1. Introduction
The range of applications associated with the submillimeter wave and terahertz bands (300 GHz–3 THz) is very extensive, such as spectroscopy, imaging and communications[1]. However, it has historically been extremely difficult to access this band due to a lack of high frequency transistors with bandwidths of above 300 GHz. The bandwidth of InP-based transistors, such as InP HBTs and InP HEMTs, has increased rapidly in recent years. To date, InP HBTs and InP HEMTs have both been demonstrated with maximum oscillation frequency exceeding 1 THz[2, 3]. Compared with InP HEMTs, InP HBTs have some key advantages in submillimeter and THz applications, such as a high breakdown voltage, high threshold uniformity, low 1/
For submillimeter and THz applications, the circuits operate near the cutoff frequency of the DHBT, where the power gain of the DHBT decreases sharply with increasing frequency, so the power gain is very precious[4]. InP DHBT is commonly used in common-emitter configuration. However, common-base DHBT yields a much higher gain than common-emitter DHBT at a high frequency, especially near the cutoff frequency, because the maximum stable gain (MSG) of the common-base DHBT extends into higher frequencies than that of a common-emitter DHBT[5]. In addition, to increase the output power, the breakdown voltage, the current and the emitter area should be as high as possible. Considering these above factors, a common-base four-finger InGaAs/InP DHBT has been demonstrated in this paper. The InGaAs/InP DHBTs were fabricated with a triple mesa process and a benzocyclobutene (BCB) planarization technique. All processes were carried out on 3-inch wafers. The area of each emitter finger is 1
2. Growth and fabrication
The layer structure of the InGaAs/InP DHBTs was grown by molecular-beam epitaxy on a 3-inch semi-insulating InP substrate. The layer sequence is shown in Fig. 1. The DHBT structure includes an InGaAs cap layer (200 nm, 3
The geometry parameters of the devices are similar to those of Ref. [4]. In contrast to most recent reports in China[7-9], the InP DHBTs in this work were designed and fabricated with standard manufacturing techniques such as i-line stepper lithography, self-aligned contact and selective dry/wet etching, etc. All InP DHBT processes were on 3-inch wafers. The InP DHBTs were fabricated with conventional wet etching and metal deposition with triple mesa design. Non-alloyed ohmic Ti/Pt/Au was used as the n-type ohmic contact and Pt/Ti/Pt/Au was used as the p-type contact. After device isolation, BCB was used for device passivation and planarization. Subsequently, an RIE etch-back step was performed to expose the tops of the device contacts and then the first-level metal was deposited to form the probe pads.
3. Measurements and results
The InP DHBTs were measured on-wafer at room temperature. The DC characteristics of the DHBTs were measured by an Agilent 1500A semiconductor parameter analyzer. The common-base
The microwave performance of the fabricated InP DHBTs was characterized by on wafer
U=|Y21−Y12|2/[4(G11G22−G21G12)], |
(1) |
where
4. Conclusion
In summary, a common-base four-finger InGaAs/InP DHBT has been successfully demonstrated with standard manufacturing techniques on 3-inch wafers. The area of each finger is 1