HexaMole mascotHexaMoleInteractive Chemistry

FT-ICR and Orbitrap

Drag to turn the view. Scroll or pinch to zoom. Double-click to reset.

Cyclotron motion

The cell sits in the bore of a superconducting magnet, with the field along its axis (slides 4–6). An ion moving across the field feels the magnetic force qvB, which points to the centre of its orbit, so it circles: qvB = mv²/r gives ωc = qB/m, or νc = qB/2πm (slides 12–13). The frequency depends only on m/z and the field, not on the ion’s speed or orbit size, which is why it measures m/z so precisely. CO⁺ at 1.00 T circles 548 000 times a second (slide 14); at 4.7 T, CO⁺, tropylium and protonated caffeine turn at 2.58 MHz, 793 kHz and 370 kHz, and the 3D view replays their motion about a million times slower. Small potentials on the end plates (+1 V) stop ions drifting along the field. “Upper mass limit” shows the thermal orbit, r = √(2mkBT)/qB, growing with mass until it reaches the cell wall: 66 MDa at 4.7 T in a 4 cm cell at 298 K (slides 19–20). The 21 T end of the field slider is the FT-ICR at the National High Magnetic Field Laboratory in Tallahassee.

Excite and detect

Press “Excite and detect”. Before excitation the ions circle on tiny orbits with random phases, and no current flows. A fast sweep of radio frequency on the excitation plates then runs from high to low frequency (slide 7). Each ion absorbs energy only while the sweep matches its own frequency, spirals outward, and ends on the same radius as the others, because that radius, r = VT/2dB, contains no m/z (slide 8): the light ion makes many turns, the heavy one few. Each excited group moves as a coherent packet; as it passes a detection plate it pulls electrons onto that plate, and the electrons slosh through the resistor as an image current at the ion’s frequency (slide 9). No ion is destroyed. The packets slowly fall out of step and the signal decays like a ringing bell (slide 10). The Fourier transform of the recorded transient gives a spectrum in Hz, and m/z = Be/2πν turns it into a mass spectrum (slide 11).

Transient and resolution

Two ions only 0.044 m/z apart, [Cl–H–Cl]⁻ at 70.9461 and Cl⁻(H₂O)₂ at 70.9905 (slide 18), circle at frequencies 636 Hz apart at 4.7 T. In the cell their packets start together and drift apart by one full turn every 1.57 ms, and the transient beats with that period. Record for less than that and the Fourier transform cannot tell the frequencies apart; record longer and the peaks narrow: R ≈ νcT/1.2 grows in proportion to the acquisition time (slide 17: 1.2 s gives about 8 times the resolution of 0.15 s). Collisions with background gas end the signal sooner; for a decaying signal the peak width is Δω = 2√3/τ, which caps the resolving power at R = ωcτ/2√3 (slide 16). A stronger field raises every frequency, and with it the resolving power: the 21 T MagLab instrument reaches R above 2 000 000.

Orbitrap

The Orbitrap has no magnet (slides 21–25). A spindle-shaped inner electrode (about −3.2 kV) pulls ions inward while they orbit it, and the shape of the two electrodes pushes them back and forth along the spindle. The frequency of that axial oscillation, ωz = √(k/(m/z)), depends only on m/z, not on the ion’s energy. The outer electrode is split in two, and the oscillating ions induce an image current between the halves; its Fourier transform gives the spectrum. Bovine insulin’s 5+, 4+ and 3+ ions oscillate at about 430, 384 and 333 kHz; with a longer transient the 4+ isotopes, 0.25 m/z apart, separate (slide 27 shows R = 60 000). Because ν ∝ 1/√(m/z), Orbitrap resolving power falls more slowly with m/z than FT-ICR’s; FT-ICR still wins at low m/z and in the highest fields. The Orbitrap was invented by Alexander Makarov and became commercial in 2005 (slide 23); it is the Fourier-transform analyzer most labs use today.

FT challenge

Four kinds of rounds, all solvable by hand from the Conditions box: the cyclotron frequency of an ion, the m/z of a measured frequency, the acquisition time needed for a given resolving power, and how the Orbitrap frequency changes with m/z.

Settings used

Cell radius 4 cm (drawn as 3 units), trapping plates +1 V, 298 K. Motion in the 3D views is slowed about a million times and orbits are drawn much larger than real. Spectra are computed exactly: the Fourier transform of each decaying sinusoid recorded for time T has a closed form, and the contributions are summed as complex numbers, then shown in magnitude mode. Orbitrap frequencies use ν = K/√(m/z) with K from a real trap (m/z 106 at 1413 kHz). Insulin isotope patterns from C254H377N65O75S6. Idealizations: perfect field, no magnetron motion or space-charge shifts. Slide notes: slide 3, one attomole is 6.0×10⁵ ions, not 17 000 (17 000 ions are about 28 zeptomoles), and atto- means one quintillionth; slide 12, the magnetic force qvB is the centripetal force, pointing to the centre; slide 18, the second label should be H₂O···Cl⁻···H₂O (m/z 70.99), since H₂O–H–OH₂ has mass 37.