Introduction

Size exclusion chromatography (SEC) and nuclear magnetic resonance (NMR) have been incorporated since several decades in routine polymer structure characterization [1]. These techniques provide information on the average polymer chain length, the molar mass distribution, and chemical structure based on the measurement of hydrodynamic volumes and chemical environments of atoms, mostly protons and carbons, respectively. When these conventional techniques (e.g., SEC, NMR) are unable to resolve the presence of polymer topology or branching contaminations in samples due to signal overlap or due to signal broadenings potentially hiding less intense signals, mass spectrometric techniques (MS) may help with sample characterization.

In such cases (complex mixtures, isobaric compounds), tandem mass spectrometry (MS/MS) experiments bring help for sample characterizations. The most commonly used MS/MS ion activation technique is collision-induced dissociation (CID). For instance, cyclic and linear polymers could be distinguished [2,3,4,5,6] by using MS/MS, different isobaric polymer mixtures could be resolved [7], and different polymer types could be distinguished according to their accelerating voltage at 50% survival yield (SY) of the precursor ions (V50; using a quadrupole ion trap instrument) [4].

If the spectra are still too complex to be resolved by MS/MS only, or if the fragments or V50 values are not different enough to draw unambiguous conclusions on the presence of different polymer contaminations in a sample, then additional separation techniques can be coupled to the mass spectrometer. Ion mobility-MS (IM-MS) counts among the most recent couplings to be probed for polymer characterizations [8,9,10,11,12,13,14,15,16,17,18,19]. Even further, MS couplings such as liquid chromatography (LC) LC-IM-MS are beginning to be investigated [20].

In the polymer field, structurally similar to natural poly(peptide)s, poly(2-alkyl/aryl-2-oxazoline)s (PAOx) have attracted significant attention since the last five decades and are subject to a renewed interest, especially in the biomedical field [21, 22]. The increasing compositional and architectural complexity of PAOx and other synthetic polymers developed for advanced applications is giving rise to analytical challenges as the conventional techniques fall short of providing a complete description of the polymer structure (e.g., co-monomer sequence, polymer topologies and branching, isomer identification). PAOx have been the subject of a number of in-depth analytical studies involving mass spectrometry, shedding light on fragmentation pathways [23,24,25], polymerization mechanism [26, 27], and unveiling chain transfer reactions [28, 29] or the mechanisms associated with polymer degradation [30].

Here, we establish the proof of concept for resolving two pure side chain isomers of PAOx [31, 32], namely poly(2-n-propyl-2-oxazoline) (Pn-PrOx) and poly(2-iso-propyl-2-oxazoline) (Pi-PrOx), using mass spectrometric techniques (MS/MS and IM-MS). First, CID MS/MS experiments are undertaken on the pure samples. A Pn-PrOx and Pi-PrOx mixture is then analyzed and semi-quantitatively titrated using CID. Then, IM-MS experiments provide a second separation dimension of the two isomeric polymers. Finally, the IM-MS experiments allow understanding and interpreting the observed MS/MS behavior.

Materials and Methods

Polymers

Poly(2-n-propyl-2-oxazoline) Pn-PrOx and poly(2-isopropyl-2-oxazoline) Pi-PrOx (Figure 1) were synthesized according to the methodology previously reported [33]. The polymerization times were calculated using the reported polymerization rate constant for each monomer to reach a monomer conversion of 99%. After termination with methanolic tetramethyl ammonium hydroxide, the solvent was removed under reduced pressure resulting in a white solid. The polymers were re-dissolved in water, purified by dialysis in demineralized water (SpectrumLabs dialysis tubing, MWCO 100–500 Da) and freeze-dried. Polymer characterization was performed by 1H-NMR spectroscopy, MALDI-ToF mass spectrometry, and size exclusion chromatography (see Supplementary Information, Figure SI1). Five Pn-PrOx (DPaverage = 20, 40, 50, 2 × 100) and four Pi-PrOx (DPaverage = 20, 30, 40, 50) samples were analyzed to cover the mass ranges in this study. The samples yielded overlaps in their mass distributions (i.e., synthesized DP ranges) in order to produce multiple replicate measurements of most of the IM-MS data points.

Figure 1
figure 1

Representation of the chemical structures of (a) Pn-PrOx and (b) Pi-PrOx

Full size image

Ion Mobility-Mass Spectrometry

Pn-PrOx and Pi-PrOx samples were dissolved in pure methanol and spiked with Na+ cations (NaCl salt) in order to yield solution concentrations of 10−6–10−5 M for MS infusions. The ion mobility-mass spectrometer used for this study was a Traveling Wave (T-Wave) Synapt G2 HDMS (Waters, UK). Detailed experimental conditions can be found in the Supplementary Information.

The analysis of the IM-MS data was performed using Waters’ MassLynx 4.1 software. IM arrival time distributions (ATD) were fitted with Gaussian functions using PeakFit v.4.11 to determine accurate drift times. Data processing was performed using Excel 2011 and IgorPro 6.37.

IM-MS Collision Cross-Section Calculation

The T-Wave IM cell has to be calibrated in order to convert the drift times into CCS values. The calibration protocol from literature was followed [34] (for details on the calibrating substances and the calibration curves, see the Supplementary Information (Figure SI2 and Figure SI3)).

MS/MS (CID) Experiments

The MS/MS CID experiments were performed using the “trap bias” as accelerating voltage (set from 45 to 200 V). These electrodes are located between the end of the trap cell, which forms the ion packets for IM separation, and the entrance of the He cell, which is located in front of the IM cell [35].

Survival Yield and Relative Intensity Calculations

The SY [2, 4, 7, 36,37,38] represents the ion intensity of a selected precursor ion which is normalized to all ion signals at a given acceleration voltage (Eq. 1). The SYs can be calculated for CID experiments sampling different accelerating voltages, yielding the SY curves. They can typically be fitted with sigmoid functions (Eq. 2). In this study, only isobaric ions are compared, which allows us to directly compare accelerating voltages. From the sigmoid functions, we can calculate the accelerating voltage at 50% of SY, called V50, in order to compare the results of the two polymers.

$$ \mathrm{SY}=\frac{I_{\mathrm{precursor}}}{I_{\mathrm{precursor}}+\sum {I}_{\mathrm{fragments}}} $$
(1)

where SY represents the survival yield, Iprecursor is the signal intensity of the precursor ion, and Σ Ifragments are the signal intensities of all other ions in the spectrum (fragments).

$$ \mathrm{SY}=\mathrm{base}+\frac{\max }{1+\exp \left(\frac{xhalf-x}{rate}\right)} $$
(2)

where SY is the survival yield plotted as a function of the accelerating voltage, base sets the SY at small voltages, base + max sets the SY at large voltages, xhalf sets the voltage at value (base + max)/2, x represents the voltage, and rate sets the decrease rate of the curve (“slope”). The parameters base and max allow for small experimental SY offsets.

The calculation of the relative intensities of fragment ions when sampling different acceleration voltages is also performed using Eq. 1. In this case, Iprecursor is the intensity of any fragment ion and the denominator still represents the intensities of all ions in the spectrum. The intensities of the formed fragment ions do not follow a sigmoid function when plotted as a function of the accelerating voltage. The intensity of the fragments first increases (fragment formation at small voltages) before decreasing (high voltages), rather yielding Gaussian functions. In this study, we do not proceed to fit the fragment ion intensities according to the accelerating voltage.

Thermogravimetric Analysis

Details are reported in the Supplementary Information.

1H Nuclear Magnetic Resonance

Details are reported in the Supplementary Information.

Size Exclusion Chromatography

Details are reported in the Supplementary Information.

Results and Discussions

CID fragments and fragmentation mechanisms of PAOx have been extensively described in literature [23,24,25, 36, 39]. They ranged from mechanisms such as depolymerizations and side chain losses to McLafferty rearrangements, depending on the side chains and the chain end functions. We then first attempted to discriminate Pn-PrOx and Pi-PrOx ions on the basis of the MS/MS spectra (see Figure SI5 and Figure SI6 in the Supplementary Information). They showed, as expected, identical CID spectra due to identical chain end functions and isomeric side chains. Therefore, fragment ion identifications cannot yield much information on the identity of the precursor side chain isomer. Consequently, we analyzed fragment ion intensities as a function of the accelerating voltage, relating the CID experiments to stability titrations of the ions.

CID MS/MS: Survival Yields of Precursor Ions

The SY calculation of the precursor ions yields discriminatory information on the two side chain isomeric polymers. The SYs were monitored for the 2+ and 3+ charge states (sodiated ions) and for several degrees of polymerization (DP = 11, 15, 20, 25, 27, 29; see Figure 2).

Figure 2
figure 2

Survival yield (SY) curves (obtained from Eq. 1) of Pn-PrOx (black markers with black dotted lines) and Pi-PrOx (blue markers with blue dotted lines) sodiated precursor ions plotted as a function of the accelerating voltage. The SY curves are recorded for different charge states (2+ and 3+ precursor ions) and different degrees of polymerization (DP = 11, 15, 20, 25, 27, 29). The black and blue lines represent the sigmoid fits of the SY curves (Eq. 2)

Full size image

The first observation to be noted from Figure 2 and Table 1 (V50 values) is the increase in the accelerating voltage as a function of the DP for a given charge state and polymer (see also Figure SI7 for the evolution of V50 values as a function of the DP). This is generally observed for the center-of-mass energy when increasing the number of normal modes of vibration (or degrees of freedom DOF), in this case when the polymer chain length increases [4, 37, 40]. The more the number of normal modes of vibration (NMVs) or DOFs increases, the more internal energy can be stored in the ions. This leads to an increasing accelerating voltage before reaching the bond dissociation energy barriers.

Table 1 V50 values of Pn-PrOx and Pi-PrOx extracted from the survival yields (Eq. 1) plotted in Figure 2 using Eq. 2. The SYs were recorded for the 2+ and 3+ sodiated precursor ions at different degrees of polymerization (DP = 11, 15, 20, 25, 27, 29)
Full size table

Second, for all monitored precursor ions, [Pi-PrOx + nNa+]n+ ions are kinetically more stable (i.e., within the reaction timeframe sampled by the mass spectrometer) than their [Pn-PrOx + nNa+]n+ counterparts having identical charge states and DPs and they degrade at higher accelerating voltages (Figure 2 and Figure SI7). The difference in V50 (accelerating voltage at 50% survival yield; Eq. 2) of the two pure isomers reaches up to 7–8 V for the sampled DPs and charge states (8 V for [PAOx29 + 3Na+]3+, see Table 1).

This observation leads to the possibility of semi-quantitatively titrating a mixture of the two isobaric polymers using their SY curves and V50 values. As the Pn-PrOx and Pi-PrOx polymers are only side chain isomers, one should nevertheless not hope for the observation of a voltage range with constant SY in the SYmixture as it has been observed for polymers differing in the nature of their monomer backbone [7] or for polymers having radically different 3D structures (e.g., linear and cyclic) [3]. In such cases, one of the polymers’ SY curve reached zero intensity before the second polymer SY curve began decreasing. In the present study, the Pn/i-PrOx V50 difference is not large enough to generate such a voltage range of constant SY (or stepwise degradation) for easy calculation of the percentage of Pn-PrOx and Pi-PrOx contained in the mixture. Figure 3 represents the titration of a ~ 50/50 mixture of the two side chain isomers for DP = 20 bearing 3 sodium cations [PAOxDP=20 + 3Na+]3+. As briefly discussed in literature [7], the deconvolution of the SYmixture curve (Eq. 2) is only semi-quantitative as small variations may result in large interpretation differences. Nevertheless, we were able to deconvolute the V50 value (V50 (MixtureDP=20) = 127 V) using a mathematical equation system (Eqs. 3 and 4) based on the two pure-sample V50 values (V50 (Pn-PrOxDP=20) = 124 V and V50 (Pi-PrOxDP=20) = 131 V, see Table 1). The Pn/i-PrOxDP=20 mixture was hence titrated to a mixture containing 60% of Pn-PrOxDP=20 and 40% of Pi-PrOx DP=20. Given that the dispersities of the two mixed samples (DPaverage = 20) are not perfectly matching, the titrated isomer proportions are in good agreement with our ~ 50/50 initial mixture.

Figure 3
figure 3

Titration curve of Pn-PrOx and Pi-PrOx complexes using the example of the 3+ charge state at DP = 20 ([PAOx20 + 3Na+]3+; m/z = 787.6). The black markers represent the SY of pure Pn-PrOx precursor ions and the blue markers represent the SY of pure Pi-PrOx ions. The green plain markers represent the SY of the mixture (~ 50/50) of the two pure isomeric polymers. The color-coded lines are the sigmoid fits of the SY curves used for V50 calculation (Eq. 2)

Full size image
$$ \Big\{{\displaystyle \begin{array}{c}{V}_{50\;\mathrm{Mixture}}=n\cdot {V}_{50\ \mathrm{P}n-\mathrm{PrOx}}+i\cdot {V}_{50\ \mathrm{P}i-\mathrm{PrOx}}\\ {}1=n+i\end{array}} $$
(3)
$$ \Big\{{\displaystyle \begin{array}{c}n=\frac{V_{50\ \mathrm{Mixture}}-{V}_{50\ \mathrm{P}i-\mathrm{PrOx}}}{V_{50\ \mathrm{P}n-\mathrm{PrOx}}-{V}_{50\ \mathrm{P}i-\mathrm{PrOx}}}\\ {}i=1-n\end{array}} $$
(4)

where n and i represent the proportions of respectively Pn-PrOx and Pi-PrOx contained in the mixture.

Comparatively, deconvoluting a Pn/i-PrOx mixture can also be performed using 1H NMR spectroscopy (Figure SI8 and Figure SI9). However, 1H NMR spectroscopy only yields average-DP information, without the possibility of deconvoluting the dispersities of the constituting pure Pn-PrOx and Pi-PrOx samples, i.e., deconvoluting each DP in the sample. Furthermore, NMR only allows the deconvolution of complex mixtures (tertiary mixtures, additives…) as long as no chemical shift overlaps take place. A brief discussion and comparison between CID ion heating and macroscopic TGA degradation analyses (Figure SI4) can be found in the Supplementary Information. In short, the pure-sample polymers degrade at different temperatures (ΔT ~ 30 °C) but the mixture does not exhibit a stepwise degradation, meaning that TGA does not allow for a semi-quantitative deconvolution of a Pn/i-PrOx mixture. Neither does SEC: the presence of both polymers in an isomer mixture cannot be elucidated, since the trace of the isomer mixture resembles that of a well-defined pure polymer (see Figure SI10).

CID MS/MS: Cation Ejection Species

Using CID tandem MS, the differences between the two side chain isomeric polymers go further than the V50 quantification using the precursor ions. The relative intensities of specific ions can also help resolving a mixture of the two polymers. Such differences can be seen in the cation ejection species. The cation ejection species are constituted of the intact precursor ions not yet having undergone any fragmentation, but having ejected one or several sodium cations (sum of all the species; see Figure 4) [41].

Figure 4
figure 4

Relative signal intensities (Eq. 1) plotted as a function of the accelerating voltage. In each plot, the black markers and lines represent Pn-PrOx ions and the blue markers and lines represent Pi-PrOx ions. The precursor ions were sampled as 2+ and 3+ charge states for different DP values. The circle markers represent the SY of the precursor ions, the upwards-facing triangles represent the intact polymers having lost sodium cations (ejection) and the downwards-facing triangles represent all other ions (fragmentation) in the spectra. Panels (a)–(f) show the SY plots for different DPs of Pn-PrOx and Pi-PrOx: the [Pn/i-PrOx11 + 2Na+]2+, [Pn/i-PrOx15 + 3Na+]3+, [Pn/i-PrOx20 + 3Na+]3+, [Pn/i-PrOx25 + 3Na+]3+, [Pn/i-PrOx27 + 3Na+]3+, and [Pn/i-PrOx29 + 3Na+]3+ precursor ions, respectively

Full size image

Figure 4 monitors the relative intensities of the precursor ions (SY curves in Figure 2), of the cation ejection species and of all other ions in the MS/MS spectrum (fragment ions) for different charge states and different DPs of Pn-PrOx and Pi-PrOx. The relative intensities are calculated using Eq. 1 (see Materials and Methods). Figure 4a–c show that at smaller DPs, the Pi-PrOx ions undergo more favorably cation ejections than Pn-PrOx ions. At higher DP values (Figure 4d–f), this trend is surprisingly reversed, meaning that Pn-PrOx ions eject more intensely sodium cations than Pi-PrOx ions.

In a Pn/i-PrOx mixture, the two polymer contributions can hence also be titrated using the cation ejection species. This can be very helpful if the precursor ion m/z isolation (e.g., using a quadrupole mass filter) is not efficient enough to yield clean precursor ion selections without any other signal overlaps. This might be the case in spectra where too many ion peaks are present in small m/z regions of interest. Due to the decrease in their charge state (loss of cations) leading to higher m/z values than the precursor ions, the CID-formed cation ejection species will appear in non-convoluted regions of the spectra where fewer interfering ions are present. This simplifies (or enables) data treatment. Figure SI11 illustrates the relative intensities of the cation ejection species from pure Pn-PrOx and Pi-PrOx samples as well as from the ~ 50/50 mixture. Using the maximum of the relative intensities (Imax (Pn-PrOx) = 0.62; Imax (Pi-PrOx) = 0.82; Imax (mixture) = 0.70), the mixture can be titrated as containing 60% of Pn-PrOx and 40% of Pi-PrOx. This is in excellent agreement with the previous SY deconvolution of the precursor ions (60% Pn-PrOx and 40% Pi-PrOx).

IM-MS: CCS Evolutions

Aside from the different MS/MS deconvolutions of the two PAOx polymers, shape information acquired through IM coupled to the mass spectrometric identification of the ions (IM-MS) also yields a discriminatory descriptor of the two side chain isomeric polymers. The CCS evolutions plotted as a function of the DP of Pn-PrOx and Pi-PrOx for different charge states (1+ to 4+) are represented in Figure 5.

Figure 5
figure 5

Collision cross-section (CCS) plotted as a function of the degree of polymerization (DP) of pure Pn-PrOx (black markers) and Pi-PrOx (blue markers) samples. The interpreted charge states range from [PAOx + 1Na+]1+ to [PAOx + 4Na+]4+

Full size image

In general, the CCS evolutions of polymer ions [10, 42,43,44] are dictated by enthalpic and entropic factors. These enthalpic and entropic factors originate from, e.g., the Coulomb repulsions (cation charges), the charge solvation capacities of the chemical functions, steric hindrance of the polymer chain, and the entropy drive to yield the most entangled (compact) polymer chains.

As already reported for other polymers [15, 42,43,44,45], the CCS increases monotonically for a given charge state when the DP increases. The 1+ and 2+ ions follow the same CCS trend, which represents the most compact shape of the polymer-sodium complex, as exhibited in Figure 5. When increasing the charge state from the 2+ to the 3+ or 4+ charge states at small DP values, the CCS increases due to the increase in Coulomb repulsions which distort the shape of the polymer ions. When the DP increases for the 3+ or 4+ charge states, the CCS evolutions exhibit tipping points [44] at higher DP values. At these tipping points, the CCS does not increase at the same rate anymore and may even decrease, as exhibited by the 4+ charge state. Here, the polymer is able to stabilize (solvate) the cations more efficiently by adopting a different shape, i.e., by undergoing a structural rearrangement. After the structural rearrangements, the CCS evolutions merge with the most compact, entropically favored CCS evolution.

In the case of Pn-PrOx and Pi-PrOx, the CCS evolutions before and after the structural rearrangements are very similar. The most compact CCS evolution [44] is also very similar for the two polymers. The differences in solvation capacities due to the isomeric side chains only clearly appear during the structural rearrangements (3+ and 4+ charge states; Figure 5 or Figure SI12). [Pn-PrOx + 3Na+]3+ ions rearrange from DP 21 to 27 whereas [Pi-PrOx + 3Na+]3+ ions rearrange from DP 28 to DP 32. The structural rearrangement of Pi-PrOx occurred at higher DP values and even higher CCS values than the structural rearrangement of Pn-PrOx for the 4+ charge state compared to the lower charge state (see Figure 5). During these DP ranges, the balance between Coulomb repulsions, favoring larger ion shapes, and charge solvation and entropic factors, favoring more compact shapes, is playing a crucial role. Pn-PrOx rearranges at smaller DP values than Pi-PrOx ions, meaning that Pi-PrOx does not solvate the cations as easily as Pn-PrOx in the gas phase. Pi-PrOx thus exhibits a reduced cation solvation efficiency compared to Pn-PrOx. This result suggests that the isopropyl groups hinder the accessibility of the amide function solvating the cation charges.

Ion mobility can thus be used for Pn/i-PrOx mixture characterizations if the DP values of structural rearrangements can be sampled (for 3+, 4+ or higher charge states). Figure 6 illustrates the pure Pn-PrOx and Pi-PrOx ATDs and the convoluted ATD [46] of a Pn/i-PrOx mixture, which can be deconvoluted into the two pure-polymer ATD peaks (different Pn/i-PrOx sample mixture than for MS/MS experiments, leading to a mixture closer to 50/50 in this case when deconvoluting the mixture using Gaussian area values).

Figure 6
figure 6

Arrival time distributions (ATDs) of pure [Pn-PrOx38 + 4Na+]4+ (black trace) and pure [Pi-PrOx38 + 4Na+]4+ (blue trace) polymers. The green trace represents the ATD of a mixture of [Pn-PrOx38 + 4Na+]4+ and [Pi-PrOx38 + 4Na+]4+ and the dotted red curves represent its Gaussian fit deconvolution

Full size image

Crossing MS/MS and IM-MS Information

Pi-PrOx ions decreasing in intensity at higher V50 values (Figure 2) compared to Pn-PrOx ions could at a first glance seem contradictory. Given the increased stability of an isopropyl carbocation fragment ion compared to the n-propyl counterpart, one could expect that Pi-PrOx should degrade easier at lower accelerating voltages. However, literature fragmentation mechanisms report that the fragments originating from side chain losses also contain the oxygen atom of the amide bonds [23]. This solvating oxygen atom cancels out the carbocation stability hypothesis for the two side chain isomers. Therefore, the main factor responsible for the increased precursor ion stability of Pi-PrOx should be the inductive effect [47] of the isopropyl side chains, or simply the decreased flexibility which could shield the polymer backbone and hence increase its stability, which is subjected to the main fragmentation mechanisms according to literature [23,24,25, 36, 39].

When analyzing in detail the signal intensities of the cation ejection species (cation losses) as a function of the DP for the two polymers (Figure 4), it can be noted that Pi-PrOx ejects more favorably (higher ion intensity) sodium cations until [PAOx20 + 3Na+]3+ (Figure 4a–c), even if its precursor ions degrade at higher accelerating voltages. This means that once enough energy is conferred to the ions for Pi-PrOx to start degrading, its sodium cations are observed to be less efficiently stabilized by the polymer chain. Around the [PAOx25 + 3Na+]3+ complexes, the intensities of the cation ejection species are reversed and Pn-PrOx ejects more easily sodium cations (Figure 4d–f). This behavior can be correlated to the IM-MS behavior of the two polymers. As long as the Pn/i-PrOx CCS evolutions are almost confounded (3+ charge state, DP < 22; see Figure 5 and Figure SI12), Pi-PrOx more easily ejects its sodium cations. When the shapes of the [Pn/i-PrOx + 3Na+]3+ ions then begin to differ (DP 22–32; see Figure 5 and Figure SI12), the Pn-PrOx polymer ejects more easily the cations. If higher accelerating voltages were accessible in the mass spectrometer used for this study, we should find a second inversion in the cation ejection intensities once the two polymers follow again similar CCS evolutions (DP > 32). This behavior has its origins in the ejection dynamics which depend on the 3D structure. As already discussed in literature, when the 3D structure changes, the ejection dynamics change [48] (DP 22–32), leading to the correlation between the observed MS/MS and IM-MS behaviors.

Conclusions

In our proof of concept, MS/MS and IM-MS analyses allowed for signal deconvolutions of each DP of the side chain isomeric polymers, namely poly(2-n-propyl-2-oxazoline) Pn-PrOx and poly(2-isopropyl-2-oxazoline) Pi-PrOx, in a Pn/i-PrOx mixture. This enables the reconstruction of the individual pure-sample dispersities.

Regarding tandem MS experiments (MS/MS), given that the CID fragment ions of both polymers were identical, we focused on survival yield calculations and on the interpretation of relative intensities of specific ions. For all the investigated charge states and DP values, Pi-PrOx ions were found to degrade at higher V50 values than Pn-PrOx ions. This was attributed to the increased inductive effect of the isopropyl side chain compared to the n-propyl side chain. Isopropyl side chains stabilize or shield (through a lower flexibility) the polymer backbone, inducing an increase in the precursor ion stabilities. This was used to semi-quantitatively titrate a Pn/i-PrOx mixture by MS/MS. The V50 of the SYs of the mixture precursor ion contributions was titrated owing to the V50 values of the SYs of the pure samples.

When analyzing the MS/MS cation ejection species, i.e., intact polymer ions having ejected one or several (sodium) cations, Pi-PrOx was found to eject or lose more easily sodium cations at small DP values than Pn-PrOx. This trend was reversed at higher DP values. The cation ejection species were then also used to semi-quantitatively titrate (or deconvolute) a Pn/i-PrOx mixture. If the spectra of the precursor ions are too congested to interpret the precursor ion signal intensities, cation ejection species can allow data interpretation because they will be found in different regions of the mass spectra where no other ions interfere.

The IM-MS CCS evolutions of both polymers were very similar except in the DP regions of their structural rearrangements. Pi-PrOx exhibited structural rearrangements taking place at higher DP values than Pn-PrOx, indicating that the charge-solvating amide functions of Pi-PrOx are less accessible than the amide functions of Pn-PrOx. i-Pr shields the amide function due to a decreased flexibility. When deconvoluting a Pn/i-PrOx mixture using IM-MS, the DP values and charge states exhibiting the structural rearrangements need thus to be sampled. On the one hand, IM-MS experiments are less time-consuming than MS/MS experiments but IM-MS deconvolutions of a Pn/i-PrOx mixture are limited to the DP values of the structural rearrangements. On the other hand, the more time-consuming MS/MS experiments allow deconvoluting an isomer mixture for each DP of Pn/i-PrOx.

Finally, the IM-MS behaviors were correlated to the MS/MS intensities of the cation ejection species. At similar CCS values, Pi-PrOx ejected more cations than Pn-PrOx whereas the intensity of the cation ejection species was reversed when the two polymers differed in CCS (or shape).