Development of an Atmospheric Pressure Ion Mobility Spectrometer

Aug 22, 2013 - Recently developed ion mobility mass spectrometer is described. The instrument is based on a drift tube ion mobility spectrometer and a...
0 downloads 0 Views 2MB Size
Download PDF
Article pubs.acs.org/ac

Development of an Atmospheric Pressure Ion Mobility Spectrometer−Mass Spectrometer with an Orthogonal Acceleration Electrostatic Sector TOF Mass Analyzer Alexey A. Sysoev,*,†,‡ Denis M. Chernyshev,†,‡ Sergey S. Poteshin,† Alexander V. Karpov,† Oleg I. Fomin,‡ and Alexander A. Sysoev† †

National Research Nuclear University MEPhI, 115409, Kashirskoe shosse 31, Moscow, Russian Federation Linantec Ltd, 115409, Kashirskoe shosse 31, Moscow, Russian Federation

‡

S Supporting Information *

ABSTRACT: Recently developed ion mobility mass spectrometer is described. The instrument is based on a drift tube ion mobility spectrometer and an orthogonal acceleration electrostatic sector timeof-flight mass analyzer. Data collection is performed using a specially developed fast ADC-based recorder that allows real-time data integration in an interval between 3 and 100 s. Primary tests were done with positive ion electrospray. The tests have shown obtaining 100 ion mobility resolving power and 2000 mass resolving power. Obtained for 2,6-di-tert-butylpyridine in electrosprayed liquid samples during 100 s analysis and full IMS/MS data collection mode were 4 nM relative limits of detection and a 1 pg absolute limit of detection (S/N=3). Characteristic ion mobility/mass distributions were recorded for selected antibiotics, including amoxicillin, ampicillin, lomefloxacin, and ofloxacin. At studied conditions, lomefloxacin forms only a protonated molecule-producing reduced ion mobility peak at 1.082 cm2/(V s). Both amoxicillin and ampicillin produce [M + H]+, [M + CH3OH + H]+, and [M + CH3CN + H]+. Amoxicillin shows two peaks at 0.909 cm2/(V s) and 0.905 cm2/(V s). Ampicillin shows one peak at 0.945 cm2/(V s). Intensity of protonated methanol containing cluster for both ampicillin and amoxicillin has a clear tendency to rise with sample keeping time. Ofloxacin produces two peaks in the ion mobility distribution. A lower ion mobility peak at 1.051 cm2/(V s) is shown to be formed by [M + H]+ ions. A higher ion mobility peak appearing for samples kept more than 48 h is shown to be formed by both [M + H]+ ion and a component identified as the [M + 2H + M]+2 cluster. The cluster probably partly dissociates in the interface producing the [M + H]+ ion.

I

ion mobility2 is used to compare experimental data obtained at different temperatures and pressures:

on mobility spectrometry is based on separation of ions within a homogeneous electric field in a counter stream of drift gas flow at the atmospheric or reduced pressures. Ion mobility is an individual property of ions. It depends mainly upon collision cross sections of sample ions and drift gas molecules and external conditions, such as temperature and pressure. The mobility is defined as the ratio of the ion velocity in a drift gas to the magnitude of the electric field that pushes the ion. In the low-field approximation, the velocity is linearly proportional to the electric field strength. In this case, the ion mobility can be calculated as

L K= Et

⎛ 273 ⎞⎛ P ⎞ ⎟⎜ ⎟ K0 = K ⎜ ⎝ T ⎠⎝ 760 ⎠

where T is the temperature (in K), and P is the pressure (in Torr). Selectivity of ion mobility spectrometry is mainly determined through single peak mobility resolving power as RK =

t K = d δK δtd

(3)

where td is the drift time, δtd is the full width of the drift time peak at half-maximum, K is the ion mobility, δK is the full width of the ion mobility peak at half-maximum. The resolving power of ion mobility spectrometry reflects a minimal difference in the

(1)

where t is the drift time (m/s), L is the length of the drift region (m/s), and E is the strength of the electric field (V cm−1). The ion mobility is an experimentally measured drift time of ions within the drift area with a homogeneous electric field. Some equations for theoretical estimation of the ion mobility were given by Revercomb and Masson.1 The reduced © 2013 American Chemical Society

(2)

Received: April 22, 2013 Accepted: August 22, 2013 Published: August 22, 2013 9003

dx.doi.org/10.1021/ac401191k | Anal. Chem. 2013, 85, 9003−9012

Analytical Chemistry

Article

Figure 1. Schematic picture showing an effect of TOF analyzer geometry on sampling duty-cycle at a particular m/z. lp is the detectable length of ion beam entering orthogonal accelerator sampled to create an ion packet, and lb is the distance traveled in the beam direction by the ions of the same m/z during the period of TOF sampling repetitions.

aperture of ion mobility spectrometers ranges from 99% to 99.9%.23 The loss of ions during injection usually lies between 99.6% and 99.9%.24 Higher sensitivity was achieved by increasing the efficiency of ion injection and accumulation with the use of quadrupole and octopole ion traps25−27 and electrodynamic ion funnels.28 The sampling efficiency was also improved up to 50% by applying various multiplexing approaches based on the Fourier transform29 or the Hadamard transform.30 A dramatic increase in sensitivity was obtained by an approach based on the Hadamard transform incorporating ion trapping, which enhances the signal-to-noise ratio up to 13.31−33 An orthogonal acceleration approach34 is used mainly when combining ion mobility spectrometry with TOF mass spectrometry. The sampling efficiency in the orthogonal accelerator is m/z dependent. In the simplest terms, the sampling duty-cycle of a ion with a particular m/z is approximately equal to the lp/lb ratio (Figure 1), where lp is the detectable length of the ion beam entering the orthogonal accelerator, sampled to create an ion packet; lb is the distance, which the ions of the same m/z travel in the beam direction during the period of TOF sampling repetitions.34 The minimal value of lb is obtained for ions with the greatest m/z values and defined by the distance between the orthogonal accelerator and a detector in the ion beam direction. This value is at least twice as large as lp for the classical mass reflectron and even more higher for multireflection geometries, giving the “geometrical” sampling duty-cycle estimation of 0.1−0.5. It should be noted, that in addition to the sampling factor, the efficiency of the mass analyzer should also be estimated by the product of transmission of ions through all grids and the detection yield.34 At the same time, the “geometrical” sampling duty-cycle close to 1 could be obtained for a linear TOF instrument.34 However, nonefficient energy focusing leads to the ion signal, more widely distributed in time, which decreases the mass resolution. This would also decrease the signal-to-noise ratio. It is promising to apply electrostatic sector TOF mass analyzers in ion mobility mass spectrometry. This combination allows for the achievement of not only more efficient energy focusing but

ion mobilities required for peak resolving within a certain mobility range. A combination of ion mobility spectrometry and mass spectrometry is a powerful tool for two-dimensional (2D) sample separation.3,4 In this case, ion mobility spectrometry expands analytical abilities of mass spectrometry dramatically because of its availability of isomer separation.4−6 In addition, ion mobility spectrometry can be considered as a tool for decreasing chemical noise in analyses of complex mixtures. In general, any kind of mass analyzers can be used in a drift tube ion mobility mass spectrometer: quadrupole mass spectrometers,7−12 ion trap mass spectrometers,13,14 and Fourier transform ion cyclotron resonance mass spectrometers.15 However, for the most part, the time required of mass spectrometer to produce one spectrum is not clearly smaller than the duration of the pulses outputted by the drift tube IMS. Those mass spectrometers cannot analyze the full IMS spectrum unless to allow the passage of only one type of mobility. From this point of view, a combination of drift tube ion mobility mass spectrometry (IMS) with time-of-flight-mass spectrometry (TOF-MS)16−21 allows one to achieve higher sampling efficiency. The separation time, obtained by a TOF mass spectrometer, is 3 orders of magnitude smaller than the ion mobility separation time. Thus, a number of mass spectra can be registered for every single component separated by the IMS. In TOF-MS, the mass resolution of a single peak is determined as RM =

t M = TOF 2δt TOF δM

(4)

where tTOF is the ion time-of-flight, δtTOF is the full width of the time-of-flight peak at half-maximum, M is the ion mass, and δM is the full width of the mass peak at half-maximum. An important property of an ion mobility drift tube, usually associated with its pulsing nature, is how effectively the tube utilizes the ions from a continuous ionization source.22 The transmission of the drift tube IMS is inversely proportional to the squared resolving power of the IMS. Ion loss at the exit 9004

dx.doi.org/10.1021/ac401191k | Anal. Chem. 2013, 85, 9003−9012

Analytical Chemistry

Article

Figure 2. A schematic diagram of the ion mobility spectrometer/mass-spectrometer with the orthogonal acceleration sector TOF mass analyzer.

Ion Mobility Spectrometer. The drift tube ion mobility spectrometer we used in this study is very similar to the instrument described previuosly36,37 and which was earlier described with Faraday cup detector,36 triple quadrupole,37 and Q-TOF38 mass spectrometers. The IMS was periodically heated between the measurements up to 150 °C by a heater placed between the drift tube electrodes and the IMS body. Cooled electrospray is desirable to allow heating the instrument continuously. Pure nitrogen (99.96%) was used as a drift gas, generated by NitroGen N300DR gas station from Peak Scientific Instruments. The gas was introduced into the drift region at 2.0−2.2 l/min flow rate. Approximately 0.5 L/min of the drift gas was pumped by the differential pumping system, while the remaining flow was injected into the drift tube. The humidity was measured by a IVA-6B humidity sensor (Microfor, Russia). The gas flow was distributed between the drift gas and nebulizer gas lines by two flow controllers F201CV-10K-RAD-33-V (Bronkhorst, Netherlands). Liquid sample was injected by a syringe pump 74900−05 (ColeParmer). A membrane pump MPC201E (ILM-VAC GmbH) was connected to the exhaust line. The drift tube pressure was controlled by the DCP 3000 pressure sensor (Vacuubrand, Germany). The temperature of the drift tube was measured by the ATT-2002 sensor (Aktakom, Russia). In order to study the ultimate characteristics of our IMS/ TOF-MS, the optimal values of the gate opening time were found for each experiment. In the resolving power study, the gate opening time of 0.2 ms was found to be an optimal duration. This duration gives the narrowest drift peaks at a reasonable signal intensity. For experiments related with determination of the detection limit and reproducibility, the gate opening time of 0.5 ms was used. This value yields approximately 4-fold higher signal intensity with the resolving power decreased approximately by a factor of 2. Differential Pumping Interface. The main function of the differential pumping system is to transport ion groups with minimal losses and minimal time broadening. The ions should be delivered from the region with the atmospheric pressure, under which the ion mobility spectrometer operates, to a high vacuum region. The differential pumping interface consists of

also higher sampling efficiency, comparable with the linear TOF analyzer. In this paper, we describe a new hybrid instrument, which consists of the drift tube ion mobility spectrometer and an orthogonal acceleration electrostatic sector time-of-flight mass spectrometer. We also discuss the results of the primary tests of the instrument using selected reference compounds analyzed in the positive ESI mode. The ion mobility mass spectral data for the selected antibiotics and the values of reduced ion mobility are presented. The novelty of the described approach is based on combining the drift tube IMS with the orthogonal acceleration electrostatic sector TOF mass analyzer. This approach can potentially allow compactness and better duty cycle to be obtained, compared to existing reflectron-based designs.

■

EXPERIMENTAL SECTION A homemade ion mobility mass spectrometer designed and built in our laboratory is shown in Figure 2. A general view of the instrument is shown in Figure S1 of the Supporting Information. The instrument includes an electrospray ion source, an ion mobility spectrometer, a differential pumping interface, and a time-of-flight mass spectrometer. Preliminary tests of the instrument were done with corona discharge of vaporized methanol solution of 2,6-DtBP and described in the letter.35 Ion Source. Only an electrospray ion source (ESI) was used in this study. The ion source was a commercial Turbo IonSpray (Applied Biosystems Sciex, Canada). Three modifications were made to the ion source. First, the hot gas probe was removed. Second, the original 10 kV high-voltage connector was replaced by a 20 kV (high voltage) connector. It was done because the voltages of 10 kV were applied to the electrospray needle during IMS/MS operation. Third, inside one of the stainless steel needles, a silica capillary (inner diameter 0.1 mm) was inserted in order to obtain a stable spray. This results in a higher sensitivity at reduced sample flow rates. At sample flow rates of 150 μL/h and below, we used the needle with reduced diameter. 9005

dx.doi.org/10.1021/ac401191k | Anal. Chem. 2013, 85, 9003−9012

Analytical Chemistry

Article

two stages, operated at the pressures of 4 and 1 × 10−3 Torr, respectively. A substantial modification to the previous interface design35 was made. This allows for decreasing the peak broadening and increasing the IMS resolving power in the realtime data registration mode. A specially designed radiofrequency skew quadrupole was installed on the second stage (Figure 3a). The idea of skew quadrupole was originally

with discrete dynode ETP 14882 (SGE, Australia) is used as a detector. Instrumental parameters of the IMS/TOFMS are summarized in Table 1. Unless specified otherwise, the conditions listed in the table are used for all experiments described in this work. Table 1. Instrumental Parameters of IMS/TOF-MS part ion source

ion mobility spectrometer

differential pumping interface

TOF mass spectrometer

Figure 3. (a) Schematic representation of the differential pumping interface and (b) detailed picture of a skew quadrupole.

proposed as a compact mass filter.39 The skew quadrupole includes four rods. The opposite rods are inclined at 1 degree relative to each other, with one pair of the converging and a pair of diverging rods (Figure 3b). The rods were electrically biased in pairs at 1−3 V. The use of this skew quadrupole allows us to decrease the ion packet broadening to below 200 μs. This also gives a substantially increased ion transmission. In particular, the 2,6-DtBP peak broadening was decreased from 300 to 120 μs. These results are in good agreement with the numerical simulations done by Simion 7 and COSMOSFlowWorks. This increase in the ion transmission gives a reproducible signal with IMS gate pulse width of 0.2 ms. These modifications, and also a substantial reduction in the broadening of the ion packets within the interface, lead to a higher resolving power. The first stage of the differential pumping interface was pumped by a 6 L/min rotary vane pump (2NVR-5MD Vacuummash, Russia). The second stage was pumped by a 70 L/sec turbo pump TMH/U 071 YPN (Pfeiffer, Germany). The analyzer chamber was pumped by a TMH 262 turbo pump with a pumping speed of 250 L/sec (Pfeiffer, Germany). The rough pumping of both turbo pumps was done by 1 L/sec membrane pump MD4NT (Vacuubrand, Germany). Time-of-Flight Mass Spectrometer. The time-of-flight mass spectrometer used in this study is an electrostatic sector mass analyzer with an orthogonal accelerator and spiral movement of ions (see Figure 2). An idea of the analyzer was described in the paper.35 The secondary electron multiplier

parameter

setting

electrospray needle voltage (V) electrospray counter electrode voltage (V) sample flow rate (μL h−1) flow rate of nebulizer gas (L min−1) desolvation region length (cm)

10000 7000

desolvation region electric field (V cm−1) drift region length (cm) drift region electric field (V cm−1) flow rate of drift gas (L min−1) gate opening time (ms) first skimmer orifice diameter (mm)

298 13.5 298 2.2 0.2−0.5 0.3

first stage pressure (Torr) first skimmer voltage (V) second skimmer orifice diameter (mm) second stage pressure (Torr) second skimmer voltage (V) plate orifice diameter (mm) an inclination angle between skew quadrupole roads (degree) pushing pulse voltage (V)

4 180 0.4 1 × 10−3 20.9 0.5 1°

TOF repetition frequency (kHz) accelerating voltage (V) average ion energy (keV) average trajectory radus (mm) inner electrode radius (mm) outer electrode radius (mm) ion rotation angle (degree) average ion path length (m) linear section voltage (V) TOF MS pressure (Torr)

20 −3500 3.7 150 144 156 254°30′ 1.4 2780 2 × 10−7

64−400 0.2−1.2 7.65

400

FPGA-Based Recorder and Data Acquisition Algorithm. An important advantage of drift tube IMS/TOF-MS instrumentation is the fact that data collection can be done really fast. Novel recorder was developed for real time 2D data collection with a 200 μs drift time sampling step and a 1.5 ns TOF sampling step. A full 2D drift time/time-of-flight data matrix is collected during 51.2 ms and then a chosen number of matrixes (from 1 to 2048) can be integrated in order to increase the signal-to-noise ratio. By this way, 2D data record time can be chosen between 0.05 to 100 s without loss of information. In addition, the recorder compresses the data into a 16 Mb file that allows 3 s full 2D data transfer to the computer. A detailed description of field-programmable gate array (FPGA) recorder and data acquisition algorithm is presented in the Supporting Information. Software. The control over the ion mobility spectrometer− mass spectrometer and data processing was done by a specially developed computer code. The code provides full control of power supplies, control over ion gates, assignment of data collection algorithm and the data collection control. The code 9006

dx.doi.org/10.1021/ac401191k | Anal. Chem. 2013, 85, 9003−9012

Analytical Chemistry

Article

Figure 4. Mobility distributions derived from IMS/MS data obtained for a mixture of tetrapropylammonium iodide, tetrapentylammonium iodide, and tetraoctylammonium bromide dissolved in pure acetonitrile: (a) for full mass range selected, (b) for tetraoctylammonium bromide (m/z = 466.5, 467.5, and 468.5 selected), (c) for tetrapentylammonium iodide (m/z = 298.4 and 299.4 selected), and (d) for tetrapropylammonium iodide only (m/z = 186.1 and 187.1 selected).

methanol (Merck, Germany). Both amoxicillin and ampicillin were prepared as mixtures in acetonitrile at a 50 μM concentration level.

allows data representation and visualization as a distribution of the reduced ion mobility, mass spectrum, mass selected mobility distribution, mobility selected mass distribution, reading parameters of the experiment, calibration of the mass and reduced mobility scales, measuring the area of the peaks, and calculation of the signal-to-noise ratio and resolving power. The reduced mobility distribution was transformed from drift time distributions using the reference points. The code determined mass, reduced mobility, and drift time of a compound by calculation of the center-of-gravity for a peak. Signal intensity was calculated as a peak area. Samples. Tetrapropylammonium iodide, tetrapentylammonium iodide from Fluka (Buchs SG, Switzerland), tetraoctylammonium bromide, and 2,6-ditert-butylpyridine (2,6-DtBP) from Sigma-Aldrich (Steinheim, Germany) were used as chemical standards because they were proposed as reference compounds for IMS.40−42 Unless specified otherwise, analytes were dissolved in acetonitrile (Baker, Deventer, Holland) at a 1.5 μM concentration. Acetic acid (0.1%; Riedel-de Haën, Seelze, Germany) was added to enhance the ionization of 2,6ditert-butylpyridine in ESI. Antibiotics were lomefloxacin, ofloxacin, amoxicillin, and ampicillin from Sigma-Aldrich (Germany). Lomefloxacin (5 μM) and ofloxacin (10 μM) were prepared as mixtures in

■

SAFETY CONSIDERATIONS High voltages are applied to the ion mobility spectrometer and mass spectrometer. Care must be taken to avoid electric discharges and injury.

■

RESULTS AND DISCUSSION

In this study, we show the analytical performance obtained for the developed ion mobility spectrometer/TOF mass spectrometer. We also present precise values of the reduced ion mobility for selected antibiotics obtained in fast IMS/TOF-MS measurements. Performance. Resolving Power. The experimental setup, described here, routinely operated with the mass resolving power of 2000. This resolving power does not represent a limit for the mass analyzer. This is a sensible compromise between the analytical parameters experimentally obtained in the current stage of our research. The ion transmission of 75% between the exit of first skimmer and the analyzer has been obtained in the preliminary tests. The further increase in the mass resolving 9007

dx.doi.org/10.1021/ac401191k | Anal. Chem. 2013, 85, 9003−9012

Analytical Chemistry

Article

following way. It is a time between IMS entrance gating and a start pulse which initiates the TOF transient in which the corresponding ion is detected. For atmospheric pressure drift tube IMS, ion packets spend some time at the differential pumping interface after separation in ion mobility and before entering the orthogonal accelerator. The ions movements cannot be described as a drift at the differential pumping interface. Existence of this region makes the direct determination of reduced mobility more complex and requires the use of internal mobility standards. For this reason, the accuracy of reduced mobility measurements is affected by both reproducibility of drift time and the error of mobility calibration. Another approach for the accurate ion mobility measurements45 is based on plotting mobility drift time versus inversed voltage. This allows determining the residence time of an ion in a mass spectrometer after the mobility separation. In the first experiments, reproducibility of drift time was studied at stabilized pressure, temperature, and drift gas humidity. Due to the optimization of data collection algorithm, 4 consequent TOF transients are summed and IMS sampling step is defacto equal to 0.2 ms. In order to clarify how this IMS sampling step could affect reproducibility of measured drift time, the 20 consequent measurements were done for an acetonitrile solution containing 1.5 μM tetrapropylammonium iodide, 1.5 μM tetrapentylammonium iodide, 1.5 μM tetraoctylammonium bromide, 0.1 nM 2,6-di-tert-butylpyridine, and 0.5% acetic acid. Reproducibility of measured drift time was determined as a maximum value from the RSDs of drift time determined for tetrapropylammonium iodide, tetrapentylammonium iodide, tetraoctylammonium bromide, and 2,6-di-tertbutylpyridine from 20 consequent measurements done for a mixture containing all four compounds. Every measurement takes 100 s. Drift time of a compound was determined by the corresponding drift time peak center-of-gravity position. Reproducibility of measured drift time was determined as 0.1%. Then calibration error was estimated. The following approach of calibration was used in order to derive unknown reduced mobility values for studied compounds. Drift times (tdi) were obtained as the peak’s center-of-gravity for N standards (N > 2) with known reduced mobilities K0i and were used to plot calibration line K0 as a function of 1/td. The mobility calibration plot is schematically shown in Figure S-4. Relatively good linearity 0.99997 was obtained for the calibration line for N = 4 when 0.5 μM tetrapropylammonium iodide [K01 = 1.47 cm2/(V s)], 1.5 μM 2,6-DtBP [K02 = 1.475 cm2/(V s)], 1 μM tetrapentylammonium iodide (K03 = 1.10 cm2/(V s)], and 1 μM tetraoctylammonium bromide [K04 = 0.80 cm2/(V s)] were used as internal standards with known reduced mobility. For every measurement, the unknown reduced mobility of a studied compound was determined by a calibration line obtained from the inverse drift time of the compound. Both calibration line and calibration error are determined by the least-squares method. Calculated unknown reduced ion mobility was obtained by averaging values from a few repetitive measurements. The N = 4 value of the calibration error was found to be 0.2−0.5% and, thus, is considered as the main factor affecting accuracy of measured reduced mobility in fast IMS/TOF-MS measurements. The main factor of calibration error was found to be an accuracy of available reduced mobility data for used standards. Precise Reduced Mobility Values for Selected Antibiotics Obtained in Fast IMS/TOF-MS Measurements. Importance of fast identification of antibiotics is caused by their

power was also affected by a 1.5 ns digitization step of a used PLIC-based recorder. For real-time data collection by atmosphere pressure drift tube IMS/TOF mass spectrometer, the increase of the ion mobility resolving power is significantly affected by ion packet broadening in the differential pumping interface. For the current setup of the interface, an ion packet broadening was experimentally found to be below 200 μs. Tetraalkylammonium halides were used for experimental estimation of the upper limit of the ion mobility resolving power. The resolving power was calculated for peaks of mass selective mobility distributions derived from IMS/TOF-MS data. Measurements were done at a 400 μL/h liquid sample flow, a 1.2 L/min nebulizer gas flow, a 2.2 L/min drift gas flow, a 0.2 ms gate opening time, and a 100 s data collection time. Figure 4 shows typical mobility distributions derived from IMS/MS data obtained for mixture of tetrapropylammonium iodide, tetrapentylammonium iodide, and tetraoctylammonium bromide dissolved in acetonitrile. Mobility resolving power calculated from mass-selected mobility distribution peaks amounted to 79 (RSD = 10%) for tetrapropylammonium iodide (m/z = 186.1 and 187.1 selected), 80 (RSD = 11%) for tetrapentylammonium iodide (m/z = 298.4 and 299.4 selected), and 110 (RSD = 22%) for tetraoctylammonium bromide (m/z = 466.5, 467.5, and 468.5 selected). The data were obtained as a result of statistical data treatment from 22 sequential IMS/TOF-MS measurements. The data approximately correspond to theoretical estimations of resolving power obtained in accordance with the method43,44 for a 0.2 ms initial ion packet taking into account diffusion contribution and 0.1 ms half-height widening inside the differential pumping interface. Theoretical values of resolving power amount to 77 for tetrapropylammonium iodide, 91 for tetrapentylammonium iodide, and 103 for tetraoctylammonium bromide. Detection Limit. The limit of detection (S/N = 3) for 2,6-ditert-butylpyridine was determined. This compound was chosen because it is relatively easy to dissolve in acetonitrile and forms a stable protonated molecule in positive-ion mode electrospray. Measurements were done at a 64 μL/hour liquid sample flow, 1.2 L/min nebulizer gas flow, 2.2 L/min drift gas flow, 0.5 ms gate opening time, and 100 s data collection time. LC−MSgrade acetonitrile was used as a solvent. First, the blank sample (0.5% solution of acetic acid in acetonitrile) has been repeatedly measured 15 times. An example of background spectrum is shown in Figure S-3 of the Supporting Information. Intensity of the background value was estimated as an average value of ion response in the area of drift time of 2,6-di-tert-butylpyridine in mass selective (m/z = 191 and 192) mobility distribution for the blank sample. The noise level was estimated as RSD for values of ion response in the area of drift time of 2,6-di-tert-butylpyridine in mass selective (m/z = 191 and 192) mobility distributions for the blank sample. Then 2,6-di-tert-butylpyridine (solution in acetonitrile with 0.5% acetic acid) has been measured 15 times for every concentration at 5, 10, 50, and 100 nM. Signal level was estimated as an average value of the peak area for 2,6-di-tertbutylpyridine in mass selective (m/z = 191,192) mobility distributions. The routinely available relative limit of detection amounted to 4 nM, and the routinely available absolute limit of detection amounted to 1 × 10−12 g. Accuracy of Measured Reduced Mobility in Fast IMS/TOFMS Measurements. In a fast real-time drift tube IMS/TOFMS, the measured ion’s drift time can be determined in the 9008

dx.doi.org/10.1021/ac401191k | Anal. Chem. 2013, 85, 9003−9012

Analytical Chemistry

Article

presence in the aquatic environment at trace level. This fact can have an effect on ecological and human health due to breeding more disease-causing bacteria that are resistant to even high doses of drugs. Antibiotics have been mainly analyzed by liquid chromatography/mass spectrometry or tandem mass spectrometry,46,47 and we have found only one paper where antibiotics were analyzed from liquid samples by ion mobility spectrometry with the Faraday cup detector.48 Reduced ion mobilities of the characteristic ions produced by selected antibiotics were measured at a temperature of 27 °C, pressure of 750 Torr, humidity of 0.05%, sample flow of 150 μL/hour, nebulizer gas flow of 0.2 L/min, drift gas flow of 2 L/ min, 0.5 ms gate opening time, and 100 s data collection time. In every sample, four internal mobility standards (0.5 μM tetrapropylammonium iodide, 1 μM tetrapentylammonium iodide, 1 μM tetraoctylammonium bromide, and 1.5 μM 2,6DtBP) were added. Reduced ion mobility and an error of reduced mobility determination were obtained in accordance with the method described in the previous section. Three through five series of measurement were done separated by a few months time intervals for every studied antibiotics. Relatively low RSD (below 0.2%) was observed for every series of measurements. Calibration error for different series of measurements was between 0.2−0.5%. Inter series reproducibility estimated as RSD of average-reduced mobility value obtained in different series amounted to 0.3−0.6% for different antibiotics. The reported value of reduced mobility was an average of all values obtained for the considered antibiotic. The reported error was calculated as the square root from a sum of squared calibration error and RSD. Lomefloxacin. Lomefloxacin has not shown dependence of ion mobility mass spectra upon sample freshness in the 0−96 h timed range. Only one well-resolved peak with reduced mobility at 1.082 cm2/(V s) is observed on the ion mobility distribution for lomefloxacin. The peak is formed by a compound with a monoisotopic mass of 352.15 Da and identified as a protonated molecule of lomefloxacin [M + H]+. The error of reduced mobility determination was found to be 0.4%. Relative reduced mobility of 0.734 was calculated relative to 2,6-ditert-butylpyridine for a protonated molecule of lomefloxacin. Ofloxacin. Ion mobility mass spectra of all studied antibiotics, excluding lomefloxacin, were shown to be dependent upon sample storage time. Recently prepared (used in 4 h) samples of ofloxacin has shown one ion mobility peak with 1.051 cm2/(V s). The peak is formed by a compound with a monoisotopic mass of 362.15 Da, which is identified as a protonated molecule of ofloxacin [M + H]+. For a sample stored for 48 h, two intense mobility peaks were observed (Figure 5a). The peak with the lower ion mobility of 1.051 cm2/(V s) was formed by a compound with a monoisotopic mass of 362.15 Da, which was identified as a protonated molecule of ofloxacin [M + H]+. It was verified by good agreement of [M + H]+ isotopic composition with relative intensity of corresponding peaks. The peak with higher mobility was mainly observed at 1.303 cm2/(V s). A few series of sporadic shifts of reduced mobility values were observed for this peak. The peak was formed by (1) a component with masses of 362.15 Da, 363.15 Da, and 364.15 Da, which was identified as a protonated molecule of ofloxacin of the most abundant isotope composition [M361 + H]+, [M362 + H]+, and [M363 + H]+, (2) a compound with masses of 362.15 Da,

Figure 5. ESI-IMS/MS data for a mixture of 0.5 μM 2,6-DtBP and 10 μM ofloxacin dissolved in methanol with 2% acetic acid: (a) mass selective mobility distribution for ofloxacin (m/z = 362.2 and 363.2 Da), mass peaks at (b) m/z = 362.2, 362.7, and 363.2 Da [K0 = 1.303 cm2/(V s)] produced by protonated molecule and doubly charged dimer, and (c) m/z = 723.4, 724.4, and 725.4 Da [K0 = 0.816 cm2/(V s)] produced by proton-bound dimer.

362.65 Da, and 363.15 Da, which was identified as a doubly charged dimer of a molecule of ofloxacin of the most abundant isotope composition [M362 + 2H + M362]+2, [M362 + 2H + M363]+2, and [M363 + 2H + M363]+2 (Figure 5b). The proposed composition is verified by the relative intensity of peaks 362.15, 363.15, and 364.15 Da after subtraction of the contribution of [M + 2H + M]+2 reconstructed by the peak of 362.65 Da. Relative intensities of the peaks are in good agreement with isotopic composition of [M+H]+. The mass spectrum also contained low-intensity peaks of 723.30 Da, 724.30 Da, 725.30 Da, which were identified as a singly charged protonated dimer of molecule of ofloxacin [M362 + H + M362]+, [M362 + H + M363]+, and [M363 + H + M363]+ (Figure 5c). This dimer formed a low intensity mobility peak of 0.816 cm2/(V s). Apparently the peak with higher mobility [1.303 cm2/(V s)] 9009

dx.doi.org/10.1021/ac401191k | Anal. Chem. 2013, 85, 9003−9012

Analytical Chemistry

Article

Table 2. Reduced Mobilities of Characteristic Ions for Studied Antibiotics antibiotic

monoisotopic mass (Da)

K0 (lit.48), cm2/(V s)

K0 (exp.), cm2/(V s)

mass peaks (Da)

identification

lomefloxacin ofloxacin

351.14 361.14

1,19

1.082a 0.816b,c 1.051a 1.303b,c,d

352.15 723.30 362.15 362.15 362.66 363.15 366.11 398.14 407.14 350.12 382.14 391.14

[M + H]+ [M361 + H + M361]+ [M + H]+ [M361 + H]+ [M361 + 2H + M361]+2 [M361 + 2H + M362]+2 [M362 + H]+ [M362 + 2H + M362]+2 [M + H]+ [M + CH3OH + H]+ [M + CH3CN + H]+ [M + H]+ [M + CH3OH + H]+ [M + CH3CN + H]+

1,21

amoxicillin

365.10

1,11

ampicillin

349.11

0,96

0.909a 0.905a 0.909a,c 0.945a 0.945a,c 0.945a

a

Peaks observed for all the samples independently upon their freshness. bPeaks observed only for the samples kept more than 48 h. cPeak intensity rise with the time of sample storing. dSporadic shift of mobility values was observed for the peaks measured in different series.

intensity than for acetonitrile-containing adducts. For samples older than 48 h, the peak formed by a protonated methanolcontaining cluster dominated and it was 2 orders of magnitude more intense than a peak of a protonated molecule and an order of magnitude more intense than a peak of a protonated acetonitrile-containing cluster. Relative reduced mobility of 0.641 was calculated relative to 2,6-ditert-butylpyridine for a protonated molecule, protonated methanol-containing cluster, and protonated acetonitrile-containing cluster of ampicillin. The obtained values of the reduced mobility for the components of the mixtures are summarized in Table 2. Through the use of the mass-selective detection, it became possible to identify groups separated by ion mobility, which are specific to a particular antibiotic.

was formed by a doubly charged diprotonated dimer of a molecule of ofloxacin, which partly dissociated into two protonated molecules of ofloxacin in the interface. The error of reduced mobility determination was found to be 0.4%. Relative reduced mobility of 0.713 was calculated relative to 2,6-ditert-butylpyridine for protonated molecules of ofloxacin. Amoxicillin. On the mobility distribution for amoxicillin, two peaks had similar ion mobility: 0.909 cm2/(V s) and 0.905 cm2/(V s). The more intense peak with mobility 0.905 cm2/(V s) is formed by a compound with monoisotopic mass of 398.14 Da, identified as a protonated methanol-containing cluster of amoxicillin [M + CH3OH + H]+. Formation of an amoxicillin adduct with methanol was earlier observed in electrospray ion trap tandem mass spectrometry.49 The minor peak with mobility 0.909 cm2/(V s) is formed by a compound with a monoisotopic mass of 366.11 Da identified as a protonated molecule of amoxicillin and a compound with monoisotopic mass 407.14 Da identified as a protonated acetonitrilecontaining cluster of amoxicillin [M + CH3CN + H]+. Intensity of the amoxicillin methanol-containing adduct has a clear tendency to rise, and then both intensities of the acetonitrilecontaining adduct of amoxicillin and the protonated molecules fall after a few days of sample storage compared with those of fresh samples. The error of reduced mobility determination was found to be 0.5%. Relative reduced mobility of 0.616, 0.614, and 0.616 were calculated relative to 2,6-ditert-butylpyridine for protonated molecules, protonated methanol-containing clusters, and protonated acetonitrile-containing clusters of amoxicillin. Ampicillin. Only one ion mobility peak [0.945 cm2/(V s)] is observed on the distribution of ion mobility for ampicillin. The peak is formed by a compound with a monoisotopic mass of 382.14 Da, identified as a protonated methanol-containing cluster of ampicillin [M + CH3OH + H]+, a compound with a monoisotopic mass of 391.14 Da identified as a protonated acetonitrile-containing cluster of ampicillin [M + CH3CN + H]+, and a compound with a monoisotopic mass of 350.12 Da, identified as a protonated molecule of ampicillin [M + H]+. The error of reduced mobility determination was found to be 0.6%. The protonated methanol-containing cluster for ampicillin has been demonstrated as being similar to amoxicillin kinetics of formation. For fresh samples, the peaks produced by the methanol-adduct cluster and protonated molecule were observed as having similar intensities with 5 times lower

■

CONCLUSION The hybrid instrument based on the drift tube IMS and the orthogonal acceleration sector time-of-flight mass analyzer could become a relatively fast and selective analytical tool with a data collection time of 100 s. The instrument demonstrated reproducibly the mobility resolving power of 100 and the mass resolving power of 2000. The relative detection limit of 4 nM was obtained, and the absolute detection limits of 1 × 10−12 g was shown for 2,6-DtBP. The developed instrument was used to define characteristic mobility/mass distributions for amoxicillin, ampicillin, lomefloxacin, and ofloxacin. Mentioned antibiotics were shown to be easily identifiable in the ESI-IMS/TOFMS mode. At studied conditions, lomefloxacin forms only a protonated molecule, producing a reduced mobility peak at 1.082 cm2/(V s). Both amoxicillin and ampicillin produce [M + H]+, [M + CH3OH + H]+, and [M + CH3CN + H]+. Amoxicillin shows two reduced mobility peaks at 0.909 cm2/(V s) for [M + H]+ and [M + CH3CN + H]+ and 0.905 cm2/(V s) for [M + CH3OH + H]+. Ampicillin shows one reduced mobility peak at 0.945 cm2/(V s) for [M + H]+, [M + CH3OH + H]+, and [M + CH3CN + H]+. The intensity of protonated methanol-containing clusters for both ampicillin and amoxicillin, has a clear tendency to rise with increasing the sample storage time. Ofloxacin produces two intense peaks in the mobility distribution. Lower mobility peak at 1.051 cm2/(V s) is shown to be formed by [M + H]+ ions. Higher mobility peak observed between 1.303 cm2/(V s) for samples kept more than 48 h is shown to be formed by both the [M + H]+ ion and a component identified as a [M + 2H + M]+2 cluster. The cluster probably partly dissociates in the 9010

dx.doi.org/10.1021/ac401191k | Anal. Chem. 2013, 85, 9003−9012

Analytical Chemistry

Article

interface, producing a [M + H]+ ion. Low intensity peaks identified as a [M + H + M]+ cluster are also present in the mass spectra. The following relative reduced mobilities are suggested for identification of studied antibiotics relative to 2,6-ditertbutylpyridine: 0.734 for [M + H]+ of lomefloxacin, 0.713 for [M + H]+ of ofloxacin, 0.616 for [M + H]+ and [M + CH3CN + H]+ of amoxicillin, 0.614 for [M + CH3OH + H]+ of amoxicillin, 0.641 for [M + H]+, [M + CH3OH + H]+, and [M + CH3CN + H]+ of ampicillin.

■

M.; Schepmoes, A. A.; Hopkins, D. F.; Tang, K.; Smith, R. D.; Belov, M. E. J. Proteome Res. 2010, 9, 997−1006. (18) Woods, A. S.; Ugarov, M.; Jackson, S. N.; Egan, T.; Wang, H.-Y. J.; Murray, K. K.; Schultz, J. A. J. Proteome Res. 2006, 5, 1484−1487. (19) Kwasnik, M.; Fuhrer, K.; Gonin, M.; Barbeau, K.; Fernandez, F. M. Anal. Chem. 2007, 79, 7782−7791. (20) Merenbloom, S. I.; Koeniger, S. L.; Valentine, S. J.; Plasencia, M. D.; Clemmer, D. E. Anal. Chem. 2006, 78, 2802−2809. (21) Bohrer, B. C.; Clemmer, D. E. Anal. Chem. 2011, 83, 5377− 5385. (22) Lapthorn, C.; Pullen, F.; Chowdhry, B. Z. Mass Spectrom. Rev. 2012, DOI: 10.1002/mas.21349. (23) Tang, K.; Shvartsburg, A. A.; Lee, H-N; Prior, D. C.; Buschbach, M. A.; Li, F.; Tolmachev, A. V.; Anderson, G. A.; Smith, R. D. Anal. Chem. 2005, 77, 3330−3339. (24) Belov, M. E.; Clowers, B. H.; Prior, D. C.; Danielson, W. F., III; Liyu, A. V.; Petritis, B. O.; Smith, R. D. Anal. Chem. 2008, 80, 5873− 5883. (25) Henderson, S. C.; Valentine, S. J.; Counterman, A. E.; Clemmer, D. E. Anal. Chem. 1999, 71, 291−301. (26) Creaser, C. S.; Benyezzar, M.; Griffiths, J. R.; Stygall, J. W. Anal. Chem. 2000, 72, 2724−2729. (27) Myung, S.; Lee, Y. J.; Moon, M. H.; Taraszka, J.; Sowell, R.; Koeniger, S.; Hilderbrand, A. E.; Valentine, S. J.; Cherbas, L.; Cherbas, P.; Kaufmann, T. C.; Miller, D. F.; Mechref, Y.; Novotny, M. V.; Ewing, M. A.; Sporleder, C. R.; Clemmer, D. E. Anal. Chem. 2003, 75, 5137−5145. (28) Wyttenbach, T.; Kemper, P. R.; Bowers, M. T. Int. J. Mass Spectrom. 2001, 212, 13−23. (29) Tarver, E. E. Sensors 2004, 4, 1−13. (30) Clowers, B. H.; Siems, W. F.; Hill, H. H.; Massick, S. M. Anal. Chem. 2006, 78, 44−51. (31) Belov, M. E.; Buschbach, M. A.; Prior, D. C.; Tang, K.; Smith, R. D. Anal. Chem. 2007, 79, 2451−2462. (32) Clowers, B. H.; Ibrahim, Y.; Prior, D. C.; Danielson, W. F., III; Belov, M. E.; Smith, R. D. Anal. Chem. 2008, 80, 612−623. (33) Clowers, B. H.; Belov, M. E.; Prior, D. C.; Danielson, W. F., III; Ibrahim, Y.; Smith, R. D. Anal. Chem. 2008, 80, 2464−2473. (34) Guilhaus, M.; Selby, D.; Mlynski, V. Mass Spectrom. Rev. 2000, 19, 65−107. (35) Chernyshev, D. M.; Poteshin, S. S.; Sysoev, A. A.; Sysoev, A. A. J. Anal. Chem. 2012, 67, 1093−1095. (36) Sysoev, A.; Adamov, A.; Viidanoja, J.; Ketola, R. A.; Kostiainen, R.; Kotiaho, T. Rapid Commun. Mass Spectrom. 2004, 18, 3131−3139. (37) Adamov, A.; Mauriala, T.; Teplov, V.; Laakia, J.; Pederson, C.; Kotiaho, T.; Sysoev, A. A. Int. J. Mass Spectrom. 2010, 298, 24−29. (38) Chernyshev, D. M.; Frolov, I. S.; Frolov, A. S.; Mukhanov, M. S.; Sysoev, A. A. J. Anal. Chem. 2011, 66, 5−9. (39) Sheretov, E. P., Buill Izobret, Inventor’s certificate USSR No. 711647, 1980; Vol. 3. (40) Eiceman, G. A.; Nazarov, E. G.; Stone, J. A. Anal. Chim. Acta 2003, 493, 185−194. (41) Viidanoja, J.; Sysoev, A.; Adamov, A.; Kotiaho, T. Rapid Commun. Mass Spectrom. 2005, 19, 3051−3055. (42) Viitanen, A.; Mauriala, T.; Mattila, T.; Adamov, A.; Pedersen, C. S.; Mäkelä, J. M.; Marjamäki, M.; Sysoev, A.; Keskinen, J.; Kotiaho, T. Talanta 2008, 76, 1218−1223. (43) Steiner, W. E.; English, W. A.; Hill, H. H., Jr. Anal. Chim. Acta 2005, 532, 37−45. (44) Kanu, A. B.; Gribb, M. M.; Hill, H. H., Jr. Anal. Chem. 2008, 80, 6610−6619. (45) Crawford, C. L.; Hauck, B. C.; Tufariello, J. A.; Harden, C. S.; McHugh, V.; Siems, W. F.; Hill, H. H., Jr. Talanta 2012, 101, 161− 170. (46) Hao, C. Y.; Clement, R.; Yang, P. Anal. Bioanal. Chem. 2007, 387, 1247−1257. (47) Hernandez, F.; Sancho, J. V.; Ibanez, M.; Guerrero, C. TrAC, Trends Anal. Chem. 2007, 26, 466−485.

ASSOCIATED CONTENT

S Supporting Information *

Additional information as noted in text. This material is available free of charge via the Internet at http://pubs.acs.org.

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS This work was financially supported by the International Science and Technology Center (project number 3623), the Moscow Department for Support and Development of Small Business (contract number 123/08-VP), the Department of Science and Industrial Policy of Moscow (contract number 8/ 3-14n-10), as well as a part of the Federal Program “Scientific and Scientific-Pedagogical Personnel of Innovative Russia 20092013” (contracts 14.740.11.0342, 14.740.11.0721, 14.740.11.0540).

■

REFERENCES

(1) Revercomb, H. E.; Mason, E. A. Anal. Chem. 1975, 47, 970−983. (2) Hill, H. H.; Siems, W. F.; St. Louis, R. H.; McMinn, D. G. Anal. Chem. 1990, 62, 1201A−1209A. (3) Collins, D.; Lee, M. Anal. Bioanal. Chem. 2002, 372, 66−73. (4) Kanu, A. B.; Dwivedi, P.; Tam, M.; Matz, L.; Hill, H. H., Jr. J. Mass Spectrom. 2008, 43, 1−22. (5) Eiceman, G. A.; Karpas, Z. Ion Mobility Spectrometry; CRC Press: Boca Raton, 2005. (6) Borsdorf, H.; Eiceman, G. A. Appl. Spectrosc. Rev. 2006, 41, 323. (7) Fernandez-Maestre, R.; Hill, H. H., Jr. Int. J. Ion Mobility Spectrom. 2009, 12, 91−102. (8) Pollard, M. J.; Hilton, C. K.; Li, H.; Kaplan, K.; Yost, R. A.; Hill, H. H., Jr. Int. J. Ion Mobility Spectrom. 2011, 14, 15−22. (9) Hill, C. A.; Thomas, C. L. Analyst 2003, 128, 55−60. (10) Wu, C.; Siems, W. F.; Asbury, G. R.; Hill, H. H., Jr. Anal. Chem. 1998, 70, 4929−4938. (11) Dugourd, Ph.; Hudgins, R. R.; Clemmer, D. E.; Jarrold, M. F. Rev. Sci. Instrum. 1997, 68, 1122−1129. (12) Clowers, B. H.; Hill, H. H., Jr. Anal. Chem. 2005, 77, 5877− 5885. (13) Tadjimukhamedov, F. K.; Jackson, A. U.; Nazarov, E. G.; Ouyang, Z.; Cooks, R. G. J. Am. Soc. Mass Spectrom. 2010, 21, 1477− 1481. (14) Hoaglund, C. S.; Valentine, S. J.; Clemmer, D. E. Anal. Chem. 1997, 69, 4156−4161. (15) Tang, X.; Bruce, J. E.; Hill, H. H., Jr. Rapid Commun. Mass Spectrom. 2007, 21, 1115−1122. (16) Crawford, C.; Graf, S.; Gonin, M.; Fuhrer, K.; Zhang, X.; Hill, H. Int. J. Ion Mobility Spectrom. 2011, 14, 23−30. (17) Baker, E. S.; Livesay, E. A.; Orton, D. J.; Moore, R. J.; Danielson, W. F.; Prior, D. C.; Ibrahim, Y. M.; LaMarche, B. L.; Mayampurath, A. 9011

dx.doi.org/10.1021/ac401191k | Anal. Chem. 2013, 85, 9003−9012

Analytical Chemistry

Article

(48) Li, S.; Jia, J.; Gao, X.; He, X.; Li, J. Anal. Chim. Acta 2012, 720, 97−103. (49) Grujic, S.; Vasiljevic, T.; Lausevic, M.; Ast, T. Rapid Commun. Mass Spectrom. 2008, 22, 67−74.

9012

dx.doi.org/10.1021/ac401191k | Anal. Chem. 2013, 85, 9003−9012