Expand this Topic clickable element to expand a topic
Skip to content
Optica Publishing Group

Enhancement in speed and responsivity of uni-traveling carrier photodiodes with GaAs0.5Sb0.5/In0.53Ga0.47As type-II hybrid absorbers

Open Access Open Access

Abstract

We demonstrate a top-illuminated high-speed uni-traveling carrier photodiode (UTC-PD) with a novel design in the p-type absorber, which can effectively shorten the photon absorption depth at telecommunication wavelengths (1.31~1.55 μm) and further enhance the bandwidth-efficiency product of UTC-PD. In our proposed new UTC-PD structure, the p-type In0.53Ga0.47As absorption layer is replaced by the type-II GaAs0.5Sb0.5 (p)/In0.53Ga0.47As (i) hybrid absorber. Due to the narrowing of the bandgap and enhancement of the photo-absorption process at the type-II interface between the GaAs0.5Sb0.5 and In0.53Ga0.47As layers, our device shows an over 16.7% improvement in the responsivity compared with that of UTC-PD with the same thickness of pure In0.53Ga0.47As absorber (0.7 μm) and a zero optical coupling loss. Our demonstrated device with a simple top-illuminated structure offers a large active mesa (25 μm), a wide optical-to-electrical (O-E) bandwidth (33 GHz), a high responsivity (0.7 A/W), and a high saturation current (>5 mA) under 1.31 µm optical wavelength. These promising results suggest that our proposed PD structure can fundamentally overcome the trade-off among bandwidth, efficiency, and device active diameter of high-speed PDs.

© 2019 Optical Society of America under the terms of the OSA Open Access Publishing Agreement

1. Introduction

The progress of data rate per channel in the optical interconnect (OI) system drives the increase of required bandwidth for both transmitter and receiver. In the newly developed 400 Gbit/sec Ethernet system with 4 fiber channels, the corresponding data rate per channel is 112 Gbit/sec. In order to attain such high data rate, one of the possible solutions is utilizing the pulse-amplitude modulation-4 (PAM-4) format with a 56 Gbaud per channel. A high-responsivity p-i-n photodiode operated at 1.31 µm wavelength with a 3-dB optical-to-electrical (O-E) bandwidth over 30 GHz and a large aperture size, serves as the key component in the receiving-end of the aforementioned Ethernet system. Commercially available 10 Gbit/sec InP based PD chips at 1.31-1.55 µm wavelength typically have around 50 μm active diameters, providing the sufficient alignment tolerance for ROSA package. However, when the required bandwidth of PD increases, a decrease of the p-i-n PD’s active diameter and a reduction of absorption layer thickness are both necessary in order to balance the trade-off between RC-limited and carrier transient time limited bandwidths [1–3]. This leads to penalties of smaller responsivity, larger coupling loss and/or tighter alignment tolerances, and higher packaging costs in ROSA module.

By use of the waveguide [1,4–6] or edge-coupled PD structures [7], the trade-off between bandwidth and efficiency (responsivity) in high-speed PD can be overcome. In the waveguide PD structure (WGPD) with a thin absorption layer thickness, the high-efficiency performance can be sustained by properly increasing the device absorption length. Furthermore, the thin absorber can also lead to a wide internal carrier transient time limited bandwidth. WGPD [1,4–6] and refracting-facet PD [7] with bandwidth-efficiency product as high as 50 GHz have been previously demonstrated. Those WGPD devices exhibited 3-dB O-E bandwidth (BW) of 50 GHz and quantum efficiency (QE) of 80% or BW of 110 GHz and QE of 50%. However, the WGPD structure usually has a much smaller alignment tolerance and higher fabrication cost, compared to the top-illuminated PD structure. To fundamentally overcome the trade-offs among bandwidth, efficiency, and coupling alignment of PD, adopting the photo-absorption layer with an absorption coefficient larger than that of the bulk In0.53Ga0.47As material [2,3], appears to be one of the most effective approaches over the WGPD design. The increase of photo-absorption coefficient leads to the decrease in the required absorption layer thickness for the desired responsivity performance and the enhancement of carrier transient time limited bandwidth.

In our previous studies [2,3], we have demonstrated that the InP/In0.53Ga0.47As-based PDs can achieve superior performance, including higher bandwidth-efficiency product and the enlarged device active area, compared to those of the GaAs-based ones for 850 nm wavelength. The performance advantage of In0.53Ga0.47As-based PD is attributed to the larger absorption coefficient of the In0.53Ga0.47As layer with respect to the GaAs (~3 vs. ~0.8 μm−1) at such wavelength. On the other hand, for the operations at telecommunication wavelengths (1.31-1.55 μm), in order to have the photo-absorption layer, such as In0.8Ga0.2As, with a narrower bandgap (0.5 vs. 0.75 eV) than that of bulk In0.53Ga0.47As material, the thick lattice-mismatched buffer layer grown on the InP substrate is necessary [8,9]. However, this leads to increased dark current and reliability issues of fabricated PDs. In this paper, we demonstrate a top-illuminated high-speed uni-traveling carrier photodiode (UTC-PD) with a novel design in the lattice-matched p-type absorber with InP substrate, which can effectively shorten the photon absorption depth at 1.31~1.55 μm wavelength and further enhance the bandwidth-efficiency product of UTC-PD. In our proposed new UTC-PD structure, the p-type In0.53Ga0.47As absorption layer is replaced by the type-II GaAs0.5Sb0.5 (p)/In0.53Ga0.47As (i) hybrid absorber. Owing to the narrowing of the bandgap and enhancement of the photo-absorption process at the type-II interface between the GaAs0.5Sb0.5 and In0.53Ga0.47As layers [10,11], our device shows an over 16.7% improvement in the responsivity as compared to that of UTC-PD with the pure In0.53Ga0.47As absorber and the same total thickness (0.7 μm). Furthermore, our demonstrated device with a simple top-illuminated structure can offer a large active mesa (25 μm), a wide optical-to-electrical (O-E) bandwidth (33 GHz), a high responsivity (0.7 A/W), and a high saturation current (> 5 mA) under 1.31 µm optical wavelength. These excellent performances are comparable with those of high-performance 56 Gbaud PD [12] with the same top-illuminated structure and a smaller mesa size (~16 µm). This suggests that our proposed PD structure can fundamentally overcome the trade-off among bandwidth, efficiency, and device active diameter of high-speed PDs.

2. Design of device structure and fabrication

Figure 1(a) shows the simulated band diagram of our demonstrated device structure. Our epilayer structure was grown on the semi-insulating (S.I.) InP substrate in a molecular-beam epitaxy (MBE) chamber [13]. From top to bottom, it is composed of the p+-In0.53Ga0.47As contact layer, p+-Al0.3Ga0.7As0.5Sb0.5 electron blocking layer, p-type hybrid absorber, InP collector layer, and n+ InP contact layer. The detailed material composition, thickness, and doping density of each layer are given in Table 1. Here, the hybrid absorption region with a type-II band alignment (GaAs0.5Sb0.5/In0.53Ga0.47As) is comprised of two major parts. The first is the p-type GaSb0.5As0.5 absorption layer with a thickness of 300 nm and a graded doping profile (top: 1 × 1019 cm−3 to bottom: 5 × 1016 cm−3) designed to induce built-in electric (E-) field and accelerate the electron diffusion process. As shown in Fig. 1(a), we can clearly see that there is a small built-in E-field in this GaAs0.5Sb0.5 absorption layer due to the graded doping profile. The second is an intrinsic In0.53Ga0.47As layer with a thickness of 400 nm. As compared to the device structure reported in our previous work [11], which is targeted on THz O-E bandwidth, the thickness of type-II hybrid absorption layer in this work is greatly increased from 160 to 700 nm in order to enhance the responsivity performance for around 30 GHz O-E bandwidth. The un-doped In0.53Ga0.47As layer is inserted in the p-type absorption region of the traditional UTC-PD to minimize the electron recombination process. Although such an intrinsic layer may contribute to undesired hole transport, the corresponding transient time limited bandwidth of the holes is still as high as around 55 GHz, larger than the measured net O-E bandwidth of our device (around 33 GHz), which will be discussed in greater detail later. As shown in Fig. 1(a), the type-II band alignment between the interface of the GaAs0.5Sb0.5 and In0.53Ga0.47As layers reduces the effective bandgap to 0.5 eV (~2.4 µm cut-off wavelength [10]), which can further enhance the absorption process under 1.3 to 1.55 µm wavelengths excitation. Our collector layer consists of an intrinsic InP layer with an ultra-low background doping density in order to reduce the required bias voltage for the full depletion in collector layer. In addition, its thickness (1.2 μm) is optimized to balance the RC-limited and internal transient time limited bandwidths for a 30 μm diameter of active mesa. Figure 1(b) shows the picture of top-view of demonstrated PD. The fabricated device has a typical top-illuminated structure with a 25 µm diameter of active mesa and is integrated with a coplanar ground-signal (GS) pads for on-wafer high-speed measurement. An anti-reflection coating film, which is composed of Si3N4 with an around 200 nm thickness, is deposited on top of our device to minimize the reflection loss of the launched optical signal at 1.31 or 1.55 µm wavelength.

 figure: Fig. 1

Fig. 1 (a). Simulated band diagram of the demonstrated hybrid-absorber UTC-PDs under −2 V bias voltage. The unit for doping density in each layer is cm−3. (b) Photo of top-view of the fabricated devices.

Download Full Size | PDF

Tables Icon

Table 1. Epitaxy structures of demonstrated devices

3. Measurement results

Here our PDs of Device-A and Device-B were fabricated with the active diameters of 25 and 30 µm, respectively. The corresponding aperture size of optical window is 15 and 20 µm, respectively. The measured dark currents of device A and B under −5 V bias is in the range of 90-110 nA. Under −2V bias, the measured DC responsivity of our demonstrated PD at 1.31 and 1.55 µm wavelength is around 0.7 and 0.53 A/W, respectively. The reported absorption coefficients of bulk GaAs0.5Sb0.5 layers (lattice-matched on InP substrate) at 1.31 and 1.55 µm wavelength are 0.9 and 0.68 µm−1, respectively [14]. They are comparable with those of bulk In0.53Ga0.47As layer (1.2 and 0.68 µm−1 at 1.31 and 1.55 µm wavelength, respectively) [14]. Based on these reported absorption coefficients and the assumption of single-pass optical path in the absorber, the theoretical maximum responsivity is 0.56 (0.47) A/W for our device with a zero optical coupling/reflection loss and a 0.7 µm absorption layer thickness at the 1.31(1.55) µm wavelength. The validness of single-pass assumption is due to that the reflected optical beam from substrate side is far from the absorption region, which would suffer significant diffraction loss and can’t contribute to the photocurrent. Such measurement result clearly indicates that the type-II interface in our hybrid absorber can contribute to around 26 and 12% enhancements in the DC responsivity under 1.31 and 1.55 µm wavelength excitations, respectively, due to its narrower bandgap in the type-II interface as discussed. Furthermore, as compared to the PD with a pure In0.53Ga0.47As absorber and the same total thickness (0.7 µm), our device can still have >16.7% enhancement in responsivity under the 1.31 µm wavelength excitation and zero optical coupling loss. Figures 2 and 3 show the measured bias-dependent O-E frequency responses of Device-A and -B under low (0.1 mA) and high (1 mA) output photocurrents at 1.31 µm wavelength, respectively. We can clearly see that when the reverse bias is over −2 V, the measured 3-dB O-E bandwidths of Device-A and -B are around 33 and 28 GHz, respectively. Furthermore, regardless of high or low output current, the measured 3-dB O-E bandwidths of both devices remain unchanged. This measurement result clearly indicates the good linearity of our device, which is a critical issue in the performance of PAM-4 receiver [15].

 figure: Fig. 2

Fig. 2 Bias dependent O-E frequency responses measured under different output photocurrents: (a) 0.1 mA and (b) 1 mA for Device A with a 25 µm active diameter.

Download Full Size | PDF

 figure: Fig. 3

Fig. 3 Bias dependent O-E frequency responses measured under different output photocurrents: (a) 0.1 mA and (b) 1 mA for Device B with a 30 µm active diameter.

Download Full Size | PDF

The overall O-E 3-dB bandwidth (f3dB) of a PD is determined by the carrier transport time (1ft) and the RC time constant (1fRC). In order to investigate the internal carrier transport time inside our device, the following equation has been adopted [1,2].

1f3dB2=1fRC2+1ft2=(2πRC)2+1ft2

Where R is the sum of the parasitic resistance and the load resistance (50 Ω) and C is the total capacitance. Here, the RC-limited bandwidth can be extracted by use of the measured scattering parameters of microwave reflection coefficients (S22) [2]. Figure 4 shows the adopted equivalent circuit models for fitting of S22 parameters of devices A and B. The fitted values of each circuit elements of devices A and B are shown in the Table 2. During our device modeling process for the extraction of extrinsic fRC of PD chips, the two artificial circuit elements; RT and CT, have been removed. This is because that they are used to mimic the low-pass frequency response of internal carrier transient time [16,17]. Figure 5(a) and (b) shows the fitted and measured frequency responses of S22 parameters on Smith Chart of Device- A and B, respectively. Clearly, there is a good match between the simulated and measured results from 40 MHz to 40 GHz. By use of the extracted equivalent circuit model and choose the proper values of RT and CT, as shown in Table 2, to fit the measured O-E frequency response, we can then determine the internal transient time limited frequency responses of our device.

 figure: Fig. 4

Fig. 4 Equivalent-circuit-model. VCCS: voltage controlled current source.

Download Full Size | PDF

Tables Icon

Table 2. Values of the circuit elements

 figure: Fig. 5

Fig. 5 Measured (blue line) and fitted (red line) S22 parameters from near dc to 40 GHz under a fixed dc bias (−2 V) for (a) Device-A and (b) Device-B.

Download Full Size | PDF

Figure 6(a) and (b) shows the RC-limited, measured O-E, fitted O-E, and transient time limited frequency responses of Device-A and -B, respectively. We note that Device-B exhibited the RC-limited bandwidth is close to the transient-time-limited value of around 35 GHz. This indicates that our collector layer thickness has been optimized to balance the ftandfRC. The other way to extract the internal carrier transient time is based on Eq. (1) as discussed. Figure 6(c) shows the (1fRC)2vs. (103f3dB2)values for Devices A and B. With Eq. (1), the internal carrier transient time in our device is determined using the intercepts shown along the y-axis of this Fig [1,2]. The obtained ftis around 33.5 GHz. Such number is consistent with the transient time limited bandwidth (around 35 GHz) obtained by the equivalent circuit modeling technique, as shown in Figs. 6(a) and (b). The drift-diffusion model is used to estimate the transient time and corresponding electron/hole drift-velocity across each layers [18]. The extracted electron drift-velocity through the 1.2 µm InP collector layer is around 11×105 m/sec. The dominating factor for the total transient time inside our proposed epi-layer structure should be the drift-time of the photo-generated holes through the depleted 400 nm thick In0.53Ga0.47As absorption layer. The average drift-velocity of electrons and holes across this layer is around 5×104 m/sec, which is close to the value reported (5.3×104 m/sec) for In0.53Ga0.47As p-i-n PDs [1]. The electron minority mobility in the p-type GaAs0.5Sb0.5 layer with a graded doping profile is 4000 cm2/V-sec. This number is pretty close to that in the p-type In0.53Ga0.47As absorption layer of UTC-PD (~5000 cm2/V-sec) [19], which indicates the good material quality of our thick (300 nm) GaAs0.5Sb0.5 layer.

 figure: Fig. 6

Fig. 6 The measured O-E, extracted RC-limited, transient time, and fitted O-E frequency responses of devices (a) A and (b) B. (c) The extracted (1fRC)2 versus the measured (103f3dB2) for devices A and B.

Download Full Size | PDF

The PDs with high-linearity performance is required for PAM-4 modulation format. The two-laser heterodyne-beating setup at 1.55 µm wavelengths is used to test the output saturation current of our demonstrated PD [5]. Figure 7(a) and (b) shows the measured photo-generated microwave power versus output photocurrent for Device-A and -B obtained under different reverse biases (−2 to −5V). The measurement frequency chosen was around the 3-dB O-E bandwidth at 30 GHz. The ideal relation between the microwave power and averaged photocurrent (solid line) with a 100% optical modulation depth under a 50 Ω load is also plotted for reference. We can clearly see that the saturation currents of Device-A and -B under −5 V bias are around 5 and 13 mA, respectively. Such high output current can far exceed the required overload optical power ( + 4.5 dBm; around 2 mA photocurrent) for the 400 Gbit/sec receiver with PAM-4 modulation format. Based on the aforementioned dynamic and static measurements results, our demonstrated PD (Device A) with a larger active mesa (25 vs. 16 µm) can achieve a comparable 3-dB O-E bandwidth (~33 vs. 38 GHz), responsivity (0.7 vs. 0.8 A/W), and a higher saturation current as compared to those of high-performance and commercial available p-i-n PD with the same top-illuminated structure for 56 Gbaud application [12].

 figure: Fig. 7

Fig. 7 The measured photo-generated microwave power versus photocurrent for device (a) A and (b) B under sinusoidal signal excitation and with different reverse biases at operating frequencies of 30 GHz. The open symbol line shows the ideal trace for a 100% optical modulation depth and 50 Ω load.

Download Full Size | PDF

4. Conclusion

In conclusion, we demonstrate the UTC-PD structure with a novel design in its hybrid absorber. Using the narrow bandgap (~0.5 eV) of the type-II interface between p-type GaAs0.5Sb0.5 and intrinsic In0.53Ga0.47As absorbers, we can achieve an enhanced responsivity by over 16.7% (at 1.31 µm wavelength) as compared to that of the PD based on a pure In0.53Ga0.47As absorber with the same total thickness (700 nm) of absorption region, and a zero optical coupling loss. Our demonstrated device with a simple top-illuminated structure and a large active mesa (25 µm) can provide a wide optical-to-electrical (O-E) bandwidth as 33 GHz, high responsivity (0.7 A/W), and a high saturation current (>5 mA) under 1.31 µm optical wavelength. These promising results suggest that our proposed hybrid absorber structure can fundamentally overcome the trade-off among bandwidth, efficiency, and device active diameter of high-speed PDs due to its stronger photo-absorption process at telecommunication wavelengths (1.3 to 1.55 µm) than that of bulk In0.53Ga0.47As layer. In our future work, we will experiment with the packaging of such novel PDs with high-performance trans-impedance amplifiers to study the sensitivity performance for 56 Gbaud applications.

Funding

Ministry of Science and Technology in Taiwan (107-2622-E-008 −002 -CC2 and 106-2221-E-008 −063 -MY3).

References

1. K. Kato, “Ultrawide-band/high-frequency photodetectors,” IEEE Trans. Microw. Theory Tech. 47(7), 1265–1281 (1999). [CrossRef]  

2. J.-W. Shi, K.-L. Chi, C.-Y. Li, and J.-M. Wun, “Dynamic analysis of high-efficiency InP based photodiode for 40 Gbit/sec optical interconnect across a wide optical window (0.85 to 1.55 μm),” IEEE/OSA J. Lightwave Technol. 33(4), 921–927 (2015). [CrossRef]  

3. J.-W. Shi, C.-Y. Li, K.-L. Chi, J.-M. Wun, Y.-M. Hsin, and S. D. Benjamin, “Large-area p-i-n photodiode with high-speed and high-efficiency across a wide optical operation window (0.85 to 1.55 μm),” IEEE J. Sel. Top. Quantum Electron. 20(6), 3800807 (2014).

4. S. Demiguel, N. Li, X. Li, X. Zheng, J. Kim, J. C. Campbell, H. Lu, and A. Anselm, “Very high-responsivity evanescently coupled photodiodes integrating a short planar multimode waveguide for high-speed applications,” IEEE Photonics Technol. Lett. 15(12), 1761–1763 (2003). [CrossRef]  

5. J.-W. Shi, C.-Y. Wu, Y.-S. Wu, P.-H. Chiu, and C.-C. Hong, “High-speed, high-responsivity, and high-power performance of near-ballistic uni-traveling-carrier photodiode at 1.55μm wavelength,” IEEE Photonics Technol. Lett. 17(9), 1929–1931 (2005). [CrossRef]  

6. X. Xie, Q. Zhou, E. Norberg, M. Jacob-Mitos, Y. Chen, Z. Yang, A. Ramaswamy, G. Fish, J. C. Campbell, and A. Beling, “High-power and high-speed heterogeneously integrated waveguide-coupled photodiodes on silicon-on-insulator,” IEEE/OSA J. Lightwave Technol. 34(1), 73–78 (2016). [CrossRef]  

7. Y. Muramoto, H. Fukano, and T. Furuta, “A polarization-independent refracting-facet uni-traveling-carrier photodiode with high efficiency and large bandwidth,” IEEE/OSA J. Lightwave Technol. 24(10), 3830–3834 (2006). [CrossRef]  

8. A. Joshi and S. Datta, “High-speed, large-area, p-i-n InGaAs photodiode linear array at 2-micron wavelength,” Proc. SPIE 8353, 83533D (2012).

9. N. Ye, H. Yang, M. Gleeson, N. Pavarelli, H. Zhang, J. O’Callaghan, W. Han, N. Nudds, S. Collins, A. Gocalinska, and E. Pelucchi, “InGaAs surface normal photodiode for 2 μm optical communication systems,” IEEE Photonics Technol. Lett. 27(14), 1469–1472 (2015). [CrossRef]  

10. R. Sidhu, L. Zhang, N. Tan, N. Duan, J. C. Campbell, A. L. Holmes, D.-F. Hsu, and M. A. Itzler, “2.4 μm cutoff wavelength avalanche photodiode on InP substrate,” Electron. Lett. 42(3), 181–182 (2006). [CrossRef]  

11. J.-M. Wun, Y.-W. Wang, and J.-W. Shi, “Ultra-fast uni-traveling carrier photodiodes with GaAs0.5Sb0.5/In0.53Ga0.47As type-II hybrid absorbers for high-power operation at THz frequencies,” IEEE J. Sel. Top. Quantum Electron. 24(2), 8500207 (2018). [CrossRef]  

12. Albis Optoelectronics AG, “PD40C1: 56 Gbaud Photodiode with Enhanced Responsivity,” http://www.albisopto.com/albis_product/pd40c1-56-gbaud-photodiode-with-enhanced-responsivity/.

13. Intelligent Epitaxy Technology, Inc., “IntelliEPI,” http://intelliepi.com.

14. M. S. Park and J. H. Jang, “GaAs0.5Sb0.5 lattice matched to InP for 1.55 μm photo-detection,” Electron. Lett. 44(8), 549–551 (2008). [CrossRef]  

15. M. Nada, Y. Muramoto, H. Yokoyama, and H. Matsuzaki, “High-speed high-power-tolerant avalanche photodiode for 100-Gb/s applications,” in Proceedings of IEEE Photonic Society Meeting (IEEE, 2014) pp. 172–173. [CrossRef]  

16. G. Wang, T. Tokumitsu, I. Hanawa, Y. Yoneda, K. Sato, and M. Kobayashi, “A time-delay equivalent-circuit model of ultrafast p-i-n photodiodes,” IEEE Trans. Microw. Theory Tech. 51(4), 1227–1233 (2003). [CrossRef]  

17. Y.-S. Wu, J.-W. Shi, and P.-H. Chiu, “Analytical modeling of a high-performance near-ballistic uni-traveling-carrier photodiode at a 1.55μm wavelength,” IEEE Photonics Technol. Lett. 18(8), 938–940 (2006).

18. J.-W. Shi, C.-W. Liu, and C.-W. Liu, “Design and analysis of separate-absorption-transport-charge-multiplication traveling-wave avalanche photodetectors,” IEEE/OSA J. Lightwave Technol. 22(6), 1583–1590 (2004). [CrossRef]  

19. N. Shimizu, N. Watanabe, T. Furuta, and T. Ishibashi, “InP-InGaAs uni-traveling-carrier photodiode with improved 3-dB bandwidth of over 150GHz,” IEEE Photonics Technol. Lett. 10(3), 412–414 (1998). [CrossRef]  

Cited By

Optica participates in Crossref's Cited-By Linking service. Citing articles from Optica Publishing Group journals and other participating publishers are listed here.

Alert me when this article is cited.


Figures (7)

Fig. 1
Fig. 1 (a). Simulated band diagram of the demonstrated hybrid-absorber UTC-PDs under −2 V bias voltage. The unit for doping density in each layer is cm−3. (b) Photo of top-view of the fabricated devices.
Fig. 2
Fig. 2 Bias dependent O-E frequency responses measured under different output photocurrents: (a) 0.1 mA and (b) 1 mA for Device A with a 25 µm active diameter.
Fig. 3
Fig. 3 Bias dependent O-E frequency responses measured under different output photocurrents: (a) 0.1 mA and (b) 1 mA for Device B with a 30 µm active diameter.
Fig. 4
Fig. 4 Equivalent-circuit-model. VCCS: voltage controlled current source.
Fig. 5
Fig. 5 Measured (blue line) and fitted (red line) S22 parameters from near dc to 40 GHz under a fixed dc bias (−2 V) for (a) Device-A and (b) Device-B.
Fig. 6
Fig. 6 The measured O-E, extracted RC-limited, transient time, and fitted O-E frequency responses of devices (a) A and (b) B. (c) The extracted ( 1 f RC ) 2 versus the measured ( 10 3 f 3dB 2 ) for devices A and B.
Fig. 7
Fig. 7 The measured photo-generated microwave power versus photocurrent for device (a) A and (b) B under sinusoidal signal excitation and with different reverse biases at operating frequencies of 30 GHz. The open symbol line shows the ideal trace for a 100% optical modulation depth and 50 Ω load.

Tables (2)

Tables Icon

Table 1 Epitaxy structures of demonstrated devices

Tables Icon

Table 2 Values of the circuit elements

Equations (1)

Equations on this page are rendered with MathJax. Learn more.

1 f 3dB 2 = 1 f RC 2 + 1 f t 2 = ( 2πRC ) 2 + 1 f t 2
Select as filters


Select Topics Cancel
© Copyright 2024 | Optica Publishing Group. All rights reserved, including rights for text and data mining and training of artificial technologies or similar technologies.