Roughened quartz surfaces and Teflon as small anglediffusers and depolarizers between 200 and 400 nm

Robert D. Saunders and Henry J. Kostkowski

Laboratory tests of transmission diffusers are reported for quartz and Teflon. These tests were conducted toselect transmission diffusers for the NOAA rocket flight UV spectroradiometers.

1. Introduction

As part of a National Oceanographic and Atmo-spheric Administration (NOAA) program to improveultraviolet (UV) spectral irradiance rocket measure-ments of the sun, a suitable diffuser is being sought forthe 200-400-nm region. Its purpose is to eliminatesignificant differences in the irradiance responsivity'of our spectroradiometer when viewing the sun andsource standards such as a tungsten type FEL lamp, anargon arc, and the NIST Synchrotron (SURF II). Dif-ferences in the irradiance responsivity arise because ofdifferences in the angle of incidence and of polariza-tion of the radiation from these sources and of thestrong sensitivity of our spectroradiometer-without adiffuser at its entrance aperture-to these quantities.

Selection of a suitable diffuser will depend on therange of angles of incidence and on the polarization ofthe incident radiation, on the maximum exiting anglefrom the diffuser, and on the attenuation, stability,and uniformity of the diffuser. The angle of incidenceof the radiation from the above mentioned sourcesvaries from 0 to -2, and the exiting angle (from thediffuser) for our f/5 system varies from 0 to -5.7 asindicated in Fig. 1. The degree of polarization of theradiation from the above sources varies from zero forthe argon arc to 3% for an FEL tungsten irradiancestandard2 and is almost 100% for SURF II.3 There-fore, we are seeking diffusers that obey the cosine law(incident flux is transmitted the same independent ofincident angle) for incident angles up to 2 and exitingangles up to 5.7° and the exitant radiation is depolar-ized. Because of signal-to-noise limitations, the dif-

Robert Saunders is with U.S. National Institute of Standards &Technology, Gaithersburg, Maryland 20899, and Henry Kostkowskiis with Spectroradiometry Consulting, Route 1, Box 69, CharlotteHall, Maryland 20622.

Received 23 February 1989.

fuser must have a receiving area of at least a squarecentimeter, and its attenuation should not be greaterthan -500. Attenuation for this investigation is de-fined as the ratio of the output signal of the spectrora-diometer without a diffuser to that with a diffuser,keeping (or correcting) the irradiance at the entranceslit when the diffuser is not used to be the same as theirradiance at the diffuser when the latter is used. Be-cause of our goal of limiting the uncertainty of the solarspectral irradiance measurements to a few percent, astability of -1% between calibration and use in solarmeasurements is required. Since the irradiance over asquare centimeter from the above sources is highlyuniform at the distances normally used, uniformity ofthe diffuser is not a critical requirement.

We have not been able to obtain information fromthe archival literature or from our colleagues needed toselect a suitable diffuser. From our own limited expe-rience, there are three possibilities: (1) an averaging/integrating sphere, (2) a roughened quartz surface or astack of such surfaces, and (3) sheet Teflon. Becausethe solid angle used by most radiometers is small, anaveraging/integrating sphere is expected to have toohigh an attenuation, particularly at the shorter wave-lengths. Reflecting plaques have similar problems.Therefore, we have concentrated on the remaining twopossibilities. The current paper reports on an investi-gation of the small angle diffusing, depolarizing, andattenuating properties of roughened quartz surfacesand virgin sheet Teflon at wavelengths between 200and 400 nm. Stability measurements will be conduct-ed later during the characterization and calibration ofthe rocket-borne instrument.

11. Experimental Setup

Figure 2 is a schematic drawing of our experimentalsetup. The spectroradiometer includes an f/15 Ebert-Fastie double monochromator with a 0.1- X 5.0-mmentrance slit, a spectral bandpass of 0.2 nm, and abialkali photomultiplier operating in a photon count-ing mode.'

3242 APPLIED OPTICS / Vol. 28, No. 15 / 1 August 1989

1 cmDiameter Diffuser

0.25'---

Sun

5 70

X EntranceSlit

f/5

I _ 0.8°

Tungsten 1 c X 3

50 cm

.57,.1 cm _5__

Argon ,Arc *

0.026

Surft t -0.01 cm x 0.2 cm)

|i 1100 cm

Fig. 1. Angular configurations for the sun and various source stan-dards.

A low power He-Ne laser (shown in Fig. 2) is used insetting the position of the 4 X 4 aperture for a particu-lar value of 0 and setting the position of the source sothat 02 = 0° for that aperture setting. The laser itselfis positioned so that its beam is on the optic axis orparallel to the optic axis striking the diffuser underinvestigation at the point where the 4- X 4-mm aper-ture is placed. A laser mount with micrometer-typetranslational and angular motions provides the re-quired adjustments.

The sources used consist of either a 12-mm longsection of a 0.25-in. diam by 2.125-in. long low pressurecold cathode mercury discharge lamp or a 30-W deute-rium arc lamp with a 1-mm diam discharge-constric-tion aperture. The deuterium lamp is used primarilyfor measurements at wavelengths below 250 nm. Theirradiance (at the slit and with no diffuser) of the UVspectral lines of the mercury lamp has the angularvariation (about the lamp axis) shown in Fig. 3. Be-fore each measurement, the mercury lamp is rotatedabout its own axis so that the irradiance at the 4- X 4-mm aperture is always at the minimum value indicatedin Fig. 3.

Ill. Transmission Diffusers

The quartz elements being investigated consist ofSuprasil II disks either 2 or 2.5 mm thick and 25 mm indiameter. One or both surfaces of the disks have beenground so that they have an 8-, 30- or 100-Atm diam (pitsize) finish. In use, the disks are stacked to produce asample containing 1-3 surfaces with the same finish.The disks are spaced 2-2.5 mm apart. The Teflondisks being investigated are also -25 mm in diameterwith thicknesses of 0.003, 0.005, 0.01, 0.014, and 0.031in. The above disks are positioned and oriented sothat their centers are on the optic axis and their sur-faces are normal to the optic axis as indicated in Fig. 2.To test for the desired normality, the laser beam is

I Laser Position (5)

Optic Axis .I Source (Mercury Lamp or D2 Arc)

53 cm 1024 mm x 4 mm Aperture

Diffuser Under12 cm I 8 ~ Investigation

Entrance Slit

Spectroradiometer

Fig. 2. Schematic drawing of the experimental setup.

a).I_

EC

00

C.)Cu

W

co

600

400

200

0

End-on View ofLamp Discharge

0 180 360

6 (Degrees)

Fig. 3. Variation of irradiance of the 254-nm mercury spectral lineas a function of angle of rotation about the axis of the lamp

envelope.

retroreflected from a piece of microscope glass placedon the upper disk.

A 4- X 4-mm aperture, made from black mat paper,together with the entrance slit and the source fix theangles 0l and 02 in the measurements involving thecosine law. The 4 X 4 aperture is replaced by a 1-cmdiam aperture, with its center on the optic axis, for theattenuation measurements. For polarization mea-surements, the full f/5 system is used along with adichroic polarizer having an extinction coefficient(when polarization axes are crossed) of a few percent at254 nm.

IV. Measurements and Results

Since the attenuation factor must be limited to-500, and approximate measurements of attenuationare easy to make, this measurement was carried outfirst.

Table I gives the results of the attenuation measure-ments, corrected for an f/5 system, for a number ofroughened quartz surfaces and various thicknesses ofTeflon. The uncertainty of the values in Table I isvery large, -25%. The reason for this is that theirradiance responsivity without a diffuser dependsgreatly on the angle of the incident radiation as men-

1 August 1989 / Vol. 28, No. 15 / APPLIED OPTICS 3243

: r

tioned earlier. However, the large uncertainty is not aproblem because the major purpose for the attenua-tion measurements is only to rule out some of thediffusers and reduce the number of time-consumingcosine law experiments that have to be conducted.Based on the results shown in Table I, Teflon with athickness >0.010 in. was excluded from further mea-surements.

To check how well the cosine law is obeyed for therange of incident and exiting angles that exist in ourcase (see Fig. 1), two sets of values were assigned to 01and 02. These are shown in Figs. 4(a) and 4(b).

For the measurement configuration shown in Fig.4(a) where 01 = 0° and 02 = +2° and -2°, the percentdeviation from the cosine law, designated D(4a) isdefined by the expression

Ft[ S 2 \/___ \1 D(4a) = 100 [ 2- So cos2)+ 2 - So cos2o /2S0

where Si is the output signal for the ith angular posi-tion. The cosine squared terms are corrections due tothe inverse square law because the source was at agreater distance when in the +2° or -2° position.Results for roughened quartz surfaces in configurationFig. 4(a) are given in Table II for the 254- and 365-nmmercury spectral lines. Without any diffuser present,the deviation is -28%. A few measurements were alsomade for 02 = +5' to observe how rapidly D(4a) in-creased with 02-

Results for the second group [Fig. 4(b)] where 02 =

+2° and -1° for each of three values of 01 are presentedin Table III. Here, 01 has values of 2.90, 4.30, and 50corresponding to the maximum 01 for an f/10, f/6.6, andf/5.7 system. The percent deviation in this case iscalculated by the equation

Di(4b) = 100 ( Si - SO cosi(COS2i

where i is either 2° or -1°. Again Si is the outputsignal in the ith degree position and cos2i term is aninverse square law correction. For this case, it is seen

Table I. Sample Attenuations for an F/S System

Diffuser type Wavelength (nm)201 210 220 254 365

Quartz surfaces1 side 8-,gm pit size 7

30 10

2 sides 8 2130 32

3 sides 8 5730 94

100 159 135

Teflon sheeta0.031 1012 4310.014 2301 1399 988 5470.010 860 625 374 2780.005 643 441 2630.003 228

a Thickness in inches.

02 = 2 02=

1201 mm 4 mm_ -a Aperture

Diffuser '

- - Entrance Slit

4A

2 0 2=1

0 = 2.9° (f/10)

4.3° (f16.6)

5.0 (11/5.7)

4B

Fig. 4. Configurations for measurements on deviation from thecosine law.

that the cosine law deviation is >1% for all the quartzconfigurations. Teflon is a significant improvement,but even here the maximum deviation for an f/5.7system is -1.7%.

The major source of uncertainty for the results inTables II and III is the drift (due to temperature) of themercury lamp output. The drift averaged -0.7% per 5min which was about the maximum time interval be-tween repeats at the same angle position. Enoughdata were taken so that the uncertainty of the results inTables II and III is estimated to not exceed a few tenthsof a percent.

Table II. Sample Deviations from Cosine Law for Roughened QuartzSurfaces; 01 =0, 02 = ±2o, and ±50 [see Fig. 4(a)]

RoughnessNumber (pit size) Percent deviation

02 of surfaces (M) 254 nm 365 nm

None - -28.01 8 -4.7 -5.0

30 -4.3 -3.7

+20 2 8 -1.3 -2.830 -1.0 -1.4

3 8 - 1.0 -1.130 -0.1 -0.8

{1 8 -19.030 -21.0

+50

3 8 -5.030 -3.3

Table Ill. Sample Deviations from Cosine Law at 254 nm for RoughenedQuartz Surfaces and Teflon; 01 = 2.9o(f/10) to 5.0(f15.7), 02 = 2° and

-1° [see Fig. 4(b)]

Percent deviationDiffuser type 01 02 = 20 02 =-1°

Quartz surfaces3 sides/30 gm 2.90(f/10) -4.8 +2.23 sides/30 ,um 4.3o (f/6.6) -6.2 +3.03 sides/100 gm 4.3o(f/6.6) -3.9 +1.6

Teflon sheet0.010-in. thick 4.30(f/6.6) -0.9 +0.7

5.00(f/5.7) -1.7 +1.5

3244 APPLIED OPTICS / Vol. 28, No. 15 / 1 August 1989

Table IV. Polarization Responsivity of a Spectroradiometer With andWithout Various Roughened Quartz and Teflon Diffusers

Percent polarization responsivityDiffuser 254 nm 365 nm

None 89 84

Quartz surfaces8 gum

3 sides 81

30 gm3 sides 84 784 sides 72

100 im3 sides 52

Teflon sheeta0.031 4.3 3.50.014 5.70.01 4.3 1.2

a Thickness in inches.

Table IV gives the results of the polarization mea-surements. The polarization responsivity for thespectroradiometer is defined as

p S(max) - S(min) X 100S(max) + S(min)

where S(max) is the maximum output signal andS(min) is the minimum output signal relative to therotation of the dichroic polarizer. One sees from Ta-ble IV that the roughened quartz surfaces are very poordepolarizers and that the Teflon is significantly better.The uncertainty of the values in Table IV is a fewpercent.

V. Summary and Conclusions

Tables I-IV give the measured attenuations, cosinelaw departures, and depolarizations for a number ofroughened quartz diffusers and for several thicknessesof sheet Teflon. It is seen that one to three stackedroughened quartz surfaces with pit sizes varying from 8to 100 um are not suitable diffusers for our applicationrelative to either the cosine law or to depolarization in

the 200-400-nm wavelength range. Sheet Teflon issignificantly better. A 0.01-in. thick sheet of Teflonobeys the cosine law at 254 nm to better than 1% forincident angles of 2 and exit angles of 4.3° (f/6.6), andit depolarizes plane polarized radiation at 254 nm by-95%. The attenuation is significantly greater forTeflon than for roughened quartz but is acceptable inour application for Teflon thicknesses of 0.005 in. andpossibly 0.010 in.

References1. Henry J. Kostkowski, Robert D. Saunders, John F. Ward, Charles

H. Popenoe, and A. E. S. Green, "Measurement of Solar Terres-trial Spectral Irradiance in the Ozone Cut-Off Region," in Self-Study Manual on Optical Radiation Measurements: Part 3Applications, Natl. Bur. Stand. (U.S.), Tech. Note 910-5 (1982),Chap. 1.

2. Raymond K. Kostuk, "Polarization Characteristics of a 100-WFEL Type Filament Lamp," Appl. Opt. 20, 2181-2182 (1981).

3. E. B. Saloman, S. C. Ebner, and L. R. Hughey, "Vacuum Ultravio-let and Extreme Ultraviolet Radiometry Using Synchrotron Ra-diation at the National Bureau of Standards," Opt. Eng. 21, 951-000 (1982).

The experimental work performed in this investiga-tion was carried out at the National Bureau of Stan-dards, Gaithersburg, MD (now the National Instituteof Standards & Technology) and was supported by theNational Oceanographic & Atmospheric Administra-tion, Boulder, CO, under contract NA82RAC00028.

1 August 1989 / Vol. 28, No. 15 / APPLIED OPTICS 3245