function [11]. Figure 5 shows the experimental decay curve for the conductance of a
Ag 2 S-based atomic switch in STM mode: a power function was used to fit the
conductance curve to analyze the psychological behavior. As the number of
rehearsals increased, the decay speeds, m, became smaller and the memory retention
ratio increased. This result for the STM mode is reminiscent of the psychological
forgetting curve.
As shown in Fig. 2, we can see the formation of SM prior to the STM formation,
in which the conductance of the atomic switch increased little during the pulse
application, followed by a sharp decay back to its initial value immediately after the
pulse. In this operation, more than one input pulse is required to facilitate the
transition from SM to STM when a small bias voltage is used, and the number of
input pulses required for the first STM formation strongly depends on the amplitude
and width of the input voltage pulse. This is shown in Fig. 6 [12]. Increasing the
pulse amplitude decreases the number of input pulses required for the first STM
formation (Fig. 6a). In addition, the cumulative probability of STM formation was
found to depend on the amplitude of the input. For example, six input pulses with
V ¼ 80 mV brought 91% of switches into STM, whereas only 13% of switches were
brought into STM by the same number (six) of pulses with V ¼ 70 mV. Input pulses
with a broader width also facilitate the first STM formation with a smaller number of
input pulses (Fig. 6b). These results resemble the memorization process in the
human brain.
2.2 Cu 2 S-Based Atomic Switch
Similar to the Ag 2 S system presented in the previous subsection, the Cu 2 S-based
switch also exhibits three different memory behaviors analogous to the SM, STM and
LTM modes of the human brain. While Ag 2 S is an n-type material, Cu 2 S is a p-type
material. In Cu 2 S, Cu vacancies act as electron acceptors, giving rise to free holes for
Fig. 5 Forgetting curves
demonstrated by a Ag 2 Sbased atomic switch.
Copyright 2011, NPG
Artificial Synapses Realized by Atomic Switch Technology
181
Ag 2 S-based atomic switch in STM mode: a power function was used to fit the
conductance curve to analyze the psychological behavior. As the number of
rehearsals increased, the decay speeds, m, became smaller and the memory retention
ratio increased. This result for the STM mode is reminiscent of the psychological
forgetting curve.
As shown in Fig. 2, we can see the formation of SM prior to the STM formation,
in which the conductance of the atomic switch increased little during the pulse
application, followed by a sharp decay back to its initial value immediately after the
pulse. In this operation, more than one input pulse is required to facilitate the
transition from SM to STM when a small bias voltage is used, and the number of
input pulses required for the first STM formation strongly depends on the amplitude
and width of the input voltage pulse. This is shown in Fig. 6 [12]. Increasing the
pulse amplitude decreases the number of input pulses required for the first STM
formation (Fig. 6a). In addition, the cumulative probability of STM formation was
found to depend on the amplitude of the input. For example, six input pulses with
V ¼ 80 mV brought 91% of switches into STM, whereas only 13% of switches were
brought into STM by the same number (six) of pulses with V ¼ 70 mV. Input pulses
with a broader width also facilitate the first STM formation with a smaller number of
input pulses (Fig. 6b). These results resemble the memorization process in the
human brain.
2.2 Cu 2 S-Based Atomic Switch
Similar to the Ag 2 S system presented in the previous subsection, the Cu 2 S-based
switch also exhibits three different memory behaviors analogous to the SM, STM and
LTM modes of the human brain. While Ag 2 S is an n-type material, Cu 2 S is a p-type
material. In Cu 2 S, Cu vacancies act as electron acceptors, giving rise to free holes for
Fig. 5 Forgetting curves
demonstrated by a Ag 2 Sbased atomic switch.
Copyright 2011, NPG
Artificial Synapses Realized by Atomic Switch Technology
181