Abstract—We compare the photocarrier lifetime measured in Br-irradiated InGaAs and cold Fe-implanted InGaAsP. We also demonstrate the possibility of a two-photon absorption (TPA) process in ErAs:GaAs. The lifetime and the TPA were measured with a fiber-based 1550 nm time-resolved differential transmission (ΔT) set-up. The InGaAs-based materials show a positive ΔT with sub-picosecond lifetime, whereas ErAs:GaAs shows a negative ΔT consistent with a two-photon absorption process. I. INTRODUCTION AND BACKGROUND HOTOCONDUCTIVE (PC) switches and photomixers are commonly used for pulsed and CW THz generation. These devices are based on an ultrafast PC material, which has to possess a sub-picosecond carrier lifetime to get a short photocurrent pulse, a high photocarrier mobility to enhance the emitted THz power, and a high resistivity to allow large bias voltage in transmitters and reduce the detected noise in receivers. To achieve such lifetime with standard semiconductors, defects must be introduced during the growth or after. Annealed low-temperature-grown (LTG) GaAs is probably the most widely used THz PC material since it combines subpicosecond photocarrier lifetime, high resistivity, and good photocarrier mobility. However, not only is this material difficult to precisely reproduce, but as a PC switch it also requires the use of expensive and bulky Ti:sapphire lasers. GaAs with embedded nanoparticles of erbium (ErAs:GaAs) is one alternative approach to the LTG GaAs [1], but is usually driven by lasers near the GaAs band-gap wavelength, Ȝ  850 nm. There are great advantages to operate THz PC devices at 1550 nm where lasers are much more affordable than at 800 nm, and there are catalogues of low-cost fiber- optic components already well developed for telecom applications. Nevertheless, the ultrafast PC materials designed for 1550 nm have generally been inferior to their counterparts at 800 nm largely because they have been based on In 0.53 Ga 0.47 As or a similar alloy lattice matched to InP. With half of the band-gap energy of GaAs, this material generally suffers from much lower resistivity and breakdown field than GaAs-based materials. In recent years, however, researchers have been developing improved 1550-nm materials, such as Br-irradiated InGaAs [2], ultrafast InGaAs/InAlAs superlattices [3], and more recently cold Fe-implanted InGaAsP [4] and 1550-nm-driven ErAs:GaAs [5]. However, these materials have not been compared in a controlled fashion. So in this work, we measure and contrast the 1550 nm pump-probe phototransmission signature and photocarrier relaxation time for three of these materials: (1) Br-irradiated InGaAs, (2) cold Fe-implanted InGaAsP, and (3) ErAs:GaAs. II. SET-UP AND RESULTS All the measurements were made with the same set-up shown in figure 1. It consists of a 1550 nm EDFA fiber mode- locked laser and an optical set-up designed to do time-resolved differential phototransmission with the same incident power on each of the samples. The pulse train is delivered by the fiber laser and split into two parts using a 90/10 fiber splitter. The pump beam is sent through a U-bench to be chopped at a frequency around 400 Hz, whereas the probe beam is sent through a delay line having a maximum relative pump-probe time delay of 300 ps. After both pump and probe beams pass through a path-length compensating fiber patch cord, the beams are focused onto the sample under test using a fiber-to- free-space coupler. The pulse width at the output of the fiber- to-free-space coupler (according to optical autocorrelation measurements) is 250 fs. This is the temporal resolution of the experiment. The pump power is approximately 140 ȝW, and the ratio between the pump and probe is about 10. Fig. 1. Time resolved differential transmission set-up. The pump and the probe path are the same in time. Figure 2 shows the differential transmission for the 1-ȝm- thick Br-irradiated InGaAs sample (on SI-InP substrate) and the 1.5-ȝm-thick cold-implanted InGaAsP sample (also on SI- InP substrate) versus the time delay between the pump and the probe. The data plot of the Br-irradiated InGaAs sample can be fit by a mono-exponential, yielding τ 1 =400 fs. The cold- implanted InGaAsP sample is much better fit by two exponential functions, yielding τ 1 =400 fs and τ 2 =2.2 ps. Nevertheless, the ratio of the amplitudes of the τ 1 and τ 2 functions is ≈30, which indicates that the τ 2 process M. Martin a , E. R. Brown a , J. Mangeney b , A. Fekecs c , R. Arès c , and D. Morris c a Department of Physics, Wright State University, Dayton, OH, 45435, USA b IEF, CNRS UMR 8622, Université Paris-Sud, 91405 Orsay Cedex, France c Institut Interdisciplinaire d’Innovation Technologique (3iT), Université de Sherbrooke, Sherbrooke, Canada Email: matthieu.martin@wright.edu Critical Comparison of Carrier Lifetime at 1.55 ȝm of Ion-Irradiated InGaAs, Cold-Implanted InGaAsP, and ErAs:GaAs P 978-1-4673-1597-5/12/$31.00 ©2012 IEEE