A Practical Guide To HPLC Merck PDF
A Practical Guide To HPLC Merck PDF
High Performance
Liquid Chromatography
intensity
tR3
tR2
tR1
• Adsorption chromatography: retention of analytes
by reversible binding on a stationary phase (e.g.,
normal phase; for example allows for the separation
of isomers by specific adsorption).
• Partition chromatography: retention by reversible
t0
dissolution in a stationary (quasi-) liquid 3-dimensional
layer; different solubility or different partition coeffi-
cients of samples in liquid stationary and liquid mobile
phase, e.g., in reversed phase (RP) LC, hydrophilic t’R3 time
interaction liquid chromatography (HILIC). t’R2
t’R1
• Ion exchange chromatography: retention by Figure 1. Schematic drawing of a chromatogram displaying the isocratic
electrostatic interaction with the stationary phase. separation of three analytes.
• Size exclusion chromatography: retention by • The correlation between column back pressure and the
permeation into the pores of a stationary phase and particle size of a particulate column can be calculated
steric exclusion of analytes (e.g., gel permeation from:
chromatography, gel chromatography, gel filtration).
ηFL
Δp =
The following information is generated during a chro- K0 π r2 dp2
matographic run (Figure 1):
Δp change in pressure (bar)
• Back pressure p (bar or psi) is determined during the
η eluent viscosity (units?)
analysis for the HPLC utilized and is a combination of
pressures generated by the column and the system. F flow rate (mL/min)
K0 column permeability (units?)
• Dead time or hold-up time t0 (s) of a compound not
r column radius (mm)
being retained under the given chromatographic
conditions (i.e. a compound that is not interacting dp particle diameter (µm)
with the stationary phase). Often also referred to
as tm. • Linear velocity or linear flow rate u (mm/s):
1
• Net retention time or reduced retention time t’R: • Plate height H (μm) corresponds to the quotient of L
(in µm) and N (L is the length of the column utilized
t’R = tr – t0 for the determination of N).
• Chromatographic resolution R can be calculated from:
• Capacity factor or retention factor k is calculated for
every analyte utilizing the respective retention time R = 2 [(tB – tA)/(WA + WB)]
and the dead time of the separation:
It is a measure of the quality of a separation of two
k = (tR – t0)/t0 analytes A and B (baseline separation has been achieved
when R ≥ 1.4).
• Selectivity coefficient or separation factor α describes
the quality of a separation of two compounds A and B: Classical HPLC columns (particle diameter 5 μm) are
normally used at rather high flow rates and above the
α = kB/kA minimum of their van Deemter plot (see details below).
As a result of this, the resolution of such a setup can
The selectivity of a chromatographic column is dictated be improved by a decreased flow rate (increase of tr).
by the physical properties of the eluent and the sta- By contrast, UHPLC (Ultra High Performance Liquid
tionary phase (modified or unmodified). Chromatography) columns (particle diameter less than
2 μm) are operated at comparably low flow rates, and
• Plate count N is calculated from: at the minimum of of their van Deemter plot. In this
case, an increase of flow rate will improve resolution.
N = 16 (tR/w)2 (US Pharmacopoeia—USP—figure 2)
• Peak symmetry is described by the symmetry factor
and describes the performance of a column in isocratic or tailing factor AS (or Tusp) and is utilized by USP (USP
mode. It is given as an absolute value (with respect to chapter 621, see Figure 3):
a specific column dimension) or per meter of a column
(unit: m-1). As = (A + B)/2A or As = 0.5 (1 + B/A)
Alternatively N can be calculated from: A and B are determined at 5% of the peak height.
N = 5.54 (tR/w1/2)2 (Figure 2) Alternatively the peak shape can be described by the
asymmetry factor calculated from the ratio of B and A.
The base line width w, and the peak width at half In this case B and A are determined at 10% of the
height w1/2 are determined by the software utilized. peak height.
The plate count depends on column length, geometric For symmetrical peaks, the symmetry or asymmetry
properties of the stationary phase (particle size dis- factor is equal to 1. Peak tailing is observed when the
tribution or monolithic skeleton diameter distribution), factor is >1, whereas peak fronting will be visible when
column packing parameters, eluent flow rate and proper the factor is <1.
setup of the system used (avoidance of dead volume,
see also chapter 2).
tangent line
0.05 h
A B
tR baseline
h w1/2
Figure 3. Schematic drawing of data acquired for the calculation of peak
symmetry AS. A and B are determined at 5% of the peak height h.
tR
w baseline
2
The shape of a chromatographic peak is ideally narrow The correlation between flow rate, plate height and the
and symmetric, but in a real system, peak broadening three terms A, B, and C is described by the van Deemter
(and peak asymmetry) can be caused by the column, or plot (Figure 5) and the following equation:
from the tubing and connectors. The following phenomena
H = A + B/u + C*u
are responsible for this type of peak deterioration
(Figure 4): H Height equivalent to theoretical plate (column length/efficiency)
• Eddy diffusion (A term): Mixing of the sample with A Eddy diffusion
surrounding mobile phase; peak broadening by different B Longitudinal diffusion
analyte flow paths/path lengths through the column
C Resistance to Mass Transfer
bed. Eddy diffusion is nearly independent from the
flow rate and is influenced by particle size and size u Mobile phase linear velocity
distribution, as well as by the homogeneity of the
column packing.
plate height H
• Longitudinal (axial) diffusion (B term) along both column
axis and sample concentration gradient. This effect
is rather weak and only strong at low (and therefore
impractical) flow rates of the mobile phase. mass transfer
• Mass transfer (C term) between stationary and mobile van Deemter plot
phase caused by repeated interaction of sample mole-
cules with the stationary phase. The strength of this
effect is proportional to the flow rate of the mobile eddy diffusion
phase.
axial diffusion
20
The contribution of axial diffusion to peak broadening is
influenced by the viscosity of the mobile phase and the 15
time of analysis.
Peak broadening caused by mass transfer increases with 10
increasing flow rate, larger particle size (larger depth
of pores) and increased viscosity of the mobile phase 5
(decreased velocity of diffusion).
0
0 1 2 3 4 5 6
Linear flow rate u (mm/s)
Figure 6. Van Deemter plots for Purospher™ STAR RP-18 endcapped
10 cm x 2.1 mm I.D. columns (top: 3 μm particle size, bottom: 2 μm)
using anthracene as an analyte.
Chromatographic conditions: Detection: UV (254 nm); mobile phase:
acetonitrile/water 75/25 (v/v); sample: anthracene. The capacity factor
is 4.1 (3 μm particle size) and 4.5 (2 μm particle size), respectively.
3
Columns: Discovery™ C18,
5.0 cm x 4.6 mm,
1.
2.
Barbital
Phenobarbital Quantification experiments in
analytical HPLC
5 µm particles 3. Butabarbital
Mobile Phase: CH3OH:H2O 4. Mephobarbital
1 5. Pentabarbital
(45:55)
Temp.: 25 °C 6. Secobarbital
Det.: UV, 214 nm The task of analytical HPLC is the chromatographic sepa-
ration of analytes within a sample and their subsequent
2
4 mL/min
1 2
3
identification. Next to these qualification experiments, the
4
quantification of compounds, for example in trace analysis,
1.0 2.0 3.0 4.0 5.0 6.0 7.0
3 mL/min
can be an additional challenge.
5
2 mL/min
The aim of setting up a proper chromatographic system
used in quantification experiments has to be a maximi-
1.0 2.0 3.0 4.0 5.0 6.0 7.0
Time (min)
both optimizes the sensitivity of a separation and hence
Figure 7: Influence of Flow rate on resolution. increases the signal to noise ratio (S/N). The former
can best be achieved by utilizing short columns with the
smallest possible internal diameter (I.D.) (Table 1) in
In practice, a high flow rate speeds up separations signifi- combination with high injection volumes and high flow
cantly, but can result in loss of resolution as demonstrated rates in gradient mode. Of course, packing quality has
with a 5 µm particulate column in Figure 7. to be maintained as consistently high and independent
In Figure 8, the detailed view shows a significant loss of from column I.D.. At the same time, minimization of
resolution at a flow rate of 4 mL/min in comparison to background noise can be ensured by using the purest
a flow rate of 1 mL/min using the same HPLC column solvents and reagents and using proper sample prepa-
(Discovery® C18 5 µm, 50 x 4.6 mm) ration and handling.
Table 1. Column I.D., typical flow rate and calculated relative sensitivity.
Smaller particles show a very flat Van Deemter plot,
which makes them better suitable for fast separations Column I.D. Typical flow rate Relative
using UHPLC. (mm) (µL/min) sensitivity
4.6 1000–6000 1
2.0 200–800 5.3
0.2 0.5–20 530
0.1 0.4–3 2100
0.05 0.1–0.8 8500
4 mL/min
1.0
In qualitative analysis, the intensity of the smallest
detectable analyte signal should be three times higher
1 mL/min than the noise signal, while in quantitative analyses, the
1.0 2.0 3.0 4.0 5.0 6.0 7.0
ratio should be approximately 10:1. In this context, the
Time (min)
limit of quantification (LOQ) is referred to as the lowest
Figure 8. Loss of Resolution Under Non-Optimal Flow Rate amount of analyte allowing for a quantitative analysis
with predefined accuracy (low deviation between “true”
and detected value), and repeatability (repeated analysis
In addition to column-related peak broadening, wider
on the same instrument, with the same settings and
peaks can also be caused by effects attributed to the LC
with a low standard deviation). See also the chapter
system used. These effects can be due to axial diffusion
on method development. The limit of detection (LOD)
within the instrument, hydrodynamic flow profiles being
is calculated at 30% of LOQ. As an example, Figure 9
inconsistent across the tubing cross section, or turbulences
combines the chromatogram and the calibration curves
or mixing at corners, edges or in dead volumes.
from the separation of seven pesticides on a monolithic
silica capillary column. LODs were obtained from a serial
dilution of the analyte mixture.
4
4.9E+07 5.0E+06
R2 = 0.9990
2.9E+07 3.0E+06
1.9E+07
2.0E+06
1.0E+06
9.0E+06
0.0E+00
-1.0E+06
0 1 2 3 4 5 6 7 8 9
0 1 2 3 4 5 6 7
Retention time (minutes) Injected mass (pg)
4.0E+06 open circles, left) and metolachlor (peak 7, closed triangles, right). The
insets show the complete concentration range covered (including the
non-linear region typical for ion trap overload) while the large diagrams
3.0E+06 display the linear regions of the curves. The limits of detection for the
setup utilized are 0.24 pg (0.16 ng/μL) for metamitron and 0.59 pg
(0.40 ng/μL) for metolachlor.
2.0E+06
Chromatographic conditions: Chromolith® CapRod® RP-18 endcapped
15 cm x 0.1 mm I.D. monolithic silica capillary column. Base peak
1.0E+06 chromatogram (m/z 200–290). Mobile phase A: water + 0.1% formic
acid, mobile phase B: acetonitrile + 0.1% formic acid; gradient: 20% B
to 80% B in 10 minutes.
0.0E+00
0 1 2 3 4 5 6 7
Injected mass (pg)
5
2. System setup and settings
Every single HPLC system has its own characteristics,
and these can change over time—either due to wear
of instrument parts, or by a variation of the setup
by the operator. The following chapter describes
important system parameters and their influence on
Intensity
chromatographic results and gives recommendations
for system optimization. Change from
solvent A 1/ hmax
Dwell volume and dead volume to B 2
6
The internal diameter of the connecting tubing that can Detector response time
be used in a chromatographic setup depends on column
properties (e.g. back pressure issues), chromatographic Most HPLC detectors have a variable response time or
conditions applied (eluent viscosity, flow rate, temper- time constant. If the response time is too slow (the time
ature), as well as the desired performance and resolution. constant is then too long, e.g., 2 seconds), peaks may
It also must be bigger when matrix rich samples are appear broad and show tailing. For narrow peak widths,
injected (at least between injector and column). In good integration of the peak area, and good optical
general, the smallest possible tubing i.d. should be used presentation of the chromatogram, the data system
in order to gain the highest column efficiency. For nano settings must enable approximately 20 data points to be
and capillary-LC (column i.d. 50–500 μm) a tubing i.d. of acquired during the peak width time. Therefore columns
0.064 mm should be chosen, whereas in micro, narrow- require a fast detector time constant, such as 0.01 seconds
bore, and conventional analytical LC (column I.D. (100 Hz, i.e. 100 data points per second, which means
1–4.6 mm) a tubing i.d. of 0.13 mm is recommended. In 1 data measurement in 10 milliseconds). As a rule of
addition, a detector with a micro flow cell is suggested thumb, the time constant should be set to 1/10 of the peak
(check pressure tolerance of the detector cell). Larger width. Always choose the lowest response time possible.
detector cells can be used, but peak widths will be By reducing the time constant from 2 to 0.1 seconds, the
wider (or the intensity will be lower; see Figure 12 and plate count can be improved significantly (Figure 13).
Table 2). In the given examples, a change from an Keep in mind that a decreased time constant improves
analytical to a micro detector cell leads to an increase in peak shape and column performance, but that at the same
performance of 20–25%. time the baseline noise is increased (because of less signal
averaging). This plays a role when performing quantitative
50
analyses, and when high sensitivity is important. Under
these prerequisites, there will be a tradeoff between high
Intensity (mAU)
50
20
40
Intensity (mV)
10
30
0
0 1 2 3 4 5 6 20
Retention time (minutes)
10
50
0
Intensity (mAU)
40
0.0 0.4 0.8 1.2 1.6 2.0
Retention time (min)
30
20 300
250
Intensity (mV)
10
200
0
150
0 1 2 3 4 5 6
Retention time (minutes) 100
Figure 12. Influence of UV detector cell volume and column type on
the peak intensity and peak width. 1.4 μL UV micro cell (green line), 50
11 μL UV analytical cell (yellow line). Top: Purospher™ STAR RP-18
endcapped (3 µm) Hibar® HR 5 cm x 2.1 mm I.D., bottom: Purospher™ 0
STAR RP-18 endcapped (3 µm) LiChroCART® HR 5.5 cm x 2.0 mm I.D.. 0.0 0.4 0.8 1.2 1.6 2.0
Chromatographic conditions: Mobile phase: acetonitrile/water 70/30 (v/v),
flow rate: 0.21 mL/min, detection 254 nm, injection 0.2 µL, sample
Retention time (min)
(in order of elution): uracil 12 mg/L, ethylbenzene 839 mg/L,
Figure 13. Setting the proper detector response time (top: 2 s,
propylbenzene 922 mg/L, butylbenzene 1006 mg/L.
bottom: 0.1 s) in fast separations.
Chromatographic conditions: Chromolith Performance RP-18 endcapped
Table 2. Change of chromatographic parameters when switching from
10 cm x 4.6 mm I.D. HPLC column, mobile phase: acetonitrile/water
an 11 μL to a 1.4 μL UV cell (example: butylbenzene peak).
40/60 (v/v), flow rate: 5.0 mL/min, injection volume 10 μL, detection
Cell Plate Plate 254 nm, cell: 11 μL, back pressure 55 bar, sample (in order of elution):
Column volume/μL Tusp count/m count/% uracil, pyridine, aniline, 4-ethylaniline, benzene.
Purospher™ STAR RP-18e 1.4 1.15 43.000 119
(3 µm) Hibar® HR
5 cm x 2.1 mm I.D. 11 1.29 36.000 100
Purospher™ STAR RP-18e 1.4 1.24 44.000 126
(3 µm) LiChroCART® HR
5.5 cm x 2.0 mm I.D. 11 1.31 35.000 100
7
Sample injection, mass and Mass overload effects depend on the sample complexity,
solubility, and retentivity and are commonly observed
volume overload when trace compounds are analyzed in complex mixtures.
Quantitative analyses need a careful system setup, as In this situation, a desirable increase in sensitivity is
an improper combination of injection needle and sample achieved by simply increasing the injection volume. Con-
vial, as well as a high sample viscosity can cause false sequently, a mass overload with the main component is
sample drawing. For example, a combination of a high most likely. As long as the peak shape of trace compounds
draw speed and a highly viscous sample can cause sample is not negatively affected and no overlap of the trace and
underdosing, because filling of the injector needle with the main component peaks is observed, this phenomenon can
sample solution can be somewhat slower in comparison to be tolerated. A visible phenomenon of mass overload is
the draw speed of the dosing syringe. In this situation, the the detection of triangular shaped peaks displaying tailing
draw speed has to be decreased in order to enable proper or fronting. In fact, the effect of mass overload on peak
sampling. A similar error can occur when the septum of shape is influenced by the interaction between analyte,
the sample vial is completely airtight and large sample mobile phase, and stationary phase and is a direct result
volumes (e.g., pharma samples) need to be taken. The of the adsorption isotherm of a chromatographic system
use of split septa, specially designed needles or puncture (analyte/mobile phase/stationary phase). The dependency
offset tuning via the system software can be of help in of the sample concentration in the mobile phase on the
this case. sample concentration in the stationary phase is described
by a Langmuir or anti-Langmuir isotherm, where peak tail-
Make sure that the autosampler syringe is free of any ing (Langmuir behavior) or peak fronting (anti-Langmuir)
air bubbles in order to avoid any negative effects on the can be observed. Depending on the characteristics of the
drawn sample volume. In order to avoid air bubbles, detector, mass overload can lead to a detector overload,
flush the syringe manually and always degas washing resulting in flat-topped peaks. This can compromise the
solution prior to use. determination of plate count or resolution, for example.
The stability of, e.g., a UV detector signal varies and In such a situation the linear range of the detector can
is influenced by both temperature (temperature drift) be increased (and an overload can be avoided) when the
and ageing of the UV lamp applied. Make sure the detection is not performed at the absorption maximum
lamp is thermally equilibrated (stable baseline signal) of an analyte.
prior to an analysis, and that standards are run directly Figure 14 displays a simulated, and rather weak, mass
before quantification experiments (in order of increasing and volume overload experiment conducted by injecting
concentration). identical volumes of caffeine at different concentrations
An increase of the volume or mass of a sample injected onto a particle packed column.
onto a column—or a decrease of column i.d. while main-
taining the injection volume constant—has a positive
influence on the sensitivity of an analysis. However, this
approach is limited, as at a certain point both mass and/
or volume overload can be observed in a chromatographic
separation. Under these prerequisites, retention capacities
and peak widths as well as resolution are no longer inde-
pendent from the amount of sample injected. As a rule
of thumb, when the injection volume does not exceed 1%
of the total column volume the maximum separation effi-
ciency of a column can be preserved. Table 3 combines
typical flow rates and sample mass and volume amounts
for the loading capacities of various column formats.
Table 3. Guidelines for typical flow rates and sample mass and
tR tR
volume amounts for the loading capacities of analytical and
Figure 14. Mass (left) and volume overload (right) effect, injection
semi-preparative columns.
of caffeine samples. Injected mass of caffeine and volume of caffeine
Column Dimension Typical Flow Sample Sample increasing from bottom to top chromatogram. For better comparability
(length x i.d. in mm) Rates (mL/min) Amount(mg) Volume (μL) of peak shapes, the peak intensity was not normalized.
Chromatographic conditions: Purospher™ STAR RP-18 endcapped (3 μm)
150 x 1 0.06 ≈ 0.05 0.05–1 Hibar® HR 10 cm x 4.6 mm I.D.; mobile phase: acetonitrile/water
250 x 2 0.25 ≈ 0.2 0.2–10 90/10 (v/v); detection 300 nm; flow rate 1.0 mL/min; temperature:
25 °C mass overload: injection volume 5 μL (sample concentrations
250 x 3 0.6 ≈1 1–20 0.2, 0.5, 1, 2, 5, 10, 20, and 50 μg/μL caffeine); volume overload:
concentration 50 μg/μL caffeine (injection volume: 0.2, 0.5, 1, 2, 5, 10,
250 x 4 1 ≈5 5–80
20, and 50 μL).
250 x 10 6 ≈ 30 30–500
250 x 25 39 ≈ 200 200–3000
8
In contrast to mass overload, volume overload is observed The detection of traces of a previous sample injection in a
when the sample concentration is kept constant but the chromatogram is referred to as sample carry-over. Carry-
injected sample volume is increased. Volume overload over can negatively affect quantitative and qualitative
results in band broadening, flat-topped peaks, or peak analyses, hence chromatographic columns should display
splitting, and negatively affects resolution as well as plate a low carry-over tendency. When calibration curves are
count (see Figures 15 and 16). A second effect of prepared running dilution series, lowest concentrated
volume overload is a decrease in separation efficiency. samples need to be analyzed first. System consumables
An additional visible phenomenon can be the so-called (seals, needle) as well as system parts such as connecting
detector overload correlated to the detector properties. capillaries, sample loops and especially the injection valve
In Figure 15, such a detector overload can be observed (stator and rotor seal) can be a source of sample carry-
for seven out of eight analytes (cropped peak tips). over. Although less likely, the chromatographic column
itself can also be the cause of carry over. If carry-
One distinct difference between mass and volume
over persists after replacement of these components,
overload is evident: In a mass overload experiment, plate
check for the compatibility of both the autosampler
count drops by one order of magnitude when increasing
wash solution composition and the sample properties.
mass load by one order of magnitude, whereas in a
Other options to overcome a carry-over effect can be
volume overload experiment plate count drops by two
a change of column (selectivity), an adaption of the
orders of magnitude when increasing volume load by one
chromatographic method, the application of washing
order of magnitude. For this reason, it is recommended
steps between chromatographic runs or a needle wash
to work with concentrated sample solutions rather than
procedure. Biological samples (e.g., proteins, peptides)
with large volumes of a diluted sample. If injection of
should be analyzed in biocompatible HPLC systems, in
large sample volumes is necessary, it has to be made sure
which all metal parts are replaced by inert polymeric
that the sample is dissolved in a weak solvent in order to
components (e.g., made out of PEEK). Combining
achieve peak focusing on a chromatographic column.
biological samples with standard HPLC systems will
most likely cause sample carry-over. Figure 17 displays
the sample carry-over in an analysis of UV filters in sun
lotions on a particulate type HPLC column. After one UV
filter analysis run, three blank runs were performed. A
washing step is necessary in order to remove residual
Tinosorb S (chromatogram peak 4) from the particle
packed column after an overloading experiment. The
sample carry over observed in the first blank run is
caused by system parts between injector and column
including stainless steel frits of the column itself.
900
UV intensity (mV)
0 2 4 6 8 10 12 14
Retention time (min)
Figure 15. Effect of volume overload on peak shape. From bottom to 600
top: 10, 25, 50, and 100 μL injection volume.
Chromatographic conditions: Purospher™ Star RP-18 endcapped (5 µm)
Hibar® HR 15 cm x 4.6 mm I.D.; detection: 247 nm; temperature: 40 °C;
flow rate: 1.3 mL/min; mobile phase A: water, mobile phase B: acetonitrile; 300
gradient conditions: 0 min 45% B, 2.5 min 95% B; sample (in order of
elution): acetanilide, acetophenone, propiophenone, butyrophenone,
benzophenone, valerophenone, hexanophenone, heptanophenone.
0
100,000 0 5 10 15 20 24
1% of column volume 10% of column volume
80,000
Retention time (minutes)
Efficiency (N)
9
Thermostatting mobile phase temperature becomes large, a temperature
gradient may appear inside columns with PEEK housing
The influence of temperature on a chromatographic sepa- that negatively affects peak shape and separation effi-
ration cannot be generalized and an increase of tempera- ciency. Figure 18 shows the influence of temperature
ture can, but does not necessarily lead to, an improve- on the plate height achievable with an RP-18 endcapped
ment of column performance, or a change of selectivity. 50–2 monolithic silica column. As a reason of this it is
Higher temperatures decrease the viscosity of the mobile recommended to install a metal capillary with 1/16‘‘ outer
phase, which can improve mass transfer and separation diameter and 0.2 mm inner diameter (length 30 cm) in
performance, but on the other hand, a loss in performance front of columns with PEEK housing. The higher the flow
or undesired changes in selectivity are possible. Working rate of the mobile phase, the more important a pre-heated
at elevated temperatures also reduces analysis time, mobile phase is. Another option for thermostatting of elu-
because an increased diffusion coefficient facilitates ents is the combined use of metal capillaries and a water
higher flow rates of the mobile phase. When working bath and a column oven. Depending on the HPLC system,
with highly viscous eluents, an increase in temperature mobile phase pre-heating modules can be a general alter-
might allow for a chromatographic run that is not possible native to the use of metal capillaries (and a water bath).
under ambient room temperature conditions, as in the
case of elevated back pressures, for example. Low flow 30
40 °C w/o capillary 25 °C w/o capillary
rates, a reproducible and sufficient heat transfer from 40 °C with capillary 25 °C with capillary
25
column oven to column, as well as a pre-column eluent
10
System care and maintenance Pump debris is collected in the pump inlet filter. These
compounds might not be visible using UV detection, but
A system suitability test (SST) should be performed it is likely that they can be detected via MS. The filter
with any HPLC system and needs to be independent should therefore be replaced every 1–2 months, or after
of its use. The SST delivers information about the changing from acetonitrile to methanol (or vice versa)
suitability of a combination of a chromatographic in order to obtain lower baseline noise and to generally
column and an HPLC system in terms of selectivity or protect the system including column and detector.
other predefined criteria for a specific type of sample
under the given chromatographic and instrumental HPLC systems are equipped with a solvent mixture bottle
conditions. In a pharmaceutical environment, such an dedicated to autosampler washing. This mixture normally
SST is mandatory and has to be conducted regularly contains water and approximately 5–10% of an organic
(every workday morning), or prior to a new analysis solvent such as isopropanol for better wettability. It is
sequence. The chromatographic properties of the common practice in many laboratories to keep the compo-
sample utilized in such a test have to be similar to the sition of this mixture constant and utilize it independent
real samples analyzed in subsequent runs. of the type of method (isocratic or gradient) as well as
stationary phase (reversed or normal phase as well as
Knowledge about the system back pressure under HILIC). After an autosampler needle-washing step, part of
specific conditions allows for a calculation of net the washing solution is transferred to the chromatographic
column back pressures and a good comparability of system. As long as chromatography is done in reversed
different chromatographic columns under identical phase mode (the organic solvent is the strong solvent),
chromatographic conditions. Changes in system back low amounts of organic solvent will hardly affect the result
pressure can help to identify system wear. The system of an isocratic or gradient run. By contrast, water is the
back pressure can be determined by performing an strong eluent under HILIC conditions. If the autosampler
HPLC run without installed chromatographic column. washing solution composition is not adapted, the effect
The HPLC instrument has to be flushed with organic will be comparable to the injection of a sample dissolved
solvent regularly to prevent microbial contamination in a highly aqueous solvent mixture. This can cause peak
and its negative effects, mainly on highly sensitive mass shape issues such as peak splitting or broadening, as well
spectrometric detection. Use alcohols such as methanol as decreased retention. Therefore, the amount of strong
or isopropanol for this means. To a certain degree, solvent in the autosampler washing solution (as well as in
acetonitrile can be contaminated with amines, these the sample solvent itself) has to be kept lower than the
adsorb on the sapphire seats and balls of check valves, content in the mobile phase starting composition.
subsequently polymerize and can block inlet valves. When preparing a sample for reversed phase analysis,
Ceramic check valves do not seem to be affected in a the sample has to be dissolved completely; ideally in
similar manner. The frequency of flushing depends on the mobile phase (gradient runs: initial composition).
the utilized eluents and buffer concentration, and should Solubility of the sample in the eluent has to be tested
be between two to four weeks. If possible, at least 5% of prior to injection. If the sample cannot be dissolved in
organic solvent should be added to the aqueous mobile the eluent, a solution in less than 50% acetonitrile is
phase. An inversion of eluent channels can also help to the best choice. The second choice is 100% water (used
avoid microbial contamination. If a buffer was used prior up to 500 µL) and avoid (close to) 100% acetonitrile.
to instrument flushing, make sure the salt is soluble in
the organic solvent or that water is used for flushing Under HILIC conditions, the sample has to be dissolved
before switching to an organic solvent. If an HPLC is in 60–100% organic solvent or the initial mobile phase
not used for prolonged periods of time, flush the entire composition. The sample should not be diluted in water!
instrument with alcohol. Again, always test sample solubility in the mobile phase
prior to injection. When performing a gradient run, the
steepness should not exceed 2–5% eluent composition
change per minute in order to keep the system in
equilibrium.
11
Figure 20 illustrates the effect of improper sample amounts of sample dissolved in the mobile phase
solvent composition in the separation of the antiviral (acetonitrile/ammonium acetate 10 mM 90/10 v/v) were
drug oseltamivir and oseltamivir carboxylate on a injected with no negative effect to the chromatographic
zwitterionic HILIC column. As water is the strong results. By contrast, injection of samples dissolved in
solvent in this chromatography mode, water content of acetonitrile/ammonium acetate 10 mM 50/50 (v/v) as
the sample solution should be kept as low as possible. a solvent under otherwise identical conditions, led to
In the example shown in the left column different the elution of broad and even splitting peaks.
1.5e5 1.5e5
MS intensity
MS intensity
(counts)
(counts)
1.0e5 1.0e5
0.5e5 0.5e5
0.0e5 0.0e5
0 1 2 3 4 0 1 2 3 4
Retention time (minutes) Retention time (minutes)
1.5e5 1.5e5
MS intensity
MS intensity
(counts)
(counts)
1.0e5 1.0e5
0.5e5 0.5e5
0.0e5 0.0e5
0 1 2 3 4 0 1 2 3 4
Retention time (minutes) Retention time (minutes)
1.5e5 1.5e5
MS intensity
MS intensity
(counts)
(counts)
1.0e5 1.0e5
0.5e5 0.5e5
0.0e5 0.0e5
0 1 2 3 4 0 1 2 3 4
Retention time (minutes) Retention time (minutes)
Figure 20. Multiple reaction monitoring mode analysis of oseltamivir and oseltamivir carboxylate on a zwitterionic HILIC column. Sample diluent:
Acetonitrile/Ammonium acetate 90/10 (v/v) (left), acetonitrile/NH4OAc 50/50 (v/v) (right). Injection volumes: 2.5 μL (top), 5.0 μL (middle),
7.5 μL (bottom).
Chromatographic conditions: SeQuant® ZIC®-HILIC (5 μm, 200Å) PEEK 5 cm x 2.1 mm I.D.; detection: MS (multiple reaction monitoring mode,
extracted ion chromatogram overlay of m/z 313.3 and 225.2); mobile phase: acetonitrile/Ammonium acetate 10 mM 90/10 (v/v); flow rate
500 µL/min; sample: oseltamivir, oseltamivir carboxylate.
12
3. Method development & optimization
The overall process is influenced by the nature of the These initial questions will direct the chromatographers
analytes and generally follows the following steps: to define the method goal, and to find out requirements
of the new method. Is high resolution (in separation and
1. Selection of the HPLC method and system
detection), short analysis time, maximum sensitivity, long
2. Sample preparation column lifetime, and/or a column with wide pH stability a
3. Selection of the detector need, or will the method be used at neutral pH and under
4. Selection of initial conditions non-aggressive conditions? True optimization of a method
is a balance between selectivity, speed, and efficiency
5. Choose the right HPLC Column
in order to produce results that are the purpose of the
6. Mobile phase selection application. Ideally, the development should result in a
7. Mobile phase preparation robust method that gives the laboratory a low, overall,
8. Selectivity optimization price-per-injection and ultimately a cost-efficient assay.
9. System optimization
Common mistakes in analytical method
10. Method validation
development
11. Scaling of HPLC methods
• Inadequate formulation of method goals
• Insufficient knowledge of chemistry
Selection of the HPLC method • Use of whatever reversed phase HPLC column is
and system available in the lab
Method development is not difficult when a literature • Use of wrong instrument set-up
reference can be found for the same or similar needs.
Methods are published in pharmacopeia, in column • Trial and error with different columns and mobile phases
manufacturer application databases, and as published These mistakes often result in laborious, time-consuming
scientific studies. These sources can provide good projects that lead to methods that fail to meet the needs
guidance for the planned work, but what happens when of the laboratory.
references to the compounds of interest do not exist?
Different approaches are possible, and trial and error is Getting started
the least successful way forward. A chromatographer After defining the goal of the method development,
normally has access to a wide variety of equipment, specific information of the sample and the analytes
columns, mobile phase compositions and operational should be sought. Different sources are available:
parameters which make high performance liquid e.g. scientific journals, chemical databases such as
chromatography method development seem complex. www.pubmed.org (small molecules), ExPASy Proteomics
In this chapter, direction will be given to make method Server http://expasy.org (large biomolecules), and
development intuitive and successful, with emphasis
reference books. Listed below are some of the most
on column selection.
common parameters.
13
Sample Preparation Overview of Techniques
At a glance, sample preparation enables the following
advantages: Filtration
Filtration is a rather simple but essential component of
• Removal of undissolved particulates that could lead
high-quality separation and purification processes for the
to instrument down-time
removal of undissolved matter and particles from samples
• Decrease all/most of impurities that would interfere prior to HPLC or UHPLC analysis. Therefore, a filtration
with the analyte during analysis step has to be part of every sample preparation procedure
• Increase the sensitivity or enrichment of the desired and depending on the throughput of a laboratory syringe
analyte(s) filters, complete filtration systems or filter plates are
available. The type of filtration membrane (pore diameter
Sample preparation is generally the most time-consuming 0.20 or 0.45 μm) should be chosen according to specific
and tedious portion in the analysis for small molecules. sample properties:
But it is one of the most, if not the most, important
• LCR (hydrophilic PTFE): Aqueous or mild organic
considerations in the process. The ideal sample prep-
solutions; low binding and extractables, filtration of
aration prior to analysis increases the compatibility of
protein-containing solutions.
the analyte with the detection technique and removes
interfering impurities from the matrix. Addressing these • Durapore® (PVDF): Aqueous or mild organic
factors allow for cleaner spectra which enables greater solutions; low binding and extractables, clarifying
sensitivity. Further, certain analysis techniques require protein-containing solutions.
the analyzed sample be particle-free meanwhile
the prepared sample needs to be miscible with the • Nylon: Aqueous or organic solutions
resultant technique. • Express™ (polyethersulfone—PES): Fast flow and low
An important note about the matrix of the analyte sample. protein binding
The matrix tolerance of chromatographic columns and • Fluoropore® (hydrophobic PTFE): biologically inert
detectors differs depending on the type and properties with broad compatibility. Compatible with acids,
of the carrier material (monolithic or particulate), and bases, and solvents
on the type of detector used (UV, MS etc.). Especially
• Mixed Cellulose esters (MCE): clarification of water,
when utilizing highly sensitive mass spectrometric
detection and quantification, the sample preparation buffers, or aqueous solutions
process has to be performed thoroughly, as matrix
components can cause signal suppression and/or Syringe filters with high-density polyethylene or
adduct formation with target molecules and therefore polypropylene housing are suitable for the fast filtration
decrease sensitivity (signal-to-noise ratio) and/or of a rather limited number of samples (1–10 per day).
increase complexity of the mass spectrum. Due to Their chemical compatibility is broad and low holdup
this, the chromatographic analysis of samples with volumes make them ideal candidates for the preparation
high matrix load can make several different selective of low sample volumes. In addition, e.g. greater than
and specific sample pretreatment steps necessary. 90% drug recovery in the first mL of filtrate is possible
During the sample preparation procedure, undesirable when utilizing filters with PTFE membrane, indicating
compounds such as particles, lipids or dissolved matter low drug binding to PTFE, making this membrane type
should be removed. Extraction and purification steps as an ideal candidate for sample preparation prior to
well as preconcentration procedures of low abundance quantitative analyses. Syringe filters are also available for
analytes can be part of the clean-up procedure. Depending robotic systems or workstations. In that case, filters are
on the physical state of the sample, various sample used in combination with an automated filter changing
preparation techniques are available: system, and for applications such as dissolution testing
(evaluation of the dissolution rate of solid dosage forms
Liquid samples:
in the digestive tract) or HPLC sample preparation.
• Filtration Next to the above-mentioned membranes, a glass fiber
• Liquid-liquid extraction (LLE) membrane for the clarification of aqueous or organic
• Solid phase extraction (SPE) solutions with high particulate levels is available.
• Solid phase microextraction (SPME) For hard-to-filter samples (samples containing particulate
• Restricted access materials (RAMs) materials as well as samples that are viscous), syringe
filtration of food and beverage samples such as, e.g.,
Solid samples: juices, honey, soups or salad dressing can be difficult
and the particle load can easily clog the syringe filter.
• Soxhlet extraction, batch extraction Other hard-to-filter samples include many pharmaceutical
• Matrix solid phase extraction suspensions, shampoos, conditioners, creams and other
household products. Under these circumstances, vacuum-
After extraction of the analytes, the sample may be driven filtration of samples for liquid chromatography with
concentrated by evaporation of the solvent. This allows specifically designed filtration systems and filters can be
the sample to be resuspended in a desired solvent, i.e. an option. Same as for the syringe filters, low hold-up
the initial mobile phase for liquid chromatography. volumes of such devices deliver high analyte recoveries.
14
In contrast to syringe filtration, many samples can be • Easy handling
treated at the same time, making filtration systems a
• No formation of emulsions
good solution for non-automated, medium throughput
(10–100 samples per day). Recommended fields of use • High recovery rate and cleaner eluates in comparison
are, for example, drug dissolution testing, food safety to “bulk” LLE
(unknown and known toxin), cosmetics (ingredients and
formulations) and pharmacokinetics/pharmacodynamics • Lower solvent consumption
(PK/PD, drug interaction with the human body). • High batch-to-batch purity and reproducibility
15
Table 4. Flowchart to help decide which sorbents are best for analyte(s) separation.
Sample Matrix Aqueous (polar, buffer, water) Organic (slightly polar to non-polar)
Retention Mechanism Reverse Phase Ion Exchange Normal Phase
Strong Ions Weak Ions
Analyte Property Slightly polar to non-polar Polar to moderately polar compound
Cation Anions Cations Anions
C18, C8, C4, Ph, CN,
Sorbent Phase SCX SAX, NH2 WCX NH2, PSA Si, CN, Diol, NH2, PSA, Florisil, Alumina
DPA-6S, HLB
Sorbents
Available Description Applications
Extractions of nonpolar to moderately polar compounds including antibotics,
C18 Octadecyl bonded, end capped silica
barbiturates, essential oils, steroids, surfactants
C8 Octyl bonded, end capped silica Similar to the C18 but slightly less hydrophobic.
Reverse Phase
C4 Butyldimethyl bonded, end capped silica Less hydrophobic than C8 and C18, extraction of peptides and proteins
Used to adsorb polar compounds (hydroxyl and phenolic compunds) from
DPA-6S Polymeric hexamide bonded
aqueous and methanolic solutions through strong hydrogen bonding.
Ph Phenyl bonded silica Less retention than C8 or C18, slightly more specific for aromatic compounds
For very hydrophobic analytes that may be irreversibly retained on more
CN Cyanopropyl bonded
hydrophobic sorbents such as C18
Hydrophobic surface enclaved by
HLB For wide range of polar to non-polar analytes.
hydrophilic network
Polymer bonded Used to extract polar analytes through hydrogen bonding from organic
Diol
2,3-dihydroxypropoxypropyl solvents, oils, and lipids
Used to extract polar analytes or analytes with hydroxyl, amines, or
Si Silica gel
heteroatoms.
NH2 Aminopropyl bonded silica Used to extract moderately polar to polar compounds
Normal Phase
QuEChERS aqueous solvents in the presence of high amounts of salts and/or buffering
different salts and/or buffering agents
agents followed by SPE clean up.
Molecular MIPs
These included sorbents for NSAIDS, nitroimidazoles, fluoroquinolones,
Imprinted Specialized sorbents whose surfaces has
Beta-agonist, riboflavin, aminoglyconsides, and bisphenol A.
Polymers been designed to mimic specific analytes.
Designed for the resolution of cis/trans isomers. The cis isomer is better
Chiral Ag-Ion retained on the sorbent compared to the trans isomer. One application is the
fractionation of fatty acids methyl esters (FAMEs)
16
As seen in Table 4, there are three major types of achieved by use of a polar solvent that disrupts these
retention methods: reverse phase (RP), normal phase interactions. Solvents that are commonly used include
(NP), and ion exchange (IE). We will discuss these acetic acid, water, and methanol. Similar to RP, secondary
methods in greater detail. Listed in Table 6 are interactions involving silanols and its deprotonated state
solvents commonly used for SPE in order from most may exist and similar strategies may need to be explored
non-polar to polar in nature as well as the type of if recovery is lower than expected.
separation they are preferred for.
Reverse Phase (RP): RP separations involve a polar Ion Exchange (IE)
matrix/solution (usually aqueous but can also be biological Ion exchange sorbents are used for analytes that are
samples such as plasma or urine) and non-polar or mildly charged in solution (usually aqueous, but sometimes
polar analyte(s). The analyte(s) are retained on the polar organic). Depending on the charge of the analyte
appropriate non-polar sorbent primarily due to Van der in solution this will dictate which type of sorbent is
Waals forces or dispersion forces. Recovery of the analyte required. If the analyte in solution is an anion (negatively
is achieved by using a non-polar solvent that the analyte charge) then a strong anion exchange (SAX) or amine
is miscible in. Common solvents for elution range from (NH2 or PSA sometimes referred to a WAX) sorbent
hexane, tetrahydrofuran, ethyl acetate, or acetonitrile is required. If the analyte instead is positively charge
depending on the analyte(s). When working with RP (cation) than the sorbent to use is either a strong
bonded silicas, some polar secondary interactions with cation exchange (SCX) or weak cation exchange (WCX).
the residual silanols may occur if the elution with the Retention for all four of these sorbents is primarily
nonpolar solvent does not efficiently elute the analyte. electrostatic interactions where the analyte is attracted to
The addition of a more polar solvent (methanol) may the opposite charge on the sorbent. The pH of the analyte
aid in disrupting these secondary interactions. Further solution and the sorbent is a critically important criteria
considerations could include using an organic modifier to when performing IE. Table 7 provides an approximation
either increase or decrease the pH. The silanol (Si-OH) will of the pH’s to use while performing IE. At the same time,
deprotonate above pH ~4. Under these circumstances, the the ionic strength of the sample and loading solution are
column may have slight cation exchange properties. This important. If the ionic strength is too high, then poor
can be overcome by using a mixture of acidic methanol or retention of the analyte(s) to the sorbent will occur.
basic methanol or by mixing these with a more nonpolar,
In general, strong ion exchange sorbents (SAX or SCX),
methanol-miscible solvent. A good suggestion is to keep
should only be used with strong anions/cations when
the organic modifier to less than 2% by volume. If the
recovery and elution is not important. If recovery of
sample recovered is going to be used for analysis by
these species are important, than a weak ion exchange
mass spectrometry, consider using a modifier that is
sorbent is recommended. Strong ion exchange sorbents
compatible (e.g. ammonium formate for acids, ammonium
are good for working with analytes that can have the
hydroxide for bases).
charge changed depending on pH. This is highlighted
Normal Phase (NP): NP procedures involve polar ana- in Table 7. The strong ion exchange sorbents have a
lyte(s) in a mid-to non-polar matrix/solution (acetone, loading capacity ~0.2 milliequivalence/gram of sorbent.
chlorinated solvents, hexanes). In this case, the sorbents Equivalence is defined as the concentration multipled
are usually silica or modified silica (CN, NH2, or Diols) by the charge of the ion. For example, a solution of
that enable either hydrogen bonding, dipole-dipole, or 0.2 mM sodium ions and 0.2 mM magnesium ions
dipole-dipole-induced interactions with the polar (more would have equivalence of 0.2 meq (0.2 mM x [+1])
hydrophilic) analyte(s). Recovery of the analyte(s) is and 0.4 meq (0.2 x [+2])
17
Table 7. Methodology of using Ion Exchange sorbents and relative pH’s to use.
Analyte Weak acid (-) pKa ~5 Strong anions Weak bases (+) pKa ~10 Strong cations
Sorbent SAX NH2/PSA/WAX SCX WCX
Approx. pkA pKa ~1 pKa ~10 pKa ~14 pKa ~ 5
charge state always + + at pH < 8 always – – at pH >7
Loading Soln pH >7 pH <8 pH <8 pH >7
Absorption
Analyte (charge) Neg (–) Neg (–) Pos (+) Pos (+)
condition
Sorbent (charge) Pos (+) Pos (+) Neg (–) Neg (–)
Loading Soln pH < 3 pH > 12 pH > 12 pH < 3
Desorption
Analyte (charge) Neutral (0) Neg (–) Neutral (0) Pos (+)
condition
Sorbent (charge) Pos (+) Neutral (0) Neg (–) Neutral (0)
Summary Analyte(s) changes charge Sorbent changes charge Analyte(s) changes charge Sorbent changes charge
Conversely, the use of weak ion exchange columns applicable, centrifuge and/or prefilter samples prior
are generally better for strong ions (whose charge is to loading. Additional details about pretreatment of
mildly effected by pH) as retention and release of the samples will be discussed later.
analyte(s) is controlled by charge manipulation of the
3. Washing. This step helps eliminate impurities that
sorbent. This is not to say that weak ion exchange
would impact the results of the desired analyte(s). It
sorbents cannot be used for weak ions. Consideration
is important to have an idea of what the impurities
of the sorbent’s and analyte’s charge state with the
might be and how their properties differ from the
elution solution’s pH are important. Alternative, the
desired analyte(s). Example, if using a RP mechanism,
appropriate ionic strength may also be used to displace
the impurities may be less retained than the desired
the analyte. These ion exchange capacities should be
analyte(s). This is can be exploited by using a slightly
determined for each individual application.
more polar solvent to wash off the impurity, leaving
behind the desired analyte. Another example involves
Solid Phase Extraction Methodology exploiting differences in pKa’s. If the sorbent is
In general, there are four steps when performing SPE. washed with a slightly acidic/basic solution, it could
These 1. Conditioning/Priming, 2. Loading 3. Washing, be used to change the charged state on the impurity
and 4. Elution. and also the affinity for the sorbent, leaving behind
the desired analyte(s). Table 6, lists some common
1. Conditioning and priming the sorbent. Conditioning
solvents and their applicational use for SPE.
is vitally important when working with silica-based
sorbents and is encouraged when desired recoveries 4. Elution. Two smaller aliquots compared to one larger
or separations are not being achieved. Conditioning aliquot generally elutes the desired analyte(s) more
“wets” or solvates the sorbent and equilibrates the efficiently. Flow rate, once again, is an important
sorbent with similar solvent strength and/or pH to factor. The longer the elution solution/solvent is in
optimize retention. During this time, it is important contact with the sorbent, the better the expected
to not completely dry the sorbent out, otherwise it recovery of the desired analyte(s) will be. Another
will defeat the purpose. As general rule, conditioning consideration is whether or not the analyte is less
should be performed between 1 and 2 volumes if retained than the impurities. If this is the case, it is
using a cartridge (ex. Using a 10 mg/1 mL cartridge important to choose an eluent that will be selective for
would suggest 1 to 2 mL of conditioning). the analyte(s) and leave the impurities on the column.
The elution process also provides an opportunity to
2. Loading the sorbent. Depending on the sample,
increase the concentration of the analyte. This is
it may be loaded directly on the column. The flow
achieved if a smaller volume if feasible is used than
rate is important. The ideal flow rate is one drop
the volume of sample used during the loading step.
per second when time is not a factor and should
not exceed 2 mL/min in most cases. As a general With an overview of the different steps that are common,
rule, the loading capacity of the analyte should there are three different schemes that are most often
not exceed 5% of the bed weight of the sorbent implored. These are selective extraction, selective
(this does not apply for ion exchange sorbents as washing, and selective elution as seen in Figure 22.
it depends on the meq/g of sorbent, see earlier
discussion). As an example, if the bed weight is In selective extraction (Figure 22, left), the desired
100 mg of sorbent, this implies no more than 5 mg sorbent will bind the selected components of the sample,
of analyte is recommended to be loaded on the which can be the analyte(s) of interest or the impurities.
sorbent. If the sample is rather viscous or the pH is The selected component will be retained on the sorbent
not amenable to the column, presample preparation while the effluent will contain the rest of the solution.
should be considered. This may involve diluting If the desired analyte is retained, then the effluent is
the sample to reduce the viscosity or adjusting the discarded, and the analyte is collected separately in a
pH to the appropriate range (especially important new collection vessel when released from the sorbent.
for ion exchange). To avoid clogging frits or disks if However, if the impurities are retained and the analyte is
in the effluent, this is called the pass-through method.
18
Loading Loading Wash Wash Elution
Figure 22. Visualization of Selective Extraction (left), Selective Washing (Middle) and Selective Elution (Right). The legend of the symbols is
located below the image.
In selective washing (Figure 22, middle), the sorbent will available such as glass. Glass is recommended when
bind both the analyte(s) of interest and impurities when there is concern about trace amounts of leachables
loaded. During the wash step, an appropriate solution is from polypropylene.
added where the impurities are selectively removed, and
the analyte(s) is retained. The effluent is then discarded,
Well Plates. Similarly
and the analyte is collected separately when released
to cartridges, these are
from the sorbent.
available in a variety of
In selective elution (Figure 22, right), the sorbent will sorbents and bed weights.
bond the analyte(s) of interest and the impurities once These are recommended for
again. Washing steps are selected to remove unwanted processing many samples at a
matrix, while leaving the impurities and analyte(s) of time and are amenable to high throughput automation
interest on the sorbent. The effluent up to this point is liquid handling systems. These are tailored for smaller
discarded. At this time, a solution/solvent is selective for volumes (less than 2 mL).
the analyte(s) while leaving the impurities on the sorbent.
Once this is completed, further post-elute treatment may
Disks. Available in a variety of
perform at this time. This may include concentration or
sorbents embedded in glass
drying the sample or diluting/resuspending the sample
fiber membranes. Enables
in starting mobile phase.
faster flow rates and less
clogging than PTFE disks
Format of sorbent while extracting organic
There are several different formats in which the sorbent analytes/contaminants from
is available, including free sorbent, columns/cartridges, larger aqueous samples.
disks, 96-well plates, and online/inline cartridges. In
addition to the format, the sorbents are also available
in different bed weights that allow for different sample Online/Inline. An alternative
volume capacities. to manual SPE. These are
designed for direct injection of
Columns/cartridges. These come in a variety untreated samples for analysis by
of sizes with different amounts of sorbents LC-MS. These are available in different
loaded in the cartridge. Bed weights typically sorbents, including one designed to remove
range between 30 mg up to 10 g (see Table 8). phospholipids from biological samples. This is an
The smaller the bed weight, the smaller amount automated SPE which leads to high reproducibility and
of elution volumes required. This is beneficial consistent results by removing the human factor. Initial
for sensitive analyses where concentration of configuration of the LC is required.
analytes may be quite small. SPE is achieved
through manual processing. Variations are
Bulk Sorbent. This is available in
available where either positive or negative
different sorbents for those either
(vacuum) pressure can be applied to aid in
interested in packing their own cartridges
extraction. Cartridges are primarily made of
or performing dispersive solid phase
polypropylene, however, other mediums are
extraction (dSPE) or QuEChERs method.
19
Table 8. SPE Configuration Guide for Cartridge Selection
*B
ed capacity depends on the type of sorbent and analytes being used. Five percent of the bed weight is a general rule and represents the values
in the table.
Figure 23. Dispersive Solid Phase Extraction (dSPE) Formats. Shown on the left are traditional Quechers tubes whose typical method is outlined above.
On the right, a newer format shown is dispersive SPE tips that is comprised of loose sorbent within a micropipette tip, allowing the extraction to be
performed entirely within the tip in a method that entails simple pipetting steps.
20
Most often analytes may be protein bound and these interactions will decreased recoveries. There are several methods
to disrupt these including:
1. S
hift the pH of the sample to an extreme (pH <3 or >9). By shifting the pH, this can cause proteins to precipitate and/or
change the charge of the protein and/or analyte. If a precipitation occurs, continue working with the supernatant.
Serum, Plasma, and
Whole Blood 2. P
recipitate the proteins. The most common way to precipitate proteins is to add three equivalents of a polar solvent,
usually acetonitrile. After mixing, the sample is centrifuge to remove the protein pellet and the supernatant is used
for further analysis.
3. S
onication. This is achieved by sonicating the biological fluid for 15 minutes, followed by diluting with either water or
buffer and finally centrifugation.
This may not require pretreatment, however, it is common to dilute at least 1:1 with water or buffer adjusted to the
appropriate pH. A strong acid (ex. 12 M HCl) or base (10 M KOH) is added and heated for 15–20 minutes when interested in
Urine
acid or base hydrolysis. After the solution is cooled it is diluted in an appropriate solution. Alternatively, enzyme hydrolysis
to free bound compounds may be used.
It is common practice to dilute the media with water or buffer at the appropriate pH. If the sample is particulate-laden,
Cell Culture Media
it will either need to be filtered or vortexed and centrifuged prior to extraction.
Milk is often diluted with water or a mixture of water and a polar solvent (such as methanol). Proteins can be precipitated
Milk
by treating with an acid. If precipation occurs, the sample will require centrifugation.
Usually extraction can occur without prior treatment. One exception if it is laden with solid particulates and it clogs the
Water Samples cartridge. In this case filtration may be required. Note, that filtration may reduce recoveries if the analytes are bound to
the particulates.
These may be able to be extracted without pretreatment usually. One consideration is if the beverage contains alcohol. If RP
Wine, Beer and
is being used, dilution of the sample so the alcohol percent is below 10% may be required. If particulates are present,
Aqueous Beverages
filtering or centrifugation may be required.
Usually processed without pretreatment. An exception is if the juice has high amounts of pulp, which will need centrifuged
Fruit Juices or filtered. A second exception is if the sample is viscous, in which case the sample should be diluted. Finally, fruit juices
are acidic and may need to have the pH adjusted depending upon the analyte(s) of interest.
Since these are typically aqueous solutions, they are usually processed by RP or IE. If the sample is viscous, the sample
Liquid Pharmaceutical
will need to be diluted. Organic extracts of the preparation may be processed by NP.
Hydrocarbons or fatty oils are generally processed by NP. They cannot be diluted with water (immiscible) and if dilution
Oils
is required, a mid- to non-polar solvent is used.
These matrices are typically extracted with mid- to non-polar solvents via a Soxhlet extraction or sonication. Resulting
Soil and Sediment extracts are normally processed by NP to remove interferences. At this time, the sample may be dried and resuspended
in an appropriate solvent/solution for choice of SPE methodology.
Plant tissue, fruit, For analytes that can be extracted by RP or IE, the material is homogenized in water and/or a polar organic solvent.
vegetables, After centrifugation or filtration, the pH of the sample may be adjusted. If the analyte(s) are not as miscible with polar
and grains solvents, the sample can be extracted by NP by homogenizing in a less polar solvent.
Meat, Fish, and These can be processed similar to the plant, fruit, and vegetables. In addition, hydrolysis or digestion of meat or tissue
Animal Tissue can be achieved with strong acids (HCl), bases (NaOH, KOH), or enzymatic.
Tablets and Other These solids should be crushed into a fine powder and extracted with the appropriate solvent/solution depending upon
Solid Pharmaceutical the SPE method being used.
Preparations
21
BioSPME Selection of initial conditions
BioSPME, or bioanalytical solid phase microextraction
devices are comprised of particles embedded in an inert Mode of separation
binder and immobilized onto a durable base, such as a
When selecting the most suitable mode of separation, it
polypropylene 96-pin device.
is dependent on sample solubility and how the analytes
The BioSPME extraction does not require pretreatment of interest differ from other compounds or matrix in the
of the sample. The coated portion of the BioSPME pin sample. The type of surface modification used influences
is preconditioned and quickly rinsed with water prior to the selectivity of a stationary phase. There are several
immersion directly into the sample for extraction. While different modes of separation to be distinguished:
immersed, the binder selectively allows small analytes
of interest to bind to the adsorbent particles, while larger Reversed phase
macromolecules are excluded. The adsorption mechanism In reversed phase mode, the mobile phase is polar and
for BioSPME is based on partitioning of analytes between the stationary phase is less polar. The major distinction
the solution (sample) and the pin coating. The rate of this between analytes is their hydrophobicity where samples
partitioning is dependent upon the affinity of the analyte should be soluble in water or a polar organic solvent.
for the phase coating compared to the affinity for the
sample matrix. BioSPME is not an exhaustive technique, The retention mechanism is based on partition of the
rather, after a given amount of time, equilibrium is analytes between quasi-liquid stationary phase layer
achieved between the concentration of analytes in the and mobile phase. As a modification of the packing
sample matrix and the pin coating. material for example n-octadecyl (RP-18), n-octyl (RP-8),
n-butyl (RP-4), or phenyl are applied—a combination
This allows for a robust and selective non-exhaustive
with or without an endcapping step is possible. Amino-,
extraction of free analyte that can be employed in both
cyano- and diol-modified, Pentafluorophenyl (PFP or F5)
qualitative and quantitative applications. The 96-pin
and RP-Amide stationary phases can also be operated
configuration allows for direct sampling from 96 well plates
under reversed phase conditions. Furthermore, a so-called
and is compatible with robotic liquid handling systems
providing a fully automated high-throughput methodology. PAH material is available, designed for the selective
separation of polycyclic aromatic hydrocarbons.
22
Normal phase (polar) and medium polar phase Chiral separations
In normal phase, the mobile phase is non-polar while In chiral separations, the chromatographic result is
the stationary phase is more polar. This is the same for determined by the three dimensional structure of a solute.
hydrophilic interaction liquid chromatography. In normal Chiral selectors bonded to a base particle display a broad
phase, the major distinction between analytes is NOT their enantioselectivity and can be used for the chiral separation
hydrophobicity, and where the samples should be soluble of enantiomers of numerous different substance classes
in a hydrophobic solvent like hexane and the mobile phase (hydrocarbons, steroids, phenol esters and derivatives,
is a weak to moderate solvent for the sample. aromatic amines, heterocycles with 5 to 7-membered
rings) using typical RP or NP eluents (depending on the
An interaction of the active sites of the polar stationary
type of selector).
phase surface with polar parts of the sample molecules
leads to retention in normal phase chromatography. One widespread chiral selector is β-cyclodextrin covalently
linked to a base particle. Cyclodextrins are cyclic oligosac-
Polar substances such as unmodified SiO2 (ideally suited
charides consisting of α-1,4-glycosidically linked D-glucose
for the separation of fat-soluble vitamins, small organic
units. β-cyclodextrin consists of 7 glucose units, respec-
molecules or pharmaceuticals) or Al2O3 act as a stationary
tively. Geometrically seen, cyclodextrins can be described
phase. In addition, a modification yielding medium polar
as truncated cones, where all secondary hydroxyl groups
phases is possible:
are directed towards the larger opening, whereas primary
• NH2: The (silica) surface is derivatized with an hydroxyl groups form the smaller opening at the other
aminopropyl silane. The resulting phase can be end. The result is a hydrophobic inner cavity, in contrast
applied in the separation of carbohydrates or with the two hydrophilic openings. Since cyclodextrins
oligosaccharides in normal or reversed phase mode. are made up of chiral D-glucose units, the structure
may be regarded as a chiral selector. The enantiomers
• CN: Derivatization with polar cyano entities. For of a racemic substance mixture, due to their opposite
the separation of charged, unpolar to semipolar configurations, can now be associated—to different
substances: Tetracycline antibiotics, steroids, organic degrees—with the cyclodextrin molecule. Consequently,
acids, peptides, proteins. diastereomeric “inclusion complexes” are formed based
• Diol: Diol modified surface with specific selectivity for on hydrophobic interaction (between cavity and guest
compounds with double bonds (azo dyes), peptides, molecule) and stereoselective hydrogen bonds (between
proteins and malto-oligosaccharides. the C2 and C3 hydrogen groups of glucose molecules
and the guest molecule).
Medium polar phases (“polar bonded phases”) can Other available selectors are (macrocyclic) glycoproteins,
generally be utilized both in normal phase and reversed (3,5-dimethylphenylcarbamate)-derivatized cellulose,
phase mode (HILIC) when the eluent properties copper ligands, polycyclic amine polymers or cyclofructans.
are adjusted accordingly. They possess both polar
properties from their functional groups and non-polar Glycoprotein Chiral Phases (CHIROBIOTIC®)
(hydrophobic) properties by -(CH2)n-spacers. A cyano CHIROBIOTIC® phases are based on covalently bonding
modified packing material can act as a normal phase macrocyclic glycoproteins to a high purity 5 micron
system when combined with an non-polar mobile silica gel in such a way as to establish it’s stability while
phase, while the addition of water (or buffer) to the retaining essential components for chiral recognition.
eluent increases its polarity and allows for separations in CHIROBIOTIC® V and V2 are based on bonding
reversed phase mode. The main application for the diol Vancomycin, which contains 18 chiral centers surrounding
phase is normal phase chromatography, and reversed three pockets or cavities. Five aromatic ring structures
phase chromatography for the amino bonded phase. bridge these strategic cavities. Hydrogen donor acceptor
In normal phase chromatography, the polar bonded sites are readily available close to the ring structures.
phases have the advantage of a shorter conditioning CHIROBIOTIC® V has demonstrated selectivity similar
time than silica phases. to glycoprotein phases except it is stable from 0–100%
Typical analytes in normal phase chromatography are organic modifier and exhibits high sample capacity.
non-ionic unpolar to medium-polar compounds such For CHIROBIOTIC® V2, changes to the linkage chemistry
as hydrocarbons, ethers, esters, alcohols, amines or and silica offer improvements for preparative LC and
carboxylic acids and derivatives. for more demanding chiral separations. CHIROBIOTIC®
Normal phase chromatography works best with isopro- T, T2, and TAG are based on bonding the amphoteric
panol, ethyl acetate, tetrahydrofuran, t-butylmethyl glycopeptide, Teicoplanin, which contains 23 chiral centers
ether, dichloromethane or hexane. These solvents need surrounding four pockets or cavities. For CHIROBIOTIC®
to be mixed according to the characteristics of both T2, changes to the linkage chemistry and silica offer
analytes and stationary phase characteristics in order to improvements for preparative LC and for more demanding
provide a suitable elution strength. chiral separations. CHIROBIOTIC® TAG has the sugars
removed from the macrocyclic glycopeptide to produce
Depending on the mode in which they are used, an aglycone structure as a variant of CHIROBIOTIC® T.
medium polar phases can be combined with any of the CHIROBIOTIC® R is based on bonding Ristocetin A to high
aforementioned reversed and normal phase eluents purity 5 micron silica.
and solvents.
23
Cellulose DMP Chiral Columns Protein-Based Chiral Columns
Cellulose DMP is a chiral stationary phase (CSP) Hermansson described the use of natural proteins
comprising spherical, high-purity porous silica coated immobilized onto a silica support for chiral separations
with DMPC (3,5-dimethylphenyl carbamate)-derivatized in 1983. Proteins contain a large number of chiral
cellulose, and packed in analytical to preparative size centers of one configuration, and many other sites
HPLC columns. It separates a wide range of chiral that contribute to the general retention process. We
compounds under normal phase, polar organic, and offer three CSPs with proteins as the chiral selectors,
SFC conditions, with high efficiency, high loading CHIRALPAK AGP (a1-acid glycoprotein), CHIRALPAK
capacity, and excellent column lifetime. Performance is CBH (cellobiohydrolase) and CHIRALPAK HSA (human
comparable to other DMPC-derivatized cellulose CSPs. serum albumin). They are typically used in reversed-
phase mode, and perform a wide variety of chiral
Key Features and Application Areas:
separations. CHIRALPAK HSA is also used for drug-
• Classic DMPC-cellulose chiral selectivity binding studies. Solutes are retained by three types of
interactions: ionic (for charged solutes), hydrophobic
• Efficient, rugged, reproducible, and scalable and hydrogen bonding. The relative contribution of
• Low backpressure the different forces to solute retention depends on the
nature of the analyte.
• Ideal for chiral analysis in the pharmaceutical
industry and for small analytes in chemical and
environmental areas Choose the right HPLC column
• Routine chiral column method development Chromatographic resolution is mainly affected by the
screening protocols selectivity (α), as can be seen in Figure 26 in the
stationary phase selection section of this guide.
• Approximately one-half the cost of most DMPC- Changing the mobile phase composition or the stationary
cellulose columns phase is the most powerful way of optimizing selectivity,
whereas the particle size, pore size, length of the column,
temperature, mobile phase strength have much less
Copper Ligand Exchange (CLC) Chiral Columns effect. Therefore, if satisfactory results are not met, or
The CLC phases are based on coupling an enantiomeric no retention is achieved, it is better to change to another
form of an amine to a proprietary derivative to create selectivity using a different column type and/or a different
an appropriate distance for copper coupling. Using the mobile phase.
copper ligand concept, this phase resolves hydroxy
acids like lactic, malic, tartaric and mandelic. This The column backbone
phase can also resolve amino acids and other amines The column backbone of a chromatography column
by the same mechanism. It has been reported that, in can consist of either a particle packing, or a monolithic
addition to amino acids, other bifunctional racemates structure. Independent from the chemical characteristics
like amino alcohols can be resolved. In theory, any of the silica or polymeric backbone, the chromatographic
analyte that can complete the coordination with the properties of a column are strongly influenced by the
copper ion can be resolved. For the CLC-D column,the resulting particle packed or monolithic structure.
L enantiomer generally elutes before D with the
exception of tartaric acid where the D elutes first. The Monolithic HPLC columns
CLC-L column has the opposite elution order and the D
enantiomer elutes before L. Monolithic HPLC columns consist of a single piece of
high-purity and metal free monolithic silica gel (see
scanning electron microscopic image in Figure 26).
The mesopores within the silica skeleton form the
Figure 26. Scanning electron microscopic image of the cross section of a monolithic silica column. The total porosity of the monolith is > 90%.
Left image: Mesoporous silica skeleton, right image: Macroporous transport pore.
24
fine porous structure and create the large uniform and make the use of dedicated HPLC systems necessary.
surface area on which adsorption takes place, thereby This needs to be taken into account when considering
enabling high performance chromatographic separation. the application of 2 μm or sub-2 μm particles.
Macroporous transport pores enable rapid flow of
Note that UHPLC equipment and columns are not
the mobile phase and low back pressure analyses.
necessarily required in order to perform ultra-high
The diameter of both types of pores, as well as the
performance separations—it is more a matter of time
diameter of the skeleton can be independently fine-
required to achieve a specific separation efficiency. As
tuned. For this reason, monolithic silica columns can
long as a column, such as a monolithic silica column,
be tailor-made for a large variety of applications such
allows for high flow rates at a tolerable back pressure,
as small molecule analysis, as well as separation of
such separations can also be performed on robust
large biomolecules (see summarized pore size data of
and economically priced standard HPLC equipment.
monolithic silica analytical HPLC columns in Table 9).
In addition, low-pressure gradients can typically be
Currently, silica monolithic and organic polymer run with up to four eluents without extra costs. In
monolithic columns are available, the latter can often UHPLC separations, and under high-pressure gradient
lack sufficient capacity and mechanical stability. In this conditions, additional pumps would be necessary to
respect, silica monoliths are advantageous, as they do perform the same experiment.
not swell or shrink in organic solvents. They show much
If the plan is to speed up the separation processes and
higher capacities due to their high surface area and do
accelerate analyses in general, then increasing the flow
not exhibit micropores that can lead to peak tailing.
rate of an existing chromatographic method utilizing a
In contrast, a strongly basic eluent pH can be an issue
particle packed column can quickly exceed both column
when working with silica based monoliths. Make sure
and system pressure limits. Monolithic (silica) columns can
that mobile phase pH matches the characteristics of
be an option to overcome those limitations of particulate
your column backbone.
columns and to make the method quicker, as their back
pressure is generally lower. Figure 27 shows an example
Table 9. Pore size data of selected monolithic silica analytical HPLC columns. of speeding up a method using a monolithic column.
Diameter Macropore Mesopore
Product (mm) Modification size (µm) size (Å)
Chromolith® 0.05 RP-18e 2 130
CapRod® 0.2
1 mL/min, 39 bar
25
As can be seen from van Deemter plots of monolithic O H
H 3C N
O CH3 3 O NH2 total number
H 3C N N N
silica columns in Figure 28, the influence of an of injections
O N N O N N
increased flow rate does not significantly decrease CH3
1 2
CH3
Cl
10059
separation efficiency. This fact means that one can
work flexibly and in a wide range of flow rates with 9405
35 7836
Chromolith® HR—4.6 mm
MS Intensity
Chromolith® 100—4.6 mm
30 7022
Chromolith® 100—3 mm
Chromolith® 100—2 mm 6049
25
5186
H (µm)
20
4371
15 3533
10 2739
1760
5
902
0
0 1 2 3 4 8
Linear velocity (mm/s)
0.00 0.50 1.00 1.50 2.00 2.50
Figure 28. van Deemter plots of selected monolithic silica HPLC columns. Retntion time (min)
26
Particulate HPLC columns (see also Figure 31: p/u curves for different particulate
Particle packed columns are well known and established in columns in comparison with a monolithic column).
laboratories worldwide. Particles used for column packing
can either be spherically or irregularly shaped with the
former being available as fully or superficially porous
material. An inherent property of particle-packed columns
is the mutual dependence of particle diameter and back
pressure. In other words, a decrease of particle diameter
is desirable in order to increase column efficiency, but
there is always a tradeoff with back pressure, especially
when it comes to high speed separations. Currently, a
lot of effort is put into the development of sub-2 μm
particle packed columns to be used in combination with
special UHPLC equipment. Particle size reduction leads to
a decreased diffusion path length as well as a minimized
C term (mass transfer, see chapter 1). As a consequence,
the gain in plate count is substantial, while the increase Figure 31. p/u curves for Purospher™ STAR (2 μm) RP-18 endcapped
in back pressure is increasing by a factor of four when 5 cm x 2.1 mm I.D. (top), Purospher™ STAR (3 μm) RP-18 endcapped
5 cm x 2.1 mm I.D (middle) and Chromolith® FastGradient RP-18
reducing the particle size by half (Figure 30). endcapped 5 cm x 2.0 mm I.D. (bottom) utilizing acetonitrile/water
60/40 (v/v) as a mobile phase and thiourea as a dead time marker.
30,000
1 Regardless of particle diameter, two types of spherical
Efficiency (N)
25,000 N
20,000
dP particles are available: fully porous particles and
superficially porous particles (FPPs and SPPs, respectively).
15,000
FPPs (or “fully porous particles”) are well known and
10,000 established in the chromatographic community, whereas
5,000 SPPs (“fused core particles”) are a relatively new devel-
0 opment. SPPs are prepared by depositing a mesoporous
0 5 10 15 20 shell onto a solid and nonporous core. The particle size
distribution of the core shell particles is narrow, and eddy
400
1 diffusion (A term) is reduced. Due to the short diffusion
Pressure (bar)
dp (µm)
0.5 µm
Doubling the efficiency by halving the particle size 0.5 µm ~5 µm 3.3 µm Solid Core
results in a pressure increase by a factor of four. 2.7 µm 1.7 µm Solid Core
~5 µm 3.3 µm Solid Core
Figure 30. Effect of Particle Size on Efficiency and Pressure Fused-Core Particles
Fused-Core Particles Traditional Porous
Traditional
Particles
Porous Particles
column back pressure and allow the use of faster flow Figure 32. Structure of SPPs (top) and its influence on peak shape:
rates, thereby considerably reducing analysis time. Minimized peak broadening by short diffusion path (reduced axial
dispersion of solutes, middle) and narrow particle size distribution
(reduced eddy diffusion, bottom).
27
Superficially porous particles (SPP) provide smaller, Currently, particle packed columns are based on high-
reduced plate heights leading to higher efficiencies purity, metal-free silica produced from tetraalkoxysilane
narrower and taller peaks, for improved resolution and in a sol-gel process (type B silica). Due to the absence
lower detection limits (LODs and LOQs)(Figure 33). A flat of metals in the silica structure, this column family can
van Deemter plot and higher linear velocity optimum allow be used for the analysis of acidic, basic, and chelating
higher flow rates with minimal column efficiency loss. compounds. In contrast, older type A (or acidic) silica
contains metal ions and the resulting chromatograms
show tailing peaks for basic solutes (Figure 35).
O
O SiOH
O
O SiOH
O
O SiOH
O
Type A Type B
–Si – O – Si – O – –Si – O – Si – O –
O O O O
SPP HPLC and UHPLC columns provide about 40% –Si – O – Si – O –
n
–Si – O – Si – O –
n
2
After setting the method goals and careful investigation
4 of the analyte structures (hydrophobicity/hydrophilicity;
functional groups and potential detection possibilities),
select a few appropriate bonded phases along with a
viable detector. The initial column selection is important
3
5 and the chromatographer is advised not to use the first
reversed phase HPLC column available. Reversed phase
0 2
Min
4 6 liquid chromatography is indeed a workhorse in most
laboratories, and a particulate RP-18 column is often
Chromatographic conditions the first choice for many chromatographers, but many
Columns Ascentis® Express C18, 10 cm x 2.1 mm I.D., methods are often developed without utilizing the best
2.7 µm particles (53823-U) and sub-2 µm
particle colum (same dimensions)
or most appropriate selectivity. If the sample is mainly
Mobile phase water/acetonitrile 49:51 (for Ascentis® Express);
of hydrophobic character, having positive log P values
water/acetonitrile 55:45 (for sub-2 µm) and predominantly hydrophobic functional groups, a
Column temp ambient reversed phase column is advisable. Select a C18 or C8
Detector UV, 200 nm bonded phase for good retention and resolution. If the
Injection 1 µL
28
sample molecules have aromatic backbones and C18 The column selectivity has the highest influence on
or C8 columns are unable to resolve all components, resolution in chromatography. Therefore, the selection
then a change to a phenyl column where, in addition to of the best suitable column chemistry for the target
hydrophobicity, π-π interactions between the stationary analytes is an important selection parameter. C18 column
phase and sample molecules provide a different selectivity. chemistries are typically the first choice. Nevertheless,
As previously mentioned it is, however, important to when a C18 doesn't give the desired separation or the
use alcohols as the mobile phase organic modifier while sample contains compounds that are known to be difficult
working with phenyl columns. Acetonitrile, or any solvent to retain or resolve on a C18, consider other stationary
with double or triple bonds (π-π bonds) in the backbone, phase chemistries.
will diminish the interaction and make the phenyl column
If the method is intended for bioanalysis, analysis of
interact only with hydrophobicity.
matrix-rich samples in general, or where proper sample
preparation is unwanted or not possible, a monolithic
Resolution is mainly controlled by selectivity reversed phase column (e.g. Chromolith®) is a supe-
Resolution, R, can be expressed in terms of three rior choice over a particulate column. For example,
parameters (k, α, and N) which are directly related Chromolith® RP-18 endcapped is a better choice over
to experimental conditions. The parameters k and Purospher™ STAR RP-18e or any other particulate
α, are determined by the experimental conditions HPLC Column for such purposes as it has good matrix
(composition of the mobile phase; stationary phase tolerability and long column lifetime.
chemistry and temperature; see also chapter 1), and
where N is affected by column length L, particle size If samples are clean and/or good sample preparation
and pore size. Figure 36. will be included in the final method, and high peak
capacity is needed, a particulate column with small
particles and small pores may be more useful. Choose
[R5] columns which are known to have long lifetime at the
3
operating mobile phase pH. Choose bonded phases
α
2.5 based on high-purity, low-acidity silica for best peak
shape. If the sample consists of polar and hydrophilic
2 analytes, an orthogonal selectivity to reversed phase
should be selected. If chiral resolution is defined in
1.5 N
the method goal, then a suitable chiral column should
1 be chosen. Use analyte specific structure information
k (chemical structure, log P values etc.) to choose a proper
0.5 stationary phase (Figure 37 and Table 10). If acidic or
0
basic analytes are present in the sample; reversed phase
1 1.05 1.1 1.15 1.2 1.25 ion suppression (for weak acids or bases), reversed
0 5000 10000 15000 20000 25000
phase ion-pairing (for strong acids or bases) or HILIC
0 5 10 15 20 25
should be used. For low/medium polarity analytes,
Figure 36. Key parameters that control resolution and their overall normal phase HPLC or HILIC are viable techniques.
contribution to changes in resolution
Hydrophobic
Hydrophilic
Log P
8 0 -8
Too much Poor Peak
Aromatic
Isomers retention Shape (Basic Retention too short or inadequate Separation on RP-18
Compounds) compounds
on C18
Polar compounds
Closely related Pi-Pi
with high water HILIC
compounds Interactions content
Reversed phase
HILIC
RP- F5 ZIC
C30 C18 C8 Amide (PFP) Phenyl AQ C18 Amide Diol Cyano Amino OH 5 Si
USP
USP L62 USP L1 USP L7 USP L60 USP L43 USP L11 USP L1 USP L68 USP L20 USP L10 USP L8 USP L95 USP L3 L114
Figure 37.
29
Table 10. Polarity scale of analyte functional groups.
Table 11.
Column dimension
(length x i.d. in mm) Application Reason
4x4 Guard-column Protection from mechanical contamination
5 x 2/3/4.6 Sample contaminated to low extent
10 x 4.6/10/25
25 x 4 Precolumn High capacity precolumn
30 x 2/2.1/3/4 Method development Short retention time
55 x 2/2.1/3/4 Rapid HPLC and UHPLC Rapid equilibration
75 x 4 (if pressure stable) Low solvent consumption (small i.d.)
Low pressure drop
100 x 2.1 High detection sensitivity Semi-micro column for low injection volumes and low peak dispersion
125 x 2/3 (mass selectivity) Low solvent consumption
150 x 2.1/3
100 x 4.6 Standard column Adequate performance for most applications (average performance
125 x 4/4.6 8000–10000 N/columns)
150 x 4.6
250 x 2/2.1/3 High detection sensitivity Semi-micro column for low injection volumes and low peak dispersion
High performance separation Low solvent consumption
For complex samples
250 x 4/4.6 High performance separation For very complex samples
250 x 10 Semi-preparative For mg quantities of pure substance on lab scale
250 x 25 Preparative For g quantities of pure substance
Guidelines for typical flow rate and orientation values for the loading capacities of analytical and semi-preparative columns
Column dimension
(length x i.d. in mm) Typical flow ratea Sample amount Sample volume
150 x 1 0.06 mL/min ~0.05 mg 0.05–1 µL
250 x 2 0.25 mL/min ~0.2 mg 0.2–5 µL
250 x 3 0.6 mL/min ~1 mg 1–20 µL
250 x 4 1 mL/min ~5 mg 5–80 µL
250 x 10 6 mL/min ~30 mg 30–500 µL
250 x 25 39 mL/min ~200 mg 200–3000 µL
30
If high mass loadability is needed, or if the sample is 50 50 50
31
Mobile phase selection true buffer, but rather a pH adjustable salt) are viable
alternatives depending on detection mode. At high
In previous sections, insight has been given to mobile pH, dipotassium hydrogen phosphate and ammonium
phase recommendations, solvent properties, and carbonate can be used as buffers to maintain pH above
buffer components. A summary of starting conditions 8. Keep in mind that when working at high pH, only
is presented in this section, along with a discussion columns with wider pH tolerability should be used,
about the difference between isocratic and gradient and for this purpose Purospher™ STAR is an excellent
elution. The mobile phase solvent strength is a measure choice. Inorganic buffers are not recommended with
of its ability to elute analytes off the column. Solvent MS, ELS and CA detectors. These buffers are not
strength is generally controlled by the concentration volatile and may precipitate, and this is also the case
of the solvent with the highest strength; for example, when high percentages of organic solvent are used in
in reverse phase HPLC with aqueous mobile phases, the mobile phase (>70%).
the strong solvent would be the organic modifier but
in normal phase and HILIC, the strong solvent would Gradient elution
be the most polar. It is worth pointing out that cyano-
bonded phases are easier to work with than plain silica Often it is not possible to elute all analytes with a single
for normal phase separations. The aim is to find the mobile phase (isocratic) in the desired k’ (2–5) range.
correct concentration of the strong solvent. With many It is therefore advisable to use gradient elution where
samples, there will be a range of solvent strengths the mobile phase strength, and sometimes also pH
that can be used within the aforesaid capacity limits. and ionic strength, will change over time. Effectively,
Other factors (such as pH) may also affect the overall this means that early in the gradient the mobile phase
retention of analytes. elution strength is low, and where the elution strength
is increasing with time according to a defined program
that maximizes the number of peaks that can be
Isocratic elution resolved with a given resolution. This method results in
In partition chromatography, the mobile phase should the constant peak width observed in gradient elution,
be a moderate to weak solvent for the samples to compared to isocratic elution where the peak width
achieve peak focusing and not to compromise the increases in proportion to retention time. Gradient
actual separation. A good rule of thumb is to achieve elution is used to solve the general elution problem
a capacity factor (retention factor, k) of 2 to 5 for an for samples containing mixtures of analytes with a
isocratic method. In both RP and HILIC mode, the wide range of polarities. Gradient elution will also give
preferred organic solvent is acetonitrile for several greater sensitivity, particularly for analytes with longer
reasons: favorable UV transmittance, low viscosity, retention times, because of the more constant peak
and being easy to volatilize (important for MS, ELS width (for a given peak area, peak height is inversely
and corona discharge, CA, detectors). Methanol is a proportional to peak width). Common practice in
reasonable alternative, hence it may be worth changing method development is to run a scouting gradient first
the organic solvent if resolution is not achieved, to decide whether to use isocratic or gradient elution.
and adjust the percentage organic solvent in the
mobile phase to accomplish maximum resolution and If Δt/tG ≥0.25 use gradient elution If Δt/tG ≤0.25 use
retention. Methanol and other alcohols are also the isocratic elution
preferred choice as the mobile phase organic modifier Where Δt is difference in the retention time between
while working with phenyl columns. Acetonitrile or any the first peak and last peak in the chromatogram, tG
solvent with double or triple bonds in the backbone is the gradient time; the time over which the solvent
will diminish the π-π interaction and make the phenyl composition is changed. For most samples (unless they
column interact only with hydrophobicity. are extremely complex), short columns (10–15 cm) are
In reversed phase mode, the initial mobile phase pH recommended to reduce method development time. Such
should be selected with two considerations. Low pH columns afford shorter retention and equilibration times.
that protonates column silanol groups and reduces A flow rate of 1–1.5 mL/min should be used initially.
their chromatographic activity is generally preferred, The direct disadvantages with gradient elution are
especially with non-endcapped columns. Mobile phases the need of a more complex HPLC system, and the
having pH 1 to 3 with 20–50 mM buffer (potassium column requires re-equilibration after every analysis,
dihydrogen phosphate, TFA or formic acid in water) which makes injection-to-injection lengthier than for
is advisable depending on detection mode, and to an isocratic method. Gradient mode is not compatible
increase temperature to reduce analysis time. with all detectors (i.e. RI and EC) and more variables
Mobile phase solvents should be water miscible, have need to be controlled to ensure method reproducibility.
low viscosity, low UV cut-off, be non-reactive, and for System dwell volume (gradient delay volume) becomes
these reasons, acetonitrile, methanol and THF are used important especially in scaling a separation or
with RP columns. Not all RP methods are suitable under whenever transferring a method between instruments
acidic conditions, and other pH intervals may provide and/or laboratories. Be aware that delay volumes will
different selectivity. Around neutral pH, dipotassium vary from instrument to instrument.
hydrogen phosphate or ammonium acetate (not a
32
Gradient method development When performing a separation under isocratic conditions,
Good laboratory practice does not allow gradients from a premixed mobile phase has to be utilized. Shifts in
100% aqueous to 100% organic. For method robustness retention times caused by irreproducible mixing of
(for better mixing, to prevent precipitation of salt, avoid a mobile phase by the HPLC pump unit are avoided.
column dewetting, and to provide more robust gradient Premixed eluents have to be prepared by separately
profiles), it is advisable to keep minimum 5% of each measuring the appropriate volumes of each solvent in
component in all mobile phase bottles. In practice for order to avoid any volume contraction effects during
a reversed phase method, this means mobile phase A mixing.
contains 5% organic solvent and 95% aqueous, while Under gradient conditions, the mixer of the pump unit
mobile phase B contains 95% organic solvent and of the HPLC system is responsible for a proper mixing
5% aqueous. of the mobile phase. Either static or dynamic mixing
Initially, run a wide scouting gradient (5-95 % B) chambers exist; make sure that the mixer is switched
over 40–60 minutes. From this run, decide whether on when working with a dynamic mixing chamber. The
isocratic or gradient elution is best for the application. If shift in retention time in a separation of alkyl phenones
gradient mode is a more appropriate alternative, eliminate with a dynamic mixer switched on and switched off can
sections of the gradient and try to compress the analyte be seen in Figure 39.
peaks in space as much as possible prior to the first and 200
last eluting peak. To further improve the gradient profile
and shorten overall cycle times (including re-equilibration),
UV Intensity (mV)
try to reduce the gradient and total run time. Keep in 150
mind that a segmented gradient can be an effective tool
to improve the separation. If there is a need to improve 100
the separation of two closely eluting peaks; change
the solvent strength by varying the fraction of each
solvent (gradient shape and steepness); change column 50
temperature; change the mobile phase pH (in small
units); use different mobile phase solvents and/or buffer
0
components; and/or use a different selectivity by 0.0 0.5 1.0 1.5 2.0 2.5 3.0
changing the stationary phase.
Retention time (min)
Figure 39. Shift in retention time in a separation of alkyl phenones
with a dynamic mixer switched on and switched off (green and yellow
33
Polypropylene membranes exhibit poor particle 1.2E+07
UV Intensity (mV)
back pressure buildup. In contrast, filtering the eluent 8.0E+06
through polytetrafluoroethylene (PTFE) or PVDF
membrane filters (e.g., Omnipore® and Durapore®), 6.0E+06
enables the UHPLC system to run without significant
back pressure buildup. 4.0E+06
34
Mobile phase composition and temperature Mobile phase properties
Verify that solvents are miscible when changing mobile Chromatography on normal phase columns is based on
phases, and that no buffer precipitation will occur. the interaction of polar functional groups of analytes
with the polar stationary phase surface. In addition, the
In reversed phase chromatography, water-miscible organic
physical properties of the solvents/eluents play a role,
solvents such as acetonitrile, methanol, isopropanol
as these compete with the analyte molecules during the
and water, or an aqueous buffer serve as eluents, and
retention process. Depending on the strength of dipole-
additives such as tetrahydrofuran or dioxane can be used.
dipole and hydrogen bonding interactions of solvents with
Buffers such as phosphate, borate, acetate, carbonate,
a normal phase stationary phase, their elution strength
organic modifiers and ion pair reagents present no
can be considered as weak or strong. An eluent showing
problems as long as the appropriate pH stability range
weak interaction with the stationary phase is only capable
of the stationary phase and packing material is not
of eluting weakly bonded analytes from the column,
exceeded. Ion pair reagents are often difficult to flush
whereas a strong interaction causes elution of strongly
completely from the column. Therefore, columns used
bonded sample molecules. The elution or solvent strength
with these reagents should be dedicated to the particular
of various solvents depends on the type of stationary
analysis involved.
phase used. Table 12 summarizes the solvent strength
Normal phase and medium polar phase columns are ε0 of various solvents commonly used in normal phase
generally used with unpolar to semipolar solvents or chromatography mode on silica gel. On alumina the
mixtures of n-heptane, hexane, cyclohexane, dioxane, obtained values for ε0 are somewhat higher, but show the
ethyl acetate, methanol, chloroform or dichloromethane. same trend. In order to fine-tune the elution strength of
n-heptane and dioxane are typical solvents for adsorption a mobile phase, different solvents can be mixed, either
chromatography. static in isocratic separations or dynamic in gradient runs.
When adjusting temperature, one has to keep in mind Diisopropyl ether 0.22
that the chemistry of most chromatographic columns is Toluene 0.22
based on silica, and one negative aspect of this chemistry Diethyl ether 0.29
in an aqueous environment is that raising temperature
Methylene chloride 0.30
dramatically increases silica solubility. Due to this specific
temperature, limits exist for the operation of different Methyl ethyl ketone
stationary phases (for details see respective columns Acetone 0.43
instructions). Keep in mind that exceeding these limits Dioxane 0.43
will cause a loss of capacity, column performance or even
Methyl acetate 0.46
a breakdown of the column bed. As a rule of thumb, the
operating temperature for silica based HPLC columns Tetrahydrofuran 0.48
should not exceed 60 °C. Depending on the surface tert-Butylmethyl ether 0.48
chemistry of each type of column this value can be lower Ethyl acetate 0.48
or higher. For detailed information, please always refer to Dimethyl sulfoxide 0.48
the column manual.
Diethyl amine
An appropriate method for increasing the lifetime of Nitromethane 0.49
analytical columns is the use of guard columns or guard
Acetonitrile 0.50
cartridges. This principle is based on the saturation of the
mobile phase with silica before entering the analytical Isopropanol 0.60
column and is especially recommended when working Ethanol 0.68
with amino bonded phases. In addition, guard columns Methanol 0.73
protect the analytical columns from solid matter and
prevent clogging. An alternative can be columns with a Ref: [table reproduced from V.R. Meyer, Praxis der Hochleistungs-
polymeric column bed; such columns can tolerate higher Flüssigchromatographie]
temperatures.
Next to the elution strength, the viscosity and UV ab-
For proper thermostatting of columns, not only at elevated sorbance of mobile phase solvents and additives play
temperature, please also see the specific section in an important role in terms of their suitability for use in
chapter 2. HPLC analyses.
35
The viscosity of an eluent is directly correlated to the 3.5 acetonitrile isopropanol
methanol THF
back pressure of a given HPLC column and system ethanol
3.0
combination. The flow rate range under which a
Viscosity (cP)
separation can be performed—and in turn, the speed 2.5
of an analysis—is enlarged when using low-viscosity.
2.0
Another benefit of low solvent viscosity is enhanced
mass transfer also attributing to the speed of analysis. 1.5
Separation at increased temperature can balance the
negative effects of high mobile phase viscosity. The 1.0
viscosity of various aqueous mixtures of important .5
organic solvents is displayed in Figure 41. In general,
low viscosity eluents are favorable for the reasons given 0
0 20 40 60 80 100
above. However, in mass spectrometry detection, it can Percent organic (v/v)
be helpful to substitute 20–30% of a low viscous solvent
Figure 41. Viscosity of aqueous mixtures of acetonitrile, methanol,
such as acetonitrile by isopropanol in order to improve ethanol, isopropanol and tetrahydrofuran.
the sensitivity of an analysis. Under these prerequisites,
back pressure issues can be avoided by utilizing monolithic
type columns. 1.0
acetonitrile
methanol
The UV transmittance or absorbance is another key issue ethanol
0.8 isopropanol
when working with organic solvents. The transmittance
Viscosity (cP)
THF
is influenced by the inherent properties of a solvent,
but also by any contamination which can significantly 0.6
decrease sensitivity. While the absorbance is negligible in
the visible range of light, the difference is rather large in 0.4
the UV wavelength range from 190 to 300 nm (Figure 42).
Acetonitrile displays the lowest UV absorbance and can 0.2
be used down to a wavelength of 200 nm, while the
application of other organic solvents such as methanol, 0.0
ethanol, is restricted to UV wave-lengths of approximately 190 210 230 250 270 290 310
≥ 220 nm isopropanol > 230 nm, THF > 250 nm. In Percent organic (v/v)
addition, buffers or additives can also affect mobile phase Figure 42. UV absorbance of acetonitrile, methanol, ethanol,
transmittance in a negative manner. isopropanol and tetrahydrofuran.
36
Table 13. Buffer ranges of selected buffer systems multiple wavelengths simultaneously. The downside is
Buffer
that a UV detector is not analyte specific and requires
that the analyte absorb more light than sample matrix
TFA
at the set wavelength. Choose a detection wavelength
Sulfonic acid/NaHSO3
that maximizes sensitivity and specificity, but keep
Glycine/HCI in mind that the mobile phase solvents and buffer
H3PO4/KH2PO4 components may cause slight shifts in UVmax from
Formic acid reference values. Therefore, it is advisable to check the
Citric acid/Na citrate analyte absorbance in the mobile phase. Mobile phase
Acetic acid/Na acetate solvents and buffer components also have UV cut-off;
Pyridin/formic acid
therefore, make sure to work well above these levels.
Otherwise, there are likely to be problems with reduced
Formic acid/Na formate
sensitivity and increased system noise (unstable and
Pyridin/acetic acid
drifting baseline noise). UV wavelengths below 200 nm
KH2PO4/Na2HPO4 should be avoided because detector noise increases in
BIS-TRIS this region. Higher wavelengths give greater selectivity.
NaHSO3/Na2SO3
TRIS/HCI Refractive index (RI)
Trimethylamine/HCI Refractive index is also a common detection technique,
Na borate/HCI and measures the difference in the refractive index of a
Trimethylamine/CO2 sample cell versus a reference cell. This detector is also
NH4HCO3 a non-selective detection technique, being concentration
(NH4)2CO3/NH3 dependent. The sensitivity is typically 100–1000 times
lower than a UV/Vis detector. The benefit over a UV
NH3/acetic acid
detector is the possibility to quantify analytes with no
Ethanolamine/HCI
chromophores in the molecular backbone. The drawback is
Na2CO3/NaHCO3 the sensitivity and the fact that RI detectors are typilcally
Na borate/NaOH used in isocratic mode only.
Triethylamine/HCI
Pyrollidine Fluorescence (FL)
Na2HPO4/Na3PO4 Fluorescence detection is specific and measures only
pH 0 1 2 3 4 5 6 7 8 9 10 11 12 13 compounds that fluoresce; hence, a requirement of this
technique. The operation is similar to a UV/Vis detector but
Detector choice where the detector flow cell is used as the sensor through
which excitation light passes axially. A photocell is located
The choice of the detection is critical in HPLC as only at the side of the cell to receive radially emitted light. The
compounds can be analyses if they are detected. Using cell wall is made of special glass to prevent the excitation
a not suitable detector for the compounds of interest the light or other stray light from reaching the photocell.
chromatographic information to this compound will get When a solute that fluoresces in the excitation light flows
lost. To select the most appropriate detection mode, four through the cell, the molecule excites and fluorescent light
important parameters should be taken into consideration; passes through the walls of the cell onto the photocell. The
chemical nature of the analytes, potential interferences, excitation light may be light of any wavelength selected
LOD and LOQ required, linearity range, availability and/ from the light source using a monochrometer. Another
or cost of detector. Below are some of the most common monochrometer may also be used to selectively analyze
detection techniques for liquid chromatography presented. the fluorescent light and thus, a fluorescent spectrum can
Fluorescence, electrochemical or mass detectors should be produced for excitation light of any specific wavelength
be used for trace analysis. For preparative HPLC, and an excitation spectrum produced for fluorescent light
refractive index is preferred because it can handle high of any specific wavelength. To improve specificity of an LC
concentrations without overloading the detector. analysis, a fluorescent derivatization reagent can be added
(either pre-column or post-column) to form a fluorescent
Ultraviolet/Visible Absorbance (UV/Vis) derivative of the substance of interest. This derivative may
then be selectively detected from other solutes, which,
UV detectors are most commonly used in HPLC. This
(if they do not fluoresce) need not be resolved from each
detector is a robust, inexpensive and versatile detection
other by the separation column. Fluorescence detection is
technique since most compounds absorb light,
up to 1000 times more sensitive than UV/Vis, and is also
especially at low UV wavelengths. It is possible to use
concentration sensitive.
a diode array detector (DAD) and allow monitoring at
37
Evaporative light scattering (ELS) and alternating radio frequency (RF) potentials applied to
ELS is also a non-selective detection technique, but them. The HPLC system handles dissolved analytes under
where the ELS detector (ELSD) is mass sensitive and not ambient pressure (760 Torr) and delivers the sample
concentration dependent. It is an ideal technique for high to the MS, where the detection of the gaseous, ionized
molecular weight compounds, sugars, and less volatile samples is performed under high vacuum conditions
acids. The detector measures the light scattering and (10-5-10-6 Torr). The transfer of the analyte solution
where the amount of scattering is related to the molecular from the LC to the MS is accomplished via an interface.
mass of the analyte, i.e. the more mass the more The interface converts the sample stepwise to an aerosol,
scattering will be seen measured. In the detector, there ionizes it, and removes the solvent. Ions are then focused
are three processes; nebulization of the mobile phase (1), and passed along the middle of the quadrupoles. Their
evaporation of the mobile phase (2) and light scattering movement will depend on the electric fields so that only
by analyte particles. In contrast to RI, it works well in ions of a particular mass to charge ratio (m/z) will have
gradient mode. Keep in mind that mobile phase solvents a stable path to the detector. The RF is varied to bring
should be volatile for best performance. ions of different m/z into focus on the detector and thus
build up a mass spectrum. Depending on the physical
properties and the molecular mass of the molecules,
Electrochemical (EC) different types of interfaces are used, which vary among
An electrochemical detector requires that the analytes each other by how they ionize the molecules and the
can be oxidized or reduced by an electrical current. pressure applied during this process. At present, all the
The detector output is an electron flow generated by a common ionization techniques operate under ambient
reaction that takes place at the surface of electrodes. If pressure; i.e. electrospray ionization (ESI), atmospheric
this reaction is complete (exhausting all the analyte), the pressure chemical ionization (APCI), matrix assisted
current becomes zero and the generated total charge laser desorption/ionization (MALDI), and atmospheric
is proportional to total mass of material that has been pressure photo ionization (APPI). ESI and APCI are by far
reacted. This process is called coulometric detection. If the the most widely used in LC-MS hyphenation. The more
mobile phase is continuously flowing past the electrodes, esoteric techniques, electron ionization (EI) and chemical
the reacting analyte is continuously replaced in the ionization (CI) work under high vacuum conditions with
detector. As long as the analyte is present between the the advantage of being suitable for GC-MS hyphenation.
electrodes, a current will be maintained, albeit varying Quadrupole mass spectrometers commonly have two
in magnitude, and is called amperometric detection. An configurations when used with liquid-chromatography,
electrochemical detector requires three electrodes, the either as a simple single quadrupole system or placed in
working electrode (where oxidation or reduction takes tandem. The latter principle, the triple quadrupole mass
place), the auxiliary electrode and the reference electrode spectrometer, enables ion fragmentation studies (tandem
(compensates for changes in the background conductivity mass spectrometry or MS/MS) to be performed.
of the mobile phase). Electrochemical detection is more
sensitive than fluorescence detection, but commonly not Electrospray Ionization (ESI)
as selective as fluorescence and generally not compatible
with gradient elution. In ESI mode, liquid solutions of charged or polar
substances, delivered with an HPLC system, are
sprayed utilizing a metal capillary (“spray needle”)
Mass spectrometer (MS) and a nebulizer gas (nitrogen) in the MS. Resulting
Mass spectrometry is regarded as an established, droplets are dried (desolvatization) and volatilized,
routine, detection technique. MS detectors can be isolated, analyte ions are transferred to the detector.
coupled to various separation techniques such as liquid Thermal stress is low so the analyte molecules do not
chromatography (LC), thin layer chromatography (TLC), decompose. ESI is almost unlimited regarding molecule
or gas chromatography (GC), where the hyphenation size and suitable for medium to strong polar molecules,
with LC is by far the most frequent setup. In contrast e.g., amines, carboxylic acids, heteroaromatics, and
to more simple detectors, i.e. UV, RI, FL etc., MS sulfonic acids. ESI is applied when fragmentations are
generates data about molecular masses and detailed unwanted and molecular masses of biomolecules have
structural parameters and thereby offers the possibility to be determined. ESI-MS is well suited for hyphenation
to discriminate between co-eluting peaks in selected ion with LC, and as long as flow rates do not exceed
monitoring mode. The latter reduces the requirement for maximum 1–2 mL/min (depending on instrumentation),
chromatographic retention and resolution before detection, attainable sensitivity is very high; however, flow rates
yet it is always better to have retained and completely of between 1–500 μL/min are more common. In liquid
resolved peaks to prevent ion suppression or ion solution, molecules are either already ionized, or will
enhancement effects. Mass analyzers can be quadrupole, become protonated or deprotonated by additives in the
magnetic sector, time-of-flight, ion trap, or ion cyclotron sample solution and the mobile phase. To achieve best
resonance type. A quadrupole mass analyzer consists sensitivity, the mobile phases used should be set at
of four parallel rods that have fixed direct current (DC) a pH where analytes are ionized, and a rule of thumb
38
is to use neutral to basic pH (7–9) for acids, whereas Buffers do not only adjust the pH of the eluent and
more acidic pH (3–4) is advisable for basic compounds. lead to ionization of a target molecule, they can also
If the analytes of interest have multiple pKa values and form adducts with the analyte. Adducts [M + buffer],
may change their ionization state, other pH values may e.g. with ammonium, alkali, halogens, formate or
be more beneficial both in terms of ionization of the acetate, will lead to the detection of an additional peak
analyte and behavior in the column. Thus, depending in the MS spectrum; even a complete suppression of
on the choice of solvent and additives, either positive the analyte signal is possible when the vapor pressure
and/or negative ESI mode can be used. Typically, of the resulting adduct (mainly alkali) is decreased
positive mode is applied in combination with more significantly. Due to this, and in order to keep the
basic molecules, while acid compounds are analyzed in ESI source clean, volatile buffers are recommended.
negative mode. 0.1% formic acid is commonly added Non-volatile salts like phosphates, borates, sulfates or
to the mobile phase in positive ESI mode to provide citrates will precipitate in the MS source, block it, and
a low pH (≈3) and to protonate the analyte(s). Acidic cause tedious cleaning procedures.
analytes will be neutralized under such conditions,
accordingly negative ESI mode is preferred and higher Atmospheric pressure chemical ionization
mobile phase pH is recommended. Volatile buffers like (APCI)
ammonium acetate or ammonium formate are used
in the pH range 4.5–7 to deprotonate the analyte(s), This technique is complementary to ESI and also useful
and for high pH, it is possible to use either ammonium for LC-MS hyphenation. It does not require a mobile
carbonate or ammonium hydroxide (aqueous ammonia). phase with conducting properties where acetone or
For both negative and positive ESI, it is a prerequisite acetic acid esters can be used as solvents and thus
that all mobile phase solvents and additives are allows for a coupling of APCI with normal phase
volatile in order to avoid contamination of the mass chromatography. In APCI mode, the analyte solution
spectrometer, and that the total mobile phase ionic is vaporized prior to the ionization. Subsequently
strength is adequate (generally 2–25 mM) to prevent solvent molecules (aqueous-organic, e.g. methanol,
unnecessary down-time for cleaning of the detector. propanol, acetonitrile, acetone etc., combined with
Strong acids like hydrochloric acid or nitric acid are 2–20 mM of a volatile organic buffer such as formic or
unsuitable for two reasons: they form ion pairs with acetic acid, ammonium acetate, ammonium formate
analyte molecules (analyte signal suppression) and or triethylamine) become ionized with a corona needle
display strong oxidizing properties. where their charge is then transferred to the analyte
molecules via proton transfer or abstraction. APCI is
Trifluoroacetic acid (TFA) is a special case: It is widely suitable for the analysis of less polar, weakly ionizable
used as an ion-pairing reagent to improve the liquid substances with small or medium molecular weight
chromatographic separation of peptides or proteins. On (analytes without acidic or basic functional groups,
the other hand, TFA can cause strong ion suppression in e.g. hydrocarbons, alcohols, aldehydes, ketones,
mass spectrometry (mainly in negative ESI mode) and esters) and is therefore complementary to ESI, as long
contaminates the LC-MS system. ’A good compromise as the sample is thermally stable and vaporizable.
here would be through the use of difluoroacetic acid Fragmentations are generally observed with APCI.
(DFA). DFA provides the same excellent increase in Highest sensitivity is achieved using acetonitrile,
efficiency as TFA but without as much ion suppression methanol or water as solvents, and where the degree
nor does it contaminate the MS system as readily as of analyte ionization can be optimized via mobile phase
TFA. Unfortunately, both a quantitative estimation of pH. As for ESI, flow rates up to maximum 1–2 mL/min
these effects as well as general recommendations is can be tolerated. There are other less commonly used
not possible as their strength strongly depends on detection techniques possible to combine with liquid
the MS system used. Triethylamine as an alternative chromatography, such as chemiluminescence nitrogen
additive behaves in a similar manner. If the use of TFA (CLND), radio detectors, charged aerosol (CA, inductive
is unavoidable, a weak acid such as propanoic acid, or coupled plasma (ICP), nuclear magnetic resonance
isopropanol can be added to the mobile phase in order (NMR), but these are not dealt with here.
to decrease a signal suppression effect.
39
General recommendations Both metal and polymeric fittings, tubing, and ferules are
available for a proper column installation. Depending on
the application and the setup, an operator can use either
Column hardware
material. The depth of the drill hole of Supelco®-branded
HPLC and UHPLC columns come in a variety of different HPLC column end fittings is 2 mm [Parker standard
column hardware formats and materials for different format, used by most column manufactures] (as is the
applications (Table 14). Depending on the material visible length of tubing after unmounting of a fitting/
(stainless steel, PEEK) the pressure stability can tubing unit, see Figure 44), while for few other column
vary significantly. manufacturers it is 3 mm. In order to avoid the creation
All Supelco® columns have 10–32 UNF female end fittings of large dead volumes or improper column installation,
that connect to 1/16” capillary tubing. Pre-installed end this fact makes careful handling necessary when working
fittings of any type of chromatographic column should with columns from different manufacturers. In addition,
not be removed from HPLC columns, because the column the end fittings of some HPLC columns (e.g., Chromolith®
bed might be damaged and the performance reduced. columns) consist of PEEK, while others are made out
For further handling instructions see also the section on of stainless steel. Mounting metal capillaries with 1/16‘‘
column installation. outer diameter and a metal cutting ring fixed to a 3 mm
drill hole length can damage the PEEK hardware (both
The column hardware of Chromolith® monolithic silica column housing and end fitting) and the silica bed of the
columns consists of a mechanically stable and chemically aforementioned columns. To avoid any damage, use either
robust polymer (PEEK), with the end fittings made from flexible metal capillaries (0.25 mm outer diameter) with
the same material. SeQuant® particulte columns are a polyvinylidene fluoride (PVDF) cone or PEEK capillaries
packed in a PEEK lined stainless-steel column hardware. with PEEK screws and adjustable plastic ferrules.
Particulate silica for reversed phase and normal phase
HPLC is delivered in stainless steel column hardware with
stainless steel frits to keep the stationary phase particles
in place. The different columns hardwares provide dif-
ferent pressure stabilities (Table 14). Hibar® columns
as well as Chromolith®, SeQuant®, Ascentis® Express,
Discovery® and Supelcosil™ columns require a separate
pre-column holder for the use of precolumns. LiChroCART®
cartridges allow the direct integration of 4-4 guard-
columns in the cartridge holder “manuCART” which can
be mounted without tools (finger tight).
Drillhole
Cone
Installation of columns with PEEK end fittings
Most HPLC columns are connected to all standard HPLC and UNF thread 10/32”
UHPLC systems with standard 1/16’’ fittings. Short capillary Figure 44. Schematic drawing of a Chromolith® end fitting.
tubing is recommended to minimize extra-column volumes.
Table 15
Trademark Pressure
Hardware Trademark Sorbent Column Use Precolumn Material stability
Purospher™ STAR
LiChrospher® Requires
HPLC direct integration of precolumn—
LiChoCART® Superspher® manuCART® Stainless Steel 250 bar
Cartridge no separate holder needed
LiChrosorb® to use
Aluspher®
Purospher™ STAR
LiChrospher®
Hibar® RT HPLC column ready to use column separate precolumn holder required Stainless Steel 400 bar
Superspher®
LiChrosorb®
Hibar® HR Purospher™ STAR UHPLC Colum ready to use column No precolumns available Stainless Steel 1000 bar
SeQuant® U/HPLC Colum ready to use column separate precolumn holder required PEEK lined Stainless Steel 200–550 bar
Chromolith® U/HPLC Colum ready to use column separate precolumn holder required PEEK 200 bar
Discovery®
Ascentis®
Supelcosil™ HPLC Colum ready to use column separate precolumn holder required Stainless Steel 400 bar
[LiChrospher®]
[LiChorsorb®]
Ascentis® Express
HPLC Colum ready to use column separate precolumn holder required Stainless Steel 600 bar
BIOshell™
Ascentis® Express (2 µm)
BIOshell™ (2 µm) UHPLC Colum ready to use column separate precolumn holder required Stainless Steel 1000 bar
Titan™
40
Equilibrating the Column Reversed phase columns are shipped in acetonitrile/water
In order to conduct reproducible results, every (e.g., 60/40 v/v). As the columns can dry out during
chromatographic column has to be flushed with eluent stocking and shipping (this is of special importance with
under the starting conditions of a chromatographic run. cartridges, because the columns are closed but not tight
Isocratic separations require flushing with 5–20 column enough and can dry out easily), the column packing has
volumes of mobile phase, depending on whether a purged to be wetting thoroughly by flushing with 10–20 column
system is used and whether the column is new or has volumes of pure organic solvent (e.g., acetonitrile or
been used before, plus instrument dead volume. Under methanol). Then conditioning of the column with mobile
gradient conditions flushing with 5–10 column volumes phase has to be continued until the baseline has stabilized.
of the mobile phase will be sufficient for reproducible Of course, the miscibility of mobile phases is a prerequisite
runs. Without equilibration, low reproducibility, a shift for successful equilibration.
in retention times, and peak overlap can be observed Normal phase columns (Si, NH2, CN, Diol) are typically
(Figure 45). Note that Table 15 summarizes gross shipped with n-heptane/dioxane (99/1 or 95/5 v/v). It
column volumes. Depending on the stationary phase, the is recommended to perform equilibration with dioxane,
net column volume will be in the range of approximately followed by the mobile phase. If columns are going to be
60–80% of the listed value. The flow rate utilized for used with aqueous eluents, flush the column with ethanol
column equilibration is limited by the back pressure or 2-propanol before equilibration with the mobile phase.
generated by the packing material. Due to this, columns Monolithic silica NH2 columns are shipped in acetonitrile/
with a monolithic silica backbone creating a low back water (90/10, v/v). The shipment solvent is described in
pressure can be equilibrated significantly faster (and at the Column Care & Use sheet, it is recommended to start
much higher flow rates) compared to particle packed the equilibration of the column in this solvent mixture,
columns with particle sizes of 5 μm or less. followed by the mobile phase.
Table 15. Gross column volumes of various column formats (length: Column coupling
100 mm).
The separation efficiency of HPLC columns is mostly
Column I.D. (mm) Gross column volume (µL) reported as relative separation efficiency in plates per
4.6 1,662 meter (N/m). For complex separations it can often be
3 707
necessary to use long columns in order to provide the
separation efficiency required for the resolution of all
2 314
compounds of interest. With respect to column coupling,
0.3 7.1 the use of particle packed columns is restricted due to
0.2 3.1 back pressure issues.
0.1 0.79 The typical relative separation efficiency of most
0.075 0.44 monolithic silica HPLC columns (Chromolith® Performance)
is similar to 5 µm particles—about 100,000 N/m. Due to
the low back pressure of the monolithic column bed, these
500 columns can be connected in series up to a length of 1 m,
resulting in more than 80,000 plates (absolute, although
400 somewhat less than in theory due to the influence of the
Intensity (mAU)
100
3
0 4
2
0.0 0.5 1.0 1.5 2.0 2.5 3.0
Retention time (min)
Figure 45. Separation of alkyl phenones without and after proper
column equilibration (blue and black chromatogram, respectively).
Chromatographic conditions: Purospher™ STAR RP-18 endcapped (2 µm) 1
Hibar® HR 5 cm x 2.1 mm I.D.; mobile phase A: water, mobile phase
B: acetonitrile; gradient conditions: 0 min 45% B, 2.5 min 95% B;
injection volume 2.1 μL; detection: UV (247 nm); temperature: 40 °C; 0 5 10 15 20
flow rate: 0.54 mL/min; sample: 1 urea, 2 acetanilide, 3 acetophenone, Retention time (min)
4 propiophenone, 5 butyrophenone, 6 benzophenone, 7 valerophenone,
8 hexanophenone, 9 heptanophenone. Figure 46. Isocratic separation of alkyl benzenes on ten coupled
monolithic silica columns yielding an absolute efficiency of 83,000 plates
at a back pressure of 70 bars. Columns were connected utilizing specific
column connectors.
Chromatographic conditions: Ten pieces Chromolith® Performance
RP-18e 10 cm x 4.6 mm I.D. mm plus specific column couplers; mobile
phase: acetonitrile/water 80/20 (v/v); flow rate 2 mL/min; detection
UV 210 nm; sample: 1 uracil, 2 toluene, 3 ethyl benzene, 4 propyl
benzene, 5 butyl benzene, 6 pentyl benzene, 7 hexyl benzene.
41
Column lifetime, cleaning and regeneration Exposure of a column to samples or solvents containing
Column lifetime is highly dependent on the sample highly adsorptive components (CRUD; Chromatograph-
properties (particle load, impurities) and the HPLC method ically Retained Undesired Debris) or even particulate
used (eluent pH, temperature), and cannot be generalized. matter will result in increased back pressure, a change in
selectivity, and a drop of separation efficiency. In detail,
In certain situations 50,000 injections on one column are possible reasons are microbiological impurities in the
possible and in other cases only a few (less than 10). This mobile phase, mechanical abrasion from components
fact depends on the sample properties, method criteria, of the HPLC system, matrix components from biological
and the method and system suitability criteria defined: samples, or particles in the mobile phase. Reverse the flow
periodically to prevent particles and non-eluting sample
• plate numbers (efficiency/sensitivity) components from accumulating on the column. This is
• selectivity (how well separated peaks are) only allowed for columns which have the same frits on the
top and the bottom. The column can often be restored to
• tailing factor (how symmetric peaks you expect) original performance by using suitable wash protocols (see
above). When performing solvent rinse regeneration, the
column should be transferred from the analytical HPLC
Column back pressure (how stable the method is and
system to a simple, inexpensive pump. Alternatively,
how little matrix accumulates on the column with time)
disconnect the column from the detector and rinse directly
can also be a critical issue here. to waste. Rinse with a minimum of 20, preferably 30,
Many injections on one column can be expected if one column volumes of each solvent. Every column has insert
or more of the following criteria are fulfilled: sheet with washing recommendations, and users must
follow insert sheet guidelines.
• big gap in retention time between peaks (high resolution)
If a reversed phase column (RP-18, RP-8, Diol, CN,
• low concentration and volumes of sample are injected NH2 if used in RP mode) is strongly contaminated,
onto the column, solvents are of high quality and filtered then pump the following solvents one after the other
(if buffers are added) through the column for 5 minutes at the upper limit
of the corresponding column optimal flow-rate range:
• peak shape changes are not too relevant
water, acetonitrile, 2-propanol, n-heptane, 2-propanol,
• software can still integrate peak area adequately acetonitrile, water, mobile phase.
For cleaning and regeneration of polar phases (Si,
If many peaks elute within a short time and with big Diol, CN, NH2) connect the column in the reverse flow
differences in peak amplitude, and sharp symmetrical direction, then pump the following solvents one after the
peaks with no co-elution are expected, a column might other through the column under the same conditions
need to be changed more regularly. Therefore, if a as described above: n-heptane, chloroform, ethanol
method is set up, and it is planned to run the method or 2-propanol, chloroform, heptane, mobile phase. For
over a long period, it is highly recommended to define normal phase columns use n-heptane, n-heptane/dioxane
limits of the criteria listed. This fact makes it easy to 50/50, dioxane, n-heptane/dioxane 50/50 as solvents/
decide when to change the column. solvent mixtures. Every column has insert sheet with
washing recommendations, and users must follow insert
For samples with large quantities of contaminants, it is sheet guidelines.
recommended to apply one or more sample preparation
When re-reversing the flow, flush the column before
methods prior to separation (e.g. solid phase extraction,
connecting it to the detector.
filtration, centrifugation; see also chapter 5). An
alternative or additional option is to apply guard columns. A documentation of HPLC column history makes sense in
It is generally good practice to protect the analytical order to trace any issues possibly affecting column lifetime
column with such a pre-column in order to ensure in a negative manner. This makes sense especially when
maximum column lifetime. Use of a pre-column might a column is run under very different conditions (sample
result in a slight shift of the chromatographic parameters. load, contaminated sample, extreme pH).
Make sure that the samples and the mobile phases are
clean and particulate free by using HPLC grade solvents Use with mass spectrometers—procedure to
and reagents. If buffers or other salts are used, a final maintain low column bleeding
filtration of the mobile phase should be carried out with Unless an LC run is performed utilizing a normal phase
a membrane filter. column, the stationary phase of every HPLC column
contains covalently bound organic entities altering its
To extend the lifetime of the column, “wash” the
physical properties. Depending on the quality of both
column after use and before storage to remove traces
phase modification and a subsequent washing step
of samples and buffers from the column. For cleaning
these entities (e.g. octadecyl, cyano, phenyl) can be
of non-polar phases (RP-18, RP-8, Diol, CN, NH2 if used
washed off the column during a chromatographic run
in RP mode), connect the column in the reverse flow
and cause weak to severe interfering signals. This
direction. The simplest procedure is to pump 100%
unwanted phenomenon is referred to as “column
methanol or acetonitrile for 5 min at middle optimal
bleeding” and leads to decreased sensitivity in MS.
flow rate. If buffers have been used, first pump 100%
water and then methanol. Note: Ion pairing reagents are A minimization of column bleed is possible by flushing of
not completely removable from a stationary phase. an RP-18 or RP-8 column prior to connection to an LC/MS
42
instrument using a mixture of isopropanol and 0.1% formic Na+
at half optimum flow rate for approximately 30 minutes.
This process cleans any trace organic material out of the N– O
F
column and hence increases sensitivity by decreasing S N
background noise. Plain Si columns should be washed with
dioxane at middle optimal flow rate for the same period of
F O N
time. As an alternative, 2–3 gradient runs from strongly
aqueous to strongly organic conditions can also be utilized
to wash a column prior to connecting it to the MS detector. Figure 47. Chemical structure of
pantoprazole sodium. O O
Column bleed will generally be low when the maximum
solvent strength of the mobile phase used is equivalent The original method was developed on a Purospher™ STAR
to methanol or acetonitrile. If stronger solvents are used RP-18 endcapped 15 cm x 4.6 mm I.D. column with 5 μm
in the mobile phase, e.g. tetrahydrofuran or DMSO, then particles with a total cycle time of 50 minutes (Figure 48).
it is recommended first to pump approximately 10 mL By changing to a Purospher™ STAR RP-18 endcapped
of this stronger mobile phase plus 0.1% formic acid 5 cm x 2.1 mm I.D. column with 2 μm particles, lowering
through the column before connecting to the detector. the flow-rate, and altering the gradient profile, it was
possible to reduce the total analysis time from 50 to
5 minutes, maintaining sample peak profile (with
Storage improved resolution), at low back pressure and high
Reversed phase columns should be stored in pure separation efficiency (Figure 49).
acetonitrile or a mixture of organic solvent (e.g.
50
acetonitrile or methanol) and water (e.g. 50/50 v/v).
Polar phase columns are best stored in heptane/
11.914/Pantopraole Na
18.025/ImpB
40
dioxane (80/20).
Storing the column in buffers or at unsuitable pH for a 30
prolonged time will shorten the lifetime of the column.
Before extended storage (i.e. over the weekend or 20
long term storage), the columns should be thoroughly
14.407/Imp D
15.848/IUnknown
16.514/IUnknown
18.420/IUnknown
14.861/Imp E
13.799/Imp F
10.791/Imp A
8.088/Imp C
rinsed-out from buffer salts, or ion-pair reagents 10
which can cause bacterial growth or precipitate in the
stationary phase or the HPLC system. The following 0
procedure is recommended, first flush the column with
0 5 10 15 20 25 30
10–20 column volumes of mobile phase minus buffer, Retention time (min)
then with 10–20 column volumes of the shipping
Figure 48. Chromatogram showing the purity profile of pantoprazole
solvent mixture (acetonitrile/water, 75/25). sodium on a Purospher™ STAR RP-18 endcapped 15 cm x 4.6 mm I.D.
column with 5 μm particles.
Alternatively, follow this protocol if the mobile phase
Chromatographic conditions: Purospher™ STAR RP-18 endcapped (5 μm)
contains a buffer: Flush with 5–10 column volumes 15 cm x 4.6 mm I.D.; mobile phase A: 1.74 g dipotassium hydrogen
water, then with 20–50 column volumes organic solvent/ phosphate in 1000 mL water, adjusted to pH 7.0 with dilute phosphoric
water (e.g. acetonitrile/water, 1/1). After flushing acid (330 g/L), mobile phase B: acetonitrile; gradient: 80% A to 20% A in
40 min, 10 min reequilibration at 80% A; flow rate 1.0 mL/min; temper-
with 20 column volumes of the storage solvent (e.g. ature 40 °C; injection: 20 μL; detection: UV 290 nm; sample: 460 ppm of
preferably acetonitrile), the column can be easily stored. pantoprazole sodium in 1:1 mixture of ACN and 0.001 N NaOH.
By rinsing with acetonitrile, aprotic impurities can also
be removed from the column. It's recommended to users 100,000
2.369
System optimization
4.047
50,000
3.344
2.860
3.477
4.482
3.611
43
Method validation but depending if it is a pharmaceutical assay or a bio-
analytical method, different acceptance criteria govern.
Proper validation of an analytical method is important to In pharmaceutical quality control there are much more
ensure that it will provide similar results, today, tomorrow, stringent method requirements and less variation
next week, next year, i.e. over a long period of time, in amongst samples compared to analysis of patient
different laboratories and independent of the analyst. Not plasma or serum samples. For any assay, the relative
only because of requirements from regulatory authorities, standard deviation (RSD) or coefficient of variation (CV)
but rather to ensure good manufacturing practice (GMP) is used as indication of the imprecision of the method.
and good laboratory practice (GLP). It is especially From a practical perspective, six to ten replicate
important for pharmaceutical analysis when assurance of injections will give you a good idea of the precision of
the continuing efficacy and safety of each manufactured the method. An analytical method can be accurate but
batch relies solely on the determination of quality. not precise, precise but not accurate, neither, or both.
Guidelines for the validation of analytical methods can be
found at the International Council on Harmonization (ICH). Specificity
The US Food and Drug Administration (FDA) and USP both
refer to ICH guidelines. Specificity is an important parameter to test in a
validation program as it verifies the ability of the
Keep in mind that analytical method validation should method to accurately measure the analyte response in
be isolated from the initial selection and development, the presence of all potential sample components. The
which actually are only the first steps in establishing analyte response from a solution containing only the
a routine analytical method. Validation means testing analyte is compared with test samples containing the
of the method to find out allowed variability for each analyte and all potential sample components (placebo,
method parameter. Routine quality control methods synthesis intermediates, excipients, degradation
should guarantee that the analytical results of raw products and impurities). For pharmaceuticals, stress
materials, excipients, intermediates, bulk products or conditions such as heat, light, acid, base and oxidant
finished products are viable. are typical. For formulated products, heat, light and
humidity are commonly used to stress the samples. The
In this section the most widely applied validation char-
analyte peak is evaluated in all test samples for peak
acteristics are explained; accuracy, precision (repeat-
purity and resolution from the nearest eluting peak.
ability and reproducibility/intermediate precision),
specificity, limit of detection, limit of quantitation,
linearity, robustness and stability of analytical solutions. Limits of detection and quantitation
The LOD is defined as the least amount of an analyte
Accuracy in a sample that can be detected, and commonly
expressed as the concentration level that is able to
An analytical method is accurate if it gives the right
provide a signal-to-noise ratio of three (S/N=3). LOQ is
numerical value for the analyte (either mass or
defined as the lowest analyte concentration level that
concentration) and can be described as the degree
can be quantified with good precision and accuracy,
of closeness of measurements of a quantity to its
and providing a signal-to-noise ratio of ten (S/N=10).
actual value. A method almost never gives the exact
LOD and LOQ can also be determined based on the
same results for replicate analyses, which means that
standard deviation of the response and the slope of
the result is presented as the mean or average. A
the calibration curve.
pragmatic way to express accuracy is to present it in
terms of the standard error, which is the difference
between the observed and the expected concentrations Linearity
of the analyte. To determine accuracy, a common The linearity of an analytical method is the capability
practice is to analyze a known amount of standard to generate results that are directly proportional to the
material under different conditions in a formulation, concentration of analyte in the sample. It is commonly
bulk material or intermediate product to ensure that illustrated as the interval between the upper and lower
nothing interferes with the method. analyte concentration levels that may be determined
with precision and accuracy. Linearity data is often calcu-
Precision lated using the calibration curve correlation coefficient
and the y-intercept. The RSD, intercept and slope of the
Precision is the ability of repeatedly performing an
calibration curve should also be calculated.
analysis with a low standard deviation. A differentiation
is made between repeatability and reproducibility.
Repeatability is the measure of how easy it is for Robustness
an analyst in a given laboratory to attain the same The method robustness is a measure on how well
result for the same batch of samples (normally by an analytical method remains unaffected by small
injecting the same samples repeatedly at different variations in the experimental conditions, but also how
concentration levels) using the same method and using reliable the method is during normal use. Important
the same equipment and reagents. Reproducibility or parameters to monitor are changes in the mobile phase
intermediate precision measures the variations within composition, mobile phase pH, changes in the gradient
or between days, analysts and equipment. Highly profile, changes in the buffer concentration, column
reproducible quantitative results should be expected, temperature and injection volume.
44
Analytical solution stability Scaling the flow rate
Analytical solution stability can be divided into different Decreasing the internal diameter of the column (e.g. from
sections; recovery, dilution, internal standard addition 4.6 mm to 2.1 mm) requires recalculation of column flow
etc. If an extraction process is used (either liquid-liquid rate in order to maintain linear velocity. Linear velocity is
or solid phase extraction), it must provide proper analyte defined as the distance which mobile phase travels over
recovery. A method with low analyte recovery and/ time (cm/min), whereas flow rate is the volume of mobile
or where the analyte is degraded during the sample phase that travels over time (mL/min). To maintain the
preparation is not tolerable for routine quality control. same linear velocity through a column with a smaller
Internal or external standards (reference materials) internal diameter, the flow rate must be decreased
should be prepared in such a way that they maintain their proportionally to the column internal diameter according
potency, and produce same response over time. Samples to the equation below.
and standards should be tested for stability to verify
stability over a normal analysis cycle. A rule of thumb is
Flow rate f2 = f1 x d22/d12
that the sample solutions, standard solutions and HPLC
mobile phases should be stable for minimum 24 hours f1 = HPLC flow rate (mL/min)
under defined storage conditions. f2 = UHPLC flow rate
d1 = HPLC column i.d. (mm)
d2 = UHPLC column i.d.
45
Scaling an HPLC method 0 0
0
C
Purospher™ STAR 5 μm column H2C N
from H2N NH2 H
dimension 15 cm x 4.6 mm I.D.
1. Urea 2. Acetanilide 3. Acetophenone
Purospher™ STAR 2 μm column
to
dimension 5 cm x 2.1 mm I.D.
0 0
0
1,500
1,000 1,000
500
0 0
–200
0 2 4 6 8 10 12 25 0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5 6.0 6.5 7.0 7.5
Retention time (min) Retention time (min)
Column Purospher™ STAR RP-18e (5 µm) Column Purospher™ STAR RP-18e (2 µm)
LiChroCART® 150 x 4.6 mm Hibar® HR 50 x 2.1 mm
Mobile phase A: Milli-Q® Ultrapure Water Mobile phase A: Milli-Q® Ultrapure Water
B: Acetonitrile B: Acetonitrile
Gradient Time (min) %A %B Gradient Time (min) %A %B
0.0 55 45 0.0 55 45
15.0 55 45 5.0 55 45
Flow rate 1.3 Flow rate 0.27
Detection UV 247 nm Detection UV 247 nm
Temperature 40 °C Temperature 40 °C
Equilitration 7.5 min Equilitration 2.5 min
Injection 10 µL Injection 0.7 µL
volume volume
Sample Alkylphenone standard Sample Alkylphenone standard
1. Urea 1. Urea
2. Acetanilide 2. Acetanilide
3. Acetophenone 3. Acetophenone
4. Propiophenone 4. Propiophenone
5. Butyrophenone 5. Butyrophenone
6. Benzophenone 6. Benzophenone
7. Valerophenone 7. Valerophenone
8. Hexanophenone 8. Hexanophenone
9. Heptanophenone 9. Heptanophenone
1,500
1,000
1,000
500
500
0 0
0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5 6.0 6.5 7.0 7.5 0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6
Retention time (min) Retention time (min)
Column Purospher™ STAR RP-18e (2 µm) Column Purospher™ STAR RP-18e (2 µm)
Hibar® HR 50 x 2.1 mm Hibar® HR 50 x 2.1 mm
Mobile phase A: Milli-Q® Ultrapure Water Mobile phase A: Milli-Q® Ultrapure Water
B: Acetonitrile B: Acetonitrile
Gradient Time (min) %A %B Gradient Time (min) %A %B
0.0 55 45 0.0 55 45
5.0 55 45 1.25 55 45
Flow rate 0.27 Flow rate 1.08
Detection UV 247 nm Detection UV 247 nm
Temperature 40 °C Temperature 40 °C
Equilitration 2.5 min Equilitration 0.625 min
Injection 0.7 µL Injection 0.7 µL
volume volume
Sample Alkylphenone standard Sample Alkylphenone standard
47
Ramipril and Related Substances— Ramipril and Related Substances—Purospher™ STAR
from HPLC to UHPLC RP-18 endcapped (HPLC)
The benefit of scaling from HPLC to UHPLC is illustrated 240
with the USP 36–NF 31 monograph method for Ramipril O
related compounds, where the liquid chromatograph O
Intensity (mV)
250–4.0 mm column that contains 3 μm packing L1
Impurity B
O H
(RP-18) and is maintained at a temperature of 65 °C.
120
O N=
Within the scope of allowed monograph method changes, HO =
and only to perform partial revalidation, this method can H
Impurity A
be changed by: 60
48
Ramipril and Related Substances—Purospher™ STAR Ramipril and Related Substances—
RP-18 endcapped (HPLC) from HPLC to UHPLC
240 As can be seen on the left page, both columns meet the
O
performance criteria in terms of:
O
180 • The resolution, R, between ramipril related compound
HN
A and ramipril (not less than 3.0)
Intensity (mV)
O H
O N= • The relative retention time between ramipril related
Impurity B
120
compound A (ramipril RS A), ramipril, and ramipril
HO =
H related compound B (ramipril RS B)
60 Impurity A
• The tailing factor for the ramipril peak (between 0.8
and 2.0).
0 • The application using HPLC conditions also meet the
0 2 4 6 8 10 12
Retention time (min) retention time requirement for ramipril
The UHPLC column, Purospher™ STAR RP-18
Column Purospher™ STAR RP-18e (2 µm) endcapped (2 μm) 100 x 2.1 mm, thus seems to
Hibar® RT 100 x 2.1 mm comply with the monograph method and leads to the
Mobile phase A: Dissolve 2.0 g of sodium perchlorate in a following benefits:
mixture of 800 mL of Milli-Q® Ultrapure Water and
0.5 mL of triethylamine. Adjust pH to 3.6 with 1. Faster method (Time-saving: 40 minutes per
phosphoric acid. Add 200 mL acetonitrile and mix. sample or 360%)
B: Dissolve 2.0 g of sodium perchlorate in a
mixture of 300 mL of Milli-Q® Ultrapure Water and (yes... the column length is 60% shorter and this
0.5 mL of triethylamine. Adjust pH to 2.6 with
phosphoric acid. Add 700 mL acetonitrile and mix.
provides 60% time saving but the real gain is to
scale the method to a column with smaller particle
Gradient Time (min) %A %B
0.0 90 10
size and not having to keep same linear velocity).
1.66 90 10
1.93 75 25
2. Higher chromatographic resolution and efficiency
5.54 65 35
8.31 25 75
11.08 25 75 … but this is not true. The retention time requirement
12.46 90 10 for ramipril is NOT between 16 and 19 minutes. In
15.23 90 10 addition, the flow rate has not been scaled to maintain
Flow rate 0.3 mL/min same linear velocity.
Detection UV 210 nm
The monograph method is documented at 1.0 mL/min
Cell 2.5 µL (use 0.1 mm tubing) on 4.6 mm column, and thus the flow rate should be
Temperature 65 °C reduced by a factor or 4.8 for the 2.1 mm i.d. UHPLC
Diluent Solution A
column (for calculations, see page 18). A flow rate of
0.2 mL/min should have been used instead of 0.3 mL/
Injection volume 2 µL
min. With the current experimental conditions, this would
Pressure drop 196 to 164 Bar (2827 to 2378 psi) give comments from an auditor and very likely a request
Sample Dissolve 25 mg of sample in diluent and dilute for method change.
to 25 mL with same solvent.
The larger Purospher™ STAR RP-18 endcapped (5 μm)
Chromatographic Data
250 x 4.6 mm column can definitely not be used. The
particle size is larger than monograph method and
Compound RT (min) RRT Asymmetry would require complete revalidation and discussion with
1. Ramipril RS A 5.6 0.81 1.1 auditor and authorities. Most likely it would not be an
2. Ramipril 6.9 1.00 1.1
accepted method.
3. Ramipril RS B 9.5 1.38 1.1
Figure 53. Scaling USP method for Ramipril from HPLC conditions to
UHPLC conditions
49
4. C
hromatographic Separation of
Large Molecules
Introduction which is responsible for the effector function, and the
antigen binding domain (Fab), which, as the name
Biomacromolecules (in particular monoclonal implies, is solely dedicated to binding the antigen to
antibodies;mAbs) have seen a renewed interest in the the antibody. mAbs are glycoproteins that have two
pharmaceutical and biotechnology industry. The reason conserved N-glycosylation sites in the Fc domain. A variety
for this high level of interest resides in the number of of other chemical and enzymatic modifications can further
benefits these biological molecules have for patients modify a mAb including methylation, phosphorylation,
including, but not limited to, high efficacy in treating an deamidation, oxidation, and conjugation to cytotoxic,
illness, high specificity for a target receptor or antigen, small molecule drugs (thus resulting in an antibody-drug
wide therapeutic range, and limited undesirable side conjugate; ADC), among many others.3 Therefore, there
effects.1 However, due to the fact that these molecules are thousands of potential variant combinations in a single
are complex and are often produced in host cell lines, mAb formulation, and some of these may elicit a lethal
bacteria, or fermentation reactors, these potential response in a patient. In addition to the above-mentioned
therapeutics have significant heterogeneity which needs modifications to the primary structure of the mAb,
to be evaluated and characterized using analytical modifications to the higher-order structure of the mAb
techniques.2 may occur, such as aggregation or clipping, which can also
affect the safety and efficacy of the therapeutic.4
The complexity of such biomacromolecules can be
easily illustrated by examining the structure of a typical Due to the above-mentioned, inherent danger associated
mAb as depicted in Figure 54. with the possible heterogeneity of these biologics, the
Food and Drug Administration (FDA) and the European
Medicines Agency (EMA) now require all biopharmaceutical
IgG1
Va
n le
L)
R Va n
bl VL
io iab
ha
L gio
(
eg r
R
ig n R
e
(C nt t C
ht
io ns ig
ai sta (CL
L)
gi
n nt )
r
ai le H)
eg
(V e R
ia
eg o
) gi
R C
n
Va
on
ea Fa
n
C
H egio
on
t )
an H1
st
av
R
b
ha
H Re
a
st (C
1) g
on
Hinge Region
N-Linked
Glycan N-Linked Glycan
(CH2) (CH2)
Constant Regions
Constant Regions
Heavy Chain
Heavy Chain
Fc
Size Exclusion Chromatography
Size exclusion chromatography (SEC) is a mode
of chromatography that separates molecules by
(CH3) (CH3) their size (i.e. hydrodynamic radius). This mode of
chromatography does not rely on the interaction of
the analytes with a stationary phase ligand; it is an
Figure 54. Graphic depiction of an IgG1 antibody. Note the structural entropic process meaning that it relies on the random
complexity of the different domains of the mAb. flow of the analytes through the stationary phase
particles. For most practical purposes, this can be
mAbs are large, tetrameric immunoglobulin G (IgG) envisioned as analytes with a higher molecular weight
molecules with a molecular weight of approximately will elute earlier in the run, since these analytes
150 kDa (150,000 g/mol). These molecules form are fully or partially excluded from the pores of the
a Y-shape composed of four peptide chains: two stationary phase particles, while lower molecular weight
identical light (L) chains with a molecular weight of analytes will elute later in the run, since these analytes
approximately 25 kDa each, and two identical heavy (H) will spend more time navigating the torturous path
chains with a molecular weight of approximately 50 kDa through the particles.5
each. To form the Y-shape, these four polypeptide The driving force in SEC is the exclusion effect based
chains associate with each other through the creation of on molecular weight and size. This exclusion effect is
inter- and intra-chain disulfide bonds. From a structural dictated by the pore diameter and geometry of the
biology perspective, a mAb can be broken down into stationary phase particle. Therefore, when selecting
two domains: the fragment crystallizable (Fc) domain, an SEC column (or any column) for a separation
50
experiment, one needs to be cognizant of the particle There have been some developments in the manu-
pore size. If the pore size is too small, the analyte(s) of facturing and application of small particle size SEC
interest will not enter the particle and will elute in the columns for biomolecule separations. These new
void volume. For SEC of smaller proteins and peptides, columns exist as 2.0 μm and sub-2 μm silica particles,
particles with a pore diameter of 150 Å will provide a enabling the biochromatographer to perform high
good linear separation range when creating a molecular efficiency separations in half the time compared to 3.0
weight calibration curve, shown in Figure 55. Table 16 and 4.0 μm particles. Recent studies (7) have shown
details the molecular weights of the analytes used in this the advantages of using small particle SEC column
study. As seen in Figure 55, proteins with molecular technology.
weights larger than approximately 50 kDa (analytes 1–3)
In addition to recent advances in column technology,
begin to be excluded from the pores of the column,
the mobile phase is also important to consider when
resulting in premature elution, whereas analytes with
performing SEC of biomolecules. One area of focus
molecular weights less than 5 kDa (analytes 8–11)
where the composition of the mobile phase will
interact more with the stationary phase and elute later
dictate success or failure in an SEC experiment is in
in the run. To achieve the best chromatographic results
characterizing ADCs by SEC. ADCs tend to exhibit
with this column, one should operate in the linear
secondary interactions with the stationary phase due
region of the curve (analytes 5–7; molecular weights
to the added cytotoxic drug payload. These interactions
17 kDa–6 kDa).6
will result in broad peak tailing and lower sensitivity
of higher-order aggregates. One way to minimize
1000000 1
these interactions is to add an organic alcohol (i.e.
2 isopropanol, 1-butanol, 1-propanol, etc.) to the mobile
Molecular weight
51
Hydrophobic Interaction Chromatography Figure 4 also shows the use of isopropyl alcohol in
Hydrophobic interaction chromatography (HIC) is a mode both mobile phases, essentially setting up an alcohol
of chromatography that separates analytes based on the gradient. The reason for this inclusion is the same
degree of interaction between hydrophobic moieties on as was discussed with SEC: some analytes tend to
the analyte and hydrophobic ligands on the stationary interact too strongly with the stationary phase and
phase. Under conditions of high concentrations of salt, need the extra, organic component to promote elution
the hydration layer around a protein may be disrupted from the column. It should be noted here, though, that
enough that it becomes entropically favorable for both SEC and HIC are considered native techniques,
hydrophobic regions of the protein’s surface to interface meaning that the structure and activity of the protein
with the non-polar stationary phase. This phenomenon is preserved. However, employing alcohols in the
is similar to the classical biochemical technique of mobile phases increase the risk of these analytes being
protein salting out, but in HIC’s case, the interactions denatured. One should always check the stability
are between protein-stationary phase ligand rather than of their analytes in dilute, organic solutions prior to
between different protein molecules. Due to the lower incorporation in a chromatographic method (by using
molecular weight and lower propensity for folding, HIC is a technique like sodium dodecyl sulfate polyacrylamide
not often used for separating peptides. Salt selection in gel electrophoresis; SDS-PAGE). Most proteins will be
HIC is dictated by the Hofmeister series, which classifies stable in solutions up to 35% organic solvent.13
cations and anions in terms of their ability to disrupt the
hydration layer around a protein (chaotropic) or promote
the formation of a hydration layer (kosmotropic). Typical
salts in HIC are ammonium sulfate, potassium sulfate, Ion Exchange Chromatography
and sodium sulfate. Ion exchange chromatography (IEX) is a mode of
The biggest application area currently under investigation chromatography that separates analytes by charge.
for HIC is in determining the DAR profile of an ADC. The Proteins and peptides are amphoteric, meaning that they
DAR profile is one critical quality attribute (CQA) that have both acidic and basic functionalities. The acidic
a biopharmaceutical company needs to determine for portions of a protein include aspartic acid, glutamic
approval by the FDA or EMA. An overly conjugated ADC acid, cysteine, tyrosine, and the α-carboxylate on the
can kill both healthy and carcinogenic cells, whereas an C-terminus. The basic portions of a protein include
under-conjugated ADC may not be effiective in killing arginine, histidine, lysine, and the α-amine on the
carcinogenic cells. In cysteine-linked ADCs, where the N-terminus. Charge variants of a biotherapeutic, another
linker and cytotoxic payload is conjugated through the CQA that regulatory bodies require manufacturers
sulfhydryl moiety of a cysteine amino acid, separation to monitor, can be detected and resolved by IEX.
by HIC leads to a profile of peaks corresponding to 0, 2, These charge variants can arise from mistranslation
4, 6, and 8 drugs attached to the antibody (Figure 57). of messenger RNA (mRNA) transcripts and/or post-
The keen reader will note the presence of a peak between translational modifications such as deamidation,
peaks 2 and 3; this corresponds to a DAR 3 species. It oxidation, or glycosylation, among others.14
is possible that not being able to detect and quantify When choosing a column for an IEX experiment, one
these odd-numbered DAR species could result in an needs to be aware of the isoelectric point (pI) of the
underestimation of the average DAR for an ADC by as native state of the protein. For example, most mAbs
much as 3%.12 have a pI of ~7.4. If the pH of the mobile phase is
lower than 7.4, the mAb will be positively charged and
Elution Order: 3 bind to a cation exchange column. If the pH of the
1. DAR 0 (nativ mAb)
2. DAR 2 mobile phase is above 7.4, the mAb will be negatively
3. DAR 4 charged and bind to an anion exchange column. In
4. DAR 6 addition to cation versus anion exchangers, these can
5. DAR 8
be broken down into weak and strong exchangers
2 based on the pKa for the exchanger.
4 There are two methods for eluting analytes off of an
5 IEX column: using a salt gradient and using a pH
gradient. The salt gradient approach employs a linear
gradient of a salt (i.e. sodium chloride, potassium
0 10 20 30 40 50 chloride, etc.) to essentially compete with the analyte
min binding to the column. Figure 58 shows a separation
of different mAb charge variants, from a series of
Figure 57. HIC/UV analysis of native SigmaMAb ADC mimic. Conditions:
Column: TSKgel® Butyl-NPR, 10 cm x 4.6 mm I.D., 2.5 µm; Mobile phase: proprietary mAbs, by IEX using a salt gradient.15
[A] 50 mM Potassium phosphate, 1.5 M Ammonium sulfate, pH 7.0 plus
5% (v/v) Isopropyl alcohol, [B] 50 mM Potassium phosphate, pH 7.0
plus 20% (v/v) Isopropyl alcohol; Gradient: 0% B to 100% B in 50 min;
Flow rate: 1.0 mL/min; Column temp.: 35 °C; Detector: UV, 215 nm;
Injection: 5.0 µL; Sample: ADC mimic, 100 µg/mL, 50 mM Potassium
phosphate, pH 7.0. Adapted from Reference 12.
52
40 To determine the pattern of glycosylation on a potential
biotherapeutic, generally one first releases the glycans
from the protein using an enzyme such as PNGase F.
30
Afterwards, the released glycans are labeled with a
Intensity (mV)
53
Absorbance at 280 nm
4
on hydrophobicity. Unlike HIC, RPC employs a water/ 40000000
2
5
organic mixture for the mobile phase. There are several 3
1. Insulin-Glargine
parameters that can be tuned in an RPC experiment to get 30000000 2. Insulin-Bovine
3. Insulin-Asp
satisfactory peak shape and resolution; these have been 4. Insulin-LisPro
5. Insulin-Human
reviewed in several recent publications.18,19 20000000
6. Insulin-Porcine
20000000
Another key aspect in RPC of biomacromolecules is
the use of an ion pair reagent in the mobile phase. 0
3 3.5 4 4.5 5 5.5 6 6.5 7
Generally, trifluoroacetic acid (TFA) has been used
as a good ion pair reagent as it masks interactions Retention time (min)
between the analytes and free, active silanols on the Figure 62. Separation of insulin variants by LC/MS using different
stationary phase. However, this ion pair reagent can mobile phase modifiers. Conditions: Column: BIOshell™ A160 Peptide
cause severe ion suppression when coupled to MS C18, 15 cm x 2.1 mm I.D., 2.7 µm; Mobile Phase: [A] 75:25 10 mM
ammonium formate, pH 2.6 with formic acid or water (0.1% (v/v)
detection. Therefore, alternatives, like difluoroacetic DFA): acetonitrile; [B] 50:50 10 mM ammonium formate, pH 2.6
acid (DFA) and ammonium formate, should be used for with formic acid or water (0.1% (v/v) DFA): acetonitrile (77:23 A:B);
MS analyses as they tend to provide good ion pairing Flow Rate: 0.2 mL/min; Column Temperature: 75 °C; Detector: MSD,
ESI-(+), TIC 100–3000 m/z; Injection: 0.5 µL; Sample: Mixture of six
capacity while minimizing ion suppression. Figure 62
insulin variants, 100 µg/mL, 10 mM ammonium formate, pH 2.6 with
shows a comparative study of these two ion pairing formic acid. Adapted from Reference 20.
reagents in separating a series of insulin variants. Note
the ~2.7-fold increase in sensitivity using ammonium
formate.20
54
References
1. A. J. Chirino, A. Mire-Sluis, Nat. Biotechnol 22, 12. B. Bobaly, G. M. Randazzo, S. Rudaz, D. Guillarme,
1383 – 1391 (2004). S. Fekete, J. Chromatogr. A 1481, 82 – 91, (2017).
2. K. Sandra, I. Vandenheede, P. Sandra, J. 13. C. N. Pace, S. Trevino, E. Prabhakaran, J. M.
Chromatogr. A 1335, 81 – 103 (2014). Scholtz, Phil. Trans. R. Soc. Lond. B 359, 1225 –
1235 (2004).
3. H. Liu, G. Ponniah, H. M. Zhang, C. Nowak, A. Neill,
N. Gonzalez-Lopez, R. Patel, G. Cheng, A. Z. Kita, 14. E. Wagner-Rousset, S. Fekete, L. Morel-Chevillet,
B. Andrien, mAbs 6, 1145 – 1154 (2014). O. Colas, N. Corvaia, S. Cianferani, D. Guillarme, A.
Beck, J. Chromatogr. A 1498, 147 – 154 (2017).
4. A. Beck, E. Wagner-Rousset, D. Ayoub, A. Van
Dorsselaer, S. Sanglier-Cianferani, Anal. Chem. 85, 15. R. Romling, Reporter, 29.3, 20 – 21 (2011).
715 – 736 (2013).
16. V. D’Atri, S. Fekete, A. Beck, M. Lauber, D.
5. J. H. Knox, High Performance Liquid Guillarme, Anal. Chem. 89, 2086 – 2092 (2017).
Chromatography, Edinburgh University Press,
17. A. Periat, S. Fekete, A. Cusumano, J. Veuthey, A.
1978.
Beck, M. Lauber, D. Guillarme, J. Chromatogr. A
6. R. A. Henry, S. Schuster, “Impact of Pore Exclusion 1448, 81 – 92 (2016).
on Reversed-Phase HPLC Column Performance,”
18. A. Beck, S. Cianferani, A. Van Dorsselaer, Anal.
presented at the Eastern Analytical Symposium
Chem. 84, 4637 – 4646 (2012).
(EAS 2016), Somerset, New Jersey, 2016.
19. A. Beck, G. Terral, F. Debaene, E. Wagner-Rousset,
7. H. K. Brandes, S. Shollenberger, C. E. Muraco,
J. Marcoux, M. C. Janin-Bussat, O. Colas, A. Van
Reporter 34.3, 33 – 34 (2016).
Dorsselaer, S. Cianferani, Expert Rev. Proteomics
8. X. M. Lu, K. Benedek, B. L. Karger, J. Chromatogr. 13, 1 – 26 (2016).
A 359, 19 – 29 (1986).
20. C. E. Muraco, S. Shollenberger, H. K. Brandes,
9. S. Fekete, S. Rudaz, J. Veuthey, D. Guillarme, J. “Reversed-Phase HPLC Analysis of Insulin Variants
Sep. Sci. 35, 3113 – 3123 (2012). and Analogs by UV and Mass Spectral Detection,”
presented at the American Association of
10. A. Goyon, A. Beck, J. Veuthey, D. Guillarme, S.
Pharmaceutical Scientists National Biotechnology
Fekete, J. Pharm. Biomed. Anal. 144, 242 – 251
Conference (AAPS NBC 2016), Boston,
(2017).
Massachusetts, 2016.
11. C. E. Muraco, K. Ray, G. Oden, D. S. Bell, “Using
All the Tools in the Toolbox: Characterization of A
Novel Antibody-Drug Conjugate Mimic by Several
Modes of Chromatography,” presented at the
International Symposium on the Separation of
Proteins, Peptides, and Polynucleotides (ISPPP
2017), Philadelphia, Pennsylvania, 2017.
55
5. Troubleshooting Common HPLC Issues
Introduction Problem No. 2: No Flow
Although HPLC method development has been improved Normal
by advances in column technology and instrumentation,
problems still arise. In this chapter, a systematic means
of isolating, identifying, and correcting many typical
Detector Responce
problems encountered in the practice of HPLC will be
given. The important segments of an HPLC system are
the same, whether using a modular system or “cutting-
edge” UHPLC instruments. Problems affecting overall
system performance can arise in each component.
Some common problems are discussed here.
Problem
Responce
Detector
Time
Time
Probable Causes:
1. Pump off.
1. Start pump.
2. Check mobile phase level in reservoir(s). Check
Time flow throughout system. Examine sample loop for
obstruction or air lock. Make sure mobile phase
components are miscible and mobile phase is
Probable Causes: properly degassed.
1. Detector off or not sufficiently “warmed up”.
3. Check system for loose fittings. Check pump for
2. Poor/no connection between instrument and computer. leaks, salt buildup, or unusual noises. Change pump
3. No mobile phase flow. seals if necessary.
4. No sample/deteriorated sample/wrong sample.
4. Disconnect tubing at guard column (if present) or
5. Settings incorrect on detector. analytical column inlet. Check for flow. Purge pump
at high flow rate (e.g., 5–10 mL/min), prime system
Possible Solutions: if necessary (prime each pump head separately). If
1. Turn detector on or allow sufficient time for detector system has check valve, loosen valve to allow air to
to “warm up”. escape. If problem persists, flush system with 100%
2. Check connectivity between instrument and computer. methanol or isopropanol. If problem still persists,
3. See “No Flow” (Problem No. 2). contact system manufacturer.
4. Be sure autosampler vials have sufficient liquid
and no air bubbles in the sample. Evaluate system
performance with fresh standard to confirm sample
as source of problem.
5. Check detector status/settings. Auto-zero if necessary.
56
Problem No. 3: No Pressure/Pressure eliminating system components, starting with
Lower Than Usual detector, then in-line filter, and working back to
Probable Causes: pump. Replace filter in pump if present.
1. Leak. 2. Remove guard column (if present) and check
pressure. Replace guard column if necessary. If
2. Mobile phase flow interrupted/obstructed. analytical column is obstructed, reverse and flush
3. Air trapped in pump head (revealed by pressure the column, while disconnected from the detector.
fluctuations). If problem persists, column may be clogged with
strongly retained contaminants. Use appropriate
4. Leak at column inlet end fitting. restoration procedure (Appendix). If problem still
5. Air trapped elsewhere in system. persists, replace column.
Detector Responce
Normal
Possible Solutions:
1. Check system for loose fittings. Check pump for
leaks, salt buildup, or unusual noises. Change pump
seals if necessary.
2. Check mobile phase level in reservoir(s). Check
flow throughout system. Examine sample loop for
obstruction or air lock. Make sure mobile phase Time
components are miscible and mobile phase is
properly degassed. Detector Responce
Problem
3. Disconnect tubing at guard column (if present) or
analytical column inlet. Check for flow. Purge pump
at high flow rate (e.g., 10 mL/min), prime system
if necessary (prime each pump head separately). If
system has check valve, loosen valve to allow air to
escape.
4. Reconnect column and pump solvent at double the Time
flow rate. If pressure is still low, check for leaks at Problem
inlet fitting or column end fitting.
Responce
Detector
57
Possible Solutions: Problem No. 7: Split Peaks
1. Check system for loose fittings. Check pump for
leaks, salt buildup, or unusual noises. Change pump
Normal
Detector Responce
seals if necessary.
2. Check make-up of mobile phase. If mobile phase
is machine mixed using proportioning values, hand
mix and supply from one reservoir.
3. Purge air from pump head or check valves. Change
pump seals if necessary. Be sure mobile phase is
degassed. Time
4. Use a reliable column oven. Note that higher
column temperatures increase column efficiency. Problem
For optimum results, heat eluant before introducing
it onto column.
Detector Responce
5. Inject smaller volume (e.g., 1.0 μL vs. 10.0 μL)
or inject the same volume after 1:10 or 1:100
dilutions of sample.
6. Adjust solvent. Whenever possible, inject samples
in starting mobile phase.
7. Substitute new column of same type to confirm
column as cause. Discard old column if restoration
procedures fail.
Time
Probable Causes:
1. Mobile phase contaminated/evaporated (causing
retention times and/or selectivity to change).
2. Obstructed guard or analytical column.
Possible Solutions:
1. Prepare fresh mobile phase.
2. Remove guard column (if present) and attempt
analysis. Replace guard column if necessary. If
analytical column is obstructed, reverse and flush.
If problem persists, column may be clogged with
strongly retained contaminants. Use appropriate
restoration procedure (Appendix). If problem still
persists, replace column.
58
Problem No. 8: Peaks Tail on Initial and Problem No. 9: Tailing Peaks
Later Injections
Normal
Normal
Detector Responce
Detector Responce
Time
Time
Problem
Detector Responce
Problem
Responce
Detector
Time
Time
Probable Causes:
Probable Causes: 1. Guard or analytical column contaminated/worn out.
1. Sample reacting with active sites. 2. Mobile phase contaminated/evaporated.
2. Wrong mobile phase pH. 3. Interfering components in sample.
3. Wrong column type.
4. Small (uneven) void at column inlet. Possible Solutions:
1. Remove guard column (if present) and attempt
5. Wrong injection solvent.
analysis. Replace guard column if necessary.
If analytical column is source of problem, use
Possible Solutions: appropriate restoration procedure (Appendix). If
1. First check column performance with standard problem persists, replace column.
column test mix. If results for test mix are good, 2. Check make-up of mobile phase.
add ion pairing reagent or competing base or acid
modifier. 3. Check column performance with a column test mix.
59
Problem No. 10: Fronting Peaks Problem No. 11: Rounded Peaks
Normal Normal
Detector Responce
Detector Responce
Time
Time
Problem
Problem
Responce
Detector
Detector Responce
Time
Probable Causes:
1. Detector operating outside linear dynamic range.
Time
2. Column overloaded.
3. Sample-column interaction.
Probable Causes:
4. Detector time constants are set too high.
1. Column overloaded.
2. Sample solvent incompatible with mobile phase. Possible Solutions:
3. Shoulder or gradual baseline rise before a main 1. Reduce sample volume and/or concentration.
peak may be another sample component. 2. Inject smaller volume (e.g., 1.0 μL vs. 10.0 μL) or
1:10 or 1:100 dilution of sample.
Possible Solutions:
3. Change buffer strength, pH, or mobile phase
1. Inject smaller volume (e.g., 1.0 μL vs. 10.0 μL). composition. If necessary, raise column
Dilute the sample 1:10 or 1:100 fold in case of temperature or change column type. (Analysis of
mass overload. solute structure may help predict interaction.)
2. Adjust solvent. Whenever possible, inject samples 4. Reduce settings to lowest values or values at which
in mobile phase. Flush polar bonded phase column no further improvements are seen.
with 50 column volumes HPLC grade ethyl acetate
at 2–3 times the standard flow rate, then with
intermediate polarity solvent prior to analysis.
3. Increase efficiency or change selectivity of system
to improve resolution. Try another column type if
necessary (e.g., switch from nonpolar C18 to polar
cyano phase).
60
Problem No. 12: Baseline Drift Problem No. 13: Baseline Noise (regular)
Normal
Responce
Detector
Normal
Responce
Detector
Time
Time
Problem
Responce
Detector
Problem
Responce
Detector
Time
Time
Probable Causes:
1. Column temperature fluctuation. Even small changes Probable Causes:
can cause cyclic baseline rise and fall. Most often affects 1. Air in mobile phase, detector cell, or pump.
refractive index and conductivity detectors, UV detectors
at high sensitivity or in indirect photometric mode. 2. Pump pulsations.
2. Nonhomogeneous mobile phase. Drift is usually to 3. Incomplete mobile phase mixing.
higher absorbance, rather than cyclic pattern from
4. Temperature effect (column at high temperature,
temperature fluctuation.
detector unheated).
3. Contaminant or air buildup in detector cell.
4. Plugged outlet line after detector. High pressure 5. Leak.
cracks cell window, producing noisy baseline.
5. Mobile phase mixing problem or change in flow rate. Possible Solutions:
6. Slow column equilibration, especially when 1. Degas mobile phase. Flush system to remove air
changing mobile phase. from detector cell or pump.
7. Mobile phase contaminated, deteriorated, or not 2. Incorporate pulse damper into system.
prepared from high quality chemicals.
8. Strongly retained materials in sample (high 3. Mix mobile phase by hand or use less viscous solvent.
retention factor (k)) can elute as very broad 4. Reduce differential or add post-column cooler.
peaks and appear to be a rising baseline. Gradient
analyses can aggravate problem. 5. Check system for loose fittings. Check pump for
9. Detector (UV) not set at absorbance maximum but leaks, salt buildup, or unusual noises. Change pump
at slope of curve. seals if necessary.
Time
Probable Causes:
1. Mobile phase composition changed.
2. Mobile phase flow rate too low.
3. Leak (especially between column and detector).
4. Detector settings incorrect.
62
Problem No. 16: Change in Peak Height Problem No. 17: Change in Selectivity
(one or more peaks)
Normal
Normal
Detector Responce
Detector Responce
Time
Problem Time
Detector Responce
Problem
Detector Responce
Time
Probable Causes:
1. One or more sample components deteriorated or
column activity changed. Time
63
Problem No. 18: Negative Peak(s) Problem No. 19: Ghost Peak
Detector Responce
Time Time
Problem Normal
Responce
Detector
Detector Responce
Time
(solvent injected after sample)
Problem
Responce
Detector
Time
Time
(solvent injected after sample)
Probable Causes:
1. Refractive index of solute less than that of mobile
phase (RI detector).
Probable Causes:
2. Sample solvent and mobile phase differ greatly in
1. Contamination in injector or column.
composition (vacancy peaks).
2. Late eluting peak (usually broad) present in sample.
3. Mobile phase more absorptive than sample
components to UV wavelength.
Possible Solutions:
Possible Solutions: 1. Flush injector between analyses (a good routine
practice). If necessary, run strong solvent through
1. Use mobile phase with lower refractive index.
column to remove late eluting compounds. Include
2. Adjust or change sample solvent. Dilute sample in final wash step in gradient analyses, to remove
mobile phase whenever possible. strongly retained compounds.
3. Change UV wavelength or use mobile phase that 2. a. Check sample preparation.
does not adsorb at chosen wavelength.
b. Include (step) gradient to quickly elute component.
64
Common Column Restoration Part 3: Polar-Bonded Reversed Phase
Columns (Amino, Cyano, Diol, Chiral)
Procedures
To restore these types of columns, the following
*Note: The below volumes are based on optimized procedure should be employed:
procedures for columns that are 4.6 mm I.D. For
different I.D. columns, please convert volumes by taking 1. For a column used in the reversed phase mode
the ratio of the square of the new I.D. to (4.6 mm I.D.)2 (e.g., organic solvent/aqueous buffer mobile
and multiplying this conversion factor to the volume. phase), follow the same cleanup procedure as for
Please check the column Care and Use Guide to ensure silica-based reversed phase columns. For a column
compatibility with these solvents prior to employing these used with nonaqueous mobile phases, use the
strategies to restore the column. following scheme:
2. Flush with the following:
Part 1: Bare Silica Columns
a. 50 mL chloroform
To restore a bare silica column, the following procedure
should be used: b. 50 mL methanol
1. 50 mL hexane c. 50 mL acetonitrile
2. 50 mL methylene chloride d. 25 mL methylene chloride
3. 50 mL 2-propanol e. 25 mL methanol
4. 50 mL methanol f. 25 mL mobile phase
5. 25 mL methylene chloride
6. 25 mL mobile phase Part 4: Silica-Based Ion Exchange Columns
To restore these ion exchange columns, the following
procedure should be employed:
Part 2a: Silica-Based Reversed Phase Columns
1. 50 mL hot (40–60 °C) distilled water
Analyzing Water-Soluble Compounds
To restore a reversed-phase column that was used 2. 50 mL methanol
in analyzing water-soluble compounds, the following 3. 50 mL acetonitrile
procedure should be employed:
4. 25 mL methylene chloride
1. Flush with warm (60 °C) ultrapure water
5. 25 mL methanol
2. 50 mL methanol
6. 25 mL mobile phase
3. 50 mL acetonitrile
4. 25 mL methanol
5. 25 mL mobile phase
65
6. R
eference Materials in LC-MS/MS
Applications: Quality Grades &
Selection Considerations
Choosing the correct reference Mass—
Kilogram
material quality grade for Luminous
your needs Intensity—
Candela kg Length—
Metre/Meter
m
cd
Who uses reference materials?
Reference materials are a critical component of the
analytical testing workflow. Through calibration of
mol
measurement systems, validation of methods, and Amount of
s
Time—
quality control programs, reference materials ensure Substance—
Second
accuracy in testing. From certified reference materials Mole
(CRMs) and other quality grades of reference materials,
to certificates of analysis, metrological traceability, and K A
other concepts, the world of reference materials is vast,
and can be confusing. Temperature— Electric Current—
Kelvin Ampere
This chapter presents critical reference material topics,
offering information on metrological traceability, the Figure 63. Metrological Traceability—SI Unit of Measurement
hierarchy of reference materials, certificates of analysis,
reference material formats and uses, as well as fit-for-
purpose selection considerations. Proper selection of ISO 17034 and quality grades of standards,
the right reference material for a laboratory’s testing reference materials, and certified reference
application is important because results are only as materials
accurate as your reference.
The reference material hierarchy includes five major
quality grades from national metrology and other
Metrological traceability and SI Units of your primary standards, to CRMs, reference materials
reference materials (RMs), analytical standards, and research grade or
Metrological traceability is an important concept in the research chemicals. Higher levels of certification
world of reference materials. A fundamental term in and traceability are required with increasing levels
metrological traceability is the International System of quality grade. While national governments give
of Units (SI) unit of measurement. The SI defines the standardization to the top level, specific ISO guidelines
seven units of measure as the basic set from which all provide standardization for CRMs and RMs. These ISO
other SI units can be derived. The two most common SI requirements include ISO 17034, ISO/IEC 17025, and
units of measure for traceability of reference materials ISO Guide 31.
are the kilogram and mole.
Reference material producers must meet these ISO
Metrological traceability means measurements can requirements to manufacture CRMs or RMs. For both of
be meaningfully compared across difference places, these quality grades, Certificates of Analysis must be
at different times, by different people, using different provided and the information contained within is defined
equipment. The measurement result must be related to a by the aforementioned ISO guidelines. The quality
reference through a documented and unbroken chain of specifications for the last two levels are defined by each
calibrations, tracing back to the SI unit of measurement. individual producer rather than by a national government
or ISO accreditations specific to CRMs and RMs.
66
National Metrology Standard (e.g. NIST, JRC, NMI Australia)
Compendial Standard (e.g. USP, EP, BP, JP, IP)
• Issued by an authorized body
• Considered to provide the highest level of accuracy &
traceability
for CRMs
ilit
Figure 64. The Hierarchy of Reference Materials—What are the Different Types?
What is measured in each grade of Homogeneity is required for the primary standards,
reference material? CRMs, and RMs, but this parameter will not be found
with the lower quality grades. Uncertainty and
Purity and identity of the material are typically included
traceability information are limited to only primary
in the Certificate of Analysis for each of the five quality
standards and CRMs. In the pharmaceutical world,
grades. Content and stability are required for primary
secondary standards can be CRMs or RMs, but here,
standards or ISO-defined CRMs and RMs.
there are two different types of traceability: to the
Analytical standards and research chemicals may or SI unit of measurement for ISO-defined CRMs, and
may not include these two parameters as their inclusion traceability to the primary compendial standard, which
is dependent on the producer. Analytical standards is a requirement specific to pharmaceutical secondary
can also in some cases be quality control materials standards.
compliant with ISO Guide 80.
67
Understanding your reference material Be sure to examine the CoA for the producer’s quality
Certificate of Analysis systems, the reference material’s certification process,
and supporting information on traceability for a CRM.
With the CRM or RM grade comes a Certificate of
The CoA is important since it can give the laboratory
Analysis (CoA). Within the CoA, there are several
information which ensures the reference material’s
quality parameters which are critical to understand:
certification is fit for purpose within the testing method
accuracy, consistency, homogeneity, purity, and
or application.
stability. Also, which property is being certified is
important to understand, whether it be concentration,
potency, or content.
Accuracy
Comparison to a primary source
or certified second source—curve/
calibration standard. Comparison of
Certificate of Analysis – Certified Reference Material multiple independent preparations.
Metrological traceability: Certified values are traceable to the International System of units (SI)
through a metrologically valid weighing process. Details see “Details on
Purity
metrological traceability”. [3] Consistent with the neat material.
Measurement method: The certified value is determined by high-precision weighing of thoroughly
characterized starting materials and verified by measurement against NIST No contamination or degradation.
SRMs or similar CRMs
Intended use: Calibration of ICP, AAS, spectrophotometry or any other analytical technique.
Instructions for handling This reference material shall be stored in the original closed bag between 5°C
and correct use: and 30°C. Before every use of the material the bottle must be shaken well
and its temperature has to be 20°C. If storage of a partially used bottle is
necessary, the cap should be tightly sealed and the bottle should be stored at
reduced temperature (e.g. refrigerator) to minimize transpiration rate. We
highly recommend using this reference material no longer than 15 months
after the aluminum bag was opened.
Health and safety Please refer to the Safety Data Sheet for detailed information about the
information: nature of any hazard and appropriate precautions to be taken.
Accreditation: Sigma-Aldrich Production GmbH is accredited by the Swiss accreditation
authority SAS as registered reference material producer SRMS 0001 in
accordance with ISO 17034 and registered testing laboratory STS 0490
according to ISO/IEC 17025.[4][5]
Certificate issue date: 04 APR 2019
Packaging: 100 mL HDPE bottle sealed with an aluminized bag
ISO 17034 ISO/IEC 17025 ISO 9001 S. Matt – CRM Operations Dr. P. Zell – Approving Officer
SRMS 0001 STS 0490 005356 QM08
68
Reference material formats: do you need a Choosing the correct reference material for
neat, solution, or matrix material? your testing purpose
Reference materials can be used in different formats in For instrument qualifications and calibrations, establishing
the testing laboratory depending on product availability and maintaining traceability is key, and the selected
and method requirements. The three formats for reference reference material should help the laboratory achieve
materials are a neat or powder form, in solution, or this. In daily routine system suitability applications, it may
matrix. The Supelco® family of reference materials be important to qualify something that is practical and
include CRMs, RMs, or analytical standards in each easy-to-use, yet reliable and cost-effective for everyday
format depending on the testing laboratory’s needs applications In method validation, it is critical to use
and application. highly accurate and precise materials to ensure a method
maintains these parameters. For identity and screening
purposes, proven authenticity and identity are important
attributes of reference materials. For quantitation, assays,
or stability assessment, stable and accurate reference
materials are needed.
Table 18. Different Formats—How Reference Materials are Used in the Testing Laboratory
Type of test Use of Ref. Mat. Examples Requirements of the Ref. Mat.
Instrument qualification/ Establish system performance Annual qualifications Traceable
Calibration
Measurement accuracy Routine balance calibrations
Routine calibration/ Daily/weekly Pre-use balance calibrations Qualify as suitable for use
System suitability
System/method specific System performance checks for
LC-UV/MS; GC-FID…
Establish routine performance
Method validation Accuracy Pharma QC; Environmental testing Accurate
Precision Standards of the analyte(s), Traceable
interferences, impurities
Specificity & interferences
LOD/LOQ & Linearity
Identity Comparison of unknown to Incoming raw materials in Authenticity
known pharma, food etc.
Screening tests
Content or assay Quantitation of analytes Pesticide/toxin limits Certified content
Pharma QC—API content Traceable
Stability assessment Monitor product stability Pharma QC Stable, homogenous
Internal Quality Control Method accuracy Routine quantitation of analytes Certified content
—pharma/pesticides/diagnostics
Traceable
69
Which quality grade is the best fit for purpose? Keep balances and pipettors maintained and calibrated.
Fit for purpose decisions in selection of reference For all equipment, balances, and pipettors in particular, read the
materials can depend on several factors, from regulatory manufacturer documentation and user guides to understand best
requirements, availability, and type of testing application, practices.
to level of accuracy and sample matrix. A fit for purpose To obtain best precision and accuracy, do not use volumes that are
guidance in standard selection can be found in Table 20. less than 20% of the total volume of the pipettor.
If in doubt about the performance of a pipettor, or to develop your
technique, try dispensing water into a container and assessing
gravimetrically.
Best Practice for Preparation of Be sure that solutions are neither hot nor cold to minimize volumetric
errors.
Calibration Curves in Matrix
Considerations when preparing Preparing Calibration Standards:
calibration standards from CRMs Keep organic content in the control matrix to a minimum—2% or
less is recommended when preparing calibrators. One easy way
to prepare a set of calibration standards is to first prepare several
Recommendations for preparing calibration standards working solutions (WS) from a higher concentration stock solution.
Prepare each working solution such that, when added to the blank
Store compounds under the recommended conditions to ensure stability.
matrix (plasma or serum for example) at a volume of 2% or less of
After removing stock solutions from storage check, that analytes have the final concentration, the resulting solution provides the desired
gone back into solution as in some cases compounds may fall out of concentration.
solution under cold storage conditions. In this case, it may be necessary
— For example, combine 20 µL, of a working solution at 10 µg/mL,
to mix or sonicate solutions for periods of time to aid dissolution.
to 980 µL of control plasma to yield 1 mL of a 200 ng/mL
For neat materials, allow time for the container to warm to room calibration standard. Prepare additional working solutions at
temperature before opening in order to eliminate condensation from appropriate concentrations so that the same volumes can be
forming inside the container. used to prepare a complete calibrator series.
In weighing out neat compounds, it is generally best to weigh out a The strategy of preparing one high calibration standard from a stock
larger mass of material if possible (though this may not always be or working solution and then diluting this standard further is often
economical). considered less desirable, as any inaccuracy in the first solution will
be carried through the entire calibration series.
To help reduce transfer losses, consider using small volumetric flasks
(1 to 10 mL), weigh material into a small aluminum weighing pan, Have a second individual weigh out and prepare solutions of the
and drop the entire pan into the flask. same compounds separately. Then, check solutions against each
other using an appropriate analytical technique. Solutions should
For organic solutions, use positive displacement pipettors to get show agreement within a few percent of each other.
accurate dispensing of high vapor pressure solvents. Be sure that air
bubbles have been removed from the tip before dispensing.
70
Appendix
Abbreviations Links and Literature for Download
• LC-MS Resource guide
APCI Atmospheric pressure chemical ionization
CA Charged aerosol (detector) • HPLC Columns selection guide for small molecules
separation wallchart
DAD Diode array detector
• Chiral selection guide wallchart
DMSO Dimethyl sulfoxide
EC Electrochemical (detector) • HPLC Troubleshooting Guide “Untangle your liquid
chromatography problems”
ELS Evaporative light scattering (detector)
• Ascentis® Express columns brochure
ESI Electrospray Ionization
FL Fluorescence (detector) • BIOshell™ columns brochure
71
72
Merck KGaA
Frankfurter Strasse 250
64293 Darmstadt, Germany
MerckMillipore.com
To place an order or receive technical assistance in Europe, please call Customer Service:
France: 0825 045 645 Spain: 901 516 645 Option 1
Germany: 069 86798021 Switzerland: 0848 645 645
Italy: 848 845 645 United Kingdom: 0870 900 4645
© 2021 Merck KGaA, Darmstadt, Germany and/or its affiliates. All Rights Reserved. Merck, the vibrant M, Supelco,
Sigma-Aldrich, Chromolith, Ascentis, BIOshell, Purospher, Milli-Q, Supel, Discovery, Titan, SeQuant, LiChroCART,
Hibar, ZIC, Durapore, Express Plus, Extrelut, CHIROBIOTIC, CapRod, and Millex are trademarks of Merck
KGaA, Darmstadt, Germany or its affiliates. All other trademarks are the property of their respective
owners. Detailed information on trademarks is available via publicly accessible resources.