Supplemental Information

Optimizing Electrospray Interfaces Using Slowly Diverging Conical Duct (ConDuct) Electrodes

Andrew N. Krutchinsky, Júlio C. Padovan, Herbert Cohen, Brian T. Chait

Laboratory for Mass Spectrometry and Gaseous Ion Chemistry, The Rockefeller University, New York, NY 10065, USA

Experimental

Metal ConDuct electrodes.A series of 25.4-mm long ConDuct electrodes with channel divergence of 0º, 1º, 2º, 3º, 5º, 8º, 13º, and 21º, respectively were produced by Midwest Precision Tool & Die Inc. from grade 440 stainless steel using electrical discharge machining (EDM) [26]. One of these ConDuct electrodes (with a channel divergence 2º) was cut in half to expose its cross-section (Figure 1A).

At first, we had five ConDuct electrodes custom-manufactured with channel divergence angles of 0º, 1º, 2º, 3º, and 5º. After initial experiments, we decided to have several additional ConDuct electrodes manufactured with channel divergence angles of 8, 13 and 21° to further our assessment of the emerging patterns that appeared to arise from increasing angles. Together with the previous batch, we created a set of ConDuct electrodes with angles that follow the Fibonacci series sequence [30].

Solution preparation details for beam diameter measurements

For measurements of the diameter of the beam, we used a solution containing the fluorescent dye SYPRO Ruby (Molecular Probes, protein gel stain) as well as a solution of Nanogold® particles (Nanoprobes, 5nm average size, cat. 2010-30nmol). The electrospray solution of the dye was prepared by acidifying 1000 l of the SYPRO Ruby dye proprietary solution with 700 l of 0.1% aqueous trifluoroacetic acid (TFA, Thermo) in the presence of Poros 20 R2 beads (Applied Biosystems) added in 20 µl of a suspension in methanol (1/1 v/v). After the vial was rotated for 10-15 min, the Poros R2 beads were collected by centrifugation at 14,000 rpm and the supernatant discarded. The beads were washed several times with 1 ml of 0.1% aqueous TFA. After this thorough washing, the dye was eluted from the Poros R2 beads with 100 μl of 100% pure methanol. The eluent was further diluted by 60% MeOH, 1% acetic acid in water to give a total of 200 μl of the SYPRO Ruby solution, which was used in the electrospray experiments. The solution of gold nano-clusters was prepared by resuspending 30 nmol of non-functionalized gold particles in 10 ml of 60% MeOH, 1% acetic acid in water.

The 2-degreeConDuct atmosphere-to-vacuum interface. Our new ConDuct interface was attached to the right arm of the T-quadrupole as shown in Figure 5. It consisted of the two-degreeConDuct electrode inserted halfway into a hollow stainless 47mm tube with dimensions 6.35 OD × 3 mm ID, and vacuum-sealed using Kalrezperfluoroelastomer O-rings (DuPont). The temperature of the assembly could be raised to 350ºC by a heater encompassing the tube in vacuum. The ion beam formed by the ConDuct electrode and heatable assembly was directed into a 15-cm long quadrupole ion guide qConDuct made of 6.35-mm diameter rods, operating in RF-only mode and driven by an in-house constructed power supply (RF amplitude 150 V0–peak at 500 kHz). Ions focused by this guide were collimated by a skimmer with a 1-mm diameter orifice, obtained from an LCQ DECA XP mass spectrometer, before entering the right arm of the T-quadrupole (Figure 5). The vacuum chamber was evacuated by a single Edwards 30 rotary pump to a pressure of 2.0–2.2 Torr. The pressure in the chamber was measured using a PRH 10 Pirani type pressure sensor (Edwards) connected to an Edwards 1011 controller.

The Thermo S-lens atmosphere-to-vacuum interface.To the other arm of the T-quadrupole, we attached an atmosphere-to-vacuum ion interface based on the one utilized in the commercial Velos Orbitrap and Q Exactivemass spectrometers produced by Thermo (henceforth referred to it as the S-lens interface). It consisted of a 59-mm long metal capillary, which can be heated to 350ºC, a stacked ring ion guide “funnel” (commercially named the S-lens electrode [28, 29]), a flat electrodess with an orifice diameter of 1.9 mm that separates the first vacuum region from the second, which houses a short quadrupole (referred to as “multipole 00”), followed by yet another flat electrode with an orifice diameter 4 mm. Here, we increased the diameter of this latter orifice to 8 mm so as to decrease the pressure in the chamber housing q0 to approximately 100 mTorr, which is close to the nominal operating pressure in this region in the commercial instruments. The first vacuum chamber was evacuated with two Edwards 30 mechanical pumps as in the original mass spectrometers implementing this interface. The pressure in this chamber was 1.0–1.2 Torr, measured by a PRH 10 Pirani type pressure sensor (Edwards) connected to an Edwards 1011 controller.

Other experimental apparatus. Ions were produced using one of the identical home-made electrospray ion sources. These sources were described in detail in our accompanying work [20].

We used two types of ESI emitters: (i) 15μm quartz tips from New Objective emitters (FS360-75-15-N-20) and (ii) stainless steel nano-bore emitters (ES301, TK190204, Proxeon).

The electrosprayed solution was delivered by a continuous double-flow syringe pump (PHD/Ultra, Harvard Apparatus) at flow rates of typically 400–600nl/min.

Photographs were taken with a Canon 6D digital camera with a 1–5x magnification macro lens (Canon MP-E 65mm).

Supplementary Figures

Figure S1. Spectra of peptides obtained when the ESI sources were positioned at the inlet of either the ConDuct or the S-lens atmosphere-to-vacuum interface. (A) Spectra of two unlabeled peptides, angiotensin I (MMavg= 1296.5u, triply charge ions at m/z 433.60 Th, and doubly charged ions at m/z 649.20 Th) and β-amyloid peptide, fragment 1-15 (MMavg= 1826.9u, quadruply charge ions at m/z 458.07 Th, and triply charged ions at m/z 610.07 Th) obtained by ESIof a 100 fmol/μl solution at a flow rate of 500nl/minat the ConDuct interface. (B) Spectra of the same two peptides but with some amino acids labeled with heavy isotopes of carbon and nitrogen. The MMavgof heavy labeled angiotensin was 1309.5u and the MMavgof heavy labeled β-amyloid peptide was 1843.9u. A 100 fmol/μl solution was electrosprayed at a flow rate of 500 nl/min at the S-lens interface. (C) Spectrum of unlabeled peptides obtained by quickly re-positioning the ESI source at the inlet of the S-lens interface. (D) Spectrum of the heavy labeled peptides obtained by quickly re-positioning the ESI source at the inlet of the ConDuct interface. The charge states of the unlabeled peptides are shown in blue, and the labeled peptides in red. Some prominent impurity peaks are labeled in light gray.

Figure S2. Fragmentation of angiotensin I peptides obtained at the increased RF amplitude of the amplitude of the q0 quadrupole ion guide. (A) unlabeled (MMavg= 1296.5u) and (B) labeled with heavy isotopes (MMavg= 1309.5u). (C) MS/MS spectrum of the triply charged ion species selected at m/z 435 ± 4 Th in the precursor spectrum (Fig. 7A). The unlabeled (light) peptides were introduced through the ConDuct interface and the labeled (heavy) peptides were introduced through the Thermo interface. After the spectrum was collected, the electrospray ion sources were quickly swapped and the data were collected again. (D) MS/MS spectrum of the triply charged ion species selectedatm/z 435±4 Thin the precursor spectrum (Fig. 7B). Unlabeled peptides were transmitted through the Thermo interface and the labeled ones through the ConDuct interface. From the intensity ratios of the b5fragments of the unlabeled and labeled peptides in the two spectra, we calculated the relative transmission efficiencies of both interfaces. The ion transmission of the ConDuct interface was 6 times better (from )than the ion transmission of the Thermo interface.

Figure S3. Spectra of peptides obtained when the ESI sources were positioned at the inlet of either the ConDuct or the S-lens atmosphere-to-vacuum interface at the different voltage settings then the ones used to obtain spectra shown in Fig 1S. In particular, we used higher amplitude of the RF voltage in the q0 (see Fig. 5A). The ratio of ion transmissions of ion interfaces become 6 (see Fig 2S) at the expense of considerable worsening of signal-to-noise ratio of the spectra of ions passed through the Themo S-lens interface.

1