Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Ion mobility spectrometry-tandem mass spectrometry strategies for the on-line monitoring of a continuous microflow reaction

Ion mobility spectrometry-tandem mass spectrometry strategies for the on-line monitoring of a... Continuous flow chemistry is an efficient, sustainable and green approach for chemical synthesis that surpasses some of the limitations of the traditional batch chemistry. Along with the multiple advantages of a flow reactor, it could be directly connected to the analytical techniques for on-line monitoring of a chemical reaction and ensure the quality by design. Here, we aim to use ion mobility, mass and tandem mass spectrometry (IMS-MS and MS/MS) for the on-line analysis of a phar- maceutically relevant chemical reaction. We carried out a model hetero-Diels Alder reaction in a microflow reactor directly connected to the IMS-MS and MS/MS using either electrospray or atmospheric pressure photo ionization methods. We were able to monitor the reaction mechanism of the Diels Alder reaction and structurally characterize the reaction product and synthesis side-products. The chosen approach enabled identification of two isomers of the main reaction product. A new strategy to annotate the ion mobility spectrum in the absence of standard molecules was introduced and tested for its valid- ity. This was achieved by determining the survival yield of each isomer upon ion mobility separation and density functional theory calculations. This approach was verified by comparing the theoretically driven collision cross section values to the experimental data. In this paper, we demonstrated the potential of combined IMS-MS and MS/MS on-line analysis platform to investigate, monitor and characterize structural isomers in the millisecond time scale. Article Highlights Direct connection of a microflow reactor to the ion mobility-mass spectrometry and tandem mass spectrometry analytical technique and data comparison with the batch chemistry. On-line monitoring of the hetero-Diels Alder reaction and analysis of the reaction mechanism and elements. • Annotating the ion mobility peaks without the standard molecules Keywords Hetero-Diels Alder reaction · Ion mobility spectrometry · Tandem mass spectrometry · Atmospheric pressure photo ionization · Flow reaction · Annotating ion mobility spectrum Introduction Over the last decades, flow chemistry has garnered much attention in the pharmaceutical industry as it facilitates the efficient synthesis of Active Pharmaceutical Ingredients (API)s. Running the chemical reactions via continuous flow Darya Hadavi and Peiliang Han are shared co-first authors. stream through milli- or micro-sized channels has a pleth- ora of advantages compared to batch chemistry [1, 2]. Most * Darya Hadavi importantly, the narrow channels of a flow reactor provide d.hadavi@maastrichtuniversity.nl a larger surface-to-volume ratio, which leads to the efficient Division of Imaging Mass Spectrometry Maastricht transfer of energy, improved control over the reaction pres- University, Maastricht Multimodal Molecular imaging sure and enhanced mixing efficiency [3 –5]. These features of (M4i) Institute, Universiteitssingel 50, 6229ER Maastricht, a flow reactor make it feasible to run reactions at high or low The Netherlands Vol.:(0123456789) 1 3 176 Journal of Flow Chemistry (2022) 12:175–184 temperatures and promote sustainability by mitigating the The (hetero) Diels-Alder (DA) cycloaddition reactions risk of working with hazardous materials and reducing the are frequently applied for the synthesis of medicinal drugs costs and environmental footprints. Moreover, flow chemis- due to their relatively high stereo selectivity and efficiency try is considered as Green Analytical Chemistry (GAC) due [20]. Despite numerous benefits and the increased interest to decreasing the use of energy and reagents and eliminating in utilizing flow reactors to run a (hetero-) DA reaction, wastes [6, 7]. Meanwhile, precise control over the reaction pharma and chemical industries have been very slow in the conditions raises the consistency of manufacturing high- transition from batch to flow processing. As mentioned, quality products and facilitates the upscaling process [8, 9]. this is also due to the scarcity of suitable Process Analytical Coupling of flow reactors with analytical detection tech- Technologies (PAT)s [18]. To accelerate this transition, a nologies allows the real-time and on-line reaction monitor- model hetero-DA reaction carried out in a microflow reactor ing, facilitating process control. However, application of and connected to the IMS-MS and MS/MS for the on-line process analytical techniques such as Liquid Chromatog- monitoring of the reaction. Using this analytical technique, raphy (LC), UltraViolet-Visible (UV-Vis) spectroscopy and the reaction mechanism and structural isomers formed in Nuclear Magnetic Resonance (NMR) poses many chal- the hetero-DA reaction were assessed and the potential of lenges. Techniques such as NMR and UV-Vis fall short to combined IMS-MS and MS/MS experimentation were pro- study complex reaction mixtures over a large concentration foundly investigated. range. In addition, the selective detection of molecules lack- ing UV-Vis chromophores becomes a challenge. With regard to the separation techniques, such as LC, the whole analysis Results and discussion cycle is relatively long and the reaction mixture is at the risk of decomposing within the column [10]. More importantly, A model hetero-DA reaction was carried out in the Chem- it is challenging and rather time consuming to separate iso- trix Labtrix® microflow reactor directly connected to the mers, in particular stereoisomers, without a chiral column IMS-MS and MS/MS for the on-line monitoring of the reac- on LC [11]. Therefore, for the on-line reaction monitoring, tion (Fig. 1). The molecular identifications and separations there is a need for universal analytical detectors being sensi- were studied by Trapped Ion Mobility Spectrometry (TIMS) tive, selective, while requiring short (micro to millisecond) and Traveling Wave Ion Mobility Spectrometry (TWIMS) analysis time. incorporated in Quadrupole Time-Of-Flight (Q-TOF) mass Mass Spectrometry (MS) coupled with tandem mass spectrometer. The reagents and synthesized molecules were spectrometry (MS/MS) and Ion Mobility Spectrometry ionized by an Electrospray Ionization (ESI) and Atmos- (IMS) outperform other technologies by meeting the major- pheric Pressure Photo Ionization (APPI). The reaction was ity of the requirements for high quality reaction monitoring monitored at different temperatures and residence times. [12]. Unlike NMR and UV-vis, MS simultaneously sepa- rates reagents, impurities, intermediates, and products of a complex reaction in microseconds based on their mass over Hetero‑Diels alder reaction charge ratio (m/z). The IMS can surpass the limitations of LC by rapid (millisecond timescale) separation of isomeric In organic chemistry, a wide range of biologically active compounds are synthesised by (hetero-) DA reactions. In compounds with high selectivity and improved signal-to- noise ratio [13]. Moreover, the separation of molecular ions particular, the formation of 1,2-oxazine compounds is of special interest to synthesise bioactive products or stabilize based on their Collision Cross Section (CCS), size, shape, and charge by IMS enables elucidating structural informa- them and selectively functionalize natural products [21–23]. The 1,2-oxazine compounds are formed as a result of nitroso tion [14, 15]. Combined with MS/MS, the fragmentation pathways add to the potential to unambiguously assign the dienophile reaction with conjugated diene. Considering the importance of nitroso hetero-DA reactions in pharma [23], molecular structure of unknowns. Collision Induced Dis- sociation (CID) is one of the commonly used approaches in in this study, nitrosobenzene 1, as a dienophile, was exposed to react with methyl 2, 4-pentadienoate 2, as a diene, through tandem mass spectrometry for survival yield analysis [16]. Hence, connecting a continuous flow chemical reactor to a hetero-DA mechanism to form product 3 (methyl 2-phenyl- 3,6-dihydro-2H-1,2-oxazine-6-carboxylate) (Fig. 2a). The the IMS-MS and MS/MS, as an analytical tool, provides the possibility to monitor a reaction in real-time without stall- reaction was performed in 4-methyl-2-pentanone (4M2P) solvent with high boiling point (116 °C) to prevent over ing the production to collect samples in different steps of a reaction. In addition, the real-time information improves the pressuring in the microflow reactor within the range of tested temperatures (25–120 °C). The reaction efficiency in process control by reducing chemical waste, measurement time and most importantly, leads to a new dimension to qual- the microflow reactor was compared with the batch chem- istry reaction. ity by design [17–19]. 1 3 Journal of Flow Chemistry (2022) 12:175–184 177 Inner diluon 1 2 6 Third inlet port Main inlets to introduce reagent 4 8 Second stac mixer Stac mixer Outlet port 5 9 Residence me loop 8 cm 8 cm Outer diluon 2h10 Pa T-connector 8.5cm ID 0.004" BPR 30 cm 15cm Chemtrix R3227 ESI/APPI-IMS-MS and MS/MS Fig. 1 Schematic of the reaction set up to visualize the microreactor, main inlets, dilution inlets and connection to the MS analyzer Fig. 2 Hetero-DA reaction sche- (a) matics. a Reaction schematic of reagent nitrosobenzene (1) with reagent Methyl 2,4-penta- dienoate (2) in 4-methyl-2-pen- tanone (4M2P) or acetonitrile N N O Methyl 2,4-pentadienoate (2) (ACN) solvent. b Ring opening O O Mₘᵢ: 112.0524 Da of molecule (3) to form iso- meric molecule (4) and its dehy- 4M2P/ACN drated molecular structure (5). Nitrosobenzene (1) Reaction product (3) c Two possible reaction mecha- Mₘᵢ: 107.03711Da Mₘᵢ: 219.0895 Da nisms of α-oxyamination and α-aminoxyation for the reaction of nitrosobenzene (1) with the (b) solvent molecules (4M2P) -H₂O N N O N O O HO O O (3)(4) (5) Mₘᵢ: 201.0789 Da (c) OOH α-oxyamination Ph (6) α-aminoxyation (1)4M2P O Ph NH (7) M ᵢ: 207.1259 Da 1 3 M2 ethyl , 4-pentadienoate in 4M2P solvent Nitrosobenzene in 4M2P solvent 178 Journal of Flow Chemistry (2022) 12:175–184 indicated that the investigated hetero-DA reaction formed The characterization of the reaction elements by ESI and APPI – MS and MS/MS two structurally isomeric molecules (3 and 4). The peak at m/z 225 was annotated as the protonated Connecting the microflow reactor to the IMS-MS and dimer ([2  M + H] ) of the reaction reagent 2. The identity of this peak was confirmed by fragmenting the precursor MS/MS enabled the tentative structural identification of the reaction products, including side-products. The mass ion (m/z 225) in collision cell and annotating the fragment peaks (online resource Fig. S3a). This reaction outcome was spectrum of the reaction outcome generated after ionizing by ESI and APPI sources are depicted in online resource formed by self-dimerization of methyl 2,4-pentadienoate (2) and could be regarded as an impurity. It is because the dimer Fig. S1a, b. The ESI-MS contained the ions at m/z 220 and 242, being respectively assigned as the [M + H] and was already present in the commercial reagent prior to run- ning the reaction. This observation was also verified by the [M + Na] ions of the reaction product 3. Although the reaction outcome ionized under soft conditions (Table. impurities on NMR (online resource Fig. S3b). The peaks at m/z 199 and 230 were identified as the side- S1), the reaction mass spectrum included several fragment ions caused by upfront CID processes [24]. According to products of hetero-DA reaction. Unlike the protonated ion of the main product (3), m/z 230 required higher collision the tandem mass spectrometry of product 3, ions with m/z 202, 190, 170,160, 132 and 94 were being directly energy (20 eV) to yield 70% disassociation. This observation implies that this molecule is more stable compared to the originated from the parent ion m/z 220 (online resource Fig. S1c). main reaction product and likely having a different molecu- lar structure. Merino et al., suggests that ketone groups tend To ensure regioselectivity of the hetero-DA reaction to form product 3, the reagents were reacted in batch with ace- to react with nitroso groups by either α-oxyamination or α-aminoxyation mechanism (Fig. 2c) [27]. Hence, the nitros- tonitrile solvent, similar to a previous study [25]. According 1 13 to the H and C NMR spectra (online resource Fig. S1d obenzene reagent (1) could also react with the solvent 4M2P and form the side-products 6 and/or 7. This reaction pathway and 1e) the outcome of this reaction was consistent with the outcome of the microflow reaction. was confirmed by running an additional batch reaction with reagent 1 and 4M2P (without reagent 2) at 60 °C overnight. The (fragment) ion at m/z 202 was formed by the loss of H O from product 3. The tandem mass spectrum of the The mass spectrum of this overnight reaction was com- parative to the spectrum of the hetero-DA reaction (online precursor ion m/z 202 showed a base peak at m/z 170 (online resource Fig. S2a). This fragment ion was formed due to the resource Fig. S4); in a way that both spectra included peaks at m/z 208 and 190. Based on the MS/MS analysis, these loss of methanol. This fragmentation pathway indicated that + + the loss of H O from the molecule 3 did not occur from the peaks were respectively [M + H] and [M + H-H O] ions of the side-products 6 and/or 7. Consequently, the peak at m/z ester group. The only remaining source of oxygen for water loss was in the oxazine ring. The oxazine could go through 230 is the sodium adduct of the side-products. The tendency of the side-product molecules to lose water suggested that the ring opening process in the microflow reactor to form the molecule 4, subsequently lose a water molecule through the reaction occurred only through the α-oxyamination path- way to form side-product 6 rather than 7. The sodium adduct the cyclization reaction to form a compound 5 (Fig.  2b). To further confirm this mechanism, the reaction product in formation of the side-product 6 in the gas phase explains the stability of the peak at m/z 230 owing to sharing the positive the liquid phase was isolated and purified using preparative Thin Layer Chromatography (TLC) and tested by NMR. The charge of Na with oxygen of ketone and hydroxyl groups. The identity of the low abundant peak at m/z 199 was also NMR spectrum of the peak at m/z 202 showed a methyl group at 3.7 ppm as well as three protons of the pyrrole determined by MS/MS and verified by NMR analysis (online resource Fig. S5). As reported earlier [28], nitrosobenzene ring with sp2 environment, unlike the molecule 3 that had only two sp2 protons (online resource Fig. S2b). In addi- molecules can react with each other to form a dimer, which explains the peak at m/z 199 as the side-product of the tion, the absence of a proton at 5 ppm close to the ester group indicated the pyrrole shape of the molecule (Fig. 2b). reaction. The APPI-MS data were comparative to the ESI-MS. Hence, the peak at m/z 202 was annotated as the molecule 5, which could be formed in the liquid phase due to the con- The APPI source resulted in the formation of both [M]˙ and [M + H] ions of the reaction products at m/z 219 and densation reaction and/or in the gas phase due to the water loss from the molecule 4. Consistent with this finding, the 220 respectively (online resource Fig. S1b). The APPI forms radical cations after being excited to high energic states or N-O bond energy of molecule 3 is 201 kJ/mol, which is a smaller value than other bonds in the ring (C-C 347 kJ/mol, by electron transfer. In the latter case a neutral species such as toluene, known as Dopant (D), absorbs photon energy C-N 305 kJ/mol, C-O 358 kJ/mol) [26]. The relatively weak bond energy of N-O explains the position of ring-opening from the light source in the ionization chamber. The inter- nally excited dopant (D ) ionizes a neutral target Molecule reaction to form molecule 4 from molecule 3. These results 1 3 Journal of Flow Chemistry (2022) 12:175–184 179 (M) that has lower ionization energy by electron transfer. indicated the presence of at least two compounds with the Consequently, the radical cations are formed (Eq. 1). same mass but different 3D structures. Therefore, the hetero- DA reaction produced two isomeric compounds (molecules ∗ +∙ − D + M → D + M + e (1) 3 and 4) with the same m/z value but different mobility. This observation is in line with the represented mechanism for Since the hetero-DA reaction was performed in the the reaction (Fig. 2b). The IMS of the radical cation at m/z 4M2P solvent, the reaction products were mainly pro- 219, which was formed by APPI source, demonstrated better tonated ([M + H] ). This is due to the high proton affin- resolution than the protonated ion at m/z 220 (Fig. 3b). This ity (≈832.7 kJ/mol at 300 K) of the 4M2P solvent, which could be due to the formation of intramolecular H bridges in imposes the protonation of molecules through the Eq. 2. protonated molecules, which increases structural similarity + + S + H + M → S + M + H of 3 and 4. The IMS of the reaction elements were consistent n n + (2) S + M + H → nS + [M + H] regardless of the type of ionization source (APPI or ESI) and the ion mobility instrument (TWIMS or TIMS), even though When the solvent of the reaction was exchanged with TIMS resulted in better resolution data. dichloromethane, having a lower proton affinity (628.0 kJ/ Apart from the isomeric separation, it is also essential to mol at 300 K) compared to 4M2P, the reaction products annotate each ion mobility peak to a correct molecular struc- formed radical cations. This process demonstrated the ver- ture. To this end, different methods were employed. One of satility of the APPI source to utilize in a diverse range of the unique features of TWIMS-TOF is that a mass analyser organic synthesis with no restrictions on the solvent choice. quadrupole is placed before the ion mobility cell, which Regardless of the solvent type, molecules are ionized either enables performing MS/MS on the isolated precursor ions by proton or electron transfer. The ESI and APPI -MS-MS/ after IMS (IMS-MS/MS). Since the internal energy of an ion MS data of the isolated batch reaction product were consist- and its decomposition reaction rate are structural dependent ent with the acquired data from microflow reactor (online features, it is known that even isomeric chemical structures resource Fig. S6). show unique stability and survival yield under CID [30]. Hence, for the first time, we employed IMS-MS/MS method ESI and APPI‑IMS‑MS/MS to define the CE of the isomeric reaction products. The m/z 220 was first isolated in the quadrupole and then separated The ion mobility spectrometry of the reaction products was in the ion mobility cell followed by CID. When the precur- studied by TWIMS and TIMS instruments. Each technique sor ions exposed to the increasing CE after IMS separa- has different separation mechanism as explained before tion, isomeric ions’ survival decreased in a sigmoidal form [29]. Annotating the IMS of isomeric compounds becomes (Fig. 3c). This graph transformed to a linear survival yield a challenge when there are no standard molecules for IMS (Fig. 3d) to find the CE as reported earlier [30]. The results comparison. This challenge is even more pronounced for the identified unique CE of 14.89 and 15.82 eV for the isomer reactions leading to unknown isomeric forms of the main with the short drift time (td 4–4.5 ms) and long drift time product. In such instances, even if each isomer is fragmented (td 4.5-5 ms) respectively. The molecule 4 could have higher by MS/MS through a distinctive pathway, the tandem mass stability due to its lower Gibbs free energy after opening the spectrum contains the fragment ions of all isomers, with- ring, caused by the formation of stronger bond (smaller ΔH) holding the separation of isomers based on one/multiple and increased flexibility (larger ΔS). To verify this assump- unique product ion(s). The combination of ion mobility tion the minimum energy of both compounds was calcu- with tandem mass spectrometry brings another dimension lated by Density Functional Theory (DFT). Regardless of to molecular ion identifications. By applying IMS-MS/MS the protonation site, the molecular ion 4 always had lower approach, the generated fragment ions of each isomer will global minimum energy than the protonated molecular ion have the same arrival time as their parent isomer. Isomers 3 (Table. S2). These findings indicated the higher stability could be separated and identified based on a unique daugh- of the molecular ion 4 compared to 3. Therefore, based on ter ion under different mobility peaks or unique Collision CE and DFT calculations the drift time td 4–4.5 ms and Energy (CE) that results in 50% survival yield (CE ). Ion 4.5–5 ms were respectively annotated to the molecular ions mobility spectrometry also provides additional information 3 and 4. regarding the 3D structure of each ion. This is possible by To further verify the identity of the annotated ion mobil- TWIMS experimentally measuring the ion-neutral gas CCS and com- ity peaks, the experimentally derived CCS was com- N2 paring it to the theoretically calculated CCS values. pared to the theoretically calculated CCS of the molecular The extracted ion mobility (EIM) of the ion at m/z 220 ions 3 and 4. Table. S3 shows the calculated CCS values was reconstructed to plot its ion mobility spectrum. The plot- of the molecules 3 and 4 that were protonated on nitrogen ted IMS showed a bimodal peak (Fig. 3a). The bimodal peak or ester group. Based on the theoretically calculated data, 1 3 180 Journal of Flow Chemistry (2022) 12:175–184 Fig. 3 IMS and IMS-MS/MS (a) 1.2E+6 (c) 1.2 EIM of m/z 220 analysis of the reaction prod- td 4-4.5 ms ucts. a EIM of the protonated td 4.5-5 ms reaction products at m/z 220 analysed by ESI-TWIMS and ESI-TIMS. b EIM of the radical cation of reaction products at m/z 219 analysed by APPI- 0.0E+0 TIMS. c) Experimental survival 3.5 4.5 5.5 yield of isomers upon ion td (ms) (b) (d) CE (eV) mobility separation and CID at 4.8E+4 EIM of m/z 219 CE of 5, 10, 15, 20, 25, 30 eV. y = 0.4999x -7.912 Filled blue circle corresponds to R² = 0.9933 the isomer with short drift time (td 4-4.5ms) and filled orange y = 0.294x -4.3793 triangle corresponds to the R² = 0.9775 isomer with longer drift time (td 4.5-5ms). d Linear survival yield of the isomer at td 4-4.5 0.0E+0 td 4-4.5 ms ms in filled blue circle and the 0.45 0.55 0.65 isomer at td 4.5-5 ms in filled (e) Mobility, 1/K0 [V·s/cm²] td 4.5-5 ms -4 orange triangle. e Comparing CE (eV) 1.4E+6 the theoretically calculated CCS of the protonated molecules (3) and (4) on Nitrogen and ester groups with the experimentally TWIMS derived CCS for the 0.0E+0 N2 detected ion mobility peaks 3.54 4.555.5 6 (dashed red lines) N O N O td (ms) O HO O O (3) (4) CCS= 157.0717 (Å2) CCS= 153.5967 (Å2) CCS= 163.3064 (Å2) CCS = 158.8419 (Å2) Protonated on N Protonated on ester Protonated on N Protonated on ester TWIMS TWIMS CCSN2= 147.044 ±0.79 (Å2) CCSN2= 152.3177 ±0.80 (Å2) products before the ion mobility cell, during the traverse the CCS of the protonated ion 4, regardless of the proto- nation site, was always higher than the protonated mol- from the quadrupole until the IMS cell. Ions with the identical drift time and increasing m/z value are related to ecule 3, which was in line with the open structure of 4. TWIMS Hence, the experimentally derived CCS value of fragment ions generated after the ion mobility cell, from N2 the transfer cell till the detector (white dashed line ellipse- 152.3177 ± 0.80 Å corresponds to the molecular ion 4 and TWIMS 2 the smaller CCS value of 147.044 ± 0.79 Å cor- online resource Fig. S7). N2 All these findings highlighted the added value of IMS- responds to the more compact molecular ion 3 (Fig.  3e). These molecules were more likely protonated on the ester MS and MS/MS experiments on providing in depth knowl- edge on the molecular structure (identity) of (un)known group, as the theoretical and experimental CCS values are in less than 5% agreement upon protonation on the ester group. reaction products in the millisecond time scale. These process analytical technologies and reaction monitoring With these data, td 4–4.5 ms was again annotated to be the protonated molecule 3 and the td 4.5-5 ms was related the strategies not only enabled separation, but also enabled the tentative identification of isomeric compounds that protonated molecule 4. The drift time/mass-to-charge heatmap of the reaction are not distinguishable by MS or MS/MS. It also enabled annotation of IMS peaks without using a standard com- outcome also gave a quick insight regarding the stabil- ity of the reaction. Online resource Fig. S7 presented the pound. Ionizing molecules with APPI combined with dif- ferent IMS-MS/MS scanning strategies shed further light low stability of the reaction products in TWIMS-MS. The heatmap of the isolated m/z 220 in the quadrupole, dem- to characterize molecules. With the combination of ESI/ APPI-IMS-MS/MS the reaction products and intermedi- onstrated the presence of ions with the raising m/z and drift time values (white solid line ellipse- online resource ates of the model hetero- DA reaction were characterised. Fig. S7). This trend showed the dissociation of the reaction 1 3 Counts Counts Counts Survival Yield In{(1-SY)/SY} Journal of Flow Chemistry (2022) 12:175–184 181 Monitoring the microflow reaction by ESI‑IMS‑MS/ Conclusion MS The ion mobility-mass spectrometry and tandem mass The reaction of nitrosobenzene with methyl 2, 4-pentadi- spectrometry technique was used for the on-line moni- enoate was studied by directly connecting the Chemtrix toring the model hetero-DA reaction carried out in the Labtrix® microflow reactor to ESI/APPI-IMS-MS/MS microflow reactor and compared to batch reaction. When analytical technique and compared to the batch chemistry utilizing IMS-MS and MS/MS strategies, the elements results. The model hetero-DA reaction was investigated of the reaction, including the isomeric reaction products, under various temperatures (25, 40, 60, 80, 100, 120 °C) impurities and side-products were identified, character - and residence times (5, 10, 20, 28, 40, 80 min) with constant ized and separated. By systematically increasing the col- pressure (2 × 10  Pa) and concentrations (50 mM methyl 2, lision energy after the ion mobility separation, the survival 4-pentadienoate and nitrosobenzene (1:1)). The MS spectra yield of each isomer was calculated. The acquired sur- were normalised to the intensity of the internal standard. vival yield was compared to the global minimum energy Figure 4a and b presented the impact of residence time of ions calculated by DFT, which enabled annotating the on the formation of reaction products (3 and 4) performed ion mobility spectrum of isomers in the absence of stand- in the microflow versus batch reactor. The formation of the ard molecules. The efficiency of the reaction was higher product molecules in both reactors were increased by raising when it was carried out in the microflow reactor com- the residence time. However, the reaction efficiency in the pared to the batch reaction. The introduced strategies of microflow reactor was higher than the batch reactor. In the hyphenating IMS-MS and MS/MS to a microflow reactor batch reactor the abundance of the product ions increased 7 showed possible potentials to simultaneously detect differ - times after 7.38 h of reaction, while in the microflow reac- ent reaction products and above all their molecular struc- tor this was done in only 40 min. The online monitoring ture (identity). The potential of APPI to act as a universal of the reaction whilst raising the temperature indicated a ionization technique, allowing the application of a wide systematic increase of the reaction products (Fig. 4c). These variety of organic solvents, was introduced. The generation results demonstrated the potentials of microflow reactors and of radical ions through APPI and their specific fragment on-line measurements in increasing the reaction efficiency, ions add another dimension to the structural identification improving reproducibility, reducing the number of experi- process. It is anticipated that this approach be a valuable ments and lowering chemical consumptions compared to technology platform for the development of efficient and running multiple parallel experiments at different conditions green chemical synthesis routes, readily linking reaction required by batch chemistry. conditions to the presence of different reaction pathways. Fig. 4 a Residence time effect (a) (c) Microflow reaction Microflow reaction on the formation of the reaction products after a) microflow 0.5 reaction, b batch reaction. c Temperature effect on the formation of reaction products carried out in the microflow reactor 20 70 120 Temperature (oC) Residence time (min) (b) Batch reaction 4.2 0500 1000 1500 Residence time (min) 1 3 Normalized Normalized intensity intensity Normalized intensity 182 Journal of Flow Chemistry (2022) 12:175–184 (1 ml- Hamilton, Amsterdam, The Netherlands), flow rate Experimental controllers (Chemyx Fusion 200 syringe pump- Stafford, TX, USA), temperature controller, Back Pressure Regulator Materials and sample preparations (BPR), outer dilution inlet and a start unit, which accom- modated the glass microreactor (Chemtrix R3227 type). The Methyl-2,4-pentadienoate, 4-methyl-2-pentanone, nitroso- Chemtrix R3227 microflow reactor consisted of two main benzene, 4-chloropiperidine hydrochloride, Dichlorometh- inlets, static mixer, residence time loop, a third inlet port ane (DCM), Acetonitrile (ACN), toluene, chloroform-d, to introduce a dilution solvent (inner dilution inlet), a sec- Methanol (MeOH), acetic acid and water (H O) with >98% ond static mixer and an outlet port. This microreactor has purity, ESI-L low concentrations tuning mix, APPI/APCI the effective reaction volume of 19.5 μl and it provides the tuning mix and polyalanine purchased from Sigma-Aldrich residence times from 0.78 to 97.5 min (for A + B reaction) (Zwijndrecht, The Netherlands). with the (total) flow rate ranging between 0.2 and 25 μl/ min. The outlet port directly connected to a BPR to keep The reaction condition in batch and isolation the pressure at constant value of 2 × 10  Pa. Right after the process BPR, a T-connector was placed to mix the reaction out- come with the dilution solvent mixture (Fig. 1 and online To perform the hetero-DA reaction in 4-methyl-2-pentanone resource Fig. S8). The line coming out of the T-connector solvent, a solution of nitrosobenzene (53.2 mg, 0.497 mmol) was directly connected to ESI or APPI ionization source, and methyl 2,4-pentadienoate (58.1  μL, 0.499  mmol) in placed in front of the MS orifice. 20 ml of 4M2P was stirred at 60 °C for 25 h. To study the The reaction was conducted by continuous injection of impact of residence time on the production efficiency, 50 μl nitrosobenzene (50  mM) and methyl-2,4-pentadienoate of the reaction mixture was collected after 0.39, 1.63, 4.65, (50 mM) solutions dissolved in 4M2P through main inlets. 7.38, and 25.1 h. Samples were diluted with 150 μl of a Dilution solvent mixture consisted of MeOH/ H O/ACN dilution solvent mixture, including 1 mM of internal stand- (2:1:1) and 0.1% acetic acid introduced to the reaction out- ard and 0.1% acetic acid to make it comparable to the flow come through inner and outer dilution inlets. As an internal reaction. When acetonitrile was used as the reaction solvent standard 1 mM of 4-chloropiperidine hydrochloride added instead of 4-methyl-2-pentanone, a solution of nitrosoben- to the reaction outcome through the inner dilution inlet. To zene (80.1 mg, 0.748 mmol) and methyl 2,4-pentadienoate monitor the impact of residence time on hetero-DA reaction (87.0 μL, 0.749 mmol) in acetonitrile (10 ml) was stirred at the temperature was set to the constant optimised value of 60 °C for 8 h. The same method was applied for the reac- 60 °C. The flow rate of each starting materials set to 2, 1, tion of nitrosobenzene (53.2 mg, 0.497 mmol) with 20 mL 0.5, 0.35, 0.25 and 0.12 μl/min to reach the residence times 4-methyl-2-pentanone for the specificity experiment. The of approximately 5, 10, 20, 28, 40 and 80 min (respectively). outcome of each batch reaction either analysed as a complex The flow rate of the dilution solvent on inner dilution inlet sample or purified by TLC preparative plate (Sigma-Aldrich, always set as the sum flow rate of each starting material. The Zwijndrecht, the Netherlands) prior to analysing. For the outer dilution inlet was used to further dilute the reaction reaction in 4M2P, the reaction mixture diluted with water outcome. Its flow rate was set to 8, 4, 2, 1.4, 1 and 0.5 μl/min (20 ml) and extracted with ethyl acetate (3 × 15 ml). For the respectively for shortest to longest residence time. To moni- reaction in acetonitrile, 15 ml of water and 3 × 10 ml of ethyl tor the effect of temperature, the reaction residence time acetate was used for dilution and extraction. Subsequently, was set to the constant optimised value of 20 min. The reac- the combined organic layer was washed with brine (10 ml) tion was monitored under increasing temperatures of 25, 40, and dried with anhydrous Na SO . After filtration, the solu - 60, 80, 100 and 120 °C. For APPI experiments, DCM was 2 4 tion was concentrated under vacuum and purified using TLC used instead of 4M2P as the reaction solvent. For this set (eluted with 1:1 dichloromethane/ n-hexane) to afford the of experiments, dopants including pure toluene and DCM/ product as a white solid (32.9 mg, 30% for the reaction in toluene (1:1) were respectively injected through the inner 4M2P and 74 mg, 45% for the reaction in acetonitrile). and outer dilution inlets. ESI/APPI‑IMS‑MS/MS parameters The microflow reactor platform and the reaction condition Synapt HDMS G2Si (Waters, Milford, MA, USA) equipped with a TWIMS, as well as a Trapped-Ion Mobility Spec- The reaction was carried out in the Chemtrix Labtrix® trometry-Q-TOF (Bruker Daltonics Inc., Billerica, MA, microflow reactor set up (Fig.1 and online resource Fig. S8) USA) were used for the ion mobility and tandem mass spec- consisted of 4 gas tight glass syringes with luer lock fitting trometry analysis. Table. S1 presents the set parameters for 1 3 Journal of Flow Chemistry (2022) 12:175–184 183 1 13 15 1 13 IMS measurements on each instrument. Nitrogen gas was channel [ H, C, N] TCI probe. The H and C spectra used in the ion mobility cell and helium gas was used before were recorded as 250 mM in CDCl solutions with 0.03% the ion mobility cell for collisional cooling. Argon was used TMS as internal chemical shift reference standard. Spectra as the collision gas in the trap and transfer cells. The ESI were processed using Topspin 3.2 softwareand analyzed and APPI experiments performed in positive ion mode under by the program Sparky 3.114 (Goddard & Kneller) and the optimized ionization conditions listed in Table. S1. The MestReNova. mass spectra acquisition range was set between m/z 50 and 300 Da. ESI-L low concentrations tuning mix and APPI/ Data evaluations APCI tuning mix were used for the mass calibration and polyalanine was used to the IMS calibration. On the TIMS The acquired data by Synapt G2Si analysed by MassLynx instrument, the massive ions with lower mobility (k) elute v4.1. and DriftScope v2.8 software. The measured data by first compared to the lighter and compact ions, however the TIMS-TOF analysed by Compass DataAnalysis v5.0 soft- IMS graph is constructed based on 1/k (V.s/cm ) values, ware (Bruker). The extracted ion mobility spectra extracted making it comparable to the IMS generated by the TWIMS and plotted in excel for further analysis. To correct for the instrument. ion suppression effects, the intensity of the reaction prod- For IMS-MS/MS experiments, the precursor ion was iso- uct ions normalised with the intensity of 4-chloropiperidin lated in the quadrupole and fragmented via CID after the (the internal standard). For CCS calculations, the molecules ion mobility cell in the transfer cell. The collision energy structurally optimized by DFT on Spartan software (Wave- was increased from 5 to 30 eV in 5 eV increments (total of function, Inc. & Q-Chem, Irvine, CA, USA). Conformer dis- 6 collision energies). The MS/MS results were the average tribution calculations were performed to obtain the global of 2 min acquisition time per collision energy. To extract the energy minimum with Density Functional ωB97X-D and CID spectrum of each isomer from its ion mobility, first the basis set 6-31G* using Molecular Force Field (MMFF) ion mobility of the precursor ion at m/z 220 was extracted. geometry. The software considered 100 conformers, while Next, the mobility range of interest were selected (e.g. from keeping the E ≤ 40 kJ/mol and the energy of maximum 10 4 until 4.5 ms). The peak picking then performed in the conformers calculated while keeping the E ≤ 15 kJ/mol. The non-chromatic mode using the following parameters: min IMoS software used for theoretical CCS calculations with drift time peak width of 0.2 ms, MS resolution of 35,000 Trajectory Method (TM) with 300.000  N gas molecules per and minimum intensity threshold of 1000 counts. The CID orientation. Mulliken charges were used for partial atomic spectra were copied to excel for further analysis. Previously charges, they were automatically calculated by Spartan. explained methods [30] were employed to plot the experi- Calculations performed for each conformer with Boltzmann mental and linear survival yield graphs. The experimental weight above 1%. survival yield was plotted based on the ratio of the intensity of the Precursor ion (P) to the sum of the intensities of the Supplementary Information The online version contains supplemen- precursor and Fragment ions (F) (Eq.  3). This graph was tary material available at https://doi. or g/10. 1007/ s41981- 021- 00209-7 . transformed to linear form by Eq. 4 and 5. Acknowledgements We would like to thank Hans Ippel (CARIM, IntensityP Maastricht University) for the execution and data interpretation of the Experimental survival yield = NMR experiments and Yuandi Zhao (M4i, Maastricht University) for IntensityP + (IntensityF) his artistic assistance. (3) Sigmoidal survival yield = Declarations (4) −b.CE 1 + c ∗ e Conflict of interest The authors declare that they have no financial or where b is the slope of the linear segment, and the intercept non-financial conflict of interest. of the linear portion is the natural logarithm of c. Open Access This article is licensed under a Creative Commons Attri- 1 − Survival yeild bution 4.0 International License, which permits use, sharing, adapta- ln = ln(c)− b(CE) (5) Survival yeild tion, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are NMR included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in The Bruker (Rheinstetten, Germany) Avance III HD the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will NMR spectrometer operated at the H frequency of 700 MHz and equipped with a cryogenically cooled triple 1 3 184 Journal of Flow Chemistry (2022) 12:175–184 need to obtain permission directly from the copyright holder. To view a wave- and trapped- ion mobility spectrometry. Rapid Commun copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. Mass Spectrom 33. https:// doi. org/ 10. 1002/ rcm. 8419 16. Iacobucci C et al (2014) Insight into the mechanisms of the mul- ticomponent Ugi and Ugi–smiles reactions by ESI-MS(/MS). Eur J Org Chem 2014(32):7087–7090. https:// doi. org/ 10. 1002/ ejoc. 20140 3179 References 17. Vargas JM et al (2018) Process analytical technology in continu- ous manufacturing of a commercial pharmaceutical product. Int 1. Bogdan AR, Dombrowski AW (2019) Emerging trends in flow J Pharm 538(1):167–178. https://doi. or g/10. 1016/j. i jphar m.2018. chemistry and applications to the pharmaceutical industry. J Med 01. 003 Chem 62(14):6422–6468. https://doi. or g/10. 1021/ acs. jmedc hem. 18. Price GA, Mallik D, Organ MG (2017) Process analytical tools for 8b017 60 flow analysis: a perspective. Journal of Flow Chemistry 7(3):82– 2. Baumann M et al (2020) A perspective on continuous flow chem- 86. https:// doi. org/ 10. 1556/ 1846. 2017. 00032 istry in the pharmaceutical industry. Org Process Res Dev. https:// 19. Djuris J, Ibric S, Djuric Z (2013) 1 - Quality-by-design in pharma- doi. org/ 10. 1021/ acs. oprd. 9b005 24 ceutical development, in Computer-Aided Applications in Phar- 3. Reckamp JM et  al (2017) Mixing performance evaluation for maceutical Technology, J. Djuris, Editor. Woodhead Publishing. commercially available micromixers using Villermaux–Dushman p. 1–16. https:// doi. org/ 10. 1533/ 97819 08818 324.1 reaction scheme with the interaction by exchange with the mean 20. Nicolaou KC et al (2002) The Diels–Alder reaction in Total syn- model. Org Process Res Dev 21(6):816–820. https:// doi. org/ 10. thesis. Angew Chem Int Ed 41(10):1668–1698. https:// doi. org/ 1021/ acs. oprd. 6b003 32 10. 1002/ 1521- 3773(20020 517) 41: 10< 1668:: Aid- anie1 668>3. 0. 4. Schwolow S et al (2012) Application-oriented analysis of mixing Co;2-z performance in microreactors. Org Process Res Dev 16(9):1513– 21. Brulíková L et al (2016) Stereo- and regioselectivity of the hetero- 1522. https:// doi. org/ 10. 1021/ op300 107z Diels-Alder reaction of nitroso derivatives with conjugated dienes. 5. Schwolow S et al (2016) Design and application of a millistruc- Beilstein J Org Chem 12:1949–1980. https://doi. or g/10. 3762/ bjoc. tured heat exchanger reactor for an energy-efficient process. Chem 12. 184 Eng Process Process Intensif 108:109–116. https:// doi. org/ 10. 22. Gamenara D (2001) Hetero Diels-Alder adduct formation between 1016/j. cep. 2016. 07. 017 nitrosobenzene and tetra-methyl purpurogallin and its retro-Diels- 6. Elvira KS et al (2013) The past, present and potential for micro- Alder reaction. J Braz Chem Soc 12(4):489 fluidic reactor technology in chemical synthesis. Nat Chem 23. Carosso S, Miller MJ (2014) Nitroso Diels-Alder (NDA) reaction 5(11):905–915. https:// doi. org/ 10. 1038/ nchem. 1753 as an efficient tool for the functionalization of diene-containing 7. de la Guardia M, Garrigues S (2020) Chapter 1 past, present and natural products. Organic & biomolecular chemistry 12(38):7445– future of green analytical chemistry, in challenges in green ana- 7468. https:// doi. org/ 10. 1039/ c4ob0 1033g lytical chemistry (2). The Royal Society of Chemistry. p. 1-18. 24. Oberacher H et al (2012) On the inter-instrument and the inter- https:// doi. org/ 10. 1039/ 97817 88016 148- 00001 laboratory transferability of a tandem mass spectral reference 8. Campbell Brewer A et al (2019) Development and scale-up of a library. 3. Focus on ion trap and upfront CID. J Mass Spectrom continuous aerobic oxidative Chan–lam coupling. Org Process 47(2):263–270. https:// doi. org/ 10. 1002/ jms. 2961 Res Dev 23(8):1484–1498. https:// doi. or g/ 10. 1021/ acs. opr d. 25. Monbaliu J-CMR et al (2010) Straightforward hetero Diels–Alder 9b001 25 reactions of nitroso dienophiles by microreactor technology. Tet- 9. Schwolow S et  al (2014) Kinetic and scale-up investigations rahedron Lett 51(44):5830–5833. https:// doi. org/ 10. 1016/j. tetlet. of a Michael addition in microreactors. Org Process Res Dev 2010. 08. 117 18(11):1535–1544. https:// doi. org/ 10. 1021/ op500 2758 26. Cottrell TL (1958) The strengths of chemical bonds. Butterworths 10. Fiers T et al (2012) Development of a highly sensitive method Scientific Publications, London for the quantification of estrone and estradiol in serum by liquid 27. Merino P et al (2016) Recent advances on asymmetric Nitroso chromatography tandem mass spectrometry without derivatiza- aldol reaction. Synthesis 48. https://doi. or g/10. 1055/s- 0035- 15615 tion. J Chromatogr B 893-894:57–62. https:// doi. org/ 10. 1016/j. jchro mb. 2012. 02. 034 28. Taft RW, Klingensmith G, Ehrenson S (1965) Multipolar com- 11. Kuki Á et al (2012) Identification of Silymarin constituents: an plexes. I. the dimerization of nitrobenzene1. J Am Chem Soc improved HPLC–MS method. Chromatographia 75(3):175–180. 87(16):3620–3626 https:// doi. org/ 10. 1007/ s10337- 011- 2163-7 29. Cumeras R et al (2015) Review on ion mobility spectrometry. Part 12. Emwas A-HM (2015) The strengths and weaknesses of NMR 1: current instrumentation. Analyst 140(5):1376–1390. https://d oi. spectroscopy and mass spectrometry with particular focus on org/ 10. 1039/ c4an0 1100g metabolomics research. In: Bjerrum JT (ed) Metabonomics: meth- 30. Kertesz TM et al (2009) CE50: quantifying collision induced dis- ods and protocols. Springer New York, New York, pp 161–193. sociation energy for small molecule characterization and identi- https:// doi. org/ 10. 1007/ 978-1- 4939- 2377-9_ 13 fication. J Am Soc Mass Spectrom 20(9):1759–1767. https:// doi. 13. Tadjimukhamedov FK et al (2008) Liquid chromatography/elec- org/ 10. 1016/j. jasms. 2009. 06. 002 trospray ionization/ion mobility spectrometry of chlorophenols with full flow from large bore LC columns. Int J Ion Mobil Spec- Publisher’s note Springer Nature remains neutral with regard to trom 11(1):51–60. https:// doi. org/ 10. 1007/ s12127- 008- 0004-7 jurisdictional claims in published maps and institutional affiliations. 14. Sorribes-Soriano A et al (2018) Trace analysis by ion mobility spectrometry: from conventional to smart sample preconcentration methods. A review. Anal Chim Acta 1026:37–50. https://doi. or g/ 10. 1016/j. aca. 2018. 03. 059 15. Hadavi D et  al (2019) Adduct ion formation as a tool for the molecular structure assessment of ten isomers in traveling 1 3 http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Journal of Flow Chemistry Springer Journals

Ion mobility spectrometry-tandem mass spectrometry strategies for the on-line monitoring of a continuous microflow reaction

Loading next page...
 
/lp/springer-journals/ion-mobility-spectrometry-tandem-mass-spectrometry-strategies-for-the-iFNSLJk2uk
Publisher
Springer Journals
Copyright
Copyright © The Author(s) 2021
ISSN
2062-249X
eISSN
2063-0212
DOI
10.1007/s41981-021-00209-7
Publisher site
See Article on Publisher Site

Abstract

Continuous flow chemistry is an efficient, sustainable and green approach for chemical synthesis that surpasses some of the limitations of the traditional batch chemistry. Along with the multiple advantages of a flow reactor, it could be directly connected to the analytical techniques for on-line monitoring of a chemical reaction and ensure the quality by design. Here, we aim to use ion mobility, mass and tandem mass spectrometry (IMS-MS and MS/MS) for the on-line analysis of a phar- maceutically relevant chemical reaction. We carried out a model hetero-Diels Alder reaction in a microflow reactor directly connected to the IMS-MS and MS/MS using either electrospray or atmospheric pressure photo ionization methods. We were able to monitor the reaction mechanism of the Diels Alder reaction and structurally characterize the reaction product and synthesis side-products. The chosen approach enabled identification of two isomers of the main reaction product. A new strategy to annotate the ion mobility spectrum in the absence of standard molecules was introduced and tested for its valid- ity. This was achieved by determining the survival yield of each isomer upon ion mobility separation and density functional theory calculations. This approach was verified by comparing the theoretically driven collision cross section values to the experimental data. In this paper, we demonstrated the potential of combined IMS-MS and MS/MS on-line analysis platform to investigate, monitor and characterize structural isomers in the millisecond time scale. Article Highlights Direct connection of a microflow reactor to the ion mobility-mass spectrometry and tandem mass spectrometry analytical technique and data comparison with the batch chemistry. On-line monitoring of the hetero-Diels Alder reaction and analysis of the reaction mechanism and elements. • Annotating the ion mobility peaks without the standard molecules Keywords Hetero-Diels Alder reaction · Ion mobility spectrometry · Tandem mass spectrometry · Atmospheric pressure photo ionization · Flow reaction · Annotating ion mobility spectrum Introduction Over the last decades, flow chemistry has garnered much attention in the pharmaceutical industry as it facilitates the efficient synthesis of Active Pharmaceutical Ingredients (API)s. Running the chemical reactions via continuous flow Darya Hadavi and Peiliang Han are shared co-first authors. stream through milli- or micro-sized channels has a pleth- ora of advantages compared to batch chemistry [1, 2]. Most * Darya Hadavi importantly, the narrow channels of a flow reactor provide d.hadavi@maastrichtuniversity.nl a larger surface-to-volume ratio, which leads to the efficient Division of Imaging Mass Spectrometry Maastricht transfer of energy, improved control over the reaction pres- University, Maastricht Multimodal Molecular imaging sure and enhanced mixing efficiency [3 –5]. These features of (M4i) Institute, Universiteitssingel 50, 6229ER Maastricht, a flow reactor make it feasible to run reactions at high or low The Netherlands Vol.:(0123456789) 1 3 176 Journal of Flow Chemistry (2022) 12:175–184 temperatures and promote sustainability by mitigating the The (hetero) Diels-Alder (DA) cycloaddition reactions risk of working with hazardous materials and reducing the are frequently applied for the synthesis of medicinal drugs costs and environmental footprints. Moreover, flow chemis- due to their relatively high stereo selectivity and efficiency try is considered as Green Analytical Chemistry (GAC) due [20]. Despite numerous benefits and the increased interest to decreasing the use of energy and reagents and eliminating in utilizing flow reactors to run a (hetero-) DA reaction, wastes [6, 7]. Meanwhile, precise control over the reaction pharma and chemical industries have been very slow in the conditions raises the consistency of manufacturing high- transition from batch to flow processing. As mentioned, quality products and facilitates the upscaling process [8, 9]. this is also due to the scarcity of suitable Process Analytical Coupling of flow reactors with analytical detection tech- Technologies (PAT)s [18]. To accelerate this transition, a nologies allows the real-time and on-line reaction monitor- model hetero-DA reaction carried out in a microflow reactor ing, facilitating process control. However, application of and connected to the IMS-MS and MS/MS for the on-line process analytical techniques such as Liquid Chromatog- monitoring of the reaction. Using this analytical technique, raphy (LC), UltraViolet-Visible (UV-Vis) spectroscopy and the reaction mechanism and structural isomers formed in Nuclear Magnetic Resonance (NMR) poses many chal- the hetero-DA reaction were assessed and the potential of lenges. Techniques such as NMR and UV-Vis fall short to combined IMS-MS and MS/MS experimentation were pro- study complex reaction mixtures over a large concentration foundly investigated. range. In addition, the selective detection of molecules lack- ing UV-Vis chromophores becomes a challenge. With regard to the separation techniques, such as LC, the whole analysis Results and discussion cycle is relatively long and the reaction mixture is at the risk of decomposing within the column [10]. More importantly, A model hetero-DA reaction was carried out in the Chem- it is challenging and rather time consuming to separate iso- trix Labtrix® microflow reactor directly connected to the mers, in particular stereoisomers, without a chiral column IMS-MS and MS/MS for the on-line monitoring of the reac- on LC [11]. Therefore, for the on-line reaction monitoring, tion (Fig. 1). The molecular identifications and separations there is a need for universal analytical detectors being sensi- were studied by Trapped Ion Mobility Spectrometry (TIMS) tive, selective, while requiring short (micro to millisecond) and Traveling Wave Ion Mobility Spectrometry (TWIMS) analysis time. incorporated in Quadrupole Time-Of-Flight (Q-TOF) mass Mass Spectrometry (MS) coupled with tandem mass spectrometer. The reagents and synthesized molecules were spectrometry (MS/MS) and Ion Mobility Spectrometry ionized by an Electrospray Ionization (ESI) and Atmos- (IMS) outperform other technologies by meeting the major- pheric Pressure Photo Ionization (APPI). The reaction was ity of the requirements for high quality reaction monitoring monitored at different temperatures and residence times. [12]. Unlike NMR and UV-vis, MS simultaneously sepa- rates reagents, impurities, intermediates, and products of a complex reaction in microseconds based on their mass over Hetero‑Diels alder reaction charge ratio (m/z). The IMS can surpass the limitations of LC by rapid (millisecond timescale) separation of isomeric In organic chemistry, a wide range of biologically active compounds are synthesised by (hetero-) DA reactions. In compounds with high selectivity and improved signal-to- noise ratio [13]. Moreover, the separation of molecular ions particular, the formation of 1,2-oxazine compounds is of special interest to synthesise bioactive products or stabilize based on their Collision Cross Section (CCS), size, shape, and charge by IMS enables elucidating structural informa- them and selectively functionalize natural products [21–23]. The 1,2-oxazine compounds are formed as a result of nitroso tion [14, 15]. Combined with MS/MS, the fragmentation pathways add to the potential to unambiguously assign the dienophile reaction with conjugated diene. Considering the importance of nitroso hetero-DA reactions in pharma [23], molecular structure of unknowns. Collision Induced Dis- sociation (CID) is one of the commonly used approaches in in this study, nitrosobenzene 1, as a dienophile, was exposed to react with methyl 2, 4-pentadienoate 2, as a diene, through tandem mass spectrometry for survival yield analysis [16]. Hence, connecting a continuous flow chemical reactor to a hetero-DA mechanism to form product 3 (methyl 2-phenyl- 3,6-dihydro-2H-1,2-oxazine-6-carboxylate) (Fig. 2a). The the IMS-MS and MS/MS, as an analytical tool, provides the possibility to monitor a reaction in real-time without stall- reaction was performed in 4-methyl-2-pentanone (4M2P) solvent with high boiling point (116 °C) to prevent over ing the production to collect samples in different steps of a reaction. In addition, the real-time information improves the pressuring in the microflow reactor within the range of tested temperatures (25–120 °C). The reaction efficiency in process control by reducing chemical waste, measurement time and most importantly, leads to a new dimension to qual- the microflow reactor was compared with the batch chem- istry reaction. ity by design [17–19]. 1 3 Journal of Flow Chemistry (2022) 12:175–184 177 Inner diluon 1 2 6 Third inlet port Main inlets to introduce reagent 4 8 Second stac mixer Stac mixer Outlet port 5 9 Residence me loop 8 cm 8 cm Outer diluon 2h10 Pa T-connector 8.5cm ID 0.004" BPR 30 cm 15cm Chemtrix R3227 ESI/APPI-IMS-MS and MS/MS Fig. 1 Schematic of the reaction set up to visualize the microreactor, main inlets, dilution inlets and connection to the MS analyzer Fig. 2 Hetero-DA reaction sche- (a) matics. a Reaction schematic of reagent nitrosobenzene (1) with reagent Methyl 2,4-penta- dienoate (2) in 4-methyl-2-pen- tanone (4M2P) or acetonitrile N N O Methyl 2,4-pentadienoate (2) (ACN) solvent. b Ring opening O O Mₘᵢ: 112.0524 Da of molecule (3) to form iso- meric molecule (4) and its dehy- 4M2P/ACN drated molecular structure (5). Nitrosobenzene (1) Reaction product (3) c Two possible reaction mecha- Mₘᵢ: 107.03711Da Mₘᵢ: 219.0895 Da nisms of α-oxyamination and α-aminoxyation for the reaction of nitrosobenzene (1) with the (b) solvent molecules (4M2P) -H₂O N N O N O O HO O O (3)(4) (5) Mₘᵢ: 201.0789 Da (c) OOH α-oxyamination Ph (6) α-aminoxyation (1)4M2P O Ph NH (7) M ᵢ: 207.1259 Da 1 3 M2 ethyl , 4-pentadienoate in 4M2P solvent Nitrosobenzene in 4M2P solvent 178 Journal of Flow Chemistry (2022) 12:175–184 indicated that the investigated hetero-DA reaction formed The characterization of the reaction elements by ESI and APPI – MS and MS/MS two structurally isomeric molecules (3 and 4). The peak at m/z 225 was annotated as the protonated Connecting the microflow reactor to the IMS-MS and dimer ([2  M + H] ) of the reaction reagent 2. The identity of this peak was confirmed by fragmenting the precursor MS/MS enabled the tentative structural identification of the reaction products, including side-products. The mass ion (m/z 225) in collision cell and annotating the fragment peaks (online resource Fig. S3a). This reaction outcome was spectrum of the reaction outcome generated after ionizing by ESI and APPI sources are depicted in online resource formed by self-dimerization of methyl 2,4-pentadienoate (2) and could be regarded as an impurity. It is because the dimer Fig. S1a, b. The ESI-MS contained the ions at m/z 220 and 242, being respectively assigned as the [M + H] and was already present in the commercial reagent prior to run- ning the reaction. This observation was also verified by the [M + Na] ions of the reaction product 3. Although the reaction outcome ionized under soft conditions (Table. impurities on NMR (online resource Fig. S3b). The peaks at m/z 199 and 230 were identified as the side- S1), the reaction mass spectrum included several fragment ions caused by upfront CID processes [24]. According to products of hetero-DA reaction. Unlike the protonated ion of the main product (3), m/z 230 required higher collision the tandem mass spectrometry of product 3, ions with m/z 202, 190, 170,160, 132 and 94 were being directly energy (20 eV) to yield 70% disassociation. This observation implies that this molecule is more stable compared to the originated from the parent ion m/z 220 (online resource Fig. S1c). main reaction product and likely having a different molecu- lar structure. Merino et al., suggests that ketone groups tend To ensure regioselectivity of the hetero-DA reaction to form product 3, the reagents were reacted in batch with ace- to react with nitroso groups by either α-oxyamination or α-aminoxyation mechanism (Fig. 2c) [27]. Hence, the nitros- tonitrile solvent, similar to a previous study [25]. According 1 13 to the H and C NMR spectra (online resource Fig. S1d obenzene reagent (1) could also react with the solvent 4M2P and form the side-products 6 and/or 7. This reaction pathway and 1e) the outcome of this reaction was consistent with the outcome of the microflow reaction. was confirmed by running an additional batch reaction with reagent 1 and 4M2P (without reagent 2) at 60 °C overnight. The (fragment) ion at m/z 202 was formed by the loss of H O from product 3. The tandem mass spectrum of the The mass spectrum of this overnight reaction was com- parative to the spectrum of the hetero-DA reaction (online precursor ion m/z 202 showed a base peak at m/z 170 (online resource Fig. S2a). This fragment ion was formed due to the resource Fig. S4); in a way that both spectra included peaks at m/z 208 and 190. Based on the MS/MS analysis, these loss of methanol. This fragmentation pathway indicated that + + the loss of H O from the molecule 3 did not occur from the peaks were respectively [M + H] and [M + H-H O] ions of the side-products 6 and/or 7. Consequently, the peak at m/z ester group. The only remaining source of oxygen for water loss was in the oxazine ring. The oxazine could go through 230 is the sodium adduct of the side-products. The tendency of the side-product molecules to lose water suggested that the ring opening process in the microflow reactor to form the molecule 4, subsequently lose a water molecule through the reaction occurred only through the α-oxyamination path- way to form side-product 6 rather than 7. The sodium adduct the cyclization reaction to form a compound 5 (Fig.  2b). To further confirm this mechanism, the reaction product in formation of the side-product 6 in the gas phase explains the stability of the peak at m/z 230 owing to sharing the positive the liquid phase was isolated and purified using preparative Thin Layer Chromatography (TLC) and tested by NMR. The charge of Na with oxygen of ketone and hydroxyl groups. The identity of the low abundant peak at m/z 199 was also NMR spectrum of the peak at m/z 202 showed a methyl group at 3.7 ppm as well as three protons of the pyrrole determined by MS/MS and verified by NMR analysis (online resource Fig. S5). As reported earlier [28], nitrosobenzene ring with sp2 environment, unlike the molecule 3 that had only two sp2 protons (online resource Fig. S2b). In addi- molecules can react with each other to form a dimer, which explains the peak at m/z 199 as the side-product of the tion, the absence of a proton at 5 ppm close to the ester group indicated the pyrrole shape of the molecule (Fig. 2b). reaction. The APPI-MS data were comparative to the ESI-MS. Hence, the peak at m/z 202 was annotated as the molecule 5, which could be formed in the liquid phase due to the con- The APPI source resulted in the formation of both [M]˙ and [M + H] ions of the reaction products at m/z 219 and densation reaction and/or in the gas phase due to the water loss from the molecule 4. Consistent with this finding, the 220 respectively (online resource Fig. S1b). The APPI forms radical cations after being excited to high energic states or N-O bond energy of molecule 3 is 201 kJ/mol, which is a smaller value than other bonds in the ring (C-C 347 kJ/mol, by electron transfer. In the latter case a neutral species such as toluene, known as Dopant (D), absorbs photon energy C-N 305 kJ/mol, C-O 358 kJ/mol) [26]. The relatively weak bond energy of N-O explains the position of ring-opening from the light source in the ionization chamber. The inter- nally excited dopant (D ) ionizes a neutral target Molecule reaction to form molecule 4 from molecule 3. These results 1 3 Journal of Flow Chemistry (2022) 12:175–184 179 (M) that has lower ionization energy by electron transfer. indicated the presence of at least two compounds with the Consequently, the radical cations are formed (Eq. 1). same mass but different 3D structures. Therefore, the hetero- DA reaction produced two isomeric compounds (molecules ∗ +∙ − D + M → D + M + e (1) 3 and 4) with the same m/z value but different mobility. This observation is in line with the represented mechanism for Since the hetero-DA reaction was performed in the the reaction (Fig. 2b). The IMS of the radical cation at m/z 4M2P solvent, the reaction products were mainly pro- 219, which was formed by APPI source, demonstrated better tonated ([M + H] ). This is due to the high proton affin- resolution than the protonated ion at m/z 220 (Fig. 3b). This ity (≈832.7 kJ/mol at 300 K) of the 4M2P solvent, which could be due to the formation of intramolecular H bridges in imposes the protonation of molecules through the Eq. 2. protonated molecules, which increases structural similarity + + S + H + M → S + M + H of 3 and 4. The IMS of the reaction elements were consistent n n + (2) S + M + H → nS + [M + H] regardless of the type of ionization source (APPI or ESI) and the ion mobility instrument (TWIMS or TIMS), even though When the solvent of the reaction was exchanged with TIMS resulted in better resolution data. dichloromethane, having a lower proton affinity (628.0 kJ/ Apart from the isomeric separation, it is also essential to mol at 300 K) compared to 4M2P, the reaction products annotate each ion mobility peak to a correct molecular struc- formed radical cations. This process demonstrated the ver- ture. To this end, different methods were employed. One of satility of the APPI source to utilize in a diverse range of the unique features of TWIMS-TOF is that a mass analyser organic synthesis with no restrictions on the solvent choice. quadrupole is placed before the ion mobility cell, which Regardless of the solvent type, molecules are ionized either enables performing MS/MS on the isolated precursor ions by proton or electron transfer. The ESI and APPI -MS-MS/ after IMS (IMS-MS/MS). Since the internal energy of an ion MS data of the isolated batch reaction product were consist- and its decomposition reaction rate are structural dependent ent with the acquired data from microflow reactor (online features, it is known that even isomeric chemical structures resource Fig. S6). show unique stability and survival yield under CID [30]. Hence, for the first time, we employed IMS-MS/MS method ESI and APPI‑IMS‑MS/MS to define the CE of the isomeric reaction products. The m/z 220 was first isolated in the quadrupole and then separated The ion mobility spectrometry of the reaction products was in the ion mobility cell followed by CID. When the precur- studied by TWIMS and TIMS instruments. Each technique sor ions exposed to the increasing CE after IMS separa- has different separation mechanism as explained before tion, isomeric ions’ survival decreased in a sigmoidal form [29]. Annotating the IMS of isomeric compounds becomes (Fig. 3c). This graph transformed to a linear survival yield a challenge when there are no standard molecules for IMS (Fig. 3d) to find the CE as reported earlier [30]. The results comparison. This challenge is even more pronounced for the identified unique CE of 14.89 and 15.82 eV for the isomer reactions leading to unknown isomeric forms of the main with the short drift time (td 4–4.5 ms) and long drift time product. In such instances, even if each isomer is fragmented (td 4.5-5 ms) respectively. The molecule 4 could have higher by MS/MS through a distinctive pathway, the tandem mass stability due to its lower Gibbs free energy after opening the spectrum contains the fragment ions of all isomers, with- ring, caused by the formation of stronger bond (smaller ΔH) holding the separation of isomers based on one/multiple and increased flexibility (larger ΔS). To verify this assump- unique product ion(s). The combination of ion mobility tion the minimum energy of both compounds was calcu- with tandem mass spectrometry brings another dimension lated by Density Functional Theory (DFT). Regardless of to molecular ion identifications. By applying IMS-MS/MS the protonation site, the molecular ion 4 always had lower approach, the generated fragment ions of each isomer will global minimum energy than the protonated molecular ion have the same arrival time as their parent isomer. Isomers 3 (Table. S2). These findings indicated the higher stability could be separated and identified based on a unique daugh- of the molecular ion 4 compared to 3. Therefore, based on ter ion under different mobility peaks or unique Collision CE and DFT calculations the drift time td 4–4.5 ms and Energy (CE) that results in 50% survival yield (CE ). Ion 4.5–5 ms were respectively annotated to the molecular ions mobility spectrometry also provides additional information 3 and 4. regarding the 3D structure of each ion. This is possible by To further verify the identity of the annotated ion mobil- TWIMS experimentally measuring the ion-neutral gas CCS and com- ity peaks, the experimentally derived CCS was com- N2 paring it to the theoretically calculated CCS values. pared to the theoretically calculated CCS of the molecular The extracted ion mobility (EIM) of the ion at m/z 220 ions 3 and 4. Table. S3 shows the calculated CCS values was reconstructed to plot its ion mobility spectrum. The plot- of the molecules 3 and 4 that were protonated on nitrogen ted IMS showed a bimodal peak (Fig. 3a). The bimodal peak or ester group. Based on the theoretically calculated data, 1 3 180 Journal of Flow Chemistry (2022) 12:175–184 Fig. 3 IMS and IMS-MS/MS (a) 1.2E+6 (c) 1.2 EIM of m/z 220 analysis of the reaction prod- td 4-4.5 ms ucts. a EIM of the protonated td 4.5-5 ms reaction products at m/z 220 analysed by ESI-TWIMS and ESI-TIMS. b EIM of the radical cation of reaction products at m/z 219 analysed by APPI- 0.0E+0 TIMS. c) Experimental survival 3.5 4.5 5.5 yield of isomers upon ion td (ms) (b) (d) CE (eV) mobility separation and CID at 4.8E+4 EIM of m/z 219 CE of 5, 10, 15, 20, 25, 30 eV. y = 0.4999x -7.912 Filled blue circle corresponds to R² = 0.9933 the isomer with short drift time (td 4-4.5ms) and filled orange y = 0.294x -4.3793 triangle corresponds to the R² = 0.9775 isomer with longer drift time (td 4.5-5ms). d Linear survival yield of the isomer at td 4-4.5 0.0E+0 td 4-4.5 ms ms in filled blue circle and the 0.45 0.55 0.65 isomer at td 4.5-5 ms in filled (e) Mobility, 1/K0 [V·s/cm²] td 4.5-5 ms -4 orange triangle. e Comparing CE (eV) 1.4E+6 the theoretically calculated CCS of the protonated molecules (3) and (4) on Nitrogen and ester groups with the experimentally TWIMS derived CCS for the 0.0E+0 N2 detected ion mobility peaks 3.54 4.555.5 6 (dashed red lines) N O N O td (ms) O HO O O (3) (4) CCS= 157.0717 (Å2) CCS= 153.5967 (Å2) CCS= 163.3064 (Å2) CCS = 158.8419 (Å2) Protonated on N Protonated on ester Protonated on N Protonated on ester TWIMS TWIMS CCSN2= 147.044 ±0.79 (Å2) CCSN2= 152.3177 ±0.80 (Å2) products before the ion mobility cell, during the traverse the CCS of the protonated ion 4, regardless of the proto- nation site, was always higher than the protonated mol- from the quadrupole until the IMS cell. Ions with the identical drift time and increasing m/z value are related to ecule 3, which was in line with the open structure of 4. TWIMS Hence, the experimentally derived CCS value of fragment ions generated after the ion mobility cell, from N2 the transfer cell till the detector (white dashed line ellipse- 152.3177 ± 0.80 Å corresponds to the molecular ion 4 and TWIMS 2 the smaller CCS value of 147.044 ± 0.79 Å cor- online resource Fig. S7). N2 All these findings highlighted the added value of IMS- responds to the more compact molecular ion 3 (Fig.  3e). These molecules were more likely protonated on the ester MS and MS/MS experiments on providing in depth knowl- edge on the molecular structure (identity) of (un)known group, as the theoretical and experimental CCS values are in less than 5% agreement upon protonation on the ester group. reaction products in the millisecond time scale. These process analytical technologies and reaction monitoring With these data, td 4–4.5 ms was again annotated to be the protonated molecule 3 and the td 4.5-5 ms was related the strategies not only enabled separation, but also enabled the tentative identification of isomeric compounds that protonated molecule 4. The drift time/mass-to-charge heatmap of the reaction are not distinguishable by MS or MS/MS. It also enabled annotation of IMS peaks without using a standard com- outcome also gave a quick insight regarding the stabil- ity of the reaction. Online resource Fig. S7 presented the pound. Ionizing molecules with APPI combined with dif- ferent IMS-MS/MS scanning strategies shed further light low stability of the reaction products in TWIMS-MS. The heatmap of the isolated m/z 220 in the quadrupole, dem- to characterize molecules. With the combination of ESI/ APPI-IMS-MS/MS the reaction products and intermedi- onstrated the presence of ions with the raising m/z and drift time values (white solid line ellipse- online resource ates of the model hetero- DA reaction were characterised. Fig. S7). This trend showed the dissociation of the reaction 1 3 Counts Counts Counts Survival Yield In{(1-SY)/SY} Journal of Flow Chemistry (2022) 12:175–184 181 Monitoring the microflow reaction by ESI‑IMS‑MS/ Conclusion MS The ion mobility-mass spectrometry and tandem mass The reaction of nitrosobenzene with methyl 2, 4-pentadi- spectrometry technique was used for the on-line moni- enoate was studied by directly connecting the Chemtrix toring the model hetero-DA reaction carried out in the Labtrix® microflow reactor to ESI/APPI-IMS-MS/MS microflow reactor and compared to batch reaction. When analytical technique and compared to the batch chemistry utilizing IMS-MS and MS/MS strategies, the elements results. The model hetero-DA reaction was investigated of the reaction, including the isomeric reaction products, under various temperatures (25, 40, 60, 80, 100, 120 °C) impurities and side-products were identified, character - and residence times (5, 10, 20, 28, 40, 80 min) with constant ized and separated. By systematically increasing the col- pressure (2 × 10  Pa) and concentrations (50 mM methyl 2, lision energy after the ion mobility separation, the survival 4-pentadienoate and nitrosobenzene (1:1)). The MS spectra yield of each isomer was calculated. The acquired sur- were normalised to the intensity of the internal standard. vival yield was compared to the global minimum energy Figure 4a and b presented the impact of residence time of ions calculated by DFT, which enabled annotating the on the formation of reaction products (3 and 4) performed ion mobility spectrum of isomers in the absence of stand- in the microflow versus batch reactor. The formation of the ard molecules. The efficiency of the reaction was higher product molecules in both reactors were increased by raising when it was carried out in the microflow reactor com- the residence time. However, the reaction efficiency in the pared to the batch reaction. The introduced strategies of microflow reactor was higher than the batch reactor. In the hyphenating IMS-MS and MS/MS to a microflow reactor batch reactor the abundance of the product ions increased 7 showed possible potentials to simultaneously detect differ - times after 7.38 h of reaction, while in the microflow reac- ent reaction products and above all their molecular struc- tor this was done in only 40 min. The online monitoring ture (identity). The potential of APPI to act as a universal of the reaction whilst raising the temperature indicated a ionization technique, allowing the application of a wide systematic increase of the reaction products (Fig. 4c). These variety of organic solvents, was introduced. The generation results demonstrated the potentials of microflow reactors and of radical ions through APPI and their specific fragment on-line measurements in increasing the reaction efficiency, ions add another dimension to the structural identification improving reproducibility, reducing the number of experi- process. It is anticipated that this approach be a valuable ments and lowering chemical consumptions compared to technology platform for the development of efficient and running multiple parallel experiments at different conditions green chemical synthesis routes, readily linking reaction required by batch chemistry. conditions to the presence of different reaction pathways. Fig. 4 a Residence time effect (a) (c) Microflow reaction Microflow reaction on the formation of the reaction products after a) microflow 0.5 reaction, b batch reaction. c Temperature effect on the formation of reaction products carried out in the microflow reactor 20 70 120 Temperature (oC) Residence time (min) (b) Batch reaction 4.2 0500 1000 1500 Residence time (min) 1 3 Normalized Normalized intensity intensity Normalized intensity 182 Journal of Flow Chemistry (2022) 12:175–184 (1 ml- Hamilton, Amsterdam, The Netherlands), flow rate Experimental controllers (Chemyx Fusion 200 syringe pump- Stafford, TX, USA), temperature controller, Back Pressure Regulator Materials and sample preparations (BPR), outer dilution inlet and a start unit, which accom- modated the glass microreactor (Chemtrix R3227 type). The Methyl-2,4-pentadienoate, 4-methyl-2-pentanone, nitroso- Chemtrix R3227 microflow reactor consisted of two main benzene, 4-chloropiperidine hydrochloride, Dichlorometh- inlets, static mixer, residence time loop, a third inlet port ane (DCM), Acetonitrile (ACN), toluene, chloroform-d, to introduce a dilution solvent (inner dilution inlet), a sec- Methanol (MeOH), acetic acid and water (H O) with >98% ond static mixer and an outlet port. This microreactor has purity, ESI-L low concentrations tuning mix, APPI/APCI the effective reaction volume of 19.5 μl and it provides the tuning mix and polyalanine purchased from Sigma-Aldrich residence times from 0.78 to 97.5 min (for A + B reaction) (Zwijndrecht, The Netherlands). with the (total) flow rate ranging between 0.2 and 25 μl/ min. The outlet port directly connected to a BPR to keep The reaction condition in batch and isolation the pressure at constant value of 2 × 10  Pa. Right after the process BPR, a T-connector was placed to mix the reaction out- come with the dilution solvent mixture (Fig. 1 and online To perform the hetero-DA reaction in 4-methyl-2-pentanone resource Fig. S8). The line coming out of the T-connector solvent, a solution of nitrosobenzene (53.2 mg, 0.497 mmol) was directly connected to ESI or APPI ionization source, and methyl 2,4-pentadienoate (58.1  μL, 0.499  mmol) in placed in front of the MS orifice. 20 ml of 4M2P was stirred at 60 °C for 25 h. To study the The reaction was conducted by continuous injection of impact of residence time on the production efficiency, 50 μl nitrosobenzene (50  mM) and methyl-2,4-pentadienoate of the reaction mixture was collected after 0.39, 1.63, 4.65, (50 mM) solutions dissolved in 4M2P through main inlets. 7.38, and 25.1 h. Samples were diluted with 150 μl of a Dilution solvent mixture consisted of MeOH/ H O/ACN dilution solvent mixture, including 1 mM of internal stand- (2:1:1) and 0.1% acetic acid introduced to the reaction out- ard and 0.1% acetic acid to make it comparable to the flow come through inner and outer dilution inlets. As an internal reaction. When acetonitrile was used as the reaction solvent standard 1 mM of 4-chloropiperidine hydrochloride added instead of 4-methyl-2-pentanone, a solution of nitrosoben- to the reaction outcome through the inner dilution inlet. To zene (80.1 mg, 0.748 mmol) and methyl 2,4-pentadienoate monitor the impact of residence time on hetero-DA reaction (87.0 μL, 0.749 mmol) in acetonitrile (10 ml) was stirred at the temperature was set to the constant optimised value of 60 °C for 8 h. The same method was applied for the reac- 60 °C. The flow rate of each starting materials set to 2, 1, tion of nitrosobenzene (53.2 mg, 0.497 mmol) with 20 mL 0.5, 0.35, 0.25 and 0.12 μl/min to reach the residence times 4-methyl-2-pentanone for the specificity experiment. The of approximately 5, 10, 20, 28, 40 and 80 min (respectively). outcome of each batch reaction either analysed as a complex The flow rate of the dilution solvent on inner dilution inlet sample or purified by TLC preparative plate (Sigma-Aldrich, always set as the sum flow rate of each starting material. The Zwijndrecht, the Netherlands) prior to analysing. For the outer dilution inlet was used to further dilute the reaction reaction in 4M2P, the reaction mixture diluted with water outcome. Its flow rate was set to 8, 4, 2, 1.4, 1 and 0.5 μl/min (20 ml) and extracted with ethyl acetate (3 × 15 ml). For the respectively for shortest to longest residence time. To moni- reaction in acetonitrile, 15 ml of water and 3 × 10 ml of ethyl tor the effect of temperature, the reaction residence time acetate was used for dilution and extraction. Subsequently, was set to the constant optimised value of 20 min. The reac- the combined organic layer was washed with brine (10 ml) tion was monitored under increasing temperatures of 25, 40, and dried with anhydrous Na SO . After filtration, the solu - 60, 80, 100 and 120 °C. For APPI experiments, DCM was 2 4 tion was concentrated under vacuum and purified using TLC used instead of 4M2P as the reaction solvent. For this set (eluted with 1:1 dichloromethane/ n-hexane) to afford the of experiments, dopants including pure toluene and DCM/ product as a white solid (32.9 mg, 30% for the reaction in toluene (1:1) were respectively injected through the inner 4M2P and 74 mg, 45% for the reaction in acetonitrile). and outer dilution inlets. ESI/APPI‑IMS‑MS/MS parameters The microflow reactor platform and the reaction condition Synapt HDMS G2Si (Waters, Milford, MA, USA) equipped with a TWIMS, as well as a Trapped-Ion Mobility Spec- The reaction was carried out in the Chemtrix Labtrix® trometry-Q-TOF (Bruker Daltonics Inc., Billerica, MA, microflow reactor set up (Fig.1 and online resource Fig. S8) USA) were used for the ion mobility and tandem mass spec- consisted of 4 gas tight glass syringes with luer lock fitting trometry analysis. Table. S1 presents the set parameters for 1 3 Journal of Flow Chemistry (2022) 12:175–184 183 1 13 15 1 13 IMS measurements on each instrument. Nitrogen gas was channel [ H, C, N] TCI probe. The H and C spectra used in the ion mobility cell and helium gas was used before were recorded as 250 mM in CDCl solutions with 0.03% the ion mobility cell for collisional cooling. Argon was used TMS as internal chemical shift reference standard. Spectra as the collision gas in the trap and transfer cells. The ESI were processed using Topspin 3.2 softwareand analyzed and APPI experiments performed in positive ion mode under by the program Sparky 3.114 (Goddard & Kneller) and the optimized ionization conditions listed in Table. S1. The MestReNova. mass spectra acquisition range was set between m/z 50 and 300 Da. ESI-L low concentrations tuning mix and APPI/ Data evaluations APCI tuning mix were used for the mass calibration and polyalanine was used to the IMS calibration. On the TIMS The acquired data by Synapt G2Si analysed by MassLynx instrument, the massive ions with lower mobility (k) elute v4.1. and DriftScope v2.8 software. The measured data by first compared to the lighter and compact ions, however the TIMS-TOF analysed by Compass DataAnalysis v5.0 soft- IMS graph is constructed based on 1/k (V.s/cm ) values, ware (Bruker). The extracted ion mobility spectra extracted making it comparable to the IMS generated by the TWIMS and plotted in excel for further analysis. To correct for the instrument. ion suppression effects, the intensity of the reaction prod- For IMS-MS/MS experiments, the precursor ion was iso- uct ions normalised with the intensity of 4-chloropiperidin lated in the quadrupole and fragmented via CID after the (the internal standard). For CCS calculations, the molecules ion mobility cell in the transfer cell. The collision energy structurally optimized by DFT on Spartan software (Wave- was increased from 5 to 30 eV in 5 eV increments (total of function, Inc. & Q-Chem, Irvine, CA, USA). Conformer dis- 6 collision energies). The MS/MS results were the average tribution calculations were performed to obtain the global of 2 min acquisition time per collision energy. To extract the energy minimum with Density Functional ωB97X-D and CID spectrum of each isomer from its ion mobility, first the basis set 6-31G* using Molecular Force Field (MMFF) ion mobility of the precursor ion at m/z 220 was extracted. geometry. The software considered 100 conformers, while Next, the mobility range of interest were selected (e.g. from keeping the E ≤ 40 kJ/mol and the energy of maximum 10 4 until 4.5 ms). The peak picking then performed in the conformers calculated while keeping the E ≤ 15 kJ/mol. The non-chromatic mode using the following parameters: min IMoS software used for theoretical CCS calculations with drift time peak width of 0.2 ms, MS resolution of 35,000 Trajectory Method (TM) with 300.000  N gas molecules per and minimum intensity threshold of 1000 counts. The CID orientation. Mulliken charges were used for partial atomic spectra were copied to excel for further analysis. Previously charges, they were automatically calculated by Spartan. explained methods [30] were employed to plot the experi- Calculations performed for each conformer with Boltzmann mental and linear survival yield graphs. The experimental weight above 1%. survival yield was plotted based on the ratio of the intensity of the Precursor ion (P) to the sum of the intensities of the Supplementary Information The online version contains supplemen- precursor and Fragment ions (F) (Eq.  3). This graph was tary material available at https://doi. or g/10. 1007/ s41981- 021- 00209-7 . transformed to linear form by Eq. 4 and 5. Acknowledgements We would like to thank Hans Ippel (CARIM, IntensityP Maastricht University) for the execution and data interpretation of the Experimental survival yield = NMR experiments and Yuandi Zhao (M4i, Maastricht University) for IntensityP + (IntensityF) his artistic assistance. (3) Sigmoidal survival yield = Declarations (4) −b.CE 1 + c ∗ e Conflict of interest The authors declare that they have no financial or where b is the slope of the linear segment, and the intercept non-financial conflict of interest. of the linear portion is the natural logarithm of c. Open Access This article is licensed under a Creative Commons Attri- 1 − Survival yeild bution 4.0 International License, which permits use, sharing, adapta- ln = ln(c)− b(CE) (5) Survival yeild tion, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are NMR included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in The Bruker (Rheinstetten, Germany) Avance III HD the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will NMR spectrometer operated at the H frequency of 700 MHz and equipped with a cryogenically cooled triple 1 3 184 Journal of Flow Chemistry (2022) 12:175–184 need to obtain permission directly from the copyright holder. To view a wave- and trapped- ion mobility spectrometry. Rapid Commun copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. Mass Spectrom 33. https:// doi. org/ 10. 1002/ rcm. 8419 16. Iacobucci C et al (2014) Insight into the mechanisms of the mul- ticomponent Ugi and Ugi–smiles reactions by ESI-MS(/MS). Eur J Org Chem 2014(32):7087–7090. https:// doi. org/ 10. 1002/ ejoc. 20140 3179 References 17. Vargas JM et al (2018) Process analytical technology in continu- ous manufacturing of a commercial pharmaceutical product. Int 1. Bogdan AR, Dombrowski AW (2019) Emerging trends in flow J Pharm 538(1):167–178. https://doi. or g/10. 1016/j. i jphar m.2018. chemistry and applications to the pharmaceutical industry. J Med 01. 003 Chem 62(14):6422–6468. https://doi. or g/10. 1021/ acs. jmedc hem. 18. Price GA, Mallik D, Organ MG (2017) Process analytical tools for 8b017 60 flow analysis: a perspective. Journal of Flow Chemistry 7(3):82– 2. Baumann M et al (2020) A perspective on continuous flow chem- 86. https:// doi. org/ 10. 1556/ 1846. 2017. 00032 istry in the pharmaceutical industry. Org Process Res Dev. https:// 19. Djuris J, Ibric S, Djuric Z (2013) 1 - Quality-by-design in pharma- doi. org/ 10. 1021/ acs. oprd. 9b005 24 ceutical development, in Computer-Aided Applications in Phar- 3. Reckamp JM et  al (2017) Mixing performance evaluation for maceutical Technology, J. Djuris, Editor. Woodhead Publishing. commercially available micromixers using Villermaux–Dushman p. 1–16. https:// doi. org/ 10. 1533/ 97819 08818 324.1 reaction scheme with the interaction by exchange with the mean 20. Nicolaou KC et al (2002) The Diels–Alder reaction in Total syn- model. Org Process Res Dev 21(6):816–820. https:// doi. org/ 10. thesis. Angew Chem Int Ed 41(10):1668–1698. https:// doi. org/ 1021/ acs. oprd. 6b003 32 10. 1002/ 1521- 3773(20020 517) 41: 10< 1668:: Aid- anie1 668>3. 0. 4. Schwolow S et al (2012) Application-oriented analysis of mixing Co;2-z performance in microreactors. Org Process Res Dev 16(9):1513– 21. Brulíková L et al (2016) Stereo- and regioselectivity of the hetero- 1522. https:// doi. org/ 10. 1021/ op300 107z Diels-Alder reaction of nitroso derivatives with conjugated dienes. 5. Schwolow S et al (2016) Design and application of a millistruc- Beilstein J Org Chem 12:1949–1980. https://doi. or g/10. 3762/ bjoc. tured heat exchanger reactor for an energy-efficient process. Chem 12. 184 Eng Process Process Intensif 108:109–116. https:// doi. org/ 10. 22. Gamenara D (2001) Hetero Diels-Alder adduct formation between 1016/j. cep. 2016. 07. 017 nitrosobenzene and tetra-methyl purpurogallin and its retro-Diels- 6. Elvira KS et al (2013) The past, present and potential for micro- Alder reaction. J Braz Chem Soc 12(4):489 fluidic reactor technology in chemical synthesis. Nat Chem 23. Carosso S, Miller MJ (2014) Nitroso Diels-Alder (NDA) reaction 5(11):905–915. https:// doi. org/ 10. 1038/ nchem. 1753 as an efficient tool for the functionalization of diene-containing 7. de la Guardia M, Garrigues S (2020) Chapter 1 past, present and natural products. Organic & biomolecular chemistry 12(38):7445– future of green analytical chemistry, in challenges in green ana- 7468. https:// doi. org/ 10. 1039/ c4ob0 1033g lytical chemistry (2). The Royal Society of Chemistry. p. 1-18. 24. Oberacher H et al (2012) On the inter-instrument and the inter- https:// doi. org/ 10. 1039/ 97817 88016 148- 00001 laboratory transferability of a tandem mass spectral reference 8. Campbell Brewer A et al (2019) Development and scale-up of a library. 3. Focus on ion trap and upfront CID. J Mass Spectrom continuous aerobic oxidative Chan–lam coupling. Org Process 47(2):263–270. https:// doi. org/ 10. 1002/ jms. 2961 Res Dev 23(8):1484–1498. https:// doi. or g/ 10. 1021/ acs. opr d. 25. Monbaliu J-CMR et al (2010) Straightforward hetero Diels–Alder 9b001 25 reactions of nitroso dienophiles by microreactor technology. Tet- 9. Schwolow S et  al (2014) Kinetic and scale-up investigations rahedron Lett 51(44):5830–5833. https:// doi. org/ 10. 1016/j. tetlet. of a Michael addition in microreactors. Org Process Res Dev 2010. 08. 117 18(11):1535–1544. https:// doi. org/ 10. 1021/ op500 2758 26. Cottrell TL (1958) The strengths of chemical bonds. Butterworths 10. Fiers T et al (2012) Development of a highly sensitive method Scientific Publications, London for the quantification of estrone and estradiol in serum by liquid 27. Merino P et al (2016) Recent advances on asymmetric Nitroso chromatography tandem mass spectrometry without derivatiza- aldol reaction. Synthesis 48. https://doi. or g/10. 1055/s- 0035- 15615 tion. J Chromatogr B 893-894:57–62. https:// doi. org/ 10. 1016/j. jchro mb. 2012. 02. 034 28. Taft RW, Klingensmith G, Ehrenson S (1965) Multipolar com- 11. Kuki Á et al (2012) Identification of Silymarin constituents: an plexes. I. the dimerization of nitrobenzene1. J Am Chem Soc improved HPLC–MS method. Chromatographia 75(3):175–180. 87(16):3620–3626 https:// doi. org/ 10. 1007/ s10337- 011- 2163-7 29. Cumeras R et al (2015) Review on ion mobility spectrometry. Part 12. Emwas A-HM (2015) The strengths and weaknesses of NMR 1: current instrumentation. Analyst 140(5):1376–1390. https://d oi. spectroscopy and mass spectrometry with particular focus on org/ 10. 1039/ c4an0 1100g metabolomics research. In: Bjerrum JT (ed) Metabonomics: meth- 30. Kertesz TM et al (2009) CE50: quantifying collision induced dis- ods and protocols. Springer New York, New York, pp 161–193. sociation energy for small molecule characterization and identi- https:// doi. org/ 10. 1007/ 978-1- 4939- 2377-9_ 13 fication. J Am Soc Mass Spectrom 20(9):1759–1767. https:// doi. 13. Tadjimukhamedov FK et al (2008) Liquid chromatography/elec- org/ 10. 1016/j. jasms. 2009. 06. 002 trospray ionization/ion mobility spectrometry of chlorophenols with full flow from large bore LC columns. Int J Ion Mobil Spec- Publisher’s note Springer Nature remains neutral with regard to trom 11(1):51–60. https:// doi. org/ 10. 1007/ s12127- 008- 0004-7 jurisdictional claims in published maps and institutional affiliations. 14. Sorribes-Soriano A et al (2018) Trace analysis by ion mobility spectrometry: from conventional to smart sample preconcentration methods. A review. Anal Chim Acta 1026:37–50. https://doi. or g/ 10. 1016/j. aca. 2018. 03. 059 15. Hadavi D et  al (2019) Adduct ion formation as a tool for the molecular structure assessment of ten isomers in traveling 1 3

Journal

Journal of Flow ChemistrySpringer Journals

Published: Jun 1, 2022

Keywords: Hetero-Diels Alder reaction; Ion mobility spectrometry; Tandem mass spectrometry; Atmospheric pressure photo ionization; Flow reaction; Annotating ion mobility spectrum

References