Atmospheric Optics18 October 1972 Freeman F. Hall

The present understanding of the atmospheric optical index-of-refraction structure parameter, Cn2, as determined by the tem-perature structure, CT2, was the subject considered by an invitedpanel and later during the audience participation session. JackL. Bufton (NASA Goddard) described the results of CT2 mea-surements made with balloon-borne, fine platinum wire sensors.The detailed results obtained on several flights were publishedrecently in the Journal of the Optical Society of America. He ex-plained that the sensors were spaced 1 m apart horizontally, andthat CT2 values were computed on board with 5-sec averagingtime. The electronic computation was accurate to within 10%.Balloon ascent was at 5 msec-1 with parachute descent at twicethis vertical velocity. Therefore, thin layers of high CT2 mayhave been underestimated because of the vertical distance aver-aging. There is also the question if measurements at one separa-tion distance are suitable for thin shear layers where turbulencestatistics may not follow a Kolmogorov spectrum. In general,daytime values of CT2 were larger than nighttime values by a fac-tor of 3. Values decreased with altitude, but large spikes weremeasured at many altitudes and not only at the tropopause aspredicted from the earlier Hufnagel model.

Robert S. Lawrence (NOAA Boulder) reported on the mea-surements of CT2 using a single fine wire probe on an aircraftflying above the Pacific ocean near San Diego, both within andabove the marine layer. The values within the marine layer wereas large as those found over land. The airborne probe was cali-brated against spaced fine wire sensors on a tower, and the agree-ment was found to be good in preliminary calibration tests. Itwas pointed out that the aircraft does not provide information onthe continuity of horizontal strata of high CT2, because of the dif-ficulty of following such thin layers. In repeated spiral climbsand descents through and above the marine layer, it was foundthat CT2 values are roughly the same over water and land, toheights of 3 km.

Theodore Blanc (NRL) reported on measures with spaced finewire temperature sensors separated by 10 cm on a boom some 8 min front of and above the surface of the water during a Navy shipcruise from the Carribean Sea to Bermuda. Values were aver-aged over 15 sec with maximum values of 2.5 x 10-13 m-1 /3 forCn2 in the Carribean to 2.5 x 10-'7 at the northern end of thecruise. Maximum values were measured at sunrise and twohours before sunset with minimum values at noon. This type ofmeasurement is fraught with difficulties, such as keeping salt de-posits off the fine wire sensors.

Darryl Greenwood (Griffiss AFB) described instrumentationat the RADC propagation range where phase, arrival angle, andatmospheric absorption at several wavelengths will be measuredsimultaneously with CT2 determined by spaced fine wire sensorson towers along the path. He demonstrated the packaging andconfiguration of an improved sensor with good low frequency re-sponse. Two such sensors tested side-by-side accurately trackeach other. The objective of their program is to measure CT2 VSaltitude over land and over the ocean from the planetary bounda-ry layer to high altitudes. In situ measurements will be supple-mented with acoustic echo sounding studies. -

Robert Hufnagel (Perkin-Elmer) commented that a calcula-tion of C,2 based on sunny day heat flux assumptions agrees wellwith measurements made in the convectively unstable atmo-sphere. Although his proposed profile of Cn2 is now over six yearsold, the general characteristics seem to be correct but with muchadditional fine scale, high peak values to be expected throughoutthe troposphere. David Fried pointed out that the higher mo-ments of CT2 must also be reduced from the measured data to ob-

tain a better understanding of the values to be expected alonghorizontal and slant paths through the atmosphere.

A poll of the attendees indicated that, with the improved mea-surement techniques now available and with the more sophisti-cated modeling that has developed during the past several years,now may be the appropriate time to schedule a Topical Meetingon optical propagation through turbulence. A probable date forthis meeting will be in the summer of 1974, with the location tobe Boulder, Colorado.

Thin Films and Interferometry13 March 1973 Philip Baumeister

A symposium was held on spectral filters with narrow band-widths. This included both the quartz and mica spacer typeFabry-Perot (F.P.) filter, the multilayer all-dielectric interferencefilter, and the Lyot-Ohman (L-O) type polarization interferencefilter. The panelists were Edgar Barr (Barr Associates), PhilipBaumeister surveyed some of the attributes that are used toento Peak Observatory), J. A. Dobrowolski (National ResearchCouncil of Canada), and David Markle (Perkin-Elmer).

Baumeister surveyed some of the attributes that are used tocharacterize a spectral bandpass filter. These include the maxi-mum radiant transmittance Tmax in the passband, which is cen-tered at X, the spectral half-width AX1/2 at 0.5 max, the deci-width AX,/lo at 0.1 Tmax, the offband rejection, the shift of thepassband as the angle of incidence changes from normal (angleshift), the shift of the passband with temperature, and finally itsstability. The offband rejection can be estimated by requiringthat the total flux in the offband region of the spectrum must beseveral orders of magnitude less than the flux in the passband.This means that filters with a narrow spectral bandwidth musthave a much lower transmittance in the offband part of the spec-trum.

Barr described the evolution of the all-dielectric interferencefilter. He recognized early that the single cavity type filter, theF.P., had poor offband rejection. In such a filter the ratioAXI/2 /AX1/1o is independent of AX, 2. This led him to designmultiple cavity filters using a method described by Turner.'This type of filter has a spectral passband shape that moreclosely approximates a rectangle than the single cavity filter. Hehas produced filters with a AX1/2 as small as 2-3 A and anoffband optical density of 5. This offband rejection is further en-hanced by adding blocking filters, which contain up to sevenmetal layers. He relied on the theory of induced transmission fil-ters2 to design these multilayers.

Beckers then compared the attributes of the L-O filter to theF.P. The L-O filter he described3 contains achromatic retarda-tion plates4 that permit it to be rapidly tuned over a wide spec-tral range while maintaining a large angular acceptance cone.The attributes of the L-O and F.P. filters are compared in TableI.

Dobrowolski then reviewed the technique of fabricating a micafilter, 5' 6 which consists of a thin mica sheet coated on both sideswith multilayer dielectric reflectors. The mica is first cleavedand then mounted in a holder, which keeps it in tension. Thethickness of the mica is preferably chosen so it is a half-waveplate. In some instances the mica spacer is not of the requiredthickness to locate the passband at a given wavelength. In thiscase, it is necessary to augment its thickness by evaporating afilm, a correcting layer, on its surface. The resulting filter has ahalf-width in the range from 0.1 A to 1.5 A and a typical peaktransmittance in the red spectral region is 0.35. The absorptionof muscovite mica makes it unusable at wavelengths below 400nm.

The mica filters are quite stable, and Dobrowolski has found

August 1973 / Vol. 12, No. 8 / APPLIED OPTICS 1993