We have investigated the transport properties of one-dimensional (1D) constrictions defined by split-gates in high quality GaAs/AlGaAs heterostructures. In addition to the usual quantized conductance plateaus, the equilibrium conductance shows a structure close to 0.7(2e 2 /h), and in consolidating our previous work [K. J. Thomas et al., Phys. Rev. Lett. 77, 135 (1996)] this 0.7 structure has been investigated in a wide range of samples as a function of temperature, carrier density, in-plane magnetic field B and source-drain voltage V sd . We show that the 0.7 structure is not due to transmission or resonance effects, nor does it arise from the asymmetry of the heterojunction in the growth direction. All the 1D subbands show Zeeman splitting at high B , and in the wide channel limit the g-factor is | g |≈ 0.4, close to that of bulk GaAs. As the channel is progressively narrowed we measure an exchange-enhanced g-factor. The measurements establish that the 0.7 structure is related to spin, and that electron-electron interactions become important for the last few conducting 1D subbands.

The absorption spectrum of the explosive 1,3,5-trinitro-1,3,5-triazacyclohexane (RDX) has been measured using a conventional Fourier transform infrared spectroscopy and by terahertz pulsed spectroscopy. Seven absorption features in the spectral range of 5–120cm−1 have been observed and identified as the fingerprint of RDX. Furthermore, the spatial distribution of individual chemical substances including RDX, has been mapped out using reflection terahertz spectroscopic imaging in combination with component spatial pattern analysis. This is the terahertz spectroscopy and chemical mapping of explosives obtained using reflection terahertz measurement, and represents a significant advance toward developing a terahertz pulsed imaging system for security screening of explosives.

No abstract

We report conductance measurements of a ballistic one-dimensional (1D) wire defined in the lower two-dimensional electron gas of a GaAs/AlGaAs double quantum well. At low temperatures there is an additional structure at 0.7(2e 2 /h) in the conductance, which tends to e 2 /h as the electron density is decreased. We find evidence for complete spin polarization in a weakly disorderd 1D wire at zero magnetic field through the observation of a conductance plateau at e 2 /h, which strengthens in an in-plane magnetic field and disappears with increasing electron density. In all cases studied, with increasing temperature structure occurs at 0.6(2e 2 /h). We suggest that the 0.7 structure is a many-body spin state excited out of, either the spin-polarized electron gas at low densities, or the spin-degenerate electron gas at high densities.One-dimensional (1D) semiconductor systems can be fabricated by a variety of techniques. Some of the best quality devices, as determined by the clarity of the quantized plateaus in the conductance characteristics, are obtained by electrostatically squeezing a two-dimensional electron gas (2DEG) at a GaAs/AlGaAs interface using a split-gate defined by electron-beam lithography. 1The conductance, measured as a function of the split-gate voltage, exhibits plateaus quantized at integer multiples of 2e2 /h, a result that is well understood as the adiabatic transmission of spin-degenerate 1D subbands. However, after the last 1D subband has been depopulated, an additional structure in the conductance has been measured at 0.7(2e 2 /h). One of the most revealing properties of this so-called 0.7 structure is its evolution into the spin-split plateau at e 2 /h in a strong in-plane magnetic field. There is also an enhancement of the g-factor as the 1D carrier density is reduced. Both results suggest that there is a possible spin polarization of the 1D electron gas at zero magnetic field. Hartree-Fock calculations 3 of electrons confined in a cylindrical wire show that correlation effects are weak, and that at low electron densities exchange interactions will drive a spontaneous spin polarization. A spin polarization at zero magnetic field would give an extra plateau in the conductance at e 2 /h rather than 0.7(2e 2 /h). To explain this discrepancy various theories 4-8 invoking spin have been put forward. Recent quantum Monte Carlo calculations 9 show that in 1D the paramagnetic state is always lower in energy than the ferromagnetic state, so it is not clear whether the Hartree-Fock calculations are in conflict with the Lieb-Mattis prediction 10 that there is no ferromagnetic order in a 1D system. The role of disorder in 1D systems is little understood, but it has been shown 11 within mean-field theory that for dimensions d ≤ 2 a disordered system may exhibit a partial spin polarization, even though the system without disorder is paramagnetic.The 0.7 structure is distinctly different from the conductance plateaus measured 12 at multiples of α(2e 2 /h) in long wires fabricated by overgrowth on a clea...

We measure the temperature of a 2D electron gas in GaAs from the thermopower of a onedimensional ballistic constriction, using the Mott relation to confirm the calibration from the electrical conductance. Under hot electron conditions, this technique shows that the power loss by the electrons follows a T 5 dependence in the Gruneisen-Bloch regime, as predicted for acoustic phonon emission with a screened piezoelectric interaction. An independent measurement using conventional thermometry based on Shubnikov-de Haas oscillations gives a T 3 loss rate; we discuss reasons for this discrepancy.Accurate electron thermometry is needed in many aspects of low-dimensional semiconductor physics, particularly given the increasing importance of hot electron effects 1 as mesoscopic device dimensions are reduced and electron mobility increases. Surplus heat energy in a two-dimensional electron gas (2DEG) is rapidly shared amongst the carriers through electron-electron interactions, and an effective electron temperature T e is established which may be considerably higher than the crystal lattice temperature T l , to which both external thermometry and refrigeration are coupled. A measurement of the electron temperature is therefore needed to determine how an electron gas thermalizes with its surroundings. Measurements of the thermoelectric response and thermal conductivity of mesoscopic devices are also interesting in their own right, as they provide fundamental information about electronic properties which is not available from electrical transport measurements alone.Although many techniques allow the measurement of bulk lattice temperatures, the weak coupling of the electrons to their surroundings has hampered accurate measurement of the electron temperature T e . Previous techniques have employed the visibility of features in the electrical transport, notably Shubnikov-de Haas (SdH) oscillations 1-6 , but also using the zero field resistance and weak localization corrections. 7,8 Mesoscopic effects such as Coulomb blockade have also been applied as electron thermometers. 9 In this letter, we introduce a novel technique, where the thermoelectric response (thermopower) of a one-dimensional constriction is used to measure the electron temperature. Self-consistent checks confirm the validity of the technique, which we then employ to deduce the energy relaxation rate of heated electrons in a 2DEG in a GaAs/AlGaAs heterostructure, obtaining good agreement with the theory of acoustic phonon emission in the Gruneisen-Bloch regime.A schematic of the device is shown in Fig. 1, and is similar to those used in previous thermopower measurements. 10,11 The area shaded grey shows the etched mesa containing a 2DEG defined at a GaAs/AlGaAs heterojunction, 2770Å below the sur-face of a structure grown by molecular beam epitaxy. 12 From measurements of the low-field SdH oscillations and the zero field resistivity, the 2DEG has an electron density n e = 2.1 × 10 11 cm −2 and mobility of µ = 4.5 × 10 6 cm 2 V −1 s −1 at 1.5 K. Electrons in the he...