[asac] New material on the 20 micron water vapour system

[Christine Wilson]

Hi, everyone,

Here is a short writeup on the new results from the 20 micron water 
vapour radiometer ("IRMA"). I'm sorry it has no figures attached; I'll 
have figures to show at the meeting.

Chris

------------------------------------------------

Recent Results with the Infrared Radiometer for Millimetre Astronomy (IRMA)
(summarized by Christine Wilson from Graeme Smith's M.Sc. thesis)


IRMA has been designed and constructed by David Naylor and his group
at the University of Lethbridge in collaboration with the JCMT group
at HIA/NRC. A prototype device was constructed in 1999 and first tested
at the JCMT in December 1999. The data from that test run have been 
written up and analyzed in Graeme Smith's M.Sc. thesis. A preliminary analysis
of those data was presented by Chris Wilson at the ASAC meeting
in Leiden in March 2000. This report summarizes the results from a more
complete and final analysis of the same data. Figures illustrating
the results from IRMA will be shown at the February 2001 ASAC meeting in
Florence.

IRMA monitors emission from water vapour in the Earth's atmosphere at
a wavelength of 20 microns. It is equipped with optics to produce a
diverging beam which is designed to match the footprint of the JCMT
at an altitude of approximately 1 km. It can observe an ambient load,
a cold load (LN2), or the sky. The optical input was chopped at a rate
of 200 Hz and the data were sampled at a rate of 10 Hz. The prototype
device can operate in either stare or continuous scan mode. In stare
mode, a single elevation and azimuth is monitored continuously, while
in continuous scan mode, the device is scanned from 1 to 3 airmasses
in steps of 0.18 degrees.

The instrumental error was derived from the standard deviation
of the signal from the LN2 load from 16 sets of continuous scan observations
(each containing typically 15 individual cycles of a continuous scan
plus an observation of each load). The instrumental error was 5.8 mV
in 0.1 s; assuming the noise will integrate down with the square root
of the integration time, the expected noise in a 1 second integration
would be 1.8 mV.

The continuous scan data give measurements of the instrumental voltage
as a function of airmass. These curves were calibrated to give voltage
as a function of mm of precipitable water vapour (mm pwv) using two
observations with the 183 GHz radiometer at the JCMT that were obtained
simultaneously with observations with IRMA. The 183 GHz data were
used to derive the pwv at the time the data were taken. (Similar simultaneous
measurements with the 350 micron taumeter at the CSO gave a similarly good
but slightly lower calibration. Attempts to calibrate the data with
pwv determined from radiosonde launches from Hilo were unsuccessful due to
the large scatter in the radiosonde measurements.)

Once the curve of IRMA voltage versus mm pwv has been obtained, it can
be combined with the instrumental error to derive the sensitivity of
IRMA to changes in the pwv (or changes in the optical path). The table
below gives the derived IRMA sensitivity for values of 0.5 mm and 1.0 mm
for the atmospheric pwv. (The optical path was derived from the pwv by
multiplying by a factor of 6.5.)

Atmospheric	Resolution of IRMA in 1 s	ALMA specs in 1 s
pwv (mm)	pwv (microns)	path (microns)	  path (microns)

0.5 mm		1.8			12		15
1.0 mm		3.0			20		20

>From the table, it can be seen that the raw sensitivity of the 
prototype device met the ALMA specifications for measuring the 
optical path above a single antenna.

A Phase 2 instrument was built and installed at the JCMT in the summer of
2000. This device has a more sensitive detector, an infrared
filter with improved efficiency, and improved digital electronics.
This device is currently operating at the JCMT where it is used to
take many continuous scan measurements whenever SCUBA does a sky dip.
The idea is to correlate the pwv derived from the SCUBA estimates of
the optical depth at 850 and 450 microns with the IRMA measurements.
The data are currently being analyzed, and the new device seems likely
to be an order of magnitude more sensitive than the prototype.
Naylor's group are also researching the effects of cirrus at 20 microns. 
So far there are no large observable effects that might be attributable
to cirrus, although this will be difficult to quantify without an independent 
measure of cirrus. A Phase 3 device which will be completely stand-alone
and will use a Stirling cooler and two point warm blackbodies is currently
being built.

Since IRMA represents a new application of infrared technology to
radio phase correction, there are of course a number of uncertainties
and questions that would need to be resolved before IRMA could be adopted
as the baseline design for a water vapour radiometer for ALMA.
Some of these uncertainties were listed in the September 2000 ASAC report;
I have summarized some of the issues I could think of as well as an
appendix at the end of this report. However, the results from the 
prototype IRMA device are extremely promising, and it is important
to work as quickly as possible to resolve some of the remaining
uncertainties. Ultimately, installation of two IRMA devices on
a millimeter array will be necessary to prove the concept.



Appendix: A summary of areas of concern and future work with IRMA


Uncertainties (possible sources of systematics):

-- IRMA beam can't match radio beam at all heights in the atmosphere; is
this a problem?

-- IRMA bandpass samples mixture of optically thin and thick lines

-- effects of clouds

-- sensitivity of IRMA to changes of pressure, temperature, distribution
of water vapour (height)

-- how easy will it be to align IRMA with the radio beam? where would
we put IRMA? i.e. weight loading on edge of dish?


Needed in the short term:

-- more direct comparisons of IRMA signals with 183 GHz system

-- further calibration of V vs airmass to V vs pwv (i.e. more than 2 points
should be used to calibrate it)

-- refinements of the original design, including conversion to a closed-cycle
cooler (on-going)


Not yet tested/investigated:

-- direct comparison of IRMA signals with phase on an astronomical source

-- stability of IRMA over 5 minutes 

-- stability of IRMA over changes in zenith angle of 1 degree

-- is 1 s a short enough timescale to be deriving pwv over? (a more general
question than just IRMA)

-- could IRMA be used to correct for anomalous refraction?