Chemical ionization mass spectrometry utilizing benzene cations for measurements of volatile organic compounds and nitric oxide
Abstract. We evaluate the capability of chemical ionization mass spectrometry (CIMS) using benzene cations as reagent ions (benzene CIMS) for detecting atmospheric trace gases. We characterize the ionization pathways and product ion distributions for 27 analytes spanning diverse chemical classes. To interpret the complex ion chemistry involving two reagent ions (C6H6+ and (C6H6)2+) and multiple ionization pathways (charge transfer, proton transfer, adduct formation, and hydride abstraction), we introduce a thermodynamics-based framework that classifies analytes into three categories based on their ionization energy (IE), relative to those of benzene monomer (9.24 eV) and dimer (8.69 eV). Each class exhibits distinct ionization mechanisms and product ions. Analytes with IE smaller than 8.69 eV (low IE class) undergo charge transfer with both reagent ions; analytes with IE between 8.69 and 9.24 eV (mid IE class) undergo charge transfer with C6H6+ and potential adduct formation with (C6H6)2+; analytes with IE larger than 9.24 eV (high IE class) could undergo adduct formation, proton transfer, or hydride abstraction. Analytes within each class also show similar sensitivity, enabling sensitivity estimation for compounds lacking calibration standards. In addition to volatile organic compounds (VOCs), benzene CIMS detects nitric oxide (NO) with a 1-minute detection limit of 5 pptv, exceeding the performance of most commercial NOx analyzers. Field deployments in Chicago and St. Louis demonstrate good agreement with reference NO measurements. Isoprene measurements show good agreement with a co-located gas chromatography–photoionization detector (GC-PID) in St. Louis, but exhibit substantial positive bias in Chicago, likely due to interferences from anthropogenic VOCs in the polluted urban environment. These results highlight the potential of benzene CIMS for concurrent measurements of NO, VOCs, and their oxidation products using a single instrument, while also underscoring challenges in complex atmospheric conditions.
General Comments
Puttu and coauthors present a very comprehensive characterization of chemical ionization mass spectrometry using benzene cluster cation chemistry, highlighting renewed potential utility of this method. Their work not only corroborates and extends prior benzene cluster cation CIMS characterization studies for VOCs but also identifies previously uncharacterized species detectable by benzene CIMS (e.g., NO), which could be highly valuable for future measurements. The methods in the manuscript are detailed and thorough and the presentation is clear and well-structured, with only minor methodological clarifications needed. After addressing these points, I believe this study is well-suited for publication in AMT.
Specific Comments
Line 97 – 98: Although larger benzene clusters are not detected because they undergo declustering before reaching the detector, this does not imply that they are absent in the IMR or that they do not influence reagent ion chemistry. While I agree that the key factors determining product ions are limited to ion affinities and IE with respect to the benzene ion monomer or dimer, it should at least be noted that larger clusters may affect reagent ion chemistry unless authors believe that even in the IMR they make up a very small fraction of the reagent ion distribution.
Line 144 – 146: This experiment introduces relatively high analyte concentrations. For example, the vapor pressure of alpha-pinene is about 3 torr at room temperature, so dilution gives approximately (3 torr/760 torr) x (0.01 SLPM / 3.01 SLPM) ≈ 13 ppm of analyte. Could this affect ion chemistry via titration? Additionally, line 361 notes that higher analyte mixing ratios lead to increased water ion signals, suggesting some effect. Please clarify why the observed product ion distribution is representative of lower analyte concentrations.
Line 185 – 187: The observed effect of O2 from zero air on the isoprene product ion distribution is interesting. Does a similar distribution occur for other species in the “Mid IE” category? Clarifying whether this behavior is class-specific or species-specific would be helpful.
Line 354 – 355: Lavi et al. (2018) observed that for isoprene, the isoprene-benzene cluster signal decreases with increasing humidity, while the charge transfer signal increases, resulting in near conservation of total signal. Is similar behavior observed in these experiments for “mid-IE” species like isoprene or others? If so, how does this reconcile with Figure S10a, where the benzene dimer concentration decreases with increasing water, potentially shifting product ions toward adducts? Since no mechanism is proposed for the sensitivity changes with water (Figure 4), discussing this could help illuminate the ion chemistry.
Line 421: Could potential interferences in the CIMS isoprene signal be verified using the GC measurements? For example, can 1,3-pentadiene be confirmed by GC, and can its ionization energy or ion affinities be compared to assess detectability with benzene CIMS?
Figure S8 and SI Lines 138-140: Why does the analyte signal in Fig. S8a rise to a maximum and then decrease? The authors suggest that increased flow raises reagent ion concentration but reduces ion-molecule reaction time, which then reduces product ions but I am a bit confused by this explanation. Since the IMR pressure is held constant, is the residence time in the IMR is primarily not set by its volume and the entrance and exit orifice, so it should not change with reagent ion concentration? If so, could the observed decrease in signal instead reflect shifts in reagent ion chemistry, such as changes in benzene cluster distribution or secondary reactions? Additionally, what does the analyte signal look like beyond the normalized maximum in Fig. S8b? Is there a similar decline that might indicate changes in residence time or reagent ion chemistry? Can you please clarify this in the text and if I am misunderstanding?
Technical Corrections
Line 42: Consider revising to “The hydronium ion…” or “Hydronium ions are…”