We report the first demonstration of a terahertz photomixer made of ion-irradiated In0.53Ga0.47As lattice-matched to InP and fiber-optic coupled with the drive lasers. A continuous-wave radiation is generated at frequencies up to 0.8 THz by photomixing two continuous-wave laser diodes around 1.55 μm. The measured 3dB-down bandwidth of 300 GHz yields a carrier lifetime of 0.53 ps, in agreement with the value of 0.41 ps measured in pump probe experiments. The detected signal is at the most 15 dB lower than the one obtained from similar photomixers fabricated from low-temperature-grown GaAs.
©2006 Optical Society of America
Photomixing is a useful technique for the generation of continuous high-purity terahertz (THz) radiation, which entails the pumping of an ultrafast optoelectronic device with two frequency-offset lasers. Photomixers can operate at room temperature with high frequency tunability. They can be used either as local oscillators in THz receivers or as swept coherent sources in THz spectroscopy. The most common photomixers consist of low-temperature-grown (LTG) GaAs [1, 2] or ErAs:GaAs  photoconductors that operate at short wavelengths (λ ~ 0.8 μm). High output power and high-frequency performances (up to 5 THz) in these cases are achieved thanks to subpicosecond carrier lifetimes and good photo-carrier mobilities (-200 cm2 V-1s-1). An LT-In0.3Ga0.7As continuous-wave terahertz photomixing system using ~ 0.8 μm optical excitation wavelengths was also recently demonstrated. A 15 dB discrepancy power compared to LTG-GaAs was obtained . The authors expect an improvement in THz power with a 1.06 μm excitation closest to the band edge. The challenge is to extend the operating wavelengths to the telecommunication wavelength domain in such a way as to take advantage of the compact and low cost optical amplifiers as well as of the wavelength-tunable laser diodes and of the superior fiber technology available at 1.55 μm. For this purpose, novel devices are required since the LTG-GaAs gap energy does not fit with the 1.55 μm wavelength. To date, THz generation from photomixing at 1.55 μm has only been reported in uni-travelling-carrier photodiodes , in ErAs: In0.53Ga0.47As photoconductors . An output-power of 2.6 μW was achieved at 1 THz in the first work and a frequency bandwidth of 500 GHz was reached in the second work with an output power of 10 nW. In this paper, we propose an alternative way for photomixers at 1.55 μm using heavy ion-irradiated In0.53Ga0.47As photoconductors. Heavy ion irradiation has been previously shown to reduce the carrier lifetime down to sub-picosecond values (< 500 fs) in In0.53Ga0.47As  while preserving good electrical properties of the material.
2. Samples and DC characteristics
Undoped 1-μm-thick n-type In0.53Ga0.47As layers were epitaxially grown by gas-source MBE on semi-insulating InP:Fe substrates. A mesa etching process was used to define an In0.53Ga0.47As absorbing area of 7 × 7 μm2 on the InP substrate. The layers were then bombarded by heavy ions (Br+) of high initial energy (11 MeV) at 4 × 1011 cm-2. Titanium and gold films of 200 Å and 3500 Å thickness, respectively, were deposited onto the In0.53Ga0.47As surface after patterning in polymethyl methacrylate to form four 0.2-μm-wide interdigitated electrodes with gap spacing of 1.8 μm and a composite spiral-dipole antenna. The quasi-static capacitance of the interdigitated-electrode structure was calculated to be 1 fF . An additional pad capacitance of 1 fF has to be considered for the total capacitance. The spiral antenna is a broadband antenna whose radiation resistance is given by: RA =60π/(εeff )1/2 for frequencies higher than 100 GHz. In the previous formula, εeff )1/2 is an effective dielectric constant given by (1 + ε)/2, where ε is the relative permittivity of the semiconductor substrate. For InP, ε=12.35 and thus RA = 73 Ω.
The dark current-voltage characteristic of the fabricated device is represented in Fig. 1. It shows a linear evolution with bias voltages up to 2.6 volts. The dark resistivity of the device was deduced to be ρ ~ 1.9 Ω.cm from Hall measurements. As seen in Fig. 1, the dark current increases superlinearly above 2.6 V. The maximum attainable bias voltage was ~ 3.1 V, which corresponds to a breakdown field of ~ 0.17 × 105 V/cm for a 1.8 μm gap between electrodes. This value is one order of magnitude lower than typical breakdown fields for non-irradiated In0.53Ga0.47As layers. An electron lifetime of 0.41 ps was measured for the irradiated samples in pump-probe experiments. The static mobility of electrons, measured in Hall experiments, was found to be 490 cm2.V-1.s-1. The dc responsivity at 1.55 μm measured versus bias voltage is represented in Fig. 1 for an optical power of 30 mW. Two conduction regimes appear. The first regime extending from zero bias to 1V is linear as expected for an ohmic behavior. The second regime at higher voltages is quadratic. This second regime is attributed to space-charge effects that arise under high injection conditions when the space charge in the electric field profile is controlled by injected carriers . The device responsivity reaches 0.031 A/W at 1.7 V.
3. Experimental set-up
Two laser diodes operating near 1550 nm were used to characterize the photomixer performances. Both lasers were operated at room temperature, one of them being frequency tunable by steps of 12.5 MHz. The laser beams were coupled into a single-mode fiber-optic combiner, and the total optical power was amplified in an erbium-doped amplifier whose maximum available output power was 40 mW. The single-mode fiber output was then focused onto the active In0.53Ga0.47As layer. The optical beam polarization was adjusted in such a way as to maximize the device photocurrent. A 210 Hz square modulation voltage was applied to the photomixer. The ac current generated by photomixing fed the dipole-spiral antenna, which was coupled to free space via a hemispherical Si lens. The radiated power was focused onto a 4.2 K Si composite bolometer with a f/50 paraboloidal mirror. A polytetra-fluoro-ethylene plate was placed in front of the bolometer to block the 1.55 μm signal. A synchronous detection scheme was used to measure the amplified output of the bolometer. The full width at half maximum of the beat signal was preliminary measured at 5 GHz by focusing the two laser beams onto a fast photodiode connected to a rf spectrum analyzer. The linewidth was smaller than 3 MHz, but the central frequency was found to fluctuate around 10 MHz for 50 ms recording time.
Figure 2 shows the bolometer output voltage measured as a function of the frequency difference between the two lasers. Two different bias voltages (1.2 and 1.7 V) are applied for an optical power of 15 mW at each laser. For comparison, the THz power detected from an identical device fabricated from LTG-GaAs grown on semi-insulating GaAs and excited by 800 nm optical excitation wavelengths is also presented in Fig. 2. The optical excitation power is 35 mW and the applied bias is 5 volts.
The detected maxima from ion-irradiated-In0.53Ga0.47As are 1.4 and 2.64 mV for the lower and higher photomixer biases respectively. The sensitivity of the bolometer coupled with an amplifier is estimated to be 12.1 × 105 V/W. The maximum detected powers ion-irradiated-In0.53Ga0.47As are then 1.6 and 3.1 nW for the 1.2 and 1.7 V bias voltages respectively. Following the work of Brown et al., the mean power of the terahertz radiation is expressed as:
where Vb is the photomixer bias voltage, G 0 ∝P 0 is the mean photoconductance, P 0 the total optical power and t the carrier lifetime. It is worth noticing that in agreement with Eq. (1), these maximum powers are in the same ratio as the squares of the photomixer voltages. From 190 to 800 GHz, the detected power of ion-irradiated In0.53Ga0.47As drops with a slope around 6 dB/octave. This roll-off is attributed only to a limitation by the carrier lifetime. Indeed, the 3-dB-down frequency is determined to be 300 GHz. The RAC time constant cannot be incriminated in this frequency roll-off since it is theoretically ~ 0.14 ps. The measured 3-dB-down frequency yields a carrier lifetime of 0.53 ps, which value is close to that measured in pump-probe experiments (0.41 ps). This result also demonstrates that the hole lifetime in heavy ion-irradiated In0.53Ga0.47As can be reduced down to sub-picosecond values as the electron lifetime. As far as we know, the hole lifetime has not been reported yet for ion-irradiated In0.53Ga0.47As. Our measurements confirm that ion irradiation produces efficient recombination centers for both types of free carriers. A plateau is then observed between 600 and 800 GHz, whose origin is not presently understood. Around 1 THz, the detected power rolls off dramatically for both devices and we attribute this to water absorption.
The 3-dB-down frequency of the LTG-GaAs photomixer of 600 GHz is higher than the one of irradiated-In0.53Ga0.47As photomixer. This is the result of the shorter carrier lifetime estimated to 0.3 ps in LTG-GaAs. Similarly, the RAC time constant cannot be incriminated in this frequency roll-off since it is equal to ~ 0.1 ps. Across the entire frequency range, the detected signal from ion-irradiated-In0.53Ga0.47As device is at the most 15 dB lower than LTG-GaAs one. This difference can be explained in part by the higher voltage applied to the LTG-GaAs photomixer as the radiated power depends quadratically with the bias voltage. The 1.7 volts is close to the maximum voltage which can be applied to the In0.53Ga0.47As photomixer without device failure. This weak breakdown field value is the result of the relatively low resistivity inherent to small band gap materials.
Figure 3 (left) shows the evolution of the detected power at 0.5 THz measured versus the photomixer bias voltage for a total incident power of 30 mW. The detected power varies as the square of the bias voltage as predicted from Eq. (1). The dashed line is a parabolic fit (∝). No saturation of the detected power is observed up to 1.7 V, thereby indicating that the carrier lifetime does not vary within the range of bias voltages. Figure 3 (right) shows the detected power at 0.5 THz measured as a function of the total optical power, which is equally distributed amongst the two laser frequencies. The bias voltage is 1.2 V. Once more, the detected power exhibits a quadratic dependence in agreement with the photomixing theory for identical laser powers. This result indicates that physical parameters such as the carrier mobility are not affected by the high optical power dissipated into the device.
Experiments reported in this work are the first demonstration of THz generation from photomixers based on heavy ion-irradiated In0.53Ga0.47As. However, the powers emitted are around 30 dB lower than the best achieved with LTG-GaAs photomixers . So, further improvements are needed to reach the required powers for practical applications. First, annealing stages should increase the dark resistivity of ion-irradiated layers  and higher breakdown field can then be expected. Second, the use of an antireflection coating at the top surface and of a distributed Bragg reflector beneath the active region should strongly enhance the photo mixer responsivity. Finally, further improvements can be expected from a vertical integration of the photoconductive detector  and a good thermal management of the whole structure .
In conclusion, a room temperature THz source has been reported for the first time using the 1.55 μm technology with heavy ion-irradiated In0.53Ga0.47As photo mixers. The measured evolutions of the detected THz power versus different system parameters have been found in perfect agreement with photo mixing theory. We found that the detected signal is at the most 15 dB lower than the one obtained from similar photomixers fabricated from low-temperature-grown GaAs. Different solutions have been proposed for further improvements.
The authors would like to thank S. Henry from CSNSM in Orsay for ion irradiation. This work has been carried out in the frame of the french RTB network.
References and links
1. E. R. Brown, F. W. Smith, and K. A. McIntosh, “Coherent millimetre-wave generation by heterodyne conversion in low-temperature-grown GaAs photoconductors,” Appl. Phys. Lett. 73, 1480 (1993).
2. E. R. Brown, K. A. McIntosh, K. B. Nichols, and C. L. Dennis, “Photomiwing up to 3.8 THz in low-temperature GaAs,” Appl. Phys. Lett. 66, 285 (1995). [CrossRef]
3. J. E. Bjarnason, T. L. J. Chan, A. W. M. Lee, E. R. Brown, D. C. Driscoll, M. Hanson, A. C. Gossard, and R. E. Muller, “ErAs:GaAs photomixer with two-decade tenability and 12 μW peak output,” Appl. Phys. Lett. 85, 3983 (2004). [CrossRef]
4. C. Baker, I. S. Gregory, W. R. Tribe, E. H. Linfield, and M. Missous, “All-optoelectronic terahertz system using low-temperature-grown InGaAs photomixer,” Opt. Express 13, 9639, (2005). http://www.opticsexpress.org/abstract.cfm?id=86252 [CrossRef] [PubMed]
5. M. Sukhotin, E. R. Brown, A. C. Gossard, D. Driscoll, M. Hanson, P. Maker, and R. Muller, “Photomixing and photoconductor measurements on ErAs/InGaAs at 1.55μm,” Appl. Phys. Lett. 82, 3116 (2003). [CrossRef]
6. H. Ito, S. Kodama, Y. Muramoto, T. Furuta, T. Nagatsuma, and T. Ishibashi, “High-speed and high output InP-InGaAs unitravelling-carrier photodiodes,” IEEE J. Sel. Top. Quantum Electron. 10, 709 (2004). [CrossRef]
7. J. Mangeney, L. Joulaud, J.-M. Lourtioz, P. Crozat, and G. Patriarche, “Ultrafast response (~2.2 ps) of ion-irradiated InGaAs photoconductive switch at 1.55 μm,” Appl. Phys. Lett. 82, 856 (2003).
8. Y. C. Lim and R. A. Moor, “Properties of alternately charged coplanar parallel strips by conformal mappings,” Trans. Electron. Dev. 15, 173 (1968). [CrossRef]
9. M. A. Lampert and P. Mark, Current Injection in Solids (Academic, New York, 1970).
10. E. R. Brown, F. W. Smith, and K. A. McIntosh “Coherent millimeter-wave generation by heterodyne conversion in low-temperature-grown GaAs photoconductors,” J. Appl. Phys. 73, 1480 (1993). [CrossRef]
11. C. Carmody, H. H. Tan, and C. Jagadish, “Electrical isolation of n- and p-In0.53Ga0.47As epilayers using ion irradiation,” J. Appl. Phys. Lett. 94, 6616 (2003).
12. E. Peytavit, S. Arcott, D. Lippens, G. Mouret, S. Matton, P. Masselin, R. Bocquet, J. F. Lampin, L. Desplanque, and F. Mollt, “Terahertz frequency difference from vertically integrated low-temperature-grown GaAs photodetector,” Appl. Phys. Lett. 81, 1174 (2002). [CrossRef]
13. I. S. Gregory, C. Baker, W. R. Tribe, I. V. Bradley, M. J. Evans, E. H. Linfield, A. G. Davies, and M. Missous, “Continuous-wave terahertz system with a 60 dB dynamic range,” Appl. Phys. Lett. 41, 717 (2005).