Determination of halide impurities in ionic liquids by total reflection X-ray fluorescence spectrometry

Article abstract of DOI:10.1021/ac5000812

The determination and quantification of halide impurities in ionic liquids is highly important because halide ions can significantly influence the chemical and physical properties of ionic liquids. The use of impure ionic liquids in fundamental studies on solvent extraction or catalytic reactions can lead to incorrect experimental data. The detection of halide ions in solution by total reflection X-ray fluorescence (TXRF) has been problematic because volatile hydrogen halide (HX) compounds are formed when the sample is mixed with the acidic metal standard solution. The loss of HX during the drying step of the sample preparation procedure gives imprecise and inaccurate results. A new method based on an alkaline copper standard Cu(NH₃) ₄(NO₃)₂ is presented for the determination of chloride, bromide, and iodide impurities in ionic liquids. The 1-butyl-3-methylimidazolium ([C₄mim]) ionic liquids with the anions acetate ([C₄mim][OAc]), nitrate ([C₄mim][NO₃]), trifluoromethanesulfonate ([C₄mim][OTf]), and bis(trifluoromethylsulfonyl)imide ([C₄mim][Tf₂N]) were synthesized via a halide-free route and contaminated on purpose with known amounts of [C₄mim]Cl, [C₄mim]Br, [C₄mim]I, or potassium halide salts in order to validate the new method and standard.

Full text of DOI:10.1021/ac5000812

Article
pubs.acs.org/ac
Determination of Halide Impurities in Ionic Liquids by Total
Reflection X‑ray Fluorescence Spectrometry
Tom Vander Hoogerstraete, Steven Jamar, Sil Wellens, and Koen Binnemans*
Department of Chemistry, KU Leuven, Celestijnenlaan 200F, P.O. Box 2404, Heverlee, B-3001, Belgium
ABSTRACT: The determination and quantification of halide
impurities in ionic liquids is highly important because halide
ions can significantly influence the chemical and physical
properties of ionic liquids. The use of impure ionic liquids in
fundamental studies on solvent extraction or catalytic reactions
can lead to incorrect experimental data. The detection of
halide ions in solution by total reflection X-ray fluorescence
(
TXRF) has been problematic because volatile hydrogen
halide (HX) compounds are formed when the sample is mixed
with the acidic metal standard solution. The loss of HX during
the drying step of the sample preparation procedure gives
imprecise and inaccurate results. A new method based on an
alkaline copper standard Cu(NH ) (NO ) is presented for the determination of chloride, bromide, and iodide impurities in
3
4
3 2
ionic liquids. The 1-butyl-3-methylimidazolium ([C mim]) ionic liquids with the anions acetate ([C mim][OAc]), nitrate
4
4
(
[C mim][NO ]), trifluoromethanesulfonate ([C mim][OTf]), and bis(trifluoromethylsulfonyl)imide ([C mim][Tf N]) were
4
3
4
4
2
synthesized via a halide-free route and contaminated on purpose with known amounts of [C mim]Cl, [C mim]Br, [C mim]I, or
4
4
4
potassium halide salts in order to validate the new method and standard.
onic liquids (ILs) are solvents that consist entirely of
ions. Typical cations are 1-alkyl-3-methylimidazolium, N-
alkylpyridinium, tetraalkyl ammonium, tetraalkyl phosphonium,
and N-alkyl-N-methylpyrrolidinium ions. Typical anions of
ionic liquids are chloride, bromide, thiocyanate, acetate,
tetrafluoroborate, hexafluorophosphate, triflate, and bis-
Depending on the experimental procedure and the nature of
the anion, newly synthesized ionic liquids prepared from halide
ionic liquids via a metathesis reaction will always contain
smaller or larger quantities of halide impurities. It is known that
the halide impurities strongly influence the physicochemical
1
−5
I
30
properties of ionic liquids. For instance, chloride impurities
30
increase the viscosity of tetrafluoroborate ionic liquids. Halide
(
trifluoromethylsulfonyl)imide ions. Ionic liquids can have
impurities also have a negative effect on the reactivity of
unique properties such as an extremely low vapor pressure, an
intrinsic ionic conductivity, a low flammability, and a wide
electrochemical window and can act as good solvents for many
classes of organic and inorganic compounds. Depending on
their chemical structure, ionic liquids can be either miscible or
immiscible with water. Ionic liquids are used as solvents for
reactions with organometallic catalysts in ionic liquids and on
31−33
enzymatic reactions.
For these reasons, it is important to
quantify the halide impurities in ionic liquids. Synthetic
chemists often rely on the “AgNO test” to check the purity
3
of their new ionic liquids: a few drops of a silver(I) nitrate
solution are added to an aqueous solution of the ionic liquid
6
,7
(
for hydrophilic ionic liquids) or to the washing water that has
29,34,35
catalytic reactions,
ymers, as electrolytes for electrodeposition of metals,
as solvent for processing of biopol-
8
,9
10−12
been used to rinse a hydrophobic ionic liquid.
The
as
13−15
absence of a silver halide precipitate is then accepted as a proof
of purity. However, this method is unreliable and cannot be
used to quantify the concentration of halide impurities. At best,
an upper limit of halide concentrations can be estimated. More
quantitative results can be obtained by argentometric titrations
the organic phase in solvent extraction,
and as engineering
1
6,17
fluids,
chemistry.
and they find many applications in analytical
1
8−21
Although halide-free routes have been
22−24
developed for the synthesis of ionic liquids,
most ionic
liquids are still prepared from chloride or bromide ionic liquids
(
Volhard method) or by turbidometric measurements using
25−29
36
via a metathesis reaction (Figure 1).
Nessler cylinders. Chloride- and bromide-selective electrodes
have been tested for the determination of halide impurities in
ionic liquids, but they are not sensitive enough and the
30
calibration procedure is difficult. For each ionic liquid cation,
Figure 1. General metathesis reaction for the synthesis of
imidazolium-based ionic liquids. A = desired anion, M = alkali metal
or silver, and X = Cl, Br, or I.
Received: January 7, 2014
Accepted: March 14, 2014
Published: March 17, 2014
©
2014 American Chemical Society
3931
dx.doi.org/10.1021/ac5000812 | Anal. Chem. 2014, 86, 3931−3938

Analytical Chemistry
Article
a different calibration curve needs to be constructed. Moreover,
this technique cannot be used for ionic liquids that are not
miscible with water. By far, the most often used method for
determination of halide impurities in ionic liquids is ion
3
7−41
chromatography (IC).
This analytical technique is
generally applicable to the determination of many different
anions and is not restricted to halide ions. However, some ions
such as bis(trifluoromethylsulfonyl)imide show very long
retention times on the IC columns with strong tailing as a
result. Moreover, the development of analysis methods can be
time-consuming, especially when different nonaqueous solvents
need to be used for dissolution of the ionic liquid samples.
Other methods that have been described in the literature for
the determination of halide impurities in ionic liquids include
42
capillary electrophoresis (CE), inductively coupled plasma
43
mass spectrometry (ICPMS), UV−vis absorption spectros-
4
4
45,46
Figure 2. Typical TXRF spectrum, showing the peaks of chlorine,
bromine, iodine, and copper. The molybdenum peaks are due to
Rayleigh and Compton scattering of photons of the molybdenum X-
ray source; potassium was the counterion of the halides, and the silicon
peaks come from the quartz sample carrier and its pretreatment.
copy, and electroanalytical methods.
Each of these
methods has its specific advantages and disadvantages.
In this paper, we describe for the first time the use of total
reflection X-ray fluorescence (TXRF) spectrometry for the
determination of halide impurities in ionic liquids. TXRF is
related to energy-dispersive X-ray fluorescence (EDX), but the
X-ray beam is directed at a very small angle to a smooth
substrate (less than 0.5°) and is under these circumstances
7
1,73
to the sample on the sample holder.
Another option is to
work in alkaline conditions, but not all standards are stable
under these conditions and metal hydroxides could precipitate
47−54
totally reflected on the surface.
The total reflection leads
72
to the formation of standing waves above the surface and down
to a few nanometers just below the surface, so that all X-ray
fluorescence is generated by atoms close to the surface. Samples
are prepared on optically flat carriers, typically polished quartz
glass disks. Solutions are pipetted onto these sample carriers
and, after evaporation, the dry residues are analyzed. A major
advantage of TXRF is that an easy calibration can be carried out
by addition of a single internal standard element. Other
advantages include the simple sample preparation, the fast
analysis, the small amounts of sample needed, the very wide
dynamic range (from ppb to 100%), and the low maintenance
costs. TXRF is a trace analysis technique, and in the cases of
homogeneous samples without organic materials such as water
samples, detection limits of close to 1 ppb can be reached for
the most X-ray sensitive elements. It is also a multielement
technique, and many different elements can be determined
simultaneously. A typical TXRF spectrum containing copper
and halides is given as an illustration in Figure 2.
at higher pH values. A better approach is probably to use a
neutral standard such as NH VO which has been used for the
4
3
7
4
determination of fluorine. Recently, we introduced Cu-
(NH ) (NO ) as a new internal standard which can be used
3
4
3 2
for the determination of halides in aqueous solutions. This
standard is stable under alkaline conditions, leaves no absorbing
matrix behind on the carrier, and is soluble in organic solvents.
Given the great potential of TXRF for determination of trace
amounts of halogens and the absence of reliable analysis
methods for halogens by TXRF, we explored the applicability of
TXRF for the determination of halide ion impurities in ionic
liquids. In this paper, we describe analytical protocols for the
determination of chloride, bromide, and iodide impurities in
ionic liquids by total-reflection X-ray fluorescence (TXRF). It
should be mentioned that TXRF is an element-specific
analytical method, so that chlorine, bromine, and iodine
concentrations are measured, rather than chloride, bromide,
and iodide concentrations.
Applications of TXRF include the determination of
5
5,56
EXPERIMENTAL SECTION
Chemicals. 1-Butyl-3-methylimidazolium chloride
impurities on silicon wafers in the microelectronic industry
■
57
and the analysis of surface water samples, wastewater
5
8,59
samples,
biological fluids (blood, blood serum,
([C mim]Cl, 99%), 1-butyl-3-methylimidazolium bromide
4
6
0−62
63
64
urine),
pharmaceutics, geological material, and aero-
([C mim]Br, 99%), 1-butyl-3-methylimidazolium iodide
4
6
5
sols, as well as the identification of artist pigments on
([C mim]I, 99%), a 30 wt % 1-butyl-3-methylimidazolium
4
6
6
67
paintings and for forensic investigations. TXRF has already
methyl carbonate solution in methanol ([C mim][MeCO ]),
trifluoromethanesulfonic acid (HOTf, 98%), and an 80 wt %
aqueous solution of hydrogen bis(trifluoromethylsulfonyl)-
4
3
68
successfully been used for the determination of bromate,
6
9
70
perchlorate, and bromide present in solid material. The
determination of dissolved halide salts is more challenging.
Different authors have reported the possibility of hydrogen
halide (HX) evaporation during the drying step when an acidic
imide (HTf N) (80 wt %) were purchased from IoLiTec
2
(Heilbronn, Germany). Nitric acid (65 wt %) and the ammonia
solution (25 wt %) were purchased from Chem-Lab NV
(Zedelgem, Belgium). Acetic acid (99%) and ethanol were
obtained from VWR (Leuven, Belgium). Potassium chloride
(KCl, 99.9%), potassium iodide (KI, 99.9%) and potassium
bromide (KBr, 99.9%) were obtained from Applichem
(Darmstadt, Germany). The detergent RBS 50 pF was obtained
from Chemical Products R. Borghgraef (Brussels, Belgium).
The copper(II) nitrate standard (1000 ppm) was purchased
from SCP Science (Courtaboeuf, France), and the silicone
71−75
internal standard is used for the TXRF measurements.
Almost all metal standards are diluted in 2−5 wt % of acid to
obtain a stable standard without the risk of metal hydrolysis and
precipitation. The acid can react with the halide and form HX
compounds, which are volatile and then evaporate during the
7
5
drying step. One way to bypass this problem is to perform the
measurement indirectly by precipitation of silver halides with
AgNO and measuring the excess of silver or by adding AgNO3
3
3
932
dx.doi.org/10.1021/ac5000812 | Anal. Chem. 2014, 86, 3931−3938

Analytical Chemistry
Article
solution in isopropanol was obtained from SERVA Electro-
phoresis GmbH (Heidelberg, Germany). The halide purity of
all chemicals were first checked by a standardless TXRF
method and approved to be useable without further
purification.
amount of ionic liquid to the pure ionic liquid. The reported
concentrations are not the halide concentrations present after
dilution with water or ethanol, but the halide concentrations in
the ionic liquid when no solvent would be present. Weighing of
the ionic liquid samples was performed in a glovebag under
nitrogen atmosphere.
Synthesis of 1-Butyl-3-methylimidazolium Acetate,
[
C4mim][OAc]. Acetic acid (10 mL, 180 mmol) was added
Preparation of Chloride Contaminated Ionic Liquid
slowly to a 30 wt % methanol solution of [C mim][MeCO ]
Solutions. [C mim]Cl was dried for 48 h in a vacuum oven at
4
3
4
(
140 mL, 175 mmol) and stirred for 1 h under a dynamic
50 °C. 0.986 g (27.8 mmol) and 0.138 g (7.0 mmol) of the salt
were dissolved in 50 and 500 mL of Milli-Q water, respectively,
to obtain solutions of 4000 and 100 ppm chloride. 2.5, 1, 0.5,
vacuum (Schlenk line) to obtain [C mim][OAc] (33.072 g;
1
line for 48 h at 60 °C. H NMR (300.13 MHz, DMSO-d , d/
ppm): 10.18 (s, 1H), 7.91 (s, 1H), 7.84 (s, 1H), 4.21 (t, 2H, J =
4
67 mmol; 95% yield) which was further dried on a Schlenk
1
and 0.25 mL of the 4000 ppm chloride solution of [C mim]Cl
6
4
were added to 1 g of [C mim][OAc] or [C mim][NO ] to
4
4
3
7
1
.2 Hz), 3.89 (s, 3H), 1.77 (m, 2H, J = 7.4 Hz), 1.59 (s, 3H),
obtain concentrations of 10 000, 4000, 2000, and 1000 ppm
.23 (m, 2H, J = 7.5 Hz), 0.87 (t, 3H, J = 7.2 Hz); 13C NMR
chloride in the contaminated [C mim][OAc]. 4, 2, and 1 mL of
4
(
75.48 MHz, DMSO-d , d/ppm): 173.2, 136.9, 123.6, 122.2,
a 100 ppm chloride of [C mim]Cl were added to 1 g of
6
4
4
8.4, 35.6, 31.3, 23.3, 18.7, 13.2.
Synthesis of 1-Butyl-3-methylimidazolium Bis-
[C mim][OAc] or [C mim][NO ] to obtain solutions of 400,
4
4
3
200, and 100 ppm chloride in the contaminated ionic liquids.
The solutions were further diluted with Milli-Q water to a final
volume of 50 mL.
(
trifluoromethylsulfonyl)imide, [C mim][Tf N]. A 80 wt
4 2
%
HTf2N solution (40 mL, 147 mmol) was added slowly to a
3
0 wt % methanol solution of [C mim][MeCO ] (100 mL, 125
Preparation of Bromide Contaminated [C mim][Tf N]
4
3
4
2
mmol) and stirred for 1 h under a dynamic vacuum (Schlenk
line) to obtain [C mim][Tf N] (50.012 g; 119 mmol; 96%
yield) which was washed three times with Milli-Q water and
Solutions. [C mim]Br was dried for 48 h in a vacuum oven at
4
50 °C. 0.548 g (2.5 mmol) and 0.138 g (0.6 mmol) of the salt
were dissolved in 50 and 500 mL of ethanol, respectively, to
obtain solutions of 5000 and 125 ppm bromide. 2.5, 1, 0.5, and
4
2
1
further dried on a Schlenk line for 48 h at 60 °C. H NMR
(
300.13 MHz, DMSO-d , d/ppm): 9.10 (s,1H), 7.75 (s, 1H),
0.25 mL of the 5000 ppm bromide solution of [C mim]Br were
6
4
7
.68 (s, 1H), 4.16 (t, 2H, J = 6.9 Hz), 3.86 (s, 3H), 1.78 (m,
added to 1 g of [C mim][Tf N] to obtain concentrations of 10
4
2
2H, J = 6.9 Hz), 1.26 (m, 2H, J = 7.4 Hz), 0.91 (t, 3H, J = 7.4);
000, 4000, 2000, and 1000 ppm bromide in [C mim][Tf N].
4
2
1
3
C NMR (75.48 MHz, DMSO-d , d/ppm): 136.4, 123.6,
One mL of a 125 ppm bromide of [C mim]Br was added to 1 g
6
4
1
22.2, 119.5 (q, J = 322 Hz), 48.5, 35.7, 31.3, 30.6, 18.7, 13.2.
Synthesis of 1-Butyl-3-methylimidazolium Trifluoro-
of [C mim][Tf N] to obtain a solution of 100 ppm bromide in
4 2
the ionic liquid. The solutions were further diluted with ethanol
methanesulfonate, [C mim][OTf]. A HOTf solution (6.8
mL, 76 mmol) was added slowly to a 30 wt % methanol
solution of [C mim][MeCO ] (18.2 mL, 76 mmol) and stirred
to a final volume of 50 mL.
4
Preparation of Bromide Contaminated [C
mim][OTf]
4
Solutions. [C mim]Br was dried for 48 h in a vacuum oven at
4
3
4
for 1 h under a dynamic vacuum (Schlenk line) to obtain
C mim][OTf] (16.555 g; 58 mmol; 76% yield) and further
50 °C. 0.689 g (3.1 mmol) and 0.170 g (0.8 mmol) of the salt
were dissolved in 50 and 500 mL of Milli-Q water, respectively,
to obtain solutions of 5000 and 125 ppm bromide. 2, 1, 0.4, and
[
4
1
dried on a Schlenk line for 48 h at 60 °C. H NMR (300.13
MHz, DMSO-d , δ/ppm): 9.06 (s, 1H), 7.74 (s, 1H), 7.67 (s,
0.2 mL of the 5000 ppm bromide solution of [C
mim]Br was
4
6
1
H), 4.15 (t, 2H, J = 7.3 Hz), 3.86 (s, 3H), 1.78 (m, 2H, J = 7.7
added to 1 g of [C mim][OTf] to obtain concentrations of 10
000, 5000, 2000, and 1000 ppm bromide in the contaminated
4
13
Hz), 1.27 (m, 2H, J = 7.7 Hz), 0.91 (m, 3H, J = 7.3 Hz);
C
NMR (75.48 MHz, DMSO-d , δ/ppm): 136.3, 122.7, 122.1,
ionic liquid [C mim][OTf]. 5, 2, and 1 mL of a 125 ppm
6
4
116.4 (q, J = 375 Hz), 48.5, 35.5, 31.2, 18.6, 13.0.
bromide solution of [C mim]Br was added to 1 g of
4
Synthesis of 1-Butyl-3-imidazolium Nitrate, [C mim]-
[C mim][OTf] to obtain a 500, 200, and 100 ppm bromide
4
4
[
NO ]. A 65 wt % HNO solution (12.0 mL, 173 mmol) was
contaminated [C mim][OTf] ionic liquid. All solutions were
3
3
4
added slowly to a 30 wt % methanol solution of [C mim]-
further diluted with Milli-Q water to a final volume of 50 mL.
4
[
MeCO ] (140.0 mL, 175 mmol) and stirred for 1 h under
Preparation of Iodide Contaminated Ionic Liquid
3
vacuum to obtain [C mim][NO ] (34.248 g; 170 mmol; 98%
Solutions. [C mim]I was dried for 48 h in a vacuum oven
at 50 °C. 0.528 g (2.0 mmol) of the salt was dissolved in 50 mL
of Milli-Q water to obtain solutions of 5000 ppm iodide. 2, 1,
4
3
4
1
yield) and further dried on a Schlenk line for 48 h at 60 °C. H
NMR (300.13 MHz, DMSO-d , δ/ppm): 9.13 (s, 1H), 7.77 (s,
6
1
H), 7.70 (s, 1H), 4.16 (t, 2H, J = 7.2 Hz), 3.84 (s, 3H), 1.76
0.4, and 0.2 mL of the 5000 ppm iodide solution of [C mim]I
4
13
(
m, 2H, J = 7.4 Hz), 1,27 (m, 2H), 0.90 (t, 3H, J = 7.2 Hz); C
were added to 1 g of [C mim][OTf], [C mim][NO ] or
4
4
3
NMR (75.48 MHz, DMSO-d , δ/ppm): 136.6, 123.6, 122.2,
[C mim][OAc] to obtain concentrations of 10 000, 5000, 2000,
6
4
4
8.5, 35.6, 31.3, 18.7, 13.2.
Preparation of the Alkaline Cu(NH ) (NO ) Standard.
and 1000 ppm iodide in the ionic liquids. The solutions were
further diluted with Milli-Q water to a final volume of 50 mL.
Preparation of Ionic Liquids Contaminated with KBr
Salts. The halide salt KBr was first dried in a vacuum oven at
50 °C for 48 h. Afterward, 0.074 g (0.6 mmol) of KBr was
dissolved in 500 mL of Milli-Q water to obtain a solution of
3
4
3 2
2
.5 mL of a 25 wt % NH OH aqueous solution was added to 5
4
mL of a 1000 ppm copper standard containing 5 wt % of nitric
acid. The solution was further diluted to 50 mL in order to
obtain a copper standard with a concentration of 100 ppm.
Preparation of the Halide Contaminated Ionic
100 ppm Br. The contaminated ionic liquid [C mim][OTf]
4
Liquids. The contaminants [C mim]X or KX were dissolved
was prepared in the same way as described above for the ionic
liquid contaminants.
4
in water or ethanol, and small amounts of the solutions were
added to the halide-free ionic liquid. This makes the addition
quantitatively more correct and easier than adding a weighed
1
Instrumentation and Analysis Methods. The H and
13
C NMR spectra were recorded with a Bruker Avance 300
3
933
dx.doi.org/10.1021/ac5000812 | Anal. Chem. 2014, 86, 3931−3938

Analytical Chemistry
spectrometer, operating at 300 MHz for H and 75 MHz for
C. 18.2 MΩ·cm Milli-Q water was prepared with a Synergy
Article
1
Chloride Impurities in Ionic Liquids. The most common
method for the synthesis of [C mim][OAc] or [C mim][NO ]
is by the addition of KOAc or KNO to [C mim]Cl in dry
1
3
4
4
3
UV water purification system. The halide anion concentrations
were determined with a benchtop total reflection X-ray
fluorescence (TXRF) spectrometer (Picofox S2, Bruker).
Calibration curves and calculated concentrations of chlorine,
bromine, and iodine were made with the elemental sensitivities
supplied by Bruker. Glasswork and sample carriers were washed
for 30 min with hot RBS 50 pF and for 2 h with hot nitric acid
to remove all impurities. Afterward, they have been rinsed with
Milli-Q water and acetone and dried in an oven at 60 °C for 2
h. All dilutions were prepared with Milli-Q water to avoid
chloride or bromide contamination during the sample or
standard preparation. 0.5 mL of each contaminated sample was
added to 0.5 mL of the alkaline copper standard and mixed to
obtain a homogeneous solution. If necessary, dilution of the
internal standard was made in order to obtain equal halide and
copper concentrations during measurements. The sample
carriers were first pretreated by adding 20 μL of silicone
solution. After 2 min, 10 μL of the mixture containing the
sample and internal standard solution was added onto the
sample carrier and dried for 60 min in an oven at 40 °C. The
halide concentrations in the contaminated ionic liquids were
measured 6 to 10 times. The effect of the matrix on the
recovery rates was studied by mixing 0.5 mL of a 10 000 ppm
chloride or bromide contaminated ionic liquid with 0.5 mL of a
3 4
acetone. In this reaction, the formed KCl is insoluble and it
precipitates and can be removed by filtration. The main
impurity in the ionic liquids obtained from this synthesis
method is [C mim]Cl. Therefore, to reproduce these
4
conditions, different amounts of [C mim]Cl were added to
4
[
C mim][OAc] or [C mim][NO ] and measured with an
4 4 3
alkaline copper standard. Although we observed in our recent
study on the determination of halides in aqueous solution by
TXRF that concentrated matrices are harmful for the
75
quantification of chloride, this issue was not observed in
75
the case of chloride determination in ionic liquids. Recovery
rates (RR) between 85% and 110% were obtained in the case of
[
C mim][NO ] for chlorine concentrations above 400 ppm
4 3
and between 90% and 98% in the case of [C mim][OAc] over
4
the whole concentration range (Table 1). The difference
Table 1. Recovery Rates (RR) and Relative Standard
Deviations (RSD) of the Chloride Concentrations (Added as
[
[
C mim]Cl) in the Ionic Liquids [C mim][OAc] and
4
4
C mim][NO ]
4
3
[
C mim][OAc]
4
[C mim][NO ]
4
3
[
Cl] (ppm)
0 000
RR (%)
92
RR (%)
7
RSD (%)
RSD (%)
1
85
89
10
6
10 and a 100 ppm alkaline copper standard, respectively.
5
4
2
1
000
000
000
000
Samples which gave an ionic liquid droplet with a diameter
smaller than 3 mm on the carrier after the drying process were
not measured and were prepared again.
90
94
97
91
96
98
8
14
5
100
98
9
12
6
4
00
10
10
13
110
137
191
RESULTS AND DISCUSSION
■
200
100
5
Choice of Sample Preparation. Measuring elemental
concentrations and in particular of elements emitting low
energy X-rays, such as chlorine, in nonvolatile samples like ionic
liquids is a challenge. In general, sample masses lower than 10
μg after drying of the aqueous samples are advised. However,
solid samples with higher sample masses can be measured with
10
between chlorine determination in aqueous solutions and ionic
liquids could be due to the difference in distribution of chlorine
and copper in both samples after the drying process. Copper
and chlorine are dissolved homogeneously in the ionic liquid
droplet which remains on the sample carrier after the drying
procedure. This is not always the case when a solvent
evaporates and crystallization processes of the remaining salts
start to occur. The formation of crystals is dependent on the
solubility product of the corresponding salts. This means that
some salts crystallize faster than others and form aggregates,
giving a less homogeneous distribution of all elements on the
sample carrier. The relative standard deviations (RSDs) were
smaller than 15% in all cases (Table 1, Figure 3). We also
observed that concentrations of 200 ppm were difficult to
measure. This is because of the 100-fold dilutions in which only
2 ppm of chloride is actually measured. Overlapping peak tails
of, for example, argon, and small chloride contaminations on
the carrier, in the starting products or on the glassware can
therefore have a large influence on the recalculated end results.
The same observation was also made for determination of
7
6
TXRF as well. The sample amount is kept low in order to
avoid matrix effects. In the case of ionic liquids, a small, viscous
droplet will be present on the sample carrier in which all the
elements are dissolved. Therefore, some precautions were taken
to reduce the matrix effects as much as possible. The ionic
liquids were diluted at least 100 times in water or ethanol. In
this way, 10 μL of the solution gives a maximum of 100 μg of a
sample on the sample carrier. This 100-fold dilution also has a
significant influence on the detection limits which increase by
the same factor. Twenty μL of silicone solution was used in the
case of aqueous samples, and sampling was performed 2 min
after the addition of the silicone solution. This fast sampling
procedure was used because it ensures flat and thin distributed
ionic liquid droplets on the carrier. Isopropanol from the
silicone solution which is not yet evaporated after 2 min
distributes the aqueous sample over the carrier whereas
silicones ensure that the droplet remains fixed in the center
of the carrier. A sample was not measured in the rare case that a
droplet smaller than 3 mm was obtained after drying. Ideal
samples are flat, homogeneously distributed over the carrier,
and between 0.3 and 0.8 mm in diameter. The samples were
dried slowly at 40 °C for 1 h. Higher temperatures decrease the
viscosity of the ionic liquid and could lead to shrinkage of the
ionic liquid droplet on the carrier and thus to a thicker droplet.
75
chloride in water samples.
Bromide Impurities in Ionic Liquids. The hydrophobic
ionic liquid [C mim][Tf N] was selected to demonstrate that
4
2
the method can also be applied to water insoluble ionic liquids.
In general, [C mim][Tf N] can be easily synthesized by a
4
2
metathesis reaction between an alkali salt such as LiTf N and a
2
halide precursor such as [C mim]Br or [C mim]Cl. [C mim]
4
4
4
Cl was not chosen as an impurity because there is an overlap
3
934
dx.doi.org/10.1021/ac5000812 | Anal. Chem. 2014, 86, 3931−3938

Analytical Chemistry
Article
Table 2. Recovery Rates (RR) and Relative Standard
Deviations (RSD) of the Bromide Concentrations (Added as
[
C mim]Br) in [C mim][Tf N] and [C mim][OTf]
4
4
2
4
[
C mim][Tf N]
[C mim][OTf]
4
4
2
[
Br] (ppm)
0 000
RR (%)
105
RSD (%)
6
RR (%)
RSD (%)
1
100
103
5
3
5
4
2
1
000
000
000
000
105
113
100
9
4
6
96
105
101
102
97
8
6
6
9
4
5
2
1
00
00
00
100
6
Figure 3. Plot of the recovered chloride concentrations (added as
[
C mim]Cl) in [C mim][OAc] and [C mim][NO ].
4 4 4 3
smaller than 5% (Table 4). By correcting the sensitivity factor
for iodine, a new and good fitting calibration curve was
obtained. However, it has to be mentioned that this sensitivity
factor was very different from the one used for a pure water
phase. Depending on the matrix, different sensitivity factors for
one element are possible in TXRF, but deviations of about 70%
are strange. The values found for [C mim]I in [C mim][OTf]
between the chlorine and sulfur peak, present in the Tf N
2
anion, in the TXRF spectrum. After synthesis and several
washing steps with water, almost all LiBr is removed from the
hydrophobic ionic liquid phase and the main impurity
remaining in the ionic liquid is [C mim]Br. This impurity
4
was imitated by adding a known amount of [C mim]Br to the
4
4
4
[
C mim][Tf N] ionic liquid, which was synthesized via a
4 2
were quite scattered and did not lie on one straight calibration
line. For a concentration of 10 000 ppm, a recovery rate of
131% was found and the recovery rates decreased to 37% with
decreasing concentrations of iodide in the ionic liquids. Also,
the RSD values are much higher than in [C mim][NO ] or
halide-free method. The reliability of the copper standard
Cu(NH ) (NO ) was also tested for the presence of [C mim]
3
4
3 2
4
Br impurities in the hydrophilic ionic liquid [C mim][OTf].
4
The recovery rates of [C mim]Br in both ionic liquids were
4
4
3
found to be in between 96% and 113%, and the relative
standard deviations (RSDs) were smaller than 10% (Figure 4,
[
C mim][OAc] (15−26%) (Table 3). At high iodide
4
concentrations, small crystals were observed in the ionic liquid
on the sample carrier. It is possible that the solubility of the
iodide salt in the ionic liquid is not high enough after drying
and that the distribution of iodide and copper in the sample is
no longer homogeneous, resulting in wrong results. Moreover,
the sensitivity of iodine is lower than that of chlorine when a
TXRF apparatus with a molybdenum source is used. For
concentrations below 500 ppm and a 100-fold dilution, no
iodine peaks could clearly be detected. A TXRF apparatus with
a tungsten source could be a solution for this problem because
the iodine sensitivity of the peak is much higher.
KX Impurities. Although less KX impurities are present in
synthesized ionic liquids, they are also investigated as possible
contamination. Due to the lower solubility of KX salts in ionic
77
liquids, a lower concentration range was selected. KBr
concentrations of 100, 200, and 500 ppm were added to
[
C mim][OTf]. Also KI and KCl salts were tested but at higher
4
Figure 4. Plot of the recovered bromide concentrations (added as
concentration in order to obtain a higher signal-to-noise ratio of
the characteristic peaks after dilution. Formation of crystals was
observed in the ionic liquid, and the recovery rates were far
lower than 100%. Again, a homogeneous distribution of the
halide salt and copper standard is necessary to obtain equal
matrix effects for both elements and in order to find correct
concentration values. Due to the much higher sensitivity of
bromine for primary X-rays (iodine is measured on the low
energetic L-line), significantly lower concentrations could be
tested. Recovery rates of 88%, 89%, and 94% were obtained for
100, 200, and 500 ppm bromide, respectively. The RSD values
were below 10% (Table 4).
[
C mim]Br) in [C mim][Tf N] and [C mim][OTf].
4 4 2 4
Table 2). In general, the recovery rates and RSD values for
bromine are better than those for chlorine. This is due to the
lower absorption of the secondary, higher energy X-rays of
bromine by the ionic liquid matrix.
Furthermore, an acidic copper standard was tested in the
lower bromine concentration range with [C mim][OTf] as
4
contaminated ionic liquid. The recovery rates of 100, 200, and
500 ppm bromine in the ionic liquid were 15%, 8%, and 6%,
respectively, with RSD values between 10% and 20%.
Iodide Impurities in Ionic Liquids. The ionic liquids
Matrix Effects. One of the main advantages of using TXRF
in comparison with EDX is the lower matrix effect. This
advantage is valid if the sample is embedded in small amounts
of matrix and if heavier, less matrix sensitive elements are
measured. There were significant matrix effects when larger
sample amounts were used or when light elements were
[
C mim][NO ], [C mim][OAc], and [C mim][OTf] were all
4 3 4 4
contaminated on purpose with various amounts of [C mim]I.
4
The measured iodide concentrations for the ionic liquids
[
C mim][OAc] and [C mim][NO ] were about 70% higher
4 4 3
than the expected values although the RSD values were all
3
935
dx.doi.org/10.1021/ac5000812 | Anal. Chem. 2014, 86, 3931−3938

Analytical Chemistry
Article
Table 3. Recovery Rates (RR) and Relative Standard Deviations (RSD) of the Iodide Concentrations (Added as [C mim]I) in
4
[
C mim][NO ], [C mim][OTf], and [C mim][OAc]
4
3
4
4
[
C mim][NO ]
[C mim][OTf]
4
[C mim][OAc]
4
4
3
[I] (ppm)
RR (%)
RSD (%)
RR (%)
RSD (%)
RR (%)
RSD (%)
1
0 000
163
165
152
139
4
4
3
4
131
94
12
15
26
15
169
171
178
175
1
3
4
3
5
2
1
000
000
000
39
37
Table 4. Recovery Rates (RR) and Relative Standard
Deviations (RSD) of the Bromide Concentrations (KBr) in
[
C mim][OTf]
4
[
C mim][OTf]
4
[Br] (ppm)
RR (%)
RSD (%)
5
2
1
00
00
00
94
89
88
5
5
6
measured. For conventional TXRF measurements, it is advised
to add a maximum of 10 μg onto the sample carrier. For ionic
liquids, this is a problem because they form a nonvolatile matrix
on the carrier, which cannot be removed during the drying
process. In our measurements, a maximum sample weight of
1
00 μg was used. Working with lower amounts is possible by
Figure 6. Influence of the sample amount on the bromine recovery
rate and number of counts. Sample: 0.5 mL of 100 ppm alkaline
copper standard and 0.5 mL of a 10 000 ppm bromide impurity
further diluting the ionic liquid. The drawback of dilution is
that the number of counts and thus the detection limit of
halides in ionic liquids will decrease. Working with high sample
amounts can give larger errors on the recovery rates (Figures 5
(
added as [C mim]Br) in [C mim][Tf N] diluted in 50 mL of
4 4 2
ethanol. Measured concentrations after dilution were 50 ppm Cu and
1
00 ppm Br.
sample. The count ratio between chlorine or bromine and the
copper standard is changing and has an influence on the
calculated concentration. The effect is very clear when chlorine
is determined with the copper standard. The recovery rates for
chlorine in 50 or 100 μg of ionic liquid on the sample carrier
are still above 85%. The chlorine recovery rates drastically
decrease when the sample amount is further increased. The
number of counts for copper is increasing with increasing
sample amount. In addition, the counts originating from
chlorine are even decreasing due to absorption of the matrix
when more sample is loaded on the sample carrier (Figure 5).
For the same reason, the shape of the droplet is of crucial
importance when measuring chlorine concentrations. Multiple
thicker droplets on the sample carrier give lower recovery rates
than a flatter homogeneous droplet with a large diameter
because the chlorine X-rays have to travel longer distances
through the matrix. Shorter drying times for instance give larger
solvent matrices, while higher drying temperatures or the
absence of a silicone coating on the sample carrier give often
multiple thick droplets. Furthermore, thick droplets can
contaminate the detector, which is positioned very close to
the sample. Therefore, it is recommended to follow the sample
preparation instructions very carefully in case of chlorine
measurements. For an ionic liquid contaminated with bromide,
the effect is reversed but much less pronounced (Figure 6).
Copper is a lighter element than bromine, which means that the
secondary X-rays of copper are more strongly absorbed by the
matrix than those of bromine.
Figure 5. Influence of the sample amount on the chlorine recovery
rate (■) and number of counts for copper (×) and chlorine (●).
Sample: 0.5 mL of a 10 ppm alkaline copper standard and 0.5 mL of a
1
0 000 ppm chloride impurity (added as [C mim]Cl) in 1 g
4
[
C mim][OAc] diluted in 50 mL H O. Measured concentrations
4 2
after dilution were 5 ppm Cu and 100 ppm Cl.
and 6). The main scattering process for low energy secondary
X-rays is a photoelectric effect. Here, a secondary X-ray,
produced by a lighter element in the matrix, totally loses its
energy by ejecting again an electron from a specific matrix
element. The element-specific mass attenuation coefficient,
defined as the absorption of radiation of a specific wavelength
per unit mass, is significantly higher for the energy X-rays of the
7
8
lighter elements. Large amounts of organic matrices, such as
ionic liquids, will significantly absorb secondary X-rays and
decrease the number of counts for the lighter elements. The
absorption is less pronounced for X-rays coming from heavier
elements and is thus not equal to all elements present in the
Due to the very small angle of the incident primary X-ray
beam, TXRF is only analyzing the surface of the ionic liquid
droplet. Therefore, the penetration depth, which is defined as
3
936
dx.doi.org/10.1021/ac5000812 | Anal. Chem. 2014, 86, 3931−3938

Analytical Chemistry
Article
the depth at which the X-ray intensity is reduced to 1/e, is very
low. The penetration depth depends on the energy of the X-
rays, the type of matrix, the shape of the remaining ionic liquid
droplet, and the incident angle of the primary X-rays.
the ionic liquid. The detection limits of halides are in the ppb
range but increase 100-fold due to the required dilution to 2,
20, and 15 ppm for bromine, chlorine, and iodine, respectively.
Klockenkam
̈
per reported penetration depths in TXRF measure-
AUTHOR INFORMATION
Corresponding Author
E-mail: Koen.Binnemans@chem.kuleuven.be.
■
54
ments ranging between 5 and 320 nm. An inhomogeneous
accumulation of copper or the halides at the ionic liquid surface
could lead to wrong results. Recent articles reported the
enrichment of the larger anions and cations on the ionic
*
Notes
The authors declare no competing financial interest.
79−81
liquid−air interface.
first few nanometers of the water-free ionic liquid [EMIM]-
EtOSO ] (where [EtOSO ] is ethyl sulfate) were enriched
Moreover, Maier et al. found that the
ACKNOWLEDGMENTS
■
[
3
3
2
+
The authors thank the KU Leuven (GOA/13/008 and IOF-KP
with the larger cation [Pt(NH ) ] and depleted with the
3
4
3
82
RARE ) and the FWO Flanders (project G.0900.13) for
smaller chlorine anion of the complex [Pt(NH ) ]Cl
A
3
4
2.
financial support.
platinum enrichment factor of 3 was found in comparison with
the bulk patina concentration. The enrichment of metal
complexes on the ionic liquid−air surface could lead to
determination of halide concentrations which are too low. In
our measurements, it was expected that copper would segregate
more to the surface than the halides ions. However, the effect
was not observed which could be due to four different reasons.
First, the penetration depth of the X-rays is deeper than the
surface which was studied in the cited references. Second, our
studies were performed with ionic liquids which were not
extremely dry. Surface effects in wet ionic liquids can be
different from surface effects in dry ionic liquids, but such
comparative studies have not been described in the literature
yet. Third, the error which is due to absorption is much larger
than the error which is due to enrichment of copper and
depletion of the halide at the surface. Anyhow, it is advisable to
use sample amounts which are similar to those used for
construction of the calibration curves.
REFERENCES
■
(
(
(
1) Welton, T. Chem. Rev. 1999, 99, 2071−2083.
2) Hallett, J. P.; Welton, T. Chem. Rev. 2011, 111, 3508−3576.
3) Wasserscheid, P.; Keim, W. Angew. Chem., Int. Ed. 2000, 39,
3
(
(
772−3789.
4) Seddon, K. R. J. Chem. Technol. Biotechnol. 1997, 68, 351−356.
5) Plechkova, N. V.; Seddon, K. R. Chem. Soc. Rev. 2008, 37, 123−
1
50.
(
6) Welton, T. Coord. Chem. Rev. 2004, 248, 2459−2477.
(7) Parvulescu, V. I.; Hardacre, C. Chem. Rev. 2007, 107, 2615−2665.
(8) Swatloski, R. P.; Spear, S. K.; Holbrey, J. D.; Rogers, R. D. J. Am.
Chem. Soc. 2002, 124, 4974−4975.
9) Fort, D. A.; Remsing, R. C.; Swatloski, R. P.; Moyna, P.; Moyna,
(
G.; Rogers, R. D. Green Chem. 2007, 9, 63−69.
(
(
10) Endres, F. Z. Phys. Chem. 2004, 218, 255−283.
11) Armand, M.; Endres, F.; Macfarlane, D. R.; Ohno, H.; Scrosati,
B. Nat. Mater. 2009, 8, 621−629.
12) Abbott, A. P.; McKenzie, K. J. Phys. Chem. Chem. Phys. 2006, 8,
265−4279.
13) Vander Hoogerstraete, T.; Wellens, S.; Verachtert, K.;
Binnemans, K. Green Chem. 2013, 15, 919−927.
(14) Vander Hoogerstraete, T.; Onghena, B.; Binnemans, K. J. Phys.
Chem. Lett. 2013, 4, 1659−1663.
15) Huddleston, J. G.; Willauer, H. D.; Swatloski, R. P.; Visser, A. E.;
Rogers, R. D. Chem. Commun. 1998, 1765−1766.
16) Werner, S.; Haumann, M.; Wasserscheid, P. Annu. Rev. Chem.
(
4
(
Detection limits. The detection limits of TXRF depend on
different parameters such as the concentration, matrix,
measuring time, etc. Detection limits between 0.100 and
0
.200 ppm were found for measured chlorine concentrations
below 100 ppm after ionic liquid dilution. The detection limits
in the same concentration range were between 0.005 and 0.020
ppm for bromine and 0.100 and 0.150 ppm for iodine.
Unfortunately, it was necessary to dilute the ionic liquid 100
times to avoid matrix effects and contamination of the detector.
This also means that the actual detection limits for halides in
ionic liquid increase 100-fold to a maximum of 2 ppm bromine,
(
(
Biomol. Eng. 2010, 1, 203−230.
(
(
(
17) Zhao, H. Chem. Eng. Commun. 2006, 193, 1660−1677.
18) Koel, M. Crit. Rev. Anal. Chem. 2005, 35, 177−192.
19) Liu, J. F.; Jonsson, J. A.; Jiang, G. B. TrAC, Trends Anal. Chem.
20 ppm chlorine, and 15 ppm iodine. Another problem is
2
005, 24, 20−27.
measuring chlorine in ionic liquids containing sulfur because
there is a significant overlap between the characteristic X-ray
peaks of these elements. Future research will focus on the
removal of the matrix from the impurities to reach the very low
detection limits of the equipment.
(20) Sun, P.; Armstrong, D. W. Anal. Chim. Acta 2010, 661, 1−16.
(21) Berthod, A.; Ruiz-Angel, M.; Carda-Broch, S. J. Chromatogr., A
2008, 1184, 6−18.
(
22) Alcalde, E.; Dinares, I.; Ibanez, A.; Mesquida, N. Molecules 2012,
1
(
7, 4007−4027.
23) Gao, J.; Liu, J. G.; Liu, W. M.; Li, B.; Xin, Y. C.; Yin, Y.; Gu, J.;
Zou, Z. G. Int. J. Hydrogen Energy 2012, 37, 13167−13177.
(24) Holbrey, J. D.; Reichert, W. M.; Swatloski, R. P.; Broker, G. A.;
Pitner, W. R.; Seddon, K. R.; Rogers, R. D. Green Chem. 2002, 4, 407−
413.
CONCLUSIONS
■
It has been proven that chloride and bromide contaminations
in ionic liquids could be detected and quantified accurately and
precisely by measuring the chlorine or bromine content with
TXRF as long as the total amount and the shape of ionic liquid
droplet on the sample carrier were carefully controlled. The
results were not as good for iodide impurities. Two calibration
curves could be drawn through the different data points, but the
sensitivity factor was lying about 70% above what was expected.
Moreover, a precipitate and an inhomogeneous distribution
were observed on the sample carrier which was giving incorrect
results and was probably due to a lower solubility. A similar
observation was made for high concentrations of halide salts in
(25) Wilkes, J. S.; Zaworotko, M. J. J. Chem. Soc. Chem. Comm. 1992,
9
(
65−967.
26) Bonhote, P.; Dias, A. P.; Papageorgiou, N.; Kalyanasundaram,
K.; Gratzel, M. Inorg. Chem. 1996, 35, 1168−1178.
27) Nockemann, P.; Thijs, B.; Driesen, K.; Janssen, C. R.; Van
(
Hecke, K.; Van Meervelt, L.; Kossmann, S.; Kirchner, B.; Binnemans,
K. J. Phys. Chem. B 2007, 111, 5254−5263.
(28) Nockemann, P.; Binnemans, K.; Thijs, B.; Parac-Vogt, T. N.;
Merz, K.; Mudring, A. V.; Menon, P. C.; Rajesh, R. N.; Cordoyiannis,
G.; Thoen, J.; Leys, J.; Glorieux, C. J. Phys. Chem. B 2009, 113, 1429−
1437.
3
937
dx.doi.org/10.1021/ac5000812 | Anal. Chem. 2014, 86, 3931−3938

Analytical Chemistry
Article
(
29) Lethesh, K. C.; Van Hecke, K.; Van Meervelt, L.; Nockemann,
(64) Muia, L.; Vangrieken, R. Anal. Chim. Acta 1991, 251, 177−181.
(65) Bontempi, E.; Zacco, A.; Benedetti, D.; Borgese, L.; Colombi,
P.; Stosnach, H.; Finzi, G.; Apostoli, P.; Buttini, P.; Depero, L. E.
Environ. Technol. 2010, 31, 467−477.
P.; Kirchner, B.; Zahn, S.; Parac-Vogt, T. N.; Dehaen, W.; Binnemans,
K. J. Phys. Chem. B 2011, 115, 8424−8438.
(
30) Seddon, K. R.; Stark, A.; Torres, M. J. Pure Appl. Chem. 2000,
7
(
2, 2275−2287.
31) Stark, A.; Ajam, M.; Green, M.; Raubenheimer, H. G.; Ranwell,
A.; Ondruschka, B. Adv. Synth. Catal. 2006, 348, 1934−1941.
32) Lee, S. H.; Ha, S. H.; Lee, S. B.; Koo, Y. M. Biotechnol. Lett.
006, 28, 1335−1339.
33) Gallo, V.; Mastrorilli, P.; Nobile, C. F.; Romanazzi, G.; Suranna,
G. P. J. Chem. Soc. Dalton 2002, 4339−4342.
34) Nikitenko, S. I.; Cannes, C.; Le Naour, C.; Moisy, P.; Trubert,
D. Inorg. Chem. 2005, 44, 9497−9505.
35) Lunstroot, K.; Driesen, K.; Nockemann, P.; Gorller-Walrand, C.;
Binnemans, K.; Bellayer, S.; Le Bideau, J.; Vioux, A. Chem. Mater.
006, 18, 5711−5715.
36) Stark, A.; Behrend, P.; Braun, O.; Muller, A.; Ranke, J.;
Ondruschka, B.; Jastorff, B. Green Chem. 2008, 10, 1152−1161.
37) Villagran, C.; Deetlefs, M.; Pitner, W. R.; Hardacre, C. Anal.
Chem. 2004, 76, 2118−2123.
38) Villagran, C.; Banks, C. E.; Deetlefs, M.; Driver, G.; Pitner, W.
(66) Moens, L.; Devos, W.; Klockenkamper, R.; von Bohlen, A.
TrAC, Trends Anal. Chem. 1994, 13, 198−205.
(67) Dhara, S.; Misra, N. L.; Maind, S. D.; Kumar, S. A.;
Chattopadhyay, N.; Aggarwal, S. K. Spectrochim. Acta B 2010, 65,
167−170.
(
2
(
(68) Hatzistavros, V. S.; Koulouridakis, P. E.; Aretaki, I. I.;
Kallithrakas-Kontos, N. G. Anal. Chem. 2007, 79, 2827−2832.
(69) Hatzistavros, V. S.; Kallithrakas-Kontos, N. G. Anal. Chem.
2011, 83, 3386−3391.
(
(
(70) Fink, H.; Panne, U.; Theisen, M.; Niessner, R.; Probst, T.; Lin,
X. Fresenius J. Anal. Chem. 2000, 368, 235−239.
(71) Misra, N. L.; Varga, I.; Dhara, S.; Aggarwal, S. K. X-Ray
2
(
Spectrom. 2009, 38, 182−185.
72) Dhara, S.; Misra, N. L.; Thakur, U. K.; Shah, D.; Sawant, R. M.;
Ramakumar, K. L.; Aggarwal, S. K. X-Ray Spectrom. 2012, 41, 316−
20.
73) Varga, I. Microchem. J. 2007, 85, 127−131.
74) Tarsoly, G.; Ovari, M.; Zaray, G. Spectrochim. Acta B 2010, 65,
87−290.
75) Vander Hoogerstraete, T.; Jamar, S.; Wellens, S.; Binnemans, K.
Anal. Chem. 2014, 86, 1391−1394.
76) Towett, E. K.; Shepherd, K. D.; Cadisch, G. Sci. Total Environ.
013, 463−464, 374−388.
77) Pereiro, A. B.; Araujo, J. M. M.; Oliveira, F. S.; Esperanca, J. M.
(
(
3
(
(
2
(
R.; Compton, R. G.; Hardacre, C. Ionic Liquids IIIB: Fundamentals,
Progress, Challenges and Opportunities: Transformations and Processes;
Rogers, R. D., Seddon, K. R., Eds.; American Chemical Society:
Washington, DC, 2005, Volume 902.
(
(
2
(
(
39) Li, S. W.; Yu, H.; Zhang, X. Chin. J. Anal. Chem. 2010, 38,
665−1669.
40) Hao, F. P.; Haddad, P. R.; Ruther, T. Chromatographia 2008, 67,
95−498.
41) Stepnowski, P.; Markowska, A. Aust. J. Chem. 2008, 61, 409−
13.
42) Berthier, D.; Varenne, A.; Gareil, P.; Digne, M.; Lienemann, C.
P.; Magna, L.; Olivier-Bourbigou, H. Analyst 2004, 129, 1257−1261.
43) McCamley, K.; Warner, N. A.; Lamoureux, M. M.; Scammells,
P. J.; Singer, R. D. Green Chem. 2004, 6, 341−344.
44) Bartosik, J.; Mudring, A. V. Phys. Chem. Chem. Phys. 2010, 12,
005−4011.
45) Ge, R.; Allen, R. W. K.; Aldous, L.; Bown, M. R.; Doy, N.;
Hardacre, C.; MacInnes, J. M.; McHale, G.; Newton, M. I. Anal. Chem.
009, 81, 1628−1637.
46) Villagran, C.; Banks, C. E.; Hardacre, C.; Compton, R. G. Anal.
Chem. 2004, 76, 1998−2003.
47) Klockenkamper, R.; Knoth, J.; Prange, A.; Schwenke, H. Anal.
1
(
4
(
4
(
S. S.; Canongia Lopes, J. N.; Marrucho, I. M.; Rebelo, L. P. N. J. Chem.
Thermodyn. 2012, 55, 29−36.
(78) Hubbel, J. H.; Seltzer, S. M. Tables of X-Ray Mass Attenuation
Coefficients and Mass Energy-Absorption Coefficients (version 1.4),
National Institute of Standards and Technology: Gaithersburg, MD,
2
004; Available: http://physics.nist.gov/xaamdi.
(
(79) Nakajima, K.; Oshima, S.; Suzuki, M.; Kimura, K. Surf. Sci. 2012,
6
(
(
06, 1693−1699.
(
4
(
80) Souda, R. Surf. Sci. 2010, 604, 1694−1697.
81) Nakajima, K.; Miyashita, M.; Suzuki, M.; Kimura, K. J. Chem.
Phys. 2013, 139, 224701.
(82) Maier, F.; Cremer, T.; Kolbeck, C.; Lovelock, K. R. J.; Paape, N.;
2
(
Schulz, P. S.; Wasserscheid, P.; Steinruck, H. P. Phys. Chem. Chem.
Phys. 2010, 12, 1905−1915.
(
Chem. 1992, 64, 1115−1122.
(
(
(
48) Alov, N. V. Inorg. Mater. 2011, 47, 1487−1499.
49) von Bohlen, A. Spectrochim. Acta B 2009, 64, 821−832.
50) Klockenkamper, R.; von Bohlen, A. Spectrochim. Acta B 2001,
5
(
(
(
6, 2005−2018.
51) Wobrauschek, P. X-Ray Spectrom. 2007, 36, 289−300.
52) Stosnach, H. Powder Diffr. 2005, 20, 141−145.
53) Tolg, G.; Klockenkamper, R. Spectrochim. Acta B 1993, 48,
1
11−127.
54) Klockenkamper, R. Total reflection X-ray fluorescence analysis;
John Wiley & Sons: New York, 1997.
55) Schwenke, H.; Beaven, P. A.; Knoth, J. Fresenius J. Anal. Chem.
999, 365, 19−27.
56) Misra, N. L.; Mudher, K. D. S. Prog. Cryst. Growth Charact.
(
(
1
(
2
(
(
002, 45, 65−74.
57) Stosnach, H. Anal. Sci. 2005, 21, 873−876.
58) Klockenkamper, R.; von Bohlen, A. X-Ray Spectrom. 1996, 25,
1
(
(
56−162.
59) Cataldo, F. J. Radioanal. Nucl. Chem. 2012, 293, 119−126.
60) Telgmann, L.; Holtkamp, M.; Kunnemeyer, J.; Gelhard, C.;
Hartmann, M.; Klose, A.; Sperling, M.; Karst, U. Metallomics 2011, 3,
1
(
(
(
035−1040.
61) Stosnach, H. Spectrochim. Acta B 2010, 65, 859−863.
62) Stosnach, H.; Mages, M. Spectrochim. Acta B 2009, 64, 354−356.
63) Shaw, B. J.; Semin, D. J.; Rider, M. E.; Beebe, M. R. J.
Pharamaceut. Biomed. 2012, 63, 151−159.
3
938
dx.doi.org/10.1021/ac5000812 | Anal. Chem. 2014, 86, 3931−3938