Open Access
Add to BibSonomy
Abstract
Electromagnetic waves in the terahertz (THz) frequency range are widely used in spectroscopy, imaging and sensing. However, commercial, table-top systems covering the entire frequency range from 100 GHz to 10 THz are not available today. Fiber-coupled spectrometers, which employ photoconductive antennas as emitters and receivers, show a bandwidth limited to 6.5 THz and some suffer from spectral artifacts above 4 THz. For these systems, we identify THz absorption in the polar substrate of the photoconductive antenna as the main reason for these limitations. To overcome them, we developed photoconductive membrane (PCM) antennas, which consist of a 1.2 µm-thin InGaAs layer bonded on a Si substrate. These antennas combine efficient THz generation and detection in InGaAs with absorption-free THz transmission through a Si substrate. With these devices, we demonstrate a fiber-coupled THz spectrometer with a total bandwidth of 10 THz and an artifact-free spectrum up to 6 THz. The PCM antennas present a promising path toward fiber-coupled, ultrabroadband THz spectrometers.
© 2022 Optica Publishing Group under the terms of the Optica Open Access Publishing Agreement
1. Introduction
Within the last decades, THz time-domain spectroscopy (TDS) has evolved from a purely scientific tool to a widely used technique in spectroscopy, imaging, quality control and industrial non-destructive testing [1–4]. On the one hand, intense terahertz transients allow for controlling light and matter resonantly and non-resonantly with high spatial and temporal resolution [5–7]. On the other, THz TDS can be used to monitor thin layers of dielectric materials in the pharmaceutical, automotive and plastics industry to improve the respective production processes [8–11]. A challenge that remains, however, is the development of compact, energy-efficient and ultrabroadband THz systems.
Historically, photoconductive antennas (PCAs) were the first devices used for generating and coherently detecting THz pulses [12–14]. For compact and broadband THz systems, PCAs have essential advantages. First, THz generation in a PCA is linear in the optical excitation power and thus compatible with low pulse energies <1 nJ [2,15], which are easily achievable by compact fiber lasers. Furthermore, phase matching between the optical excitation and the generated THz radiation does not have to be taken into account [16,17]. In contrast, non-linear THz generation in optical crystals in combination with electro-optic (EO) detection requires higher pulse energies and careful control of the phase mismatch between the optical and THz pulse [18–20]. In addition, PCAs based on the III-V semiconductor indium phosphide (InP) are compatible with 1550 nm excitation, and thus fiber-based telecommunication technology [21–23]. In combination with erbium-doped fiber lasers, this enabled compact, all fiber-coupled TDS systems, which are used both in fundamental science and increasingly in industrial process control [11,24–27].
Despite many attempts, however, the maximum bandwidth of fiber-coupled TDS systems remains limited to 6.5 THz [28,29]. To increase the bandwidth, research has mainly focused on optimizing the photoconductive material [30,31]. Although ultrafast photoconductors with electron lifetimes of 100 fs, high mobility >1000 cm2/Vs and breakdown fields higher than 100 kV/cm were developed, the THz bandwidth could not be increased beyond 6.5 THz. So far, THz spectra with bandwidths exceeding 10 THz were only demonstrated with ultrashort optical pulses provided by free-space laser systems [20,32]. For such short pulses, optical fiber-coupling becomes increasingly challenging due to nonlinear processes, which cannot be compensated by linear dispersion management. This argument is independent of the type of THz emitter and applies to PCAs [33,34], spintronic emitters [35,36], non-linear optical crystals [37,38], as well as gas plasma THz generation [17,39,40]. Thus, often their use is restricted to free-space setups in laboratory environments.
In this Article, we revisit PCAs in a state-of-the-art, fiber-coupled TDS system and show that THz absorption in the InP substrate of the PCAs is the limiting factor of the spectral bandwidth. From previous studies, it is already known that the reststrahlen band of InP ranges from approx. 8.5 THz to 10.5 THz [19,41]. However, THz absorption starts from frequencies as low as 2 THz. In good agreement with the literature [42], we can link prominent dips in our measured THz spectra, which limit the usable bandwidth of today’s TDS systems, to impurity and phonon absorption in the InP substrate. In order to overcome these limitations, we developed a wafer-scale substrate transfer process from InP to silicon with a µm-thin bonding layer. The result is a 1.2 µm-thin photoconductive membrane (PCM) on a silicon (Si) substrate, which combines efficient THz generation and detection in InGaAs with the spectrally flat transmittance of Si. With these PCM antennas we were able to increase the bandwidth of our fiber-coupled THz TDS system from 6.5 to 10 THz. In addition, we could minimize spectral artifacts caused by impurity and phonon absorption in the substrate, which extends the artifact-free bandwidth of fiber-coupled THz TDS systems from 4 THz to more than 6 THz.
2. InP and Si as substrate materials for PCAs
Before we discuss the properties of the PCM antennas, we investigate THz absorption in iron-doped InP (InP:Fe), which is the established substrate material for PCAs for 1550 nm excitation. Figure 1 depicts a schematic of a typical THz TDS setup (a) and a PCA (b). Both PCA emitter and receiver are excited by a femtosecond laser source, which is split into two optical arms. The optical delay line in the receiver arm of the spectrometer allows for scanning the electric field of the incoming THz pulse as a function of time. For details we refer to the literature [2,43,44]. For the following discussion, the configuration of the PCAs is important, which is shown in more detail in Fig. 1(b). The PCA comprises a thin photoconductive layer grown on top of a semiconductor substrate. Gold contact electrodes are used to apply a bias voltage to the emitter and to extract the photocurrent from the receiver, respectively. Both emitter and receiver are illuminated by the femtosecond laser from the top side, where the photoconductor lies. A substrate lens is used to couple the THz radiation from the PCA chip into free space. Such a lens is needed since the difference in refractive indices between substrate and air is large and the THz pulse exhibits a large bandwidth of a few THz. Typically, the material of the lens is high resistivity float zone (HRFZ) silicon as the refractive index of silicon matches the refractive index of the substrate in the THz frequency range. The THz receiver features an identical geometry. Thus, the THz pulse propagates twice through the substrate from the emitter to the receiver. Hence, THz absorption in the substrate would influence the propagation significantly.
Fig. 1. Details of a THz TDS spectrometer with photoconductive THz antennas. (a) Schematic of the setup. The output of a femtosecond laser is split into two optical arms. One arm is used to excite the biased emitter antenna, which generates a THz pulse. The other arm includes an optical delay unit to compensate the THz path length and allows sampling of the THz pulse as a function of time. (b) Details of a photoconductive antenna. The antenna chip itself is placed on a hyper-hemispherical HRFZ-Si lens. The photoconductive antenna consists of the photoconductive layer with an antenna structure, which lies on top of an appropriate substrate. In this configuration, the antenna is optically excited from the top.
Commercial, fiber-coupled THz spectrometers commonly use femtosecond optical pulses centered around 1550 nm for excitation. The established photoconductor with a band gap energy compatible with 1550 nm excitation is In0.53Ga0.47As, which can be grown lattice matched on an InP substrate. All PCAs presented in this Article are based on InGaAs. In 2017, Alyabyeva and co-workers investigated the dielectric properties of semi-insulating InP:Fe in the terahertz frequency range [42]. Interestingly, they found that the THz absorption starts to increase significantly before the well-known reststrahlen band around 8.5 THz, namely for frequencies as low as 2 THz. Based on these results, we characterized the dielectric properties of our own InP:Fe substrates (AXT Inc.) with a Bruker Vertex 80v Fourier transform infrared (FTIR) spectrometer by measuring the transmittance T and reflectance R. Details of this specific FTIR spectrometer and a comparison to a THz TDS system can be found in the literature [45,46]. The reflectance, absorbance (1-T-R) and transmittance of our 350 µm-thick InP:Fe substrates are shown as blue lines in Fig. 2(a), 2(b) and 2(c), respectively. The periodic oscillations for frequencies lower than 4 THz are due to Fabry-Pérot resonances inside the substrate. For higher frequencies, these oscillations decrease in amplitude since the surface roughness of the substrate, which was polished from one side only, becomes larger relative to the wavelength. In Fig. 2(a) and 2(b) the reststrahlen band of InP between 8.5 THz and 10.5 THz is clearly visible. Very importantly for application as THz antennas, the absorbance of InP:Fe starts to increase for frequencies around 2 THz and reaches values above 0.5 for frequencies between 5 THz and the reststrahlen band [see Fig. 2(b)]. In addition, prominent dips in the spectrum can be found around 4.0, 5.0 and 5.7 THz. Since we ensured that the beam path of the FTIR spectrometer was evacuated adequately, these absorption features are not caused by residual water vapor absorption. In addition, Alyabyeva and co-workers observed similar spectral dips at these frequencies in their FTIR measurements on InP:Fe substrates from AXT and attributed them to multi-phonon resonances and absorption from unintended background doping. Around 4.0 THz it could be absorption by two transversal acoustic phonons (2TA), around 5 THz absorption by unintended germanium (Ge) background doping. The 5.7 THz dip may be caused by a differential longitudinal optic-longitudinal acoustic phonon (LO-LA) process [42]. Overall, our FTIR measurements confirm that THz absorption in the InP:Fe substrate is non-negligible for frequencies as low as 2 THz and limits the gap-less transmittance to a maximum frequency of 6.5 THz.
Fig. 2. Optical properties of SI-InP:Fe (blue) and HRFZ-Si (green) between 1 THz and 20 THz. The reflectance R (a), absorbance (b) and transmittance T (c) of both samples are shown. The optical properties were measured with a Fourier transform infrared spectrometer (FTIR). (d) Comparison of the attenuation [T/(1-R)]2 of InP (blue) with the measured THz spectrum (gray) of a state-of-the-art fiber-coupled THz time-domain spectrometer. Characteristic absorption dips appear at 4.0, 5.0 and 5.7 THz, which are caused by absorption by two transversal acoustic phonons (2TA), germanium background doping (Ge) and differential absorption by a longitudinal optical and a longitudinal acoustic phonon (LO-LA) [42], respectively. In addition, the fundamental bandwidth limit at 6.5 THz can be observed in both spectra.
It is well-known that HRFZ-Si offers low THz absorption and high transmittance. Therefore, we characterized the optical properties of a HRFZ-Si wafer with a thickness of
380 µm with the same FTIR spectrometer. Reflectance, absorbance and transmittance of Si are depicted as green lines in Fig. 2(a)–(c). As expected, the transmittance is almost constant up to 17 THz. This is mainly due to the fact that Si is a monoatomic crystal, which does not feature optical phonons with resonances in the THz frequency range. Since the Si wafer was polished on both sides, the Fabry-Pérot oscillations are visible up to 20 THz.
In order to demonstrate that the THz transmission of the InP:Fe substrate determines the spectrum of state-of-the-art TDS spectrometers using PCAs as emitter and receiver, Fig. 2(d) compares the square of the internal transmittance of the InP:Fe substrate (blue) with a measured TDS spectrum (gray). We use the square of the internal transmittance [T / (1-R)]2 since both PCA emitter and receiver were grown on an InP:Fe substrate. Hence, the THz pulse propagates through 350 µm of InP:Fe in both devices. Note that the frequency roll-off and the position of the dips in the measured spectrum agree well with the transmission of the substrate. The pronounced dips at 4.0, 5.0 and 5.7 THz appear in the measured THz spectrum as well as in the absorbance of InP:Fe. In addition, the fundamental bandwidth limit of 6.5 THz manifests itself in both spectra. These results indicate that THz absorption in the InP substrate determines the spectral roll-off of state-of-the-art THz TDS systems and limits the spectral bandwidth to 6.5 THz.
To evaluate the benefit of an antenna design without an absorbing substrate, we performed numerical simulations of THz generation, propagation and detection in a TDS system using PCAs within the framework of a classical Drude transport model [2,43]. Figure 3(a) compares the simulation results for THz antennas without accounting for the absorption in the InP substrate (blue) with measurement results obtained with state-of-the-art PCAs on InP:Fe (green). While spectral components with the highest power in the range between 300 GHz and 2 THz are relatively well reproduced by the simulation, a larger deviation is visible for the higher frequency components above 2 THz. Importantly, this simulation suggests that the large bandwidth of the optical fs-pulse should allow for generating and detecting THz spectra with more than 6.5 THz bandwidth.
Fig. 3. Comparison of simulated and measured THz spectra for state-of-the-art photoconductive antennas. (a) Simulation results (blue) without accounting for the InP substrate. Without an absorbing substrate, the large optical bandwidth should allow for detecting a THz spectrum up to 10 THz. (b) Simulation results including the effect of a 350 µm thick InP substrate. Now, measurement (green) and simulation (blue) are in good agreement. The maximum measured THz bandwidth of 6.5 THz is reproduced by the simulation and limited by absorption in the InP substrate.
In Fig. 3(b) we extended the numerical model to include the effect of the substrate by adding a transfer matrix model to account for the 350 µm thick InP:Fe substrate. The dielectric function of InP:Fe was modeled as a sum of Lorentzian oscillators with parameters taken from the literature [42]. Now, simulation and measurement are in good agreement. Distinct absorption features of the InP substrate at 4.0, 5.0 and 5.7 THz can be identified both in the simulation and the measurement. In addition, the model now correctly predicts the maximum bandwidth of PCAs, which is limited to 6.5 THz by absorption in the InP substrate.
3. III-V photoconductive membrane antennas on silicon
3.1 Microfabrication process
The optical properties of HRFZ-Si and InP:Fe presented in Fig. 2 suggest that HRFZ-Si would be the preferred substrate material for broadband THz PCAs. However, there is no simple way to grow III-V semiconductors directly on silicon due to the large difference of their respective lattice constants. The integration of III-V semiconductors on silicon substrates has been an important scientific topic for decades. In this regard, epitaxial lift-off techniques are an alternative to lattice-mismatched heteroepitaxy [47,48]. For GaAs-based THz receivers for 800 nm excitation, epitaxial lift-off techniques have been investigated before [49,50]. We followed this line of work and developed an epitaxial lift-off process for the InGaAs photoconductor from InP:Fe to a HRFZ-Si substrate. The overall goal was a reproducible wafer-scale fabrication process.
Before the epitaxial lift-off process, gas source molecular beam epitaxy was used to grow the photoconductor lattice matched on 2” semi-insulating InP:Fe substrates. The photoconductive membrane antennas are based on InGaAs:Fe with a Fe doping concentration of approx. 5×1019 cm-3. For this material, a 1.2 µm thick, homogenously doped In0.53Ga0.47As:Fe layer was grown on top of a 0.7 µm In0.52Al0.48As buffer layer after native oxide desorption [51]. The growth temperature was approx. 400°C. Please note that the described epitaxial lift-off process is compatible with any photoconductive material. For the future, we intend to investigate other photoconductors such as rhodium-doped InGaAs, as well [29,53].
PCA emitters used for comparison to the PCM devices were made of InGaAs:Fe. PCA receivers were based on Be-doped InGaAs/InAlAs superlattices [23,52]. Low-temperature-grown, Be-doped InGaAs/InAlAs superlattices are the standard photoconductor used for some commercially available THz receivers. In addition, it was one of the first photoconductive materials featuring an ultrashort electron lifetime in combination with a high resistivity [23,39].
Figure 4(a) gives an overview of the microfabrication steps of a THz PCM on Si. After epitaxy, the InP wafer with the photoconductive layer on top is bonded onto an unstructured
Fig. 4. Microfabrication of III-V photoconductive membranes on Si. (a), Overview of the microfabrication steps. First, the InP-substrate with the epitaxially grown III-V photoconductor is bonded with BCB onto an unstructured HRFZ-Si wafer. Second, the InP is removed via lapping and wet etching. Third, photoconductor and BCB are mesa structured in two dry etching processes. Finally, the antenna structure is sputtered onto the wafer and the photoconductive membrane antennas are separated by sawing. (b), Scanning electron micrograph of a photoconductive membrane antenna. The III-V photoconductor and BCB mesa are visible in the center. (c), Differential transmission signal of optical femtosecond pulses centered around 1550 nm through the InGaAs photoconductive layer before and after PCM processing.
HRFZ-Si wafer with the same thickness. For wafer bonding, we use benzocyclobutene (BCB), since BCB is an electrical isolator and widely used in microfabrication processes. In order to avoid reflections of the generated THz pulse at the BCB-Si interface, it is critical to employ thin BCB layers. A frequency of 10 THz corresponds to a wavelength of 30 µm in air and 8.8 µm in Si (n = 3.43). Thus, the thickness of the BCB bonding layer should be smaller than 8.8 µm in order to avoid detrimental reflections. In our samples, the thickness of the BCB layer measures approx. 4 µm. After waferbonding, the InP:Fe substrate is removed in a two-step process. First, lapping reduces the substrate thickness from 350 µm to about 70 µm. Afterwards, wet etching with hydrochloric acid (HCl) and phosphoric acid (H3PO4) removes the InP substrate completely. At this stage, a 1.2 µm-thin photoconductive membrane remains bonded to the BCB layer. Following the removal of the InP substrate, the THz antennas are photolithographically fabricated. An InGaAs and BCB mesa are formed in two dry etching processes [54]. Mesa structuring of the BCB layer allows to apply the electrical contact lines on the Si substrate, which improves the transfer of heat from the photoconductor into the Si substrate. Finally, gold contacts and the antenna structure are formed by metal sputtering. Figure 4(b) shows a scanning electron micrograph of a PCM. The InGaAs and BCB mesa are visible in the center. The presented device is a THz emitter that was structured as a strip-line antenna with a photoconductive gap of 25 µm width. Processed THz receivers feature a dipole-type antenna with a 10 µm-wide photoconductive gap. The λ/2 dipole length of the antenna is 25 µm.
A critical property of the photoconductor is the lifetime of optically excited carriers in the conduction band. An ultrashort carrier lifetime below 1 ps is needed for broadband detection capability, otherwise the receiver current corresponds to a temporal integral over the THz signal [2,43]. To verify that the processing steps did not deteriorate the ultrafast trapping capability of the photoconductor, we performed optical pump-optical probe measurements before and after microfabrication of the PCMs [55]. As for the THz TDS experiments, an Er-doped fiber laser emitting pulses with a duration of 90 fs centered on 1550 nm is used as the optical source. Excitation by the pump will lead to an increased transmission of the probe due to Pauli blocking. The transmission decreases again as excited carriers are trapped. Therefore, it is convenient to trace the differential transmission (DT) of the probe as a function of time, which is depicted in Fig. 4(c). As can be seen, the ultrashort lifetime of excited electrons is maintained in the PCM structure.
3.2 THz TDS results
For all THz TDS measurements, the individual antennas were packaged as fiber-coupled THz modules. A fine placer was used to position the THz antennas in the center of a hyper-hemispherical HRFZ-Si lens with a diameter of 10 mm and a thickness of 6.1 mm. The fiber pigtail of the THz module was actively aligned to the position of maximum photocurrent and fixated on the housing. In this way, the THz modules can be easily exchanged with the fiber-coupled TDS system.
A commercially available THz TDS system was used for all measurements (TeraFlash pro, TOPTICA Photonics AG). This system includes an Er-doped fiber laser, which emits optical pulses with a duration of approx. 90 fs centered on 1550 nm at a repetition rate of 100 MHz. The available average optical power in the emitter and receiver arm is 20 mW. A fast linear optical delay line is used to sample the THz pulse in the time-domain. In this way, a single 70 ps long pulse trace can be recorded in 60 ms. The time-domain step size is fixed to 50 fs resulting in a maximum measurable frequency of 10 THz. No lock-in scheme is needed. Instead, the position of the optical delay unit is correlated with the measured current induced in the receiver by a field programmable gate array. In comparison to free-space laboratory THz systems, this fiber-coupled spectrometer is compact, robust and energy efficient. For all measurements shown in this section, we employed fiber-coupled THz modules.
In Fig. 5(a), the effect of replacing the PCA receiver by a PCM antenna is shown. For both measurements, the same setup with the same InGaAs:Fe based PCA emitter was used. As can be seen, the prominent absorption features of the InP:Fe substrate are significantly reduced with the PCM receiver. This supports our hypothesis that these absorption features stem from the substrate. The remaining dips around 4.0, 5.0, and 5.7 THz are attributed to THz absorption in the substrate of the PCA emitter. In addition, one observes that the transmission window of InP:Fe around 8 THz, which is visible in Fig. 3(b), also appears in the THz spectrum recorded with the PCM receiver. However, the maximum gap-less spectral bandwidth remains limited to 6.5 THz due to THz absorption in the substrate of the PCA emitter. Clearly, to convert the full available optical bandwidth into the THz-domain, it is necessary to apply PCM antennas as both emitter and receiver. These results are presented in Fig. 5(b) with measurement in green and simulation results in blue. The gray curve indicates the noise spectrum of the PCM receiver, which was obtained by blocking the THz path such that no THz signal was incident on the receiver. In this configuration, the measured spectrum in Fig. 5(b) has a bandwidth of 10 THz. This is by far the highest bandwidth detected with an all fiber-coupled THz TDS system and constitutes an increase in bandwidth of 3.5 THz in comparison to the best results obtained with fiber-coupled PCAs so far. Although the overall spectral bandwidth of the PCM antennas agrees well with the predictions of the numerical simulation, the measured spectrum shows some remaining absorption features between 6.5 THz and 8 THz. We attribute these to THz absorption by the InGaAs photoconductor itself, since InGaAs is a polar semiconductor featuring TO and LO phonons. However, the thickness of the InGaAs layer measures 1.2 µm only, so that the THz signal remains above the noise floor in this spectral region, too.
Fig. 5. PCM antennas as THz emitters and receivers. (a) Influence of the THz receiver. The combination of state-of-the-art PCA emitter and receiver is shown in gray. The green line corresponds to the combination of the same emitter with a PCM receiver. While the latter combination reduces spectral artifacts caused by absorption in the InP:Fe substrate, the maximum bandwidth remains limited to 6.5 THz due to absorption in the InP:Fe substrate of the emitter. (b) THz spectra with PCM antennas, only. Measurement and simulation are depicted as green and blue lines. The gray line represents the noise level obtained by blocking the THz beam. The measured spectrum features a bandwidth of 10 THz in agreement with the simulation. This corresponds to an increase of 3.5 THz compared to standard PCA emitters operated with the same TDS system. For a further increase from 10 to 40 kV/cm in the bias field at the PCM emitter, we expect a 12 dB increase in spectral power (dashed blue line).
Although the spectrum detected with our PCM antennas features already a bandwidth of 10 THz, the potential of this approach is even higher. In comparison to state-of-the-art THz emitters made of InGaAs:Fe, which were optimized to withstand bias fields of more than 40 kV/cm, [51] the presented PCM emitter was biased at 10 kV/cm, only. The reason of the lower breakdown field strength of PCM emitters is not fully understood; yet. We are currently investigating if the BCB layer between photoconductive gap and Si substrate may act as a thermal barrier that leads to excessive heating and premature thermal breakdown during operation. In order to increase the breakdown field strength, we are evaluating a BCB layer with a thickness of a few hundred nm only. If the PCM emitters could be operated with a 40 kV/cm bias, we would expect a 4-fold increase of the emitted pulse amplitude, corresponding to a 12 dB increase in spectral power [56]. Since the noise level of a TDS measurement is determined by the receiver noise, the higher THz power emitted from a PCM operated at 40 kV/cm would directly translate into a higher dynamic range of the THz spectrum. Hence, a PCM emitter operated at 40 kV/cm would enable an even higher bandwidth than 10 THz. This potential increase in dynamic range and bandwidth is illustrated by the dashed blue line in Fig. 5(b). Overall, the results presented in Fig. 5 suggest that PCM based THz systems are very promising candidates for table-top, ultrabroadband THz spectroscopy.
4. Discussion
The presented improvements in bandwidth and spectral quality directly impact THz applications in science and technology. For spectroscopy, the elimination of impurity absorption in the InP substrate increases the usable bandwidth of TDS systems employing PCM antennas. In order to demonstrate the advantages of the increased usable THz bandwidth and the elimination of InP related absorption features, we measured the transfer function of a 380 µm-thick silicon wafer in reflection geometry. For these measurements we employed the same THz TDS system and the same THz modules as for the measurements shown in Fig. 5. The only difference between PCA and PCM measurements was the bias field applied to the emitter, which measured 40 kV/cm (corresponding to 100 V bias voltage) for PCAs and only 10 kV/cm (corresponding to 25 V) for PCMs. As shown in Fig. 6, the full bandwidth extends up to 6.5 THz for the PCAs (a) and approaches 10 THz for the PCMs (b). By dividing the Si spectrum by a reference spectrum reflected from a planar metallic mirror, we calculated the transfer function of the Si wafer. The results are shown in Fig. 6(c) and 6(d) for frequencies up to 8 THz. According to the FTIR measurements presented in Fig. 2(c), the transmission of silicon is expected to be spectrally flat. However, the presented results indicate that only the measurement with PCM antennas shows a spectrally flat transfer function. In contrast, the spectral features of the PCAs caused by THz absorption in the InP substrate lead to inaccuracies and even to an unphysical transfer function >1 around 5 THz. This means that the usable bandwidth of a TDS system employing PCAs is limited to frequencies below 4 THz, although the overall bandwidth is significantly higher. By eliminating the InP related absorption features through PCM antennas, the usable bandwidth increases to 6.5 THz. Hence, PCM antennas enable spectroscopic measurements with significantly improved quality, accuracy and bandwidth compared to the current state-of-the-art, which may be especially important for applications in material science, chemistry and biology where complex FTIR spectrometers may be replaced by compact, fiber-coupled THz TDS systems.
Fig. 6. Transfer function of a 380 µm thick HRFZ-Si wafer as measured by THz TDS. (a), (b) THz spectra from a polished Si wafer measured in reflection geometry with state-of-the-art PCAs and PCM antennas, respectively. We used the same antennas as for the measurements shown in Fig. 5. Due to the minimized absorption in the substrate for the PCM antennas, the total bandwidth is increased from 6.5 THz to almost 10 THz. (c), (d) Transfer function of the Si wafer calculated by dividing the spectra shown in (a) and (b) by a reference spectrum reflected from a planar metallic mirror. For the PCAs, absorption features of the InP substrate cause artifacts in the transfer function and limit the usable THz bandwidth to approx. 4 THz. With PCM antennas, these artifacts are minimized resulting in a spectrally flat transfer function and a usable bandwidth of 6.5 THz.
Regarding industrial applications of THz TDS, one of the most prominent examples are non-contact layer thickness measurements of paint layers and coatings. Today, the minimal resolvable layer thickness with THz TDS lies around 5 to 10 µm, depending on the refractive index of the material. This limit is mainly set by the spectral bandwidth of the THz pulse [57]. Thus, the increased bandwidth of the PCM antennas will enable layer thickness measurements of µm-thin layers with low refractive index contrast, as they are typically employed in plastic foils and thermal barrier coatings. Thus, it is likely that the presented development will open additional industrial markets for THz sensing. Overall, the PCM antennas presented in this Article have the potential to enable both ultrabroadband spectroscopy and precise industrial sensing with a compact, fiber-coupled THz system.
5. Conclusion
In this Article, we presented InGaAs based photoconductive membrane (PCM) antennas as ultrabroadband emitters and receivers for THz TDS and achieved a 10 THz bandwidth with a commercial, all fiber-coupled spectrometer. This is a bandwidth increase of 3.5 THz compared to state-of-the-art photoconductive antennas (PCAs). In numerical simulations of THz TDS spectra we found that the spectral bandwidth of PCAs should be essentially higher than 6.5 THz, which was the limit in all previous publications. In order to explain this, we performed FTIR measurements on InP:Fe, which is the established substrate material for PCAs for 1550 nm excitation. We found significant THz absorption for frequencies as low as 2 THz. In addition, we could link prominent dips in the THz spectra of standard PCAs with impurity and two phonon absorption in InP:Fe. Hence, we concluded that THz absorption in the InP:Fe substrate is the key factor limiting the bandwidth of today’s fiber-coupled TDS systems. In order to overcome this, we developed a wafer-scale, substrate transfer process of the III-V photoconductor InGaAs from InP:Fe to silicon. FTIR measurements on silicon revealed an absorption-free and spectrally flat transmission for frequencies up to 17 THz. The resulting PCM antennas combine the advantages of III-V photoconductors as efficient THz emitters and receivers with the low THz absorption of silicon. The PCM antennas were packaged as fiber-coupled modules and operated with a commercial TDS system. Thereby, we achieved a bandwidth of 10 THz and minimized unwanted spectral artifacts caused by phonon and impurity absorption in the InP:Fe substrate. As an illustration of the improved quality and bandwidth of the TDS system, we measured the transfer function of a silicon wafer in reflection geometry. This measurement demonstrated an artifact-free spectrum with a spectrally flat transmission function up to 6.5 THz and an overall bandwidth approaching 10 THz. In the future, PCM antennas could enable TDS spectrometers with even higher bandwidth, if the pulse duration of today’s Er-doped fiber lasers with fiber delivery could be further reduced. Overall, the PCM approach opens a promising route to compact and robust, ultrabroadband THz spectrometers for applications in science and industry.
Funding
Deutsche Forschungsgemeinschaft (GL 958/1-2).
Acknowledgments
We thank Elliott Brown, Roman J. B. Dietz and Andreas Steiger for careful proofreading of the manuscript and valuable discussions.
Disclosures
The authors declare no conflicts of interest.
Data availability
Data underlying the results presented in this paper are not publicly available at this time, but may be obtained from the authors upon reasonable request.
Supplemental document
See Supplement 1 for supporting content.
References
1. M. Tonouchi, “Cutting-edge terahertz technology,” Nat. Photonics 1(2), 97–105 (2007). [CrossRef]
2. P. U. Jepsen, D. G. Cooke, and M. Koch, “Terahertz spectroscopy and imaging - Modern techniques and applications,” Laser Photonics Rev. 5(1), 124–166 (2011). [CrossRef]
3. S. S. Dhillon, M. S. Vitiello, E. H. Linfield, A. G. Davies, M. C. Hoffmann, J. Booske, C. Paoloni, M. Gensch, P. Weightman, G. P. Williams, E. Castro-Camus, D. R. S. Cumming, F. Simoens, I. Escorcia-Carranza, J. Grant, S. Lucyszyn, M. Kuwata-Gonokami, K. Konishi, M. Koch, C. A. Schmuttenmaer, T. L. Cocker, R. Huber, A. G. Markelz, Z. D. Taylor, V. P. Wallace, J. Axel Zeitler, J. Sibik, T. M. Korter, B. Ellison, S. Rea, P. Goldsmith, K. B. Cooper, R. Appleby, D. Pardo, P. G. Huggard, V. Krozer, H. Shams, M. Fice, C. Renaud, A. Seeds, A. Stöhr, M. Naftaly, N. Ridler, R. Clarke, J. E. Cunningham, and M. B. Johnston, “The 2017 terahertz science and technology roadmap,” J. Phys. D: Appl. Phys. 50(4), 043001 (2017). [CrossRef]
4. D. M. Mittleman, “Twenty years of terahertz imaging,” Opt. Express 26(8), 9417 (2018). [CrossRef]
5. T. Kampfrath, K. Tanaka, and K. A. Nelson, “Resonant and nonresonant control over matter and light by intense terahertz transients,” Nat. Photonics 7(9), 680–690 (2013). [CrossRef]
6. T. L. Cocker, V. Jelic, M. Gupta, S. J. Molesky, J. A. J. Burgess, G. D. L. Reyes, L. V. Titova, Y. Y. Tsui, M. R. Freeman, and F. A. Hegmann, “An ultrafast terahertz scanning tunnelling microscope,” Nat. Photonics 7(8), 620–625 (2013). [CrossRef]
7. H. A. Hafez, S. Kovalev, J. C. Deinert, Z. Mics, B. Green, N. Awari, M. Chen, S. Germanskiy, U. Lehnert, J. Teichert, Z. Wang, K. J. Tielrooij, Z. Liu, Z. Chen, A. Narita, K. Müllen, M. Bonn, M. Gensch, and D. Turchinovich, “Extremely efficient terahertz high-harmonic generation in graphene by hot Dirac fermions,” Nature 561(7724), 507–511 (2018). [CrossRef]
8. M. Naftaly, N. Vieweg, and A. Deninger, “Industrial applications of terahertz sensing: State of play,” Sensors 19(19), 4203 (2019). [CrossRef]
9. D. Markl, P. Bawuah, C. Ridgway, S. van den Ban, D. J. Goodwin, J. Ketolainen, P. Gane, K. E. Peiponen, and J. A. Zeitler, “Fast and non-destructive pore structure analysis using terahertz time-domain spectroscopy,” Int. J. Pharm. 537(1-2), 102–110 (2018). [CrossRef]
10. K. Su, Y. C. Shen, and J. A. Zeitler, “Terahertz sensor for non-contact thickness and quality measurement of automobile paints of varying complexity,” IEEE Trans. Terahertz Sci. Technol. 4(4), 432–439 (2014). [CrossRef]
11. M. Yahyapour, A. Jahn, K. Dutzi, T. Puppe, P. Leisching, B. Schmauss, N. Vieweg, and A. Deninger, “Fastest thickness measurements with a Terahertz time-domain system based on electronically controlled optical sampling,” Appl. Sci. 9(7), 1283 (2019). [CrossRef]
12. D. H. Auston, K. P. Cheung, and P. R. Smith, “Picosecond photoconducting Hertzian dipoles,” Appl. Phys. Lett. 45(3), 284–286 (1984). [CrossRef]
13. P. R. Smith, D. H. Auston, and M. C. Nuss, “Subpicosecond Photoconducting Dipole Antennas,” IEEE J. Quantum Electron. 24(2), 255–260 (1988). [CrossRef]
14. C. Fattinger and D. Grischkowsky, “Terahertz beams,” Appl. Phys. Lett. 54(6), 490–492 (1989). [CrossRef]
15. F. W. Smith, H. Q. Le, V. Diadiuk, M. A. Hollis, A. R. Calawa, S. Gupta, M. Frankel, D. R. Dykaar, G. A. Mourou, and T. Y. Hsiang, “Picosecond GaAs-based photoconductive optoelectronic detectors,” Appl. Phys. Lett. 54(10), 890–892 (1989). [CrossRef]
16. J. Shan and T. F. Heinz, “Terahertz radiation from semiconductors,” in Ultrafast Dynamical Processes in Semiconductors, Topics Appl. Phys. (Springer, 2004).
17. F. Blanchard, G. Sharma, L. Razzari, X. Ropagnol, H. Bandulet, F. Vidal, R. Morandotti, J. Kieffer, T. Ozaki, H. Tiedje, H. Haugen, M. Reid, and F. Hegmann, “Generation of Intense Terahertz Radiation via Optical Methods,” IEEE J. Sel. Top. Quantum Electron. 17(1), 5–16 (2011). [CrossRef]
18. H. J. Bakker, G. C. Cho, H. Kurz, Q. Wu, and X. C. Zhang, “Distortion of terahertz pulses in electro-optic sampling,” J. Opt. Soc. Am. B 15(6), 1795–1801 (1998). [CrossRef]
19. A. Leitenstorfer, S. Hunsche, J. Shah, M. Nuss, and W. Knox, “Femtosecond Charge Transport in Polar Semiconductors,” Phys. Rev. Lett. 82(25), 5140–5143 (1999). [CrossRef]
20. C. Kübler, R. Huber, and A. Leitenstorfer, “Ultrabroadband terahertz pulses: generation and field-resolved detection,” Semicond. Sci. Technol. 20(7), S128–S133 (2005). [CrossRef]
21. M. Suzuki and M. Tonouchi, “Fe-implanted InGaAs terahertz emitters for 1.56 µm wavelength excitation,” Appl. Phys. Lett. 86(5), 051104 (2005). [CrossRef]
22. J. Sigmund, C. Sydlo, H. L. Hartnagel, N. Benker, H. Fuess, F. Rutz, T. Kleine-Ostmann, and M. Koch, “Structure investigation of low-temperature-grown GaAsSb, a material for photoconductive terahertz antennas,” Appl. Phys. Lett. 87(25), 252103 (2005). [CrossRef]
23. R. J. B. Dietz, B. Globisch, H. Roehle, D. Stanze, T. Göbel, and M. Schell, “Influence and adjustment of carrier lifetimes in InGaAs/InAlAs photoconductive pulsed terahertz detectors: 6 THz bandwidth and 90 dB dynamic range,” Opt. Express 22(16), 19411–19422 (2014). [CrossRef]
24. B. Sartorius, H. Roehle, H. Künzel, J. Böttcher, M. Schlak, D. Stanze, H. Venghaus, and M. Schell, “All-fiber terahertz time-domain spectrometer operating at 1.5 µm telecom wavelengths,” Opt. Express 16(13), 9565–9570 (2008). [CrossRef]
25. N. Vieweg, F. Rettich, A. Deninger, H. Roehle, R. Dietz, T. Göbel, and M. Schell, “Terahertz-time domain spectrometer with 90 dB peak dynamic range,” J. Infrared, Millimeter, Terahertz Waves 35(10), 823–832 (2014). [CrossRef]
26. R. J. B. Dietz, N. Vieweg, T. Puppe, A. Zach, B. Globisch, T. Göbel, P. Leisching, and M. Schell, “All fiber-coupled THz-TDS system with kHz measurement rate based on electronically controlled optical sampling,” Opt. Lett. 39(22), 6482–6485 (2014). [CrossRef]
27. A. Bartels, R. Cerna, C. Kistner, A. Thoma, F. Hudert, C. Janke, and T. Dekorsy, “Ultrafast time-domain spectroscopy based on high-speed asynchronous optical sampling,” Rev. Sci. Instrum. 78(3), 035107 (2007). [CrossRef]
28. U. Nandi, K. Dutzi, A. Deninger, H. Lu, J. Norman, A. Gossard, N. Vieweg, and S. Preu, “ErAs:In(Al)GaAs photoconductor-based time domain system with 4.5 THz single shot bandwidth and emitted terahertz power of 164 µW,” Opt. Lett. 45(10), 2812–2815 (2020). [CrossRef]
29. R. B. Kohlhaas, S. Breuer, L. Liebermeister, S. Nellen, M. Deumer, M. Schell, M. P. Semtsiv, W. T. Masselink, and B. Globisch, “637 µW emitted terahertz power from photoconductive antennas based on rhodium doped InGaAs,” Appl. Phys. Lett. 117(13), 131105 (2020). [CrossRef]
30. A. Mingardi, W.-D. Zhang, E. R. Brown, A. D. Feldman, T. E. Harvey, and R. P. Mirin, “High power generation of THz from 1550-nm photoconductive emitters,” Opt. Express 26(11), 14472–14478 (2018). [CrossRef]
31. U. Nandi, J. C. Norman, A. C. Gossard, H. Lu, and S. Preu, “1550-nm Driven ErAs:In(Al)GaAs Photoconductor-Based Terahertz Time Domain System with 6.5 THz Bandwidth,” J. Infrared, Millimeter, Terahertz Waves 39(4), 340–348 (2018). [CrossRef]
32. A. Sell, R. Scheu, A. Leitenstorfer, and R. Huber, “Field-resolved detection of phase-locked infrared transients from a compact Er:fiber system tunable between 55 and 107 THz,” Appl. Phys. Lett. 93(25), 251107 (2008). [CrossRef]
33. A. Singh, A. Pashkin, S. Winnerl, M. Helm, and H. Schneider, “Gapless Broadband Terahertz Emission from a Germanium Photoconductive Emitter,” ACS Photonics 5(7), 2718–2723 (2018). [CrossRef]
34. A. Singh, A. Pashkin, S. Winnerl, M. Welsch, C. Beckh, P. Sulzer, A. Leitenstorfer, M. Helm, and H. Schneider, “Up to 70 THz bandwidth from an implanted Ge photoconductive antenna excited by a femtosecond Er:fibre laser,” Light: Sci. Appl. 9(1), 30 (2020). [CrossRef]
35. T. Kampfrath, M. Battiato, P. Maldonado, G. Eilers, J. Nötzold, S. Mährlein, V. Zbarsky, F. Freimuth, Y. Mokrousov, S. Blügel, M. Wolf, I. Radu, P. M. Oppeneer, and M. Münzenberg, “Terahertz spin current pulses controlled by magnetic heterostructures,” Nat. Nanotechnol. 8(4), 256–260 (2013). [CrossRef]
36. T. Seifert, S. Jaiswal, U. Martens, J. Hannegan, L. Braun, P. Maldonado, F. Freimuth, A. Kronenberg, J. Henrizi, I. Radu, E. Beaurepaire, Y. Mokrousov, P. M. Oppeneer, M. Jourdan, G. Jakob, D. Turchinovich, L. M. Hayden, M. Wolf, M. Münzenberg, M. Kläui, and T. Kampfrath, “Efficient metallic spintronic emitters of ultrabroadband terahertz radiation,” Nat. Photonics 10(7), 483–488 (2016). [CrossRef]
37. R. Huber, A. Brodschelm, F. Tauser, and A. Leitenstorfer, “Generation and field-resolved detection of femtosecond electromagnetic pulses tunable up to 41 THz,” Appl. Phys. Lett. 76(22), 3191–3193 (2000). [CrossRef]
38. Y. C. Shen, P. C. Upadhya, E. H. Linfield, H. E. Beere, and A. G. Davies, “Ultrabroadband terahertz radiation from low-temperature-grown GaAs photoconductive emitters,” Appl. Phys. Lett. 83(15), 3117–3119 (2003). [CrossRef]
39. K. Biermann, D. Nickel, K. Reimann, M. Woerner, T. Elsaesser, and H. Künzel, “Ultrafast optical nonlinearity of low-temperature-grown GaInAs/AlInAs quantum wells at wavelengths around 1.55µm,” Appl. Phys. Lett. 80(11), 1936–1938 (2002). [CrossRef]
40. H. G. Roskos, M. D. Thomson, M. Kreß, and T. Löffler, “Broadband THz emission from gas plasmas induced by femtosecond optical pulses: From fundamentals to applications,” Laser Photonics Rev. 1(4), 349–368 (2007). [CrossRef]
41. R. Huber, C. Kübler, S. Tübel, A. Leitenstorfer, Q. T. Vu, H. Haug, F. Köhler, and M. C. Amann, “Femtosecond formation of coupled phonon-plasmon modes in InP: Ultrabroadband THz experiment and quantum kinetic theory,” Phys. Rev. Lett. 94(2), 027401 (2005). [CrossRef]
42. L. N. Alyabyeva, E. S. Zhukova, M. A. Belkin, and B. P. Gorshunov, “Dielectric properties of semi- insulating Fe-doped InP in the terahertz spectral region,” Sci. Rep. 7(1), 7360 (2017). [CrossRef]
43. P. U. Jepsen, R. H. Jacobsen, and S. R. Keiding, “Generation and detection of terahertz pulses from biased semiconductor antennas,” J. Opt. Soc. Am. B 13(11), 2424–2436 (1996). [CrossRef]
44. J. Neu and C. A. Schmuttenmaer, “Tutorial: An introduction to terahertz time domain spectroscopy (THz-TDS),” J. Appl. Phys. 124(23), 231101 (2018). [CrossRef]
45. L. Oberto, M. Bisi, A. Kazemipour, A. Steiger, T. Kleine-Ostmann, and T. Schrader, “Measurement comparison among time-domain, FTIR and VNA-based spectrometers in the THz frequency range,” Metrologia 54(1), 77–84 (2017). [CrossRef]
46. M. Kehrt, C. Monte, A. Steiger, A. Hoehl, J. Hollandt, H.-P. Gemünd, A. Brömel, F. Hänschke, T. May, N. Deßmann, H.-W. Hübers, R. Mientus, and E. Reck, “Intercomparison of far-infrared transmittance measurements,” Opt. Express 26(26), 34484–34496 (2018). [CrossRef]
47. E. Yablonovitch, T. Gmitter, J. P. Harbison, and R. Bhat, “Extreme selectivity in the lift-off of epitaxial GaAs films,” Appl. Phys. Lett. 51(26), 2222–2224 (1987). [CrossRef]
48. H. Kum, D. Lee, W. Kong, H. Kim, Y. Park, Y. Kim, Y. Baek, S.-H. Bae, K. Lee, and J. Kim, “Epitaxial growth and layer-transfer techniques for heterogeneous integration of materials for electronic and photonic devices,” Nat. Electron. 2(10), 439–450 (2019). [CrossRef]
49. S. Kasai, T. Katagiri, J. Takayanagi, K. Kawase, and T. Ouchi, “Reduction of phonon resonant terahertz wave absorption in photoconductive switches using epitaxial layer transfer,” Appl. Phys. Lett. 94(11), 113505 (2009). [CrossRef]
50. M. Yamashita, H. Takahashi, T. Ouchi, and C. Otani, “Ultra-broadband terahertz time-domain ellipsometric spectroscopy utilizing GaP and GaSe emitters and an epitaxial layer transferred photoconductive detector,” Appl. Phys. Lett. 104(5), 051103 (2014). [CrossRef]
51. B. Globisch, R. J. B. Dietz, R. B. Kohlhaas, T. Göbel, M. Schell, D. Alcer, M. Semtsiv, and W. T. Masselink, “Iron doped InGaAs: Competitive THz emitters and detectors fabricated from the same photoconductor,” J. Appl. Phys. 121(5), 053102 (2017). [CrossRef]
52. B. Globisch, R. J. B. Dietz, S. Nellen, T. Göbel, and M. Schell, “Terahertz detectors from Be-doped low-temperature grown InGaAs / InAlAs : Interplay of annealing and terahertz performance,” AIP Adv. 6(12), 125011 (2016). [CrossRef]
53. R. B. Kohlhaas, S. Breuer, S. Nellen, L. Liebermeister, M. Schell, M. P. Semtsiv, W. T. Masselink, and B. Globisch, “Photoconductive terahertz detectors with 105 dB peak dynamic range made of rhodium doped InGaAs,” Appl. Phys. Lett. 114(22), 221103 (2019). [CrossRef]
54. H. Roehle, R. J. B. Dietz, H. J. Hensel, J. Böttcher, H. Künzel, D. Stanze, M. Schell, and B. Sartorius, “Next generation 1.5 µm terahertz antennas: mesa-structuring of InGaAs/InAlAs photoconductive layers,” Opt. Express 18(3), 2296–2301 (2010). [CrossRef]
55. B. Globisch, R. J. B. Dietz, D. Stanze, T. Göbel, and M. Schell, “Carrier dynamics in beryllium doped low-temperature-grown InGaAs/InAlAs,” Appl. Phys. Lett. 104(17), 172103 (2014). [CrossRef]
56. J. Y. Suen, W. Li, Z. D. Taylor, and E. R. Brown, “Characterization and modeling of a terahertz photoconductive switch,” Appl. Phys. Lett. 96(14), 141103 (2010). [CrossRef]
57. S. Krimi, J. Klier, J. Jonuscheit, G. Von Freymann, R. Urbansky, and R. Beigang, “Highly accurate thickness measurement of multi-layered automotive paints using terahertz technology,” Appl. Phys. Lett. 109(2), 021105 (2016). [CrossRef]
