com/eula.html.
18
Valve Body Rotor
Pre-load Assembly
Connecting Slot
Courtesy of Valco Instruments Inc.
Figure 14. A Simple Form of the LC Sample Valve
After injection, the valve is rotated to position (B) and the mobile phase
flushes the sample directly onto the column. The sample is actually
forced out of the beginning of the loop so it does not have to flow
through the entire length of the loop. This type of injection system is
ideally suited for quantitative LC, and is probably by far the most
popular injection system in use. Sample valves based on this design are
available from a number of manufacturers.
As a consequence of the high pressures that must be tolerated, LC
sample valves are usually made from stainless steel. The exception to the
use of stainless steel will arise in biochemical applications where the
materials of construction may need to be biocompatible. In such cases
the valves may be made from titanium or some other appropriate
biocompatible material. It should be stressed that only those surfaces
that actually come in contact with the sample need to be biocompatible
and the major parts of the valve can still be manufactured from stainless
steel. The actual structure of the valve varies a little from one
manufacturer to another but all are modifications of the basic sample
valve shown in figure 14.
This eBook is protected by Copyright law and International Treaties. All rights are reserved. This book is covered by an End User Licensee
Agreement (EULA). The full EULA may be seen at http://www.library4science.com/eula.html.
19
The valve usually consists of five parts. Firstly there is the control knob
or handle that allows the valve selector to be rotated and thus determines
the load and sample positions. Secondly, a connecting device that
communicates the rotary movement to the rotor. Thirdly the valve body
that contains the different ports necessary to provide connections to the
mobile phase supply, the column, the sample loop if one is available, the
sample injection port and finally a port to waste. Then there is the rotor
that actually selects the mode of operation of the valve and contains slots
that can connect the alternate ports in the valve body to provide loading
and sampling functions. Finally there is a pre-load assembly that
furnishes an adequate pressure between the faces of the rotor and the
valve body to ensure a leak tight seal.
Column Switching
The technique of column switching can increase the versatility of the
liquid chromatograph significantly. An example of a six-port valve
arranged for column switching is given in figure 15. The arrangement
incorporates the same valve as that used for the external loop sampling
system. It is seen that column (1) is connected between ports (5) and (6)
and column (2) is connected between ports (2) and (3). Mobile phase
from a sample valve, or more usually from another column, enters port
(1) and the detector is connected to port (4). In the initial position of the
rotor shown in the diagram on the left hand side, the rotor slots connect
ports (1) and (6), (2) and (3) and (4) and (5).
This results in mobile phase passing from port (1) to port (6), through
column (1) to port (5), from port (5) to port (4) and out to the detector.
Thus, the separation will take place in column (1). The ports connected
to column (2) are themselves connected by the third slot and thus
isolated.
This eBook is protected by Copyright law and International Treaties. All rights are reserved. This book is covered by an End User Licensee
Agreement (EULA). The full EULA may be seen at http://www.library4science.com/eula.html.
20
Column 1
Column 1
Column 2
Column 2
Mobile Phase
Mobile Phase
To Detector To Detector
112 2
3344
5 5
6 6
Courtesy of Valco Instruments Inc.
Figure 16. Valve Arrangement for Column Switching
When the valve is rotated, the situation is depicted on the right hand side
of figure (7); port (1) is connected to port (2), port (3) connected to port
(4) and port (5) connected to port (6). This results in the mobile phase
from either a sample valve or another column entering port (1) passing
to port (2) through column (2) to port (3), then to port (4) and then to the
detector. The ports (5) and (6) are connected, this time isolating column
(1). This arrangement allows either one of two columns to be selected
for an analysis or part of the eluent from another column pass to column
(1) for separation and the rest passed to column (2). This system,
although increasing the complexity of the column system renders the
chromatographic process far more versatile. The number applications
that require such a complex chromatographic arrangement are relatively
small, nevertheless, when required, column switching can provide a
simple solution to certain difficult separation problems.
Column Ovens
The effect of temperature on LC separations is often not nearly so
profound as its effect in GC separations, but can be critical when closely
similar substances are being separated. In LC a change in temperature
will change the free energy of the solute in both phases, (generally in a
This eBook is protected by Copyright law and International Treaties. All rights are reserved. This book is covered by an End User Licensee
Agreement (EULA). The full EULA may be seen at http://www.library4science.com/eula.html.
21
commensurate manner) and so the net change in the free energy
difference with temperature, which controls the magnitude of the
absolute retention, can be relatively small. Its effect on relative
retention, however, can be very significant and, in fact, be the
determining factor in achieving a satisfactory resolution. (5-7) The
effect of temperature on diffusivity will be similar in both GC and LC. An
increase in temperature will increase the diffusivity of the solute in both
phases and thus increase the dispersion due to longitudinal diffusion
and decrease dispersion due to resistance to mass transfer. As a result, at
the optimum velocity, the efficiency of both the LC and GC column will
be largely independent of temperature, however, the optimum velocity
will be higher at higher temperatures and provide the potential for faster
analyses. Due to the lesser effect of temperature on solute retention in
LC (compared to that in GC), temperature is not nearly so critical in
governing absolute retention time but is often essential in achieving
adequate resolution, particularly between closely eluting solutes such as
isomers. In contrast to the GC column, the thermal capacity of an LC
column is much higher as the specific heats of liquids are much greater
than those of a gas. As a consequence, a high heat capacity
thermostatting fluid is necessary and if retention measurements need to
be precise, air ovens would not be ideal for thermostatting LC columns.
On the other hand, liquid thermostatting media are rather messy to use
and tend to make column changing difficult and lengthy. However, if
accurate data is required, good temperature control may be essential. If
precise retention measurements are not required, an air thermostatting
oven might be a reasonable compromise.
Detectors
A large number of LC detectors have been developed over the past thirty
years based on a variety of different sensing principles. However, only
about twelve of them can be used effectively for LC analyses and, of
those twelve, only four are in common use. The four dominant detectors
used in LC analysis are the UV detector (fixed and variable wavelength)
the electrical conductivity detector, the fluorescence detector and the
refractive index detector. These detectors are employed in over 95% of
This eBook is protected by Copyright law and International Treaties. All rights are reserved. This book is covered by an End User Licensee
Agreement (EULA). The full EULA may be seen at http://www.library4science.com/eula.html.
22
all LC analytical applications. These four detectors will be described and
for those readers requiring more information on detectors are referred to
Book 5 on Liquid Chromatography Detectors. The subject of detector
specifications will not be discussed here but will also be dealt with in
detail in Books 4 and 5. Detector sensitivities and detector linearity will,
however, be given for each of the four detectors.
The UV Detector
The UV detector is by far the most popular and useful LC detector that
is available to the analyst at this time. This is particularly true if multi-
wavelength technology is included in this class of detectors. Although
the UV detector has some definite limitations (particularly for the
detection of non polar solutes that do not possess a UV chromaphores) it
has the best combination of sensitivity, linearity, versatility and
reliability of all the LC detectors so far developed .
Most compounds adsorb UV light in the range of 200-350Å including
all substances having one or more double bonds ( electrons) and all
substances that have unshared (non bonded) electrons; e.g. all olefins, all
aromatics and all substances containing >CO, >CS, -N=O and
N N groups. The relationship between the intensity of UV light
transmitted through the detector cell and solute concentration is given by
"Beers' Law,
IT Io ekLc
where, (Io) is the intensity of the light entering the cell,
(IT) is the intensity of the transmitted light,
(L) is the path length of the cell,
(c) is the concentration of the solute,
(k) is the molar extinction coefficient of the solute for the
specific wavelength of the UV light.
or, Ln IT ln Io kLc
This eBook is protected by Copyright law and International Treaties. All rights are reserved. This book is covered by an End User Licensee
Agreement (EULA). The full EULA may be seen at http://www.library4science.com/eula.html.
23
Rearranging,
IT Io 10k'Lc
where (k') is the molar extinction coefficient of the solute.
Differentiating,
LnIT
Io
ckL
It is seen that there are two factors that control the detector sensitivity,
the magnitude of the extinction coefficient of the solute being detected
(which will depend on the wavelength of the UV light that is used) and
the path length of the light passing through the cell. Thus, although the
minimum detectable concentration can be changed by selecting a light
source of different wavelength, the cell length can not be increased
indefinitely to provide higher sensitivity as long cells will provide
excessive peak dispersion with consequent loss of column resolution. It
follows, that the optimum detector cell design involves the determination
of the cell length that will provide the maximum sensitivity and at the
same time constrain detector dispersion to a minimum so that there is
minimum loss in resolution.
The Fixed Wavelength Detector
There are two types of UV detector the fixed wavelength detector and
the multi-wavelength detector. A diagram of a Fixed Wavelength UV
Detector is shown in figure 17.
The detector consists of a small cylindrical cell (2.0 to 10.0 l in volume)
through which flows the eluent from the column. UV light from an
appropriate UV lamp passes through the cell and falls on a UV photo
electric cell. In the fixed wavelength detector the wavelength of the light
depends on the type of lamp that is used. There are a number of lamps
This eBook is protected by Copyright law and International Treaties. All rights are reserved. This book is covered by an End User Licensee
Agreement (EULA). The full EULA may be seen at http://www.library4science.com/eula.html.
24
available the at provide of wavelengths ranging from about 210 nm to
280 nm.
UV Lamp
Quartz Windows
Quartz Windows
Sample Cell
Reference Cell
Lens
Photo Cells
Figure 17. The Fixed Wavelength UV Detector
The common lamps that are commercially available at the time of
writing this book are as follows:-
Lamp Type Emission Wavelengths
Mercury Vapor Lamp 253.7 nm
Zinc Vapor Lamp 2123.9 nm and 307.6 nm
Cadmium Vapor Lamp 228.8, 326.1,340.3, and 346.6 nm
The mercury vapor lamp is the most popular as it has an emission
wavelength that allows the detector to sense a wide range of solute
types. The detector usually contains both a sample and reference cell and
the output from the reference cell is compared to that from the sample
cell The difference is fed to a non linear amplifier that converts the
signal to one that is linearly related to concentration of solute in the
sample cell. The fixed wavelength detector is the least expensive and, as
the majority of the light is emitted at a specific wavelength(s) it has a
This eBook is protected by Copyright law and International Treaties. All rights are reserved. This book is covered by an End User Licensee
Agreement (EULA). The full EULA may be seen at http://www.library4science.com/eula.html.
25
high intensity, and thus, a higher intrinsic sensitivity than the multi-
wavelength UV detectors. However, the multi-wavelength detector can
often compensate for the lower sensitivity by choosing a wavelength that
has the highest extinction coefficient for the solutes of interest. Average
specifications for commercially available fixed wavelength UV detectors
are as follows:-
Typical Specifications for a Fixed Wavelength UV Detector
Sensitivity (toluene) 5x 10-8 g/ml
Linear Dynamic Range 5 x 10-8 to 5 x 10-4 g/ml
Response Index 0.98 - 1.02
Figure 18. The High Speed Separation of a Two Component
Mixture
By the use of very small sensing cells and electronic systems with very
small time constants, the fixed wavelength detector can be designed to
give a very fast response at high sensitivity and very low dispersion and
for this reason it can be used for high speed separations. An example of
the rapid separation of a two component mixture (8) is shown in figure
18. The reason for separating benzene and benzyl alcohol in 2.6
seconds remains (to say the least) obscure and figure 1 is obviously and
This eBook is protected by Copyright law and International Treaties. All rights are reserved. This book is covered by an End User Licensee
Agreement (EULA). The full EULA may be seen at http://www.library4science.com/eula.html.
26
example of "Chromatography Show Biz". Nevertheless, it does
demonstrate that columns can be designed and detectors developed that
can provide extremely fast analyses.
The Multi-Wavelength Detectors
Multi-Wavelength UV detectors utilize a single (perhaps more
accurately a narrow range) of wavelengths to detect the solute. Most
multi wavelength UV detectors can also provide a UV spectrum of the
eluted solute if appropriately arranged. There are two types of multi-
wavelength detectors the dispersion detector that monitors the eluent at
one wavelength only and the diode array detector that monitors the
eluted solute over a range of wavelengths simultaneously. The former
passes the light from a broad emission light source through a
monochrometer, selects a specific wavelength and allows it to pass
through the detecting cell. The second, also uses a broad emission light
source, but all the light is allowed to pass through the sensing cell and
subsequently the light is dispersed by means of a holographic grating
and the dispersed light allowed to fall on an array of diodes.
The Multi-Wavelength Dispersive Detector
A diagram of the Multi-Wavelength Dispersive Detector is shown in
figure 19.
Photo CellLens
Flow
Cel lLens
Plane
Mi rror
Curved
Mi rror
Curved
Mi rror
MONOC HROMATOR
Grating
Curved Mirror
Deuterium
Lamp
This eBook is protected by Copyright law and International Treaties. All rights are reserved. This book is covered by an End User Licensee
Agreement (EULA). The full EULA may be seen at http://www.library4science.com/eula.html.
27
Figure 19. The Multiwave Length Dispersive Detector
Light from a broad wavelength source such as a deuterium or xenon
discharge lamp is collimated by two curved mirrors onto a holographic
diffraction grating. The dispersed light is focused by means of a curved
mirror, on to a plane mirror and light of specific wavelength selected by
appropriately positioning the angle of the plane mirror. Light of the
selected wavelength is then focused by means of a lens through the flow
cell and consequently, through the column eluent. The exit beam from
the cell is focused by another lens onto a photocell which gives a
response that is some function of the intensity of the transmitted light.
The detector is usually fitted with a scanning facility that, by arresting
the flow of mobile phase, allows the spectrum of the solute contained in
the cell to be obtained.
Due to the limited information provided by UV spectra and the
similarity between many spectra of widely different types of compound,
UV spectra are not very reliable for solute identification. The technique
is useful, however, for determining the homogeneity of a peak by
obtaining spectra from a sample on both sides of the peak. The
technique is to normalize both spectra, then either subtract one, from the
other, and show that the difference is close to zero or take the ratio and
show it is constant throughout the peak.
A more common use of the multi-wavelength detector is to select a
wavelength that is characteristically absorbed by a particular component
or components of a mixture. This can be done to either enhance the
sensitivity of the detector to those particular solutes, or render the
detector more specific and consequently, not give a significant response
to other substances in the mixture.
The multi-wavelength dispersive detector more useful than the fixed
wavelength type of UV detector providing adequate sensitivity,
versatility and a linear response. It is however somewhat bulky, due to
the need for an relatively large internal 'optical bench', has mechanically
operated wavelength selection and requires a stop/flow procedure to
This eBook is protected by Copyright law and International Treaties. All rights are reserved. This book is covered by an End User Licensee
Agreement (EULA). The full EULA may be seen at http://www.library4science.com/eula.html.
28
obtain spectra "on-the-fly". The Diode Array Detector has the same
advantages but none of the disadvantages, though, as one might expect,
is somewhat more expensive.
The Diode Array Detector
The diode array detector, although offering detection over a range of UV
wavelength, functions in an entirely different way from that of the
dispersive instrument. A diagram of a diode array detector is shown in
figure 20. Deuterium
Lamp
Achromatic Lens System
Shutter
Flow Cell
Inlet
Outlet
Photo Diode Array
Holographic Grat ing
Figure 20. The Diode Array Detector
Light from a broad emission source such as a deuterium lamp is
collimated by an achromatic lens system so that the total light passes
through the detector cell onto a holographic grating. In this way the
sample is subjected to light of all wavelengths generated by the lamp.
The dispersed light from the grating is allowed to fall on to a diode
array. The array may contain many hundreds of diodes and the output
from each diode is regularly sampled by a computer and stored on a hard
disc. At the end of the run, the output from any diode can be selected
This eBook is protected by Copyright law and International Treaties. All rights are reserved. This book is covered by an End User Licensee
Agreement (EULA). The full EULA may be seen at http://www.library4science.com/eula.html.
29
and a chromatogram produced employing the UV wavelength that was
falling on that particular diode. Most instruments will permit the
monitoring of a least one diode in real time so that the chromatogram
can be followed as the separation develops. This system is ideal in that
by noting the time of a particular peak, a spectrum of the solute can be
obtained by recalling from memory the output of all the diodes at that
particular time. This gives directly the spectrum of the solute, i.e., a
curve relating adsorption against wavelength.
The performance of both types of multi-wavelength detectors are very
similar and typical values for their more important specifications are as
follows.
Typical Specifications for a Multi-Wavelength UV Detector
Sensitivity 1 x 10-7g/ml
Linear Dynamic Range 5 x 10-7 to 5 x 10-4 g/ml
Response Index 0.97 - 1.03
The Electrical Conductivity Detector
The electrical conductivity detector can only detect those substances that
ionize and consequently, is largely used in the analysis of inorganic
acids, bases and salts. It has also found particular use in the detection of
organic acids and bases that are frequently required in environmental
studies and in biotechnology applications. The sensor is the simplest of
all the detectors consisting of only two electrodes situated in a suitable
flow cell.
An example of an electrical conductivity sensing cell is shown in figure
21. It consists of two electrodes situated in a suitable flow cell as
depicted in the upper diagram. The electrodes are arranged to constitute
one arm of a Wheatstone Bridge. When ions enter the detector cell, the
electrical resistance changes and the out of balance signal is fed to a
suitable amplifier. The output from the amplifier is either digitized, and
This eBook is protected by Copyright law and International Treaties. All rights are reserved. This book is covered by an End User Licensee
Agreement (EULA). The full EULA may be seen at http://www.library4science.com/eula.html.
30
the binary number sent to a computer for storage, or the output is passed
directly to a potentiometric recorder. The detector actually measures the
electrical resistance between the electrodes, which by suitable non-linear
amplification, can be made to provide an output that is linearly related to
solute concentration. It is essential that an AC voltage is used across the
electrodes to measure the cell impedance to avoid electrode polarization.
The frequency of the AC potential across the electrodes is usually
around 10 kHz
Electrodes
Inlet Outlet
Conduct ivity Cell
Amplifier
Inlet Outlet
AmplifierPT FE Insulation
Figure 21. An Electrical Conductivity Detector Sensing Cell
A more practical system shown in the lower part of the diagram consists
of short lengths of stainless steel tube insulated from each other by a
PTFE connecting sleeves. For convenience, the first tube (that connected
to the column) is usually grounded (earthed). The resistance between the
inlet tube and the center tube is continuously monitored which will
constitute the resistance across the tiny gap between the tubes contained
in the first PTFE sleeve. The volume of eluent in this gap can be
extremely small and thus, the peak dispersion can also be made very
small. The resistance of the solution situated between the tubes is
inversely proportional to the electric conductivity of the solution, which
in turn, is related to the ion concentration in mobile phase. Some typical
specifications for an electrical conductivity detector are as follows.
This eBook is protected by Copyright law and International Treaties. All rights are reserved. This book is covered by an End User Licensee
Agreement (EULA). The full EULA may be seen at http://www.library4science.com/eula.html.
31
Typical Specifications for an Electrical Conductivity Detector
Sensitivity 5x 10-9 g/ml
Linear Dynamic Range 5 x 10-9 to 1 x 10-6 g/ml
Response Index 0.97 - 1.03
The separation of a mixture of alkali and alkaline earth cations at levels
of a few parts per million, shown in figure 22 gives an example of the
use of the electrical conductivity detector. The cations lithium, sodium,
ammonium, potassium, magnesium and calcium were present in the
original mixture at concentrations of 1, 4, 10, 10, 5 and 10 ppm
respectively.
Courtesy of Dionex Inc
Figure 22. Determination of Alkali and Alkaline Earth Cations
The column used was the IonPacCS12 (a proprietary ion exchange
column) with a mobile phase consisting of a 20 nM methanesulphonic
acid solution in water at a flow rate of 1 ml/min. The sample volume
was 25 l and the separation is also an interesting example of the use of
the ion suppression technique. If the methanesulphonic acid solution
This eBook is protected by Copyright law and International Treaties. All rights are reserved. This book is covered by an End User Licensee
Agreement (EULA). The full EULA may be seen at http://www.library4science.com/eula.html.
32
was passed directly from the column through the detector it would have
a high electrical conductivity and give a large detector base current
which would swamp the signal from the ions being monitored. Thus,
subsequent to the mobile phase leaving the column (and after the
methane sulphonic acid has achieved its purpose in producing the
desired separation) the reagent must be removed to ensure that the
mobile phase entering the detector only contains those ions of interest
and minimal background conductivity. The methane sulphonic acid was
removed by passing the mobile phase through a short reverse phase
column before it entered the detector. The reverse phase will remove all
organic material by adsorption due to the strong dispersive forces that
will occur between the hydrocarbon chains of the reverse phase and the
methyl group of the methanesulphonic acid. The ion suppression column
eventually saturates and require regeneration by desorbing the methane
sulphonic acid with a strong dispersive solvent that is miscible with
water such as acetonitrile. This technique of ion suppression is
frequently used in ion exchange chromatography when using the
electrical conductivity detector. A wide variety of different types of ion
suppression columns are available but it should be pointed out that, any
suppresser system introduced between the column and the detector, will
cause some degree band spreading and consequently reduce the
resolving power of the system. It follows, that the connecting tubes and
suppression column itself must be very carefully designed to eliminate
or reduce this dispersion to an absolute minimum.
The Fluorescence Detector
The fluorescence detector is one of the most sensitive LC detectors and
for this reason is often used for trace analysis. Unfortunately, although
the detector is very sensitive, its response is only linear over a relatively
limited concentration range. In fact, the response of the detector can
only be assumed to be linear over a concentration range of two orders of
magnitude. Unfortunately, the majority of substances do not naturally
fluoresce which is a serious disadvantage to this type of detector. It
follows, that in many instances fluorescent derivatives must be
synthesized to render the substances of interest detectable. There are a
This eBook is protected by Copyright law and International Treaties. All rights are reserved. This book is covered by an End User Licensee
Agreement (EULA). The full EULA may be seen at http://www.library4science.com/eula.html.
33
number of regents that have been developed specifically for this purpose
but derivatizing procedures will be discussed in detail in Book 18. A
diagram of the Fluorescence Detector is shown in figure 23. In its
simplest form, light from a fixed wavelength UV lamp passes through a
cell, through which the column eluent flows and acts as the excitation
source. Any fluorescent light that is emitted is sensed by a photo electric
cell positioned normal to the direction of exciting UV light.
Fluorescent Light
Outlet
Inlet
Fluorescent Cell
UV Lamp
UV Light
Photo Cell
Figure 23. The Fluorescence Detector
The photocell senses fluorescent light of all wavelengths but the
wavelength of the excitation light can only be changed by use of an
alternative lamp. This simple type of fluorescence detector was the first
to be developed; it is relatively inexpensive and for certain compounds
can be extremely sensitive. Typical specifications for a fluorescence
detector are as follows:-
Typical Specifications for a Fluorescence Detector
Sensitivity (Anthracene) 1x 10-9 g/ml
Linear Dynamic Range 1 x 10-9 to 5 x 10-6 g/ml
Response Index 0.96 - 1.04
This eBook is protected by Copyright law and International Treaties. All rights are reserved. This book is covered by an End User Licensee
Agreement (EULA). The full EULA may be seen at http://www.library4science.com/eula.html.
34
A more elaborate form of fluorescence detector uses a monochromator
to select the excitation wavelength and a second monochromator to
select the wavelength of the fluorescent light. This instrument gives the
maximum versatility and allows the maximum sensitivity to be realized
for any type of solute. The system can also provide a fluorescence
spectra by arresting the flow of mobile phase when the solute resides in
the detecting cell and scanning the fluorescent light.
The Refractive Index Detector
The refractive index detector is one of the least sensitive LC detectors. It
is very sensitive to changes in ambient temperature, pressure changes,
flow-rate changes and cannot be used for gradient elution. Despite these
many disadvantages, this detector is extremely useful for detecting those
compounds that are nonionic, do not adsorb in the UV, and do not
fluoresce. There are many optical systems used in refractive index
detectors (9) but one of the most common is the differential refractive
index detector shown diagrammatically in figure 24.
Light SourceMask
Le ns
Sample
Re ference
Amplifi er and
Power S uppl y Zero Adjust
Se nsor
Mi rror
Re corde r
Re ference
Sample
Courtesy of the Waters Chromatography
This eBook is protected by Copyright law and International Treaties. All rights are reserved. This book is covered by an End User Licensee
Agreement (EULA). The full EULA may be seen at http://www.library4science.com/eula.html.
35
Figure 24. The Refractive Index Detector
The refractometer shown in figure 24 monitors the deflection of a light
beam caused by the refractive index difference between the contents of
the sample cell and that of the reference cell. A beam of light (usually
from an incandescent lamp) passes through an optical mask that
confines the beam to the region of the cell. The lens collimates the light
beam, which then passes through both the sample and reference cells to
a plane mirror. The mirror reflects the beam back through the sample
and reference cells to a lens, which focuses it onto a photo cell. The
location of the beam, rather than its intensity, is determined by the
angular deflection of the beam resulting from the refractive index
difference between the contents of the two cells. As the beam changes
its position of focus on the photoelectric cell, the output changes and the
difference signal is electronically modified to provide a signal
proportional to the concentration of solute in the sample cell.
The refractive index detector can often be a 'choice of last resort' and is
selected for those applications where, for one reason or another, all other
detectors are inappropriate or impractical. However, the detector has one
particular and unique area of application and that is in the separation
and analysis of polymers. For those polymers that contain more than ten
monomer units, the refractive index is directly proportional to the
concentration of the polymer and is practically independent of the
molecular weight. A quantitative analysis of a polymer mixture can,
therefore, be obtained by the simple normalization of the peak areas in
the chromatogram (there being no need for the use of individual
response factors). Some typical specifications for the refractive index
detector are as follows:-
Typical Specifications for a Refractive Index Detector
Sensitivity (benzene) 1x 10-6 g/ml
Linear Dynamic Range 1 x 10-6 to 1 x 10-4 g/ml
Response Index 0.97 - 1.03
This eBook is protected by Copyright law and International Treaties. All rights are reserved. This book is covered by an End User Licensee
Agreement (EULA). The full EULA may be seen at http://www.library4science.com/eula.html.
36
A typical application of the RI detector is for carbohydrate analysis.
Carbohydrates do not adsorb in the UV, do not ionize and although
fluorescent derivatives can be made, the procedure is tedious and time
consuming. An example of such an application is shown in figure 25 by
the separation of the products of cyclodextrin hydrolysis.
Courtesy of TOYO SODA Manufacturing Co. LTD
Figure 25. The Separation of Hydrolyzed cyclodextrin
A TSK gel G-Oligo-PW column 7.8 mm I.D.. and 30 cm long was used
for the separation which was carried out at 60˚C and a flow rate of 1
ml/min. The TSK gel packing is a vinyl polymer based material suitable
for separation by size exclusion using aqueous solvents. It is seen that
the products of the hydrolysis are well separated and almost all of the
oligomers are resolved.
The Tridet Multi Functional Detector
This eBook is protected by Copyright law and International Treaties. All rights are reserved. This book is covered by an End User Licensee
Agreement (EULA). The full EULA may be seen at http://www.library4science.com/eula.html.
37
The popularity of the UV detector, the electrical conductivity detector
and the fluorescence detector motivated Schmidt and Scott (10,11) to
develop a trifunctional detector that detected solutes by all three method
simultaneously in a single low volume cell. A diagram of their detector
is shown in figure 26. The UV adsorption system consists of a low-
pressure mercury lamp (major emission at 254 nm) and a solid state
photo cell with quartz windows so it responds to light in the UV region.
The cell is 3 mm long and is terminated at on end by a cylindrical quartz
window and at the other by a quartz lens.
UV Lamp
Fluorescence Photo Cell UV Adsorpti on
Photo Cell
Quartz Lens
Quartz Lens
Flow Ce ll
Inl et
Outlet
Conduit Tubes also Se rving as Electrodes
Re taini ng Springs
Figure 26. The Trifunctional Detector
The lens focuses the transmitted light on to the photocell. There are two
stainless steel discs separated by a 1 cm length of Pyrex tube adjacent to
the quartz windows. Mobile phase enters and leaves the detector cell
through radial holes in the periphery of the stainless steel discs. The steel
discs act as the electrodes for conductivity detection. Normal to the
Pyrex tube is another photocell that receives fluorescent light emitted at
This eBook is protected by Copyright law and International Treaties. All rights are reserved. This book is covered by an End User Licensee
Agreement (EULA). The full EULA may be seen at http://www.library4science.com/eula.html.
38
right angles to the UV excitation light. The output from each sensor is
processed by an appropriate amplifier to provide an output that is
linearly related solute concentration.
The column eluent is, thus, continuously and simultaneously monitored
by UV adsorption, fluorescence and electrical conductivity. This
detector is very versatile and is supplied with a column, sample valve,
pump and recorder. Chromatograms showing the simultaneous use of all
three detector functions are depicted in figure 27.
UV Adsorption
Function
Electrical.Conductivity
Function
Fluorescence
Function
The column used was a Pecosphere™ 3 mm in diameter and 3 cm long carrying a
C18 stationary phase. The mobile phase was a mixture of methanol(75%) and water
(25%) at a flow rate of 2 ml/min. The solutes were 1 benzene, 2 toluene, 3 ethyl
benzene, 4 isopropyl benzene, 5 t-butylbenzene, 6 anthracene, and 7 sodium
chloride.
Courtesy of Bacharach Inc.
Figure 27. Chromatograms Demonstrating the Simultaneous
Monitoring of a Mixture by all Three Detector Functions
This eBook is protected by Copyright law and International Treaties. All rights are reserved. This book is covered by an End User Licensee
Agreement (EULA). The full EULA may be seen at http://www.library4science.com/eula.html.
39
It is seen that the anthracene is clearly picked out from the mixture of
aromatics by the fluorescence detector and the chloride ion, not shown
at all by the UV adsorption or fluorescence detectors, is monitored by
the electrical conductivity detector. The simultaneous use of all detector
functions make this detector very useful but, the real advantage of the
trifunctional detector is that it allows the analyst a choice of the three
most useful detector function in one detecting system. In addition, any
of the three functions can be chosen at the touch of a switch and without
any changes in hardware. An example of the use of the three individual
detector function in the analysis of three quite different types of sample
demonstrates this advantage.
Column-PecoSphere Column Pecosphere Column Pecosphere 3 Size 3 mm x 3 cm C18 Size 4.6 mm x 15 cm C18 Size 3 mm x 3 cm C18 17% Methanol/Water 90% Acetonitrile/Water 1nMtetrabutyl-
ammonium hydroxide and buffer
Flow-Rate 3.0 ml/min. Flow-Rate 2 ml/min. 1.5 ml/min.
UV Detector Fluorescence Detector Conductivity Detector
Sample Composition
1 Theobromine 1 Naphthalene 1 Solvent peak
2 Theophyline 2 Fluorene 2 Chloride ions
3 hydroxyethyltheophyline 3 Acenaphthene 3 Nitrite ions
4 Caffeine 4 Phenanthrene 4 Bromide ions
5 Anthracene 5 Nitrate ions
6 Fluoranthene 6 Phosphate ions
7 pyrene 7 Phosphite ions
This eBook is protected by Copyright law and International Treaties. All rights are reserved. This book is covered by an End User Licensee
Agreement (EULA). The full EULA may be seen at http://www.library4science.com/eula.html.
40
8 Benzo(a)anthracene 8 Sulfate ions
9 Chrysene 9 Iodide ions
10 Benzo(b) Fluoranthene
11 Benzo(k) Fluoranthene
12 Benzo(k) Fluorantmene
13 Dibenz(a,h)anthracene
14 Idenol(1,2,3-cd))pyrene
15 Benzo(ghi)perylene
Figure 28. Chromatograms from the Trifunctional Detector
Data Acquisition and Processing
The data acquisition and processing system has been discussed in Book
1 and Book 2 and is dealt with in further detail in Books 4 and 5.
However, a general outline of the data system layout is included here
and is given in figure 29 in the form of a block diagram that shows the
individual steps employed.
Detector Output
Scaling Amplifier
Analogue to Digital Converter
Interface
Computer
Presentation of Results
Figure 29. A Typical Data Acquisition and Processing System
The output from the detector, usually in millivolts, is passed to a scaling
amplifier that converts the signal to a voltage that is acceptable to the
This eBook is protected by Copyright law and International Treaties. All rights are reserved. This book is covered by an End User Licensee
Agreement (EULA). The full EULA may be seen at http://www.library4science.com/eula.html.
41
analog to digital (A/D) converter The A/D converter changes the voltage
output to a binary number which is temporarily stored in a register. This
process is continuously repeated at a defined rate, called the 'sampling
rate'. The current binary number, stored in the register is regularly
sampled by the computer and stored (usually on hard disk). On
completion of the analysis the computer accesses all the data from store,
calculates the retention report, compares peak heights or peak areas to
provide the quantitative analysis according to the processing program
that is used and finally prints out the results in tabulated form.
Modern data processing software often includes routines that can
process chromatograms where the components of the sample are
incompletely resolved. The routines deconvolute the individual peaks
from the composite envelope and calculate the area of the individual de-
convoluted peaks. Such algorithms can be used very effectively on peaks
that are entrained in the tail of a major peak but are not so accurate for
composite envelopes containing many unresolved peaks.
It should be emphasized that clever algorithms or subroutines are a
poor substitute for good chromatography.
The chromatographic system should, wherever possible, be optimized to
obtain complete resolution of the mixture and not place too much
reliance on mathematical techniques to aid in the analysis.
Liquid Chromatography Stationary Phases
Traditionally the stationary phase used in LC has been silica gel, which
separates solutes largely on the basis of polarity, although, due to its
unique structure, silica gel also exhibits strong exclusion characteristics.
The bonded phases were introduced to provide a material that would
separate solutes by dispersive interactions and also to provide some semi
polar stationary phases. The bonded phases were also based on silica
gel. Subsequently, polymeric stationary phases were introduced to
This eBook is protected by Copyright law and International Treaties. All rights are reserved. This book is covered by an End User Licensee
Agreement (EULA). The full EULA may be seen at http://www.library4science.com/eula.html.
42
provide materials that were insoluble in water and that were stable at
extremes of pH.
Silica Gel
Silica gel is manufactured by releasing silicic acid from a strong solution
of sodium silicate by hydrochloric acid. (Sodium silicate is prepared by
heating sand at a high temperature in contact with caustic soda or
sodium carbonate).
Initially, silicic acid is released,
Na2SiO3 +H2O + 2HCl = Si(OH)4 + 2NaCl
and then the free acid quickly starts to condense with itself with the
elimination of water to form dimers, trimers and eventually polymeric
silicic acid. The polymer grows, initially forming polymer aggregates
and then polymer spheres, a few Angstrom in diameter. These polymeric
spheres are called primary silica particles. These primary particles
continue to grow until, at a particular size, the surface silanol groups on
adjacent primary polymer particles, condense with the elimination of
water. This condensation causes the primary particles to adhere to one
another and at this stage the solution begins to gel. During this process,
the primary particles of silica gel will have diameters ranging from a few
Angstrom to many thousands of Angstrom depending on the conditions
of formation.
The size of the primary particles depend, among other factors, on the
conditions of synthesis, (e.g., reaction temperature, the pH of the
mixture at the time of gelling and even the subsequent treatment of the
gel, including the conditions of washing). The product at this stage is
called a 'hydrogel'.
If a neutralized silicate solution is allowed to age with gentle stirring to
prevent gel formation, Ostwald Ripening occurs, causing the larger
primary particles to grow in size at the expense of the formation of
smaller particles. In general, the aging and ripening of silica solutions
This eBook is protected by Copyright law and International Treaties. All rights are reserved. This book is covered by an End User Licensee
Agreement (EULA). The full EULA may be seen at http://www.library4science.com/eula.html.
43
will increase the size of the larger particles, which, in turn, will decrease
the surface area of the silica and increase its porosity.
It is apparent that the formation of the primary particles of condensed
silica, and their subsequent fusion by gelling, that confers on silica gel
its high porosity and high surface area, which are so important in its use
as a stationary phase in LC. Furthermore, as has already been stated, it is
the condensation of the surface silanol groups of the primary particles
that causes their adhesion and the onset of gel formation and,
consequently, the mechanical strength of the gel.
After the gel has first set, it is a very soft and is usually transferred to
vats or trays where it is then allowed to stand for a number of days.
During this period, condensation between the primary particles
continues to take place and the gel shrinks and exudes salinated water.
This shrinking process accompanied by saline elimination is called
sinerisis and eventually a firm, almost rigid gel is produced which is still
called the hydrogel. The compact hydrogel is then well washed under
controlled conditions to eliminate the last of the sodium chloride and
then heated for a few hours at 120oC. The resulting product is a hard
amorphous mass called the xerogel. The xerogel, ground and graded is
the material that is used for packing LC columns and manufacturing
bonded phases. The product, prepared in this way, is called irregular
silica gel, to differentiate it from spherical silica gel, which is prepared
by employing an entirely different synthetic procedure. Irregular silica
gel is one of the basic materials from which bonded phases can be
prepared.
The Preparation of Spherical Silica Gel
Spherical particles of silica can be prepared by spraying a neutralized
silicate solution (the colloidal silica sol.) into fine droplets before gelling
has taken place and subsequently drying the droplets in a stream of hot
air. It has also been shown possible to disperse a silica sol in the form of
an emulsion in a suitable organic solvent where the droplets gel in
spherical form (12). Unfortunately, details for the preparation of
This eBook is protected by Copyright law and International Treaties. All rights are reserved. This book is covered by an End User Licensee
Agreement (EULA). The full EULA may be seen at http://www.library4science.com/eula.html.
44
spherical silica have tended to be kept very confidential for commercial
reasons and so information is a little sparse. One of the first methods
reported was that of Le Page et al (13,14). A stable silica sol. (generated
at low pH so that gelling does not take place) is passed through a non
aqueous solvent in such a manner as to produce droplets. These droplets
rapidly solidify and are then filtered off, dried, and heated to 400˚C to
800˚C to form a rigid xerogel. The structure of the resulting silica is
strongly affected by the alkali metal content and the calcining
temperature. The higher the temperature employed the lower the surface
area and pore volume of the resulting silica.
An alternative method devised and patented by Unger (15) called the
polyethoxy silane procedure involved a two-stage process. Firstly
tetraethoxy-silane is partially hydrolyzed to polyethoxysiloxane, a
viscous liquid, which is then emulsified in an ethanol water mixture by
vigorous stirring. The stirring produces spheres of polyethoxysiloxane
which by hydrolytic condensation, initiated by a catalyst, are changed to
silica hydrogel. The hydrogel spheres are then washed and converted to
the xerogel by heating. Spherical silica gel is readily available
commercially, both in the form of silica and as different types of bonded
phase. Today the majority of silica-based LC column packings are
spherical although with modern methods of grinding, the so-called
irregular particles are more rounded and the difference between the
performance of spherical and irregular particles is less clearly defined.
The Structure of Silica Gel
The matrix of the primary silica gel particle consists of a core of silicon
atoms joined together with oxygen atoms by siloxane bonds (silicon-
oxygen-silicon bonds). On the surface of each primary particle some
residual, uncondensed hydroxyl groups from the original polymeric
silicic acid remain. These residual hydroxyl groups confer upon silica
gel its polar properties. These hydroxyl groups react with the silane
reagents to form bonded phases. The silica surface is quite complex and
contains more than one type of hydroxyl group, strongly bound or
'chemically' adsorbed water and loosely bound or 'physically adsorbed'
This eBook is protected by Copyright law and International Treaties. All rights are reserved. This book is covered by an End User Licensee
Agreement (EULA). The full EULA may be seen at http://www.library4science.com/eula.html.
45
water. There are three types of hydroxyl group. The first is a single
hydroxyl group attached to a silicon atom, which has three siloxane
bonds joining it to the gel matrix. The second is one of two hydroxyl
groups attached to the same silicon atom, which, in turn, is joined to the
matrix by only two siloxane bonds. These twin hydroxyl groups are
called Geminal hydroxyl groups. The third is one of three hydroxyl
groups attached to a silicon atom, which is now only joined to the silica
matrix by only a single siloxane bond. An example of each type of
hydroxyl bond is shown in Figure 30.
Employing NMR techniques Sindorf and Maciel, (16,17) has shown
that the single hydroxyl group is likely to be the most prolific. The next
most common is the geminal hydroxyl groups followed by the tertiary
hydroxyl group.
O
O
H
Si O H
O
H
O
Si
SiSiSi
Si
OO
OO
O O
O SiO
H H
O
H
O H
S i ngle S i lanol Group
Two Sil anol Groups (Geminal Groups)
Three S il anol Groups
Figure 30. Different Forms of Hydroxyl Group that can Occur on
the Surface of Silica Gel.
The silica surface, however, has additional complexities. Water can be
hydrogen bonded to the hydroxyl groups and multi-layers of water
physically adsorbed on top of these. Water can be hydrogen bonded to
the silica gel surface in a number of different ways, which are depicted
in figure 31.
This eBook is protected by Copyright law and International Treaties. All rights are reserved. This book is covered by an End User Licensee
Agreement (EULA). The full EULA may be seen at http://www.library4science.com/eula.html.
46
Si Si Si Si SiSi
O O O O O
O O O O O O
H H H H H
OH
H
H H
O
H
O
H
Figure 31. Different Ways in Which Water May be Hydrogen
Bonded to Silica Gel Hydroxyl Groups
None of the above structures has been confirmed in an unambiguous
manner but all are reasonably possible. The center and right hand side
structures contain a type of double hydrogen bond and would have high
energies of formation and, thus, more stable than the simple hydrogen
bond depicted on the left. The right hand structure would be particularly
stable as it constitutes a four membered hydrogen bonded ring similar to
that, which might be expected to form in the strong association of water
with itself.
Silica gel adsorbs relatively large quantities of water, which was
explained on the basis of multi-layer adsorption. This concept was
supported by Vleeskens (18,19) and experimentally validated by
gravimetric measurements (20). An example of one type of multi-layer
adsorption is shown in figure 32.
This eBook is protected by Copyright law and International Treaties. All rights are reserved. This book is covered by an End User Licensee
Agreement (EULA). The full EULA may be seen at http://www.library4science.com/eula.html.
47
SiSi
O
O O
H
OH
H
OH
H
OH
H
H
Figure 32. Multi-Layers of Physically Adsorbed Water
The multi-layer adsorption depicted in figure 32 is much over
simplified, as adsorption could also take place on the surface of siloxane
bonds as well.
Probably the most informative experimental procedure to help in the
elucidation of the structure of the silica surface is thermogravimetric
analysis (TGA). Formally, the silica was heated for known times at
known temperatures and the loss in weight noted (21). With the advent
of the programmed TGA apparatus this procedure has been simplified a
great deal and the basic thermogram can now provide considerable
information on the nature of the surface hydroxyl groups and adsorbed
water (22).
The Thermogravimetric Analysis of Silica Gel
The sample is suspended from the arm of a continuously recording
micro balance in a temperature controlled furnace and is heated from a
defined starting temperature to a specified final temperature at a
designated heating rate usually given as temperature change per unit
time. The temperature and the sample weight are continuously digitized
and the data stored. The results can then be printed out or presented as
This eBook is protected by Copyright law and International Treaties. All rights are reserved. This book is covered by an End User Licensee
Agreement (EULA). The full EULA may be seen at http://www.library4science.com/eula.html.
48
an appropriate graph relating sample mass to temperature. To help
identify the desorption of different species, derivative curves can also be
produced. The results obtained from a sample of Matrex 20 LC silica
gel taken directly from the Perkin Elmer TGA instrument is shown in
figure 33.
Figure 33. Thermogram of Silica Gel
The derivative curve is seen to show three distinctly different desorption
processes. The first takes place from about 30oC to 130oC; the second
between about 200oC and 450oC and the third between about 400oC
and 900oC. The three different desorption processes are distinct and
unambiguous and are similar to those previously identified by Scott and
Traiman (22). The total loss from the sample was about 5%w/w but it
would appear from the TGA curve that the desorption may not be
entirely complete at the temperature of 900oC.
The curve relating mass of water lost (obtained by subtracting each data
point from the initial total mass of sample) against temperature in a
TGA analysis, was a type of desorption isotherm that could be described
by assuming three distinct and separate desorbable species on the silica
surface, each evolving water over a specific temperature range.
This eBook is protected by Copyright law and International Treaties. All rights are reserved. This book is covered by an End User Licensee
Agreement (EULA). The full EULA may be seen at http://www.library4science.com/eula.html.
49
The first loss of water between 50˚C and 150˚C can be attributed to
physically adsorbed water. The second source that is lost between 150˚C
and 400˚C appears to be strongly held (probably hydrogen bonded)
water. The third and last loss appears to come from the condensation of
the silanol groups to siloxane bonds with the elimination of water. The
amount of water lost during the three desorption processes shown as
desorption curves in figure 34.
Bonded Phases
Bonded phases are formed by reacting the surface hydroxyl groups with
an appropriate reagent to chemical link an organic moiety to the silica
surface. The nature of the organic moiety will determine the type of
interaction that will take place between the solute and the surface. If the
moiety is a hydrocarbon chain, then the interaction will be dispersive, if,
for example, it contains a cyano group, it will be polar and if it carries a
group that can dissociate into ions, the interactions will be ionic.
0 200 400 600 800 1000
0.0
0.2
0.4
0.6
0.8
TE MPERATURE (C)
Wt. Loss 1 Calc
Wt. Loss 2 Calc
Wt. Loss 3 Calc
Figure 34. Graph of Calculated Weight Loss/Temperature
for Each Desorption Stage
The Synthesis of Bonded Phases
This eBook is protected by Copyright law and International Treaties. All rights are reserved. This book is covered by an End User Licensee
Agreement (EULA). The full EULA may be seen at http://www.library4science.com/eula.html.
50
The particle size will be determined by the nature of the separation and
will vary with the complexity of the mixture and the instrument
characteristics. For example a difficult separation will require high
efficiencies and thus a small particle diameter (e.g. 3 m). If however,
the available column pressure is limited then the particle diameter made
need to be relatively large (e.g. 10 m) to provide adequate flow rate.
The choice of the pore size and surface area of the silica is more
complex. The surface area of silica tends to vary inversely as the pore
size, so the larger the pore size the smaller the surface area. Thus, when
one is defined the other, to some extent, is also specified. The most
efficient bonded phase has the maximum surface coverage. It is
understood, that due to stearic hindrance from the bonded moiety itself,
only a proportion of the silanol groups can be bonded and there is little
that can be done to avoid this problem. However, there are other reasons
for incomplete silanization of the silica. Incomplete silanization can
result from the reagent molecule being excluded from the smaller pores
of the silica. Exclusion can be a particular problem when bonding
relatively large molecular weight materials such as long chain
hydrocarbons onto the silica surface. It is therefore, important to choose
a silica gel that has a relatively large pore size (e.g., a mean pore
diameter of 150Å), which may limit the surface area to between 150 and
250 sq.m per gram and thus, reduce the retentive capacity of the
stationary phase. Nevertheless, this will ensure that the vast majority of
the pores will be accessible to the silanizing reagent. Care must be taken
to ensure that the pore size is not chosen to be so large that the silica gel
is mechanically weak and collapses under pressure during packing or
even during normal use. An alternative approach is to subject the silica
to hydrothermal treatment. Hydrothermal treatment removes silica from
the surface of large pores and deposits it on the smaller pores, which
eventually become blocked and thus, impermeable. This would remove
the small pores, and consequently possible sites for future erosion and
make the large pores even larger and more accessible to the silanizing
reagents.
Bonded Phase Synthesis by Reaction in a Solvent.
This eBook is protected by Copyright law and International Treaties. All rights are reserved. This book is covered by an End User Licensee
Agreement (EULA). The full EULA may be seen at http://www.library4science.com/eula.html.
51
The solvents normally used in bonded phase synthesis are aromatic
hydrocarbons e.g., toluene that boils at 110˚C or mixed xylenes that boil
138-140˚C. The procedure varies a little depending on the size of the
batch and the type of silanizing reagent. An example of a laboratory
scale synthesis using a chlorosilane is as follows.
10 g of the chosen silica is dried at 250˚C for about 2 hours and
dispersed in a flask containing 100 ml of toluene dried over sodium. If a
mono-chlorosilane reagent is used a small trace of water in the toluene
can be tolerated and can be eventually eliminated by the use of excess
chlorosilane reagent. Under such circumstances the toluene need not be
dried over sodium. If, a dichlorosilane is used however, for example in
the initial step in the synthesis of an oligomeric phase, the presence of
water may cause linear polymerization. Consequently, stringent
precautions must be taken to eliminate all traces of water. A slight
excess of the chlorosilane is then added to the silica dispersion together
with 5 ml of pyridine. The pyridine acts as scavenger for the
hydrochloric acid released during the reaction. The mixture is refluxed
for about 5 hours and the product is then filtered on a sintered glass
filter, washed sequentially with toluene, tetrahydrofuran (THF),
methanol, methanol water (50:50 v/v) and finally with methanol and
dried under suction. The bonded phase now needs end-capping; that is,
any unreacted silanol groups are treated with a small molecular weight
silanizing reagent to react with those hydroxyl groups that were
stearically unavailable to the larger reagent due to exclusion. To end-cap
the product, the bonded phase is refluxed for two hours in a mixture of
100 ml of toluene and 25 ml of hexamethydisilazane. The product is
again filtered free of the reaction liquid mixture and washed sequentially
with, toluene tetrahydrofuran, methanol, methanol water (50:50 v/v) and
finally with methanol and then dried under vacuum. It should be pointed
out that end-capping cannot eliminate the hydroxyl groups that are
stearically hindered by the bonded moiety, or at best, only a small
proportion of them will be removed. It can, however, react with any
readily available hydroxyl groups, particularly those contained in pores
that the original reagent could not enter but to which the smaller capping
This eBook is protected by Copyright law and International Treaties. All rights are reserved. This book is covered by an End User Licensee
Agreement (EULA). The full EULA may be seen at http://www.library4science.com/eula.html.
52
reagent has access. Capping can also eliminate any hydroxyl groups
attached to the bonded moiety resulting from the presence of dichloro or
tricoloro impurities in the silanizing reagent. After capping, the carbon
content of the product can then be determined to estimate the extent of
reaction.
The carbon content of a bonded phase is often used to determine the
efficacy of bonding, but its value must be used in conjunction with a
knowledge of the surface area of the native silica in order to arrive at a
meaningful conclusion. The amount of material bonded to the silica will
depend, not only on the efficiency of the reaction, but also on the
number of silanol groups that were available with which it could react;
ipso facto it will also depend on the surface area of the parent silica. The
carbon content of the bonded phase is usually determined by micro-
analysis and the result expressed as %w/w of the combined bonded
organic material and the silica gel. Consider a bonded phase where the
carbon content is (y)%w/w coated with a hydrocarbon moiety having (n)
carbon atoms per aliphatic chain (e.g., for the dimethyl octyl brush
phase, n=10). Then, the concentration of aliphatic chains in mols. per
gram of silica (m')will be,
m'
y
100
1
12 n
Consequently, if the surface area of the silica is (A) m2g-1, the number
of mols of aliphatic chains (m) in micromols per square meter will be
given by,
m
y
100
1
12 n
106
A
or, m
833 y
nA (1)
The amount of material bonded per square meter, which indicates the
efficacy of the synthetic process, will be directly proportional to the
carbon content of the product and inversely proportional to the surface
area of the original silica gel.
This eBook is protected by Copyright law and International Treaties. All rights are reserved. This book is covered by an End User Licensee
Agreement (EULA). The full EULA may be seen at http://www.library4science.com/eula.html.
53
The method of synthesis is very similar for the alkoxysilane reagents.
The same solvents can be used but, as no hydrochloric acid is generated,
there is no need to have pyridine present as a scavenger. The most
reactive alkoxy reagents are the methoxy and ethoxysilanes and their
reaction with a hydroxyl group is accompanied by the release of
methanol or ethanol.
Si CnH2n+ 1(CH3)2
CnH2n+ 1(CH3)2 Si 0
Si O
+HOSi CH3
CH3OH+
The reaction is best carried out in a distillation flask and the methanol
removed as it is formed, the heating rate is adjusted to ensure that the
aromatic solvent is not removed as well. The reaction is allowed to
proceed for about 5 hours and the product then filtered through a
sintered glass filter and washed with the same sequence of solvents as
those used in the chlorosilane synthesis. The product is then refluxed
with the THF/water mixture, filtered and again washed with the
appropriate solvents and dried. The final capping process is also the
same as that employed in the method using the chlorosilanes reagents,
utilizing hexamethyldisilazane as the capping reagent. The alkoxy-
silanes are almost as readily available as the chlorosilanes and are easier
and more pleasant to handle. They are, however, just as hygroscopic as
the chlorosilanes and must be kept under the same anhydrous
conditions. Approximately the same yields can be expected and the
same degree of bonding.
The Fluidized Bed Method for Bonded Phase Synthesis
The concept of synthesizing bonded phases by means of a fluidized bed
reactor was first suggested by Unger (23). The term fluidization
This eBook is protected by Copyright law and International Treaties. All rights are reserved. This book is covered by an End User Licensee
Agreement (EULA). The full EULA may be seen at http://www.library4science.com/eula.html.
54
describes a contact process in which particulate matter is transformed
into a fluid-like state by a stream of gas or liquid. This state is achieved
by the upward passage of a fluid through the bed of particles to a
velocity at which the drag force acting on the surface of the particles is
equal to the gravitational force downwards. At this point the particles
move apart and become suspended by the fluid flow and the bed is said
to be fluidized.
The fluidized bed apparatus can be operated in such a way that it
allowed far more complex multi-stage syntheses to be carried out rapidly
and with relative ease; e.g., the synthesis of oligomeric phases. The
fluidized bed apparatus is shown in figure 34. The fluidized bed is
enclosed in a tube about 25 cm long, 4.5 cm in diameter situated in a
heating jacket. The temperature of the bed is monitored by three
thermocouples placed in the center and at either end of the bed. A
condenser is situated at the top of the bed, which returns unreacted
silanizing reagent to the vapor generator. The vapor generator consists of
a simple boiling flask that can be provided with a nitrogen stream if the
silanizing reagent does not boil at reasonable temperatures. The
silanizing reagent vapor passes from the vaporizer to a preheater and
then to the base of the fluidizer.
This eBook is protected by Copyright law and International Treaties. All rights are reserved. This book is covered by an End User Licensee
Agreement (EULA). The full EULA may be seen at http://www.library4science.com/eula.html.
55
Condenser
To Waste
Thermocoup les
Heater Jacket
Silica Gel
Preheater
Sily l Chloride Vaporiz er
St eam Generat er
Valves
Valve
Nit rogen
Flow Meter
Coolant
Coolant Exit
Unreact ed Reagent Return
Figure 34. The Fluidized Bed Apparatus for the Synthesis of
Bonded Phases
A steam generator, similar in form to the silanizing reagent vaporizer, is
also attached to the preheater and consequently, can supply steam to the
fluidized bed if, and when required. This means that the fluidized bed
This eBook is protected by Copyright law and International Treaties. All rights are reserved. This book is covered by an End User Licensee
Agreement (EULA). The full EULA may be seen at http://www.library4science.com/eula.html.
56
can be used to hydrothermally treat the silica prior to silanization. The
nitrogen stream is continuously monitored by a flow meter. All vapor
and nitrogen streams are adequately fitted with valves to provide the
maximum operating flexibility. The fluidizer is normally loaded with
about 25 g of silica (which provides a fluidized bed about 5 cm long) and
heated to 200oC for about 3 hr. in a fluidizing stream of preheated
nitrogen.
The silanizing reagent is then heated to its boiling point in the vaporizer
or to a temperature where it has a significant vapor pressure and the
reagent blown through the fluidized bed by a stream of nitrogen. Excess,
unreacted reagent is condensed and passed directly back to the
vaporizer. The reaction is allowed to proceed for about 6 hours and then
the reagent vapor flow stopped and replaced by a stream of pure
nitrogen. When reagent is no longer returning to the vaporizer, the bed is
allowed to cool to room temperature with the nitrogen still flowing.
When cool, the nitrogen flow is stopped, and the bonded phase can be
removed from the fluidizer. Alternatively, the silanizing reagent can be
replaced with a capping reagent and the material capped by a similar
procedure and the product then removed.
The fluidized bed synthesis gives a more reproducible product and, at
the same time, eliminates many of the tedious operations that are
involved in the alternative method, such as solvent removal and
recovery, washing procedures and other manipulations. It also offers the
possibility of carrying out very complicated syntheses involving multiple
steps by a relatively simple procedure. However, the technique is more
complicated, requires more elaborate apparatus and needs to be operated
by an experienced technician.
Choosing a Bonded Phase
Different organic moieties can be bonded to all types of silica gel
particles including the very popular spherical variety. One of the most
important features to consider when choosing a bonded phase is the
reproducibility of the product. The packing must obviously achieve the
separation that is required, but unfortunately any bonded phase has a
This eBook is protected by Copyright law and International Treaties. All rights are reserved. This book is covered by an End User Licensee
Agreement (EULA). The full EULA may be seen at http://www.library4science.com/eula.html.
57
limited lifetime that may range from a few hours, if operated at extremes
of pH, to several months if operated under mild conditions. It follows,
that all columns will eventually need replacing and if a specific
analytical procedure has been established on a particular column, then
the replacement must have as near identical chromatographic properties
as possible. The reproducibility of the column is particularly important
in forensic analysis or for analyses that are carried out to ensure
adherence to regulatory standards as in environmental and pollution
studies. The following tests are recommended as minimal for a new
column and the results should be compared with the data obtained from
the previous column as received.
1/ The column permeability should be measured i.e., the pressure
required to produce a given flow rate e.g., a flow rate of 1 ml per
minute.
2/ The column dead volume should be measured by determining the
retention volume of an unretained solute
3/ The column efficiency should be measured for a set of standard
solutes. If possible, the solutes should be chosen, from those likely to be
present in the samples to be analyzed. Solutes eluting at (k') values of 2,
5 and 10 would be appropriate.
2/ The corrected retention volumes of a series of solutes spanning a (k')
range of 1 to 20 should be determined and their retention ratios
calculated.
All the measurements should fall within 5% of the specified values.
However, these criteria may not be sufficiently stringent for some
forensic purposes and, consequently, the tests given above should be
considered as minimum requirements for litigation purposes..
Types of Bonded Phase
There are three basic types of bonded phases, the brush phases, the bulk
phases and the oligomeric phases. The different phases are produced by
the use of the mono, di- and tri- substituted silanes respectively in the
This eBook is protected by Copyright law and International Treaties. All rights are reserved. This book is covered by an End User Licensee
Agreement (EULA). The full EULA may be seen at http://www.library4science.com/eula.html.
58
bonding process, e.g. the monochloro, dichloro and tricoloro silanes.
The monochlorosilanes, for instance octyldimethyl-chlorosilane, react
with the hydroxyl groups on the silica surface to produce
dimethyloctylsilyl chains attached to the silica.
CH3CH2CH2CH2CH2CH2CH2CH2 3 2-SiCl + H-O-Si-
CH3CH2CH2CH2CH2CH2CH2CH2 3 2-Si-O-Si-HCl +
)(CH
)(CH
The alkyl chains are thought to stand out from the surface like bristles of
a brush, hence the term brush phase. The extent to which the silanyl
groups are reacted is a subject of debate at this time. It is thought that
the two methyl groups next to the silicon atom hinder the reaction of
adjacent hydroxyl groups with the reagent and thus there will be a
considerable amount of unreacted hydroxyl groups remaining even after
capping. In the extreme, it has been suggested that there is a hydroxyl
group situated between each bonded chain. There is certainly evidence
of some polar interactions with reverse phases, which if completely
covered with hydrocarbon chains should only exhibit dispersive
interactions. However, reverse phases are predominantly dispersive in
character and it would appear that if there are any hydroxyl groups still
present on the surface it is likely that they would be relatively few in
number compared with the bonded moieties.
Oligomeric Phases
The di-substituted silanes such as methyloctyldichlorosilanes can be
used to synthesis the oligomeric bonded phases, which is a far more
complicated procedure (24). The silica is first reacted with the
methyloctyldichlorosilane to link methyloctylchlorosilyl groups to the
surface. The product is then treated with water, which hydrolyses the
methyloctylchlorosilyl groups to methyloctylhydroxysilyl groups with
This eBook is protected by Copyright law and International Treaties. All rights are reserved. This book is covered by an End User Licensee
Agreement (EULA). The full EULA may be seen at http://www.library4science.com/eula.html.
59
the elimination of hydrochloric acid. The "hydroxy" product is then
reacted with more methyloctyldichlorosilane, attaching another
methyloctylchlorosilyl group to the previous groups.
This process of alternately treating the product with the silane reagent
and then water can be repeated until, in the original synthesis, eight or
ten oligomers are linked to each other and attached to each stearically
available hydroxyl group on the surface. However, it should be noted
that the hydroxylation of the bonded moiety is not the only possible
reaction that can take place with the water. There will be situations
where it will be stearically possible for a water molecule to react with
two adjacent chains and thus produce some cross-linking. However, a
small amount of cross-linking would indeed strengthen the bonded
system and could be advantageous.
The product is finally treated with trimethylchlorosilane or some other
capping reagent to eliminate the last hydroxyl groups formed at the end
of the oligomer. The oligomers are layered over the surface making the
product extremely stable exhibiting almost no polar characteristics
whatsoever. However, due to the complexity of the synthesis [which
needs to be carried out in a fluidized bed for efficient reaction]
oligomeric phases are expensive to manufacture and, consequently, are
not often used and, at this time, are not commercially available.
Nevertheless, this type of procedure, if used to synthesize a reversed
phase, would produce material that would probably be completely
dispersive in character with no measurable polarity.
Bulk Phases
If the silica surface is saturated with water and octyltrichlorosilane is
used as the reagent, reaction occurs with both the hydroxyls of the silica
surface and the adsorbed water causing a cross-linking reaction, and an
octylsilyl polymer can be built up on the surface. The same procedure
can be used as that in the synthesis of oligomeric phases and the
material can be alternatively treated with water and the trichlorosilane
reagent. Layers of bonded phase are built up on the surface but, in this
case, due to the tricoloro function of the reagent, extensive cross-linking
This eBook is protected by Copyright law and International Treaties. All rights are reserved. This book is covered by an End User Licensee
Agreement (EULA). The full EULA may be seen at http://www.library4science.com/eula.html.
60
occurs. As a result of the polymerization process, the stationary phase
has a chemically cross-linked, multi-layer character and, consequently,
is termed a "bulk" phase. The "bulk" phases are almost as popular as the
"brush" phases as they tend to have a higher carbon content (more
organic material bonded to the surface) and thus provide a little greater
retention and selectivity. "Bulk" phases have about the same stability to
aqueous solvents and pH as the "brush" phases.
Using appropriate organic chlorosilanes, polar or polarizable groups
such as nitriles or aromatic rings can be bonded to the silica to provide
stationary phases covering a wide range of polarities. Bonded ion
exchange materials have also been synthesized, although they are not as
stable to salt solutions and extremes of pH as the ion exchange resins.
Interactions Between 'Brush' and 'Bulk' Reverse Phases and Aqueous
Solvents
The interactions between aqueous solvents and brush reverse phases
differ very significantly from those with a bulk reverse phase at very low
concentrations of solvent.
This difference has been investigated by a number of workers (25-27)
and the basic difference between the two types of phase are shown in
the curves relating retention volume of methanol to the concentration of
methanol in the mobile phase in figure 35. The phases shown are the
RP-18 brush, reverse phase manufactured by E. M. Laboratories, which
had a C18 (dimethyloctadecyl) chain and ODS-3 a bulk reverse phase
which had a C18 (octadecyl) chain and was manufactured by Whatman
Inc. The curves relating retention volume with solvent composition for
the two phases show very different behavior patterns.
The ODS-3 bulk reverse phase behaves in the expected manner, as the
concentration of methanol increases the retention volume of the ethanol
decreases smoothly and continuously up to a concentration of 10%w/v
of methanol. The brush phase, however, behaves in a very unexpected
fashion. The retention volume of ethanol at first increases as the solvent
This eBook is protected by Copyright law and International Treaties. All rights are reserved. This book is covered by an End User Licensee
Agreement (EULA). The full EULA may be seen at http://www.library4science.com/eula.html.
61
concentration increases, to a maximum at a concentration of about
3%w/v of methanol.
Figure 35. A Graph of the Corrected Retention Volume of Ethanol
against the Concentration of Methanol in the Mobile Phase for a
Bulk and Brush Reverse Phase
At methanol concentrations above 3%w/v the retention values begin to
fall and eventually follow a curve parallel to that for the bulk phase but
somewhat higher. Lochmüller and Wilder (25) carried out some similar
experiments and obtained the same results while Gilpin and Squires (26)
carried out some thermodynamic measurements on the two systems and
showed an anomalous behavior by the brush type phase at very low
concentrations of solvent. It was suggested that the behavior of the brush
phase could be explained as a result of the free chains interacting with
themselves in preference to interacting with the aqueous solvent. In
effect, this was the same phenomena of immiscibility that occurs
between water and a liquid hydrocarbon. The dispersive forces between
the hydrocarbon chains themselves are greater than the forces between
the hydrocarbon chains and the aqueous solvent. As a result the chains
This eBook is protected by Copyright law and International Treaties. All rights are reserved. This book is covered by an End User Licensee
Agreement (EULA). The full EULA may be seen at http://www.library4science.com/eula.html.
62
interact with themselves and collapse on the surface of the silica. When
the methanol content of the mixture is increased, the solvent becomes
more dispersive in nature and the hydrocarbon chains can then interact
with the solvent and no longer exist in a collapsed state. The situation is
depicted in figure 36.
Brush Phase in Contact wi th Methanol Water Mixture
Brush Reverse Phase in Contact with Pure Water
Figure 36. Effect of Solvent Composition on the Orientation of the
Brush Phase.
The relationship between the retention of ethanol on the brush phase
with water/methanol mixtures, as shown in figure 35, can now be
explained. In pure water the hydrocarbon chains of the brush phase are
collapsed on the surface and thus, the effective surface area of the
stationary phase is much reduced. Consequently, the retention volume of
the solute, being proportional to the available surface area, is also
reduced. As methanol is added to the solvent mixture, the solvent
becomes more dispersive and the hydrocarbon chains can begin to
interact with it and, as a consequence, begin to unfold. The liberation of
the chains from the surface results in an increase in the effective surface
area of the stationary phase and the retention of the solute also starts to
increase. This process continues until there is sufficient methanol in the
solvent for the hydrocarbon chains to totally interact with the solvent
and be completely released from the surface. At this concentration
(about 3%w/v of methanol) the retention volume of the ethanol reaches
This eBook is protected by Copyright law and International Treaties. All rights are reserved. This book is covered by an End User Licensee
Agreement (EULA). The full EULA may be seen at http://www.library4science.com/eula.html.
63
a maximum. Subsequent increase in methanol concentration merely
increases the interactions of the ethanol with the mobile phase and, by
adsorption of the solvent onto the surface of the reverse phase, reduces
the interactive forces with the reverse phase. Consequently, the retention
volume steadily decreases in the expected manner.
Having explained the behavior of the brush phase it is now interesting to
consider the behavior of the bulk phase. There appears to be no
interaction between the hydrocarbon chains themselves and no change in
surface area at high water concentrations. In fact, the retention of
ethanol falls steadily as the methanol content increases in the expected
manner. The explanation given for this is that the cross-linking that
takes place when the bulk phase is synthesized, keeps the polymeric
chain system rigid and does not allow the individual hydrocarbon chains
to collapse on the surface. As a consequence, the surface area is not
reduced and the retention behavior is normal. Another aspect of the
behavior of the bulk phase at low solvent concentrations is that it does,
in fact, confirm the cross-linked nature of the bulk phase. It would also
appear that for certain solutes, the best reverse phase for operation with
aqueous mixtures containing very little solvent might be a bulk reverse
phase. The retention mechanism on brush type phases under these
conditions might be anomalous.
The Retention Properties of Bulk and Brush Phases.
The equation for the corrected retention volume of a solute (V'r ) (see
Book 6) is as follows,
V'r = KVS or V'r = KAS
where (K) is the distribution coefficient of the solute between the two
phases,
(VS) is the effective volume of stationary phase in the column,
and (AS) is the effective surface area of the packing.
This eBook is protected by Copyright law and International Treaties. All rights are reserved. This book is covered by an End User Licensee
Agreement (EULA). The full EULA may be seen at http://www.library4science.com/eula.html.
64
The two equations are given to illustrate that the reverse phase system
may be considered as a liquid/liquid or liquid/solid distribution system
where,
VS = dfAS
and (df) is the effective film thickness of the bonded material and
any solvent that may be adsorbed on its surface
It follows that the retention of the solute will depend only on the volume
or surface area of the bonded material. Thus, providing all the bonded
phase is available for solute interaction, the retention volume will be
proportional to the carbon content of the phase. Scott and Kucera (28)
examined a series of commercially available reverse phases and
determined the carbon content of each phase and the retention volume of
a series of solutes on columns packed with each adsorbent. The retentive
properties of the five reverse phase are shown in figure 37 where the
corrected retention volume (V'r) of 2-ethyl anthraquinone is plotted
against carbon content of the reverse phase. It is seen, somewhat
surprisingly, that there is a linear relationship between retention volume
and carbon content of the brush phases (R2, R8, R18). This relationship
can only be expected to occur if all the stationary phase is available to
the solute and the packing procedure is very reproducible so that each
column contains the same amount of packing.
Carbon Content (%w/w)
RP2
RP8
RP18
ODS
ODS2
V'
(ml)
0
5
10
15
20
0 10 20
This eBook is protected by Copyright law and International Treaties. All rights are reserved. This book is covered by an End User Licensee
Agreement (EULA). The full EULA may be seen at http://www.library4science.com/eula.html.
65
Figure 37. Graph of Retention Volume against Carbon Content
(%w/w)
It should again be stressed that all three reverse phases were produced
from base silicas of very different surface areas and, despite this, the
linear relationship between carbon content and corrected retention
volume remained. This relationship may will depend, not only on the
type of silica gel that is used but also on h bonding process and his
relationship has not been proved generally.
The relative resolution obtained from different reverse phases carrying
diverse carbon contents and extreme chain lengths is shown by the
chromatograms in figure 38. The higher retentive capacity of the bulk
phase ODS2 phase is again clearly demonstrated.
Figure 38. Chromatograms from Brush Phases Carrying Different
Carbon Loads and Different Chain Lengths
This eBook is protected by Copyright law and International Treaties. All rights are reserved. This book is covered by an End User Licensee
Agreement (EULA). The full EULA may be seen at http://www.library4science.com/eula.html.
66
Short chain reverse phases reduce the extent to which proteins are
denatured in the separation of substances of biological origin, it is seen
by the chromatogram from the C2 reverse phase, that a serious price
must be paid in loss of resolution if the nature of the separation demands
the use of such material. However, the development of the polymer
packings have, at last, partly solved this problem.
In general, because the brush type phases can be synthesized in a more
reproducible manner, particularly if carried out in a fluidized bed, the
brush phases are generally recommended for the majority of
applications. For high retentive capacity and for systems that will be
operated with aqueous solvent mixtures having a very high water
content, the bulk phases might be preferred. The partially reacted, low
carbon content bulk phase may also have special areas of application
particularly in sample preparation.
Macroporous Polymers
Polymeric ion exchange materials were developed for chromatography
in the early sixties resulting in the introduction of macro-porous
polymers (29-31). The advantages of this material lay in the macro-
porous nature of the resin packing, which consisted of resin particles a
few microns in diameter, which, in turn, comprised of a fused mass of
polymer micro-spheres a few Angstroms in diameter. The resin polymer
micro-spheres play the same part as the silica gel primary particles, and
confer on the polymer a relatively high surface area together with a high
porosity. The high surface area provided increased solute retention and
selectivity together with a superior loading capacity and, consequently, a
wide dynamic range of analysis. The material consists of a highly cross-
linked polystyrene resin with about a 50Å pore diameter. In the case of
the ion exchange materials, inorganic groups of appropriate charge were
chemically attached (e.g., by sulfonation). The initial resins
manufactured by Rohm and Haas were called Amberlite but were
largely employed in production processes and, consequently, had very
large particle diameters. Modern macroporous resin-type stationary
This eBook is protected by Copyright law and International Treaties. All rights are reserved. This book is covered by an End User Licensee
Agreement (EULA). The full EULA may be seen at http://www.library4science.com/eula.html.
67
phases have a range of particle diameters, which can be as small as 2.5
or 3.0 microns. The more popular resin based packings are based on the
co-polymerization of polystyrene and divinylbenzene. The degree of
cross-linking determines the rigidity of the resin and the greater the
cross-linking the harder the resin becomes. Ultimately, at extremely high
cross-linking, the resin becomes brittle. To produce an ion exchange
resin, the surface of the polystyrene-divinylbenzene copolymer is reacted
with suitable reagents and covered with the required ionogenic
interacting groups.
Most cross-linked polystyrene resins employed in LC are the macro-
reticular type and can be produced with almost any desired pore size,
ranging from 20Å to 5,000Å. Underivatized, they exhibit strong
dispersive type interactions with solvents and solutes together with some
induced polarizability arising from the aromatic nuclei in the polymer if
strongly polar solutes are being separated. Consequently, the untreated
resin has found some use as an alternative to the C8 and C18 reverse
phase columns based on silica.
This eBook is protected by Copyright law and International Treaties. All rights are reserved. This book is covered by an End User Licensee
Agreement (EULA). The full EULA may be seen at http://www.library4science.com/eula.html.
68
Courtesy of TOSOHAAS Inc.
Figure 39. The Separation of a Crude Protein Extract by Exclusion
on a Micro-Reticulated Resin Column
Due to their being stable to extremes of pH, their use for the separation
of peptide and proteins at both high and low pH has been well
established. An example of a macro-reticulated resin phase used as an
exclusion medium in the separation of a crude protein extract is shown
in figure 39. The column, 9TSKgel QC-PAK GFC 399GL0, was 15 cm
long and 8 mm in diameter and the separation was carried out at a flow
rate of 1 ml/min. The mobile phase consisted of 0.5M NaCl in 0.05M
sodium phosphate buffer at pH 7.0. The sample was 5 l of a rat liver
extract. An excellent separation based on molecular size is obtained
from which the molecular weight range of the mixture and even that of
the individual components could be estimated.
LC Mobile Phases
The choice of phase system can be very complex, particularly if
multicomponent mixtures are to be separated. In the first instance the
type of stationary phase needs to be chosen and this choice must be
based on the interactive character of the solutes to be separated. If the
solutes are predominantly dispersive then the stationary phase must also
be dispersive (a reversed phase) to promote dispersive interaction with
the solutes and provide adequate retention and selectivity. If the solutes
are strongly polar then a polarizable stationary phase (one containing
aromatic rings or cyano groups) would be appropriate to separate the
solutes by polar and induced polar interactions. If the solutes are weakly
polar then a strong polar stationary phase would be required (such as
silica gel) to separate the solute by polar interactions.
The mobile phase must be chosen to complement the stationary phase so
that the selected interactions are concentrated in the stationary phase.
Thus, a reversed phase having strong dispersive interactions would be
used with a strongly polar mobile phase (e.g., mixtures of methanol and
This eBook is protected by Copyright law and International Treaties. All rights are reserved. This book is covered by an End User Licensee
Agreement (EULA). The full EULA may be seen at http://www.library4science.com/eula.html.
69
water acetonitrile and water or tetrahydrofuran and water). In contrast, if
the strongly polar silica gel is selected for the stationary phase then a
strongly dispersive mobile phase would be appropriate (e.g., n-heptane,
n-heptane/methylene chloride or n-heptane with a small quantity of n-
propanol or ethanol). In general the mobile phase must be chosen so that
the selected interactions strongly dominate in the stationary phase and
are minimized in the mobile phase.
Solvent/Solute Interactions with the Silica Gel Surface
In all chromatography systems both the solvent and the solutes interact
with the stationary phase. It follows that, when the silica surface is in
contact with a solvent, the surface is covered with a layer of the solvent
molecules. If the mobile phase consists of a mixture of solvents the
surface is partly covered by one solvent and partly with the other (32).
Thus, any solute interacting with the stationary phase may well be
presented with two, quite different types of surface with which to
interact.
The probability that a solute molecule will interact with one particular
type of surface will be statistically controlled by the proportion of the
total surface area that is covered by that particular solvent.
Dispersive solvents appear to be adsorbed from a solvent mixture on the
surface of silica gel according to the Langmuir adsorption isotherm (33).
Examples of mono-layer adsorption isotherms obtained for benzene,
chloroform and butyl chloride are shown in figure 40.
This eBook is protected by Copyright law and International Treaties. All rights are reserved. This book is covered by an End User Licensee
Agreement (EULA). The full EULA may be seen at http://www.library4science.com/eula.html.
70
Figure 40. Langmuir Adsorption Isotherms for Benzene, Butyl
Chloride and Chloroform
The adsorption isotherms of the more polar solvents, ethyl acetate,
isopropanol and tetrahydrofuran from n-heptane solutions on to the
silica gel surface did not fit the simple mono-layer adsorption equation
but did fit the bi-layer adsorption isotherm which is a simple extension
of the monolayer formation process. The bi-layer adsorption isotherm
for ethyl acetate on silica gel is shown in figure 41. The curve is
theoretical and the points experimental.
The individual isotherms for the two adsorbed layers of ethyl acetate are
included in figure 41. The two curves, although of the same form, are
quite different in magnitude. The first layer is very strongly held to the
surface and is complete when the concentration of ethyl acetate in the
mobile phase is no more than 1%w/w. As the concentration of ethyl
acetate starts to rise above 1%w/w the second layer is only just being
formed. The formation of the second layer of ethyl acetate is much
slower and obviously the interactions between the solvent molecules
with those already adsorbed on the surface are much weaker than their
interaction with the silica gel silanol groups.
This eBook is protected by Copyright law and International Treaties. All rights are reserved. This book is covered by an End User Licensee
Agreement (EULA). The full EULA may be seen at http://www.library4science.com/eula.html.
71
0.00
0.05
0.10
0.15
Calculated
Experimental
Firs t Lay er
Sec ond Lay er
Concentration of Ethyl Acetate (% w/v)
Co
nce
ntr
ati
on
on
Sil
ica
(g/g
)
0 1 2 3 4
First Layer
Se cond Laye r
Combined Layer
Figure 41. The Individual and Combined Adsorption Isotherms for
Ethyl Acetate on Silica Gel
If it is assumed that the total area covered by the first layer of ethyl
acetate will be very similar to the area covered by the second layer, then
only about one third of the second layer is complete at an ethyl acetate
concentration of about 4%w/v. In contrast, the first layer is virtually
complete at an ethyl acetate concentration of 1%w/v. It should be
pointed out that solvent layer formation on the surface of the silica is not
necessarily restricted to binary systems and multi-layers are quite
feasible.
Solute Stationary Phase Interactions
There are basically two types of interaction that can take place between
a solute and the silica gel surface. Firstly, the solute molecule can
interact with the adsorbed solvent layer and rest on the top of it. This
type of interaction is called sorption interaction and occurs when the
molecular forces between the solute and the silica are relatively weak
compared with the forces between the solvent molecules and the silica.
The second type of interaction is where the solute molecules displace the
This eBook is protected by Copyright law and International Treaties. All rights are reserved. This book is covered by an End User Licensee
Agreement (EULA). The full EULA may be seen at http://www.library4science.com/eula.html.
72
solvent molecules from the surface and interact directly with the silica
gel itself, for example, the silanol groups.
SORPTION
SO LUTE
CHLOROFORM n-HEPTANE
DISPLCACEMENT
CHLOROFORM n-HEPTANESOLUTE
SOLUTE
Figure 42. Diagram Depicting Sorption and Displacement
Occurring on the Silica Surface.
These two types of interaction are shown in figure 42. Displacement
would occur if the solute was strongly polar such as an alcohol, which
would interact more strongly with the polar silanol group than the
dispersive chloroform layer. Sorption is depicted as a solute molecule on
interacting with each solvent layer and can not interact strongly enough
with the silica gel surface to displace the solvent..
Mobile phases consisting of mixtures of polar and dispersive solvents
frequently produce surface bi-layers when used with silica gel as a
stationary phase and therefore a far more complicated set of interactive
possibilities exist. These possibilities are depicted in figure 43.
This eBook is protected by Copyright law and International Treaties. All rights are reserved. This book is covered by an End User Licensee
Agreement (EULA). The full EULA may be seen at http://www.library4science.com/eula.html.
73
So lute i ntera cting
with second l ayer of
sol vent (B) SORPT ION
So lute i ntera cting
with first laye r of
sol vent (B) by
displaci ng so lvent
fro m second layer
DISPLACEM ENT
Displace d
So lvent (B)
Displace d
So lvent (B)Displace d
So lvent (A)
So lute i ntera cting
directly with f irst la yer
of solve nt (B) SORPTION
So lute i ntera cting
with sil ica surface by
displaci ng so lvent (A)
fro m first laye r
DISPLACEM ENT
So lute i ntera cting
with sil ica surface by
displaci ng so lvent (B)
fro m first laye r
DISPLACEM ENT So lute i ntera cting
directly with l ayer of
sol vent (A) SORPT ION
Figure 43. Different Types of Solute Interaction that can occur on
Silica Surfaces Covered with a Solvent Bi-layer
The surface offers the opportunity of a number of sorption and
displacement processes that can take place between the solute and the
stationary phase surface. There are three different surfaces on which a
molecule can interact by sorption and three different surfaces from
which molecules of solvent can be displaced and allow the solute
molecule to penetrate. In any separation all the alternatives are possible
but it is more likely that for one particular solute, one type of interaction
will dominate. the various types of interaction are depicted in figure 43.
Where there are multi-layers of solvent, the solvent that interacts directly
with the silica surface is the most polar, and consequently constitutes the
first layer. Depending on the concentration of the polar solvent the next
layer may be a second layer of the same polar solvent as in the case of
ethyl acetate. If, however, the quantity of polar solvent is limited, then
This eBook is protected by Copyright law and International Treaties. All rights are reserved. This book is covered by an End User Licensee
Agreement (EULA). The full EULA may be seen at http://www.library4science.com/eula.html.
74
the second layer might consist of a less polar component of the solvent
mixture. If a ternary mixture of solvents is used, the nature of the
surface, and the solute interactions with the surface can become very
complex indeed. In general the stronger the polarity of the solute the
more likely it is to interact with the surface by displacement even to the
extent of displacing both layers of solvent (one of the alternative
processes that is not depicted).
Solvent/Solute Interactions with the Reversed Phase Surface
Solvents interact with the surface of a reverse phase in a similar manner
to the surface of silica gel. In figure 44, the adsorption isotherms for a
series of aliphatic alcohols are shown. The effect of the carbon chain-
length of the alcohol on the strength of the adsorption is clearly seen
from the shape of the curves. The most strongly adsorbed alcohol,
butanol, (the alcohol with the longest chain and, thus, the most
dispersive) has only a four carbon chain and yet the surface is
completely covered when the butanol concentration is only about
2%w/v. Thus, any component of the mobile phase with a hydrocarbon
chain length of four or more, will be rapidly adsorbed and modify the
reverse phase surface extensively and, consequently, the magnitude of
solute retention.
The curves shown in figure 44 only cover a range of 0 to 0.05 g.ml-1. In
order to show the shapes of the adsorption isotherms for the higher
alcohols in proportion to those of the lower alcohols with reasonable
clarity, the same curves are shown in figure 45 for an alcohol
concentration range of 0-100% (which is approximately 0-0.8g/ml).
This eBook is protected by Copyright law and International Treaties. All rights are reserved. This book is covered by an End User Licensee
Agreement (EULA). The full EULA may be seen at http://www.library4science.com/eula.html.
75
0 0.01 0.02 0.03 0.04 0.05
0
Concentrati on of Alcohol (g/ml)
Methanol
Ethanol
n-Propanoln-Butanol
Ma
ss o
f So
lvent
on
Surf
ace
(g
/cm
)
2
It should be noted that the mass adsorbed is expressed as g.cm-2
Figure 44. The Adsorption Isotherms of a Homologous Series of
Aliphatic Alcohols over the Concentration Range of 0 to 0.0.5 g.ml-
1
The weak nature of the methanol adsorption, relative to the other
alcohols is clear and it is seen that the surface of the reverse phase is
being modified over one third of the methanol concentration range. The
reverse phase surface can be modified in a controlled manner, over the
range of 0 to about 40 % methanol, but between methanol
concentrations of 40% and 100 % the nature of the reverse phase
surface remains sensibly constant and it is the solute interactions in the
mobile phase that are progressively modified. Acetonitrile and
tetrahydrofuran behave in a similar manner but their adsorption
isotherms are closer in magnitude to those of ethanol than of methanol.
The types of interactions that can take place between the solute and the
reverse phase are similar to those that can take place between the solute
and the silica gel surface. Solutes can interact by the sorption process,
the displacement process or a combination of both. The same rules
apply; if the solvent interacts more strongly with the surface than the
This eBook is protected by Copyright law and International Treaties. All rights are reserved. This book is covered by an End User Licensee
Agreement (EULA). The full EULA may be seen at http://www.library4science.com/eula.html.
76
solute then the solute interacts with the adsorbed layer of solvent by
sorption.
3
4
0 0.16 0.32 0.48 0.800.64
Concentration of Alcohol (g/ml)
0
1 Butanol 2 Propanol
3 Ethanol 4 M ethanol
1
2
Mass
of
Solv
ent
on
Su
rface
(cm
)
-2
Figure 45. The Adsorption Isotherms of a Homologous Series of
Aliphatic Alcohols
If, on the other hand, the solute interacts more strongly with the reverse
phase than the layer of solvent molecules then the solute will displace
the solvent and interact directly with the surface by displacement. In,
general, those solutes that elute early in the chromatogram will interact
by sorption, those that elute late in the chromatogram will interact by
displacement and at some intermediate point in the elution scale, solute
stationary phase interactions will probably involve both sorption and
displacement. Bi-layer adsorption is also possible with reverse phases
but, at this time, experimental evidence of this does not appear to be
available in the literature.
Molecular Interactions in the Mobile Phase
The 'elutive' capacity of the mobile phase (as opposed to the 'retentive'
capacity of the stationary phase) depends on the strength of the different
interactions that can take place in the mobile phase and the probability
of a particular interaction occurring. Purnel and Laub (34)
This eBook is protected by Copyright law and International Treaties. All rights are reserved. This book is covered by an End User Licensee
Agreement (EULA). The full EULA may be seen at http://www.library4science.com/eula.html.
77
experimentally demonstrated in GC, that for a stationary phase
consisting of a pair of non-associating liquids, the distribution
coefficient of a solute was linearly related to the volume fraction of
either liquid. This relationship indicated that the volume fraction of a
solvent in a liquid mixture determined the probability of interaction.
Much the same as the partial pressure of gas determines the probability
of collision. This relationship was challenged by a number of workers
who also demonstrated that for certain solvent pairs this linear
relationship broke down. However, it was also shown (35,36) that the
nonlinear relationship occurred when there was strong association
between the components of the mixture which resulted in a ternary
mixture containing the two components and the associate of the two
components the concentration of which depended on the equilibrium
constant and the experimental conditions.
Thus, for a mobile phase consisting of a binary mixture of solvents, as
the retention volume will be inversely proportion to the elutive capacity
of the mobile phase it will be also inversely proportional to the volume
fraction of either component providing there is no strong association
between the components. This was experimentally demonstrated by Katz
et al. (35) who employed a liquid/liquid distribution system using water
and a series of immiscible solvent mixtures as the two phases and
measured the absolute distribution coefficient of a solute for different
mixtures.
The solute they used was n-pentanol and the immiscible solvent
consisted of mixtures of n-heptane and chloroheptane, n-heptane and
toluene and n-heptane and heptyl acetate. The two-phase system was
thermostatted at 25oC and, after equilibrium had been established, the
concentration of solute in the two phases was determined by GC
analysis. The results they obtained are shown in figure 46. It is seen that
linear relationships between solvent composition and distribution
coefficient was obtained for all three solvent mixtures simulating the
results that Purnell and Laub obtained in their gas chromatography
experiments.
This eBook is protected by Copyright law and International Treaties. All rights are reserved. This book is covered by an End User Licensee
Agreement (EULA). The full EULA may be seen at http://www.library4science.com/eula.html.
78
Figure 46. Graph of Distribution Coefficient against Solvent
Composition.
Aqueous Solvent Mixtures
When the relationship between the distribution coefficient of a solute
and solvent composition, or the corrected retention volume and the
solvent composition, was tested with aqueous solvent mixtures it was
found that the relationship identified by Purnell and Laub and Katz et al
failed. It was suspected that the failure was due to the solvent strongly
associating with the water and, in fact, an aqueous solution of methanol,
for example, contained methanol, water and methanol associated with
water. The solvent mixture was thus, a ternary system and thus the
linear relation ship between the volume fraction (before mixing) and
retention would not be expected to hold.
The association of methanol and water was examined by Katz,
Lochmüller and Scott (36) using volume change on mixing and
refractive index data and established that the methanol/water solvent
This eBook is protected by Copyright law and International Treaties. All rights are reserved. This book is covered by an End User Licensee
Agreement (EULA). The full EULA may be seen at http://www.library4science.com/eula.html.
79
system was indeed a complex ternary system. They calculated both the
association equilibrium constant and the distribution of the different
components of a methanol water mixture form zero to 100% methanol.
The curves they obtained are shown in figure 47.
Figure 47. Diagram of the Ternary Solvent System for
Methanol/Water Mixtures
It is seen from figure 47 that there are three distinct ranges of methanol
concentration where the solvent will behave very differently. From zero
to 40%v/v of methanol in the original mixture, the solvent will largely
behave as though it were a binary mixture of water and methanol
associated with water. From 40%v/v to 80%v/v of methanol in the
original mixture, the solvent will predominantly behave as though it
were a ternary mixture of water, methanol and water associated with
methanol. From 80%v/v to 100%v/v of methanol in the original mixture,
This eBook is protected by Copyright law and International Treaties. All rights are reserved. This book is covered by an End User Licensee
Agreement (EULA). The full EULA may be seen at http://www.library4science.com/eula.html.
80
the solvent will again behave as though it were again a binary mixture
but this time a mixture of methanol and water associated with methanol.
The curves shown in figure 47 explain some of the unique
characteristics of mobile phases consisting of methanol water mixtures
when used in reversed phase LC. From figure 47 it is seen that when the
original mixture contains 50%v/v of methanol there is little free
methanol available in the mobile phase to elute the solutes as it is mostly
associated with water. Subsequently, however, the amount of methanol
unassociated with water increases rapidly in the solvent mixture and this
rapid increase must be accommodated by the use of a convex gradient
profile when employing gradient elution. The convex gradient will
compensate for the strongly concave form of the unassociated methanol
concentration profile shown in figure 47 which will be the strongest
eluting component of the mobile phase. The strong association of
methanol with water could also account for the fact that proteins can
tolerate a significant amount of methanol in the mobile phase before
they become denatured. It is clear that this is because there is virtually
no unassociated methanol present in the mixture, which could cause
protein denaturation since all the methanol is in a deactivated state by
association with water.
Katz, Lochmüller and Scott also examined acetonitrile/water, and
tetrahydrofuran(THF)/water mixtures in the same way and showed that
there was significant association between the water and both solvents
but not to the same extent as methanol/water. At the point of maximum
association for methanol, the solvent mixture contained nearly sixty
percent of the methanol/water associate. In contrast the maximum
amount of THF associate that was formed amounted to only about 17%
and for acetonitrile the maximum amount of associate that was formed
was as little as 8%. It follows that acetonitrile water mixtures would be
expected to behave more nearly as binary mixtures than methanol/water
or THF/water mixtures.
Katz et al. measured the distribution coefficient of benzene between n-
hexadecane and some methanol/water mixtures. Using the data from
This eBook is protected by Copyright law and International Treaties. All rights are reserved. This book is covered by an End User Licensee
Agreement (EULA). The full EULA may be seen at http://www.library4science.com/eula.html.
81
figure 47, they plotted the distribution coefficient of benzene against the
volume fraction of methanol unassociated with water. The results they
obtained are shown in figure 48.
Figure 48. Graph of Distribution Coefficient of Benzene Between
Methanol and Water Mixtures and n-Hexadecane against Volume
Fraction of Methanol Unassociated with Water in the Aqueous
Mixture
From the results in figure 48, benzene does not appear to interact with
methanol associated with water or water itself but solely with methanol.
The linear curve is obtained with zero intercept confirming the validity
of this dependence. In fact, the methanol associated with water plays no
significant part in competing for the benzene against the dispersive
interactions of the n-hexadecane.
Chiral Stationary Phases
This eBook is protected by Copyright law and International Treaties. All rights are reserved. This book is covered by an End User Licensee
Agreement (EULA). The full EULA may be seen at http://www.library4science.com/eula.html.
82
There are basically five general types of chiral stationary phase in
common use in LC. The first is the protein based stationary phase.
These stationary phases usually take the form of natural proteins bonded
to a silica matrix. As they are proteins, they contain a large number of
chiral centers and are known to interact strongly with small analytes
exhibiting strong chiral selectivity. There are specific interactive sites
that provide chiral selectivity, but there are many more sites that only
contribute to general retention. These other sites can be deactivated with
mobile phase additives (e.g. octylamine), which reduces the overall
retention and increases the chiral selectivity.
The second type consists of relatively small molecular weight chiral
substances bonded to silica Pirkle (37). Each bonded group has a limited
number of chiral centers available but, due to their small size, there can
be a large number of groups bonded to the silica (as opposed to much
larger complex chiral moieties). It follows, that a relatively high
probability is maintained of the solute interacting with a chiral center.
The advantage of the Pirkle chiral phases is that, as the overall
interacting molecule is small, the solutes are not strongly retained and
thus the chiral selectivity becomes the dominant factor. The third type is
based on polymers of cellulose and amylose, which were developed by
Okamato (38). These are derivatized to link appropriate interactive
groups to the cellulose polymer, which is then coated onto a silica
support. The fourth type is based on the macrocyclic glycopeptides
introduced by Armstrong (39). These are materials that also contain a
large number of chiral centers, together with molecular cavities in which
solute molecules can enter and interact with neighboring groups.
The spatial character of the solute will determine the degree of entry and
consequently the proximity of interaction, which, in turn, will determine
the energy of interaction and the magnitude of the retention. Finally, the
fifth group contains the cyclodextrin-based materials that control
retention in a similar manner to that previously described for GC. In LC,
the cyclodextrin stationary phases are bonded to a support such as silica
and are prepared using similar techniques to those for making reverse
phases. The more recent and most effective stationary phases are
This eBook is protected by Copyright law and International Treaties. All rights are reserved. This book is covered by an End User Licensee
Agreement (EULA). The full EULA may be seen at http://www.library4science.com/eula.html.
83
without doubt those based on the macrocyclic glycopeptides and the
cyclodextrins.
Macrocyclic Glycopeptide Phases
The concept of using macrocyclic glycopeptides as chiral stationary
phases was first introduced by Armstrong [39]. One method of
preparation is to covalently bond Vancomycin to the surface of silica gel
particles. Vancomycin contains 18 chiral centers surrounding three
'pockets' or 'cavities' which are bridged by five aromatic rings. Strong
polar groups are proximate to the ring structures to offer strong polar
interactions with the solutes. This type of stationary phase is stable in
mobile phases containing 0–100% organic solvent. The proposed
structure of Vancomycin is shown in figure 49.
O
CH3
NH2
CH3
OH
OO
NN
N
NN
N
O
OH
H
HOH
HOOC
HH
OO
O
OO
H
H H2NC
O
Cl
NHCH3
H H
O
O
O
OHHO
HOCH2
OHOH HO
H
Cl
A
BC
A, B and C are inclusion cavities. Molecular weight 1449. Chiral centers 18. pK's
2.9, 7.2, 8.6, 9.6, 10.4, 11.7. Isoelectric point 7.2
Courtesy of ASTEC Inc.
Figure 49 The Proposed Structure of Vancomycin
This eBook is protected by Copyright law and International Treaties. All rights are reserved. This book is covered by an End User Licensee
Agreement (EULA). The full EULA may be seen at http://www.library4science.com/eula.html.
84
Vancomycin is very stable with a relatively high sample capacity, and,
when covalently bonded silica gel has multiple linkages to the silica gel
surface. It can be used as a reversed phase, with mobile phases having a
high water content, or, alternatively, as polar stationary phase with a
mobile phase of high solvent content (e.g., when used as a reversed
phase, strongly polar THF–water mixtures are very effective mobile
phases). Conversely, when used as a polar stationary phase, n-hexane–
ethanol mixtures are appropriate. Vancomycin has a number of ionizing
groups and thus can be used over a range of different pH values (pH 4.0
to 7.0) and exhibit a wide range of retention characteristics and chiral
selectivities. Ammonium nitrate, triethylammonium acetate and sodium
citrate buffers have all been used satisfactorily with this stationary
phase. An example of the use of the stationary phase to separate the
enantiomers of 3-methyl-5-phenylhydantoin is shown in figure 50.
to 1.85 min. to 2.80 min.
k1 0.78 k1 1.57
k2 1.32 k2 2.13
1.69 1.35 R 2.18 R 3.0
This eBook is protected by Copyright law and International Treaties. All rights are reserved. This book is covered by an End User Licensee
Agreement (EULA). The full EULA may be seen at http://www.library4science.com/eula.html.
85
Courtesy of ASTEC Inc.
Figure 50. The Separation of the Enantiomers of 3-Methyl-5-
Phenylhydantoin Using Polar and Dispersive Interactions
The separation was carried out under two conditions, the first used pure
ethanol as the mobile phase, which is strongly dispersive, and in the
second, the mobile phase that contains 90% of water. In the first case,
the ethanol provides strong dispersive interactions in the mobile phase,
which would significantly exceed any dispersive interactions involved
with the stationary phase. Consequently, the remaining dominant
retentive forces will be polar or ionic. In the second case, the mobile
phase is predominantly water and thus provides strong polar interactions
with the solute but weak dispersive interactions. It also follows, that the
dispersive forces will dominate in the stationary phase. These two
examples demonstrate the useful flexibility of Vancomycin.
Another macrolytic glycopeptide used in chiral chromatography is the
amphoteric glycopeptide Teicoplanin which is commercially available
under the trade name of CHIROBIOTIC T.
This eBook is protected by Copyright law and International Treaties. All rights are reserved. This book is covered by an End User Licensee
Agreement (EULA). The full EULA may be seen at http://www.library4science.com/eula.html.
86
HO
O
NH2
O
HO
Cl
HHO
H
O
H
O
N NNH
O
H
B
A
OCl
N
HO
HHO
C
OHOHO
NHH
NH
O
HOOC
HD
HO
HO
HO NHR
CH2OH O
CH2OH O
HO
HOHNCOCH3
OH
CH2OHOH
OH
O
A, B, C and D are inclusion cavities. Molecular weight 1885. Chiral centers 20,
Sugar moieties 3, and R is CH3-decanoic acid
Courtesy of ASTEC Inc.
Figure 51. The Proposed Structure of Teicoplanin
This material can also be bonded to 5 m silica gel particles by multiple
covalent linkages. Teicoplanin contains 20 chiral centers surrounding
four molecular 'pockets' or 'cavities'. Neighboring groups are strongly
polar and aromatic rings provide ready polarizability. The proposed
structure of Teicoplanin is shown in figure 51. This stationary phase is
claimed to be complementary to the Vancomycin phase and can be used
with the same types of mobile phase, one often providing chiral
selectivity, when the other does not. Teicoplanin can be used in a
reversed phase mode using strongly polar mixtures such as
acetonitrile/aqueous buffer : 10/90 v/v, THF/aqueous buffer 10/90 : v/v,
and ethanol/aqueous buffer : 20/80 v/v). It can also be used as a polar
stationary phase using n-hexane/ethanol mixtures as the mobile phase. In
some cases it is advisable to control the pH even when the solutes are
not ionic, suitable buffers being ammonium nitrate and triethylamine
This eBook is protected by Copyright law and International Treaties. All rights are reserved. This book is covered by an End User Licensee
Agreement (EULA). The full EULA may be seen at http://www.library4science.com/eula.html.
87
acetate. The separation of the Propranolol enantiomers on Teicoplanin is
shown in figure 52.
Figure 52 The Separation of the Enantiomers of Propranolol
Employing Different Acid/Base Ratios
The column used was 25 cm long, 4.6 mm I.D., packed with Chirobiotic
T. The mobile phase was methanol containing acetic acid and
triethylamine in the concentrations shown in figure 52. The column was
operated at room temperature and at a flow rate of 2 ml/min.
Teicoplanin is stable over a pH range of 3.8 to 6.5 although it can be
used for limited periods of time outside this range. By suitable choice of
mobile phase, Teicoplanin can be used in the reversed phase mode.
Another macrolytic glycopeptide that has been used as a chiral
stationary phase, is the glycopeptide, Avoparcin [40]. This stationary
phase is available as CHIROBIOTIC A. Avoparcin is an antibiotic
complex produced by Streptomysces candidus. In particular, it is used to
prevent necrotic enteritus in chickens. The structure of avoparcin is
shown in figure 53.
This eBook is protected by Copyright law and International Treaties. All rights are reserved. This book is covered by an End User Licensee
Agreement (EULA). The full EULA may be seen at http://www.library4science.com/eula.html.
88
Figure 53. The Structure of Avoparcin
There are two forms of Avoparcin the unsubstituted -Avoparcin
structure and the chlorinated structure -Avoparcin, the molecular
weights being 1909 and 1944 respectively. The ratio of -Avoparcin to
-Avoparcins is about 1:4. The aglycon portion of Avoparcin contains
three connected semie rigid macrocyclic rings (one 12-membered, and
two 16-membered), which form a pocket providing possible solute
inclusion. The glycopeptide contains seven aromatic rings with four
phenol moieties, four carbohydrate chains, 16 hydroxyl groups, one
carboxylic acid, two primary amines, one secondary amine, six amide
linkages, two chlorine atoms for -Avoparcin (only one for -
Avoparcin) and 32 stereogenic centers.
It is clear that, there is a wide diversity of interactive possibilities
ranging from weak and strong dispersive interactions, to polar
interactions that span from induced dipole interaction, through dipole–
This eBook is protected by Copyright law and International Treaties. All rights are reserved. This book is covered by an End User Licensee
Agreement (EULA). The full EULA may be seen at http://www.library4science.com/eula.html.
89
dipole interaction, to strong hydrogen bonding. In addition, at the right
pK, basic and acidic ionic interactions can also be invoked. More
importantly, with 32 stereogenic centers the probability of interaction
between chiral centers of solute and stationary phase is relatively high.
Cyclodextrin
The cyclodextrin based chiral stationary phases are some of the more
popular materials used for contemporary chiral separations. One of their
advantages lies in their use with all types of solvent. They can be used
very effectively in the reversed phase mode and, as well as being usable
as a normal phase. The cyclodextrins and their derivatives have been
widely used for all types of chiral separations and can often be used for
preparative separations. Cyclodextrin-based phases are readily available,
covalently bonded to spherical silica gel particles 5 m in diameter. The
cyclodextrins are produced by the partial degradation of starch followed
by the enzymatic coupling of the glucose units into crystalline,
homogeneous toroidal structures of different molecular size. The
molecular structure of , , and cyclodextrins are shown in figure 54.
The alpha-, beta- and gamma-cyclodextrins and have been shown to
contain 6 (cyclohexamylose), 7 (cycloheptamylose) and 8
(cyclooctamylose) glucose units, respectively. These cyclic, chiral, torus
shaped macromolecules contain the D(+)-glucose residues bonded
through -(1-4)glycosidic linkages.
This eBook is protected by Copyright law and International Treaties. All rights are reserved. This book is covered by an End User Licensee
Agreement (EULA). The full EULA may be seen at http://www.library4science.com/eula.html.
90
Courtesy of ASTEC Inc.
Figure 54 The Molecular Structure of , , and Cyclodextrins
The mouth of the torus-shaped cyclodextrin molecule has a larger
circumference than at the base and is linked to secondary hydroxyl
groups of the C2 and C3 atoms of each glucose unit (see figure 55). The
primary hydroxyl groups are located at the base of the torus on the C6
atoms. As these hydroxyl groups are free to rotate, they partially block
the base aperture. The size of the cavity increases with increasing
number of glucose units (figure 54). The secondary hydroxyl groups can
be reacted with appropriate reagents to introduce further interactive
character to the cyclodextrin molecule. The very effective chiral
characteristics of the cyclodextrins structures arise from the many chiral
This eBook is protected by Copyright law and International Treaties. All rights are reserved. This book is covered by an End User Licensee
Agreement (EULA). The full EULA may be seen at http://www.library4science.com/eula.html.
91
centers they contain, for example, -cyclodextrin has 35 stereogenic
centers.
Courtesy of Supelco
Figure 55 A Molecular Model of Cyclodextrin
When the , or cyclodextrins are derivatized, the hydroxyl group on
the 2-position reacts first. However, the derivative is still size selective
and interaction will be determined by the size and functional groups
contained by the interacting molecule. Derivatizing the 6-hydroxyl
position has little or no effect on chiral selectivity but does enhance the
loading capacity of the stationary phase. This position is used for
anchoring the cyclodextrins to silica gel in the preparation of LC
stationary phases.
This eBook is protected by Copyright law and International Treaties. All rights are reserved. This book is covered by an End User Licensee
Agreement (EULA). The full EULA may be seen at http://www.library4science.com/eula.html.
92
The cyclodextrins have found a wide field of application in chiral
chromatography and there arc many applications in the literature The
following is a simple example of the use of this stationary phase.
In order to carry out in vivo pharmacological profiling of enantiomeric
drugs, the direct analysis of biological fluids is required to reduce
sample preparation time, and the chance of enantiomeric change.
Unfortunately, many chiral stationary phases, such as the Pirkle types
phases and the derivatized cellulose phases, demand the use of mobile
phases that are incompatible with the biological fluids. Haginaka and
Wakai [41] suggested that the silica should first be reacted with (3-
glycidoxypropyl)trimethoxysilane, to cover a significant part of the silica
surface with 'spacers', and then the remaining silanol groups reacted
with cyclodextrin–carbamoylated triethoxysilane to attach the chiral
agent.
Courtesy of J. Liq. Chromatogr. [Ref.17]
Figure 56. The Separation of the Hexobarbital Enantiomers
Contained in Blood Serum by Direct Injection on a Cyclobond
Phase
This eBook is protected by Copyright law and International Treaties. All rights are reserved. This book is covered by an End User Licensee
Agreement (EULA). The full EULA may be seen at http://www.library4science.com/eula.html.
93
Subsequent treatment, would convert the epoxy group to diols that
would insulate the analytes from the silica surface. Stalcup and Williams
[42] analyzed a series of these type of materials and found that there was
about 10 times as much spacer on the surface (ca. 2 mol/m2) as there
was derivatized cyclodextrin (ca. 0.2 mol/m2). This type of chiral
stationary proved to be very effective. The separation of the enantiomers
of hexabarbital on this stationary phase by the direct injection of blood
serum is shown in figure 56. Chromatogram A was obtained after 20
injections of serum and chromatogram B after 60 consecutive injections
of blood serum. It is seen that here is very little column deterioration and
that, although the tail of the major peak has become a little extended
after 60 injections, the column could still be used very effectively for the
analysis.
Liquid Chromatography Applications
Liquid chromatography has been used in an extremely wide range of
analyses and it is impossible to give a comprehensive set of examples
that would illustrate its wide applicability. The following are a few LC
analyses that may indicate the scope of the technique and give the reader
some idea of its importance and versatility.
An example of the use of reversed phase chromatography (employing a
C8 column) for the separation of some benzodiazepines is shown in
figure 57. The column used was 25 cm long, 4.6 mm in diameter
packed with silica based, C8 reverse phase packing particle size 5 .
The mobile phase consisted of 26.5% v/v of methanol, 16.5%v/v
acetonitrile and 57.05v/v of 0.1M ammonium acetate adjusted to a pH
of 6.0 with glacial acetic acid and the flow-rate was 2 ml/min.
The column efficiency available at the optimum velocity would be about
15,000 theoretical plates. The retention time of the last peak is about 12
minutes (i.e., a retention volume of 24 ml). At a flow rate of 2 ml/min.,
the mobile phase velocity will be well above the optimum and so the
maximum efficiency has not been realized. The general technique used
when there are more theoretical plates available than required is to
This eBook is protected by Copyright law and International Treaties. All rights are reserved. This book is covered by an End User Licensee
Agreement (EULA). The full EULA may be seen at http://www.library4science.com/eula.html.
94
increase the flow rate until the separation required is just realized. This
procedure trades efficiency for time and allows the separation to be
achieved in the minimum time given the column and phase system that
has been chosen.
Courtesy of Supelco Inc
.
Figure 57. The Separation of Eight Diazepines Employing a C8
Reverse Phase Column
The mobile phase is buffered appropriately to complement the
dissociation constants of the solutes. A mixture of methanol and
acetonitrile is employed, the acetonitrile being used to increase the
dispersive interactions in the mobile phase. The reason for the particular
solvent mixture is not clear and it would appear that the separation
might be achieved equally well by using a stronger solution of methanol
alone or a more dilute solution acetonitrile alone. There is no particular
This eBook is protected by Copyright law and International Treaties. All rights are reserved. This book is covered by an End User Licensee
Agreement (EULA). The full EULA may be seen at http://www.library4science.com/eula.html.
95
advantage to one solvent mixture over another except for the fact that
'waste' acetonitrile produces greater solvent disposal problems than
methanol. Another example of the use of reversed phase
chromatography for the separation of mixture of growth regulators (a
C18 reverse phase column) is shown in figure 58.
Courtesy of Supelco Inc.
Figure 58. The Separation of a Mixture of Growth Regulators on a
C18 Reverse Phase C18 Column
The packing is silica based but is contained in a short column 3.3 cm
long, 4.6 mm in diameter and packed with particles 3 m in diameter.
The expected efficiency of the column (when operated at the optimum
velocity) would be about 5,500 theoretical plates. This is not a
This eBook is protected by Copyright law and International Treaties. All rights are reserved. This book is covered by an End User Licensee
Agreement (EULA). The full EULA may be seen at http://www.library4science.com/eula.html.
96
particularly high efficiency and so the separation depends on the phases
chosen to provide the necessary selectivity and an appropriate gradient
program.
The selectivity was achieved using a complex mixture of ionic and
dispersive interactions between the solutes and the stationary phase and
ionic, polar and dispersive forces between the solutes and the mobile
phase. The initial solvent in the gradient program was a 1% acetic acid
and 1 mM tetrabutyl ammonium phosphate buffered to a pH of 2.8. The
tetrabutyl ammonium salt would be adsorbed strongly on the reverse
phase and thus acted as an adsorbed ion exchanger. During the program,
acetonitrile was added to the solvent and initially this increased the
dispersive interactions between the solute and the mobile phase.
As the acetonitrile concentration became higher, however, the tetrabutyl
ammonium salt would be desorbed from the reverse phase reducing the
ionic interactions of the solutes with the stationary phase. At even higher
concentrations of acetonitrile, the tetrabutyl ammonium salt would be
completely desorbed and the interactions of the solutes with the
stationary phase would become almost exclusively dispersive. This is an
example where a complex phase system was necessary because there
was limited column efficiency available. It is likely that a column with
intrinsically more efficiency might achieve the separation with a much
simpler solvent system and a more straightforward solvent program.
An example of the use of native silica is given by for the analysis of
Darvocet® and its generic equivalent formulation. Darvocet® is an
acetaminophen product in which the active ingredient (and other
material in the medicine) are weakly polar and, consequently, lend
themselves to separation on a strongly polar stationary phase such as
silica gel. The analysis is depicted in figure 59. The analysis is
completed in less than 4 minutes using a short column 3.3 cm long and
4.6 mm in diameter. The silica packing had a particle size of 3
providing a maximum efficiency of about 5,500 theoretical plates.
This eBook is protected by Copyright law and International Treaties. All rights are reserved. This book is covered by an End User Licensee
Agreement (EULA). The full EULA may be seen at http://www.library4science.com/eula.html.
97
In order to identify the impurities, the column had to be significantly
overloaded. Despite this, the impurities were well separated from the
main component and a substance was shown to be present in the generic
formulation that was not in the Darvocet®. The mobile phase was
98.5% dichloromethane with 1.5% v/v of methanol containing 3.3%
ammonium hydroxide. The ammoniacal methanol deactivated the silica
gel but the interaction of the solutes with the stationary phase would still
be polar in nature. In contrast solute interactions with the methylene
dichloride would be exclusively dispersive.
Courtesy of Supelco Inc.
Figure 59. The Analysis of Acetaminophen Formulations
Ion chromatography can be used in a number of novel ways and
employing the appropriate conditions can even be used to separate
mixtures where the components are not ionic or do not normally
produce interactive ions in aqueous solution. An example of this type of
separation is the analysis of saccharide mixtures using ion exchange
This eBook is protected by Copyright law and International Treaties. All rights are reserved. This book is covered by an End User Licensee
Agreement (EULA). The full EULA may be seen at http://www.library4science.com/eula.html.
98
interactions. An illustration of such a separation is given in figure 60.
The saccharides are reacted with a borate with which saccharides
readily forms complex anions. The procedure for making the complex is
simple and is achieve by merely including a borate buffer in the mobile
phase. The process is, in fact, a form of 'in-line' derivatization.
Courtesy of TOYO SODA Manufacturing Co., Ltd.
Figure 60. The Separation of a Saccharide Mixture by Ion
Exchange Chromatography
The column packing was a strong anion exchange resin designated as
TSKgel Sugar AXG. It had a particle diameter of 10 and contained
quaternary ammonium ions as the ion exchange moiety. The column
This eBook is protected by Copyright law and International Treaties. All rights are reserved. This book is covered by an End User Licensee
Agreement (EULA). The full EULA may be seen at http://www.library4science.com/eula.html.
99
was 15 cm long, 4.6 mm in diameter and had a potential efficiency of
about 7,500 theoretical plates. The mobile phase consisted of three
borate buffer solutions, which were used in a stepwise gradient. The first
buffer solution was a 0.5 M borate buffer (pH 7.7), the second a 0.7 M
borate buffer (pH 7.3) and the last a 0.7 M borate buffer (pH 8.7) and
the flow rate was 4 ml/min.
The last example clearly broaches a decidedly different approach to the
sample separation. It is just as feasible for the solutes to be modified to
suit a particular phase system as it is to choose or modify a phase system
to suit the solutes. This emphasizes the wide range of variables and
alternative approaches that liquid chromatography provides for the
analyst.
Before ending an example needs to be included that utilizes Micro-
reticulated polystyrene gels as a stationary phase. Micro-reticulated
polystyrene gels, as already discussed, are formed from cross-linked
styrene-divinyl benzene polymers and can be manufactured with a wide
range of different pore sizes. Micro-reticulated polystyrene gels,
however, are not used solely for the separation of large multi-interactive
molecules (e.g., protein and polypeptide mixtures), but can be used for
the same type of analyses as silica gel.
This eBook is protected by Copyright law and International Treaties. All rights are reserved. This book is covered by an End User Licensee
Agreement (EULA). The full EULA may be seen at http://www.library4science.com/eula.html.
100
Courtesy of Supelco Inc.
Figure 61. Separation of Some Phthalate Esters by Exclusion
Chromatography on Styrene-Divinyl Benzene Based Gel
Due to their method of manufacture, the pore volume of the micro-
reticulated polystyrene gels may be significantly lower than that of silica
gel. Consequently, some gels may have neither the peak capacity nor the
loading capacity normally experienced with silica. However, this is
generally not a problem in analytical LC. An example of the separation
of some phthalate esters is shown in figure 61. The column was 30 cm
long and 7.8 mm diameter. In exclusion chromatography, a large
diameter column is necessary to provide adequate peak capacity (see
Book 6). The particle diameter of the packing was 6 m and thus, at the
optimum velocity, an efficiency of about 25000 theoretical plates should
be produced. It is seen in figure 61 that the dead volume time is about
19 minutes and a flow-rate of 1.0 ml/min this would be equivalent to a
dead volume of 19.5 ml.
Courtesy of Asahipak Inc.
This eBook is protected by Copyright law and International Treaties. All rights are reserved. This book is covered by an End User Licensee
Agreement (EULA). The full EULA may be seen at http://www.library4science.com/eula.html.
101
Figure 62. The Separation of a Standard Protein Mixture by
Exclusion Chromatography on a Vinyl Alcohol-Styrene Co-
Polymer Hard Gel
A more typical application of micro-reticulate resins in exclusion
chromatography is shown by the separation of a standard protein
mixture depicted in figure 62.
The column was 50 cm long and 7.6 mm wide giving a total column
volume of about 22.7 ml. This conflicts with the value for the dead
volume, which appears to be about 22 ml (the retention time of 22 min.
at a flow rate of 1 ml/min.). Consequently, it would appear that the resin
occupied no volume in the column. This paradox could be explained on
the basis that the separation was not achieved solely by exclusion. The
dispersive and polarizable nature of the resin was contributing dispersive
and possibly slight polar interactions with the solutes increasing their
retention beyond that expected from exclusion alone. This would be
supported by the fact that the mobile phase contained no solvent and
consisted of a buffer solution containing 0.1 M sodium phosphate and
0.3 M sodium chloride (pH 7.0), which would have minimum dispersive
interactions with the solute. In contrast, the hydrocarbon matrix of the
styrene gel would be capable of exerting strong dispersive interactions
with the dispersive groups present in the proteins. The use of mixed
retention systems to achieve the separation is perfectly acceptable and
occurs in most distribution systems. The stationary phase was GFA-50,
the number 50 referring to the columns length. The pore size was
defined in terms of the minimum molecular weight of a totally excluded
solute which was given as 300,000 and the material was specifically
prepared for exclusion chromatography.
The analyst should be aware that generally, when working with large
molecules (particularly proteins), column efficiencies may, under some
circumstances, be as little as 10% of the expected value and this must
be taken into account when choosing the column and the phase system.
References
This eBook is protected by Copyright law and International Treaties. All rights are reserved. This book is covered by an End User Licensee
Agreement (EULA). The full EULA may be seen at http://www.library4science.com/eula.html.
102
1. M.S.Tswett, Tr .Protok . Varshav . Obshch . Estestvoispyt Otd .
Biol . 14(1905).
2. R. Willstatter and A. Stoll, Utersuchungenuber Chlorophy ,
Springer, Berlin 1913
3. A. J. P. Martin and R. L. M. Synge , Biochem . J . 35(1941)1358.
4. A. Tiselius and D. Claesson, Ark. Kemi. Mineral. Geol.
15B(No 18)(1942).
5. T. E. Beesley and R. P. W. Scott, Chiral Chromatography, John
Wiley and Sons, Chichester-New York, (1998)39.
6 R. P. W. Scott and T. E. Beesley, Analyst, 124(1999)713
7. R. P. W. Scott, Analyst, 124(1999).
8. E. Katz and R. P. W. Scott, J. Chromatogr., 253(1982)159.
9. R.P.W.Scott, Chromatography Detectors: Design. Function and
Operation, Marcel Dekker Inc., New York-Basle (1997).
10. G.J.Schmidt and R.P.W.Scott Analyst, 110(1985)757.
11. G.J.Schmidt and R.P.W.Scott, U.S. Patent 4,555,936 (1983)
12. K. K. Unger, Angew . Chem . Int . Engl . 11(1972)267.
13. M. Le Page, R. Beau, and J. Duchene, Fr. Pat.,
No 1.473,240 (1967).
14. M. Le Page and A. de Vries, Fr. Pat., No 1.475,929 (1967).
15. K. Unger and B. Scharf, Ger.. Pat., No 2,155,281 (1971).
16. D. W. Sindorf and G. E. Maciel, J. AM. Chem. Soc.
105(1983)1487.
17. D. W. Sindorf and G. E. Maciel, J. Phys. Chem. 87(1983)5516.
18. J.M.Vleeskens, Thesis, University of Technology , Delft,
The Netherlands,(1959).
19. J.H.de Boer and J.M.Vleeskens,Koninkl.ned.Akad.Wet.,
Proc.B61,2,(1957)85.
20. J.H.de Boer,J.H.Linsen and C.Okkerse,Koninkl.ned.Akad.Wet.,
Proc.B63,(1958)360.
21. J. Erkelens and B. G. Linsen J . Colloid Interface Sci .,
29(1969)464.
22. R. P. W. Scott and S. Traiman, J. Chromatogr .,196(1980)193.
23. K.K.Unger,"Porous Silica", J.Chromatogr. Libr.16(1979)99.
24. C. F. Simpson, U K Patent 1,310,8712,(1975).
25. C.H.Lochmüller and D.R.Wilder,J.Chromatogr. Sci.,
This eBook is protected by Copyright law and International Treaties. All rights are reserved. This book is covered by an End User Licensee
Agreement (EULA). The full EULA may be seen at http://www.library4science.com/eula.html.
103
17(1979)574.
26. R.K.Gilpin and J.A.Squires, J.Chromatogr. Sci.,17(1979)574.
27. R.P.W.Scott and C.F.Simpson, J.Chromatogr. 197(1980)11.
28. R.P.W.Scott, J.Chromatogr., 142(1977)213.
29. R. Kunin, E. F. Meitzner and N. Bostnick, J. Am. Chem. Soc.
84(1962)84.
30. J. R. Millar, D. G. Smith, E. E. Marr and T. R. E. Kressman,
J. Chem. Soc.,218(1963).
31. R. Kunin, E. F. Meitzner, J. A. Oline, S. A. Fischer and N.W.
Frisch, Ind. Eng. Chem. Prod. Res. Develop.,1(1962)140.
32. K.K.Unger,"Porous Silica", Elsevier Scientific Publishing
Company, Amsterdam-Oxford-New York (1979)187.
33. S.G.Glasstone,"Textbook of Physical Chemistry, D.Van Nostrand,
New York,(1946)1198.
34. R. J. Laub and J. H. Purnell, J. Chromatogr., 112(1975)71.
35.E. D. Katz, K. Ogan and R. P. W. Scott, J.Chromatogr.
352(1986)67.
36. E. D. Katz, C. H. Lochmüller and R. P. W. Scott, Anal Chem.
61(1981)344.
37. W. H. Pirkle and D. W. House, J. Org. Chem., 44(1979)1957.
38. Y. Okamato, Chem. Lett., (1986)1237.
39. D. W. Armstrong, Y. Tang, S. Chen, Y. Zhou, C. Bagwill, and
J. R. Chen, Anal. Chem., 66(9)(1994)1473.
40. K. H. Ekborg-Ott, Chirality, (1998).
41. J. Haginaka and J. Wakai, Anal. Chem., 62 (1990)997.
42. A. M. Stalcup and K. L. Williams, J. Liq. Chromatogr.,
15(1)(1992)29.