In general, the ferroelectric fatigue is the dominant failure mechanism in FeRAMs, and thus must be minimized to prevent unintended memory malfunction during the specified lifetime. As shown in Figure 1796, Fatigue in ferroelectrics is a type of degradation phenomena in ferroelectric materials and is defined as the change of ferroelectric properties (e.g. the remanent polarization becomes small) with load cycles, including polarization reversals. Therefore, the fatigue also measures the loss of charge on repetitive switching. For FeRAM technology, the actual cycles for the fatigue test are ~10¹².

With fatigue in ferroelectrics, due to the decrease of the remanent polarization, the charge difference between logic “0” and “1” for FeRAM devices becomes smaller, which may cause failure. The degree of fatigue is quantified by the decrease of switchable polarization with respect to the number of switching cycles.

In practice, fatigue can be induced by both bulk- and interface-related phenomena as follows:
         i) In the bulks of ferroelectric films, domain walls can be pinned by charged defects, resulting in a reduction of switchable polarization. [1]
         ii) The domain walls can also be pinned by the interface between the ferroelectric film and the electrode due to the electromigration of charged defects and/or oxygen vacancies parallel to the interface. [2] For instance, it is proposed that the presence of space charges at the interfaces between Pt electrode and PZT (PbZr_xTi_1−xO₃) is responsible for the fatigue in PZT capacitors [4, 5].
         iii) Domain nucleation is suppressed by the trapped charges in the ferroelectric films. The trapped charges are injected, from the electrode into the film, by the large electric field during polarization switching. [3]
         iv) The fatigue can be induced by diffusion ions.

Some efforts have been taken to overcome the fatigue problem:
         i) Improvement of the film fabrication process.
         ii) Use new ferroelectric materials (e.g. see page1805). A significant improvement of stability (see page1789) is achieved with the application of multi-layer interstitial ferroelectric compounds, e.g. (Bi₂O₂)²⁺(A_m−1B_mO_3m+1)²⁻, here A = Ca²⁺, Sr²⁺, Ba²⁺, Pb²⁺, Bi²⁺, or La³⁺ and B = Ti⁴⁺, Ta⁵⁺, or Nb⁵⁺. For instance, for m = 2, Bi₂O₂ layers are alternated with perovskite-like AB₂O₇ layers with double oxygen octahedral units. [6,7]
         iii) Improvement of electrode materials (e.g. see page1805).
         

[1] E. L. Colla, D.V. Taylor, A.K. Taganstev, N. Setter: Appl. Phys. Lett. 72, 2478 (1998).
[2] M. Dawber, J. F. Scott: Appl. Phys. Lett. 76, 1060 (2000).
[3] I. Stolichnov, A.K. Tagantsev, E. L. Colla, N. Setter: Appl. Phys. Lett. 73, 1361 (1998).
[4] J. F. Scott, C.A. Araujo, B.M. Melnick, L.D. McMillan, R. Zuleeg: J. Appl. Phys. 70, 382 (1991).
[5] J.M. Benedetto, R.A. Moore, F.B. McLean: J. Appl. Phys. 75, 460 (1994).
[6] C.A. Paz de Araujo, J.D. Cuchiaro, L.D. McMillan, M. C. Scott, J. F. Scott: Nature 374, 627 (1995).
[7] T. Sumi, N. Moriwaki, G. Nakane, T. Nakakuma, Y. Judai, Y. Uemoto, Y. Nagano, S. Hayashi, M. Azuma, E. Fujii, S. Katsu, T. Otsuki, L. McMillan, C. A. Paz de Araujo, G. Kano: Tech. Dig. Int. Solid-State Circuits Conf., San Francisco (1994) p. 268.