Electrospray ionization-ion mobility spectrometry: a rapid
analytical method for aqueous nitrate and nitrite analysis
Prabha Dwivedi,ab Laura M. Matz,a David A. Atkinson†c and Herbert H. Hill, Jr*ab
a Department of Chemistry, Washington State University, Pullman, WA 99164-4630, USA.
E-mail: [email protected]; Fax: 509-335-8867; Tel: 509-335-5648
b Center for Multiphase Environmental Research, Washington State University, Pullman, WA
99164, USA
c Chemistry Department, Idaho National Engineering and Environmental Laboratory, Idaho
Falls, Idaho 83415, USA
Received 11th September 2003, Accepted 15th December 2003
First published as an Advance Article on the web 12th January 2004
This paper reports the first example of electrospray ionization (ESI) for the separation and detection of anions in aqueous
solutions by ion mobility spectrometry (IMS). Standard solutions of arsenate, phosphate, sulfate, nitrate, nitrite, chloride,
formate, and acetate were analyzed using ESI-IMS and distinct peak patterns and reduced mobility constants (K0) were
observed for respective anions. Real world water samples were analyzed for nitrate and nitrite to determine the
feasibility of using ESI-IMS as a rapid analytical method for monitoring nitrate and nitrite in water systems. The data
showed satisfactory correlation between the measured value ( ~ 0.16 ppm) and the reported maximum nitrate-nitrogen
concentration (0.2 ppm) found in a local drinking water system. For on-site measurement applications, direct sample
introduction and air as an alternate drift gas to nitrogen were evaluated. The identities of the nitrite and nitrate mobility
peaks were verified by comparison of reduced mobility constants with mass identified nitrate and nitrite ions reported in
literature. In the mixing ratio, a linear dynamic range of 3 orders of magnitude and instrument detection limits of 10 ppb
for nitrate and 40 ppb for nitrite were obtained. The calibration curves showed r2 value of 0.98 and slope of 0.06 for
nitrate and r2 value of 0.99 and slope of 0.11 for nitrite.
DOI: 10.1039/b311098b
Introduction
Ion mobility spectrometry (IMS) with electrospray ionization was
first reported for ions such as peptides and other non-volatile
compounds from aqueous solutions in the 1980’s by Shumate and
Hill.1 Later, in 1994, Wittmer et al. demonstrated that whole
proteins could be electrosprayed under atmospheric conditions and
their multi-charged states separated by IMS.2 Other applications of
ESI-IMS have included the determination of polar and non-volatile
organic compounds such as pesticides in aqueous samples.
Recently, Dion et al., have demonstrated that electrospray ionization with ion mobility spectrometry (ESI-IMS) could be used to
detect inorganic cations such as metal ions from aqueous samples.3
The potential for ESI/IMS to determine inorganic ionic species in
aqueous samples opened a new and potentially huge application
area for IMS. Because IMS has a resolving power significantly
greater than that of liquid chromatography, it may be possible to
perform similar analyses to those requiring ion chromatographs but
in only a fraction of the time and with higher resolution and
sensitivity. In this paper, selected inorganic anions were used to
characterize the response of an ion mobility spectrometer. Nitrate
and nitrite ions were selected as the primary test ions because their
gas phase mobilities had previously been measured as product ions
from explosive vapours.4
Past studies have suggested the involvement of nitrate and nitrite
in the pathogenesis of metheamoglomia, infertility, cancer, tumor,
still birth in livestock, through a mechanism involving reduction of
nitrate to nitrite and subsequent formation of potentially mutagenic
nitroso-compounds.5–10 Regular sampling of drinking water systems is necessary to prevent health risks from these contaminants.
Excessive use of fertilizers (artificial or natural), biodegradation of
nitrogenous organic matter, and waste streams from industries
(explosives, pharmaceuticals, food processing) are the main
sources of nitrate–nitrite contamination in natural water re† Current address: Pacific Northwest National Laboratory, (PNNL), 902
Battelle Blvd., Richland, WA 99352, USA.
sources.11 For these reasons, a rapid and sensitive method capable
of determining low levels of nitrate and nitrite directly from
aqueous samples is desired.
In routine water quality analysis, determination of nitrate and
nitrite is achieved by traditional analytical methods, including
spectrophotometry, cadmium reduction, and ion chromatography.12 Traditional methods are limited for on-site field measurements due to their high cost and maintenance, interference from
matrix effects, low sensitivity and/or extensive sample preparation
required prior to analysis.13 For today’s environmental analytical
applications, methods that are low cost, fast, accurate, sensitive,
and portable enough for field applications are desirable. In addition,
these analytical methods should require minimum instrumental
maintenance and sample preparation procedures. When compared
to methods currently used for nitrate–nitrite analysis, ESI-IMS
method has an analysis time advantage over traditional methods
since separation by IMS is achieved in milliseconds. Furthermore,
the ESI-IMS instrument is low cost, easy to handle and requires
minimal maintenance. This manuscript reports the results and
conclusions of using ESI-IMS for the direct detection of anions
such as nitrate and nitrite ions in natural water samples.
1.0 Experimental
1.1 Chemicals, solvents and samples
The chemicals used in this study were sodium nitrate (NaNO3),
sodium nitrite (NaNO2), calcium nitrate (Ca(NO3)2), potassium
nitrite (KNO2), formic acid (HCOOH), acetic acid (CH3COOH),
sodium chloride (NaCl), sodium arsenate (Na2HAsO4.7H2O),
sodium phosphate monobasic (NaH2PO4.H2O), and magnesium
sulfate 7-hydrate (MgSO4.7H2O) purchased from Aldrich Chemical Company (Milwaukee, WI). Solvents used were HPLC grade
methanol (CH3OH), acetonitrile (CH3CN) and water (H2O), and
were purchased from J.T. Baker (Phillipsburg, NJ). Aqueous
environmental water samples were collected from a river, a creek
and a household water tap located in Palouse region at the state of
Washington, USA.
This journal is © The Royal Society of Chemistry 2004
Analyst, 2004, 129, 139–144
139
1.2 Instrumentation
The water cooled electrospray ionization source used in this
research was developed in our laboratory and has been described in
detail elsewhere.2 A 73 mm inner diameter silica capillary tube and
a dual port syringe pump (Brownlee Labs. Micro Gradient 6 port
system) were used for solvent delivery to the ESI needle. Injections
were performed with a six port injection valve (Valco Industries,
Houston, TX) equipped with a 50 mL external injection loop. The
ESI solvent used was a solution of 90% methanol and 10% water.
All salt solutions were made in this solvent unless otherwise
mentioned. The solvent flow rate was maintained at 5 mL min21
throughout the study. An electric potential of 210.0 kV was
applied to the electrospray needle. Overheating of the ESI needle
was prevented by water-cooled nitrogen gas flowing over the
needle at a flow rate of 50 mL min21. Water-cooling the ESI source
was required because the electrospray unit was kept in close
proximity (less than 1 cm) to the high temperature (200–250 °C)
IMS unit. Cooling of the ESI needle was necessary to avoid solvent
evaporation inside the needle before the sample was electrosprayed
for better performance.
The IMS tube, used in this study was 23.6 cm in total length, with
a 9.4 cm desolvation region and a 14.2 cm drift region. The IMS
drift tube was maintained at 240 °C temperature. A Bradbury–
Nielsen type ion gate14 separated the two regions and was closed by
applying an electric potential of +41 V at adjacent wires resulting
in an orthogonal field of ~ 960 V cm21. The ion gate was at an
electric potential of 25.26 or 25.40 kV when the gate was open.
The target screen (first ring of the desolvation region) was
maintained at a constant electric potential of 27.00 or 27.14 kV
resulting in an electric field of 370 and 380 V cm21 in the drift
region. The stacked stainless steel rings were separated from each
other by insulating alumina rings and were connected externally by
resistors creating a potential gradient to guide the ions towards the
collector plate. A counter current flow of heated drift gas (flow rate
= 900 mL min21) was introduced from the back of the Faraday
plate to facilitate desolvation of ions and sweeping out any neutral
molecules.
A Keithley 427 amplifier (Keithley Instruments, Cleveland, OH)
was utilized for signal amplification (109). A Labview based data
acquisition system (DAQ) from National Instruments, (Houston,
TX) was used for signal acquisition. The gating electronics is
described in detail elsewhere.2 Two high voltage supply boxes
obtained from Bertan (Hicksville, New York) were used for
constant voltage supply to the ESI and IMS units.
1.3 Calculation
All reduced mobility values (K0) were calculated using the equation
shown below.15
where, l is the length of the drift region (14.2 cm), td is the
experimentally determined drift time of the ion (s), E is the electric
field in the IMS drift region (370–380 V cm21), T is the
temperature of the drift tube (513 K) and P is atmospheric pressure
(694–703 Torr).
2.0 Results and discussion
2.1 Identification of solvent ions
Previously, in our lab, ESI conditions in negative mode were
investigated with an IMS interfaced to a 150 QC ABB Extrel
quadrupole mass spectrometer (QMS).16 Solvent ion peaks observed by IMS for 9:1 (v/v) methanol–water solution (ESI
condition) were mass identified using the QMS. The presence of
nitrite, chloride, and nitrate ions in the solvent was suggested to be
due to impurities present in the solvent. When Asbury and Hill
140
Analyst, 2004, 129, 139–144
increased the percentage of water from 10% to 40% in the ESI
solvent (9:1, v/v) methanol–water solutions the total ion current
(TIC) decreased. Water percentages higher than 40% water (v/v) in
the ESI solvent lead to ionization due to corona discharge instead of
electrospray ionization.16 The onset of corona discharge above the
40% water level correlated with a sharp increase in the TIC and
intensities of the nitrate and nitrite ion peaks while, intensities of
the chloride, formate and acetate ion peaks were reduced. Thus,
initially a solution of 90% methanol and 10% water was chosen as
the ESI solvent.
Fig. 1 illustrates the electrospray ionization-ion mobility spectrum of the ESI solvent. All measurements were made at conditions
outlined in Table 1. Five prominent peaks observed in the
experiment and were identified, by comparison with the previous
ion mobility mass spectrometry (IMMS) study, as: chloride at 6.29
ms drift time (Peak number 1), nitrite at 6.78 ms drift time (Peak
number 2), formate at 7.01 ms drift time (Peak number 3), nitrate at
7.54 ms drift time (Peak number 4) and acetate at 8.01 ms drift time
(Peak number 5) at 370 V cm21 electric field. At 380 V cm21, the
drift time values were 6.02, 6.53, 6.72, 7.31, and 7.73 ms for
chloride, nitrite, formate, nitrate, and acetate ions respectively. The
identity of each peak was confirmed by cross referencing the K0
values measured in the current study to the K0 values obtained in
Fig. 1 Electrospray ionization-ion mobility spectrum of the ESI solvent:
The figure shows the ion mobility spectrum of different solvent ions
observed when 9:1 mixture of methanol and water was used as the
electrospray solvent. The inset in the spectrum shows the peak numbers
assigned to each peak in column 1 (maintained throughout the manuscript
except in Fig. 2), the drift times of respective ion in column 2, and the
identity of each ion in column 3.
Table 1 Experimental parameters and conditions
Experimental parameter
Conditions
Diameter of ESI capillary
ESI flow rate
Voltage on ESI needle
Drift tube voltage
Voltage on the ion gate
Electric field
Mode of operation
Desolvation region length
Drift region length
Drift gas
Drift gas flow rates
Cooling gas
Cooling gas flow rate
IMS drift tube temperature
Atmospheric pressure (Pullman, WA)
73 mm
5 mL min21
210 kV
27.00/27.14 kV
25.26/25.40 kV
2370/2380 V cm21
Negative mode
9.4 cm
14.2 cm
Nitrogen
900 mL min21
Nitrogen
50 mL min21
240 °C
694–703 Torr
nitrogen as the drift gas by other researchers.17–22 Though the
literature K0 values referenced in Table 2 were obtained at different
temperature and atmospheric pressure conditions than those used in
the experiments, they are comparable to the reported values
because all K0 values are normalized to standard temperature and
pressure. Table 2 shows the identity of each solvent ion peak in
column A, the peak number assigned to each peak in column B, the
molecular weight of the anions observed in column C, the drift time
of respective anion with nitrogen as drift gas at 370 V cm21 electric
field in column D, the drift time of respective anion with nitrogen
as drift gas at 380 V cm21 electric field in column E, the drift time
of respective anion with air as drift gas at 370 V cm21 in column F,
experimental K0 values in column G, K0 values from other
references in column H and reference numbers in column I.
2.2 Ion mobility spectra of anions
Fig. 2 shows the ion mobility spectra obtained from standard
solutions (mixing ratio: 15 ppm) of sodium arsenate, magnesium
sulfate, sodium phosphate, sodium nitrate, sodium nitrite, trinitrotoluene (TNT), and a mixture of TNT and sodium sulfate.
Trinitrotoluene was used as a marker since the K0 value for TNT is
well established. For each anion, distinct peak pattern was observed
and the experimental K0 values for nitrate, nitrite, and TNT
matched well with the values reported in literature.23 Peaks
observed for sulfate and TNT solutions were reproduced in the
mixture of TNT and sulfate shown in peaks marked (*) of Table 3.
Table 3 lists the different anions analyzed with their respective drift
times and measured K0 values at 370 V cm21 electric field.
Multiple prominent peaks were observed for arsenate, sulfate and
phosphate. The presence of multiple peaks for arsenate, sulfate, and
phosphate could be attributed to the existence of the anions in
different oxidation state (viz. arsenate/arsenite) at atmospheric
conditions and/or different degrees of anion hydration. The peak
patterns can be used as anion identification since they were unique
and reproducible for each anion. However, mass identification of
the peaks will be necessary to identify these anions. Future work
will be focused on mass identification of the unidentified peaks.
From the nitrate–nitrite data alone, this study demonstrates that
there is potentially a huge application area for ESI-IMS to
determine inorganic ionic species in aqueous samples.
mixing ratio of nitrate ion in both nitrate salt solutions was 15 ppm.
Nitrite ion mixing ratios in both the nitrite salt solutions was 10
ppm. In Fig. 3 (spectra b and c) the drift time of the nitrate ion (peak
number 4) was 7.35 ± 0.05 ms when either of the salt solutions was
analyzed. The reduced mobility values (K0) measured for nitrate
ion, in both cases were 2.47 ± 0.02 cm2 V21 s21, showing
satisfactory consistency with the literature value listed in Table 2.
Similarly, ion mobility spectra (e) and (f) of Fig. 3 showed that the
presence of different counter ions in the nitrite salt solutions had no
effect either on peak intensity or K0 value (2.75 ± 0.02 cm2 V21
s21) for the nitrite ion (peak number 2, drift time 6.72 ± 0.05 ms).
Intensity of the nitrate and nitrite ion peaks and the drift times of the
nitrate and nitrite ions were reproducible with a deviation of 1–2%
on a day-to-day basis. An increase in the intensity of the nitrate and
nitrite ion peaks with the introduction of respective standards
confirmed the peak identities. Also, Fig. 3 demonstrated that the
nitrate and nitrite responses were independent of the cationic
composition of the sample.
2.3 Nitrate and nitrite
Fig. 3 shows the mobility spectra of: (a) solvent only, (b) NaNO3,
(c) Ca(NO3)2, (d) solvent only, (e) KNO2, and (f) NaNO2 at 380 V
cm21 electric field. Ion mobility spectra of the ESI solvent (spectra
a and d) are shown twice for convenient comparison of the
background spectrum with that of the salt solutions. All spectra in
Fig. 3 are shown at same scale except spectrum (a) where the y-axis
was expanded for better observation of solvent ion peaks. The
Fig. 2 Ion mobility spectra of different anions. The figure shows the
different peak patterns observed for different anions. Distinct peak(s)
observed for each anion have been numbered and their identity given in
Table 3. Each anion was prepared in the ESI solvent at a mixing ratio of 10
ppm.
Table 2 ESI-IMS of electrospray solvent: Table 2 shows the identity of each solvent ion peak in column A, the peak number assigned to each peak in column
B, the molecular weight of the anions observed in column C, the drift time of respective anion with nitrogen as drift gas at 370 V cm21 electric field in column
D, the drift time of respective anion with nitrogen as drift gas at 380 V cm21 electric field in column E, the drift time of respective anion with air as drift
gas at 370 V cm21 in column F, experimental K0 values in column G, K0 values from other references in column H and reference numbers in column I
B
C
D 370
V cm21
E 380
V cm21
F
G
H
I
Solvent ions
Peak #
Molecular
weight/u
Drift time
(±0.05 ms)
(nitrogen)
Drift time
(±0.05 ms)
(nitrogen)
Drift time
(±0.05ms)
(air)
Measured K0
value (N2)/
cm2 V21 s21
Literature K0
value/
cm2 V21 s21
Reference #
Chloride (Cl2)
1
35.45
6.29
6.02
6.27
2.99 ± 0.02
2.94 3.01
Nitrite (NO22)
2
46.01
6.78
6.53
6.76
2.75 ± 0.02
2.70 2.76
Formate (HCOO2)
Nitrate (NO32)
3
4
45.02
62.00
7.01
7.54
6.72
7.32
6.96
7.53
2.66 ± 0.02
2.47 ± 0.02
NA
2.46 2.48
Acetate (CH3COO2) 5
59.04
8.01
7.73
7.99
2.32 ± 0.02
NA
A
Spangler and Carrico18;
Griffin et al.20
Spangler and Lawless19;
Carr21
NA
Lawrence and
Neudrofl17; Asselin22
NA
Analyst, 2004, 129, 139–144
141
3.0 Parametric investigation
3.1 Electrospray solvent
Fig. 4a and b are ion mobility spectra obtained for (a) ESI solvent
only and (b) a 9:1 (v/v) solutions of acetonitrile and water at 370 V
cm21. In both cases all five solvent ions peaks were observed.
Comparison of spectrum (a) with spectrum (b) shows that, when
acetonitrile was used as the solvent instead of methanol, intensities
for the formate and acetate ion peaks were reduced, however, the
intensity of the nitrite ion peak increased. There was no significant
Table 3 ESI-IMS of different anions: with their drift times and
experimental K0 values: Table 3 shows the peak number assigned to each
peak in column 1, the identity of anion peak in column 2, the drift time of
respective anion with nitrogen as drift gas in column 3, experimental K0
values in column 4. Previously mass identified molecular anions are marked
(*)
Peak number
ESI anions from
Drift time
(±0.05 ms)
(1)
(2)
(3)
(4)
(5)
(6)
(7)
(8)
(9)
(10)
(11)
(12)
(13)
(14)
Chloride*
Nitrite*
Formate*
Nitrate*
Acetate*
Arsenate
Arsenate
Sulfate
Sulfate
Sulfate
TNT*
Phosphate
Phosphate
Phosphate
6.28
6.77
7.02
7.52
7.96
9.01
9.71
8.52
9.11
10.34
12.31
8.65
9.81
10.94
Measured K0
Values (±0.02
cm2 V21 s21)
2.99
2.76
2.67
2.49
2.32
2.08
1.92
2.19
2.05
1.81
1.52
2.16
1.91
1.71
Fig. 3 Effect of different nitrate–nitrite salts on the nitrate–nitrite peak
response. In the figure are shown the mobility spectra of: (b) NaNO3 and (c)
Ca(NO3)2, (e) KNO2, (f) NaNO2. The ion mobility spectra of the ESI
solvent (part a. and d.) are shown twice for convenient comparison of the
background spectrum with that of the salt solutions. All spectra in Fig. 3 are
shown at same scale except spectrum (a) where y-axis is expanded for better
observation of solvent ion peaks.
142
Analyst, 2004, 129, 139–144
change in the nitrate ion peak height. Fig. 4c and d illustrate the ion
mobility spectra of 1 ppm sodium nitrate solution in (c) the ESI
solvent and (d) the 9:1 (v/v) acetonitrile/water solution. Nitrate was
successfully detected when acetonitrile was used as solvent instead
of methanol. However, since no advantage was observed using
acetonitrile instead of methanol as the ESI solvent, methanol was
chosen as the solvent for direct analysis.
3.2 Solvent modifiers
Fig. 5 illustrates the effect of various solvent modifiers on the
response of the nitrate and nitrite ions at 380 V cm21. Solvent
modifiers studied were 5% acetic acid, 1% formic acid and 1 mM
sodium chloride in the 9:1 methanol–water solvent. In Fig. 5 ion
mobility spectra of (a) the ESI solvent, (b) 5% acetic acid in the ESI
solvent, (c) 1% formic acid in the ESI solvent and (d) 1 mM sodium
chloride in the ESI solvent are shown. Comparison of the solvent
ion peaks of the ESI solvent without modifiers (spectrum 5a) with
solvent ion peaks of the ESI solvent with modifiers (spectra 5b, 5c
and 5d) showed that the addition of modifiers in the ESI solvent
suppressed the formate ion peak intensity (peak #3) but increased
the nitrate ion peak intensity (peak #4). Although, the nitrite ion
peak intensity (peak #2) was suppressed with acetic acid as solvent
modifier (Fig. 5b) the intensity of this peak increased with the
addition of formic acid and sodium chloride. While it may be
possible to increase the sensitivity of ESI-IMS towards nitrate and
nitrite by addition of selected modifiers, it appears that the
background response of nitrate and nitrite will also be increased.
Thus, more extensive investigations are needed before the optimal
modifier conditions can be determined.
3.3 Sample introduction
For on-site measurement applications of aqueous environmental
samples, minimum sample preparation is desired. Since corona
discharge ionization occurred with high (greater than 40%) water
concentration in the ESI solvent,16 experiments were conducted
Fig. 4 Effect of organic solvent on solvent ion response. The figure shows
the ion mobility spectra observed of (a) 90:10 (v/v) solutions of methanol
and water and (b) 90:10 (v/v) solutions of acetonitrile and water. Part (c) and
(d) illustrates the ion mobility spectra of 1 ppm sodium nitrate solution in (c)
90:10 (v/v) solutions of methanol and water and (d) 90:10 (v/v) solution of
acetonitrile and water.
where 9:1 (v/v) methanol–water solution was used as the carrier
solvent and injections of 100% water solutions (with and without
nitrate ions) were made. The ion mobility spectra thus obtained
were then compared to ion mobility spectra obtained with injections
of 9:1 (v/v) methanol–water solutions (with and without nitrate
ion). Fig. 6 represents a series of ion mobility spectra at 380 V cm21
with injections of (a) 9:1 (v/v) methanol–water solution, (b) 100%
water and (c) 20 ppm nitrate in 100% water. Comparison of
spectrum (a) with spectrum (b) of Fig. 6 suggested that corona
discharge could be avoided if the carrier solvent used is 9:1 (v/v)
methanol–water solution and only injections of 100% water are
made. Since, the analysis time required by this method is in
milliseconds, amount of methanol used for a run is negligible (4.5
mL min21). Solutions of nitrate salts in 100% water with varying
nitrate ion concentrations (0.1 ppm to 25 ppm) were analyzed and
direct correlation was observed between the nitrate ion concentration and the IMS response. Fig. 6c) represents the ion mobility
spectrum of a solution of 20 ppm nitrate in 100% water illustrating
the increase in the response of the nitrate ion peak (number 4) with
increase in the nitrate concentration.
3.4 Drift gas
Because air is a readily available gas compared to nitrogen for onsite measurement applications, experiments were performed to
evaluate the feasibility of using air as the drift gas instead of
nitrogen at an electric field of 380 V cm21. Fig. 6d is the ion
mobility spectra of solvent ions observed of 9:1 methanol–water
solution when air was used as the drift gas and 6e is the ion mobility
spectra of 20 ppm nitrate solution in 9:1 methanol–water (v/v) with
air as the drift gas. As expected, because air contains 79% nitrogen,
the same solvent ions were formed by the electrospray process
when either drift gas was used. Table 2 reports the drift times of
solvent ions in air and in nitrogen as the drift gas. From the results
of the experiments in Fig. 6, it was concluded that direct sample
introduction and use of air as the drift gas is possible with ESI-IMS.
These results were promising in establishing ESI-IMS as a simple,
Fig. 5 Effect of solvent modifiers in the ESI solvent [9:1 (v/v) solutions of
methanol and water] on the solvent ions. The figure shows the ion mobility
spectrum obtained of (a) the ESI solvent, (b) 5% acetic acid in ESI solvent,
(c) 1% formic acid in the ESI solvent and (d) 1% sodium chloride (1M) in
the ESI solvent.
rapid, sensitive, and quantitative method for direct aqueous nitrate
and nitrite analysis.
4.0 Analytical figures of merit
Solutions of sodium nitrate and sodium nitrite with nitrate and
nitrite ion mixing ratios ranging from 1 ppb to 30 ppm in the ESI
solvent, were used to obtain calibration curves (ESI solvent: 9:1
methanol–water solution, ESI solvent flow rate: 5 mL min21, ESI
voltage: 210.0 kV, drift gas flow rate 900 mL min21, cooling gas
flow rate 50 mL min21, IMS tube temperature: 240 °C, IMS target
screen voltage: 27.00 kV, IMS gate voltage: 25.26 kV). Ion
current intensity of the nitrate and nitrite ion peak was plotted
against concentration of nitrate and nitrite ion in the analyzed
solution and equations for calibration curves were developed by
least-squares technique. The calibration curves for nitrate and
nitrite showed linear relationship between the analyte concentration
and IMS response, over the concentration range studied with an
instrumental detection limit of 10 ppb for nitrate and 40 ppb for
nitrite ion concentration. The limit of detection for each ion was
calculated at three times the signal to chemical noise ratio. The
calibration curves showed r2 value of 0.98 and slope of 0.06 for
nitrate and r2 value of 0.99 and slope of 0.11 for nitrite. Each point
on the calibration curve was an average of three replicate
measurements.
5.0 Analysis of environmental water samples
To demonstrate the potential of ESI-IMS for the determination of
nitrate–nitrogen in real world samples, environmental water
samples were collected and analyzed. Collected samples were (A)
tap water, (B) river water, (C) creek water and (D) creek water
collected 6 miles down stream from site C. All samples were
filtered with a 0.2 mm filter and methanol was added to each to
Fig. 6 Effect of drift gas on IMS response. Ion mobility spectra illustrating
the effect of 100% water injections and change in drift gas on the solvent
ions and nitrate detection. The figure shows ion mobility spectra obtained of
injections of (a) 9:1 (v/v) methanol–water solution with nitrogen as drift
gas, (b) 100% water with nitrogen as drift gas, (c) 20 ppm nitrate in 100%
water with nitrogen as drift gas, (d) 9:1 (v/v) methanol–water solution with
air as drift gas and (e) 20 ppm nitrate in 9:1 (v/v) methanol–water solution
with air as drift gas.
Analyst, 2004, 129, 139–144
143
make a final solution of 9:1 methanol–water (v/v) compositions.
Thus, based on the instrumental detection limits given above, the
sample detection limit for nitrate was 0.01 ppm and 0.04 ppm for
nitrite, well below the limits required for monitoring nitrate/nitrite
in natural water samples.
Three replicate measurements for each sample were made by
ESI-IMS. IMS spectra of the above water samples at 370 V cm21
are shown in Fig. 7. The intensity for both nitrate and nitrite peaks
were seen to increase over the background with the introduction of
the collected samples, indicating that both nitrate and nitrite ions
were present in the samples. The average intensity of the nitrate ion
peak for each sample was determined from the ion mobility
spectrum. Then, using the calibration curve and correcting for
dilution, nitrate-nitrogen mixing ratio in each sample was determined. Nitrate-nitrogen concentration was the least in drinking
water ( ~ 0.16 ppm) followed by river water ( ~ 0.36 ppm) creek
water ( ~ 0.51 ppm) and creek water 6 miles down stream from the
previous sample site ( ~ 0.72 ppm). These data are consistent with
nitrate runoffs from heavily fertilized agricultural fields.24,25
Experimentally measured nitrate-nitrogen mixing ratio in drinking
water (0.16 ppm) compared well with that reported for the
maximum nitrate-nitrogen concentration found in the local water
system (maximum 0.2 ppm)26,27 well below the maximum
permissible limit of nitrate-nitrogen concentration in drinking
water (10 ppm) set by the US Environmental Protection Agency
(EPA).
Fig. 7 Ion mobility spectra of different real world water samples. These
spectra demonstrate sensitive, selective and rapid response for nitrate and
nitrite in various real water samples.
144
Analyst, 2004, 129, 139–144
Conclusions
ESI-APIMS appears suitable for the rapid determination of nitrate
and nitrite in real water samples with instrumental detection limits
as low as 10 ppb for nitrate and 40 ppb for nitrite. Distinguishing
peak patterns observed for arsenate, phosphate, sulfate, chloride,
formate, and acetate point to the potential ESI-IMS anion analysis
in general. The ability to make direct water injection and to use air
as the drift gas lead to the conclusion that a portable ESI-IMS
instrument could be constructed. In addition, the high resolution of
the separation along with its response to many anions with little
interference indicates that ESI-IMS may serve as an alternate
method to ion chromatography for the determination of a variety of
anions in aqueous samples.
Acknowledgements
The authors thank the Idaho National Engineering and Environmental Laboratory, (INEEL) Idaho, and the US Army Research
Office for partial financial support of this project.
References
1 C. B. Shumate and H. H. Hill, Jr, Anal. Chem., 1989, 61, 601–606.
2 D. Wittmer, B. K. Luckenbill, H. H. Hill and Y. H. Chen, Anal. Chem.,
1994, 66, 2348–2355.
3 H. M. Dion, L. K. Ackerman and H. H. Hill, Talanta, 2002, 57,
1161–1171.
4 G. A. Eiceman, Crit. Rev. Anal. Chem., 1991, 22, 17–36.
5 J. H. Weisburger, Nutr. Cancer, 1986, 8, 47–62.
6 T. L. Carson, Vet. Clin. North. Am. Food. Anim. Pract., 2000, 16,
455–464.
7 E. Graham, Clin. Biochem., 1998, 31, 195–220.
8 D. Connolly, Anal. Chim. Acta, 2001, 441, 53–62.
9 W. H. O. Tech. Rep. Ser., 1995, 859, pp. 1–54.
10 C. Oldreive and C. Rice-Evans, Free Radical Res., 2001, 35,
215–231.
11 B. M. Peyton and J. N. Petersen, Water Res., 2001, 35, 4237–4242.
12 A. E. Greenberg, L. S. Clesceri and M. A. Eaton, American Public
Health Association, 18th edn. 1992, pp. 1–4.
13 A. A. Okemgbo, H. H. Hill, S. G. Metcalf and M. Bachelor, J.
Microcolumn Sep., 2000, 12, 48–56.
14 N. E. Bradbury and N. R. Neilson, Phys. Rev., 1936, 49, 388.
15 C. Wu, W. F. Siems, G. R. Asbury and H. H. Hill, Anal. Chem., 1998,
70, 4929–4938.
16 G. R. Asbury and H. H. Hill, Int. J. Ion Mobility Spectrom., 1999, 2,
1–8.
17 P. A. Lawrence and P. Neudrofl, Anal. Chem., 1988, 60(2), 104–109.
18 G. E. Spangler and J. P. Carrico, Int. J. Mass Spectrom. Ion Processes,
1983, 267.
19 G. E. Spangler and P. A. Lawless, Anal. Chem., 1978, 50(7), 884.
20 G. W. Griffin, I. Dzidic, D. I. Carroll, R. N. Stillwell and E. C. Horning,
Anal. Chem., 1973, 45(7), 1204.
21 T. W. Carr, Anal. Chem., 1979, 51(6), 705.
22 M. Asselin, New Concepts Symposium and Workshop on Detection and
Identification of Explosives., 1978, p. 177.
23 J. I. Baumbach and J. Stach, 6th International Conference for Ion
Mobility Spectrometry, Bastei, Germany, 1997.
24 R. F. Follett and J. L. Hatfield, Sci. World J., 2001, 1, 920–926.
25 R. E. Casey and S. J. Klaine, J. Environ. Qual., 2001, 30, 1720–1731.
26 Palouse Basin Aquifer Commmittee Report, University of Idaho,
2000.
27 Columbia Plateau Aquifer System Annual Report, 2000.