A full-epitaxial bulk acoustic wave (BAW) resonator is attractive because of its high Q and high-power handling capability. An epitaxial technique is difficult to be employed due to the amorphous SiO2 low acoustic impedance layer in the solidly mounted resonator (SMR), which consists of a piezoelectric thin film on an acoustic Bragg reflector. In this study, we report a full-epitaxial ScAlN or MgZnO SMR based on an epitaxial Ti/Pt or ZnO/Pt acoustic Bragg reflector. The pole figure of epitaxial ScAlN and MgZnO piezoelectric layers showed clear sixfold symmetry. The epitaxial Sc0.43Al0.57N SMR exhibits keff2 of 13.6%. Moreover, the full-epitaxial metal Bragg reflector can act as a thick bottom electrode. This thick electrode is attractive for high frequency operation above 10 GHz in which BAW filter suffers from Q degradation due to the increase in the resistivity of extremely thin electrode.
High sharpness (Q factors) is required for RF filters to prevent interference between neighboring frequency bands. In the bulk acoustic wave (BAW) filters, lower dielectric and mechanical losses in the single crystalline piezoelectric thin films should contribute to the increase in Q factors and high-power handling. The available methods for obtaining single crystalline piezoelectric thin films are either the bulk crystal slicing technique or the epitaxial growth technique. In recent years, BAW resonators with single crystalline LiNbO3 (Refs. 1–6) or LiTaO3 (Refs. 5–8) films sliced from the bulk crystals have been reported. In addition, the film transfer techniques have demonstrated success in fabricating the periodically polarized piezoelectric film (P3F) resonators. The P3F based on the film transfer of sliced LiNbO3 single crystalline thin films,9–12 epitaxial ScAlN thin films,13,14 or polarization inverted epitaxial ScAlN thin films by applying voltage15 have been proposed. In order to obtain a thin single crystalline layer in large areas (over 8 in.), the epitaxial growth technique should be better.
Two types of BAW resonators, FBAR (film bulk acoustic resonator) and SMR16 (solidly mounted resonator), have been employed in the RF front-end.17–19 FBAR has self-standing thin film structures using air-gap between the resonance layer and the substrate. On the other hand, acoustic Bragg reflector allows acoustic isolation of the resonance layer and the substrate in the SMR. Power durability of the SMR should be higher than the FBAR because of heat dissipation into the supporting substrate. However, the epitaxial growth technique is difficult to be used to obtain a single crystalline piezoelectric thin film on the acoustic Bragg reflector because amorphous SiO2 is commonly used for a low acoustic impedance layer. In this study, we propose full-epitaxial piezoelectric layers and an acoustic Bragg reflector on a single crystal substrate using Ti or ZnO thin films instead of the amorphous SiO2 as a low acoustic impedance layer.
All epitaxial thin films were grown by RF magnetron sputtering. First, the acoustic Bragg reflector with five pairs of epitaxial (0001) Ti/(111) Pt or six pairs of epitaxial (0001) ZnO/(111) Pt was grown on the (0001) sapphire substrate. Next, the (0001) ScAlN or (0001) MgZnO piezoelectric layer was epitaxially grown on the epitaxial acoustic Bragg reflector, respectively. For the fabrication of the epitaxial ScAlN thin films, a 3 in. Al metal target with embedded Sc grains or a 3 in. Sc0.43Al0.57 alloy target was used.20,21 A ceramic ZnO target with 4 in. was used for the epitaxial ZnO thin film growth, and two powder target (ZnO:MgO = 70:30) with the diameter of 2 in. was used for the epitaxial MgZnO thin film growth. In the epitaxial Sc0.43Al0.57N SMR with five pairs of Ti/Pt, a Sc0.12Al0.88N thin film was inserted as a buffer layer between the epitaxial Pt bottom electrode and the Sc0.43Al0.57N piezoelectric layer. Finally, the patterned Au top electrodes were deposited on epitaxial piezoelectric layers. The growth conditions of the epitaxial thin films are shown in Table I.
. | Ti . | ZnO . | Pt . | Sc0.12Al0.88N . | Sc0.43Al0.57N . | Mg0.30Zn0.70O . |
---|---|---|---|---|---|---|
Sputtering gas | Ar | Ar/O2 = 3 | Ar | Ar/N2 = 4 | Ar/N2 = 4 | Ar/O2 = 3 |
Deposition pressure (Pa) | 0.5 | 1.0 | 0.5 | 0.5 | 0.5 | 1.0 |
Target-substrate distance (mm) | 30 | 30 | 30 | 20 | 20 | 30 |
Substrate temperature (°C) | 350 | 500 | 500 | 450 | ⋯ | 450 |
RF power (W) | 100 | 150 | 80 | 100 | 200 | 80 |
Deposition time (min) | 30 | 13 | 10 | 50 | 10 | 40 |
Thickness | 760 nm | 540 nm | 450 nm | 1.6 μm | 1.3 μm | 850 nm |
. | Ti . | ZnO . | Pt . | Sc0.12Al0.88N . | Sc0.43Al0.57N . | Mg0.30Zn0.70O . |
---|---|---|---|---|---|---|
Sputtering gas | Ar | Ar/O2 = 3 | Ar | Ar/N2 = 4 | Ar/N2 = 4 | Ar/O2 = 3 |
Deposition pressure (Pa) | 0.5 | 1.0 | 0.5 | 0.5 | 0.5 | 1.0 |
Target-substrate distance (mm) | 30 | 30 | 30 | 20 | 20 | 30 |
Substrate temperature (°C) | 350 | 500 | 500 | 450 | ⋯ | 450 |
RF power (W) | 100 | 150 | 80 | 100 | 200 | 80 |
Deposition time (min) | 30 | 13 | 10 | 50 | 10 | 40 |
Thickness | 760 nm | 540 nm | 450 nm | 1.6 μm | 1.3 μm | 850 nm |
As shown in Fig. 1, the epitaxial ScAlN or MgZnO piezoelectric layer was fabricated on the epitaxial acoustic Bragg reflector with five pairs of Ti/Pt or six pairs of ZnO/Pt. However, we can see the large roughness at the interfaces of ZnO/Pt in Figs. 1(c) and 1(d). The roughness can be improved using a CMP process.
The crystal orientation of epitaxial thin films was evaluated by x-ray diffraction (X'Pert PRO, PANalytical). Figure 2 shows the 2θ-ω scan XRD pattern of the epitaxial SMR. We can observe some diffraction peaks corresponding to the (0001) ScAlN or (0001) MgZnO piezoelectric film and the (111) Pt/(0001) Ti or (0001) ZnO acoustic Bragg reflector. FWHM values of the ω-scan rocking curves of the epitaxial acoustic Bragg reflector with five pairs of (0001) Ti/(111) Pt or six pairs of (0001) ZnO/(111) Pt were determined as 0.38°, 0.47°, 1.7°, and 1.6°, respectively. Figure 3 shows the XRD pole figures of the acoustic Bragg reflector. All layers exhibit good crystal orientation and sixfold symmetry, indicating epitaxial growth. Next, the out-of-plane and in-plane orientation of the piezoelectric thin films was evaluated. FWHM values of ω-scan rocking curves of the epitaxial (0002) Sc0.12Al0.88N and (0002) Sc0.43Al0.57N on five pairs of (0001) Ti/(111) Pt were measured to be 2.2° and 2.7°, respectively. On the other hand, FWHM values of ω-scan rocking curves of epitaxial (0002) Sc0.20Al0.80N and (0002) Mg0.30Zn0.70O on six pairs of (0001) ZnO/(111) Pt were measured to be 7.4° and 2.0°, respectively. In general, the crystallinity of epitaxial ScAlN films tends to degrade when Sc concentration increases. In this experiment, although sixfold symmetry was observed in Sc0.43Al0.57N, crystallinity of both out-of-plane and in-plane films was poorer than that of Sc0.12Al0.88N, as seen in the ω rocking curve FWHM or pole figures. On the other hand, we confirmed that ω rocking curve FWHM of (0001) Mg0.30Zn0.70O epitaxial films on unroughed (111) Pt surface was observed to be 1.6° in the additional experiment. In contract, (0001) Mg0.30Zn0.70O on roughed (0001) ZnO/(111) Pt Bragg reflector showed the ω rocking curve with the FWHM value of 2.0°. The degradation of the crystal orientation of these piezoelectric layers on ZnO/Pt may be due to the large roughness of the ZnO layer. As shown in Fig. 4, a sixfold symmetry in the (10 1) plane pole figure indicates the epitaxial growth of ScAlN and MgZnO piezoelectric layers on the epitaxial acoustic Bragg reflector.
The crystal orientation of local point of Sc0.12Al0.88N SMR with five pairs of Ti/Pt were observed by TEM (JEM-2010, JEOL). Figure 5(a) shows the cross-sectional TEM image. Figures 5(b) and 5(c) show the electron diffraction pattern of the fifth pair of Ti and Pt, respectively. Figure 5(d) shows the Sc0.12Al0.88N piezoelectric layer. Spot electron diffraction pattern of the fifth pair of Ti/Pt acoustic Bragg reflector implies a single crystal-like structure. Moreover, the Sc0.12Al0.88N layer exhibits spot electron diffraction pattern on the [11 0] zone axis. These results indicate that the Sc0.12Al0.88N SMR with five pairs of Ti/Pt is epitaxially grown from the sapphire substrate to the piezoelectric layer.
Piezoelectric layer . | Sc0.12Al0.88N . | Sc0.43Al0.57N . | Mg0.30Zn0.70O . |
---|---|---|---|
Bragg reflector | Five pairs of Ti/Pt | Five pairs of Ti/Pt | Six pairs of ZnO/Pt |
Qs | 206 | 130 | 39 |
Qp | 167 | 38 | 47 |
Qmax | 180 | 153 | 24 |
Piezoelectric layer . | Sc0.12Al0.88N . | Sc0.43Al0.57N . | Mg0.30Zn0.70O . |
---|---|---|---|
Bragg reflector | Five pairs of Ti/Pt | Five pairs of Ti/Pt | Six pairs of ZnO/Pt |
Qs | 206 | 130 | 39 |
Qp | 167 | 38 | 47 |
Qmax | 180 | 153 | 24 |
In conclusion, the SMR, where epitaxial ScAlN or MgZnO piezoelectric layer was deposited on epitaxial Ti/Pt or ZnO/Pt acoustic Bragg reflector, was characterized. Epitaxial ScAlN and MgZnO piezoelectric layers show good crystal orientation and clear sixfold symmetry. keff2 of epitaxial Sc0.12Al0.88N and Sc0.43Al0.57N SMR with five pairs of Ti/Pt was determined to be 5.1% and 13.6% by resonance–antiresonance method, respectively. The full-epitaxial metal Bragg reflector can be used as a thick bottom electrode, which might improve series resistivity of the resonator operating above 10 GHz. This should contribute to avoid Q degradation due to the extremely thin bottom layer in the high frequency region.
This work was supported by the JST CREST (No. JPMJCR20Q1), JST FOREST (No. JPMJFR212L), JST A-STEP (No. JPMJTR231C), and KAKENHI (No.21K18734).
AUTHOR DECLARATIONS
Conflict of Interest
The authors have no conflicts to disclose.
Author Contributions
Satoshi Tokai: Conceptualization (supporting); Data curation (lead); Formal analysis (lead); Investigation (lead); Methodology (equal); Software (supporting); Validation (equal); Visualization (equal); Writing – original draft (equal); Writing – review & editing (equal). Takahiko Yanagitani: Conceptualization (lead); Data curation (supporting); Formal analysis (supporting); Funding acquisition (lead); Investigation (supporting); Methodology (equal); Project administration (lead); Resources (lead); Software (lead); Supervision (lead); Validation (equal); Visualization (equal); Writing – original draft (equal); Writing – review & editing (equal).
DATA AVAILABILITY
The data that support the findings of this study are available from the corresponding author upon reasonable request.