J. Micro/Nanolith.

MEMS MOEMS 73, 033010 JulSep 2008

Optical properties of Teon AF amorphous uoropolymers

Min K. Yang Roger H. French DuPont Co. Central Research Experimental Station Wilmington, Delaware 19880-0400 E-mail: [email protected] Abstract. The optical properties of three grades of Teon AF AF1300, AF1601, and AF2400were investigated using a J.A. Woollam VUV-VASE spectroscopic ellipsometry system. The refractive indices for each grade were obtained from multiple measurements with different lm thicknesses on Si substrates. The absorbances of Teon AF lms were determined by measuring the transmission intensity of Teon AF lms on CaF2 substrates. In addition to the refractive index and absorbance per cm base 10, the extinction coefcient k, and absorption coefcient per cm base e, Urbach parameters of absorption edge position and edge width, and two-pole Sellmeier parameters were determined for the three grades of Teon AF. We found that the optical properties of the three grades of Teon AF varied systematically with the AF TFE/PDD composition. The indices of refraction, extinction coefcient k, absorption coefcient , and absorbance A increased, as did the TFE content, while the PDD content decreased. In addition, the Urbach edge position moved to a longer wavelength, and the Urbach edge width became wider. 2008 Society of Photo-Optical Instrumentation Engineers.
DOI: 10.1117/1.2965541

Edward W. Tokarsky DuPont Fluoropolymer Solutions Chestnut Run Plaza Wilmington, Delaware 19880-0713

Subject terms: uoropolymer; ellipsometry.

absorbance

absorption

coefcient;

VUV

Paper 07086R received Oct. 24, 2007; revised manuscript received May 16, 2008; accepted for publication Jun. 6, 2008; published online Aug. 13, 2008.

1 Introduction Teon AF uoropolymers are widely used in industrial and commercial applications. Teon AF chemical name: uorinated ethylenic-cyclo oxyaliphatic substituted ethylenic copolymer is a family of amorphous uoropolymers based on copolymers of 2,2-bistriuoromethyl-4,5diuoro-1,3-dioxole PDD.1 Its chemical structure is shown here:2,3
F F CF2CF2 x O O y

CF3

CF3

The PDD dioxole monomer in Teon AF yields polymers that have unexpected properties. Teon AF is a copolymer of PDD and tetrauoroethylene TFE. The principle difference in three grades of Teon AF polymers AF1300, AF1601, and AF2400 is based solely on the relative amount of the dioxole monomer TFE:PDD in the basic polymer chain Table 1. In addition to outstanding chemical, thermal, and surface properties, Teon AF uoropolymers have unique electrical, optical, and solubility characteristics. The Teon AF uoropolymers are distinct from other uoropolymers in that they are soluble in selected solvents and have high gas permeability, high compressibility, high creep resistance, and low thermal conductivity. Teon AF
1537-1646/2008/$25.00 2008 SPIE

uoropolymers have the lowest dielectric constant of any known solid polymer, and this makes them attractive for demanding electronics applications. Teon AF polymers have the lowest index of refraction of any known polymer, making them very suitable as low-index claddings for waveguide applications. Because of the combination of electric and optical properties, the use of Teon AF uoropolymers is very attractive in photonics and photovoltaics. Teon AF is the rst polymer to demonstrate a transparency possibility of at least 40% of incident light at 157 nm making it to the bottom of the resist layer for purposes of sharp pattern development.4 This paper describes the optical properties such as refractive index, extinction coefcient k, absorbance, absorption coefcient , and absorption edge that we have extensively studied. We have determined these optical properties and found the relationship between those propTable 1 Physical properties of Teon AF grades 1300, 1601, and 2400. TFE/PDD Composition Ratio mol % 50/ 50 35/ 65 11/ 89 225 M 600 M

TAF Grade AF1300 AF1601 AF2400

Specic Gravity ASTM D792 1.88 1.78 1.67

MN, Nominal

J. Micro/Nanolith. MEMS MOEMS

033010-1

JulSep 2008/Vol. 73

Yang, French, and Tokarsky: Optical properties of Teon

Spin Curves for Teflon AF
180 160 140 Film Thickness (nm) 120 100 80 60 40 20 0 0 500 1000 1500 2000 2500 3000 3500 4000 4500 5000 Spin Speed (RPM) AF1300 Film Thickness (nm) AF1601 Film Thickness (nm) AF2400 Film Thickness (nm)

CaF2 substrates that were also vapor primed with HMDS and dried under the same conditions as the other two grades. Film thicknesses were measured using a stylus prolometer, Dektak 3ST Veeco, Woodbury, NY. 2.2 Spectroscopic Ellipsometry and Analysis The variable-angle spectroscopic ellipsometry VASE measurements were performed with the VUV-VASE VU302 instrument J.A. Woollam Co., Lincoln, Nebraska, which has a range from 0.69 to 8.55 eV 1800 to 145 nm and employs MgF2 polarizers and analyzers rather than the more common calcite optics. The spectrometer was upgraded from single beam to dual beam for more accurate transmission measurements and better instrument stability. The instrument has an MgF2 auto-retarder and is fully nitrogen purged to avoid absorption of VUV light by ambient oxygen and water vapor, which is important at wavelengths below 200 nm. Light from both the deuterium lamp and the xenon lamp passes through a double-chamber CzernyTurner-type monochromator to provide wavelength selection and stray-light rejection. The spot diameter of the light source on the surface of the sample is 2 mm. Computercontrolled slit widths can adjust the bandwidth to insure adequate spectral resolution of optical features in the data such as the closely spaced interference oscillations, which arise in very thick lm. A photomultiplier tube is utilized for signal detection in the ultraviolet. A stacked Si / InGaAs photodiode detector is used for longer wavelengths. Ellipsometric measurements were conducted using light incident at angles of 55 deg to 80 deg relative to normal on the front surface of the sample, the back of which was roughened with coarse polishing paper. The instrument measures the ellipsometric parameters and , which are dened5 by Eq. 1: tanei =

Fig. 1 Experimentally determined spin curves for Teon AF grades 1300, 1601, and 2400, spun on silicon wafers vapor-primed with HMDS.

erties for three grades of Teon AF uoropolymers. 2 Experimental and Analytical Techniques 2.1 Sample Preparation For VUV spectroscopic ellipsometry, a spin coating method was used for obtaining thin uniform coatings on silicon wafers. A Brewer Science spincoater/hotplate Brewer Science, Rolla, Missouri was used to spin Teon AF solution onto hexamethyldisilazane HMDS vapor-primed silicon wafers. In order to obtain various thicknesses of lm, the spinning speed range was set to 1000 RPM, 1500 RPM, 2000 RPM, 2500 RPM, 3500 RPM, and 5000 RPM, respectively, for 60 s. Different thicknesses of lm improves condence in the results, as light travels different paths through the lm. The bake temperature was 140 for 60 s. In order to obtain a workable spinning solution, samples of 6.5 wt% Teon AF1300 and AF1601 were diluted by adding FC-40 surfactant Fluorinert electronic liquid manufactured by 3M to the Teon AF spinning solution. The solids ratio of FC-40 to Teon AF polymer for AF1300 is 1 to 0.022 and for AF1601 is 1 to 0.020. Various lm thicknesses between 60 and 160 nm were obtained by use of a rotation speed range of 1500 to 5000 RPM and a spinning time of 60 s for AF1300. Various lm thicknesses between 50 and 100 nm were obtained by use of a rotation speed range of 1000 to 2000 RPM and a spinning time of 60 s for AF1601. A sample Teon AF 400S1-100-1 Teon AF2400 was diluted by adding FC-40 surfactant to the AF2400 for a workable solution. The solids ratio of FC-40 to Teon AF polymer for AF2400 is 1 to 0.008. Various lm thicknesses between 70 and 120 nm were obtained by use of a rotation speed range of 1000 to 2500 RPM and a spinning time of 60 s. Figure 1 shows the experimentally determined spin curves for these solutions of Teon AF grades 1300, 1601, and 2400, determined on HMDS-vaporprimed silicon wafers. For transmission samples, droplets of undiluted original Teon AF1300 and AF1601 solutions were dispensed on the 13 1 mm bare CaF2 substrate Corning, Brookeld, Massachusetts and slowly dried in a closed metal sample box for several days in order to obtain a crack-free thick lm. For improved wetting for the undiluted original Teon AF2400 solution, it was deposited on 13 1 mm
J. Micro/Nanolith. MEMS MOEMS

RP , RS

where R P / RS is the complex ratio of the p- and s-polarized components of the reected amplitudes. These parameters are analyzed using the Fresnel equations6,7 in a computerbased modeling technique, including a surface roughness layer to directly determine the optical constants.8,9 VUVVASE VU-302 measurements for this experiment were taken from wavelength range 145 nm to 1550 nm and at multiple angles 55 deg to 0 deg. Variable angles improve condence, as light travels different paths through the lm. The ellipsometry data, taken from the lm on silicon substrate, was t to determine the polymer lm thickness, roughness, thickness nonuniformity, and complex refractive index10 Fig. 2. Initially a parameterized model was used to describe the Teon AF lm optical constants over the wide spectral range. The Teon AF lms were modeled using tabulated optical constants from Woollam for silicon and SiO2. A thin surface layer was described in the optical model using roughness, which is an effective medium approximation EMA consisting of 50 / 50 mix of organic lm and void n = 1. In the initial stage, a Cauchy dispersion model was used to t the transparent region of 450 nm to 1550 nm. The nal model was determined by tting the optical constants on a point-by-point basis over the full spectral range
JulSep 2008/Vol. 73

033010-2

Yang, French, and Tokarsky: Optical properties of Teon

100 80

in degrees

60 40 20 0 145

Model Fit Exp E 60 Exp E 70 Exp E 80

1 UV and 1 IR Eq. 3 and Eq. 4 and one 3-pole 2 UV and 1 IR Eq. 5 Sellmeier equations tted to the three grades of Teon AF to parameterize the wavelength dispersion of the index of refraction: n 2 = 1 + B 1 2 B 2 2 + 2 , 2 C1 C2 B 1 2 B 2 2 + , 2 C1 2 C2 B 1 2 B 2 2 B 3 2 + + , 2 C1 2 C2 2 C3 3

380

615

850 1085 Wavelength (nm)

1320

1555

n 2 = A +

300
Model Fit Exp E 60 Exp E 70 Exp E 80

n 2 = 1 +

200

100

-100 145

380

615

850 1085 Wavelength (nm)

1320

1555

Fig. 2 Ellipsometric data and measured on a Teon AF1300 grade thin lm on a silicon wafer and the model t results.

where is wavelength, n2 is the square of the refractive index, the coefcient A is an approximation of the shortwavelength UV absorption contributions to the refractive index at longer wavelengths, and B1,2,3 and C1,2,3 are the parametrically tted Sellmeier coefcients associated with a given optical material. This equation was deduced in 1871 by Sellmeier and was developed by Cauchy as Cauchys equation for modeling dispersion.13 The Sellmeier equation is usually valid only in high-transparency spectral regions. We nd that the two-pole Sellmeier equation given in Eq. 4 gives the minimum standard error of the t for these samples, and these Sellmeier parameters are summarized in Table 2. 2.4 Optical Transmission and Absorption Coefcient The optical absorbance of the Teon AF lm was determined by measuring the transmission intensity of CaF2 substrates and lms of different thicknesses on CaF2 substrates. Once the transmission spectra of the substrate and the lm coated on that substrate were determined, the lm optical absorbance per cm base 10 was then determined by Eq. 6: A/cm = log10Tsubstrat/T film , t film 6

in which the data in each single wavelength are t separately. A typical application for this data class was chosen as displaying wavelength-by-wavelength t parameter condence limits for optical constants. This method does not insist on Kramers-Kronig consistency. Teon AF1300 is used as an example of the ellipsometric analysis. The ellipsometry data Fig. 2, taken from the lm on silicon substrate, was t, in a linear regression sense, to a thin-lm optical model11 to determine its lm thickness roughness and thickness nonuniformity, and complex refractive index Eq. 2:12 = n + ik . n 2

in degrees

2.3 Sellmeier Equations Sellmeier equations and their tted parameters have been employed for many years as a convenient way to permit the calculation of the index of refraction of transparent materials at arbitrary wavelengths. We have evaluated two 2-pole

where T is the transmission, t is the thickness of lm, and A / cm is the absorbance per centimeter base 10. Using multiple samples of different lm thickness, one can then solve the resulting system of equations and determine the reproducibility and a standard deviation of the optical absorbance per centimeter. Transmission-based measurements also require that the lm thickness of the sample on the substrate be optimized

Table 2 Two-pole Sellmeier parameters. TAF Grade AF1300 AF1601 AF2400 A 1.517 1.461 1.456 B1 0.184 0.226 0.181 C1 0.016 0.014 0.014 B2 1 1 1 C2 104.66 119.03 159.28

J. Micro/Nanolith. MEMS MOEMS

033010-3

JulSep 2008/Vol. 73

Yang, French, and Tokarsky: Optical properties of Teon

Fig. 3 Extinction coefcient k determined from transmission measurement for three grades of Teon AF.

Fig. 4 Optical absorption edge absorbance ts of three grades of Teon AF with their corresponding Urbach edge ts dashed lines.

for the dynamic range of the technique so that the transmittance of the lm and substrate falls in the range from 3 to 90%. If the transmittance falls much below 1%, the accuracy of the measurement is severely degraded and erroneous results appear.14 Much of the uncertainty comes from the antireective effect of thin polymer lm when coated on CaF2. That is, the small amount of light absorbed by a highly transparent lm is lost in comparison to the transmission increase resulting from decreased reectivity from the antireective effect. In the worst such case, Tsample is greater than Tsubstrate, and the calculated absorbance goes negative.15 As in all experimental measurements, the accuracy of the measured values is a function of the sample and measurement apparatus. The inherent sensitivity of spectral transmission and absorbance measurements is affected by the optical path length of the sample and the transmission drop that occurs as light transmits through the sample in the measurement. As the transmission drop decreases, the accuracy of the absorbance measurement decreases. A transmission difference of 0.1% is near the limit of the measurement method. Once the extinction coefcient k from ellipsometry, Fig. 3 and the base-10 absorption coefcient A Fig. 4 have been determined, the base-e optical absorption parameter Fig. 5 can be determined using Eq. 7. The absorption coefcient, , corresponds to the attenuation of the light transmitted through the sample and is calculated on a natural logarithm basis. Since the absorbance per cm, A, is determined from the base-10 logarithm of the optical density given in Eq. 6, a value of ln10, or 2.302585, is introduced into Eq. 6 and Eq. 7. Both k and are inherent optical properties of the material. On the other hand, absorbance/cm is frequently based only on transmission measurements and thus neglects effects arising from the index mismatch between the lm and substrate, thin lm interference effects, and lm nonuniformity effects. The fundamental absorption edge spectra have been determined16 by Eq. 7:
J. Micro/Nanolith. MEMS MOEMS

= 4k/ ,

where is the optical absorption coefcient base e, is the wavelength of the light source, and k is the extinction coefcient. The relationship between the base-e absorption coefcient and the base-10 absorption coefcient A is given in Eq. 8: A= = log10T1 log10T2 1 lnT1 lnT2 = ln10 t t 1 , ln10 8

where T1 is the transmission of the substrate, T2 is the transmission of the lm, and t is the thickness of the lm.

Fig. 5 Optical absorption edge absorbance coefcient ts of three grades of Teon AF with their corresponding Urbach edge ts dashed lines. JulSep 2008/Vol. 73

033010-4

Yang, French, and Tokarsky: Optical properties of Teon

2.5 Urbach Analysis of Absorption Edges Urbach edge analysis is a useful way to parametrically characterize a lms optical absorption edge and to potentially distinguish intrinsic and extrinsic contributions to the 17 absorbance. Urbach18 originally observed that a materials optical absorption for energies below the fundamental optical absorption edge is exponential in nature and can be characterized by the Urbach edge energy E0 and the Urbach width W, which is related to the slope of the Urbach edge using the nm eV conversion nm = 1239.8 / E eV.

lnA = mA E + bA .

15

The intercept energies b energy at which ln = 1, and bA energy at which lnA = 1 are the quantities of interest. Using the conversion between and A, we can nd the relationship between the slopes and intercepts for A versus . What one nds is that the computed Urbach edge widths will be the same for ts to A or : W = WA . 16

2.5.1 Urbach analysis of the absorption coefcient and the absorbance A When Urbach ts an exponential to , the equations used are:

The intercept energies E0A and E0, as reported by Urbach, are related by: E0 = E0A + ln2.302585/mA , E0 = E0A + ln2.302585 WA . 17 18

E = H exp

E E 0 , W

ln =

E E 0 h W , W

3 10

Results

3.1 Complex Index of Refraction n + ik The refractive index n and k, lm thickness of Teon AF lms, and transmission intensity of Teon AF lms were determined over the wavelength range 145 nm to 1550 nm. In order to obtain the best quality of uniform and crackfree thin lms by spin coating, Teon AF lms with thicknesses in the range of 50 to 160 nm were measured instead of the 200-nm-thick lms commonly studied in ellipsometry. Therefore, these samples are a good candidate for multiple sample analysis performed by simultaneously tting the data from two or more samples. This is a very powerful method for determining optical constants because t parameters are coupled between the models to take advantage of the additional information provided by the use of multiple samples. A good multiple sample analysis is effective only if all lms have the same optical constants. The complex indices of refraction for the three grades of Teon AF, AF1300, AF1601, and AF2400 are shown in Fig. 6 and Table 3. A plot of the extinction coefcient k determined from transmission measurement of the lm on the CaF2 substrates is shown in Fig. 3. The results of the indices of refraction calculated from the Sellmeier equations are quite similar to those determined from the Cauchy model. The index of refraction versus wavelength data can be provided by the Sellmeier equation without the need for extensive tables. 3.2 Transmission and Absorption Coefcient For the determination of the optical absorption of the lm micron-thick lms were measured. These thicker lms are more susceptible to the effects of nonuniformity and nite spectrometer bandwidth. Ellipsometers that include a retarding element are capable of measuring percent depolarization.10 Both lm nonuniformity and nitebandwidth can depolarize the measurement beam. We quantify these nonideal effects and model their behavior during analysis of experimental data. In this manner, the optical constants can be determined with a higher degree of certainty.
JulSep 2008/Vol. 73

where the absorption coefcient as a function E is characterized by the Urbach edge energy E0 and the Urbach width W. The intercept E0 energy at which = 1 is the quantity of interest. When Urbach ts an exponential to A, the equations used are: AE = HA exp

E E 0A , WA

11

lnA =

E E 0A h AW A . WA

12

The intercept E0A energy at which A = 1 is the quantity of interest. The relationship between E0 and E0A is obtained as follows: AE = HA exp

E E 0 E E 0A H = exp = . WA ln10 ln10 W 13

2.5.2 Relationships between slopes and intercepts for Urbach ts to A versus Note that Urbach actually ts the exponential equations shown earlier. This is done because the error statistics are known in data space A and , not in lndata space. The equations given here are the easiest to use to understand the slope and intercepts that are provided in the solution. The preceding exponential equations that are used by Urbach for tting can be transformed to a linear form by taking the logarithm of the data. When we do this, we obtain the following: straight line equations, with slopes m and intercepts b. ln = m E + b ,
J. Micro/Nanolith. MEMS MOEMS

14

033010-5

Yang, French, and Tokarsky: Optical properties of Teon

1.44 1.41 Index of refraction ' n' 1.38 1.35 1.32 1.29 1.26 145
AF1300 AF1601 AF2400

380

615

850 1085 Wavelength (nm)

1320

1555

1.44 1.41 Index of refraction ' n' 1.38 1.35 1.32 1.29 1.26 140
AF1300 AF1601 AF2400

193 nm. This variability requires that each substrate be measured before the thin lm is deposited and then used in calculation of the lms absorbance/cm calculations. In more well established optical materials such as ultrahighpurity fused silica or single-crystal sapphire, the assumption that the substrates are all comparable is an acceptable approximation. For substrates of CaF2, transmission can vary19 from 71 to 86%. Knowledge of the lm thickness is also required to determine the absorbance/cm. Determination of the thickness of a soft polymer lm that is nearly index matched to a transparent substrate can be very challenging. This thickness uncertainty enters directly into the accuracy of the resulting absorbance/cm values. The fundamental absorbance, A base 10, and absorption coefcient, base e, spectra versus wavelength for the three grades of Teon AF are shown in Figs. 4 and 5, respectively.

3.3 Urbach ts results

150 160 170 180 Wavelength (nm) 190 200

Fig. 6 Index of refraction n determined from ellipsometric data for three grades of Teon AF.

The accurate measurement of transmitted light intensities is inherently difcult to attain and reproduce. The use of dual-beam spectrometers has improved accuracy and reproducibility, but the instrument relies on the equivalence of optical elements and detectors in multiple optical paths and/or the stability of the instrument over time. Since transmission-based absorbance/cm measurements result directly from the difference in transmission, T, of the substrate and the lm on the substrate, the accuracy to which T can be known is a major limitation. In addition to the variability in the measurements, the quality of the CaF2 substrates is highly variable due to bulk absorptions and surface contamination issues. These issues are more critical at 193 nm than at longer wavelengths, since materials are much more strongly absorbing at

To compare with the experimental data, the Urbach edge t to the optical absorption edge of absorbance A and absorption coefcient of the three grades of Teon AF are shown in Figs. 4 and 5. The relationship between these two ways of Urbach tting to A or was described in Secs. 2.5.1 and 2.5.2. The Urbach edge ts underneath the measured absorption edge for energies below the fundamental absorption edge, and the Urbach edge positions are presented in Table 4 for the three grades. Typically, we found that the standard deviation of the Urbach edge position is 1.6 nm and of the Urbach edge width is 0.3 nm. There is a correlation of the Urbach edge parameters with the range of optical absorbance of the sample and, therefore, Urbach parameters should be compared for samples over wavelength ranges exhibiting comparable ranges of optical absorbance. We can also compare the absorbance at 193 nm for the measured absorbance and the Urbach absorption edge. In this case, the broad, diffuse nature of the extrinsic absorbance allows the Urbach edge t to distinguish the extrinsic contributions from the intrinsic contributions to the optical absorbance.

Table 3 Indices of refraction at lithographic wavelengths. AF1300 nm 157.0 172.2 185.1 248.0 n 1.432 1.380 1.364 1.329 k 0.031 0.004 0.002 n 1.417 1.375 1.357 1.325 AF1601 k 0.026 0.007 0.002 n 1.377 1.344 1.333 1.303 AF2400 k 0.006 0.000 0.000

J. Micro/Nanolith. MEMS MOEMS

033010-6

JulSep 2008/Vol. 73

Yang, French, and Tokarsky: Optical properties of Teon Table 4 Urbach edge t parameters. Urbach Edge Position nm from Absorbance Fits 205 199 176 Urbach Edge Position nm from Absorbance Coefcient Fits 211 204 178 Urbach Edge Width nm from Absorbance 5.9 5.5 2.8 Urbach Edge Width nm from Absorbance Coefcient Fits 6.0 5.5 2.8

TAF Grade AF1300 AF1601 AF2400

4 Discussion 4.1 Complex Index of Refraction The index values of Teon AF1300 presented in Table 3 show that
it has a higher index of refraction than Teon AF1601 and Teon AF2400 uoropolymers. Refractive indices at other wavelengths can be determined from the Sellmeier parameters given in Table 2. But Fig. 6 shows that the index values dropped from the wavelength of 155 nm to shorter wavelengths for both AF1300 and AF2400. Also, Teon AF1300 shows a higher extinction coefcient k at 193.4 nm Fig. 3 than Teon AF1601 and Teon AF2400. The index values and extinction coefcient k of Teon AF1601 are intermediate to those of Teon AF1300 and Teon AF2400. Teon AF2400 has the lowest index values and extinction coefcient k. This is consistent with the earlier observation that the refractive index increases as the TFE content increases and the PDD content decreases.1

sequences of F-C-F units down the chain is bad for absorption, while the introduction of the PDD disrupts these TFE runs and reduces their length, thereby reducing the optical absorption. Nonetheless, Teon AF2400, which shows the lowest absorbance, has a stiff and sterically congested chain, as evidenced by its high glass transition temperature 240 C.1 We assume that the absorption of CF2 units is also dependent upon rotational angle and that the PDD monomer forces conformations are unfavorable for 193.4-nm absorption. 4.3 Urbach Edge Analysis and Polymer Composition We have seen that increasing TFE content increases absorption at 193.4 nm. Therefore, the Urbach optical absorption edge positions moved to longer wavelengths and the Urbach edge width became wider as the absorption increases. Teon AF1300 contains 50 mol% TFE, while AF1601 and AF2400 have lower TFE contents, at 35 mol% and 11 mol%, respectively. Therefore, the same TFE runlength mechanism discussed earlier can be associated with the dramatic shift of the absorption edge position with the associated change in TFE/PDD ratio. 4.3.1 Urbach edge position In Urbach analysis of the absorbance A, Teon AF1300 shows the optical absorption edge at a wavelength of 205 nm and an Urbach edge width of 5.9 nm, while analysis of the absorption coefcient shows a wavelength of 211 nm and an Urbach edge width of 6.0 nm. The Urbach edge width is supposed to be the same in both analyses; however, we see a 0.1-nm difference here. Both Urbach analysis parameters are greater than with the other two grades of Teon AF. For Teon AF1601, both Urbach parameters of optical absorption edge 199 nm and 204 nm and Urbach edge width 5.5 nm were sandwiched between Teon AF1300 and Teon AF2400. Additionally, the Urbach ts of Teon AF2400 show that both Urbach parameters of optical absorption edge 176 nm and 178 nm and Urbach edge width 2.8 nm in analyses of absorbance A and absorption coefcient were less than the other two grades of Teon AF. 4.3.2 Urbach edge absorbance at 193.4 nm We can also compare the absorbance at 193.4 nm for the measured absorbance 93 / cm and the Urbach absorption

4.2 Optical Absorbance The absorbance and absorption coefcient spectra of Teon AF1300 presented in Table 5 show that it is more strongly absorbing at 193.4 nm 93 / cm than Teon AF1601 and Teon AF2400 uoropolymers. This is consistent with our earlier observations.19 Increasing the TFE content in the polymer increases the average CF2 CF2n run length in the polymer chain, and this is associated with a noticeable increase in 193.4-nm absorption. Apparently, PDD interferes with the absorptive effect of long runs of adjacent CF2n groups, presumably by breaking up the electron conjugation down the polymer chain backbone in the CF2 run segments. Looking at the structure of PDD, it can be seen that the PDD monomer unit adds two adjacent O-C-F bonds as opposed to the F-C-F of TFE to a polymer chain. In a chain consisting of PDD and TFE monomers, such as in Teon AF, the presence of uninterrupted
Table 5 Absorbance/cm values at 193.4 nm from experiment and Urbach edge analysis. Urbach Absorption Edge/cm 7.9 2.5 0.6

TAF Grade AF1300 AF1601 AF2400

Measured Absorbance/cm 93 41 28

J. Micro/Nanolith. MEMS MOEMS

033010-7

JulSep 2008/Vol. 73

Yang, French, and Tokarsky: Optical properties of Teon

edge less than 10 / cm for Teon AF1300. For Teon AF1601,

the Urbach ts absorbance curve shows it to have a measured absorbance of 41 / cm and Urbach absorption edge less than 3 / cm at 193.4 nm. The Urbach ts absorbance curve of Teon AF2400, shows it to have a measured absorbance of 28 / cm and an Urbach absorption edge less than 1 / cm at 193.4 nm Table 5. So in these cases, the broad, diffuse nature of the extrinsic absorbers allows the Urbach edge t to distinguish the extrinsic contributions from the intrinsic contributions to the optical absorption.

11.

12. 13.

5 Conclusions Three grades of Teon AF uoropolymers have been studied using ellipsometry and transmission measurements to determine their optical properties. These three Teon AF grades have differing copolymer compositions the ratio of TFE/PDD in mol% and compositional differences that give rise to systematic changes in the optical properties such as the index of refraction and absorption coefcient. A comparison of the absorbance at 193.4 nm and the Urbach t absorption edge of the three grades of Teon AF uoropolymers suggests that extrinsic absorbers may account for a small increase in absorbance over what might be the intrinsic absorbance of the polymer. We found that the optical properties of the three grades of Teon AF varied systematically with the AF TFE/PDD composition. The indices of refraction, extinction coefcient k, absorption coefcient , and absorbance A increased, as did the TFE content, while the PDD content decreased. In addition, the Urbach edge position moved to a longer wavelength, and the Urbach edge width became wider. Acknowledgments The authors acknowledge the assistance of Dr. Lin K. DeNoyer Spectrum Square Associates, Inc. for Urbach analysis programming and Sellmereir parameters determination, B. B. French for editing the manuscript, and Mike Lemon, Bob Balback, and Mike Reilly for assistance in the measurements. References
1. P. R. Resnick and W. H. Buck, in Modern Fluoropolymers, J. Schiers, Ed., pp. 397419, Wiley, West Sussex, England 1997. 2. G. Belanger, P. Sauvageau, and C. Sandorfy, The far-ultraviolet spectra of peruoro-normal-parafns, Chem. Phys. Lett. 38, 649 1969. 3. K. Seki, H. Tanaka, T. Ohta, Y. Aoki, A. Imamura, H. Fujimoto, H. Yamamoto, and H. Inokuchi, Electronic structure of polytetrauoroethylene studied by UPS,VUV absorption, and band calculations, Phys. Scr. 41, 167 1990. 4. T. M. Bloomstein, M. Rothschild, R. R. Kunz, D. E. Hardy, R. B. Goodman, and S. T. Palmacci, Critical issues in 157 nm lithography, J. Vac. Sci. Technol. B 166, 3154 1998. 5. B. Johs, R. H. French, F. D. Kalk, W. A. McGahan, J. A. Woollam, Optical interference coatings, Proc. SPIE 2253, 1098 1994. 6. J. A. Woollam, B. Johs, C. M. Herzinger, J. Hilker, R. Synowicki, C. L. Bungay, Overview of variable angle spectroscopic ellipsometry VASE, part I: basic theory and typical applications, Proc. SPIE CR72, 328 1999. 7. R. H. French, J. S. Meth, J. R. G. Thorne, R. M. Hochstrasser, and R. D. Miller, Vacuum ultraviolet spectroscopy of the optical properties and electronic structure of seven polydi-alkylsilanes, Synth. Met. 5013, 499508 1992. 8. M. E. Innocenzi, R. T. Swimm, M. Bass, R. H. French, and M. R. Kokta, Optical absorption in undoped yttrium aluminum garnet, J. Appl. Phys. 683, 12001204 1990. 9. M. E. Innocenzi, R. T. Swimm, M. Bass, R. H. French, A. B. Villaverde, and M. R. Kokta, Room temperature optical absorption in undoped -Al2O3, J. Appl. Phys. 6712, 75427546 1990. 10. J. A. Woollam, B. Johs, C. M. Herzinger, J. Hilker, R. Synowicki, and C. L. Bungay, Overview of variable angle spectroscopic ellip-

14.

15.

16.

17.

18. 19.

sometry VASE, part II: advanced applications, Proc. SPIE CR72, 2958 1999. B. Johs, R. H. French, F. D. Kalk, W. A. McGahan, and J. A. Woollam, Optical analysis of complex multilayer structures using multiple data types, in Optical Interference Coatings, F. Abeles, Ed., Proc. SPIE 2253, 10981106 1994. F. Wooten, Optical Properties of Solids, p. 49, Academic Press, New York 1972. W. Sellmeier, II. Regarding the sympathetic oscillations excited in particles by oscillations of the ether and their feedback to the latter, particularly as a means of explaining dispersion and its anomalies, in German, Ann. Phys. Chem. 147, 52454 1872; http:// en.wikipedia.org/wiki/Sellmeier_equation. R. H. French, R. C. Wheland, D. J. Jones, J. N. Hilker, R. A. Synowicki, F. C. Zumsteg, J. Feldman, and A. E. Feiring, Fluoropolymers for 157-nm lithography: optical properties from VUV absorbance and ellipsometry measurements, in Optical Microlithography XIII, C. J. Progler, Ed., Proc. SPIE 4000, 14911502 2000. R. H. French, J. Gordon, D. J. Jones, M. F. Lemon, R. C. Wheland, E. Zhang, F. C. Zumsteg, K. G. Sharp, and W. Qiu, Materials design and development of uoropolymers for use as pellicles in 157-nm photolithography, in Optical Microlithography XIV, C. J. Progler, Ed., Proc. SPIE 4346, 8997 2001. R. H. French, K. I. Winey, M. K. Yang, and W. Qiu, Optical properties and van der Waals-London dispersion interactions of polystyrene determined by vacuum ultraviolet spectroscopy and spectroscopic ellipsometry, Aust. J. Chem. 60, 251263 2007. R. H. French, H. Sewell, M. K. Yang, S. Peng, D. McCafferty, W. Qiu, R. C. Wheland, M. F. Lemon, L. Markoya, and M. K. Crawford, Imaging Of 32-nm 1:1 lines and spaces using 193-nm immersion interference lithography with second-generation immersion uids to achieve a numerical aperture of 1.5 and a k1 of 0.25, J. Microlithogr., Microfabr., Microsyst. 43, 031103 2005. F. Urbach, The long-wavelength edge of photographic sensitivity and of the electronic absorption of solids, Phys. Rev. 92, 1324 1953. R. H. French, R. C. Wheland, W. Qiu, M. F. Lemon, E. Zhang, J. Gordon, V. A. Petrov, V. F. Cherstkov, and N. I. Delaygina, Novel hydrouorocarbon polymers for use as pellicles in 157-nm semiconductor photolithography, J. Fluorine Chem. 122, 6380 2003.

Min K. Yang is a staff technologist in Central Research at the DuPont Company in Wilmington, Delaware. She received her masters degree in materials science and engineering from Virginia Tech in 1994. She joined DuPont in 2000 and has been involved in several projects. Currently, she works on immersion lithography, where she focuses on optical properties characterization for immersion uids. She has contributed to several papers and patents in the area of immersion lithography. Roger H. French holds a PhD in materials science from MIT, where his work involved vacuum ultraviolet spectroscopy on Al2O3 from 80 to 800 nm. He is a research fellow in materials science in Central Research at the DuPont Company in Wilmington, Delaware, and also an adjunct professor of materials science at the University of Pennsylvania. His research is in optical properties and the electronic structure of ceramics, optical materials, and polymers. His work on new materials for optical lithography for integrated circuit fabrication has produced attenuating phase shift photomasks, photomask pellicles, semiconductor photoresists, and most recently, lithographic immersion uids. He also works on the origins and applications of London dispersion forces and the electronic structure and wetting of interfaces, such as intergranular and surcial lms. He studies neareld optics and scattering by particulate dispersions and complex microstructures using computational solutions to Maxwells equations. His work has produced 17 issued patents 135 published papers. JulSep 2008/Vol. 73

J. Micro/Nanolith. MEMS MOEMS

033010-8

Yang, French, and Tokarsky: Optical properties of Teon Edward W. Tokarsky received a BS in ceramic science and technology from Penn State University and a PhD in materials engineering from Rensselaer Polytechnic Institute, with thesis work on carbon ber funded by the US Air Force. After graduate school, he joined Hercules, Inc., where he developed advanced carbon bers for a wide variety of applications including carbon-carbon composites, thermosetting acetylenic polymers, and high-toughness propylene copolymer systems. Moving to DuPontAdvanced Fibers Systems in 1979, he made signicant contributions to new business

developments for Kevlar and Nomex in structural aircraft composites, printed circuit board reinforcements, and advanced honeycomb materials. He led the Aracon metal-clad aramid bers team in developing lightweight, high-strength electromagnetic shielding and signal conductors, with key adoptions in the U-2 eet and JSF. He joined Fluoropolymer Solutions in 2001 as a technical consultant supporting developments of uoropolymer wire insulation systems in aircraft/aerospace. His current responsibilities involve technical service and growth initiatives for amorphous uoropolymers in a wide range of applications. He is the author of numerous technical publications related to the various business developments over the years, has over 12 patents.

J. Micro/Nanolith. MEMS MOEMS

033010-9

JulSep 2008/Vol. 73