TERAHERTZ DETECTORS: Quantum dots enable integrated terahertz imager

A highly sensitive and frequency tunable terahertz detector based on carbon nanotube quantum dots is the main ingredient for a near-field terahertz imaging system in which all components—an aperture, a probe, and a detector—are integrated on one semiconductor chip.

Th Quantum 01

A highly sensitive and frequency tunable terahertz detector based on carbon nanotube quantum dots is the main ingredient for a near-field terahertz imaging system in which all components—an aperture, a probe, and a detector—are integrated on one semiconductor chip.

YUKIO KAWANO

Technologies based on terahertz waves are currently in strong demand in many fields including biochemistry, medicine, environmental science, and security. Advantageous properties such as terahertz transmission through objects opaque to visible light are attracting much attention for spectroscopy and imaging applications. Terahertz measurements inevitably require a highly sensitive detector to obtain distinct spectra and images. Nevertheless, the photon energy of the terahertz wave, on the order of millielectron volts (meV), is two to three magnitudes lower than that of the visible light, making the development of a high-performance terahertz detector a difficult task. Another problem with terahertz detection is low spatial resolution of terahertz imaging, which results from the longer wavelengths of terahertz radiation compared to that of visible light. Fortunately, the application of nanoscale materials and devices is opening up new opportunities to overcome these difficulties.

Th Quantum 01
FIGURE 1. Source-drain current vs. gate voltage is observed with and without terahertz irradiation for a carbon nanotube quantum-dot (CNT-QD) device (right; device structure in upper right). The energy spacing (¿ΔVG) between the original and satellite peaks as a function of the photon energy of the terahertz wave is also plotted (above). (Courtesy of RIKEN)
Click here to enlarge image

Two types of emerging terahertz detectors are based on novel nanoelectronic technologies. The first is a highly sensitive and frequency tunable terahertz detector based on a carbon nanotube (CNT) quantum dot (QD).1 We found that terahertz irradiation generates new CNT currents and its peak position relative to the gate voltage linearly depends on the photon energy of the incident terahertz wave. These observations provide direct evidence of terahertz photon-assisted tunneling, demonstrating that the CNT-QDs can be used as a frequency-tunable terahertz photon detector. The second type of detector is a near-field terahertz detector for high-resolution imaging. Contrary to the situation in the microwave and visible-light region, the development of near-field imaging in the terahertz region has not been well established. We have developed a new device for near-field terahertz imaging in which all components—an aperture, a probe, and a detector—are integrated on one gallium arsenide/aluminum gallium arsenide (GaAs/AlGaAs) chip.2 This scheme allows highly sensitive detection of the terahertz evanescent field alone, without requiring optical or mechanical alignment.

CNT frequency-tunable detector

The device structure for a terahertz detector fabricated from single-wall CNTs consists of source, drain, and gate electrodes made from titanium/gold (Ti/Au) films (see Fig. 1). These CNT-QD devices were immersed into a cryostat and conductivity measurements performed at 1.5 K with and without terahertz irradiation. We observed periodic oscillations of source-drain current versus gate voltage, indicating that the devices work as single-electron transistors.

Terahertz irradiation of these devices cause the generation of new-satellite peaks in the Coulomb blockade regime, and the energy spacing between the new-satellite and original peaks is proportional to the photon energy of the incident terahertz wave. These observations are clear evidence of electron tunneling via terahertz-photon detection, called photon-assisted tunneling; this result means that the CNT-QD structure can be utilized as a frequency tunable terahertz detector.

We observed that the CNT-QD detector functions properly up to approximately 7 K. This performance eliminates the use of a dilution or 3He refrigerator with complex systems and low cooling capacity. The CNT-QD detector, therefore, can be used in a compact cryo-free mechanical refrigerator with much greater ease of use and much higher cooling capacity, allowing extended temperature-range operation for ultrasensitive terahertz measurements. Furthermore, the CNT-QD device has the advantage of room-temperature operation of a single-electron transistor, thanks to its large single-electron charging energy. We anticipate that higher-temperature operation of the CNT-QD terahertz detector is also possible with more refined fabrication techniques.

The next important step is to improve detector performance in two important ways: sensitivity and frequency selectivity. A much more sensitive readout of the terahertz-detected signal could be achieved by capacitively coupling a CNT-QD with a quantum point contact device on a GaAs/AlGaAs heterostructure, which makes it possible to observe single-electron dynamics. And frequency selectivity could be improved by using a double-coupled CNT-QD, in which photon-assisted tunneling takes place as a result of electron transitions between two well-defined discrete levels.

On-chip terahertz detection

Two approaches can be used to achieve high spatial resolution in optical imaging: a solid immersion lens and near-field imaging. Though we have previously constructed a terahertz imaging setup based on a solid immersion lens, its resolution is restricted by the diffraction limit.3 A powerful method for overcoming the diffraction limit is the use of near-field imaging. This technique has been well established in visible and microwave regions using either a tapered, metal-coated optical fiber or a metal tip, and either a waveguide or a coaxial cable. However, the development of near-field imaging in the terahertz region has been hindered by the lack of terahertz fibers or other bulk terahertz-transparent media suitable for generating near-field waves, as well as the low sensitivity of commonly used detectors in the terahertz region.

Th Quantum 02
FIGURE 2. An optical micrograph (left) and a schematic representation (right) shows the design of a highly sensitive on-chip near-field THz detector. The 8-µm-diameter aperture and planar metallic probe, each of which is insulated by a 50-nm-thick silicon dioxide (SiO2) layer, are deposited on the surface of a GaAs/AlGaAs heterostructure chip. (Courtesy of RIKEN)
Click here to enlarge image

In conventional near-field imaging systems, the propagation field arising from the scattering of the near-field (evanescent) wave is measured with a distant detector, which requires detecting very weak waves (and the influence of far-field waves is unavoidable). In contrast, our near-field terahertz imager places the aperture, probe, and detector in close proximity. The 8-µm-diameter aperture and planar probe, each of which is insulated by a 50-nm-thick silicon dioxide (SiO2) layer, are deposited on the surface of a GaAs/AlGaAs heterostructure chip (see Fig. 2). Because integration with the CNT-QD detector requires improvements in the device fabrication process (specifically, by using higher-performance electron-beam lithography equipment), a two-dimensional electron gas (2DEG)—located only 60 nm below the chip surface—is used as the terahertz detector. This detector utilizes conductivity changes (photoconductivity) induced by the terahertz absorption of the high mobility 2DEG.4 The presence of the planar probe changes the distribution profile of the evanescent field, enhancing the coupling of the evanescent field to the 2DEG detector. This leads us to expect that the presence of the probe will result in strongly enhanced terahertz transmission through the small aperture, increasing the detection sensitivity. Moreover, the 2DEG detector is not affected by the far-field wave owing to the close vicinity to the aperture and the probe, allowing the detection of the evanescent field alone.

Th Quauntum 03
FIGURE 3. Comparisons are made between two types of near-field terahertz imagers: one with just the aperture alone (top left), and the other with aperture plus a probe (top right). Analysis of the electric field distributions near the aperture as the detector is scanned across a structure consisting of alternate layers of terahertz-opaque and terahertz-transparent materials (bottom) show that for the aperture plus the probe, a clear square-wave profile corresponding to the material structures is observed, while for the aperture only, no signal is observed. These results indicate significant enhancement in the detection sensitivity of the terahertz evanescent field due to the existence of the planar probe. (Courtesy of RIKEN)
Click here to enlarge image

We first made calculations of terahertz electric field distributions near the aperture using a finite element method. The spatial distribution of the evanescent field is strongly enhanced by the probe, confirming that detection sensitivity is enhanced. We then measured the terahertz transmission profile by scanning the device across a sample. The sample is made up of a terahertz-transparent substrate, the surface of which is covered at regular intervals by terahertz opaque gold films. The widths of terahertz-opaque and terahertz-transparent regions across the scan direction are 80 and 50 µm, respectively. For comparison, we used two kinds of near-field terahertz devices; the aperture plus the probe and the aperture alone (see Fig. 3). For the aperture plus the probe, a clear square-wave profile is observed corresponding to the terahertz-transparent and terahertz-opaque regions of the sample. For the aperture alone, no signal is observed. These results show proof of significant enhancement in the detection sensitivity of the terahertz evanescent field due to the existence of the planar probe.

From a signal decay curve, we obtain a spatial resolution of 9 µm. The resolution does not depend on the wavelength of the incident terahertz wave, almost matches the aperture diameter of 8 µm, and corresponds to λ/24 for the wavelength λ = 214.6 µm. These facts clearly demonstrate that the device properly functions as a near-field terahertz detector.

Future prospects

One of the challenges for future terahertz sensing technology is to achieve high detection sensitivity and high spatial resolution simultaneously. To realize this, we are now trying to combine the two techniques described above; namely to modify the CNT-QD terahertz detector into a similar structure for near-field detection. Compared to the 2DEG detector, the CNT detector exhibits much higher sensitivity and has a much smaller sensing area (approximately 200 nm compared to 8 µm for the 2DEG detector). This detector, integrated with an aperture and a probe, would show ultrahigh sensitivity and nanometer resolution simultaneously.

We further expect that when many CNTs are integrated in a two-dimensional configuration, the resulting device will serve as a real-time, high-resolution terahertz imaging detector; in effect, a terahertz video camera.

REFERENCES

  1. Y. Kawano et al., J. Appl. Phy. 103, p. 034307-1-5 (2008).
  2. Y. Kawano and K. Ishibashi, Nature Photonics 2, p. 618 (2008).
  3. Y. Kawano and T. Okamoto, Phys. Rev. Lett. 95, 166801-1-4 (2005).
  4. Y. Kawano et al., J. App. Phys. 89, p. 4037 (2001).

Yukio Kawano is a research scientist at RIKEN, The Institute of Physical and Chemical Research, 2-1 Hirosawa, Wako, Saitama 351-0198, Japan; e-mail: ykawano@riken.jp; www.riken.jp/lab-www/adv_device/kawano/eng/index.html.


Tell us what you think about this article. Send an e-mail to LFWFeedback@pennwell.com.

More in Detectors & Imaging