Terahertz quantum cascade lasers have been investigated as a turn-key terahertz source for widespread applications. Two lasers were mounted in a small liquid nitrogen-cooled dewar and combined with a sophisticated pulse driver. We present a detailed analysis in respect to current-voltage characteristics, emission wavelengths, polarization, pulse lengths and repetition rates. We have measured the laser power with a germanium photoconductor and compared the results to a Golay detector evaluating potential artifacts. We have studied mode profiles in the far-field which mirror the internal mode structure. Potential applications have been illustrated by imaging optical elements and a simple test object. Video rate room temperature imaging has been demonstrated in concept.
©2006 Optical Society of America
The development of powerful terahertz (THz) radiation sources and new detection schemes  has opened up new research fields and applications [2, 3, 4, 5, 6] between the microwave and infrared regions, in the so-called THz-gap. Semiconductor based table-top THz sources include time-domain THz systems , quantum cascade lasers (QCLs) , germanium lasers  and THz generation by photomixing .
In chemistry, THz spectroscopy could answer important questions. THz frequencies correspond to relevant energies in molecular rotation and intermolecular modes of large biomolecules in the gas and liquid phase, i.e., in water . The spectroscopic THz fingerprint of molecules, substances and tissue allowed new applications in pharmacy and medical imaging .
Widespread applications of THz laser devices ideally require low-cost turn-key and room temperature operation. Photomixers driven by optical lasers are potential sources but they decrease rapidly in intensity above 1 THz . However, new photomixing concepts based on four-wave mixing in two-color lasers promise increasing output up to 4 THz .
Semiconductor THz lasers directly convert applied electrical power into THz photons, resulting in both small and turn-key devices. QCLs have been recently established as laser sources in the infrared  and THz spectral range [7, 12, 13]. Recently, single line emission and linewidths of less than 100 kHz were reported for continuous wave THz QCLs operated at less than 15 K (Ref.  and references within). They could be used as local oscillators in heterodyne receivers for high resolution spectroscopy replacing gas laser systems. The detection of beat signals in multimode QCLs enables THz imaging applications with a high dynamic range of up to 60 dB . The THz image acquisition time is typically long due to raster-scanning methods. Recently, THz imaging with a gas laser was demonstrated by using a microbolometer focal plane array (FPA) . In future, scanning methods may become obsolete if a microbolometer FPA can be used with a QCL light source, thus allowing THz imaging with an all solid state device at video rates.
Peltier cooling is sufficient to maintain pulsed and even continuous wave laser output in the infrared but not yet in the THz frequency range. Nevertheless, operation temperatures up to 164 K have been demonstrated . Increasing output power, e.g., in excess of 100 mW at 80 K , promises even higher operation temperatures. Device performance is expected to depend on both design and MBE growth. A detailed discussion is given in Ref. .
Although THz QCLs require liquid nitrogen (LN2) cooling it is an easily accessible and cheap commodity in laboratory, biochemical, medical and clinical laboratory environments. To ease access to THz frequencies a small LN2-cooled, turn-key THz QCL system containing two QCLs has been developed (Fig. 1). We demonstrate that the power output of high temperature THz QCLs, although the lasers have not been tailored for best performance, is sufficient for applications with a room temperature detector. We will show that it is a convenient-to-use device avoiding the substantial costs of a closed-cycle machine. As a reference, we have used a liquid helium (LHe) germanium photoconductor (Ge PC) to study single laser pulses. The emission frequencies of both THz QCLs have been studied with a grating spectrometer. Imaging was used to determine the mode profiles and the beam spread. To illustrate potential applications we have imaged the exit aperture of a reflective cone and a test object. In addition, we simulated single pixel video rate room temperature detection.
2. Experimental set-up
The investigated Quanta-Tera system from Laser Components (Fig. 1) consists of two THz QCLs fabricated by the University of Neuchâtel. Later designated as laser L1 and L2. The sample N314 has 120 repetitions of the active region and is described in detail elsewhere . However, the samples used here have no high reflectivity (HR) coating which will result in lower emission power levels in comparison to Ref. . The total thickness from the top of the ridge to the lateral contact is 14.62 μm. The ridge width of L1 is 110 μm wide. L2 has a ridge width of 210 μm. The lasers are cleaved to 1.7 mm length. L1 and L2 are installed side by side on the cold stage of the dewar centered at a 25.4 mm diameter, 1 mm thick quartz window. The distance between the laser facet and the window was 4.6 mm.
The pulse driver is connected to the dewar via a SMA connector. The driver generates voltage pulses with lengths of up to 300 ns and repetition rates from 0.87 to 200 kHz. The dewar is filled with LN2. Precooling and filling needs 300 ml LN2 and takes less than 5 minutes. The lasers have been used with the default limit settings for current, voltage and pulse width. The voltage monitor, current monitor and the trigger signal are accessible through SMB connectors. An external frequency trigger signal can be provided in the trigger line if one of the control lines of the driver is set to a LOW (0 V) signal. Unless otherwise stated, the lasers were operated at maximum output power with a maximum excitation pulse length Δt of 300 ns. Maximum power was obtained at a voltage U of 10.7 V and current I of 1.18 A for L1. L2 required a voltage of 12.2 V and a current of 2.20 A related to the wider ridge width. These optimized values were initially fixed at the driver allowing turn-key operation for the imaging experiments on a day-to-day basis by simply switching on the power cord adapter of the pulse driver.
A room temperature pneumatic detector (QMC Instruments Ltd.), originally invented by M.J.E. Golay in 1947, was used to detect the THz emission, since the Golay cell can achieve thermal background limited performance with a typical noise equivalent power (NEP) of 2×10-10 W Hz-1/2. A white high density polyethylene (HD-PE) window of the Golay cell defines an aperture of 6 mm diameter. The HD-PE window transmission decays linear in the region of interest from 81% at 3.5 THz to 79% at 4.0 THz. The QCLs were pulsed at very low repetition rates to measure single pulse intensities. We used electronic chopping by low frequency modulation (50% duty) of a high frequency laser pulse train to obtain a high responsivity and signal-to-noise ratio (SNR) for imaging.
LHe cooled photoconductors and composite bolometers are very sensitive direct detectors with a NEP of typically 10-14 W Hz-1/2 at 4.2 K. We chose a Ge PC to achieve a sub-μs time resolution for the study of single laser pulses. The Ge detector crystal with a cross-section of 4×4 mm2 was mounted in a LHe cryostat. The THz emission incident on the Ge PC was filtered by a white 3 mm thick HD-PE vacuum window and two 100 μm thick cold filters of black low density PE at 77 K and 4.2 K. The LHe container has a volume of 1.2 liters which allows three days’ continuous operation. The detector is biased in the linear part of its I-V-curve (UBIAS = 0.39 V, IBIAS = 0.32 mA).
We have used the Ge PC if the distance between source and detector exceeded 50 mm. A high speed variable voltage amplifier (FEMTO DHPVA-200, 200 MHz, 10-60 dB), typically set at 60 dB, amplifies the signal further for source-detector distances up to 500 mm. The full pulse trace is read-out from a high speed oscilloscope (HP Infinium 1.5 GHz, 8 GSa/s) via a National Instruments’ GPIB PCI-card. A LabView program handles the data acquisition and controls the DC motor-driven stages for the grating spectrometer (Physik Instrumente M-038.DG) as well as the x-y imaging stages (two PI M-415.DG). The small Quanta-Tera system was mounted on the x-y stages for all imaging experiments scanning the laser with the detector fixed. We have also used phase sensitive detection by employing a lock in amplifier (Ametek DSP 7265). The QCL trigger frequency f or the electronic chopping frequency fM served as the reference signal.
3. Single pulse detection
Single pulses can be immediately visualized on the oscilloscope without further amplification with a Ge PC if the distance between the QCL and detector is small (Fig. 2(a)). The used Ge PC detector is too slow (time constant ≈ 300 ns) to monitor the pulses directly on a ns-scale. However, we are able to determine the pulse shape by analyzing the integral of the laser pulse trace I(t) versus different voltage pulse lengths U(t). For convenient data reduction we have used the integrals of each near-square voltage pulse U(t) as a measure for the excitation time. The data points in Fig. 2(b) have been calculated from single pulses. A linear function is found with a slight oscillation around the slope. This is caused by our data reduction process due to ringing effects in the voltage raw data pulse U(t). The linear behavior of the integral indicates a constant intensity versus time and a square-like shape of the laser pulse. Zero intensity is obtained at 13 ns. We assume that the current in the laser reaches operating conditions only 13 ns after the excitation. This is indicated by the ringing in the voltage and current pulse. A better impedance matching of driver and QCL can reduce the delay time. However, the laser pulse can be shortened to a few nanoseconds and less by adjusting the driving pulse to be slightly longer than 13 ns.
From the single pulses measured with the Ge PC we can immediately estimate the collected power. We obtain 0.14 mW due to the Ge PC responsivity of 300 V/W and the measured integral of 12 nVs (Fig. 2) corresponding to 42 mV × 287 ns. For comparison, we have used the Golay detector to estimate the power level of a single pulse. In this case we did not use electronic chopping to avoid time constant artifacts observed previously with a germanium laser. A Ge laser has a tuning range from 1 to 4 THz for a single device [20, 21, 22] and is typically much larger than QCLs. The laser delivers output powers of up to 1 W for a moderate crystal size of 30×4×3 mm3 while operated at I = 27 A (30 A/cm2), U = 400 V (1 kV/cm) and below 25 K in a closed-cycle machine . With the same Golay cell we obtain, in close proximity to the Ge laser, a peak signal of up to 30 mV for a single emission pulse. This signal can be directly viewed on an oscilloscope. Although the responsivity of the Golay cell decays with repetition frequency we can still observe Ge laser pulses up to several 100 Hz without electronic chopping. Using electronic chopping with pulse repetition rates up to several kHz showed that the detected signal decays after each pulse even before the next pulse arrives. This effect may underestimate the peak power if electronic chopping is used. The detector rise time for each pulse is approximately 0.7 ms (10% to 90% value) and the decay time approximately 100 ms. Therefore, L1 was triggered at 10 Hz emitting single pulses at 100 ms intervals. The Golay cell with a white HD-PE window detects both infrared and THz radiation due to its broad spectral coverage. To avoid further artifacts due to thermal effects we have raster-scanned the laser emission in a square area of 20 mm × 16.5 mm with a step-size of 0.25 mm corresponding to 80 × 66 pixel. The output of the Golay cell amplifier is connected to the input of the lock in amplifier (LIA). The LIA integration time was set to two seconds. A data point was read-out after a delay of 2.6 seconds following each step. This allows to eliminate artifacts by possible mechanical and acoustic vibrations. The LIA output was fed through a National Instruments’ analog/digital converter (PCI-6013). Figure 3(a) shows two measurements, with just the Golay cell (upper row of images) and with an additional black LD-PE filter (lower row of images). The laser intensity distribution corresponds to the 6 mm diameter Golay aperture (Fig. 3(a)). We have virtually increased the integration time by binning 4 by 4 pixel (Fig. 3(b)) to give a total integration time of 32 seconds per pixel, thereby reducing the resolution to 1 mm. Left and right of the laser emission we observed additional bright spots. The spots can be completely removed by inserting a black LD-PE filter (cf. images in lower row of Fig. 3). Consequently we attribute these hot spots to thermal infrared radiation, due to power dissipation in the adjacent laser circuitry. The filter efficiently suppresses the infrared thermal radiation, reducing the detected THz signal by 15% to 20%. In all further imaging experiments we have attached the black PE filter in front of the Golay detector.
The detected THz power per pulse can now be estimated. The background signal is approximately 5 μV, giving a THz signal of 17 μV due to the laser emission of L1. With the Golay cell responsivity of 32 kV/W at 10 Hz (NEP of 3.2×10-10 W Hz-1/2), this value corresponds to 530 pW. From Fig. 2 we obtained a square-like pulse of 287 ns length corresponding to duty cycle of 2.9× 10-6. In the best case the whole pulse energy is absorbed in the detector. We obtained 0.19 mW for the laser pulse at an applied electrical power of 12.6 W. The ratio of the collected to applied power is 1.5×10-5.
4. Heat management
We have studied single laser pulses of L1 as a function of the repetition rate f (Fig. 4(a)) and the applied electrical power respectively. For comparison, we measured modulated pulse trains with the Golay cell. The modulation frequency was 10 Hz. The average applied electrical power is given by Pel = U × I × f × Δt. The maximum duty cycle of the laser dc = f × Δt is determined by the heat management influenced by the device geometry  and its connection to the LN2 bath. The intensity is approximately constant up to dc = 0.03% and decreases to 1/2 at dc = 0.9%. At a dc = 2.1% the slope changes sharply, indicating that the laser is too hot to maintain laser operation. As a consequence only spontaneous emission is detected.
Devices cooled with cryogenic liquids are only convenient if the refilling intervals are long. They can be stable in performance if the device can run over several hours to establish thermal equilibrium with the mount and the LN2 bath (Fig. 4(b)). With an enthalpie of evaporation for LN2 of 5.6 kJ/mol at 77 K we evaporate 22 ml/Whr. At a maximum duty cycle of ≈ 3% with an average electrical power dissipation of 0.38 W we evaporate 8 ml/hr, due to operation of L1. Therefore a container (200 ml) of liquid nitrogen supports the laser operation for 24 hours. In practice, the dewar of the Quanta-Tera was cold for at least 12 hours for one LN2 filling at typical operating duty cycles of 0.03% (870 Hz × 300 ns). The evaporation is mainly determined by thermal radiation of the dewar walls. The 295 K black body radiation of the 20 mm window incident on the 77 K laser mount already introduces 0.13 W.
5. Spectral measurements
The spectra of both lasers have been studied with a grating spectrometer (Fig. 5). The divergent laser beams were formed into parallel beams by attaching a plane-convex HD-PE lens to the exit window of the small LN2 dewar, securing an even illumination of the diffraction grating. The focal length of the lens was 50 mm. The grating was blazed at an angle of 20° with a grating constant of 0.1 mm . Both distances, between laser and grating and between detector and grating, were approximately 200 mm. The spectral resolution is limited by the 4 mm wide Ge detector crystal serving as an entrance slit. In one measurement we inserted an additional slit. The slit increases the resolution of the spectrometer to less than 1 cm-1 or 0.03 THz. The detected laser signal was amplified by 60 dB and 256 single pulses were averaged in the oscilloscope before read-out. The lasers emit in the frequency range from 3.5 to 4 THz with L1 covering the high frequency part. With increasing voltage and current we observe that the laser emission shifts to higher frequencies. The emission is attenuated by water vapor absorption lines. For comparison, the theoretical water line spectrum , with appropriate broadening for 1000 hPa at room temperature, is displayed in Fig. 5. The spectrum with higher resolution using a narrower slit leads to a laser mode spacing of 24.0 GHz. This spacing corresponds to a length of 1.74 mm and a refractive index of 3.6. A strong absorption line of quartz at 76 μm = 132 cm-1 may suppress higher frequency modes. We tested another quartz window identical to the window installed in the dewar. The transmission in the laser emission range is approximately 75% due to the refractive index of quartz n = 2.1 and subsequent reflection losses.
In chemistry, a polarized light source is desirable. For example, two beam spectrometers for small difference measurements use polarizers to split and recombine polarized beams. The efficiency of a grating spectrometer also depends on the polarization of the light source. We have measured the polarization of both lasers by inserting a polarizing wire grid between the lasers and detector (Fig. 6). The free standing wire grid consists of 25 μm diameter stainless steel wires. We have plotted an ideal polarizing curve. Removing the wire grid gives the maximum corresponding to 100% of the laser intensity. Blocking the laser path gives the 0% intensity. L2 follows the ideal curve more closely than L1. We attribute this to longer emitted wavelengths of L2 (cf. Fig. 5); in this case the polarizing function of the grid is better. The emission of both lasers is highly polarized, since the deviations from the ideal curve are mainly due to the non-ideal wire grid.
7. Mode profiles
We have analyzed the beam profiles by imaging the emission at different distances between the laser and detector (Fig. 7). The emission in plane of the laser facet for different distances and both lasers is shown. The far-field intensity distribution follows closely the internal mode structure (Fig. 7(a)). The highest intensity should appear at the interface of the active layer and the substrate . Two beams from the active region and the substrate originate from a near point source. The beams are diffraction limited and depend on the laser dimensions and the emitted wavelengths . The emission is disturbed with increasing distance to the facet. We attribute these artifacts to the laser and detector windows. The line data for the mode profiles of Fig. 7(e) has been calculated by integrating the image points along the y direction of the image. It can be seen that the maximum beam intensity decreases due to spreading of the laser pulse energy in space. The profile integral over the whole area decreases only slightly. We can attribute this intensity reduction to water vapor absorption along the path due to air humidity. Although this absorption can be very strong it only affects specific frequencies (cf. Fig. 5), due to the quantized nature of the rotational transition frequencies of the water molecules . Since we measure the integral over a wide frequency range we have an average absorption coefficient of 0.03 cm-1. The integral values are given in Table 1. Analyzing the beam spread of the multimode emission of each image we can calculate a half angle at which the intensity drops to 1/e of 16° and 21° for L1 and L2, respectively. These values correspond to angles obtained for QCLs in Ref. . The image features at a distance D of 133 mm (Fig. 7(c) and (f)) are due to multiple reflections within and between the dewar windows. Nevertheless, the general shape of the internal mode profile is recovered. Arrows indicate the corresponding features. The wider ridge of L2 seems to result in a smaller separation of the two beams.
8. THz imaging applications
Reflective optics are preferred in THz imaging applications. They avoid additional losses and standing waves typically found with lenses. Cones focus light along the optical axis. In astronomy, they are often used to map light from a telescope onto a THz detector array. Introducing a small wire into the cone forms a coaxial waveguide which allows sub-wavelength near-field imaging . Winston cones avoid the critical off-axis distance of a linear cone . We have used L1 to evaluate a fabricated Winston cone. In general, Winston cones are used to collect all wavelengths entering the entrance aperture through the exit aperture . The Winston cone formula is given here for reference as we noticed a different representation in Ref. . We chose an entrance aperture diameter of 2rin = 10 mm and an exit aperture of 2rout = 2.4 mm. For simplification we define the ratio R = rout/rin = 0.24. The geometric field of view is 2θ = 2arcsin(R) = 28°. By solving Eq. 1 we can find the necessary curve r(z) for the cone:
The curve has been transferred to a CNC-lathe with a diameter 2r coordinate resolution of 1 μm. The length of the cone is 25 mm. The entrance aperture of the cone was fixed to the quartz window to collect the entire intensity of L1. We scanned L1 and subsequently the exit aperture of the Winston cone while the Ge PC was fixed in space. A 1 mm diameter aperture was fixed to the Ge PC window to improve resolution. Figure 8(a) shows an image of the exit aperture of the Winston cone. We used the lock in amplifier at a sensitivity of 5 μV with an integration time of 5 ms per pixel operating L1 at 10 kHz. The complicated fine structure of the transformed THz laser far-field pattern contains dimensions of the order of one wavelength. Analyzing the Fourier-transform of the image gives characteristic frequencies corresponding to dimensions within the interference structure: 2.4 and 1 mm equal to cone exit and 1 mm diameter aperture, vertical lines tilted by -24° from the horizontal axis with period 200 μm attributed to non-perfect window wedges and finally, concentric rings with 110 μm period which coincide with the ridge width of L1, respectively.
Another application is THz imaging of objects. Typical imaging set-ups raster-scan the sample while the focal point is fixed. The advantages are diffraction limited resolution and, more importantly, a high intensity in the focal point. Subsequently a good SNR is obtained. However, if THz lasers are considered as potential sources for medical imaging of patients, source and detector have to be compact and mobile. Our dewar and pulse driver are small and have been mounted on the translation stage for all experiments presented here. Therefore, we evaluate in the following THz imaging of immobile objects.
The detector and the object have been fixed in place. One HD-PE lens was fixed in front of the Ge PC detector to increase the field of view. Another lens was fixed to the window of L1 to form a wide parallel laser beam. Detector and laser were separated by approximately 500 mm.
As a test object we chose letters printed on a Xerox laser printer transparency foil. First, the intensity distribution I 0(x,y) without the test object was imaged by scanning the laser system (Fig. 9(a)). It corresponds to the far-field of the mode profile disturbed by the optical properties of the lenses. The profile is mirrored vertically in comparison to the profiles displayed in Fig. (7 (c)), since we set the viewing direction to the detector and not to the laser. As an indicator we reversed the horizontal arrow in the inset in comparison to previous figures. The object (Fig. 9(d)) was fixed between the laser and the detector. Again, the intensity distribution I(x,y) was measured. It resembles the mode profile but attenuated by the object (Fig. 9(b)). Calculating I(x,y)/I 0(x,y) gives the transmission of the object at each coordinate (x,y) (Fig. 9(c)). The 100 μm thick Xerox foil transmits approximately 55%. The toner layer of the printed letters absorbs 5%. The letters have a size of 500 μm and less. Due to the large wavelength we observe diffraction at the borderline of the printed lines. A double line feature appears, whereas the object has a single circle line.
The last experiment, in principle, demonstrates that room temperature detectors allow video rate imaging when using high temperature QCLs. The Golay cell can be considered as a single pixel. A signal can be integrated for 20 ms at a video rate of 50 Hz. First, we used electronic chopping of a 20 kHz pulse train to evaluate the response of the Golay cell up to 1 kHz (Fig. 10(a)). Even at a modulation frequency fM of 1 kHz a signal of 41 μV is still detectable.
Finally, we chopped a 90 kHz pulse train at a modulation frequency fM = 51 Hz. The Golay cell is placed 9 mm from the laser facet. The narrow laser beam of L1 (cf. Fig. 7(a)) is scanned across the entrance aperture of the Golay cell. We essentially obtain the detector window area of 6 mm diameter. The Golay window is placed at the end of a short (5 mm) cone as part of the manufactured detector. This leads to a smooth onset of detected laser light outside the 6 mm aperture. The peak detected power is approximately 36 times larger than the value of the beam profile measured in Fig. 7(a) if we scale by the ratio of the used modulation frequencies 50:10. This value corresponds to the reduced aperture of 1 mm diameter which was used for beam profiling (Fig. 7(a)). The SNR at 20 ms integration time is 85. The total data acquisition time was 46 seconds. Focal plane arrays with similar or better characteristics per pixel in comparison to the Golay cell should allow video rate imaging with our small LN2-cooled QCLs.
Two high temperature THz QCLs have been mounted in a small liquid nitrogen-cooleddewar. A careful estimate of the power level has been performed. We obtained an emission peak power of approximately 0.2 mW at 77 K for the smaller laser L1 with no high reflection coating leaving room for improvements. A fair coverage of the THz spectrum from 3.5 to 4 THz has been obtained by using both lasers. The spectrum consists of modes which are spaced by 24 GHz. We observed attenuation due to broad spectral water lines for which the 24 GHz frequency comb can be ignored. The THz emission power is sufficient for a variety of THz applications. We demonstrated the characterization of optical elements, detector systems and imaging. The lasers show ns-scale pulses which could be used to test fast THz detector circuits. The laser mode profile has been found to follow closely the internal mode structure in intensity and orientation. Beam distortion due to optical elements has been observed for larger distances from the facet. The concept of video rate imaging with the small dewar and a room temperature Golay cell has been demonstrated at a rate of 50 Hz. All THz imaging experiments have been performed by moving and scanning the compact laser system. The device has been proven to be small, light-weight and mobile enabling imaging applications on immobile larger objects. A future challenge is the integration of source and detector in a single compact measurement tool to allow THz reflection experiments, for example, to evaluate THz contrasts between normal and cancerous tissue in-vivo. Recent progress in the design and MBE growth promise increasing THz laser power at elevated temperatures. Thus integration times for a good signal-to-noise ratio are reduced, which may allow THz room temperature emission and detection at video rates, in future.
We would like to thank A. Lindner and his team from the precision engineering of the chemistry faculty workshop at Bochum, who fabricated the HD-PE lenses and the Winston cone. In addition, we acknowledge financial support from the Ministry of Science and Research of North Rhine Westfalia. The work has been performed within the UZMT (university center for medical technology) of the Ruhr-University Bochum.
References and links
1. P. H. Siegel, “Terahertz Technology,” IEEE Trans. Microwave Theory Tech. , MTT-50, 910–928 (2002). [CrossRef]
3. S. Ebbinghaus, K. Schröck, J.C. Schauer, E. Bründermann, M. Heyden, G. Schwaab, M. Böke, J. Winter, M. Tani, and M. Havenith, “Terahertz time-domain spectroscopy as a new tool for the characterisation of dusty plasmas,” Plasma Sources Sci. Technol. 15, 72–77 (2006), [CrossRef]
4. T. Löffler, T. Bauer, K. J. Siebert, H. G. Roskos, A. Fitzgerald, and S. Czasch, “Terahertz dark-field imaging of biomedical tissue,” Opt. Express 9, 616–621 (2001), http://www.opticsexpress.org/abstract.cfm?URI=OPEX-9-12-616. [CrossRef] [PubMed]
5. J. Darmo, V. Tamosiunas, G. J. Kröll, K. Unterrainer, M. Beck, M. Giovannini, J. Faist, C. Kremser, and P. Debbage, “Imaging with a Terahertz quantum cascade laser,” Opt. Express 12, 1879–1884 (2004), http://www.opticsexpress.org/abstract.cfm?URI=OPEX-12-9-1879. [CrossRef] [PubMed]
6. M. Yamashita, K. Kawase, C. Otani, T. Kiwa, and M. Tonouchi, “Imaging of large-scale integrated circuits using laser terahertz emission microscopy,” Opt. Express 13, 115–120 (2005), http://www.opticsexpress.org/abstract.cfm?URI=OPEX-13-1-115. [CrossRef] [PubMed]
7. R. Köhler, A. Tredicucci, F. Beltram, H. E. Beere, E. H. Linfield, A. G. Davies, D. A. Ritchie, R. C. Iotti, and F. Rossi, “Terahertz semiconductor-heterostructure laser,” Nature 417, 156–159 (2002). [CrossRef] [PubMed]
8. A. Bergner, U. Heugen, E. Bründermann, G. Schwaab, M. Havenith, D. R. Chamberlin, and E. E. Haller, “New p-Ge THz spectrometer for the study of solutions: THz absorption spectroscopy of water,” Rev. Sci. Instrum. 76, 06310–999 (2005). [CrossRef]
9. C. Baker, I. S. Gregory, M. J. Evans, W. R. Tribe, E. H. Linfield, and M. Missous. “All-optoelectronic terahertz system using low-temperature-grown InGaAs photomixers,” Opt. Express 13, 9639–9644 (2005), http://www.opticsexpress.org/abstract.cfm?URI=OPEX-13-23-9639. [CrossRef] [PubMed]
10. S. Hoffmann, M. Hoffmann, E. Bründermann, M. Havenith, M. Matus, J. V. Moloney, A. S. Moskalenko, M. Kira, S. W. Koch, S. Saito, and K. Sakai, “Four-wave mixing and direct terahertz emission with two-color semiconductor lasers,” Appl. Phys. Lett. 84, 3585–3587 (2004). [CrossRef]
12. L. Ajili, G. Scalari, J. Faist, H. E. Beere, J. Fowler, E. H. Linfield, D. A. Ritchie, and A. G. Davies, “High power quantum cascade lasers operating at λ ≈ 87μm and 130 μm,” Appl. Phys. Lett. 84, 3986–3988 (2004). [CrossRef]
13. C. Worrall, J. Alton, M. Houghton, S. Barbieri, H. E. Beere, D. Ritchie, and C. Sirtori, “Continuous wave operation of a superlattice quantum cascade laser emitting at 2 THz,” Opt. Express 14, 171–181 (2006), http://www.opticsexpress.org/abstract.cfm?URI=OPEX-14-1-171. [CrossRef] [PubMed]
14. H.-W. Hübers, S. G. Pavlov, A. D. Semenov, R. Köhler, L. Mahler, A. Tredicucci, H. E. Beere, D. A. Ritchie, and E. H. Linfield, “Terahertz quantum cascade laser as local oscillator in a heterodyne receiver,” Opt. Express 13, 5890–5896 (2005), http://www.opticsexpress.org/abstract.cfm?URI=OPEX-13-15-5890. [CrossRef] [PubMed]
15. S. Barbieri, J. Alton, C. Baker, T. Lo, H. E. Beere, and D. Ritchie, “Imaging with THz quantum cascade lasers using a Schottky diode mixer,” Opt. Express 13, 6497–6503 (2005), http://www.opticsexpress.org/abstract.cfm?URI=OPEX-13-17-6497. [CrossRef] [PubMed]
17. B. S. Williams, S. Kumar, Q. Hu, and J. L. Reno, “Operation of terahertz quantum-cascade lasers at 164 K in pulsed mode and at 117 K in continuous-wave mode,” Opt. Express 13, 3331–3339 (2005), http://www.opticsexpress.org/abstract.cfm?URI=OPEX-13-9-3331. [CrossRef] [PubMed]
18. B. S. Williams, S. Kumar, Q. Hu, and J. L. Reno, “High-power terahertz quantum-cascade lasers,” Elec. Lett. 42, 3331–3339 (2006). [CrossRef]
19. G. Scalari, N. Hoyler, M. Giovannini, and J. Faist, Terahertz bound-to-continuum quantum-cascade lasers based on optical-phonon scattering extraction,” Appl. Phys. Lett. 84, 3585–3587 (2005).
20. E. Bründermann, H. P. Röser, A. V. Muravjov, S. G. Pavlov, and V. N. ShastinMode fine structure of the FIR p-Ge Intervalenceband Laser measured by Heterodyne Mixing Spectroscopy with an optically pumped ring gas laser,” Infrared Phys. Technol. 36, 59–69 (1995). [CrossRef]
21. E. E. Haller and E. Bründermann, “Doping of germanium and silicon crystals with non-hydrogenic acceptors for far infrared lasers,” U.S. Patent No. 6,011,810 (January 4, 2000).
22. E. Bründermann, “Widely Tunable Far Infrared Hot Hole Semiconductor Lasers”, in Long-wavelength Infrared Semiconductor Lasers, edited by Hong K. Choi (Wiley& Sons, New York, 2004), Chapter 6, pp. 279–350.
23. F. Matsushima, H. Odashima, T. Iwasaki, S. Tsunekawa, and K. Takagi, “Frequency measurement of pure rotational transitions of H2O from 0.5 to 5 THz,” J. Molecular Structure 352/353, 371–378 (1995) [CrossRef]
24. R. Sachs and H. G. Roskos, “Mode Calculations for a Terahertz Quantum Cascade Laser,” Opt. Express 12, 2062–2069 (2004), http://www.opticsexpress.org/abstract.cfm?URI=OPEX-12-10-2062. [CrossRef] [PubMed]
25. F. Keilmann, “FIR microscopy,” Infrared Phys. Technol. 36, 217–224 (1995). [CrossRef]
26. E. W. Weisstein, “Winston Cone,” http://scienceworld.wolfram.com/physics/WinstonCone.html.