Anal. Chem. 2004, 76, 2531-2536
Characterization of a Focal Plane Camera Fitted to a Mattauch-Herzog Geometry Mass Spectrograph. 2. Use with an Inductively Coupled Plasma James H. Barnes, IV,† Gregory D. Schilling,† Roger Sperline,‡ M. Bonner Denton,‡ Erick T. Young,§ Charles J. Barinaga,| David W. Koppenaal,| and Gary M. Hieftje*,†
Department of Chemistry, Indiana University, Bloomington, Indiana 47405, Department of Chemistry and Steward Observatory, University of Arizona, Tucson, Arizona 85721, and Pacific Northwest National Laboratory, Richland, Washington 99352
A novel charge-sensitive detector array, termed the focal plane camera (FPC), has been coupled to a MattauchHerzog mass spectrograph (MHMS) with an inductively coupled plasma ionization source. The FPC employs an array of gold Faraday cups, each with its own chargeintegrating circuit that allows the simultaneous detection of several m/z ratios. The ion-sampling interface of the MHMS has been redesigned to provide better heat transfer away from the sampler and skimmer cones and to reduce the negative effects of turbulent gas flows around the plasma. The instrument has produced limits of detection in the tens to hundreds of parts per quadrillion regime and isotope ratio accuracy and precision of 5% error and 0.007% RSD, respectively. Limits of detection with the FPC are comparable to those obtained with a single-channel secondary electron multiplier (SEM). However, the isotope ratio accuracy and precision are better with the FPC than when the SEM is employed. The dynamic range has been shown to be linear over 7 orders of magnitude. Since the initial coupling of an inductively coupled plasma (ICP) to a mass spectrometer,1 the field of elemental analysis has experienced a shift from traditional analysis techniques such as atomic absorption and flame emission spectroscopy to ICPMS. This shift is due to the numerous beneficial characteristics of the ICP. The ICP is a very robust ionization source, capable of handling high levels of solvent or solid material without significant degradation in its operating characteristics. This property arises from the annular shape of the plasma, which permits samples to be introduced axially through the central channel with little influence on energy coupling into the plasma. Additionally, the annular shape aids in sample desolvation and volatilization. Because of this property, and a high electron number density and gas kinetic temperature, typically on the order of 1015 cm-3 and * To whom correspondence should be addressed. E-mail:
[email protected]. † Indiana University. ‡ Department of Chemistry, University of Arizona. § Steward Observatory, University of Arizona. | Pacific Northwest National Laborator. (1) Houk, R. S.; Fassel, V. A.; Flesch, G. D.; Svec, H. J.; Gray, A. L.; Taylor, C. E. Anal. Chem. 1980, 52, 2283-2289. 10.1021/ac030337u CCC: $27.50 Published on Web 03/27/2004
© 2004 American Chemical Society
6000 K, respectively,2 most elements are ionized with near-unity efficiency in the ICP, a benefit leading to high sensitivities and a fairly uniform multielemental response. A final benefit of the ICP is that ions are formed in a relatively inert environment, leading to limited molecular interferences. One of the most significant limitations of the ICP is the noise level it introduces into a measurement. Although there are numerous contributors to this noise, the majority is correlated or multiplicative in nature. Such a noise source varies with time and affects all signals, or m/z ratios, identically and in a manner proportional to the signal amplitude. Mathematically, the correlated-noise modified signal (Sn) can be expressed as
Sn ) N(t)S0
(1)
where S0 is the original signal and N(t) is the correlated noise function. To eliminate this noise from a measurement, one can attempt to determine N(t) and divide the noise-modified signal by the correlated noise function to obtain the original signal. This is often an uncertain and time-consuming process. Another method to eliminate correlated noise is through a ratio method. Since N(t) varies with time and not with signal level, the correlated noise contribution can be eliminated by recording two m/z signals (Sn1 and Sn2) simultaneously and taking the ratio of them. This process can be expressed mathematically for two signals, S01 and S02:
Sn1 Sn2
)
N(t)S01 N(t)S02
)
S01 S02
(2)
To take advantage of this method of noise reduction, virtually simultaneous detection must be employed. Since m/z ratios are measured at different times in scanning mass spectrometers, it is not possible to completely eliminate this source of noise. At best, noise can be reduced by hopping back and forth between two m/z values, ideally at a rate that is fast compared to the highest multiplicative noise frequency. Additionally, since simul(2) Lehn, S. A.; Warner, K. A.; Huang, M.; Hieftje, G. M. Spectrochim. Acta, Part B 2002, 57B, 1739-1751.
Analytical Chemistry, Vol. 76, No. 9, May 1, 2004 2531
taneous extraction methods such as time-of-flight MS (TOFMS) and ion trap mass spectrometry (ITMS) really detect ions in a sequential fashion, correlated noise that affects the analyzer portion of the instrument cannot be eliminated through a ratio procedure, though ICP source noise can be reduced. Simultaneous detection has several other benefits as well. By detection of all m/z ratios at once, the duty cycle of the measurement increases to unity. In scanning-based measurements, such as those performed with quadrupole or many sectorfield mass spectrometers (SFMS), the duty cycle (DC) is approximated by
DC ) 1/n
(3)
where n is the number of isotopes that must be measured. In reality, moreover, the duty cycle is lower, due to the finite length of time that is required to move between m/z values. Similarly, in TOFMS, the duty cycle is significantly less than unity, on the order of 10%, since the technique is inherently pulsed and one must wait for a full spectrum to be acquired before the next ion injection pulse occurs.3 The duty cycle in TOFMS can be approximated by
DC ) tfill/tms
(4)
where tfill is the time required for ions to fill the extraction region of the TOFMS and tms is the time required to record a mass spectrum. When the duty cycle is improved to unity, the volume of sample necessary for a complete analysis, as well as the total time needed, decreases. These benefits combine to improve absolute (mass-based) detection limits. A final benefit of simultaneous detection is the ability to accurately monitor fast transient signals. When speciation is desired, for example, separation methods such as gas chromatography or capillary electrophoresis are often coupled to ICPMS. In modern separation methods, the analyte concentration can change rapidly with time. With scanning-based methods, there will be an error in the concentration measurement if a mass spectrum is not acquired fast enough. This phenomenon, known as spectral skew, can be eliminated either by simultaneous extraction of a packet of ions, as can be accomplished with TOFMS, FT-ICR, and ITMS, or through simultaneous detection. Clearly, there is a need to develop improved mass spectrometers capable of simultaneous detection. In turn, the most direct route to simultaneous detection is to use a spatially dispersive mass spectrometer. Of the available mass spectrometer geometries, magnetic sector instruments are the only spatially dispersive instruments and, therefore, the best ones capable of performing truly simultaneous and rapid detection. A review of SFMS geometries can be found in the literature.4 In SFMS, instruments with a magnetic sector before an electrostatic analyzer (ESA) are known as reverse geometry, while instruments with the ESA before the magnetic sector are known as forward geometry. Only forward-geometry SFMS instruments focus a broad mass range at once and are therefore capable of simultaneous detection. This (3) Dawson, J. H. J.; Guilhaus, M. Rapid Commun. Mass Spectrom. 1989, 3, 155-159. (4) Burgoyne, T. W.; Hleftje, G. M. Mass Spectrom. Rev. 1997, 15, 241-259.
2532
Analytical Chemistry, Vol. 76, No. 9, May 1, 2004
group of acceptable geometries is limited further by the need to have a flat focal plane and the requirement to simultaneously focus both ion energy and mass. Only a few forward geometries, such as the Matsuda5,6 and Mattauch-Herzog,7 possess flat focal planes and simultaneously focus both ion energy and mass. At Indiana University, a Mattauch-Herzog geometry mass spectrograph (MHMS) has been developed for use as a simultaneous detection mass spectrometer. The initial work with the MHMS focused on construction and improving its analytical performance.4,8-11 Once the MHMS was operational at a satisfactory level, several ionization sources were characterized with a conventional single-channel mass spectrometry detector, a continuous-dynode secondary electron multiplier (SEM), to establish a baseline figure-of-merit set.12-14 Next, a conventional mass spectrometry array detector, the electrooptical imaging detector,15 was installed in the instrument. Results from this coupling were significantly poorer than those obtainable with the SEM.16 Recently, a new mass spectrometry array detector, termed the focal plane camera (FPC), has been coupled to the MHMS.17 With a glow discharge ionization source, figures of merit were comparable to those obtained with the SEM and much improved over those obtained with the EOID.18 Though a detailed description of the FPC can be found in ref 17, a brief summary will be given here. The FPC consists of an array of 31 Faraday cups, each electrically connected to its own high-gain integrating amplifier. With this configuration, up to 16 m/z values can be simultaneously resolved and detected. Although the array of Faraday cup/amplifiers integrates ion current for all channels simultaneously, they are interrogated sequentially by means of a low-noise multiplexer unit. Each Faraday cup is a gold pad, 145 µm × 5 mm, that has been fabricated by standard lithographic techniques on a glass substrate. Due to initial design considerations, some of the Faraday cups were divided in half or in thirds to test the effects of Faraday cup length on the integrator response. Since the collection efficiency of a Faraday cup is approximately equal to its area, two-thirds of each cup was masked off, to give each cup an effectively equal collection area. The amplifier circuit is described as a capacitive transimpedance amplifier. Simply put, the gain of the amplifier is inversely (5) Matsuda, H. J. Phys. Soc. Jpn. 1956, 11, 183-191. (6) Matsuda, H. Int. J. Mass Spectrom. Ion Phys. 1974, 14, 219-233. (7) Mattauch, J.; Herzog, R. Z. Phys. 1934, 89, 786-795. (8) Burgoyne, T. W.; Hieftje, G. M.; Hites, R. A., Proc. 41st ASMS Conf., San Francisco, 1993; WP200. (9) Burgoyne, T. W.; Hieftje, G. M.; Hites, R. A. J. Anal. At. Spectrom. 1997, 12, 1149-1153. (10) Burgoyne, T. W.; Hieftje, G. M.; Hites, R. A. J. Am. Soc. Mass Spectrom. 1997, 8, 307-318. (11) Burgoyne, T. W.; Hieftje, G. M.; Hites, R. A. Anal. Chem. 1997, 69, 485489. (12) Barnes, J. H., IV; Gron, O. A.; Hieftje, G. M. J. Anal. At. Spectrom. 2002, 17, 1132-1136. (13) Solyom, D. A.; Gron, O. A.; Barnes, J. H., IV; Hieftje, G. M. Spectrochim. Acta, Part B 2001, 56B, 1717-1729. (14) Solyom, D. A.; Hieftje, G. M. J. Anal. At. Spectrom. 2002, 17, 329-333. (15) Giffin, C. E.; Boettger, H. G.; Norris, D. D. Int. J. Mass Spectrom. Ion Phys. 1974, 15, 437-449. (16) Solyom, D. A.; Barnes, J. H., IV; Gron, O. A.; Hieftje, G. M. Spectrosc. Spectral Anal. 2002, 22, 828-837. (17) Knight, A. K.; Sperline, R. P.; Hieftje, G. M.; Young, E.; Barinaga, C. J.; Koppenaal, D. W.; Denton, M. B. Int. J. Mass Spectrom. 2002, 215, 131139. (18) Barnes, J. H., IV; Sperline, R.; Denton, M. B.; Barinaga, C. J.; Koppenaal, D.; Young, E. T.; Hieftje, G. M. Anal. Chem. 2002, 74, 5327-5332.
Figure 1. Newly designed MHMS interface specifically designed for ICP ion sampling.
proportional to the size of a feedback capacitor. The FPC was designed to have a feedback capacitor of 36 fF, leading to an output voltage of 4.4 µV per incident singly charged ion. The following text will describe the coupling of an ICP source to the MHMS-FPC instrument. Details will be given on the implementation of a newly designed ICP sampling interface and extraction optic. Additionally, figures of merit, including limits of detection, isotope ratio precision and accuracy, and dynamic range will be presented and compared to results obtained with an SEM. EXPERIMENTAL SECTION Mattauch-Herzog Mass Spectrograph. An in-house-built MHMS was used in all experimentation and has been described elsewhere.8-11 In the present study, the MHMS was equipped with a locally fabricated three-stage differentially pumped interface designed specifically for use with an ICP source. This interface has been substantially modified from that used previously with this instrument.13 In the previous design, the sampler plate leading to the first vacuum stage incorporated both the sampling orifice and a cooling channel. As a result, changing of the sampling cone was not possible without replacing the entire sampler plate. Additionally, the surface of the sampler plate was not planar, leading to a recirculating gas flow directed back toward the ICP and destabilizing it. Moreover, this earlier design did not have an integrated cooling channel for the skimmer cone. The newly designed interface, shown in Figure 1, incorporates several improvements. The sampler plate has been divided into two separate parts: the main body, which houses a cooling channel, and a sampling cone insert. The main body has been constructed from copper, which provides good heat transfer from the sampling cone to the cooling channel. In the past, we have experienced problems with similar two-piece assemblies caused by the thermal and electrical isolation of the two components by the O-ring seal. Limited thermal conductivity throughout the assembly manifests itself in short cone lifetimes and O-ring degradation. The new design eliminates those problems by moving the O-ring seal nearer the cooling channel and far from the ICP. Additionally, the sampling cone inserts have been machined to fit into the assembly with a clearance of ∼10 µm. The thermal expansion of the inserts during operation of the ICP is sufficient to ensure good thermal contact between the two pieces. This contact also improves electrical continuity between the two pieces. Moreover, to ensure electrical continuity, the two pieces have been physically connected by means of a thin copper sheet placed between them. Another benefit of the new interface
Figure 2. Redesigned EQEL optic composed of an inner, slotted cylindrical element (a) and an outer solid cylinder (b). The EQEL optic has been modified so less of the ion beam is intercepted by the inner unit of the assembly (c).
is that the front surface of the sampler plate is planar, which results in less recirculating gas flow near the ICP. The final advantage of this design is the incorporation of a cooling channel in the skimmer-cone mount, which aids in cooling and prolongs the life of the cone. Electrostatic Quadrupole Extraction Lens. A new extraction optic, termed the electrostatic quadrupole extraction lens (EQEL), has recently been reported for use in the MHMS.19 Although a similar optic has been reported in the literature for use as an ion guide,20 the EQEL is the first attempt to use the device as an extraction element. This lens consists of a modified conventional extraction optic surrounded by an electrically isolated barrel unit. The conventional extraction optic was modified by machining into it two sets of slots, each consisting of four rectangular holes spaced 90° apart, with the two sets offset by 45°. A problem with the initial EQEL design19 was that the tapered tip of the slotted inner cylinder intercepted a large portion of the ion beam. To alleviate this problem, the tapered inner cylinder was replaced with a similarly slotted straight cylinder, as shown in Figure 2. Inductively Coupled Plasma. The ICP is housed in a locally built copper torch box that effectively shields the rest of the instrument from the rf and photon radiation emitted by the plasma. Power is delivered to the ICP from a PlasmaTherm (St. Petersburg, FL) 2500D power supply and 2500E impedance matching unit. To reduce the plasma offset potential and eliminate a secondary discharge, a copper torch shield is employed between the torch and load coil.21,22 The maximum analyte signal was (19) Barnes, J. H., IV; Schilling, G. D.; Denton, M. B.; Koppenaal, D. W.; Hieftje, G. M. J. Anal. At. Spectrom. 2003, 18, 1015-1018. (20) Takada, Y.; Sakairi, M.; Ose, Y. Rev. Sci. Instrum. 1996, 67, 2139-2141. (21) Nonose, N. S.; Matsuda, N.; Fudagawa, N.; Kubota, M. Spectrochim. Acta, Part B 1994, 49B, 955-974. (22) Appelblad, P. K.; Rodushkin, I.; Baxter, D. C. J. Anal. At. Spectrom. 2000, 15, 359-364.
Analytical Chemistry, Vol. 76, No. 9, May 1, 2004
2533
obtained at a sampling depth of 12-mm ALC with central, intermediate, and outer argon flows of 1.20, 1.35, and 16.0 L min-1, respectively. The flow of argon (99.998%, Air Products, Allentown, PA) was controlled with MKS mass flow controllers (1179A, Andover, MD) and digital readout (247A). Sample Introduction. Aqueous samples were introduced into the ICP by a CETAC ultrasonic nebulizer (USN) (U-6000AT+, Omaha, NE). The evaporation and condensation portions of the USN were maintained at 160 and 5 °C, respectively. A 1-m heated transfer line, operating at 150 °C, was used to ensure complete aerosol evaporation before the plasma. The sample introduction and drain lines were regulated with a four-head peristaltic pump (RP-1, Rainin, Emeryville, CA), operating at a sample uptake of 0.80 mL min-1. To reduce pulsations caused by the action of the peristaltic pump, the unit was operated with small-bore tubing (0.76-mm i.d., Rainin) and a high pump speed (17.5 rpm). With this pump head and pump speed, pulsations occurred every 0.3 s, significantly shorter than the integration times used for data acquisition. Analytical Standards. Analytical standards were produced by serial dilution of 100 µg mL-1 multielemental standards obtained from SPEX Certiprep (Metuchen, NJ). Dilutions were performed using 0.1 M HNO3 prepared from reagent-grade nitric acid (70%, EM Science, Gibbstown, NJ). Before use, the reagent-grade nitric acid was purified by sub-boiling distillation in a Teflon still (SBS108, Savillex, Minnetonka, MN). For dynamic range determination, single-element solutions were prepared by serial dilution of singleelement atomic absorption standards (1000 µg mL-1, Aldrich, Milwaukee, WI). All results reported here have been backgroundsubtracted against the 0.1 M HNO3 blank. Data Acquisition. When the MHMS was operated in scanning mode, ions were detected with a Channeltron model 4772 continuous-dynode SEM (Burle, Sturbridge, MA) operating in analog mode. With the continuum background observed in this instrument, little benefit is gained through pulse counting. The signal from the SEM was amplified with a fast current preamplifier (SRS570, Stanford Research Systems, Sunnyvale, CA) and recorded with LabVIEW 6.1 (National Instruments, Austin, TX) software and a National Instruments data acquisition system (PCI6035E). For limit-of-detection (LOD) determinations, 10 replicate measurements, each 1 s in length, were made for both the signal and blank and summed for a 10-s integration period. For both the FPC and SEM data, the m/z peak of interest was centered on either a particular Faraday cup or the SEM. The LOD values were defined as the analyte concentration necessary to produce a signal-to-noise ratio of 3. Since the majority of each m/z peak spanned only a single Faraday cup, the signal for a given m/z peak was taken as the Faraday cup response associated with the m/z peak of interest. The rms background noise level was determined on peak for each m/z by using the 0.1 M HNO3 blank. Pertinent information regarding the operation of the data acquisition systems, as well as the rest of the instrument, can be found in Table 1. RESULTS AND DISCUSSION Limits of Detection. To determine limits of detection with both the FPC and SEM, 10 integrations of 1 s were performed with the m/z of interest centered either on a Faraday cup, in the case of the FPC, or on the exit slit, in the case of the SEM. All 2534 Analytical Chemistry, Vol. 76, No. 9, May 1, 2004
Table 1. Operating Conditions of the ICP-MHMS-FPC sample introduction system sample uptake rate USN chamber temperature USN condenser temperature inductively coupled plasma source forward power reflected power central argon flow intermediate argon flow outer argon flow sampling depth mass spectrograph sampler-cone aperture skimmer-cone aperture third-stage aperture first-stage pressure second-stage pressure third-stage pressure entrance slit width acceleration potential secondary electron multiplier operating potential exit slit width exit slit height focal plane camera collector width collector height active collectors operating temperature
0.80 mL min-1 160 °C 5 °C 1.25 kW
