Posted by : kaushik zala
Friday, December 16, 2011
|
High-Performance
Liquid Chromatography (HPLC) is one of the premier analytical techniques
|
|
widely
used in analytical laboratories. Numerous analytical HPLC analyses have been developed
|
|
for
pharmaceutical, chemical, food, cosmetic and environmental applications. The popularity
of
|
|
HPLC
analysis can be attributed to its powerful combination of separation and quantitation
|
|
capabilities.
HPLC instrumentation has reached a state of maturity such that the majority of
|
|
vendors
are capable of supplying highly automated and sophisticated systems to meet user
needs.
|
|
In
order to provide a high level of assurance that the data generated from the HPLC
analysis are
|
|
reliable,
the performance of the HPLC system should be monitored at regular intervals. In
this
|
|
article,
some of the key performance attributes for a typical HPLC system (consisting of
a
|
|
quaternary
pump, an auto-injector, a UV-Visible detector, and a temperature-controlled column
|
|
compartment)
will be discussed.
|
|
The
overall performance of the HPLC system can be evaluated by examining the key functions
|
|
of
the different modules that comprise the system, followed by a holistic testing,
which tests the
|
|
performance
of the LC components as an integrated unit for its intended use. Modular testing
can
|
|
provide
specific information related to the performance of the individual components of
the LC.
|
|
Information,
such as the wavelength accuracy of the UV detector and the gradient accuracy of
|
|
the
pump, cannot be obtained by holistic testing alone. The holistic test can be as
simple as
|
|
running
a frequently used HPLC method in the operating laboratory. This frequently used
|
|
method
can also be used as a means to compare the overall performance of different HPLC
|
|
systems
in the laboratory. The common performance attributes for each HPLC module, and
the
|
|
general
expectations for each, are listed in Table 1
.
|
|
Table
1: Performance Attributes for HPLC Modules
|
|
One
of the key performance requirements for the pump module is the ability to maintain
accurate
|
|
and
consistent flow of the mobile phase, which will be necessary to provide stable
and repeatable
|
|
interactions
between the analytes and the stationary phase. Poor flow-rate accuracy will affect
|
|
the
retention time of the separation.
|
 The flow-rate accuracy of the
pump can be evaluated simply by calculating the time required to
|
|
collect
a predetermined volume of mobile phase at different flow-rate settings. For example,
the
|
|
flow-rate
accuracy at 2 mL/min. can be verified by using a calibrated stopwatch to measure
the
|
|
time
it takes to collect 25 mL of effluent from the pump into a 25 mL volumetric flask.
The
|
|
typical
acceptance of the flow-rate accuracy is listed in Table 1.
|
|
Gradient Accuracy and Linearity
|
|
When
it comes to gradient analysis, the ability of the pump to deliver the mobile phase
at various
|
|
solvent
strengths over time by varying the composition of the mobile phase accurately
in linear
|
|
steps
is crucial to achieve the proper chromatographic resolution and reproducibility.
|
|
Even
though lower-pressure gradient LC pumps are usually equipped with quaternary
|
|
proportioning
valves, which can handle up to four solvents, typical low- and high-pressure
|
|
gradient
runs involve two solvent systems. The accuracy and linearity of the gradient solvent
|
|
delivery
can be verified indirectly by monitoring the absorbancechange as the binary
|
|
composition
of the two solvents changes from two different channels. For example, an LC
|
|
gradient
has four channels: A, B, C and D. The test will be performed for two channels
at a time.
|
|
Channel
A is filled with a pure solvent such as methanol, while channel B is filled with
a solvent
|
|
containing
a UV-active tracer such as caffeine. The gradient profile is programmed to vary
the
|
|
composition
of the mixture from 100% A to 100% B in a short period of time, and changed back
|
|
to
100% A in a stepwise manner (See Fig. 1). The absorbancechange from 100% A (baseline)
to
|
|
100%
B is measured and expressed as height H in the plot of absorbanceversus solvent
|
|
composition.
As the percentage of solvent B decreases in the solvent mixture, the UV
|
|
absorbanceof
the mixture should decrease accordingly. If the composition of the 20% A and 80%
|
|
B is
accurate, the height B1, which corresponds to the absorbanceat 80% B, should be
close to
|
|
80%
of H. Similarly, accuracy verifications can be determined at 60%, 40%, 20% and
0% B. The
|
|
linearity
of the gradient delivery can be verified by plotting the absorbanceat various
mobile-
|
|
phase
compositions versus the theoretical composition. The entire process can be repeated
for
|
|
Figure
1: Gradient accuracy and linearity measurement
|
 Pressure Test
|
|
The
performance of the LC pump depends on the proper functioning of the check valves
and the
|
|
proper
connection of the tubing. Properly functioning check valves and tubing connections
|
|
(seals)
are important in maintaining stable mobile-phase flow and system pressure. For
pump
|
|
systems
that output the pressure reading in the pump head over time, a simple pressure
test can
|
|
be
a useful qualitative test to check the condition of the check valves and to determine
whether
|
|
or
not there are any leaks in the system. The first step of the pressure test is
to plug the outlet of
|
|
the
pump using a dead-nut and by setting the automatic pump shutdown pressure to 6,000
psi.
|
|
The
pump-head pressure signal output is connected to a recorder. Pressurize the pump
by
|
|
pumping
methanol at 1 mL/min. The pressure inside the pump head increases quickly as the
|
|
outlet
of the pump is blocked. As the pressure increases to about 3,000 psi, the flow
rate is
|
|
reduced
to 0.1 mL/min. The pressure will gradually rise to the shutdown pressure if the
check
|
|
valves
are able to hold the mobile phase in the pump chamber as would be normally expected
|
|
(Fig.
2). If the check valve is not functioning properly, the pressure will fluctuate
at about 3,000
|
|
psi
instead of reaching the shutdown pressure. The pressure in the pump head decreases
slowly
|
|
over
time after the automatic shutdown. A steep decrease in pressure over time implies
poor
|
|
check-valve
performance or leaks within the pumping system.
|
|
Figure
2: Pressure test of the pump module
|
|
The
ability of the injector to draw the same amount of sample in replicate injections
is crucial to
|
|
the
precision and accuracy for peak-area or peak-height comparison for external standard
|
|
quantitation.
If the variability of the sample and standard being injected into the column is
not
|
|
controlled
tightly, the basic principle of external standard quantitation is seriously compromised.
|
|
No
meaningful comparison between the responses of the sample and the standard can
be made.
|
|
The
absolute accuracy of the injection volume is not critical as long as the same
amount of
|
|
standard
and sample is injected.
|
|
The
precision of the injector can be demonstrated by making at least six replicate
injections from
|
|
a sample.
The relative standard deviation (RSD) of the response of the injections is then
|
|
calculated
to evaluate the precision.
|
|
Most
of the automated LC injectors are capable of varying the injection volume without
|
|
changing
the injection loop. Variable volume of sample will be drawn into a sample injection
|
|
loop
by a syringe or other metering device. The uniformity of the sample loop and the
ability of
|
|
the
metering device to draw different amounts of sample in proper proportion will
affect the
|
|
linearity
of the injection volume. The linearity is important for methods that require the
use of
|
|
variable
injection volumes, such as the high-low method in the quantitation of impurities.
|
|
The
linearity of the injector can be demonstrated by making injections to cover a
range of 0 to
|
|
100
µL. The response of the injections at each
injection volume is plotted against the injection
|
|
volume.
The correlation coefficient of the plot will be used in the evaluation of the
injection
|
|
Carryover (Not all vendors do
this test because it is very dependent on the analyte.)
|
|
Small
amounts of analyte may get carried over from the previous injection and contaminate
the
|
|
next
sample to be injected. The carryover will affect the accurate quantitation of
the subsequent
|
|
sample.
The problem is more serious when a dilute sample is injected after a concentrated
|
|
sample.
In order to avoid cross contamination from the previous sample injection, all
the parts in
|
|
the
injector that come into contact with the sample (the injection loop, the injection
needle and
|
|
the
needle seat) have to be cleaned effectively after the injection. The effectiveness
of the
|
|
cleaning
can be evaluated by injecting a blank after a sample that contains a high concentration
|
|
of
analyte. The response of the analyte found in the blank sample expressed as a
percentage of
|
|
the
response of the concentrated sample can be used to determine the level of carryover.
|
|
UV-Visible Detector Module:
|
|
Wavelength
accuracy is defined as the deviation of the wavelength reading at an absorption
or
|
|
emission
band from the known wavelength of the band. The detrimental effects of wavelength
|
|
deviation
on the qualitative and quantitative UV-Vis measurements have been discussed in
detail
|
|
previously
in an article on the performance of UV-Vis spectrophotometer (Laboratory Focus-
|
|
Gazette Edition, April
2000, pg. 8). In short, the accuracy and sensitivity of the measurement
|
|
will
be compromised if there is a wavelength accuracy problem.
|
|
There
are many ways to check the wavelength accuracy of a UV-Vis detector. For the built-in
|
|
wavelength
verification, the deuterium line at 656 nm and the absorption bands at 360, 418,
453
|
|
and
536 nm in a holmium oxide filter are often used. The deuterium line and the holmium
oxide
|
|
bands
are easy to use, but are restricted to the visible range. The wavelength verification
of the
|
|
UV
range, where most quantitative analysis is done, is performed by filling a flow
cell with a
|
|
solution
of a compound with a well-known UV absorption profile, and scanning the solution
for
|
|
absorption
maxima and minima. The ëmax or
ëmin from the scan profile is then compared to the
|
|
known
ëmax or ëmin of
the compound to determine the wavelength accuracy. Solutions of
|
|
potassium
dichromate in perchloric acid and holmium oxide in perchloric acid can be used.
|
|
However,
these acidic solutions are difficult to work with as the flow cell has to be thoroughly
|
|
cleaned
after the measurement to remove any traces of fluorescence from the potassium
|
|
dichromate
solution. Aqueous caffeine solution, which is easy to prepare and handle, with
ëmax at
|
|
272
nm and 205 nm, and ëmin at
244 nm, can also be used.
|
|
Since
the analytes of interest may vary in concentration, the ability of a detector
to produce a
|
|
linear
response to concentration variation within a reasonable range is crucial to the
accuracy for
|
|
peak-area
and peak-height comparison between standards and samples. The linearity of the
|
|
detector
response can be checked by injecting or by filling the flow cell with a series
of standard
|
|
solutions
of various concentrations. Aqueous caffeine solutions are convenient for the linearity
|
|
measurement.
The concentration range typically should generate responses from zero to at least
|
|
1.0
AU. Absorbencies beyond 1.5 AU are more prone to deviation due to stray light.
From the
|
|
plot
of response versus the concentration of the solutions, the correlation coefficient
between
|
|
sample
concentration and response can be calculated to determine the linearity.
|
|
Noise and Drift (Not
all vendors perform this test. Older systems may not be able to meet the
|
|
same
signal-to-noise ratio specified for the new equipment.)
|
|
Electronic,
pump and photometric noise, poor lamp intensity, dirty flow cell, and thermal
|
|
instability
contribute to the overall noise and drift in the detector. Excessive noise can
reduce the
|
|
sensitivity
of the detector and hence affect the quantitation of low-level analytes. Detector
drift
|
|
may
affect the baseline determination and peak integration. Many procedures for detector
noise
|
|
and
drift estimation are based on the ASTM (American Society for Testing and Material)
|
|
Method
E 685. Nowadays, most chromatographic software is capable of calculating the detector
|
|
noise
and drift. Typically, the detector should be warmed up prior to the test, and
any
|
|
temperature
fluctuations should be avoided during the test. For a dynamic testing condition,
|
|
methanol
is passed through the flow cell at 1 mL/min. A backpressure of about 500 psi is
|
|
maintained
to prevent bubble formation.
|
|
In
reality, the performance of the LC system will deteriorate over time. If the performance
|
|
verification
tests do not pass the predetermined acceptance criteria, an impact assessment
should
|
|
be
done to evaluate the effect of the failure on the quality of the data generated
by the system.
|
|
The
impact assessment should cover all the analyses done on the system since the last
|
|
performance
verification, as there is no effective way of determining when the failure occurred.
|
|
The
system suitability data generated together with the analyses will be very useful
in
|
Great web page! I don`t imagine I`ve seen every one of the angles of this theme the way in which you`ve pointed them out. You`re a accurate star, a rock star guy. You`ve got a great deal to say and know so much about the subject that i think you ought to just teach a class about it ....uv detector
ReplyDelete