68
W. C. Law and S. De W. Wong
MTJ nanopillars has also evolved from various methods, where etchback becomes
the norm as it can allow for MTJ features to go below 40x. Due to the hydrophilic
nature of MgO, in-situ encapsulation is done after etching to prevent the formation
of Mg(OH)2. Since the free layer consisting of CoFeB alloys achieves PMA through
interfacial effect and out-diffusion of the interstitial B atoms, a MTJ stack optimized
for annealing 400 °C process does not necessarily yield the equivalent or better
performance at lower processing temperature.
5 State-of-the-Art MTJ Stack Design
We have presented Fig. 2 as an ideal MTJ configuration earlier during the introduction, which has eliminated a lot of considerations for simplicity. However, the MTJ
stack should consist of appropriate materials to address the concerns as mentioned
in Sect. 4. As summarized in Fig. 10a, the key considerations in the design of pMTJ
stacks revolve not only around material, magnetic and electrical properties, but must
also contemplate on how each decision intertwines with the subsequent overlayers.
In addition, certain magnetostatic effects are only desirable in a certain scenarios to
achieve an objective, often at an expense to other parameters.
For instance, the hard layer has to be extremely stable in its magnetic configuration,
but a high stray field may be induced. The stray field may offset the coercivity field
of the soft layer via magnetostatic coupling, leading to an additional energy bias
required in order to induce magnetization reversal. Biasing is undesirable as it leads
to increase in energy consumption, decreased reliability (unintentional switching),
decrease in TMR (voltage biasing), and additional stress to the MTJ stack (e.g.
time-dependent dielectric breakdown). To resolve this problem, a synthetic antiferromagnetic (SAF) structure consisting of two hard layers coupled by ultrathin
Fig. 10 a Summary of stack design considerations. b State-of-the-art MTJ stack design
W. C. Law and S. De W. Wong
MTJ nanopillars has also evolved from various methods, where etchback becomes
the norm as it can allow for MTJ features to go below 40x. Due to the hydrophilic
nature of MgO, in-situ encapsulation is done after etching to prevent the formation
of Mg(OH)2. Since the free layer consisting of CoFeB alloys achieves PMA through
interfacial effect and out-diffusion of the interstitial B atoms, a MTJ stack optimized
for annealing 400 °C process does not necessarily yield the equivalent or better
performance at lower processing temperature.
5 State-of-the-Art MTJ Stack Design
We have presented Fig. 2 as an ideal MTJ configuration earlier during the introduction, which has eliminated a lot of considerations for simplicity. However, the MTJ
stack should consist of appropriate materials to address the concerns as mentioned
in Sect. 4. As summarized in Fig. 10a, the key considerations in the design of pMTJ
stacks revolve not only around material, magnetic and electrical properties, but must
also contemplate on how each decision intertwines with the subsequent overlayers.
In addition, certain magnetostatic effects are only desirable in a certain scenarios to
achieve an objective, often at an expense to other parameters.
For instance, the hard layer has to be extremely stable in its magnetic configuration,
but a high stray field may be induced. The stray field may offset the coercivity field
of the soft layer via magnetostatic coupling, leading to an additional energy bias
required in order to induce magnetization reversal. Biasing is undesirable as it leads
to increase in energy consumption, decreased reliability (unintentional switching),
decrease in TMR (voltage biasing), and additional stress to the MTJ stack (e.g.
time-dependent dielectric breakdown). To resolve this problem, a synthetic antiferromagnetic (SAF) structure consisting of two hard layers coupled by ultrathin
Fig. 10 a Summary of stack design considerations. b State-of-the-art MTJ stack design