* The reactions in the lower section of the table yield a standard product ion. For displacement and elimination mechanisms, Cl⁻ and Br⁻ product ions have been most commonly encountered to date, so they are used for illustrative purposes. ** The reagent ion m/z is defined as a negative value here. Hence for O^−• (m/z = −16) the product ion following hydrogen atom transfer is OH⁻ (m/z = −17). Since a substantial breadth of literature describes volatile compounds that are detectable using SIFT-MS (see, for example, very convenient review articles ^[6][23][24][6,9,10]), several examples of important compounds that are not analyzable, or have significant constraints, are discussed here. First, consider methane which is of significant concern for climate due to its high global warming potential ^[25][27]. In positive ion mode, the O₂^+• reagent ion reacts with methane, as shown in Equation (1), whereas H₃O⁺ and NO⁺ are too inefficient to be useful ^[26][27][28,29]. However, the sensitivity of SIFT-MS is lower than for most volatiles due to the slow reaction rate coefficient (k)—over 200-fold lower than typical k for VOCs. Nevertheless, methane analysis has been demonstrated for breath ^[27][29] and results for landfill and biogas analysis were reported to commercial customers by the present researcheuthor over a 14-year period. Due to the low sensitivity, great care needs to be taken to ensure that the methane product ion does not suffer from interference. O₂^+•₄₂₂^+•k^{−123−1−1} (1) Interestingly, O^−• reacts with methane somewhat more rapidly (Equation (2)) ^[28][30] via the hydrogen atom transfer mechanism (Table 1). However, the product ion is also formed from reaction of O^−• with water ^[28][30] and other low-molecular-weight hydrocarbons ^[28][30], so it is not of practical use for specific detection of methane. O^−•₄⁻₃^•k^{−103−1−1} (2) Second, carbon monoxide is toxic ^[29][31] and exposure is of significant concern in confined spaces and in the urban environment. Unfortunately, of the standard SIFT-MS reagent ions only O^−• reacts with CO ^[28][30] and this solely via the associative detachment mechanism (Table 1). As shown in Equation (3), no useful ion is obtained since free electrons (e⁻) are not detectable in SIFT-MS instruments. O^−•₂⁻k^{−103−1−1} (3) Finally, nitrous oxide (N₂O)—another important greenhouse gas ^[25][27]—reacts with O⁻ as shown in Equation (4) ^[28][30]. In principle, this reaction rate coefficient provides adequate sensitivity for trace gas analysis. However, in air the NO⁻ product ion reacts with O₂ to form O₂^−• (Equation (5)) ^[28][30], which cannot be distinguished from O₂^−• formed at trace levels from other sources. N₂O is hence detectable, but it is only practicable to do so in oxygen-free matrices (e.g., in specialty gas applications). O^−•₂⁻k^{−103−1−1} (4) NO⁻₂₂^−•k^{−103−1−1} (5) Despite these limitations, SIFT-MS has found broad applications ^[6][23][24][6,9,10]. Later sections will review applications that are more closely aligned with chemosensor technologies.

2.3. Specificity of Analysis

Achieving specific analysis in moderately complex matrixes can be challenging for low-cost chemosensors. Eminent flavor chemist, Gary Reineccius wrote of the application of electronic noses (sensor arrays) in food flavor analysis that “one has no clear idea of what the instrument is responding to in making a judgement” ^[30][32]. In contrast, SIFT-MS achieves specific analysis in real-time by the combination of highly controlled, ultra-soft CI and MS, producing signals that are attributable to specific analytes. Table 1 illustrates the variety of common reaction mechanisms that occur for the standard SIFT-MS reagent ions and indicates the relative shifts from parent product ion m/z, or standard product ion m/z, for these mechanisms. Combining this behavior with the observation that different functional groups typically react in different ways ^[5][6][6,7], at least one unique ion for a given analyte is usually realized. It is important to reemphasize that when using the phrase “real-time specificity” in the context of SIFT-MS, a user is not limited to a single reagent ion in each analysis. By means of the first QMF (Figure 1), combined with software control, reagent ions generated in a given ion source setting (e.g., positive ions with H₃O⁺, NO⁺, and O₂^+•) are switchable in tens of milliseconds. This means that target compounds can be analyzed specifically across different reagent ions, if necessary. This contrasts with the other DIMS techniques [3]. Note that where an ion source setting change is required—e.g., from positive ions to either of the negative ion settings—this is still near real time at approximately 10 s. Figure 2 summarizes the elements that contribute to specific analysis in SIFT-MS. Addition of a time-of-flight (TOF) mass spectrometer to SIFT-MS instruments (as has been done successfully for PTR-MS [3]) could further improve specificity, but at additional cost. However, regardless of the way that mass filtering is accomplished in SIFT-MS or PTR-MS, it is not reasonable to expect that the specificity of DIMS techniques will approach that of GC/MS with its temporal separation of analytes. GC/MS techniques—especially those utilizing high-resolution mass spectrometry—will remain the technique of choice for compound identification and analysis of very complex matrices. Figure 1. A schematic illustration of the SIFT-MS technique, showing the three zones in this continuous-analysis approach. Used with permission of Syft Technologies. Since detailed examples of SIFT-MS specificity are given in Refs. ^{[31][32][33][34]}[33,34,35,36], one simple example suffices here. Consider two structural isomers of molecular formula C₂H₆O and molar mass 58.0791 g mol⁻¹: acetone (propan-2-one, (CH₃)₂CO) and propanal (propionaldehyde, CH₃CH₂CHO). Table 3 shows the primary product ions formed when acetone and propanal react with the three positively charged SIFT-MS reagent ions ^[10][14]. Use of the NO⁺ reagent ion provides the most convenient resolution of these isomers. The H₃O⁺ reagent ion cannot resolve the compounds, and matrix effects are encountered frequently for the unique O₂^+• product ions. Table 3 ignores the so-called secondary chemistry of acetone and propanal—only the primary ions formed because of reagent ion reaction with the analytes are shown. These primary product ions can react with matrix species—most commonly water, though carbon dioxide is common for negative ions—to form secondary product ions ^[5][6][6,7]. Secondary reactions are commonly association reactions (Table 1), requiring that a third molecule (a so-called “third body”) is present in the reactive collision because it carries away sufficient kinetic energy for adduct formation to occur. The association mechanism is hence sensitive to the temperature and pressure in the SIFT-MS instrument flow tube ^[5][6][6,7]. The secondary reactions pertinent to these analytes occur for the H₃O⁺ product (the most common situation): In these equations, M represents the necessary third body and the secondary product ions are located +18 units higher than the primary product ion (i.e., at m/z 77). When they occur, it is necessary to measure the signals of secondary product ions for correct quantitation (see the next subsection). Method development should also ensure that they are not interfering with other analytes nor interfered with by other analytes.

2.4. Quantitation in SIFT-MS

The preceding subsections have described how SIFT-MS can analyze a wide range of compounds and generally do so with high specificity. It remains, then, to outline how quantitative results are obtained in SIFT-MS. Since the details of the gas-phase kinetic theory have been described elsewhere ^[6][35][6,37], and the detailed calculation of concentration presented ^[36][38], only a brief sketch will be given here. A future article will describe the calculation, as currently adapted in commercial SIFT-MS instruments, in more detail. In this articlentry, the use of non-SI pressure units (torr) follows that of the SIFT-MS literature ^[36][38]. The master equation used in commercial instruments to determine the concentration of analyte A in air ([A]_i in units of parts-per-billion by volume, ppbV) for each primary product ion (i) of A utilized in the analytical method is: Here:

• T_FT is the temperature in the flow tube in Kelvin (K)
• P_FT is the pressure in the flow tube in torr
• k_j is the rate coefficient for reaction of reagent ion R_j⁺ with the analyte (in cm³ molecule⁻¹ s⁻¹)
• t_r is the reaction time (in seconds)
• φ_carr is the carrier gas flow in torr L s⁻¹
• φ_samp is the sample flow in torr L s⁻¹
• ) is the transmission factor for the reagent ion
• R
• _j
• ⁺
• (dimensionless)
• br_i is the branching ratio for primary product ion i (0 < i ≤ 1 for calculation purposes, but ordinarily tabulated as a percentage; see, e.g., Table 3).

The coefficient (1.035 × 10⁻¹⁰) in Equation (8) is derived from several physical constants and unit conversions and has units of Torr cm³ molecule⁻¹ K⁻¹. Note that the concentration [A]_i is directly proportional to the ratio of measured product ion signal to measured reagent ion signal, which overcomes any ion source drift, should that occur, and confers measurement stability compared to MS detection coupled with chromatographic techniques. Clearly for correct calculation of [A]_i the appropriate secondary ions need to be included in the analytical method (e.g., Equations (6) and (7)), otherwise the concentration will underreport. The parameters in Equation (8) are obtained from various sources:

• Instrument operating parameters: T_FT, P_FT, φ_carr, and φ_samp
• The instrument’s automated performance check on a certified gas standard: t_r, TF_Prod(i), TF_Prod(ki), and TF_Reag(j)
• Software library: k_j and br_i (clearly, together with the m/z at which the relevant reagent and product ions will be located)
• Measurement of sample: P_i⁺, P_ki⁺, and R_j⁺.
• P
• _i⁺ is the primary product ion signal (in counts per second, cps) for primary product ion i counted by the particle multiplier detector
• TF_Prod(i) is the transmission factor for the primary product ion P_i⁺ (dimensionless)
• P_ki⁺ is the secondary product ion signal (in cps) for secondary product ion k derived from primary product ion i
• TF_Prod(ki) is the transmission factor for the secondary product ion P_ki⁺ (dimensionless)
• R_j⁺ is the reagent ion signal (in cps) for the injected reagent ion (j = 0) and its water cluster ions (if appropriate; j = 1, 2, 3)
• TF_Reag(j

When more than one concentration determination is made for analyte A (i.e., several primary product ions are used in the method), then instrument software supports automated interference rejection through comparison of the [A]_i values (within a user-chosen tolerance; by default, 20%) ^[37][39]. The most common failure mode for this approach is when secondary chemistry is omitted from the analytical method. Several brief comments must be made to conclude the brief overview of SIFT-MS. First, SIFT-MS is quantifying compounds based on unique reagent ion-product ion pairs, not on pattern recognition from a sensor array—the analyst knows which compound is being detected using SIFT-MS, in contrast to sensors ^[30][32]. This confidence is evident in the extensive SIFT-MS literature reviewed elsewhere ^[5][23][24][7,9,10]. Second, several publications have compared SIFT-MS with gold-standard chromatography-based techniques and have reported favorable results, for example, Refs. ^[38][39][40][40,41,42]. Third, SIFT-MS analysis can be validated according to industry standard procedures (e.g., ^[41][42][43,44]), with only minor adaptations required despite the significant change in analytical approach from chromatographic separation to DIMS. Finally, the CI approach underpinning the SIFT-MS technique provides a wide linear range ^[6][43][6,45]. Ensuring that the instrument remains in the linear range is important for reliable concentration determination [6], although calibration and related approaches can circumvent this restriction when used with care ^[44][5]. Generally, however, SIFT-MS measurements do not drift with changes in matrix as chemosensors may. Hence SIFT-MS might provide an alternative to chromatographic methods for benchmarking chemosensors since it provides high-quality analysis in real-time.