Evaluating Dynamic Traveling Wave Profiles for the Enhancement of Separation and Sensitivity in Traveling Wave Structures for Lossless Ion Manipulations

Abstract

The amenability of traveling wave ion mobility spectrometry (TWIMS) to extended separation pathlengths has prompted a recent surge of interest concerning the technique. While promising, the optimization of ion transmission, particularly when analyzing increasingly disparate species, remains an obstacle in TWIMS. To address this issue, we evaluated a suite of dynamic TW profiles using an original TW structures for lossless ion manipulations (TW-SLIM) platform developed at Washington State University. Inspired by the range of gradient elution profiles used in traditional chromatography, three distinct square TW profiles were evaluated: a static approach which represents a traditional waveform, a dual approach which consists of two distinct TW profiles within a given separation event; and a ramp approach which varies TW speed and amplitude at a fixed rate during separation. The three waveform profiles were evaluated in terms of their impact on separation (quantified as resolution) and sensitivity (quantified using signal-to-noise ratio (SNR), and ion abundance). Concerning separation, the highest resolution (R) was observed when operating with the static waveform ($\mathrm{R}=7.92$); however, the ramp waveform performed comparably ($\mathrm{R}=7.70$) under similar conditions. Regarding SNR, optimum waveform profiles were species dependent. Bradykinin²⁺ displayed the largest gains in SNR (36.6% increase) when ramping TW speed, while the gains were greatest (33.5% increase) for tetraoctylammonium when modulating TW amplitude with the static waveform. Lastly, significant (>10%) increases in the abundance of tetraoctylammonium ions were observed exclusively when utilizing a ramped waveform. The present set of experiments outline the results and challenges related to optimizing separations using alternative TW profiles and provides insight concerning TW-SLIM method development which may be tailored to enhance select analytical metrics.

1. Introduction

Ion mobility spectrometry (IMS) is a gas-phase analytical technique that separates analytes based on their ionic mobilities as they are propelled through a gas under the influence of an externally applied electric field.^1,2 Lending to relatively simple construction and rapid analysis, much of the early implementations of IMS concerned matters of threat detection and field deployment.^3–5 However, developments in the field of IMS, particularly those concerning its coupling with mass spectrometry, have led to an expansion of IMS applications to include the investigation of biomolecules and even the separation of isomers.^6–8

While the most commonly deployed IMS approach is drift tube IMS (DTIMS) which separates ions using linear, homogeneous fields, other iterations of IMS instrumentation exist, such as field asymmetric IMS (FAIMS)⁹, trapped IMS (TIMS),^10,11 and traveling wave IMS (TWIMS).^12,13 TWIMS in particular has seen a recent spike in interest in part due to the development of traveling wave structures for lossless ion manipulations (TW-SLIM), which utilizes planar printed circuit boards (PCB) to effectively confine and separate ions.¹⁴ While TW-SLIM differs from other TWIMS approaches in terms of construction, the nature of ion separation via traveling waves remains relatively consistent across the techniques. In-depth explanations of ion behavior in response to traveling waves are outlined elsewhere.^15–17 Briefly, ions exposed to dynamic electric fields, which are in essence transient potential wells, attempt to “keep up” with the wells which are moving at a user-defined rate. This rate is often referred to as TW speed ($\mathrm{s}$). The ability of ions to remain within a given well and not fall to a preceding well is dependent upon their velocity ($\mathrm{v}$), which is a function of the ions’ mobilities and the depth of the well. Well-depth is defined by the user and is often referred to as TW amplitude. The ratio of ion velocity to traveling wave speed is often expressed as $\mathrm{c}$ (eq. 1):

$$c=\frac{v}{s}$$ (1)

Ions whose velocities approach the TW speed will fall over to preceding wells less frequently than those with lower velocities (i.e., lower mobility species), which leads to mobility-based separations that are related to the rate of ion rollover.

A profound benefit of TWIMS is its use of electric fields which are defined by localized minima and maxima, as opposed to the absolute maximum and minimum field strengths found in DTIMS systems, which ultimately limit their length and subsequent separation power. Recent work has seen this unique flexibility with regard to field strengths leveraged to achieve separation pathlengths in excess of 10 m, far exceeding the length of a typical DTIMS system.^18,19 Additionally, TWIMS experiments using cyclic separations continue to grow in popularity, with both the Waters Cyclic-IMS and TW-SLIM platforms enabling theoretically infinite pathlengths for separation.^20,21 While the unique nature of TWIMS may appeal to those interested in performing mobility separations, it is important to note that the technique is by no means a solution to all IMS shortcomings and presents its own inherent issues. Concerning issues inherent to TWIMS, the concurrent transmission of increasingly disparate species remains an obstacle. This issue stems from the use of RF confinement^22,23 and the dependence of ion motion (i.e., bobbing or riding wavefronts) on mobility (eq 1). In the greater context of IMS, an issue analogous to the general elution problem often associated with chromatography exists as static conditions which enable the separation of early “eluting” species do not necessarily provide for optimum transmission of later “eluting” species.

Recognizing the aforementioned issues concerning TWIMS, we sought to evaluate a suite of TW profiles that seek to simultaneously remedy issues of ion transmission as well as address the general elution problem. Taking inspiration from the gradient elutions employed in chromatography, we implemented a ramped waveform in which TW speed and amplitude were varied at a fixed rate during separation in hopes of providing optimum conditions for transmission and separation of both early and later arriving (i.e., higher and lower mobility) species. In order to evaluate the performance of the ramped waveform against a more traditional waveform, we performed similar experiments with a static waveform analogous to isocratic approaches in chromatography.

Lastly, a third waveform was implemented in which a sudden change in conditions was implemented immediately following the arrival of the first species. This was theorized as an approach that would enable the optimum transmission of one species before modulating to conditions that provide ideal transmission of another species.

The following discussion provides insight into the impact of the various waveforms on matters of both separation and sensitivity. Acknowledging issues associated with resolving power as a metric for evaluating TWIMS separations,^14,24 resolution was used instead (eq. 2):

$$Resolution=1.18\mathrm{*}(t2-t1)\left(\begin{array}{lll}w& 2+w& 1\end{array}\right)$$ (2)

where ${\mathrm{t}}_2$ and ${\mathrm{t}}_1$ correspond to the arrival times of two species, and ${\mathrm{w}}_2$ and ${\mathrm{w}}_1$ are the species’ full widths at half maximum (FWHM). The optimum resolution was obtained by operating with the static waveform and modulating TW amplitude, though it is worth noting that the ramped waveform did not perform significantly (>10%) worse than its static counterpart. With regard to sensitivity, signal-to-noise ratio (SNR) and ion abundance (which relates to peak area) were used to characterize the performance of the waveforms. For the two compounds used throughout the presented analyses, bradykinin²⁺ and tetraoctylammonium (T8), a combination of a ramped waveform and TW speed modulation provided optimal SNR for T8, whereas static TW amplitude modulation performed best for bradykinin²⁺. Akin to observations made regarding SNR, ideal conditions for ion abundance were not consistent amongst the analytes, with T8 benefiting most from the use of the ramped waveform and bradykinin²⁺ reporting the lowest ion abundance when using the ramp.

2. Experimental

2.1. Materials

Tetraoctyl ammonium bromide and bradykinin acetate salt were purchased from Sigma-Aldrich (St. Louis, MO). All data presented were collected using a solution of 1 μM each T8 and bradykinin prepared in methanol (HPLC grade, Fisher Chemical, Fair Lawn, NJ) with 0.1% formic acid (Honeywell-Fluka, Muskegon, MI). T8 was selected in order to gauge how the various experimental conditions would impact simpler analytes (singly charged, lack of conformers), which are popular within the field of ion mobility (e.g., explosives and narcotics).^3,4,25 Bradykinin was chosen as a representative biomolecule, though it is important to note that the peptide is neither particularly large nor complex and that further experiments would be necessary to explore how the presented methods impact increasingly complex species.

2.2. Instrumentation: Components and Conditions

A schematic of the instrument used throughout the presented work is shown in Figure 1. The instrument used throughout the presented analyses is analogous to one reported in detail elsewhere.²⁶ Briefly, ions were generated via electrospray ionization (ESI) using a 75 μm ID glass capillary emitter biased 2.2 kV relative to an adjacent heated (160 °C) stainless-steel inlet capillary (560 μm i.d., 1600 μm o.d.) purchased from McMaster-Carr (Elmhurst, IL). The sample flow rate used throughout the analyses was 3 μL/min, controlled by a Model 11 Syringe Pump from Harvard Apparatus (Holliston, MA). After traversing the inlet capillary, ions were focused by an ion funnel (897 kHz, 240 V_p-p) housed in a low-pressure (2.45 torr, ambient air) chamber before being transported to a second, smaller ion funnel (1.1 MHz, 140 V_p-p) housed in a slightly higher pressure (2.50 torr, ambient air) chamber. The pressure in each chamber was monitored using Setra Model 730 capacitance manometers (Boxborough, MA).

Figure 1.

Schematic of the instrument used throughout the presented work. A photo of the instrument is shown in Figure S1.

After traveling through the second ion funnel, ions arrived at the first TW-SLIM segment, referred to as the “drag” board, which was used for ion packet injection. The injection of stored ion packets was achieved utilizing a series of electrodes capable of switching between a blocking state, in which they were biased above the adjacent electrodes using a static DC potential, and a TW state which enabled ion transmission (i.e., injection). Throughout all experiments, the injection width was 1 ms. Immediately following the drag board was a second, distinct SLIM board dubbed the “separation board.” An investigation concerning the impact of two distinct sets of SLIM boards, as opposed to the traditional single set, will be expounded upon in future publications. While capable of enabling more advanced separation techniques, such as multi-pass and compression ratio ion mobility programming (CRIMP)^27,28, the secondary track of the separation board was not used in the presented analyses. The total path length used for separation was approximately 1.5 m. A custom Faraday plate served as the primary detector.

2.3. Data Proccessing

Current detected at the Faraday plate was first amplified using a Keithley Model 427 current amplifier (10⁹ Gain, Rise time: 1 ms, Cleveland, OH) and then subsequently recorded as a function of time using an Analog Discovery 2 (Digilent Inc., U.S). Data were processed using Python notebooks similar to those released in a previous publication.²⁶ Arrival time distributions (ATD) were averaged 75 times, and each presented data point represents an average of 5 ATD replicates. The standard deviation of the 5 replicates is representative of the measurement’s uncertainty. “Noise,” as it pertains to SNR calculations, was defined as the baseline immediately preceding and following each peak. “Ion abundance” refers to the total number of charges associated with each species and was calculated via integration of the ion signal detected at the Faraday plate. Peak identities were assigned by analyzing the individual standards of each analyte.

TW-SLIM conditions were controlled using a modular intelligent power source (MIPS) from GAA Custom Electronics (Kennewick, WA). Ions were confined laterally by guard electrodes biased to the TW at half its amplitude. Radial confinement was achieved using RF electrodes operating at 1.07 MHz, 300 V_p-p. The square TW was generated using a 4 high, 4 low electrode configuration. Control conditions for the TW were defined as 23 V_p-p, 183 m/s. Further details regarding the configuration of the SLIM boards are reported elsewhere.²⁶

2.4. Traveling Wave Profiles

Three waveform profiles were evaluated as a means of optimizing peak profiles within TW-SLIM: a static waveform ( static), a waveform with two distinct segments (dual), and a ramped waveform (ramp). A visual representation of the three waveform profiles is provided in Figure 2. The static waveform represents the simplest TW profile in which all species reach the detector under identical conditions (analogous to isocratic elutions in LC). The dual waveform consists of two distinct profiles: control conditions (maintained until after the first species (bradykinin²⁺) reaches the detector) and conditions in which either the TW amplitude was increased or the TW speed was decreased. The transition between the conditions in the dual profile occurred as rapidly as the MIPS would allow (≈300 μs). Regarding the ramp, 5 ms after injection, TW conditions were ramped to either higher amplitudes or lower speeds over the course of 120 ms. Information regarding ramp rates can be found in Table S1. Across all waveform profiles, the TW was operated at conditions that enhanced ion accumulation (TW amplitude was reduced to 16 Vp-p) for 60 ms prior to injection.

Figure 2.

Visualization of the waveform profiles evaluated. The orange portion of each waveform represents control conditions. The “ static” approach is representative of a typical TW in which all species arrive at the detector under identical conditions. The “dual” waveform was theorized as an approach that could enable the optimum transmission of one species before switching to the conditions that provide the idealized transmission of another species. Lastly, the “ramp” waveform modulates TW speed or amplitude at a fixed rate during separation and was hypothesized as a middle-ground for the transmission of disparate species.

Regardless of the waveform profile implemented, the TW-SLIM operated under control conditions for the first 5 ms from the point of injection. This was done to avoid any confounding effects stemming from the impact of TW parameters on ion injection, which have been reported elsewhere.²⁶ Additionally, despite preliminary modeling efforts suggesting that varied TW parameters do not significantly impact ions traversing the gap between board sets, the 5 ms control period enabled ions to traverse this gap prior to any waveform modulation, once again avoiding any potential confounding effects.

3. Results and Discussion

3.1. Separation

Prior to evaluating the dual and ramp waveforms, preliminary experiments were conducted to validate results reported elsewhere¹⁴ concerning the relationship between resolution and TW conditions. In agreement with Hamid et al.,¹⁴ over a given range of values, decreasing TW speed and increasing TW amplitude coincide with gains in resolution between T8 and bradykinin²⁺ (Figure 3). Having established that the presented TW-SLIM platform performed as expected when operating with the static waveform, the previous analysis was repeated, this time with the dual and ramp waveforms (Figure 3). For both the amplitude and speed experiments, similar trends were observed for the static and ramp waveforms which demonstrated gains in resolution associated with decreases in TW speed and increases in TW amplitude, while the dual waveform behaved inversely. Interestingly, the experiments varying amplitude provided greater gains in resolution for the static and ramp waveforms and mitigated the decrease in resolution associated with the dual approach. While a direct comparison cannot be made between the presented results and those published by Hamid et al.,¹⁴ it is interesting to note that in both cases, modulating TW amplitude appears to have a more profound impact on separation than TW speed. With exception to the TW speed experiments concerning the static and ramp waveforms, the data presented in Figure 3 show distinct trends associated with resolution as a function of waveform profiles ( static, dual, and ramp). It is worth noting, however, that aside from select conditions (TW speed: 110 m/s, TW amplitude: 29–31 V_p-p), no significant (>10%) variation between resolution values was observed.

Figure 3.

Variations in resolution between T8 and bradykinin²⁺ as a function of TW speed (left) and amplitude (right). Values on the x-axis represent conditions that followed the control period of the static and dual waveforms, and the value that was ramped to in the ramp waveform.

Despite the comparable resolutions observed across various conditions, a more in-depth analysis of the data associated with Figure 3 reveals substantial differences between the ATDs collected using the different waveform profiles (Figure 4). A qualitative assessment of the ATDs presented in Figures 4b and 4c suggests that each of the three waveform profiles has a distinct impact on factors concerning both separation and sensitivity. Regarding separation, a quantitative evaluation of peak widths and the temporal separation between T8 and bradikinin²⁺ is presented in Figure 5. Recalling that gains in resolution were observed when increasing TW amplitude and decreasing TW speed, it is interesting to note that these gains in resolution coincide with decreases in temporal separation (i.e., more tightly spaced peaks). Ultimately, so long as the reduction in peak width is more significant than the loss of temporal separation (eq. 2), the quality of separation will improve. While similar results and discussions have been published elsewhere,^14,29 it is important to acknowledge this behavior as it provides insight regarding TW method development, suggesting that optimal separation is achieved within some range of rollover events in which ions are sufficiently separated but not subject to excessive diffusion. Regardless of whether TW speed or amplitude is being modulated, the rates of peak width reduction and loss of temporal separation are more profound when operating in static mode relative to ramp mode (values associated with the lines of best fit are provided in Table S2 ), as ion populations are exposed to conditions which discourage ion rollover, for longer periods. Interestingly, despite comparable behavior between the speed and amplitude data (Fig 5), a clear disparity in the nature of these trends (linear for speed, quadratic for amplitude) is apparent. We attribute the linear nature of the speed data to a steady-state separation mechanism proposed by May and Mclean,²⁹ which occurs as a result of ions undergoing numerous rollover events and thus experience a range of field strengths. Regarding the amplitude data, we ascribe this nonlinear behavior to the quadratic relationship between drift time and KE proposed by Shvartsburg and Smith,¹⁶ where K is an ion’s mobility and E is the field intensity (which relates to TW amplitude) experienced by the ion.

Figure 4.

Baseline corrected ATD of bradykinin²⁺ (first peak) and T8 (second peak) under control conditions (a) and conditions in which TW speed (b) and amplitude (c) were modulated. The TW speed was modulated to 110 m/s for all ATD shown in (b). The TW amplitude was modulated to 31 V_p-p for all ATD shown in (c).

Figure 5.

The average FWHM of bradykinin²⁺ and T8 and temporal separation between the two species as a function of TW speed (left) and amplitude (right). Values on the x-axis represent conditions that followed the control period of the static and dual waveforms, and the value that was ramped to in the ramp waveform.

In the context of method development, the data presented in Figures 3 and 5 illustrate the tradeoffs when choosing different TW profiles. Beginning with a simple comparison between the modulation of TW speed and amplitude, the presented data suggest that modulating TW amplitude has a greater impact on resolution. It is important to note that the full range of TW amplitudes evaluated represented a 35% deviation from the control value (Control: 23 V_p-p, highest amplitude evaluated: 31 V_p-p) compared to a 40% deviation for the TW speed experiments (Control: 183 m/s, lowest speed evaluated: 110 m/s), which further emphasizes the impact that TW amplitude has on resolution relative to TW speed. Regarding the three waveform profiles, should a user opt to modulate TW amplitude, the static elution method provides the greatest gains in resolution, though it is worth noting that no significant (>10%) differences were observed between the static and ramp waveforms. Lastly, the dual waveform expectedly performed the worst as the arrival time of bradykinin²⁺ remains stagnant, while that of T8 decreases as a function of TW speed and amplitude.

3.2.1. Sensitivity (SNR)

While the information provided in Figure 5 enhances our understanding of how each waveform impacts separation, further investigation into how these waveforms impact sensitivity is required to fully understand their analytical impact. A key distinction between the ATD presented in Figures 4b and 4c are apparent variations in peak height. While a quantitative overview of peak height variation is provided in the supplementary material (Figure S2 ), this discussion will instead focus on SNR, which, while related to peak height, is more relevant to matters of sensitivity. Fluctuations in SNR as a function of TW speed and amplitude are presented in Figure 6. Unsurprisingly, the SNR of bradyikinin²⁺ does not vary significantly (>10%) for any conditions in which either the dual or ramp waveforms were used. While the cause of this minimal variation is straightforward for the dual waveform (conditions are only varied after bradykinin²⁺ exits the TW-SLIM), it is important to recognize that the ramped approach does not affect all species equally, instead having a more profound impact on later arriving species. We attribute this greater impact in part to the disparity in conditions during which ions reach the detector, leading to some degree of temporal compression as ions exit the TW-SLIM faster as the ramp progresses. To be clear, unlike the gains in peak intensity and SNR observed by Williamson and Nagy,³⁰ contributions from temporal compression may contribute to, but do not fully explain, the observed behavior which is also impacted by variations in c throughout the separation process.

Figure 6.

SNR of T8 and bradykinin²⁺ relative to control. An offset was added in the x-dimension for visual clarity. To be clear, all data for the static, dual, and ramp waveforms were collected using identical TW speeds and amplitudes. Values on the x-axis represent conditions that followed the control period of the static and dual waveforms and the value that was ramped to in the ramp waveform.

Unlike the ramp and dual waveforms, the static approach has a greater impact on the SNR of bradykinin²⁺, particularly when modulating TW amplitude. Interestingly, the rate at which the SNR of bradykinin²⁺ increases is similar to that of T8 when modulating TW amplitude, but shows a far greater disparity when modulating TW speed. The drastic deviations associated with the SNR of bradykinin²⁺ when modulating to 110 and 119 m/s are attributed to bradykinin³⁺. Bradykinin³⁺ increases the noise preceding bradykinin²⁺ for the 119 m/s data (lower SNR) and increases the signal attributed to bradykinin²⁺ as the two species reach the detector simultaneously (greater SNR) at 110 m/s.

In contrast to bradykinin²⁺, the behavior displayed by T8 in Figure 6 shows less disparity between the performance of the three waveform profiles, though variations in trends between the TW speed and amplitude experiments persist. Regarding TW speed modulation, the ramp waveform consistently outperforms the other two, particularly at lower speeds, with the highest disparity observed at 174 m/s between the static (1.0 ± 0.1 relative SNR) and ramp (1.2 ± 0.1 relative SNR) waveforms. In contrast, when modulating TW amplitude, the ramp waveform outperforms the other two at lower amplitudes (<27 V_p-p) before being overtaken by the static waveform as TW amplitude is increased. It is also worth noting that the amplitude experiments produced the least variance between waveforms, with the greatest disparity observed at 25 V_p-p between the ramp (1.19 ± 0.09) and dual (1.11 ± 0.09) waveforms.

The results presented in Figures 6 and S2 provide valuable guidance for those interested in optimizing the sensitivity of their TW-SLIM methods. Recalling the earlier discussion regarding separation, the modulation of TW amplitude and implementation of a static waveform proved the most effective means of increasing resolution. In contrast, when optimizing for SNR, there is no single approach that provides superior gains in SNR for both species. For T8, the ramped TW speed approach provided the greatest gains in SNR over the same range of parameters that saw this method fall short regarding separation. It is important to note however, that the performance of the ramped waveform did not differ significantly from that of the other two. When evaluating the data associated with T8, a simplistic argument could be made that the ramped TW speed approach should be used when gains in sensitivity are required, and the static TW amplitude approach could be implemented when greater resolution is desired. However, this argument ultimately falls short as it neglects the SNR of bradykinin²⁺, whose greatest gains did not coincide with those of T8, instead benefiting from the static TW amplitude approach. A more complete interpretation of the data presented thus far suggests that a user must consider whether earlier or later arriving species are in greater need of SNR enhancement and then select a method accordingly while ensuring adequate separation is maintained.

3.2.2. Sensitivity (Ion Abundance)

While metrics such as SNR are often at the forefront of discussions pertaining to sensitivity, it is important to note that SNR alone inadequately describes a method’s sensitivity.³¹ Figure 7 presents the ion abundance of both T8 (Figures 7 and 7b) and bradykinin²⁺ (Figures 7c and 7d) across various conditions, normalized to those observed under control conditions. Beginning with bradykinin²⁺, no significant (>10%) variation in ion abundance was observed, with the exception of data associated with static TW speed modulation at 110 m/s and ramped TW amplitude modulation at 27 V_p-p. We attribute the spike in ion abundance observed at 110 m/s to the simultaneous arrival of bradykinin³⁺ (see earlier discussion regarding SNR) and the “loss” of ion abundance at 27 V_p-p to issues of peak fitting, which arise from abnormalities in peak shape under those conditions. It is interesting to note that across all conditions, the ramp waveform performed the worst with regard to the ion abundance of bradykinin²⁺, but provided optimal performance for T8. Aside from the performance of the ramp waveform, differences between the behavior of T8 and bradykinin²⁺ displayed in Figure 7 persist for all waveform profiles, as the trends for bradykinin²⁺ are relatively static in comparison to those of T8. While the exact cause of this behavior requires further investigation, we attribute the observed disparities between T8 and bradykinin²⁺ to the fact that T8 spends a greater amount of time within the TW-SLIM during separation and is thus more sensitive to perturbations to experimental conditions.

Figure 7.

Relative (to control) ion abundances of T8 when modulating TW speed (a) and amplitude (b). Relative (to control) ion abundances of bradykinin²⁺ when modulating TW speed (c) and amplitude (d). Values on the x-axis represent conditions that followed the control period of the static and dual waveforms, and the value that was ramped to in the ramp waveform. An offset was added in the x-dimension for visual clarity.

One last point of emphasis for the data presented in Figure 7 is the disparate performances of the three waveforms with regard to T8. As mentioned previously, the ramp waveform provided the optimal transmission of T8 ions through the SLIM (i.e., enabled the greatest number of charges to reach the detector). Not only did the ramp outperform both the dual and static waveforms in this regard, but it also increased the observed T8 ion population relative to the control, particularly when implementing more gradual ramps wherein the value being ramped to is closer in value to control conditions. In alignment with results associated with the ramp, the static and dual waveforms provided optimal transmission of T8 ions under conditions that were closer to those of the control (TW speed: 174–146, TW amplitude: 24–27 V_p-p) with greater ion losses observed as conditions deviated more profoundly from the control. A possible explanation for the loss of ions when implementing the static and dual waveforms is ion activation and or collisions with SLIM surfaces during the rapid transition from control conditions (Figure 2). While these two events cannot be ruled out entirely at this time, the authors would like to suggest that if these events were truly the singular cause of the observed ion loss, the relative T8 ion abundances for the static and dual waveforms would be near indistinguishable across all conditions evaluated. In actuality, barring the modulation of TW amplitude at 24 and 25 V_p-p, the observed ion abundance of T8 is consistently less when operating with the static waveform as opposed to the dual waveform. An alternative explanation for the behavior of T8 in Figure 7 is that higher speed/lower amplitude conditions simply provide more optimal transmission of T8 ions based on their motion relative to the passing waves. While ion activation is not considered a significant factor in the presented work, future investigations utilizing more labile analytes should be completed prior to expanding the presented methods to a wider range of analytes.

3.3. Considerations for Analyzing Disparate Species

Aside from an enhanced perspective on the three waveforms presented, the results shown in Figure 7 give rise to a greater discussion of how best to simultaneously transmit ions of disparate size within TW-SLIM. Given that “ideal” confinement (namely radial confinement via RF)^22,23 and ion motion with regards to separation (eq. 1) are dependent upon mobility, it is unsurprising that a singular set of experimental parameters utilized throughout the presented work was unable to provid e optimum analytical performance across all metrics and analytes. Though an absolute solution to optimizing separation and sensitivity for all ions present in a system is most likely unattainable, the waveform profiles implemented here, namely the dual and ramp waveforms, offer compromises that may mitigate the negative consequences of performance optimization. While not explored in this preliminary work, one could theorize a dual waveform profile which, having determined optimum transmission conditions for each analyte within a mixture, can be used to achieve optimum transmission of one species before transitioning to optimum conditions for another species. While this would inevitably limit the transmission of later arriving species, it would allow a user to enhance the performance of a particular species which may prove difficult to observe. Interestingly, the ramp waveform, which was originally hypothesized to be a middle ground for disparate ion transmission, proved to provide optimal transmission for one species (T8) and relatively poor transmission for another (bradykinin²⁺). Future work will evaluate more flexible implementations of the ramp waveform in which both the start of the ramp is varied (e.g., after the arrival of higher mobility species) as well as the steepness of the ramp. Additionally, the simultaneous ramping of TW speed and amplitude will be evaluated as well.

4. Conclusion

The presented analysis of three distinct TW waveform profiles provides insight regarding approaches to TW-SLIM method development which may be tailored to enhance select analytical metrics concerning matters of separation (resolution) and sensitivity (SNR and ion abundance). Concerning separation, the static waveform demonstrated the greatest impact on peak width and temporal separation and ultimately provided the greatest gains in resolution across all conditions analyzed when modulating TW amplitude. It is worth noting that while static TW amplitude modulation provided the greatest increase in resolution, these gains were not significantly (>10%) greater than those provided by the ramp waveform. Analyses of results pertaining to sensitivity were unable to provide a singular “optimal” approach and suggest a more nuanced approach to method development is required to achieve the desired result. Beginning with SNR, TW speed modulation utilizing the ramp waveform provided the greatest gains in SNR for T8 (though the other waveforms did perform comparably), but fell short for bradykinin^2+, which benefited most from the static TW amplitude approach. Leveraging the use of a Faraday plate as a detector, measurements of ion abundance were conducted which provided insight into how various conditions impacted ion transmission through the TW-SLIM. Similar to observations made concerning SNR, no singular set of parameters provided optimal ion transmission for both T8 and bradykinin²⁺, with the former benefiting most from the ramp waveform and the latter suffering the greatest ion loss when operating with the ramp.

While not explored in this work, it is important to note that dynamic waveform profiles may complicate TW-SLIM experiments which incorporate time-sensitive events such as multi-pass separations or the targeted manipulation of select ion species.^21,32,33 Time-sensitive events are often defined by ion velocities which allow a user to predict the location of ions at a particular moment. The use of dynamic waveform profiles limits this predictive ability as ion velocity is no longer constant throughout a given separation. It’s important to note that the use of dynamic waveform profiles may also complicate efforts to determine ion-neutral collisional cross sections (CCS), particularly when ion velocity is used for calibration.³⁴Given the nature of TW-SLIM separations, it is unsurprising that a one size fits all approach is unable to provide ideal conditions for separation and ion transmission when analyzing ions of disparate size. That being said, given the unique capabilities of TW-SLIM,^21,27,35 much of the work concerning the technique has thus far focused on hard-to-separate isomers (i.e., ions of similar size). However, as TW-SLIM continues to grow as a technique and becomes more widely available,^26,36 the scope of TW-SLIM analyses will inevitably expand and approaches that address the issue of disparate ion analysis must expand in turn. Arguably, the results associated with our analysis of species with only modest differences in reduced mobility (i.e., T8 = 0.816 cm²/Vs vs. bradykinin²⁺ = 1.189 cm²/Vs)^37,38 suggest that addressing this issue is of more immediate relevance.

With that in mind, the presented waveforms provide alternatives to traditional TW-SLIM methodologies and establish a point of comparison for future TW profiles that adopt a widening degree of profiles (e.g., square vs. sinusoidal vs. sawtooth).¹⁷ Most importantly, these concepts will aid those implementing the technique with added flexibility which will be necessary as the technique adapts to meet an ever-increasing number of analytical demands.

Highlights:

• The performance of multiple dynamic traveling wave profiles was evaluated

• Optimal resolution was obtained using a static TW, but the ramp performed similarly

• Regarding SNR, optimum waveform profiles were species dependent

• Increases in ion abundance were observed when using a ramped TW for select species

• Tailored waveforms to optimize TW-SLIM separations require continued refinement