(Color online) Gas pulsing sequence and the course of the controlled total process pressure (solid line) during one complete ALD cycle (55 s).
High-resolution transmission electron microscopy cross-section of a (10.7 ± 0.6) nm thick and 0.87 nm rms-rough Ru ALD film indicating an underlying silicon silicon dioxide interface that was supposed to be formed during the ALD growth initiation. The thickness and roughness of the optical Ru layer stack model constituted in Table I as well as the nominal Ru film thickness calculated from Eq. (1) are illustrated aside.
(Color online) Distribution of the GPC data points which interpolated the slope of the nominal Ru film thickness over ten ALD cycles under standard process conditions listed in Table II. The line represents an appropriate fitted GAUSSIAN normal distribution.
(Color online) (a) Optical thicknesses of the Ru layer and the EMA layer (i.e., the surface roughness) fitted on the basis of the optical model in Table I as well as the mean squared error of the regression analysis in the course of ALD cycles. (b) The growth per cycle, interpolating the slope of the nominal Ru film thickness over ten cycles, in the progress of ALD cycles. A fairly stable growth region with negligible statistical spread resulted between 100 and 230 ALD cycles.
(Color online) Growth per cycle for varying ECPR pulsing times. Each dot represents the interpolated slope of the nominal Ru film thickness over ten ALD cycles in which a respective ECPR pulsing time was set. The solid line shows the kinetic growth model of Eq. (2) fitted to the individual data points.
(Color online) Growth per cycle for varying O2 pulsing times. Each dot represents the interpolated slope of the nominal Ru film thickness over ten ALD cycles in which a respective O2 pulsing time was set. The advanced kinetic growth model of Eq. (3) (solid line) relates these individual data points to each other. The dashed line represents the first growth process without superposition of the second.
(Color online) Growth per cycle as a function of the Ar purging times after ECPR and O2 pulses.
(Color online) Growth per cycle as a function of the real substrate temperature. Each dot represents the interpolated slope of the nominal Ru film thickness over ten ALD cycles in which a respective stabilized substrate temperature was set. The solid line serves a guide to the eye, relating the individual data points to each other.
(Color online) Growth per cycle for a varied total working pressure, which was controlled to a fairly constant value employing a Baratron pressure gauge in combination with a butterfly valve.
Scanning electron microscopy pictures: (a) Delamination of a 25 nm thick Ru film. (b) Improved adhesion of a comparable film produced with an intermediate annealing step after 5 nm.
(Color online) Microsection of through silicon via structures with an aspect ratio of 10:1 after the direct electrochemical deposition of copper on an ALD Ru seed layer.
This optical layer stack model translates the ellipsometric spectra into structural film dimensions.
Overview of our process parameters for the thermal-activated ALD of Ru from ECPR and O2, revealed from a process development in the linear homogeneous Ru-on-Ru film growth regime.
Article metrics loading...
Full text loading...