Non-Destructive Ion Trap Mass Analysis at High Pressure. Supporting information

Transcription

1 Non-Destructive Ion Trap Mass Analysis at High Pressure Supporting information Wei Xu 1, Jeffrey Maas 2, Frank Boudreau 3, William J. Chappell 2 and Zheng Ouyang 1,2,3* 1. Weldon School of Biomedical Engineering, Purdue University, West Lafayette, IN 47907, USA 2. Department of Electrical and Computer Engineering, Purdue University, West Lafayette, IN 47907, USA 3. Center for Analytical Instrumentation Development, Purdue University, West Lafayette, IN 47907, USA *Corresponding Author: Prof. Zheng Ouyang Weldon School of Biomedical Engineering Purdue University, West Lafayette, IN Ouyang@purdue.edu Phone: Fax:

2 Additional information for instrumentation design Pressure measurement The pressure variation inside the vacuum chamber was measured using a micro-pirani vacuum gauge (MKS 925C, MKS Instruments, Inc. Wilmington, MA). Unlike linear traps used in commercial instruments, which have sealed configurations, the linear trap used in this study has an open configuration (Figure 1) with 1 mm gaps between the end electrodes and the ion trap. This allows the gas to leak fast from the ion trap and a pressure equilibrium to be reached relatively fast between the inside and outside of the trap. Differential preamplifier The preamplifier used for the image current measurement was home built. It has a differential input and was designed to reject the common mode noises, such as the coupled RF noise. Two compound amplifiers, which are composed of precision operational amplifiers (OPA637 from Texas instruments, Dallas, TX) and low-distortion broadband operational amplifiers (THS3092D from Texas instruments, Dallas, TX), were used as the differential inputs of the preamplifier. The outputs of these compound amplifiers are fed into an OPA627 (Texas instruments, Dallas, TX) to reject the common mode signals. Low-pass RC filters with a cutoff frequency of 500 khz were also built in the circuits. Ion motions under dipolar excitation Under a constant dipolar excitation, the motion of an ion trapped in a quadrupole ion trap has multiple frequency components as shown in Figure S1. In addition to the motions at the secular frequency (f 0 ) and its harmonics (f RF ±f 0 ), there are motions at the excitation frequency (f ac ) and

3 its harmonics (f RF ±f ac ). Under collisions with buffer gas, the motions at the secular frequency and its harmonics (f 0, f RF ±f 0 ) decay while those at the dipolar excitation frequency and its harmonics (f ac, f RF ±f ac ) are sustained. 1 Physically, the AC excitation effect and the damping effect by the collisions with background gas are balanced after the initial stage of excitation Excitation f ac f 0 Related f AC Related Amplitude f 0 f rf -f ac Detection f rf -f 0 f rf + f 0 f rf + f ac Ion secular frequency (khz) Figure S1 Ion motion frequency components under dipolar excitation. Determination of the RF amplitude for setting the ion secular motion frequency to the target value (excitation frequency) Due to the imperfection in the electric field inside the ion trap and other experimental errors, the RF amplitude (V RF ) required for setting an ion (of an m/z value) to a target value f 0, which is the excitation frequency (f ac ) in the image current measurement, was experimentally determined. Based on the Mathieu's equation, a relationship of V RF = a(m/z) +b was assumed.

4 With two sets of (m/z, V RF ) experimentally identified, the constant parameters a and b can be obtained through linear fitting. Experimentally, the AC excitation at a frequency is applied at relatively low amplitude before a mass selective instability scan to acquire the full spectrum. If the secular frequency of the ions at a particular m/z value coincides with the excitation frequency, some or all of the ions will be excited and ejected, and the abundance of the corresponding peak would vary in the spectra acquired with mass selective instability scan. For example, two nano-esi spectra are shown in Figure S2 for a polyethylene glycol (PEG) mixture. The ion m/z 305, is missing in the spectrum in Figure S2b, for which the RF amplitude was adjusted to have the secular frequency of ion m/z 305 coincided with the excitation frequency. (a) Signal intensity (Arb Unit) (b) Signal intensity (Arb Unit) m/z m/z Figure S2 (a) A full mass spectrum of PEG. (b) Mass spectrum of PEG with the selected ion (m/z=305) ejected due to dipolar excitation. This procedure was applied with the experimental condition used for image current measurement with the dipolar excitation frequency f AC = 125 khz. Two sets of (m/z, V RF ) were

5 determined using cocaine ion (m/z 304) and its fragment ion (m/z 182) and the function for calculating RF amplitude was determined to be V RF = 0.64 (m/z) RF Amplitude (V) Excitation Frequency f AC =125 khz V RF = (m/z) m/z Figure S3 Prediction of the RF amplitudes as a function of m/z values for an excitation at 125 khz. Calculation of the q value based on ion secular frequency and RF frequency Due to the imperfection in the electric field inside a real ion trap device, the q value cannot be accurately calculated using Mathieu s equation unless the quadrupole coefficient A 2 can be accurately determined. Practically, once the ion secular frequency f 0 is known (for example, determined using the procedure described above), the parameter β can be calculated using equation f 0 =β f RF /2; 2 the q value can then be calculated by numerically solving the following equation: 2 ( a 1) q (5a + 7) q (9a + 58a + 29) q β = au 2( a 1) q 32( a 1) ( a 4) 64( a 1) ( a 4)( a 9) u u u u u u u u u u u u u u where a u = 0 for the experimental conditions used for the image current measurement.

6 Gain correction All the components in the detection circuit are designed with a constant gain over the frequency bandwidth (455±2.85 khz), except for the band-pass filter, which has ripples in the gain of the mechanical filter. A frequency spectrum was recorded for the background noise, as shown in Figure S4a, which shows the false peaks due to the gain variation. Assuming the background noise is Gaussian white noise, the transfer function of the filter (H) was obtained by (H=S out /S in ) and used for the gain correction. Figure S4b shows frequency spectrum of the background after gain correction. Figure S4 c and d show the effects of the gain correction on frequency spectrum recorded for cocaine m/z 304. (a) 6.0x10-4 Noise before gain correction (b) 6.0x10-4 Noise after gain correction Intensity (Arb unit) 4.0x x10-4 Intensity (Arb unit) 4.0x x Frequency (khz) Frequency (khz) (c) Signal before gain correction (d) Signal after gain correction Intensity (Arb unit) Intensity (Arb unit) Frequency (khz) Frequency (khz) Figure S4 Gain correction effects on Gaussian white noise (a,b) and a signal (c,d).

7 Calculation of image current amplitude The potential inside an ion trap due to the dipolar AC signal can be approximated as, Φ 2 0 where V is the amplitude of the AC applied on the electrodes and x 0 is the ion trap dimension (center to electrode) in the x direction. Based on the Green's Reciprocity theory, 2, 3 the charge induced on the same set of electrodes by an ion with charge q can be expressed as, 2 0 where x is the location of the ion inside the ion trap in x direction, and its variation under dipolar excitation has been modeled and simulated. 1 Under a constant excitation, ion motion at the AC frequency can be treated as a harmonic oscillator with a constant amplitude, so that the image current on the electrodes (I) can be calculated as: sin 2 0 where A is the ion motion amplitude at angular frequency ω. In this study, two line electrodes were used for measuring the image current, and the amount of charge induced on the line electrodes (Q') should be a fraction of the total induced charge (Q) on the x electrodes and the same applies for the image current. Therefore, the image current due to n ions moving coherently is: 2 0 sin where n is the number of ions and K is the fraction constant. Based on the theoretical modeling, 1 for an ion with a maximum motion amplitude of x 0 /(1+q z /2) at secular frequency of f 0 = 125 khz, the image current (I') induced at (f rf -f ac ) is 48.5 n K A. The image current measured for

8 the spectrum shown in Figure 2a is about A. If K = 0.1 is assumed based on the area ratio between the line and x electrode, it is estimated that n = ions contributed to the signal measured. In practice, ion motion amplitude was much smaller than x 0 /(1+q z /2) at high pressure to avoid CID; hence, more ions could have been trapped in the ion trap to produce the signal shown in Figure 2a. Peak broadening effects The theoretical study predicts a peak broadening at higher pressures; however, this was not observed in the experiments when the background pressure changed from 1-2 mtorr to mtorr during mass analysis using image current measurements (Figure 2c). Presumably, the pressure effect is suppressed by other effects on peak broadening. Most likely, the dominant effect is space charge, which is supported by both calculation and experimental results. As discussed in the previous section, it is estimated that more than ions were trapped for obtaining the image current signal shown in Figure 2a. A mass spectrum was recorded for the cocaine ion (m/z 30) at the same experimental conditions as with mass selective instability scan using an electron multiplier (Figure S5a), which shows a broad peak, also indicating a large number of ions trapped inside the ion trap. The peak widths are calculated for the cocaine ion (m/z 304) at different pressures (Figure S5b), without considering space charge effect, 1 and a FWHM = 2.5 khz should be observed with non-destructive mass analysis at 50 mtorr with minimum space charge effect. However, the peak width at 1 mtorr is already 3.5 khz (Figure 2c), which indicates a dominant effect due to space charge.

9 (a) Intensity (arb unit) 12.0k 9.0k 6.0k 3.0k m/z (b) Th 2.5 khz mtorr 10 mtorr 50 mtorr 100 mtorr x 10 5 Frequency (Hz) Figure S5 (a) Mass spectrum of the cocaine ions recorded using mass selective instability scan with an electron multiplier. (b) Calculated peak widths for cocaine ion (m/z 304), at the same working conditions as shown in Figure 2c, such as the RF frequency and voltage, but without considering space charge effect. 1 Ion motion frequency profile mapping Based on the theoretical study, 1 ion secular motion frequency distribution is dependent on the pressure. The amplitude of the ion motion at (f RF - f ac ) is a function of the ion secular motion amplitude. To map the frequency profile for the ion motion response under a dipolar excitation, the frequency of excitation can be scanned while the response at (f RF - f ac ) is measured, as shown in Figure S6.

10 Ion secular motion frequency profile f 0 Ion Motion Measured at f RF -f ac Excitation f ac Scan f ac Frequency Frequency Figure S6, Method for mapping the ion motion profile. Scan function for tandem mass spectrometry Ion introduction Isolation CID Image current Detection MS Selective Instability Scan DAPI RF AC 200ms Figure S7 Scan function of tandem mass spectrometry References (1) Xu, W.; Chappell, W.; Ouyang, Z. submitted for publication. (2) Dunbar, R. C. International Journal of Mass Spectrometry and Ion Processes 1984, 56, 1-9. Deleted: to JASMS 2010

11 (3) Goeringer, D. E.; Crutcher, R. I.; McLuckey, S. A. Analytical Chemistry 1995, 67,