The experimental setup (shown schematically at the left) supplies pulses from an AWG to the device under test using two oscilloscope channels to measure the device voltage and current independently. To the right, a SEM image of the fabricated bridge device shows the narrow line of phase change material bridging two planarized TiN contacts separated by insulating . The inset shows the active device volume located above the narrow dielectric gap between the two electrodes (Ref. 12 ).
(a) Resistance of GeSb bridge devices as a function of applied voltage pulse starting from the as-deposited amorphous phase for different plateau durations. Here , , and denote leading edge time, plateau time, and trailing edge time of the pulses. The measured resistance after a voltage pulse drops to an intermediate level at about 1.2 V for all pulse lengths (threshold voltage). However, the fully crystalline set state is reached at different pulse voltages depending on the pulse width. The inset shows the same resistance data from part (a) but plotted as a function of applied power. (b) The current-voltage characteristic for a pulse that produces the transition from as-deposited amorphous to the intermediate state. (c) Current-voltage characteristic for a pulse that transforms the intermediate state to the set state.
Sketch of a phase change bridge device illustrating the crystalline (blue) and amorphous (yellow) portions of the device during the (a) initial Joule heating of the device, (b) intermediate state, in which the central region has finally crystallized yet has not grown to overlap the electrodes, the (c) set state, once the central crystalline region overlaps the electrodes at which point the high peak temperature in the bridge center can potentially induce melting of a small region, and the (d) reset state in which this central crystalline region is blocked by a plug of amorphous-as-melt-quenched material.
Resistance of a GeSb bridge device as a function of device power for both the set-to-reset transition (open circles, 14 ns pulses of varying amplitude) and the reset-to-set transition (filled squares, 560 ns pulses of varying amplitude). A trailing edge of 50 ns is sufficient to prevent the device from quenching into the reset state.
Resistance of GeSb bridge devices during the reset-to-set transition as a function of applied voltage for various pulse durations and trailing edge times. For a pulse longer than 260 ns, the device can be set with an applied voltage of . For shorter pulses, higher applied voltages and thus higher switching power are required.
(a) Resistance of a GeSb bridge device during the set-reset cycling with 10 ns pulses. Device (b) voltage and (c) current for the set (black solid curve) and reset (red dashed curve) pulses.
Comparison of the pulse power-duration parameter space for [(a) and (c)] electrically- and [(b) and (d)] optically-induced crystallizations of GeSb starting from the [(a) and (b)] amorphous-as-deposited and [(c) and (d)] amorphous-as-melt-quenched phases (Ref. 12 ). The color scale signifies (a) the resistance after an electrical pulse, (c) the resistance change after an electrical pulse normalized to the initial resistance , or [(b) and (d)] the reflectivity change measured after a laser pulse normalized to the initial reflectivity . Please note the different time axes in the upper two (microsecond) and the lower two (nanosecond) plots. Power comparisons between the electrical and optical experiments are complicated by the differences in the heating and cooling environments.
Article metrics loading...
Full text loading...