Technical Note pubs.acs.org/ac
Gas Chromatography Interfaced with Atmospheric Pressure Ionization-Quadrupole Time-of-Flight-Mass Spectrometry by LowTemperature Plasma Ionization Asger W. Nørgaard,*,† Vivi Kofoed-Sørensen,† Bo Svensmark,‡ Peder Wolkoff,† and Per Axel Clausen† †
The National Research Centre for the Working Environment, DK-2100 Copenhagen, Denmark Department of Basic Sciences and Environment, University of Copenhagen, DK-1870, Frederiksberg, Denmark
‡
S Supporting Information *
ABSTRACT: A low temperature plasma (LTP) ionization interface between a gas chromatograph (GC) and an atmospheric pressure inlet mass spectrometer, was constructed. This enabled time-of-flight mass spectrometric detection of GC-eluting compounds. The performance of the setup was evaluated by injection of mixtures of common volatile organic compounds. Amounts down to ca. 0.5 ng (on column) could be detected for most compounds and with a chromatographic performance comparable to that of GC/EIMS. In the positive mode, LTP ionization resulted in a compound specific formation of molecular ions M+•, protonated molecules [M + H]+, and adduct ions such as [(M + O) + H]+ and [M + NO]+. The ion patterns seemed unique for each of the analyzed compound classes and can therefore be useful for identification of functional groups. A total of 20 different compounds within 8 functional groups were analyzed.
A
radical cations, M+•, and various adduct ions may also be produced.1,6 Negative LTP produces O2−, and clusters containing water and oxygen are probably the dominating ions under ambient conditions leading to the formation of [M − H]−, M−•, and various adduct ions.5,6 The desorption mechanism is governed by thermal processes,3 thus LTP works best for volatile and semivolatile compounds (i.e., heating of the sample may improve the sensitivity for compounds of low volatility). Therefore, compounds eluted from a GC should be ideal for LTP ionization. We have tested a GC/LTP-MS setup on a range of compound classes that represent common indoor volatile organic compounds (VOCs), including alkanes, alkenes, alcohols, aromatic compounds, aldehydes, PAHs, phenols, and terpene alcohols.
ccurate mass spectral data are required for reliable sum formulas. This is an important supplement to the structural information obtained by electron ionization (EI), especially for compounds that are unaccessible in EI library databases or as standards. In general, dedicated and expensive instruments (usually sector instruments) are necessary for accurate mass determination of eluting compounds from a gas chromatograph (GC). The introduction of atmospheric pressure inlet (API) mass spectrometers (MS) has facilitated the setup of different methods for ionization. Ions simply need to be generated near the MS inlet in order to be guided by an electric field into the MS for mass separation and detection. This flexibility has prompted us to develop and explore a GC interface to an API quadrupole time-of-flight MS by use of lowtemperature plasma (LTP) ionization. LTP is a recently developed ionization technique where cold plasma (∼30 °C) is utilized for ionization followed by mass spectrometric analysis of liquid and solid samples and compounds in the gas.1 LTP is formed by a series of microdischarges that induce a cascade of reactions producing a nonequilibrium LTP, rich in metastable neutrals and ionized species.2,3 When He is used as a discharge gas, He2+ is the dominant positive ion in the discharge region. As the plasma enters the open atmosphere (after-glow region), N2+•, O2+•, and subsequently water clusters, [(H2O)n + H]+, are formed by charge-transfer reactions.3 Thus, LTP ionization probably resembles DART, where similar ions are formed.2,4,5 In the positive mode, LTP ionization under ambient conditions generally yields protonated molecules, [M + H]+. However, © 2012 American Chemical Society
■
EXPERIMENTAL SECTION A LTP probe, based on the experimental setup by Ouyang and collaborators,1 was mounted in a modified APCI source and carefully aligned with a GC column in front of the inlet orifice (see Figure 1). A heated transfer line kept the column temperature at 225 °C from the GC to the ion source. Sample volumes (1 μL) were introduced via splitless injection onto a 0.25 mm × 30 m Agilent VF-5 column. A Bruker microTOF-Q Received: July 4, 2012 Accepted: December 3, 2012 Published: December 3, 2012 28
dx.doi.org/10.1021/ac301859r | Anal. Chem. 2013, 85, 28−32
Analytical Chemistry
Technical Note
separated with GC peak widths of 7 s for the first eluting compounds and up to 14 s for the last eluting compounds (αterpineol and geraniol). The responses of toluene and decane are significantly lower due to their low proton affinities (see below). For comparison, the same solution was analyzed by GC/EIMS resulting in the TIC shown in Figure 2b. Here, all chromatographic peaks have similar widths (6−8 s), and the responses of the eluting species are more uniform as expected for EI. The calibration curves of the compounds leveled off; however, they could be approximated by two linear ranges below and above 10 ng on-column, as shown for α-pinene and α-terpineol in Figure 3 (entire range ca. 0.5−170 ng on column). The curvature may be caused by changes in the ionization at an increasing analyte concentration. For instance, α-pinene forms dimers [2M + H]+ that appear in the spectra at concentrations above 1 ng, while the relative abundance of the major fragment ions decreases significantly with an increase of the concentration (see the Supporting Information). LTP ionization under ambient pressure occurs by a complex series of competing reactions that depend on flow of discharge gas, flow and composition of the dry gas, distance between the LTP and MS inlet orifice, and the properties of the analyte. A continuous flow of hot, dry (250 °C) N2, that contains a maximum of 2% O2, ensures stable conditions in the GC/LTP ion source for the eluting species (see Figure 1). In the positive mode, protonated molecules [M + H]+ were the most commonly formed ions from the analyzed compounds (see Table 1). Furthermore, M+• ions were observed in several cases in addition to various adduct ions {e.g., [(M + O) + H]+, [(M + O2) + H]+, [M + NO]+, and [M + NO2]+}. [M − H]+ ions were also observed; they may be formed by hydride abstraction or more likely, by the loss of water from [(M + O) + H]+ ions. Fragment ions were observed to some extent, especially for terpenes, terpenoids, and compounds with aliphatic chains. The extent of fragmentation depended on the analyte concentration in the ion source (i.e., a high concentration showed less fragmentation). Overall, LTP ionization carried out in the water- and oxygen-depleted conditions used in this study are considered to be relatively hard. However, molecular ions or protonated molecules were observed for the analyzed compounds with the exception of hydrocarbons. Ionization in the negative mode was, however, mostly unsuccessful, probably due to the low oxygen conditions and thus low concentration of ionizing species in the negative mode (e.g., O2−).5 The combination of several ionizing species enables LTP ionization of a wide range of compound classes. Compounds with high proton affinity (e.g., terpenes and aldehydes) may easily form [M + H]+, while compounds with low proton affinity such as alkanes, that do not form protonated molecules, can interact with O2+• to form a range of different oxygenated product ions. Examples of how different compound classes ionize in LTP are summarized in Table 1. Most of the eight compound classes showed characteristic and distinguishable patterns of ions. For instance, terpenes (see spectrum of αpinene in Figure 3b) form abundant [M + H]+, [M − H]+ ions, and [(M + O) + H]+, [(M + O2) + H]+, [M + NO]+, and [M + NO2]+ adducts, while terpene alcohols, in addition to [M + H]+, [M − H]+, [(M + O) + H]+, and [(M + O2) + H]+, showed an abundant [(M − O) + H]+ ion (see spectrum of αterpineol in Figure 3d). The spectra of decane and dodecane, that showed various oxidation products {e.g., [(M + O) − H]+,
Figure 1. Schematic of the experimental setup.
was operated in the positive mode, and spectra were recorded at a rate of 1 Hz. A mass error of ±5 mDa was obtained for abundant ions by lock mass calibration using phthalic anhydride (m/z 149.0223) as the calibrant. GC/EIMS experiments were carried out using a Perkin-Elmer Autosystem XL (GC)/ Turbomass (MS) system equipped with a 0.25 mm × 30 m Agilent VF-5 column. Further experimental details may be found in the Supporting Information.
■
RESULTS AND DISCUSSION The performance of the GC/LTP-MS setup was tested by the analysis of CH2Cl2 solutions of alkanes, aromatic compounds, PAHs, terpenes, and terpenols in different concentrations. Figure 2a shows the total ion chromatogram (TIC) of a solution containing nine common indoor air VOCs in concentrations of about 30 ng/μL. All VOCs are baseline
Figure 2. Total ion chromatograms of a mixture of 9 VOCs in dichloromethane; ca. 30 ng of each compound was injected. (a) GC/ LTP-MS and (b) GC/EI-MS. Unassigned peaks are mainly stabilizing agents from the solvent. 29
dx.doi.org/10.1021/ac301859r | Anal. Chem. 2013, 85, 28−32
Analytical Chemistry
Technical Note
Figure 3. Standard curves and mass spectra of α-pinene (a and b) and α-terpineol (c and d). Note that there are two linear ranges for both compounds: above and below 10 ng. The curves are based on the extracted ion chromatograms of m/z 137 and 153 for α-pinene and α-terpineol, respectively.
[(M + O) − 3H]+, and [CnH2n−1]+ fragment ions}, were somewhat similar to previously reported DART spectra of alkanes.4 The low abundance observed for the alkanes is probably caused by a low O2 level in the ion source and thus low concentration of the O2+• ions, which are necessary for their ionization. Mass spectra of the 20 analyzed VOCs are presented in the Supporting Information. The compound specific patterns can be useful additional information for the identification of functional groups present in the analytes. A more detailed work based on the atmospheric sampling Townsend discharge ionization (ASTDI), carried out on a triple quadrupole MS, has been published previously by Nøjgaard et al.7 Here, 65 different VOCs within 12 functional groups were evaluated, and identification of functional groups was based on the observed ion patterns. On the basis of the results in the present study, it is plausible that a similar compilation of ion patterns can be done for LTP. Since the LTP ionization is sensitive to a range of factors, the performance of the setup highly depends on the ability to reproduce all experimental parameters (concentration of analyte, discharge gas flow, dry gas composition and flow, temperature, etc.). Therefore, addition of humid or dry auxiliary air to the ion-source region, will probably have a high impact in the observed ion patterns (as described previously for DART4
and ASTDI7). Thus, the ionization conditions (e.g., proton or charge transfer) can be optimized for the relevant analytes. There exists examples of API mass spectrometers coupled with GC by the use of ionization techniques such as Atmospheric Pressure Chemical Ionization9 and Atmospheric Pressure Photo Ionization.10 Table 2 gives an overview of the different techniques used to couple GC with MS. It is clear that the best performance, regarding sensitivity and linearity, is still obtained by the “traditional” vacuum-based techniques. However, the flexibility of the API inlet enables the use of any type of mass analyzer with a GC system. In addition, appropriate ionization methods can be applied, depending on the application at hand.
■
CONCLUSION The GC/LTP-MS interface can be used with any API-MS and is a relatively simple and inexpensive alternative to other interfaces based on, for example, APCI or APPI. The LTP ionization supplements EI data by information about molecular weight and may provide qualitative information of functional groups and molecular structure from ion patterns. In addition, LTP ionization yields accurate mass data when used with highresolution instruments (Orbitrap or QTOF) and can thus be used for identification of unknown compounds. Future 30
dx.doi.org/10.1021/ac301859r | Anal. Chem. 2013, 85, 28−32
Analytical Chemistry
Technical Note
Table 1. Relative Abundances of Selected Ions for 20 Different VOCs (ca. 30 ng On-Column) in Positive Modea selected ions [M − 3H]+ aldehydes: nonanal (142)b,c decanal (156)b,c alkanes: decane (142)c dodecane (170)c aliphatic alcohol octanol (130)b,c aromatic compounds: toluene (92)b m-xylene (106)b styrene (104)b PAHs: naphthalene (128) acenaphthene (155)b acenaphthylene (153) fluorene (166) anthracene (178) phenols: phenol o-cresol terpenes: α-pinene (136)b limonene (136)b terpene alcohols: linalool (152)b α-terpineol (152)b geraniol (152)b
[M − H]+
M+•
[M + H]+
5 5
[(M − O) + H]+
[(M + O) − 3H]+/ [(M + O) − H]+
[(M + O) + H]+/ [(M + O2) + H]+
60 65
[M + NO]+/ [M + NO2]+
[M + H2O2]+
40/0 50/0
5 5
100/40 100/75 50
30
85
100 15 40
30/20 30 60 95
100 100
80 35 30 50 45
100 100 100 100 100
10/0 5/0 10/0
15 15
100 100
20/5 5/0
35 65
100 50
35/0 100/10
35 35 30
75 100 45
15
10
45/10 20/5 75/0
10/10
15/5
100 50 100
15/5 30/15
10/0 10/20 10/0
a Assignments are generally based on mass measurements within ±5 mDa of the theoretical values. Abundances below 5% are not reported. Nominal compound masses are shown in brackets. bObserved [M − H]+ ions can be formed by hydride abstraction or loss of H2O from the observed [(M + O) + H]+ ions. cUndergo extensive fragmentation.
■
Table 2. Comparison of Ionization Techniques Used for the Coupling of GC with MS8−10
sensitivity linearity pressure fragmentation adduct formation Mw information polarity library search commercially available
APCI/ APLIb
PIa
LTPb
EIa
CIa
pg 4−5
