Perspective Chapter: Advantages of Ion Mobility Coupled with HPLC/UPLC

Abstract

Ion mobility is a new separation technique that can be coupled with high performance liquid chromatography (HPLC) or ultra-performance liquid chromatography (UPLC). Variances in cross-sectional ionic areas of different molecules create differential speeds through a gas allowing for millisecond separations. Combining ion mobility with both liquid chromatography and mass spectrometry with fragmentation, separations can be achieved on the second (HPLC), millisecond (ion mobility), and microsecond (mass spectrometry) timescales. This orthogonal separation greatly cleans up mass spectral data of co-eluting peaks from the liquid chromatography and adds to the descriptive data of each ion. With descriptive data such as retention time, cross-sectional area, m/z ratio, and mass spectral fragmentation, many options become available for analytical analysis. Options ranging from descriptive data collation into instrument libraries to sensitivity enhancement for trace analysis will be explored in this chapter along with the description of different forms of ion mobility.

1. Introduction

This chapter will explain the basic principles around ion mobility along with some different forms of ion mobility and how they function. Advantages and disadvantages of each technique will be offered. Case studies will demonstrate the effectiveness of using ultra-performance liquid chromatography (UPLC) paired with ion mobility and mass spectrometry. The drift time or the time required to traverse the ion mobility cell can be used to align and differentiate the mass spectra of isobaric species in complex matrices. This differentiation may allow for better quantitation and/or sensitivity enhancement of those features. In addition, assignment of descriptive data such as collision cross sections can be used to form databases for targeted and untargeted analysis.

2. Ion mobility explained

One of the most common applications of ion mobility in everyday life is swabbing luggage and hands at airports for nitrate-based explosives [1]. It is portable and fast with high volume and highly reproducible results. The fundamental concept of ion mobility is ionic separation in a gas with an applied electric field. As the ions are propelled down a drift tube by an electrical field, the ions hit the gas molecules, and their velocity slows down based on the number of gas molecules that interact with the ions [2]. Figure 1 shows the basic concept behind the ability of different sized ions moving at different rates through a drift tube. Drift time or the time it takes ions to travel the entire length of the drift tube would be the observed measurement from the ion mobility spectrometer. The drift time and other parameters can be used to calculate the collision cross section (CCS) of each compound to be used as compound-specific descriptive data. The larger the cross-sectional area, the more interaction there are with the gas molecules (lower mobility) and the slower the ion moves through the drift tube. When the cross-sectional area of the ion is smaller, there is less interaction with the gas molecules (higher mobility), and the ion can move faster compared with the larger ions [2, 3]. The next sections will describe different types of ion mobility spectrometers along with reported advantages and disadvantages of each technique. The details of each technique allow for a more informed decision when purchasing a system with a specific ion mobility separation technique. Important facts to think about would be isomer separation, ion mobility aligned spectra, and CCS fingerprinting for databases [2].

Figure 1.

Ion mobility chamber separating ions based on mobility.

3. Types of ion mobility

3.1 DTIMS

Drift tube ion mobility spectrometry (DTIMS) is potentially the simplest of the ion mobility techniques based on a simplified structural and parameter interplay [2, 4]. The drift tube is filled with a defined volume of buffer gas with no directional flow, which acts as a velocity regulator to moving ions [4]. The ions are propelled through the drift tube by a uniformly applied, static electric field with a decreasing voltage as the ions traverse the drift tube [2, 4]. As seen in Figure 1, the drift tube is made up of stacked electrodes that allow for the electric field to be applied over the length of the tube. The static electric field, defined drift tube length, and time it takes to traverse the entire length of the drift tube allow for ionic separation and CCS calculations [4]. One of the disadvantages of the DTIMS system is the voltage drop across the length of the tube. In order to create a good separation, the tube length can be increased, but that would require a higher electric field at the beginning of the drift tube to propel the ions a greater distance [2, 5].

There are many advantages to DTIMS. Unlike traveling wave ion mobility spectrometry (TWIMS), DTIMS does not require calibration with a complex mixture of compounds with well-defined CCS values in order to reproducibly measure CCS values [6]. The well-defined CCS values for TWIMS calibration were first acquired on a DTIMS instrument. In some instances, CCS calibration is required, but no complex mixture is needed to increase CCS reproducibility in DTIMS. Even without calibration, a comprehensive drift time library of many analytes can be collected in one experiment [2].

This comprehensive drift time library is limited by the duty cycle of the DTIMS systems. The duty cycle or the time in between ion trapping, ion separation, ion detection, and another cycle is shorter compared with ion mobility techniques that do not trap [6]. This trapping and analysis of ion groups limit how many ions that can be separated at any one time versus a continuous stream of ions entering the ion mobility system [7]. Instead of waiting for the duty cycle to finish, some vendors have been sending packets of ions into the drift tube one after the other before the end of the duty cycle in order to measure overlapping drift time experiments [8].

3.2 TWIMS

The traveling wave ion mobility spectrometry (TWIMS) drift tube is structurally similar to the DTIMS tube with stack electrodes propelling an ion with an electric field through a fixed volume of buffer gas [2, 6]. The electrical field oscillates continuously with no voltage drop versus a static electrical field with decreasing voltage over the length of the tube. This oscillation creates oscillating voltage waves that push the ion through the drift tube in a manner similar to surfers on the top of waves in the ocean [6]. The top of the wave carries the ion for a period until the mobility of the ion slows its velocity and allows the wave to pass it. The ion continues to be propelled by subsequent waves until it reaches the detector. An ion with a small collision cross section will ride the wave farther than ions with larger collision cross sections. This also means that ions with larger collision cross sections will require more waves to push it through the drift tube due to the slower velocity [6].

Unlike the DTIMS, the TWIMS needs to be calibrated with a complex mixture of ions that have known drift times. This allows for a continuous measurement of the ions with predetermined parameters based on the calibrated standards. Most likely the standards would include compounds with a wide range of CCS values along with a wide mass range [6, 9]. The ion mobility pressure chamber is locked before calibration, and if pressure changes occur after calibration, then the calibration will have to be repeated to account for the slight pressure change [9]. If there are slight differences in CCS values between instruments, the fundamental parameters of ion mobility should be compared.

Some important advantages of TWIMS are low-voltage requirements due to a constant wave height in an oscillating electrical field and voltage modulation over long drift tube lengths to maintain ion flow [2, 10]. Both of these qualities allow for the movement of ions across longer distances, which increase the interactions with the buffer gas and increase peak resolution. Longer path lengths would not be possible without the low-voltage requirements. Remember that the DTIMS has a static electric field with a linearly decreasing voltage over the length of the tube. If a DTIMS tube length was increased, the voltage would also increase in order provide enough momentum to push the ions through the drift tube into the detector [11].

Cyclic ion mobility systems are an extension of the TWIMS system except the length of the drift tube has the potential to vary based on how many circular passes it takes in the cyclic mobility tube. The hardware has ion guides directing ions to the circular ion mobility cell perpendicular to the main body of the system. It is similar to a trap and allows the user to choose how many circular revolutions it will take. This ability to customize the drift tube length allows for better separation/customization and has the potential to create ion mobility separation far surpassing previous instrument resolution [12].

3.3 TIMS

Trapped ion mobility spectrometry (TIMS) is a relatively new commercialized product. It uses a nonuniform electrical field unlike DTIMS to trap ions. Both DTIMS and TWIMS have constant gas volumes with no flow with the only movement caused by the electric field and voltage changes. TIMS uses both buffer gas flow toward the detector and electric field changes over the length of the ion mobility cell to propel the ions to the MS detector [13, 14]. The electric field can be tuned to guide the ions to the detector.

TIMS is composed of three regions: entrance funnel, ion mobility tube, and the exit funnel. The entrance funnel focuses and compacts the ions using an ion funnel into the ion mobility cell [15]. When the ions enter the mobility cell after focusing, the DC electric field at the exit is set higher than the potential in the mobility cell and at a 180-degree angle. This creates a field, which repeals the ions near the exit. The ions are trapped because the air flow pushes them toward the exit, and the DC field pushes them away from the exit [15]. In addition, a low-intensity electric field starts to gradually increase over the length of the mobility cell. The ions are trapped and separated based on their size-to-charge ratio [13, 15]. The ions with the lowest mobility or the largest size-to-charge ratio will congregate toward the exit with the highest electric field. This is caused by the gas molecules interacting more with the largest compounds, thus pushing them closer to the exit. The compounds with the highest mobility or the smallest size-to-charge ratio will be farther from the exit because of the opposing field at the exit pushing it away [13]. The smaller ions interact less with the gas molecules causing them to move less downstream and more away from the exit. The ions can exit the mobility cell by lowering the electric field intensity gradually at the exit. This will cause the lowest mobility ions to exit first due to the gas flow and the highest-mobility compounds last. The gradual decrease in the electric field intensity allows for a segmented elution of compounds based on the parameters of the experiment. Having the lower-mobility compounds come out first is opposite to the behavior of DTIMS and TWIMS [15].

Unlike DTIMS and TWIMS, which allow for scanning to see multiple ion mobilities with the same experimental conditions, TIMS requires sequential experimental parameter changes to see multiple ions injected into the mobility chamber. As stated above, there is a gradual decrease in the electric field at the exit, which causes a segmented release of the ions [13]. This change in voltage only allows certain mobilities to exit at a time without allowing higher-mobility ions to exit at the same voltage. TIMS is a highly selective technique relating to separation efficiency, but the ability to scan is lost or needs extra parameter changes. The trapping ability allows the TIMS to maintain peak separation despite short path lengths and short lab instrument footprints [15].

3.4 FAIMS

Field asymmetric ion mobility spectrometry (FAIMS) is an atmospheric pressure ion mobility technique. This technique uses both high and low electric fields to separate ions in the gas-phase mobility cell. The FAIMS device is small, and it can potentially be placed in between the ion source and the vacuum inlet of the mass spectrometer [2, 16]. The integration onto MS systems without ion mobility is achieved by small accessories to maintain the required negative pressure in the mass spectrometer. Like TIMS, gas flow is used for ion movement in addition to the electric field. In FAIMS, the electric field alternates both in strength and polarity to separate ion according to field strength changes and not drift time mobilities over the length of the mobility cell [16].

FAIMS cannot provide CCS values due to a lack of mobility measurement and change in ion structure due to the changes in electric field strength and polarity. This technique cannot scan for ions with multiple electric field strength changes but rather specific electric field changes. This can allow a continuous monitoring of ions with the same electric field change response. For compounds belonging to similar chemical classes, this would be beneficial because the signal-to-noise ratio would increase [2, 17].

4. Case studies

4.1 Case study 1

The Yassin et al. [18]study pairs ion mobility, UPLC and mass spectrometry together allowing for the characterization of many polyphenolic compounds in tea. The reversed-phase UPLC separation yielded chromatographic peaks, but when coupled with IMS separation, additional peaks were revealed. This section will summarize how ion mobility, liquid chromatography, and MS fragmentation can be used to characterize structural features of unknowns.

The data were acquired in negative mode using MS^e mode to collect low- and high-collision energy spectra during the same acquisition. The traveling wave ion mobility was used to collect drift times and calculate collision cross sections. The drift time, pseudomolecular ion m/z, and mass spectral fragments were then aligned to the LC elution time. This is very useful in complex mixtures because mass fragments of two isobaric species may overlap with no indication to which pseuodmolecular ion the fragments belong.

Theasinensin C (TS) and proanthocyanidin B (PA) were used as examples of polyphenolic isomeric species in this study (Figure 2). Assam and Sri Lanka teas were analyzed for the presence of TS isomers. The levels of PA were compared in extracts of Ziziphus spina-christi and Rhododendron. The extracted ion chromatograms of each extract used a nominal mass search of 609 Da with resulting accurate masses ranging from 609.156 Da to 609.183 Da. Characteristic fragmentation patterns of TS and PA standards were created for family classification of any potential new isomers.

Figure 2.

Structural differences in isomers Theasinensin C and Proanthocyanidin.

The Assam tea extract benefited from having UPLC and ion mobility together. Ion mobility was able to show that a 609 Da feature seen at 1.28 min contained two compounds with distinct drift times of 5.18 ms and 5.56 ms and two distinct fragmentation patterns characteristic of the TS family. In this case, the ion mobility was able to separate the features the UPLC could not. In addition, the UPLC separation showed additional 609 Da features at 10.09 min and 10.22 min with the same drift time of 5.67 min. Without the UPLC separation, these two compounds would have been classified as one isomer and not two. The same features were found in the Sri Lanka tea.

According to this investigation, drift times close in value and at the same LC retention time were investigated further. Under normal conditions, the gas is heated in the ion mobility tube, and this elevated temperature may cause structural changes between two isomeric forms. There would be an equilibrium between both diasterioisomers each with its own similar but distinct collision cross section, which would cause two peaks to appear in the ion mobility plot. The investigators looked at the MS^e fragmentation of the TS isomers at both 5.18 ms and 5.56 ms. The isomer at 5.56 ms had a loss of 18 Da characteristic of water, whereas the TS isomer at 5.18 ms did not lose water. PAs, which are similar in structure to TSs, have been shown to lose water from epicatechin moieties, whereas PAs with catechin moieties do not loose water. To test whether this was also true with the TS family, the investigators employed molecular modeling. A correlation was established between the isomer drift times and their calculated collisional cross sections. The data showed that the isomer with 5.18 ms drift time had a collisional cross section of 154 Å² and a trans stereochemistry from a catechin building block. The isomer at 5.56 ms had a calculated collisional cross section of 157 Å² with a cis stereochemistry from an epicatechin moiety. The confirmed epimerization was a drawback of ion mobility in this study because the TS could not definitively be assigned a cis or trans designation at retention time of 1.28 min.

The proanthocyanidins (PAs) in both the Ziziphus spinae-christi and Rhododendron extracts showed interesting results for UPLC separation with ion mobility. The Ziziphus spinae-christi extract showed two 609 Da features at 1.67 min and 10.22 min with specific drift times at 5.35 ms and 5.62 ms. Fragmentation confirmed PA isomers. The similar drift times would have required further investigation like with the example above, but since the peaks were chromatographically well resolved, the peaks were not artifacts of the ion mobility. The Rhododendron extract showed 609 Da features at 2.35 min and 4.98 min with specific drift times of 5.24 ms and 8.37 ms. Fragmentation confirmed very similar PA isomeric structures with different drift times, and this suggested the presence of regioisomers. Putative identifications of regioisomers epigallocatechin-(4,8)-epigallocatechin (PA) and epigallocatechin- (4,6)-epigallocatechin were suggested. Computational modeling data agreed with the experimentally observed collisional cross sections of 154 Å² (5.24 ms drift time) for the epigallocatechin-(4,8)-epigallocatechin and 178 Å² (8.37 ms drift time) for the epigallocatechin-(4,6)-epigallocatechin (Figure 3).

Figure 3.

Connectivity of proanthocyanidin isomers and different CCS values.

The combination of ion mobility, UPLC chromatography, and mass spectral fragmentation allowed the separation of isomers that may not have been separated in any single technique. Excess heat in the ion mobility drift tube was shown to be one drawback to the technique because it caused epimerization. In addition, descriptive data provided by each technique helped guide structural confirmation.

4.2 Case study 2

A further look into pairing liquid chromatography, ion mobility, and mass spectrometry together for chemical profiling is described below in a work by McCullagh et al. [19]. Analytical approaches that provide the maximum amount of descriptive are needed for complex herbal extracts for product authentication, extract profiling, stability/degradation, and purity analysis. This study investigated the genus Passiflora for the many flavanoid derivatives such as C-glycosylflavone specifically apigenin/luteolin derivatives, which may have medicinal properties. Main areas of emphasis will be the analysis of 6-C and 8-C glycosides to establish exact mass fragmentation, LC chromatographic separation, and ion mobility collision cross sections (CCS) to decrease sample complexity in a herbal extract matrix. An additional goal was to establish a CCS searchable database with attached fragmentation and LC data. Fragmentation and LC data are more variable between systems, whereas the CCS values should be more reproducible.

The Passiflora species evaluated include P. edulis, P. alata, P. incarnata, and P. caerula. Each sample was extracted, and the flavone fraction was purified with solid-phase extraction. Reversed-phase chromatography was performed using C18 chromatography. The data were acquired in both positive and negative modes using MS^e to collect low- and high-collision energy spectra during the same acquisition. The low energy was 4 eV and the high energy was a ramp from 30 to 75 eV. Traveling wave ion mobility was used to collect drift times and calculate collision cross sections. The pseudomolecular ion m/z and mass spectral fragments were then aligned to the retention time and drift time to create a database with calculated CCS values.

This study incorporated targeted profiling of flavanoid derivatives in the genus Passiflora to establish a non-targeted approach. The 6-C and 8-C-glycosylflavone isomers of orientin/isoorientin and vitexin/isovitexin had similar isomeric structures (Figure 4). Each standard was characterized both in pure solvent and plant extract matrix. All CCS values were based on data from replicate injections of pure standards at high and low concentrations to simulate different amounts in extract and different ionization responses in plant matrix. The CCS values created in the study were used as metrics to confirm the presence of isomeric flavanoids similar to the standards. The identity of these unknown flavanoids would not need to be known since the descriptive data were assigned to each unknown during this study.

Figure 4.

Structure differentiation between C6 and C8 pairs isoorientin/orientin and isovitexin/vitexin.

The customized database created in this study aligned accurate mass of the pseudomoleulcar ions and fragment ions to retention time and drift time. This drift time alignment helped assign fragments to the correct pseudomolecular ion. Collision cross section values were created to use with a delta ^TWCCSN₂ metric to help to differentiate flavanoids despite co-elution. Based on this study, this delta ^TWCCSN₂ metric can be used even on trace components when fragmentation data would be hard to attain.

The CCS values and retention time of each isomeric pair orientin/isoorientin and vitexin/isovitexin are displayed in Table 1. Notice that the retention times of the isomers within each pair are very close, and without drift time alignment, the mass spectral fragmentation would be blended without specific assignment to either isomer. These values were then used to calculate delta ^TWCCSN₂ taking the difference between the CCS values of each isomer within each pair. The compounds with glycosides attached at carbon-8 had smaller CCS values compared with the compounds with glycosides attached at carbon-6 (Table 1). These values are determined by their 3D structure either being more compact or spread out to change the interaction with the gas molecules as they travel through the drift tube. The CCS values for the negative mode were more different than those in the positive mode (Table 1). In the positive mode, the close chromatographic retention times and close CCS values would cause little to no baseline resolution in both dimensions. In this case, the negative mode data would allow for better differentiation of the isomers despite possible co-elution. In addition, the drift time aligned data would allow the correct quantitation of each isomer. The delta ^TWCCSN₂ method was tested with P. alata, P. incarnata, and P. caerulea, and the isomers that were found were within 0.75% error. In addition, these results were replicated over 3 years, and the delta ^TWCCSN₂ values were within 0.8% error for that time period. When this metric is used with accurate mass of parent and daughter ions and retention time, the database can be used to better analyze complex herbal extracts.

Flavanoid	Orientin	Isoorientin	Vitexin	Isovitexin
Retention Time (min)	7.88	7.83	8.52	8.40
Negative Polarity CCS (A^2)	187.7	198.1	188.8	195.5
Δ^TWCCSN2 (A^2)	10.4		6.7
Positive Polarity CCS (A^2)	200.7	203.4	198.2	199.3
^TWCCSN2 (A^2)	2.7		1.1

Table 1.

Retention times and CCS values for negative and positive MS experiments.

In the Passiflora matrix, the retention times of vitexin and isovitexin were based on the apexes of the peaks with no chromatographic baseline resolution. This was where the ion mobility aided in differentiating vitexin and isovitexin. With no baseline separation, the fragmentation patterns of both isomers were overlapping with no distinction between them. When the accurate masses of the parent and fragment ions were drift time aligned, this allowed the fragmentation patterns to be established for each of the isomers using the collision energy ramp of 30 eV–75 eV. The distinctive parts of the fragmentation were the ratios of m/z 281/282/282/284 to distinguish between vitexin and isovitexin. For vitexin, the characteristic ion ratios were m/z 281 < 282 and m/z 282 < 284. For isovitexin, the ion ratios were m/z 281 > 282 and m/z 282 > 284. The less than or greater than refers to the ion intensities. When these ions were retention time and drift time aligned, the identification of specific isomers was successful.

The retention times of orientin and isoorientin were within 0.05 min; therefore, there was no chromatographic baseline resolution. Again, the ion mobility aided in differentiating the isomers along with the elevated collision energy ramp of 30 eV–75 eV. The fragmentation patterns of both isomers were overlapping with no distinction between orientin and isoorientin. The accurate masses of the parent and fragment ions were drift time aligned allowing the fragmentation patterns to be established for each of the isomers. The distinctive parts of the fragmentation were the ratios of 284/285 and m/z 297/298/299 to distinguish between orientin and isoorientin. For orientin, the characteristic ion ratios were m/z 284 > 285, m/z 297 > 298, m/z 298 > 299. For isoorientin, the ion ratios were m/z 284 < 285, m/z 297 < 298, m/z 298 < 299. The fragments could be used as another metric to differentiate between orientin and isoorientin with retention time and drift time alignment. The increased specificity from CCS libraries may create better future characterization protocols for herbal extracts in consumer products.

Combining accurate mass, fragmentation data, IMS separation, and CCS measurements created what the authors of this study called “known-unknown” fingerprinting. The normal protocol for herbal extract characterization is to test for a small number of pure active compounds, but often these standards are in limited supply or are cost prohibitive. Using this “known-unknown” fingerprinting technique, features can be cataloged. As more descriptive data are assigned to unknown features, putative identifications may be assigned. The unknowns can also be cataloged into the “known-unknown” database to be tracked from sample to sample even if there is no confirmed identification. This technique was used in the P. caerulea species with unknown analytes present with accurate masses ranging from 431.0958 Da to 431.0983 Da with CCS values ranging from 185.5 Å² to 188.0 Å². Collecting these data along with the fragmentation data added descriptive data to this type of non-targeted analysis for future comparison to other Passiflora species.

The “known-unknown” workflow was applied to compounds with the masses of 431.09 Da and 447.09 Da or compounds similar to the orientin/isoorietin and vitexin/isovitexin isomer pairs. Nineteen different candidates in P. caerula, P. edulis, P. alata, and P. incarnata were identified with similar accurate masses, fragmentation, and CCS values as the C6/C8 glycoside isomers. All the abovementioned descriptive data allowed putative identification of C6 or C8 glycosidic isomers. As an example, two isomers with similar retention times shared accurate mass of 431.098 Da, but one had a higher CCS value than the other. The higher CCS value would suggest a C6 glycoside, and to confirm, the fragmentation data were evaluated. The C6 glycoside had fragment ion ratios with m/z 284 < 285 and m/z 297 < m/z 298 > m/z 299. The C8 glycoside had fragment ratios with m/z 284 > m/z 285, and m/z 297 > m/z 298 < m/z 299. These unknowns were similar to isovitexin/vitexin, but not the same since the retention times were different than the standards.

Other than the “known-unknown” characterization, a system with ultra-performance liquid chromatography, ion mobility, and mass spectrometry (UPLC-IM-MS) can be used to better quantify convoluted peaks. When isobaric species are present, the peak areas overlap with no demarcation of either peak, but when ions are drift time aligned, isomeric quantitation can occur. The drift time alignment separated the peak areas of each isomer, allowing for the quantitation of each isomer even in the presence of a complicated matrix. Using this concept, the authors quantitated isoorientin, orientin, isovitexin, and vitexin in P. caerula, P. edulis, P. alata, and P. incarnata. To further show this concept, the authors calculated the concentrations with and without drift time alignment. As expected, the isobaric species had greatly different calculated concentrations with and without drift time alignment.

In this study, the authors demonstrated the effectiveness of using UPLC paired with ion mobility and mass spectrometry. The drift time alignments allowed for differentiation of mass spectra of isobaric species. This differentiation also allowed for better quantitation of isobaric species. The calculated CCS values aided in the formation of a database of “known-unknowns” that could be tracked between herbal extracts despite the unconfirmed identities of some compounds.

4.3 Case study 3

Adams et al. [20] created an liquid chromatography trapped ion mobility spectrometry with mass spectrometry (LC-TIMS-MS) technique to provide a high-throughput orthogonal separation technique for isomeric opioids in the complex matrix of urine. Three groups of isomeric opiods and deuterated analogs were monitored at trace levels in human urine despite the possible matrix interferences from the urine. As with previous studies, retention time, CCS, and accurate mass were descriptive data used to identify and monitor the compounds of interest in the analyses. The high selectivity of the TIMS when paired with the LC and MS allowed for low detection levels comparable to liquid chromatography with tandem mass spectrometry (LC-MS/MS) and even with potentially less false-positives based on shared multiple reaction monitoring transitions between the isomers.

The opioids used in this study were as follows: 6-acetylmorphine (6-AM), naloxone, codeine, hydrocodone, morphine, hydromorphone, norcodeine, norhydrocodone, and the deuterated versions of each of these opioids (Figure 5). Calibration curves of the standard mixes and internal standard mixes were diluted with urine to create matrix match standards. This calibration curve in urine was created using liquid chromatography trapped ion mobility spectrometry with mass spectrometry (LC-TIMS-MS). LC separation was performed used a reversed-phase monolithic C18 column. Positive-mode electrospray ionization was used to ionize the opioids. The TIMS unit required specific parameters for gas flow and voltages at the entrance and exit funnels along with the mobility cell. An important parameter for the TIMS was the voltage at the exit funnel. As the voltage is decreased gradually at the exit funnel, the ions would exit based on the lowest mobility (largest CCS) first and the highest mobility (smallest CCS) last. Calibration standards for the instrument were used to calibrate voltage to CCS values. The theoretical CCS values of the opioids were calculated to predict the appropriate voltages needed to eject the ions from the mobility chamber.

Figure 5.

Structures of opiod and opiod derivatives.

The ion mobility of each compound was measured based on the calibration standards. 6-AM and naloxone had the highest CCS values of 176.7 and 171.1 Å². The orientation of the acetyl group on the 6-AM increased the CCS compared with the carbonyl group on the naloxone. The aliphatic group on the naloxone also extended out from the opioid body allowing the CCS to be larger compared with the rest of the opioids. In addition, computational modeling showed the tertiary amine having different orientations between 6-AM and naloxone. Even with this small difference in CCS values, the TIMS was able to separate these ions. Based on computational modeling, codeine and hydrocodone had CCS values of 168.2 and 167.8 Å². The hydroxyl versus carbonyl on carbon 6 was the only structural difference between these compounds, which led to small differences in CCS values and no separation in TIMS. There was no separation in the TIMS for morphine and hydromorphone due to the small difference in CCS values (162.9 and 163.3 Å²) caused by small structural differences in the hydroxyl and carbonyl groups. Norcodeine and norhydrocodone were not separated in the TIMS due to the small difference in CCS values (167.9 and 167.4 Å²) caused again by small structural differences in the hydroxyl and carbonyl groups. The opioid pairs of morphine/norcodeine and hydromorphone/norhydrocodone can be separated on TIMS due to a 5 Å² difference caused by a difference in secondary and tertiary amine orientation. The specificity in TIMS allowed for good peak resolution. The baseline mobility separations between 6-AM and naloxone, hydromorphone and norhydrocodone, and morphine and norcodeine were achieved with fast and slow scan TIMS. Despite the baseline separations of some opioid pairs, not all opioids were separated from each other in the calibration mix.

The study added liquid chromatography to the separation to obtain separation of the opioids with similar CCS values. Matrix interferences with water and urine in the TIMS with and without the LC were evaluated. Some matrix interferences could not be resolved with TIMS alone. Spiking the standards and internal standards into the urine and water showed increased limits of detection caused by interfering compounds. The liquid chromatography was successfully implemented to separate most matrix interferences with the standards and internal standards when run with TIMS. There were still slight LOD increases in urine using LC-TIMS-MS, but those increases were also seen while using LC–MS.

The LC runtime allowed for separation of the opioids. The internal standards of each opioid were deuterated and thus had a higher m/z when separated in the mass spectrometer despite having the same retention time and CCS values of their non-deuterated analytes. The internal standards were used as quality control checks of the retention times with the deuterated standards of each analyte containing different quantities of deuterium. Naloxone and 6-AM were successfully separated with LC and TIMS with retention times of 6.85 and 7.00 minutes. Hydrocodone and codeine were not successfully separated using TIMS, but the LC produced near-baseline separation with retention times of 6.8 and 7.0 minutes. Norhydrocodone and norcodeine had near-baseline resolution on the LC with retention times of 6.9 and 7.0 minutes despite having no separation in the TIMS. The retention times for morphine and hydromorphone were not mentioned in the LC-TIMS-MS analysis. Compounds with the same retention times can be separated in the TIMS based on previously mentioned CCS values. There was good reproducibility of CCS, LC retention times, and m/z values in between experiments and among different calibration levels of the calibration curve. The relative percent deviation (RPD) was <0.5% for CCS values with and without urine. Since the CCS values did not change based on concentration, this helped emphasize that CCS values can be used effectively for qualitative analysis in addition to retention time and accurate mass. Both additional data points showed low variability between water and urine, low and high concentrations, and inter-day performance. The LC-TIMS-MS protocol showed successful separation and quantitation for low ng/mL concentrations.

Combining LC with TIMS-MS improved peak resolution for instances when TIMS was not enough to separate the opioids based on very similar CCS values. In addition, the CCS values were very consistent with the drift tube ion mobility measurements and were consistent between experiments. This high reproducibility and specificity of the TIMS were very important when the opiods co-eluted in the LC dimension. Both the LC and TIMS techniques were crucial in order to separate compounds in this study.

5. Conclusions

This chapter has introduced the reader to basic ion mobility along with benefits and disadvantages of some ion mobility techniques. In addition, three studies were evaluated to show the benefits of coupling UPLC to ion mobility-MS techniques. This coupling allowed the separation of isobaric isomers that required an orthogonal separation. Descriptive data such as CCS and mass spectral fragmentation with drift time alignment helped guide structural confirmation and the formation of databases for tracking even if the compound identity was not confirmed. The drift time alignment also allowed for better quantitation of isobaric species. Ion mobility is not a perfect technique as seen in some of the above studies, but when paired with LC chromatography, the analytical strength of a method significantly increases.

Conflict of interest

The authors declare no conflict of interest.