14
Once the data are acquired, the next task is the formation of an image. A full
data set consists of a complete THz time-domain waveform (see Figure 2.2) corre-
sponding to each pixel of the image. These waveforms obviously contain a great deal
of information: the amplitude and phase of the transmitted terahertz pulse, for many
spectral components. A two-dimensional false-color image can be formed using any
subset of this large data set. Typically, images formed using different portions of the
data contain different types of information about the sample. For example, peak-to-
peak amplitude of the THz time-domain pulse conveys the amount of THz absorption
at different parts of the imaged object, whereas spectral phase and time delay of the
transmitted THz pulse encode the thickness of the imaged object.
2.1.3 Other imaging methods
Most imaging techniques using THz-TDS relies on a raster-scanning method, to ac-
quire image data pixel by pixel. Other more sophisticated approaches which are not
discussed here, such as THz tomography and Interferometrix imaging [5-7], also face
similar speed limitations. This can be a limiting factor in some applications, since
serial acquisition of image data can be quite slow. One solution would be the de-
velopment of a sensitive focal plane array detector, an area of active research [33].
Ultrafast optics offers another solution to this problem, using the free-space electro-
optic sensing technique described above. Among the two, we will first discuss the
latter.
As noted, electro-optic sampling relies on an induced polarization rotation in
a femtosecond probe pulse, which depends on the amplitude and direction of the
terahertz field. Because the detected signal is an optical one (i.e. polarization rotation
of an optical probe beam) rather than an electrical one (photo-induced current in a