017_CHROMTODAY_May_2011 Ferguson_Huet_Article:ChromatographyToday
24/5/11
14:34
Page 17
17
Correlation of Analyte Retention in
Organic and Inorganic Mobile Phases
to aid Liquid Chromatography
Method Development
Paul Ferguson1* and Ronan Huet2
1
Research Analytics, Pfizer Global Research & Development, Sandwich, Kent, CT13 9NJ. *Corresponding author: [Link]@[Link]
2
Devices Centre of Emphasis, Pfizer Global Research & Development, Sandwich, Kent, CT13 9NJ.
Many liquid chromatography methods are developed using organic mobile phase additives which allow compatibility with mass-spectrometric
(MS) detection. However, these types of additives often give high UV absorbance which can lead to low level impurity quantitation issues.
Additionally, these additives often have little or no buffering capacity at the pH they are typically used, which in turn can lead to variability in
analyte retention time. A rational approach for the selection of phosphate buffers from organic based mobile phase additives of the same pH
(acidic or neutral) in liquid chromatography stability indicating method (SIM) development may provide a solution to this problem. Excellent
correlation was observed for analyte retention (33 test analytes) in switching from an organic based mobile phase additive to an appropriate
potassium phosphate buffer at low and mid pH. This approach provides a basis for developing SIM methods under mass-spectrometer
friendly conditions and converting them directly to phosphate methods (or vice-versa) which typically provide higher UV sensitivity and
retention robustness while maintaining the elution order and chromatographic resolution observed with the organic mobile phase additives.
Keywords: UHPLC, phosphate buffers, retention correlation, ion-pairing
Introduction
In the pharmaceutical industry, early clinical
phase analytical method development
requires a balance between spending
sufficient time and effort in developing a fitfor-purpose method against the very real
possibility that the compound may be halted
before the next project milestone is reached
(e.g. due to toxicology or compound
absorbance issues). Accordingly, the starting
point for chromatographic method
development typically involves screening
relevant samples through a set of generic
methods and choosing the conditions offering
best retention, analyte resolution and peak
shape [1-4]. The method providing best global
resolution is generally selected and further
optimisation undertaken. This process is
often supplemented using other key
information (both measured and predicted insilico e.g. physical-chemical parameters such
as pKa and/or Log D) and chromatographic
predictive software such as DrylabTM (Molnar
Instiut, Berlin, Germany) [5,6], ChromSwordTM
(Software Entwicklung, Muehltal, Germany) [7]
or LC SimulatorTM (ACDLabs, Toronto,
Canada) [8]. This approach often greatly
reduces both the amount of resource and
time required to develop a suitable method.
The generic screening systems may be based
impurities, while UV detection (often diode-
on HPLC or UHPLC instrumentation, but are
array detection (DAD), also known as photo
typically hyphenated to both UV and mass
diode array (PDA)) is typically used for
spectrometric (MS) detectors. The use of MS
quantitation of impurities by area
detection, particularly in early-phase
normalisation. The use of MS detection
pharmaceutical development, is essential for
requires LC mobile phases which are both
early characterisation of new and unknown
volatile and promote ionisation of the sample
Effective pH range or
Additive
pKa
1.9
0.1% trifluoroacetic acid (TFA)
<1.0
2.8 4.8
formic acid
3.8 (HCOO-)
3.8 5.8
acetic acid
4.8 (CH3COO-)
6.8 11.3
ammonium bicarbonate
7.8 (H2CO32-)
9.2 (NH3+)
10.3 (HCO3-)
3.8 5.8
8.2 10.2
ammonium (acetate)
4.8 (CH3COO-)
9.2 (NH3+)
2.8 4.8
8.2 10.2
ammonium (formate)
3.8 (HCOO-)
9.2 (NH3+)
9.7 11.7
Triethylamine (TEA) (acetate)
4.8 (CH3COO-)
10.7 (TEA+)
10.6
0.1% ammonium hydroxide
(NH4OH)
9.2 (NH3+)
10.3 - 12.3
pyrrolidine
11.3 [9,10]
commonly used pH
Table 1. Some common mobile phase additives employed in LC/MS work. Note higher pH mobile phases require suitably
stable columns such as Agilent Zorbax Extend, Phenomenex Gemini or Waters BEH/Acquity phases.
017_CHROMTODAY_May_2011 Ferguson_Huet_Article:ChromatographyToday
18
24/5/11
14:34
Page 18
May/June 2011
analytes. One benefit of these mobile phases
Method
Column
Mobile Phase
Organic Solvent
(A)
(B)
is they are typically quick and easy to prepare
(for example they are not often pH adjusted
Waters Acquity
Shield RP18
0.1% formic acid
(pH 2.6)
acetonitrile
0-1.7 min 5% B
1.7-8.7 min linear
to 95% B
8.7-10.4 min 95%
B
Waters Acquity
C18
10 mM
ammonium
acetate (pH 6.8)
methanol
0-1.7 min 5% B
1.7-8.7 min linear
to 95% B
8.7-10.4 min 95%
B
Waters Acquity
Phenyl
0.1% ammonium
hydroxide (pH
10.6)
methanol
0-1.7 min 5% B
1.7-8.7 min linear
to 95% B
8.7-10.4 min 95%
B
and are used as is). However, as these
additives are organic molecules (in
comparison to inorganic buffers such as
phosphate or borate), they often have
appreciable background absorbance and (in
our experience) often contain impurities which
can negatively impact chromatographic
gradient profile, peak identification and limits
of detection. Additionally, these additives are
used to simply adjust the pH of the mobile
phase and typically offer little buffering
capacity. For example, 10 mM ammonium
acetate has an aqueous pH of 6.8 which has
no buffering capacity. If these mobile phases
are used for project progression activities such
Gradient
Table 2. Organic buffers, solvents and columns investigated in this study.
as SIMs, this may lead to issues with irregular
If correlations were identified, then this
separations were investigated in an effort to
and non-reproducible analyte retention times,
approach could be used as part of a method
identify alternative mobile phase additives
poor peak-shapes and ultimately poor
development strategy whereby a method
offering higher UV sensitivity and reduced
method robustness. A list of some organic
developed with MS compatible mobile
baseline artefacts.
additives typically used in LC/MS analyses are
phases for impurity identification could be
listed in Table 1.
transferred directly to an appropriate
During method development if lower limits of
detection or increased robustness are
required, then inorganic additives should be
used. Most often, the additive chosen is a
phosphate based salt. Phosphate has a wide
buffering capacity (pKa values of = 2.15, 7.20
and 12.30) and excellent UV opacity down to
phosphate method for increased UV
sensitivity and method robustness. This could
be incorporated into a Quality-by-Design
(QbD [16]) framework for SIMs or used alone
e.g. for transferring phosphate based
methods to MS-friendly buffers for new
impurity identification.
Dihydrogen potassium phosphate and
dipotassium hydrogen phosphate (both
anhydrous) were purchased from Fluka
Analytical (HPLC grade). Phosphoric acid
(Fluka Analytical puriss 85% in water, HPLC
grade) was used to pH adjust the mobile
phases to the identical pH of the organic
buffers they were being compared with.
Purified water was obtained from a Millipore
wavelengths as low as 200 nm [11]. This often
Additionally, high pH buffers were also
makes phosphate the buffer of choice for
investigated for SIM development. A
method development, particularly in late-
comparison of analyte retention in alternative
stage pharmaceutical development where
buffers to a commonly utilised generic
method transferability is very important and
method employing ammonium hydroxide was
The following organic additives and buffers
many, if not all, process impurities and
examined. Baseline characteristics (such as
were utilised for the high pH investigation;
prevalence of system peaks and effect on peak
ammonium formate, acetate and bicarbonate,
sensitivity) of the alternative buffers were noted.
pyrrolidine, 1-methylpiperidine, triethylamine
degradents are known [e.g.
12,13
]. The downside
of utilising phosphate buffers are obviously
incompatibility with MS detection, they are
more aggressive on silica based columns
[14]
and also have lower solubility in organic
Experimental
Samples & Reagents
MilliQ Gradient A10 system producing water
of 18.2 [Link] and < 3 ppb total organic
carbon quality.
acetate and the Goods buffers (BioXtra grade)
- CAPS, CAPSO and glycine which were all
obtained from either Fluka Analytical (puriss
solvents meaning that gradients to high
33 compounds were utilised in this study
LC-MS grade) or Sigma (SigmaUltra grade)
organic fractions are not possible (typically
which were families of proprietary Pfizer
except pyrrolidine (Alfa-Aeser). These
<80% (v/v) solvent depending on organic solvent
molecules including the active pharmaceutical
additives were compared to a mobile phase of
and salt type, salt concentration and pH).
ingredients (API), synthetic precursors and
0.1% ammonium hydroxide (unadjusted pH =
process related impurities (PRIs) covering a
10.6). The additives for comparison were
range of acid, basic, neutral and zwitterionic
prepared at 10 mM (solids) or 0.1% (liquids)
character. These were typically combined in
concentration and adjusted to a pH of 10.6
groups of related compounds at 0.1 mg/mL
with ammonium hydroxide (Fisher) or glacial
each in methanol - water (50:50, v/v).
acetic acid (Sigma Aldrich) as appropriate.
same pH was important in this work as pH is
Acetonitrile (Sigma-Aldrich Chromasolv) and
the dominant contributor to analyte retention
methanol (Sigma-Aldrich Chromasolv) were
for ionisable compounds under reversed-
used in this study. Formic acid (HCOOH
Instruments
UPLC experiments were performed on a
Waters Acquity UPLC (with PDA detection).
The system was controlled through Waters
Empower 2 software. Both organic mobile
phase additive and phosphate experiments
were performed on the same system with the
same column, thus mitigating any effects from
different system dwell volumes or column
differences and aging in the correlations. This
The primary focus of this investigation was to
identify if retention correlations existed
between organic buffers and their potassium
phosphate alternatives at the same pH,
solvent and column types. Analysis at the
phase conditions [15] and could lead to
Fluka Analytical puriss LC-MS grade
significant deviations in retention and
ampoules) and ammonium acetate
selectivity if not held constant. Correlations
(NH4OOCCH3 Fluka Analytical puriss HPLC
were assessed by comparing retention times
grade) were investigated in this study.
of the analytes under both sets of conditions
Additionally, alternatives to ammonium
and assessing any changes in
hydroxide (NH4OH Fluka Analytical >25% in
chromatographic selectivity.
water ampoules LC-MS grade) for high pH
017_CHROMTODAY_May_2011 Ferguson_Huet_Article:ChromatographyToday
24/5/11
14:35
Page 19
19
allows direct comparison of retention times in
this work (rather than a more formal use of
retention factors, k, to be used).
Methods
The generic methods listed in Table 2 were
used in this investigation. Columns were 100
x 2.1 mm i.d. (1.7 m) and utilised a flow rate
of 0.4 mL/min. Columns were thermostatted
at 30oC. 2 L of test mixes (typically 0.1
mg/mL of each analyte) were injected, as well
as individual stock solutions for peak tracking.
While PDA detection was used, data was
reprocessed and compared at 210 nm as a
worst-case scenario for observing system
peaks and other baseline artefacts.
Potassium phosphate buffers have lower
solubility in organic solvent-water mixtures
than organic buffers. 10 mM dipotassium
hydrogen phosphate (K2HPO4 pH 6.8) was
found to be compatible to a maximum
methanol volume of 80% (v/v), while 10 mM
potassium dihydrogen phosphate (KH2PO4
pH 2.6) was soluble up to 85% (v/v) acetonitrile
(both at room temperature). The same
gradient was mimicked in both sets of
experiments i.e. the same gradient slope was
used in the phosphate experiments as the
organic buffer experiments. However, when
the maximum organic solvent level was
reached in the phosphate gradients, the
mobile phase was held isocratically for the
remainder of the analysis.
Figure 1. A comparison of UV spectra for (a) 10 mM potassium dihydrogen phosphate and 0.1% formic acid (pH 2.6), and (b) 10
mM dipotassium hydrogen phosphate and 10 mM ammonium acetate (pH 6.8). In both cases, the phosphate spectra (dashed
line) exhibits much lower absorbance from wavelengths above 225 250 nm.
Results and Discussion
UV spectra
A comparison of the UV spectra for potassium
dihydrogen phosphate and formic acid (pH
2.6), and dipotassium hydrogen phosphate
and ammonium acetate (pH 6.8) are shown in
Figure 1. Quite clearly the spectra for
equivalent pH phosphate buffer exhibits much
lower absorbance than the respective organic
buffers. Significant absorbance is observed
for formic acid from 240 nm and below, while
ammonium acetate shows high levels of
absorbance from 225 nm and below. The
obvious result of this is significantly better
signal to noise and baseline characteristics
when using phosphate allowing lower levels of
analyte quantitation. For example, up to 2530% increase in analyte response in the
dipotassium phosphate buffer was observed
compared to the ammonium acetate mobile
phase. This exemplifies one of the reasons
why phosphate is often the preferred buffer
for chromatographic method development.
Comparisons of separations using the
different buffers are shown in Figures 2 and 3.
Retention
The correlation coefficients for the test
analytes retention and applying a linear
relationship between the organic and
Figure 2. Overlay of one set of test analytes using method 1 with 0.1% formic acid (pH 2.6 black line) and 10 mM potassium
phosphate (pH 2.6 blue line). The sloping nature of the formic acid baseline can make quantitation difficult and the increase
in analyte response with the phosphate buffer is obvious. The signals were both collected at 210 nm and the same sample
injected with both mobile phases.
Figure 3. Overlay of one set of test analytes using method 2 with 10 mM ammonium acetate (pH 6.8 black line) and 10 mM
potassium phosphate (pH 6.8 blue line). This is an extreme case where retention (for one analyte) is significantly different
between the two methods (corresponding to extreme top-right data point in Figure 5).
017_CHROMTODAY_May_2011 Ferguson_Huet_Article:ChromatographyToday
20
24/5/11
14:35
Page 20
May/June 2011
phosphate buffers is shown in Table 3.
Methods
Organic additive
Correlation coefficients for the low and mid-
Phosphate
buffer
Correlation
coefficient (r )
Maximum
retention time
variation/min
pH data demonstrate excellent correlation
(see Figures 4 and 5 for the low and mid-pH
0.1% Formic
acid (pH 2.6)
10 mM
KH2PO4 (pH 2.6)
0.9492
-1.09 to +1.18
10 mM
Ammonium
acetate (pH 6.8)
10 mM
K2HPO4 (pH 6.8)
0.9826
-1.20 to +0.08
10 mM
Ammonium
hydroxide
(pH 10.6)
N/A (see text)
0.8290 (0.9837
see text)
N/A (see text)
plots respectively). However, the absolute
retention times between the two methods can
vary. In the largest variation observed (1.18
and 1.20 min under method 1 and 2
conditions respectively), this equates to
approximately an 11% deviation in retention
over the length of the whole analysis.
The lower pH formic acid buffer demonstrates
a larger maximum retention range variation
Table 3. Correlation coefficients for test analyte retention of MS compatible and phosphate methods.
than the pH 6.8 systems. This is most often
observed with early eluting polar or charged
analytes. This is due to significant ion-pairing
of basic analytes with the phosphate counterion leading to longer retention. For acidic
and neutral analytes (18 out of 33 compounds)
at pH 2.6 where no ion-pairing with phosphate
would be anticipated to occur, near perfect
correlation of retention was observed (r2 =
0.9997) supporting this conclusion.
In the low pH mobile phase comparison, only
two analytes (from all 33 analytes analysed)
were found to change elution order when
switching from formic acid to potassium
dihydrogen phosphate. At pH 6.8, no
changes in elution order were observed and
the observed retention correlation is superior
to that at low pH. As may be expected, ion
pairing effects for the basic analytes were
much less pronounced at this higher pH,
possibly due to less ionisation of the bases
under these conditions. Again, this reinforces
Figure 4. Comparison of analyte retention times for method 1 with 0.1% formic acid (pH 2.6) and 10 mM potassium phosphate
the premise that retention order and peak
(pH 2.6). Note the higher retention with phosphate buffer for early eluting analytes which is believed to be due to ion-pairing
selectivity is largely maintained when
effects. Full method details are provided in the Experimental section. The correlation coefficient was 0.9492.
switching mobile phases.
Peak shape
A qualitative assessment of peak shape was
undertaken as part of this study. Generally
peak shapes were consistent when switching
between organic and phosphate buffers. On
occasion where different peak shapes were
noted, the organic additives were generally
found to give worse peak shapes than their
phosphate equivalent. At low pH, this might
be slightly surprising as the potassium
diphosphate buffer has lower ionic strength
than the formic acid mobile phase (Table 4)
and higher ionic strength mobile phases is
one factor that generally provides better peak
shapes [17,18]. As noted above, a likely factor for
the improved peak shape with the phosphate
buffer is phosphate ion-pairing with the basic
analytes. This prevents secondary interactions
with silanol groups on the stationary phase
and leads to improvements in peak shape.
At pH 6.8, peak shape for the dipotassium
Figure 5. Comparison of analyte retention times for method 2 with 10 mM ammonium acetate (pH 6.8) and 10 mM potassium
phosphate (pH 6.8). Full method details are provided in the Experimental section. The correlation coefficient was 0.9826.
017_CHROMTODAY_May_2011 Ferguson_Huet_Article:ChromatographyToday
24/5/11
14:35
Page 21
21
phosphate buffer was nearly always better
than the ammonium acetate (up to a 20%
Mobile phase
Ionic Strength / mol dm-3
0.1% HCOOH (unadjusted pH 2.6)
0.025
10 mM NH4OOCCH3 (unadjusted pH 6.8)
0.010
10 mM KH2PO4 (adjusted to pH 2.6
with H3PO4)
0.013
10 mM K2HPO4 (adjusted to pH 6.8
with H3PO4)
0.026
decrease in peak tailing was typically
observed). As reported in Table 3, the analyte
retention correlation at this pH is very good
and indicates a smaller phosphate ion-pairing
effect than at lower pH. The peak shape in
this case can be rationalised through the
significantly higher ionic strength of the
Table 4. Ionic strengths of aqueous portion of low and mid-pH mobile phases used in this study.
mobile phase compared to the ammonium
acetate mobile phase. In this instance it is
Concentration
Correlation coefficient (r2)
Ammonium hydroxide
0.1% (v/v)
N/A
cation (K+) competition for these sites.
Triethylamine acetate
0.1% (v/v)
0.8290 (0.9837)
Baseline noise & artefacts
In this investigation, it was found that shifting
to phosphate buffers not only increased analyte
signal-to-noise, but also led to a decrease in
baseline artefacts when comparing organic
and phosphate buffer mobile phases at both
low and mid-pH. An example of the improved
baseline with phosphate compared to
ammonium acetate at 210 nm is shown in
Figure 6. It should be reiterated that
phosphate buffer solubility is significantly
lower in organic solvents than alternative
organic additives. If the maximum organic
solvent levels in the gradient are breached,
this can lead to the phosphate salt crashingout in the instrument or column which in turn
may lead to increased baseline noise and
instrument back-pressure issues.
Pyrrolidine
0.1% (v/v)
0.5639
1-Methyl piperidine
0.1% (v/v)
0.7786
Ammonium acetate
10 mM
0.6272
Ammonium bicarbonate
10 mM
0.6860
Ammonium formate
10 mM
0.6685
CAPS
10 mM
0.8992
CAPSO
10 mM
0.8541
Glycine
10 mM
0.7466
believed that residual analyte-silanol
interactions are decreased by mobile phase
Additives at high pH
The use of phosphate buffers at high pH is
deleterious to silica based column stability
which at high buffering pHs (>11.3) leads to
rapid silica dissolution and collapse of the
phase [14,19]. For this reason, the majority of
mobile phases used in high pH
High pH additive
Table 5. Correlation coefficients for test analytes comparing retention in the specified mobile phase with the ammonium
hydroxide mobile phase. All solutions were titrated to pH 10.6 with ammonium hydroxide or acetic acid as required except
ammonium hydroxide which was naturally pH 10.6 in aqueous solution at this concentration.
chromatography utilise organic based buffers
very limited buffering capacity which can lead
which are less aggressive on the column. A
to separation reproducibility issues (the pH of
buffer commonly used for high pH work is
0.1% ammonium hydroxide is around 10.6.
ammonium hydroxide as it is both easy to
The pKa of the ammonium ion is 9.2).
prepare and MS compatible. However, in our
experience ammonium hydroxide tends to
degrade in aqueous solution quickly and this
can be observed as baseline artefacts in the
chromatogram. Additionally, at 0.1% (v/v)
concentration, ammonium hydroxide offers
One other point of note is that at high pH, oncolumn degradation of analyte molecules may
be observed. This is usually confirmed by
comparing a high pH separation with analysis
of the analyte at low pH. In the higher pH
separation, additional peaks may be observed
which are not present in the low pH
separation (e.g. identified by mass tracking
using MS detection) and care must be taken
not to misidentify these high pH impurities as
originating from the sample under analysis.
A number of high pH buffers were therefore
investigated as alternatives to ammonium
hydroxide (see Experimental section for full
details). The ammonium hydroxide mobile
phase was chosen as a baseline generic
method having been previously found to
provide a high degree of orthogonality [20] to
other generic methods employed in our
department. Note that not all of the buffers
investigated are MS compatible, but were
chosen as potential alternative mobile phase
additives that might exhibit improved
Figure 6. UV baselines of blank injections with the dipotassium phosphate buffer and ammonium acetate mobile phases for
sensitivity characteristics. The main criteria
method 2. Clearly the ammonium acetate baseline (black trace) exhibits a greater number of baseline artefacts than the
assessed was the ability to provide similar
phosphate buffer (blue trace) which can lead to peak quantitation issues when low level analyte related impurities are present.
analyte retention to ammonium hydroxide
017_CHROMTODAY_May_2011 Ferguson_Huet_Article:ChromatographyToday
22
24/5/11
14:36
Page 22
May/June 2011
while providing reduced baseline noise (and
hopefully additional buffering capacity to
increase separation robustness). The reason
for requiring a cleaner baseline is simply that
many of these impurities can co-elute with
real compound related impurities (which are
often present at similar levels) making
quantitation difficult.
High pH additive UV performance
A UV spectral overlay of two of the
comparator mobile phase additives (0.1% (v/v)
1-methyl piperidine and 0.1% (v/v) TEA) with
0.1% (v/v) ammonium hydroxide is shown in
Figure 7. The figure clearly shows that
ammonium hydroxide has a much reduced
absorbance at the lower end of the UV
spectrum compared to the other two
solutions. This was also observed for the
other high pH mobile phase additives and
illustrates that from a signal-to-noise and
sensitivity perspective, ammonium hydroxide
is a good choice as a mobile phase additive.
Additionally, with these solutions and all the
other mobile phase additives investigated,
none gave superior sensitivity and at best
only a similar baseline artefact count to
ammonium hydroxide.
High pH additive retention correlation
Correlation coefficients for the test analyte
retention in the ammonium hydroxide and the
other high pH mobile phases were calculated
and are listed in Table 5. Of the alternative
mobile phases, only triethylamine (as acetate
salt) gave similar retention to the baseline
ammonium hydroxide method. Two values for
this particular correlation are listed in Table 5.
The first value (0.8290) is for all 33 test
analytes. The second value is the correlation
from 32 analytes as one data point exhibited
much reduced retention (over 3 minutes less
retention) in the triethylamine acetate mobile
phase which significantly skewed the
correlation. No reason can be attributed for
this observation at this time. Significant
deviations from linearity were observed for all
other mobile phases. An example
comparison of one of the test analyte sets
with 10 mM triethylamine as acetate salt and
0.1% ammonium hydroxide is shown in Figure 8.
Figure 7. Comparison of some UV spectra for high pH buffers investigated in this study. Ammonium hydroxide is the additive
used as the reference mobile phase to which all other high pH systems were compared.
After triethylamine acetate, CAPS provided the
closest retention correlation to ammonium
hydroxide, but exhibited poorer analyte
sensitivity and peak shape. While the baseline
artefact count for CAPS was equivalent to
ammonium hydroxide, CAPS is not MS
compatible and is therefore a poorer choice for
SIM development. This was a similar finding for
Figure 8. Comparison of analyte response in 0.1% ammonium hydroxide (black line) with 0.1% triethylamine acetate (pH 10.6
all other high pH mobile phases tested.
blue line). While triethylamine acetate provided the most similar orthogonality to ammonium hydroxide and a slight
The findings of this study suggest that from a
the test analytes.
improvement in baseline artefacts, this mobile phase gave a much lower analyte response and generally poorer peak shape for
017_CHROMTODAY_May_2011 Ferguson_Huet_Article:ChromatographyToday
24/5/11
14:36
Page 23
23
sensitivity, peak shape, baseline artefact and
temperature or a combination of both to
retention perspective, none of the alternative
achieve the separation.
high pH mobile phases tested offered better
properties to ammonium hydroxide for MS
friendly SIM development work.
Conclusions
Excellent correlation of analyte retention when
transferring from organic mobile phase
additives to phosphate buffers at low and midpH was observed. In most cases absolute
elution order and resolution was maintained
while significantly increased UV sensitivity and
decreased baseline slope were observed with
phosphate buffers.
Data from this comparison suggests that many
SIMs (or other methods) could be developed
from an MS compatible method and
transferred directly to an appropriate
phosphate method for increased UV sensitivity
and method robustness. Alternatively, a
6.
alternative additive provided similar properties
majority of the alternative mobile phases.
Alternative approaches to clean ammonium
hydroxide mobile phases in order to decrease
the artefact level are therefore required e.g.
through use of appropriate mobile phase filters.
Acknowledgements
The authors would like to thank Adrian Davis
(Research Analytics, Pfizer Global Research &
Development, Sandwich) for useful discussions
regarding ionic strength determination.
References
1.
Van Gyseghem E., Van Hemelryck S., Daszykowski M.,
Questier F., Massart D. L., Vander Heyden Y. J.
transferred to an MS friendly alternative for
Chromatogr A (2003), 988, 77.
2.
Van Gyseghem, E., Deconinck, E., Dumarey, M.,
Jimidar, M., Sneyers, R., Redlich, D., Verhoeven, E.,
of confidence.
Peys, W., De Smet, M., Vander Heyden, Y. J.
In the few cases where a change in selectivity is
observed and resolution decreases dramatically,
7.
Chromatogr. Sci. (2008), 46, 793.
3.
Van Gyseghem, E.; Jimidar, M.; Sneyers, R.; Redlich,
Hewitt E.F., Lukulay P., Galushko S. J. Chromatogr. A
(2006), 1107, 79.
8.
McBrien M. Chromatography Today May/June
(2010), 30.
artefacts, poorer peak shapes and lower analyte
sensitivity were typically observed with the
I. Molnr I., Rieger H.-J., Monks K.E. J. Chromatogr. A
(2010), 1217, 3193.
to the ammonium hydroxide generic method.
Low retention correlation, increased baseline
Jupille T. H., Dolan J. W., Snyder L. R., Molnar I. J.
Chromatogr. A (2002), 948, 35.
In the high pH mobile phase investigation, no
phosphate method could be directly
identification of unknowns with a high degree
5.
9.
Apollonio L.G., Pianca D.J., Whittall I.R., Maher W.A.,
Kyd J.M. J. Chromatogr. B (2006), 836, 111.
10. Guo M.X., Wrisley L., Maygoo E Anal. Chim. Acta
(2006), 571, 12.
11. Shao Y., Alluri R., Mummert M., Koetter U., Lech S. J.
Pharm. & Biomed. Anal. (2004), 35, 625.
12. Mohammadi A., Rezanour N., Ansari Dogaheh M.,
Ghorbani Bidkorbeh F., Hashem M., Walker R.B. J.
Chromatogr. B (2007), 846, 215.
13. Grosa G., Del Grosso E., Russo R., Allegrone G., J.
Pharm. & Biomed. Anal. (2006), 41, 798.
14. Claessens H. A., van Straten M. A., Kirkland J.J. J.
Chromatogr. A (1996) 728, 259.
15. Neue, U.D; Phoebe, C.H.; Tran, K.; Cheng, Y-F.; Lu, Z.
J. Chromatogr. A (2001), 925, 49.
16. Li Y, Terfloth G.J., Kord A.S. American Pharmaceutical
Review Apr May (2009), 87.
17. Gritti F, Guiochon G., J. Chromatogr. A (2004),
1033, 43.
18. Gritti F, Guiochon G., J. Chromatogr. A (2009),
1216, 3175.
19. Kirkland J. J., Henderson J. W., DeStefano J. J., van
it is suggested that experimental modelling is
D.; Verhoeven, E.; Massart, D. L.; Vander Heyden, Y J.
Straten M. A., Claessens H. A. J. Chromatogr. A
investigated (e.g. utilising Drylab or ACDLabs
Chromatogr. A (2004), 1042, 69.
(1997), 762, 97.
LC Simulator) to re-optimise the separation.
This may involve varying gradient slope or
4.
Xiao K.P., Xiong Y., Liu F.Z., Rustum A.M. J.
Chromatogr. A, (2007), 1163, 145.
20. Slonecker P.J., Li X.D., Ridgeway T.H., Dorsey J.G.
Anal. Chem. (1996), 68, 682.