Advertisement

Chromatographia

, Volume 82, Issue 1, pp 251–260 | Cite as

Critical Comparison of Liquid Chromatography Coupled to Mass Spectrometry and Three Different Ion Mobility Spectrometry Systems on Their Separation Capability for Small Isomeric Compounds

  • Tobias Werres
  • Juri Leonhardt
  • Martin Jäger
  • Thorsten TeutenbergEmail author
Article
  • 143 Downloads
Part of the following topical collections:
  1. 50th Anniversary Commemorative Issue

Abstract

The very fast separation and identification of isomeric small substances with a molecular weight under 800 Da is still a challenge for high-throughput analysis. Inadequate chromatographic or spectrometric separation hampers an unequivocal identification of isomers, which may be recorded as a sum parameter. Reversed-phase liquid chromatography can be considered a generic method for the separation of isomers. However, a separation is usually achieved within minutes and not milliseconds as is typical for ion mobility spectrometry. The aim of this study, therefore, was to investigate the potential of IMS to separate small isomeric compounds. Hence, 23 substances divided into 11 isomeric groups have been selected. Among them, cancer drugs, hormones, pain relievers and others are contained. Three ion mobility spectrometers with different separation principles were compared with respect to their resolving performance. These systems comprised a traveling wave ion mobility spectrometer, a differential ion mobility spectrometer and a differential mobility analyzer. For reference, the chromatographic resolution and peak capacity was determined by high-performance liquid chromatography using a reversed phase.

Keywords

Micro HPLC Generic method Ion mobility spectrometry Small molecules Isomeric separation 

Introduction

The very fast separation and identification of isomeric substances with a molecular weight under 800 Da is still a challenge for high-throughput analysis. This is especially true for very complex samples containing many target analytes and a variety of unknown matrix components. Isomers might have different structural entities after fragmentation that allow a differentiation by mass spectrometry (MS). However, compounds with the same exact mass and the same fragmentation pattern cannot be distinguished by MS regardless of the resolution power. For example, the anthracyclines doxorubicin and epirubicin, which are one of the most effective available chemotherapeutic agents for the treatment of breast cancer, only differ in the spatial orientation of one of the hydroxyl groups as is shown in Fig. 1a, b [1]. While both drugs have the same side effects such as myelosuppression and cardiotoxicity, the epimers exhibit different equimolar dose ratio, which is 1:1.2 and 1:1.7–2.0 for doxorubicin and epirubicin, respectively [2, 3]. This means that, unlike doxorubicin, twice the amount of epirubicin can usually be administered before cardiotoxicity occurs. Since the preparations for chemotherapy are individually adjusted for each patient, confusion of the substances or an incorrect preparation can lead to undesirable side effects for the patient.

Fig. 1

Structural formulae of a epirubicin, b doxorubicin, c trenbolone, and d estrone. The circle in a and b highlights the spatial orientation of the hydroxyl group

The necessity of a separation of isomers can also be demonstrated by the example of trenbolone and estrone. Both hormones are strong endocrine disruptors and constitutional isomers. Due to their potential to exert a massive influence on the organism even at low concentrations, estrone has been added to the watch list of the European Union [4]. A maximum acceptable method detection limit of 0.4 ng/L is envisaged, whereas trenbolone is not mentioned in that regulation. The combination of faulty sample preparation and insufficient isomer separation can lead to the determination of the concentration of estrone and trenbolone as a sum parameter. Due to the very low limit value for estrone, this error can quickly lead to a large number of environmental samples being exceeded.

In general, when two compounds have the same exact mass there are three different possible ways to separate them: (i) different fragmentation in MS–MS experiments (ii); different retention times in chromatography; (iii) different collision cross section (CCS) in ion mobility spectrometry. It is often assumed that when using an MS technique, the need for a chromatographic separation is obsolete, because co-eluting substances can be distinguished due to the difference in their accurate masses, different mass transitions or fragmentation patterns. Since isomers have the same monoisotopic mass, a differentiation by high-resolution mass spectrometry is impossible. If different fragmentation patterns are observed in MS/MS experiments, a differentiation is possible. However, ion suppression might aggravate the quantitation.

The second and probably most generic approach for the separation of isomeric species is the use of high-performance liquid chromatography (HPLC) coupled to mass spectrometry [5]. However, the separation is usually achieved within minutes [6]. The possible increase in speed and the gain in resolution when using a sub 2 µm phase for chromatographic separation could be the key factor to obtain the necessary peak capacity in a very short time [7].

The third approach is the use of ion mobility spectrometry (IMS). Its principle is based on the difference in time that ions need to travel through a drift area in a gas atmosphere. The time needed is based on the average drift velocity in an electric field [8]. Current investigations point out that IMS has a great potential to separate isobaric compounds within milliseconds [9]. The aim of this study, therefore, was to investigate the potential of IMS as a generic method to separate small isomeric compounds. If IMS can achieve sufficient resolution, it will outperform chromatography in high-throughput analysis.

For the investigation, 10 isomeric pairs and one isomeric set of three were chosen to compare different analytical systems with respect to their performance in separation resolution. A traveling wave IMS (TWIMS), a differential IMS (DMS) and a differential mobility analyzer (DMA) have been selected. Different dopants and drift gases were used to investigate their influence on the resolution between critical peak pairs. Structure formulas of the critical peak pairs can be found in the supplementary file.

The second part of the work should answer the question what peak capacity is needed to separate the majority of the critical peak pairs with a resolution higher or equal than 1.5. For this investigation, a 1.9 µm fully porous reverse phase stationary phase was chosen. The gradient slope was varied between 30 s and 30 min. Besides that, no further optimization of the mobile phase, its pH or the column temperature was made.

Theoretical Background

Systems Used in this Work

Several IMS systems were used in this work. The first was a traveling-wave IMS (TWIMS). The drift region consists of a stacked-ring ion guide on which the wave voltage is applied. The ions are accelerated through the gas-filled separation region by wave-shaped voltages, which are generated by pulsing on two adjacent ring electrodes. After a predefined dwell time, the next pair is pulsed. By influencing the speed and amplitude of the TW, separation by mobility can be achieved. The compounds are separated by fast successive waves [10, 11].

Second, a differential mobility spectrometer (DMS) was used. The separation by DMS is based on the same principles than the better-known field asymmetric ion mobility spectrometry (FAIMS) systems. The separation area is composed of two flat parallel electrodes. Across these electrodes, high alternating electric fields, known as dispersion voltage (DV), are applied. To achieve stable trajectories for the respective ions, a compensation voltage (CV) must be applied. Otherwise, the ions would be discharged at the electrodes with the exception of ions that pass at CV = 0 [12]. To achieve high field strengths at moderate voltages, the distance between the electrodes should be small. This allows achieving a higher resolution without causing an electric spark between both electrodes. The ions travel with the gas flow perpendicular to the electric fields. In contrast to the TWIMS the DMS functions as an ion filter. Through serially scanning the CV only ions with matching differential mobility pass the drift area.

The third type of IMS was a differential mobility analyzer (DMA). Unlike other IMS, the DMA system separates the ions only in space and not by drift time. All ions have to travel the same distance to the detector. The DMA system can also be considered as a variation of an aspiration condenser IMS (A-IMS) [13]. The separation takes place between two electrodes with different potentials. A sheath gas is fed parallel to these electrodes. The ions are introduced orthogonally in the laminar gas flow. Higher flow rates usually result in better resolution. The DMA is also a filter method. To scan a sample mixture with polydispersive mobility, the voltage of one of the electrodes is gradually increased. Depending on the current voltage, only one type of ions with a specific mobility reaches the detector. All other components are discharged at the electrode [14]. A schematic representation of the systems used can be found in the supplementary material provided.

Peak Capacity and Resolution

The peak capacity \({n_{\text{c}}}\) describes the efficiency of a column and indicates how many components can be separated theoretically during a given gradient time and with a defined resolution. The peak capacity depends on the gradient time and the column length. It was calculated in this work according to Eq. 1 [15]:
$${n_{\text{c}}}=1+\frac{{{t_{\text{g}}}}}{{{w_{\text{b}}}}} \approx 1+\frac{{2354 \times {t_{\text{g}}}}}{{4 \times {w_{\text{h}}}}},$$
(1)
where \({w_{\text{b}}}\) is the peak base width, \({w_{\text{h}}}\) is the peak width at full width at half maximum (FWHM) and \({t_{\text{g}}}\) the gradient time. The peak capacity production rate \(~{n_{\text{p}}}\) is obtained by dividing the peak capacity by the gradient time.
$${n_{\text{p}}}=\frac{{{n_{\text{c}}}}}{{{t_{\text{g}}}}}.$$
(2)
At a resolution of 1.5, the peaks are baseline separated. At a resolution of 1.0, they share a 2.3% mutual overlap, assuming there is no tailing or fronting and the peak areas are identical. In this paper, the resolution \({R_{{\text{FWHM}}}}\) was calculated at full width at half maximum (FWHM) using the following equation:
$${R_{{\text{FWHM}}}}=1.18 \times \left( {\frac{{{t_{{\text{r2}}}} - {t_{{\text{r1}}}}}}{{{w_{{\text{h1}}}}+{w_{{\text{h2}}}}}}} \right),$$
(3)
where \({t_{{\text{r1}}}}\) is the retention time of the first and \({t_{{\text{r2}}}}\) the retention time of the second component. To calculate the resolution for the IMS separation, the retention times in Eq. 3 were substituted by the drift times or dispersion voltages.

Experimental Section

Chemicals

Water, acetonitrile and methanol were of HPLC grade quality purchased from Th. Geyer-Chemsolute (Renningen, Germany). Formic acid was used as a solvent additive to adjust an acidic pH of the mobile phases and was purchased from Sigma-Aldrich (Seelze, Germany).

Table 1 lists the compounds that were selected in this study to determine the resolution for IMS and RP-LC. The mass range extends from m/z 133.064 to m/z 776.854. With the exception of tramadol from Heumann PCS (Nürnberg, Germany) and the three components trenbolone, iomeprol and iopamidol from Dr. Ehrenstorfer (Augsburg, Germany), all analytes were purchased from Sigma-Aldrich (Missouri, USA).

Table 1

Isomeric compounds sorted by groups with the same monoisotopic mass

Group

Substance

Mass

Isomerism

MIX 1

MIX 2

MIX 3

MIX 4

1

5-Methyl-1H-benzotriazole

4-Methyl-1H-benzotriazole

133.06

Constitutional

X

X

  

X

 

X

 

2

Cyclophosphamide

Ifosfamide

260.03

Constitutional

X

X

  

X

 

X

 

3

Tramadol

Desvenlafaxine

263.19

Constitutional

X

X

  

X

 

X

 

4

Estrone

Trenbolone

270.16

Constitutional

X

X

  

X

 

X

 

5

Dehydroepiandrosterone

Testosterone

288.21

Constitutional

X

X

  

X

 

X

 

6

Dihydrotestosterone

Etiocholanolone

Androsterone

290.23

Configurational

X

  

X

X

X

  

X

 

X

 

7

Sucrose

Lactose

342.12

Constitutional

X

X

  

X

 

X

 

8

Doxorubicin

Epirubicin

543.17

Constitutional

X

X

  

X

 

X

 

9

Iomeprol

Iopamidol

776.85

Constitutional

X

X

  

X

 

X

 

10

Corticosterone

Cortexolone

346.21

Configurational

X

X

  

X

 

X

 

11

Cortisone

Prednisolone

360.19

Constitutional

X

X

  

X

 

X

 

Four mixtures were prepared. Sample mixture 1 containing all 23 substances, sample mixture 2 containing the first substance of a group and sample mixture 3 containing the second substance of a group. Mixture 4 only contained dihydrotestosterone.

Instrumentals

The SYNAPT G2-Si High (Waters, Milford, USA) was the TWIMS and HRMS system. This consisted of the ZSpray™ ion source, the StepWave device for elimination of neutral molecules, a quadrupole, the TriWave ion mobility separation unit and the time-of-flight mass spectrometer QuantTOF for detection. A linear wave velocity ramp started from 800 m/s and ended at 200 m/s. The starting voltage of the wave was set to 40 V and the final voltage to 5 V. If not stated otherwise, carbon dioxide was used as drift gas for all IMS-systems.

The planar DMA cell (SEADM, Valladolid, Spain) was coupled with a Triple Quad 3500 (Sciex, Dublin, USA) for identification of the ions. When possible, a multiple reaction monitoring (MRM) was performed. The ionization was carried out at ambient temperature via an electrospray ionization (ESI) source. The samples were fed through a passive sample changer. The carrier gas was controlled by the DMA blower unit, which was set to 15,000 rpm for all measurements. The range from 0 to 6000 V was scanned. The residence time for each mass transient was 300 ms with a step size of 5 V. To evaluate whether there were any changes in the quality of the separation by use of dopants, experiments have been carried out with and without acetonitrile (ACN), which was added via an additional input in the blower unit at a concentration of 3 vol-%.

The differential ion mobility spectrometer (DMS) SelexION® (Sciex, Dublin, USA) was coupled with a Triple Quad 6500 mass spectrometer (Sciex, Dublin, USA) for identification of the ions. For all experiments, the separation voltage was set to 3500 V and the temperature inside the separation region to 498 K. The compensation voltage that serves to stabilize the trajectories for the ions between both electrodes in the DMS was scanned from − 100 V to 100 V in 0.1 V increments. The ions were produced in an ESI source and the detector operated in positive ion mode. The total scan time for the identification method amounted to 0.23 s, the time of the whole method was 7.59 min with 2000 cycles. A declustering potential of 100 V, an entrance potential of 10 V, an ion spray voltage of 5500 V and a curtain gas voltage of 20 V were chosen after optimization. For the investigation of the separation of the isobaric groups, sample mixture 1 was directly introduced into the DMS. For identification of the single components, sample mixture 2 and sample mixture 3 were measured afterwards. The triple quadrupole MS was operated in single ion monitoring mode. Moreover, 2-propanol was used as dopant. Without the dopant, none of the pairs could be separated from the DMS-system.

Chromatographic Separation

An Eksigent ExpressLC Ultra system (Sciex, Dublin, CA, USA) was coupled to a QTrap 3200 mass spectrometer (Sciex, Dublin, CA, USA) for identification of the ions. The system consisted of a binary pneumatic pump that was able to generate a maximum pressure of 690 bar and a flow module with a flow rate range from 5 to 50 µL/min. A PEEKSil sample loop with dimensions of 75 µm × 10 cm, resulting in a volume of 442 nL, was used. The built-in six-port valve was used in the two positions load / inject mode for sample introduction. Sample introduction was achieved by an HTS PAL autosampler (CTC Analytics, Zwingen, Switzerland). The column was heated to 323 K using the built-in column oven. All other tubing consisted of fused-silica capillaries with an inner diameter of 50 µm and an outer diameter of 360 µm. The length of the capillary for the connection between the six-port valve and the column was adjusted to 10 cm using a diamond cutter. The connection to the ESI source was achieved with 50 µm fused-silica capillaries. To minimize the band broadening after the column, an emitter tip with an inner diameter of 50 µm was installed instead of the classical 100 µm ID emitter tip. While the classical tip is made of stainless steel, the modified miniaturized emitter tip is based on a PEEKSil capillary. To ensure the ionization, at the top of the PEEKSil capillary a stainless steel tip with the respective inner diameter is installed. Regarding the connection technique, it should be mentioned that the PEEKSil emitter is designed for 1/32″ fittings. High-pressure resistant fittings are screwed to a 1/32″ union. This union additionally offers the advantage of being able to be used as a grounding point.

A silica-based reversed phase stationary phase was purchased from YMC Europe GmbH (Dinslaken, Germany). The column with a length of 50 mm and an ID of 0.3 mm was packed with fully porous 1.9 µm particles with a pore size of 120 Å.

Software

Data acquisition for the TWIMS was performed using the software MasLynx 4.1 (Waters, Milford, USA). For all the other analytical systems, Analyst V. 1.6.3 (Sciex, Dublin, CA, USA) was used. Further data processing was performed using Origin 2018 V. 9.5 (OriginLab, Massachusetts, USA), Microsoft Office Excel 2010 and UNIFI V. 1.8.2.0 (Waters, Milford, USA).

Results and Discussion

Results of the IMS Experiments

Table 2 summarizes the calculated resolutions for all three IMS systems. Two different approaches were applied for the determination of the resolution. The first approach is based on the calculation of the resolution that could be achieved if the substances do not interfere with each other. This so-called virtual resolution was calculated from the reconstructed chromatogram resulting from two different samples in which only one component of a group was present. The second approach is based on the calculation of the real resolution that was achieved when both substances of a group were present in the same sample.

Table 2

Overview of the IMS resolution-data

Group

Substance

TWIMS

DMS

DMA

Virtual

Real

Virtual

Real

Virtual

Real

1

5-Methyl-1H-benzotriazole

4-Methyl-1H-benzotriazole

X

X

X

X

X

X

2

Cyclophosphamide

Ifosfamide

X

X

X

X

X

X

3

Tramadol

Desvenlafaxine

0.5

0.5

1.6

1.4

1.4

0.9

4

Estrone

Trenbolone

X

X

X

X

1.2

X

5

Dehydroepiandrosterone

Testosterone

X

X

X

X

X

X

6

Dihydrotestosterone

Etiocholanolone

Androsterone

X

X

X

X

X

X

7

Lactose

Sucrose

X

X

X

2.3

6.4

X

8

Doxorubicin

Epirubicin

X

X

0.5

X

X

X

9

Iomeprol

Iopamidol

0.7

0.7

X

X

X

X

10

Corticosterone

Cortexolone

X

X

0.6

X

3.2

3.1

11

Cortisone

Prednisolone

2.2

2.5

X

X

2.8

2.1

Only values for which a resolution higher or equal than 0.5 was obtained are shown. DMS data for the measurements with 2-propanol. DMA data gathered with 15,000 rpm and without dopants

X not separated

A baseline separation with a real resolution of more than 1.5 was only obtained for three groups (pairs 7, 10 and 11). With the DMA system, two peak pairs with a resolution > 1.5 were separated (pairs 10 and 11). With the DMS system, a baseline separation was obtained for group seven. Furthermore, a resolution of 1.4 is reached for the peak pair three. The highest resolution of 3.1 was achieved for corticosterone and cortexolone (pair 10) using the DMA system. TWIMS was able to separate iomeprol and iopamidol (pair 9) with a resolution of 0.7 and pair 11 with a resolution of 2.5.

A higher resolution was usually obtained when the components were injected separately and the virtual resolution was determined. This is especially noticeable for the DMA and DMS systems. In contrast, when using the TWIMS system, the resolution did not depend on whether the substances were introduced as a single compound or as a pair. For cortisone and prednisolone, an even higher resolution was observed when both compounds were introduced simultaneously. One possible explanation for this phenomenon is the interaction between the Coulomb forces and the travel wave forces [16]. When cortisone and prednisolone are introduced into the drift tube at the same time, an increased ion density results. Therefore, stronger repulsive Coulomb forces will be observed, leading to peak broadening. As a result of this repulsion, the analyte ions of both components are partially mixed. The focusing effect of the sequence of symmetric potential waves continually propagating through a tube and hitting the ions could be the main reason for the observation of a higher resolution when using the TWIMS. By the impact of the waves, the ions experience a kind of inelastic collision which leads to a focusing effect. This effect then counteracts the repulsive Coulomb forces. The focused ion cloud of each analyte can now be considered as a single charge. The two charges of the different analyte ions repel each other and thus the resolution is increased.

For the DMA system, the resolution is always lower if the compounds are introduced simultaneously into the drift region. The effect of mutual influence of the interaction between estrone and trenbolone is shown in Fig. 2. The black and grey marked peaks represent the virtual separation. Here, a resolution of 1.2 was achieved. When both components are introduced simultaneously into the ionization chamber, only one single peak is observed. The peak area of this single peak is approximately as large as the sum of the peak areas of the individual peaks. By increasing the ion density through the simultaneous introduction of both ions, the signals merge resulting in a single peak. A possible explanation for this phenomenon is the formation of heterodimers as has recently been demonstrated by Morrison et al. [17]. In this instance, a formation of almost identical estrone and trenbolone dimeres would have occurred.

Fig. 2

Separation of estrone and trenbolone, recorded on the DMA device at 15,000 rpm and a scan time per mass of 300 ms

Using ACN as dopant, no improvement in the resolution was achieved. The change of the transport gas from nitrogen to carbon dioxide did not lead to an increase in resolution, either. The combination of ACN and carbon dioxide did not result in more separated compounds. Only a shift of the entire spectrum to higher deflection voltages was found. This is due to an increase in the CCS due to the formation of ACN-analyte clusters [18].

Using the DMS device, the separation of lactose and sucrose could be achieved due to the formation of a sodium adduct. A separation of the protonated species was not possible. In addition, the influence of dopants was investigated. Using 2-propanol, the separation of tramadol and desvenlafaxine was achieved, as shown in Fig. 3. Without dopant, no separation could be obtained. By injecting 2-propanol, the nonlinear interaction in the gas phase changes resulting in the signal shift towards negative compensation voltages.

Fig. 3

Mobility spectra of tramadol and desvenlafaxine using the DMS-system with 2-propanol (grey) and without dopant (white). Peak corresponding to desvenlafaxine (1), tramadol, (2) and both compounds (1/2)

For pair 10 and 11, a resolution of 3.1 and 2.1 was obtained using the DMA device. Even if both substances of these groups are present, the differences in their three-dimensional structure lead to a sufficient difference in their CCS, thus enabling the separation. This is 5.7 Å2 or 2% for cortisone and prednisolone. However, the high-resolution for corticosterone and cortexolone cannot be explained by the differences of their CCS values alone, as this is only 0.5 Å2. In addition, the effects of charge absorption must play a role here. All four analytes are steroids whose basic element is gonane, which consists of three six-membered (A, B, C) and one five-membered ring (D). For corticosterone, an OH group is located at position 11 and thus at the C-ring of the sterol. In the case of cortexolone, the OH substituent is in position 17, which also influences the orientation of the alpha-hydroxy ketone. Prednisolone differs from cortisone in that there is an OH group at position 11 and an additional double bond in the A ring.

When entering the vacuum of the MS, the ion clusters are no longer stable. Therefore, in terms of resolution, it is only of minor relevance whether a low resolution or a high-resolution mass spectrometer is used. A substance in IMS does not necessarily produce a single signal. Since multimers or clusters may be formed, several peaks may be observed. The intensity ratio of these peaks supports or even enables the identification of the compounds. Tramadol and desvenlafaxine serve as an example in this study cf. Fig. 4. For tramadol, the ratio between the monomer peak and dimer peak is roughly 1:1, while for desvenlafaxine this ratio is 1:3.

Fig. 4

Difference in the intensity ratio of tramadol and desvenlafaxine using the DMA-system. Comparison between monomer and dimer peaks

Results of the LC Experiment

Table 3 lists the resolution and Fig. 5 shows the peak capacity and peak capacity production rate for the different gradient times. Peak capacity was calculated according to Eq. 1. If the gradient time is 30 s, 24 peaks with a resolution of 1 can be separated within the gradient window. A gradient time of 30 s represents the range of ultra-fast separations and is interesting for high-throughput applications. For comprehensive 2D-LC, the modulation time and thus cycle time is usually 30 s–60 s [19]. Against this background, gradient times with a running time of 45 s and 60 s were selected, as these completely cover the range for “comprehensive” 2D-LC. The data summarized in Fig. 5 shows that the peak capacity increases from 32 to 38 when the gradient time is increased from 45 to 60 s.

Table 3

Values of the individual resolution for different gradient times determined by RP-HPLC

Group

Substance

Gradient time

30 s

45 s

60 s

4 min

10 min

30 min

1

5-Methyl-1H-benzotriazole

4-Methyl-1H-benzotriazole

X

X

X

X

X

X

2

Cyclophosphamide

Ifosfamide

X

0.5

0.5

1.5

2.4

2.5

3

Tramadol

Desvenlafaxine

0.8

1.2

1.6

3.3

4.4

10.5

4

Estrone

Trenbolone

1.4

1.6

2.0

6.7

10.6

15.1

5

Dehydroepiandrosterone

Testosterone

X

X

X

3.3

3.8

7.2

6a

Dihydrotestosterone

Etiocholanolone

X

X

X

2.8

3.0

3.2

6b

Dihydrotestosterone

Androsterone

X

X

X

2.8

3.0

3.2

6c

Etiocholanolone

Androsterone

X

1.8

2.6

4.2

6.0

7.6

7

Lactose

Sucrose

X

X

X

X

X

1.3

8

Doxorubicin

Epirubicin

X

X

X

1.7

2.6

5.5

9

Iomeprol

Iopamidol

0.7

1.1

1.6

2.0

2.6

2.3

10

Corticosterone

Cortexolone

X

X

X

1.2

2.2

3.5

11

Cortisone

Prednisolone

X

X

X

1.4

1.8

1.8

Fig. 5

Comparison of peak capacity (black) and peak capacity production rate (light grey). Characteristically, the peak capacity increases with longer gradient times. The peak capacity production rate reaches a maximum at around 60 s

A gradient time of 4 min was selected because this time represents the range of fast separations. Furthermore, this time corresponds to the modulation time for the LC + LC concept developed by Schmitz et al. [20]. If the gradient time is increased to 4 min, the peak capacity increases significantly to 119.

The gradient times of 10 min and 30 min were selected because they cover the range of typical screening applications for complex samples. The peak capacity increases from 220 with a gradient time of 10 min to 412 with a gradient time of 30 min.

For ultra-fast separations of 30 s, the same peak capacity production rate of 12 will be obtained as if the gradient time is 30 min. This is due to the fact that a considerably higher absolute peak capacity is achieved with a gradient time of 30 min. The highest peak capacity production rate is achieved with a gradient time of 1 min. This is accompanied by the minimization of the peak width. If very low detection limits must be achieved for a quantification method, a gradient time of 1 min should be aimed for from a chromatographic point of view. However, this does not mean if a mass spectrometric detector is used the highest signal-to-noise ratio is obtained, since coeluting components of the matrix can lead to a significant signal reduction. In this case, the degree of signal reduction as a function of the gradient time must be determined using the so-called matrix effect chromatograms. For less complex matrices, however, a gradient time at which the highest peak capacity production rate is obtained is always ideal.

At a gradient duration of 30 s, a baseline separation is not achieved for any of the critical peak pairs. Only peak pair 4 (estrone, trenbolone) is separated with a resolution of 1.4, peak pairs 3 and 9 are separated with a resolution < 1.0. This means that for short analysis times of 30 s, as is typical for comprehensive 2D-LC, no sufficient separation efficiency is achieved for critical peak pairs. If the gradient time is increased to 45 s or 60 s, two or four peak pairs with a resolution > 1.5 can be separated. As already mentioned above, a gradient time of 60 s is the upper limit for the LC x LC mode.

By increasing the gradient time to 4 min, nine peak pairs with a minimum resolution of 1.5 can be separated, two further peak pairs are separated with a resolution of 1.2 and 1.4. At a gradient time of 30 min, 12 peak pairs with a resolution higher than 1.5 are separated, for the two sugars lactose and sucrose the resolution is 1.3. Hence, a resolution of over 1.5 is achieved for the majority of the peak pairs. This means that reversed phase chromatography is a generic method for the separation of isobars and isomers. However, this requires a peak capacity higher than 400, which in turn results in a correspondingly long gradient time.

Conclusion

In summary, IMS is not a generic method for the separation of small isobaric and isomeric compounds < 800 Da regardless of the specific technique, even if this is postulated in scientific publications or product brochures of the instrument manufacturers. However, it is possible to separate individual peak pairs with a high resolution (> 2.0) within the time frame of milliseconds. This approach, therefore, has great potential for high-throughput applications, since chromatographic (pre-)separation is not required in principle. For corticosterone/cortexolone and cortisone/prednisolone, a chromatographic baseline separation was obtained for gradient times between 4 and 10 min. With the DMA system, it was possible to achieve a resolution above 2.0 within a few milliseconds for these peak pairs. In addition, IMS needs no time for equilibration after sample measurement. This means that, with the exception of the sample introduction, no delay time between the individual measurements is required.

For LC separations, the peak capacity is a function of the gradient time. RP-LC, therefore, is a generic approach to enhance the resolution for critical peak pairs simply by increasing the gradient time. Longer gradient times generally result in higher peak capacities. In techniques such as DMS and DMA, the peak capacity is based on the simplified formula \({n_{\text{c}}}={\text{CV/FHWM}}\) and depends mainly on the applicable CV range. By changing the architecture of the electrodes, higher CV could be achieved. With changes in selectivity using dopants, the FWHM can be reduced.

As demonstrated with tramadol and desvenlafaxine, separation was feasible using 2-propanol as a dopant. One approach for increasing the number of separated groups could, therefore, be the use of various dopants. Their impact on resolution should be studied. While it will be impossible to achieve optimal conditions for the separation of all analytes in a single method, it might be practicable to select those dopants that will have the highest impact on the separation of the respective groups.

Notes

Acknowledgements

The authors would like to thank for financial aid support the German Federal Ministry for Economic Affairs and Energy within the agenda for the promotion of industrial cooperative research and development (IGF) on the basis of a decision by the German Bundestag. The access was opened by the German Federation of Industrial Research Association—AiF—and its member organisation Environmental Technology in short- member organization Environmental Technology (IGF Project No. 18861N). Special thanks to the technological SME SEADM, Dr. Marcus Winkler from Waters Corporation and Dr. Michael Schlüsener from the Bundesanstalt für Gewässerkunde for the kind opportunity to perform the resolution measurements on their systems. Additional thanks for the scientific exchange to Dr. Michaela Wirtz, Dr. Stefan Zimmermann and Dr. Terence Hetzel.

Compliance with ethical standards

Conflict of Interest

The authors declare that they have no conflict of interest.

Supplementary material

10337_2018_3640_MOESM1_ESM.pdf (1.3 mb)
Supplementary material 1 (PDF 1297 KB)

References

  1. 1.
    O’Shaughnessy J, Miles D, Vukelja S et al (2002) Superior survival with capecitabine plus docetaxel combination therapy in anthracycline-pretreated patients with advanced breast cancer: phase III trial results. J Clin Oncol 20:2812–2823.  https://doi.org/10.1200/JCO.2002.09.002 CrossRefGoogle Scholar
  2. 2.
    Monsuez JJ, Charniot JC, Vignat N, Artigou JY (2010) Cardiac side-effects of cancer chemotherapy. Int J Cardiol 144:3–15.  https://doi.org/10.1016/j.ijcard.2010.03.003 CrossRefGoogle Scholar
  3. 3.
    Kaklamani VG, Gradishar WJ (2003) Epirubicin versus doxorubicin: which is the anthracycline of choice for the treatment of breast cancer? Clin Breast Cancer 4(suppl 1):S26–S33.  https://doi.org/10.3816/CBC.2003.s.012 CrossRefGoogle Scholar
  4. 4.
    European Commission (2015) Commission Implementing Regulation (EU) 2015/495 of 20 March 2015 establishing a watch list of substances for Union-wide monitoring in the field of water policy pursuant to Directive 2008/105/EC of the European Parliament and of the Council. Off J Eur Union L78/40:20–30. http://eur-lex.europa.eu/pri/en/oj/dat/2003/l_285/l_28520031101en00330037.pdf. Accessed 15 June 2018
  5. 5.
    Aalberg L, Clark CR, Deruiter J (2004) Chromatographic and mass spectral studies on isobaric and isomeric substances related to 3,4-methylenedioxymethamphetamine. J Chromatogr Sci 42:464–469.  https://doi.org/10.1093/chromsci/42.9.464 CrossRefGoogle Scholar
  6. 6.
    Kocic D, Farrell W, Dennis GR, Shalliker RA (2016) Ultra-fast HPLC MS analyses using active flow technology columns. Microchem J 127:160–164.  https://doi.org/10.1016/j.microc.2016.03.002 CrossRefGoogle Scholar
  7. 7.
    Hetzel T, Blaesing C, Jaeger M et al (2017) Characterization of peak capacity of microbore liquid chromatography columns using gradient kinetic plots. J Chromatogr A 1485:62–69.  https://doi.org/10.1016/j.chroma.2017.01.018 CrossRefGoogle Scholar
  8. 8.
    Eiceman GA, Karpas Z, Hill Herbert H, Jr (2016) Ion mobility spectrometry, 3rd edn. CRC Press, Boca Raton, FloridaGoogle Scholar
  9. 9.
    Pu Y, Ridgeway ME, Glaskin RS et al (2016) Separation and identification of isomeric glycans by selected accumulation-trapped ion mobility spectrometry-electron activated dissociation tandem mass spectrometry. Anal Chem 88:3440–3443.  https://doi.org/10.1021/acs.analchem.6b00041 CrossRefGoogle Scholar
  10. 10.
    Mason EA, McDaniel EW (1988) Transport properties of ions in gases. Wiley, New YorkCrossRefGoogle Scholar
  11. 11.
    Kasner E, Hunter CA, Ph D et al (2013) The influence of drift gas composition on the separation mechanism in traveling wave ion mobility spectrometry: insight from electrodynamic simulations. Int J Ion Mob Spectrom 70:646–656.  https://doi.org/10.1002/ana.22528.Toll-like Google Scholar
  12. 12.
    Shvartsburg AA, Li F, Tang K, Smith RD (2006) High-resolution field asymmetric waveform ion mobility spectrometry using new planar geometry analyzers. Anal Chem 78:3706–3714.  https://doi.org/10.1021/ac052020v CrossRefGoogle Scholar
  13. 13.
    Zimmermann S, Barth S, Baether WKM, Ringer J (2008) Miniaturized low-cost ion mobility spectrometer for fast detection of chemical warfare agents. Anal Chem 80:6671–6676.  https://doi.org/10.1021/ac800559h CrossRefGoogle Scholar
  14. 14.
    Oberreit D, Rawat VK, Larriba-Andaluz C et al (2015) Analysis of heterogeneous water vapor uptake by metal iodide cluster ions via differential mobility analysis-mass spectrometry. J Chem Phys 143:.  https://doi.org/10.1063/1.4930278
  15. 15.
    Dolan JW, Snyder LR, Djordjevic NM et al (1999) Reversed-phase liquid chromatographic separation of complex samples by optimizing temperature and gradient time: I. Peak capacity limitations. J Chromatogr A 857:1–20.  https://doi.org/10.1016/S0021-9673(99)00765-7 CrossRefGoogle Scholar
  16. 16.
    Tolmachev AV, Clowers BH, Belov ME, Smith RD (2009) Coulombic effects in ion mobility spectrometry. Anal Chem 81:4778–4787.  https://doi.org/10.1021/ac900329x CrossRefGoogle Scholar
  17. 17.
    Morrison KA, Bendiak BK, Clowers BH (2018) Assessment of dimeric metal-glycan adducts via isotopic labeling and ion mobility-mass spectrometry. J Am Soc Mass Spectrom 29:1638–1649.  https://doi.org/10.1007/s13361-018-1982-2 CrossRefGoogle Scholar
  18. 18.
    Waraksa E, Perycz U, Namieśnik J et al (2016) Dopants and gas modifiers in ion mobility spectrometry. TrAC Trends Anal Chem 82:237–249.  https://doi.org/10.1016/j.trac.2016.06.009 CrossRefGoogle Scholar
  19. 19.
    Haun J, Leonhardt J, Portner C et al (2013) Online and splitless NanoLC × CapillaryLC with quadrupole/time-of- flight mass spectrometric detection for comprehensive screening analysis of complex samples. Anal Chem 85:10083–10090.  https://doi.org/10.1021/ac402002m CrossRefGoogle Scholar
  20. 20.
    Stephan S, Jakob C, Hippler J, Schmitz OJ (2016) A novel four-dimensional analytical approach for analysis of complex samples. Anal Bioanal Chem 408:3751–3759.  https://doi.org/10.1007/s00216-016-9460-9 CrossRefGoogle Scholar

Copyright information

© Springer-Verlag GmbH Germany, part of Springer Nature 2018

Authors and Affiliations

  1. 1.Institut für Energie- und Umwelttechnik e. V.DuisburgGermany
  2. 2.Instrumental AnalysisNiederrhein University of Applied ScienceKrefeldGermany
  3. 3.CURRENTA GmbH & Co. OHGDormagenGermany

Personalised recommendations