Time-Resolved Ion Mobility Spectrometry with a Stop Flow Confined Volume Reaction Region

An ion source concept is described where the sample flow is stopped in a confined volume of an ion mobility spectrometer creating time-dependent patterns of ion patterns of signal intensities for ions from mixtures of volatile organic compounds and improved signal-to-noise rate compared to conventional unidirectional drift gas flow. Hydrated protons from a corona discharge were introduced continuously into the confined volume with the sample in air at ambient pressure, and product ions were extracted continuously using an electric field for subsequent mobility analysis. Ion signal intensities for protonated monomers and proton bound dimers were measured and computationally extracted using mobilities from mobility spectra and exhibited distinct times of appearance over 30 s or more after sample injection. Models, and experimental findings with a ternary mixture, suggest that the separation of vapors as ions over time was consistent with differences in the reaction rate for reactions between primary ions from hydrated protons and constituents and from cross-reactions that follow the initial step of ionization. The findings suggest that the concept of stopped flow, introduced here for the first time, may provide a method for the temporal separation of atmospheric pressure ions. This separation relies on ion kinetics and does not require chromatographic technology.


API reactions
Atmospheric pressure ionization is a complex chain of reactions resulting to hydrated protons, called here reactant ions (H + (H 2 O) n or R + ). 35With reactant ions API of a vapor mixture can be complex and comprised of initial formation of ions which is followed in time by secondary (or cross) reactions in a vapor cloud of substances in air.Some cross-reactions will be more favored than others based on ionization properties of individual substances.
A descriptive example of possible reactions and their cross-reaction paths in case of two substances are presented in Table S2.It is notable, that primary ionization of any neutral is probable as long as reactant ions are present.Also, the cross-reactions do not become possible until at least one kind of product ion exist.
Primary ionization Cross-reactions With the generig loss terms, simplifying notation, and considering the tiny volume the reactions in Table S2 can be described in more generic way and in dynamic form as: Equation S1 describe the generation S R + and loss L R + of reactant ions via reactions with neutrals, ion extraction, or for example losses with unintended convection, Equation S2reduction of quantity of some neutral molecules via direct reactions of reactant ions or crossreactions with some product ions, and Equation S3 dynamic development of product ions S-5 via reactions of neutrals and reactant ions and via cross-reactions to and from the said product ions, and the losses L N i R + .Equations S1 to S3 can be applied to FEM-models where source and loss terms are given as initial values and boundary conditions, and where set of equations are linked to simulated coordinates.Because the quantity of reactant ions R + can be considered given, the kinetics is not defined only by the reaction coefficients k N i but rather by the product of the reaction coefficient and concentration (k ).This means that the order of ionized substances in the temporal separation can not be predicted until values of both rate coefficients and neutral concentrations are known.Furthermore, the duration of temporal separation overall relies on the total concentration of neutrals and occurs most rapidly at low concentrations.The extracted portion represent ions with the highest formation probability at the extraction moment and can be observed as a temporal separation of the the mixture components.This process is called Gas Ion Distillation, GID.

Spectrometer
Figure S2 presents the method of Flow control.In the first stage, the entire system is cleaned with purified flow through the drift region and reaction region, called as GID volume, and Corona discharge (CD).At the second stage, the drift flow is driven out from the flow drift region and sample is feeded into GID.At the third stage, sample feed is stopped, leaving the GID volume in the Stop Flow condition and reactions take place in the GID volume.
Ions are gated to the drift region and their signal intensity is measured.Once the predefined residence time has passed, the process starts over again with purge stage.

S-8
Single component spectra of sulcatone and 2-butanone samples Figure S3: Spectra in one second interval of 2-butanone (30 ppbv) referred to elapsed time 0. Peak at 2.1 cm 2 V •s is reactant ion, peak at 1.92 cm 2 V •s 2-butanone monomer, and peak at 1.64 cm 2 V •s butanone dimer.Shutter was operated at 500 µs.
Figure S3 presents the response of IMS when 2-butanone was injected into the reaction region.The control (reactant ions) is presented at t=0 s.At t=1s the sample has filled the reaction region, and both monomers and dimers has been formed.At t=2 s, both monomer and dimer peaks reach maxima and reactant ion peak is not detected.At t=3 s, the sample has diluted in the reaction region, monomer and dimer signals have reduced from the maxima and reactant ions can be observed again.This process continues until t=8 s, where no dimer peak can be observed, and monomer and reactant peaks overlap without any separation.
Figure S4: Spectra in one second interval of Sulcatone (20 ppbv) referred to elapsed time 0. Peak at 2.1 cm 2 V •s is reactant ions and peak at 1.72 cm 2 V •s is sulcatone monomer.Intensity of octane-2-7-dione (sulcatone impurity) monomer at 1.90 cm 2 V •s is not peak separated.Shutter was operated at 500 µs.V •s is reactant ions, peak at 2.2 cm 2 V •s is assumed to be ammonia and is unresolved when shutter is operated at 500 µs, unresolved peaks at 1.93 cm 2 V •s and 1.88 cm 2 V •s are 2-butanone monomer and sulcatone impurity monomer respectively, and unresolved peaks at 1.72 cm 2 V •s and 1.64 cm 2 V •s sulcatone monomer and 2-butanone dimer respectively.Shutter was operated at 250 µs.V •s is reactant ions, unresolved peaks at 1.93 cm 2 V •s and 1.88 cm 2 V •s are 2-butanone monomer and sulcatone impurity respectively, and peaks at 1.72 cm 2 V •s and 1.64 cm 2 V •s sulcatone monomer and 2-butanone dimer respectively.The shutter was operated at 250 µs.
Five component mixture measured in SFCV mode at NMSU, USA

Computational modelling
Reaction-based simulation predicts that separation of equal concentrations of neutrals M and N is a possible if there is difference in reaction rate coefficients. 37Plain reaction-based simulation does not account for effects from simulation space geometry, dynamic or spatial effects of sample introduction, gas flow in the reaction region, or transportation of ions in the electric field.

Testing the separation hypothesis with simulation
To study the realism of separation, while still idealized environment, a computational test model of reaction space followed by small drift-tube was simulated with Comsol multiphysics.The concentration of both M and N was equal, and thus the probability of ionization depends only on the reaction rate coefficient.The simulation predicts that at least equimolar binary mixture can be separated in time, if the residence time of reactant ions with the sample is long enough and there is a difference in their rate coefficients.In generic case, the separation is effective only when the ratio k rM [M ]  k rN [N ] ̸ = 1.

Figure S1 :
Figure S1: Top: Modified drift tube with thermal insulator interfaced with GC-inlet port.Bottom: Cut-off view of mechanical design of ion mobility spectrometer with flow control interface between extended reaction and drift regions.A: corona ionization and drift gas exit in UDF-mode, B: reaction region and sample injection, C: drift gas exit in SFCV-mode, D: ion shutter, E: drift region, F ion detector.

Figure S2 :
Figure S2: Method of flow control.

Figure
FigureS4presents the response of IMS when sulcatone was injected into the reaction region.The control (reactant ions) is presented at t=0 s.At t=1s, sulcatone sample has filled the reaction region, but has not yet reached the maximum.At t=2 s, and t=3s, the sulcatone peak has reached maximum.At t=4 s and t=5 s, the octane-2,7-dione (impurity in sulcatone) response is seen, and the return towards the baseline can be observed when t=8 s.

Figure S5 :
Figure S5: Selected spectra at selected moments referred to elapsed time 0 from Figure 5 top panel (UDF-mode), with 20 ppbv of 2-butanone and 50 ppbv of sulcatone sample.Peak at 2.1 cm 2V •s is reactant ions, peak at 2.2 cm 2 V •s is assumed to be ammonia and is unresolved when shutter is operated at 500 µs, unresolved peaks at 1.93 cm 2 V •s and 1.88 cm 2 V •s are 2-butanone monomer and sulcatone impurity monomer respectively, and unresolved peaks at 1.72 cm 2

Figure S6 :
Figure S6: Selected spectra at selected moments referred to elapsed time 0 from Figure 5 bottom panel (SFCV-mode), with 20 ppbv of 2-butanone and 50 ppbv of sulcatone sample.Peak at 2.1 cm 2V •s is reactant ions, unresolved peaks at 1.93 cm 2 V •s and 1.88 cm 2 V •s are 2-butanone monomer and sulcatone impurity respectively, and peaks at 1.72 cm 2 V •s and 1.64 cm 2 V •s sulcatone monomer and 2-butanone dimer respectively.The shutter was operated at 250 µs.

Five
component mixture was measured with an equivalent instrument as used for ternary and multi-mixture experiments.Measurements were performed at NMSU Department of Chemistry and Biochemistry, Las Cruces, USA by Oliver Hecht and Gary Eiceman.Instrumentation GID enhanced IMS was used.IMS drift voltage 6 kV .Corona needle at 9 kV .Sample was introduced over the side port into the GID region via a gas tight syringe.The drift tube and drift gas were heated to 50 • C.

Figure S7 :
Figure S7: Topographical plot of 5-component mixture averaged over 5 sample repetitions vs elapsed time.Colors show the relative signal intensity of ions.

Figure S8 :
Figure S8: Ion signal intensities of averaged intensities of selected ion mobilities as detector response vs elapsed time of 5 component mixture.The error bars show 4-sigma deviation between the averaged samples.The legend letters refer to Figure S7.

Figure S9 :
Figure S9: Hypothetical arrangement of GID as simulation model geometry to study effects of geometry and construction, and simulated concentration of reactant ions while counterflow was kept zero.From left to right: axisymmetric section of the model, R + , product ions M R + and product ions N R + at 60 ms when drift-field potential was set to 500 V, and sample concentration of M and N was 100 ppb.The sample was injected as 0.1 ms pulse at flow rate of 5 mL • min −1 .

Table S1 :
SFCV drift IMS configuration.Distances are relative from ring to ring.The first distance is measured from the inner surface of the supporting structure.