Evolution of crystallographic structure and ferroelectricity of Hf0.5Zr0.5O2 films with different deposition rate

Hf0.5Zr0.5O2 films are one of the most attractive HfO2-based ferroelectric films because of good ferroelectricity, extreme thinness, and excellent compatibility with silicon devices. The origin of the ferroelectricity of Hf0.5Zr0.5O2 films is the noncentrosymmetric orthorhombic phase (space group Pca21). The effects of process temperature, annealing temperature, thickness, and doping to increase the portion of the orthorhombic phase, which contributes to ferroelectricity, have been studied extensively. However, although most studies have used atomic layer deposition, no study has been reported on the effect of the deposition rate on the ferroelectricity of Hf0.5Zr0.5O2 films. In this work, the influences of the deposition rate on the ferroelectricity and crystal structure of Hf0.5Zr0.5O2 films were examined. In order to conduct systematic and quantitative analysis, measurements of switching transient current, ferroelectric P-E curve, dielectric constant, deconvolution of grazing angle incidence X-ray diffraction, and piezoresponse force microscopy were performed. Hf0.5Zr0.5O2 films with a deposition rate of 1.1 A/cycle have a more ideal hysteresis curve shape, higher remanent polarization (initial state: 16 μC/cm2, wake up state: 22 μC/cm2), and a higher orthorhombic phase portion than other deposition rates.


I. INTRODUCTION
Recently, several doped distorted fluorite structured HfO 2 ferroelectric materials have been extensively investigated. [1][2][3][4][5][6][7][8][9] In particular, the admixture of 50 mol. % HfO 2 -50 mol. % ZrO 2 solid-solution (Hf 0.5 Zr 0.5 O 2 ) has drawn a lot of interest for semiconductor memory applications, thanks to their high remanent polarization (Pr), wide tunability, and reliability. [10][11][12][13][14][15][16][17][18] In general, it is accepted that the origin of the ferroelectricity of HfO 2 -based thin films is the formation of the noncentrosymmetric orthorhombic phase (o-phase, space group Pca21). 6,19 Even though the noncentrosymmetric o-phase is desirable for the ferroelectricity of Hf 0.5 Zr 0.5 O 2 films, the o-phase is a thermodynamically unstable phase. 20,21 Several recent theoretical and experimental studies demonstrated the origin of the ferroelectric o-phase and the influences of the various conditions such as film thickness, annealing temperature, process temperature, and doping cations on the formation and portion of the ferroelectric o-phase. 5,15,[20][21][22] Despite recent active studies, the focus of most studies on the emergence origin of the o-phase is tailored to subsequent processes. 5,15,[20][21][22] The films could have in situ crystallized nuclei of about 1 nm during atomic layer deposition (ALD), and the initial nuclei act as seeds to affect the final grain of more than 5 nm in size, which crystallizes during postmetallization annealing (PMA). Although previous studies have suggested that the o-phase is thermodynamically stabilized at a particular grain size, 15,23,26 the effect of the initial nuclei on the as-deposited state of the grain size accompanying the ferroelectricity-induced crystal structure has been overlooked.
In this work, the influence of the deposition rate for the ferroelectric properties of Hf 0.5 Zr 0.5 O 2 films from the viewpoint of the crystallographic structures and interface properties was carefully examined. To perform quantitative analysis, deconvolution of glancing incidence X-ray diffraction (GIXRD) measurement in the 2 theta range of 25 ○ -35 ○ of Hf 0.5 Zr 0.5 O 2 with various deposition rates was conducted by fitting to the superposition of Gaussian functions to determine the integrated peak areas and based on peak positions reported in previous studies. 10,15,24,25 It was found that ferroelectric properties and crystal structure composition are strongly dependent on the film deposition rate. Hf 0.5 Zr 0.5 O 2 films with a deposition rate of 1.1 Å/cycle have a more ideal hysteresis curve shape and higher Pr (initial state: ∼16 μC/cm 2 , wake up state: ∼22 μC/cm 2 ) than other deposition rates with the highest o-phase portion (0.48). Therefore, our research demonstrates that the grain size accompanying ferroelectricity can be controlled by the deposition rate with the initial nuclei at the as-deposited state by quantitatively analyzing the ferroelectric properties and corresponding crystallographic structures that depend on the film deposition rate. This new approach to controlling the grain size by controlling the deposition rate provides more opportunities to improve the performance of high-quality HfO 2 -based ferroelectric films.

II. EXPERIMENTS
The 10 nm thick Hf 0.5 Zr 0.5 O 2 films were deposited on an oxidized silicon substrate deposited with a 50 nm thick TiN metal electrode, using atomic layer deposition (ALD) at a process temperature of 300 ○ C. A single source of Tetrakis(ethylmethylamido) hafnium(IV) (TEMAHf) and Tetrakis(ethylmethylamido)zirconium (IV) (TEMAZr) was used as a hafnium-precursor and a zirconiumprecursor, respectively. O 3 was used as an oxygen reactant. The Hf 0.5 Zr 0.5 O 2 films were deposited by depositing HfO 2 (0.9 Å/cycle, 1.1 Å/cycle, and 1.4 Å/cycle) and ZrO 2 (0.9 Å/cycle, 1.1 Å/cycle, and 1.4 Å/cycle) with an almost identical deposition rate at a 1:1 cycle ratio. Deposition rates of HfO 2 and ZrO 2 were controlled by controlling the respective source feeding time. A 50 nm thick TiN metal was deposited using DC sputtering as a top electrode. Then, a 30 nm thick Pt metal was deposited by using an e-beam evaporator through a shadow mask as contact pads. Subsequently, crystallization of the Hf 0.5 Zr 0.5 O 2 films was performed through rapid thermal annealing (RTA) in an N 2 atmosphere for 30 s at 500 ○ C. For electrical characterization, the ferroelectric P-E curves and capacitance vs electric field curves, for dielectric constant vs electric field (εr-E) curves, were measured by using a semiconductor parameter analyzer (Keithley 4200A). The evolution of the crystal structural properties of the Hf 0.5 Zr 0.5 O 2 films with various deposition rates was examined by GIXRD.

III. RESULTS AND DISCUSSION
The deposition rate-dependent electrical properties are witnessed in Figs. 1(a) and 1(b), and a summary of the Pr values of the Hf 0.5 Zr 0.5 O 2 films with various deposition rates is depicted in Fig. 1(c). Generally, when a ferroelectric P-E curve is obtained, a triangular bipolar pulse is applied. Figure 1(a) shows the direct observation of the switching instantaneous current during a triangular bipolar pulse excitation, which is the change in charge with the change in time between each point, and the corresponding hysteresis curves are shown in Fig. 1 Fig. 1, Hf 0.5 Zr 0.5 O 2 with a deposition rate of 1.1 Å/cycle has an ideal hysteresis curve with a high Pr value of ∼16 μC/cm 2 and a switching current despite the initial state. Hf 0.5 Zr 0.5 O 2 with a deposition rate of 0.9 Å/cycle has a slightly more tilt and the most distorted hysteresis curve (P r,initial value: ∼9 μC/cm 2 ), and the hysteresis curve of Hf 0.5 Zr 0.5 O 2 with a deposition rate of 1.4 Å/cycle is the flattest (P r,initial value: ∼5 μC/cm 2 ).

(b). As shown in
To examine the variation of ferroelectric properties with electric field cycles (the wake-up state), hysteresis curves (left panel) and

ARTICLE scitation.org/journal/adv
εr-E curves (right panel) were measured with an increase in the number of electric field cycles, as shown in Fig. 2. When the Hf 0.5 Zr 0.5 O 2 films are in the initial state, the hysteresis curve shapes were tilted with small humps, especially in the deposition rates of 0.9 and 1.4 Å/cycle. However, as the number of electric cycles increases, the shape of the hysteresis curves becomes more ideally shaped (squarelike). The εr-E curves (right panel) show the butterfly-like shape, which is a typical ferroelectric characteristic. As the number of electric cycles increases, the dielectric constant values of Hf 0.5 Zr 0.5 O 2 films approach 30. The variation in the dielectric constant value makes it easier to analyze the crystal structure than the shape of the hysteresis curve. The dielectric constant values of t-, o-, and m-phases are 17-20, ∼30, and 35-40, respectively. As shown in the right panel of Fig. 2 and in Table I 23 Assuming that the V O effect is the same in Hf 0.5 Zr 0.5 O 2 if 10% of V O are present in Hf 0.5 Zr 0.5 O 2 , the free energy of the t-phase is further reduced by 7meV and 31meV compared to the o-phase and m-phase, respectively. Therefore, the V O concentration in Hf 0.5 Zr 0.5 O 2 should be higher when the Hf 0.5 Zr 0.5 O 2 film is deposited with a higher deposition rate, and a higher t-phase portion is expected, which is inconsistent with the observed experimental result. Therefore, the ferroelectric properties of each deposition rate cannot be explained by the influence of oxygen vacancies alone. Thus, in order to understand the ferroelectric properties of each deposition rate, various approaches are required such as the viewpoint of crystallographic structures and the interface   15 In addition, by using the Gaussian function, the diffraction peak could be deconvoluted by (m-111), (o111), (t011), and (m111) located at 28.54 ○ , 30.4 ○ , 30.8 ○ , and 31.64 ○ , as shown in Fig. 3. The relative ratio of each phase that is obtained by integrating the area of each Gaussian peak is summarized in Table I  Hoffmann et al. suggested a qualitative model for film crystallinity of doped-HfO2, which depends on film thickness, dopant concentration, and thermal budget. 23 According to this model, the thicker film leads to more m-phase and the thinner film leads to more t-phase. The o-phase exists somewhere at the boundary of the phase between t-and m-phases. Park et al. reported that the Hf 0.5 Zr 0.5 O 2 films become crystallized as the film thickness increases with the accompanying increase in grain size. 15 The authors reported that a 5.5 nm Hf 0.5 Zr 0.5 O 2 film with a small grain size is crystallized mostly with the t-phase, and the crystallization occurs mostly with the o-phase when the film thickness reaches to 10 nm with a bigger grain size. Then, there is an increasing portion of the m-phase with the increasing film thickness so that the m-phase grows on top of the o-phase and t-phase. 15 Materlik et al. suggested that the o-phase can be thermodynamically stabilized at the appropriate grain size because the surface energy and bulk free energy of the o-phase are between the t-phase and the m-phase. 26 Deposited state films could have in situ crystallized nuclei of about 1 nm in size during ALD, and the initial nuclei affect the final grain of more than 5 nm in size, which crystallizes during PMA. Therefore, the initial nuclei according to the deposition rate must be considered to control grain size. It can be deduced that the initial nuclei of low deposition rates lead to small grain sizes with the t-phase during crystallization, while the initial nuclei of high deposition rates lead to a large grain size with the m-phase during crystallization. The o-phase requires a proper grain size, the grain size can be controlled by the deposition rate, and 1.1 Å/cycle is the most suitable deposition rate.
In addition, the film thickness effect was observed in Hf 0.5 Zr 0.5 O 2 films with various deposition rates, as shown in Fig. 4. HZO films exhibit the same trend under all deposition rate conditions, depending on the thickness effect: a 6 nm thick film with a small grain size has a distorted hysteresis curve dominated by the t-phase, a 10 nm thick film with a proper grain size has an ideal ferroelectric P-E curve dominated by the o-phase, and as the thickness increases to 13 nm with a large grain size, the m-phase portion increases and the remanent polarization has drastically reduced. This observation suggests that the effect of the deposition rate inducing the change in grain size discussed above is valid. tip. The bidomain phase images present dark (−3 V) and bright (+3 V) contrast regions with the near 180 ○ switching phase difference, which stands for the downward and upward polarization states, respectively. Pulse switching current measurements were performed to evaluate the interfacial properties with ferroelectric switching kinetics. In this measurement, first, a polling pulse (10 μs) is applied in the opposite direction of the switching pulse, and then, a switching pulse (10 μs) is applied to measure the switching current with various magnitudes of the applied switching pulse. 18,27 As shown in Figs. 6(a) and 6(b), the switching current was evaluated in both the positive (up to down) and negative (down to up) switching pulses. According to the polarization reversal theory, the switching current can be expressed as follows: where t0 is the starting time of ferroelectric switching, RL is the total resistance of the measurement system, Ci is the interfacial capacitance, tsw is the time at which the switching process is finished, and I 0 sw is the current when the switching process is started. According to the proposed model of Kim et al., the value of Ci includes the dead-layer capacitance of the interface and the nonferroelectric phase capacitance of the interface. 27 As can be seen from the variation values of Ci in Fig. 6(c), the interface property of the Hf 0.5 Zr 0.5 O 2 film with 1.1 Å/cycle is also the best.

IV. CONCLUSION
In this work, the influence of the deposition rate on Hf 0.5 Zr 0.5 O 2 films was systematically examined from the viewpoint of the crystallographic structures and interface properties. It was shown that the Hf 0.5 Zr 0.5 O 2 film with 0.9 Å/cycle has a distorted hysteresis curve (P r, initial value: ∼9 μC/cm 2 ), ε r, initial value of 34, and t-phase with the highest relative ratio (0.47); the Hf 0.5 Zr 0.5 O 2 film with 1.1 Å/cycle has an ideal hysteresis curve (P r, initial value: ∼16 μC/cm 2 ), ε r, initial value of 31.5, and o-phase with the highest relative ratio (0.48); and the Hf 0.5 Zr 0.5 O 2 film with 1.4 Å/cycle has a flat hysteresis curve (P r, initial value: ∼5 μC/cm 2 ), ε r, initial value of 25, and m-phase with the highest relative ratio (0.44). This experimental observation suggests that the grain size can be controlled by controlling the deposition rate, and the low deposition rate leads to a small grain size with the t-phase and the high deposition rate leads to a large grain size with the m-phase. Even though this is the first experimental evidence of the ferroelectric properties of Hf 0.5 Zr 0.5 O 2 films with deposition rates, the material characterization is very limited to demonstrate the substantial property variation in Hf 0.5 Zr 0.5 O 2 films with exact process parameters. Therefore, a systematic and extensive study (i.e., XPS, HRTEM, and in situ monitoring) is needed as future work from the viewpoint of material engineering. Nevertheless, this work also gives hints that Hf 0.5 Zr 0.5 O 2 films with an appropriate grain size with the o-phase with high-quality ferroelectricity can be obtained from the initial film deposition stage by controlling the deposition rate.