LINEAR QUADRUPOLES IN MASS SPECTROMETRY D.J. Douglas* Department of Chemistry, University of British Columbia, 2036 Main Mall, Vancouver, BC, Canada V6T 1Z1 Received 4 October 2008; received (revised) 14 November 2008; accepted 14 November 2008 Published online 2 June 2009 in Wiley InterScience (www.interscience.wiley.com) DOI 10.1002/mas.20249 The use of linear quadrupoles in mass spectrometry as mass filters and ion guides is reviewed. Following a tutorial review of the principles of mass filter operation, methods of mass analysis are reviewed. Discussed are extensions of quadrupole mass filters to higher masses, scanning with frequency sweeps of the quadrupole waveform, operation in higher stability regions, and operation with rectangular or other periodic waveforms. Two relatively new methods of mass analysis the use of ‘‘islands of stability’’ and ‘‘mass selective axial ejection’’ are then reviewed. The optimal electrode geometry for a quadrupole mass filter constructed with round rods is discussed. The use of collisional cooling in quadrupole ion guides is discussed along with ion guides that have axial fields. Finally, mass analysis with quadrupoles that have large distortions to the geometry and fields is discussed. An Appendix gives a brief tutorial review of definitions of electrical potentials and fields, as well as the units used in this article. # 2009 Wiley Periodicals, Inc., Mass Spec Rev 28:937–960, 2009 Keywords: linear; quadrupole; mass filter; resolution; stability; mass scans; quadrupole excitation; ion guide; axial fields; distorted fields; multipoles I. INTRODUCTION The linear quadrupole, originally developed for mass spectrometry by Paul and co-workers (Paul, Reinhard, & von Zahn, 1958), is widely used in many areas as a mass filter, an ion guide, or a linear ion trap. The history of the first 21 years of quadrupole devices has been given by Dawson (1976, Ch. 1). His classic book describing theory and applications of quadrupoles, both two-dimensional and three-dimensional, was published more than 30 years ago (Dawson, 1976). Since that time, there has been much progress in the development and use of quadrupoles in new ways. The book by March and Hughes (1989), while mostly concerned with three-dimensional ion traps, contains much useful information about the properties of ion motion in quadrupole fields that can be directly applied to two-dimensional linear quadrupoles. The use of quadrupoles as linear ion traps was reviewed by Douglas, Frank, and Mao (2005). This article attempts to review the use of linear quadrupoles as mass analyzers and ion guides. Section II is a review of the ———— Contract grant sponsor: Natural Sciences and Engineering Research Council of Canada; Contract grant sponsor: MDS-Analytical Technologies. *Correspondence to: D.J. Douglas, Department of Chemistry, University of British Columbia, 2036 Main Mall, Vancouver, BC, Canada V6T 1Z1. E-mail: douglas@chem.ubc.ca Mass Spectrometry Reviews, 2009, 28, 937– 960 # 2009 by Wiley Periodicals, Inc. equations of ion motion and stability regions. Section III describes the use of linear quadrupoles as mass analyzers with conventional and new methods of operation including the development of high mass range quadrupoles, scanning with frequency sweeps, operation in higher stability regions, and operation with rectangular waveforms. Two new methods of mass analysis the use of ‘‘islands of stability’’ and axial ejection are then reviewed. The optimal electrode geometry for a quadrupole mass filter constructed with round rods is discussed. The use of collisional cooling in quadrupole ion guides is discussed along with ion guides that have axial fields. Section IV discusses mass analysis with quadrupoles that have large distortions to the geometry and fields. An Appendix gives a brief review of definitions of electrical potentials and fields, as well as the units used in this article. Table 1 shows the units used in this article. Many areas of quadrupole research could not be reviewed here. These include special applications of quadrupoles, development of hybrid instruments, novel electrode geometries, and micro-machined and miniature quadrupoles. Mathematical methods, such as matrix methods to calculate stability diagrams and boundaries, properties of solutions of the Mathieu equation, and mathematical methods for calculating ion motion in distorted fields are not covered. The patent literature is not generally included. Relevant publications up to June 2008 have been considered. II. THE LINEAR QUADRUPOLE A. The Quadrupole Potential The electric potential of the linear quadrupole, f(x, y), is given by 2 x y2 0 ð1Þ ðx; yÞ ¼ r02 where x and y are Cartesian co-ordinates. This potential is produced by four parallel electrodes with hyperbolic shapes, as shown in Figure 1. The point (x, y) ¼ (0, 0) is the center of the quadrupole. The parameter r0, called the ‘‘field radius,’’ is the distance from the center to an electrode. Balanced potentials are normally applied between the x and y electrodes, that is, a potential þf0 is applied to the two electrodes in the x direction and a potential f0 is applied to the two electrodes in the y direction. In this case the potential at the center of the quadrupole is 0. It is readily verified that the potential of Equation (1) satisfies Laplace’s equation. In fact, for a two-dimensional potential where electric fields vary linearly with position (the simplest variation with position) Equation (1) can be derived from Laplace’s equation (Dawson, 1976, pp. 9–10). & DOUGLAS TABLE 1. SI units used in this article B. Other Two-Dimensional Potentials If a quadrupole does not have an ideal field with the potential given by Equation (1), the mathematical description of the potential is more complicated. The potential can be described as a superposition of two-dimensional multipole fields. The quadrupole potential is just one of the many possible solutions to Laplace’s equation in two dimensions. In general, the spatial part of any two-dimensional potential F(r, y) can be written in polar co-ordinates (r, y) as a sum of two-dimensional multipoles as follows: 1 X ’N ðr; Þ ð3Þ Fðr; Þ ¼ ½ðA þ BÞðC ln r þ DÞ þ N¼1 The terms jN(r, y) are referred to as ‘‘spatial harmonics’’ or ‘‘circular harmonics’’ (Smythe, 1939, p. 62) and are given in the most general form when N 1 as ’N ðr; Þ ¼ ðAN cos N þ BN sin NÞðCN rN þ DN r N Þ If unbalanced voltages are applied to the x and y rods a quadrupole potential is still formed, but there is an axis potential. Suppose the magnitude of the voltage applied to the x rods is f0x and to the y rods f0y. The potential can be written 2 0x þ 0y x y2 0x 0y ðx; yÞ ¼ ð2Þ þ 2 2 r02 where ((f0x f0y)/2) is the axis potential. Round electrodes are often used to provide an approximate quadrupole field because it is less expensive to manufacture round rods with high precision. The optimum choice of the ratio of the radius of a round rod r to the field radius r0 is discussed in Section IIIJ. ð4Þ For the field within a linear quadrupole only the terms in rN contribute, so all DN are zero and without loss of generality the spatial harmonics can be taken as N r ’N ðr; Þ ¼ cos N ð5Þ r0 In Cartesian coordinates, used in this review, a two-dimensional potential can be expanded in multipoles fN(x, y) as Fðx; yÞ ¼ þ1 X AN N ðx; yÞ ð6Þ N¼0 where AN is the dimensionless amplitude of the two-dimensional multipole fN(x, y). Each of the two-dimensional multipoles is also a solution to Laplace’s equation. The analytical form of a two-dimensional multipole can be calculated from N ðx; yÞ ¼ Imððx þ iyÞN Þ ð7Þ N ðx; yÞ ¼ Reððx þ iyÞN Þ ð8Þ or where Im(f(z)) is the imaginary part of the complex function f(z), Re(f(z)) is the real part of the complex function f(z), and i2 ¼ 1 (Smythe, 1939, p. 70; Feynmann, Leighton, & Sands, 1963). The set Re(x þ iy)N can be used without loss of generality. The term f0 represents a potential that is independent of position, f1 a dipole potential, f2 a quadrupole potential (Eq. 1), f3 a hexapole potential, f4 an octopole potential, and so on. Analytical forms of the two-dimensional multipoles have been given by Szylagyi (1988). The analytical forms of the first five multipoles are ð9Þ 0 ¼ constant 1 ðx; yÞ ¼ FIGURE 1. The quadrupole electrodes with applied potentials, showing equipotential lines. Reproduced from Du, Douglas, and Konenkov (1999a). Copyright the Royal Society of Chemistry. 938 2 ðx; yÞ ¼ x r0 x2 y2 r02 ð10Þ ð11Þ Mass Spectrometry Reviews DOI 10.1002/mas LINEAR QUADRUPOLES IN MS 3 ðx; yÞ ¼ 4 ðx; yÞ ¼ x3 3xy2 r03 x4 6x2 y2 þ y4 r04 ð12Þ & independent because the potential does not contain terms of the form xmyn. As a result, ion motion in the x and y directions can be considered independently. ð13Þ The hexapole potential corresponds to the potential in a region between six rods, the octopole potential to the potential in the region between eight rods, and so on. The multipole fN(x, y) gives the potential from an array of 2N electrodes with shapes given by N ðx; yÞ ¼ 1 2. Ion Motion and Stability in the x Direction The equation of ion motion for the x direction can be written as m ð21Þ Introducing the variables ð14Þ The electrode geometries and applied potentials producing these multipoles are shown by Szabo (1986). d2 x 2zex ¼ 2 ðU VRF cos OtÞ dt2 r0 x¼ Ot ; 2 ax ¼ 8zeU ; mr02 O2 qx ¼ 4zeVRF mr02 O2 ð22Þ Equation (21) becomes d2 x þ ðax 2qx cos 2xÞx ¼ 0 dx2 C. Equations of Motion and the Stability Diagram 1. Electric Fields in the x and y Directions For mass analysis, the potential applied to the quadrupole rods has both DC and time-varying parts, and is given by 0 ðtÞ ¼ U VRF cos Ot A typical operating frequency is approximately 1.0 MHz (O ¼ 2 p 1 106 s1) which is a radio frequency (RF). Equation (16) has a periodic cosine variation of the potential with time. In fact, any periodic function may be used (see below). Ion motion is determined by Newton’s law d~ u ~ f ¼m dt ð17Þ where ~ f is the force on an ion, m is the mass of the ion,~ u is the ion velocity, and t is the time. The force on a positive ion is ~ f ¼ zerFðx; y; tÞ ð18Þ where z is the number of charges on the ion. The electric field in the x direction is @Fðx; y; tÞ 2x ¼ 2 ðU VRF cos OtÞ Ex ¼ @x r0 @Fðx; y; tÞ 2y ¼ 2 ðU VRF cos OtÞ @y r0 ð20Þ The electric field in x depends only on x and the electric field in y depends only on y. The electric fields in the x and y directions are Mass Spectrometry Reviews DOI 10.1002/mas !n ¼ ð2n þ Þ O 2 ð24Þ where n ¼ 0, 1, 2,. . . and b is a function of the ax and qx parameters. Many of the methods of operating linear quadrupoles can be understood in terms of the stability of ion trajectories and stability diagrams. Solutions to the Mathieu equation are classified as ‘‘stable’’ or ‘‘unstable.’’ Ions oscillate in the x direction in a complex way. A solution is ‘‘stable’’ if the amplitude of oscillation remains finite as t ! 1. Conversely, a solution is ‘‘unstable’’ if the amplitude of oscillation increases exponentially with time. In this case the ion will eventually have an oscillation amplitude equal to r0 and will strike a rod. Examples of stable and unstable trajectories are given by Dawson (1976, p. 17). The stability of ion motion depends on the parameters ax and qx. Combinations of ax and qx that give stable ion motion in the x direction are shown by the shaded regions labeled ‘‘stable x’’ in Figure 2. 3. Ion Motion and Stability in the y Direction The equation of ion motion for the y direction is m ð19Þ and the electric field in the y direction is Ey ¼ Equation (23) is the Mathieu equation (Émile Léonard Mathieu, 1835–1890). Ion motion consists of oscillations with combinations of many different frequencies. The angular frequencies of ion oscillation are given by ð15Þ where U is a DC voltage applied pole to ground, and VRF is a zero to peak alternating voltage applied pole to ground. In this case the quadrupole potential is 2 x y2 ðU VRF cos OtÞ ð16Þ Fðx; y; tÞ ¼ r02 ð23Þ d2 y 2zey ¼ þ 2 ðU VRF cos OtÞ dt2 r0 ð25Þ Equation (25) is similar to Equation (21), but differs in the sign of the right side (because of the different signs of the potential in x and y, Eq. 1). Equation (25) can also be written as the Mathieu equation with the change of variables: x¼ Ot ; 2 ay ¼ 8zeU ; mr02 O2 qy ¼ 4zeVRF mr02 O2 ð26Þ 939 & DOUGLAS FIGURE 2. Regions of stability for the x and y motions in a linear quadrupole. Reproduced from Du, Douglas, and Konenkov (1999a) with permission. Copyright the Royal Society of Chemistry. With these variables, the equation of motion for y becomes 2 dy þ ðay 2qy cos 2xÞx ¼ 0 dx2 ð27Þ The stability of ion motion in the y direction can be considered. Because Equation (27) is the same as Equation (23), the y motion will be stable if ay and qy have same values as those shown by the shaded regions labeled ‘‘stable x’’ in Figure 2 for ax and qx. 4. Combined Stability for Ion Motion in the x and y Directions What combinations of ax,qx and ay,qy make ion motion stable in both the x and y directions? The key is to note that ay ¼ ax ; qy ¼ qx ð28Þ What values of ax and qx make the y motion stable? For the y motion to be stable the values of ay and qy must be in the shaded regions ‘‘stable x’’ of Figure 2. If ax and qx have the same magnitudes but with opposite sign, ay and qy will have these 940 values. The stability diagram Figure 2 is symmetric about the a axis; replacing qx by qx makes no difference. However, for the y motion to be stable when expressed in terms of the x parameters, the sign of ax for the stable regions must be changed. This gives the regions in Figure 2 labeled ‘‘stable y.’’ Regions where both the x and y motions are stable are called ‘‘combined stability regions’’ or simply ‘‘stability regions.’’ Eight stability regions can be found in Figure 2. In fact, there are an infinite number of stability regions. The a and q values of the tips of the first six regions have been given by Konenkov, Sudakov, and Douglas (2002). Almost all commercial mass filters use the stability region labeled ‘‘1’’ in Figure 2. Operation of quadrupole mass filters in other regions is reviewed in Section IIIC. 5. Hexapole and Octopole Fields As discussed below, mass analysis takes advantage of the stability regions. Can stability diagrams be derived for other multipole rod sets, such as hexapoles or octopoles? The stability diagram of Figure 2 was derived by considering first the ion motion in the x direction and then the ion motion in the y Mass Spectrometry Reviews DOI 10.1002/mas LINEAR QUADRUPOLES IN MS direction. Separate stability diagrams for each direction could be calculated because the ion motions in the x and y directions are independent. For a higher multipole such as a hexapole or octopole this is not possible. Consider the octopole. From Equation (13) the electric field in the x direction is given by 3 4x 12xy2 ð29Þ Ex ¼ r04 and the electric field in the y direction by 12x2 y þ 4y3 Ey ¼ r04 ð30Þ The electric field in x depends on y, and the electric field in y depends on x. Thus, the x and y motions are coupled and cannot be considered independently. No stability diagram can be derived. Stability boundaries are not well defined and become diffuse (Haag & Szabo, 1986; Gerlich, 1992). Thus, higher order multipole rod sets have not been used as mass filters. D. Acceptance and Emittance Not all combinations of initial position and initial radial velocity are transmitted at a given operating point (a, q), in the stability diagram. For a given RF phase, combinations of initial position (x, y) and radial velocity, x_ ¼ dx=dt, y_ ¼ dy=dt, that are transmitted, fall within ellipses in the x, x_ and the y, y_ planes (the ‘‘phase’’ planes) (Dawson, 1976, p. 32) as shown schematically in Figure 3. There are separate ellipses for the x and y directions. The area of an ellipse is the ‘‘acceptance’’ for that particular phase. The ellipses rotate about the origin of the phase plane at the RF frequency. The region where the ellipses overlap for all phases gives the acceptance in x or y for 100% transmission regardless of RF phase. The acceptance can also be described FIGURE 3. An acceptance ellipse for the x direction. Reproduced from Du, Douglas, and Konenkov (1999a) with permission. Copyright the Royal Society of Chemistry. Mass Spectrometry Reviews DOI 10.1002/mas & for 50% or some other value of transmission. The area of the acceptance ellipses scales as r02 f for each of the x and y directions, so that the combined acceptance in x and y is proportional to r04 f 2 . The positions and radial velocities of ions from a source will occupy regions in the x, x_ and the y, y_ planes. In general, these positions will not form an ellipse. The areas of these regions are called the source emittance in x and y. The transmission of a quadrupole will be determined by the overlap of the quadrupole acceptance and source emittance (Dawson, 1990). As the resolution of a quadrupole increases, the combined acceptance in x and y decreases with resolution (R) as 1/R (Dawson, 1980, 1990). III. MASS ANALYSIS WITH LINEAR QUADRUPOLES A. Mass Analysis at the Tip of the First Stability Region Figure 4 shows the first stability region. The boundaries correspond to the lines bx ¼ 0, bx ¼ 1, by ¼ 0, and by ¼ 1. Lines of constant bx and by are shown within the stability region (‘‘iso-beta lines’’). At the tip of the region ax ¼ 0.23699 and qx ¼ 0.70600 (Dawson, 1976, p. 21). The stability boundaries when ax ¼ 0 are at qx ¼ 0.000 and qx ¼ 0.908. Consider the positions on the stability diagram of three ions with masses (more accurately, mass-to-charge ratios) m1 < m2 < m3 for fixed values of U and VRF. For mass analysis the values of U and VRF are set so an ion, m2 in this example, is just within the stability region at the tip of the stability diagram. For the same voltages heavier ions m3 have lower a and q values, and lighter ions m1 have greater a and q values. The ions lie on an ‘‘operating line’’ or ‘‘scan line.’’ Ions m2 have stable motion within the quadrupole, and ions m1 and m3 have unstable motion. Comparison to Figure 2 shows m1 has motion unstable in the x direction, and m3 has motion unstable in the y direction. Thus, if an ion source is placed at the entrance of a quadrupole and a detector or some other device at the exit of the quadrupole, ions of mass m2 will be transmitted and ions of other masses, m1 and m3 in this example, will become unstable, will collide with the electrodes of the quadrupole, and will not be transmitted. Thus the quadrupole acts as a ‘‘mass filter’’ for ions of mass m2. To transmit another mass the voltages U and VRF are adjusted to place that mass at the tip of the stability region. If U and VRF are scanned together with a constant ratio, ions of increasing mass will reach the tip of the stability region in order of their mass and will be transmitted sequentially to produce a mass spectrum. Scans by decreasing U and VRF are also possible. Alternatively, the voltages U and VRF can be switched so that the quadrupole transmits predetermined masses to provide ‘‘peak hopping’’ or single or multiple ion monitoring. The scanning is purely electronic and is amenable to computer control. The mass resolution of a quadrupole can be changed by changing the ratio of DC (U) to RF voltage (VRF) to change the slope of the scan line. Lower ratios provide lower mass resolution. If a quadrupole is operated with a constant ratio U/VRF the resolution will be constant across a mass scan. It is more common to operate a quadrupole so that the peak width is constant across a spectrum (e.g., to provide unit resolution 941 & DOUGLAS 0.3 m3 0.2 βx a 0.1 Operating Line 0.8 m2 0.2 m1 0.4 0.6 βy 0.6 0.4 0.8 0.2 0.0 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 q FIGURE 4. The first stability region, showing an operating line. Reproduced from Moradian (2007). Copyright A. Moradian. across a spectrum). This can be done by changing the ratio U/VRF with mass (Marchand & Marmet, 1964). The simplest, although somewhat approximate, method to do this for positive ions is to apply a constant negative DC voltage between the rods in addition to the positive DC voltage U which changes with mass. In principle, the resolution can be increased without limit by placing ions closer to the tip of the stability region. In practice, at least two things limit the resolution—the residence time of ions in the quadrupole field and field imperfections in the quadrupoles. The effects of field imperfections are complex and not well understood and are discussed below. The effects of residence time are introduced here. If ions do not spend a sufficiently long time in the quadrupole field they may be transmitted even though they have a,q values outside the stability region. The stability boundaries correspond to the amplitude of ion oscillation increasing exponentially and are for an infinitely long quadrupole (i.e., t ! 1). If an ion which would otherwise be unstable reaches the end of the quadrupole before its amplitude of oscillation reaches r0, it will be transmitted. When the mass resolution is limited by the ion residence time in the quadrupole, the resolution measured at half height, R1/2, is given by R1=2 n2 ¼ RF h ð31Þ where nRF is the number of RF cycles an ion spends in the quadrupole field, and h is a constant. For operation in the first stability region Paul, Reinhard, and von Zahn (1958) calculated h ¼ 12.25. Austin, Holme, and Leck (1976) found h ¼ 20. For this reason quadrupole mass filters, at least when operated in the first stability region, provide the highest mass resolution with low energy ions. High energy ions give poor resolution with tails on peaks, particularly the low mass side. If an ion beam has a distribution of energies with a high energy tail, the ions with high energies will not be well resolved and will produce peak 942 broadening. Thus, quadrupoles benefit from ion beams that have both low energies and low energy spreads. Equation (31) shows that the resolution of a quadrupole might be increased by running at a higher RF frequency. This, however, requires higher RF voltages for a given mass range. Manufacturers make choices between resolution, sensitivity, and mass range. Higher frequencies generally give higher resolution and somewhat higher sensitivity (because of the greater acceptance), but at the expense of lower mass range. Conversely, lower frequencies give higher mass range but at some expense of sensitivity and resolution. Where only a limited mass range is required, such as for inductively coupled plasma mass spectrometry (ICP-MS) (maximum m/z ca. 250), quadrupoles operated at frequencies of 2.0–3.0 MHz are commonly used. Another approach to increase nRF is to use a longer quadrupole. Von Zahn (1962) used a quadrupole with a length of 5.82 m to obtain a resolution of several thousands with Xeþ ions. Amad and Houk (1998, 2000) increased nRF by reflecting ions to give multiple passes through a quadrupole. A resolution of R1/2 & 11,000 was obtained at m/z ¼ 28. B. Fringe Fields The field of a quadrupole mass filter does not stop abruptly with its full value at the ends of the rods. At each end of the rods there are regions of length approximately r0 both outside and inside the quadrupole where the field grows from 0 to its maximum value within the quadrupole. The ‘‘fringing fields’’ in this region can be defocusing. Consider an ion approaching the tip of the stability region of Figure 4 along the scan line. If the field grows linearly the ion will be in a region of unstable motion for the y direction and may be rejected before it reaches the mass filter (Dawson, 1980). The fringe field can change the acceptance of the quadrupole. The fringe field is not necessarily defocusing. The transmission depends on the number of RF cycles nf that ions Mass Spectrometry Reviews DOI 10.1002/mas LINEAR QUADRUPOLES IN MS spend in the fringing field. There is an optimum transit time of nf & 2 cycles which gives the maximum transmission with a quadrupole operated in the first stability region (Dawson, 1980). In general, the fringing field is complex. Ion motion is no longer independent in the x, y, and z directions. Dawson (1980) proposed a linear model of the potential in the fringing field as follows: ðx; y; zÞ ¼ ðx2 y2 Þz r03 ð32Þ Hunter and McIntosh (1989) calculated numerically the fringing field of a quadrupole with an aperture plate at a distance d from the end of the rods and fit the fields to ðx; y; zÞ ¼ x2 y2 f ðzÞ r02 ð33Þ with f ðzÞ ¼ 1 expðaz bz2 Þ ð34Þ where the distance z is measured from the aperture plate in units of r0. Values of a and b were given for different distances of the end plate from the quadrupole. In a second article, McIntosh and Hunter (1989) calculated the effects of this more realistic fringe field model on the quadrupole acceptance. The defocusing effect of the fringing field can be eliminated with a delayed DC ramp, short RF-only quadrupole, which has some fraction of the main RF applied but no DC between the rods (Brubaker, 1968). The short quadrupole is sometimes called a ‘‘prefilter.’’ As an ion approaches the quadrupole it first experiences an RF field in which it is stable and then the DC is applied, but in such a way that the ion retains stable values of a and q. The fringing field at the exit of a quadrupole can also be defocusing. If ions leaving the quadrupole enter a detector floated at a high potential the defocusing effects at the exit are overcome by the strong attractive field from the detector, and a delayed DC ramp is not used at the exit of the quadrupole. C. Quadrupole Mass Filters for High Mass Ions For mass analysis an ion is placed near the tip of the stability diagram where a ¼ 0.23699 and q ¼ 0.70600. For a quadrupole power supply with a maximum RF voltage output of VRFmax, there will be a maximum value of the mass-to-charge ratio that can be mass analyzed, given by ðm=zÞmax ¼ 4eVRFmax 0:70600r02 O2 ð35Þ For given rod size r0, the maximum mass can be increased by lowering the operating frequency O. This was demonstrated for a 3D trap by Wuerker, Shelton, and Langmuir (1959) who trapped small (ca. 1 mm) particles of aluminum in a trap operated at 50 Hz. With a lower operating frequency the number of RF cycles that an ion spends in a quadrupole mass filter is reduced, so somewhat lower resolution can be expected. As well, because the acceptance of the quadrupole scales as O2, the sensitivity can be expected to be somewhat lower for a quadrupole operated at lower frequencies. Beuhler and Friedman (1982) described a Mass Spectrometry Reviews DOI 10.1002/mas & quadrupole operated at 292 kHz with a maximum m/z of approximately 80,000, used to study cluster ions. Labastie and Doy (1988) described a quadrupole mass filter operated at 560 kHz, with a maximum mass range of 9,000, also used to study cluster ions. A resolution of approximately 80 was used, sufficient for the cluster experiments. With the development of electrospray ionization (ESI) there has been renewed interest in quadrupole mass filters with high mass ranges. Winger et al. (1993) described a quadrupole similar to that of Beuhler and Friedman (1982), operated at 292 kHz to give a maximum m/z of 45,000. The quadrupole was used to examine low charge states of protein ions and protein–protein complexes (Light-Wahl, Schwartz, & Smith, 1993, 1994). The resolution was surprisingly low, ca. 100, and there was a severe loss of sensitivity for ions with m/z <500. Collings and Douglas (1997) described a quadrupole operated at 683 kHz with a mass range of 8,585. Unit resolution was possible at m/z ¼ 5,000, and below m/z 3,000 the sensitivity was comparable to the same system operated at 1.00 MHz with a mass range of 4,000, as expected from the slightly lower acceptance with the higher mass range quadrupole. High mass range quadrupoles have also been used with hybrid quadrupole time-of-flight (TOF) instruments. Sobott et al. (2002) described a system with a quadrupole with a calculated mass range of 32,000, used to study protein–protein complexes. A resolution of ca. 1,000 at m/z 22,226 was demonstrated. Interestingly, to collisionally cool (see below) ions of large protein complexes requires a higher pressure in a quadrupole ion guide than is used for more highly charged protein ions or peptides (Chernusevich & Thomson, 2004). D. Scanning with Frequency Sweeps If the voltages U and VRF are kept constant and the frequency of the quadrupole power supply is scanned, ions of different massto-charge ratio will reach the tip of the stability region and a mass spectrum can be produced. This was first demonstrated by Paul and Raether (1955) who swept the frequency of a quadrupole from 2.38 to 2.54 MHz to produce a mass spectrum of Rb isotopes (see also Paul, 1990). A quadrupole power supply capable of frequency scans was described by Mellor (1971). Marmet and Proulx (1982) described a frequency swept quadrupole that was intended to give greater stability than a conventional quadrupole with the electronics available then. A frequency scan from 500 to 350 kHz was used to produce a spectrum from m/z ¼ 10 to 40 with better than unit resolution. A spectrum of Xeþ ions was demonstrated where the isotopic peaks were baseline resolved. Landais et al. (1998) described a quadrupole with a frequency that could be scanned from 0.4 to 1.1 MHz. Resolution of approximately 100–200 was obtained at m/z 79–176. Nie et al. (2006) described a quadrupole with the frequency scanned from 100 to 500 kHz. The quadrupole was operated in the third stability region (see below) to improve resolution. The resolution with ions of cytochrome c was approximately 30–100. There are two potential difficulties with a frequency-swept quadrupole. First, at the lower frequency necessary for higher mass ions, resolution is limited by the number of RF cycles in the quadrupole. Second, because the acceptance of the quadrupole is proportional to the square of the frequency, as the quadrupole 943 & DOUGLAS scans to higher masses (i.e., to lower frequency) the sensitivity might be expected to decrease. E. Operation in Other Stability Regions Most commercial quadrupole systems operate in the first stability region. However, there have been several investigations of quadrupole operation in other stability regions with higher a and q values. In general, operation in higher stability regions provides higher resolution and the ability to obtain unit resolution with higher kinetic energy ions because the value of h in Equation (31) is considerably lower than with operation in the first region. Different authors use different notations for the stability regions. The notation used here is from Dawson (1976). Earlier reviews of operation in higher stability regions have been given by Konenkov and Kratenko (1991) and by Du, Douglas, and Konenkov (1999a). 1. The Second Stability Region The second stability region, Figure 5, is centered at q ¼ 7.547 with a tip at a ¼ 0.0295. Values of bx and by vary from 1 to 2 across the region. The width of the stability diagram at the base gives a resolution q/Dq of 114 for a quadrupole operated in RFonly mode. A quadrupole that requires VRF ¼ 5,000 V to reach q ¼ 7.7457 requires only U ¼ 20 V to reach the tip of the region. A scan line that passes through the tip of the second region will also pass through the first region. When ions are transmitted at the center of the second region, ions of m/z 8.3 times greater or more are transmitted in the first region. Practical operation in the second region will require a second auxiliary mass analyzer to prevent transmission of ions in the first region. However, only very low resolution is required from the second analyzer. Dawson and Bingqi (1984a, b) first studied this region and showed that considerably higher resolution is possible than with a quadrupole operated in the first region. The value of h in Equation (31) is lower in this region. A resolution of several thousands at m/z 44 and 129 was shown. With 300 eV ions, a FIGURE 5. The second stability region. Reproduced from Ying and Douglas (1996) with permission. Copyright John Wiley and Sons, Ltd. 944 resolution of ca. 1,500, sufficient to separate 131Xeþ from C3 Fþ 5, was demonstrated. Dawson and Bingqi (1984a) calculated that in theory R ¼ 180n2RF and measured R ¼ 19n2RF (Dawson & Bingqi, 1984b). Titov (1998) calculated R ¼ 24n2RF . Du, Douglas, and Konenkov (1999a) measured R1=2 ¼ 26n2RF and Konenkov and co-workers measured R1=2 ¼ 24n2RF (Shagimuratov et al., 1990; Konenkov & Kratenko, 1991). Because of the lower value of h compared to the first region, ions of higher energy can be mass analyzed with unit resolution in the second stability region. Hiroki, Kanenko, and Murakami (1995) showed resolution of 48 at m/z ¼ 32 with a quadrupole operated in the second region with no DC between the poles (i.e., a ¼ 0) and an ion energy of 3,000 eV. Du, Douglas, and Konenkov (1999a) showed unit resolution at m/z ¼ 39 with 1,000 eV Kþ ions. Because of the higher RF voltages required for scanning in the second region, it may be most useful for mass analysis of relatively low m/z ions. Ying and Douglas (1996) and Du, Douglas, and Konenkov (1999a) discussed the possible advantages for ICP-MS. The ability to mass analyze high energy ion beams may reduce space charge problems, a source of nonspectroscopic interferences in ICP-MS. With low energy beams a resolution R1/2 & 5,000–9,000 is sufficient in many cases to separate atomic ions from molecular ion interferences below m/z ¼ 80. Ying and Douglas (1996) demonstrated a resolution at half height of R1/2 ¼ 5,000 at m/z ¼ 56, sufficient to separate ArOþ from Feþ. Low energy ions limit the maximum scan speed that can be achieved with a quadrupole. If the quadrupole mass setting changes before an ion passes through the quadrupole, the ion will not be transmitted. With high energy ion beams a quadrupole can, in principle, scan more quickly. For ICP-MS this will increase the number of elements that can be detected in a transient sample. Grimm, Clawson, and Short (1997) demonstrated the use of the second stability region to mass analyze ions with kinetic energies up to 1,000 eV for fast GC/MS scans. Scan rates of 1,000 scans/sec over 80 mass units were demonstrated. The transmission with operation in the second region is expected to be considerably less than in the first region for two reasons: (1) the acceptance of the quadrupole is lower in the second region, and (2) the fringe fields are more defocusing. When a quadrupole is operated in the second region at a resolution of 10,000 the acceptance is approximately 103 of the quadrupole operated in the first region at low resolution (Douglas & Konenkov, 1998). With operation at low resolution (114) in the second region, the fringing field can decrease the acceptance by three orders of magnitude if the ions spend too long in the fringing field (Dawson & Bingqi, 1984a; Douglas & Konenkov, 1998). However, these effects are not prohibitive. Ions can be tightly focused into the quadrupole so that the decrease in transmission between the first and second regions is less than calculated from the differences in acceptance (Douglas & Konenkov, 1998). The fringing fields do not defocus if the ion residence time is less than approximately 0.5 RF cycles (Dawson & Bingqi, 1984a; Douglas & Konenkov, 1998). Titov (1998) calculated a maximum in the transmission when nf ¼ 0.1 for a quadrupole with a distorted field. These short times can be achieved by accelerating the ions in the fringe field and decelerating the ions in the quadrupole. A comparison between Mass Spectrometry Reviews DOI 10.1002/mas LINEAR QUADRUPOLES IN MS & an experimentally measured curve of transmission versus resolution and a curve calculated from the overlap of the source emittance with the quadrupole acceptance showed good agreement provided the emittance corresponded to the ions being tightly focused into the quadrupole (Douglas & Konenkov, 1998). Under these conditions the transmission at low resolution in the second region is approximately 10 times less than in the first region. Increasing the resolution to 5,000 causes about another factor of 10 loss of transmission (Ying & Douglas, 1996). Du, Douglas, and Konenkov (1999b) observed considerable structure on the peaks with a quadrupole operated in the second stability region at resolution of ca. 100–500 and ion energies of 20–100 eV. Two possible sources of the structure were considered: nonlinear resonances caused by field imperfections and ion collection effects. It was shown that the structure was caused by ion collection effects. Ions entering the quadrupole near the axis can be focused or defocused at the exit of the quadrupole. When defocused, the ions are collected less efficiently and a dip appears on the peak. It was shown that nonlinear resonances caused by hexapole and octopole fields cannot cause structure on a peak when the quadrupole is operated at a resolution greater than 800 because the resonance lines do not pass through the tip of the stability region. Similar observations of focusing effects were reported for a quadrupole operated in the third stability region (Du et al., 2000). 2. The Third Stability Region The third stability region (Fig. 2) has a somewhat rectangular shape and is near a ¼ 3, q ¼ 3. The upper and lower tips are shown in Figure 6a and b, respectively. Values of bx vary from 1 to 2 and by from 0 to 1. A scan line that passes through the center of the region gives a resolution of 22 (Dawson, 1974). Mass analysis is possible with scan lines that pass through the upper tip (a, q) ¼ (3.164, 3.234) or through the lower tip (a, q) ¼ (2.521, 2.815). Scan lines that pass through these tips do not pass through the first stability region and so a mass filter operated in the third region does not require any additional device to reject higher mass ions. Operation of a mass filter in the third stability region can provide higher resolution than operation in the first stability region. Konenkov et al. (1990) first showed the separation of COþ from Nþ 2 (m/Dm ¼ 2,500 at m/z 28) (reproduced in the review of Konenkov and Kratenko, 1991), and separation of 40Arþ2 from 20 Neþ (m/Dm ¼ 177) and D2Oþ (m/Dm ¼ 656). Shagimuratov et al. (1990) demonstrated baseline separation of isotopic peaks of Krþþ differing in mass by 0.5 Th. Konenkov, Silakov, and Mogilchenko (1991) demonstrated separation of 4Heþ from Dþ 2 (m/Dm ¼ 157). Hiroki, Abe, and Murakami (1991) noted the resolution is higher than with operation in the first region and that the peaks have much less tailing, allowing the detection of a minor peak beside a major peak. Separation of Heþ from Dþ 2 was demonstrated (Hiroki, Abe, & Murakami, 1992) and separation of 3Heþ from HDþ (m/Dm ¼ 518) (Hiroki et al., 1994). Detection of 4He in a large excess (104) of D2 was described (Hiroki, Abe, & Murakami, 1994a) and applied to leak detection (Hiroki, Abe & Murakami, 1996). In all these studies operation at Mass Spectrometry Reviews DOI 10.1002/mas FIGURE 6. The third stability region showing (a) the upper tip and (b) the lower tip. Reproduced from Du, Olney, and Douglas (1997). Copyright the American Society for Mass Spectrometry. the upper tip of the stability diagram was preferred because peaks had less tailing on the low mass side. Fringing fields of different lengths were investigated (Hiroki, Abe, &Murakami, 1994b) as well as use of a prefilter to improve sensitivity by a factor of 2 (Hiroki et al., 1995). Du, Olney, and Douglas (1997) investigated the use of the third stability region for ICP-MS. Resolution of up to 4,000 was obtained at m/z ¼ 59 (59Coþ) with operation at the upper tip, and ca. 1,000 with operation at the lower tip. These resolutions are not generally high enough to separate atomic ions from molecular ion interferences. However, at a resolution of ca. 600–1,000, readily obtained with operation at the upper tip, the peaks were found to be remarkably free of tails on the low and high mass 945 & DOUGLAS sides compared to operation in the first stability region. An abundance sensitivity of greater than 107 less than 0.5 Th from the peak center was demonstrated at m/z 59. This lack of peak tails was also seen in later computer simulations of peak shapes by Turner, Taylor, and Gibson (2005), and Hogan and Taylor (2008). Somewhat higher energy ions can be analyzed with operation in the third region. Hiroki, Abe, and Murakami (1994a) used peak shapes from computer simulations to calculate that at the upper tip R ¼ 1:7n2RF and at the lower tip R ¼ 0:69n2RF . The same simulations showed R ¼ n2RF =10:9 for the first region, in reasonable agreement with the calculation of Paul, Reinhard, and von Zahn (1958) which gave R ¼ n2RF =12:25. Konenkov and co-workers found R ¼ 0:7n2RF (Konenkov et al., 1990; Konenkov & Kratenko, 1991). Dawson (1974) estimated that approximately five times fewer cycles were required than in the first region so R 0:7n2RF . Titov (1998) calculated R ¼ 1:4n2RF . Du, Douglas, and Konenkov (1999a) measured R1=2 ¼ 1:0n2RF for operation at the upper tip. These calculations and measurements suggest that ions of somewhat higher energy, a few tens of eV, might be mass analyzed with operation in the third region. Unit resolution with 63 eV Coþ ions was shown by Du et al. (2000), and resolution of 275 with 120 eV Coþ ions was demonstrated by Du and Douglas (1999). Acceptance ellipses for the third region with operation at the upper and lower tips have been calculated by Dawson (1974), Konenkov, Mogilchenko, and Silakov (1992), Konenkov and Dowell (1997), Konenkov (1997), Titov (1998), and Du, Douglas, and Konenkov (1999a ). The acceptance in the x direction is much greater than in the y direction. The combined acceptances (x and y) at the upper and lower tips are similar. Du, Douglas, and Konenkov (1999a) showed a calculated transmission versus resolution curve for a realistic emittance and an ion residence time in the fringe field of one RF cycle. Similar transmissions were calculated for operation at the upper and lower tips. In contrast, the experiments of Du, Olney, and Douglas (1997) showed remarkably different transmissions at the upper and lower tips. The differences between experiment and modeling remain unexplained, but may be related to the complex emittance of the ‘‘hollow’’ beam used in the experiments. Du, Olney, and Douglas (1997) showed that peaks have remarkably little tailing provided ion energies remain low (<10 eV). With higher energy ions, peaks form tails. Hiroki, Abe, and Murakami (1994a) used computer simulations of peak shapes to show that, with operation at the upper tip, peaks tail on the high mass side but remain relatively sharp on the low mass side. Conversely, with operation at the lower tip, peaks tail on the low mass side but remain sharp on the high mass side. Du and Douglas (1999) demonstrated this with 120 eV Coþ ions. Du and Douglas (1999) then described a tandem quadrupole mass analyzer that takes advantage of these peak shapes. Two quadrupole mass analyzers are operated in tandem, the first at the upper tip and the second at the lower tip and each at relatively low resolution. The quadrupoles scan together, with each eliminating tails on one side of the peak. The resulting peak is free of tails, even with relatively high energy (100 eV) ions. The sensitivity was found to be highest with the two quadrupoles placed ca. 2 mm apart with no intervening lens. 946 Several authors have considered the effects of the fringing field with operation in the third region. In comparison to the first stability region, fewer cycles in the fringing field give the optimum transmission. Konenkov (1993) showed that the transmission is highest if ions spend 1.2 cycles in the fringing field. Konenkov (1997) showed that with the optimum number of cycles in the fringe field, ca.1.2 at the upper and lower tips, the acceptance increased ca. 3 times over that with no fringing field so that, at a resolution of 100, the transmission was similar to that of a quadrupole operated in the first region. Konenkov and Dowell (1997) showed that the optimum residence times are about nf ¼ 1.0 with operation at the upper tip and nf ¼ 1.2 with operation at the lower tip. Titov (1998) calculated the optimum nf at the upper tip is 1.3, and at the lower tip 1.2. Hiroki, Abe, and Murakami (1994b) studied experimentally the effects of the inlet and outlet fringing field lengths by changing the separation between the quadrupole and a source and detector, and concluded short fringing fields were optimal. Hiroki et al. (1995) added a short RF-only quadrupole ‘‘prefilter,’’ with an RF voltage approximately 0.2 of that of the mass analyzing quadrupole, to a quadrupole operated at the center of the third region, and gained a factor of 2 in transmission. This was attributed to the ions remaining more stable in the y direction as they approached the quadrupole, although x motion was still unstable in the fringe field. 3. The Fourth Stability Region The fourth stability region is shown in Figure 7. The q values on the a ¼ 0 axis range from 21.298631 and 21.303174. At the tip a ¼ 2.06 103. The values of bx and by vary from 2 to 3. When ions are transmitted in this region, ions with m/z 2.82 times greater are simultaneously transmitted in the second region, and ions with m/z 23.46 times greater or more are transmitted in the first region. With no DC applied to a quadrupole the width of the region gives a nominal resolution q/Dq of 4,700. Chen, Collings, and Douglas (2000) investigated the ion optical properties of this region. With no DC applied between the rods a resolution of R1/2 ¼ 13,900 was obtained with 40 eV 39Kþ ions, as expected from the width of the stability region. When DC was added between the rods, the resolution did not increase. The resolution was found to vary approximately as R1=2 ¼ 829n2RF , and it was calculated that unit resolution should be possible at m/z 39 with 10,000 eV ions. High resolution is possible with relatively high energy ions. A resolution of R1/2 ¼ 5,000 was demonstrated with 750 eV ions and a resolution of R1/2 ¼ 1,400 with 4,000 eV ions. The acceptance of the quadrupole was calculated for various residence times in the fringing field, and was found to drop severely for times of more than 0.3 RF cycles. The acceptance was calculated to be 102 to 104 of the acceptance of a quadrupole operated in the second region (depending on resolution). Measurements of the transmission in the fourth region were compared to measurements of the transmission in the second region when the quadrupole was operated at resolutions of 1,500–5,000. The transmission in the fourth region was two to three times less, because high ion energies (600–4,000 eV) could be used to increase the transmission in the fourth region. Mass Spectrometry Reviews DOI 10.1002/mas LINEAR QUADRUPOLES IN MS a & G 2.9 0.0020 2.1 2.8 2.2 2.7 2.3 0.0015 2.6 2.4 βy βx 2.5 0.0010 2.5 2.4 0.0005 2.6 2.3 2.7 2.8 2.2 0.0000 2.1 2.9 21.299 21.300 21.301 21.302 21.303 q FIGURE 7. The fourth stability region. Reproduced from Chen, Collings, and Douglas (2000) with permission. Copyright the American Chemical Society. F. Mass Filter Operation with Rectangular or Other Periodic Waveforms Quadrupoles need not be operated with the sinusoidal waveform described by Equation (15). Any periodic waveform may be used. In this case, the equation of motion becomes d2 u þ CðtÞu ¼ 0 dt2 ð36Þ where u is x or y and C(t) is any periodic function of time. This equation is known as the Hill equation (Dawson, 1976, p. 14). Richards, Huey, and Hiller (1973) showed that a quadrupole mass filter can be operated with a periodic rectangular waveform, like that shown in Figure 8. They considered the case where jV1 j ¼ jV2 j. The claimed advantages are that more precise control of the waveform may be possible and, during the periods of constant voltage, the solutions to the equation of motion can be obtained analytically. Matrix methods (Pipes, 1953; Dawson, 1976, p. 83; Konenkov, Sudakov, & Douglas, 2002) can then be used to calculate the stability diagram, which differs somewhat from the more familiar diagrams of Figures 2 and 4. Provided the waveform has a net DC component on average, parameters a and q, analogous to the Mathieu parameters, can be calculated and mass analysis is possible by placing an ion at the tip of a stability region. Richards, Huey, and Hiller showed that if the ratio t/T is set to 0.390, the stability region shrinks to a small region on the q axis and so mass analysis is possible with no DC component to the waveform. They demonstrated a resolution of up to 300 at m/z ¼ 84 (Krþ). When operated with the waveform of Figure 8a, Mass Spectrometry Reviews DOI 10.1002/mas a quadrupole will have constant resolution over its mass scan. As discussed above quadrupoles are more often operated at constant peak width (often unit resolution). Richards (1977) show that with a rectangular waveform the resolution can be adjusted during a scan by changing the duty cycle t/T. Any periodic waveform can be used to operate a quadrupole. Richards, Huey, and Hiller (1973) also discussed the use of a trapezoidal waveform. More complex waveforms can be used, provided they are periodic. The stability diagrams for a variety of rectangular waveforms were described by Konenkov, Sudakov, and Douglas (2002). Sheretov (2000) derived properties of the FIGURE 8. A rectangular waveform for use with a quadrupole mass filter. In general, V1 6¼ V2 and the ratio t/T can be adjusted. 947 & DOUGLAS ion motion for a rectangular waveform and a harmonic waveform of the form f0(t) ¼ U VRF(cos Ot þ k cos jOt), where j is an integer (usually 2 or 3). Sheretov et al. (2000) calculated stability diagrams for 3D traps and quadrupole mass filters operated with these waveforms, and described experimental investigations of the use of such waveforms with a quadrupole mass filter and three-dimensional ion traps. A quadrupole mass filter was operated with a rectangular waveform in two of the higher stability regions. Resolution matched that expected from the width of the stability regions and scan lines used. Sheretov and co-workers (Sheretov, 2002; Sheretov et al., 2002) later showed methods to calculate the properties of ion motion for any periodic waveform, including various rectangular and harmonic waveforms. In any real system, the rectangular waveform will have finite rise times and possibly ringing. Sudakov and Nikolaev (2002) showed that these distortions of the waveform make only minor changes to the stability diagram. stability region along lines with bu values determined from Equation (42). Islands of stability are formed between the bands of instability. Figure 9 shows, for example, the stability islands and bands of instability near the tip of the first stability region formed with n ¼ 1/10 and q0 ¼ 0.0050. As the excitation amplitude q0 increases, the width of the lines of instability increases. Mass analysis can be done with a scan line that passes through the tip of a stability island. The excitation parameters and scan line must be chosen so that the scan line does not pass through any other G. Mass Analysis with Islands of Stability Another method of mass analysis takes advantages of ‘‘islands’’ of stability formed by quadrupole excitation. An auxiliary quadrupole excitation voltage with amplitude V0 and frequency oex is added, so that the quadrupole potential becomes 2 x y2 ðU VRF cos Ot V 0 cosð!ex t gÞÞ Fðx; y; tÞ ¼ r02 ð37Þ where g is the phase of the excitation which can be taken as 0. The excitation is applied at a frequency which is a rational fraction of the main RF frequency, so that oex ¼ nO where n ¼ Q/P where Q and P are integers. The equations of motion become d2 x þ ða 2q cos½2x 2q0 cos½2nxÞx ¼ 0 dx2 ð38Þ d2 y ða 2q cos½2x 2q0 cos½2nxÞy ¼ 0 dx2 ð39Þ where x, a, and q are as in Equation (22) and q0 ¼ 4ezV 0 V0 ¼q 2 2 VRF mO r0 ð40Þ The angular frequencies of oscillation in the quadrupole field are given by Equation (24). With quadrupole excitation, resonances are excited when !ex ¼ jl þ u j O K ð41Þ where u is x or y, K ¼ 1, 2, 3,. . . and l ¼ 0, 1, 2, 3,. . . (Collings & Douglas, 2000; Sudakov et al., 2000). When n ¼ Q/P, the b values of the quadrupole resonances are determined by Q jl þ u j ¼ P K ð42Þ The stability diagram forms P 1 relatively strong resonance lines for the x and y motions. Bands of instability form in the 948 FIGURE 9. The first stability region with quadrupole excitation with n ¼ 1/10 and (a) q0 ¼ 0.0025, (b) q0 ¼ 0.0050, and (c) q0 ¼ 0.0100. Reproduced from Konenkov et al. (2001) with permission. Copyright Elsevier Science B. V. Mass Spectrometry Reviews DOI 10.1002/mas LINEAR QUADRUPOLES IN MS islands. The upper or lower tip of an island can be used for mass analysis (e.g., island A in Fig. 9c). The use of auxiliary excitation to improve peak shapes was first described by Devant et al. (1989). It was shown that addition of an auxiliary excitation waveform with frequency approximately 101 of the main quadrupole frequency and amplitude approximately 102 of the main quadrupole RF voltage could remove tails from peaks. However, this was not described in terms of bands of instability or islands of stability. The formation and use of islands of stability to improve peak shape was described by Miseki (1993) who again noted that islands can be used to remove tails on the low and high mass sides of a peak, and therefore improve the resolution. Konenkov et al. (2001) reported extensive calculations of island positions and described experiments with an ICP-MS system to map the island positions. The island positions and boundaries were calculated using a matrix method. For any values of a and q two ion trajectories, calculated over one period of the applied waveforms (PT where T ¼ 2p/O) can be used to determine if the ion motion is stable or unstable (Konenkov, Sudakov, & Douglas, 2002). Trajectories were calculated for a large number of a and q values and used to map the bands of instability and islands of stability. Measured island positions agreed well with the calculated positions. While the use of islands for mass analysis did not improve the limiting resolution of the quadrupole, the abundance sensitivity was improved from 106 –107 to 109 –1010. Baranov, Konenkov, and Tanner (2001) described mass analysis with islands of stability with an ICP-MS system. They showed that use of an island prevented peak tails from forming when the ion energy in the quadrupole was increased from approximately 5 to 13–15 eV. The increased ion energy improved the sensitivity by approximately two times. Glebova and Konenkov (2002) modeled peak shapes and resolution with operation at the tip of island ‘‘A’’ (notation of Fig. 9). For a quadrupole constructed with round rods (r/r0 ¼ 1.130), operation with an island with q0 ¼ 0.012 and n ¼ 9/10 removed tails on the low mass side of a peak. This effect was not seen with an ideal quadrupole field, and it was concluded that use of an island of stability can improve peak shape for weakly distorted quadrupole fields. Sheretov, Gurov, and Kolotilin (1999) considered the general case of periodic variations of the quadrupole operating parameters that produce parametric resonances and hence bands of instability in the stability diagram. They pointed out that bands of instability can be formed by modulating both the applied RF and DC together, or by modulating just the DC or just the RF amplitudes, and calculated that the number of RF cycles required in the field for a given resolution is decreased when an island of stability is used for mass analysis. Glebova and Konenkov (2002) compared conventional mass analysis with an ideal quadrupole field and mass analysis using an island of stability at a resolution of approximately 100, and concluded there were no advantages to using an island of stability. Later, however, Konenkov, Korolkov, and Machmudov (2005) used matrix methods to calculate island positions and boundaries, and ion trajectory calculations to determine peak shapes and resolution with mass analysis in islands. The effects of modulating the RF amplitude, and both the RF and DC amplitudes, were compared. Ideal fields were compared to results with a round rod set with ratio of rod diameter to field radius r/r0 ¼ 1.13. With amplitude modulation the Mass Spectrometry Reviews DOI 10.1002/mas & resolution can be adjusted in three ways: (1) by changing the slope of the scan line at the upper or lower tip of an island, (2) by setting the scan line to pass through the center of an island and adjusting the modulation amplitude, or (3) by setting the scan line to pass through the center of an island and adjusting the modulation frequency. It was concluded that the preferred (and simplest) method is to fix the modulation amplitude and frequency and adjust the slope of the scan line. Modulation of the RF gave similar results to modulation of the RF and DC. Because modulation of the RF alone is simpler, it is preferred. With an ideal quadrupole field, amplitude modulation of the RF was found to remove peak tails at a resolution of 390 with 100 cycles of the field. This contrasts with the finding of Glebova and Konenkov (2002) that for an ideal field there are no benefits from using islands of stability, because the earlier study considered only low resolution (100). Konenkov, Korolkov, and Machmudov (2005) also found that use of islands removed peak tails in a quadrupole constructed with round rods. Finally, it was concluded that the number of RF cycles in the field required to reach a resolution of 400 was decreased from approximately 150 without modulation, to 75 with modulation of the RF amplitude. H. Mass Analysis with Radial Ejection If a quadrupole is used as a linear ion trap, ions may be ejected in order of their mass radially through a slot in a rod or rods to produce a mass spectrum. This can be done by scanning the trapping RF so that ions reach the stability boundary at a ¼ 0, q ¼ 0.908, or by applying dipole excitation between one of the rod pairs to excite ions for ejection (Schwartz, Senko, & Syka, 2002). This is not discussed in detail here because the method is used exclusively with trapped ions (reviewed by Douglas, Frank, & Mao, 2005). I. Mass Analysis with Axial Ejection A quadrupole mass filter may also be operated in RF-only mode for mass analysis. Ions of a broad range of mass-to-charge ratios are transmitted through the quadrupole. A stopping potential is applied to an electrode or electrodes at the quadrupole exit. Ions of a selected mass-to-charge ratio are excited in the radial directions. In the fringing field at the exit of the quadrupole the x and y motions are mixed with axial motion in the z direction. Some of the increased energy in the radial direction is converted to axial kinetic energy so that the excited ions overcome the potential barrier and are transmitted to a detector. The original motivation for this work was to overcome the defocusing fringing fields at the entrance to a quadrupole mass filter operated conventionally with applied RF/DC voltages, to increase the acceptance over that of a conventional mass filter, and to overcome deleterious effects of field imperfections. Brinkmann (1972) first described this concept. A quadrupole was operated in RF-only mode with a potential barrier at the exit. The quadrupole was scanned so that ions of increasing m/z reached the stability limit at q ¼ 0.908, where they gained kinetic energy in the radial directions. A resolution of more than 1,400 was obtained at m/z ¼ 500, and the transmission was approximately 10 times greater than that of the same quadrupole 949 & DOUGLAS operated conventionally. This work was extended by Holme and co-workers (Holme, 1976; Holme, Sayyid, & Leck, 1978) who reported that there is less dependence of the performance on mechanical imperfections, and less defocusing of ions at the quadrupole entrance. Yang and Leck (1982) extended the mass range to 600 and reported this mode of operation provided uniform transmission over the mass range, although with low resolution. They also reported a relative lack of susceptibility of the performance of the RF-only mass spectrometer to contamination and construction errors. Yang and Leck (1984) showed that the performance was only slightly affected by small spurious potentials applied to the rods, and that therefore an RF-only mass analyzer might be more suitable for use in a dirty vacuum system. Ross and Leck (1983) modified the collector assembly and reported the transmission depended less critically on operating parameters in comparison to conventional operation of the quadrupole, and that for a given transmission, the resolution was higher than with conventional operation. Dawson (1985) reported a detailed investigation of peak shapes, peak tailing, and resolution. The quadrupole was operated at various frequencies up to 3.0 MHz. The major finding was that reasonable peak shape and resolution were possible provided the ions spent sufficient time in the RF field. In all these studies there were difficulties with a continuum background caused by relatively high-velocity low q-value ions. Hager (1999) reported operation of a quadrupole in RF-only mode with a greatly reduced continuum background. Ions formed by electrospray were collisionally cooled in a 20-cm long RFonly quadrupole operated at 7 mTorr, and then entered a 24-mm long RF-only quadrupole for mass analysis. Because it is the fringing fields that cause ions to gain energy in the axial direction when they are excited at q ¼ 0.908, a very short quadrupole (24 mm) could be used. A stopping potential, typically 7 Vabove the rod offset of the quadrupole, was applied to an aperture lens at the exit of the quadrupole. The RF of the quadrupole was scanned to bring ions to the stability boundary at q ¼ 0.908. The continuum background was reduced by applying unbalanced RF to the rods and 3.0 Vof DC between the rods. It was found that applying some of the quadrupole RF voltage to the exit lens improved the transmission by nearly an order of magnitude. It is possible that the improved performance in this work derives partially from the collisional cooling of ions before they enter the quadrupole. This provides an ion source with smaller spatial and energy spreads than were used in prior studies (see below). Hager (2002) later extended this work to a study with a triple-quadrupole mass spectrometer system. Precursor ions, mass selected in a first quadrupole (Q1), formed fragment ions in a quadrupole collision cell (Q2). The fragment ions could be trapped and accumulated in either Q2 (operated at 7 103 Torr) or a downstream quadrupole (Q3) (3 105 Torr). For axial ejection ions were excited by auxiliary dipole or quadrupole fields. The trapping RF voltage was scanned to bring ions of different m/z values into resonance with the excitation. Higher scan speed was possible with trapping of ions at the lower pressure in Q3. With ejection from Q2 a resolution of 6,000 at m/z 609 was possible with very slow scans (5 Th/sec), and a resolution of 1,000 was possible at 1,000 Th/sec. With ejection from Q3, resolution of 6,000 was possible at a scan speed of 100 Th/sec. Higher scan speeds gave lower resolution. The major 950 advantage of this technique is that fragment ions can be accumulated in Q2 or Q3 operated as a linear trap to dramatically improve the sensitivity for tandem mass spectrometry. Hager (2002) reported an improvement in sensitivity of 16 times for MS/MS of reserpine ions in comparison to a conventional fragment ion scan with Q3. Hopfgartner, Hussel, and Zell (2003) reported a sensitivity improvement of 60 times for MS/MS analysis of drug metabolites. Londry and Hager (2003) have used analytical and numerical calculations in a detailed investigation of forces on ions in the fringing field of a quadrupole used for axial ejection. The potential in this region consists of three parts: (1) the quadrupole RF potential which decreases towards the exit lens (the positive z direction), (2) the potential from the exit lens which decreases in the quadrupole, and (3) the potential from an auxiliary dipole excitation field. The decreasing quadrupole potential causes a net force (averaged over one RF cycle) towards the exit lens. This force increases in x and y as the square of the distance from the quadrupole center. The stopping potential on the exit lens produces a force in the opposite direction. At some point these two forces balance each other. Ions on the exit lens side of the balance point are ejected. Ions on the quadrupole side of the balance point are reflected back into the quadrupole. A map of the points where ions are reflected gives a cone shaped region near the quadrupole exit. Excitation of ions moves ions from regions near the quadrupole center, where they are reflected, to regions closer to the quadrupole rods where they are ejected. J. The Ratio r/r0 While hyperbolic electrodes can produce the best approximation to a quadrupole field (aside from construction errors and the finite truncation of the electrodes), round rods are often used because they are easier to manufacture and mount to high precision. If the quadrupole electrodes have fourfold rotational symmetry, the symmetry of the potential requires that the amplitudes of higher multipoles are 0 except A2, A6, A10, A14,. . . (Denison, 1971). Several authors have discussed the ratio of rod radius r to field radius r0 that is optimum for a mass filter. With round rods, additional multipoles are added to the potential (Eq. 6). Dayton, Shoemaker, and Mozley (1954) showed that a ratio r/r0 ¼ 1.148 makes the amplitude of next highest multipole in the expansion, A6, 0. Lee-Whiting and Yamazaki (1971) used conformal mapping to calculate the potential from four parallel rods and showed that A6 ¼ 0 when r/r0 ¼ 1.14511. Denison (1971) calculated the field of a quadrupole in a cylindrical housing of radius 3.44r0 and found a ratio r/r0 ¼ 1.1468 makes A6 ¼ 0. Reuben et al. (1994, 1996) again used conformal mapping to determine the potential of a quadrupole in a grounded case and also found the ratio r/r0 ¼ 1.14511 makes A6 ¼ 0. This ratio was found to be independent of the radius of the case. The same ratio was calculated by Douglas et al. (1999). In these studies it was assumed that the ratio r/r0 that makes A6 ¼ 0 would produce the best mass filter. However, manufacturers did not necessarily use this ratio. It was shown in 1999 that a ratio that balances the effects of A6 and A10 to decrease the effects of nonlinear resonances is preferred (Schulte, Shevchenko, & Radchik, 1999). A ratio r/r0 ¼ 1.10 was proposed Mass Spectrometry Reviews DOI 10.1002/mas LINEAR QUADRUPOLES IN MS to minimize losses from nonlinear resonances. Gibson and Taylor (2000) used trajectory calculations to simulate peak shapes for quadrupoles with hyperbolic and round rods with a ratio r/r0 ¼ 1.148 and found that the round rods produced tails on the peaks, especially on the low mass side. Gibson and Taylor (2001) subsequently investigated the effects of the ratio r/r0 on the performance of a quadrupole and proposed that a ratio 1.12r0 to 1.13r0 would give the highest resolution and transmission. Douglas and Konenkov (2002) used simulations to show that a ratio r/r0 ¼ 1.128–1.130 gives peak shapes and transmission similar to those of an ideal quadrupole field. When r/r0 ¼ 1.130, A6 ¼ 1.00 103 and A10 ¼ 2.44 103. The A6 and A10 terms have similar magnitudes but opposite signs and compensate for each other. The A6 term shifts the pass band to low mass. The A10 term shifts the pass band to high mass. When both are present with the correct amplitude, the pass band corresponds to that of an ideal field. Thus, a ratio r/r0 that makes A6 ¼ 0 does not produce the best mass filter. A ratio r/r0 & 1.128 is preferred. K. Ion Guides 1. Collisional Cooling and Focusing As described above, to obtain high transmission with a quadrupole mass filter, it is desirable to match the emittance of the ion source to the acceptance of the quadrupole. As the resolution of the quadrupole increases, the acceptance decreases (as 1/R) for a quadrupole operated in the first stability region (Dawson, 1980, 1990). As mass spectrometry has pushed to higher mass for applications in life sciences, the demands for higher resolution and higher sensitivity have increased. The shape of the emittance from the source can be changed with an ion lens (Regenstreif, 1967; Dawson, 1980). For example, a nearly parallel beam with a spread of x positions but a small spread of radial velocities can be focused so it has a smaller spread of positions but a greater spread of radial velocities. However, there is a limit to using this to match a source to a mass analyzer. Liouville’s theorem (Joseph Liouville, 1809–1882) states that the area of the beam in phase space before and after an ion lens remains constant (Regenstreif, 1967; Dawson, 1980). Radial velocities can be traded for large radial positions, or vice versa, but an ion lens or a quadrupole ion guide cannot reduce the area in phase space occupied by the beam. The use of RF ion guides, and particularly quadrupole ion guides, has provided a method to overcome this limitation. If a quadrupole is operated in RF-only mode, that is, with only RF between the rods, the Mathieu parameter a will be 0, ions of a broad range of m/z values will lie on the a ¼ 0 axis and have stable trajectories in the quadrupole. The quadrupole can be used as an ‘‘ion guide’’ to transport ions from a source to an analyzer. Initially, it was believed that the pressure in the ion guide should be as low as possible to prevent losses of ions by scattering (see the review by Thomson, 1998). However, Douglas and French (1992) reported that the ion transmission through an ion guide and into a quadrupole mass filter increased as the pressure was increased from 5 104 to 5 103 Torr (see also Douglas & French, 1990). Collisions between ions and the gas in the ion guide cause the ions to move to the center of the ion guide and Mass Spectrometry Reviews DOI 10.1002/mas & also to lose radial energy. Ions are ‘‘focused’’ to the center of the ion guide with a small spread of positions and small radial energies. Thus, the emittance of ions at the exit of the ion guide is less than at the entrance. At the same time the kinetic energy distributions of ions at the exit of the ion guide narrow to distributions with energies and energy spreads of approximately 1 eV per charge or less. The beam quality at the exit of the higher pressure ion guide is greatly improved for transmission into a downstream mass analyzer, particularly a quadrupole. Collisional focusing requires that the ions be injected into the ion guide with relatively low energies, ca. 10 eV/charge. At higher energies collisions cause ions to scatter or fragment. Subsequently, Xu et al. (1993) reported measurements of the transmission of Oþ 2 ions through a hexapole ion guide with collisions with He. They noted the ions at the exit of the ion guide had energies and energy spreads of approximately 0.8 eV. Calculations of ion trajectories showed the ions moved to a region of approximately 0.5 mm in diameter at the center of the guide and therefore also showed collisional focusing. Collisional cooling of the ion beam in these experiments apparently overcomes Liouville’s theorem because the region occupied by the ions in phase space shrinks. However, Liouville’s theorem is not violated because the phase space of the gas molecules is not considered. The use of collisional cooling or collisional focusing in ion guides was soon extended to ESI orthogonal injection TOF systems (Krutchinsky et al., 1998a) and vacuum MALDI orthogonal injection TOF systems (Krutchinsky et al., 1998b). With MALDI a nearly continuous beam of ions is obtained at the exit of the ion guide, and the properties of the ions are largely independent of the source parameters. The improvements for orthogonal injection TOF, higher sensitivity and higher resolution, have been described by Standing (2000). Ion guides with collisional cooling have since found widespread application in many areas such as ESI-FTICR (Jebanathirajah et al., 2005), ICPMS (Tanner, Baranov, & Bandura, 2002), atmospheric pressure MALDI (Schneider, Lock, & Covey, 2005), and nuclear physics (see, for example, Nieminen et al., 2001; Kellerbauer et al., 2002). A complete review is beyond the scope of this article. Collisional cooling and focusing have also found applications in tandem mass spectrometry. In triple-quadrupole mass spectrometers, mass selected precursor ions are fragmented by energetic collisions with gas in a quadrupole ion guide. If the ion guide is operated near single collision conditions the kinetic energies of fragment ions Ef are related approximately to the kinetic energies of the precursor ions Ep by mf Ep ð43Þ Ef mp Thus, the fragment ions have mass-dependent kinetic energies. To obtain high mass resolution in the downstream quadrupole mass analyzer Q3, it is desirable that the fragment ions have low kinetic energies within Q3 (Section IIIA). Ions can be slowed by increasing the rod offset potential of Q3. However, this requires knowing the kinetic energies of the fragments. The problem is that Equation (43) is only approximate. If the rod offset of Q3 is too low, the resolution is degraded; if too high, the ions are not transmitted. If the kinetic energies of the ions are measured, then the rod offset of Q3 can be set correctly to provide high 951 & DOUGLAS transmission and resolution (Shushan et al., 1983) but this is not practical. In many applications, to avoid any loss of signal the rod offset of Q3 was commonly set equal to the rod offset of Q2. In this case fragment ions had relatively high kinetic energies in Q3 and the mass resolution was degraded. A solution to this problem was found when it was realized that collisional cooling could be applied to the fragment ions (Thomson et al., 1995). If the collision cell pressure is increased from ca. 5 104 to ca. 5 103 Torr (Ar), all fragments at the cell exit have kinetic energies of 1 eV or less. Thus improved mass resolution is possible in Q3 which can be operated with a fixed rod offset voltage, slightly below that of Q2. At the same time, because collisional cooling causes all ions to move to the center of Q2 with a reduced emittance, the transmission through Q3 increases greatly. The improvements to MS/MS spectra are dramatic and have been described by Thomson et al. (1995) and in more detail by Morris, Thibault, and Boyd (1994). The benefits of increased sensitivity and mass resolution are also seen where Q3 is replaced by a TOF system (Standing, 2000). 2. Ion Guides with Axial Fields The use of collisional focusing in Q2 of triple quadrupoles provided an ion beam of greatly improved quality for Q3 but had one drawback. With pressures in the collision cell of 5 103 Torr or greater, the transit time of the ions was increased and this caused problems with fast scans in some modes of operation. Suppose Q3 is set to a fixed product ion mass, and Q1 switches rapidly between two precursor masses, ions of mass m1 which produce a fragment at the mass setting of Q3, and ions of mass m2 which do not. When Q1 switches from m1 to m2, some of the fragments formed from m1 are still draining from the collision cell and appear to have been formed from m2. Thus, ‘‘crosstalk’’ problems were encountered with early versions of the highpressure collision cell. A simple solution was to apply a voltage pulse to empty the collision cell between measurements. However, this did not address the slow transit of ions that also caused problems in precursor and neutral loss scan modes. A more complete solution was to apply a small axial field of the order of 1 V/m along the axis of the cell to cause the ions to drift more rapidly to the exit (Thomson, Jolliffe & Javahery, 1996; Thomson, 1998). Initially, this was done with a quadrupole with rods made of many segments. Each segment was given a different DC bias to provide a net axial field along the quadrupole. Many other methods of providing an axial field are possible (Thomson, Jolliffe, & Javahery, 1996; Thomson & Jolliffe, 1998). A system with conical shaped rods, which is both mechanically and electrically simple, has been described (Thomson, 1998; Loboda et al., 2000). The conical rods are arranged so that a first opposing pair of tapered rods have larger diameters at the entrance, and a second pair of rods have larger diameters at the exit. A DC potential difference is applied between the rod pairs. At any point along the quadrupole, the axis potential is determined by the larger rods. If a positive DC potential is applied to the first pair, and a negative DC potential to the second pair, the quadrupole axis potential will be positive at the entrance, zero at the center, and negative at the exit, so that an axial field is produced. Properties of a collision cell with an axial field produced this way were investigated in detail by Mansoori et al. (1998) who 952 concluded that crosstalk problems were eliminated and at the same time, the benefits of collisional cooling were retained. An even simpler version that used cylindrical rods that are tilted with respect to the axis was eventually implemented in commercial systems. The DC applied between the rods limits the transmission of Q2 to a band pass. For triple-quadrupole instruments where Q3 scans this is not a problem because the pass band can be scanned along with Q3 to transmit the ions of interest. For instruments where Q3 is replaced by a TOF system, the pass band becomes more of a problem. Loboda et al. (2000) described a method of adding an axial field by placing additional electrodes between the rods of Q2. The same potential is applied to all electrodes. The field on axis is determined by the spacing of the electrodes from the quadrupole center which varies along the length of the quadrupole. The pass band with this method can be up to 10 times that of the quadrupole with tapered rods. The use of quadrupole ion guides with axial fields has been extended to other applications such as the measurement of collision cross sections (Javahery & Thomson, 1997; Guo et al., 2005), the fragmentation of ions in an ESI-TOF system (Dodonov et al., 1997), improved scan speed and controlled ion molecule chemistry with an ICP-MS system (Bandura, Baranov, &Tanner, 2002), and an ion-molecule reactor with heating of ions (Guo, Siu, & Baranov, 2005). Most importantly, however, the use of collisional cooling to couple high-pressure sources to a quadrupole mass analyzer, and to improve the emittance of an ion beam from a collision cell, has contributed to the development of bench top mass spectrometers with sensitivity and resolution exceeding those of earlier much larger systems. IV. LINEAR QUADRUPOLES WITH ADDED MULTIPOLE FIELDS Most quadrupole mass filters are constructed with high precision to make the closest approximation to the desired field. It has long been argued that small construction errors can degrade the performance of a mass filter, although the effects of various errors have not been well described in the literature. Austin, Holme, and Leck (1976) concluded that to obtain a resolution of 900 required mechanical errors in general of less than 103. Titov (1995) calculated that a mechanical error of 10 mm in the rod diameters reduced the acceptance and hence the transmission by a factor of nearly 10. Taylor and Gibson (2008) used trajectory calculations of peak shapes to show that if the y rod of a quadrupole (the rod with the negative DC with positive ions) is moved in by 0.005r0 (20 mm for a quadrupole with r0 ¼ 4.0 mm), the transmission decreases by ca. 2 times and structure is formed on a peak (nominal resolution 300–400). Dawson (1980) calculated the limiting resolution with various mechanical distortions, and stated ‘‘quadrupole rods are said to be manufactured to tolerances in radius of 105 cm and to parallelism of 104 cm.’’ As discussed, changing the ratio r/r0 of a quadrupole with round rods from 1.12 to 1.13 significantly changes the peak shape and transmission. This corresponds to a change in rod radius of 40 mm for 4.0 mm radius rod. The field of a quadrupole with distorted electrode geometries contains additional multipoles. Generally these have amplitudes AN which are 103 or less of A2. For example, when an Mass Spectrometry Reviews DOI 10.1002/mas LINEAR QUADRUPOLES IN MS & x rod is moved in 0.005r0, the highest multipole amplitudes are A1 ¼ þ0.0042, A3 ¼ þ0.0011, and A6 ¼ þ0.0011. (A dipole term is introduced because the field center moves from the center of the quadrupole.) Recently, Douglas and co-workers have described quadrupoles that have much greater amplitudes of added multipoles; added octopole fields with A4 ¼ 0.02–0.04 (Sudakov & Douglas, 2003), and added hexapole fields with A3 ¼ 0.040– 0.012, Konenkov et al., 2006). These fields were added to improve the fragmentation efficiencies of ions for MS/MS when the quadrupoles are used as linear ion traps (Douglas, Frank, & Mao, 2005). In view of the literature describing the effects of small field distortions on mass filter performance, it would not be expected that these rod sets would be capable of mass analysis at all. Yet in some cases the transmission and resolution with mass analysis are surprisingly high. As well, the new methods of scanning quadrupoles using islands of stability or axial ejection have been found to at least partially overcome the otherwise deleterious effects of the much larger multipole amplitudes of these rod sets. These results are reviewed here. A. Linear Quadrupoles with Added Octopole Fields A linear quadrupole with an added octopole field has a potential of the form: 2 4 x y2 x 6x2 y2 y4 þ A4 Fðx; y; tÞ ¼ A2 r02 r04 ð44Þ ðU VRF cos OtÞ where A2 & 1.0 and A4 is the amplitude of the added octopole field. Sudakov and Douglas (2003) showed that an octopole field can be added to a linear quadrupole constructed with round rods by placing the two y rods closer to the central axis than the x rods. However, this method adds significant amplitudes of other multipole terms. A preferred method is to make the radius of the y rods (Ry) greater than the radius of the x rods (Rx). This adds an octopole term and only low levels of other higher multipoles. An axis potential is formed, but this can be removed by unbalancing the RF and DC applied to the rods. The changes in rod radius are much greater than those caused by manufacturing tolerances. For example, to make A4 ¼ 0.026 requires Ry/Rx ¼1.130, a difference of rod diameters of approximately 1.0 mm for a quadrupole with Rx ¼ 4.0 mm. Figure 10 shows a photograph of a rod set with a 2.0% added octopole field (Rx/Ry ¼ 1.220). A detailed description of trapped ion motion and fragmentation efficiencies of these rod sets used as linear ion traps has been given by Michaud et al. (2005). Given the discussion above, it would not be expected that such rod sets would be capable of conventional mass analysis. Ding, Konenkov, and Douglas (2003) investigated mass analysis with quadrupoles with added octopole fields of 2.0– 4.0% (i.e., A4/A2 ¼ 0.020–0.040). Conventional mass analysis was done with RF and DC voltages applied to the rods. With the positive DC applied to the y rods (the rods of greater diameter) so that for positive ions the Mathieu parameter ax < 0, only very broad peaks with low transmission were seen. The added octopole field prevents mass analysis. However, when the positive DC was applied to the x rods, so that ax > 0, the resolution improved dramatically. Figure 11 shows peak shapes obtained with a quadrupole with a 2.6% added octopole field Mass Spectrometry Reviews DOI 10.1002/mas FIGURE 10. A linear quadrupole with 2.0% added octopole field. The radius of the smaller rods is 4.50 mm and the radius of the larger rods is 5.49 mm. Reproduced from Ding, Konenkov, and Douglas (2003) with permission. Copyright John Wiley and Sons, Ltd. [Color figure can be viewed in the online issue, which is available at www.interscience. wiley.com.] (Ry/Rx ¼ 1.300). With the positive DC applied to the larger rods, Figure 11a, a broad peak with nominal resolution of approximately 150 is produced. With the positive DC applied to the smaller rods, Figure 11b, the isotopic peaks are resolved and R1/2 & 5,890. Measurements of transmission versus resolution showed that at R1/2 & 1,000, the transmission of the quadrupole with 2.0% octopole field equaled or exceeded that of a conventional quadrupole. At higher resolution the transmission of the quadrupole with the octopole field dropped more rapidly with increasing resolution, so that at R1/2 ¼ 3,000 it was approximately 10 times less than a conventional quadrupole. With distorted fields, split peaks might be expected, caused by nonlinear resonances (Dawson & Whetton, 1969; Dawson, 1976, 1980). The octopole field causes a nonlinear resonance when bx þ by ¼ 1. This line runs through the tip of the stability region. However, split peaks are not observed (Fig. 11b) because the residence time of the ions in the quadrupole is too short to allow nonlinear resonances to build up. Ding, Konenkov, and Douglas (2003) used trajectory calculations to determine the stability boundaries for a quadrupole with a 2.0% added octopole field. With positive ions, when the positive DC is applied to the x rods (ax > 0), the boundaries are very similar to those of a pure quadrupole field, and the tip of the stability diagram remains sharp and well defined. When the positive DC is applied to the y rods (ax < 0), the boundaries move out and become diffuse. This 953 & DOUGLAS FIGURE 11. Peak shapes with mass analysis with a quadrupole with a 2.6% added octopole field with the positive DC applied to (a) the larger and (b) the smaller rods. Reproduced from Ding, Konenkov, and Douglas (2003) with permission. Copyright John Wiley and Sons, Ltd. is consistent with the experimental observations, but does not explain why the boundaries become diffuse or sharp. Dawson (1980) estimated that, for a limiting resolution of 2,000, A4 must be less than 3.5 104. The results of Ding, Konenkov, and Douglas show that mass analysis is possible with values of A4 nearly two orders of magnitude greater. The use of islands of stability with quadrupoles with added octopole fields has also been studied for the cases ax > 0 and 954 ax < 0. Konenkov et al. (2007) used trajectory calculations to determine island positions and peak shapes for quadrupoles that have added octopole fields of 2.0–4.0%. Rod sets that have only an added octopole field, and round rod sets with added octopole fields and multipoles up to N ¼ 10 were modeled. Operation in the upper stability island at the tip with the greatest jaj gave resolution peak shape and transmission comparable to conventional operation with a > 0. In this case there is no advantage to using the island of stability. With the polarity of the applied DC reversed so a < 0, only low resolution and transmission are possible with conventional analysis. However, with a < 0 and use of the uppermost island operated at the tip with the lesser jaj, the simulations showed that peak shapes and resolution similar to those with conventional operation with a > 0 are possible. These simulations were confirmed in experiments by Moradian and Douglas (2007). An island of stability was formed by using q0 ¼0.020 and n ¼ 9/10. These values are not necessarily optimum but were chosen to match the calculations of Konenkov et al. (2007). With a quadrupole with 2.0% added octopole field, measured positions of the island tips agreed well with the calculations of Konenkov et al. With ax < 0 at least unit resolution was obtained at m/z ¼ 609, sufficient to separate the isotopic peaks of ions of reserpine. Without the use of the island, under the same conditions, the peak was ca. 6 Th wide, and isotopic peaks could not be resolved. In this case use of the island overcomes the adverse effects of the added octopole field. Mass analysis with mass selective axial ejection is also possible with quadrupoles with added octopole fields and was investigated by Moradian and Douglas (2008). Axial ejection at q ¼ 0.908 with a quadrupole with 2.6% added octopole field gave only broad peaks with low resolution, both with ions flowing continuously through the quadrupole and with trapped ions. The low resolution was attributed to changes in the x and y stability boundaries. When a ¼ 0, the x stability boundary is at q ¼ 0.908, but the y boundary moves out to q & 0.931. It was proposed that between these q values, ions excited in x lead to excitation in y where the motion is stable, possibly delaying ion ejection. Axial ejection of trapped ions was investigated with a quadrupole with 2.0% added octopole field, with ions excited by dipole excitation in the x or y directions, quadrupole excitation, or simultaneous dipole excitation in the x and y directions. With a linear quadrupole with added multipole fields, the resolution and scan speed possible depend on the scan direction, as with 3D traps (Makarov, 1996). Best resolution and sensitivity were obtained with forward scans from low to high mass, dipole excitation applied to the smaller rods, and ion excitation and ejection at high q (0.80). Ejection efficiencies were similar to those obtained with a conventional quadrupole and decreased from approximately 60% at a scan speed of 15 Th/sec to approximately 1% at 4,000 Th/sec. B. Linear Quadrupoles with Added Hexapole Fields A linear quadrupole may also be constructed with an added hexapole field with the potential given by 2 3 x y2 x 3xy2 þ A Fðx; y; tÞ ¼ A2 3 r02 r03 ð45Þ ðU VRF cos OtÞ Mass Spectrometry Reviews DOI 10.1002/mas LINEAR QUADRUPOLES IN MS x y r02 2 þ A3 3 x 3xy r03 2 0.8 0.6 0.4 0.2 0.0 606 608 610 612 614 Apparent Mass (m/z) ¼ c ð46Þ where c is a constant. With a quadrupole with a 2% added hexapole field, and no other higher multipoles, a resolution of ca. 1,130 can be obtained with positive ions. The peak shape and transmission are similar to those with a pure quadrupole field, provided the positive DC is applied to the x rods so the Mathieu parameter a > 0. If the polarity of the DC is reversed so a < 0, only very low resolution and poor peak shape are obtained. Increasing the amplitude of the hexapole field causes the peak to broaden because of changes to the stability diagram, although resolution of several hundreds could be obtained with quadrupoles with 8% and 12% added hexapole fields. Mass analysis with quadrupoles with added hexapole fields of 2–12%, constructed with round rods, was then modeled. Multipoles up to N ¼ 10 were included. Poor resolution and peaks with considerable structure were obtained, showing that the higher multipoles strongly affected the transmission and resolution. It was shown that the octopole term that is added is largely responsible for the poor performance. The octopole term can be removed by making both x rods greater in diameter than the y rods so that A4 ¼ 0. Modeling the transmission and resolution with these rod sets showed greatly improved transmission and resolution. Resolution of R1/2 & 700–800 was calculated for quadrupoles with 8–12% added hexapole field, and resolution of 1,400 for a quadrupole with 6% added hexapole. At low and high resolution, peaks were free of structure. However, at intermediate resolution peaks showed some structure. The stability diagram was calculated for a quadrupole with 2% hexapole and no other multipoles, and round-rod quadrupoles with 2–12% added hexapole fields. With a > 0, the boundaries remained sharp. Increasing A3 causes the x stability boundary to move out towards larger q values. With a < 0, only a diffuse region of stability with low transmission was seen, consistent with the low resolution and transmission seen in simulations of mass analysis when a < 0. Preliminary experimental tests of mass analysis with quadrupoles with added hexapole fields have appeared (Zhao et al., 2007; Xiao, Zhao, & Douglas, 2008). Conventional mass analysis, mass analysis with islands of stability, and mass analysis with axial ejection have been tested. Figure 12 shows Mass Spectrometry Reviews DOI 10.1002/mas b 1.0 Relative Transmission A2 2 1.0 0.8 0.6 0.4 0.2 0.0 606 608 610 612 614 Apparent Mass (m/z) 2 c 7.5x10 Relative Transmission a Relative Transmission where A3 is the amplitude of the hexapole field. The properties of quadrupoles with added hexapole fields have been modeled in detail by Konenkov et al. (2006). It was shown that a hexapole field can be added to a quadrupole constructed with round rods by rotating the two y rods through an angle y towards an x rod. With round rods, other multipoles are added to the potential, especially an octopole term. Geometries that produce nominal hexapole fields of 4%, 8%, and 12% were described. Rotating the y rods towards an x rod moves the field center closer to the x rod. Mass analysis was modeled by running many trajectories with a thermal distribution of speeds (300 K) and a Gaussian distribution of positions in x and y with a standard deviation s ¼ 0.002r0. Quadrupoles with added hexapole fields and no other multipoles were modeled first. These can be constructed with electrodes that have the shapes given by & 6.0x10 2 2 4.5x10 2 3.0x10 2 1.5x10 0.0 613 614 615 616 617 618 619 620 621 Apparent Mass (m/z) FIGURE 12. Mass analysis with a quadrupole with a 4.0% added hexapole field and unequal diameter rods R y ¼ r 0 ¼ 4.500 mm, Rx ¼ 5.279 mm. a: Conventional mass analysis with the positive DC applied to the x rods, R1/2 ¼ 2,070, relative sensitivity 9 104 ions/sec. b: Mass analysis with an island of stability with n ¼ 1/20 and q0 ¼ 0.002, R1/2 ¼ 2,170, relative sensitivity 2.8 105 ions/sec. c: Mass selective axial ejection with dipole excitation between the x rods at q ¼ 0.72, R1/2 ¼ 2,250, ejection efficiency 2%. 955 & DOUGLAS peak shapes obtained with a rod set with 4% added hexapole field and rod diameters chosen to make A4 ¼ 0. With conventional mass analysis, Figure 12a, the resolution is R1/2 & 2,000, but there are tails on the peaks. With an island of stability and at similar resolution, Figure 12b, the tails of the peaks are removed. The transmission is approximately five times lower than that of a conventional quadrupole operated at the same resolution. With axial ejection of trapped ions, Figure 12c, peak shapes and resolution comparable to those obtained with axial ejection from a conventional round-rod quadrupole are possible. These results show that mass analysis with rod sets with large geometric and field distortions is sometimes possible. The use of islands of stability can in some cases overcome the effects of both small and large field distortions. It appears that our understanding of the effects of field distortions is incomplete. Dawson (1976) concluded that ‘‘There remains, therefore, a great deal of scope for further advance in design, performance and applications, building upon the established technological foundation . . ..’’ This statement, subsequently proven correct, would seem to be equally valid in 2009. ACKNOWLEDGMENTS This work was supported by the Natural Sciences and Engineering Research Council of Canada and MDS-Analytical Technologies through an Industrial Research Chair. It follows that the components of the electric field are given by ~ f ðx; y; zÞ ¼ Q ~ Eðx; y; zÞ ðA:1Þ where Q is the charge of the ion and ~ Eðx; y; zÞ is the electric field at that point. The electric field at a point is the force per unit charge. Force and electric field are vectors; they have both magnitude and direction. In a Cartesian co-ordinate system they have components along the x, y, and z axes, fx, fy, fz, and Ex, Ey, and Ez, respectively. If a charged particle is moved from a position s1 to a position s2 through an electric field, the mechanical work done on the particle is Z s2 ~ w¼ f ~ dl ðA:2Þ s1 The work done per unit charge is the change in electrical potential DF(x, y, z) between the points s1 and s2 so that Z s2 ~ DF ¼ E ~ dl ðA:3Þ s1 956 @Fðx; y; zÞ @x ðA:4Þ Ey ¼ @Fðx; y; zÞ @y ðA:5Þ Ez ¼ @Fðx; y; zÞ @z ðA:6Þ Equations (4)–(6) can be written as ~ E ¼ rFðx; y; zÞ ðA:7Þ where the gradient operator, r, is given by r¼ @^ @^ @^ iþ jþ k @x @y @z ðA:8Þ where ^i, ^j, and ^k are unit vectors in the x, y, and z directions, respectively. In free space where there are no charges, such as the region between the electrodes of a quadrupole mass filter, ~ E must satisfy r~ E¼0 ðA:9Þ @Ex @Ey @Ez þ þ ¼0 @x @y @z ðA:10Þ or equivalently V. APPENDIX: ELECTRIC POTENTIALS AND FIELDS: UNITS This appendix gives a brief review of definitions of electrical potentials and fields, as well as the units used in this article. More detail can be found in text books on electromagnetic theory (Smythe, 1939; Kip, 1969). In linear quadrupoles, as well as many other devices for manipulating ions, electric potentials are applied to electrodes to produce forces on ions. The force ~ f on an ion at a point in space with Cartesian co-ordinates (x, y, z) is given by Ex ¼ It follows from Equations (A.10) and (A.7) that: r2 Fðx; y; zÞ ¼ 0 ðA:11Þ or equivalently @ 2 Fðx; y; zÞ @ 2 Fðx; y; zÞ @ 2 Fðx; y; zÞ þ þ ¼0 @x2 @y2 @z2 ðA:12Þ Equation (A.12) is known as Laplace’s equation (Pierre Simon Laplace, 1749–1824). A. Units This article uses SI units. Mass is in kilograms, distance in meters, time in seconds, force in newtons. The unit of electrical potential is the volt (the practical unit). Because electric field is force per unit charge, it has units of newton per coulomb. However, because electric field is also derivative of potential with position, it also has units of volts per meter. These units are summarized in Table 1. In these units, if a charge of 1 C is moved through a potential difference of 1 V, the change in potential energy of the charge is 1 J. If a charge of 1 C is at a point where the field is 1 V/m, the force on the charge is 1 N. If a charge on an electron or proton of 1 1019 C is moved through a potential difference of 1 V, the change in potential energy is QDV ¼ 1.6 1019 J. This unit of energy is called the electron volt. Mass Spectrometry Reviews DOI 10.1002/mas LINEAR QUADRUPOLES IN MS REFERENCES Amad MH, Houk RS. 1998. High-resolution mass spectrometry with a multiple pass quadrupole mass analyzer. Anal Chem 70:4885–4889. Amad MH, Houk RS. 2000. Mass resolution of 11,000 to 22,000 with a multiple pass quadrupole mass analyzer. J Am Soc Mass Spectrom 11:407–415. Austin WE, Holme AE, Leck JH. 1976. The mass filter: Design and performance. In: Dawson PH, editor. Quadrupole mass spectrometry and its applications. Amsterdam: Elsevier, Reprinted 1995. Woodbury: AIP Press. p. 122. Bandura DR, Vladimir I, Baranov VI, Tanner SD. 2002. Inductively coupled plasma mass spectrometer with axial field in a quadrupole reaction cell. J Am Soc Mass Spectrom 13:1176–1185. Baranov VI, Konenkov NV, Tanner SD. 2001. In: Holland G, Tanner SD, editors. Plasma source mass spectrometry: The new millennium. Cambridge: The Royal Society of Chemistry. pp. 63–72. Beuhler RJ, Friedman L. 1982. A study of the formation of high molecular weight water cluster ions (m/e<59000) in expansion of ionized gas mixtures. J Chem Phys 77:2549–2557. Brinkmann U. 1972. A modified quadrupole mass filter for the separation of ions of higher masses with high transmission. Int J Mass Spectrom Ion Phys 9:161–166. Brubaker WM. 1968. An improved quadrupole mass analyzer. Adv Mass Spectrom 4:293–299. Chen W, Collings BA, Douglas DJ. 2000. High resolution mass spectrometry with a quadrupole operated in the fourth stability region. Anal Chem 72:540–545. Chernusevich IV, Thomson BA. 2004. Collisional cooling of large ions in electrospray mass spectrometry. Anal Chem 76:1754–1760. Collings BA, Douglas DJ. 1997. An extended mass range quadrupole for electrospray mass spectrometry. Int J Mass Spectrom Ion Proc 162: 121–127. Collings BA, Douglas DJ. 2000. Observation of higher order quadrupole excitation frequencies in a linear ion trap. J Am Soc Mass Spectrom 11:1016–1022. Dawson PH. 1974. Higher zones of stability for the quadrupole mass filter. J Vac Sci Technol 11:1151–1153. Dawson PH. 1976. Quadrupole mass spectrometry and its applications. Amsterdam: Elsevier. Reprinted 1995. Woodbury: AIP Press. Dawson PH. 1980. Ion optical properties of quadrupole mass filters. Adv Electron Electron Opt 53:153–208. Dawson PH. 1985. Performance characteristics of an rf-only quadrupole. Int J Mass Spectrom Ion Proc 67:267–276. Dawson PH. 1990. Source-analyser coupling in the quadrupole mass filter. Int J Mass Spectrom Ion Proc 100:41–50. Dawson PH, Bingqi Y. 1984a. The second stability region of the quadrupole mass filter. I. Ion optical properties. Int J Mass Spectrom Ion Proc 56:25–39. Dawson PH, Bingqi Y. 1984b. The second stability region of the quadrupole mass filter. II. Experimental results. Int J Mass Spectrom Ion Proc 56: 41–50. Dawson PH, Whetton NR. 1969. Non-linear resonances in quadrupole mass spectrometers due to imperfect fields. II. The quadrupole mass filter and monopole mass spectrometer. Int J Mass Spectrom Ion Phys 3:1–12. Dayton IE, Shoemaker FC, Mozley RF. 1954. The measurement of two dimensional fields. Part II: Study of a quadrupole magnet. Rev Sci Instrum 25:485–489. Denison DR. 1971. Operating parameters of a quadrupole in a grounded cylindrical housing. J Vac Sci Technol 8:266–269. Devant G, Fercoq P, Lepetit G, Maulat O. 1989. Procede d’alimentation en tension des spectrographs de mass du type quadripolaire. Patent. Fr. No 2,620,568. Mass Spectrometry Reviews DOI 10.1002/mas & Ding C, Konenkov NV, Douglas DJ. 2003. Quadrupole mass filters with octopole fields. Rapid Commun Mass Spectrom 17:2495–2502. Dodonov A, Kozlovsky V, Loboda A, Raznikov V, Sulimenkov I, Tolmachev A, Kraft A, Wollnik H. 1997. A new technique for decomposition of selected ions in molecule ion reactor coupled with ortho-time-offlight mass spectrometry. Rapid Commun Mass Spectrom 11:1649– 1656. Douglas DJ, French JB. 1990. Mass spectrometer and method and improved ion transmission. US Patent 4,963,736. Douglas DJ, French JB. 1992. Collisional focusing effects in radio frequency quadrupoles. J Am Soc Mass Spectrom 3:398–408. Douglas DJ, Konenkov NV. 1998. Ion source emittance influence on the transmission of a quadrupole operated in the second stability region. J Am Soc Mass Spectrom 9:1074–1080. Douglas DJ, Konenkov NV. 2002. Influence of the 6th and 10th spatial harmonics on the peak shape of a quadrupole mass filter with round rods. Rapid Commun Mass Spectrom 16:1425–1431. Douglas DJ, Glebova TA, Konenkov NV, Sudakov MY. 1999. Spatial harmonics of the field in a quadrupole mass filter with circular electrodes. Tech Phys 44:1215–1219. Douglas DJ, Frank AJ, Mao D. 2005. Linear ion traps in mass spectrometry. Mass Spec Rev 24:1–29. Du Z, Douglas DJ. 1999. A novel tandem quadrupole mass analyzer. J Am Soc Mass Spectrom 10:1053–1066. Du Z, Olney T, Douglas DJ. 1997. Inductively coupled plasma mass spectrometry with a quadrupole mass filter operated in the third stability region. J Am Soc Mass Spectrom 8:1230–1236. Du Z, Douglas DJ, Konenkov NV. 1999a. Elemental analysis with quadrupole mass filters operated in higher stability regions. J Anal At Spectrom 14:1111–1119. Du Z, Douglas DJ, Konenkov N. 1999b. Peak splitting with a quadrupole mass filter operated in the second stability region. J Am Soc Mass Spectrom 10:1262–1270. Du Z, Douglas DJ, Glebova T, Konenkov NV. 2000. Peak structure with a quadrupole mass filter operated in the third stability region. Int J Mass Spectrom 197:113–121. Feynmann RP, Leighton RB, Sands M. 1963. The Feynmann lectures on physics V2: The electromagnetic field. Reading: Addison-Wesley. pp. 7-2–7-5. Gerlich D. 1992. Advances in Chemical Physics LXXXII. Inhomogeneous RF fields: Aversatile tool for the study of processes with slow ions. New York: John Wiley and Sons. pp. 1–176. Gibson JR, Taylor S. 2000. Prediction of quadrupole mass filter performance for hyperbolic and circular cross section electrodes. Rapid Commun Mass Spectrom 14:1669–1673. Gibson JR, Taylor S. 2001. Numerical investigation of the effects of electrode size on the behaviour of quadrupole mass filters. Rapid Commun Mass Spectrom 15:1960–1964. Glebova T, Konenkov NV. 2002. Quadrupole mass filter transmission in island A of the first stability region with quadrupolar excitation. Eur J Mass Spectrom 8:201–205. Grimm CC, Clawson R, Short R. 1997. Use of region II of the a/q stability diagram for fast scanning of a linear quadrupole mass spectrometer. J Am Soc Mass Spectrom 8:539–544. Guo Y, Siu KWM, Baranov VI. 2005. The high-pressure segmented quadrupole collision cell (HP-SQCC) as an ion molecule reactor. J Am Soc Mass Spectrom 16:957–966. Guo Y, Wang J, Javahery G, Thomson BA, Siu KWM. 2005. Ion mobility spectrometer with radial collisional focusing. Anal Chem 77:266– 275. Haag C, Szabo I. 1986. New ion-optical devices utilizing oscillatory electric fields. II. Stability of ion motion in a two dimensional hexapole field. Int J Mass Spectrom Ion Proc 73:237–275. 957 & DOUGLAS Hager JW. 1999. Performance optimization and fringing field modifications of a 24-mm long rf-only quadrupole mass spectrometer. Rapid Commun Mass Spectrom 13:740–748. Hager JW. 2002. A new linear ion trap mass spectrometer. Rapid Commun Mass Spectrom 16:512–526. Hiroki S, Abe T, Murakami Y. 1991. Development of a quadrupole mass spectrometer using the second stable zone in Mathieu’s stability diagram. Rev Sci Instrum 62:2121–2124. Hiroki S, Abe T, Murakami Y. 1992. Separation of helium and deuterium peaks with a quadrupole mass spectrometer by using the second stability zone in the Mathieu diagram. Rev Sci Instrum 63:3874–3876. Hiroki S, Abe T, Murakami Y, Yanagishita K, Nakamura S. 1994. Development of a high resolution quadrupole mass spectrometer capable of detecting 3He and 4He is a hydrogen isotope atmosphere. J Vac Sci Technol A 12:2711–2715. Hiroki S, Abe T, Murakami Y. 1994a. Detection of a 104 He peak in a deuterium atmosphere using a modified high-resolution quadrupole mass spectrometer. Rev Sci Instrum 65:1912–1917. Hiroki S, Abe T, Murakami Y. 1994b. Influence of the fringing field length on the separated 4He/D2 peak shape of a high resolution quadrupole mass spectrometer. Int J Mass Spectrom Ion Proc 136:85–89. Hiroki S, Kanenko K, Murakami Y. 1995. Applicability of the rf-only mode in a quadrupole mass spectrometer for direct analysis of keV ions. Vacuum 46:1445–1447. Hiroki S, Sakata K, Sugiyama N, Muramoto S, Abe T, Murakami Y. 1995. Effects of a pre-filter on the sensitivity of a high resolution quadrupole mass spectrometer. Vacuum 46:681–683. Hiroki S, Abe T, Murakami Y. 1996. Sensitive helium leak detection in a deuterium atmosphere using a high-resolution quadrupole mass spectrometer. Vacuum 47:767–769. Hogan TJ, Taylor S. 2008. Performance simulation of a quadrupole mass filter operating in the first and third stability zones. IEEE Trans Instr Meas 57:498–508. Holme AE. 1976. Operation of a quadrupole mass filter with only an R.F. voltage component applied to the rod system. Int J Mass Spectrom Ion Phys 22:1–5. Holme AE, Sayyid S, Leck JH. 1978. A detailed study of the quadrupole mass filter operating with no d.c. component of the rods. Int J Mass Spectrom Ion Phys 26:191–204. Hopfgartner G, Husser C, Zell M. 2003. The rapid screening and characterization of drug metabolites using a new quadrupole-linear trap mass spectrometer. J Mass Spectrom 38:138–150. Hunter KL, McIntosh BJ. 1989. An improved model of the fringing fields of a quadrupole mass filter. Int J Mass Spectrom Ion Process 87:157–164. Javahery G, Thomson BA. 1997. A segmented radiofrequency-only quadrupole collision cell for measurements of ion collision cross section on a triple quadrupole mass spectrometer. J Am Soc Mass Spectrom 8:697– 702. Jebanathirajah JA, Pittman JL, Thomson BA, Budnik BA, Kaur P, Rape M, Kirschner M, Costello C, O’Conner PB. 2005. Characterization of a new qQq-FTICR mass spectrometer for post-translational modification analysis and top-down tandem mass spectrometry of whole proteins. J Am Soc Mass Spectrom 16:1985–1999. Kellerbauer A, Bollen G, Dilling J, Henry S, Herfurth F, Kluge H-J, Lamour E, Moore RB, Scheidenberger C, Schwarz S, Sikler G, Szerypo J. 2002. A linear radiofrequency quadrupole ion trap for the cooling and bunching of radioactive ion beams. Nucl Phys A 701:565c–569c. Kip AF. 1969. Fundamentals of electricity and magnetism. New York: McGraw-Hill. Konenkov NV. 1993. Influence of fringing fields on the acceptance of a quadrupole mass filter operated in the separation mode of the intermediate stability region. Int J Mass Spectrom Ion Proc 123:101– 105. 958 Konenkov NV. 1997. Effect of an edge field on the acceptance of a quadrupole mass filter operating at the bottom vertex of the stability rectangle. Tech Phys 42:1220–1222. Konenkov NV, Dowell JT. 1997. Transmission of the quadrupole mass filter in the intermediate stability zone separation mode. Int J Mass Spectrom Ion Proc 164:201–206. Konenkov NV, Kratenko VI. 1991. Characteristics of a quadrupole mass filter operated in the separation mode of a few stability regions. Int J Mass Spectrom Ion Proc 108:115–136. Konenkov NV, Mogilchenko GA, Silakov SS, Shagimuratov GI. 1990. Quadrupole mass spectrometer with high resolution ability. Prib Tekh Eksp 2:179–181. Konenkov NV, Silakov SS, Mogilchenko GA. 1991. Isotopic analysis of hydrogen with a quadrupole mass spectrometer with implantation of deuterium in titanium. J Tech Phys 17:21–24. Konenkov NV, Mogilchenko GA, Silakov SS. 1992. The intermediate domain of QMF stability-sorting-out time and high resolution. J Tech Phys (Leningrad) 60:148–152. Konenkov NV, Cousins LM, Baranov VI, Sudakov MY. 2001. Quadrupole mass filter operation with auxiliary quadrupolar excitation: Theory and experiment. Int J Mass Spectrom 208:17–27. Konenkov NV, Sudakov M, Douglas DJ. 2002. Matrix methods for the calculation of stability diagrams in quadrupole mass spectrometry. J Am Soc Mass Spectrom 13:597–613. Konenkov NV, Korolkov AN, Machmudov M. 2005. Upper stability island of the quadrupole mass filter with amplitude modulation of the applied voltages. J Am Soc Mass Spectrom 16:379–387. Konenkov NV, Londry F, Ding C, Douglas DJ. 2006. Linear quadrupoles with added hexapole fields. J Am Soc Mass Spectrom 17:1063–1073. Konenkov NV, Zhao X, Xiao Z, Douglas DJ. 2007. Mass analysis in islands of stability with linear quadrupoles with added octopole fields. J Am Soc Mass Spectrom 18:826–834. Krutchinsky AN, Chernusevich IV, Spicer VL, Ens W, Standing KG. 1998a. Collisional damping interface for an electrospray ionization time-offlight mass spectrometer. J Am Soc Mass Spectrom 9:569–579. Krutchinsky AN, Loboda AV, Spicer VL, Dworschak R, Ens W, Standing KG. 1998b. Orthogonal injection of matrix-assisted laser desorption/ ionization ions into a time-of-flight spectrometer through a collisional damping interface. Rapid Commun Mass Spectrom 12:508–518. Labastie P, Doy M. 1988. A high mass range quadrupole spectrometer for cluster studies. Int J Mass Spectrom Ion Proc 91:105–112. Landais B, Beaugrand C, Capron-Dukan L, Sablier M, Somonneau G, Rolando C. 1998. Varying the radio frequency: A new scanning mode for quadrupole mass analyzers. Rapid Commun Mass Spectrom 12: 302–306. Lee-Whiting GE, Yamazaki L. 1971. Semi-analytical calculations for circular quadrupoles. Nucl Instrum Methods 94:319–332. Light-Wahl KJ, Schwartz BL, Smith RD. 1993. Observation of the multimeric forms of concavalin A by electrospray ionization mass spectrometry. J Am Chem Soc 115:5869–5870. Light-Wahl KJ, Schwartz BL, Smith RD. 1994. Observation of the noncovalent quaternary associations of proteins by electrospray ionization mass spectrometry. J Am Chem Soc 116:5271–5278. Loboda A, Krutchinsky A, Loboda O, McNabb J, Spicer V, Ens W, Standing K. 2000. Novel Linac II electrode geometry for creating an axial field in a multipole ion guide. Eur J Mass Spectrom 6:531–536. Londry FA, Hager JW. 2003. Mass selective axial ion ejection from a linear quadrupole ion trap. J Am Soc Mass Spectrom 14:1130–1147. Makarov AA. 1996. Resonance ejection from the Paul trap: A theoretical treatment incorporating a weak octapole field. Anal Chem 68:4257– 4263. Mansoori BA, Dyer EW, Lock CM, Bateman K, Boyd RK, Thomson BA. 1998. Analytical performance of a high-pressure radio frequency-only Mass Spectrometry Reviews DOI 10.1002/mas LINEAR QUADRUPOLES IN MS quadrupole collision cell with an axial field applied by using conical rods. J Am Soc Mass Spectrom 9:775–788. March RE, Hughes RJ. 1989. Quadrupole storage mass spectrometry. New York: John Wiley. Marchand P, Marmet P. 1964. Operation d’un filtre de mass quadrupolaire en condition optimum. Can J Phys 42:1914–1919. Marmet P, Proulx M. 1982. A frequency-swept quadrupole mass filter. Int J Mass Spectrom Ion Phys 42:3–10. McIntosh BJ, Hunter KL. 1989. Influence of realistic fringing fields on the acceptance of a quadrupole mass filter. Int J Mass Spectrom Ion Proc 87:165–179. Mellor DJ. 1971. A variable frequency supply for a quadrupole massspectrometer. Vacuum 21:607. Michaud AL, Frank AJ, Ding C, Zhao X, Douglas DJ. 2005. Ion excitation in a linear quadrupole ion trap with an added octopole field. J Am Soc Mass Spectrom 16:835–849. Miseki K. 1993. Quadrupole mass spectrometer. US Patent 5,227,629. Moradian A. 2007. New methods of mass analysis with quadrupoles with added octopole fields. Ph.D. Thesis, University of British Columbia, 2007. Moradian A, Douglas DJ. 2007. Experimental investigation of mass analysis using an island of stability with a quadrupole with 2.0% added octopole field. Rapid Commun Mass Spectrom 21:3306–3310. Moradian A, Douglas DJ. 2008. Mass selective axial ejection from linear quadrupoles with added octopole fields. J Am Soc Mass Spectrom 19:270–280. Morris M, Thibault P, Boyd RK. 1994. Characterization of a high-pressure quadrupole collision cell for low-energy collision-induced dissociation. J Am Soc Mass Spectrom 5:1042–1063. Nie ZX, Kin C-W, Peng W-P, Lee YT, Chang HC. 2006. Frequency scan of a quadrupole mass analyzer in the third stability region for protein analysis. J Chin Chem Soc 53:47–52. Nieminen A, Huikari J, Jokinen A, Äystö J, Campbell P, Cochrane ECA. 2001. Beam cooler for low-energy radioactive ions. Nuclear Instrum Methods Phys Res A 469:244–253. Paul W. 1990. Electromagnetic traps for charged and neutral particles. Rev Mod Phys 62:531–540. Paul W, Raether M. 1955. Das elektrische massenfilter. Z Phys 140:262–273. Paul W, Reinhard HP, von Zahn U. 1958. Das elektrische massenfilter als massenspektrometer und isotopentrenner. Z Phys 152:143–182. Pipes LA. 1953. Matrix solution of equations of the Mathieu-Hill type. J Appl Phys 24:902–910. Regenstreif E. 1967. In: Septier A, editor. Focusing of charged particles V1. New York: Academic Press. p. 401. Reuben AJ, Radchik AV, Smith GB, Vagov AV. 1994. Exact fields in multielectrode ion optics. Rapid Commun Mass Spectrom 8:939–941. Reuben AJ, Smith GB, Moses P, Vagov AV, Woods MD, Gordon DB, Munn RW. 1996. Ion trajectories in exactly determined quadrupole fields. Int J Mass Spectrom Ion Proc 154:43–59. Richards JA. 1977. Quadrupole mass filter spectrum control using pulse width modulation. Int J Mass Spectrom Ion Phys 24:219–224. Richards JA, Huey RM, Hiller J. 1973. A new operating mode for the quadrupole mass filter. Int J Mass Spectrom Ion Phys 12:317–339. Ross DN, Leck JH. 1983. A novel quadrupole mass spectrometer operating with no DC voltage components on the rods. Int J Mass Spectrom Ion Phys 49:1–9. Schneider BB, Lock C, Covey TR. 2005. AP and vacuum MALDI on a QqLIT instrument. J Am Soc Mass Spectrom 16:176–182. Schulte J, Shevchenko PV, Radchik AV. 1999. Nonlinear field effects in quadrupole mass filters. Rev Sci Instrum 70:3566–3571. Schwartz JC, Senko MW, Syka JEP. 2002. A two-dimensional quadrupole ion trap mass spectrometer. J Am Soc Mass Spectrom 13:659–669. Mass Spectrometry Reviews DOI 10.1002/mas & Shagimuratov G, Konenkov NV, Mogilchenko GA, Silakov SS. 1990. Method of separation of ions in the II stability region of the quadrupole mass filter. J Tech Phys (Leningrad) 60:112–116. Sheretov EP. 2000. Opportunities for optimization of the RF signal applied to electrodes of quadrupole mass spectrometers. Part 1. General theory. Int J Mass Spectrom 198:83–96. Sheretov EP. 2002. Some properties of charged particle trajectories in quadrupole mass spectrometers. Part I. General theory. Int J Mass Spectrom 219:315–324. Sheretov EP, Gurov VS, Kolotilin BI. 1999. Modulation parametric resonances and their influence on stability diagram structure. Int J Mass Spectrom 184:207–216. Sheretov EP, Kolotilin BI, Veselkin NV, Brykov AV, Fedosov E. 2000. Opportunities for optimization of the RF signal applied to electrodes of quadrupole mass spectrometers. Part II. The results of trajectory analysis. Int J Mass Spectrom 198:97–111. Sheretov EP, Philippov IV, Karnav TB, Fedosov EV, Ivanov VW. 2002. Some properties of charged particle trajectories in quadrupole mass spectrometers. Part II. The results of trajectory analysis. Int J Mass Spectrom 219:325–341. Shushan B, Douglas DJ, Davidson WR, Nacson S. 1983. The role of kinetic energy in triple quadrupole collision induced dissociation (CID) experiments. Int J Mass Spectrom Ion Phys 46:71–74. Smythe WR. 1939. Static and dynamic electricity. New York: McGraw-Hill Book Company. Sobott F, Hernandez H, McGammon MG, Tito MA, Robinson CV. 2002. A tandem mass spectrometer for improved transmission and analysis of large macromolecular assemblies. Anal Chem 74:1402–1407. Standing KG. 2000. Timing the flight of biomolecules: A personal perspective. Int J Mass Spectrom 200:597–610. Sudakov M, Douglas DJ. 2003. Linear quadrupoles with added octopole fields. Rapid Commun Mass Spectrom 17:2290–2294. Sudakov M, Nikolaev E. 2002. Ion motion stability diagram for distorted square waveform trapping voltage. Eur J Mass Spectrom 8:191–199. Sudakov M, Konenkov N, Douglas DJ, Glebova T. 2000. Excitation frequencies of ions confined in a quadrupole field with quadrupole excitation. J Am Soc Mass Spectrom 11:10–18. Szabo I. 1986. New ion-optical devices utilizing oscillatory electric fields. I. Principle of operation and analytical theory of multipole devices with two dimensional fields. Int J Mass Spectrom Ion Proc 73:197–235. Szylagyi M. 1988. Electron and ion optics. New York: Plenum Press. p. 62. Tanner SD, Baranov VI, Bandura DR. 2002. Reaction cells and collision cells for ICP-MS: A tutorial review. Spectrochim Acta B 57:1361–1452. Taylor S, Gibson JR. 2008. Prediction of the effects of imperfect construction of a QMS filter. J Mass Spectrom 43:609–616. Thomson BA. 1998. 1997 McBryde medal award lecture. Radio frequency quadrupole ion guides in modern mass spectrometry. Can J Chem 76:499–505. Thomson BA, Jolliffe C. 1998. Spectrometer with axial field. US Patent 5,847,386. Thomson BA, Douglas DJ, Corr JJ, Hager JW, Jolliffe C. 1995. Improved collisionally activated dissociation efficiency and mass resolution on a triple quadrupole mass spectrometer system. Anal Chem 67:1696– 1704. Thomson B, Jolliffe C, Javahery R. 1996. Proceedings of the 44th ASMS Conference on Mass Spectrometry and Allied Topics, May 12–16, Portland, OR, p. 1155. Titov VV. 1995. Ion separation in imperfect fields of the quadrupole mass analyzer Part III. Transmission and optimal ion injection. Int J Mass Spectrom Ion Proc 141:37–43. Titov VV. 1998. Detailed study of the quadrupole mass analyzer operating within the first, second and third (intermediate) stability region. II. Transmission and resolution. J Am Soc Mass Spectrom 9:70–87. 959 & DOUGLAS Turner P, Taylor S, Gibson JR. 2005. Effect of ion entry acceptance conditions on the performance of a quadrupole mass spectrometer operated in upper and lower stability regions. J Vac Sci Technol 23:480–487. Von Zahn U. 1962. Prazisions-massenbestimmungen mit dem elektrischen massenfilter. Z Phys 168:129–142. Winger BE, Light-Wahl KJ, Loo RR, Udseth HR, Smith RD. 1993. Observation and implications of high mass-to-charge ratio ions from electrospray ionization mass spectrometry. J Am Soc Mass Spectrom 4:536–545. Wuerker RF, Shelton H, Langmuir RV. 1959. Electrodynamic containment of charged particles. J Appl Phys 30:342–349. Xiao Z, Zhao X, Douglas DJ. 2008. Mass selective axial ion ejection from a linear quadrupole with an 8% added hexapole field. Presented at the 56th ASMS Conference on Mass Spectrometry and Allied Topics, Denver, CO. 960 Xu HJ, Wada M, Tanaka J, Kawakami H, Katayama I, Ohtani S. 1993. A new cooling and focusing device for ion guide. Nucl Instrum Methods Phys Res A 333:274–281. Yang J, Leck JH. 1982. Operation of the quadrupole mass spectrometer in the ac mode. Vacuum 32:691–694. Yang J, Leck JH. 1984. The operation of a novel quadrupole mass spectrometer in hostile atmospheres. Int J Mass Spectrom Ion Proc 60:127–136. Ying J-F, Douglas DJ. 1996. High resolution inductively coupled plasma mass spectra with a quadrupole mass filter. Rapid Commun Mass Spectrom 10:649–652. Zhao X, Xiao Z, Moradian A, Konenkov NV, Douglas DJ. 2007. Experimental investigation of mass analysis with linear quadrupoles with added multipole fields operated in islands of stability. Presented at the 55th ASMS Conference on Mass Spectrometry and Allied Topics, Indianapolis, IN. Mass Spectrometry Reviews DOI 10.1002/mas