LIF deconvolution example. The uncorrected LIF measured distribution is i (data were taken at 42 mm above the substrate, in the presheath region), the spread function due to saturation broadening is s and the deconvolved IVDF is o.
(Color online) Schematic diagram of the Laser Room optics showing the lasers, iodine calibration system, and beam shaping optics. The abbreviations are BD for beam dump, BS for beam splitter, L for lens, M for mirror, ND for neutral density filter, and PMT for photomultiplier tube. The upper optical system conditions the beam profile and couples the beam into the optical fiber. As discussed in the text, the iodine vapor system provides an absolute wavelength calibration of the dye laser.
(Color online) Chamber schematic: The laser beam is incident from the top of the machine such that the vertical IVDF can be measured. The silicon substrate was supported by a ceramic chuck with an embedded RF electrode referenced to the grounded chamber walls. The laser radiation transmitted from the Laser Room by the optical fiber is collimated into a 1 cm diameter beam by lens L1. Beam splitter BS1 splits a portion of the beam into a laser energy meter. The beam passes through a quartz vacuum window on the top of the chamber and strikes the wafer at normal incidence to measure the vertical (z component) of the IVDF. The CCD camera is carefully aligned such that its line of sight grazes the surface of the wafer, allowing LIF measurement down to 1 mm above the wafer surface. The fluorescence light passes through another quartz window on the side of the chamber, and is focused onto the CCD detector by a standard F/1.4 camera lens L2. F1 is a narrow (1 nm bandwidth) interference filter centered on the fluorescence wavelength of 461 nm to block background plasma light and scattered laser radiation.
Circuit diagram of the RF bias circuit and hardware connections. A 2.2 MHz low impedance RF supply45 transmits power through a resonant tank circuit. The RF voltage V RF and current I RF are monitored after the tank circuit. The connections leading up to the capacitive plasma applicator have an effective inductance L HW. The series capacitances of the RF electrode embedded in the ceramic chuck and the wafer itself were calculated from first principles. The complex plasma impedance was treated as a variable parameter in the circuit analysis.
Timing and triggering diagram showing the equipment and connections used to phase lock the laser pulse to the RF waveform, as well as to synchronize the CCD camera exposure gate to the laser (and thus the LIF signal).
Timing and triggering graph illustrating how the 10 Hz YAG laser output and camera exposure gate are phase-locked to the 2.2 MHz bias waveform. PG1 is internally triggered at 10 Hz, acting as the clock for the experiment. PG1 outputs a trigger to the YAG flash lamps, and then a 100 μs long gate signal to the phase lock unit. The phase lock unit gets a continuous RF bias signal from the bias current monitor. Upon receiving the gate signal, the phase lock unit will output a trigger pulse upon the next rising zero-crossing of the RF waveform. This output pulse triggers PG2, phase-locking it to the RF waveform. PG2 can be adjusted to select the RF phase at which the LIF envelope occurs.
(Color online) Oscilloscope traces showing RF phase locking of laser pulses, horizontal divisions are 100 ns for a total displayed time of 1 μs. The oscilloscope was operated in analog persist mode: over 1000 overlapping data acquisitions are shown. The three different channels are the RF bias current monitor, the photodiode signal (laser pulses), and the camera gate signal. The actual CCD exposure (and, therefore, laser pulse envelope) is delayed ∼60 ns from the gate signal. Laser pulses arrive within a time envelope ∼40 ns wide, in a well-defined phase interval of the 2.2 MHz RF waveform.
Bulk plasma electron temperature T e vs time t afer ICP power turn-off (solid curve). Data were taken with a swept (τsweep = 2 μs). Langmuir probe at several times in the plasma discharge after-glow. The dashed cuve is an exponential fit to the decaying electron temperature. Extrapolation of the fit to t = 0, gives the electron temperature during an ICP pulse to be T e ≈ 1.6 eV.
Ion density n i as a function of height z above the substrate as measured by a Langmuir probe biased to collect ion saturation current (V bias = −45 V). Data were taken at a radial position r = 7 cm from the center of the 30 cm diameter substrate.
Ion density n i as a function of radial position r across the 30 cm diameter substrate as measured by a Langmuir probe biased to collect ion saturation current (V bias = −45 V). Data were taken at a vertical position z = 3 cm above the 30 cm diameter substrate. Fill pressure was 0.5 mTorr.
Camera images with laser wavelength tuned (a) far from resonance λ0 and (b) on resonance λ0. Background images are subtracted from each frame (a) is dark, indicating the low noise level within the wings of the LIF distribution and the very small ion population with v = 60 km/s; (b) shows a large fluorescence signal from ions with v = 0 km/s.
Phase reference diagram. One cycle of the RF sheath voltage waveform is shown. Ions experience a maximum attractive force toward the substrate at the time when the voltage is most negative (phase 0).
Time-resolved normalized IVDFs at phase π/2 (see Fig. 12) acquired by LIF at two different heights above the substrate surface: z = 1.06 mm (dashed curve) and z = 42 mm (solid curve).
Time-averaged vertical ion energy distribution measured by LIF at z = 1.06 mm. The energy spltting of the two peaks is ΔE i = 380 eV at this location. The expected ΔE i at z = 0, the substrate surface is 1100 V.
Collection of vertical IVDFs measured by LIF at z = 42 mm.
Collection of IVDFs measured by LIF at z = 1.06 mm.
(Color) 3D surface plots of IVDF as a function of height above the substrate taken at four phases (see Fig. 12). The white line at z = s = 3.6 mm shows the location of the sheath boundary.
Vertical ion flux (Γi) vs height (z) above the substrate. (a) The flux at various times in the first half of an RF period and (b) the second. Flux profiles for all eight phases and the time-averaged flux profile (bold curve) is shown in both panels. The time-averaged flux profile is approximately flat throughout the sheath region (z < 3.6 mm), as expected from ion continuity in a collisionless sheath. The broken horizontal line indicates the value of the calculated Bohm flux . The observed position of the sheath edge z = s ≈ 3.6 mm is shown.
Vertical ion flow velocity u i vs height z above the substrate. Profiles at all eight phases are shown as well as the time-average (bold curve). A horizontal line indicates the Bohm speed u B = 1.93 km/s. The position of the observed sheath edge z = s ≈ 3.6 mm is shown.
Vertical heat flux Q vs height z above the substrate. Profiles at all eight phases are shown as well as the time-average (bold curve). The position of the observed sheath edge z = s ≈ 3.6 mm is shown.
Article metrics loading...
Full text loading...