- Email: [email protected]
The use of ultra-high pressure liquid chromatography with tandem mass spectrometric detection in the analysis of agrochemical residues and mycotoxins in food – Challenges and applications
Accepted Manuscript Title: The Use of Ultra-High Pressure Liquid Chromatography with Tandem Mass Spectrometric Detection in the Analysis of Agrochemical Residues and Mycotoxins in Food–Challenges and Applications Authors: John O’Mahony, Lesa Clarke, Michelle Whelan, Richard O’Kennedy, Steven J. Lehotay, Martin Danaher PII: DOI: Reference:
S0021-9673(13)00047-2 doi:10.1016/j.chroma.2013.01.007 CHROMA 353950
To appear in:
Journal of Chromatography A
Received date: Revised date: Accepted date:
16-8-2012 19-12-2012 4-1-2013
Please cite this article as: J. O’Mahony, L. Clarke, M. Whelan, R. O’Kennedy, S.J. Lehotay, M. Danaher, The Use of Ultra-High Pressure Liquid Chromatography with Tandem Mass Spectrometric Detection in the Analysis of Agrochemical Residues and Mycotoxins in FoodndashChallenges and Applications, Journal of Chromatography A (2010), doi:10.1016/j.chroma.2013.01.007 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
*Highlights (for review)
Highlights
Ac ce p
te
d
M
an
us
cr
This review examines the role of UHPLC-MS/MS technology in agrochemical analysis. The emergence of UHPLC technology is discussed, and the implications for veterinary drug residue testing are documented. Current challenges associated with UHPLC-MS methods are reported on. An overview is given of multi-residue analytical methods, where UHPLC-MS technology has been successfully applied.
ip t
Page 1 of 47
1
The Use of Ultra-High Pressure Liquid Chromatography with Tandem Mass
2
Spectrometric Detection in the Analysis of Agrochemical Residues and Mycotoxins
3
in Food – Challenges and Applications
4 John O’Mahonya, Lesa Clarkea, b, Michelle Whelana, Richard O’Kennedyb, Steven
6
J. Lehotayc,Martin Danaher*, a
ip t
5 7 8
a
9
b
cr
Food Safety Department, Teagasc Food Research Centre, Ashtown, Dublin 15, Ireland
us
School of Biotechnology and National Centre for Sensor Research, Dublin City
10
University, Dublin 9, Ireland
11
c
12
Research Center, 600 East Mermaid Lane, Wyndmoor, PA 19038, USA
an
US Department of Agriculture, Agricultural Research Service, Eastern Regional
M
13 14
*Corresponding author: Tel. +353 1 8059500; fax: +353 1 8059550
16
Email address: [email protected]
ed
15
pt
17 18 Abstract
20
In the field of food contaminant analysis, the most significant development of recent
21
years has been the integration of ultra-high pressure liquid chromatography (UHPLC),
22
coupled to tandem quadrupole mass spectrometry (MS/MS), into analytical applications.
23
In this review, we describe the emergence of UHPLC through technological advances.
24
The implications of this new chromatographic technology for MS detection are discussed,
25
as well as some of the remaining challenges in exploiting it for chemical residue
26
applications. Finally, a comprehensive overview of published applications of UHPLC-
27
MS in food contaminant analysis is presented, with a particular focus on veterinary drug
28
residues.
Ac ce
19
1
Page 2 of 47
29 30
Keywords: Mass spectrometry (MS); Ultra-high pressure liquid chromatography
31
(UHPLC); veterinary drug residue analysis; matrix effects; frictional heating
32
Ac ce
pt
ed
M
an
us
cr
ip t
33
2
Page 3 of 47
33
Contents
34 1. Introduction p.2
36
2. The introduction of UHPLC and impact on residue analysis p.5
37
3. Coupling UHPLC with MS/MS detection p.7
38
4. Challenges inherent in UHPLC-MS/MS techniques p.11
39
5. Recent applications of UHPLC-MS technology in the analysis of agrochemical
40
residues and mycotoxins p.18
41
6. Conclusion p. 29
43
References p.30
cr
Acknowledgements p. 30
us
42
an
44 45
ip t
35
1. Introduction
46
Within the past decade, the introduction of ultra-high pressure liquid chromatography
48
(UHPLC) and rapid-scan and sensitive mass spectrometery (MS instruments has resulted
49
in a seismic shift away from traditional chromatographic techniques in the field of food
50
contaminant analysis, towards multi-class, multi-residue methods with short injection
51
cycle times and minimal sample preparation. The competing pressures of improving
52
sample throughput, reducing analysis costs and complying with regulatory requirements
53
have led to the continued adoption of UHPLC-MS/MS as the technique of choice in the
54
analysis of veterinary drug residues and related substances. In the present review, a
55
comprehensive overview of published UHPLC-MS/MS food contaminant applications is
56
presented, with a particular focus on veterinary drug residues. However, as this is still a
57
relatively new development, challenges associated with the technology are still emerging,
58
and some of these will also be detailed and discussed.
ed
pt
Ac ce
59
M
47
60
Analytical strategies applied in this field have evolved, not only to respond to both
61
statutory requirements for the scope of analytes detected and the needed method detection
62
limits to meet enforcement action levels, such as those stipulated by the European
63
Commission in 2002/657/EC [1], but also in response to the adoption of MS as the most
3
Page 4 of 47
frequently used detection method. The improved selectivity offered by MS/MS methods
65
means that the analysis can still be conducted with a lesser degree of chromatographic
66
resolution between target analytes, or between matrix components and analytes. LC-
67
MS/MS also offers greater versatility than that of GC-ECD (gas chromatography –
68
electron capture detection) or GC-MS methods, where the analyte must be volatile, or
69
else put through a derivatisation process to permit volatilisation of the analyte. With the
70
specificity of detection greatly improved and sample preparation time reduced,
71
innovation in the field of chromatographic separation became an area of research focus.
72
Since 2003, more commercial UHPLC systems continue to become available (Table 1).
cr
ip t
64
74
us
73 Table 1
an
75 76
While the development of ever-more efficient monolithic silica HPLC columns has been
78
an area of extensive chromatographic research activity, the development and refinement
79
of UHPLC, employing columns packed with sub-2 m particles, has had the largest
80
impact on analytical separation science. The Van Deemter equation for describing band-
81
broadening mechanisms relates separation efficiency of a method to the linear velocity of
82
the mobile phase, which changes according to the size of the stationary phase carrier
83
particles. Thus, lower particle size permits an increase in separation efficiency, and
84
consequently, conventional HPLC separation time can be reduced, or resolution between
85
analytes increased. This has beneficial ramifications for the field of chemical residue
86
analysis, where frequently, analytical methods test for analytes that are similar in
87
chemical nature.
ed
pt
Ac ce
88
M
77
89
For example, the epimeric compounds dexamethasone and betamethasone (Fig. 1) are
90
licensed for therapeutic use against inflammation in bovine animals in the European
91
Union (although use as growth promoters is banned). Chromatographic resolution of
92
these compounds can be difficult to achieve, due to their structural similarity [1].
93
Separation of these compounds has been performed using a special column containing
94
Hypercarb™ porous graphite [2]. Applying UHPLC conditions to this separation,
4
Page 5 of 47
Deceuninck and co-workers [3] attained resolution of the dexamethasone/betamethasone
96
pair using an Acquity™ BEH C18 column (1.7 m, 2.1 x 100 mm), in a method validated
97
according to Commission Decision 2002/657/EC [1]. The same column chemistry was
98
employed by Touber et al. in their earlier work [4] on developing a multi-residue method
99
for corticosteroids, including dexamethasone and betamethasone. Chromatographic
100
resolution of dexamethasone and betamethasone is only one example of where the power
101
of UHPLC has been successfully applied to solve challenges in multi-residue analyses.
102
Elsewhere, it has been shown that UHPLC-MS/MS can be used to separate and identify
103
as many as 107 pesticide compounds in strawberries, using a run time of just 6.5 min [5].
104
Further examples of successful application of UHPLC-MS/MS in food contaminant
105
applications will be discussed in the review.
us
cr
ip t
95
107
an
106 Fig. 1
M
108
The number of published methods exploiting UHPLC-MS/MS continues to grow each
110
year as the technology matures (see Fig. 2). However, there remain outstanding
111
challenges to be addressed. The reduced sample preparation approach which has become
112
standard in analytical strategies (e.g. QuEChERS) poses questions as to what extent the
113
matrix effect influences the outcome of analyses, and whether generic clean-up methods
114
are necessarily compatible with MS detection. The application of UHPLC-type separation
115
can also lead to chromatographic peaks which have a width of 1 – 3 s. The analyst must
116
then examine whether the data acquisition rate is appropriate for compliance with quality
117
control requirements, especially in applications where analytes co-elute or elute within a
118
time segment where data is collected for several analytes. The application of high
119
pressure to separations that is the core of UHPLC technology can also result in the
120
creation of “frictional heating” effects, with possible ramifications for chromatographic
121
methods. This review will examine these aspects of UHPLC-MS/MS application,
122
followed by a detailed examination of how UHPLC-MS/MS has been exploited in
123
various food contaminant analyses, through an overview of work published in the field,
124
as well as work of general relevance which is of importance to method development
125
using this new technology. However, we first describe the emergence and development of
Ac ce
pt
ed
109
5
Page 6 of 47
126
UHPLC, after two decades of the predominance of HPLC in fields such as veterinary
127
drug residue analysis, environmental analysis and pesticide residue testing.
128 Fig. 2.
130 2. The introduction of UHPLC and its impact on residue analysis
cr
131
ip t
129
us
132
The utility of very high pressures (in combination with small particles) in improving
134
chromatographic efficiency, as well as reducing analysis time, was first explored by
135
Jorgenson and colleagues in the 1990’s [6], [7]. Soon after this work, the first commercial
136
UHPLC systems came on to the market in 2004, along with Agilent RRHT columns (1.8
137
m particle size) and Waters Acquity BEH columns (1.7 m). The successful uptake of
138
this technology was attributed to the benefits associated with the sub-2 m column
139
technology. In order to utilise sub-2 m columns, new pumping technology had to be
140
developed, which could pump mobile phase at pressures up to 1000 Bar and above. This
141
represented a major improvement in pumping technology, as until then LC systems were
142
rated up to approximately 400 Bar and separations were typically run on columns
143
containing 5 m particles at 170 Bar to prolong pump seal lifetimes. As indicated in
144
Table 1, several manufacturers have entered the UHPLC market, producing systems
145
which operate at ever greater pressures and thus providing ever greater chromatographic
146
power.
M
ed
pt
Ac ce
147
an
133
148
Many (UHP)LC-MS/MS applications to date have been in the area of food safety because
149
of the requirement to analyse a wide range of residues down to low ng/g levels. HPLC
150
coupled with UV or fluorescence detection continues to be used in some laboratories, but
151
most groups are progressing to (UHP)LC-MS/MS platforms. Although it is difficult to
152
measure the number of laboratories that are using UHPLC-MS/MS compared to LC-
153
MS/MS without a comprehensive survey of end-users, the FAPAS proficiency test (PT) 6
Page 7 of 47
reports issued by the Food and Environmental Research Agency in the United Kingdom
155
offer an indication of UHPLC-MS/MS implementation among food safety laboratories.
156
In the 2012 anticoccidial PT study 02183, it was shown that 48 laboratories participated.
157
Only two of the 38 laboratories that completed the questionnaire reported that they used
158
UHPLC [8]. Similarly, in the 2012 Nitrofurans PT study 02179, two out of 35
159
laboratories reported that they used UHPLC-MS/MS [9]. More interestingly, a PT study
160
was carried out on avermectins in corned beef (FAPAS Round 02175), where 17
161
laboratories used HPLC-fluorescence and 13 reported the use of LC-MS [10]. This can be
162
ascribed to the sensitivity offered by HPLC-Fluorescence towards avermectin
163
compounds, while analysis by MS poses challenges due to their propensity to form
164
sodium adducts. UHPLC-MS/MS was reported to be used in two laboratories in the
165
avermectins PT study. Cumulatively, these results offer some insight on the adoption of
166
UHPLC in residue laboratories, although reported usage of “LC” as the separation
167
technique is ambiguous, and may be either HPLC or UHPLC. In the future, it is expected
168
that more laboratories will move to UHPLC-MS/MS because the technique can be
169
applied to multiple analytes in shorter run times compared to traditional HPLC-UV or
170
fluorescence systems.
171
MS/MS, less complicated sample preparation procedures can be used. We anticipate
172
additional developments in the area of online SPE combined with UHPLC-MS methods
173
to further improve method reproducibility and speed of analysis. Some examples of such
174
applications have already been published [11,12]. The peak broadening inherent in such
175
methods may be addressed by using advanced chromatographic techniques such as the
176
solvent plug injection technique described by Gode et al.[13]. Moreover, the application
177
of UHPLC-MS/MS can reduce solvent waste by a factor of 10 and increase the sample
178
throughput by a factor of 5 [14]. Typically, the cost per analyte can be reduced by a factor
179
of 15 and can be further reduced by including additional analytes. In the future, it is
180
expected that UHPLC will be implemented in more residue control laboratories in order
181
to reduce cost and hazardous waste. The many advantages offered by UHPLC are thus
182
expected to result in widespread use of the technology in the all areas of chemical residue
183
analysis.
M
an
us
cr
ip t
154
Ac ce
pt
ed
In addition, because of the increased selectivity of UHPLC-
7
Page 8 of 47
184 185
3. Coupling UHPLC with MS/MS detection
ip t
186 UHPLC coupled with MS/MS is becoming an increasingly used instrumental set-up for
188
researchers due to the advantages such as increased sensitivity, selectivity, resolution and
189
throughput compared with conventional HPLC instruments. The sub-2 µm particles used
190
in UHPLC columns offer increased efficiency and peak capacity. Peaks can be as narrow
191
as 1-2 s wide and coupling with modern, fast scanning mass spectrometers allows for the
192
detection of hundreds of analytes in a short retention window, even in complicated
193
biological matrices. Modern MS/MS analysers are capable of full scan acquisition rates
194
of up to 10,000 m/z per s, dwell times of 1 ms and polarity switching in 30 ms or less
195
(Table 2). Although the focus of this review is the combination of UHPLC technology
196
with tandem mass spectrometers, it is important to note that high resolution mass
197
spectrometry (HRMS) continues to evolve as a powerful alternative to MS/MS detection.
198
A careful study by Kaufmann and colleagues, comparing LC-MS/MS and LC-HRMS for
199
the analysis of seven veterinary drugs in a variety of matrices, showed that once HRMS
200
resolution attains 50,000 full width at half maximum (FWHM), selectivity for the two
201
techniques is comparable [15]. MS/MS has been widely adopted in the food analysis and
202
pharmaceutical industries, due to advantages such as its robust tolerance of different
203
matrices [16], but HRMS retains distinct advantages such as the ability to retrospectively
204
analyse historical data for non-preselected compounds [17]. Kaufmann [15] also noted
205
that at very high resolution, HRMS could identify a matrix component which behaved
206
similarly to a banned nitroimidazole compound, which in MS/MS analysis produced a
207
false positive. HRMS detection may eventually be the dominant detection method in
208
pesticide testing in fruits and vegetables, mainly due to its potential ability to identify
209
unknown chemicals [18].
Ac ce
pt
ed
M
an
us
cr
187
210 211
Table 2.
212
8
Page 9 of 47
213
3.1. Acquisition rate and method sectoring
214 215 The acquisition rate in SIM (selected ion monitoring) and SRM (selected reaction
217
monitoring) modes is determined by the switching time and the dwell time per SIM or
218
SRM channel. Dwell time is the amount of time spent collecting a data point at a set
219
transition or peak before switching to the next value in the SIM or SRM method; this
220
dwell time is set by the analyst. If necessary, dwell time can be increased to collect fewer
221
data points. Conversely, if a paucity of data points is collected, dwell time may be
222
decreased to improve peak definition. Dwell time represents a compromise between
223
signal-to-noise (S/N) ratio and the number of data points acquired across a
224
chromatographic peak.
225
sensitivity, as less time is required to acquire adequate mass spectral data. Background
226
noise is decreased as a result of the increased peak height of the analyte in UHPLC, and
227
its increased chromatographic resolving power may offer a reduction in ion suppression,
228
also leading to an increase in the S/N ratio compared with HPLC.
cr
us
M
an
Generally, decreased acquisition time results in increased
ed
229
ip t
216
Time-sectoring is another important factor when setting up UHPLC-MS/MS methods. If
231
the S/N ratio is poor for an analyte, it will require an increase in dwell time to increase
232
the S/N ratio. However, increasing dwell time may result in the detection of too few data
233
points across the peak, due to insufficient acquisition or cycle time. If this is the case, the
234
number of transitions within a SIM or SRM should be reduced to increase the cycle time
235
and obtain more points across the chromatographic peak. Alternatively, if the analytes are
236
sufficiently separated, the SIM or SRM channel can be time-sectored to increase the
237
cycle time for that analyte and acquire sufficient number of points across that
238
chromatographic peak. Time-sectoring also places less pressure on the instrument and
239
allows more time for the raw data to be processed [19].
Ac ce
pt
230
240 241
As well as sectoring the method and setting adequate dwell times, the inter-channel and
242
inter-scan delays are important. Inter-scan delay must be set to leave enough time for the
243
instrument electronics to switch from one transition to the next within an SIM or SRM
9
Page 10 of 47
and inter-channel delay to leave enough time between successive SIM or SRM channels.
245
If the method contains analytes that need different ionisation polarities, extra time is
246
required for the electronics to switch between one channel and the next, as well as change
247
polarities. Typically with newer MS/MS instruments, inter-scan and inter-channel delays
248
are set to 5 ms when switching between successive channels and 20 ms is required for
249
inter-channel delay when switching from positive ionisation to negative ionisation mode
250
between successive channels.
ip t
244
252
cr
251 3.2 UHPLC-MS/MS vs. HPLC-MS/MS
us
253
UHPLC-MS/MS yields many advantages over HPLC and HPLC-MS/MS. It is more
255
environmentally friendly, as it reduces the volume of chemicals used and waste produced.
256
It also offers increased sensitivity, resolution, S/N ratio and shorter injection cycle times.
257
UHPLC-MS/MS is capable of detecting analytes at a lower limit of quantitation (LOQ)
258
due to the increased resolution of the system. Yu et al. demonstrated the advantages of
259
UHPLC-MS/MS approach over HPLC-MS/MS in a quantitative analytical method for
260
five drug compounds in rat plasma. A shorter run time and narrower peaks were achieved
261
with UHPLC, which lowered the LOQ and increased the sensitivity, as well as
262
considerably improving accuracy and precision compared with conventional HPLC [20].
263
The authors proposed that this was due to a narrow band of analyte entering the mass
264
spectrometer, resulting in increased concentration of analyte and in turn increased
265
ratios/N. The throughput of samples is dramatically increased for quantitative analysis
266
due to the short separation time of the UHPLC coupled with the faster acquisition
267
capability of MS/MS. Recent examples of high throughput methods include that of
268
Whelan et al., who developed a multi-residue method for the determination of 38
269
anthelmintic drug residues in milk (as well as 20 internal standards) using UHPLC-
270
MS/MS with rapid polarity switching; all analytes were eluted within 8 min. [21].
271
Ventura et al. detected 34 forbidden drugs and metabolites in a total run time of 5 min by
272
UHPLC-MS/MS [22]. A comparison of UHPLC-MS/MS and HPLC-MS/MS methods for
273
the determination of pesticide residues in baby foods demonstrated that the UHPLC
274
approach was 2.5 times faster than the corresponding HPLC method [23]. Aside from the
Ac ce
pt
ed
M
an
254
10
Page 11 of 47
clear benefit of shorter run times, it has also been noted that use of UHPLC in place of
276
HPLC can also improve method performance via reducing matrix effects, such as in the
277
study of pharmaceutical drugs in surface water described by Van de Steene and Lambert
278
[24]. Here, the reduction in observed matrix effect permitted the authors to replace
279
standard addition with a simpler internal standardisation approach for analyte
280
quantification. Other examples of methods that have been improved by use of UHPLC
281
have been well described in a recent review by Guillarme et al.[25].
cr
282
ip t
275
A major consideration in selecting UHPLC in place of conventional HPLC is that the
284
higher throughput obtained from using UHPLC coupled to MS/MS increases the
285
requirement to perform regular preventative maintenance on the instrument to ensure
286
continued optimal performance. This is especially important when working with more
287
complex biological samples; if the orifice of the MS/MS is not routinely cleaned,
288
accumulated residue on the front end could potentially enter the quadrupoles, which
289
would lead to reduction in sensitivity or the drop-off of less-intense analytes. A full
290
system clean would then be required by engineers to regain sensitivity; this may not be
291
covered under service contracts and will also result in unnecessary down-time of the
292
instrument due to negligence. If the instrument is analysing a large number of samples
293
daily, the sample cone of the MS/MS should be cleaned daily at the end of the sequence,
294
although it should be noted that UHPLC methods typically involve lower injection
295
volumes than HPLC methods.
an
M
ed
pt
Ac ce
296
us
283
297
However, there are some disadvantages to using UHPLC-MS/MS over HPLC-MS/MS.
298
For one, a significant disadvantage is the high cost of the sub-2 µm columns. Also, there
299
is a much smaller selection of UHPLC column chemistries available - although the range
300
is steadily increasing each year, this could prevent a straight transfer from HPLC to
301
UHPLC-MS/MS for some methods.
302
volumes inherent in UHPLC systems have practical implications. For example, samples
303
must be filtered through 0.2 m filters prior to use, as solid material may easily
304
precipitate out of solution at column frits, especially when a steep gradient is employed.
305
Mobile phases may also require filtration, to exclude particulate matter, although care
Furthermore, the narrow plumbing and small
11
Page 12 of 47
must be taken to avoid introducing additional contamination from filtration glassware,
307
filters, etc. Although the instrumentation can be capital-intensive initially, increased
308
market competition is reducing prices, while gains in productivity can offset high
309
instrument costs [26]. However, the benefits far outweigh the disadvantages, e.g. the
310
columns may cost more than traditional HPLC columns but the increased throughput
311
obtained counteracts the increased column cost.
312 4. Challenges inherent in UHPLC-MS/MS techniques
us
313
an
314 315
cr
ip t
306
4.1 Frictional Heating Effects
M
316
As observed by Churchwell et al. [27], UHPLC offers improvements in both the duration
318
of a separation and in sensitivity – although this is compound dependent and not
319
necessarily straightforward. One of the issues which may present difficulty in transferring
320
a method from HPLC to UHPLC is frictional heating [28]. This is a physical
321
phenomenon whereby the interaction between the mobile phase and the stationary phase
322
creates heat. This occurs with liquid chromatography in general, but is far more
323
pronounced with UHPLC. Potentially, this phenomenon could mean molecules travelling
324
in a band through the column experience greater heat at the centre of the column than at
325
the walls, with the net result being band broadening (fig. 3).
327
pt
Ac ce
326
ed
317
Fig. 3.
328
12
Page 13 of 47
The effect of the heat can also have an impact on method performance – de Villiers and
330
co-workers determined that when a radial temperature gradient is prevalent within a
331
column, the efficiency may be seriously affected, particularly with (relatively) larger
332
columns (internal diameter (ID) = 2.1 mm). However, in their work, it was observed that
333
use of a still-air column heater meant that a longitudinal temperature gradient was
334
prevalent along the length of the column, with no considerable impact on separation
335
efficiency. For routine use, the implication is that pre-heating of mobile phase and use of
336
a column oven are essential in controlling heat flow. Work by Nováková and colleagues
337
[29], using a selection of pharmaceutically relevant molecules, investigated the
338
importance of the frictional heating effect for method selectivity. Using a column of ID =
339
2.1 mm, a pressure drop of 1000 bar over the column resulted in an observed temperature
340
increase of 16°C. Changes in retention time were not considered significant however,
341
with changes of 5-10% observed for most of the compounds analysed in the range 30°C –
342
60°C. The effect was more pronounced for acidic compounds, with changes of 25% for
343
ascorbic acid and 41% gallic acid. The authors commented that the effects of frictional
344
heating may be mitigated through several means, including use of narrow-bore columns;
345
use of several shorter columns in place of one long column; a decrease in column oven
346
temperature in UHPLC when transferred from HPLC; or modification to the method
347
gradient or mobile phase pH to compensate for changes in resolution between HPLC and
348
UHPLC. In a study comparing the efficiency of sub-2 micron porous and sub-3 micron
349
shell particles in reverse phase separations, McCalley noted that when using 100%
350
acetonitrile with sub-2 micron columns, the detrimental effects of frictional heating are
351
enhanced [30]. However, aqueous content in the mobile phase was found to moderate
352
this effect, due to greater thermal conductivity. For larger (3 – 5 micron) particles, the
353
effects of frictional heating were also significant [31]. The authors observed that the
354
impact of heating effects were proportional to the capacity factor of a compound and the
355
column inner diameter, but inversely proportional to column particle size and the set
356
column oven temperature.
Ac ce
pt
ed
M
an
us
cr
ip t
329
357 358
4.2 Band Broadening Effects 13
Page 14 of 47
359 The use of UHPLC-MS/MS technology has afforded, among other advantages, a
361
particular benefit when it comes to sample injection volumes. The greater sensitivity
362
achievable with MS/MS combined with the chromatographic power of UHPLC
363
technology means that the analyst may use smaller injection volumes than was previously
364
required, to attain acceptable detection limits. The small diameter of the columns
365
typically employed in UHPLC means that a small injection volume is required, to avoid
366
possible column overload. Smaller quantities of matrix are thus loaded on to the column
367
than with standard HPLC, meaning matrix suppression effects are minimised. However,
368
to ensure good chromatographic performance, the system must contain the absolute
369
minimum amount of dead volume to minimise extra-column broadening effects. As
370
system volume is low, extra-column volume can represent a large percentage of overall
371
system volume. Consequently, extra-column volume becomes a significant source of
372
band broadening. As noted by McCalley [32], the influence of dead volume has an
373
inversely proportional relationship to column internal diameter. Dead volume effects
374
have been observed to be more detrimental to performance in narrow bore (2.1 mm)
375
columns employed with a UHPLC system than with a standard bore (4.6 mm) column
376
operated on a typical HPLC system [30].
cr
us
an
M
ed
pt Ac ce
377
ip t
360
378
As well as tubing diameter, connectors for tubing must be carefully selected so as to
379
avoid contributing extra-column volume. The narrow peaks found in UHPLC analyses
380
mean that band broadening can have a large effect on resolution and overall performance.
381
Gritti and Guichon [33] have observed that for some modern high-pressure LC systems,
382
the largest contributing factor to band broadening was the inner diameter of the needle
383
seat capillary tube (along with detector cell volume). One possible path to reducing extra-
384
column band broadening due to sample injection is described in detail by Sanchez et al.
385
[34]. Their use of a “performance optimising injection sequence” (POISe) allowed the
386
reduction or elimination of the effect of sample injection on the chromatographic
387
performance. This method involves the injection of a defined amount of a weak solvent
14
Page 15 of 47
(relative to the mobile phase) along with the sample. This has the effect of minimising the
389
extra-column dispersion taking place ahead of the analytical column, while having the
390
advantage of not requiring any physical modification to the chromatographic system.
391
Analyte bands are compressed at the top of the analytical column, in proportion to their
392
retention factor. This is referred to as isocratic focussing. It was found that band
393
compression could be achieved in both isocratic and gradient elution modes.
ip t
388
4.3 Narrow Analyte Peaks
us
395
cr
394
an
396
The very short dwell times possible with modern MS instruments (as low as 1 ms, as
398
indicated in Table 2) mean that many analytes can be rapidly detected, e.g. 55 anabolic
399
and androgenic steroids in a 5-min injection cycle [35]. Software-based optimisation of
400
dwell times, cycle times and multiple reaction monitoring windows (dynamic multiple
401
reaction monitoring) has permitted the analysis of up to 11 pesticides which elute within
402
a one-min time window [36]. However, as well as possible peak overlap,
403
chromatographic systems may also produce peaks of very narrow width, particularly in
404
UHPLC [37]. Improvements in data-handling capabilities permit the collection of
405
detector data from compounds where the UHPLC peak profile is only about 2 s across. A
406
multi-residue qualitative MS/MS method for the identification of drugs of abuse in urine
407
by Maquille and co-workers [38] clearly indicates the challenge this can pose. The
408
application of a small-particle (1.7 m) column and high pressures caused all compounds
409
to elute within approximately 2 min. However, as all compounds exhibited narrow peaks
410
(a few seconds), the authors employed rapid data collection parameters – selected
411
reaction monitoring dwell-time was set at 5 ms, while inter-channel delay was also set to
412
5 ms. Time-segmenting of the compounds according to their retention time was also
413
essential to achieving adequate numbers of points across peaks. Although not defined in
414
the relevant legislation, a general reference for peak definition is that a minimum of 5-6
415
data points should be achieved per peak, to ensure reliable integration [39]. Martínez
Ac ce
pt
ed
M
397
15
Page 16 of 47
Vidal [40] et al. exploited an alternative means of data collection in their work on
417
determination of veterinary drug residues in milk, which permits rapid screening of
418
samples. A MS/MS screening method was designed based on the selection of a neutral
419
loss or product ion which is selective for a family of compounds. If a positive sample is
420
detected, the sample can then be retested using a confirmatory (MRM) method. The run
421
time for this particular screening method was 3 min, and 21 veterinary drug residues were
422
included in its scope.
cr
ip t
416
us
423
Kovalczuk et al. have presented a comprehensive study of the chromatographic output of
425
a UHPLC-MS/MS multi-residue analysis [41]. This represents one of the few examples
426
of multi-residue methods where the authors have assessed chromatographic parameters
427
such as peak capacity, peak symmetry and height of theoretical plates. Direct comparison
428
of HPLC and UHPLC analysis (with MS/MS detection) of a multi-pesticide mixture
429
highlighted some of the key chromatographic considerations when selecting a UHPLC
430
approach over a HPLC approach. Peak compression on the UHPLC system meant that
431
S/N at low concentrations improved, with a consequent reduction in the method limit of
432
quantification (LOQ) versus the HPLC method. For the compound tebuconazole (the
433
matrix analysed was apple), S/N improved by a factor of 10 going from HPLC to
434
UHPLC, while for several compounds, the LOQ was as much as four times lower using
435
UHPLC. Peak capacity was comparable in both techniques under optimised conditions,
436
although it is noted that UHPLC offers shorter analysis time (and reduced solvent
437
consumption). Height of theoretical plates was greater in UHPLC, but with significant
438
variability.
440
M
ed
pt
Ac ce
439
an
424
4.4. Matrix effects
441 442
Matrix effects are a common phenomenon when using MS for chromatographic
443
detection, as well as for screening methods. These effects can result from a number of 16
Page 17 of 47
factors, generally the presence or co-elution of interfering compounds, resulting in
445
enhancement or suppression of the analyte of interest. Interfering analytes or compounds
446
can be extracted from the original sample matrix or from extraction equipment such as
447
plastic tubing. The presence of interfering analytes along with the analyte of interest in
448
the ionisation source increases the competition for ionisation and typically reduces ion
449
efficiency of the analyte. The effect of ion suppression in residue analysis applications
450
has been investigated in detail by Antignac and colleagues [42]. Using a method for the
451
analysis of the beta-agonist isoproterenol in bovine meat samples, it was illustrated how a
452
polar analyte that eluted early in a typical reverse-phase separation may potentially co-
453
elute with matrix interferences which come with the solvent front. The signal for this
454
compound was shown to reduce by as much as 95% from the anticipated response due to
455
matrix interference, indicating how severe the effects may be. The choice of ionisation
456
technique can be a major factor in eventual matrix effects in an analysis – it has been
457
observed that atmospheric pressure chemical ionisation (APCI) is, in some cases, less
458
prone to matrix effects than electrospray ionisation (ESI) [43], although ESI is more
459
widely applicable. This has been attributed to the differences in ionisation techniques, in
460
that using ESI, the charged analyte is formed in the liquid phase, while with APCI, the
461
analyte is transferred into the gas phase in the neutral form before ionisation occurs [44].
462
Several possible mechanisms that lead to matrix effects in ESI are described by Trufelli
463
et al. [44]. High analyte concentration (>10-5 M) can lead to a loss of sensitivity, possibly
464
due to saturation of the electrospray ionisation droplet surface with analyte. Other
465
possible causes include the formation of analyte-inclusion particles from non-volatile
466
additives and analytes of interest, with resultant hindrance of ionisation. It has also been
467
postulated that the presence of non-volatile interferents affects the viscosity and surface
468
tension of ESI droplets, which thus affects ionisation efficiency and the eventual signal
469
recorded [45]. Hydrophilic analytes may also be significantly suppressed in the presence
470
of surfactants in the sample matrix. The presence of lipophilic components in the sample
471
matrix is another factor which can have a large impact on method response, as observed
472
in a method for the analysis of veterinary drug residues in meat-based baby food [46].
473
Another known issue in the LC-MS/MS analysis of samples of biological origin is the
474
effect of phospholipids on ionisation efficiency [47]; these compounds can cause
Ac ce
pt
ed
M
an
us
cr
ip t
444
17
Page 18 of 47
significant matrix ionization effects in positive electrospray ionisation mode.
476
Phospholipids represent a main constituent of cell membranes. These compounds can
477
present very broad elution profiles, requiring the analyst to either monitor these
478
compounds chromatographically to ensure they do not co-elute with analytes [48], or
479
employ advanced sample preparation techniques to remove the interfering compounds
480
[49].
ip t
475
cr
481
Highly selective sample preparation techniques are important for removing compounds
483
from the matrix that produce matrix effects – although more elaborate sample preparation
484
procedures do have a bearing on sample throughput, and are more costly. Chambers et
485
al. carried out a systematic and comprehensive strategy for reducing matrix effects in LC-
486
MS/MS as applied in bioanalytical assays. A paired t-test illustrated a statistically
487
significant improvement in matrix effects using UHPLC when compared with traditional
488
HPLC [50]. The evolution of UHPLC-MS/MS, which has permitted the analysis of
489
multiple drug residues from multiple classes, as well as the development of fast generic
490
clean-up and extraction techniques, means that highly selective sample preparation
491
techniques are no longer critical. A consequence of employing these generic methods is
492
that matrix effects can have a large bearing on method performance. This is particularly
493
the case when using solvent standard calibration curves, and where possible, matrix
494
matched calibration curves offer a more accurate means of quantification. Alternatively,
495
stable isotopically labelled standards can be employed to negate matrix effects to improve
496
both accuracy and precision. This is demonstrated in work by Varga et al. [51], on the
497
determination of various mycotoxins in maize, where the use of
498
analogues permitted the development of the first multi-target method capable of
499
determining all mycotoxins regulated in the EU for maize.
an
M
ed
pt
Ac ce
500
us
482
13
C-labelled mycotoxin
501
The likelihood of matrix effects is high with the analysis of samples of biological origin,
502
such as food of animal origin in veterinary drug residue analysis. These effects may be
503
somewhat mitigated when using UHPLC - as mentioned in section 4.2, UHPLC methods
504
require smaller injection volume than comparable HPLC methods, to avoid overloading
18
Page 19 of 47
the column. A major advantage of this reduced injection size is that matrix effects are
506
reduced, as smaller quantities of matrix are injected onto the source. This, in turn, leads to
507
less competition for ionisation in the ESI droplets, increased resolution between the
508
analytes of interest, and a much cleaner mass spectrum compared to that of HPLC-
509
MS/MS. The smaller quantity of matrix passing through the system means less
510
accumulation of non-volatile components and other contaminants on the sample cone
511
ahead of the quadrupoles, as well as on the quadrupoles themselves. Such accumulations
512
can lead to charging effects with a resultant loss of sensitivity, so this is a further benefit
513
of UHPLC interfaced with MS/MS. Given the extreme impact on analyte signal which is
514
possible with matrix effects, quantification/ evaluation of matrix effects in a given
515
analysis can provide the analyst with essential detail in ensuring consistent method
516
performance. This may be assessed either through spiking of samples with analyte
517
following extraction, or through infusion of analyte post-column into the mass
518
spectrometer detector system, and examining the effect on signal when a negative control
519
sample is injected (see Fig. 4). The possibility of matrix interference with analyte signal
520
is one reason why validation procedures (such as those indicated in 2002/657/EC, [1])
521
require that a method be validated for each compound in a multi-analyte method, rather
522
than selecting one representative compound.
ed
M
an
us
cr
ip t
505
pt
523
525 526 527
Ac ce
524
Fig. 4. Adapted from [42] with permission from Elsevier.
4.5 Column Blockage
528 529
Due to the narrow plumbing employed in UHPLC, a high level of chromatographic
530
hygiene is required to avoid problems with blockages. In addition, sample matrix and
19
Page 20 of 47
analytes can contribute to column blockage. In an application for detection of steroid
532
esters, columns were susceptible to blocking when solvent standards were injected onto
533
the column in a gradient that started with a high aqueous mobile phase, going to high
534
organic content [52]. It was proposed that the waxy ester substances crashed out of
535
solution, when mixed with the mobile phase during the injection cycle. The solution to
536
the problem was simple; by elevating column temperature to 60°C, and including
537
isopropyl alcohol in mobile phase and needle washes, the problem was circumvented. In
538
the authors’ laboratory, a number of methods have been developed using high
539
temperature [14,21,53,54]. However, increasing temperature may not be suitable for all
540
substances, e.g. tetracycline antibiotics have to be separated at 40°C or less, as they can
541
form epimers [55].
an
us
cr
ip t
531
542
5. Recent applications of UHPLC-MS technology in the analysis of agrochemical
544
residues and mycotoxins.
ed
545
M
543
As noted in the introduction, there has been significant uptake of UHPLC-MS/MS
547
technology in areas of analytical research such as pesticide testing, veterinary drug
548
residue analysis and environmental contaminant detection, and Fig. 2 clearly indicates
549
how widely adopted the technique has become. In this section, an overview of some of
550
these published applications is given. Methods have been published for many classes of
551
compound; these may be broadly grouped into substances with an established permitted
552
limit [e.g. maximum residue limits (MRLs)], and substances which are prohibited. In
553
veterinary drug residue testing, Council Directive 96/23/EC [56] lists prohibited
554
compounds as Group A substances, covering stilbenes, antithyroid agents, steroids,
555
resorcylic acid lactones, beta-agonists and banned veterinary drugs. Other substances
556
with a permitted limit are found in Group B substances (anthelmintics, anticoccidials,
557
nitroimidazoles, carbamates, pyrethroids, sedatives, non-steroidal anti-inflammatory
558
drugs, sulphonamides, quinolones and others). Example UHPLC-MS/MS methods for
Ac ce
pt
546
20
Page 21 of 47
different veterinary drug classes are summarised in table 3. For pesticides, there are 1289
560
active substances that are controlled under EU regulation 1107/2009 [57]. A total of 511
561
substances have been given MRLs under EU regulation 396/2005 [58,59], while for
562
pesticides without an established MRL, a default limit of 0.01 mg kg-1 is set [60]. The
563
sheer diversity of pesticides in use underlines the need for a robust, versatile analytical
564
approach, as offered by UHPLC-MS techniques. Examples of such methods are described
565
in the following sections.
cr
ip t
559
us
566 567
pt
ed
M
an
Table 3
Ac ce
568
21
Page 22 of 47
ip t cr
5.1 UHPLC-MS/MS methods for Prohibited Substances
572
Analysis for banned compounds represents an area of continued research focus, as these are typically chemical entities which pose a direct
573
risk of harm to human health. UHPLC-MS/MS presents a very attractive option for the analysis of complex biological samples for residual
574
quantities of zero-tolerance substances. This can be attributed to the inherent resolving power, permitting resolution of chemically similar
575
compounds, and sensitivity. This is crucial in areas such as environmental analysis, where the ability to specifically detect small polar
576
organic molecules within complex sample matrices poses a particular challenge [61]. In the area of veterinary drug analysis, UHPLC-
577
MS/MS analytical power has been applied to most banned substances including thyreostats [62], as well as stilbenes and resorcylic acid
578
lactones [63]. The bulk of published work, however, has focussed on two main sub-categories of Group A substances, namely the steroids
579
and -agonists.
M
d
ep te
580
an
571
us
570
Steroid hormones are banned from use in food-producing animals, and fall into three distinct groups: estrogens, gestagens and androgens
582
(EGAs) [64]. Exploitation of these compounds for the purposes of growth promotion in food-producing animals is strictly prohibited within
583
the European Union by Directive 96/22/EC [65]. Application of UHPLC-MS/MS to determine the presence of these compounds in bovine
584
hair has been described by Duffy et al. [52], who found that the method performed comparably to the more conventionally-employed GC-
585
MS/MS methods. In their method for the detection of boldenone (an androgen) and metabolites in calf urine, Van Poucke et al. [66] used a
586
UHPLC-MS/MS, with a short column (50mm) containing particles of 1.7 m diameter. Gradient optimisation permitted the polar analyte
587
6-hydroxy--boldenone to be separated from the solvent front. The authors noted that the limiting factor in speeding up their method
Ac c
581
22
Page 23 of 47
beyond the 5-min injection cycle achieved was the number of data points being acquired
589
for peaks. Covering a broader range of compounds, Malone and colleagues analysed for
590
13 synthetic growth promoters in bovine muscle, applying a 1.8 m particle-size column
591
(2.1 x 50 mm) and achieving separation of all compounds in 14 min [67]. Wang et al.
592
described a modified QuEChERS (Quick, Easy, Cheap, Effective, Rugged, Safe)
593
extraction procedure to extract nine estrogens from milk powder. Two UHPLC C18
594
columns were evaluated, the first was an ethylene bridged hybrid (BEH) column, and the
595
second was a HSS T3 column. Results from this study showed that the BEH-C18 column
596
was not suitable as it was not capable of separating all nine estrogens [68], but the HSS
597
T3 column was successfully employed, using an appropriate gradient.
598
extracted 10 anabolic steroids (AASs) from equine plasma using methyl tert-butyl ether
599
(MTBE) prior to separation on a C18 reversed-phase column [69]. This extraction
600
procedure was again employed by Guan et al., who developed a method to isolate 55
601
anabolic and androgenic steroids in equine plasma. Separation was performed using a
602
Hypersil Gold™ C18 column. Amongst the 55 AASs were epimer pairs (testosterone and
603
epitestosterone, boldenone and epiboldenone, nanadrolone and epinandrolone) which
604
were chromatographically base-line separated. This high-throughput method had a cycle
605
time of 5 min per sample [35]. Finally, Wong and colleagues [70] described a UHPLC-
606
MS/MS method for the analysis of 140 compounds in equine urine, with an injection
607
cycle time of 13 min – including anabolic steroids. The authors noted how a combination
608
of fast chromatography and rapid polarity switching in MS detection have allowed them
609
to condense what was formerly four individual test methods into one comprehensive
610
method.
cr
us
pt
ed
M
an
You et al.
Ac ce
611
ip t
588
612
Also falling into the category of banned (Group A) substances, β-agonists are synthetic
613
derivatives of catecholamines, used at lower quantities as tocolytics and bronchodilators
614
in veterinary medicine [71]. However, for the last 20 years there has been frequent
615
reported misuse as growth-promoting agents in food producing animals [72]. These
616
compounds may be divided into two groups: the substituted anilines, including
617
clenbuterol, and the substituted phenols, which include salbutamol [64]. Yang et al. 23
Page 24 of 47
developed a method capable of detecting 13 β-agonists in milk. Ion suppression was
619
overcome by adding perchloric acid during the sample extraction step to precipitate
620
protein out of the samples [72]. Separation was carried out using an Acquity™ BEH C18
621
column. Shao et al. isolated 16 β-agonists from porcine tissue using enzymatic
622
hydrolysis. Maximum sensitivity in this method was achieved by optimising the solvent
623
type and pH of the mobile phase. Methanol-water containing 0.1% formic acid provided
624
good separation and resolution when compared to a comparable acetonitrile-water mobile
625
phase [73]. A similar method was used to determine 19 -blockers and 11 sedatives in
626
animal tissue, with an injection cycle time of 14 min [74]. Nielen et al. developed a
627
method for 22 β-agonists, which was sufficiently versatile to be applied to the analysis of
628
bovine and porcine urine, feed and hair using a mixed-mode solid-phase extraction
629
procedure. The method was performed by conventional LC-MS/MS and UHPLC-
630
MS/MS. Due to the diverse structure of β-agonists, developing a single analytical method
631
which encompasses multiple compounds is challenging. This issue was overcome by
632
performing two injections of each sample using the same LC gradient whilst acquiring a
633
different set of ion transitions. Run time was reduced from 20.5 min to less than 15 min
634
using UHPLC-MS/MS.
635
ratio improved with the introduction of a high-resolution column [75].
638
cr
us
an
M
ed
pt
637
Chromatographic resolution, peak shape and signal-to-noise
5.2 UHPLC-MS/MS methods for substances with an established permitted limit
Ac ce
636
ip t
618
639
The presence of a MRL, ‘tolerance’ or recommended limit (RL) for a compound interest
640
has somewhat different implications – methods have a well-defined target concentration
641
for analysis, rather than the criterion of ‘as low as possible’ as with banned substances.
642
An example of a class of veterinary drug which possess MRLs is the quinolone group.
643
These are potent antibacterials used in the therapeutic treatment of infections in both
644
human and veterinary medicine [76]. MRLs to monitor quinolone residues in edible
645
animal tissue are controlled by European Union Directive 2377/90/EC [77]. Lombardo-
24
Page 25 of 47
Agüí et al. described the determination of eight quinolones in bee products. The
647
separation capabilities of three different C18 columns were compared, where the particle
648
size of each varied from 1.7 μm to 1.9 μm. An intermediate partial size of 1.8 µm
649
delivered slightly better separation of two of the quinolones [78]. Zhang et al. extracted
650
22 fluoroquinolones for milk using an EDTA-McIlvaine buffer solution and purified
651
using Bond Elut™ Plexa SPE cartridges. Zhang examined the separation capabilities of
652
four different reversed-phase chromatographic columns which varied in particle size and
653
bonded phase chemistry.
654
observed with a Waters Acquity™ Shield RP18 Column (2.1 mm x 100 mm, 1.7 µm).
655
Recoveries of 61.9 – 94.6 % were observed for 19 of the 22 fluoroquinolones included in
656
the study [76].
ip t
646
an
us
cr
The highest sensitivity and acceptable peak shaped was
657
Sulphonamides represent a class of drugs used to prevent and treat bacterial infections.
659
They are amongst the most widely used veterinary drugs because of their broad-spectrum
660
antimicrobial activity, low cost, and their effectiveness as growth promoters in livestock
661
[79]. The MRL for individual sulphonamides in food of animal origin has been set at 100
662
µg kg-1 by the European Union [77]. Careful selection of gradient conditions by She and
663
co-workers permitted the separation of 24 sulphonamides in bovine milk in 10 min [80].
664
Although C8 and phenyl stationary phases were investigated, an Acquity™ BEH C18
665
column proved most suitable. Zhao et al. employed an immunoaffinity chromatography
666
(IAC) technique to purify muscle and milk extracts containing 12 sulphonamides [79].
667
McDonald et al. developed a rapid method to isolate 19 sulphonamides from muscle
668
matrix using 0.1 M EDTA/acetonitrile mix, with a total injection cycle time of 15 min.
669
Two sets of sulphonamides had similar molecular weights and parent-daughter transitions
670
(tetracycline/ doxycycline and sulfamethoxypyridazine/ sulfamonomethoxine).
671
authors distinguished between these compounds using retention time [81]. Tamošiūnas et
672
al. compared the performance capabilities of LC-MS/MS to UHLC-MS/MS. A simple
673
acetonitrile extraction procedure was carried out to extract ten sulphonamides from egg
674
and honey matrix, followed by SPE clean-up with Strata-X™ cartridges.
675
recoveries (80 – 110 %) were obtained by HPLC and UHPLC-MS/MS determination for
Ac ce
pt
ed
M
658
The
Similar
25
Page 26 of 47
676
all analytes.
However, narrower peaks, and consequently better chromatographic
677
resolution was observed with UHPLC-MS/MS [82].
678
method for isolating 24 sulphonamides in muscle followed by analysis via UHPLC–
679
MS/MS. Three columns (C18 5 µm, C18 1.7 µm, and C8 1.7 µm) were tested for signal
680
intensity and separation efficiency.
681
displayed improved sensitivity and separation efficiency, when compared to the C18 5 µm
682
column [83].
Cai et al. developed a simple
cr
ip t
The columns containing smaller particle size
us
683
A large number of anthelmintic drugs are licensed for the treatment of gastrointestinal
685
nematodes and trematodes in food producing animals [84]. They include nematicides,
686
flukicides, endectocides, benzimidazoles and macrocyclic lactones [14], [21]. The EU,
687
through Council Regulation 2010/37, established MRLs for a limited number of
688
anthelmintics in food of animal origin [85], [21]. Zhang et al. described a procedure for
689
isolating albendazole (ABZ) and three of its metabolites in fish muscle using UHPLC-
690
MS/MS equipped with an electrospray ionisation (ESI) source [86]. The separation
691
capabilities of three columns (BEH C8, BEH C18, and CSH C18) were compared.
692
Identical retention times (BEH C8), poor separation and double-humped peaks (BEH C18)
693
were observed for two of the three columns.
694
profile, the CSH C18 column was selected and used for the study. A study of extraction
695
behaviour was necessary due to the variation in lipophilicity and pKa among ABZ and its
696
metabolites. Optimum peak shape and sensitivity were obtained through optimisation of
697
the reconstitution solution. Subsequent to a series of optimisation experiments where
698
mean recoveries were monitored, ethyl acetate was selected as an extraction solvent.
699
Whelan et al. extracted 38 anthelmintic drugs from milk using a modified QuEChERS-
700
type extraction method. Factorial design experiments were carried to determine the
701
optimum mobile phase composition.
702
ionisation of analytes improved when the mobile phase contained fewer additives.
703
However, it was necessary to add ammonium formate to the mobile phase at a low
704
concentration to improve the sensitivity of several of the anthelmintic drugs [21]. Whelan
705
et al. further optimised this method by including an enzymatic hydrolysis step to
Following optimisation of the gradient
Ac ce
pt
ed
M
an
684
Results from these experiments indicated that
26
Page 27 of 47
investigate the prevalence of nitroxynil and its conjugate forms in lactating dairy cows
707
[87]. Xia et al. describe the rapid chromatographic separation of 13 benzimidazoles in 8
708
min using UHPLC-MS/MS [88]. The analysis time for each injection was 8 min, which
709
was substantially lower than an earlier HPLC-MS/MS method [89]. UHPLC-MS/MS
710
methods for other members of group B2 have also been published, including
711
anticoccidials [90] and non-steroidal anti-inflammatory drugs (NSAIDs) [91].Hu and co-
712
workers have recently described a UHPLC-MS/MS method for the determination of
713
thirty NSAIDs in swine muscle, in a run time of 16 min [92].
cr
us
714
ip t
706
Other possible contaminants and substances of environmental origin are covered within
716
group B3 in Council Directive 96/23/EC [56], such as mycotoxins. As is commonly the
717
case with multi-residue methods, the chemical diversity found within this group of
718
compounds represents a challenge for analysis. However, combined with an effective
719
clean-up technique, UHPLC-MS/MS has provided a means of rapidly and sensitively
720
determining these contaminants. Van Pamel et al. developed a UHPLC-MS/MS method
721
for 26 different mycotoxins in maize silage [93]. The polarity-switching capability of the
722
mass spectrometer system used (a Waters Xevo™ TQ mass spectrometer) permitted
723
detection of positively and negatively ionised analytes within the same run, avoiding the
724
need for two separate injections of each sample and giving a total run time of 9 min.
726 727
M
ed
pt
Ac ce
725
an
715
5.3 Multi-class multi-residue analysis
728
Finally, the expansion of methods to cover ever greater numbers of compounds means
729
that certain methods cannot be simply categorised as being specific towards banned or
730
MRL compounds. We here examine recent methods which are capable of simultaneously
731
determining several classes of analyte. The advantages of such an approach are high
732
sample throughput, simplicity and reduced cost. The bulk of methods reported use
27
Page 28 of 47
tandem mass spectrometry (MS/MS), as multiple transitions can be monitored for
734
compounds simultaneously, providing unambiguous identification. Perhaps one of the
735
most accomplished methods published to date is the generic extraction and analysis
736
procedure published by Mol and colleagues [94]. With careful examination of extraction
737
parameters, as well as matrix effects, existing multi-analyte methods were combined to
738
give a method capable of analysing for 172 pesticides, mycotoxins and plant toxins.
739
Separation and detection of the analytes was performed via UHPLC-MS/MS. Hermann et
740
al also demonstrated how a generic sample extraction procedure combined with UHPLC-
741
MS/MS analysis could be used to test for a range of 108 compounds (out of 127 targeted
742
compounds) related to emergency situations [95]. One of the methods displaying the
743
widest scope of analytes to date is the UHPLC-MS/MS method for analysis of
744
contaminants in milk published by Zhan et al [96].The method covered 255 veterinary
745
drug residues, pesticide residues, mycotoxins and other potential chemical contaminants,
746
using a generic extraction technique based on a two-step precipitation, followed by
747
sample drying and then reconstitution in injection solvent.
M
an
us
cr
ip t
733
ed
748
For screening of large numbers of compounds, time-of-flight (ToF) mass spectrometry
750
has also been successfully applied. Kaufmann et al. developed a method for isolating 100
751
veterinary drug residues from muscle, liver and kidney prior to UHPLC-ToF MS
752
analysis.
753
extraction), a technique which is known for its ability to recover polar, and medium polar
754
compounds.
755
analytes. Evaluation of the method indicated that these polar analytes were adsorbing on
756
proteins which precipitated during the extraction stage. This issue was overcome by
757
rinsing the sample and glassware with DMSO and complexing buffer. Poor recovery of at
758
high spiking concentration was attributed to detector saturation and a mass shift, a
759
phenomenon not observed at lower concentrations [97].
760
modified and improved this procedure to isolate over 100 residues in muscle, kidney,
761
liver, fish and honey using a single stage Orbitrap MS. Co-eluting proteins or peptides
762
resulted in extensive signal suppression effects, which is an issue that had not been
pt
749
Ac ce
Samples were extracted using liquid-liquid-solid extraction (bi-polarity Kaufmann encountered difficulties when attempting to isolate polar
Recently, Kaufmann et al.
28
Page 29 of 47
763
typically encountered when using ToF [18]. In the area of environmental contaminant
764
analysis, UHPLC-ToF-MS has been successfully applied to the screening of organic
765
pollutants in water [98].
ip t
766 Also selecting ToF MS for detection, Ortelli et al. reported a method for extracting 150
768
residues in milk which involved protein precipitation using acetonitrile and ultrafiltration.
769
The group encountered difficulties in the isolation of avermectin residues, where
770
recoveries (≤ 10%) did not fulfil the criteria of validation. This result was attributed to
771
losses during ultrafiltration. A post-column infusion system was used to evaluate matrix
772
effects on response. Signal response was sample-dependent for two of the compounds
773
(febantel and erythromycin), while a significant signal enhancement (1300 %) was
774
observed for enrofloxacin [99]. Peters et al. developed a simple extraction procedure to
775
isolate 100 veterinary drug residues in egg, fish and muscle by extracting with
776
acetonitrile and purifying using Strata-X™ SPE cartridges. Separation was performed
777
using a Waters Acquity™ UPLC BEH C18 column, and detection was via ToF-MS. The
778
authors constructed the method using retention times, specific accurate mass and isotope
779
pattern ratios calculated by SigmaFit™ [100]. Rapid quantification and accurate mass
780
measurement for 166 pesticides in fruits and vegetables was reported by Wang and co-
781
workers [60]. Following sample extraction via a QuEChERS approach, samples were
782
tested using a full-scan MS detection strategy for quantification, with confirmation via a
783
UHPLC-Q-Orbitrap MS/MS data-dependent scan. This represents an efficient means of
784
confirmatory analysis of a large number of pesticides, with the additional benefit of
785
circumventing the need for lengthy compound-dependent parameter optimisation as is
786
necessary with MRM approaches. QuEChERS in particular has proven to be an excellent
787
means of sample preparation for pesticides, providing clean sample extracts for analysis
788
by UHPLC-MS/MS, such as in the recent work described by Arienzo et al on the testing
789
of pesticides in fresh cut vegetables [101].
Ac ce
pt
ed
M
an
us
cr
767
790
29
Page 30 of 47
791
Using the more commonly-employed LC-MS/MS approach, Rezende et al. described a
792
simple extraction procedure for 12 β-lactams and tetracyclines (in bovine muscle) which
793
traditionally rely on time-consuming, expensive SPE approaches. Tetracyclines can be
794
difficult to analyse for, due to the ability of tetracycline to epimerise to 4-epitetracycline
795
at pH 4 - 8.
796
purification with dSPE (dispersive solid phase extraction) C18 and finally defatting with
797
hexane. Rezendes studied the matrix effect using an F-test to examine the
798
homoscedasticity of data and a t-test to compare average results obtained in matrix
799
samples and solvent. Results from these tests suggested there was a matrix effect. To
800
overcome this issue, samples were quantified using matrix-matched calibration curves
801
[102]. Lehotay and co-workers have recently reported a streamlined method designed to
802
detect residues of as many as 60 priority veterinary drugs in bovine kidney, using
803
UHPLC-MS/MS [103], with an injection cycle time of under 10 min. The authors
804
determined that their method was suitable for screening of 54 of the 62 drugs studied,
805
qualitative identification of 50 compounds, and quantification of 30. For bovine muscle
806
[104], a multi-class, multi-residue UHPLC-MS/MS method was designed for the purpose
807
of monitoring over 100 veterinary drug residues. The importance of the choice of the
808
dispersive solid-phase extraction sorbent used was investigated, as was the influence of
809
post-column infusion of ammonium formate on ionisation enhancement for anthelmintic
810
compounds. The final method presented was capable of screening for 113 compounds,
811
identification of 98 compounds, and quantification of 87.
ip t
cr
us
an
M
ed
pt
Ac ce
812
The method involved extraction with water/acetonitrile followed by
813
When MRLs have not been established for compounds, researchers typically attempt to
814
develop analytical methods capable of detecting these compounds at extremely low
815
levels. Aguilera-Luiz et al. encountered this problem when attempting to develop a
816
multiclass method for 39 veterinary drugs in meat-based baby food and powdered milk-
817
based infant formulae [46]. Separation was carried out on a C18 column (100 mm × 2.1
818
mm, 1.7 μm particle size) using a previously developed chromatographic procedure [94].
819
Chromatographic separation was optimised by evaluating two solvents, acetonitrile and
820
methanol.
Retention time decreased with acetonitrile; however sensitivity was 30
Page 31 of 47
821
significantly better when methanol was used in the mobile phase. The authors also
822
compared two additives, formic acid and acetic acid. Ionisation efficiency improved with
823
the addition of formic acid at a concentration of 0.05%.
ip t
824 Matrix effects associated with milk can pose a significant challenge in creating multi-
826
residue methods. Tang et al. described the difficulties encountered when attempting to
827
extract veterinary drugs from milk. A relatively high level of acetonitrile was used to
828
isolate 23 veterinary drugs from milk.
829
matrix-matched calibration curves were employed for accurate quantitation [105]. Vidal
830
et al. developed two screening methods to identify 21 veterinary drug residues in milk.
831
The first method used parent ion and neutral loss scan whereas the second method only
832
used one transition per compound. Both methods were separately validated but it was
833
ultimately decided that the second method was more suitable as it was capable of
834
screening at lower concentration levels [40]. Wang et al. compared the separation of ten
835
antibiotic residues in milk using three kinds of columns (C18, C8 and phenyl). The
836
investigation indicated that satisfactory separation and resolution could only be achieved
837
using a C18 column. Samples were extracted using McIlvaine buffer: methanol (8:2), and
838
purified with HLB SPE cartridges [106].
pt
ed
M
an
us
Matrix effects were quantified in milk, and
Ac ce
839
cr
825
840
The QuEChERS technique has proved suitable when it comes to cleaning up milk
841
samples prior to UHPLC-MS/MS analysis. Aguilera-Luiz et al. developed a method for
842
the analysis of 18 veterinary drugs in milk. Initially, extraction was carried out using
843
acidified acetonitrile followed by clean-up using QuEChERS dSPE. Separation was
844
performed using an Acquity™ UPLC BEH C18 column. Poor recoveries were attributed
845
to the propensity macrolides have to form chelation complexes with cations present in
846
solution. To overcome this problem, EDTA was added to the extractant solution prior to
847
QuEChERS clean-up [107]. Vidal et al. utilised this method to isolate 17 antibiotic
848
residues in honey. To avoid contamination of the column from lipids and waxes present
31
Page 32 of 47
in honey, the organic solvent content was quickly raised during the gradient profile [108].
850
Frenich et al. compared several extraction procedures (solvent extraction, SPE, MSPD,
851
and a modified QuEChERs procedure) to analyse 25 veterinary drugs in eggs. A uniform
852
dwell time of 0.015 s was used for most of the compounds except for levamisole,
853
marbofloxacin and sulfadimidine which had a dwell time of 0.025 s, whereas
854
fenbendazole, emamectin and ivermectin were monitored at 0.100 s. Results suggested
855
that solvent extraction was the most reliable procedure for the simultaneous extraction of
856
several classes of veterinary drugs (tetracyclines, macrolides, quinolones, sulphonamides
857
and anthelmintics), where recoveries ranged from 60 – 119 % [109].
us
cr
ip t
849
858
As previously mentioned, the tetracycline compounds pose particular challenges for
860
multi-compound methods, as very specific analysis conditions are required. Shao et al.
861
described the difficulties encountered when developing an extraction procedure for 21
862
veterinary drugs from two classes (tetracyclines and quinolones).
863
typically extracted solvents acidified at a pH of less than 3. However under these acidic
864
conditions, tetracyclines will reversibly form 4-epi-tetracyclines and anhydro-
865
tetracyclines. An investigation was carried out, where three different extraction solvents
866
(methanol, acetonitrile and EDTA-McIlvaine buffer) were acidified at pH 4 prior to
867
sample clean-up with HLB SPE. Results indicated that acceptable repeatability was
868
achieved with the use of EDTA-McIlvaine (pH 4) [110].
870 871
M
pt
ed
Quinolones are
Ac ce
869
an
859
6. Conclusion
872
The emergence and rapid development of UHPLC-MS/MS technology which we have
873
documented in this review has had a large impact on the area of veterinary drug residue
874
analysis and related compounds. Injection cycle times have shortened, sensitivity has
875
improved, chromatographic resolution of compounds has increased and the number of
32
Page 33 of 47
compounds included in multi-residue methods continues to climb. The combination of
877
these advantages (along with peripheral advantages such as lower solvent consumption)
878
has led to the increasing adoption of UHPLC-MS/MS as the default approach for
879
confirmatory analysis of multiple compounds. In this review, we have examined the
880
technical aspects of coupling UHPLC instrumentation to mass spectrometric detection
881
systems, as well as highlighting some of the inherent challenges that come with the
882
technology, such as frictional heating effects, narrow chromatographic peaks and the
883
impact of sample matrix effects. Despite these challenges, the arenas of veterinary drug
884
residue analysis, pesticide analysis and mycotoxin testing have shown prolific uptake of
885
the technology, and we have detailed some of the most accomplished methods which
886
employ UHPLC-MS/MS as the separation and detection system. Comprehensive methods
887
for some of the more important contaminant groups in residue analysis have been
888
developed for UHPLC-MS/MS, including anthelmintics, -agonists, steroids, quinolones
889
and others. Expected future developments include evolution of methods to include larger
890
numbers of compounds and classes of compound; the use of ever higher temperatures and
891
pressures to create more effective separation methods; and further reduction in sample
892
preparation via online SPE and other techniques, to increase the speed of analysis beyond
893
current standards.
pt
894
ed
M
an
us
cr
ip t
876
Acknowledgements
896
This work was funded by the Food for Health Research Initiative (FHRI) administered by
897
the Irish Department of Agriculture, Food and the Marine and the Health Research Board
898
(Contract 7FHRIAFRC5).
Ac ce
895
899 900 901 902 903
[1] [2] [3]
Off. J. Eur. Union L221 (2002) 8. M. McDonald, K. Granelli, P. Sjoberg, Anal. Chim. Acta 588 (2007) 20. Y. Deceuninck, E. Bichon, F. Monteau, J.P. Antignac, B. Le Bizec, Anal. Chim. Acta 700 (2011) 137.
33
Page 34 of 47
[12] [13] [14] [15] [16] [17] [18] [19] [20] [21]
ip t
cr
[11]
us
[10]
an
[9]
M
[6] [7] [8]
ed
[5]
M.E. Touber, M.C. van Engelen, C. Georgakopoulus, J.A. van Rhijn, M.W.F. Nielen, Anal. Chim. Acta 586 (2007) 137. M.J. Taylor, G.A. Keenan, K.B. Reid, D.U. Fernandez, Rapid Commun. Mass Sp. 22 (2008) 2731. J.E. MacNair, K.C. Lewis, J.W. Jorgenson, Anal. Chem. 69 (1997) 983. J.E. MacNair, K.D. Patel, J.W. Jorgenson, Anal. Chem. 71 (1999) 700. M. Knaggs, FAPAS Proficiency Test Report 02183 Coccidiostats in Egg Powder February - April 2012, FAPAS, 2012. M. Knaggs, FAPAS Proficiency Test Report 02179 Nitrofuran Metabolites in Chicken Egg November - December 2011, FAPAS, 2012. M. Knaggs, FAPAS Proficiency Test Report 02175 Avermectins in Corned Beef September - October 2011, FAPAS, 2011. U. Chiuminatto, F. Gosetti, P. Dossetto, E. Mazzucco, D. Zampieri, E. Robotti, M.C. Gennaro, E. Marengo, Anal. Chem. 82 (2010) 5636. T. Vega-Morales, Z. Sosa-Ferrera, J.J. Santana-Rodriguez, J. Chromatogr. A 1230 (2012) 66. D. Gode, M.M. Martin, F. Steiner, C.G. Huber, D.A. Volmer, Anal. Bioanal. Chem. 404 (2012) 433. B. Kinsella, M. Whelan, H. Cantwell, M. McCormack, A. Furey, S.J. Lehotay, M. Danaher, Talanta 83 (2010) 14. A. Kaufmann, P. Butcher, K. Maden, S. Walker, M. Widmer, Anal. Chim. Acta 673 (2010) 60. A. Kaufmann, P. Butcher, K. Maden, S. Walker, M. Widmer, Rapid Commun. Mass Sp. 25 (2011) 979. N. Leon, M. Roca, C. Igualada, C.P.B. Martins, A. Pastor, V. Yusa, J. Chromatogr. A 1258 (2012) 55. A. Kaufmann, P. Butcher, K. Maden, S. Walker, M. Widmer, Anal. Chim. Acta 700 (2011) 86. R.K. Boyd, C. Basic, R.A. Bethem, Trace Quantitative Analysis by Mass Spectrometry, John Wiley & Sons Ltd., Chicester, 2008. K. Yu, D. Little, R. Plumb, B. Smith, Rapid Commun. Mass Sp. 20 (2006) 544. M. Whelan, B. Kinsella, A. Furey, M. Moloney, H. Cantwell, S.J. Lehotay, M. Danaher, J. Chromatogr. A 1217 (2010) 4612. R. Ventura, M. Roig, N. Monfort, P. Saez, R. Berges, J. Segura, Eur. J. Mass Spectrom. 14 (2008) 191. C.C. Leandro, P. Hancock, R.J. Fussell, B.J. Keely, J. Chromatogr. A 1103 (2006) 94. J.C. Van De Steene, W.E. Lambert, J. Am. Soc. Mass Spectrom. 19 (2008) 713. D. Guillarme, J. Ruta, S. Rudaz, J.L. Veuthey, Anal. Bioanal. Chem. 397 (2010) 1069. H.F. De Brabander, J. Vanden Bussche, W. Verbeke, L. Vanhaecke, TrAC Trend Anal. Chem. 30 (2011) 1088. M.I. Churchwell, N.C. Twaddle, L.R. Meeker, D.R. Doerge, J. Chromatogr. B 825 (2005) 134. A. de Villiers, H. Lauer, R. Szucs, S. Goodall, P. Sandra, J. Chromatogr. A 1113 (2006) 84.
pt
[4]
Ac ce
904 905 906 907 908 909 910 911 912 913 914 915 916 917 918 919 920 921 922 923 924 925 926 927 928 929 930 931 932 933 934 935 936 937 938 939 940 941 942 943 944 945 946 947 948 949
[22] [23] [24] [25] [26] [27] [28]
34
Page 35 of 47
[41] [42] [43] [44] [45] [46] [47] [48] [49]
ip t
cr
us
[39] [40]
an
[38]
M
[37]
ed
[36]
L. Novakova, J.L. Veuthey, D. Guillarme, J. Chromatogr. A 1218 (2011) 7971. D.V. McCalley, J. Chromatogr. A 1218 (2011) 2887. M.M. Fallas, M.R. Hadley, D.V. McCalley, J. Chromatogr. A 1216 (2009) 3961. D.V. McCalley, J. Chromatogr. A 1217 (2010) 4561. F. Gritti, G. Guiochon, J. Chromatogr. A 1218 (2011) 4632. A.C. Sanchez, J.A. Anspach, T. Farkas, J. Chromatogr. A 1228 (2012) 338. F.Y. Guan, C.E. Uboh, L.R. Soma, Y.W. You, Y. Liu, X.Q. Li, J. Mass Spectrom. 45 (2010) 1270. http://www.chem.agilent.com/Library/applications/5990-4253EN.pdf, Last accessed 22/06/2012. D. Guillarme, J. Schappler, S. Rudaz, J.L. Veuthey, TrAC - Trend Anal. Chem. 29 (2010) 15. A. Maquille, D. Guillarme, S. Rudaz, J.L. Veuthey, Chromatographia 70 (2009) 1373. K. Mastovska, S.J. Lehotay, J. Chromatogr. A 1000 (2003) 153. J.L.M. Vidal, A.G. Frenich, M.M. Aguilera-Luiz, R. Romero-Gonzalez, Anal. Bioanal. Chem. 397 (2010) 2777. T. Kovalczuk, M. Jech, J. Poustka, J. Hajslova, Anal. Chim. Acta 577 (2006) 8. J.P. Antignac, K. de Wasch, F. Monteau, H. De Brabander, F. Andre, B. Le Bizec, Anal. Chim. Acta 529 (2005) 129. B.K. Matuszewski, M.L. Constanzer, C.M. Chavez-Eng, Anal. Chem. 75 (2003) 3019. H. Trufelli, P. Palma, G. Famiglini, A. Cappiello, Mass Spectrom. Rev. 30 (2011) 491. R. King, R. Bonfiglio, C. Fernandez-Metzler, C. Miller-Stein, T. Olah, J. Am. Soc. Mass Spectrom. 11 (2000) 942. M.M. Aguilera-Luiz, J.L.M. Vidal, R. Romero-Gonzalez, A.G. Frenich, Food Chem. 132 (2012) 2171. J.L. Little, M.F. Wempe, C.M. Buchanan, J. Chromatogr. B 833 (2006) 219. O.A. Ismaiel, M.S. Halquist, M.Y. Elmamly, A. Shalaby, H.T. Karnes, J. Chromatogr. B 859 (2007) 84. V. Pucci, S. Di Palma, A. Alfieri, F. Bonelli, E. Monteagudo, J. Pharmaceut. Biomed. 50 (2009) 867. E. Chambers, D.M. Wagrowski-Diehl, Z.L. Lu, J.R. Mazzeo, J. Chromatogr. B 852 (2007) 22. E. Varga, T. Glauner, R. Koppen, K. Mayer, M. Sulyok, R. Schuhmacher, R. Krska, F. Berthiller, Anal. Bioanal. Chem. 402 (2012) 2675. E. Duffy, L. Rambaud, B. Le Bizec, M. O'Keeffe, Anal. Chim. Acta 637 (2009) 165. B. Kinsella, P. Byrne, H. Cantwell, M. McCormack, A. Furey, M. Danaher, J. Chromatogr. B 879 (2011) 3707. A. Radovnikovic, M. Moloney, P. Byrne, M. Danaher, J. Chromatogr. B 879 (2011) 159. B.F. Spisso, M.A.G. de Araujo, M.A. Monteiro, A.M.B. Lima, M.U. Pereira, R.A. Luiz, A.W. da Nobrega, Anal. Chim. Acta 656 (2009) 72. Off. J. Eur. Union L125 (1996) 10.
pt
[29] [30] [31] [32] [33] [34] [35]
Ac ce
950 951 952 953 954 955 956 957 958 959 960 961 962 963 964 965 966 967 968 969 970 971 972 973 974 975 976 977 978 979 980 981 982 983 984 985 986 987 988 989 990 991 992 993 994 995
[50] [51] [52] [53] [54] [55] [56]
35
Page 36 of 47
[67] [68] [69] [70] [71] [72] [73] [74] [75]
ip t
cr
us
[65] [66]
an
[63] [64]
M
[62]
http://ec.europa.eu/sanco_pesticides/public/index.cfm?event=homepage&CFID=1 404859&CFTOKEN=69040180&jsessionid=2405714dcd661a78e106TR#, Last accessed 18/12/2012. Off. J. Eur. Union L70 (2005) 1 J. Wang, W. Chow, D. Leung, J. Chang, J Agric Food Chem 60 (2012) 12088. M. Farre, L. Kantiani, M. Petrovic, S. Perez, D. Barcelo, J Chromatogr A 1259 86. J.V. Bussche, L. Vanhaecke, Y. Deceuninck, K. Verheyden, K. Wille, K. Bekaert, B. Le Bizec, H.F. De Brabander, J. Chromatogr. A 1217 (2010) 4285. K. Schmidt, C. Stachel, P. Gowik, Anal. Bioanal. Chem. 391 (2008) 1199. B. Kinsella, J. O'Mahony, E. Malone, M. Moloney, H. Cantwell, A. Furey, M. Danaher, J. Chromatogr. A 1216 (2009) 7977. Off. J. Eur. Union L125 (1996) 3. C. Van Poucke, E. Van Vossel, C. Van Peteghem, Rapid Commun. Mass Sp. 22 (2008) 2324. E.M. Malone, C.T. Elliott, D.G. Kennedy, L. Regan, Anal. Chim. Acta 637 (2009) 112. H.X. Wang, Y. Zhou, Q.W. Jiang, Chinese J. Anal. Chem. 39 (2011) 1323. Y.W. You, C.E. Uboh, L.R. Soma, F.Y. Guan, X.Q. Li, J.A. Rudy, Y. Liu, J.W. Chen, Rapid Commun. Mass Sp. 23 (2009) 2035. C.H.F. Wong, F.P.W. Tang, T.S.M. Wan, Anal. Chim. Acta 697 (2011) 48. H. Zheng, L.G. Deng, X.A. Lu, S.C. Zhao, C.Y. Guo, J.S. Mao, Y.T. Wang, G.S. Yang, H.Y. Aboul-Enein, Chromatographia 72 (2010) 79. F. Yang, Z.C. Liu, Y.H. Lin, J. Chen, G.F. Pang, G.N. Chen, Food Anal. Method 5 (2012) 138. B. Shao, X.F. Jia, J. Zhang, J. Meng, Y.N. Wu, H.J. Duan, X.M. Tu, Food Chem. 114 (2009) 1115. J. Zhang, B. Shao, J. Yin, Y.N. Wu, H.J. Duan, J. Chromatogr. B 877 (2009) 1915. M.W.F. Nielen, J.J.P. Lasaroms, M.L. Essers, J.E. Oosterink, T. Meijer, M.B. Sanders, T. Zuidema, A.A.M. Stolker, Anal. Bioanal. Chem. 391 (2008) 199. H. Zhang, Y.P. Ren, X.L. Bao, J. Pharmaceut. Biomed. 49 (2009) 367. Off. J. Eur. Comm. L224 (1990) 1. M. Lombardo-Agui, A.M. Garcia-Campana, L. Gamiz-Gracia, C. Cruces-Blanco, Talanta 93 (2012) 193. X.T. Zhao, Q.B. Lin, H. Song, Y.L. Pan, X. Wang, J. Agric. Food Chem. 59 (2011) 9800. Y.X. She, J.J. Liu, J. Wang, Y. Liu, R.Y. Wang, W.Q. Cao, Anal. Lett. 43 (2010) 2246. M. McDonald, C. Mannion, P. Rafter, J. Chromatogr. A 1216 (2009) 8110. V. Tamosiunas, A. Padarauskas, Chromatographia 67 (2008) 783. Z.X. Cai, Y. Zhang, H.F. Pan, X.W. Tie, Y.P. Ren, J. Chromatogr. A 1200 (2008) 144.
ed
[59] [60] [61]
Off. J. Eur. Union L309 (2009) 1
pt
[57] [58]
Ac ce
996 997 998 999 1000 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040
[76] [77] [78] [79] [80] [81] [82] [83]
36
Page 37 of 47
[91] [92] [93] [94] [95] [96] [97] [98] [99] [100]
ip t
cr
[90]
us
[89]
an
[88]
M
[87]
ed
[85] [86]
M. Whelan, C. Chirollo, A. Furey, M.L. Cortesi, A. Anastasio, M. Danaher, J. Agric. Food Chem. 58 (2010) 12204. Off.J.Eur.Union L15 (2010) 1. X.J. Zhang, H.X. Xu, H. Zhang, Y.M. Guo, Z.Y. Dai, X.C. Chen, Anal. Bioanal. Chem. 401 (2011) 727. M. Whelan, Y. Bloemhoff, A. Furey, R. Sayers, M. Danaher, J. Agric. Food Chem. 59 (2011) 7793. X. Xia, Y.C. Dong, P.J. Luo, X. Wang, X.W. Li, S.Y. Ding, J.Z. Shen, J. Chromatogr. B 878 (2010) 3174. H. De Ruyck, E. Daeseleire, H. De Ridder, R. Van Renterghem, J. Chromatogr. A 976 (2002) 181. B. Shao, X.Y. Wu, J. Zhang, H.J. Duan, X.G. Chu, Y.N. Wu, Chromatographia 69 (2009) 1083. E.M. Malone, G. Dowling, C.T. Elliott, D.G. Kennedy, L. Regan, J. Chromatogr. A 1216 (2009) 8132. T. Hu, T. Peng, X.J. Li, D.D. Chen, H.H. Dai, X.J. Deng, Z.F. Yue, G.M. Wang, J.Z. Shen, X. Xia, S.Y. Ding, Y.N. Zhou, A.L. Zhu, H.Y. Jiang, J. Chromatogr. A 1219 (2012) 104. E. Van Pamel, A. Verbeken, G. Vlaemynck, J. De Boever, E. Daeseleire, J. Agric. Food Chem. 59 (2011) 9747. H.G.J. Mol, P. Plaza-Bolanos, P. Zomer, T.C. de Rijk, A.A.M. Stolker, P.P.J. Mulder, Anal. Chem. 80 (2008) 9450. A. Herrmann, J. Rosen, D. Jansson, K.E. Hellenas, J. Chromatogr. A 1235 115. J. Zhan, X.J. Yu, Y.Y. Zhong, Z.T. Zhang, X.M. Cui, J.F. Peng, R. Feng, X.T. Liu, Y. Zhu, J. Chromatogr. B 906 (2012) 48. A. Kaufmann, P. Butcher, K. Maden, M. Widmer, J. Chromatogr. A 1194 (2008) 66. M. Ibanez, J.V. Sancho, D. McMillan, R. Rao, F. Hernandez, TrAC - Trend Anal. Chem. 27 (2008) 481. D. Ortelli, E. Cognard, P. Jan, P. Edder, J. Chromatogr. B 877 (2009) 2363. R.J.B. Peters, Y.J.C. Bolck, P. Rutgers, A.A.M. Stolker, M.W.F. Nielen, J. Chromatogr. A 1216 (2009) 8206. M. Arienzo, D. Cataldo, L. Ferrara, Food Control 31 (2013) 108 C.P. Rezende, M.P. Almeida, R.B. Brito, C.K. Nonaka, M.O. Leite, Food Addit. Contam. A 29 (2012) 541. S.J. Lehotay, A.R. Lightfield, L. Geis-Asteggiante, M.J. Schneider, T. Dutko, C. Ng, L. Bluhm, K. Mastovska, Drug Test. Anal. 4 (2012) 75. L. Geis-Asteggiante, S.J. Lehotay, A.R. Lightfield, T. Dutko, C. Ng, L. Bluhm, J. Chromatogr. A 1258 (2012) 43. Y.Y. Tang, H.F. Lu, H.Y. Lin, Y.C. Shih, D.F. Hwang, J. Chromatogr. B 881-882 (2012) 12. L. Wang, Y.Q. Li, Chromatographia 70 (2009) 253. M.M. Aguilera-Luiz, J.L.M. Vidal, R. Romero-Gonzalez, A.G. Frenich, J. Chromatogr. A 1205 (2008) 10. J.L.M. Vidal, M.D. Aguilera-Luiz, R. Romero-Gonzalez, A.G. Frenich, J. Agric. Food Chem. 57 (2009) 1760.
pt
[84]
Ac ce
1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055 1056 1057 1058 1059 1060 1061 1062 1063 1064 1065 1066 1067 1068 1069 1070 1071 1072 1073 1074 1075 1076 1077 1078 1079 1080 1081 1082 1083 1084 1085 1086
[101] [102] [103] [104] [105] [106] [107] [108]
37
Page 38 of 47
1087 1088 1089 1090 1091
[109] A.G. Frenich, M.D. Aguilera-Luiz, J.L.M. Vidal, R. Romero-Gonzalez, Anal. Chim. Acta 661 (2010) 150. [110] B. Shao, X. Jia, Y. Wu, J. Hu, X. Tu, J. Zhang, Rapid Commun. Mass Sp. 21 (2007) 3487.
Ac ce
pt
ed
M
an
us
cr
ip t
1092
38
Page 39 of 47
Table 1. Currently available commercial UHPLC systems and capabilities. Table 2. Examples of currently available commercial tandem mass spectrometric detection systems and capabilities.
ip t
Table 3. Application of UHPLC-MS to analysis of veterinary drug residues.
Fig. 1. Structures of epimeric synthetic glucocorticosteroids dexamethasone and
us
cr
betamethasone.
Fig. 2. Growth in the number of publications on UHPLC, and UHPLC combined with mass spectrometry, 2003 – 2011. Data was sourced from Thomson Reuters Web of
M
an
Science™ (keywords: “UHPLC” or “UPLC”; “MS” or “mass spec*”).
Fig. 3. Band broadening due to frictional heating. The top diagram shows the typical
ed
quantity of band broadening expected due to diffusion processes when the column is heated evenly. The lower diagram shows the effect of internal column heating due to
ce pt
frictional effects between mobile phase and stationary phase.
Fig. 4. Experiment for observation of suppression effects. The signal corresponds to the introduction into the mass spectrometer of a mobile phase or blank sample extract
Ac
injection (from the LC) and standard solution (via direct infusion). When just mobile phase is introduced from the LC, signal remains constant (a, b). However, upon injection of blank sample extract, although the TIC increases (c), the specific transition monitored for the analyte (d) is suppressed due to matrix effects. Adapted from [20] with permission from Elsevier.
Page 40 of 47
Ac
ce pt
ed
M
an
us
cr
ip t
Figure 1
Page 41 of 47
Figure 2
800
700
600
ip t
500
400
UHPLC
cr
300
UHPLC & MS
200
0 2005
2006
2007
2008
an
2004
2009
2010
2011
Ac
ce pt
ed
M
2003
us
100
Page 42 of 47
Ac
ce pt
ed
M
an
us
cr
ip t
Figure 3
Page 43 of 47
Ac
ce pt
ed
M
an
us
cr
ip t
Figure 4
Page 44 of 47
Table 1
Manufacturer UHPLC System
Maximum
Maximum
Comments
Pressure
Flow
(bar)
(mL min-1)*
Rate
Pump delay volume of <10 μL, Infinity 1290
1200
2
Hitachi
LaChromUltra
600
2.5
depending on solvent mixing unit.
ip t
Agilent
System delay volume as low as 266 μL.
Jasco
X-LC
1000
cr
Minimum injection interval of 30 1.35
seconds.
Minimum injection interval of 14 Nexera
1300
3
Thermo Scientific
Accela
1310
2
an
Class
1000
1
1240
Waters
nanoAcquity
690
Dionex
UltiMate
1000
PerkinElmer
Flexar
1240
Knauer
1720
ce pt
Eksigent
UltraLC
PLATINblue
1
M
Class
ed
Waters
UHP
injection interval of 30 seconds. System delay volume of < 100 μL;
Acquity I-
SSI LabAlliance
Pump delay volume of 70 μL. System delay volume < 400 μL;
Acquity HWaters
seconds
us
Shimadzu
1240
1000
0.1
injection interval of 15 seconds. System delay volume of < 1 μL.
8
Injection cycle time of 30 seconds.
5
Injection cycle time of 8 seconds. Flow rates of up to 12 mL min-1
5
possible. Mixer volume 60 μL. Injection cycle
5
time of < 60 seconds. System delay volume of 100 μL; 15
2
second injection cycle time.
Ac
* Maximum flow rate achievable at maximum pressure
Page 45 of 47
Table 2
Vendor
Dwell time (ms)
Polarity Switch (ms)
Quoted Sensitivity
Scan Speed
6460 Triple Quadrupole 6500 Triple Quadrupole Xevo TQ-S
Agilent
1
30
1 pg (Reserpine)
AB SCIEX
1
20
700 ag Alprazolam
Waters
1
20
Triple Quad 6500 8030 Triple Quadrupole TSQ Vantage Triple Quadrupole
AB SCIEX
1
20
500 pg l-1 (Raffinose) 700 ag (Verapamil)
Shimadzu
1
15
1 pg (Reserpine)
N/A
1 pg (Chloramphenicol)
5,200 amu/s 12,000 amu/s 1000 amu/s 20,000 amu/s 15,000 amu/s 5000 amu/s
cr
Ac
ce pt
ed
M
an
us
ThermoScientific 2
ip t
Mass Spectrometer
Page 46 of 47
13 β-agonists
16 22
Synthetic Steroids
55 9
Quinolones
8 22
Sulphonamides
12
Macrolides
4
NSAID
30
-Lactam
8
REF.
0.14 – 123 (CCα)
0.1
[21]
8
0.01 – 0.5 (LOD)
0.1
40
17
0.2 – 0.79 (CCα)
0.1
40
15
< 0.2 (CCα)
2.5
50
5
0.5 x 10-7
40
13
0.5 x 10-7 – 2 x 10-7 (LOD) 0.11 – 0.3 (LOD)
0.4
35
5.5
0.2 – 4.1 (LOD)
5
0.3
40
12
0.002 – 0.102 (LOD)
3 x 10-6
0.25
40
25
0.4 – 8.0 (LOD)
10
Micromass Quattro Premier XE triple quadropole MS Micromass Quattro Premier XE triple quadropole MS Micromass Quattro Premier XE triple quadropole MS Micromass Quattro Premier XE triple quadropole MS Finnigan TSQ Quantum Ultra triple stage MS Hybrid quadrupole-orthogonal acceleration TOF MS Triple quadrupole MS API 3200 Micromass Quattro Premier XE triple quadropole MS TQ Detector MS
0.3
30
9.5
12.2 – 217.8 (CCα)
1.5
[109]
0.2
30
16
0.4 - 2 (LOD)
1
Acquity TQD tandem quadrupole MS TQ MS
0.6
40
11
30.9 – 371.7 (CCα)
12.5
Micromass Quattro Premier XE triple quadropole MS
[102]
60
Injection Cycle Time (min) 13
0.3
30
0.3 0.4 0.5 0.7
LOD (µg kg-1)
us
Acquity UPLC HSS T3 (100 mm x 2.1 mm, 1.8 µm) Acquity BEH C18 (50 mm x 2.1 mm, 1.7 µm) Acquity BEH C18 (50 mm x 2.1 mm, 1.7 µm) Acquity BEH C18 (50 mm x 2.1 mm, 1.7 µm) Hypersil Gold C18 (50 mm x 2.1 mm, 1.9 µm) Acquity UPLC HSS T3 (100 mm x 2.1 mm, 1.8 µm) Zorbax Eclipse Plus HHRD (100 mm x 2.1 mm, 1.8 µm) Acquity Shield RP 18 (100 mm x 2.1 mm, 1.7 µm) Acquity BEH C18 (100 mm x 2.1 mm, 1.7 µm) Acquity BEH C18 (50 mm x 2.1 mm, 1.7 µm) Acquity BEH C18 (50 mm x 2.1 mm, 1.7 µm) Acquity BEH C18 (50 mm x 2.1 mm, 1.7 µm)
Detector
Temp (°C)
M an
38
LOQ (µg kg-1)
Flow Rate (mL min-1) 0.6
ed
Column
ce pt
Anthelmintics
Residues
Ac
Drug
cr
ip t
Tables
1
[88] [73] [75] [35] [68] [78] [76] [79]
[92]
Page 47 of 47
S0021-9673(13)00047-2 doi:10.1016/j.chroma.2013.01.007 CHROMA 353950
To appear in:
Journal of Chromatography A
Received date: Revised date: Accepted date:
16-8-2012 19-12-2012 4-1-2013
Please cite this article as: J. O’Mahony, L. Clarke, M. Whelan, R. O’Kennedy, S.J. Lehotay, M. Danaher, The Use of Ultra-High Pressure Liquid Chromatography with Tandem Mass Spectrometric Detection in the Analysis of Agrochemical Residues and Mycotoxins in FoodndashChallenges and Applications, Journal of Chromatography A (2010), doi:10.1016/j.chroma.2013.01.007 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
*Highlights (for review)
Highlights
Ac ce p
te
d
M
an
us
cr
This review examines the role of UHPLC-MS/MS technology in agrochemical analysis. The emergence of UHPLC technology is discussed, and the implications for veterinary drug residue testing are documented. Current challenges associated with UHPLC-MS methods are reported on. An overview is given of multi-residue analytical methods, where UHPLC-MS technology has been successfully applied.
ip t
Page 1 of 47
1
The Use of Ultra-High Pressure Liquid Chromatography with Tandem Mass
2
Spectrometric Detection in the Analysis of Agrochemical Residues and Mycotoxins
3
in Food – Challenges and Applications
4 John O’Mahonya, Lesa Clarkea, b, Michelle Whelana, Richard O’Kennedyb, Steven
6
J. Lehotayc,Martin Danaher*, a
ip t
5 7 8
a
9
b
cr
Food Safety Department, Teagasc Food Research Centre, Ashtown, Dublin 15, Ireland
us
School of Biotechnology and National Centre for Sensor Research, Dublin City
10
University, Dublin 9, Ireland
11
c
12
Research Center, 600 East Mermaid Lane, Wyndmoor, PA 19038, USA
an
US Department of Agriculture, Agricultural Research Service, Eastern Regional
M
13 14
*Corresponding author: Tel. +353 1 8059500; fax: +353 1 8059550
16
Email address: [email protected]
ed
15
pt
17 18 Abstract
20
In the field of food contaminant analysis, the most significant development of recent
21
years has been the integration of ultra-high pressure liquid chromatography (UHPLC),
22
coupled to tandem quadrupole mass spectrometry (MS/MS), into analytical applications.
23
In this review, we describe the emergence of UHPLC through technological advances.
24
The implications of this new chromatographic technology for MS detection are discussed,
25
as well as some of the remaining challenges in exploiting it for chemical residue
26
applications. Finally, a comprehensive overview of published applications of UHPLC-
27
MS in food contaminant analysis is presented, with a particular focus on veterinary drug
28
residues.
Ac ce
19
1
Page 2 of 47
29 30
Keywords: Mass spectrometry (MS); Ultra-high pressure liquid chromatography
31
(UHPLC); veterinary drug residue analysis; matrix effects; frictional heating
32
Ac ce
pt
ed
M
an
us
cr
ip t
33
2
Page 3 of 47
33
Contents
34 1. Introduction p.2
36
2. The introduction of UHPLC and impact on residue analysis p.5
37
3. Coupling UHPLC with MS/MS detection p.7
38
4. Challenges inherent in UHPLC-MS/MS techniques p.11
39
5. Recent applications of UHPLC-MS technology in the analysis of agrochemical
40
residues and mycotoxins p.18
41
6. Conclusion p. 29
43
References p.30
cr
Acknowledgements p. 30
us
42
an
44 45
ip t
35
1. Introduction
46
Within the past decade, the introduction of ultra-high pressure liquid chromatography
48
(UHPLC) and rapid-scan and sensitive mass spectrometery (MS instruments has resulted
49
in a seismic shift away from traditional chromatographic techniques in the field of food
50
contaminant analysis, towards multi-class, multi-residue methods with short injection
51
cycle times and minimal sample preparation. The competing pressures of improving
52
sample throughput, reducing analysis costs and complying with regulatory requirements
53
have led to the continued adoption of UHPLC-MS/MS as the technique of choice in the
54
analysis of veterinary drug residues and related substances. In the present review, a
55
comprehensive overview of published UHPLC-MS/MS food contaminant applications is
56
presented, with a particular focus on veterinary drug residues. However, as this is still a
57
relatively new development, challenges associated with the technology are still emerging,
58
and some of these will also be detailed and discussed.
ed
pt
Ac ce
59
M
47
60
Analytical strategies applied in this field have evolved, not only to respond to both
61
statutory requirements for the scope of analytes detected and the needed method detection
62
limits to meet enforcement action levels, such as those stipulated by the European
63
Commission in 2002/657/EC [1], but also in response to the adoption of MS as the most
3
Page 4 of 47
frequently used detection method. The improved selectivity offered by MS/MS methods
65
means that the analysis can still be conducted with a lesser degree of chromatographic
66
resolution between target analytes, or between matrix components and analytes. LC-
67
MS/MS also offers greater versatility than that of GC-ECD (gas chromatography –
68
electron capture detection) or GC-MS methods, where the analyte must be volatile, or
69
else put through a derivatisation process to permit volatilisation of the analyte. With the
70
specificity of detection greatly improved and sample preparation time reduced,
71
innovation in the field of chromatographic separation became an area of research focus.
72
Since 2003, more commercial UHPLC systems continue to become available (Table 1).
cr
ip t
64
74
us
73 Table 1
an
75 76
While the development of ever-more efficient monolithic silica HPLC columns has been
78
an area of extensive chromatographic research activity, the development and refinement
79
of UHPLC, employing columns packed with sub-2 m particles, has had the largest
80
impact on analytical separation science. The Van Deemter equation for describing band-
81
broadening mechanisms relates separation efficiency of a method to the linear velocity of
82
the mobile phase, which changes according to the size of the stationary phase carrier
83
particles. Thus, lower particle size permits an increase in separation efficiency, and
84
consequently, conventional HPLC separation time can be reduced, or resolution between
85
analytes increased. This has beneficial ramifications for the field of chemical residue
86
analysis, where frequently, analytical methods test for analytes that are similar in
87
chemical nature.
ed
pt
Ac ce
88
M
77
89
For example, the epimeric compounds dexamethasone and betamethasone (Fig. 1) are
90
licensed for therapeutic use against inflammation in bovine animals in the European
91
Union (although use as growth promoters is banned). Chromatographic resolution of
92
these compounds can be difficult to achieve, due to their structural similarity [1].
93
Separation of these compounds has been performed using a special column containing
94
Hypercarb™ porous graphite [2]. Applying UHPLC conditions to this separation,
4
Page 5 of 47
Deceuninck and co-workers [3] attained resolution of the dexamethasone/betamethasone
96
pair using an Acquity™ BEH C18 column (1.7 m, 2.1 x 100 mm), in a method validated
97
according to Commission Decision 2002/657/EC [1]. The same column chemistry was
98
employed by Touber et al. in their earlier work [4] on developing a multi-residue method
99
for corticosteroids, including dexamethasone and betamethasone. Chromatographic
100
resolution of dexamethasone and betamethasone is only one example of where the power
101
of UHPLC has been successfully applied to solve challenges in multi-residue analyses.
102
Elsewhere, it has been shown that UHPLC-MS/MS can be used to separate and identify
103
as many as 107 pesticide compounds in strawberries, using a run time of just 6.5 min [5].
104
Further examples of successful application of UHPLC-MS/MS in food contaminant
105
applications will be discussed in the review.
us
cr
ip t
95
107
an
106 Fig. 1
M
108
The number of published methods exploiting UHPLC-MS/MS continues to grow each
110
year as the technology matures (see Fig. 2). However, there remain outstanding
111
challenges to be addressed. The reduced sample preparation approach which has become
112
standard in analytical strategies (e.g. QuEChERS) poses questions as to what extent the
113
matrix effect influences the outcome of analyses, and whether generic clean-up methods
114
are necessarily compatible with MS detection. The application of UHPLC-type separation
115
can also lead to chromatographic peaks which have a width of 1 – 3 s. The analyst must
116
then examine whether the data acquisition rate is appropriate for compliance with quality
117
control requirements, especially in applications where analytes co-elute or elute within a
118
time segment where data is collected for several analytes. The application of high
119
pressure to separations that is the core of UHPLC technology can also result in the
120
creation of “frictional heating” effects, with possible ramifications for chromatographic
121
methods. This review will examine these aspects of UHPLC-MS/MS application,
122
followed by a detailed examination of how UHPLC-MS/MS has been exploited in
123
various food contaminant analyses, through an overview of work published in the field,
124
as well as work of general relevance which is of importance to method development
125
using this new technology. However, we first describe the emergence and development of
Ac ce
pt
ed
109
5
Page 6 of 47
126
UHPLC, after two decades of the predominance of HPLC in fields such as veterinary
127
drug residue analysis, environmental analysis and pesticide residue testing.
128 Fig. 2.
130 2. The introduction of UHPLC and its impact on residue analysis
cr
131
ip t
129
us
132
The utility of very high pressures (in combination with small particles) in improving
134
chromatographic efficiency, as well as reducing analysis time, was first explored by
135
Jorgenson and colleagues in the 1990’s [6], [7]. Soon after this work, the first commercial
136
UHPLC systems came on to the market in 2004, along with Agilent RRHT columns (1.8
137
m particle size) and Waters Acquity BEH columns (1.7 m). The successful uptake of
138
this technology was attributed to the benefits associated with the sub-2 m column
139
technology. In order to utilise sub-2 m columns, new pumping technology had to be
140
developed, which could pump mobile phase at pressures up to 1000 Bar and above. This
141
represented a major improvement in pumping technology, as until then LC systems were
142
rated up to approximately 400 Bar and separations were typically run on columns
143
containing 5 m particles at 170 Bar to prolong pump seal lifetimes. As indicated in
144
Table 1, several manufacturers have entered the UHPLC market, producing systems
145
which operate at ever greater pressures and thus providing ever greater chromatographic
146
power.
M
ed
pt
Ac ce
147
an
133
148
Many (UHP)LC-MS/MS applications to date have been in the area of food safety because
149
of the requirement to analyse a wide range of residues down to low ng/g levels. HPLC
150
coupled with UV or fluorescence detection continues to be used in some laboratories, but
151
most groups are progressing to (UHP)LC-MS/MS platforms. Although it is difficult to
152
measure the number of laboratories that are using UHPLC-MS/MS compared to LC-
153
MS/MS without a comprehensive survey of end-users, the FAPAS proficiency test (PT) 6
Page 7 of 47
reports issued by the Food and Environmental Research Agency in the United Kingdom
155
offer an indication of UHPLC-MS/MS implementation among food safety laboratories.
156
In the 2012 anticoccidial PT study 02183, it was shown that 48 laboratories participated.
157
Only two of the 38 laboratories that completed the questionnaire reported that they used
158
UHPLC [8]. Similarly, in the 2012 Nitrofurans PT study 02179, two out of 35
159
laboratories reported that they used UHPLC-MS/MS [9]. More interestingly, a PT study
160
was carried out on avermectins in corned beef (FAPAS Round 02175), where 17
161
laboratories used HPLC-fluorescence and 13 reported the use of LC-MS [10]. This can be
162
ascribed to the sensitivity offered by HPLC-Fluorescence towards avermectin
163
compounds, while analysis by MS poses challenges due to their propensity to form
164
sodium adducts. UHPLC-MS/MS was reported to be used in two laboratories in the
165
avermectins PT study. Cumulatively, these results offer some insight on the adoption of
166
UHPLC in residue laboratories, although reported usage of “LC” as the separation
167
technique is ambiguous, and may be either HPLC or UHPLC. In the future, it is expected
168
that more laboratories will move to UHPLC-MS/MS because the technique can be
169
applied to multiple analytes in shorter run times compared to traditional HPLC-UV or
170
fluorescence systems.
171
MS/MS, less complicated sample preparation procedures can be used. We anticipate
172
additional developments in the area of online SPE combined with UHPLC-MS methods
173
to further improve method reproducibility and speed of analysis. Some examples of such
174
applications have already been published [11,12]. The peak broadening inherent in such
175
methods may be addressed by using advanced chromatographic techniques such as the
176
solvent plug injection technique described by Gode et al.[13]. Moreover, the application
177
of UHPLC-MS/MS can reduce solvent waste by a factor of 10 and increase the sample
178
throughput by a factor of 5 [14]. Typically, the cost per analyte can be reduced by a factor
179
of 15 and can be further reduced by including additional analytes. In the future, it is
180
expected that UHPLC will be implemented in more residue control laboratories in order
181
to reduce cost and hazardous waste. The many advantages offered by UHPLC are thus
182
expected to result in widespread use of the technology in the all areas of chemical residue
183
analysis.
M
an
us
cr
ip t
154
Ac ce
pt
ed
In addition, because of the increased selectivity of UHPLC-
7
Page 8 of 47
184 185
3. Coupling UHPLC with MS/MS detection
ip t
186 UHPLC coupled with MS/MS is becoming an increasingly used instrumental set-up for
188
researchers due to the advantages such as increased sensitivity, selectivity, resolution and
189
throughput compared with conventional HPLC instruments. The sub-2 µm particles used
190
in UHPLC columns offer increased efficiency and peak capacity. Peaks can be as narrow
191
as 1-2 s wide and coupling with modern, fast scanning mass spectrometers allows for the
192
detection of hundreds of analytes in a short retention window, even in complicated
193
biological matrices. Modern MS/MS analysers are capable of full scan acquisition rates
194
of up to 10,000 m/z per s, dwell times of 1 ms and polarity switching in 30 ms or less
195
(Table 2). Although the focus of this review is the combination of UHPLC technology
196
with tandem mass spectrometers, it is important to note that high resolution mass
197
spectrometry (HRMS) continues to evolve as a powerful alternative to MS/MS detection.
198
A careful study by Kaufmann and colleagues, comparing LC-MS/MS and LC-HRMS for
199
the analysis of seven veterinary drugs in a variety of matrices, showed that once HRMS
200
resolution attains 50,000 full width at half maximum (FWHM), selectivity for the two
201
techniques is comparable [15]. MS/MS has been widely adopted in the food analysis and
202
pharmaceutical industries, due to advantages such as its robust tolerance of different
203
matrices [16], but HRMS retains distinct advantages such as the ability to retrospectively
204
analyse historical data for non-preselected compounds [17]. Kaufmann [15] also noted
205
that at very high resolution, HRMS could identify a matrix component which behaved
206
similarly to a banned nitroimidazole compound, which in MS/MS analysis produced a
207
false positive. HRMS detection may eventually be the dominant detection method in
208
pesticide testing in fruits and vegetables, mainly due to its potential ability to identify
209
unknown chemicals [18].
Ac ce
pt
ed
M
an
us
cr
187
210 211
Table 2.
212
8
Page 9 of 47
213
3.1. Acquisition rate and method sectoring
214 215 The acquisition rate in SIM (selected ion monitoring) and SRM (selected reaction
217
monitoring) modes is determined by the switching time and the dwell time per SIM or
218
SRM channel. Dwell time is the amount of time spent collecting a data point at a set
219
transition or peak before switching to the next value in the SIM or SRM method; this
220
dwell time is set by the analyst. If necessary, dwell time can be increased to collect fewer
221
data points. Conversely, if a paucity of data points is collected, dwell time may be
222
decreased to improve peak definition. Dwell time represents a compromise between
223
signal-to-noise (S/N) ratio and the number of data points acquired across a
224
chromatographic peak.
225
sensitivity, as less time is required to acquire adequate mass spectral data. Background
226
noise is decreased as a result of the increased peak height of the analyte in UHPLC, and
227
its increased chromatographic resolving power may offer a reduction in ion suppression,
228
also leading to an increase in the S/N ratio compared with HPLC.
cr
us
M
an
Generally, decreased acquisition time results in increased
ed
229
ip t
216
Time-sectoring is another important factor when setting up UHPLC-MS/MS methods. If
231
the S/N ratio is poor for an analyte, it will require an increase in dwell time to increase
232
the S/N ratio. However, increasing dwell time may result in the detection of too few data
233
points across the peak, due to insufficient acquisition or cycle time. If this is the case, the
234
number of transitions within a SIM or SRM should be reduced to increase the cycle time
235
and obtain more points across the chromatographic peak. Alternatively, if the analytes are
236
sufficiently separated, the SIM or SRM channel can be time-sectored to increase the
237
cycle time for that analyte and acquire sufficient number of points across that
238
chromatographic peak. Time-sectoring also places less pressure on the instrument and
239
allows more time for the raw data to be processed [19].
Ac ce
pt
230
240 241
As well as sectoring the method and setting adequate dwell times, the inter-channel and
242
inter-scan delays are important. Inter-scan delay must be set to leave enough time for the
243
instrument electronics to switch from one transition to the next within an SIM or SRM
9
Page 10 of 47
and inter-channel delay to leave enough time between successive SIM or SRM channels.
245
If the method contains analytes that need different ionisation polarities, extra time is
246
required for the electronics to switch between one channel and the next, as well as change
247
polarities. Typically with newer MS/MS instruments, inter-scan and inter-channel delays
248
are set to 5 ms when switching between successive channels and 20 ms is required for
249
inter-channel delay when switching from positive ionisation to negative ionisation mode
250
between successive channels.
ip t
244
252
cr
251 3.2 UHPLC-MS/MS vs. HPLC-MS/MS
us
253
UHPLC-MS/MS yields many advantages over HPLC and HPLC-MS/MS. It is more
255
environmentally friendly, as it reduces the volume of chemicals used and waste produced.
256
It also offers increased sensitivity, resolution, S/N ratio and shorter injection cycle times.
257
UHPLC-MS/MS is capable of detecting analytes at a lower limit of quantitation (LOQ)
258
due to the increased resolution of the system. Yu et al. demonstrated the advantages of
259
UHPLC-MS/MS approach over HPLC-MS/MS in a quantitative analytical method for
260
five drug compounds in rat plasma. A shorter run time and narrower peaks were achieved
261
with UHPLC, which lowered the LOQ and increased the sensitivity, as well as
262
considerably improving accuracy and precision compared with conventional HPLC [20].
263
The authors proposed that this was due to a narrow band of analyte entering the mass
264
spectrometer, resulting in increased concentration of analyte and in turn increased
265
ratios/N. The throughput of samples is dramatically increased for quantitative analysis
266
due to the short separation time of the UHPLC coupled with the faster acquisition
267
capability of MS/MS. Recent examples of high throughput methods include that of
268
Whelan et al., who developed a multi-residue method for the determination of 38
269
anthelmintic drug residues in milk (as well as 20 internal standards) using UHPLC-
270
MS/MS with rapid polarity switching; all analytes were eluted within 8 min. [21].
271
Ventura et al. detected 34 forbidden drugs and metabolites in a total run time of 5 min by
272
UHPLC-MS/MS [22]. A comparison of UHPLC-MS/MS and HPLC-MS/MS methods for
273
the determination of pesticide residues in baby foods demonstrated that the UHPLC
274
approach was 2.5 times faster than the corresponding HPLC method [23]. Aside from the
Ac ce
pt
ed
M
an
254
10
Page 11 of 47
clear benefit of shorter run times, it has also been noted that use of UHPLC in place of
276
HPLC can also improve method performance via reducing matrix effects, such as in the
277
study of pharmaceutical drugs in surface water described by Van de Steene and Lambert
278
[24]. Here, the reduction in observed matrix effect permitted the authors to replace
279
standard addition with a simpler internal standardisation approach for analyte
280
quantification. Other examples of methods that have been improved by use of UHPLC
281
have been well described in a recent review by Guillarme et al.[25].
cr
282
ip t
275
A major consideration in selecting UHPLC in place of conventional HPLC is that the
284
higher throughput obtained from using UHPLC coupled to MS/MS increases the
285
requirement to perform regular preventative maintenance on the instrument to ensure
286
continued optimal performance. This is especially important when working with more
287
complex biological samples; if the orifice of the MS/MS is not routinely cleaned,
288
accumulated residue on the front end could potentially enter the quadrupoles, which
289
would lead to reduction in sensitivity or the drop-off of less-intense analytes. A full
290
system clean would then be required by engineers to regain sensitivity; this may not be
291
covered under service contracts and will also result in unnecessary down-time of the
292
instrument due to negligence. If the instrument is analysing a large number of samples
293
daily, the sample cone of the MS/MS should be cleaned daily at the end of the sequence,
294
although it should be noted that UHPLC methods typically involve lower injection
295
volumes than HPLC methods.
an
M
ed
pt
Ac ce
296
us
283
297
However, there are some disadvantages to using UHPLC-MS/MS over HPLC-MS/MS.
298
For one, a significant disadvantage is the high cost of the sub-2 µm columns. Also, there
299
is a much smaller selection of UHPLC column chemistries available - although the range
300
is steadily increasing each year, this could prevent a straight transfer from HPLC to
301
UHPLC-MS/MS for some methods.
302
volumes inherent in UHPLC systems have practical implications. For example, samples
303
must be filtered through 0.2 m filters prior to use, as solid material may easily
304
precipitate out of solution at column frits, especially when a steep gradient is employed.
305
Mobile phases may also require filtration, to exclude particulate matter, although care
Furthermore, the narrow plumbing and small
11
Page 12 of 47
must be taken to avoid introducing additional contamination from filtration glassware,
307
filters, etc. Although the instrumentation can be capital-intensive initially, increased
308
market competition is reducing prices, while gains in productivity can offset high
309
instrument costs [26]. However, the benefits far outweigh the disadvantages, e.g. the
310
columns may cost more than traditional HPLC columns but the increased throughput
311
obtained counteracts the increased column cost.
312 4. Challenges inherent in UHPLC-MS/MS techniques
us
313
an
314 315
cr
ip t
306
4.1 Frictional Heating Effects
M
316
As observed by Churchwell et al. [27], UHPLC offers improvements in both the duration
318
of a separation and in sensitivity – although this is compound dependent and not
319
necessarily straightforward. One of the issues which may present difficulty in transferring
320
a method from HPLC to UHPLC is frictional heating [28]. This is a physical
321
phenomenon whereby the interaction between the mobile phase and the stationary phase
322
creates heat. This occurs with liquid chromatography in general, but is far more
323
pronounced with UHPLC. Potentially, this phenomenon could mean molecules travelling
324
in a band through the column experience greater heat at the centre of the column than at
325
the walls, with the net result being band broadening (fig. 3).
327
pt
Ac ce
326
ed
317
Fig. 3.
328
12
Page 13 of 47
The effect of the heat can also have an impact on method performance – de Villiers and
330
co-workers determined that when a radial temperature gradient is prevalent within a
331
column, the efficiency may be seriously affected, particularly with (relatively) larger
332
columns (internal diameter (ID) = 2.1 mm). However, in their work, it was observed that
333
use of a still-air column heater meant that a longitudinal temperature gradient was
334
prevalent along the length of the column, with no considerable impact on separation
335
efficiency. For routine use, the implication is that pre-heating of mobile phase and use of
336
a column oven are essential in controlling heat flow. Work by Nováková and colleagues
337
[29], using a selection of pharmaceutically relevant molecules, investigated the
338
importance of the frictional heating effect for method selectivity. Using a column of ID =
339
2.1 mm, a pressure drop of 1000 bar over the column resulted in an observed temperature
340
increase of 16°C. Changes in retention time were not considered significant however,
341
with changes of 5-10% observed for most of the compounds analysed in the range 30°C –
342
60°C. The effect was more pronounced for acidic compounds, with changes of 25% for
343
ascorbic acid and 41% gallic acid. The authors commented that the effects of frictional
344
heating may be mitigated through several means, including use of narrow-bore columns;
345
use of several shorter columns in place of one long column; a decrease in column oven
346
temperature in UHPLC when transferred from HPLC; or modification to the method
347
gradient or mobile phase pH to compensate for changes in resolution between HPLC and
348
UHPLC. In a study comparing the efficiency of sub-2 micron porous and sub-3 micron
349
shell particles in reverse phase separations, McCalley noted that when using 100%
350
acetonitrile with sub-2 micron columns, the detrimental effects of frictional heating are
351
enhanced [30]. However, aqueous content in the mobile phase was found to moderate
352
this effect, due to greater thermal conductivity. For larger (3 – 5 micron) particles, the
353
effects of frictional heating were also significant [31]. The authors observed that the
354
impact of heating effects were proportional to the capacity factor of a compound and the
355
column inner diameter, but inversely proportional to column particle size and the set
356
column oven temperature.
Ac ce
pt
ed
M
an
us
cr
ip t
329
357 358
4.2 Band Broadening Effects 13
Page 14 of 47
359 The use of UHPLC-MS/MS technology has afforded, among other advantages, a
361
particular benefit when it comes to sample injection volumes. The greater sensitivity
362
achievable with MS/MS combined with the chromatographic power of UHPLC
363
technology means that the analyst may use smaller injection volumes than was previously
364
required, to attain acceptable detection limits. The small diameter of the columns
365
typically employed in UHPLC means that a small injection volume is required, to avoid
366
possible column overload. Smaller quantities of matrix are thus loaded on to the column
367
than with standard HPLC, meaning matrix suppression effects are minimised. However,
368
to ensure good chromatographic performance, the system must contain the absolute
369
minimum amount of dead volume to minimise extra-column broadening effects. As
370
system volume is low, extra-column volume can represent a large percentage of overall
371
system volume. Consequently, extra-column volume becomes a significant source of
372
band broadening. As noted by McCalley [32], the influence of dead volume has an
373
inversely proportional relationship to column internal diameter. Dead volume effects
374
have been observed to be more detrimental to performance in narrow bore (2.1 mm)
375
columns employed with a UHPLC system than with a standard bore (4.6 mm) column
376
operated on a typical HPLC system [30].
cr
us
an
M
ed
pt Ac ce
377
ip t
360
378
As well as tubing diameter, connectors for tubing must be carefully selected so as to
379
avoid contributing extra-column volume. The narrow peaks found in UHPLC analyses
380
mean that band broadening can have a large effect on resolution and overall performance.
381
Gritti and Guichon [33] have observed that for some modern high-pressure LC systems,
382
the largest contributing factor to band broadening was the inner diameter of the needle
383
seat capillary tube (along with detector cell volume). One possible path to reducing extra-
384
column band broadening due to sample injection is described in detail by Sanchez et al.
385
[34]. Their use of a “performance optimising injection sequence” (POISe) allowed the
386
reduction or elimination of the effect of sample injection on the chromatographic
387
performance. This method involves the injection of a defined amount of a weak solvent
14
Page 15 of 47
(relative to the mobile phase) along with the sample. This has the effect of minimising the
389
extra-column dispersion taking place ahead of the analytical column, while having the
390
advantage of not requiring any physical modification to the chromatographic system.
391
Analyte bands are compressed at the top of the analytical column, in proportion to their
392
retention factor. This is referred to as isocratic focussing. It was found that band
393
compression could be achieved in both isocratic and gradient elution modes.
ip t
388
4.3 Narrow Analyte Peaks
us
395
cr
394
an
396
The very short dwell times possible with modern MS instruments (as low as 1 ms, as
398
indicated in Table 2) mean that many analytes can be rapidly detected, e.g. 55 anabolic
399
and androgenic steroids in a 5-min injection cycle [35]. Software-based optimisation of
400
dwell times, cycle times and multiple reaction monitoring windows (dynamic multiple
401
reaction monitoring) has permitted the analysis of up to 11 pesticides which elute within
402
a one-min time window [36]. However, as well as possible peak overlap,
403
chromatographic systems may also produce peaks of very narrow width, particularly in
404
UHPLC [37]. Improvements in data-handling capabilities permit the collection of
405
detector data from compounds where the UHPLC peak profile is only about 2 s across. A
406
multi-residue qualitative MS/MS method for the identification of drugs of abuse in urine
407
by Maquille and co-workers [38] clearly indicates the challenge this can pose. The
408
application of a small-particle (1.7 m) column and high pressures caused all compounds
409
to elute within approximately 2 min. However, as all compounds exhibited narrow peaks
410
(a few seconds), the authors employed rapid data collection parameters – selected
411
reaction monitoring dwell-time was set at 5 ms, while inter-channel delay was also set to
412
5 ms. Time-segmenting of the compounds according to their retention time was also
413
essential to achieving adequate numbers of points across peaks. Although not defined in
414
the relevant legislation, a general reference for peak definition is that a minimum of 5-6
415
data points should be achieved per peak, to ensure reliable integration [39]. Martínez
Ac ce
pt
ed
M
397
15
Page 16 of 47
Vidal [40] et al. exploited an alternative means of data collection in their work on
417
determination of veterinary drug residues in milk, which permits rapid screening of
418
samples. A MS/MS screening method was designed based on the selection of a neutral
419
loss or product ion which is selective for a family of compounds. If a positive sample is
420
detected, the sample can then be retested using a confirmatory (MRM) method. The run
421
time for this particular screening method was 3 min, and 21 veterinary drug residues were
422
included in its scope.
cr
ip t
416
us
423
Kovalczuk et al. have presented a comprehensive study of the chromatographic output of
425
a UHPLC-MS/MS multi-residue analysis [41]. This represents one of the few examples
426
of multi-residue methods where the authors have assessed chromatographic parameters
427
such as peak capacity, peak symmetry and height of theoretical plates. Direct comparison
428
of HPLC and UHPLC analysis (with MS/MS detection) of a multi-pesticide mixture
429
highlighted some of the key chromatographic considerations when selecting a UHPLC
430
approach over a HPLC approach. Peak compression on the UHPLC system meant that
431
S/N at low concentrations improved, with a consequent reduction in the method limit of
432
quantification (LOQ) versus the HPLC method. For the compound tebuconazole (the
433
matrix analysed was apple), S/N improved by a factor of 10 going from HPLC to
434
UHPLC, while for several compounds, the LOQ was as much as four times lower using
435
UHPLC. Peak capacity was comparable in both techniques under optimised conditions,
436
although it is noted that UHPLC offers shorter analysis time (and reduced solvent
437
consumption). Height of theoretical plates was greater in UHPLC, but with significant
438
variability.
440
M
ed
pt
Ac ce
439
an
424
4.4. Matrix effects
441 442
Matrix effects are a common phenomenon when using MS for chromatographic
443
detection, as well as for screening methods. These effects can result from a number of 16
Page 17 of 47
factors, generally the presence or co-elution of interfering compounds, resulting in
445
enhancement or suppression of the analyte of interest. Interfering analytes or compounds
446
can be extracted from the original sample matrix or from extraction equipment such as
447
plastic tubing. The presence of interfering analytes along with the analyte of interest in
448
the ionisation source increases the competition for ionisation and typically reduces ion
449
efficiency of the analyte. The effect of ion suppression in residue analysis applications
450
has been investigated in detail by Antignac and colleagues [42]. Using a method for the
451
analysis of the beta-agonist isoproterenol in bovine meat samples, it was illustrated how a
452
polar analyte that eluted early in a typical reverse-phase separation may potentially co-
453
elute with matrix interferences which come with the solvent front. The signal for this
454
compound was shown to reduce by as much as 95% from the anticipated response due to
455
matrix interference, indicating how severe the effects may be. The choice of ionisation
456
technique can be a major factor in eventual matrix effects in an analysis – it has been
457
observed that atmospheric pressure chemical ionisation (APCI) is, in some cases, less
458
prone to matrix effects than electrospray ionisation (ESI) [43], although ESI is more
459
widely applicable. This has been attributed to the differences in ionisation techniques, in
460
that using ESI, the charged analyte is formed in the liquid phase, while with APCI, the
461
analyte is transferred into the gas phase in the neutral form before ionisation occurs [44].
462
Several possible mechanisms that lead to matrix effects in ESI are described by Trufelli
463
et al. [44]. High analyte concentration (>10-5 M) can lead to a loss of sensitivity, possibly
464
due to saturation of the electrospray ionisation droplet surface with analyte. Other
465
possible causes include the formation of analyte-inclusion particles from non-volatile
466
additives and analytes of interest, with resultant hindrance of ionisation. It has also been
467
postulated that the presence of non-volatile interferents affects the viscosity and surface
468
tension of ESI droplets, which thus affects ionisation efficiency and the eventual signal
469
recorded [45]. Hydrophilic analytes may also be significantly suppressed in the presence
470
of surfactants in the sample matrix. The presence of lipophilic components in the sample
471
matrix is another factor which can have a large impact on method response, as observed
472
in a method for the analysis of veterinary drug residues in meat-based baby food [46].
473
Another known issue in the LC-MS/MS analysis of samples of biological origin is the
474
effect of phospholipids on ionisation efficiency [47]; these compounds can cause
Ac ce
pt
ed
M
an
us
cr
ip t
444
17
Page 18 of 47
significant matrix ionization effects in positive electrospray ionisation mode.
476
Phospholipids represent a main constituent of cell membranes. These compounds can
477
present very broad elution profiles, requiring the analyst to either monitor these
478
compounds chromatographically to ensure they do not co-elute with analytes [48], or
479
employ advanced sample preparation techniques to remove the interfering compounds
480
[49].
ip t
475
cr
481
Highly selective sample preparation techniques are important for removing compounds
483
from the matrix that produce matrix effects – although more elaborate sample preparation
484
procedures do have a bearing on sample throughput, and are more costly. Chambers et
485
al. carried out a systematic and comprehensive strategy for reducing matrix effects in LC-
486
MS/MS as applied in bioanalytical assays. A paired t-test illustrated a statistically
487
significant improvement in matrix effects using UHPLC when compared with traditional
488
HPLC [50]. The evolution of UHPLC-MS/MS, which has permitted the analysis of
489
multiple drug residues from multiple classes, as well as the development of fast generic
490
clean-up and extraction techniques, means that highly selective sample preparation
491
techniques are no longer critical. A consequence of employing these generic methods is
492
that matrix effects can have a large bearing on method performance. This is particularly
493
the case when using solvent standard calibration curves, and where possible, matrix
494
matched calibration curves offer a more accurate means of quantification. Alternatively,
495
stable isotopically labelled standards can be employed to negate matrix effects to improve
496
both accuracy and precision. This is demonstrated in work by Varga et al. [51], on the
497
determination of various mycotoxins in maize, where the use of
498
analogues permitted the development of the first multi-target method capable of
499
determining all mycotoxins regulated in the EU for maize.
an
M
ed
pt
Ac ce
500
us
482
13
C-labelled mycotoxin
501
The likelihood of matrix effects is high with the analysis of samples of biological origin,
502
such as food of animal origin in veterinary drug residue analysis. These effects may be
503
somewhat mitigated when using UHPLC - as mentioned in section 4.2, UHPLC methods
504
require smaller injection volume than comparable HPLC methods, to avoid overloading
18
Page 19 of 47
the column. A major advantage of this reduced injection size is that matrix effects are
506
reduced, as smaller quantities of matrix are injected onto the source. This, in turn, leads to
507
less competition for ionisation in the ESI droplets, increased resolution between the
508
analytes of interest, and a much cleaner mass spectrum compared to that of HPLC-
509
MS/MS. The smaller quantity of matrix passing through the system means less
510
accumulation of non-volatile components and other contaminants on the sample cone
511
ahead of the quadrupoles, as well as on the quadrupoles themselves. Such accumulations
512
can lead to charging effects with a resultant loss of sensitivity, so this is a further benefit
513
of UHPLC interfaced with MS/MS. Given the extreme impact on analyte signal which is
514
possible with matrix effects, quantification/ evaluation of matrix effects in a given
515
analysis can provide the analyst with essential detail in ensuring consistent method
516
performance. This may be assessed either through spiking of samples with analyte
517
following extraction, or through infusion of analyte post-column into the mass
518
spectrometer detector system, and examining the effect on signal when a negative control
519
sample is injected (see Fig. 4). The possibility of matrix interference with analyte signal
520
is one reason why validation procedures (such as those indicated in 2002/657/EC, [1])
521
require that a method be validated for each compound in a multi-analyte method, rather
522
than selecting one representative compound.
ed
M
an
us
cr
ip t
505
pt
523
525 526 527
Ac ce
524
Fig. 4. Adapted from [42] with permission from Elsevier.
4.5 Column Blockage
528 529
Due to the narrow plumbing employed in UHPLC, a high level of chromatographic
530
hygiene is required to avoid problems with blockages. In addition, sample matrix and
19
Page 20 of 47
analytes can contribute to column blockage. In an application for detection of steroid
532
esters, columns were susceptible to blocking when solvent standards were injected onto
533
the column in a gradient that started with a high aqueous mobile phase, going to high
534
organic content [52]. It was proposed that the waxy ester substances crashed out of
535
solution, when mixed with the mobile phase during the injection cycle. The solution to
536
the problem was simple; by elevating column temperature to 60°C, and including
537
isopropyl alcohol in mobile phase and needle washes, the problem was circumvented. In
538
the authors’ laboratory, a number of methods have been developed using high
539
temperature [14,21,53,54]. However, increasing temperature may not be suitable for all
540
substances, e.g. tetracycline antibiotics have to be separated at 40°C or less, as they can
541
form epimers [55].
an
us
cr
ip t
531
542
5. Recent applications of UHPLC-MS technology in the analysis of agrochemical
544
residues and mycotoxins.
ed
545
M
543
As noted in the introduction, there has been significant uptake of UHPLC-MS/MS
547
technology in areas of analytical research such as pesticide testing, veterinary drug
548
residue analysis and environmental contaminant detection, and Fig. 2 clearly indicates
549
how widely adopted the technique has become. In this section, an overview of some of
550
these published applications is given. Methods have been published for many classes of
551
compound; these may be broadly grouped into substances with an established permitted
552
limit [e.g. maximum residue limits (MRLs)], and substances which are prohibited. In
553
veterinary drug residue testing, Council Directive 96/23/EC [56] lists prohibited
554
compounds as Group A substances, covering stilbenes, antithyroid agents, steroids,
555
resorcylic acid lactones, beta-agonists and banned veterinary drugs. Other substances
556
with a permitted limit are found in Group B substances (anthelmintics, anticoccidials,
557
nitroimidazoles, carbamates, pyrethroids, sedatives, non-steroidal anti-inflammatory
558
drugs, sulphonamides, quinolones and others). Example UHPLC-MS/MS methods for
Ac ce
pt
546
20
Page 21 of 47
different veterinary drug classes are summarised in table 3. For pesticides, there are 1289
560
active substances that are controlled under EU regulation 1107/2009 [57]. A total of 511
561
substances have been given MRLs under EU regulation 396/2005 [58,59], while for
562
pesticides without an established MRL, a default limit of 0.01 mg kg-1 is set [60]. The
563
sheer diversity of pesticides in use underlines the need for a robust, versatile analytical
564
approach, as offered by UHPLC-MS techniques. Examples of such methods are described
565
in the following sections.
cr
ip t
559
us
566 567
pt
ed
M
an
Table 3
Ac ce
568
21
Page 22 of 47
ip t cr
5.1 UHPLC-MS/MS methods for Prohibited Substances
572
Analysis for banned compounds represents an area of continued research focus, as these are typically chemical entities which pose a direct
573
risk of harm to human health. UHPLC-MS/MS presents a very attractive option for the analysis of complex biological samples for residual
574
quantities of zero-tolerance substances. This can be attributed to the inherent resolving power, permitting resolution of chemically similar
575
compounds, and sensitivity. This is crucial in areas such as environmental analysis, where the ability to specifically detect small polar
576
organic molecules within complex sample matrices poses a particular challenge [61]. In the area of veterinary drug analysis, UHPLC-
577
MS/MS analytical power has been applied to most banned substances including thyreostats [62], as well as stilbenes and resorcylic acid
578
lactones [63]. The bulk of published work, however, has focussed on two main sub-categories of Group A substances, namely the steroids
579
and -agonists.
M
d
ep te
580
an
571
us
570
Steroid hormones are banned from use in food-producing animals, and fall into three distinct groups: estrogens, gestagens and androgens
582
(EGAs) [64]. Exploitation of these compounds for the purposes of growth promotion in food-producing animals is strictly prohibited within
583
the European Union by Directive 96/22/EC [65]. Application of UHPLC-MS/MS to determine the presence of these compounds in bovine
584
hair has been described by Duffy et al. [52], who found that the method performed comparably to the more conventionally-employed GC-
585
MS/MS methods. In their method for the detection of boldenone (an androgen) and metabolites in calf urine, Van Poucke et al. [66] used a
586
UHPLC-MS/MS, with a short column (50mm) containing particles of 1.7 m diameter. Gradient optimisation permitted the polar analyte
587
6-hydroxy--boldenone to be separated from the solvent front. The authors noted that the limiting factor in speeding up their method
Ac c
581
22
Page 23 of 47
beyond the 5-min injection cycle achieved was the number of data points being acquired
589
for peaks. Covering a broader range of compounds, Malone and colleagues analysed for
590
13 synthetic growth promoters in bovine muscle, applying a 1.8 m particle-size column
591
(2.1 x 50 mm) and achieving separation of all compounds in 14 min [67]. Wang et al.
592
described a modified QuEChERS (Quick, Easy, Cheap, Effective, Rugged, Safe)
593
extraction procedure to extract nine estrogens from milk powder. Two UHPLC C18
594
columns were evaluated, the first was an ethylene bridged hybrid (BEH) column, and the
595
second was a HSS T3 column. Results from this study showed that the BEH-C18 column
596
was not suitable as it was not capable of separating all nine estrogens [68], but the HSS
597
T3 column was successfully employed, using an appropriate gradient.
598
extracted 10 anabolic steroids (AASs) from equine plasma using methyl tert-butyl ether
599
(MTBE) prior to separation on a C18 reversed-phase column [69]. This extraction
600
procedure was again employed by Guan et al., who developed a method to isolate 55
601
anabolic and androgenic steroids in equine plasma. Separation was performed using a
602
Hypersil Gold™ C18 column. Amongst the 55 AASs were epimer pairs (testosterone and
603
epitestosterone, boldenone and epiboldenone, nanadrolone and epinandrolone) which
604
were chromatographically base-line separated. This high-throughput method had a cycle
605
time of 5 min per sample [35]. Finally, Wong and colleagues [70] described a UHPLC-
606
MS/MS method for the analysis of 140 compounds in equine urine, with an injection
607
cycle time of 13 min – including anabolic steroids. The authors noted how a combination
608
of fast chromatography and rapid polarity switching in MS detection have allowed them
609
to condense what was formerly four individual test methods into one comprehensive
610
method.
cr
us
pt
ed
M
an
You et al.
Ac ce
611
ip t
588
612
Also falling into the category of banned (Group A) substances, β-agonists are synthetic
613
derivatives of catecholamines, used at lower quantities as tocolytics and bronchodilators
614
in veterinary medicine [71]. However, for the last 20 years there has been frequent
615
reported misuse as growth-promoting agents in food producing animals [72]. These
616
compounds may be divided into two groups: the substituted anilines, including
617
clenbuterol, and the substituted phenols, which include salbutamol [64]. Yang et al. 23
Page 24 of 47
developed a method capable of detecting 13 β-agonists in milk. Ion suppression was
619
overcome by adding perchloric acid during the sample extraction step to precipitate
620
protein out of the samples [72]. Separation was carried out using an Acquity™ BEH C18
621
column. Shao et al. isolated 16 β-agonists from porcine tissue using enzymatic
622
hydrolysis. Maximum sensitivity in this method was achieved by optimising the solvent
623
type and pH of the mobile phase. Methanol-water containing 0.1% formic acid provided
624
good separation and resolution when compared to a comparable acetonitrile-water mobile
625
phase [73]. A similar method was used to determine 19 -blockers and 11 sedatives in
626
animal tissue, with an injection cycle time of 14 min [74]. Nielen et al. developed a
627
method for 22 β-agonists, which was sufficiently versatile to be applied to the analysis of
628
bovine and porcine urine, feed and hair using a mixed-mode solid-phase extraction
629
procedure. The method was performed by conventional LC-MS/MS and UHPLC-
630
MS/MS. Due to the diverse structure of β-agonists, developing a single analytical method
631
which encompasses multiple compounds is challenging. This issue was overcome by
632
performing two injections of each sample using the same LC gradient whilst acquiring a
633
different set of ion transitions. Run time was reduced from 20.5 min to less than 15 min
634
using UHPLC-MS/MS.
635
ratio improved with the introduction of a high-resolution column [75].
638
cr
us
an
M
ed
pt
637
Chromatographic resolution, peak shape and signal-to-noise
5.2 UHPLC-MS/MS methods for substances with an established permitted limit
Ac ce
636
ip t
618
639
The presence of a MRL, ‘tolerance’ or recommended limit (RL) for a compound interest
640
has somewhat different implications – methods have a well-defined target concentration
641
for analysis, rather than the criterion of ‘as low as possible’ as with banned substances.
642
An example of a class of veterinary drug which possess MRLs is the quinolone group.
643
These are potent antibacterials used in the therapeutic treatment of infections in both
644
human and veterinary medicine [76]. MRLs to monitor quinolone residues in edible
645
animal tissue are controlled by European Union Directive 2377/90/EC [77]. Lombardo-
24
Page 25 of 47
Agüí et al. described the determination of eight quinolones in bee products. The
647
separation capabilities of three different C18 columns were compared, where the particle
648
size of each varied from 1.7 μm to 1.9 μm. An intermediate partial size of 1.8 µm
649
delivered slightly better separation of two of the quinolones [78]. Zhang et al. extracted
650
22 fluoroquinolones for milk using an EDTA-McIlvaine buffer solution and purified
651
using Bond Elut™ Plexa SPE cartridges. Zhang examined the separation capabilities of
652
four different reversed-phase chromatographic columns which varied in particle size and
653
bonded phase chemistry.
654
observed with a Waters Acquity™ Shield RP18 Column (2.1 mm x 100 mm, 1.7 µm).
655
Recoveries of 61.9 – 94.6 % were observed for 19 of the 22 fluoroquinolones included in
656
the study [76].
ip t
646
an
us
cr
The highest sensitivity and acceptable peak shaped was
657
Sulphonamides represent a class of drugs used to prevent and treat bacterial infections.
659
They are amongst the most widely used veterinary drugs because of their broad-spectrum
660
antimicrobial activity, low cost, and their effectiveness as growth promoters in livestock
661
[79]. The MRL for individual sulphonamides in food of animal origin has been set at 100
662
µg kg-1 by the European Union [77]. Careful selection of gradient conditions by She and
663
co-workers permitted the separation of 24 sulphonamides in bovine milk in 10 min [80].
664
Although C8 and phenyl stationary phases were investigated, an Acquity™ BEH C18
665
column proved most suitable. Zhao et al. employed an immunoaffinity chromatography
666
(IAC) technique to purify muscle and milk extracts containing 12 sulphonamides [79].
667
McDonald et al. developed a rapid method to isolate 19 sulphonamides from muscle
668
matrix using 0.1 M EDTA/acetonitrile mix, with a total injection cycle time of 15 min.
669
Two sets of sulphonamides had similar molecular weights and parent-daughter transitions
670
(tetracycline/ doxycycline and sulfamethoxypyridazine/ sulfamonomethoxine).
671
authors distinguished between these compounds using retention time [81]. Tamošiūnas et
672
al. compared the performance capabilities of LC-MS/MS to UHLC-MS/MS. A simple
673
acetonitrile extraction procedure was carried out to extract ten sulphonamides from egg
674
and honey matrix, followed by SPE clean-up with Strata-X™ cartridges.
675
recoveries (80 – 110 %) were obtained by HPLC and UHPLC-MS/MS determination for
Ac ce
pt
ed
M
658
The
Similar
25
Page 26 of 47
676
all analytes.
However, narrower peaks, and consequently better chromatographic
677
resolution was observed with UHPLC-MS/MS [82].
678
method for isolating 24 sulphonamides in muscle followed by analysis via UHPLC–
679
MS/MS. Three columns (C18 5 µm, C18 1.7 µm, and C8 1.7 µm) were tested for signal
680
intensity and separation efficiency.
681
displayed improved sensitivity and separation efficiency, when compared to the C18 5 µm
682
column [83].
Cai et al. developed a simple
cr
ip t
The columns containing smaller particle size
us
683
A large number of anthelmintic drugs are licensed for the treatment of gastrointestinal
685
nematodes and trematodes in food producing animals [84]. They include nematicides,
686
flukicides, endectocides, benzimidazoles and macrocyclic lactones [14], [21]. The EU,
687
through Council Regulation 2010/37, established MRLs for a limited number of
688
anthelmintics in food of animal origin [85], [21]. Zhang et al. described a procedure for
689
isolating albendazole (ABZ) and three of its metabolites in fish muscle using UHPLC-
690
MS/MS equipped with an electrospray ionisation (ESI) source [86]. The separation
691
capabilities of three columns (BEH C8, BEH C18, and CSH C18) were compared.
692
Identical retention times (BEH C8), poor separation and double-humped peaks (BEH C18)
693
were observed for two of the three columns.
694
profile, the CSH C18 column was selected and used for the study. A study of extraction
695
behaviour was necessary due to the variation in lipophilicity and pKa among ABZ and its
696
metabolites. Optimum peak shape and sensitivity were obtained through optimisation of
697
the reconstitution solution. Subsequent to a series of optimisation experiments where
698
mean recoveries were monitored, ethyl acetate was selected as an extraction solvent.
699
Whelan et al. extracted 38 anthelmintic drugs from milk using a modified QuEChERS-
700
type extraction method. Factorial design experiments were carried to determine the
701
optimum mobile phase composition.
702
ionisation of analytes improved when the mobile phase contained fewer additives.
703
However, it was necessary to add ammonium formate to the mobile phase at a low
704
concentration to improve the sensitivity of several of the anthelmintic drugs [21]. Whelan
705
et al. further optimised this method by including an enzymatic hydrolysis step to
Following optimisation of the gradient
Ac ce
pt
ed
M
an
684
Results from these experiments indicated that
26
Page 27 of 47
investigate the prevalence of nitroxynil and its conjugate forms in lactating dairy cows
707
[87]. Xia et al. describe the rapid chromatographic separation of 13 benzimidazoles in 8
708
min using UHPLC-MS/MS [88]. The analysis time for each injection was 8 min, which
709
was substantially lower than an earlier HPLC-MS/MS method [89]. UHPLC-MS/MS
710
methods for other members of group B2 have also been published, including
711
anticoccidials [90] and non-steroidal anti-inflammatory drugs (NSAIDs) [91].Hu and co-
712
workers have recently described a UHPLC-MS/MS method for the determination of
713
thirty NSAIDs in swine muscle, in a run time of 16 min [92].
cr
us
714
ip t
706
Other possible contaminants and substances of environmental origin are covered within
716
group B3 in Council Directive 96/23/EC [56], such as mycotoxins. As is commonly the
717
case with multi-residue methods, the chemical diversity found within this group of
718
compounds represents a challenge for analysis. However, combined with an effective
719
clean-up technique, UHPLC-MS/MS has provided a means of rapidly and sensitively
720
determining these contaminants. Van Pamel et al. developed a UHPLC-MS/MS method
721
for 26 different mycotoxins in maize silage [93]. The polarity-switching capability of the
722
mass spectrometer system used (a Waters Xevo™ TQ mass spectrometer) permitted
723
detection of positively and negatively ionised analytes within the same run, avoiding the
724
need for two separate injections of each sample and giving a total run time of 9 min.
726 727
M
ed
pt
Ac ce
725
an
715
5.3 Multi-class multi-residue analysis
728
Finally, the expansion of methods to cover ever greater numbers of compounds means
729
that certain methods cannot be simply categorised as being specific towards banned or
730
MRL compounds. We here examine recent methods which are capable of simultaneously
731
determining several classes of analyte. The advantages of such an approach are high
732
sample throughput, simplicity and reduced cost. The bulk of methods reported use
27
Page 28 of 47
tandem mass spectrometry (MS/MS), as multiple transitions can be monitored for
734
compounds simultaneously, providing unambiguous identification. Perhaps one of the
735
most accomplished methods published to date is the generic extraction and analysis
736
procedure published by Mol and colleagues [94]. With careful examination of extraction
737
parameters, as well as matrix effects, existing multi-analyte methods were combined to
738
give a method capable of analysing for 172 pesticides, mycotoxins and plant toxins.
739
Separation and detection of the analytes was performed via UHPLC-MS/MS. Hermann et
740
al also demonstrated how a generic sample extraction procedure combined with UHPLC-
741
MS/MS analysis could be used to test for a range of 108 compounds (out of 127 targeted
742
compounds) related to emergency situations [95]. One of the methods displaying the
743
widest scope of analytes to date is the UHPLC-MS/MS method for analysis of
744
contaminants in milk published by Zhan et al [96].The method covered 255 veterinary
745
drug residues, pesticide residues, mycotoxins and other potential chemical contaminants,
746
using a generic extraction technique based on a two-step precipitation, followed by
747
sample drying and then reconstitution in injection solvent.
M
an
us
cr
ip t
733
ed
748
For screening of large numbers of compounds, time-of-flight (ToF) mass spectrometry
750
has also been successfully applied. Kaufmann et al. developed a method for isolating 100
751
veterinary drug residues from muscle, liver and kidney prior to UHPLC-ToF MS
752
analysis.
753
extraction), a technique which is known for its ability to recover polar, and medium polar
754
compounds.
755
analytes. Evaluation of the method indicated that these polar analytes were adsorbing on
756
proteins which precipitated during the extraction stage. This issue was overcome by
757
rinsing the sample and glassware with DMSO and complexing buffer. Poor recovery of at
758
high spiking concentration was attributed to detector saturation and a mass shift, a
759
phenomenon not observed at lower concentrations [97].
760
modified and improved this procedure to isolate over 100 residues in muscle, kidney,
761
liver, fish and honey using a single stage Orbitrap MS. Co-eluting proteins or peptides
762
resulted in extensive signal suppression effects, which is an issue that had not been
pt
749
Ac ce
Samples were extracted using liquid-liquid-solid extraction (bi-polarity Kaufmann encountered difficulties when attempting to isolate polar
Recently, Kaufmann et al.
28
Page 29 of 47
763
typically encountered when using ToF [18]. In the area of environmental contaminant
764
analysis, UHPLC-ToF-MS has been successfully applied to the screening of organic
765
pollutants in water [98].
ip t
766 Also selecting ToF MS for detection, Ortelli et al. reported a method for extracting 150
768
residues in milk which involved protein precipitation using acetonitrile and ultrafiltration.
769
The group encountered difficulties in the isolation of avermectin residues, where
770
recoveries (≤ 10%) did not fulfil the criteria of validation. This result was attributed to
771
losses during ultrafiltration. A post-column infusion system was used to evaluate matrix
772
effects on response. Signal response was sample-dependent for two of the compounds
773
(febantel and erythromycin), while a significant signal enhancement (1300 %) was
774
observed for enrofloxacin [99]. Peters et al. developed a simple extraction procedure to
775
isolate 100 veterinary drug residues in egg, fish and muscle by extracting with
776
acetonitrile and purifying using Strata-X™ SPE cartridges. Separation was performed
777
using a Waters Acquity™ UPLC BEH C18 column, and detection was via ToF-MS. The
778
authors constructed the method using retention times, specific accurate mass and isotope
779
pattern ratios calculated by SigmaFit™ [100]. Rapid quantification and accurate mass
780
measurement for 166 pesticides in fruits and vegetables was reported by Wang and co-
781
workers [60]. Following sample extraction via a QuEChERS approach, samples were
782
tested using a full-scan MS detection strategy for quantification, with confirmation via a
783
UHPLC-Q-Orbitrap MS/MS data-dependent scan. This represents an efficient means of
784
confirmatory analysis of a large number of pesticides, with the additional benefit of
785
circumventing the need for lengthy compound-dependent parameter optimisation as is
786
necessary with MRM approaches. QuEChERS in particular has proven to be an excellent
787
means of sample preparation for pesticides, providing clean sample extracts for analysis
788
by UHPLC-MS/MS, such as in the recent work described by Arienzo et al on the testing
789
of pesticides in fresh cut vegetables [101].
Ac ce
pt
ed
M
an
us
cr
767
790
29
Page 30 of 47
791
Using the more commonly-employed LC-MS/MS approach, Rezende et al. described a
792
simple extraction procedure for 12 β-lactams and tetracyclines (in bovine muscle) which
793
traditionally rely on time-consuming, expensive SPE approaches. Tetracyclines can be
794
difficult to analyse for, due to the ability of tetracycline to epimerise to 4-epitetracycline
795
at pH 4 - 8.
796
purification with dSPE (dispersive solid phase extraction) C18 and finally defatting with
797
hexane. Rezendes studied the matrix effect using an F-test to examine the
798
homoscedasticity of data and a t-test to compare average results obtained in matrix
799
samples and solvent. Results from these tests suggested there was a matrix effect. To
800
overcome this issue, samples were quantified using matrix-matched calibration curves
801
[102]. Lehotay and co-workers have recently reported a streamlined method designed to
802
detect residues of as many as 60 priority veterinary drugs in bovine kidney, using
803
UHPLC-MS/MS [103], with an injection cycle time of under 10 min. The authors
804
determined that their method was suitable for screening of 54 of the 62 drugs studied,
805
qualitative identification of 50 compounds, and quantification of 30. For bovine muscle
806
[104], a multi-class, multi-residue UHPLC-MS/MS method was designed for the purpose
807
of monitoring over 100 veterinary drug residues. The importance of the choice of the
808
dispersive solid-phase extraction sorbent used was investigated, as was the influence of
809
post-column infusion of ammonium formate on ionisation enhancement for anthelmintic
810
compounds. The final method presented was capable of screening for 113 compounds,
811
identification of 98 compounds, and quantification of 87.
ip t
cr
us
an
M
ed
pt
Ac ce
812
The method involved extraction with water/acetonitrile followed by
813
When MRLs have not been established for compounds, researchers typically attempt to
814
develop analytical methods capable of detecting these compounds at extremely low
815
levels. Aguilera-Luiz et al. encountered this problem when attempting to develop a
816
multiclass method for 39 veterinary drugs in meat-based baby food and powdered milk-
817
based infant formulae [46]. Separation was carried out on a C18 column (100 mm × 2.1
818
mm, 1.7 μm particle size) using a previously developed chromatographic procedure [94].
819
Chromatographic separation was optimised by evaluating two solvents, acetonitrile and
820
methanol.
Retention time decreased with acetonitrile; however sensitivity was 30
Page 31 of 47
821
significantly better when methanol was used in the mobile phase. The authors also
822
compared two additives, formic acid and acetic acid. Ionisation efficiency improved with
823
the addition of formic acid at a concentration of 0.05%.
ip t
824 Matrix effects associated with milk can pose a significant challenge in creating multi-
826
residue methods. Tang et al. described the difficulties encountered when attempting to
827
extract veterinary drugs from milk. A relatively high level of acetonitrile was used to
828
isolate 23 veterinary drugs from milk.
829
matrix-matched calibration curves were employed for accurate quantitation [105]. Vidal
830
et al. developed two screening methods to identify 21 veterinary drug residues in milk.
831
The first method used parent ion and neutral loss scan whereas the second method only
832
used one transition per compound. Both methods were separately validated but it was
833
ultimately decided that the second method was more suitable as it was capable of
834
screening at lower concentration levels [40]. Wang et al. compared the separation of ten
835
antibiotic residues in milk using three kinds of columns (C18, C8 and phenyl). The
836
investigation indicated that satisfactory separation and resolution could only be achieved
837
using a C18 column. Samples were extracted using McIlvaine buffer: methanol (8:2), and
838
purified with HLB SPE cartridges [106].
pt
ed
M
an
us
Matrix effects were quantified in milk, and
Ac ce
839
cr
825
840
The QuEChERS technique has proved suitable when it comes to cleaning up milk
841
samples prior to UHPLC-MS/MS analysis. Aguilera-Luiz et al. developed a method for
842
the analysis of 18 veterinary drugs in milk. Initially, extraction was carried out using
843
acidified acetonitrile followed by clean-up using QuEChERS dSPE. Separation was
844
performed using an Acquity™ UPLC BEH C18 column. Poor recoveries were attributed
845
to the propensity macrolides have to form chelation complexes with cations present in
846
solution. To overcome this problem, EDTA was added to the extractant solution prior to
847
QuEChERS clean-up [107]. Vidal et al. utilised this method to isolate 17 antibiotic
848
residues in honey. To avoid contamination of the column from lipids and waxes present
31
Page 32 of 47
in honey, the organic solvent content was quickly raised during the gradient profile [108].
850
Frenich et al. compared several extraction procedures (solvent extraction, SPE, MSPD,
851
and a modified QuEChERs procedure) to analyse 25 veterinary drugs in eggs. A uniform
852
dwell time of 0.015 s was used for most of the compounds except for levamisole,
853
marbofloxacin and sulfadimidine which had a dwell time of 0.025 s, whereas
854
fenbendazole, emamectin and ivermectin were monitored at 0.100 s. Results suggested
855
that solvent extraction was the most reliable procedure for the simultaneous extraction of
856
several classes of veterinary drugs (tetracyclines, macrolides, quinolones, sulphonamides
857
and anthelmintics), where recoveries ranged from 60 – 119 % [109].
us
cr
ip t
849
858
As previously mentioned, the tetracycline compounds pose particular challenges for
860
multi-compound methods, as very specific analysis conditions are required. Shao et al.
861
described the difficulties encountered when developing an extraction procedure for 21
862
veterinary drugs from two classes (tetracyclines and quinolones).
863
typically extracted solvents acidified at a pH of less than 3. However under these acidic
864
conditions, tetracyclines will reversibly form 4-epi-tetracyclines and anhydro-
865
tetracyclines. An investigation was carried out, where three different extraction solvents
866
(methanol, acetonitrile and EDTA-McIlvaine buffer) were acidified at pH 4 prior to
867
sample clean-up with HLB SPE. Results indicated that acceptable repeatability was
868
achieved with the use of EDTA-McIlvaine (pH 4) [110].
870 871
M
pt
ed
Quinolones are
Ac ce
869
an
859
6. Conclusion
872
The emergence and rapid development of UHPLC-MS/MS technology which we have
873
documented in this review has had a large impact on the area of veterinary drug residue
874
analysis and related compounds. Injection cycle times have shortened, sensitivity has
875
improved, chromatographic resolution of compounds has increased and the number of
32
Page 33 of 47
compounds included in multi-residue methods continues to climb. The combination of
877
these advantages (along with peripheral advantages such as lower solvent consumption)
878
has led to the increasing adoption of UHPLC-MS/MS as the default approach for
879
confirmatory analysis of multiple compounds. In this review, we have examined the
880
technical aspects of coupling UHPLC instrumentation to mass spectrometric detection
881
systems, as well as highlighting some of the inherent challenges that come with the
882
technology, such as frictional heating effects, narrow chromatographic peaks and the
883
impact of sample matrix effects. Despite these challenges, the arenas of veterinary drug
884
residue analysis, pesticide analysis and mycotoxin testing have shown prolific uptake of
885
the technology, and we have detailed some of the most accomplished methods which
886
employ UHPLC-MS/MS as the separation and detection system. Comprehensive methods
887
for some of the more important contaminant groups in residue analysis have been
888
developed for UHPLC-MS/MS, including anthelmintics, -agonists, steroids, quinolones
889
and others. Expected future developments include evolution of methods to include larger
890
numbers of compounds and classes of compound; the use of ever higher temperatures and
891
pressures to create more effective separation methods; and further reduction in sample
892
preparation via online SPE and other techniques, to increase the speed of analysis beyond
893
current standards.
pt
894
ed
M
an
us
cr
ip t
876
Acknowledgements
896
This work was funded by the Food for Health Research Initiative (FHRI) administered by
897
the Irish Department of Agriculture, Food and the Marine and the Health Research Board
898
(Contract 7FHRIAFRC5).
Ac ce
895
899 900 901 902 903
[1] [2] [3]
Off. J. Eur. Union L221 (2002) 8. M. McDonald, K. Granelli, P. Sjoberg, Anal. Chim. Acta 588 (2007) 20. Y. Deceuninck, E. Bichon, F. Monteau, J.P. Antignac, B. Le Bizec, Anal. Chim. Acta 700 (2011) 137.
33
Page 34 of 47
[12] [13] [14] [15] [16] [17] [18] [19] [20] [21]
ip t
cr
[11]
us
[10]
an
[9]
M
[6] [7] [8]
ed
[5]
M.E. Touber, M.C. van Engelen, C. Georgakopoulus, J.A. van Rhijn, M.W.F. Nielen, Anal. Chim. Acta 586 (2007) 137. M.J. Taylor, G.A. Keenan, K.B. Reid, D.U. Fernandez, Rapid Commun. Mass Sp. 22 (2008) 2731. J.E. MacNair, K.C. Lewis, J.W. Jorgenson, Anal. Chem. 69 (1997) 983. J.E. MacNair, K.D. Patel, J.W. Jorgenson, Anal. Chem. 71 (1999) 700. M. Knaggs, FAPAS Proficiency Test Report 02183 Coccidiostats in Egg Powder February - April 2012, FAPAS, 2012. M. Knaggs, FAPAS Proficiency Test Report 02179 Nitrofuran Metabolites in Chicken Egg November - December 2011, FAPAS, 2012. M. Knaggs, FAPAS Proficiency Test Report 02175 Avermectins in Corned Beef September - October 2011, FAPAS, 2011. U. Chiuminatto, F. Gosetti, P. Dossetto, E. Mazzucco, D. Zampieri, E. Robotti, M.C. Gennaro, E. Marengo, Anal. Chem. 82 (2010) 5636. T. Vega-Morales, Z. Sosa-Ferrera, J.J. Santana-Rodriguez, J. Chromatogr. A 1230 (2012) 66. D. Gode, M.M. Martin, F. Steiner, C.G. Huber, D.A. Volmer, Anal. Bioanal. Chem. 404 (2012) 433. B. Kinsella, M. Whelan, H. Cantwell, M. McCormack, A. Furey, S.J. Lehotay, M. Danaher, Talanta 83 (2010) 14. A. Kaufmann, P. Butcher, K. Maden, S. Walker, M. Widmer, Anal. Chim. Acta 673 (2010) 60. A. Kaufmann, P. Butcher, K. Maden, S. Walker, M. Widmer, Rapid Commun. Mass Sp. 25 (2011) 979. N. Leon, M. Roca, C. Igualada, C.P.B. Martins, A. Pastor, V. Yusa, J. Chromatogr. A 1258 (2012) 55. A. Kaufmann, P. Butcher, K. Maden, S. Walker, M. Widmer, Anal. Chim. Acta 700 (2011) 86. R.K. Boyd, C. Basic, R.A. Bethem, Trace Quantitative Analysis by Mass Spectrometry, John Wiley & Sons Ltd., Chicester, 2008. K. Yu, D. Little, R. Plumb, B. Smith, Rapid Commun. Mass Sp. 20 (2006) 544. M. Whelan, B. Kinsella, A. Furey, M. Moloney, H. Cantwell, S.J. Lehotay, M. Danaher, J. Chromatogr. A 1217 (2010) 4612. R. Ventura, M. Roig, N. Monfort, P. Saez, R. Berges, J. Segura, Eur. J. Mass Spectrom. 14 (2008) 191. C.C. Leandro, P. Hancock, R.J. Fussell, B.J. Keely, J. Chromatogr. A 1103 (2006) 94. J.C. Van De Steene, W.E. Lambert, J. Am. Soc. Mass Spectrom. 19 (2008) 713. D. Guillarme, J. Ruta, S. Rudaz, J.L. Veuthey, Anal. Bioanal. Chem. 397 (2010) 1069. H.F. De Brabander, J. Vanden Bussche, W. Verbeke, L. Vanhaecke, TrAC Trend Anal. Chem. 30 (2011) 1088. M.I. Churchwell, N.C. Twaddle, L.R. Meeker, D.R. Doerge, J. Chromatogr. B 825 (2005) 134. A. de Villiers, H. Lauer, R. Szucs, S. Goodall, P. Sandra, J. Chromatogr. A 1113 (2006) 84.
pt
[4]
Ac ce
904 905 906 907 908 909 910 911 912 913 914 915 916 917 918 919 920 921 922 923 924 925 926 927 928 929 930 931 932 933 934 935 936 937 938 939 940 941 942 943 944 945 946 947 948 949
[22] [23] [24] [25] [26] [27] [28]
34
Page 35 of 47
[41] [42] [43] [44] [45] [46] [47] [48] [49]
ip t
cr
us
[39] [40]
an
[38]
M
[37]
ed
[36]
L. Novakova, J.L. Veuthey, D. Guillarme, J. Chromatogr. A 1218 (2011) 7971. D.V. McCalley, J. Chromatogr. A 1218 (2011) 2887. M.M. Fallas, M.R. Hadley, D.V. McCalley, J. Chromatogr. A 1216 (2009) 3961. D.V. McCalley, J. Chromatogr. A 1217 (2010) 4561. F. Gritti, G. Guiochon, J. Chromatogr. A 1218 (2011) 4632. A.C. Sanchez, J.A. Anspach, T. Farkas, J. Chromatogr. A 1228 (2012) 338. F.Y. Guan, C.E. Uboh, L.R. Soma, Y.W. You, Y. Liu, X.Q. Li, J. Mass Spectrom. 45 (2010) 1270. http://www.chem.agilent.com/Library/applications/5990-4253EN.pdf, Last accessed 22/06/2012. D. Guillarme, J. Schappler, S. Rudaz, J.L. Veuthey, TrAC - Trend Anal. Chem. 29 (2010) 15. A. Maquille, D. Guillarme, S. Rudaz, J.L. Veuthey, Chromatographia 70 (2009) 1373. K. Mastovska, S.J. Lehotay, J. Chromatogr. A 1000 (2003) 153. J.L.M. Vidal, A.G. Frenich, M.M. Aguilera-Luiz, R. Romero-Gonzalez, Anal. Bioanal. Chem. 397 (2010) 2777. T. Kovalczuk, M. Jech, J. Poustka, J. Hajslova, Anal. Chim. Acta 577 (2006) 8. J.P. Antignac, K. de Wasch, F. Monteau, H. De Brabander, F. Andre, B. Le Bizec, Anal. Chim. Acta 529 (2005) 129. B.K. Matuszewski, M.L. Constanzer, C.M. Chavez-Eng, Anal. Chem. 75 (2003) 3019. H. Trufelli, P. Palma, G. Famiglini, A. Cappiello, Mass Spectrom. Rev. 30 (2011) 491. R. King, R. Bonfiglio, C. Fernandez-Metzler, C. Miller-Stein, T. Olah, J. Am. Soc. Mass Spectrom. 11 (2000) 942. M.M. Aguilera-Luiz, J.L.M. Vidal, R. Romero-Gonzalez, A.G. Frenich, Food Chem. 132 (2012) 2171. J.L. Little, M.F. Wempe, C.M. Buchanan, J. Chromatogr. B 833 (2006) 219. O.A. Ismaiel, M.S. Halquist, M.Y. Elmamly, A. Shalaby, H.T. Karnes, J. Chromatogr. B 859 (2007) 84. V. Pucci, S. Di Palma, A. Alfieri, F. Bonelli, E. Monteagudo, J. Pharmaceut. Biomed. 50 (2009) 867. E. Chambers, D.M. Wagrowski-Diehl, Z.L. Lu, J.R. Mazzeo, J. Chromatogr. B 852 (2007) 22. E. Varga, T. Glauner, R. Koppen, K. Mayer, M. Sulyok, R. Schuhmacher, R. Krska, F. Berthiller, Anal. Bioanal. Chem. 402 (2012) 2675. E. Duffy, L. Rambaud, B. Le Bizec, M. O'Keeffe, Anal. Chim. Acta 637 (2009) 165. B. Kinsella, P. Byrne, H. Cantwell, M. McCormack, A. Furey, M. Danaher, J. Chromatogr. B 879 (2011) 3707. A. Radovnikovic, M. Moloney, P. Byrne, M. Danaher, J. Chromatogr. B 879 (2011) 159. B.F. Spisso, M.A.G. de Araujo, M.A. Monteiro, A.M.B. Lima, M.U. Pereira, R.A. Luiz, A.W. da Nobrega, Anal. Chim. Acta 656 (2009) 72. Off. J. Eur. Union L125 (1996) 10.
pt
[29] [30] [31] [32] [33] [34] [35]
Ac ce
950 951 952 953 954 955 956 957 958 959 960 961 962 963 964 965 966 967 968 969 970 971 972 973 974 975 976 977 978 979 980 981 982 983 984 985 986 987 988 989 990 991 992 993 994 995
[50] [51] [52] [53] [54] [55] [56]
35
Page 36 of 47
[67] [68] [69] [70] [71] [72] [73] [74] [75]
ip t
cr
us
[65] [66]
an
[63] [64]
M
[62]
http://ec.europa.eu/sanco_pesticides/public/index.cfm?event=homepage&CFID=1 404859&CFTOKEN=69040180&jsessionid=2405714dcd661a78e106TR#, Last accessed 18/12/2012. Off. J. Eur. Union L70 (2005) 1 J. Wang, W. Chow, D. Leung, J. Chang, J Agric Food Chem 60 (2012) 12088. M. Farre, L. Kantiani, M. Petrovic, S. Perez, D. Barcelo, J Chromatogr A 1259 86. J.V. Bussche, L. Vanhaecke, Y. Deceuninck, K. Verheyden, K. Wille, K. Bekaert, B. Le Bizec, H.F. De Brabander, J. Chromatogr. A 1217 (2010) 4285. K. Schmidt, C. Stachel, P. Gowik, Anal. Bioanal. Chem. 391 (2008) 1199. B. Kinsella, J. O'Mahony, E. Malone, M. Moloney, H. Cantwell, A. Furey, M. Danaher, J. Chromatogr. A 1216 (2009) 7977. Off. J. Eur. Union L125 (1996) 3. C. Van Poucke, E. Van Vossel, C. Van Peteghem, Rapid Commun. Mass Sp. 22 (2008) 2324. E.M. Malone, C.T. Elliott, D.G. Kennedy, L. Regan, Anal. Chim. Acta 637 (2009) 112. H.X. Wang, Y. Zhou, Q.W. Jiang, Chinese J. Anal. Chem. 39 (2011) 1323. Y.W. You, C.E. Uboh, L.R. Soma, F.Y. Guan, X.Q. Li, J.A. Rudy, Y. Liu, J.W. Chen, Rapid Commun. Mass Sp. 23 (2009) 2035. C.H.F. Wong, F.P.W. Tang, T.S.M. Wan, Anal. Chim. Acta 697 (2011) 48. H. Zheng, L.G. Deng, X.A. Lu, S.C. Zhao, C.Y. Guo, J.S. Mao, Y.T. Wang, G.S. Yang, H.Y. Aboul-Enein, Chromatographia 72 (2010) 79. F. Yang, Z.C. Liu, Y.H. Lin, J. Chen, G.F. Pang, G.N. Chen, Food Anal. Method 5 (2012) 138. B. Shao, X.F. Jia, J. Zhang, J. Meng, Y.N. Wu, H.J. Duan, X.M. Tu, Food Chem. 114 (2009) 1115. J. Zhang, B. Shao, J. Yin, Y.N. Wu, H.J. Duan, J. Chromatogr. B 877 (2009) 1915. M.W.F. Nielen, J.J.P. Lasaroms, M.L. Essers, J.E. Oosterink, T. Meijer, M.B. Sanders, T. Zuidema, A.A.M. Stolker, Anal. Bioanal. Chem. 391 (2008) 199. H. Zhang, Y.P. Ren, X.L. Bao, J. Pharmaceut. Biomed. 49 (2009) 367. Off. J. Eur. Comm. L224 (1990) 1. M. Lombardo-Agui, A.M. Garcia-Campana, L. Gamiz-Gracia, C. Cruces-Blanco, Talanta 93 (2012) 193. X.T. Zhao, Q.B. Lin, H. Song, Y.L. Pan, X. Wang, J. Agric. Food Chem. 59 (2011) 9800. Y.X. She, J.J. Liu, J. Wang, Y. Liu, R.Y. Wang, W.Q. Cao, Anal. Lett. 43 (2010) 2246. M. McDonald, C. Mannion, P. Rafter, J. Chromatogr. A 1216 (2009) 8110. V. Tamosiunas, A. Padarauskas, Chromatographia 67 (2008) 783. Z.X. Cai, Y. Zhang, H.F. Pan, X.W. Tie, Y.P. Ren, J. Chromatogr. A 1200 (2008) 144.
ed
[59] [60] [61]
Off. J. Eur. Union L309 (2009) 1
pt
[57] [58]
Ac ce
996 997 998 999 1000 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040
[76] [77] [78] [79] [80] [81] [82] [83]
36
Page 37 of 47
[91] [92] [93] [94] [95] [96] [97] [98] [99] [100]
ip t
cr
[90]
us
[89]
an
[88]
M
[87]
ed
[85] [86]
M. Whelan, C. Chirollo, A. Furey, M.L. Cortesi, A. Anastasio, M. Danaher, J. Agric. Food Chem. 58 (2010) 12204. Off.J.Eur.Union L15 (2010) 1. X.J. Zhang, H.X. Xu, H. Zhang, Y.M. Guo, Z.Y. Dai, X.C. Chen, Anal. Bioanal. Chem. 401 (2011) 727. M. Whelan, Y. Bloemhoff, A. Furey, R. Sayers, M. Danaher, J. Agric. Food Chem. 59 (2011) 7793. X. Xia, Y.C. Dong, P.J. Luo, X. Wang, X.W. Li, S.Y. Ding, J.Z. Shen, J. Chromatogr. B 878 (2010) 3174. H. De Ruyck, E. Daeseleire, H. De Ridder, R. Van Renterghem, J. Chromatogr. A 976 (2002) 181. B. Shao, X.Y. Wu, J. Zhang, H.J. Duan, X.G. Chu, Y.N. Wu, Chromatographia 69 (2009) 1083. E.M. Malone, G. Dowling, C.T. Elliott, D.G. Kennedy, L. Regan, J. Chromatogr. A 1216 (2009) 8132. T. Hu, T. Peng, X.J. Li, D.D. Chen, H.H. Dai, X.J. Deng, Z.F. Yue, G.M. Wang, J.Z. Shen, X. Xia, S.Y. Ding, Y.N. Zhou, A.L. Zhu, H.Y. Jiang, J. Chromatogr. A 1219 (2012) 104. E. Van Pamel, A. Verbeken, G. Vlaemynck, J. De Boever, E. Daeseleire, J. Agric. Food Chem. 59 (2011) 9747. H.G.J. Mol, P. Plaza-Bolanos, P. Zomer, T.C. de Rijk, A.A.M. Stolker, P.P.J. Mulder, Anal. Chem. 80 (2008) 9450. A. Herrmann, J. Rosen, D. Jansson, K.E. Hellenas, J. Chromatogr. A 1235 115. J. Zhan, X.J. Yu, Y.Y. Zhong, Z.T. Zhang, X.M. Cui, J.F. Peng, R. Feng, X.T. Liu, Y. Zhu, J. Chromatogr. B 906 (2012) 48. A. Kaufmann, P. Butcher, K. Maden, M. Widmer, J. Chromatogr. A 1194 (2008) 66. M. Ibanez, J.V. Sancho, D. McMillan, R. Rao, F. Hernandez, TrAC - Trend Anal. Chem. 27 (2008) 481. D. Ortelli, E. Cognard, P. Jan, P. Edder, J. Chromatogr. B 877 (2009) 2363. R.J.B. Peters, Y.J.C. Bolck, P. Rutgers, A.A.M. Stolker, M.W.F. Nielen, J. Chromatogr. A 1216 (2009) 8206. M. Arienzo, D. Cataldo, L. Ferrara, Food Control 31 (2013) 108 C.P. Rezende, M.P. Almeida, R.B. Brito, C.K. Nonaka, M.O. Leite, Food Addit. Contam. A 29 (2012) 541. S.J. Lehotay, A.R. Lightfield, L. Geis-Asteggiante, M.J. Schneider, T. Dutko, C. Ng, L. Bluhm, K. Mastovska, Drug Test. Anal. 4 (2012) 75. L. Geis-Asteggiante, S.J. Lehotay, A.R. Lightfield, T. Dutko, C. Ng, L. Bluhm, J. Chromatogr. A 1258 (2012) 43. Y.Y. Tang, H.F. Lu, H.Y. Lin, Y.C. Shih, D.F. Hwang, J. Chromatogr. B 881-882 (2012) 12. L. Wang, Y.Q. Li, Chromatographia 70 (2009) 253. M.M. Aguilera-Luiz, J.L.M. Vidal, R. Romero-Gonzalez, A.G. Frenich, J. Chromatogr. A 1205 (2008) 10. J.L.M. Vidal, M.D. Aguilera-Luiz, R. Romero-Gonzalez, A.G. Frenich, J. Agric. Food Chem. 57 (2009) 1760.
pt
[84]
Ac ce
1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055 1056 1057 1058 1059 1060 1061 1062 1063 1064 1065 1066 1067 1068 1069 1070 1071 1072 1073 1074 1075 1076 1077 1078 1079 1080 1081 1082 1083 1084 1085 1086
[101] [102] [103] [104] [105] [106] [107] [108]
37
Page 38 of 47
1087 1088 1089 1090 1091
[109] A.G. Frenich, M.D. Aguilera-Luiz, J.L.M. Vidal, R. Romero-Gonzalez, Anal. Chim. Acta 661 (2010) 150. [110] B. Shao, X. Jia, Y. Wu, J. Hu, X. Tu, J. Zhang, Rapid Commun. Mass Sp. 21 (2007) 3487.
Ac ce
pt
ed
M
an
us
cr
ip t
1092
38
Page 39 of 47
Table 1. Currently available commercial UHPLC systems and capabilities. Table 2. Examples of currently available commercial tandem mass spectrometric detection systems and capabilities.
ip t
Table 3. Application of UHPLC-MS to analysis of veterinary drug residues.
Fig. 1. Structures of epimeric synthetic glucocorticosteroids dexamethasone and
us
cr
betamethasone.
Fig. 2. Growth in the number of publications on UHPLC, and UHPLC combined with mass spectrometry, 2003 – 2011. Data was sourced from Thomson Reuters Web of
M
an
Science™ (keywords: “UHPLC” or “UPLC”; “MS” or “mass spec*”).
Fig. 3. Band broadening due to frictional heating. The top diagram shows the typical
ed
quantity of band broadening expected due to diffusion processes when the column is heated evenly. The lower diagram shows the effect of internal column heating due to
ce pt
frictional effects between mobile phase and stationary phase.
Fig. 4. Experiment for observation of suppression effects. The signal corresponds to the introduction into the mass spectrometer of a mobile phase or blank sample extract
Ac
injection (from the LC) and standard solution (via direct infusion). When just mobile phase is introduced from the LC, signal remains constant (a, b). However, upon injection of blank sample extract, although the TIC increases (c), the specific transition monitored for the analyte (d) is suppressed due to matrix effects. Adapted from [20] with permission from Elsevier.
Page 40 of 47
Ac
ce pt
ed
M
an
us
cr
ip t
Figure 1
Page 41 of 47
Figure 2
800
700
600
ip t
500
400
UHPLC
cr
300
UHPLC & MS
200
0 2005
2006
2007
2008
an
2004
2009
2010
2011
Ac
ce pt
ed
M
2003
us
100
Page 42 of 47
Ac
ce pt
ed
M
an
us
cr
ip t
Figure 3
Page 43 of 47
Ac
ce pt
ed
M
an
us
cr
ip t
Figure 4
Page 44 of 47
Table 1
Manufacturer UHPLC System
Maximum
Maximum
Comments
Pressure
Flow
(bar)
(mL min-1)*
Rate
Pump delay volume of <10 μL, Infinity 1290
1200
2
Hitachi
LaChromUltra
600
2.5
depending on solvent mixing unit.
ip t
Agilent
System delay volume as low as 266 μL.
Jasco
X-LC
1000
cr
Minimum injection interval of 30 1.35
seconds.
Minimum injection interval of 14 Nexera
1300
3
Thermo Scientific
Accela
1310
2
an
Class
1000
1
1240
Waters
nanoAcquity
690
Dionex
UltiMate
1000
PerkinElmer
Flexar
1240
Knauer
1720
ce pt
Eksigent
UltraLC
PLATINblue
1
M
Class
ed
Waters
UHP
injection interval of 30 seconds. System delay volume of < 100 μL;
Acquity I-
SSI LabAlliance
Pump delay volume of 70 μL. System delay volume < 400 μL;
Acquity HWaters
seconds
us
Shimadzu
1240
1000
0.1
injection interval of 15 seconds. System delay volume of < 1 μL.
8
Injection cycle time of 30 seconds.
5
Injection cycle time of 8 seconds. Flow rates of up to 12 mL min-1
5
possible. Mixer volume 60 μL. Injection cycle
5
time of < 60 seconds. System delay volume of 100 μL; 15
2
second injection cycle time.
Ac
* Maximum flow rate achievable at maximum pressure
Page 45 of 47
Table 2
Vendor
Dwell time (ms)
Polarity Switch (ms)
Quoted Sensitivity
Scan Speed
6460 Triple Quadrupole 6500 Triple Quadrupole Xevo TQ-S
Agilent
1
30
1 pg (Reserpine)
AB SCIEX
1
20
700 ag Alprazolam
Waters
1
20
Triple Quad 6500 8030 Triple Quadrupole TSQ Vantage Triple Quadrupole
AB SCIEX
1
20
500 pg l-1 (Raffinose) 700 ag (Verapamil)
Shimadzu
1
15
1 pg (Reserpine)
N/A
1 pg (Chloramphenicol)
5,200 amu/s 12,000 amu/s 1000 amu/s 20,000 amu/s 15,000 amu/s 5000 amu/s
cr
Ac
ce pt
ed
M
an
us
ThermoScientific 2
ip t
Mass Spectrometer
Page 46 of 47
13 β-agonists
16 22
Synthetic Steroids
55 9
Quinolones
8 22
Sulphonamides
12
Macrolides
4
NSAID
30
-Lactam
8
REF.
0.14 – 123 (CCα)
0.1
[21]
8
0.01 – 0.5 (LOD)
0.1
40
17
0.2 – 0.79 (CCα)
0.1
40
15
< 0.2 (CCα)
2.5
50
5
0.5 x 10-7
40
13
0.5 x 10-7 – 2 x 10-7 (LOD) 0.11 – 0.3 (LOD)
0.4
35
5.5
0.2 – 4.1 (LOD)
5
0.3
40
12
0.002 – 0.102 (LOD)
3 x 10-6
0.25
40
25
0.4 – 8.0 (LOD)
10
Micromass Quattro Premier XE triple quadropole MS Micromass Quattro Premier XE triple quadropole MS Micromass Quattro Premier XE triple quadropole MS Micromass Quattro Premier XE triple quadropole MS Finnigan TSQ Quantum Ultra triple stage MS Hybrid quadrupole-orthogonal acceleration TOF MS Triple quadrupole MS API 3200 Micromass Quattro Premier XE triple quadropole MS TQ Detector MS
0.3
30
9.5
12.2 – 217.8 (CCα)
1.5
[109]
0.2
30
16
0.4 - 2 (LOD)
1
Acquity TQD tandem quadrupole MS TQ MS
0.6
40
11
30.9 – 371.7 (CCα)
12.5
Micromass Quattro Premier XE triple quadropole MS
[102]
60
Injection Cycle Time (min) 13
0.3
30
0.3 0.4 0.5 0.7
LOD (µg kg-1)
us
Acquity UPLC HSS T3 (100 mm x 2.1 mm, 1.8 µm) Acquity BEH C18 (50 mm x 2.1 mm, 1.7 µm) Acquity BEH C18 (50 mm x 2.1 mm, 1.7 µm) Acquity BEH C18 (50 mm x 2.1 mm, 1.7 µm) Hypersil Gold C18 (50 mm x 2.1 mm, 1.9 µm) Acquity UPLC HSS T3 (100 mm x 2.1 mm, 1.8 µm) Zorbax Eclipse Plus HHRD (100 mm x 2.1 mm, 1.8 µm) Acquity Shield RP 18 (100 mm x 2.1 mm, 1.7 µm) Acquity BEH C18 (100 mm x 2.1 mm, 1.7 µm) Acquity BEH C18 (50 mm x 2.1 mm, 1.7 µm) Acquity BEH C18 (50 mm x 2.1 mm, 1.7 µm) Acquity BEH C18 (50 mm x 2.1 mm, 1.7 µm)
Detector
Temp (°C)
M an
38
LOQ (µg kg-1)
Flow Rate (mL min-1) 0.6
ed
Column
ce pt
Anthelmintics
Residues
Ac
Drug
cr
ip t
Tables
1
[88] [73] [75] [35] [68] [78] [76] [79]
[92]
Page 47 of 47