Research Article

Received: 16 October 2015 Revised: 13 January 2016 Accepted: 19 January 2016 Published online in Wiley Online Library

Rapid Commun. Mass Spectrom. 2016, 30, 890–896

(wileyonlinelibrary.com) DOI: 10.1002/rcm.7516

Flame-induced atmospheric pressure chemical ionization mass

spectrometry
Sy-Chyi Cheng1, Yen-Ting Chen1, Siou-Sian Jhang1 and Jentaie Shiea1,2*
1
Department of Chemistry, National Sun Yat-Sen University, Kaohsiung, Taiwan
2
Department of Medicinal and Applied Chemistry, Kaohsiung Medical University, Kaohsiung, Taiwan

RATIONALE: Charged species such as formylium (CHO+), hydronium (H3O+), and water clusters [H3O+(H2O)n] are
commonly found in flames. These highly reactive species can react with analytes via ion-molecule reactions (IMRs) to
form analyte ions. A new mass spectrometric technique, named flame-induced atmospheric pressure chemical ionization
mass spectrometry (FAPCI-MS), was developed to characterize organic compounds via these mechanisms.
METHODS: A commercial corona-discharge atmospheric pressure chemical ionization (APCI) source was modified by
replacing the corona needle with a flame to make a FAPCI source. Liquid samples were vaporized in a heated tube
and delivered to the IMRs region by nitrogen to react with the charged species generated by a flame. Analytes on
surfaces were directly desorbed and ionized by a flame using the technique called desorption-FAPCI-MS (DFAPCI-MS).
RESULTS: Intact molecular ions of various chemical and biological compounds were successfully characterized by
FAPCI-MS. The FAPCI mass spectra are nearly identical to those obtained by traditional APCI-MS. The limit of detection
(LOD) of reserpine by FAPCI-MS was 50 μg L–1 with a linear calibration curve (R2 = 0.9947) from 100 μg L–1 to 10 mg L–1.
The LOD for ketamine by DFAPCI-MS was estimated to be less than 0.1 ng.
CONCLUSIONS: In FAPCI, analytes are not incinerated but vaporized and introduced into the ion source to react with
the reactive charged species generated by a flame. The features of the FAPCI source include: configuration is very simple,
operation is easy, high voltage or inert gas is unnecessary, and the source is maintenance free. Various combustible gases,
solvents and solids are useful flame fuels for FAPCI. Copyright © 2016 John Wiley & Sons, Ltd.

Atmospheric pressure chemical ionization (APCI) sources are [H3O+(H2O)n].[8] Charge and proton transfers then occur
commonly used to generate intact analyte ions for subsequent through ion-molecule reactions (IMRs) between primary ions
mass spectrometric (MS) detection. An APCI source using and analytes which own higher proton affinities than those of
63
Ni to ionize gaseous analytes was first introduced in the primary ions. Metastable molecules generated in plasma can
early 1970s.[1] A corona discharge based technique was then also react with analytes through Penning ionization processes
developed as an alternative to the 63Ni source to ionize chemical to form analyte ions.[4]
compounds eluted from liquid chromatography (LC).[2] The combustion of materials in flame plasma is an
Recently, plasma-based techniques such as atmospheric exothermic process which generates light, heat, and ions.
pressure glow discharge (APGD), direct analysis in real time The flame has been integrated with many spectroscopic and
(DART), plasma-assisted desorption ionization (PADI), detection techniques such as flame atomic absorption
dielectric barrier discharge ionization (DBDI), and low- spectroscopy (FAAS), flame emission spectroscopy (FES),
temperature plasma (LTP) have been developed to ionize flame ionization mass spectrometry (FIMS), and flame
analytes under ambient conditions.[3–7] A DC or AC high ionization detection (FID).[9–15] Unfortunately, organic
voltage is needed for either corona discharge based or plasma- compounds are thermally decomposed in the flame, so that
based APCI sources. intact molecular ions are not detected. Therefore, FAAS, FES,
The ionization mechanisms of APCI are known to be and FIMS are techniques that are only useful for detecting
complicated. Metastable atoms or molecules and primary elements in inorganic samples,[9–13] while FID measures the
charged species (protons, electrons, etc.) are involved in ion currents produced by organic analytes burned in a
reactions with gas-phase analytes. Electrons or radical ions hydrogen flame.[14,15]
generated near the corona-discharge needle or in the plasma Since charged species including formylium (CHO+),
initiate a series of chemical reactions to produce primary ions hydronium (H3O+), hydronium-water clusters, CH2O+, NO+,
such as hydronium (H3O+) and hydronium-water clusters K+, Na+, and NO2H+ have been detected in flames,[16–18] these
highly reactive charged species are potentially useful for
reacting with analytes via IMRs to form analyte ions. In this
* Correspondence to: J. Shiea, Department of Chemistry, National study, we have developed a mass spectrometric technique
Sun Yat-Sen University, 70 Lien-Hai Road, Kaohsiung called flame-induced atmospheric pressure chemical ionization
80424, Taiwan. (FAPCI) to ionize organic and biological compounds. This
890

E-mail: jetea@mail.nsysu.edu.tw new ionization method uses charged species generated in a

Rapid Commun. Mass Spectrom. 2016, 30, 890–896 Copyright © 2016 John Wiley & Sons, Ltd.
Flame-induced APCI-MS

flame to react with gas-phase analytes which are generated by measured by a gas flow meter (RMA-13-SSV, Dwyer) from
thermal desorption. Because analytes are not burned in the 0.1 to 0.3 L/min before being directed into the combustion
flame, intact molecular ions are observed. Unlike existing head. The temperature at the ionization region was measured
APCI techniques, no high voltage or inert gas is needed in by a thermocouple (HI 93530, Hanna Instruments, RI, USA)
FAPCI. The limits of detection (LODs), linearity, and placed at the center of the source.
reproducibility of FAPCI-MS, as well as direct desorption of Liquid sample was introduced into the FAPCI source
analytes from sample surfaces with a micro-flame, were through a heated nebulizer. The sample solution flowing
studied. out of a capillary (2.5 μL/min) was nebulized by a high-
velocity N2 jet (20–25 psi) ejected outside the capillary.
Sample droplets were vaporized in a heated transfer tube
EXPERIMENTAL (350 °C) and subsequently delivered into the FAPCI source
to react with the charged species generated in a flame. To
Standards, reagents, and materials detect the analyte ions formed in the FAPCI source, an
ion trap mass spectrometer (Esquire 3000 plus, Bruker
Lidocaine, reserpine, cholesterol, polyethylene glycol 1000 Daltonics, Billerica, MA, USA) was attached to the FAPCI
(PEG 1000), triethylamine, thiabendazole, and chlorhexidine source. The drying temperature and gas flow rate of the
was obtained from Sigma-Aldrich (St. Louis, MO, USA). ion trap mass spectrometer were set at 280 °C and
Ethylenediaminetetraacetic acid (EDTA) was purchased from 1 L/min, respectively.
Janssen Chimica (Belgium). 1,3,5-Trinitro-1,3,5-triazinan (RDX)
powder was obtained from Kaohsiung City Government Police Desorption flame atmospheric pressure chemical ionization
Bureau. Ketamine was obtained from Cerilliant (Round Rock, mass spectrometry (DFAPCI-MS)
TX, USA). LC-grade organic solvents such as methanol and
acetone were purchased from Merck (Darmstadt, Germany). For DFAPCI-MS analysis, a micro oxy-acetylene flame
Ethanol was purchased from Panreac (Barcelona, Spain). flowed from a stainless steel tube (O.D. 0.72 mm, I.D. 0.15
Over-the-counter drug tablets, windproof torch lighters, butane mm) was directed at sample surfaces with an incident angle
gas cartridges, and matchsticks were obtained from local between 30–60° for analyte desorption and ionization.
pharmacies and grocery stores. The flow rates of oxygen and acetylene were set at 10
and 15 mL/min, respectively. The temperature (ca. 350 °C)
of the flame located 0.6 cm from the exit of the stainless
Flame atmospheric pressure chemical ionization mass
steel tube was measured with a thermocouple. The
spectrometry (FAPCI-MS)
distances from the exit of the stainless steel tube to the
The FAPCI source was modified from a commercial corona- sample surface and from the inlet of the mass spectrometer
discharge APCI source (Bruker Daltonics) (Fig. 1). The corona to the sample surface were 8 mm and 4 mm, respectively.
needle in the commercial APCI source is located under a The samples for DFAPCI-MS analysis included: thiabendazole
heated sample introduction tube, where electric discharge of solution (200 μg/mL, 1 μL) deposited on a stainless steel plate;
the air molecules is induced by applying a current of 8000 RDX powder, ketamine solution (100 ng/mL, 1 μL), and
nA to the needle. In the FAPCI source, this needle is replaced lidocaine solution (100 μg/mL, 5 μL), each deposited on a glass
by a flame generated from a handheld windproof lighter rod; and drug tablets.
equipped with a combustion head. The flame is positioned In addition to gaseous fuels, a wooden matchstick and
2 cm in front of the orifice where the corona needle has been organic solvents such as methanol, ethanol, and acetone
disassembled. To study the influence of the flow rate of butane were used to generate flames through a lamp burner.
gas on analyte ion intensity, the gas was adjusted and The matchstick or lamp burner was placed below the

Figure 1. (a) Schematic illustration and (b) photograph of the FAPCI source.
891

Rapid Commun. Mass Spectrom. 2016, 30, 890–896 Copyright © 2016 John Wiley & Sons, Ltd. wileyonlinelibrary.com/journal/rcm
 S.-C. Cheng et al.

sample to generate a flame. The sample solution butane-fueled flame which was positioned 4 cm from the
deposited on a glass rod was moved close to the flame center of the ionization region (Fig. 2(a)).
for analysis. Analyte ions were detected either by the It was found that the intensity of reserpine ions
ion trap mass spectrometer or a quadrupole time-of-flight decreased as the butane flame was moved away from
(QTOF) mass analyzer (microTOF-Q II, Bruker Daltonics, the source. The flame was actually positioned outside the
Billerica, MA, USA). ion source so the temperature in the ion source would
not be too high (Fig. 1(b)). To prevent heat accumulation
in the source, a circle (ID 2 cm) at the bottom of the
RESULTS AND DISCUSSION source body was cut so the hot air could flow out of the
source. Under this condition, the temperature at the
The FAPCI source was constructed by modifying a ionization region (i.e., 0.5 cm under the heated tube) was
commercial corona-discharge APCI source. The corona kept at approximately 420 °C (measured by a thermocouple
needle in the APCI source was removed and replaced set in the source) during analysis. Since the temperature in
with a flame generated from a handheld windproof the ionization source was not increased with time, this
lighter equipped with a combustion head (Fig. 1(b)). prevented the analyte from thermal decomposition and
To successfully detect intact analyte ions, the analytes made long-term operation of the FAPCI source possible
must not burn in the flame or be exposed to high (e.g., >2 h). The experimental results also showed that
temperature. Therefore, the parameters affecting the more reactive primary ion species were generated in a
IMRs and thermal decomposition of the analytes in the butane flame than that in a hydrogen flame since the
FAPCI source have to be well controlled so intact analyte intensity of reserpine ions ([M+H]+, m/z 609) was 100
ions can be produced. The experimental parameters times lower with a hydrogen flame than a butane flame
include: (i) the distance from the flame to the ionization (data not shown).
region, (ii) the size of the flame, (iii) the temperature of the FAPCI-MS was used to characterize reserpine (1 μg/mL),
ionization region, and (iv) the composition of the fuel to ethylenediaminetetraacetic acid (EDTA, 1 μg/mL), and
generate flame. Typically, intact reserpine ions ([M+H]+) polyethylene glycol 1000 (PEG 1000, 1 μg/mL) standards
could be detected by FAPCI-MS using a 1.5-cm-long with a butane-air flame. By operating under the experimental

Figure 2. (a) Positive FAPCI mass spectrum of reserpine (m/z 609, [M+H]+); (b) negative
FAPCI mass spectrum of ethylenediaminetetraacetic acid (EDTA, m/z 291, [M–H]–); and
(c) positive FAPCI mass spectrum of PEG 1000; ion series at ■ (m/z 635+44n), ● (m/z 696+44n),
+
□ (m/z 657+44n), and ○ (m/z 805+44n) correspond to protonated PEG ions [M+H] ,
ammoniated PEG ions [M+NH4]+, sodium [M+Na]+, and potassium [M+K]+ adducts,
respectively. Standards prepared in MeOH (1 μg/mL each) were analyzed.
892

wileyonlinelibrary.com/journal/rcm Copyright © 2016 John Wiley & Sons, Ltd. Rapid Commun. Mass Spectrom. 2016, 30, 890–896
Flame-induced APCI-MS

parameters listed above, a protonated intact reserpine ion of conventional APCI, samples were analyzed with a
([M+H]+, m/z 609) was detected in the positive FAPCI mass commercial corona-discharge APCI mass spectrometer
spectrum and no fragment ions were found (Fig. 2(a)). A with the corona needle current set at ca. 8000 nA.
deprotonated EDTA ion ([M–H]–, m/z 291) was detected in Protonated triethylamine (m/z 102), ketamine (m/z 238),
the negative FAPCI mass spectrum (Fig. 2(b)). Figure 2(c) and chlorhexidine (m/z 505) were detected by FAPCI-MS
displays the positive FAPCI mass spectrum of the PEG 1000 (Figs. 3(a), 3(b), and 3(d)). On the other hand, the
standard, which shows signals corresponding to the dehydrated cholesterol ion [M–H2O+H]+ at m/z 369 was
protonated PEG ions ([M+H] +, 635.3+44n, ■), ammoniated predominant in the FAPCI mass spectrum, where an
PEG ions ([M+NH4]+, 696.4+44n, ●), sodium adducts additional ion signal at m/z 383 (tentatively assigned
([M+Na]+, 657.5+44n, □), and potassium adducts ([M+K]+, as 7-ketocholesterol)[19] was detected (Fig. 3(c)). Traditional
805.5+44n, ○), respectively. APCI mass spectra of the above sample solutions are shown
Standard solutions of triethylamine, ketamine, in Figs. 3(e)–3(h). As can be seen the mass spectra obtained
cholesterol, and chlorhexidine (1 μg/mL each) were by APCI-MS and FAPCI-MS were similar, showing
analyzed with both FAPCI-MS and traditional corona- predominant signals of intact analyte ions without
discharge APCI-MS. To compare FAPCI results with those fragments. Since 7-ketocholesterol was detected by both

Figure 3. Positive FAPCI (left) and APCI (right) mass spectra of (a, e)
trimethylamine, (b, f) ketamine, (c, g) cholesterol, and (d, h)
chlorhexidine. Standards prepared in MeOH (1 μg/mL each) were
893

analyzed.

Rapid Commun. Mass Spectrom. 2016, 30, 890–896 Copyright © 2016 John Wiley & Sons, Ltd. wileyonlinelibrary.com/journal/rcm
 S.-C. Cheng et al.

APCI-MS and FAPCI-MS, the compound was either an

impurity in the sample solutions or it was generated
by oxidation of cholesterol in the heated sample
introduction tube.
The relationship between the absolute ion intensity of
reserpine and the flow rate of butane gas was studied. It
was found that the abundance of background ions at a lower
mass range (m/z 150–450) increased with the butane gas flow
rate (0.1–0.3 L/min, data not shown). The flow rate of butane
gas was then maintained at 0.1 L/min throughout the study.
In addition, higher reproducibility was obtained using a low
butane gas flow. The relative standard deviation (RSD) of
the reserpine ion (n = 3) was 6.1% at 0.1 L/min, 9.4% at
0.2 L/min, and 18.2% at 0.3 L/min. The limit of detection
(LOD) for reserpine was found to be 50 ng/mL using an
FAPCI or commercial APCI source, where an ion trap mass
analyzer operated in full scan mode was used for ion
detection. The LOD of reserpine can be dramatically
improved using selective ion monitoring (SIM) or multiple
reaction monitoring (MRM) mode. In addition, the use of a
state-of-art mass analyzer will be helpful in increasing the
LOD. A linear calibration curve (y = 31.249x – 3993.2) from
100 ng/mL to 10 μg/mL with R2 = 0.9947 was obtained by
FAPCI-MS.
Based on the similarity of the mass spectra, FAPCI
mechanisms are suggested to be similar to those in
conventional corona discharge based APCI and plasma-
based APCI methods including APGD, LTP, or DBDI.
Previously, CHO+, H3O+, and H3O+(H2O)n were reported
to be the major reactive charged species observed in FID
using hydrocarbon fuel.[16,17] The mechanisms of the
formation of these charged species were suggested as in
Eqns. (1)–(3):[16,17]

CH þ O → CHO → CHOþ þ e– (1) Figure 4. (a) Photograph of an oxy-acetylene flame combined

with a mass analyzer to detect sample deposited on a stainless
CHOþ þ H2 O → H3 Oþ þ CO (2) steel plate. DFAPCI mass spectra of (b) thiabendazole on a
stainless steel plate, (c) RDX on a glass rod, and (d) a drug
H3 Oþ þ nH2 O → H3 Oþ ðH2 OÞn (3) tablet.

CHOþ þ M → CO þ MHþ (4)

and dextromethorphan (m/z 272) were detected from the
H3 Oþ ðH2 OÞn þ M → n þ 1 H2 O þ MHþ (5) surface of a drug tablet (Fig. 4(d)). Supplementary Fig. S1
(see Supporting Information) displays the positive DFAPCI
mass spectrum of a ketamine solution (100 ng/mL, 1 μL)
For positive ion formation, the proton in the reactive deposited on a glass rod. The LOD of DFAPCI-MS
species [i.e., CHO+, H3O+, and H3O+(H2O)n] is transferred determined by using ketamine solutions was estimated to
to the analytes through ion-molecule reactions (Eqns. (4) be less than 0.1 ng.
and (5)). In addition to hydrocarbon fuels such as butane and
Direct analysis of a sample on surfaces was achieved with a acetylene, combustible organic solvents and solids can also
technique named desorption-FAPCI-MS (DFAPCI-MS). The be used to generate the flame for DFAPCI-MS analysis.
analysis was done by directing a micro oxy-acetylene flame Various combustible gases, solvents and solids were
toward sample surfaces (Fig. 4(a)). The sampling and ionization demonstrated as useful flame fuels, where fuels with different
processes of DFAPCI were similar to those of desorption components generated characteristic spectra. A methanol
atmospheric pressure chemical ionization (DAPCI),[20] where flame was used for DFAPCI to characterize lidocaine
a heated gas containing reactive species was applied onto the deposited on a glass rod (see the inset in Fig. 5(a)). Protonated
sample to thermally desorb analytes and generate analyte ions lidocaine ions [M+H]+, sodium [M+Na]+, and potassium [M
through IMRs. Figure 4(b) displays the positive DFAPCI mass +K]+ adducts were detected by DFAPCI-MS using the flames
spectrum of thiabendazole, [M+H]+ at m/z 202. RDX, an generated from methanol, ethanol, and acetone, respectively
explosive, was detected as [M+NO3]– at m/z 284 by DFAPCI-MS (Figs. 5(a)–5(c)), while only protonated lidocaine ions and
(Fig. 4(c)). Protonated active ingredients including acetaminophen potassium adducts were detected by using the flame from a
(m/z 152), methylephedrine (m/z 180), caffeine (m/z 195),
894

matchstick (Fig. 5(d)).

wileyonlinelibrary.com/journal/rcm Copyright © 2016 John Wiley & Sons, Ltd. Rapid Commun. Mass Spectrom. 2016, 30, 890–896
Flame-induced APCI-MS

Figure 5. Positive DFAPCI mass spectra of lidocaine deposited on a glass rod.

Organic solvents and solid materials were used as fuels including (a)
methanol, (b) ethanol, (c) acetone, and (d) a matchstick, respectively.

CONCLUSIONS REFERENCES
Flame-induced atmospheric pressure chemical ionization mass [1] E. C. Horning, M. G. Horning, D. I. Carroll, I. Dzidic,
spectrometry (FAPCI-MS) was developed to ionize thermally R. N. Stillwell. New picogram detection system based on a
stable organic compounds for mass spectrometry. Flames of mass spectrometer with an external ionization source at
combustible gases and solvents such as butane, acetylene, atmospheric pressure. Anal. Chem. 1973, 45, 936.
methanol, ethanol, and acetone were successfully used to [2] D. I. Carroll, I. Dzidic, R. N. Stillwell, K. D. Haegele,
E. C. Horning. Atmospheric pressure ionization mass
generate primary ions to react with the analytes for ionization.
spectrometry. Corona discharge ion source for use in a
A flame can also be used to directly desorb and ionize the
liquid chromatograph-mass spectrometer-computer analytical
analytes on sample surfaces. The analytes are not incinerated system. Anal. Chem. 1975, 47, 2369.
in the flame but introduced into the ionization region, where [3] F. J. Andrade, J. T. Shelley, W. C. Wetzel, M. R. Webb,
they are ionized by reactive charged species. The combination G. Gamez, S. J. Ray, G. M. Hieftje. Atmospheric pressure
of a commercially available APCI source and a flame is easily chemical ionization source. 1. Ionization of compounds in
constructed and operated. In addition, the flame torch is the gas phase. Anal. Chem. 2008, 80, 2646.
maintenance-free, stable even for long-term use, and high [4] R. B. Cody, J. A. Laramée, H. D. Durst. Versatile new ion
voltages and inert gases commonly used in traditional APCI source for the analysis of materials in open air under
source are unnecessary. For its simple configuration and ambient conditions. Anal. Chem. 2005, 77, 2297.
operation, FAPCI may thus be suitable to be equipped on a [5] L. V. Ratcliffe, F. J. Rutten, D. A. Barrett, T. Whitmore,
D. Seymour, C. Greenwood, Y. Aranda-Gonzalvo,
portable mass spectrometer to perform in situ analysis.
S. Robinson, M. McCoustra. Surface analysis under ambient
conditions using plasma-assisted desorption/ionization
Acknowledgement mass spectrometry. Anal. Chem. 2007, 79, 6094.
[6] N. Na, M. Zhao, S. Zhang, C. Yang, X. Zhang.
This study was partially supported by grants provided by Development of a dielectric barrier discharge ion source
the Ministry of Science and Technology of Taiwan (NSC for ambient mass spectrometry. J. Am. Soc. Mass Spectrom.
895

101-2113-M-110-002-MY3). 2007, 18, 1859.

Rapid Commun. Mass Spectrom. 2016, 30, 890–896 Copyright © 2016 John Wiley & Sons, Ltd. wileyonlinelibrary.com/journal/rcm
 S.-C. Cheng et al.

[7] J. D. Harper, N. A. Charipar, C. C. Mulligan, X. Zhang, columns and flame ionization detector. Anal. Chem. 1964,
R. G. Cooks, Z. Ouyang. Low-temperature plasma probe for 36, 461.
ambient desorption ionization. Anal. Chem. 2008, 80, 9097. [16] A. J. C. Nicholson, D. L. Swingler. Ion formation in the flame
[8] I. Dzidic, D. I. Carroll, R. N. Stillwell, E. C. Horning. ionization detector. Combust. Flame 1980, 39, 43.
Comparison of positive ions formed in nickel-63 and corona [17] T. Holm. Aspects of the mechanism of the flame ionization
discharge ion sources using nitrogen, argon, isobutane, detector. J. Chromatogr. A 1999, 842, 221.
ammonia and nitric oxide as reagents in atmospheric [18] H. E. Taylor, J. R. Garbarino, S. R. Koirtyohann. Flame
pressure ionization mass spectrometry. Anal. Chem. 1976, ionization mass spectrometry: Isotope ratio determinations
48, 1763. for potassium. Appl. Spectrosc. 1991, 45, 886.
[9] A. Walsh. The application of atomic absorption spectra to [19] M. Careri, D. Ferretti, P. Manini, M. Musci. Evaluation of
chemical analysis. Spectrochim. Acta 1955, 7, 108. particle beam high-performance liquid chromatography–mass
[10] D. T. Coker, J. M. Ottaway, N. K. Pradhan. Metal atom spectrometry for analysis of cholesterol oxides. J. Chromatogr.
formation processes in flames. Nature Phys. Sci. 1971, 233, 69. A 1998, 794, 253.
[11] W. E. C. Wacker, C. Iida, K. Fuwa. Concentration of [20] F. P. M. Jjunju, S. Maher, A. Li, A. K. Badu-Tawiah, S. Taylor,
magnesium in human plasma or serum. Nature 1964, 202, 659. R. G. Cooks. Analysis of polycyclic aromatic hydrocarbons
[12] P. Wu, S. He, B. Luo, X. Hou. Flame furnace atomic using desorption atmospheric pressure chemical ionization
absorption spectrometry: a review. Appl. Spectrosc. Rev. coupled to a portable mass spectrometer. J. Am. Soc. Mass
2009, 44, 411. Spectrom. 2015, 26, 271.
[13] L. L. Yu, G. C. Turk, S. R. Koirtyohann. Characterization of a
hydrogen flame as an ion source for mass spectrometry.
J. Anal. At. Spectrom. 1999, 14, 669.
[14] I. G. Mcwilliam, R. A. Dewar. Flame ionization detector for SUPPORTING INFORMATION
gas chromatography. Nature 1958, 181, 760.
[15] H. Bruderreck, W. Schneider, I. Halasz. Quantitative gas Additional supporting information may be found in the
chromatographic analysis of hydrocarbons with capillary online version of this article at the publisher’s website.
896

wileyonlinelibrary.com/journal/rcm Copyright © 2016 John Wiley & Sons, Ltd. Rapid Commun. Mass Spectrom. 2016, 30, 890–896