Building a FT-FIR

1
Building a FT-FIR

• Towards a THz version of the Flygare

R. Braakman 1,) M.J. Kelley1), K. Cossel1),
G.A. Blake2)
1) Division of Chemistry Chemical Engineering,
California Institute of Technology 2) Division of
Geological Planetary Sciences, California
Institute of Technology ) contact
rogier_at_caltech.edu
Laser stabilization
Introduction
Recording the FID of the molecules in the time
domain followed by a Fourier transform to the
frequency domain allows for very high resolution
spectra, in principle. In the heterodyne
detection method, however, the FID signal is
mixed with a LO signal at the HEB mixer, and the
resulting line width is thus dependent on the
widest of the two incoming signals. It is
therefore essential to stabilize both input
lasers to ensure an LO signal with a narrow line
width. In addition, long term drift due to
thermal fluctuations can thus be countered. Prior
to stabilization efforts, the line width of the
beat note between DFB and Agilent lasers was 10
MHz, with a long term drift of 100 MHz. For the
DFB laser a temperature feedback circuit will be
used, which is shown below. A similar method
based on an external cavity will be used for the
Agilent.
THz radiation probes low energy states in a wide
variety of systems, including rotational states
in small molecules, phonons in solids, thermal
emission from cold sources, and the lowest lying
vibrational modes of large molecules. In our
group the main push for the development of a THz
spectrometer is the commissioning of the
SOFIA/Herschel observatories over the next 2 to 3
years, which will allow high sensitivity access
to interstellar THz spectra for the first time.
Due to technological difficulties, easily
tunable intense THz radiation sources and
sensitive detectors have been lacking. To
circumvent this problem we will use a non-linear
effect to generate THz radiation (explained
below) from two 1.55 ?m sources. The result is a
stable yet easily tunable source. The THz power
output is only on the order of 0.1 ?W, however,
which is quite low for spectroscopic purposes.
We thus propose the development of a FT-FIR
spectrometer based on the Flygare FTMW design
(Ref. 1) to increase the sensitivity of the
system.
Molecular nozzle
THz Flygare
mirrors
Wire grid polarizer (R99.9)
THz photomixer
T, I controller
Lock-in amplifier
THz generation
HDO
5050 beam splitters
A photomixing scheme will be used to generate the
THz radiation. In this scheme, two beams of light
with roughly the same frequency are
simultaneously directed onto the surface of a
photomixer chip, whose characteristics allow for
the generation of THz waves at the difference
frequency of the two lasers. A recently developed
material consisting of ErAs/InGaAs that operates
at 1.55 ?m (Ref. 3) is currently being fabricated
into
1.5 ?m Er Amplifier
DFB
EOM
e.g. P, ? meter
Heterodyne HEB THz mixer
Mirror (Rgt99.99)
3 dB coupler
mirror
Figure 5. Temperature feedback circuit for DFB
laser. EOM sideband probes the HDO transition,
which is detected as first derivative line shape
by Lock-in amplifier. LIA then feeds positive or
negative signal back to the temperature
controller of the DFB
Amplifier, filters
Figure 2. Photomixer setup. Adopted from Ref. 2
chips . This allows for a fully fiber based
setup, in which a fixed frequency DFB laser and a
tunable Agilent laser are coupled together in a 3
dB coupler and subsequently amplified
simultaneously. After decoupling, the two beams
are then directed onto the photomixer chip, where
THz generation takes place.
Fixed tuned 1.55 ?m DFB laser
Tunable 1.55 ?m Agilent laser
Summary
We propose the development of a FT-FIR instrument
as a THz analog of the Flygare system. There are
several disadvantages that must be overcome. The
input power of the cavity from the photomixer is
quite low compared with typical input powers in a
Flygare system. In addition, the noise level of
detectors is intrinsically higher at THz
frequencies than at microwave frequencies and the
higher detection bandwidth adds to this
effect. However, it is possible to take
advantage of the Fabry-Perot resonator to a
greater extent than in the case of the Flygare by
having better coupling and using longer radiation
pulses to build up a higher field. Spectroscopic
properties of molecules are also favorable at THz
frequencies. Due to the significantly higher
Einstein A coefficients, a much larger fraction
of the absorbed radiation can be recovered in
emission. In principle, these combined factors
should sufficiently compensate for the
disadvantages and allow higher sensitivity than
at the wavelengths of a traditional Flygare
system.

THz cavity
THz detection
The power output of a photomixer device is
several orders magnitude lower than the power
used in a typical Flygare experiment (0.1 ?W vs.
1 mW). However, using a Fabry-Perot resonator to
build up a radiation field and subsequently
recording the FID of the excited molecules in the
time domain allows us to take advantage of
several aspects of this system to compensate for
this large difference.
With a Q of 104 the bandwidth of the cavity will
be around 100 MHz. Therefore a detector that is
both fast and sensitive is needed in this
experiment. New Hot Electron Bolometers (HEB)
that have recently been developed in the Prober
group at Yale meet both of these requirements.
These Nb-based superconducting devices, which are
currently being tested in a joint effort, have an
electron-phonon cooling time of 1 ns, which
results in a possible detection bandwidth of at
least 150 MHz. HEBs are capable of attaining a
system temperature (Tsys) of 1000 K, which gives
a NEP of 1.4x10-20 W/Hz in heterodyne detection
mode. This compares to a Tsys of 100 K and a NEP
of 1.4x10-21 W/Hz for a typical Schottky detector
in the Flygare system. The increased bandwidth in
the THz setup adds another factor of 102 to the
noise level.

• Different source stabilization method allows for
  longer radiation pulses in THz system which help
  build up higher field (1 ms vs 0.1 ?s).
• Einstein A coefficients, and thus emitted power,
  tend to be significantly higher in the THz
  region. Comparing the fundamental transition (J
  0 1) for OCS at 12.163 GHz ( standard for
  Flygare system) and that for H2O at 1113.342 GHz,
  we find A 1.8x10-2 s-1 for H2O and A 3.6x10-9
  s-1 for OCS, a difference of almost 7 orders of
  magnitude!
• The loaded Q-factor should be equivalent even
  at higher frequencies. A test cavity at 300 GHz
  (spectrum shown on right) is indeed found to have
  a Q-factor close to that of Flygare systems. The
  semi-confocal cavity employed allows for better
  coupling into the cavity than in a Flygare FTMW
  setup, resulting in less loss.

FWHM 4 MHz, Q 7.5 x 104 !!
References
Figure 3. Cavity absorption of prototype
semi-confocal cavity at 300 GHz
Figure 4.Nb based HEB detector (above) and their
T-R response curve (below). Courtesy of Prober
group.

1. T.J. Balle W. H. Flygare, 1981, Rev. Sci.
   Instr., 52, 33-45
2. S. Matsuura et. al., 1999, Appl. Phys. Lett., 74,
   2872-2874
3. M. Sukhotin et. al., 2003, Appl. Phys. Lett., 82,
   3116-3118