A Novel 9.4 Tesla FTICR Mass Spectrometer with Improved Sensitivity, Mass Resolution, and Mass Range
- First Online:
- Cite this article as:
- Kaiser, N.K., Quinn, J.P., Blakney, G.T. et al. J. Am. Soc. Mass Spectrom. (2011) 22: 1343. doi:10.1007/s13361-011-0141-9
Fourier transform ion cyclotron resonance (FTICR) mass spectrometry provides unparalleled mass measurement accuracy and resolving power. However, propagation of the technique into new analytical fields requires continued advances in instrument speed and sensitivity. Here, we describe a substantial redesign of our custom-built 9.4 tesla FTICR mass spectrometer that improves sensitivity, acquisition speed, and provides an optimized platform for future instrumentation development. The instrument was designed around custom vacuum chambers for improved ion optical alignment, minimized distance from the external ion trap to magnetic field center, and high conductance for effective differential pumping. The length of the transfer optics is 30% shorter than the prior system, for reduced time-of-flight mass discrimination and increased ion transmission and trapping efficiency at the ICR cell. The ICR cell, electrical vacuum feedthroughs, and cabling have been improved to reduce the detection circuit capacitance (and improve detection sensitivity) 2-fold. The design simplifies access to the ICR cell, and the modular vacuum flange accommodates new ICR cell technology, including linearized excitation, high surface area detection, and tunable electrostatic trapping potential.
KeywordsFourier transform Ion cyclotron resonance ICR FTMS
Fourier transform ion cyclotron resonance mass spectrometry (FTICR MS) offers the highest resolving power m/Δm50% (Δm50% is the full peak width at half-maximum peak height) and mass measurement accuracy of any mass analyzer [1, 2]. Unparalleled peak capacity allows resolution and identification of thousands of components in a complex mixture in a single mass spectrum with sub-part per million (ppm) mass measurement accuracy . For bottom-up and top-down proteomics [4, 5], high mass measurement accuracy increases identification confidence and narrows database search space, and ultrahigh resolving power allows separation of overlapping isotopic distributions and enables charge state determination of large, multiply charged biomolecules [6, 7, 8]. FTICR is uniquely suited to analysis of complex natural organic mixtures, such as petroleum crude oil  and humic/fulvic acids . Ultrahigh mass resolving power (m/Δm50% ≥ 400,000) is needed to resolve thousands of peaks in a petroleum mass spectrum , and ultrahigh mass measurement accuracy is needed for confident elemental composition assignments . However, the number of unresolved multiplets increases at higher m/z due to decreased resolving power and increased variation in possible elemental compositions. The complexity of current applications thus requires further optimization and advancement of FTICR MS technology.
In FTICR-MS, ions are typically accumulated external to the magnetic field at a pressure many orders of magnitude greater than that required for image charge analysis in the ICR cell . The physical dimensions of the magnet and the need for multiple stages of differential pumping necessitate an ion flight path of approximately one meter between the ion accumulation trap and the ICR cell, leading to time-of-flight mass discrimination and decreased trapping efficiency during ion transfer to the ICR cell . The trapped ions are excited to a large cyclotron orbit by resonant RF voltage and subsequently observed by image charge detection. Longer time-domain transient lifetime is needed for greater resolving power, to resolve isobaric species and to assign elemental composition at high m/z [15, 16]. Coherent cyclotron motion of an ion packet is required to obtain high-resolution mass spectra. However, the number of coherently orbiting ions can diminish with time due to electric and magnetic field inhomogeneities and ion cloud interactions [17, 18].
The continuing development of higher magnetic field instruments increases several figures of merit, such as data acquisition speed, resolving power, reduced sensitivity to space charge effects, mass measurement accuracy, and dynamic range [6, 19]. However, many other instrument parameters besides higher magnetic field strength can be modified to improve instrument performance , e.g., increased ion transfer and trapping efficiency, and new ICR cell geometries that more closely approach harmonicity in the electrostatic trapping potential at large cyclotron radius [21, 22, 23]. Increased ion cyclotron radius can reduce space charge frequency shifts , and precise control of the number of trapped ions can reduce variation in space charge frequency shifts from successive ion injections . In addition, the signal magnitude is directly proportional to the proximity of the ions to the detection electrodes, so the closer the ion cloud is to the ICR cell, the greater the signal induced on the detection electrodes and the greater the sensitivity. However, deviation from ideal magnetic and electric field, ion-neutral collisions, and ion-image charge interactions increase with post-excitation cyclotron radius and limit transient lifetime and induce additional frequency shifts [26, 27, 28, 29]. The optimal post-excitation cyclotron radius depends on the required resolving power and therefore will vary with application .
Here we describe a substantial redesign and initial performance of a 9.4 tesla FTICR mass spectrometer. The prior instrument utilized the same 9.4 tesla magnet and generated numerous publications over the past 12 years [13, 31], setting several benchmarks for FTICR technology [32, 33, 34]. Prior modifications include the addition of a quadrupole mass filter and electron capture dissociation (ECD) capability, while the back end of the instrument (ion transfer optics, vacuum system, ICR cell) remained unchanged. The present modifications are designed to shorten the ion flight path from the external accumulation ion trap to ICR cell, improve optical alignment along the center of the bore of the magnet, increase accessibility to the ICR cell for future modifications, decrease detection circuitry capacitance for improved signal-to-noise and detection sensitivity, and facilitate complex experiments and future modifications.
2.1 Sample Preparation
Standard compounds (melittin, Ultramark, MRFA peptide, caffeine, and bovine apo-transferrin) were purchased from Sigma Aldrich (St. Louis, MO, USA) and used without further purification. Samples were dissolved in HPLC grade water at 1 mg/mL. Electrosprayed samples were diluted to 1–2 μM in 49:49:2 methanol/water/acetic acid or 49.75:49.75:0.5 methanol/water/formic acid. Direct infusion was performed at a flow rate of 0.5 μL/min. Middle Eastern crude oil was diluted to 0.25 mg/mL in toluene and infused at 50 μL/min into a modified ThermoFisher source  for atmospheric pressure photoionization.
2.2 Magnetic Field
The instrument is based on a passively shielded 9.4 tesla superconducting horizontal solenoid magnet (Oxford Instruments, Oxney Mead, UK) with a 225 mm room temperature bore diameter. The measured magnetic field inhomogeneity is 5 ppm peak-to-peak variation on the surface of a cylinder 40 mm in diameter and 80 mm along the symmetry (z-) axis of the magnetic field. Repeated measurements with and without the ICR cell and electrical feedthroughs installed confirmed that the chosen materials have no measurable impact on the magnetic field.
2.3 Vacuum System
2.4 Ion Transfer
Ions created at atmospheric pressure are transferred through different pressure regions by rf-only octopoles (1.6 mm diameter, 4.8 mm i.d. titanium rods) operated at 1.5 to 2.0 MHz and 190 < Vp-p < 240 rf amplitude. Ions pass through a quadrupole mass filter and accumulate in octopole 2 before transfer to the ICR cell. Helium gas is typically leaked into the accumulation region to maintain a pressure of 1 mTorr, but other gases (e.g., nitrogen or argon) are easily substituted. Appropriate acceleration of ions into octopole 2 can achieve collisionally activated dissociation (CAD). To produce more efficient fragmentation for proteins and peptides, the collision gas is switched to nitrogen. Octopole 2 has been modified with angled wire electrodes for improved ion extraction efficiency . A mini-UHV gate valve with 2 ¾ in. Conflat edges isolates the source region from the UHV region of the instrument. The distance between octopole 3 and conductance limit 4 (CL4) is set at 7 mm so that the UHV gate valve can operate freely. Based on signal magnitude, a negligible loss (<10%) in ion transmission occurs as ions traverse that gap compared to the prior 1 mm separation between octopole 3 and CL4. A single long transfer octopole has been divided into two individual transfer octopoles biased at −60 V and operated at 2.4 MHz and 250 Vp-p with a custom-built rf tuned circuit (Ardara Technologies, North Huntingdon, PA, USA). The two transfer octopoles are wired separately so that individual power supplies may be used for each rod set. The reduced capacitance of the separate octopoles compared to a single long octopole increases the available range of rf frequency and Vp-p amplitude.
2.5 ICR Cell
There are three stages of differential pumping in the UHV region of the instrument. The first two are located on the source side of the magnet, and the final stage is located on the back side of the magnet. The base pressure in the ion detection region is <2 × 10–10 Torr as measured by a Bayard-Alpert ionization gauge. The ICR cell (constructed from oxygen-free high conductivity (OFHC) copper with no additional treatment to the internal surfaces) is supported through the flange on the back UHV chamber, allowing for convenient access for modification of the ICR cell while maintaining optical alignment of the vacuum system along the center of the magnetic field. The ICR cell geometry is open cylindrical, with three equal sections of 100 mm in length and an inner diameter of 94 mm . The angular extents of the detection and excitation electrodes have been modified to be 120° and 60° . The gaps between all adjacent cell electrodes are 2.5 mm. The cell is wired externally for capacitive coupling . Ions are frequency-sweep excited (from high to low frequency) to ~40% of the cell radius. The post-excitation radius is optimized to minimize ion-ion interactions while maintaining ion cloud phase coherence for >5 seconds to achieve the resolving power required for analysis of complex mixtures. The waveform for dipolar excitation is input from the arbitrary waveform generator to a high power rf amplifier (ENI, Rochester, NY, USA) followed by a center tapped transformer (Northhills, Syosett, NY, USA) to split the waveform by 180° and provide additional rf voltage gain. Available ion fragmentation methods in the ICR cell include electron capture dissociation (ECD)  and infrared multi-photon dissociation (IRMPD) . A heated dispenser cathode is located on the magnetic field axis. A 40 W continuous IR laser (Synrad, Mukilteo, WA, USA) is mounted off-axis at an angle of 4° relative to the central axis of the system. The laser beam intersects the central axis of the system at approximately the center of the ICR cell.
2.6 Data Generation and Handling
Instrument control, data acquisition, and data analysis are carried out with a modular ICR data station . The time-domain ICR signal is Hanning apodized, zero-filled once, and Fourier transformed to produce a magnitude-mode spectrum that is converted to a mass-to-charge ratio by a two term calibration equation .
3 Results and Discussion
3.1 Ion Transfer
The user-defined flight time determines which ions are trapped and which are lost. With low voltage (5–15 V) applied to the extraction electrodes, ions initially at different axial position and velocity in the wired octopole are ejected at sufficiently different times to induce broadening of the arrival time distribution for each m/z, and it is more likely that heavy m/z ions that are immediately ejected from the accumulation trap reach the ICR cell at the same time as light ions that were ejected at a later time. However, a broader arrival time distribution reduces the trapping efficiency at the ICR cell for that species. Increased voltage on the extraction electrodes (~30 V) shortens the duration of ion ejection from the accumulation cell so that all ions are ejected at approximately the same time, thereby narrowing the arrival time distribution and the m/z range trapped in the ICR cell for a given flight time. Thus, the detected m/z range for ions will be determined by the acceleration voltage and the length of time between ion extraction from octopole 2 and the closing of the front trapping electrodes of the ICR cell. Lower voltage on the wires is preferred for detection of ions spanning a broad m/z range, whereas higher voltage is preferred for ions of narrow m/z range or for focusing of ions of a particular m/z, as for selective injection of precursor ions for in-cell fragmentation (ECD or IRMPD). Figure 2c illustrates the wider m/z distribution that is trapped when a lower voltage (15 V rather than 30 V) is applied to the extraction electrodes.
3.2 Detection Circuitry
The number of ions needed to create a detectable signal is directly related to the capacitance of the detection circuitry, including the ICR cell plates, electrical feedthroughs, and cabling . Reduction in capacitance of the system results in an increase in detected ICR signal magnitude. The new ICR cell has been designed with larger gaps between all cell plates to decrease capacitance. The wire that carries the image current from the detection plates to the back flange is shielded for noise reduction and the inner diameter of that grounded shield has been increased to 11 mm to minimize capacitance. The total capacitance of the detection circuitry is thus reduced by 2-fold from 145 pF to 70 pF, thereby doubling the ICR signal strength. Ideally, capacitance could be further reduced by placement of the preamplifier in vacuum adjacent to the ICR cell. However, outgassing, magnetic susceptibility, heat dissipation, and ease of operation limit implementation of that option.
3.3 Peptide and Protein Analysis
Mass analysis of intact proteins is inherently more difficult due to distribution of the analyte signal over a greater number of charge states, wider isotopic distribution, and adducts (cations, buffers, etc.), all of which decrease the signal magnitude of all measured species. Higher charge states also result in increased ion–ion Coulombic repulsion. MS/MS of these proteins results in a larger number of overlapping isotopic distributions, requiring high resolution to directly determine the charge states and masses of the fragment ions. Accurate determination of the masses of many fragment ions is needed to confidently assign sequence and post-translational modifications. In top-down analysis, large proteins are typically fraction-collected and analyzed off-line for target analysis, so that it is possible to signal-average a large number of acquisitions to generate useful fragment ion information . However, high resolution, high mass measurement accuracy MS and MS/MS data acquired on-line for intact proteins of 5–40 kDa have recently been reported [53, 54].
3.4 Complex Mixture Analysis
The analysis of compositional complex samples such as petroleum  or natural organic matter  requires high resolving power and mass measurement accuracy with a large dynamic range over a wide mass range. Ultrahigh resolution is needed to resolve the thousands of peaks in a single mass spectrum, resulting in >60 peaks per nominal mass , and to resolve mass doublets such as 12C3 vs. 32SH4 whose masses differ by only 3.4 mDa. Such resolution requires that the time-domain ICR signal persist essentially undamped for at least 3 s. Moreover, the fixed charge capacity of the ICR cell is distributed over thousands of m/z values, thereby severely limiting dynamic range for a single data acquisition. Thus, multiple (>100) time-domain acquisitions are typically averaged to obtain a single spectrum with acceptable signal-to-noise ratio and a dynamic range of 103–104. For complex mixtures, the new instrument achieves comparable signal-to-noise ratio at 8-fold lower sample concentration and with half the ion external accumulation time period.
The design and initial performance of our custom-built 9.4 tesla FTICR mass spectrometer are described. Improvements over the prior design increase sensitivity and overall performance of the instrument. The custom vacuum chambers reduce the ion transfer distance by 30% for decreased time-of-flight mass discrimination and improved ion transfer and trapping efficiency at the ICR cell. A reduction in the capacitance of the detection circuitry results in a 2-fold increase in sensitivity. The instrument is ideally suited for proteomic analysis of large intact proteins with the capability to isotopically resolve large intact proteins, and features fragmentation by IRMPD and ECD in the ICR cell as well as CAD in the external accumulation octopole. The instrument readily achieves the resolving power, high dynamic range, and mass measurement accuracy needed for complex mixture analysis. For the same signal-to-noise ratio, petroleum samples may now be analyzed at 1/8 the concentration and 1/2 the ion external accumulation period relative to the prior instrument. The current instrument modular configuration facilitates implementation of new technology. Future efforts will focus on further reduction in detection circuit capacitance for improved sensitivity, and exploration of new ICR cell configurations that approach ideal 3D axial quadrupolar trapping potential.
The authors thank Dr. Jeremiah D. Tipton for help with the top-down experiments and analysis, Dr. Amy M. McKenna for help with the petroleum data collection and analysis, Daniel McIntosh for fabrication of the custom components, and Saudia Aramco (Dhahran, Saudia Arabia) for providing the crude oil sample. This work was supported by NSF Division of Materials Research through DMR-0654118 and the State of Florida.