Resistive Random Access Memory Device …
323
is the barrier height at the electrode-oxide interface whereas for bulk-limited conduction mechanisms, the electrical properties of the oxide play a critical role. Typically,
RRAM devices exhibit ohmic or hopping conduction at the LRS [31]. However,
during HRS, the dominant conduction mechanisms vary even for similar MIM stacks
[37, 38] or could depend on the applied voltage regime [31].
2.2 Anion Devices
Switching Mechanisms
The active switching materials of anion RRAM devices include oxide dielectrics
such as transition metal oxides, complex metal oxides, large bandgap dielectrics, as
well as non-oxide dielectrics like chalcogenides and nitrides. In most metal oxides,
the mobile species are widely believed to be the oxygen anions which is equivalent
to the positively-charged oxygen vacancies. Hence, these RRAM devices are known
as oxygen vacancy based RRAM (OxRRAM). Additionally, the anion migration
leads to valence changes of the metal cations which leads to resistance change of
the metal oxide and thus these devices are also termed as valence change memories
(VCM) [5]. Since the resistance switching is caused by defects that modify electronic
transport that is not limited to a specific electronic structure, almost all insulating
oxides should show resistance switching behavior. In theory, the resistive switching
phenomenon should also be observable in other insulating compounds like halides,
borides, carbides and phosphides.
Since the earliest report of resistance switching in oxides by Hickmott in 1962
[7], oxides have been widely studied as anion-based switching materials, ranging
from simple binary transition metal oxides, e.g. TiO x [13, 32, 39], TaO x [30, 40, 41],
HfO x [42–45], ZrO x [46, 47], to rare-earth metal oxides, e.g. CeO x [48], EuO x [49],
to perovskite-type complex oxides such as SrTiO 3 and others [50–52]. In the even
larger set of non-oxide switching materials, the resistive switching phenomenon has
been shown in nitrides, e.g. AlN [53], selenides, e.g. ZnSe [54], tellurides, e.g. ZnTe
[55], and polymers [56, 57].
Insulating oxides can be regarded as semiconductors with native dopants. They can
be either n-type or p-type semiconductors if they are oxygen deficient or excessive,
respectively. Under applied high electric field and/or Joule heating, the electrically
and/or thermally driven motion of these native dopants leads to chemical changes
which results in resistance change. As the mobility and concentration of oxygen
vacancies or cation interstitials are adequately high in transition metal oxides [5, 58],
they are widely acknowledged as the mobile species responsible for the resistance
switching and this is supported by experimental evidence [27, 33, 59–61]. However,
more direct evidence is needed to confirm the actual mobile species even though
it is very difficult to detect and track the migration of oxygen vacancies in RRAM
devices [62–68].
323
is the barrier height at the electrode-oxide interface whereas for bulk-limited conduction mechanisms, the electrical properties of the oxide play a critical role. Typically,
RRAM devices exhibit ohmic or hopping conduction at the LRS [31]. However,
during HRS, the dominant conduction mechanisms vary even for similar MIM stacks
[37, 38] or could depend on the applied voltage regime [31].
2.2 Anion Devices
Switching Mechanisms
The active switching materials of anion RRAM devices include oxide dielectrics
such as transition metal oxides, complex metal oxides, large bandgap dielectrics, as
well as non-oxide dielectrics like chalcogenides and nitrides. In most metal oxides,
the mobile species are widely believed to be the oxygen anions which is equivalent
to the positively-charged oxygen vacancies. Hence, these RRAM devices are known
as oxygen vacancy based RRAM (OxRRAM). Additionally, the anion migration
leads to valence changes of the metal cations which leads to resistance change of
the metal oxide and thus these devices are also termed as valence change memories
(VCM) [5]. Since the resistance switching is caused by defects that modify electronic
transport that is not limited to a specific electronic structure, almost all insulating
oxides should show resistance switching behavior. In theory, the resistive switching
phenomenon should also be observable in other insulating compounds like halides,
borides, carbides and phosphides.
Since the earliest report of resistance switching in oxides by Hickmott in 1962
[7], oxides have been widely studied as anion-based switching materials, ranging
from simple binary transition metal oxides, e.g. TiO x [13, 32, 39], TaO x [30, 40, 41],
HfO x [42–45], ZrO x [46, 47], to rare-earth metal oxides, e.g. CeO x [48], EuO x [49],
to perovskite-type complex oxides such as SrTiO 3 and others [50–52]. In the even
larger set of non-oxide switching materials, the resistive switching phenomenon has
been shown in nitrides, e.g. AlN [53], selenides, e.g. ZnSe [54], tellurides, e.g. ZnTe
[55], and polymers [56, 57].
Insulating oxides can be regarded as semiconductors with native dopants. They can
be either n-type or p-type semiconductors if they are oxygen deficient or excessive,
respectively. Under applied high electric field and/or Joule heating, the electrically
and/or thermally driven motion of these native dopants leads to chemical changes
which results in resistance change. As the mobility and concentration of oxygen
vacancies or cation interstitials are adequately high in transition metal oxides [5, 58],
they are widely acknowledged as the mobile species responsible for the resistance
switching and this is supported by experimental evidence [27, 33, 59–61]. However,
more direct evidence is needed to confirm the actual mobile species even though
it is very difficult to detect and track the migration of oxygen vacancies in RRAM
devices [62–68].