We studied theoretically the hysteresis behavior of ferroelectric thin films. The anomalous ferroelectric response is discussed by use of a bilayer model. Electrical conductivities of the films have been taken into account. To model the effects of the inhomogeneity of polarization and permittivity across the interface, the film is assumed to possess a secondary dielectric/ferroelectric phase ͑a dead or passive layer͒ with asymmetric conductivity. This configuration is found to produce large shifting ͑along the field axis͒ and deformation of the measured hysteresis loop. This is a manifestation of the asymmetric conductivity of the material. Theoretical calculation based on this model shows that the observed phenomena of shifted and skewed hysteresis loop in ferroelectric thin films can be explained in this way. © 2005 American Institute of Physics. ͓DOI: 10.1063/1.1853520͔The hysteresis behaviors of ferroelectric thin films have attracted great research interest for many years. Their bistable polarization as well as small size are promising characters for developing compact nonvolatile memories. Research studies of the anomalous effects and degradation mechanisms in the hysteresis behavior of ferroelectric thin films are essential for device applications. A most notable phenomenon is the large voltage offset along the horizontal ͑electric field͒ axis found in the hysteresis loop measurements. [1][2][3][4][5][6][7][8][9][10][11] The result is a deformed hysteresis loop with asymmetric switching. 6 Many researchers attributed the shift effect to the internal field caused by trapped charge carriers. [1][2][3][4] Other explanations included domain pinning, 5,6 rectifying effects formed at the ferroelectric-electrode interface 7-9 and the effect of passive layers. [10][11][12] It seems that the definitive mechanism is still not fully understood and the voltage shift may arise from multiple sources. Theoretical models to gain deeper insight using different approaches have been proposed. [11][12][13][14][15] In the absence of a unifying theory to account for all possible origins, we believe that the current theoretical understanding of the shift phenomenon can benefit from further input in terms of ideas and models. This work attempts to model this offset phenomenon by introducing a passive dielectric/ferroelectric layer at the filmelectrode interface. In the framework of the LandauGinzburg theory, the variation of polarization, which is usually more pronounced in thinner films, is modeled by a gradient term and extrapolation lengths in the free energy expression. This approach is suggested to be effectively the same as the existence of a passive layer ͑or dead layer͒. 16 It has been demonstrated that polarization gradient across a ferroelectric film can lead to asymmetric conduction. 17 The spatially varied polarization near the thin film surfaces is then thought to possess asymmetric electrical conductivity, while asymmetric conductivity may not emerge from the "normal" region of the film since the polarization shoul...