This is a lot to cover so bear with me. A lot of our first week consisted of reading paper after paper on our project topics in order to, of course, learn about them and to be able to assemble a decent 5 min presentation summarizing what we think we will be doing in Paris. I will reiterate what I said (and also say some things I neglected to say) here. Then I will express what plans and goals are for the rest of my time here.
What is THz?
THz radiation consists of a band of frequencies overlapping microwave and far infrared regions, 0.1-30x10^12 Hz (wavelengths 0.01-3 mm, about the size of a pin head). It is utilized in spectroscopic evaluation and imaging. There are a lot of interesting optic and electronic aspects to this form of radiation. Its attractive qualities include: It's non-ionizing, because of it's low energy, which makes it safe to use around people and in open areas. It's a non-destructive means for evaluation (NDE). And because THz radiation is in a pulsed form (down to picosecond duration and 80-100MHz repition rate) when used in reflection it can provide better depth resolution when compared to say X-rays. The pulses that are reflected from a sample arrive at a receiver at different times and can be sorted out based on t and used to define depth and layers.
THz Setup
*This image was provided by Dr Jason Deibel, WSU, Dayton, OHThe image above shows a simple setup for THz transmission. I'll go through each component and tell you what I think they do. TiSaph Laser is a mode-locked laser which is able to generate fs (10^-15 s) pulses centered at about 800nm wavelength and a repetition rate of 80-100 MHz (so one pulse every 12.5-10 ns). The pulse train is split into two paths, the delay (probe) line, and the pump line. The pump line travels to the transmitter, the optoelectronic switch.
The figure to your left is my rendition of said switch. The gray wafer is made up of low-temperature grown GaAs. This material is photoconductive. Note that there are a lot different phoconducting materials that can be used here. For receivers its ideal to have fast relaxation times, but not essential for transmitters. Gold electrodes with an external DC bias are patterned on one of its faces.
The dimension of the electrodes can be 10s of μm and the gap between them is typically 10 μm. The bias can be set at 10-25V.
The fs pulses are focused between these two electrodes and supply enough energy to excite electrons from the valance band to the conduction band. These charge carriers are now able to move under the influence of the bias, accelerate, and thus radiate. This is the driving force for sub-picosecond THz pulse generation. There is a lot more to this process (e.g., material science, E&M) and I will revisit this component soon to give a more in detail look at how it works. The THz beam is focused using HDPE (high-density polyethylene), which is also used to make milk jugs and plastic cups. According to Daniel Mittleman in Sensing with Terahertz Radiation HDPE has very low absorption, (.04cm^-1 - 1 cm^-1 at 300GHz - 2 THz respectively), little dispersion, and small Fresnel reflection losses (n = 1.52, insertion loss order of 9% per lens). In the transmission setup we have radiation exiting the material and making its way into a receiver. The probe line at the receiver enables us to move the sampling gate across the waveforms we wish to sample. The probe line path length differs from the pump line length and encounters a retroreflecting mirror mounted on a mechanical scanner. The scanner makes the mirror moves along the axis of propagation of the probe beam at a fixed amplitude, and the frequency at which it does this can be either considered slow or rapid step/oscillation/scan (what frequencies constitute slow or rapid is another good question to look into). Tens of hertz seems to be the largest frequency you can get out of these scanners (but this is circa 2003). Mittleman says for 'a [sampling] window of 100 ps, the retroreflecting mirror arrangement must oscillate with an amplitude of ~1.5cm'. The way the equation goes in my head is 2*A/c= t. For a shorter window higher acquisition rates can be achieved and piezoelectric transducers can be used (which sounds neat to me). The receiver is very similar to the transmitter except there is no DC bias applied to the electrodes. What supplies the bias are the transmitted THz pulses. The current from the receiver travels to a current preamplifier and then terminates at a resistive load at the lock-in amp, which ends in a resistor. The change in potential across the resistor is then used to spit out the signal on an oscilloscope. And of course there is more to it and I will go back into it in detail some point in the near future.
What's been done thus far?
Bianca Jackson, my mentor, has conducted a lot of THz research with a wide range of applications for artwork evaluation. There is still a lot to cover but lucky for me I should be getting a hard copy of her dissertation. For one study, with the use of a transmission setup she shows how THz can be used to distinguish different pigments. Each pigment was placed in a cuvet (one reference sample, air, was taken). The graph to the left shows the waveform received as a function of time. Solving for the bulk refractive index of the material is a sort of plug and chug operation, nsample= nair +c/d*(tsample-tair). With time domain spectroscopy (TDS) temporal signals can be converted to supply spectral information. Using a reflective setup (which I will go into detail with some other time) Bianca studied three very similar looking pigments, lead white, titanium white, and Bianco San Giovanni. Here she used the spectral component of the received E field,which can be obtained using an FFT of the TDS data or simply clicking a button on a computer I imagine. Using the spectral data, reflectivity as a function of frequency, r(v), can be plotted and applied to Fresnel equations to determine the pigment's extinction coefficient (k), imaginary part of the refractive index. And with the use of more equations the phase associated with k and n can be determined and then all of these parameters can be used to output the pigment's absorption spectrum, a(v). This is the material's so called 'THz fingerprint' because of its uniqueness.
The fresco paintings above were used by Bianca to demonstrate time domain reflectivity (TDR). Reflection of the THz pulses at material and layer interfaces can be exploited to obtain spectral and spatial information of the paintings. It's similar to the processes explained above but with a pixel by pixel approach. The image furthest left is a fresco painted on plaster, the next is a THz image of that same fresco with ~5mm thick plaster above it, the third is another fresco painted onto the previous image, and to the furthest right is a THz image of the previous fresco. By the last image it is hard to see the original painting. This is attributed to losses (scattering and absorption) and also because the THz image of each layer is displayed. Steve Yalisove begin_of_the_skype_highlighting end_of_the_skype_highlighting suggested I look at what Bianca had done with dendrochronology (the science of dating based on tree rings etc) where if the pulses coming from one layer do not interfere with any other layer pulses, they can be removed from the image in order to get a better look at the other layers. These are just brief examples of the extensive work done.
What's my plan?
I will read up on THz and optics in general. I want to understand the processes fully so I can be of some real use. I know that Bianca has asked me to familiarize myself with LabView to help her streamline and sort through data. From what I can see of the program, it reminds me of Simulink and MatLab which I am a little familiar with. After visiting with John Whitaker (who elucidated a lot of the delay line information for me) and seeing his setup, it 's clear that it's the application is a vital part in the setup. So expect blog entries about that. THz evaluation and imaging really interests me so I hope I can do my best and learn a lot.
