| SUBJECT |
Anal. Chem., ASAP Article 10.1021/ac049420n S0003-2700(04)09420-X Web Release Date: July 16, 2004 Copyright © 2004 American Chemical Society Rectilinear Ion Trap: Concepts, Calculations, and Analytical Performance of a New Mass Analyzer Zheng Ouyang, Guangxiang Wu, Yishu Song, Hongyan Li, Wolfgang R. Plass, and R. Graham Cooks* Department of Chemistry, Purdue University, West Lafayette, Indiana 47907-2084, and II. Physikalisches Institut, Justus-Liebig-Universität Giessen, Heinrich-Buff-Ring 16, 35392 Giessen, Germany Received for review April 17, 2004. Accepted June 23, 2004. Abstract: A mass analyzer based on a rectilinear geometry ion trap (RIT) has been built, and its performance has been characterized. Design concepts for this type of ion trap are delineated with emphasis on the effects of electrode geometry on the calculated electric field. The Mathieu stability region was mapped experimentally. The instrument can be operated using mass-selective instability scans in both the boundary and resonance ejection versions. Comparisons of performance between different versions of the device having different dimensions allowed selection of an optimized geometry with an appropriate distribution of higher-order electric fields. Comparisons made under the same conditions between the performance of a conventional cylindrical ion trap and a RIT of 4 times greater volume show an improvement of 40 times in the signal-to-noise ratio resulting from the higher ion trapping capacity of the RIT. The demonstrated capabilities of the RIT include tandem mass spectrometry, a mass resolution in excess of 1000, and a mass/charge range of 650 Th, all in a simple structure that is only 3.5 cm3 in internal volume. -------------------------------------------------------------------------------- Mass spectrometry is perhaps the most widely used analytical technique, due principally to its high specificity in analyte identification as well as its high sensitivity and applicability to the analysis of a broad range of isotopes, elements, and chemical and biological compounds. Mass spectrometers have played an important role in isotopic analysis, in trace elemental analysis, in identification of unknown compounds in complex mixtures, and in structural elucidation, to mention just a few areas of application. Despite these successes, the mass spectrometer has disadvantages in terms of its relatively large size, weight, power consumption, and complexity of maintenance. Further extension of the areas of application of mass spectrometers requires improvements in these characteristics. Applications in sensing for explosives, chemical and biological toxic agents, in situ analysis for environment protection, in-line monitoring for process control, clinical screening, and quality assurance by covering many aspects of people's daily lives could all be greatly impacted by the availability of small, portable instruments. While the requirements for sensitivity and specificity of an analysis vary with the particular application, it is highly desirable that such compact instruments have tandem mass spectrometry (MSn) capabilities for analysis of complex samples and that their operation be simple. All these considerations have contributed to a growing interest in miniature mass spectrometers.1 Miniaturization of a mass spectrometer usually starts with simplification and reduction in size of the mass analyzer to allow easy machining on a smaller scale. In the case of ion trap, this also allows operation at lower voltages and higher pressures. Virtually all types of mass analyzers have been miniaturized,1 including sector instruments,2-4 time-of-flight (TOF) instruments,5,6 linear quadrupoles,7,8 quadrupole ion traps,9 Fourier transform ion cyclotron resonance spectrometers,10 and ion mobility instruments.11 After adequate performance is obtained with a given miniature mass analyzer, the next step in the development of a miniature mass spectrometer system is the miniaturization of the other components, especially the vacuum and electronics systems. This is a complex task and only a few miniature mass spectrometers, including those using cylindrical ion traps,12 BE sectors,13 TOF analyzers,14,15 and mass filters,16 have actually been built and tested. Among the mass analyzers that have been miniaturized, ion traps show some unique advantages. Ion traps operate at pressures above 10-3 Torr, achievable using a two-stage mechanical pump (without a high-vacuum pump) or by a miniaturized pumping system with compromised pumping capability but suitable for small portable instruments.17 As the ion trap dimensions are reduced, the ion trajectories involved in mass analysis become shorter and the trap can operate at still higher pressures. Another important capability, especially for chemical identification of components of complex mixtures, is tandem mass spectrometry, and it is noteworthy that MSn experiments (multiple stages of mass analysis) can be performed using a single ion trap.18 Ion traps are also known for their relative ease of fabrication and their low mechanical tolerances. Ion traps are also the only type of mass spectrometer that most users can disassemble and reassemble without sacrificing instrument performance. Cylindrical ion traps (CITs) have geometries that are much simpler than hyperbolic Paul traps, and for this reason, they have been developed as miniature mass analyzers19,20 (Figure 1). Portable mass spectrometers using CIT analyzers have been constructed.17,21 The performance of these instruments has been established through detection of specific compounds in water and air at trace levels (low-ppb range; unit mass/charge resolution to 300 Th)17,21 with confirmation through tandem mass spectrometry. These characterization experiments have involved remote field as well as laboratory operations. More recently, the capabilities of portable mass spectrometers have been extended by using ion/molecule reactions for the detection of specific explosives and chemical warfare agent simulants at low levels.22 The targeted chemical compounds thus can be identified from mixtures. -------------------------------------------------------------------------------- Figure 1 Conceptual evolution of the rectilinear ion trap and interrelationship to other types of ion traps. -------------------------------------------------------------------------------- All three-dimensional (3D) ion traps, including the CIT, suffer drawbacks of low trapping capacity and low trapping efficiency for externally injected ions.23 It has been reported that fewer than 1000 ions24 can be trapped when high-resolution mass analysis is performed due to space charge effects, although larger numbers are probably common for lower resolution work. Trapping efficiencies of only 5%23 are achieved when ions are injected into the trap from an external source, such as an electrospray ionization or chemical ionization source. To address the limitations of conventional quadrupole ion traps, linear ion traps,25-27 which have geometries more closely related to those of quadrupole mass filters (Figure 1), have been developed. Even with 30 000 trapped ions,28 high-quality mass spectra can be recorded without any space charge effects in these new commercial instruments. In addition, linear ion traps show much higher trapping efficiencies.26,27,29 In the present study, we combine the advantages of a linear ion trap with the geometrical simplicity of the CIT, to arrive at a new ion trap geometry-the rectilinear ion trap. Figure 1 shows the conceptual lineage of the instrument.30 The performance of this new mass analyzer is reported in this paper. Instrument Design. Rectilinear ion traps can be conceived in many different configurations.30 One of the simplest versions, shown in Figure 2, consists of two pairs (x and y) of rectangular electrodes supplied with rf potentials, and a pair of z electrodes to which only a dc voltage is applied. The rf signal is applied between the x and y electrode pairs and forms an rf trapping field in the xy plane, while the z-electrode dc voltage generates a dc trapping potential well along the z axis. The rf field in the xy plane of the RIT is approximately quadrupolar, and the secular frequency () of ion motion in the xy plane can be calculated using the following equation, which also applies to 3D traps and to the conventional 2D quadrupole mass filter:31 where can be calculated from the Mathieu parameters a and q and must have a value between 0 and 1 for both the x and y directions for ions to be trapped. Other trapping devices with similar geometries and configurations, such as a cubic trap32,33 and a square quadrupole ion guide,34 have previously been designed and used for ion storage or ion transfer, but the RIT is the first instrument of this general type to be developed as a mass analyzer although linear ion traps with rectangular electrodes have been tested for mass analysis using the image current detection method.35 Slits in the rf electrodes and apertures in the z electrodes allow the charged particles to be injected into or ejected from the RIT. The simple geometry of the RIT makes it easy to design and fabricate a massively parallel structure, a version of the instrument that should be especially useful when miniaturized for the purpose of operating the mass spectrometer at lower rf voltage and higher pressure.30 -------------------------------------------------------------------------------- Figure 2 Configuration of the rectilinear ion trap and the control signals applied to it. -------------------------------------------------------------------------------- As in the Paul trap18 and other linear ion traps, buffer gas is used in the RIT to facilitate trapping of ions through collisions. Ions can be generated inside the RIT via internal electron ionization (EI) by injecting electrons from a filament into the RIT into which vaporized sample has been introduced. Alternatively, ions can be generated outside of the RIT via any of a number of ionization methods and then transferred into the RIT. When injected through the slits in the x or y electrode, the ions will experience the strong alternating rf field upon injection. As is the case also for external ion injection into 3D traps, the trapping efficiency is expected to be low (~5%23). To maximize the trapping efficiency, the ions can be injected into the RIT along the z axis through apertures in the z electrodes. When identical rf signals, balanced in amplitude but 180 different in phase, are applied to the x and y electrodes, resistance to injection in the z direction by the rf field is minimized. The ions injected into the RIT lose kinetic energy through collisions with the buffer gas and are subsequently trapped by the field due to the combination of rf and dc potentials operating in all three directions. The trapping efficiency when the RIT is operated in this mode is expected to be as high as that for other linear ion traps, for which values as high as 100% have been reported.27,29 An increase of a factor of 60 in trapping capacity was reported for a linear ion trap with a volume just 1.4 times that of a 3D Paul trap.27,28 A similar improvement in trapping capacity is expected for the RIT in comparison with the CIT, and the trapping capacity can be increased as needed by increasing the length of the RIT.27 Mass analysis of trapped ions in the RIT is performed using standard mass-selective instability rf scans.36 The ions can be scanned out of the RIT through the slits in the x, y electrodes, as is done for some commercial linear ion traps with hyperbolic geometry.27 In addition, a supplementary ac signal can be applied between the two x electrodes to achieve resonance ejection during the rf scan. The effect of this ac dipolar field on ion ejection will be illustrated by the experimental results discussed below. As is the case for 3D ion traps and the linear quadrupole mass filter, the relationship between the mass-to-charge ratio of the ions and the voltages applied to the RIT electrodes can be expressed as follows, if the effects of the z-electric field, ion-neutral collisions, and columbic (space charge) effects are ignored: In these expressions, A2 is the quadrupole coefficient in the multipole expansion of the electric potential, rN is the corresponding normalization radius used in the expansion calculation, Vrf (0-p) and Udc are the amplitudes of the rf and the offset dc voltage applied between the x and y electrodes, a and q are the Mathieu parameters, and is the frequency of the applied rf. An alternative method of mass-selective ejection, in which the ions are ejected through the apertures in the z electrodes by scanning the rf and the auxiliary ac,26 is also expected to be effective due to the similarity between the electric field coupling occurring at the end of an RIT and at the end of a linear ion trap with round electrodes.26 The performance of any ion trap as a mass analyzer is dependent on the quality of the trapping electrical field. During the development of the RIT, the internal electric field was calculated for different geometries using a home-written program CreatePot, which calls the Laplace solver from the Poisson/Superfish package (Los Alamos National Levorotary, Los Alamos, NM).37 The effects of the composition of the electric field were analyzed by comparing the spectra obtained using RITs with different geometries. As shown in Figure 3, and extending the analogy between a CIT and a 3D Paul trap, the electric field in the RIT is similar to that in a linear quadrupole with round rod electrodes. The electrical field strength inside an ideal 3D ion trap or linear quadrupole mass filter increases linearly with the distance from its geometric center. However, the field in a real ion trap is usually an approximation to a quadrupolar field with various high-order field components arising from many factors including truncation of the electrodes, the apertures in the electrodes, and fabrication imperfections such as machining errors and misalignment. As is the case also for the CIT,38 the use of simplified (here flat) electrodes increases the high-order field contributions. The high-order fields generally degrade the analytical performance, but under particular conditions, their effects can be mutually compensated to minimize the performance loss or they can even be used to improve the performance, including the mass resolution and MS/MS efficiency.39,40 An optimized geometry with an appropriate distribution of the high-order fields is essential for obtaining adequate performance in an ion trap. Using a procedure similar to the process of optimizing the CIT, the geometric parameters of an RIT, such as the ratio of half-distances (x0 and y0) between the x, y electrodes, the gap distance between the x and y electrodes, and the size of the slits, all can be varied to change the distribution of the high-order fields. The expansion coefficients in the multipole expansion of the electric potential, A2 the quadrupolar term, A4 the octapolar, and A6 the dodecapole, have been calculated and are listed in Table 1 for three RIT geometries with x0 = 5.0 mm and different y0 values, ranging from 5.0 to 4.0 mm. The theoretical values of the expansion coefficients for the odd-order terms, such as hexapole or decapole, are equal to zero for an RIT with a geometry that is symmetrical with respect to the xz and yz planes. The effects of the superimposition of the high-order fields and of varying the x0 to y0 ratio on the RIT performance will be discussed later in this paper. -------------------------------------------------------------------------------- Figure 3 Calculated trapping fields. For (a) rectilinear ion trap: x0 = 5.0 mm and y0 = 5.0 mm. (b) Cylindrical ion trap: r0 = z0 = 5.0 mm. (c) Quadrupole mass filter with round rod electrodes: x = y = 10 mm, rod radius 11.5 mm. (d) Quadrupole ion trap with hyperbolic electrodes: r = 10 mm, z = 7.07 mm, end cap hole radius 0.6 mm. -------------------------------------------------------------------------------- To demonstrate and test the concept of using an RIT for mass analysis, a device has been constructed with the most basic configuration (Figure 2) and its analytical performance has been characterized and carefully compared with that of a CIT. Instrument Construction and Operation. An RIT with the configuration shown in Figure 2 has been fabricated and assembled as shown in Figure 4a. The x, y electrodes were machined from 304-stainless steel and assembled using machinable ceramic holders. The overall dimensional tolerance is ~0.1 mm. Each x, y electrode is 40 mm in length. The half-distance between the x pair of electrodes (x0) is 5.0 mm, and the half-distance between the y pair (y0) was varied in different embodiments from 5.0 to 4.0 mm by adding shims between the y electrodes and the holder. The gap between the x and y electrodes is 1.6 mm. Centrally located on the x electrodes are slits 15.0 mm long and 1.0 mm wide. The slit sizes, like the initial x0 and y0 dimensions, were chosen based on our experience with CITs. Two stainless steel plates without apertures are used as z electrodes located a distance of 1.6 mm from the ends of the x, y electrodes and with an interelectrode distance (z0) of 43.2 mm. The RIT mass analyzer was tested using the control electronics and vacuum system of a prototype Thermo Finnigan ITMS.41 A schematic of the instrumental setup is shown in Figure 4b. An rf signal (1.1 MHz) was applied to the y electrodes, and the x electrodes were virtually grounded to form a trapping field in the x-y plane. A positive dc voltage of up to 200 V was applied to the z electrodes (not shown in Figure 4b) to form a dc trapping potential well for positively charged ions along the z axis. Internal EI was used to ionize neutral molecules inside the RIT, and an electron gating lens was used to control the time during which electrons from a heated filament were allowed to enter the RIT through one of the x slits. Trapped ions were mass selectively ejected by scanning the rf amplitude at a rate of 16,665 Th/s, except for experiments in which slower rf scan rates were used to improve the resolution. An ac signal, generated by a WaveTek 395 arbitrary waveform generator (WaveTek, San Diego, CA) and amplified by a Balun amplifier, was applied between the x electrodes to provide the dipolar field that facilitated ion ejection during the rf scan and ion excitation during the collision-induced dissociation period. Both rf/dc and stored waveform inverse Fourier transform (SWIFT) waveform isolation42,43 were used for ion isolation in MSn experiments. The SWIFT waveforms were calculated using the ion trap simulation program (ITSIM research version),44 generated using WaveForm DSP2 software (version 2.02) and then deployed with the Wavetek 395-64k 100-MHz arbitrary waveform generator. The ions ejected from the RIT were detected using an electron multiplier, operated at -1800 V with a conversion dynode operated at -5 kV. The signal was first amplified using the preamplifier in the ITMS and then acquired using a digital oscilloscope (model TDS 540; Tektronix Beaverton, OR) at a sampling rate of 250K samples/s. -------------------------------------------------------------------------------- Figure 4 RIT assembly (a) and the RIT instrumentation (b) used for the experiments reported in this paper. RIT with x0 = 5.0 mm, y0 = 5.0-4.0 mm, and z0 = 43.2 mm. An rf signal with a floated dc level is applied to the y pair of electrodes, while the x electrodes are either grounded for boundary ejection or supplied with an ac signal for resonance ejection. -------------------------------------------------------------------------------- For the performance comparison between the RIT and CIT mass analyzers, a previously characterized and optimized CIT45 was used. The CIT was constructed with equal r0 (radius of the ring electrode) and z0 (half-distance between the end caps) of 5.0 mm, a spacing of 1.6 mm between the end cap and ring electrodes, and an end cap hole radius of 0.5 mm. CIT performance was recharacterized for this study using the ITMS instrument system under experimental conditions identical to those used to characterize the RIT. Experimentally, the RIT was replaced by the CIT and the rf coil was tuned to provide an rf amplitude high enough at 1.1 MHz to scan the mass-to-charge ratio out to 650 Th. Helium was used as buffer gas for both the RIT and CIT analyzers at an indicated pressure of ~8.5 × 10-5 Torr, as measured using a Bayert-Alpert-type ionization gauge. The headspace vapor of the organic compounds was leaked into the vacuum through a Granville-Phillips (Granville-Phillips Co., Boulder, CO) leak valve to maintain an indicated pressure of 8.0 × 10-7 Torr. Results and Discussion An RIT with x0 = y0 = 5.0 mm was used to record mass spectra for various compounds. Figure 5 shows the mass spectrum of 1,3-dichlorobezene recorded using boundary ejection (Figure 5a) and resonance ejection (Figure 5b), the latter achieved by using an auxiliary ac dipolar field (512 kHz and 2.9 V0-p) applied during the rf scan. The efficiency of ejection along the x axis was found to be 4 times higher for the resonance ejection experiment than for the boundary ejection based on measurement of the peak area. During the boundary ejection rf scan without application of the offset-dc or the supplementary dipolar ac field, the opportunities for the ions to become unstable along the y and z directions is comparable to the x direction (where the detector is located) and the number of the detected ions is much lower than the number of ions actually trapped in the RIT. To obtain the highest possible ion signal, it is highly desirable to use an auxiliary ac field in the x direction to force ions to be ejected through the slits in the x electrode. Note that this difference in sensitivity does not occur in the Paul trap because in that instrument the potential well in the r direction is deeper (by a factor of 2) than that in the z direction. -------------------------------------------------------------------------------- Figure 5 Mass spectra of 1,3-dichlorobenzene collected using an RIT (x0 = y0 = 5.0 mm and z0 = 43.2 mm) and operating under identical conditions except using (a) boundary ejection and (b) resonance ejection with a supplementary ac of 512 kHz, 2.9 V0-p. 1.1 MHz rf frequency applied. -------------------------------------------------------------------------------- The effects of z-electrode dc voltage on RIT performance were investigated by ionizing acetophenone for 5 ms, using a constant value of the ac amplitude of 2.9 V0-p and a resonance frequency of 512 kHz, but varying the dc voltage applied to the z-electrode. The intensity and the full width at half-maximum (fwhm) of the peak corresponding to the fragment ion m/z 105 of acetophenone was recorded as a function of the z-electrode dc voltage (Figure 6). The abundance increases significantly with the z electrode dc voltage up to ~60 V, presumably because the deeper dc potential well in the z direction helps to trap more ions. The fwhm of the m/z 105 signal decreases with further increases in the z electrode voltage, presumably as a result of ions being forced to the center of the RIT and being ejected within a relatively narrow band along the z axis. This improves the resolution because effects of mechanical errors of machining and assembly are minimized.27 At z-electrode dc voltages above 80 V, the ion abundance also starts to decrease and the improvement in resolution tails off due to the loss of ion trapping capacity as the ions are pushed closer to the center of the RIT. The spectral resolution was further optimized as described later in this paper. -------------------------------------------------------------------------------- Figure 6 z axis dc trapping potential effects on (a) peak intensity of m/z 105 fragment ions from acetophenone and (b) peak full width at half-maximum (fwhm). RIT x0 = y0 = 5.0 mm, z0 = 43.2 mm, ac resonance at 512 kHz and 2.9 V0-p, ionization time 5 ms. -------------------------------------------------------------------------------- Mathieu Stability Diagram. The Mathieu diagram is an important feature of quadrupole ion storage devices. The stability diagram is used to characterize various types of traps and to guide in various ion trap operation modes including tandem mass spectrometry.45,46 In this study, the stability region for RITs of several geometries was mapped (Figure 7) and the characteristic features of the stability diagram were used to select appropriate ion isolation and excitation points for tandem mass spectrometry. The m/z 105 fragment ion of acetophenone was first isolated by using appropriate amplitudes of the rf amplitude and its dc offset to make all other ions unstable (rf/dc isolation). After 10-ms cooling of the isolated ion m/z 105, the amplitude of the rf and its dc offset were varied before a spectrum was collected using a normal rf scan at 0 V dc offset. The stability boundary was taken to be defined by that combination of rf and rf-offset dc voltages at which the signal of the ion m/z 105 generated from acetophenone was just indistinguishable from background (S/N < 2). The stability diagram for the RIT with x0 = y0 = 5.0 mm is shown in Figure 7. As expected, a symmetrical profile is observed for this RIT, and the stability diagram is similar to that for a quadrupole mass filter with round or hyperbolic electrodes. -------------------------------------------------------------------------------- Figure 7 Stability diagram mapped for an RIT with x0 = y0 = 5.0 mm and z0 = 43.2 mm. Fragment ion m/z 105 was isolated and the rf and offset dc voltages were systematically varied. The boundary was taken to be reached when the signal-to-noise ratio is lower than 2. -------------------------------------------------------------------------------- Mass Resolution Improvement. Mass spectra of n-butylbenzene were collected using the RIT, and the peaks of m/z 91 and 92 were used to characterize the mass resolution. For the RIT with equal x0 and y0 of 5.0 mm, peaks m/z 91 and 92 were only partially resolved at the rf scan rate of 16 665 Th/s (Figure 8a) using an z-axis potential of 125 V. It is known that a decrease in the rf scan rate improves the mass resolution for ion traps.47,48 A home-built circuit9 was therefore used to attenuate the rf scan speed by a factor of 2 (8333 Th/s) or 4 (4166 Th/s). Partial mass spectra of n-butylbenzene showing m/z 91 and 92, collected at reduced rf scan rates with a constant z-electrode dc voltage of 125 V, are shown in Figure 8b and c. Improved mass resolution was observed at the slower rf scan rates due to the increased number of increments of the rf voltage in a given mass range and the increased time allowed for ions with adjacent m/z values to be scanned out at the threshold of their instability.9,49 The resolution is found to increase by a factor of 2 for a 4-fold decrease in the rf scan rate. Further improvements are anticipated by improving the mechanical precision of the trap, increasing the z-axis potential, and compensating for the field imperfections introduced by the slits. -------------------------------------------------------------------------------- Figure 8 Mass spectra of n-butylbenzene showing the doublet m/z 91/92, recorded at different rf scan rates using an RIT with x0 = y0 = 5.0 mm, z0 = 43.2 mm, and a z-electrode dc voltage of 125 V. (a) rf scan rate 16 665 Th/s, with applied supplementary ac of 478 kHz, 3.4 V0-p; (b) rf scan rate 8333 Th/s, with applied supplementary ac of 496 kHz, 3.18 V0-p; (c) rf scan rate 4166 Th/s, with applied supplementary ac of 516.5 kHz, 4.9 V0-p. -------------------------------------------------------------------------------- During the development of the CIT, changes in the z0 to r0 ratio were found to be an effective means to vary the mass resolution due to its effects on the higher-order field contributions.19 This method was adopted in this study and the y0 dimension was decreased from 5.0 to 4.2 mm and then to 4.0 mm to produce RITs with stretched (in the x direction) geometries. The relative strengths of the octapolar and dodecapolar fields, as calculated using Poisson, are listed in Table 1, and the analytical performance for the three geometries are compared on the basis of spectral resolution. Mass spectra of n-butylbezene were collected using RITs with x-stretched geometries under experimental conditions identical to those used for the symmetrical geometry both being operated with resonance ejection. The peak overlap of m/z 91 and 92 is compared for the nonstretched and stretched geometries in Figure 9a-c. Significantly improved resolution is observed for the stretched geometries. This phenomenon has been previously reported for Paul traps and for linear traps with hyperbolic electrodes.27,50 The negative octapole field is believed to be the main factor responsible for the ion ejection delays in quadrupole ion traps.51 The stretch in the RIT geometry, which introduces a positive octapolar field as shown in Table 1, is highly desirable for improving RIT performance. The negative dodecapolar fields, which occur for the stretched RITs, as shown in Table 1, were also observed for Paul traps and CITs with optimized geometry.19,40 The higher-order fields are augmented by stretching the trap geometry to compensate for the electric field strength weakening caused by the holes or slits in the electrodes. If not corrected, this phenomenon would allow peak broadening and mass shifts. -------------------------------------------------------------------------------- Figure 9 Partial mass spectrum of n-butylbenzene showing the doublet m/z 91/92, recorded using RITs with the same x0 = 5.0 mm and z0 = 43.2 mm and varying the y0 dimensions; z electrode dc = 125 V. (a) y0 = 5.0 mm, supplementary ac of 478 kHz and 3.4 V0-p,; (b) y0 = 4.2 mm, supplementary ac of 503 kHz and 2.9 V0-p; (c) y0 = 4.0 mm, supplementary ac of 440 kHz and 0.4 V0-p; (d) y0 = 4.0 mm, supplementary ac of 392 kHz and 0.5 V0-p, nonlinear resonance ejection. -------------------------------------------------------------------------------- The frequency of the ac used for resonance ejection was varied over a wide range using each of the geometries, and the change in the resolution was monitored as a function of the ac frequency. A significantly higher resolution was found for the stretched geometries at a resonant frequency around 392 kHz (Figure 9d). This frequency corresponds to a x value of 0.71 and a qx value of 0.81, which is close to the octapolar nonlinear resonance point = 0.70.39,52 The superposition of the higher-order fields inside the trap generates higher harmonic sideband frequencies in addition to the basic secular frequency. At this nonlinear resonance point, one of the higher-harmonic frequencies of the ion motion corresponds to a sideband frequency caused by the presence of the octapolar field. In a previous CIT study, it was found that the delay in the ion ejection and mass shift is significantly decreased at this nonlinear resonance point,52 and the present results are consistent with the same phenomenon in the RIT. Tandem Mass Spectrometry. Like other ion traps, RITs retain the capability of performing multiple-stage tandem mass spectrometry experimentally using the single mass analyzer. Product ion MS/MS spectra were acquired for selected precursor ions generated from a number of compounds, including acetophenone, n-butylbezene, and SF6. The parent ions of interest were first isolated using either SWIFT or rf/dc isolation and subsequently excited to cause collision-induced dissociation using an ac signal of appropriate frequency. The MS3 spectra of acetophenone shown in Figure 10 were recorded with the x-stretched RIT of x0 = 5.0 mm and y0 = 4.0 mm. The molecular radical cation of acetophenone, m/z 120, was isolated using rf/dc isolation. By adjusting the rf voltage and applying a negative dc bias potential to the y electrodes at the same time, the ion of m/z 120 was placed at the upper apex of the stability region. Under these conditions, all other ions except those of m/z 120 follow unstable trajectories and are ejected from the ion trap (Figure 10b). After 3-ms isolation time, the dc offset voltage on the y electrodes was turned off and the rf amplitude was ramped down to bring the mass-selected ion m/z 120 to a x value of 0.33. A resonance excitation ac signal of 183 kHz and 800 mV0-p was applied between the two x electrodes for 30 ms to fragment the selected parent ion of m/z 120. The resonance excitation frequency used was selected by calculating the rf voltage needed for trapping the product ions with a low-mass cutoff (LMCO) of 71 Th. After fragmentation and 10-ms cooling, the fragment ions were mass analyzed using an rf scan with ac resonant ejection and a fragment ion m/z 105 was observed (Figure 10c). The ion m/z 105 was then isolated again using the rf/dc procedure and subsequently fragmented at x = 0.43 (chosen to trap product ions with a LMCO of 74 Th) with a resonance excitation ac at 236 kHz and 630 mV0-p for 30 ms. The resulting sequential product spectrum53 is shown in Figure 10d. Similar MS3 spectra have been acquired for acetophenone using SWIFT waveforms for isolation. Identical fragmentation product ions and efficiency were observed using these two isolation methods. -------------------------------------------------------------------------------- Figure 10 MS3 data for acetophenone collected using an RIT with x0 = 5.0 mm, y0 = 4.0 mm, and z0 = 43.2 mm. (a) Spectrum of acetophenone; (b) molecular ion m/z 120 isolated using rf/dc; (c) product ion spectrum with resonant excitation at 183 kHz and 800 mV; (d) sequential product ion spectrum of isolated m/z 105 with resonant excitation at 236 kHz, 630 mV. -------------------------------------------------------------------------------- Comparisons with the CIT. As a type of linear ion trap, the RIT is expected to provide better performance than a cylindrical ion trap of similar size due to its increased trapping capacity. Comparisons have been made between an RIT of x0 = 5.0 mm, y0 = 4.0 mm, and z0 = 43.2 mm and a CIT of r0 = z0 = 5.0 mm. Note that the volume for the RIT is ~4.4 times that for the CIT. A mass resolution similar to that shown in Figure 9a has been reported previously for this particular CIT.45 Mass spectra of perfluorotributylamine (PFTBA) were collected using these two ion traps by applying basic rf scan functions composed of (i) an ionization at a single fixed rf level, (ii) a 10-ms cooling, and (iii) an rf amplitude scan at a rate of 16 665 Th/s. The LMCO of 45 Th was used for ionization in the scan functions applied to both the RIT and CIT. Due to the different space charge limits for the CIT and RIT, the ionization time was varied to obtain the highest peak intensity for ions in the high-mass range. The spectra with the widest ranges of observed ion masses were collected using an ionization time of 75 s for the RIT (Figure 11a), but 4000 s was required for the CIT (Figure 11b). The overall peak intensity and signal-to-noise ratio were higher for the RIT than for the CIT even though a much shorter ionization time was used for the RIT. In the spectrum collected using the CIT, peaks due to ions with m/z values higher than 300 Th are missing. The trapping efficiency for high-m/z ions is usually lower than that for the low-m/z ions since the rf trapping potential well depth is lower for high-m/z ions.18 The trapping efficiency for the high-m/z ions is generally degraded by the space charge effects, which are partly responsible for the lack of high-m/z ions in the spectrum shown in Figure 11b. In the past, this kind of discrimination effect has been compensated for by using a larger CIT,19 which has a larger trapping capacity, or by using a higher rf frequency, which results in a deeper rf trapping potential well.54 However, the problem is now solved using an RIT of similar dimensions, which has a greater ion trapping capacity and smaller space charge effects when operated using a similar rf trapping potential and frequency. The effects of better trapping efficiency over a wide mass range are shown by comparison of the RIT and CIT mass spectra (Figure 11). Relatively high intensities were obtained with the RIT for the high-mass ions m/z 414, 502, and 614. Mass resolution higher than 1000 (fwhm) was achieved as measured from the fwhm values of these peaks as illustrated for m/z 502. -------------------------------------------------------------------------------- Figure 11 Mass spectrum of PFTBA collected using (a) RIT with x0 = 5.0 mm, y0 = 4.0 mm, and z0 = 43.2 mm, 75 s ionization, resonance ejection at 394 kHz and 1.2 V0-p and (b) cylindrical ion trap with r0 = z0 = 5.0 mm, 4000 s ionization, resonance ejection at 397 kHz and 0.7 V0-p. -------------------------------------------------------------------------------- In another comparison experiment, mass spectra of perfluoropropane (C3F8) were recorded using the RIT and CIT utilizing the same basic scan functions and the same LMCO but using different ionization times. The voltages applied to the conversion dynode and electron multiplier, the signal gain by the current amplifier, and the sample pressure used were all kept constant in making this comparison. The intensity of the fragment ion m/z 69 (arbitrary units but the same for both RIT and CIT) is plotted as a function of ionization time for the RIT and CIT ion traps in Figure 12. The intensity is greater and increases more rapidly with ionization time for the RIT than for the CIT. Better peak shapes and higher mass resolutions were also observed for the RIT over a wide range of ionization times. The slit in the RIT x electrode, in comparison with the hole on the CIT end cap electrode, allows more electrons to enter this trap, resulting in better ionization. However, the electron beam is well focused; therefore, the difference in the ion intensity observed with the similar degree of deterioration in resolution due to space charge effects cannot be explained solely by the difference in the entrance aperture area. To better evaluate quantitatively the trapping capacity of these two traps, the increase in m/z 69 peak intensity was compared in conjunction with the deterioration of the peak shape caused by space charge effects. When comparing spectra collected at 1500-s ionization time to those collected using 500 s or less, peak broadening by a factor of 2 was observed for both the RIT and CIT at 1500 s. However, the ion abundance recorded using the RIT is 40 times higher than that for the CIT, and the mass resolution for the RIT is twice as great as that for the CIT. It is noteworthy that the volume for the RIT is only 4.4 times of that for the CIT. -------------------------------------------------------------------------------- Figure 12 Peak intensity of perfluoropropane (C3F8) m/z 69 as a function of ionization time using RIT (x0 = 5.0 mm, y0 = 4.0 mm, and z0 = 43.2 mm) and CIT (r0 = z0 = 5.0 mm). Sample pressure 1.0 × 10-6 Torr, RIT resonance ejection at 387 kHz and 500 mV, and CIT resonance ejection at 402 kHz and 100 mV. Inset shows the EI mass spectrum of C3F8. -------------------------------------------------------------------------------- In yet another comparison, acetophenone and 1,3-dichlorobenzene were mixed in varying proportions. The headspace vapor of the mixture was leaked into the vacuum system to test sample relative concentrations measurable using the RIT and CIT mass analyzers. The lowest vapor pressure ratio that could be observed using the CIT was 1:100 acetophenone/1,3-dichlorobenzene while in the case of the RIT it was 1:1000. This is consistent with an order of magnitude lower detection limit for the RIT compared to that of the CIT. The increased trapping capacity contributes to the higher ion intensity, better peak shapes, and better detection limit observed with the RIT compared to the CIT using as closely similar conditions as possible. Conclusions When a CIT is miniaturized to operate at a lower rf amplitude and lower power, the r0 and z0 dimensions must be decreased proportionately to retain the analytical performance. However, when size is reduced, the trapping capacity is decreased, and this occurs as the third power of r0.27,55 In a miniaturized RIT, the x0 and y0 dimensions can be decreased allowing the use of lower rf voltages while the resulting loss in trapping capacity can be compensated for by elongation of the z0 dimension. This latter step will not affect the mass range, which is determined by the rf amplitude and the dimensions in the x and y directions. Considerably better performance has been obtained using the RIT as a mass analyzer compared to a CIT of similar size and operated under similar conditions. Capabilities for tandem and multistage mass spectrometry are demonstrated. Mass resolution can be optimized by using slow scans, by optimizing the electrode geometry, and by increasing the z-axis potential well depth through adjusting the dc potential. Many of the advantages of the RIT are the result of increased trapping capacity compared with the CIT. Note that it should be possible to increase the RIT sensitivity of detection by a further factor of 2, as has been done with commercial linear ion traps,27 by using two detectors to detect the ions ejected from the slits of the x electrodes on both sides of the RIT. Some concepts for alternative configurations of the RIT are described elsewhere.30 An RIT with three rf sections can be provided with a dc trapping potential well along the z axis formed by the dc voltages in the three sections. The ions are trapped in the center section to avoid the undesirable effects caused by the fringing field distortion between the z electrodes and the ends of the x and y electrodes. Another future version of the RIT will have slits in the x, y, and z electrodes to allow ions to be injected or ejected along any of the three spatial directions. Mass selective ejection of the ions can be deployed by scanning the rf while applying dc and supplementary ac signals with appropriate amplitudes and frequencies. The direction in which the alternating ac dipolar field is applied determines the direction along which the ions are ejected. With the capabilities demonstrated and the potential improvements that can still be made, the RIT appears to be a very good candidate as a mass analyzer for miniature mass spectrometers. Acknowledgment We thank Ray E. March and John. F. J. Todd for helpful discussions. We acknowledge funding from the Office of Naval Research (N00014-02-0834) and the Indiana 21st Century Fund through a grant to the Indiana Instrumentation Institute. * Corresponding author. Tel: (765) 494-5262. Fax: (765) 494-0239. E-mail: cooks@purdue.edu. Purdue University. Justus-Liebig-Universität Giessen. 1. Badman, E. R.; Cooks, R. G. J. Mass Spectrom. 2000, 35, 659-671. [Medline] 2. Kogan, V. T.; Kazanskii, A. D.; Pavlov, A. K.; Tubol'tsev, Y. V.; Chichagov, Y. V.; Gladkov, G. Y.; Il'yasov, E. I. Instrum. Exp. Technol.-Engl. Tr. 1995, 38, 106-110. 3. Sinha, M. P.; Tomassian, A. D. Rev. Sci. Instrum. 1991, 62, 2618-2620. [CrossRef] 4. Sinha, M., 46th ASMS Conference on Mass Spectrometry and Allied Topics, Orlando, FL, May 31-June 4, 1998. 5. Bryden, W. A.; Benson, R. C.; Ecelberger, S. A.; Phillips, T. E.; Cotter, R. J.; Fenselau, C. Johns Hopkins APL Technol. Dig. 1995, 16, 296-310. 6. Cotter, R. J.; Fancher, C.; Cornish, T. J. J. Mass Spectrom. 1999, 34, 1368-1372. [Medline] 7. Ferran, R. J.; Boumsellek, S. J. Vac. Sci. Technol. A 1996, 14, 1258-1265. [CrossRef] 8. Orient, O. J.; Chutjian, A.; Garkanian, V. Rev. Sci. Instrum. 1997, 68, 1393-1397. [CrossRef] 9. Kaiser, R. E.; Cooks, R. G.; Stafford, G. C.; Syka, J. E. P.; Hemberger, P. H. Int. J. Mass Spectrom. Ion Processes 1991, 106, 79-115. [CrossRef] 10. Miller, G.; Koch, M.; Hsu, J. P.; Ozuna, F., 45th ASMS Conference on Mass Spectrometry and Allied Topics, Palm Springs, CA, June 1-5, 1997. 11. Xu, J.; Whitten, W. B.; Ramsey, J. M. Int. J. Ion Mobility Spectrom. 2002, 5, 207-214. 12. Patterson, G. E.; Guymon, A. J.; Riter, L. S.; Everly, M.; Griep-Raming, J.; Laughlin, B. C.; Ouyang, Z.; Cooks, R. G. Anal. Chem. 2002, 74, 6145-6153.[Full text - ACS] [Medline] 13. Diaz, J. A.; Giese, C. F.; Gentry, W. R. Field Anal. Chem. Technol. 2001, 5, 156-167. [CrossRef] 14. White, A. J.; Blamire, M. G.; Corlett, C. A.; Griffiths, B. W.; Martin, D. M.; Spencer, S. B.; Mullock, S. J. Rev. Sci. Instrum. 1998, 62, 565-571. 15. Syage, J. A.; Nies, B. J.; Evans, M. D.; Hanold, K. A. J. Am. Soc. Mass Spectrom. 2001, 12, 648-655. [Medline] [CrossRef] 16. Spaeder, T. A.; Walton, R. B., 226th ACS National Meeting, New York, September 7-11, 2003. 17. Riter, L. S.; Peng, Y.; Noll, R. J.; Patterson, G. E.; Aggerholm, T.; Cooks, R. G. Anal. Chem. 2002, 74, 6154-6162.[Full text - ACS] [Medline] 18. March, R. E. J. Mass Spectrom. 1997, 32, 351-369. 19. Wells, J. M.; Badman, E. R.; Cooks, R. G. Anal. Chem. 1998, 70, 438-444.[Full text - ACS] 20. Kornienko, O.; Reilly, P. T. A.; Whitten, W. B.; Ramsey, J. M. Rapid Commun. Mass Spectrom. 1999, 13, 50-53. 21. Patterson, G. E.; Grossenbacher, J. W.; Wells, J. M.; Knecht, B. A.; Rardin, B.; Barket, D. J. J., 51st ASMS Conference Mass Spectrometry and Allied Topics, Montreal, Canada, June 8-12, 2003. 22. Riter, L. S.; Meurer, E. C.; Handberg, E. S.; Laughlin, B. C.; H., C.; Patterson, G. E.; Eberlin, M. N.; Cooks, R. G. Analyst 2003, 128, 1112-1118. [CrossRef] 23. Quarmby, S. T.; Yost, R. A. Int. J. Mass Spectrom. 1999, 190/191, 81-102. 24. Schwartz, J. C., 9th Sanibel Conference on Mass Spectrometry, Sanibel Island, FL, 1997. 25. Bier, M. E.; Syka, J. E. P. U.S. Patent 5,420,425, 1995. 26. Hager, J. M. Rapid Commun. Mass Spectrom. 2002, 16, 512-526. 27. Schwartz, J. C.; Senko, M. W.; Syka, J. E. P. J. Am. Soc. Mass Spectrom. 2002, 13, 659-669. [Medline] [CrossRef] 28. Senko, M. W.; Schwartz, J. C.; Wieghaus, A. 51st ASMS Conference on Mass Spectrometry and Allied Topics, Montreal, CA, June 8-12, 2003. 29. Dolnikowski, G. G.; Kristo, M. J.; Enke, C. G.; Watson, J. T. Int. J. Mass Spectrom. Ion Processes 1988, 82, 1-15. [CrossRef] 30. Ouyang, Z.; Cooks, R. G. U.S. Patent applied. 2003. 31. March, R. E., Todd, J. F. J., Eds. Practical Aspects of Ion Trap Mass Spectrometry, Vol. I: Fundamentals of Ion Trap Mass Spectrometry; CRC Press: Boca Raton, FL, 1995. 32. Todd, J. F. J.; Lawson, G.; Bonner, R. F. In Quadrupole Mass Spectrometry and its Applications; Dawson, P. H., Ed., 1976; pp 181-224. 33. Langmuir, D. B.; Monica, R.; Langmuir, R. V.; Altadena; Shelton, H.; Hills, W.; Wuerker, R. F.: U.S. Patent 3,065,640, 1962. 34. Schwartz, J. C.; Syka, J. E. P. U.S. Patent 6,392,225, 2002. 35. Senko, M.: US Patent 6,403, 955, 2002 36. Stafford, G. C.; Kelley, P. E.; Syka, J. E. P.; Reynolds, W. E.; Todd, J. F. J. Int. J. Mass Spectrom. Ion Processes 1984, 60, 85-98. [CrossRef] 37. Billen, J. H.; Young, L. M., Proceedings of the 1993 Particle Accelerator Conference, 1993. 38. Wu, G.; Ouyang, Z.; Plass, W. R.; Cooks, R. G. 50th ASMS Conference on Mass Spectrometry and Allied Topics, Orlando, FL, June 2-6, 2002. 39. Franzen, J.; Gabling, R. H.; Schubert, M.; Wang, Y. In Practical Aspects of Ion Trap Mass Spectrometry; March, R. E., Todd, J. F. J., Eds.; CRC Press: Boca Raton, FL, 1995; Vol. 1, pp 49-167. 40. Wells, J. M.; Plass, W. R.; Patterson, G. E.; Ouyang, Z.; Badman, E. R.; Cooks, R. G. Anal. Chem. 1999, 71, 3405-3415.[Full text - ACS] 41. Louris, J. N.; Cooks, R. G.; Syka, J. E. P.; Kelley, P. E.; Stafford, G. C.; Todd, J. F. J. Anal. Chem. 1987, 59, 1677-1685. 42. March, R. E.; Hughes, R. J. In Quadrupole Storage Mass Spectrometry; John Wiley and Sons: New York, 1989; p 365. 43. Guan, S. H.; Marshall, A. G. Anal. Chem. 1993, 65, 1288-1294. 44. Plass, W. R. Ph D. Thesis, Justus-Liebig-Universität, Germany, 2001. 45. Ouyang, Z.; Badman, E. R.; Cooks, R. G. Rapid Commun. Mass Spectrom. 1999, 13, 2444-2449. [Medline] 46. Todd, J. F. J.; Waldren, R. M.; Mather, R. E.; Lawson, G. Int. J. Mass Spectrom. Ion Phys. 1978, 28, 141-151. [CrossRef] 47. Schwartz, J. C.; Syka, J. E. P.; Jardine, I. J. Am. Soc. Mass Spectrom. 1991, 2, 198-204. [CrossRef] 48. Williams, J. D.; Cox, K. A.; Cooks, R. G.; Kaiser, R. E.; Schwartz, J. C. Rapid Commun. Mass Spectrom. 1991, 5, 327-329. 49. Makarov, A. A. Anal. Chem. 1996, 68, 4257-4263.[Full text - ACS] 50. Syka, J. E. P. In Practical Aspects of Ion Trap Mass Spectrometry; March, R. E., Todd, J. F. J., Eds.; CRC Press: Boca Raton, FL, 1995; Vol. 1, p 169. 51. Plass, W. R.; Li, H.; Cooks, R. G. Int. J. Mass Spectrom. 2003, 228, 237-267. [CrossRef] 52. Wells, J. M.; Plass, W. R.; Cooks, R. G. Anal. Chem. 2000, 72, 2677-2683.[Full text - ACS] 53. Schwartz, J. C.; Wade, A. P.; Enke, C. G.; Cooks, R. G. Anal. Chem. 1990, 62, 1809-1818. [Medline] 54. Tabert, A.; Misharin, A. S.; Cooks, R. G. Analyst 2004, 129, 323-330. [CrossRef] 55. Campbell, J. M.; Collings, B. A.; Douglas, D. J. Rapid Commun. Mass Spectrom. 1998, 12, 1463-1474. 56. Bui, H. A.; Cooks, R. G. J. Mass Spectrom. 1998, 33, 297-304. Table 1. Multipole Expansion Coefficients for Different RIT Geometries Calculated Using CreatePota geometryb quadrupolar octapolar dodecapolar y0 x0/y0 A2 A4 A4/A2 (%) A6 A6/A2 (%) 5.0 mm 1.0 0.541 -0.036 -6.6 -0.052 -9.6 4.2 mm 1.19 0.633 0.025 10.4 -0.098 -15.5 4.0 mm 1.25 0.654 0.052 7.9 -0.113 -17.2 a Program in the ITSIM suite,56 which calls Laplace solver from Poisson/Superfish.37 b Slit widths for the different RITs were 1.0 mm and calculations were done including the slit effects. |
|
Want more information ? Interested in the hidden information ? Click here and do your request. |