Novel ethyl-derivatization approach for the determination of fluoride by headspace gas chromatography/mass spectrometry

Article abstract of DOI:10.1021/ac302303r

We report a novel derivatization chemistry for determination of fluoride based on the batch reaction of fluoride ions with triethyloxonium tetrachloroferrate(III) in a closed vessel to yield fluoroethane. Gaseous fluoroethane was readily separated from the matrix, sampled from the headspace, and determined by gas chromatography/mass spectrometry. The method was validated using rainwater certified reference material (IRMM CA408) and subsequently applied to the determination of fluoride in various matrixes, including tap water, seawater, and urine. An instrumental limit of detection of 3.2 μg/L with a linear range up to 50 mg/L was achieved. The proposed derivatization is a one-step reaction, requires no organic solvents, and is safe, as the derivatizing agent is nonvolatile. Determination of fluoride is affected by common fluoride-complexing agents, such as Al(III) and Fe(III). The effect of large amounts of these interferences was studied, and the adverse effect of these ions was eliminated by use of the method of standard additions.

Full text of DOI:10.1021/ac302303r

Article
pubs.acs.org/ac
Novel Ethyl-Derivatization Approach for the Determination of
Fluoride by Headspace Gas Chromatography/Mass Spectrometry
Enea Pagliano,*,†,‡ Juris Meija,† Jianfu Ding,† Ralph E. Sturgeon,† Alessandro D’Ulivo,§
and Zoltan Mester†
́
†National Research Council of Canada, 1200 Montreal Road, Ottawa, ON K1A 0R6, Canada
‡Scuola Normale Superiore, Piazza dei Cavalieri, 7, 56126, Pisa, Italy
§CNR, Consiglio Nazionale delle Ricerche, Istituto di Chimica dei Composti Organometallici, Via G. Moruzzi, 1, 56124 Pisa, Italy
ABSTRACT: We report a novel derivatization chemistry for
determination of fluoride based on the batch reaction of
fluoride ions with triethyloxonium tetrachloroferrate(III) in a
closed vessel to yield fluoroethane. Gaseous fluoroethane was
readily separated from the matrix, sampled from the
headspace, and determined by gas chromatography/mass
spectrometry. The method was validated using rainwater
certified reference material (IRMM CA408) and subsequently
applied to the determination of fluoride in various matrixes,
including tap water, seawater, and urine. An instrumental limit
of detection of 3.2 μg/L with a linear range up to 50 mg/L was achieved. The proposed derivatization is a one-step reaction,
requires no organic solvents, and is safe, as the derivatizing agent is nonvolatile. Determination of fluoride is affected by common
fluoride-complexing agents, such as Al(III) and Fe(III). The effect of large amounts of these interferences was studied, and the
adverse effect of these ions was eliminated by use of the method of standard additions.
large number of municipalities in North America
although poor sensitivity is reported (negative ion mode has
Asupplement their drinking water with fluoride as it been attempted yielding a detection limit of 110 μg/L16).
Molecular absorption from ions such as GaF has been exploited
for the determination of fluoride using high-resolution
continuum source atomic absorption spectrometry with
reported detection limits close to 1 μg/L.17,18 Similar figures
of merit are reported for electrospray MS.19,20
prevents tooth decay owing to its cariostatic activity. Ensuing
debates over the efficacy of fluoridation notwithstanding,
accurate monitoring of fluoride is of importance in countries
where water fluoridation is practiced.1−4 Because of the toxic
effects of fluoride at higher concentrations, a guideline value of
1.5 mg/L is recommended by the World Health Organization
as a level at which dental fluorosis should be minimal.5
Considering also that elevated levels of fluoride occur in natural
waters over an extensive geographical belt, reaching well above
10 mg/L in some regions of Kenya6 and India,7 the monitoring
of fluoride in water and food products is important from a
toxicological point-of-view.
Fluorine, like boron, does not enter into chemical reactions
that are selective enough to permit its direct determination in
the presence of concomitant elements.8,9 Consequently, no
direct or specific spectrophotometry or fluorescence methods
are available for its quantitation.10−13 A significant advance in
the determination of fluoride was development of the fluoride
ion-selective electrode in 1966.14 Owing to its simplicity, this
has effectively become the method of choice for field and
laboratory settings, providing a detection limit of ∼20 μg/L.15
However, ion-selective electrodes respond to the activity of
fluoride ions, not to their concentration, and are affected by the
presence of hydroxide ions.14
Derivatization of fluoride ions has traditionally been
performed using silicon-based chemistry in order to convert
F− to Me3SiF21,22 or Ph3SiF.23 The resulting volatile
fluorosilanes can be determined by gas chromatography
following either liquid extraction with organic solvents or
solid-phase microextraction from the headspace, achieving a
detection limit of 6 μg/L in the latter case with Me3SiF.21
However the Si−F bond in these derivatives is not stable in
aqueous solutions, especially at high pH. Consequently, Kage et
al.24 proposed carbon-based alkylation of fluoride with
pentafluorobenzylbromide (F5BzBr), yielding volatile F5BzF
which can be determined by gas chromatography/mass
spectrometry (GC/MS). Although stability of the resultant
derivative is achieved, the method is limited by a poor detection
limit (0.5 mg/L). Carbon-based derivatization of fluoride can
also be achieved using tertiary oxonium salts. Despite their
common use as alkylation agents in organic chemistry,25,26
oxonium salts have only recently been introduced to the field of
Modern instrumental methods can address the problem of
measurement specificity. Fluorine can be determined by
inductively coupled plasma mass spectrometry (ICPMS),
Received: August 10, 2012
Accepted: December 10, 2012
Published: December 10, 2012
© 2012 American Chemical Society
877
dx.doi.org/10.1021/ac302303r | Anal. Chem. 2013, 85, 877−881

Analytical Chemistry
Article
analytical chemistry.27−29 The commercially available triethy-
loxonium tetrafluoroborate permits otherwise difficult alkyla-
tion reactions to be undertaken in aqueous media; in addition,
Et3O+[BF4]− has been successfully applied for the derivatiza-
tion of inorganic anions,27 including chloride, bromide, iodide,
cyanide, thiocyanate, sulfide, nitrite, and nitrate.28,29 Generation
of volatile derivatives allows for the separation of the analyte
from the matrix, thereby minimizing spectral interferences and
reducing background levels in the separation and detection
phase of the analysis.30 However, Et3O+[BF4]− is not suitable
for determination of fluoride because tetrafluoroborate slowly
hydrolyzes, releasing fluoride ions. To overcome this drawback,
the method proposed here employs triethyloxonium
tetrachloroferrate(III) to convert fluoride to stable fluoroethane
which can be subjected to analysis by GC/MS.
during the reaction. During magnetic stirring, 0.1 mol (9.25 g)
of 2-(chloromethyl)oxirane was added dropwise using a
syringe. The addition of 2-(chloromethyl)oxirane was con-
trolled to keep the ether under constant boiling. After this
addition, the resultant mixture was further stirred for 5 min and
then cooled to 0 °C using an ice/water bath and held at this
temperature for 2 h with gentle magnetic stirring. The brown
crystalline product was filtered on a Buchner funnel and
̈
subsequently washed twice with dry ether. A 23.2 g mass of
Et3O+[FeCl4]− was obtained (77% yield), enough to perform
all the experiments described herein.
Although the triethyloxonium tetrachloroferrate(III) is
soluble in water, it slowly undergoes hydrolysis. Aqueous
solutions of this salt should therefore be prepared shortly
before use. In this work, the Et3O+[FeCl4]− solutions were
prepared by dissolving 0.4 g of Et3O+[FeCl4]− in 16 mL of
Milli-Q water.
EXPERIMENTAL SECTION
■
Reagents and Materials. Standard solutions of fluoride
were prepared by dissolution of sodium fluoride (Baker; w =
0.999 g/g) in deionized water generated with a Milli-Q system.
Anhydrous diethyl ether (Sigma-Aldrich), anhydrous iron(III)
chloride (Sigma-Aldrich, w ≥ 0.98 g/g), and 2-(chloromethyl)-
oxirane (Fluka, w = 0.99 g/g; also known as ( )-epichlorohy-
drin) were used for the synthesis of triethyloxonium
tetrachloroferrate(III). Triethyloxonium tetrafluoroborate
Sample Preparation. A 2 mL volume of aqueous sample
was introduced without pretreatment into five different vials,
four of which were spiked with 0.2 mL of different fluoride
standard solutions whereas 0.2 mL of Milli-Q water was added
to the remaining vial. Quantitation of fluoride was based on the
method of standard additions, and the concentration of the
standard was chosen so to obtain a maximum 4-fold increase in
the analytical signal compared to the nonspiked sample. The
samples were spiked with a 50 μL volume of ammonia solution
(w(NH3) = 0.2 g/g) and then with a 1 mL volume of the
Et3O+[FeCl4]− solution prepared as described above. After the
addition of all reagents, the vials were quickly sealed with a
screw-cap containing a PTFE/silicone septum and kept at room
temperature for 3 h before sampling the headspace with a
gastight syringe. Triethyloxonium converts F− into fluoro-
ethane which is gaseous at ambient temperature (tvap = −37
°C). Therefore, in order to avoid analyte loss, it is important
that unpunctured septa be used for sealing.
(Fluka,
w > 0.97 g/g) and triethyloxonium
hexachloroantimonate(VI) (Aldrich, w > 0.97 g/g) were also
sourced. An aqueous solution of ammonia (Tamapure AA-100;
w(NH3) = 0.2 g/g) was used for adjustment of the pH of the
sample. Rainwater, certified for fluoride content, was obtained
from the Institute for Reference Materials and Measurements
(CA408; Geel, Belgium). All other reagents were of analytical
grade. Samples of different origin were also selected for testing:
seawater (MOOS-2 certified reference material; NRC Canada),
City of Ottawa drinking water, and urine from a healthy
volunteer.
Safety Considerations. Despite widespread use of
alkylation and silylation reactions for gas chromatography, the
health risks of handling potentially toxic and volatile
compounds, such as diazomethane, chloroformates, or
trimethylsilyl chloride, must be carefully considered.31
Triethyloxonium tetrachloroferrate(III) is a strong ethylating
agent and must be handled in a fume-hood with appropriate
personal protective equipment. However, in comparison to
other reagents used for derivatization, its potential toxicity is
reduced because it is a nonvolatile crystalline salt which
undergoes complete hydrolysis within 3 h.32 Et3O+[FeCl4]− is
hygroscopic, but considerably less so than tetrafluoroborate,
and it should be stored in a freezer. In order to avoid excessive
exposure to ambient moisture, the bottle containing the reagent
should be allowed to reach room temperature before opening
and use of contents.
GC/MS Analysis. A Hewlett-Packard 6890 gas chromato-
graph equipped with a Hewlett-Packard 5973 mass selective
detector and fitted with a 60 m DB-624 column (6%
cyanopropyl-phenyl 94% dimethyl polysiloxane) was operated
at a constant flow of 0.9 mL He/min. Pulsed split (5:1)
injection mode was chosen, with a pulsed pressure of 172 kPa
for 1 min. The inlet temperature was set to 100 °C. After the
first isotherm at 30 °C (held for 7 min), the oven was heated at
20 °C/min to 200 °C. The transfer line temperature was 230
°C. The generated fluoroethane was manually sampled with a
gastight syringe from the vial headspace and a 250 μL volume
injected into the GC/MS. Fluoroethane eluted at 5 min under
the 30 °C isotherm. In addition, chromatographic performance
was improved with use of a narrow inlet liner (internal diameter
of 0.75 mm) designed for solid phase microextraction (SPME).
The mass spectrometer was operated in electron impact
mode (70 eV) using standard settings (ion source temperature
250 °C; quadrupole temperature 150 °C). The intensity of the
fluoroethane molecular ion shown in Figure 1 is low (10% of
the base peak); the most abundant fluorine-containing ions are
CH2CH2−F+ (m/z = 47 Da) and CH2−F+ (m/z = 33 Da).
Therefore, quantitation of fluoroethane was done by selected
ion monitoring (SIM) using the area of the peak extracted at
m/z = 47 Da. The fragment ion at m/z = 33 Da was monitored
for quality control purposes. Both were acquired using a 100 ms
dwell time.
Triethyloxonium Tetrachloroferrate(III). Triethyloxo-
nium tetrachloroferrate(III) was synthesized according to the
procedure of Meerwein et al:33
2Et2O·FeCl3 + C2H3(O)CH2Cl
→ Et3O+[FeCl4]− + FeCl2[(OCH(CH2Cl)CH2OEt]
For this purpose, 0.2 mol (32.4 g) of anhydrous FeCl3 was
dissolved in 150 mL of dry diethyl ether. The headspace above
the resultant solution was evacuated and purged with argon
three times using a Schlenk line while keeping argon flowing
878
dx.doi.org/10.1021/ac302303r | Anal. Chem. 2013, 85, 877−881

Analytical Chemistry
Article
triethyloxonium tetrachloroferrate is also able to ethylate other
inorganic anions.27
Figures of Merit and Applications. The determination of
fluoride was undertaken in various liquid matrixes using a
standard additions calibration approach whereby the fluoro-
ethane peak (m/z = 47 Da signal) was integrated and linearly
regressed against the mass of fluoride added. Headspace
sampling and injection into the GC/MS was manually
performed; no internal standards were used to correct for
variations that may occur during these operations. This method
was validated using IRMM CA408 rainwater certified reference
material, yielding good agreement with the certified property
value, as summarized in Table 1. Other matrixes were also
Table 1. Determination of Fluoride in IRMM CA408
a
Rainwater Certified Reference Material
b
run
γ (F−)/(mg/L)
1
0.186 0.034
0.196 0.020
0.195 0.017
0.199 0.011
0.194 0.012
6.5%
2
3
Figure 1. Electron impact mass spectrum of fluoroethane.
4
mean
ur
RESULTS AND DISCUSSION
■
a
b
Certified property value: γ(F−) = 0.194 0.008 mg/L. Uncertainty
Derivatization. The use of triethyloxonium salts as
derivatizing reagents is a recent introduction to the field of
analytical chemistry.27−29 Unlike triethyloxonium
tetrachloroferrate(III), which was selected for determination
of fluoride in this work, the two commercially available
oxonium salts, Et3O+[BF4]− and Et3O+[SbCl6]−, are unsuitable
analytical reagents for conversion of fluoride to fluoroethane,
of each standard additions experiment was evaluated according to
ISO/GUM guidelines. All uncertainties are reported here as expanded
uncertainties with coverage factor k = 1.
examined, including City of Ottawa drinking water, seawater
reference material (NRC MOOS-2), and urine from a healthy
volunteer. Results are based on the mean values of three
independent measurements derived from four point (in
addition to the unspiked sample) standard additions curves,
and the uncertainties were evaluated according to the ISO/
GUM guidelines. For each measurement, the relation between
the mass of fluoride added and the integrated signal was always
linear and characterized by a coefficient of determination R2 >
0.99. The mass concentration of fluoride determined in City of
Et3O+(aq) + F−(aq) → EtF(g) + Et2O
because Et3O+[BF4]− releases fluoride ions during hydrolysis of
the tetrafluoroborate,34 whereas Et3O+[SbCl6]− is not water-
soluble. Ethylation of fluoride with Et3O+[FeCl4]− was
performed in an alkaline medium (pH = 10) because an acidic
environment releases Fe(III) from the salt which then binds to
the fluoride ions. Additionally, the alkalinity of the sample
causes precipitation of a number of endogenous metal ions,
resulting in the removal of potential interferences in the
determination of fluoride. Moreover, alkaline conditions
prevent the conversion of fluoride to gaseous HF, a significant
advantage compared to an acidic silylation method, wherein
this may occur. The ethylation reaction stops after hydrolysis of
Et3O+ is complete. Granik et al.25 reported that complete
hydrolysis of Et3O+ occurs in 80 min. From our previous
research27 and from similar experiments repeated here for
Et3O+[FeCl4]−, we found that complete hydrolysis and
equilibration of the headspace is complete in 3 h. After this
time, the headspace can be sampled for analysis by GC/MS.
Fluoroethane is a stable compound and was detected in the
headspace up to 3 days after the derivatization. The chemical
yield for derivatization, achieved using the experimental
conditions described herein, was determined to be 8%. For
this purpose, four fluoride standard solutions were prepared
and subjected to the ethylation procedure. The headspace was
then purged with nitrogen to eliminate the fluoroethane, and
the remaining solutions were analyzed for the residual content
of fluoride.
Ottawa drinking water, γ(F−) = 0.72
0.06 mg/L, is in
agreement with the target level of 0.70 mg/L recommended by
Health Canada and adopted by the City of Ottawa.35
The results for the other two samples were as follows: γ(F−)
= 0.59 0.07 mg/L in urine and γ(F−) = 1.39 0.09 mg/L in
north Atlantic water (MOOS-2). The fluoride content in
MOOS-2 is also in good agreement with ion selective electrode
measurements of ocean water of similar salinity.36
All samples yield three phases after derivatization with
Et3O+[FeCl4]−: (1) a precipitate which contains Fe(OH)3 and
part of the matrix, (2) a homogeneous aqueous phase, and (3)
an upper gas phase (headspace) from which the fluoroethane is
sampled. Derivatization allows separation of the analyte from
the sample into the headspace and enables the realization of
extremely clean chromatograms. Figure 2 shows the results
from a urine sample; even with such a complex matrix, there are
no interferences coeluting with the analytical peak. The
instrumental limit of detection obtained by processing
calibration solutions is γ(F−)DL = 3.2 μg/L with a linear
range up to 50 mg/L. The estimation of γ(F−)DL is based on
the signal-to-noise ratio calculated from the standard deviation
of the baseline in proximity to the analyte peak (i.e., detection
limit is the concentration which produces a signal-to-noise-ratio
The purpose of this work was to demonstrate the
performance of Et3O+[FeCl4]− for derivatization of fluoride.
However, one must be mindful that, much like Et3O+[BF4]−,
879
dx.doi.org/10.1021/ac302303r | Anal. Chem. 2013, 85, 877−881

Analytical Chemistry
Article
approach compared to ion selective electrode methods, this
novel ethylation chemistry coupled with GC/MS provides
better selectivity and specificity. With respect to other GC/MS
approaches, this method offers a comparable or better detection
limit, and it is considerably safer to use than silylation or
alkylation with F5BzBr because Et3O+[FeCl4]− is a water-
soluble salt which undergoes complete hydrolysis in a few
hours. Additionally, it is not a volatile liquid such as Me3SiCl,
and the oxonium alkylation proceeds in aqueous media at room
temperature which adds to the speed and convenience of the
method. Further improvement to the analytical figures of merit
appears possible by employing an internal standard. Bromide
and iodide ions should serve as ideal internal standards since
they undergo the same derivatization chemistry as fluoride.
Figure 2. Analytical peaks for fluoroethane (at m/z = 47 Da) and the
resulting standard additions calibration plot derived from a sample of
incrementally spiked urine.
of 3). Standard solutions of γ(F−) = 123 μg/L and 531 μg/L
give signal-to-noise ratios of 115 and 492, respectively.
AUTHOR INFORMATION
■
Corresponding Author
*E-mail: e.pagliano@sns.it, enea.pagliano@nrc.ca.
Notes
Interferences. Fluoride forms strong complexes with many
metal cations, such as Al(III), and Fe(III). These interferences
are common to any method of determination of fluoride.
Therefore, the method of standard additions should be used to
overcome such matrix effects.36 Interferences arising from the
complexation of fluoride reduce the analytical signal and
deteriorate the performance of analytical methods. Therefore,
the effect of the presence of large amounts of various metal
cations on the analytical response was evaluated. Table 2
The authors declare no competing financial interest.
REFERENCES
■
(1) Verkerk, R. H. J. Toxicology 2010, 278, 27−38.
(2) Harrison, P. T. C. J. Fluorine Chem. 2005, 126, 1448−1456.
(3) Levy, S. M. J. Public Health Dent. 1999, 59, 211−223.
(4) Ekstrand, J. J. Nutr. 1989, 119, 1856−1860.
(5) WHO. Guidelines for Drinking-Water Quality. Vol. 2. Health
Criteria and Other Supporting Information.; World Health Organization:
Geneva, Switzerland, 1984.
Table 2. Interference Effect Arising from the Presence of 10
mM Metal Ion in a Standard Solution of 0.05 mM Fluoride
(6) Wambu, E. W.; Muthakia, G. K. Fluoride 2011, 44, 37−41.
(7) Meenakshi; Maheshwari, R. C. J. Hazard. Mater. 2006, 137, 456−
463.
(8) Fuwa, K. Stud. Environ. Sci. 1985, 27, 3−14.
(9) Lundell, G. E. F.; Hoffman, J. I. Outlines of Methods of Chemical
Analysis; Wiley: New York, 1948.
a
interference
S/S0
none
1.00 0.07
0.11 0.09
0.62 0.07
0.83 0.03
0.32 0.02
0.97 0.06
1.01 0.06
Al(III)
Fe(III)
Ca(II)
Mg(II)
Zn(II)
Cu(II)
(10) Williams, W. J. In Handbook of Anion Determination;
Butterworth: London, 1979; pp 335−373.
(11) Zhu, C.-Q.; Chen, J.-L.; Zheng, H.; Wu, Y.-Q.; Xu, J.-G. Anal.
Chim. Acta 2005, 539, 311−316.
a
S/S0 is the ratio between the analytical signal from fluoride in the
presence of interference (S) to that generated without interference
(S0). Three independent measurements were performed.
́
(12) Bayon, M. M.; Rodríguez Garcia, A.; García Alonso, J. I.; Sanz-
Medel, A. Analyst 1999, 124, 27−31.
́
(13) Garrido, M.; Lista, A. G.; Palomeque, M.; Fernandez Band, B. S.
Talanta 2002, 58, 849−853.
(14) Frant, M. S.; Ross, J. W., Jr. Science 1966, 154, 1553−1555.
(15) Campbell, A. D. Pure Appl. Chem. 1987, 59, 695−702.
(16) Fulford, J. E.; Quan, E. S. K. Appl. Spectrosc. 1988, 42, 425−428.
(17) Gleisner, H.; Welz, B.; Einax, J. W. Spectrochim. Acta, Part B
2010, 65, 864−869.
(18) Heitmann, U.; Becker-Ross, H.; Florek, S.; Huang, M. D.;
Okruss, M. J. Anal. At. Spectrom. 2006, 21, 1314−1320.
(19) Barnett, D. A.; Horlick, G. J. Anal. At. Spectrom. 1997, 12, 497−
501.
summarizes the interference arising from the presence of 10
mM (240−650 mg/L) levels of several cations added to
standard solutions containing 0.05 mM fluoride. The 200-fold
excess of these potential interferences over the analyte was
chosen in order to show the robustness of the proposed
method in coping with the interference of metal cations. Signal
suppression is observed in the presence of Mg(II), Ca(II),
Fe(III), and Al(III). The latter exhibits the most severe
suppression and contributes to a 10-fold decrease in the
analytical signal. Negligible effects were evident with Zn(II) and
Cu(II). The alkaline condition induced by pretreatment of the
sample with ammonia results in precipitation of many metal
cations as hydroxides, thus significantly reducing their ability to
complex fluoride.
(20) Hotta, H.; Kurihara, S.; Johno, K.; Kitazume, M.; Sato, K.;
Tsunoda, K.-I. Anal. Sci. 2011, 27, 953−956.
(21) Wejnerowska, G.; Karczmarek, A.; Gaca, J. J. Chromatogr., A
2007, 1150, 173−177.
(22) Tashkov, W.; Benchev, I.; Rizov, N.; Kolarska, A.
Chromatographia 1990, 29, 544−546.
(23) Musijowski, J.; Szostek, B.; Koc, M.; Trojanowicz, M. J. Sep. Sci.
2010, 33, 2636−2644.
(24) Kage, S.; Kudo, K.; Nishida, N.; Ikeda, H.; Yoshioka, N.; Ikeda,
N. Forensic Toxicol. 2008, 26, 23−26.
CONCLUSIONS
■
Triethyloxonium tetrachloroferrate(III) was successfully em-
ployed for derivatization of fluoride ion in aqueous solutions to
produce stable volatile fluoroethane, which is amenable to
analysis by GC/MS. Although it is a lower throughput
(25) Granik, V. G.; Pyatin, B. M.; Glushkov, R. G. Russ. Chem. Rev.
1971, 40, 747−759.
(26) Olah, G. A.; Laali, K. K.; Wang, Q.; Surya Prakash, G. K. Onium
Ions; Wiley: New York, 1998.
880
dx.doi.org/10.1021/ac302303r | Anal. Chem. 2013, 85, 877−881

Analytical Chemistry
Article
(27) D’Ulivo, A.; Pagliano, E.; Onor, M.; Pitzalis, E.; Zamboni, R.
Anal. Chem. 2009, 81, 6399−6406.
(28) Pagliano, E.; Onor, M.; Pitzalis, E.; Mester, Z.; Sturgeon, R. E.;
D’Ulivo, A. Talanta 2011, 85, 2511−2516.
(29) Pagliano, E.; Meija, J.; Sturgeon, R. E.; Mester, Z.; D’Ulivo, A.
Anal. Chem. 2012, 84, 2592−2596.
(30) Sturgeon, R. E.; Guo, X.; Mester, Z. Anal. Bioanal. Chem. 2005,
382, 881−883.
(31) Wells, R. J. J. Chromatogr., A 1999, 843, 1−18.
(32) Raber, D. J.; Gariano, P., Jr.; Brod, A. O.; Gariano, A.; Guida, W.
C.; Guida, A. R.; Herbst, M. D. J. Org. Chem. 1979, 44, 1149−1154.
(33) Meerwein, H.; Battenberg, E.; Gold, H.; Pfeil, E.; Willfang, G. J.
Prakt. Chem. 1939, 154, 83−156.
(34) Morales, R.; West, P. W. Anal. Chim. Acta 1966, 35, 526−529.
(35) Health Canada. Guidelines for Canadian Drinking Water Quality:
Guideline Technical Document - Fluoride; Health Canada: Ottawa,
Ontario, 2010.
(36) Rix, C. J.; Bond, A. M.; Smith, J. D. Anal. Chem. 1976, 48,
1236−1239.
881
dx.doi.org/10.1021/ac302303r | Anal. Chem. 2013, 85, 877−881