a) Water – One of the most restrictive challenges of THz imaging is due to the high absorption rate of polar liquids, especially water.  The absorption coefficient of water is as high as 150 cm-1 at 1 Thz.  This property limits the sensing and imaging capabilities through thick tissue. 

b) Power – The typical power output of a THz wave source is on the order of μW when using a pulse laser oscillator at 1W peak output power.  The peak THz power ranges from 100 μW to 0.1 W depending on the duty cycle of the oscillator and produces signal-to-noise ratios in excess of 10^4.  However, for real-time 2D imaging, an average of mW or higher sustained power output is preferred.

c) Signal-to-noise ratio (SNR)– THz time domain spectroscopy systems are capable of SNRs of over 100,000.  However, this ratio can be greatly reduced with an increase in acquisition speed and increased absorption of signal within biological tissue.  This can be overcome with better free-electron lasers and more sensitive sensors. 

d) Acquisition Speed – Due to the fact that conventional THz imaging systems rely on scanning in the x- and y- directions, acquisition is fairly slow (on the order of 50 pixels/sec).  Future technologies of using two-dimensional electro-optic sampling with a CCD camera may overcome this hurdle. 

e) Limited frequency bandwidth – The current standard of using photoconductive dipole antenna (PDA) to produce THz waves are limited to frequencies below 4 THz.  Optical ratification can be used to boost bandwidth in the range of 30 THz; however, this is at the expense of power and SNR.  Ideally, a system would be designed with bandwidth of 100+ THz in the infrared range, as to reduce the absorption of water significantly. 

f) Scattering – Although scattering is greatly reduced by using longer wavelengths in T-ray imaging, it is not completely eliminated.  The scattering exhibited in various clinical applications has yet to be studied thoroughly.  This will eventually lead to changes in the actual calculation algorithms for reconstructing the THz images.

g) Target reconstruction – Most of the research of THz imaging is currently performed on a thin parallel-faced samples or the reflection of relatively flat surfaces.  However, there are many applications in which an irregularly shaped 3D object needs to be imaged.  In this case, a collection of optics and algorithms are being developed for target reconstruction.  Another possibility of target reconstruction is T-ray computed tomography (T-ray CT) in which multiple transmission-mode images are obtained at different projecting angles and are reconstructed to obtain a 3D structure. 

h) Biomedical spectroscopic database – One of the primary advantages of using THz imaging is that it provides spectroscopic information of the target within a particular frequency band.  However, the responses of biological tissues in this band are currently unknown.  The problem is that there are many intra- and inter-molecular interactions that could have an effect of the frequency range, making its characterization very difficult. 

i) Size – The size of current T-ray imaging machines require a few square meters of area, most of which is occupied by the pulse laser.  In order to create an endoscopic T-ray probe, the technology must be advanced to greatly reduce the size.  There is development using a single ZnTe crystal for both emission and detection of THz pulses that may reduce the overall size of the THz system. 

j) Cost – Currently, the high costs of ultra-fast Ti:Sapphire lasers impede the access of THz imaging for many applications.  A typical T-ray sensing system and an imaging system is $100,000 and $200,000, respectively.  Nevertheless, this is comparable to x-ray, CT, MRI, and various NMR systems.  As solid-state electronic T-ray sources advance technologically, it may greatly reduce the total cost of a THz imaging system.