B American Society for Mass Spectrometry, 2016
J. Am. Soc. Mass Spectrom. (2017) 28:1066Y1078 DOI: 10.1007/s13361-016-1548-0
FOCUS: HONORING R. G. COOKS' ELECTION TO THE NATIONAL ACADEMY OF SCIENCES: RESEARCH ARTICLE
High Mass Ion Detection with Charge Detector Coupled to Rectilinear Ion Trap Mass Spectrometer Avinash A. Patil,1 Szu-Wei Chou,1,2 Pei-Yu Chang,1 Chen-Wei Lee,1 Chun-Yen Cheng,2 Ming-Lee Chu,3 Wen-Ping Peng1 1
Department of Physics, National Dong Hwa University, Shoufeng, Hualien, Taiwan 97401, Republic of China AcroMass technologies Inc., Hukou, Hsinchu, Taiwan 30352, Republic of China 3 Institute of Physics, Academia Sinica, Taipei, Taiwan, Republic of China 2
Abstract. Conventional linear ion trap mass analyzers (LIT-MS) provide high ion capacity and show their MSn ability; however, the detection of high mass ions is still challenging because LIT-MS with secondary electron detectors (SED) cannot detect high mass ions. To detect high mass ions, we coupled a charge detector (CD) to a rectilinear ion trap mass spectrometer (RIT-MS). Immunoglobulin G ions (m/z ~150,000) are measured successfully with controlled ion kinetic energy. In addition, when mass-to-charge (m/z) ratios of singly charged ions exceed 10 kTh, the detection efficiency of CD is found to be greater than that of SED. The CD can be coupled to LIT-MS to extend the detection mass range and provide the potential to perform MSn of high mass ions inside the ion trap. Keywords: Voltage-scan rectilinear ion trap mass analyzer, Charge detector, Orthogonal wavelet packet decomposition, Noise rejection, High mass proteins, LDI/MALDI ions, Scan rate, Buffer gas pressure, Quadrupole ion guide Received: 9 July 2016/Revised: 25 October 2016/Accepted: 28 October 2016/Published Online: 13 December 2016
Introduction
L
inear ion traps (LIT) have drawn much attention because of their high ion capacity and MSn ability [1]. In addition, LIT can be easily incorporated into a triple quadrupole mass spectrometer [2–4] and coupled with high resolution mass analyzers in hybrid instruments [5–8]. In the development of macromolecule ion/ion chemistry [9], there is huge demand to extend the m/z range of an ion trap mass analyzer. Especially, top-down proteomic and metabolomic analyses are expected to become critically important for the analysis of intact proteins and biomedical samples [10]. Typically an electrospray ionization (ESI) source is coupled with LIT-MS to yield mass spectra of high mass molecules with reduced m/z by increasing z value, e.g., human hemoglobin α-chain with 17 charges (m/z ~895Th) [11] and bovine cytochrome c
Electronic supplementary material The online version of this article (doi:10. 1007/s13361-016-1548-0) contains supplementary material, which is available to authorized users. Correspondence to: Wen–Ping Peng; e-mail:
[email protected]
with 16 charges (m/z ~772Th) [12]. However, multiply charged ions generated by ESI often lead to very complex multimer mass spectra of high mass biomolecules (>100,000 Da) and increase the analysis difficulties [13, 14]. The multiply charged ions also increase the space-charge effect inside ion traps and thus lower the detection sensitivity. To reduce these disadvantages observed with ESI, matrixassisted laser desorption ionization (MALDI) [15] was applied to generate singly and doubly charged ions [16]. But, the narrow m/z range (with m/z ratios less than 4000 Th) of most commercial LIT-MS prevents the use of MALDI for the measurement of large biomolecular ions [17]. Therefore, extending LIT-MS m/z range is critical to detect high mass ions with low charge state. Three key factors contribute to the extension of LIT-MS m/z range: radio frequency (rf) of ion traps, ion kinetic energy, and ion detectors [17, 18]. To analyze high mass ions with LIT-MS, the ion trap (rf) must be tuned to low values [19]. Hager et al. [20] developed a mass selective axial ejection (MSAE) method for LIT to detect doubly charged BSA (m/z 33,000 Th) ions at 10 kHz axial resonance ejection frequency. Moreover, Koizumi et al. developed a large LIT operated in MSAE mode at
A. A. Patil et al.: Charge Detector with Rectilinear Ion Trap Mass Spectrometer
axial resonance ejection frequency of 7 kHz to detect singly charged BSA (m/z ~66,000 Th) ions generated with a charge reduced electrospray (ESI) ion source [21]. In LIT-MS, reducing ion kinetic energy (K.E.) can improve ion trapping, which is achieved by increasing the ion trapping pressure. Buffer gas pressure is set at ~1 mTorr to cool down and focus ions at the center of traps [22, 23]. With such pressure, the acceleration voltage (dynode voltage) of the secondary electron detector (SED) is often restricted to ~ −10 kV and this low dynode voltage limits the detection of high mass ions (>50,000 Da) in LIT-MS with a SED [24]. To increase the dynode voltage of SED, Chen et al. designed differential pumping stages in a frequency-scan LIT-MS to maintain high trapping pressure (~60 mTorr) inside the LIT and low pressure (5 × 10−5 Torr) outside the LIT [24]. High trapping pressure reduced the K.E. of injected MALDI ions. Radially ejected ions could be accelerated with a dynode voltage of −25 kV to gain enough kinetic energy to leading successful detection of singly charged immunoglobulin G (IgG) ions. However, such differential pumping designs are too complicated to be adopted in a commercial LIT-MS instrument. Conversion dynode/electron multiplier of SED is mostly employed in LIT-MS but it cannot detect ions at high m/z range [25, 26]. Several reports indicate that ion detection efficiency of SED mainly depends on the velocity of incident ions [25–29] and the detection efficiency is close to zero when the ion velocity is less than 1 × 106 cm/s [25]. At a given ion energy, the velocity of ion is inversely proportional to the root square of the mass (z = 1) [25]. Therefore, SED is not an appropriate detector to detect high mass ions with LIT-MS and the design of a proper detector that can be incorporated with LIT-MS is necessary. The aim of this study is to detect high mass ions by coupling a charge detector (CD) [30–35] to a rectilinear ion trap (RIT) mass analyzer developed by Cooks et al. [36, 37]. RIT is a modified design of LIT, where rectangular geometry electrode pairs are used to build a trap that increases ion trapping capacity [36–38]. To achieve the aim, we reduced ion trap rf, quenched ion kinetic energy, and increased detection efficiency by the CD. CD can work in high pressure conditions with pressure from 50 to 150 mTorr [30–32]. CD combined with RIT could detect image charges induced by ions, which means the ion detection with CD is independent of ion masses and velocities [30–34, 39–41]. However, CD is prone to pick up the strong rf fields from RIT electrodes and the strong rf fields saturate CD. To reduce rf field interferences on CD, we first shielded CD by a grounded copper cap with a guarding mesh, so the rf fields could be greatly reduced and CD would not be saturated. After that we employed orthogonal wavelet packet decomposition (OWPD) theory to analyze all rf noise components and rf field interferences could be completely removed. Then we reduced the K.E. of bovine serum albumin (BSA) ions and IgG ions by introducing 40 mTorr helium gas and 30 mTorr argon gas in guiding quadrupole (Q1) region. At last singly charged BSA (66 kDa) and
1067
IgG (150 kDa) ions were detected by CD RIT-MS successfully with mass resolution of 120 and 90 (m/Δm, FWHM) and S/N ratio of 42 and 28, respectively. The detection surface area of CD could be increased considerably in a RIT-MS instrument while its noise characteristics remained at low level (~450 ions), i.e., the signal to noise (S/N) ratio is improved especially in high m/z range. When comparing the detection efficiency of CD with that of SED, we found CD was superior to SED when the ion (m/z) was over 10 kTh. In application, CD RIT-MS might potentially be used to study macromolecules ion/ion proton transfer chemistry for manipulating charge states [11], where final product ion mostly carries a single charge, which is unable to be detected with present linear ion trap mass spectrometers.
Experimental Instrument A schematic of a voltage-scan charge detection rectilinear ion trap mass spectrometer (V-Scan CD RIT-MS) is shown in Figure 1. A laser desorption ionization (LDI) ion source and a matrix assisted laser desorption/ionization (MALDI) ion source were adopted to generate ions which were guided through a linear quadrupole ion guide Q1 (P/N: 97055– 20106, a quadrupole ion guide by Thermo-Fisher Scientific, Waltham, MA, USA) with a diameter of 5.4 mm (spacing between two opposite rods) and a length of 28.6 mm. After ions exited Q1, they entered a rectilinear ion trap (RIT) which acted as a mass analyzer with two endcap lenses having 3 mm openings at the center for laser path and ion entrance. DC potentials were set on the endcap lenses for ion confinement along the z axis. The RIT was constructed with four 10 cm long stainless steel rods and each was separated from the center of the trap by a distance of 5 mm × 4 mm in the x and y directions, respectively. The slits (0.5 mm in width and 34 mm in length) were at the center of the x electrodes for the radial ejection of ions. The apparatus was comprised of two detectors, a conversion dynode/electron multiplier detector and a CD. Two detectors were placed on opposite sides of the x electrodes of RIT mass analyzer. The z-axis spacing between the adjacent optical elements was maintained at ~1 mm for all ion optical elements. Half of the volume of Q1 was located in chamber 2 (RIT chamber), which was covered by an insulating cabinet to maintain a pressure gradient between the two chambers.
Electronics The arrangement of RIT-MS with a CD, a SED, and rf power circuit is shown in Figure 1b. Two function generators (DS345; Stanford Research Systems, Sunnyvale, CA, USA), and a 1000 W power amplifier (Powertron model 1000A; Industrial Test Equipment Co. Inc., Port Washington, NY, USA ) were used to boost RIT and a 50 W power amplifier (Powertron model 50A; Industrial Test Equipment Co. Inc. Port Washington, NY, USA) was used to boost the quadrupole ion guide (Q1);
1068
A. A. Patil et al.: Charge Detector with Rectilinear Ion Trap Mass Spectrometer
Figure 1. (a) Schematic presentation of a charge detector rectilinear ion trap mass spectrometer (CD RIT-MS). The vacuum system is divided into three stages. Stage 1 is equipped with a rotary vane pump to evacuate the region at the sample probe (LDI/MALDI target) inlet. Stage 2 (chamber 1) and stage 3 (chamber 2) are equipped with a rotary vane pump and a turbo pump, respectively. (b) Schematic presentation of the arrangement of RIT-MS with a charge detector (CD), a secondary electron detector (SED), and the rf power supply. A mesh is placed in front of CD. The ion signals are acquired simultaneously with a CD and a SED. (c) The rf voltage scan function for ion trapping and mass analysis. (d) Image of a modified rectangular shape charge detector
two homemade air coil transformers, a homemade charge detector, a conversion dynode/electron multiplier detector, two synchronized data acquisition (DAQ) cards (NI PCI 6251 and NI PCI 6281; National Instrument, Austin, TX, USA), and a personal computer were adopted to control the mass scan function of CD RIT-MS and collect data. The rf sinusoidal signal was provided by a function generator and the rf signal was amplified by a 50 W power amplifier to boost the ion guide Q1 through a homemade air coil transformer and then applied to the X- and Yelectrodes of the Q1. The rf amplitude was modulated from the analog output of a DAQ card and the rf signal was boosted by a 1000 W current amplifier and an air coil transformer, and then applied to the X- and Y-electrodes of RIT (Figure 1b). The rf voltages on two pole pairs of Q1 and RIT were monitored by
Agilent’s ( Santa Clara, CA, USA) DSOX2WAVEGEN 2000 X series oscilloscope to ensure rf balancing [42] between the two pole pairs of quadrupoles. A mass selective instability scan at q= 0.908 with boundary ejection [18] was adopted to eject ions. The scan function is shown in Figure 1c. We applied dynamic rf trapping (i.e., the trapping voltage was gradually increased during ion injection) to increase the ion trapping efficiency. The rf voltage was then kept fixed for ion cooling. Finally, the trapped ions were ejected radially by ramping the rf voltage.
Vacuum The vacuum system was designed with three stages. In stage 1, the sample probe inlet evacuated by a rotary vane pump and the
A. A. Patil et al.: Charge Detector with Rectilinear Ion Trap Mass Spectrometer
atmospheric pressure region were separated. In stage 2 (chamber 1) and stage 3 (chamber 2), chamber 1 and chamber 2 were separated and vacuum was achieved by a rotary vane pump and a turbo pump (Osaka Vacuum, model TG 350FCAB, 210 L/s, Osaka, Japan), respectively. Samples were loaded on a stainless steel probe (LDI/MALDI target) and a sample probe was introduced by a transfer rod in chamber 1 and evacuated to pressure of 1 × 10−4 Torr by a turbo pump (210 L/s). To cool the protein ions, bath gas (helium or argon) was introduced in chamber 1 (Q1 region) and the pressure was raised from 10 to 40 mTorr. In a full mass scan mode, chamber 2 was evacuated to a pressure of 5 × 10−5 Torr by a turbo pump (210 L/s) and further pressure was maintained at ~0.5–1 mTorr for small molecules (m/z <1000 Th) and 1–2 mTorr for high mass proteins with helium buffer gas. When ions were injected into RIT, the kinetic energy of those ions was decreased by the buffer gas and ions were distributed near the center of trap and confined by DC potentials which were applied on the two endcap lenses.
Samples Ubiquitin from bovine erythrocytes (MW ~8.56 kDa), cytochrome c from bovine heart; MW ~ 12.30 kDa), bovine serum albumin (BSA; MW ~ 66 kDa), and immunoglobulin G from bovine serum (IgG; MW ~150 kDa) and MALDI matrices were purchased from Sigma. All analytical grade reagents and fullerene (C60, purity >99%, with C70 as major impurity) were purchased from UniRegion Bio-Tech (Taipei, Taiwan). Toluene, acetone, methanol and acetonitrile were purchased from J. T. Baker (Phill ipsburg, NJ, USA). The above compounds were used as received. Deionized water was purified to 18.2 MΩ/cm by Milli-Q water purification system (Millipore, Billerica, MA, USA). Herein, we used Li’s two-layer sample preparation method [43] to prepare protein samples in which the matrix solution with a concentration of 6 mg/mL in 60% methanol/acetone (v/v) of sinapinic acid (SA) was placed and dried on a sample probe to form microcrystals as the first layer. Then, a solution containing both the analyte and SA matrix was added to the top of the matrix layer as the second layer. A second matrix layer solution was prepared in 50% acetonitrile/water (v/v) that contained 0.1% trifluoroacetic acid with a concentration of 1 mg/100 μL for SA. Roughly, a 6 μL mixture of sample and matrix was added onto the stainless sample probe (8 mm diameter) and then air-dried. The protein-to-matrix ratio is cytochrome c/ubiquitin:SA = 1:1,000, BSA:SA = 1:30,000, IgG:SA =1:100,000, respectively. Six μL of a 1 mM solution of C60 was prepared in toluene and loaded on a sample probe and then air dried.
Ion Source A third harmonic frequency of Nd:YAG laser (model LS2139, LOTIS TII; Minsk, Republic of Belarus) of wavelength 355 nm was used to generate ions. The pulse duration was 16 ns and a pulse repetition rate of 100 Hz was adopted. The laser beam was focused on a spot diameter of ~300 μm. The UV laser was
1069
aligned to pass through RIT and the guiding quadrupole and hit the target probe with a diameter of 8 mm. The mass spectrum of low mass ions (<1000 Th) was obtained with a laser energy density ~0.75 mJ/cm2. The cytochrome c mass spectrum was obtained with a higher laser energy density ~1.4 mJ/cm2 and the laser energy was further increased to ~1.8 mJ/cm2 for BSA ion generation. Laser energy of 1.5 mJ/cm2 was used for IgG ions.
Detectors There were two detectors, a SED, and a CD, placed in the x direction of RIT. Both detectors were employed simultaneously to acquire mass spectra. Approximately 50% of the ejected ions were detected by each detector. The configuration of two detectors is shown in Figure 1a and b. Secondary Electron Detector (SED) The conversion dynode/ electron multiplier detector was set on one side of the Xelectrode of RIT. A DeTech 397 detector assembly (Detector Technology, Inc., Palmer, MA, USA) was adopted in this experiment. The detector assembly had a stainless steel case that shielded the conversion dynode and the electron multiplier, which minimized interference from the applied high voltages. An opening of about 12.5 mm diameter on the detector shielding casing allows ions to enter the detector assembly. Ions ejected from the center part of the slit were detected by SED. Ions that ejected from RIT slit at the side of SED were accelerated into the conversion dynode by DC potential. Ions were accelerated to impact the conversion dynode surface with high energy, and secondary electrons were generated in this process. Further secondary electrons were amplified exponentially by successive secondary electron generation within the electron multiplier tube. The conversion dynode was operated at −4 kV to detect low mass ions (<1 k Th) and at −8 kV to detect ubiquitin and cytochrome c ions. The electron multiplier was set at −1.8 kV to detect positive ions. The velocities of ions radially ejected by the mass selective instability method were very low compared with the ion velocities post-accelerated in a MALDI-TOF mass analyzer. Thus, the low ion velocity characteristic made ion detection challenging with SED in a RITMS. The conversion dynode voltage adopted in RIT-MS should be set very high (> −4 kV) to accelerate those ions toward the conversion dynode with high impact energy (incident ion kinetic energy) of several keV. With such a high kinetic energy, ions were converted to secondary electrons. However, the conversion efficiency of a SED detector strongly depends on the ejected ion velocity and dynode voltage. Charge detector (CD) Image charges were induced by the ejected ions and the image charge signal intensity was proportional to the number of ions. The circuitry of the charge detector was similar to that used in a cell mass spectrometer (CMS) [30, 41], but some modifications were made in the design to fit RIT geometry and RC discharging time. The modifications included a rectangular CD with a rectangular shaped Faraday plate with
A. A. Patil et al.: Charge Detector with Rectilinear Ion Trap Mass Spectrometer
Removal of rf Noise in a Voltage-Scan RIT MS We found the charge detector experiences strong rf field from RIT and the induced rf noise is very high and saturates the charge detector (Supplementary Figure S1a, in the Supporting Information). To tackle the rf field interference, we first added a shielding copper cap with an opening slit of 3 mm × 34 mm covered by a nickel mesh (98% transmission) over the charge detector to reduce the strong rf field (Supplementary Figure S1b) according to skin effect theory [44]. The shielding copper cap and mesh could mostly prevent the strong rf field from saturating the charge detector and it suppressed the induced rf noise. The mesh could not only prevent rf interference but also allowed ion transmission to the Faraday detector. Nonetheless, some portion of rf noise was observed in ion signals and the induced rf signals increased linearly when the scanning voltage was swept linearly as shown in Figure 2a and
400
600
800
1000
1200
Charge Detector Signal without Denoising 3k +
C60
12k 6k
+
C50
0
S/N ~ 6
Number of Ions
18k
S/N ~ 2
(a)
-3k 550
600
650
700
750
0
(b) 15k
Charge Detector Signal with Denoising 3k
+
C60
10k
+
C50
+
C52 C+ C56 C+ +
5k
+
C50
S/N ~ 18
54
S/N ~ 4
Number of Ions
58
0
600
650
700
750
0 0.6
Secondary Electron Detector Signal +
C60
0.4
0.2
+
C50
S/N ~ 40
(c)
S/N ~ 12
dimensions of 36 mm in length and 5 mm in width. Image charges were proportional to the number of ejected ions and they were induced while ions flew close to the rectangular shaped Faraday plate. The leading edge of the signal was due to image charge induction and the falling edge was derived from ions landing on the surface and the RC discharging time. The charge detector circuitry is comprised of two parts. The first part of the charge detector circuit includes a low-noise JFET and an operational amplifier. The discharge time constant is the time taken by CD to discharge the voltage generated across capacitor from image charge generation on the detection plate [30] and is determined by the RC value and set to 3RC time constants (RC time constant is 180 μs). The second part includes a voltage amplifier and a simple band-pass filter to filter out both high frequency and low frequency noise. The designed CD is depicted in Figure 1d. With this rectangular shape charge detector, the effective detection area was greatly increased, which allows detection of ions ejected through the slit with a width of 0.5 mm and length of 34 mm. Therefore, the signal-to-noise ratio (S/N) of detected ions was greatly improved. To reduce the rf field interference generated from RIT electrode voltage, a copper cap of 2 mm thickness with an opening 35 mm in length and 3 mm in width at the center was covered with a nickel mesh (98% transmission) set over CD. The effective detection area of the rectangular CD was increased by about a factor of 3 compared with the circular shape Faraday plate CD used in a threedimensional QIT [30, 40, 41]. In comparison with SED, CD could detect all ions ejected at the side of CD from RIT slit. Calibration of CD was carried out by applying a square pulse with known amplitude to the Btest pulse input^ connector. The test pulse calibration gave a charge conversion gain of ~28 electrons/mV. Signal intensities of the mass spectrum detected by CD were obtained in units of Volt (VCD), which was further converted into the number of ions by the charge conversion ratio. The number of ions presented in the mass spectra was calculated from the measured signal amplitude (the rising term, i.e., image charge) of the charge detector.
Intensity (a.u.)
1070
0.0
400
600
800
1000
1200
m/z (Th)
Figure 2. Mass spectra of C60 and C60 clusters obtained by a voltage scan setting from 100 to 600 Vp-p at a fixed rf frequency at 350 kHz. (a) Charge detector signal without denoising. The inset signal is a mass spectrum from 550 to 750 Th. (b) Charge detector signal with denoising. The inset signal is an enlarged þ þ þ þ mass spectrum of C þ 50 , C 52 , C 54 , C 56 , and C 58 ions. (c) Mass
spectrum obtained with a secondary electron detector
Figure 3a. Therefore, we developed an algorithm based on the OWPD theory to completely remove the rf field interference without any signal distortion [45]. Based on our previous study, we have developed an OWPD technique to remove the rf induced noise in a frequency-scan quadrupole ion trap mass spectrometer [41]. But in a voltage-scan RIT-MS, the noise feature was completely different from the frequency-scan mode. Therefore, it was necessary to develop a new voltagescan noise analysis model to remove the noise induced from the strong rf field in CD RIT-MS. To analyze signals and reduce all noise interferences (rf fields and power sources) from a voltage-scan charge detection RIT mass spectrometer (V-scan CD RIT-MS), an observation model of time domain signal S is described as S ¼ Sx þ N r f þ N h þ N w
ð1Þ
where S x is the signal induced by ions, N r f the interference induced by rf fields, N h the sum of harmonics induced by AC power sources, and N w the white noise. Both N h and N w are
A. A. Patil et al.: Charge Detector with Rectilinear Ion Trap Mass Spectrometer
42k
[M+H]
28k
S/N ~7
14k
24k
28k
32k
36k
Charge Detector Signal Before Denoising
[2M+H]
+
[M+H]
Charge Detector Signal After Denoising
S/N ~ 80
28k
+
14k
[2M+H]
+
0
[M+H]
0.8 0.6
+
Parameters for Ion Trapping and Cooling
S/N ~ 2
0.4 0.2
In the absence of buffer gas, trapping of molecular ions is determined by pseudo-potential well depth (Du=x,y) along the x, y plane of RIT [23], which can be described as [2M+H]
+
0.0
8k
12k
electrodes is l =4. As shown in Eq. 2, N r f is a superposition of interference from the rf voltage of each electrode. In the voltage-scan mode, ρi ðt Þ is a product of a straight line and a sinusoidal waveform at f r f and it can be completely removed by the OWPD method. A detailed description of wavelet packet decomposition coefficient of signal is described in Supplementary Figures S2–S4. A possible way to reduce the rf interference was to adopt single phase rf configuration, which applied rf voltage to Yelectrodes while X-electrodes were grounded. With this configuration, the rf noise interferences could be greatly reduced. But the trapping efficiency of externally injected ions was reduced with single-phase rf configuration because the kinetic energy of the ions in the x-y plane became significantly higher after entering the ion trap axially [22]. Although the single phase rf configuration could greatly reduce the induced rf noises, the OWPD algorithm is indispensable for the removal of the rf noise interferences completely.
Secondary Electron Detector Signal
S/N ~ 60
Number of Ions Intensity (a.u.)
+
20k
0
(b) 42k
(c)
16k
S/N ~ 6
Number of Ions
12k
S/N ~ 1
8k
(a)
1071
16k
20k
24k
28k
32k
36k
Du ¼
mqu 2 Ω2 r20 16ze
ð3Þ
m/z (Th) Figure 3. Mass spectra of (a) Charge detector signal without denoising; (b) charge detector signal with denoising; (c) secondary electron detector signal. Singly charged monomer and dimer of cytochrome c ions are obtained with a rf frequency at 140 kHz and the voltage is scanned from 700 to 3200 Vp-p. The S/N ratio of singly charged monomer signal with CD is better than that of SED. The dimer signal observed by CD is more intense and resolved than that of SED detected signal
common noise sources in instruments and can be reduced by signal processing methods [45–52]. The OWPD was used to decompose original signals via band-pass filters to estimate N r f and then remove the N r f to reconstruct the decomposed signals back to the original signals without distortions and rf field interference [45]. A denoising model of N r f is written as follows. N r f ðt Þ ¼
l X
ρi ðt Þ
ð2Þ
where, Ω = 2πf is the rf applied to the trap, r0 is the inner dimension in x or y plane measured from the center of the trap to the inner surface of trap electrode, qu is the Mathieu’s trapping parameter, z is the number of elementary charges carried by ions of mass (m), and e is the elementary charge (1.6 × 10−19 C). According to Eq. 3, Du correlates with the qu parameter and Eq. 3 is valid under the pseudo‐potential wellapproximation by Dehmelt for those ions with qu < 0.4 [18]. In an ion trap system, ions are generated either inside the trap in the presence of buffer gas molecules [24, 32, 53, 54] or outside of the trap [9, 20, 55, 56], and then injected into the trap with controlled K.E.) (i.e., a few electron volts. The control of K.E. is achieved by ion molecule collisions with buffer gas molecules (pressure <1 mTorr) inside of the ion trap [24, 53, 54] or in the guiding quadrupole [9, 20, 55, 56], where ions must satisfy the following condition (Eq. 4) to achieve efficient ion guiding and ion trapping [32]: K:E: ¼
1 2 mv < Ecooling þ Du 2
ð4Þ
i¼1
We assume N r f is proportional to the intensity of the rf electric fields, where t is time, ρi ðt Þ is the rf field interference which is proportional to the rf voltage of i th (i =1,2,3,4, … and so on) electrode at time t; and l is the upper limit of summation [41]. For double phase RIT rf supply, the final number of
where v is the initial velocity of ions, Ecooling is the energy from collisional cooling, and Du is pseudo-potential depth of the trap. Subsequently, Ouyang et al. [23] derived mathematical expression for maximum allowable K.E. (Dt) in the x-y plane for the externally injection of ions in a two-dimensional ion trap, which is determined by
1072
A. A. Patil et al.: Charge Detector with Rectilinear Ion Trap Mass Spectrometer
combining rf trapping and collisional cooling effects, and is expressed as u0
Dt ¼ Du γ− λ
ð5Þ
where, Du is pseudo-potential well depth, γ is the kinetic energy loss ratio of the ion for each collision, λ is mean free path of ions, and μ0 is reduced mass of the ion and neutral pair. Ions cannot be trapped when the K.E. exceeds the trapping potential depth Dt and, therefore, it is crucial to control the ion K.E. to be less than Dt (i.e., K:E: ¼ 12 mv2 < Dt ) in order to trap high mass ions. When the molecular weight of a protein is above 5 kDa, the K.E. of that protein generated in the MALDI process is over 25 eV with an assumption of initial ion velocity of protein ions close to matrix ion velocity, and this K.E. increases linearly with increasing mass [57, 58]. High K.E. greatly reduces the trapping efficiency of high mass ions in an ion trap system, especially when the ion trap is operated at low bath gas pressure. Increasing buffer gas pressure can cool high mass ions effectively, but high pressure (>10 mTorr) also causes serious discharging due to high rf voltage (V rf ) in a mass analyzer especially operated in a voltage-scan mode. High buffer gas pressure changes the trapping conditions (qu value) and reduces the mass resolution of ion trap mass analyzer as well [22]. The high pressure condition is not favorable to trap and scan high mass MALDI ions in a voltage-scan LIT mass analyzer. To solve these problems, earlier studies proposed external ion injection by a linear ion guide operated at high buffer gas pressure [54–56]. In this condition, collision of ions with buffer gas molecules (helium or argon) reduced the K.E. greatly and focused those ions at the center of ion guide, which could be injected through a small aperture at the ion trap entrance and thus improved the transmission of analytes [55, 56].
Ion Injection with Increasing rf Field: Dynamic rf Trapping The broad K.E. energy distribution of MALDI ions made trapping of protein ions in an ion trap challenging [57, 59]; therefore, the K.E. of ions must be sufficiently reduced to trap ions over a wide mass range in a RIT mass analyzer according to Eqs. 4 and 5. Trapping of high m/z ions in a RIT is qu- and potential well depth (Du)-dependent, and the low charge state characteristic of MALDI ions leads to large m/z differences between two nearby mass peaks, so that choosing proper trapping parameters (e.g., qu and Du values, bath gas pressure) is crucial to achieve effective ion cooling and trapping. The setting of trapping voltages and gas pressure in this study were similar to the conditions used in a dual-pressure LIT mass analyzer [60], where the pressure in the first ion trap was held one order of magnitude higher than in the second ion trap. The dual pressure LIT-MS design improved ion isolation, ion fragmentation at the first trap, and mass resolution in the second trap [60]. Here we kept the ion guide (Q1) at a pressure about one order of magnitude higher than that in RIT mass analyzer, which effectively cooled down the ions via collisional cooling in the rf-only quadrupole (Q1) region and improved ion transmission from the ion source to RIT mass analyzer. After ions were sufficiently cooled, we applied the dynamic rf trapping [61] technique prior to ion injection into RIT by gradually increasing rf fields. With dynamic rf trapping, the trapping potential of RIT was maintained at low qu values and then the qu value was gradually increased to a final trapping qu [62, 63]. This process made rf field penetration easy for externally injected ions (K.E. controlled) and, thus, ions could be trapped and cooled down with good efficiency. Therefore, dynamic rf trapping could inject maximum ions into RIT. Besides dynamic rf trapping, the optimization of rf amplitude, rf frequency, and collision cooling gas pressure were crucial factors to trap high mass protein ions, and those parameters were directly correlated to the trapping parameter qu and Du. The qu, Du, low-mass cutoff (LMCO) values, and secular frequencies of ions adopted by Q1 and RIT for ion injection and ion cooling were calculated and are given in Table 1. The LMCO setting of
Table 1. Trapping Parameters (qu), Psudo-Potential Well Depth (Du), Low-Mass Cutoff (LMCO), and Dynamic rf Trapping Settings in Quadrupole Ion Guide (Q1) and RIT Mass Analyzer Protein ions Guiding quadrupole (Q1) qu Du (eV) Ion secular frequency (Hz) LMCO (Th) Rectilinear ion trap (RIT) Initial trapping qu Final trapping qu Initial trapping Du (eV) Final trapping Du (eV) Ion secular frequency (Hz) Initial LMCO (Th) Final LMCO (Th)
[Cyt c + H]+
[2Cyt c + H]+
[BSA + 2H]2+
[BSA + H]+
[IgG + 2H]2+
[IgG + H]+
0.464 24.36 35796 6265
0.232 12.18 17898
0.579 30.44 24593 21078
0.289 15.22 12296
0.575 30.14 16232 47425
0.28 15.07 8116
0.463 0.589 31.82 51.55 29148 6301 8020
0.231 0.253 13.15 19.94 14574
0.385 0.47 21.65 32.34 14128 13987 17095
0.192 0.235 10.82 16.15 7064
0.404 0.454 22.75 33.98 9608 31104 38016
0.202 0.247 11.37 16.99 4803
A. A. Patil et al.: Charge Detector with Rectilinear Ion Trap Mass Spectrometer
Q1 rejects unwanted low mass ions and improves the transmission of ions of interest from Q1 to RIT. The maximum m/z range of CD RIT-MS with mass selective instability method (voltage scan) [54] could be extended by lowering rf frequency of rf supply. However, lowering trapping frequency often leads to reduction in trapping potential depth and ion secular frequency as shown in Table 1, and it leads to reduction of trapping efficiency of high m/z ions. The qu values of the guiding quadrupole Q1 were calculated by using formula, qu = 4eV0-p /mr02Ω2 [39]. The calculation of qu value of RIT was adopted from Cooks et al. with qu = A24eV0-p /mrN2Ω2, where A2 = 0.654 is the quadrupole coefficient in the multipole expansion of the electric potential and rN is the corresponding normalization radius [37].
Results and Discussion Detection of Low Mass Ions with m/z ~1 kTh In Figure 2, C60 ions were guided through Q1 at an rf frequency and amplitude of 650 kHz and 125 Vp-p (qu ~ 0.27) and the pressure at Q1 was maintained to 5 × 10−5 Torr. The mass spectra of C60 ions were acquired by scanning the rf amplitude from 100 to 600 Vp-p at a fixed rf of 350 kHz. The K.E. of C60 ions was below 5 eV [39], so it is not necessary to introduce cooling gas at the Q1 region. Helium gas pressure of 0.8 mTorr was maintained in RIT for ion cooling and trapping. Figure 2a shows the raw CD signal of C60 samples. It is observed that the rf noise increased rapidly with rf amplitude. The S/N ratios of þ Cþ 50 and C 60 ions measured 2 and 6, respectively. The inset figure is the enlarged mass spectrum from 550 to 750 Th, where strong rf interferences and rf noise bumps were observed. Figure 2b shows that with signal processing by the OWPD algorithm, the rf noise of C60 ions could be completely þ removed and the S/N ratios of C þ 50 and C 60 ions were increased to 4 and 18, respectively. Also, ~700 ions of carbon clusters, þ þ þ Cþ 52 , C 54 , C 56 , and C 58 ions, were detected as shown in the inset of Figure 2b. This indicated the charge detector was able to detect a few hundred ions. Figure 2c shows the same mass spectrum acquired simultaneously by SED with the same experimental settings. The mass resolution (m/Δm, FWHM) of C60 ions was calculated as 500 with CD and 250 with SED with a scan rate of 3020 Da/s. The S/N ratio of SED signal was ~2 to 3 times higher than that of CD signal, but the detected signals with CD can reflect the real number of ions ejected during mass analysis. In addition, in Figure 2, the response time of CD signal was observed faster than that of SED, which we thought improved the mass resolution of RIT-MS. The mass resolution of a mass spectrometer depends on scan rate, high order field of ion trap, gas pressure, and detector performance [18]. In this study, the ion trap settings and scan function in Figure 2b and c were the same, so we examined the differences in response time of CD and SED. We found the leading edge of both detectors’ signals was almost the same but the falling edge was different. The
1073
falling edge of CD signal was determined by the RC time constant, which provided good response time to detect protein ion signals. The falling edge of SED signal was correlated to the input impedance resistor of 1 MΩ and was longer than that of CD. C.-H. Chen et al. [25] found that if the detector used a high impedance resistor of 1 MΩ (a long collection time), a smooth spectrum would result. In our experiment, we adopted 1 MΩ resistor as the input impedance resistor to acquire the mass spectra from SED. Therefore, the resolution of SED was poorer than that of CD, which was because the response time of CD and of SED was electric circuitry-dependent. The poor resolution of SED could be from the interferences of secondary ions and secondary electrons. Upon the impact of heavy ions against a dynode surface, small secondary ions (typically <100 Da) were sputtered from the dynode surface and were timedispersed in the post-acceleration region, which caused a significant signal broadening and corresponding loss of mass resolution [34]. However, in CD it was the primary ions that were detected; therefore, there was no interference from secondary ions with concomitant signal broadening and loss of mass resolution.
Detection of Protein Ions with m/z ~10 kTh Ion trapping conditions of protein ions are different from those of C60 ions. To guide singly charged cytochrome c ions, the rf frequency and amplitude of Q1 were set at 220 kHz and 420 Vp-p (qu ~ 0.46) and the pressure at Q1 was maintained to 10 mTorr with helium buffer gas to quench the K.E. of cytochrome c ions before they entered RIT (pressure ~1 mTorr). In order to extend the mass range, the trapping frequency had to be set at 140 kHz and ions were injected into RIT mass analyzer with dynamic rf trapping setting from 550 Vp-p (qu ~0.46) to 700 Vp-p (qu ~ 0.59). In RIT the trapping pressure was about 1 mTorr and the rf amplitude was set ~ four times higher than the rf amplitude used for the C60 ions. The high rf amplitude setting led to three times higher noise interference in detected ion signals with CD. The mass spectrum of cytochrome c ions was acquired by scanning the rf voltage from 700 to 3200 Vp-p with a rf frequency of 140 kHz, yielding a low-mass cutoff of ~8 kTh. Figure 3 shows the mass spectra of cyt c ions with and without denoising. Figure 3a is the mass spectrum of singly charged cytochrome c monomer and dimer ions obtained by CD without any signal processing. It shows a pickup of three times higher rf noise by CD compared with that of C60 ions (see Figure 2a). The S/N ratios of singly charged cytochrome c monomer and dimer ions were 7 and 1, respectively. Figure 3b shows the rf noise of cytochrome c ions could be completely removed after signal processing by the OWPD algorithm, and the S/N ratios of singly charged cytochrome c monomer and dimer ions were greatly increased to 80 and 6. The ion number of cytochrome c monomer and of dimer measured ~36,000 and ~2600, respectively. The results shown in Figure 2 (C60 ions) and Figure 3 (cytochrome c proteins) demonstrated that the rf noise interference could be completely removed without any signal distortion by the OWPD-based
A. A. Patil et al.: Charge Detector with Rectilinear Ion Trap Mass Spectrometer
algorithm, and CD could be successfully coupled with RIT mass analyzer to detect high mass protein ions. In order to further understand the relation of mass resolution and scan rate of a RIT-MS, mass spectra of cytochrome c ions were acquired by CD and SED, both with two scan rate settings. The mass resolution (m/Δm, FWHM) of cytochrome c monomer ions was ~27 with CD and ~18 with SED at a scan rate of 67,700 Da/s (Supplementary Figure S5a). The mass resolution of cytochrome c monomer ions measured ~62 with CD and ~30 with SED at a slow scan rate of 30,400 Da/s as shown in Supplementary Figure S5b. The mass resolution of RIT-MS could be improved with a slow scan rate, but with a low scan rate, ions would be lost as well. In Figure 3c, we also found the S/N ratio of cytochrome c monomer ions with SED was 60, but SED yielded poor dimer signal (S/N ratio ~2), which was due to the different potential well depth (Du) of the monomer (Du = 51.55 eV) and dimer (Du = 19.94 eV) species. The potential well depth determined the maximum ejection energy of the ions. In other words, the deeper the potential well depth was, the more ions could gain kinetic energy during ion ejection [18]. As a result, monomer ions gained more K.E. upon ejection compared with dimer ions. The ion velocities of ejected dimer ions 2.5 times slower than those of the monomer ions according to their potential well depth are as calculated in Table 1, so poor signals were obtained with SED for dimer ions even when the dynode voltage was set at −8 kV. Moreover, we noticed the ion generation ratio of cytochrome c monomer relative to dimer ions was about 30:1 with SED (Figure 3c), which was much less than the ratio in a MALDI-TOF MS (the ratio was ~5:1 and the MALDI-TOF mass spectrum, shown in Supplementary Figure S6). This indicated that shallow potential well depth resulted in fewer dimer ions in the trap and less K.E. during ion ejection and, thus, dimer ion signals could hardly be detected by SED. By contrast, the dimer ions could be clearly observed with CD as shown in Figure 3b. The number of dimer ions was less than that of monomer ions with a ratio of 13:1, and this ratio was better than that with SED (Figure 3c). In the C60 mass spectra of Figure 2b and c, S/N ratio of SED was two times higher than that of CD. We therefore could conclude that when m/z was below 1 kTh, SED could provide better S/N ratios than that of CD; however, the S/N ratio decreased gradually when the molecular mass increased, which is consistent with other studies [26, 28, 29].
Mass Analysis of Intact Proteins with m/z from 60 to 150 kTh It was observed that SED could not detect high mass proteins such as BSA and IgG with the dynode voltage applied over −10 kV, which was because the high dynode voltage setting often caused discharging problems at high trapping pressure conditions, typically greater than 1 mTorr. CD, however, could work well with pressure up to several hundred mTorr (or higher) and had no discharging problems [32, 41]. To detect high mass ions, optimization of experimental parameters such as reduction of trapping frequency, increase of bath gas pressure in the
Q1 region, and negative floating DC potential at RIT are crucial. To guide BSA ions, the guiding quadrupole Q1 was operated at rf frequency of 120 kHz and amplitude of 420 Vp-p. This setting was equivalent to qu ~ 0.29 and a low-mass cutoff at ~21 kTh. By setting this low-mass cutoff on the quadrupole ion guide, matrix clusters and collision-induced dissociation product ions (<21 kTh) could be filtered out. To trap BSA ions, the rf of RIT was reduced to 85 kHz and ions were injected into RIT with dynamic rf trapping from 450 Vp-p (qu ~ 0.19) to 550 Vp-p (qu ~ 0.235). BSA ions are about five times heavier than cytochrome c ions, so the K.E. of BSA was almost five times higher than that of cytochrome c ions with an assumption of similar initial ion velocity during MALDI ion generation [58]. The 10 mTorr cooling gas at guiding quadrupole (Q1) used for cytochrome c ions was found insufficient to cool down the K.E. of BSA ions efficiently before they entered RIT. To effectively reduce the K.E. of BSA ions, 40 mTorr helium gas was introduced at Q1 and 1.2 mTorr pressure was maintained at RIT. Figure 4 shows the mass spectrum of BSA ions obtained by scanning rf voltage from 550 to 3600 Vp-p at a fixed rf of 85 kHz. The number of singly charged BSA ions was measured as ~18,600 ions and ~4000 doubly charged ions were observed. The high collision gas setting at Q1 efficiently cooled down the K.E. of singly charged BSA ions. We also observed that some ions underwent collision-induced dissociation and the fragmentation features appeared at m/z from 30 to 50 kTh. The doubly charged ions were guided at higher qu = 0.579 (a deeper potential well depth of Du ~ 30.44); thus, those ions might undergo more energetic collisions and gain enough internal energy to dissociate at higher gas pressure conditions (40
25k [M+H]
+
20k
Number of Ions
1074
m/ m ~ 120
15k 10k [M+2H]
+
5k 0 -5k 20k
40k
60k
80k
100k
m/z (Th) Figure 4. Mass spectrum of singly and doubly charged BSA monomer detected by a CD RIT-MS with scanning amplitude from 550 to 3600 Vp-p at 180 kHz; 40 mTorr He buffer gas is leaked to Q1 and 1.2 mTorr helium gas is introduced into RIT. The mass resolution of singly charged BSA monomer is ~120
A. A. Patil et al.: Charge Detector with Rectilinear Ion Trap Mass Spectrometer
mTorr) in Q1. The mass resolution was noted to be improved to ~120 (m/Δm, FWHM) for singly charged BSA ions with a scan rate of 900,000 Da/s. Detection of BSA ions with LIT-MS were also conducted by Hager et al. [12]. They observed high abundance +2 (m/z ~33,000 Th) and +3 charge states ions (m/z ~22,000 Th) in the acquired mass spectra, but singly charged BSA ions (m/z ~66,000 Th) were hardly detected by SED with a voltage-scan LIT-MS [12]. Moreover, Koizumi et al. designed a larger diameter LIT operated in mass selective axial ejection mode (MSAE) [21] to detect singly charged BSA ion signals but mass spectra were resolved poorly in their setup. Compared with the results of these two studies, CD RIT-MS could effectively extend the detection limit up to BSA ions (m/z ~66,000 Th) with good mass resolution in radial ejection mode. Figure 5a shows the IgG mass spectrum obtained by a CD RIT-MS when 40 mTorr helium gas was introduced at Q1 and 1.5 mTorr helium gas was maintained at RIT. This buffer gas condition was similar to the condition used in the detection of BSA ions. But the rf frequency must be tuned lower to 80 kHz in the guiding quadrupole Q1 region and the rf amplitude was set to 420 Vp-p (qu ~ 0.28 and Du ~ 15.07). The trapping rf frequency of RIT was further reduced to 55 kHz and ions were injected with dynamic rf trapping condition from 450 Vp-p (qu ~ 0.20) to 550 Vp-p (qu ~ 0.247). To acquire the mass spectrum of IgG, the amplitude was scanned from 550 to 3800 Vp-p with a voltage scan rate of 32,500 Vp-p/s. There were ~9400 doubly charged and ~4200 singly charged IgG ions detected by CD. The number of doubly charged ions was higher than singly charged ions because the buffer gas pressure and trapping conditions were more appropriate for doubly charged ions (Du = 22.75 eV) than for singly charged ions (Du = 11.37 eV). Two possibilities could explain this phenomenon. First, the low trapping frequency (55 kHz) caused few cycles of rf field (see the calculated secular frequencies in Table 1) and therefore ions were lost during ion transportation and ion trapping. Second, the helium buffer gas was not effective when cooling IgG ions in Q1 because the K.E. of IgG ions was ~2.2 times higher than that of BSA ions. In order to keep enough potential well depth (Du) (see Table 1) for ion trapping, the setting of rf frequency and amplitude cannot be changed. Therefore, introducing heavy collision gas (argon, neon, krypton) should be a feasible option to cool down the K.E. of IgG ions, and those heavier gases could offer efficient cooling via collision [64, 65]. We also found that the setting of negative floating potential on RIT was important to increase the potential well depth. With negative potential, the ion transfer and trapping efficiency were improved [66]. The setting of negative potential was observed to be important for IgG ion trapping, but not for BSA ion trapping. Figure 5b shows a mass spectrum of singly charged IgG ions obtained by introducing 30 mTorr argon gas in Q1 and with 1.5
1075
Figure 5. Mass spectra of singly and doubly charged IgG monomer detected by a CD RIT-MS. The rf voltage is scanned from 550 to 3800 Vp-p at a rf frequency of 55 kHz and 1.5 mTorr He gas is introduced into RIT. (a) Mass spectrum obtained with 40 mTorr helium buffer gas and the gas is leaked at Q1. (b) Mass spectrum obtained at −12 V DC offset on RIT with a scan rate of 5,869,275 Da/s and 30 mTorr Ar gas is leaked at Q1. (c) Mass spectrum obtained at −12 V dc offset on RIT with a scan rate of 4,695,420 Da/s and 30 mTorr argon gas is leaked at Q1
A. A. Patil et al.: Charge Detector with Rectilinear Ion Trap Mass Spectrometer
mTorr helium pressure maintained at RIT region; −12 V DC floating potential was set on RIT. We found ~4800 doubly charged ions and much improved ~12480 singly charged monomer ions. The mass resolution (m/Δm, FWHM) of singly charged IgG ions was estimated to be ~60 with a scan rate of 5,869,275 Da/s. If we further reduced the scan rate to 4,695,420 Da/s and maintained the same experimental conditions as mentioned above (see Figure 5c), mass resolution (m/Δm, FWHM) of singly charged IgG ions could be improved to ~90, which was comparable to the mass resolution with a nano-membrane TOF-MS analyzer [67, 68]. Chen et al. also coupled LIT with conventional SED, which was operated at high dynode voltages (−25 kV) and used differential pumping design to measure BSA and IgG ions [24]. The mass resolution achieved in their setup was under 30 (m/Δm, FWHM), which probably was because of high buffer gas pressure of the ion trap and long response time of a SED detector. Our CD RIT-MS did not have such problems and thus had better mass resolution. Also, the differential pumping design used in their setup was too complicated to be used in a commercial linear ion trap mass spectrometer, whereas CD can be implemented with LIT-MS without complicated instrumental design. So far, the mass resolution of 90 (m/Δm, FWHM) of IgG ions obtained by the voltage-scan CD RIT-MS in radial ejection mode is the best compared with mass resolution acquired by other linear ion trap mass spectrometers.
Comparison of the Detection Efficiency of CD and SED Approximately equal amounts (50%:50%) of ejected ions were detected by CD and SED, and then we plotted signal intensity ratios of mass spectra in units of mVolt to calculate the detection efficiency. Figure 6 shows the comparison of the detection efficiency of SED and that of CD. Typical mass spectra used to calculate the detection efficiency in Figure 6 are shown in Supplementary Figure S7. The signal intensity ratios of SED and CD (ISED/ICD) were plotted against m/z of low mass C60, C60 clusters, and high mass proteins. The left inset in Figure 6 is the plot of ISED/ ICD against m/z of C60 and C60 clusters obtained at −4 kV dynode voltage. The right inset in Figure 6 is the plot of ISED/ICD against m/z of singly charged ubiquitin and cytochrome c monomer and dimer ions obtained at −8 kV dynode voltage. It was noted the ISED/ICD ratio of low mass ions (m/z <1 kTh) was over one, implying that the signals of SED had higher intensities than that of CD and more secondary electrons were generated at dynode surface and amplified by electron multiplier. The slope of ISED/ICD was −3.37 × 10−3 (R2 = 0.97) when m/z <1 kTh. However, the slope of ISED/ICD (m/z >1 kTh) decreased to −5.57 × 10−5 (R2 = 0.86), which was about two orders of magnitude less than the slope in the
Protein ions -8 kV
C60 and C60 Clusters -4 kV 3 2
3
1
2
0 1 600
2
ISED/ICD
1076
700
800
10k 15k 20k 25k
-3
1
y=-3.37 10 x+4.04
-5
y=-5.57 10 x+1.49
2
r =0.97
2
r =0.866
0 550 600 650 700
10k 15k 20k 25k
m/z (Th) Figure 6. Comparison of the detection efficiency of a secondary electron detector (SED) and a charge detector (CD) as function of ion mass. Signal intensity ratios of SED and CD (ISED/ICD) are plotted with respect to m/z. First inset figure is the plot of ISED/ICD of C60 and C60 clusters collected at a fixed dynode voltage of −4 kV. Second inset figure is the plot of ISED/ ICD of monomer and dimer of ubiquitin and cytochrome c ions obtained at a fixed dynode voltage of −8 kV
low mass case. We also noted the ISED/ICD ratio of high mass protein ions (m/z >8 kTh) was close to one or less than one. This indicated that the signal intensities of SED decreased as m/z values increased. In other words, SED was unable to generate enough secondary electrons when the dynode voltage was set at −8 kV, and this might be due to poor secondary yield of SED. The reason for the poor detection efficiency was that the ion ejection velocity decreased rapidly when ion mass increased and the applied dynode voltage was unable to accelerate those ions to gain sufficient impact energy and generate secondary electrons efficiently. Besides, the slope of ISED/ICD of protein ions was about two orders of magnitude lower than that of low mass C60 ions, which indicated that the detection efficiency of heavier ions decreased rapidly with increasing mass with SED. Figure 6 shows the polynomial fitting of intensity ratio (ISED/ICD) of protein ions with R2 = 0.98 (blue dash line), which indicates that the major contribution came from the first order (−2.416 × 10−4; x) but not from the second order (5.327 × 10−9; x2). The above analysis confirmed signal intensity and m/z ratio was linearly correlated. Earlier studies also found that the efficiency of secondary electron generation with a microchannel plate (MCP) detector in MALDI-TOF mass analyzers decreased as m/z values increased [26–29, 67–69].
A. A. Patil et al.: Charge Detector with Rectilinear Ion Trap Mass Spectrometer
For example, Smith et al. [29] found the MCP detection efficiencies calibrated by an inductive charge detector (ICD) were very close to unity for smaller ions (e.g., angiotensin, 1046.5 Da) at high acceleration voltage (25 kV), but decreased to ~11% for the largest ions (e.g., IgG dimer, 290 kDa) even when the high acceleration voltage (25 kV) was applied. McLuckey et al. utilized Q-TOF MS to extend the dynamic range of quadrupole ion guide up to 66 kDa and performed topdown protein identification and characterization of a priori unknown proteins [70]. However, this might increase the cost of instrument complexity and size. CD RIT-MS could detect high mass ions and perform MS/MS analysis and, therefore, extends the dynamic range of linear ion trap mass spectrometers. CD RIT-MS might be used in macromolecular ion/ion proton transfer chemistry applications where output products of reaction are singly charged ions that are unable to be detected with the limited dynamic range of commercial linear ion trap mass spectrometers [9].
2. 3.
4. 5.
6.
7.
8.
9.
10. 11.
Conclusions CD RIT-MS improves the detection sensitivity for low charge state high mass MALDI ions. With the OWPD method, we successfully analyzed and removed all rf noise interferences induced from RIT and found the noise of CD could be reduced to ~450 ions, and the S/N ratio was improved by a factor of six to eight when cytochrome c ions were analyzed. Optimization of trapping parameters such as qu value, collision cooling gas, bath gas pressure, and scan rate is imperative to achieve well resolved mass spectra of BSA and IgG ions. We found that a negative floating DC potential at RIT mass analyzer was essential to improve trapping and transmission of IgG ions. The analysis of intensity ratios with SED and CD showed the detection efficiency of SED decreased linearly as ion mass increased. The detection efficiency and mass resolution of RIT-MS with CD was better than that with SED, especially for m/z over 10 kTh. The mass range of CD RIT-MS could be extended to singly charged IgG ions (m/z ~150,000 Th) with mass resolution of 90 (m/Δm, FWHM).
Acknowledgment The authors gratefully acknowledge Professor R. Graham Cooks for the supply of RIT setup and his helpful suggestions. This work is supported by grants from the Ministry of Science and Technology of Taiwan (grants MOST 102-2112-M-259003-MY3 and MOST 105-2112-M-259 -002 -MY3) (W.P.P.).
12.
13.
14.
15.
16.
17.
18. 19. 20.
21.
22.
23.
24.
References 1.
Schwartz, J.C., Senko, M.W., Syka, J.E.P.: A two-dimensional quadrupole ion trap mass spectrometer. J. Am. Soc. Mass Spectrom. 13, 659– 669 (2002)
25. 26.
1077
Hager, J.W.: A new linear ion trap mass spectrometer. Rapid Commun. Mass Spectrom. 16, 512–526 (2002) Hager, J.W., Yves Le Blanc, J.C.: Product ion scanning using a Q-qQlinear ion trap (Q TRAP) mass spectrometer. Rapid Commun. Mass Spectrom. 17, 1056–1064 (2003) Londry, F.A., Hager, J.W.: Mass selective axial ion ejection from a linear quadrupole ion trap. J. Am. Soc. Mass Spectrom. 14, 1130–1147 (2003) Peterman, S.M., Dufresne, C.P., Horning, S.: The use of a hybrid linear trap/FT-ICR mass spectrometer for on-line high resolution/high mass accuracy bottom-up sequencing. J. Biomol. Tech. 16, 112–124 (2005) Sampson, J.S., Murray, K.K., Muddiman, D.C.: Intact and top-down characterization of biomolecules and direct analysis using infrared matrix-assisted laser desorption electrospray ionization coupled to FTICR mass spectrometry. J. Am. Soc. Mass Spectrom. 20, 667–673 (2009) Hu, Q.Z., Noll, R.J., Li, H.Y., Makarov, A., Hardman, M., Cooks, R.G.: The Orbitrap: a new mass spectrometer. J. Mass Spectrom. 40, 430–443 (2005) Makarov, A., Denisov, E., Kholomeev, A., Balschun, W., Lange, O., Strupat, K., Horning, S.: Performance evaluation of a hybrid linear ion trap/orbitrap mass spectrometer. Anal. Chem. 78, 2113–2120 (2006) Xia, Y., McLuckey, S.A.: Evolution of instrumentation for the study of gas-phase ion/ion chemistry via mass spectrometry. J. Am. Soc. Mass Spectrom. 19, 173–189 (2011) Garcia, B.A.: What does the future hold for top down mass spectrometry? J. Am. Soc. Mass Spectrom. 21, 193–202 (2010) Wu, J., Hager, J.W., Xia, Y., Londry, F.A., McLuckey, S.A.: Positive ion transmission mode ion/ion reactions in a hybrid linear ion trap. Anal. Chem. 76, 5006–5015 (2004) Xia, Y., Wu, J., McLuckey, S.A., Londry, F.A., Hager, J.W.: Mutual storage mode ion/ion reactions in a hybrid linear ion trap. J. Am. Soc. Mass Spectrom. 16, 71–81 (2005) Tseng, Y.-H., Uetrecht, C., Heck, A.J.R., Peng, W.-P.: Interpreting the charge state assignment in electrospray mass spectra of bioparticles. Anal. Chem. 83, 1960–1968 (2011) Tseng, Y.-H., Uetrecht, C., Yang, S.-C., Barendregt, A., Heck, A.J.R., Peng, W.-P.: Game theory-based search engine to automate the mass assignment in complex native electrospray mass spectra. Anal. Chem. 85, 11275–11283 (2013) Karas, M., Bahr, U., Hillenkamp, F.: UV laser matrix desorption/ ionization mass spectrometry of proteins in the 100,000 dalton range. Int. J. Mass Spectrom. Ion Processes 92, 231–242 (1989) Smith, S.A., Blake, T.A., Ifa, D.R., Cooks, R.G., Ouyang, Z.: Dualsource mass spectrometer with MALDI-LIT-ESI configuration. J. Proteome Res. 6, 837–845 (2007) Peng, W.-P., Chou, S.-W., Patil, A.A.: Measuring masses of large biomolecules and bioparticles using mass spectrometric techniques. Analyst 139, 3507–3523 (2014) March, R.E., Todd, J.F.: Quadrupole ion trap mass spectrometry. John Wiley and Sons, Inc., Hoboken (2005) March, R.E.: An introduction to quadrupole ion trap mass spectrometry. J. Mass Spectrom. 32, 351–369 (1997) Xia, Y., Wu, J., McLuckey, S.A., Londry, F.A., Hager, J.W.: Mutual storage mode ion/ion reactions in a hybrid linear ion trap. J. Am. Soc. Mass Spectrom. 16, 71–81 (2005) Koizumi, H., Whitten, W.B., Reilly, P.T.A.: Trapping of intact, singlycharged, bovine serum albumin ions injected from the atmosphere with a 10-cm diameter, frequency-adjusted linear quadrupole ion trap. J. Am. Soc. Mass Spectrom. 19, 1942–1947 (2008) Song, Q., Xu, W., Smith, S.A., Gao, L., Chappell, W.J., Cooks, R.G., Ouyang, Z.: Ion trap mass analysis at high pressure: an experimental characterization. J. Mass Spectrom. 45, 26–34 (2010) Xu, W., Song, Q., Smith, S.A., Chappell, W.J., Ouyang, Z.: Ion trap mass analysis at high pressure: a theoretical view. J. Am. Soc. Mass Spectrom. 20, 2144–2153 (2009) Lu, I.C., Lin, J.L., Lai, S.-H., Chen, C.-H.: Frequency-scanning MALDI linear ion trap mass spectrometer for large biomolecular ion detection. Anal. Chem. 83, 8273–8277 (2011) Chen, C.H., Chen, C.H., Lin, J.L., Chu, M.L.: Mass spectrometer and methods for detecting large biomolecules. US Patent 8,258,464 (2012) Li, M.H., Tsai, S.T., Chen, C.H., Chen, C.W., Lee, Y.T., Wang, Y.S.: Bipolar ion detector based on sequential conversion reactions. Anal. Chem. 79, 1277–1282 (2007)
1078
27. 28. 29.
30. 31.
32. 33.
34.
35. 36. 37. 38.
39. 40. 41.
42. 43. 44. 45. 46. 47. 48. 49.
A. A. Patil et al.: Charge Detector with Rectilinear Ion Trap Mass Spectrometer
Wenzel, R.J., Matter, U., Schultheis, L., Zenobi, R.: Analysis of megadalton ions using cryodetection MALDI time-of-flight mass spectrometry. Anal. Chem. 77, 4329–4337 (2005) Chen, X., Westphall, M.S., Smith, L.M.: Mass spectrometric analysis of DNA mixtures: instrumental effects responsible for decreased sensitivity with increasing mass. Anal. Chem. 75, 5944–5952 (2003) Liu, R., Li, Q., Smith, L.M.: Detection of large ions in time-of-flight mass spectrometry: effects of ion mass and acceleration voltage on microchannel plate detector response. J. Am. Soc. Mass Spectrom. 25, 1374–1383 (2014) Peng, W.P., Lin, H.C., Chu, M.L., Chang, H.C., Lin, H.H., Yu, A.L., Chen, C.H.: Charge monitoring cell mass spectrometry. Anal. Chem. 80, 2524–2530 (2008) Peng, W.P., Lin, H.C., Lin, H.H., Chu, M., Yu, A.L., Chang, H.C., Chen, C.H.: Charge-monitoring laser-induced acoustic desorption mass spectrometry for cell and microparticle mass distribution measurement. Angew. Chem. Int. Ed. 46, 3865–3869 (2007) Chen, C.H., Lin, J.L., Chu, M.L., Chen, C.H.: MALDI ion trap mass spectrometer with charge detector for large biomolecule detection. Anal. Chem. 82, 10125–10128 (2010) Alexander, J.D., Graham, L., Calvert, C.R., Kelly, O., King, R.B., Williams, I.D., Greenwood, J.B.: Determination of absolute ion yields from a MALDI source through calibration of an image-charge detector. Meas. Sci. Technol. 21, 045802 (2010) Bahr, U., Röhling, U., Lautz, C., Strupat, K., Schürenberg, M., Hillenkamp, F.: A charge detector for time-of-flight mass analysis of high mass ions produced by matrix-assisted laser desorption/ionization (MALDI). Int. J. Mass Spectrom. Ion Processes 153, 9–21 (1996) Park, M.A., Callahan, J.H., Vertes, A.: An inductive detector for time-of-flight mass spectrometry. Rapid Commun. Mass Spectrom. 8, 317–322 (1994) Peng, W.-P., Goodwin, M.P., Nie, Z., Volný, M., Ouyang, Z., Cooks, R.G.: Ion soft landing using a rectilinear ion trap mass spectrometer. Anal. Chem. 80, 6640–6649 (2008) Ouyang, Z., Wu, G., Song, Y., Li, H., Plass, W.R., Cooks, R.G.: Rectilinear ion trap: concepts, calculations, and analytical performance of a new mass analyzer. Anal. Chem. 76, 4595–4605 (2004) Song, Q., Kothari, S., Senko, M.A., Schwartz, J.C., Amy, J.W., Stafford, G.C., Cooks, R.G., Ouyang, Z.: Rectilinear ion trap mass spectrometer with atmospheric pressure interface and electrospray ionization source. Anal. Chem. 78, 718–725 (2005) Patil, A.A., Jiang, S.-C., Yen, K.-C., Chou, S.-W., Cheng, C.-Y., Peng, W.-P.: Ion velocities of laser desorbed ions passing through quadrupole electric fields. Int. J. Mass Spectrom. 401, 46–54 (2016) Lin, H.C., Lin, H.H., Kao, C.Y., Yu, A.L., Peng, W.P., Chen, C.H.: Quantitative measurement of nano-/microparticle endocytosis by cell mass spectrometry. Angew. Chem. Int. Ed. 49, 3460–3464 (2010) Chou, S.-W., Shiu, G.-R., Chang, H.-C., Peng, W.-P.: Wavelet-based method for time-domain noise analysis and reduction in a frequencyscan ion trap mass spectrometer. J. Am. Soc. Mass Spectrom. 23, 1855–1864 (2012) Guna, M., Biesenthal, T.A.: Performance enhancements of mass selective axial ejection from a linear ion trap. J. Am. Soc. Mass Spectrom. 20, 1132–1140 (2009) Dai, Y., Whittal, R.M., Li, L.: Two-layer sample preparation: a method for MALDI-MS analysis of complex peptide and protein mixtures. Anal. Chem. 71, 1087–1091 (1999) Cheng, D.K.: Field and wave electromagnetics. Addison-Wesley, Boston (1989) Mallat, S.: A wavelet tour of signal processing, 3rd ed.; The sparse way. Academic Press, Cambridge (2008) Xu, L.J., Yan, Y.: Wavelet-based removal of sinusoidal interference from a signal. Meas. Sci. Technol. 15, 1779–1786 (2004) Zhang, S., Li, D., Sun, J.: Power system frequency tracking using an adaptive lattice notch filter. Meas. Technol. Intel. Instrum VI 295/296, 673–678 (2005) Delmas, J.P.: Adaptive harmonic jammer canceler. IEEE Trans. Signal Processes 43, 2323–2331 (1995) Fainzilberg, L.S., Glushauskene, G.A.: Narrow-band rejection filter for suppression of harmonic concentrated interference on the basis of discrete Fourier transform. J. Automat. Inform. Sci. 41, 55–70 (2009)
50.
Xu, L.J.: Cancellation of harmonic interference by baseline shifting of wavelet packet decomposition coefficients. IEEE Trans. Signal. Processes 53, 222–230 (2005) 51. Barclay, V.J., Bonner, R.F., Hamilton, I.P.: Application of wavelet transforms to experimental spectra: Smoothing, denoising, and data set compression. Anal. Chem. 69, 78–90 (1997) 52. Li, X.L., Li, J., Yao, X.: A wavelet-based data pre-processing analysis approach in mass spectrometry. Comput. Biol. Med. 37, 509–516 (2007) 53. Stafford Jr., G.C., Kelley, P.E., Syka, J.E.P., Reynolds, W.E., Todd, J.F.J.: Recent improvements in and analytical applications of advanced ion trap technology. Int. J. Mass Spectrom. Ion Processes 60, 85–98 (1984) 54. Louris, J.N., Amy, J.W., Ridley, T.Y., Cooks, R.G.: Injection of ions into a quadrupole ion trap mass spectrometer. Int. J. Mass Spectrom. Ion Processes 88, 97–111 (1989) 55. Chernushevich, I.V., Thomson, B.A.: Collisional cooling of large ions in electrospray mass spectrometry. Anal. Chem. 76, 1754–1760 (2004) 56. Douglas, D.J., French, J.B.: Collisional focusing effects in radio frequency quadrupoles. J. Am. Soc. Mass Spectrom. 3, 398–408 (1992) 57. Glückmann, M., Karas, M.: The initial ion velocity and its dependence on matrix, analyte and preparation method in ultraviolet matrix-assisted laser desorption/ionization. J. Mass Spectrom. 34, 467–477 (1999) 58. Berkenkamp, S., Menzel, C., Hillenkamp, F., Dreisewerd, K.: Measurements of mean initial velocities of analyte and matrix ions in infrared matrix-assisted laser desorption ionization mass spectrometry. J. Am. Soc. Mass Spectrom. 13, 209–220 (2002) 59. Zenobi, R., Knochenmuss, R.: Ion formation in MALDI mass spectrometry. Mass Spectrom. Rev. 17, 337–366 (1998) 60. Pekar Second, T., Blethrow, J.D., Schwartz, J.C., Merrihew, G.E., MacCoss, M.J., Swaney, D.L., Russell, J.D., Coon, J.J., Zabrouskov, V.: Dual-pressure linear ion trap mass spectrometer improving the analysis of complex protein mixtures. Anal. Chem. 81, 7757–7765 (2009) 61. Eiden, G.C., Garrett, A.W., Cisper, M.E., Nogar, N.S., Hemberger, P.H.: An improved method for capturing laser desorbed ions in an ion trap mass spectrometer: dynamic rf trapping. Int. J. Mass Spectrom. Ion Processes 136, 119–141 (1994) 62. Doroshenko, V.M., Cotter, R.J.: Injection of externally generated ions into an increasing trapping field of a quadrupole ion trap mass spectrometer. J. Mass Spectrom. 32, 602–615 (1997) 63. Qin, J., Chait, B.T.: Matrix-assisted laser desorption ion trap mass spectrometry: efficient trapping and ejection of ions. Anal. Chem. 68, 2102–2107 (1996) 64. Sobott, F., Hernandez, H., McCammon, M.G., Tito, M.A., Robinson, C.V.: A tandem mass spectrometer for improved transmission and analysis of large macromolecular assemblies. Anal. Chem. 74, 1402– 1407 (2002) 65. van den Heuvel, R.H.H., van Duijn, E., Mazon, H., Synowsky, S.A., Lorenzen, K., Versluis, C., Brouns, S.J.J., Langridge, D., van der Oost, J., Hoyes, J., Heck, A.J.R.: Improving the performance of a quadrupole timeof-flight instrument for macromolecular mass spectrometry. Anal. Chem. 78, 7473–7483 (2006) 66. Li, L., Zhou, X., Hager, J.W., Ouyang, Z.: High efficiency tandem mass spectrometry analysis using dual linear ion traps. Analyst 139, 4779–4784 (2014) 67. Park, J., Kim, H., Blick, R.H.: Quasi-dynamic mode of nanomembranes for time-of-flight mass spectrometry of proteins. Nanoscale 4, 2543–2548 (2012) 68. Park, J., Qin, H., Scalf, M., Hilger, R.T., Westphall, M.S., Smith, L.M., Blick, R.H.: A mechanical nanomembrane detector for time-of-flight mass spectrometry. Nano Lett. 11, 3681–3684 (2011) 69. Westmacott, G., Frank, M., Labov, S.E., Benner, W.H.: Using a superconducting tunnel junction detector to measure the secondary electron emission efficiency for a microchannel plate detector bombarded by large molecular ions. Rapid Commun. Mass Spectrom. 14, 1854–1861 (2000) 70. Xia, Y., Chrisman, P.A., Erickson, D.E., Liu, J., Liang, X., Londry, F.A., Yang, M.J., McLuckey, S.A.: Implementation of Ion/Ion Reactions in a Quadrupole/Time-of-Flight Tandem Mass Spectrometer. Anal. Chem. 78, 4146–4154 (2006)