Radio frequency (RF) devices are becoming more multi-band, increasing the number of filters and other front-end components while simultaneously pushing towards reduced cost, size, weight, and power (CSWaP). One approach to reducing CSWaP is to eliminate high quality factor filtering that relies on acoustic wave technologies, allowing end-to-end solutions in CMOS or compound semiconductor platforms. Yet another way would be to augment the achievable functionalities of electromechanical/acoustic filtering chips to include “active” and nonlinear functionalities, such as gain and mixing. The acoustoelectric (AE) effect could enable such active acoustic wave devices, but, to compete with the performance of today’s multi-chip architectures, any solution to this problem needs to provide high gain, large bandwidth, low noise figure, and sufficient power handling, in addition to the reduced CSWaP. Surface acoustic wave (SAW) amplifiers based on the AE effect were demonstrated as early as the 1960s. In these devices, evanescent fields associated with piezoelectric acoustic waves interact with charge carriers undergoing voltage-induced drift. The interaction causes polarization of the drifting majority charge carriers that leads to a Coulomb drag effect, resulting in acoustic wave attenuation or amplification, analogous to a traveling-wave tube amplifier. The amplifier performance is improved by utilizing a piezoelectric substrate with a large electromechanical coupling coefficient, which alone determines the maximum gain, combined with an acoustically thin semiconductor with low conductivity and high mobility to lower the required voltage for a given gain. Previous demonstrations of SAW amplifiers have resulted in over 100 dB of RF output contrast (a.k.a. electronic gain), but at voltages exceeding 1 kV. Molecular beam epitaxy (MBE) and metal-organic chemical vapor deposition (MOCVD) were either in their infancy or not yet invented at the time, making it impossible to seamlessly integrate low defectivity, thin semiconductors with low conductivity and high mobility together with high electromechanical coupling coefficient piezoelectric substrates. Thus, previous AE amplifiers consisted of essentially one of two types: 1) evaporated thin semiconductor layers with large conductivity and high defectivity on piezoelectric substrates or 2) high quality bulk crystalline semiconductors separated from the piezoelectric by a thin air gap to prevent acoustic radiation into the thick semiconductor. The first approach yielded performance limited by the sub-optimal material properties; the second approach significantly reduced the AE interaction strength and made the devices more dispersive, which prevents high frequency operation. A third approach uses charge confinement structures or the high mobility of 2D materials, but performance of these devices has never approached that of the other two. Here, we present a high-gain leaky SAW (LSAW) amplifier based on heterogenous integration of an epitaxial III-V semiconductor and lithium niobate (LiNbO3), providing a solution to the aforementioned problems and enabling compact, high-gain devices with significantly lower power consumption. YX LiNbO3 is chosen due to its exceptionally high electromechanical coupling coefficient (~25%), providing the highest electromechanical coupling for a SAW on bulk LiNbO3. We demonstrate gain of 1.9 dB/Λ in a 338 μm long device, where Λ is the acoustic wavelength, equivalent to an electronic gain of 30.3 dB and a terminal gain of 5 dB. Moreover, on account of the directly bonded semiconductor layer’s low thickness and conductivity and relatively high gain, this large peak gain occurs at only -55 V.
Read full abstract
Dec 1, 2020
Ryotaro Ozaki + 1
Open Access